Full text of "Fishery bulletin"

See other formats

Fishery Bulletin

National Oceanic and Atmospheric Administration • National Marine Fisheries Service

^^ATES 0^ ^

ORARV

'"'1

^■

Vol. 74, No. 1

January 1976

BROTHERS, EDWARD B., CHRISTOPHER Pf MAX-HEWS /and REUBEN LASKERI Daily
growth increments in otoliths from larval and adult fishes -.-*?-^?4=l.

STRUHSAKER, PAUL, and JAMES H. UCHIYAMA. Age and growth of the nehu! Stole-
phorus purpureas (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily
growth increments of sagittae

MERRINER, JOHN V. Aspects of the reproductive biology of the weakfish, Cynoscion regalis
(Sciaenidae), in North Carolina

MacGREGOR, JOHN S. DDT and its metabolites in the sediments off southern California .

SHARP, GARY D., and ROBERT C. FRANCIS. An energetics model for the exploited yellowfin
tuna, Thunnus albacares, population in the eastern Pacific Ocean

ROGERS, CAROLYN A. Effects of temperature and saUnity on the survival of winter flounder
embryos

UCHIDA, RICHARD N. Reevaluation of fishing effort and apparent abundance in the
Hawaiian fishery for skipjack tuna, Katsuwonus pelamis, 1948-70

PEARCY, WILLIAM G. Seasonal and inshore-offshore variations in the standing stocks of
micronekton and macrozooplankton off Oregon

HUNTER, JOHN R. Culture and grovii;h of northern anchovy, Engraulis mordux, larvae

CRAWFORD, L., and M. J. KRETSCH. Effects of cooking in air or in nitrogen on the develop-
ment of fishy flavor in the breast meat of turkeys fed tuna oil with and without a-tocopherol
supplement or injection

LEWIS, THOMAS C, and RALPH W. YERGER. Biology of five species of searobins (Pisces^
Triglidae) from the northeastern Gulf of Mexico

LORD, GARY, WILLIAM C. ACKER, ALLAN C. HARTT, and BRIAN J. ROTHSCHILD.

acoustic method for the high-seas assessment of migrating salmon

PRISTAS, PAUL J., ELDON J. LEVI, and ROBERT L. DRYFOOS. Analysis of returns of
tagged Gulf menhaden

TILLMAN, MICHAEL F , and DONALD STADELMAN. Development and example appUca-
tion of a simulation model of the northern anchovy fishery

MASON, J. C, and S. MACHIDORI. Populations of sympatric sculpins, Cottus aleuticus and
Cottus asper, in four adjacent salmon-producing coastal streams on Vancouver Island, B.C

BUTLER, JOHN L., and ELBERT H. AHLSTROM. Review of the deep-sea fish genus
Scopelengys (Neoscopehdae) with a description of a new species, Scopelengys clarkei, from the
central Pacific

CHITTENDEN, MARK E., JR. Weight loss, mortaUty, feeding and duration of residence of
adult American shad, Alosa sapidissima, in fresh water

18
27

70
81

89
93
104
112
118
131

142
151

(Continued on back cover)

Seattle, Washington

U.S. DEPARTMENT OF COMMERCE

Rogers C. B. Morton, Secretary

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION

Robert M. White, Administrator

NATIONAL MARINE FISHERIES SERVICE

Robert W. Schoning, Director

Fishery Bulletin

The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, a
economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries
1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the li
document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared a:
numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulle
instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodic
issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Offi
Washington, D.C. 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, a
in exchange for other scientific publications.

EDITOR

Dr. Bruce B. Collette

Scientific Editor, Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Editorial Committee

Dr. Elbert H. Ahlstrom

National Marine Fisheries Service

Dr. William H. Bayliff

Inter-American Tropical Tuna Commission

Dr. Roger F. Cressey, Jr.
U.S. National Museum

Mr. John E. Fitch

California Department of Fish and Game

Dr. William W. Fox, Jr.

National Marine Fisheries Service

Dr. Marvin D. Grosslein
National Marine Fisheries Service

Dr. Edward D. Houde
University of Miami

Dr. Merton C. Ingham

National Marine Fisheries Service

Dr. Reuben Lasker

National Marine Fisheries Service

Dr. Jay C. Quast

National Marine Fisheries Service

Dr. Paul J. Struhsaker

National Marine Fisheries Service

Dr. Austin Williams

National Marine Fisheries Service

Kiyoshi G. Fukano, Managing Editor

The Fishery Bulletin is published quarterly by Scientific Publications Staff, National Marine Fisheries Service, NCAA,
Room 450, 1107 NE 45th Street, Seattle, WA 98105. Controlled circulation postage paid at Seattle, Wash.

The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the
public business required by law of this Department. Use of funds for printing of this periodical has been approved by the
Director of the Office of Management and Budget through 31 May 1977.

Fishery Bulletin

CONTENTS

Vol. 74, No. 1 January 1976

BROTHERS, EDWARD B., CHRISTOPHER R MATHEWS, and REUBEN LASKER. Daily
growth increments in otoliths from larval and adult fishes 1

STRUHSAKER, PAUL, and JAMES H. UCHIYAMA. Age and growth of the nehu, Stole-
phorus purpureas (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily
growth increments of sagittae 9

MERRINER, JOHN V. Aspects of the reproductive biology of the weakfish, Cynoscion regalis
(Sciaenidae), in North Carolina 18

MacGREGOR, JOHN S. DDT and its metabolites in the sediments off southern California . 27

SHARP, GARY D., and ROBERT C. FRANCIS. An energetics model for the exploited yellowfin
tuna, Thunnus albacares, population in the eastern Pacific Ocean 36

ROGERS, CAROLYN A. Effects of temperature and salinity on the survival of winter flounder

embryos 52

UCHIDA, RICHARD N. Reevaluation of fishing effort and apparent abundance in the

Hawaiian fishery for skipjack tuna, Katsuwonus pelamis, 1948-70 59

PEARCY, WILLIAM G. Seasonal and inshore-offshore variations in the standing stocks of
micronekton and macrozooplankton off Oregon 70

HUNTER, JOHN R. Culture and growth of northern anchovy, Engraulis mordax, larvae . . 81

LEWIS, THOMAS C, and RALPH W YERGER. Biology of five species of searobins (Pisces,
Triglidae) from the northeastern Gulf of Mexico 93

LORD, GARY, WILLIAM C. ACKER, ALLAN C. HARTT, and BRIAN J. ROTHSCHILD. An
acoustic method for the high-seas assessment of migrating salmon 104

PRIST AS, PAUL J., ELDON J. LEVI, and ROBERT L. DRYFOOS. Analysis of returns of
tagged Gulf menhaden 112

TILLMAN, MICHAEL F, and DONALD STADELMAN. Development and example applica-
tion of a simulation model of the northern anchovy fishery 118

MASON, J. C, and S. MACHIDORI. Populations of sympatric sculpins, Cottus aleuticus and
Cottus asper, in four adjacent salmon-producing coastal streams on Vancouver Island, B.C. 131

BUTLER, JOHN L., and ELBERT H. AHLSTROM. Review of the deep-sea fish genus
Scopelengys (Neoscopelidae) with a description of a new species, Scopelengys clarkei, from the
central Pacific 142

CHITTENDEN, MARK E., JR. Weight loss, mortaUty, feeding, and duration of residence of

adult American shad, Alosa sapidissima , in fresh water 151

(Continued on next page)

Seattle, Washington

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washing-
ton, D.C. 20402 — Subscription price: $11.80 per year ($2.95 additional for foreign mail-
ing). Cost per single issue - $2.95.

Contents — continued

BRUSHER, HAROLD A., and LARRY H. OGREN. Distribution, abundance, and size of
penaeid shrimps in the St. Andrew Bay system, Florida 158

MASON, J. C. Some features of coho salmon, Oncorhynchus kisutch, fry emerging from simu-
lated redds and concurrent changes in photobehavior 167

HURLEY, ANN C. Feeding behavior, food consumption, growth, and respiration of the squid
Loligo opalescens raised in the laboratory 176

GARRISON, DAVID L. Contribution of the net plankton and nannoplankton to the standing
stocks and primary productivity in Monterey Bay, California during the upwelling season . 183

TRENT, LEE, EDWARD J. PULLEN, and RAPHAEL PROCTOR. Abundance of macrocrusta-

ceans in a natural marsh and a marsh altered by dredging, bulkheading, and filling 195

Notes

FISHER, WILLL\M S., and DANIEL W. WICKHAM. Mortalities and epibiotic fouling of eggs

from wild populations of the Dungeness crab. Cancer magister 201

MATSUMOTO, WALTER M. Second record of black skipjack, Euthynniis lineatus, from the

Hawaiian Islands 207

WEIS, JUDITH S., and PEDDRICK WEIS. Optical malformations induced by insecticides in

embryos of the Atlantic silverside, Menidia menidia 208

CHENG, LANNA, and RALPH A. LEWIN. Goose barnacles (Cirripedia: Thoracica) on flotsam

beached at La Jolla, California 212

LAURENCE, GEOFFREY C. Caloric values of some North Atlantic calanoid copepods 218

HAURY, LOREN R. Method for restraining living planktonic crustaceans 220

STILLWELL, CHARLES E., and JOHN G. CASEY. Observation on the bigeye thresher shark,

Alopias superciliosus, in the western North Atlantic 221

LEWIS, ELIZABETH G. Epizoites associated with Bathynectes superbus (Decapoda:

Portunidae) 225

Vol. 73, No. 4 was published on 11 December 1975.

The National Marine Fisheries Service (NMFS) does not approve, recommend or
endorse any proprietary product or proprietary material mentioned in this publica-
tion. No reference shall be made to MNFS, or to this publication furnished by
NMFS, in any advertising or sales promotion which would indicate or imply that
NMFS approves, recommends or endorses any proprietary product or proprietary
material mentioned herein, or which has as its purpose an intent to cause directly or
indirectly the advertised product to be used or purchased because of this NMFS
publication.

DAILY GROWTH INCREMENTS IN OTOLITHS FROM
LARVAL AND ADULT FISHES

Edward B. Brothers,^ Christopher P. Mathews,^ and Reuben Lasker^

ABSTRACT

Daily growth increments have been found in otoliths offish larvae. The daily nature of these layers
was verified by examining larval fish of known age reared in the laboratory. A simple technique for
observing these marks is described and can be used on otoliths from larvae and adults. This provides
a convenient method for determining early growth in fishes and is particularly useful for fishes which
do not lay down annual or seasonal rings.

The use of otoliths in age determination (by
means of annual marks) is well known. The
techniques used have been described by Williams
and Bedford (1974) and Blacker (1974). Recently
Pannella (1971) has suggested that daily marks
may be formed in the sagittae (the otoliths used
almost universally in age determinations) of some
temperate species, while in 1974 Pannella
claimed to have detected them in a number of
tropical species. He also studied the temperate
species — silver hake, Merluccius bilivoaris; red
hake, Urophycis chuss; Atlantic cod, Gadus mor-
hua; and winter flounder, Pseudopleuronectes
americanus — in greater detail in this latter pa-
per. For some of these temperate species, particu-
larly for the first, Pannella was able to show that
there were fortnightly, monthly, and annual pat-
terns. The annual marks detected in the conven-
tional way were shown to contain about 365 daily
units. Pannella used acetate replicas of ground
otoliths which had been previously etched with
HCl. Pannella's work appears to justify the fol-
lowing conclusions:

1 . Daily increments^ occur in certain temperate
fish, e.g., M. bilinearis.

2. Periodic variations in increment thickness
occur with fortnightly, monthly, and annual
frequencies in this species.

'Scripps Institution of Oceanography, La Jolla, CA 92038;
present address: Langmuir Laboratory, Section of Ecology and
Systematics, Cornell University, Ithaca, NY 14853.

^Department of Fisheries, Escuela Superior de Ciencias
Marinas, University of Baja California A.P. 453, Ensenada,
B.C., Mexico.

^Southwest Fisheries Center, National Marine Fisheries Ser-
vice, NOAA, P.O. Box 271, La Jolla, CA 92038.

■•The smallest visible concentric layers seen in an otolith.

3. Structural units that are similar to those
shown to be daily in their occurrence in tem-
perate species are also found in some
tropical species.

Pannella (1974) was careful to explain that the
marks present in otoliths of tropical fish that ap-
peared to be annual on the basis of conventional
criteria could be deceptive. He suggested that by
analogy with temperate species, certain struc-
tures found in otoliths of tropical fish were also
daily in occurrence. Although he found spawning
marks, he did not find any seasonal or winter
growth checks in the otoliths of tropical fish. In
view of Pannella's expressed skepticism about the
formation of annual marks and his tentative con-
clusions, further evidence is needed that daily
increments occur in tropical fish. Furthermore, no
one appears so far to have attempted to apply this
method of age determination to larval fish, yet it is
in this last area that the most accurate and useful
results might be expected. Pannella (1974) com-
mented on the great regularity of the presumably
daily marks near the center of the otoliths of both
tropical and temperate fish. In these portions of
the otoliths, no superposition of more complex
patterns (e.g., 14 day, 28 day) were found.

It is the object of this paper to show that 1) true
daily increments are found in the otoliths of the
larvae of several species, and that daily marks
may be used to determine the ages of larval fish
with great accuracy and precision, at least for
approximately the first 100 days of life; and 2) in
adults offish from a variety of habitats, including
tropical waters, daily increments may be proven
to exist, and so to confirm Pannella's work.

Manuscript accepted July 1975.

FISHERY BULLETIN: VOL. 74, NO. 1. 1976.

FISHERY BULLETIN: VOL. 74, NO. 1

METHODS

Some material was examined with a Stereo-
scan^ S4 scanning electron microscope (Cam-
bridge Scientific Instruments Ltd.). These otoliths
were prepared for viewing by embedding them in
polyester resin, grinding and polishing them
to the vertical mid-sagittal plane with a graded
series of silicon carbide or aluminum oxide com-
pounds (400, 600, and 900 grit), and finishing
with 1-yum diamond paste. The polished surface
was then etched with 0.1 N HCl before being
rotary coated in a vacuum evaporator with 150 A
of gold-palladium alloy.

Both this technique and that of Pannella (1974)
involve the use of equipment and materials that
may be inaccessible in many countries. This is
particularly true for those countries in which
daily growth increments might prove to be espe-
cially helpful in stock assessment of commercial
fish, so that an alternative practical method with
minimal equipment was also used here and found
to be successful.

Otoliths of adult fish were ground by hand on a
glass plate covered with a water-silicon carbide
powder mixture (400-600 grit). The final polish
may be administered with diamond paste, but this
step is not essential. The ground otolith was then
examined in immersion oil. The grinding was
done in the same plane as described by Pannella
(1974). It is possible that storage in oil over a long
period of time may reduce the resolution obtained
when an otolith is examined. This appears to be
particularly true for larval otoliths. The above
technique is simple and requires only a good com-
pound microscope. Magnifications used in this
work ranged to 1,800 x; at least 600 x is required
for general viewing.

Otoliths from larvae were removed by teasing
them from fresh specimens. Oven-dried material
needed only to be moistened with a drop of water
before otolith removal. The otoliths were manipu-
lated and transferred to clean slides by picking
them up on the end of a fine dissecting needle
wetted with immersion oil. No additional prep-
aration was necessary, and the otoliths were
examined in immersion oil or after being perma-
nently mounted under a cover slip in a quick-
drying, neutral mounting medium. Ground sec-
tions from juveniles and adults may be similarly

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

mounted with no apparent loss in clarity. Larval
otoliths are thin enough that only optical section-
ing (i.e., carefully focusing to the plane of
maximum clarity) is necessary to make total
increment counts.

Material from a variety of species was ex-
amined and larval material of known age was
obtained by rearing eggs that had been fertilized
in the laboratory (Lasker et al. 1970; Leong 1971).
The chronological age from these fish was known
and could be compared with the number of growth
increments observed in their otoliths. Larvae of
northern anchovy, Engraulis mordax, were kindly
made available to us by John R. Hunter of the
Southwest Fisheries Center, National Marine
Fisheries Service, NOAA, at La Jolla, Calif

RESULTS

Otoliths of 15 E. mordax, aged 6 days, were
examined. The mean total length of the fish was
4.5 mm. The yolk-sac had been absorbed by the
fifth day after hatching. Figure la shows the ap-
pearance of the otolith of one of these fish.

Only one or two daily increments were present,
suggesting that daily growth increments ap-
peared in the otoliths of E. mordax only after
completion of yolk-sac absorption. In the labora-
tory, anchovy larvae were maintained in 14 h of
light when feeding took place and 10 h of darkness
when no feeding occurred (Lasker et al. 1970).

Table 1 shows the relation between chronologi-
cal age and number of apparently daily incre-
ments for larvae of E. mordax aged 6 to 100 days.
It is clear that there is an extremely close corre-
spondence between the chronological age in days
and the number of increments. Figure lb is a
micrograph showing the daily increments in an
anchovy otolith from a larva 18 days old.

Additional data presently being collected on
laboratory and wild-caught larvae indicates that
there is some interaction between the rate of
larval growth and the rate of increment formation
which may complicate the interpretation of oto-
lith age estimates.

Figure 2 shows the structure of adult anchovy
otoliths with successively greater magnification
of the scanning electron microscope. The darker
areas in the photographs represent areas of the
otolith that were more heavily etched because
they contained a higher proportion of CaCOs,
while the lighter areas have relatively more
organic material, probably otolin (see Degens et
al. 1969). It is seen from Figure 2 that the smallest

BROTHERS ET AL.: DAILY GROWTH INCREMENTS IN OTOLITHS

Figure l. — Light microscope photographs of
otoliths from laboratory-reared northern an-
chovy: a) 8-day-old larval otolith showing two
daily growth rings; b) 18-day-old larval otolith
showing 12 daily growth rings.

Table l. — Chronological age (days from hatching) and numbers
of growth increments in otoliths of northern anchovy.

Number

Cfironological

Chronological

an number

of fish

age in days

age less 5 days

increments

Range

0- 2

2- 4

4- 8

8- 11

9- 11

12- 15

14-16

16- 19

18- 21

18- 23

95-100

cyclical units are 1 to 2 ^tm thick in this part of the
anchovy otolith and that they do not appear to
contain any smaller units. It is these units that
are counted and appear in the data in Table 1. The
daily increment would therefore appear to be the
smallest unit of growth that is formed at the
supra-molecular level and, as such, is in principle
the most natural unit to use for age estimation.

Fertilized eggs of the California grunion,
Leuresthes tenuis, were obtained and reared in the
laboratory. The larvae were maintained in a
natural light cycle at 17° to 20°C with food
{ Artemia nauplii) continuously available. Larvae
were sacrificed at intervals and their otoliths were
examined. Table 2 shows the results obtained and
Figure 3 shows a photograph of a grunion otolith.

Table 2 shows that there is a close relation be-
tween the number of growth increments and the
chronological age of the larvae. Although the
agreement between age and daily increments is
not as good as it is for the anchovy, the results are
still very good. Table 2 also shows that in L.
tenuis, daily increments appear at hatching,
rather than at yolk absorption. Prehatching
marks also occur, although they were not tallied
in Table 2. Clearly the exact timing of the initi-
ation of daily increment formation varies from

FISHERY BULLETIN: VOL. 74, NO. 1

E
a.
O

o
m

..^,. 4...4^aH>- aL.'.,.i ^.LWjk'.

BROTHERS ET AL.: DAILY GROWTH INCREMENTS IN OTOUTHS

Table 2. — Chronological age and number of growth increments
in the otoliths of the California grunion.

Number

Chronological

Mean number

of fish

age In days

increments

Range

1- 2

8-10

10-12

16-18

20-26

species to species and must be independently
determined for each one.

Young striped bass, Morone saxatilis, were col-
lected on 2 July 1974 in the Sacramento River
delta (Tracy Pumping Station), Calif. These five
fish measured 29 to 37 mm SL (standard length)
and their otoliths had 62 to 120 observable incre-
ments; i.e., a sample of striped bass which should
have been 2 to 4 mo old according to their known
spawning season (Scofield 1931) were 2 to 4 mo old

according to the presence of growth layers found
in their otoliths. The spread in the age calculated
from daily increments probably corresponds to a
considerable spread in the dates when the fish
examined were hatched.

Otoliths from two striped bass 135 and 142 mm
SL were also examined. Published information on
the growth rate of this species (Scofield 1931)
indicates that striped bass of this size taken in
July should be 14 to 16 mo old. The ages obtained
by counting the presumed daily growth marks
were 419 and 445 days respectively, i.e., 14 to 15
mo old.

Figure 4 shows the daily marks in an otolith of
striped bass. Daily increments were fairly thick
near the center, thinner in an intermediate area
corresponding to the hyaline zone, and wider
again near the edge. In one specimen the central

f»

FIGURE 3. — Daily growth rings in an otolith of a
California grunion larva. The larva was approxi-
mately 26 days old.

Figure 4. — ^Daily growth rings in a striped
bass otolith. This fish was approximately 15
mo old. Differential growth can be seen in
rings grown in adjacent seasons. F =
fall; W = winter.

FISHERY BULLETIN: VOL. 74, NO. 1

area contained 231 daily increments, the mar-
ginal area contained 120, and there were 94
thinner marks in the middle zone. Working back-
wards from the 2 July collection date, this indi-
cated the slow growth zone occurred in December,
January, and February. These figures correspond
well with the known life cycle (Scofield 1931)
which suggests a fast growth period in spring,
summer, and fall (230 days, ~ 8 mo), a short
winter of slow growth (~ 3 mo), and a spring and
early summer (~ 4 mo) of faster growth prior to
capture.

Otoliths of postlarvae of the gobies Clevelandia
ios, Ilypnus gilherti, and Quietula y-cauda were
also examined. The fish were collected in Mission
Bay, San Diego. The 2-mo larval period indicated
in the otoliths agree with several independent
estimates of the length of time between hatching
and settlement (Brothers 1975).

Otoliths of two species of hake obtained from
the Gulf of California were studied. Mathews
( 1975) has shown that annual marks (annuli) may
be detected by means of the usual discrimination
of hyaline and opaque zones in Merluccius an-
gustimanus while in Merluccius sp. (Mathews in
press) the same techniques have also been applied
successfully. The ages of hake determined by
means of annuli may be compared with age
determined from counting the number of daily
increments; these are identified by analogy with
the structures shown to be daily in their incidence
in anchovy, grunion, striped bass, and other fish
and which appear to be the same as those shown
by Pannella (1971) to be daily in M. bilinearis
(Figures 5, 6). In most cases, direct total counts
were not possible because increments were not
equally visible over a complete nucleus to margin
radius. For these otoliths measurements of incre-

50jjm

I I

Figure 5. — a) Nucleus of an otolith from aMer-
luccius sp., 7 yr old; b) daily growth increments
shown from near the center of the otolith.

BROTHERS ET AL.: DAILY GROWTH INCREMENTS IN OTOUTHS

lOjjm

Figure 6. — Daily growth increments from the
otolith of Merluccius angustimanus . Note radial
fibers crossing the growth layers.

ment width were made at five or more locations
along a radius and then total counts were calcu-
lated by extrapolation. No larval or very young
hake were available for examination.

For Merluccius sp., data were available for 22
specimens aged 1 to 7 yr from the annuli present
in their otoliths. Figure 7 shows the graph of age
by annuli against age by daily increments for this
species. The correlation coefficient was 0.91 (20 df,
P » 0.001). The slope of the regression line was
1.14 (99% confidence limits [C.L.], 0.81-1.46).
This is not significantly different from the value

7i-

0 1 2 3 4 5 6 7

AGE BY DAILY GROWTH INCREMENTS (years)

Figure 7. — Graph of age-by-annuli against age-by-daily-
growth-rings in the otoliths of Merluccius sp. The encircled point
represents two points at the same position.

of 1.00 expected if age by years and by days were
to yield identical values.

Data from seven specimens of M. angusti-
manus were available and they varied in age from
only 1 to 2 yr. Given the much narrower ranges
and the smaller sample, the results obtained were
acceptable: r = 0.74 (0.05 > P > 0.01) and the
slope of the line was 1.25 (99% C.L., 0.24-2.25);
i.e., the slope was significantly different from zero,
but not from 1.0.

The precision of estimates of age obtained for M.
angustimanus was not very good, with deviations
of up to 0.5 yr being obtained; however, for
Merluccius sp. a somewhat narrower range was
usual, with some values differing by 0. 1 yr or less.
Extreme variations occurred with fish aged 7 to 13
yr, where errors of up to 2 to 3 yr could be obtained
where daily counts were made.

The average widths of the daily bands found in
the hake otoliths were 3 to 4 /xm, with wider and
narrower bands appearing sometimes in appar-
ently weekly, fortnightly, and monthly units. The
incidence of these units has not been examined in
detail and requires further study, but preliminary
work suggests that the basic unit used in age esti-
mates should be the daily unit; the higher order
units may be of great ecological interest, but
should probably not be used in aging these hake:
Only daily increments occur with the necessary
consistency and regularity.

In addition to the species mentioned above,
apparently daily marks have been found in a wide
variety of other fish, e.g., in Tilapia zilli, T.
nilotica, and Clarias mossambicus from Lake
Victoria (examined by E. B. B. and C. P. M.;
specimens kindly collected by John Rinne and Dr.

FISHERY BULLETIN: VOL. 74. NO. 1

Peretti of the East African Freshwater Fisheries
Research Organization, Jinja, East Africa), and
the following species examined by one of the
authors (E.B.B.): in the deep living Pacific rattail
Coryphaenoides acrolepis (58 cm SL; 10 to 11 yr);
in the myctophids Stenobrachius leucopsarus,
Tarletonbeania crenularis, and Triphoturus mexi-
canus; in the freshwater fish Cottus asper and
Salmo gairdneri; in the tropical marine fish
Chromis atrilobata and Apogon retrosella; in
adults of the gobies Cleuelandia ios and Gil-
lichthys mirabilis, where clear growth checks also
occur, so that daily marks alone would lead to
distinct underestimates of age; and in four species
of rapidly growing tropical and temperate tunas.
Statoliths from the squid Loligo opalescens (both
wild caught adults and laboratory-reared juve-
niles) also show what appear to be growth layers
analogous to those in fish otoliths. The appear-
ance of growth interruptions in a number of
species, e.g., the rockfish (genus Sebastes), either
as winter checks, spawning checks, or apparently
dispersed more evenly throughout the year, may
impose a severe limitation upon the use of daily
marks to age these fish. The technique seems best
suited to larvae, juveniles, fast-growing species,
and tropical species.

It is clear from our work that some difficulties
must be overcome before age estimation by means
of daily rings can become a standard tool in fish-
eries biology. However, it is also clear that

1. Daily rings may be used to estimate the ages
of larvae of some species up to 100 days old
with very great precision and that they prob-
ably can be used for fish up to 1 yr of age,
perhaps with a smaller degree of precision.
Struhsaker and Uchiyama (1976) show simi-
lar results with the tropical engraulid Stole-
phorus purpureus.

2. Daily marks may be used as a means of ac-
curate age determination for at least some
species of fish up to 6 yr old.

3. Daily marks may be used for age determina-
tion of at least some tropical fish. Pannella's
(1974) suggestion that daily increments
might be used in tropical fish as a means of
age estimation is almost certainly true, and
should be applicable to most species.

LITERATURE CITED

Blacker, R. W.

1974. Recent advances in otolith studies. In F. R. Harden
Jones (editor), Sea fisheries research, p. 67-90. John Wiley
and Sons, N.Y.

BROTHERS, E. B.

1975. Comparative ecology and behavior of three sjonpatric
California gobies. Ph.D. Thesis, Univ. California, San
Diego, 370 p.

DEGENS, E. T, W. G. DEUSER, AND R. L. HAEDRICH.

1969. Molecular structure and composition offish otoliths.
Mar Biol. (Berl.) 2:105-113.

Lasker, R., H. M. Feder, G. H. Theilacker, AND R. C. May.

1970. Feeding, growth, and survival of Engraulis mordax
larvae reared in the laboratory. Mar. Biol. (Berl.)
5:345-353.

LEONG, R.

1971. Induced spawning of the northern anchovy, En-
graulis mordax Girard. Fish. Bull., U.S. 69:357-360.

MATHEWS, C. P.

1975. Some observations on the ecology and the population
dynamics of Merluccius angustimanus in the south Gulf
of California. J. Fish. Biol. 7:83-94.

In press. The biology, ecology and population dynamics
of the large Gulf of California hake. Symposium in
Fisheries Biology, Ensenada, B.C., Mexico. Ciencas
Marinas, Spec. Suppl.
PANNELLA, G.

1971. Fish otoliths: daily growth layers and periodical

patterns. Science (Wash., D.C.) 173:1124-1127.
1974. Otolith growth patterns: An aid in age determina-
tion in temperate and tropical fishes. In T B. Bagenal
(editor), The ageing of fish, p. 28-39. Unwin Brothers,
Ltd., Surrey.
SCOFIELD, E. C.

1931. The striped bass of California (Roccus lineatus).
Calif. Dep. Fish Game, Fish Bull. 29, 84 p.
STRUHSAKER, P., AND J. H. UCHIYAMA.

1976. Age and growth of the nehu, Stolephorus purpureus
(Pisces: Engraulidae), from the Hawaiian Islands as indi-
cated by daily growth increments of sagittae. Fish. Bull.,
U.S. 74:9-17.

WlLLUMS, T., AND B. C. BEDFORD.

1974. The use of otoliths for age determination. In T. B.
Bagenal (editor). The ageing offish, p. 114-123. Unwin
Brothers, Ltd., Surrey.

AGE AND GROWTH OF THE NEHU, STOLEPHORUS PURPUREUS
(PISCES: ENGRAULIDAE), FROM THE HAWAIIAN ISLANDS AS
INDICATED BY DAILY GROWTH INCREMENTS OF SAGITTAE

Paul Struhsaker and James H. Uchiyamai

ABSTRACT

Direct evidence is presented that the sagittae of nehu, Stolephorus purpureas, grow by discernible
daily increments. Aging by daily growth increments provides the means to establish a general growth
curve for the first 6 mo of life for this species. Adult nehu exhibit nearly linear growth between 30 and
60 mm standard length. Preliminary evidence is presented that the nehu population of Pearl Harbor
may grow more rapidly than that of Kaneohe Bay.

Attempts to age tropical fishes by conventional
methods have generally been thwarted by the
absence of well-defined annuli in calcarious
structures and protracted spawning periods
which make length-frequency mode progression
analyses difficult. Recognizing that exceptions
to the above statement exist, Pannella's work
(1971) providing indirect evidence of the pres-
ence of daily growth layers and periodical
deposition patterns in the sagittae (otoliths) of
three species of boreal fishes from the western
North Atlantic suggested a means for conducting
age and growth studies of tropical species. He
concluded in that report: "Preliminary observa-
tion of growth patterns in sagittae of other
species, living at various depths and different
climates, appears to support the idea that daily
growth may be a universal feature of fish oto-
liths." Pannella's (1974) later work in Puerto Rico
provided circumstantial evidence of daily growth
layers in sagittae of several species of tropical
fishes.

To gain direct evidence that daily growth incre-
ments exist in tropical fishes we studied the nehu,
Stolephorus purpureas Fowler, a small engraulid
endemic to the Hawaiian Islands. The nehu is the
basis of a live-bait fishery producing about 4,000
metric tons annually of skipjack tuna, Katsu-
wonus pelamis (Linnaeus), from the vicinity of the
Hawaiian Islands. Stolephorus purpureus is a
short-lived species (less than 1 yr) and has been
the subject of relatively numerous studies: Naka-
mura (1970) has summarized the biological

'Southwest Fisheries Center, Honolulu Laboratory, National
Marine Fisheries Service, NOAA, Honolulu, HI 96812.

knowledge of this species available through 1965.
Our work provides evidence of the presence of
daily growth increments in the sagittae of nehu
and permits the assembly of a growth curve for the
first 6 mo of life for this species.

Brothers et al. (1976) have recently demon-
strated the presence of daily growth increments in
larval Engraulis mordax Girard and Leuresthes
tenuis (Ayres) and presented evidence that the
phenomenon occurs in several other species of
California fishes.

METHODS AND MATERIALS

The nehu samples were taken with three types
of gear in Pearl Harbor and the southeastern end
of Kaneohe Bay, Oahu, Hawaiian Islands. Adults
and juveniles (> about 30 mm standard length
(SL) ) were sampled with commercial bait seines
(square mesh measuring 3.2 mm to a bar) in Pearl
Harbor. Postlarvae (about ^ 20 mm SL), juveniles,
and adults were obtained in Kaneohe Bay by a
similar seine having a bar mesh measurement of
1.6 mm. Larvae (< 20 mm SL) were obtained near
Coconut Island by personnel of the Hawaii Insti-
tute of Marine Biology with 0.5-m ring nets with
mesh sizes of 550 /um.

Three separate holding experiments were con-
ducted to test the hypothesis that the sagittae of
nehu grow by discernible daily increments. All
animals for these experiments were collected in
Pearl Harbor and held in tanks of 38-kl capacity
at the National Marine Fisheries Service (NMFS)
Kewalo Basin Facility. The tanks were supplied
with well sea water of 23°-24°C and 33-35%o salin-
ity at a rate of about 300 liters/min. The nehu

Manuscript accepted August 1975.
FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

FISHERY BULLETIN: VOL. 74, NO. 1

were fed with frozen and live brine shrimp,
Artemia sp., under variable regimes as described
below. Each experimental population of nehu was
sampled during placement in holding tanks, and
then subsampled at various time intervals as
described for each experiment. Otoliths were ex-
tracted from most specimens within a few hours of
sampling. The remaining samples were frozen in
seawater or preserved in 75% solution of iso-
propanol until extraction of otoliths (removal of
tissue from otoliths of alcohol preserved speci-
mens is difficult).

The first holding experiment was begun 5
April 1972. A 16-day sample (21 April) and a
34-day sample (9 May) were obtained from
this population. The animals were fed once a
day With frozen and/or live brine shrimp. The
second holding experiment was begun 15 Decem-
ber 1972. This population was initially fed once a
day. A high mortality was observed during the
first 2 wk, after which food was provided twice
daily. Samples were collected weekly after 1 mo of
captivity. We examined sagittae from animals
collected on 19 January and 26 January 1973. The
third holding experiment was begun 4 May 1973.
This population was fed two or three times daily
with frozen brine shrimp. Samples were obtained
weekly between 4 May and 6 July. We examined
sagittae from animals collected 25 May and 8
June 1973.

Wild populations of larval, juvenile, and adult
nehu were sampled 13 times in Kaneohe Bay be-
tween 19 March 1972 and 13 July 1973 to obtain
estimates of growth rates at various seasons. Al-
though a second species of Stolephorus (S. buc-
caneeri Strasburg) occurs in Hawaii, larvae of this
species have not yet been collected in the south-
eastern end of Kaneohe Bay (Watson and Leis
1974; W. Watson pers. commun.).

After extraction, the sagittae were cleaned and
etched for up to 3 min in a 1% solution of HCl, then
washed and mounted whole on glass slides with
the mounting medium EuparaP and covered with
glass cover slips. Short lengths of monofilament
line were used to prevent the contact of the
specimen by the cover slide. Although the small-
est growth increments are microscopically dis-
cernible immediately after extraction their detec-
tion was enhanced after about 30 days of clearing
in the mounting medium. Sagittae used in the

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

first holding experiment and those collected from
Kaneohe Bay and Pearl Harbor during spring
1972 were placed in glycerine on slides and
covered. Some erosion of the sagittae edges was
noted after about 5 mo, and this practice was
discontinued after the first experiment. Slides
were either labeled with date of collection and
length of fish or assigned a five digit random
number for identification.

Our initial counts were taken from thin sections
of sagittae taken on the frontal plane. After
mounting the sagittae in epoxy resin, the initial
plane of polishing was made with rough sand-
paper. As the surface approached the desired
section, fine wet silicon carbide sandpaper (400
grit) was used. Final polishing of the surface was
done with suspensions of aluminum oxide parti-
cles having diameters of 15, 5, and 0.3 /^m. The
section was thinned on the opposite side to a
practical thickness and etched in a 1% solution of
HCl for variable periods up to 3 min. A few
attempts to make acetate peels of the small nehu
sagittae sections as described by Pannella (1971)
and Pannella and MacClintock (1968) were un-
successful. We eventually abandoned the section-
ing of sagittae because of the time required and
the difficulty in obtaining a precise section from
the nucleus to the posterior edge of the sagitta.

Sagittae were obtained from larvae less than
about 20 mm SL by placing the specimen on a slide
and gently teasing the otoliths from the head re-
gion. The sagittae were then mounted in Euparal
and read immediately. These otoliths tended to
clear completely within a few hours, and photo-
graphs are the only permanent record of these
specimens.

The smallest growth increments of the mounted
sagittae were counted with a compound micro-
scope at magnifications of 400-800 x . The smallest
growth increment in all fish otoliths consists of
both an organic and an inorganic layer (Degens et
al. 1969). These two layers in the nehu otolith to-
gether measure about 1-4 ^im thick. A zoom fea-
ture of the microscope was found to be extremely
useful. Counts were maintained on a hand tally.

Enumeration of the smallest growth increment
layers in whole sagittae is tedious, and reliable
counts can be obtained only after a moderate
amount of experience has been acquired. Enu-
meration is, obviously, much easier in sagittae
from smaller fishes (Figure 1). Usually, readings
cannot be made in a direct line from the nucleus to
the selected point on the edge of the sagitta;

STRUHSAKER and UCHIYAMA: AGE AND GROWTH OF STOLEPHORUS PURPUREUS

4 ^ * I

* .4

m
A'

* , N
,1

*, <

%■

e.^

■^

Figure l.— Sagittae of larval Stolephorus purpureas. A: Portion of sagitta from a 28.8-mm SL individual with about 65 growth
increments. B: 12.6 mm SL, 14 increments. C: 7.3 mm SL, 7 increments. D: 3.9 mm SL, 1 increment.

FISHERY BULLETIN: VOL. 74, NO. 1

rather, a somewhat circuitous route must usually
be taken from one area of the sagitta to another by
following a prominent growth increment.

Each sagitta was counted several times in
succession, the number of counts (up to 10) being
proportional to the size of the sagitta. Counts were
made from the nucleus to the antirostrum, ros-
trum, and postrostrum (terminology of Messieh
1972). A consistent count for the number of lamel-
lae was then obtained. Verification counts were
then made by the same reader at a later time. Ver-
ification counts were made by a second reader on
167 otoliths from the second and third holding ex-
periments, as well as randomly selected sagittae
representing the wild populations: 26.3% of these
counts agreed with the original count; 48.5% dif-
fered by less than 1%; 72.5% differed by less than
2%; 86.9% differed by less than 3%; 92.9% differed
by less than 4%; and 95.9% differed by less than
5%. Errors of less than 5% were considered accept-
able, and the median values of the two readers
were then utilized in the analyses. In cases where
the results differed by more than 5% , the sagittae
were reexamined and either a consensus of opin-
ion reached or the data discarded.

Standard lengths were taken to the nearest 0.01
mm with dial calipers. Sagittae were measured
with a micrometer eyepiece.

RESULTS
Holding Experiments

The holding experiments were undertaken as
one means to determine if the smallest growth
increments observable in the sagittae of nehu rep-
resent daily growth increments. We examined

sagittae of specimens from samples taken at vari-
ous time periods after the initial collection to
determine if there was an increase in mean
number of increments approximating the num-
bers of days between sampling. (Length data
collected from all samples indicate that the
length-frequency distributions of most of the
captive populations studied were normally
distributed.)

The data obtained for each holding experiment
were subjected to analysis of covariance and the
results are summarized in Table 1 and Figures 2-
4. There was homogeneous variance within the
samples for each of the three experiments as indi-
cated by Bartlett's test of homogeneity (chi-square
values = 0.56, 3.59, and 0.59, respectively).

In the first experiment there were no significant
differences between the means of the independent
variable (standard length) for each of the three
samples at the P< 0.05 level. There were signifi-
cant differences between the regression coeffi-
cients and the ielevation of the regression curves
for each sample at the P<0.01 level (Table 1,
Figure 2).

The significant differences between regression
coefficients seems best explained by the effects of
captivity. Hypothetically, the regression coeffi-
cient of the initial sample of 5 April represents the
relationship between number of growth incre-
ments and standard length in the wild population.
The smaller regression coefficient value of the 21
April sample indicates a slower growth rate of the
captive population during the 16-day interval
between sampling. This is probably due to less
than optimal food supply and/or other effects of
captivity. The intermediate regression coefficient
value of the 9 May sample indicates that the

Table l. — Summary of analysis of covariance for three holding experiments.

Sampling
date

Dependent variable
(Increments)

F ratios

Unadjusted

Adjusted
7

Independent

variable

(standard lengthi)

Regression
coefficient

Elevation

5 Apr 1972
21 Apr 1972

9 May 1972
First experiment

19 Jan. 1973
26 Jan. 1973
Second experiment

25 May 1973
8 June 1973
Ttiird experiment

84.9

101.1
118.1

114.8
120.8

124.9
140.0

86.7

0.77

38.2

100.6

0.76

10.8

116.4

0.74

20.6

114.0

0.95

14.7

121.6

0.85

31.1

132.1

0.97

13.8

133.4

0.95

6 1

1.2-

0.1

34—

5.4-

1.3

1.1

206*"

31*

1.1

"P sO.01.
'"P 5^0.001.

STRUHSAKER and UCfflYAMA: AGE AND GROWTH OF STOLEPHORUS PURPUREUS

150
140
130
120
110

-1 \ r-

-T 1 1 1 T

MAY 9, 1972

33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
STANDARD LENGTH (mm )

Figure 2. — Stolephorus purpureus: First holding experiment.

170
160
ISO
140

\2 130

S 120

S no

100
90
80

JANUARY 26, 1973

JANUARY 19, 1973

_l 1 I 1 1 ] 1 I I L.

34 36 38 40 42 44 46 48 50 52 54 56 58

STANDARD LENGTH (mm)

Figure 3. — Stolephorus purpureas: Second holding experiment.

170
160
ISO

I-
z

S 130

(£

? 120

MO
100

-I 1 1 1 T

"1 1 ^ r

MAY 25,1973

>xA(

JUNE 8,1973

46 47 48 49 50 51 52 53 54 55 56 57 58
STANDARD LENGTH (mm)

Figure 4. — Stolephorus purpureus: Third holding experiment.

growth rate has increased in the captive popula-
tion after 34 days in captivity, but has not reached
the value of the wild population from which it was
taken.

In the first holding experiment, the second and
third samples were collected 16 and 34 days, re-
spectively, after the initial sample. For unadjust-
edy values, these samples differed from the initial

sample by 16.2 and 33.2 increments, whereas for
the adjusted y values, they differed from the
initial sample by 13.9 and 29.7 increments
(Table 1).

The results of the two samples (collected after
more than 30 days in captivity) collected 19 and 26
January 1973, and compared in the second hold-
ing experiment, are summarized in Table 1 and
Figure 3. There were no significant diff'erences
between the means of the independent variables
or the regression coefficients at theP ssO.OS level.
The elevations of the two regression curves are
significantly different at theP ^0.001 level. The
differences in number of increments between
unadjusted y values (6.0) and adjusted y values
(7.6) again closely approximate the expected dif-
ference of 7 days between samples.

The results of the samples of 25 May and 8 June
1973 compared from the third holding experiment
are given in Table 1 and Figure 4. In this experi-
ment there was a significant difference between
the means of the independent variable (P <0.001),
but no differences between the regression coef-
ficients and elevations of the two regression
curves at theP «0.05 level. The significant differ-
ence in mean length between the two samples is
probably attributable to the increased amount of
food provided to the captive population and the
resulting high growth rate exhibited throughout
the duration of the experiment. Because the
treatment significantly affected the independent
variable, further examination of the regression
statistics is unwarranted. However, if the two
samples are subjected to a two-group comparison,
there is a significant difference between the mean
number of increments for each sample (P <0.05).
The difference between the means for each sample
(25 May,y = 124.9; 8 June, J = 140.0) closely ap-
proximates the expected difference of 14 days be-
tween samples.

We conclude from the relatively good agree-
ment between the increase in mean number of
growth increments and the number of days be-
tween collection of samples, that these data from
the holding experiments provide direct evidence
of the presence of daily growth increments in the
sagittae of nehu.

Growth of Sagittae

The total lengths of sagittae from the 5 April
and 9 May 1972 nehu samples (the initial sample
from the wild population and the 34-day sample)

FISHERY BULLETIN: VOL. 74, NO. 1

of the first holding experiment were taken in
order to examine the effects of captivity on sagit-
tal growth. Four measurements for the 5 April
sample were arbitrarily deleted because their
values were well below the distribution of the
majority of the sample. All 24 measurements from
the 9 May sample were utilized. There are signifi-
cant relationships between sagitta length and fish
length for the two samples (P <0.001, r^ values: 5
April, 0.82; 9 May, 0.70) (Figure 5). The first
experiment demonstrated that there was a signifi-
cant increase in the mean number of increments
between the two samples. Analysis of covariance
of sagittae lengths indicated that there were no
significant differences between the means of the
independent variables, regression coefficients, or
elevations of the regression curves for the two
samples (respective F ratios: 2.5, 1.0, 1.2) presum-
ably because of intrinsic variation, limited preci-
sion of measurements, and the relatively short
time period between samples. Although there
were no statistically significant differences found
in the comparison of the two curves, the two
regression coefficients exhibit perhaps expectable
trends. The lesser regression coefficient and r^
value for the 9 May sample may be indicative of a
decreased growth rate and more variable re-
sponses of individuals in the population to the
highly variable, and probably less than optimal,
conditions of the holding facility. In addition, the
differences between the unadjusted and adjusted
means of sagittal lengths between the 5 April
(1.094 mm; 1.070 mm, respectively) and 9 May
(1.176 mm; 1.201 mm, respectively) samples of
0.082 mm and 0.131 mm are to be expected with
daily growth increments of about 3-4 /u.m.

We have noted one apparent example of pro-
visioning rates affecting the growth rates of sagit-
tae of captive nehu. Sagittae from the 19 January
sample of the second holding experiment usually
exhibited 23-24 distinctive, more widely spaced
increments on the edge of the otolith. The num-
bers of distinctive increments approximately cor-
respond to the number of days during which the
daily amount of food provided the sample popula-
tion was double the initial ration. As might be ex-
pected otoliths collected 7 days later in the 26
January sample exhibited 30-31 distinctive incre-
ments. Indeed, the wider increments observed
after provisioning rates were doubled were much
more effective in "labeling" the sagitta than our
attempts to accomplish the same objective with
Tetracyclene. Possibly, controlled experiments

MAY 9,1972

APRIL 5,1972

32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
STANDARD LENGTH (mm)

Figure 5. — Stolephorus purpureas: Growth of sagittae during
first holding experiment.

with rapidly growing fish species incorporating
this treatment would be a much more expeditious
test of the daily growth increment hypothesis.

Age and Growth in Wild Populations

We examined larval, juvenile, and adult nehu
collected in Kaneohe Bay to obtain an estimate of
age and growth of a wild population based on the
assumption that the smallest observable growth
layers in the sagittae represent daily growth in-
crements. We examined 213 specimens from 13
collections made during most seasons between
spring 1972 and summer 1973 (no collections were
made in the months November through January).
The growth curves obtained from the individual
collections are given in Figure 6. Because all
individuals in a sample have been exposed to the
vagaries of the environment during their ob-
served lifespan, a composite growth curve for all
collections is presented in Figure 6F. Although
some variation between samples is apparent, the
composite scattergram serves as a first estimate of
the growth pattern of nehu in Kaneohe Bay.

There are two well-defined segments to the
composite growth curve (Figure 6F). Young lar-
vae exhibit exponential growth to a length of
about 15-17 mm. At about 20 mm the population
enters an almost linear growth phase to about 60
mm. The composite scattergram obscures another,
lesser inflection at about 20-30 mm exhibited by
the spring 1972 collections (Figure 6A). Yama-
shita (1951) has demonstrated that nehu have
completed larval metamorphosis at about 30 mm.
The major inflection at a length of about 17 mm
appears to reflect the fact that nehu begin to
exhibit exponential growth in body depth at this

STRUHSAKER and UCHIYAMA: AGE AND GROWTH OF STOLEPHORUS PURPUREUS

E
E

o
tr
<
o

1 1

I 1

^ (A)

. a, °°

1 1

(C)

(D)

': 1 I

< ! I
o

<fi

;

1 ;

(E)

(F)

880°°

20 40 60 80 100 120 140 160

AGE (DAYS)

180 0 20 40 60 80 100 120 140 160 180

AGE (DAYS)

Figure 6. — Stolephorus purpureas: Age-length relations of 213 Individuals from 13 collections in Kaneohe Bay. A: 19 March 1972 (21
individuals, 66-189 days); April 1972 ( 13, 1-24); 1 May 1972 (11, 6-12); 26 May 1972 ( 16, 21-68). B: 26 August 1972 (all specimens). C: 7
October 1972 (8, 16-23); 14 October 1972 (23, 19-62); 19 October 1972 (13, 99-148); 25 October 1972 (9,3-9). D: 12 February 1973 (12, 60-
87), (11, 115-140); 19 March 1973 (15, 78-125). E: 5 May 1973 (8, 40-69); 13 July 1973 (27, 66-136). F: Composite scattergram of all
observations.

size (cf, Nakamura 1970, fig. 4). Thus, much
growth of individual nehu is directed to allometric
growth of body depth, rather than body length.
The growth rate of young nehu (<17 mm) indi-
cated by the composite scattergram is consistent
with the estimates of larval growth rates pre-
sented by Tester (1951) and Yamashita (1951).

The possibility that the inflection at 15-17 mm
is related to a change in diet was examined.
Burdick (1969) investigated the feeding habits of
larval nehu from hatching to a length of 25 mm in
Kaneohe Bay. He found that young nehu less than
5 mm long fed almost exclusively on copepod
nauplii. At lengths of 5-7 mm, the diet shifted to a

FISHERY BULLETIN; VOL. 74, NO. 1

preponderance of small, adult copepods represent-
ing two genera. Larvae less than 20 mm fed
exclusively during day, at 20 mm they began
occasional feeding at night, and when they at-
tained a length of 25 mm they fed regularly at
night. None of these changes in feeding habits
seem related to the 15-17 mm inflection.

Only one fish, estimated to be 189 days old at a
length of about 63 mm, indicated that the Kane-
ohe Bay population of nehu may enter an asymp-
totic growth phase at about 60 mm. Obviously,
additional collections of older fishes are required
to elucidate this portion of the growth curve.

The absence of large adults might be explained
by the heavy exploitation of this stock by commer-
cial fishermen. Another possible explanation re-
lates to the observations of Muller^ on Stole-
phorus heterolobus Riippell in the Palau Islands of
the western Pacific. He found that large spawning
adults occur in open lagoon waters 2-4 km offshore
over depths of 30-40 m during night. The daytime
distribution of these individuals is unknown, but
it is thought that they occur near bottom in the
open lagoon. In the case of nehu, however, the
explanation of an absence of adults in the asymp-
totic growth phase by invoking an offshore spawn-
ing movement is argued against by a recent study
demonstrating that this species is capable of
spawning at a length of 35-40 mm (Leary et al. in

press).

These readings of whole-mounted sagittae from
Kaneohe Bay nehu did not reveal any periodic
deposition patterns of increments or spawning
checks as reported by Pannella (1971).

Geographical Comparison
of Growth Rates

One of the more exciting aspects of being able to
accurately determine growth rates of young fishes
is the tool that it provides to examine the effects of
various environmental conditions. As an exercise,
we compared the linear segments of the growth
curves of two samples {n = 15) of nehu collected
during March and April 1972 in Pearl Harbor and
Kaneohe Bay (Figure 7). Unfortunately, the dif-
ferences in size ranges of the two samples and the
small sample sizes resulted in significant hetero-
geneity of variance (P <0.05). The analysis of co-
variance did indicate, however, that there may be

140
130
120

;;; '°°

< 80

70
60
50
40

KANEOHE BAY

PEARL HARBOR

^MuUer, R. G. Population biology of Stolephorus heterolobus
Riippell in Palau. Ph.D. Dissertation in preparation. University
of Hawaii, Honolulu, HI 96822.

-JJ ^ 28 30 32 34 36 38 40 ^ 44 46 48 50 52 54
STANDARD LENGTH (men)

FIGURE 7.— Comparison of Stolephorus purpureas growth rates
in Pearl Harbor and Kaneohe Bay, spring 1972.

significant differences between the regression co-
efficients (P <0.05) and elevations (P <0.01) of
the two population curves, the Pearl Harbor sam-
ple exhibiting a faster growth rate to a length of
about 44 mm. Similar, but more intensive, studies
should provide a wealth of insight into a variety of
aquatic situations.

ACKNOWLEDGMENTS

We are indebted to Denis C. K. Pang, Barbara
Sumida, and Kenneth M. Hatayama for furnish-
ing us with specimens of nehu from Kaneohe Bay.
Supplementary increment counts were accomp-
lished by Patricia L. Seidl and Glen H. Sugiyama.
We thank Paul M. Shiota for assistance in con-
ducting the experiments. Illustrations are by
Tamotsu Nakata.

LITERATURE CITED

BROTHERS, E. B., C. P. MATHEWS, AND R. LASKER.

1976. Daily growth increments in otoliths from larval and
adult fishes. Fish. Bull., U.S. 74:1-8.
BURDICK, J. E.

1969. The feeding habits of nehu iHawaiian anchovy)
larvae. M.S. Thesis, Univ. Hawaii, Honolulu, 54 p.
DEGENS, E. T, W. G. DEUSER, AND R. L. HAEDRICH.

1969. Molecular structure and composition offish otoliths.
Mar Biol. (Berl.) 2:105-113.

Leary, D. F, G. I. Murphy, and M. Miller.

In press. Fecundity and length at first spawning of the
Hawaiian anchovy, or nehu (Stolephorus purpureas
Fowler) in Kaneohe Bay, Oahu. Pac. Sci.

Messieh, S. N.

1972. Use of otoliths in identifying herring stocks in the

southern Gulf of St. Lawrence and adjacent waters. J.

Fish. Res. Board Can. 29:1113-1118.
NAKAMURA, E. L.

1970. Synopsis of biological data on Hawaiian species of
Stolephorus. In J. C. Marr (editor), The Kuroshio: A sym-

STRUHSAKER and UCHIYAMA: AGE AND GROWTH OF STOLEPHORUS PURPUREUS

posium on the Japan Current, p. 425-446. East- West Cen-
ter Press, Honolulu.
PANNELLA, G.

1971. Fish otoliths: Daily growth layers and periodical pat-
terns. Science (Wash., D.C.) 173:1124-1127.

1974. Otolith growth patterns: An aid in age determination
in temperate and tropical fishes. In T. B. Bagenal (editor),
The ageing offish. Proc. International Symposium on the
Ageing of Fish, Univ. Reading, Engl., p. 28-39. Unwin
Brothers, Ltd., Engl.
PANNELLA, G., AND C. MACCLINTOCK.

1968. Biological and environmental rhythms reflected in
molluscan shell growth. Paleontol. Soc. Mem. 2:64-80. (J.
Paleontol. 42 (Suppl. to No. 5) ).

Tester, a. L.

1951. The distribution of eggs and larvae of the anchovy,
Stolephorus purpureas Fowler, in Kaneohe Bay, Oahu,
with a consideration of the sampling problem. Pac.
Sci. 5:321-346.

WATSON, W., AND J. M. LEIS.

1974. Ichthyoplankton of Kaneohe Bay, Hawaii. A one-year
study of fish eggs and larvae. Sea Grant Tech. Rep.,
UNIHI-SEAGRANT-TR-75-01, 178 p.

Yamashita, D. T.

1951. The embryological and larval development of the
nehu, an engraulid baitfish of the Hawaiian Islands.
M.S. Thesis, Univ. Hawaii, Honolulu, 64 p.

ASPECTS OF THE REPRODUCTIVE BIOLOGY OF THE WEAKFISH,
CYNOSCION REGALIS (SCIAENIDAE), IN NORTH CAROLINA^^

John V. Merriner^

ABSTRACT

The weakfish, Cynoscion regalis, has an extended spawning season in North Carohna's inshore waters
(males are ripe March to August, and females are ripe April to August). Peak spawning activity occurs
from late April through June. The extended spawning season throughout the range is a major factor in
variability of size within a year class.

Published accounts cite attainment of sexual maturity at age II for males and age III for females. I
conclude that weakfish of both sexes reach sexual maturity as yearling fish, although some smaller
members of a year class do not mature until their second year.

Weight and length of weakfish are better indicators of fecundity than is age (higher correlation
coefficients). A female weakfish of 500 mm standard length produces slightly over two million eggs.

The weakfish, Cynoscion regalis, is a littoral
species of commercial and sport importance in the
middle Atlantic states from North Carolina to
New York (Bigelow and Schroeder 1953). Welsh
and Breder (1923), Higgins and Pearson (1928),
Hildebrand and Schroeder (1927), Hildebrand
and Cable (1934), Pearson (1941), Roelofs (1951),
and Harmic (1958) described portions of the re-
productive biology of weakfish. The most recent
data concerning reproductive biology of this
species in North Carolina were in Hildebrand and
Cable (1934).

The decline in commercial catch of weakfish be-
tween 1945 and the mid-1960's and speculation as
to its cause(s) (Roelofs 1951; Perlmutter 1959;
Fahy 1965a, b; Brown and McCoy 1969; Joseph
1972) indicated the need for a biological study of
the weakfish along the Atlantic coast (Nesbit
1954; Perlmutter et al. 1956; Massmann et al.
1958). I undertook a study of the weakfish in
North Carolina (1967-70) to provide biological
data from which recommendations for manage-
ment could be formulated. This paper presents
data on reproduction of weakfish pertaining to:
1) spawning season, 2) age and size at which sex-
ual maturity is attained, 3) fecundity relation-
ships, and 4) possible role of reproductive biology
in the observed population decline along the east-
ern seaboard.

'Adapted from part of a thesis submitted in partial fulfill-
ment of the requirements for Ph.D. in the Zoology Department,
North Carolina State University, Raleigh, NC 27607. Financial
support was provided by the Sport Fishing Institute.

2 Virginia Institute of Marine Science Contribution No. 699.

^Virginia Institute of Marine Science, Gloucester Point,
VA 23062.

Manuscript accepted July 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

MATERIALS AND METHODS

A total of 3,635 weakfish were obtained for
biological examination from the area bounded by
Cape Hatteras and Cape Fear, N.C. Landings of
pound nets, haul seines, gill nets, and shrimp
trawls in the vicinity of Cape Hatteras, between
June 1967 and November 1969, contributed 1,606
specimens (Figure 1). An additional 2,029 weak-
fish were obtained between June 1967 and
January 1970 from trawler landings in Morehead
City and Beaufort, and from haul seines landing
in Atlantic and Sea Level (Figure 1).

MORTH
CAROLINA

HATTERAS

« CAPE LOOKOUT

CAPE FEAR

re" 00'

78 » 00'

FIGURE 1. — Location of sampling sites included in 1967 to
1970 collections of weakfish from North Carolina waters.

MERRINER: REPRODUCTIVE BIOLOGY OF THE WEAKFISH

Scale samples were taken from under the tip of
the pectoral fin below the lateral line of 2,159
weakfish for age determination. Age-group or
age-class cited herein refers to the number of an-
nuli on scales. Weight in grams and length (total,
fork, and standard) in millimeters were recorded
from all specimens.

Sex and maturation stage of gonads were as-
signed after macroscopic examination of the
gonads using a modification of the classification
of Kesteven (1960). Histological sections of repre-
sentative gonads in each stage provided verifica-
tion of maturation class assignment (Table 1).

Gonad index indicated duration and peak of
spawning season as well as the age and size at
which weakfish attain sexual maturity. Gonads
from 571 females and 117 males from the Hat-
teras and Morehead City areas were preserved in
10% Formalin^ and used for analysis of gonad
condition. The index value equals the weight of
the preserved gonad, to the nearest 0.01 g, di-
vided by the body weight of the fish, to the
nearest 1.0 g, times 100. It represents the percent
contribution of gonads to total fish weight.

Twenty-two female weakfish with well-
developed oocytes (mature ovaries) provided the
basis for fecundity relationships. Age-groups I
through IV are represented by 20 fish collected
between 25 May and 13 June 1969, from Pamlico
Sound. Age-group 0 is represented by two females
collected near Morehead City on 4 June 1968. The
preserved ovaries were blotted dry and weighed
to the nearest 0.01 g. One ovary from each pair
was randomly selected for sampling. A thin slice
(1-2 mm) was cut from the anterior, middle, and
posterior regions of the ovary. These slices were
weighed to the nearest 0.0001 g and placed in
Gilson's solution for 8 to 12 h to facilitate egg
separation from connective tissue (Bagenal
1967). Then the sections were rinsed with tap

"Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

water and teased apart with dissecting needles.
The separated egg samples were placed in a petri
dish which was areally divided as a 6 x 6 grid and
stirred until equally distributed within the dish
before counting. Specific grid sectors were ran-
domly selected. The portion of the sample counted
ranged from one-ninth in larger ovaries to a total
count in small ovaries. Counts were made using a
dissecting microscope and included all eggs hav-
ing yolk deposition equal to or greater than the
diameter of the oil globule.

Treatment of fecundity data included analysis
of variance for age-groups 0 through IV (Steel
and Torrie 1960) and linear regression. Fecundity
was related to total length (TL) and standard
length (SL) of the fish in millimeters using the
equation,

F = aL^

where F = fecundity,

L = total or standard length of the fish,
a, b = constants for the equation.

Fecundity was related to fish weight in grams
using the equation,

F = a +bW

where F = fecundity,
W = fish weight,
a, b = constants for the equation.

RESULTS

Monthly summaries of testes maturation class-
es revealed an extended spawning season and an
early summer peak in spawning for male weak-
fish. Over one-fourth of the males sampled from
March through August were ripe running (Table
2) and over one-half were ripe running from April
through July. During September and October,

Table l.— Gonad stage designations and macroscopic condition of the male drumming muscle used in describing

weakfish maturity.

Female

Male

Gonad stage'

Immature

(1)

Mature

(II, III, and IV)

Ripe

(V)

Ripe running

(VI)

Ripe spent

(VII)

Spent

(VIII)

Spent resorbing

(VIII and II)

Resorbing

(1 for larger fish)

Immature

(1)

Mature

(II and III)

Ripe

(IV)

Ripe running

(V and VI)

Ripe spent

(VII)

Spent

(VIII)

Spent resorbing

(VIII and II)

Resorbing

(1 for larger fisti)

Condition of drumming muscle

Whiite, undeveloped

Pink, beginning to ttiicken

Red, ttiickened

Deep red, very thick

Red to deep red, thinner

Mottled red to pale red, thinner

Pink, thin

Pink to white, thin

'Roman numerals indicate the corresponding stage for the Kesteven scheme (1960).

FISHERY BULLETIN: VOL. 74, NO. 1

Table 2. — Gonad condition for male weakfish from North Carolina as a percent of the

monthly sample.

Ripe

Spent

Month

Number

Immature

Mature

Ripe

running

Spent

spent

resorbing Resorbing

January

30.8

38.4

23.1

7.7

March

201

3.5

30.3

39.8

26.4

April

100.0

May

100.0

June

121

3.3

90.1

July

137

13.1

3.7

80.3

2.9

August

173

34.7

5.3

46.8

12.1

September

189

20.1

0.5

4.2

47.6

13.8

October

26.3

2.6

46.1

7.9

14.5 2.6

November

27.3

72.7

December

27.3

72.7

Total

968

45% of the males were in the spent condition
while only 4.2% and 2.6% respectively were ripe
running. Male weakfish examined during No-
vember and December were not in the ripe run-
ning stage. Testes of fish collected in December
were developing, and the drumming muscle
was enlarging for the next spawning season
(Table 1).

The gradual progression of ovarian maturation
state from mature (dominant in March) to resorb-
ing (dominant in December) suggests an ex-
tended spawning season for female weakfish.
Females were in the ripe or ripe running stage
from March through September (Table 3). From
April through July, over one-fourth of the females
were in the ripe category. Female weakfish com-
pleted spawning by October. Over 30% of the
ovaries were in the spent resorbing condition dur-
ing September and October.

Evidence of multiple spawning during a given
season by individual fish in age-groups I and
older was found during analysis of ovarian condi-
tion. Ovaries contained mature follicles during
April and May with clusters of immature follicles
interspersed among translucent oocytes. These
ovaries were staged as ripe or ripe running de-
pending upon the extrusibility of oocytes. During
June, 26.6% of the ovaries were classified as ripe

spent (Table 3). These ovaries still possessed ma-
ture follicles, but were flaccid relative to those of
April and May, showed hemorrhage, and the
clusters of immature follicles were maturing (en-
larging). Ovaries staged as ripe or ripe running in
July and August were rather flaccid relative to
ovaries collected in the spring and did not have a
hemorrhagic appearance. In late August and Sep-
tember, the ovaries possessed spent characteris-
tics including atresia of remaining follicles. In
the fall, ovaries exhibited further resorption of
follicles and flaccid condition. The ovaries gradu-
ally resumed firmness after resorption.

The distribution of gonad index for male and
female weakfish of all age-classes was unimodal
with the greatest contribution of gonad to total
body weight occurring in the early summer. Tes-
ticular indices peaked in May at 2.6% and de-
clined to 0.5% in July (Figure 2). After July no
change in gonad index occurred until the next
spawning season. Ovarian indices reached
maximum values in May (8.3%) and declined to
an autumn low of less than 0.1% in September
(Figure 2). Both male and female indices from
April through June were considerably greater
than those of either March or July. Mean monthly
gonad indices reveal the major spawning period
to be April through June. However, some males

Table 3. — Gonad condition of female weakfish from North Carolina as a percent of the

monthly sample.

Ripe

Spent

Month

Number

Immature

Mature

Ripe

running

Spent

spent

resorbing

Resorbing

January

57.1

42.9

March

148

13.5

79.7

6.8

April

33.3

66.7

May

1.8

38.6

56.1

3.5

June

173

26.6

4.6

31.8

4.0

5.8

26.6

0.6

July

186

24.2

9.1

27.4

5.9

13,4

17.2

2.2

0.5

August

221

35.7

1.4

15.4

31 7

11.7

3.2

September

128

14.0

0.8

1.6

406

41.4

1.6

October

33.9

12.9

322

21.0

December

22.2

77.8

Total

1,000

MERRINER: REPRODUCTIVE BIOLOGY OF THE WEAKFISH
3.00-1

2.50
2.00 H
1.50
1.00
0.50
0.00 H

MALES
n= I I 7

S 900 n

8.00-
7.00-

8 6.00-

5.00-
4.00-
3.00-
2.00-
1.00-
0.00-

FEMALES
n= 577

-I—
M

-I—
A

-I—
M

n r

J J

MONTH

-r-
A

-I [—

0 N

-1
D

Figure 2. — Mean monthly gonad index for male and female
weakfish of all age-classes expressed as percent body weight.

and females were in spawning condition from
March through September.

Age 0 weakfish (no scale annulus) exhibited a
seasonal gonad index pattern similar to that of
older fish. The peak index values for age 0
females occurred in June, a month later than age
I females (Figure 3). Gonads of age 0 females ac-
counted for only 4% of the total body weight
whereas they represented over 8% of body weight
in age I females.

Over one-half of the age 0 weakfish collected
were classified as mature (Table 4). Of the 201
age 0 females, 105 or 52% were mature. Of the

Table 4. — Number of immature and mature age-group 0
weakfish from North Carolina by month.

Female

Male

Month

Immature

Mature

Immature

Mature

January

March

115

May

June

July

August

September

October

November

December

—

Total

105

100

163

4.0
3.0 H
^ 2.0

s«

■^ 1.0

'H 0.0

AGE-CLASS 0
n=52

< 8.0

^ 6.0 4

4.0

2.0
0

AGE-CLASS I
n=3l8

A M J J
MONTH

A S 0 N 0

Figure 3. — Mean monthly gonad index for female weakfish
of age-class 0 and I expressed as percent body weight.

263 age 0 males, 163 or 62% were mature. Dele-
tion of obvious young of the year fish collected
after 3 to 4 mo growth elevated the percent ma-
ture to 91 for males and 68 for females in age-
group 0.

Male weakfish attain sexual maturity at a
smaller size than do female weakfish and both
sexes attain sexual maturity at smaller size in
the vicinity of Morehead City than in Pamlico
Sound (Table 5). The standard length range in
which 50% of the weakfish were classified as ma-
ture, ripe, or ripe running was considered the size
at which sexual maturity is attained. Weakfish
less than 100 mm SL were not sexually mature in
either area. Males from the Morehead City area
reached the 50% criterion at about 130 mm SL (n
= 1). Male weakfish from Pamlico Sound fulfill
the criterion for population maturity at about 150
mm SL (n = 13, 61% mature). Female weakfish
fi-om the Morehead City area attain maturity at
about 145 mm SL (n = 11, 54% mature), while
female weakfish from the Pamlico Sound area at-
tain sexual maturity at about 190 mm SL {n = 28,
57% mature).

Size of the individual fish rather than age is the
dominant factor affecting the attainment of sex-
ual maturity by weakfish. In the vicinity of
Morehead City, male weakfish of age-group I but
less than 170 mm SL were mature (n = 12, 1
immature) (Table 6). All age II male weakfish
examined from that area were mature (n = 26).
Females of 175 mm SL or larger from the same
area with no annulus on their scales were ma-
ture. There was only one immature fish among

FISHERY BULLETIN: VOL. 74, NO. 1

Table 5. — Relationship of standard length and percent mature for weakfish from
North Carolina by sex and area (1967-69).

«100
105
110
115
120
125
130
135
140
145
150
155
160
165
170
175
180
185
190
195
200
205
210
215
220
225
230
235
240

>240

Total

Standard

length'

(mm)

Female

Male

Pamlico Sound Morehead City

Number % mature Number % mature

100

469

100

383

100

411

473

'Midpoint of length interval (102.6 to 107.5 = 105, etc.]

Table 6. — Relationship of age-group and standard length to percent sexually mature by sex for weakfish from the vicinity of

Morehead City, N.C. ( 1968-69).

slOO
110
115
120
125
130
135
140
145
150
155
160
165
170
175
180
185
190
195
200

>200

Total
Mean

Standard
length'

Age-group 0

Age-group 1

Age-group II

Male Female

(mm)

Number % mature Number % mature

42
2

203

100

122

244

229

1
1

1
29

100

100
100

100

'Midpoint of length interval.

the female weakfish of age-group I less than 175
mm SL (n = 21). All females in age-group II were
mature (n = 32). Males of age-group 0 from Pam-
lico Sound were sexually mature at 150 mm SL
while males in age-group I reached maturity at

135 mm SL (Table 7). All age-group II males
examined from this area were sexually mature {n
= 89). Females of age-group 0 in the Pamlico
Sound area reached sexual maturity at 175 mm
SL (n = 3, 1 immature) while age- group I females

MERRINER: REPRODUCTIVE BIOLOGY OF THE WEAKFISH

Table 7. — Relationship of age-group and standard length to percent sexually mature by sex for weakfish from Pamlico Sound,

N.C. (1967-69).

Standard

length'

(mm)

Age-group

Age-gro

up 1

Age-groL

P II

Male

Female

Male

Female

Male

Female

Number %

mature

umber

mature

Number

mature

Number

mature

Number

% mature

Number

% mature

105

100

125

130

135

140

145

150

100

155

160

165

100

170

100

175

100

180

100

185

100

190

100

195

100

200

100

205

100

210

100

215

100

220

100

>220

100

Total

262

300

100

Mean

100

'Midpoint of lengtfi interval.

were mature at a length of 190 mm (n = 25, 52%
mature) (Table 7).

Average estimated fecundity increased with
age from 45,000 eggs for age 0 females to
1,726,000 eggs for age IV females. The increases
in fecundity with age were significant (F = 15.64,
df = 17.4; P < 0.01; Table 8). Variation within
individual age groups was great with the stan-
dard deviation approaching one-third of the mean
estimated fecundity. Relative fecundity, the
number of eggs per gram of ovary, decreased from
37,650 at age 0 to 14,867 at age IV.

Regression analysis indicated significant rela-
tionships between fecundity and fish length and
weight. The equations describing the relation-
ships and coefficients of determination are:

F --
F --
F --

Table 8.

0.116SL2"55^^2 = 0.85;

0.152 TL2-64i8^ ^2 = 0.86 (Figure 4);

21,198 + 1,279 W, r2 = 0.88.

DISCUSSION

Weakfish spawn in or near the various inlets
along the coast of North Carolina (Welsh and
Breder 1923; Higgins and Pearson 1928; Hilde-
brand and Cable 1934) and also in Pamlico
Sound. Earlier authors did not include sounds
and bays as probable spawning sites since no
female weakfish in spawning condition had been
taken from inshore waters of North Carolina
(Roelofs 1951). Higgins and Pearson (1928) re-
ported a few weakfish with "free running ripe
eggs" in Pamlico Sound. Twenty-four female
weakfish in the ripe running condition were ob-
tained from Pamlico Sound, and this indicates
weakfish may also spawn in sounds and bays.
These areas may be at the edge of the spawning
zone, however.

Spawning activity in coastal waters north of
North Carolina is cited by Hildebrand and

-Fecundity estimates and relative fecundity for 22 weakfish from North Carolina and analysis of variance results for

age versus fecundity.

Age-
group

Number
examined

Mean
fecundity
estimates

Standard
deviation

Mean no.
of eggs
per gram
of ovary

Standard
length
range
(mm)

Anova

Source

Sum of squares

Mean square

0
1

II
III
IV

Total

2
8

7
2
3

44,880

285.740

579,660

491,700

1,725,920

10,693
105,600
302,700
186,900
614,300

37,650
21.225
19,400
15,150
14,867

145-160
190-268
245-308
292-335
395-480

Age
Error
Total

4
17
21

5.219 X 10'2
1.418 X 10'2
6.637 X 10'2

1.305 X 10'2
8.341 X 10'"

15.64*'

"Probability less than 0.01.

FISHERY BULLETIN: VOL. 74, NO. 1

O
O

26-1

24-

22-

20-

18-

16-

14-

12-

10-

FECUNDITY = 0.152 TL
0.86

2.6418

FECUNDITY = 0.116 SL^--'']^
'2 = 0.85

100 200 300 400 500

FISH LENGTH (mm)

600

FIGURE 4.-

- Relationship of weakfish fecundity to fish length
based upon data from 22 females.

Schroeder (1927), Pearson (1941), and Massman
(1963) for Chesapeake Bay; by Parr (1933),
Daiber (1954), Harmic (1958), and Thomas (1971)
for Delaware Bay; by Nesbit (1954) and Perlmut-
ter et al. (1956) for New York and New Jersey
waters; and by Bigelow and Schroeder (1953) for
the Gulf of Maine. However, the magnitude of
spawning in northern areas is unknown. Progeny
from spawning activity north of Chesapeake Bay
are considered insufficient to maintain the north-
ern stock (Harmic 1958), and young from the
Carolinas and Chesapeake Bay are thought to be
recruited to the northern population as age III or
older fish (Pearson 1941; Nesbit 1954; Perlmutter
et al. 1956; Harmic 1958). The validity of this
supposition remains to be documented.

Mature weakfish enter the inshore waters,
sounds, and bays of North Carolina in early
spring (Hildebrand and Schroeder 1927; Hilde-
brand and Cable 1934; Roelofs 1951). Fertilized
eggs have been taken in Delaware Bay when
water temperatures ranged from 17° to 26.5°C
and at salinities from 12.1 to 31.3'L (Harmic 1958).

Weakfish apparently have an extended spawn-
ing season in North Carolina waters as reported
by Welsh and Breder (1923), Higgins and Pearson
(1928), Hildebrand and Cable (1934), and Pear-
son (1941). Distributional data for weakfish eggs
and larvae are lacking in North Carolina waters.
Peak spawning activity occurs from late April

through June as indicated by gonad condition and
gonadal index. Females appear to spawn the
major portion of their eggs in May or June with a
second spawn of smaller magnitude possibly oc-
curring in late July or August. Thus, weakfish of
a given year class may vary considerably in size
due to their extended spawning season and mul-
tiple spawning by females.

Weakfish males and females probably attain
sexual maturity as 1-yr-old fish throughout
their geographic range, though some of the small-
er members of a year class may not mature until
their second year of life. Weakfish in North
Carolina waters were previously reported to
reach sexual maturity at age II for males and age
III for females (Taylor 1916; Welsh and Breder
1923; Higgins and Pearson 1928), and subsequent
papers have reiterated these ages without ver-
ification. Higgins and Pearson (1928) reported no
mature females less than 200 mm fork length
(approximately 170 mm SL) and that a fork length
of 230 mm was attained before 50% of the female
weakfish mature in Pamlico Sound. This size
group was allocated to age-group III without
examining scales for annuli. I consider their allo-
cation of age-classes to be in error on the basis of
data presented here and in Merriner (1973). I
found 21 mature female weakfish 170 mm SL in
samples from Pamlico Sound and 90 mature
female weakfish of the same size from the vicinity
of Morehead City. Over one-half of the female
weakfish were mature at 190 mm SL in samples
from Pamlico Sound, and male weakfish become
sexually mature at a smaller size than females.
Weakfish spawned in May or June would be ma-
ture the following May or June. Those fish
spawned in late July or August probably would
not be sexually mature until late summer of the
year following their hatch or the following spring.
Scrap samples from pound nets in Chesapeake
Bay contained mature female weakfish measur-
ing 170 to 250 mm TL during late spring and
summer months (McHugh 1960). Maturation at a
small size is also likely for fish from more north-
erly areas (Daiber 1954; Thomas 1971).

No evidence of alternate year spawning was
found even in the oldest specimens examined. All
of the females of age III or older were either in
spawning condition or mature during early sum-
mer. However, some of the older weakfish in the
population may not migrate inshore during
spring and summer.

Weakfish are characterized by high fecundity.

MERRI>fER: REPRODUCTIVE BIOLOGY OF THE WEAKFISH

In Delaware Bay a female weakfish, 190 mm SL,
contained a total of 267,500 eggs and would re-
lease approximately 52,000 eggs at one spawning
(Daiber 1954). My estimates of fecundity for
females of a similar size are equivalent to the
total egg production figure for Delaware Bay.
Fecundity increases by approximately 106,000
eggs for each 100 g of body weight for weakfish in
Delaware Bay, while my data indicate an in-
crease of 127,900 eggs per 100 g of body weight.

The variation in fecundity per age-group is best
explained by the size range present in the sam-
ples of each age-group. Regression analysis
showed a significant relationship between fecun-
dity and fish length (coefficient of determination
= r^ = 0.85) and between fecundity and fish
weight (r^ = 0.88). The average range of standard
length for all females in age-groups 0 to IV was
57 mm. High variability in fecundity estimates
for age-groups is expected due to the range in fish
size and variation in gonad size among fish of the
same size (Bagenal 1967).

It is highly unlikely that weakfish experienced
a synchronous failure or severe depression of em-
bryonic or larval survival in all spawning areas.
Harmic (1958) analyzed the early life history of
weakfish in Delaware Bay. Fertilized eggs are
pelagic and measure from 0.87 to 0.99 mm in
diameter. Weakfish larvae emerge after about 40
h at water temperatures of 68° to 70° F and aver-
age 1.8 mm SL. Soon after hatching, the demersal
larvae disperse into the nursery areas. Through-
out the coastal waters from North Carolina to at
least New York, anomalous water conditions
(such as rapid changes in salinity, temperature,
or dissolved oxygen) may occur in small areas due
to local weather phenomena or industrial-
domestic development. Hurricanes, however, may
affect the entire eastern seaboard (tropical storm
Agnes — 1972) or portions of it (Hurricane
Camille — 1969) with the greatest impact occur-
ring in the estuarine areas (i.e., weakfish nur-
sery). The extended spawning season of weakfish
would tend to minimize any effect of a short-term
calamity upon a local population.

Tolerance of weakfish eggs and larvae to tem-
perature, salinity, dissolved oxygen, etc., remains
poorly known. According to data compiled by
Harmic (1958), natural fluctuations in the es-
tuary approach the ranges that are detrimental
to weakfish survival. For Delaware Bay and pre-
sumably throughout its range, the variation in
water parameters due to natural phenomena

alone may largely explain fluctuations in the
weakfish population abundance and year class
strength.

LITERATURE CITED

Bagenal, T. B.

1967. A short review of fish fecundity. In S. D. Gerking
(editor), The biological basis of freshwater fish produc-
tion, p. 89-111. John Wiley and Sons Inc., N.Y.
BIGELOW, H. B., AND W. C. SCHROEDER.

1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv.,
Fish. Bull. 53, 577 p.

BRowTj, J., AND E. McCoy.

1969. A review of the North Carolina scrap fishery. N.C.
Dep. Conserv. Dev., Div., Conuner. Sport Fish., Mimeo.
Rep., 13 p.
Daiber, F. C.

1954. Fisheries research program. Mar. Lab. Dep. Biol.
Sci., Univ. Del. Biennial Rep. 1953 and 1954. Publ.
2:50-64.

fahy, W. E.

1965a. Report of trash-fish study in North Carolina in 1962.

Div. Commer. Fish., N.C. Dep. Conserv. Dev., Spec. Sci.

Rep. 5, Mimeo., 20 p.
1965b. Report of trash-fish study in North Carolina in 1964.

Div. Commer. Fish., N.C. Dep. Conserv. Dev., Spec. Sci.

Rep. 7, Mimeo., 13 p.
Harmic, J. L.

1958. Some aspects of the development and ecology of
the pelagic phase of the gray squeteague, Cynoscion
regalis (Bloch and Schneider), in the Delaware estuary.
Thesis, Univ. Delaware, Newark, 84 p.
HIGGINS, E., AND J. C. PEARSON.

1928. Examination of the summer fisheries of Pamlico and
Core sounds, N.C, with special reference to the de-
struction of undersized fish and the protection of the
gray trout Cynoscion regalis (Bloch and Schneider).
Rep. U.S. Comm. Fish., 1927 append. 2:29-65.
HILDEBRAND, S. F., AND L. E. CABLE.

1934. Reproduction and development of whitings or king-
fishes, drums, spot, croaker, and weakfishes or sea-
trouts, family Sciaenidae, of the Atlantic Coast of the
United States. U.S. Bur. Fish., Bull. 48:41-117.
HILDEBRAND, S. F., AND W. C. SCHROEDER.

1927. Fishes of Chesapeake Bay. U.S. Bur. Fish., Bull.
43:1-366.
JOSEPH, E. B.

1972. The status of the sciaenid stocks of the middle
Atlantic Coast. Chesapeake Sci. 13:87-100.
KESTEVEN, G. L. (editor).

1960. Manual of field methods in fisheries biology. FAO
Man. Fish. Sci. 1, 152 p.
MCHUGH, J. L.

1960. The pound-net fishery in Virginia. Part 2 - Species
composition of landings reported as menhaden. Com-
mer. Fish. Rev. 22(2):1-16.
MASSMANN, W. H.

1963. Age and size composition of weakfish, Cynoscion
regalis, from pound nets in Chesapeake Bay, Virginia
1954-1958. Chesapeake Sci. 4:43-51.
MASSMANN, W. H., J. P. WHITCOMB, AND A. L. PACHECO.
1958. Distribution and abundance of gray weakfish in the

FISHERY BULLETIN: VOL. 74, NO. 1

York River system, Virginia. Trans. 22nd North Aro.

Wildl. Conf. , p. 361-369.
MERRINER, J. V.

1973. Assessment of the weakfish resource, a suggested

management plan, and aspects of life history in North

Carolina. Ph.D. Thesis, North Carolina State Univ.,

Raleigh, 201 p.
NESBIT, r. a.

1954. Weakfish migration in relation to its conservation.

U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 115,

81 p.
PARR, A. E.

1933. A geographical-ecological analysis of the seasonal

changes in temperature conditions in shallow water

along the Atlantic Coast of the United States. Bull.

Bingham Oceanogr. Collect. Yale Univ. 4(3): 1-90.
PEARSON, J. C.

1941. The young of some marine fishes taken in lower

Chesapeake Bay, Virginia, with special reference to the

gray sea trout, Cynoscion regalis (Bloch). U.S. Bur.

Fish., Bull. 50:79-102.

Perlmutter, a.

1959. Changes in the populations of fishes and in their
fisheries in the Middle Atlantic and Chesapeake regions,
1930 to 1955. Trans. N.Y. Acad. Sci., Ser. H, 21:484-
496.

Perlmutter, a., W. S. miller, AND J. C. Poole.

1956. The weakfish {Cynoscion regalis) in New York
waters. N.Y. Fish Game J. 3:1-43.
ROELOFS, E. W.

1951. The edible finfishes of North Carolina. In H. F.
Taylor (editor), Survey of marine fisheries of North
Carolina, p. 109-139. Univ. N.C. Press, Chapel Hill.

Steel, R. G. d., and J. H. Torrie.

I960. Principles and procedures of statistics with special
reference to the Biological Sciences. McGraw-Hill Book
Co., N.Y., 481 p.
Taylor, H. F.

1916. The structure and growth of the scales of the
squeteague and the pigfish as indicative of life history.
U.S. Bur. Fish., Bull. 34:285-330.
THOMAS, D. L.

1971. The early life history and ecology of six species of
drum (Sciaenidae) in the lower Delaware River, a brack-
ish tidal estuary. Ichthyol. Assoc. Bull. 3. An ecological
study of the Delaware River in the vicinity of artificial
island. Delaware Progress Report for the period January-
December 1970. Part III, 247 p.

Welsh, w. W., and C. M. Breder, Jr.

1923. Contributions to life histories of Sciaenidae of the
eastern United States coast. U.S. Bur. Fish., Bull.
39:141-201.

DDT AND ITS METABOLITES IN THE SEDIMENTS
OFF SOUTHERN CALIFORNIA

John S. MacGregori

ABSTRACT

To assess the degree of DDT contamination in the marine sediments off Los Angeles, 103 stations in
the Pacific Ocean off southern Cahfomia were sampled in July and August 1971 for DDT and its
metabolites, DDD and DDE. Heavy contamination of bottom sediments in this area was expected
because of large amounts of DDT that have entered the ocean through the Los Angeles County sewer
system as waste from a DDT manufacturing plant.

From the data acquired, it was estimated that there were about 200 metric tons of DDT, DDD, and
DDE in the sediments in an area of 14 square nautical miles near the sewer outfalls and 300 metric
tons in the entire 911 square nautical mile area sampled. The heaviest concentrations of total DDT
were distributed in the relatively shallow- water area on the Palos Verdes shelf to the northwest of the
sewer outfalls in the general direction of the current flow.

Metabolism of DDT was inhibited in deepwater sediments. Ratios of DDE to DDT were low, and
DDT was more abundant than DDE at some stations. In sediments from shallow- water stations, DDE
exceeded DDT by more than 10 times.

The bottom of the ocean off Los Angeles, CaHf,
has been very heavily contaminated with the
pesticide DDT owing to the discharge of wastes
from a DDT manufacturing plant into the Los
Angeles County sewer system over a period of
about 20 yr ending in 1970 (MacGregor 1974).

The amount of DDT which entered the ocean
through the Los Angeles County sewer system
was estimated at 250 kg/day. Following the ces-
sation of DDT discharges by the manufacturer,
the amount entering the ocean dropped to 45
kg/day in December 1970 and to 11 kg/day in
October 1971. Most of these later discharges re-
sulted from sewer cleaning operations which
stirred up old deposits of DDT in the sewer lines.
The discharges resulting from the cleaning oper-
ations were primarily DDD and DDE, metabo-
lites of DDT, while the earlier discharges were
primarily DDT.

Because there has been a great deal of specula-
tion about the fate of DDT and other toxic chem-
icals released into the environment by man
(Woodwell et. al. 1971; National Academy of
Sciences 1971), this investigation was under-
taken to determine the areal distribution and
fate of these chemicals in the bottom sediments
in the ocean off Los Angeles.

^Southwest Fisheries Center, La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

MATERIALS AND METHODS

The bottom sediments were sampled from a
grid of 103 stations between lat. 33°30' and
33°58'N and long. 118°00' and 118°44'W (Figure
1). The stations were designated by four-digit
numbers, the first two indicating minutes north

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

Figure l. — Distribution of total DDT in milligrams per square
meter of bottom in the sediments of southern California. Total
DDT ranged from 6,600 mg/m^ of bottom at station 43-22 to
0.12 mg/m^ at station 30-08.

FISHERY BULLETIN: VOL. 74, NO. 1

of lat. 33°N and the second two indicating min-
utes west of long. 118°W.

The samples were taken aboard the National
Marine Fisheries Service RV David Starr Jordan
between 26 July and 3 August 1971. The Shipek
bottom sampler was used to obtain the samples of
sediment. This device obtains a block of material
equal to 400 cm^ of bottom sediment to a depth of
about 10 cm, or slightly more, in soft mud or to a
depth of half as much or less in coarse sand.

Two samples were taken at each station in
order to obtain an estimate of sampling error.
The vessel was allowed to drift while the samples
were being taken, so the sample pairs were taken
in only approximately the same location. How-
ever, agreement in the various parameters be-
tween samples from the same station was good.

The samples were placed in aluminum foil-
lined containers of approximately the same size
as the sampling bucket and were quick-frozen.
They were stored in a freezer until removed for
analysis.

In most samples, DDT was confined to the top 2
or 3 cm of the sediment. At most of the stations
where the sampler sampled to 10 cm, and at all of
the stations where it sampled to a lesser depth, it
appeared that all of the DDT under the 400 cm^
had been sampled. In this study, therefore, DDT
concentrations are given as the weight of DDT
per unit area of bottom to a depth of 10 cm. In a
few areas of rapid sedimentation, where the sam-
pler sampled to about 10 cm depth, there were
still significant amounts of DDT below 10 cm.
Estimates for the amounts of DDT below 10 cm
are based on core samples taken by other investi-
gators in this area.

The bottom sediment samples were thawed
and blended in a 1-gallon Waring^ commercial
blender. Before blending, small stones were re-
moved from the few samples that contained
them. Some samples contained a few small mol-
luscs or brittle stars, but these were not removed.
Measured amounts of distilled water were added
to some of the drier (sandy) samples to facilitate
blending.

A sample of 15 to 20 g of blended sediment was
weighed onto a watch glass, dried to constant
weight, and reweighed to obtain percent water in
the sediment. This gave an index of bottom type

^Reference to trade name does not imply endorsement by the
National Marine Fisheries Service, NOAA.

ranging from 30 to 40% water for coarse sand to
60 to 70% water for fine silt.

A second sample weighing about 30 g was
weighed into a 1-pint Mason jar for DDT deter-
mination. About four or five times as much
Na2S04 was weighed into the jar as a drying
agent. The sediment and Na2S04 were mixed
using a stainless steel spatula, and the mixture
was frozen. A cutting assembly was fitted to the
jar, and the frozen mixture was thoroughly
blended to a powder using an Osterizer blender.

About 5 g of the powder was weighed into a
tared, large disposable pipet (Matheson super
pipet) plugged with glass wool. The powder was
extracted into a 15-ml graduated centrifuge tube
with 5 ml of hexane and 5 ml of acetone. The
extract was evaporated to dryness and redis-
solved in 1 ml of hexane. This sample was eluted
through a super pipet filled with activated alu-
mina (McClure 1972) using enough hexane to
obtain a 6-ml sample.

This sample was reduced or increased in vol-
ume as required and injected into a model 402
Hewlett Packard gas chromatograph (GLC) with
a Ni^^ electron capture detector. The 6-foot glass
column contained 4% SE-30/6% QF-1 on 100/120
mesh Supelcoport.

There was evidence of a polychlorinated bi-
phenyl, Aroclor 1254, in all samples, but the
DDT peaks were so dominant in the chromato-
grams that they generally obliterated any traces
of other chlorinated hydrocarbons within their
range. Only the six peaks representing the ortho-
para and para-para forms of DDE, DDD, and
DDT were quantified. "Total DDT" is used to
designate the sum of these six analogs.

RESULTS AND DISCUSSION

Fifty-five correlations were obtained for 11
parameters to determine various DDT relation-
ships. The 55 correlations were obtained for all
103 stations (Table 1, values above 1.000 correla-
tion diagonal) and for 76 stations leaving out
those 27 stations having total DDT readings
greater than 100 mg/m^ (Table 1, values below
1.000 correlation diagonal). For 100 observations
a correlation coefficient of 0.254 indicates a
probability of 0.01. Logarithms were used for
total DDT and distance from outfall, arithmetic
values for the other nine measurements.

There is a very high negative correlation be-
tween log total DDT and log distance from the

MacGREGOR: DDT OFF SOUTHERN CALIFORNIA

Table l. — Correlation coefficients for 11 parameters relating to DDT and its metabolites in bottom sediments off southern California.
Values above 1.000 correlation diagonal are for 103 stations. Values below diagonal are for 76 stations leaving out those 27 stations
having total DDT readings greater than 100 mg/m^. For 100 observations, a correlation coefficient of 0.254 indicates a probability
of 0.01.

Log
total

Log

distance

from

Sample

% H2O
in

p,p DDD

p,p' DDE

o,p' DDE

o,p' DDD

o,p' DDT

Parameter

DDT

outfall

Depth

weight

sample

p,p' DDT

p.pDDD

P.P'DDE

p,p' DDD

p,p' DDT

Log total DDT

1.000

-0.871

-0.253

0.221

0.157

0.142

0.144

0.281

-0,040

-0,272

-0,334

Log distance

from outfall

-0.604

1.000

0.228

-0.095

-0.147

-0.032

-0.043

-0.238

0,016

0.332

0,334

Depth

0.078

-0.036

1.000

0.643

0.771

-0.443

-0.512

-0.572

0,315

0,168

0.095

Sample weight

0.245

0.041

0.761

1.000

0.743

-0.265

-0.341

-0.332

0.190

0,036

-0 151

% H2O in sample

0.095

0.002

0.921

0.756

1.000

-0.315

-0.390

-0,396

0.228

0,019

-0.013

p.p'DDD/p.p'DDT

0.123

0.066

-0,418

-0.268

-0.325

1.000

0.909

0,297

-0.168

-0.194

0.556

p,pDDE/p,pDDT

0.100

0.040

-0.492

-0.366

-0.416

0.907

1.000

0,446

-0,208

-0.171

0,463

p,p 'DDE/p,p'DDD

0.060

-0.075

-0.535

-0.418

-0.487

0.268

0.408

1.000

-0,289

-0,043

-0,011

o,p'DDE/p,pDDE

0.190

-0.234

0.296

0.225

0.248

-0.151

-0.196

-0.273

1,000

0.021

0,113

o,p 'DDD/p,pDDD

-0.129

0.275

0.106

0090

0.059

-0.177

-0.151

0.023

-0.024

1.000

0,130

o,p'DDT/p,p'DDT

-0.154

0.167

0.006

-0.127

0.025

0.630

0.537

0,084

0.068

0.060

1,000

Los Angeles County sewer outfalls (r = —0.871).
Values ranged from 6,600 mg of total DDT/m^ of
bottom near the sewer outfalls to about 1 mg/m^
at more distant stations.

The distribution of DDT was modified some-
what by currents which tended to deposit the
DDT along the coast and to the northwest more
than to the east (Figure 1). The apparent relation
between total DDT and depth results from the
fact that the sewers discharge into relatively
shallow coastal waters and the sludge tends to
remain there. The deeper waters are merely
farther from the sewer outfalls and the areas
along the coast favored by the currents.

McDermott et. al. (1974) took sediment sam-
ples from the Palos Verdes shallow-water shelf
area in the vicinity of the sewer outfalls only.
Their tables A-1 and A-4 give total DDT in parts
per million dry weight from gravity core samples
taken in 1972. I have contoured their data
(Figure 2B) for the top 10 cm of sediment to
compare with the 1971 data (Figure 2A) which
has been converted to parts per million dry
weight. Their 1973 data in their table A-5 repre-
sents parts per million dry weight of total DDT in
the top 5 cm of Shipek samples taken in the same
area (Figure 2C). In each of the 3 yr the patch of
sediment representing more than 100 ppm. total
DDT tends to retain its integrity fairly well as an
oblong area stretching to the northwest of the
sewer outfalls. The contours representing 10 to
100 ppm. seem to be expanding somewhat to the
northwest and in 1973 to the southeast also.

At the Los Angeles County sewage disposal
plant, most of the solids are removed by centri-
fuging, but the supernatant is pumped into the
ocean along with the water from the settling

tanks. This reduces the amount of particulate
matter being discharged into the ocean. Never-
theless, quantities of relatively DDT-free partic-
ulate matter have been deposited on the Palos
Verdes shelf since dumping of DDT into the sew-
er system was stopped. In time this could cause
a change in the DDT profile of the sediments.

On the other hand, most of the shallow inshore
areas along this section of coast tend to have
sandy bottoms, and the silt bottoms in the vicin-
ity of the sewer outfalls would appear to be
unstable artifacts. Storms, tides, and currents
could remove or deposit layers of bottom silts in
this shallow-water area and further change the
DDT profiles.

Based on the paired samples taken in 1971, the
variation within a sampling area for one sample
would be roughly plus 100% minus 50% at a one
standard deviation level. For an average for two
samples it would be plus 70% minus 40%. This
could account for the differences in the distribu-
ion of total DDT for the 3 yr. However, the
similarities are much more striking than the
differences.

High sample weight and high water content
both indicate samples containing more silt, while
lower weights and lower water content indicate
samples containing more sand. Both of these
measurements are related to depth, with the bot-
tom in deep basins tending to be fine silt while
shallow areas tend to be sandy. This tendency is
masked in shallow-water areas where there are
sewer outfalls which deposit large quantities of
fine material in the shallower waters. This is in-
dicated by the improvement of the correlation
coefficients from 0.643 to 0.761 for sample weight
and from 0.771 to 0.921 for percent water with

FISHERY BULLETIN: VOL. 74, NO. 1

33° 40

33° 40'

FIGURE 2.— Total DDT (parts
per million dry weight) in the
bottom sediments off Palos
Verdes in the vicinity of the
Los Angeles County sewer out-
falls. A. Shipek samples (pres-
ent paper); B. top 10 cm of
gravity cores (McDermott et al.
1974); C. top 5 cm of Shipek
samples (McDermott et al.
1974).

•n8° 20-

118° 16-

MacGREGOR: DDT OFF SOUTHERN CALIFORNIA

depth when the 27 stations of heavy sewer deposi-
tion in shallower waters are omitted.

The very high correlation coefficient (0.909) be-
tween p,p'DDD/p,p'DDT and p,p'DDE/p,p'DDT
shows that when metabolism of DDT to DDD is
high, metabolism of DDT to DDE is high also.
These high rates of metabolism are negatively
correlated with depth. Actually, they are more
probably associated with some of the conditions
prevailing at depth in the ocean off Los Angeles.
The deep areas sampled tend to be anaerobic, and
it is probably the lack of oxygen and colder water
that determines the low rate of metabolism. The
high correlations of the ratios with sample weight
and percent water are secondary effects of the
correlations of these two factors with depth.

The high negative correlation between
p,p'DDE/p,p'DDD and depth indicates that
metabolism of DDT to DDE is favored over me-
tabolism to DDD in shallower waters. However,
the positive correlation ofp,p ' DDE /p,p' DDD with
p,p'DDD/p,p'DDT (0.297) as well as with
p,p'DDE/p,p'DDT (0.446) supports the conclusion
that metabolism to both metabolites is much
greater in shallow aerobic waters than in deep
anaerobic waters. Actually, much more DDT is
probably metabolized to DDD than to DDE under
all circumstances prevailing in the study area,
but the DDE is much more persistent than the
DDD and accumulates to a greater degree while
DDD is further metabolized to DDMU and other
metabolites.

There was at least 10 times as much DDE as
DDT in the bottom sediments from stations along
the coast of the study area, while 10 stations in
deeper waters north of Santa Catalina Island had
less DDE than DDT in the bottom samples (Fig-
ures 3, 4). DDD tended to follow somewhat the
same pattern (Figure 5).

At the 10 stations the average total DDT was
19.9 mg/m2, of which 60% was DDT, 19% DDD,
and 2V/c DDE. Mean depth was 341 fathoms (623
m) and the total area represented by the 10 sta-
tions was 111 sq nautical miles containing an es-
timated 5.74 metric tons of total DDT.

It appears that most of the pesticide discharged
from the Los Angeles County sewer outfalls has
been DDT with the exception of the period of
sewer cleaning operations in 1970-71 when DDD
and DDE predominated (MacGregor 1974). Most
of the DDT settles on the bottom close to the out-
falls in shallow waters. Once the DDT becomes
part of the bottom sediment it tends to stay there

Figure 3.— Distribution of ratios ofp.p'DDE top,p'DDT. In
the shallow waters near shore the ratios exceed 10:1, while in
the deeper waters north of Santa Catalina Island the ratios
are less than 1:1.

Figure 4. — Chromatograms of: A — a deepwater sample
1274 m), station 30-40, showing highp,p'DDT peak, and D—
a shallow-water sample (36.5 m), station 40-16, showing a high
DDE peak. B and C are standards of the DDT analogs.

and metabolize in place, rapidly in shallower
waters and more slowly in deeper waters.

FISHERY BULLETIN: VOL. 74, NO. 1

Figure 5. —Distribution of ratios ofp,p'DDD to p,p' DDT. In
the shallow nearshore areas the ratios exceed 2:1, while in the
deeper waters the ratios are less than 1:1. The higher ratios
were probably enhanced by sewer cleaning operations in
1970-71.

The DDT deposits in the deeper waters must
have been transported there directly from the
sewer outlets before much metabolism could take
place. If they had originated from bottom sedi-
ments closer to the sewer outfalls and in shal-
lower waters, the DDE content would be much
higher. DDE averages about 85% of total DDT in
biological material in this area; therefore, most of
the total DDT in bottom sediments in the deeper
water could not have originated from this source.

For the time series for total DDT accumulation
in myctophid fish (MacGregor 1974), DDE was
less than DDT from 1949 to 1956, but in the
subsequent years DDE became much higher. If
the deep water with relatively low DDE had re-
sulted from biological fallout as represented by
the myctophids for 22 yr (1949-70), and if there
had been no metabolism at depth, the DDE would
have been twice as high as the DDT rather than
one-third as high.

There is either very little metabolism in deep-
water sediments, or there is no metabolism, and
the small amounts of DDD and DDE found there
are the result of fallout from material metabo-
lized in the better-oxygenated surface and inter-
mediate depths.

In commercial DDT, the ratio of p,p' DDT to
o,p'DDT is about 4:1 (i.e., o,p' DDT is about 25%
ofp,p'DDT). The distribution of these latter val-
ues for the sediment samples indicate that

o,p 'DDT is about what might be expected, while
o,p' DDD is higher and o,p' DDE is lower (Table 2).

In the case of DDT this may mean that o,p ' DDT
metabolizes as readily as p,p' DDT. The two high
positive correlations with the parameters indicat-
ing high metabolism, p,p'DDD/p,p'DDT and
p,p'DDE/p,p'DDT, may indicate that o,p' DDT
metabolizes more readily than p,p' DDT under
conditions of low metabolism of DDT to DDD and
DDE. Both the ratios of o,p' DDT to p,p' DDT and
o,p ' DDD to p,p ' DDD tend to be high in the bottom
sediments north of Santa Catalina Island and in
Santa Monica Bay, while ratios tend to be low just
south of Palos Verdes Peninsula and in the sandy
shallower waters to the east of this area (Figures
6, 7). The association of greater distance from the
sewer outfalls and lower total DDT values with
high ratios is undoubtedly fortuitous, although
the few very high ratios are associated with very
low DDT values and probably result from poorer
resulting measurements and interfering sub-
stances that are no longer completely dominated
by DDT at these very low values.

The ratios of o,p' DDE to p,p' DDE are greater
than 1.00:1.00 for 19 stations. Unlike the other
two ratios these high ratios are associated with
depth. They also tend to be concentrated in the
deeper waters just off the Palos Verdes shelf
where the sewer outfalls are located (Figure 8).
These apparent high relative values of o,p' DDE
are probably caused by interfering substances,
probably DDMU, a metabolite of DDD, which is
not being further metabolized under the condi-
tions prevailing at these stations.

Table 2. — Frequency distributions of ortho-para isomer as a
percent of para-para isomer of DDT, DDD, and DDE in bottom
sediments.

Percent

DDT

DDD

DDE

0.0-

5.0

5.1-

10.0

10 1-

15.0

15.1-

20.0

20.1-

25.0

25.1-

30.0

30.1-

35.0

35.1-

40.0

40.1-

45.0

45.1-

50.0

50.1-

55.0

55.1-

60.0

60.1-

65.0

65.1-

70.0

70.1-

75.0

75.1-

80.0

80.1-

85.0

85.1-

90.0

90.1-

95.0

95.1-

100.0

>100

MacGREGOR: DDT OFF SOUTHERN CALIFORNIA

►'«•..

«,.«•'••

O o o o

FIGURE 6.— Stations at which the ratio of o,p' DDT top.p'DDT
was greater than 0.40:1.00.

Figure 7. — Stations at which the ratio of o,p' DDD to
p,p 'DDD was greater than 0.40:1.00.

To estimate the amount of DDT stored in the
bottom sediments in the approximately 911 sq
nautical miles between long. 117°58' and
118°46'W and lat. 33°18'N and the California
coast, represented by the 103 stations, we must
assume that each station is representative of its
surrounding area. Each pair of samples from each
station showed a high correlation for all pairs of
parameters. The correlation coefficient for the
logarithms of total DDT for paired samples from
94 stations from which two samples were ob-
tained was 0.964 and the standard error of esti-
mate ±0.321.

The Shipek sampler took bottom silts only to a
depth of about 10 cm and sandy bottoms or shal-
low sediment deposits to a lesser depth. At all
stations except those where bottom deposition
was very rapid, as near sewer outfalls, all DDT in
the sediments was sampled. Near the sewer out-
falls the sample represents only DDT deposits in
the top 10 cm of sediment. The total amount of
DDT determined for the 911 sq nautical mile
sampling area was 217 metric tons in the top 10
cm of bottom sediment. Of this total, 179 metric
tons (82%) was DDE, 22 metric tons (10%) was
DDD, and 16 metric tons (8%) was DDT. McDer-
mott and Heesen (1974) found that the total DDT
in the top 5 cm of sediment consisted of 86% DDE,
11% DDD, and 3% DDT in the area of the Palos
Verdes shelf These somewhat different percent-
ages may have resulted from further metabolism
of DDT without replenishment. In addition, the
DDE percentages tend to be higher in this area,
and the DDD was increased in 1970-71 because of
sewer cleaning operations.

The total DDT ranged from an estimated 0.42
kg per sq nautical mile at station 30-08 repre-
senting 13.3 sq nautical miles to 28.6 metric tons
per sq nautical mile at station 43-22 representing
1.25 sq nautical miles.

Five stations representing 6.24 sq nautical
miles or 0.7% of the 911 sq nautical mile area
represented by the 103 stations contained 47.3%

Figure 8.— Stations at which the ratio of o,p' DDE to p,p' DDE
was greater than 1.00:1.00. The high apparent o.p'DDE values
probably were caused by DDMU which has the same retention
time as o,p 'DDE on the column used.

FISHERY BULLETIN: VOL. 74, NO. 1

of the total DDT (102.7 metric tons). Sixteen sta-
tions representing 18.1 sq nautical miles (2.0% of
total area) contained 64.0% (193 metric tons) of
the total DDT.

Subsamples taken from the tops and bottoms of
the blocks of sediment obtained with the Shipek
sampler indicated that most of the pesticide was
concentrated in the top strata of the samples ex-
cept for samples taken in the vicinity of the sewer
outfalls where deposition was very rapid. Cores
were taken from one sample taken near the sewer
outfalls and from a second taken at a greater dis-
tance from the outfalls to determine more about
vertical distribution of DDT in the sediments
(Table 3).

At station 42-36 only p,p' DDE was measured
because DDT and DDD were not readily measur-
able in the deeper sediment sections. Half of the
DDE was found in the top 2 cm, 81% in the top 4
cm, and 95% in the top 6 cm. At station 42-20,
close to the sewer outfall where sewer sediment
deposition was heavy, there was very little
change in the chlorinated hydrocarbon concen-
trations at all five depths.

Vance McClure (pers. commun.) has provided
me with a plot of the depth distribution of DDT,
DDE, DDD, and DDMU found in a box core sam-
ple taken about 1 nautical mile west-northwest of
the sewer outfall. Subsamples were taken from
the core at 3-cm intervals from 0 to 12 cm and at
6-cm intervals from 12 to 36 cm. The pesticide
values remained high through 12 cm depth and
dropped off rapidly between 12 and 18 cm. DDMU
had a deeper distribution than the other three
components and increased to a maximum at 9 cm
and was still present at 36 cm. DDE was last
measured at 24 cm, and DDD and DDT at 18 cm.
Excluding DDMU, 72% of the pesticide was found
in the column corresponding to the top 10 cm and
28% below that depth. Including DDMU, 67%
was in the top 10 cm and 33% below.

If the box core sample is typical of the stations
near the sewer undergoing rapid sedimentation,
about 30% of the pesticide was missed by sam-
pling only to a depth of 10 cm at these stations.
Because these stations near the sewer outfalls
contain most of the pesticide, the 217 metric tons
of pesticide estimated for the entire area in the
top 10 cm could be increased to roughly 300 met-
ric tons as a maximum estimate of total DDT in
the area.

In the area of the Palos Verdes shelf only,
McDermott and Heesen (1974) estimated that

Table 3. — Vertical distribution of DDT in the sediments as
determined from core samples taken at stations 42-20 and
42-36.

Stn. 42-20

Sfn. 42-36

Core

o,p' DDD

o,p' DDT

Aroclor

depth

o.p' DDE

p.p'DDE

p.p'DDD

p,p' DDT

1254

p.p'DDE

(cm)

(ppm.)

0-2

14.6

67.4

10.1

3.1

6.2

0.0233

2-4

19.4

90.7

11.4

3.5

6.0

0.0149

4-6

16.2

84.2

10.1

2.9

4.8

0.0063

6-8

348

64.4

13.2

3.5

6.2

0.00180

8-10

34.0

79.1

8.2

2.8

4.8

0.00065

there were 218 tons of total DDT under 62 km^ of
bottom. They calculated that 85% of the total
DDT was in the top 12 cm of sediment. If the
pesticide is fairly equally distributed in the top 12
cm, about 14% would be in the 10- to 12-cm layer,
and the Shipek sampler would sample about 71%
of the total DDT.

Sixteen contiguous stations on the Palos Ver-
des shelf sampled by us in 1971 represented an
area of 18.1 sq nautical miles (62.0 km^) and a
total DDT load of 139 metric tons. If this was only
71% of the total DDT in the area (the load of
the top 10 cm only), then the corrected esti-
mate including DDT below 10 cm would be 196
metric tons.

McDermott et al. (1974) using a reduced sam-
pling area of 48 km^ determined that there were
156 tons of total DDT in their revised sampling
area. In this present study the area can be ad-
justed to 48 km^ by omitting the effect of ^Vi
peripheral stations. Estimated total DDT then
would be 132 metric tons. However, McDermott
et al. (1974, table 5) give estimates of total DDT
in the area in 2-cm increments down to a depth of
30 cm of sediments. This table indicates that only
about 59% of the total DDT is in the top 10 cm in
this area. This would increase my estimate of
total DDT to 224 metric tons for the 48 km^ area.

The available data indicate that there is con-
siderable variation in the depth distribution of
total DDT in the sediments on the Palos Verdes
shelf However, the general conclusion that can
be drawn from the samples is that there are about
200 metric tons of total DDT in the bottom sedi-
ments in the 14 sq nautical mile area (48 km^) in
the vicinity of the sewer outfalls and another 100
metric tons in the remaining 897 sq nautical
miles of the 1971 survey area.

On 27-28 June 1972, 11 mo after the first sam-
ples were taken, additional samples were ob-
tained from seven of the original stations. Four of
these stations were in deeper water, between 600

MacGREGOR: DDT OFF SOUTHERN CALIFORNIA

and 890 m deep, and 5 to 11 nautical miles from
the sewer outfalls. Total DDT remained low in
these stations averaging about 30 mg/m^ of
bottom, and the composition was essentially
unchanged.

The remaining three stations, in areas of much
higher pollution within 1.3 nautical miles of the
sewer outfalls and in shallower water, showed
some apparent changes in grams per square
meter of bottom (Table 4).

Table 4. — Changes in composition (in grams per square meter
of bottom) at stations 42-21, 43-21, and 42-19 in 11 mo.

Station

Depth

o,p DDE

o.p'DDD

o.pDDT

Total

year

(m)

DDMU

p,p' DDE

p.p'DDD

p,p' DDT

DDT

42-21

1971

119

0.54

3.45

0.53

0.19

4.71

1972

2.09

3.36

0.61

0.23

6.29

43-21

1971

0.46

1.80

0.33

0.14

2.73

1972

0.99

0.80

0.11

0.05

1.95

42-19

1971

0.38

1.78

0.32

0.14

2.62

1972

0.82

0.71

0.16

0.13

1.82

Totals

1971

1.38

7.03

1.18

0.47

10.06

1972

3.90

4.87

0.88

0.41

10.06

At station 42-21, DDT, DDD, andp,p'DDE re-
mained relatively unchanged with a total of
4.2 g/m^ of bottom in both years, while the
o,p'DDE-DDMU peak increased by almost four
times. At the two shallower stations, 43-21 and
42-19, DDT, DDD, and p,p'DDE decreased in
1972 to less than half its value in 1971, while the
o,p'DDE-DDMU peak more than doubled. These
changes could be caused by metabolism, by the
addition of sewage deposits that were relatively
free of DDT combined with metabolism, or even
by the removal of a few centimeters of the
deposits in the shallow-water areas without
metabolism.

CONCLUSIONS

Total DDT in the bottom sediments in the
ocean off southern California in an area of 911 sq
nautical miles was estimated to be between 200
and 300 metric tons. Most of the total DDT
was concentrated in a relatively small area with-

in a few miles of the Los Angeles County sewer
outfalls.

Total DDT in the top 10 cm of sediment ranged
from 6,600 mg/m^ of bottom near the sewer out-
falls to about 1 mg/m^ of bottom at the more dis-
tant stations.

Eighty-two percent of the total DDT was DDE;
10%, DDD; and 8%, DDT. Metabolism of DDT to
DDD and DDE was more rapid in shallow waters
and apparently very slow or lacking in deep,
cold waters that were low in oxygen. Seven
samples taken 11 mo later tended to confirm
these findings.

ACKNOWLEDGMENTS

I am indebted to W. Rommel, G. Boehlert, and
V. McClure for their advice and help in process-
ing the samples; to G. Stauffer for programming
the data for the computer; to R. Lasker for valu-
able criticism and guidance; to the personnel of
the RV David Starr Jordan for their cooperation,
assistance, and interest; and to K. Raymond for
preparing the figures. This work was supported in
part by NOAA, Office of Sea Grant, under grant
#UCSD 2-35208 with the Institute of Marine Re-
sources, University of California.

LITERATURE CITED

MacGregor, J. S.

1974. Changes in the amount and proportions of DDT
and its metaboHtes, DDE and DDD, in the marine en-
vironment off southern California, 1949-72. Fish. Bull.,
U.S. 72:275-293.
McCLURE, V. E.

1972. Precisely deactivated adsorbents applied to the
separation of chlorinated hydrocarbons. J. Chro-
matogr. 70:168-170.
MCDERMOTT, D. J., AND T. C. HEESEN.

1974. Inventory of DDT in sediments. Annual report for
the year ended 30 June 1974. Southern California
Coastal Water Research Project, p. 123-127.
MCDERMOTT, D. J., T. C. HEESEN, AND D. R. YOUNG.

1974. DDT in bottom sediments around five southern
California outfall systems. TM 217. Southern California
Coastal Water Research Project, 54 p.
NATIONAL ACADEMY OF SCIENCES.

1971. Chlorinated hydrocarbons in the marine environ-
ment. Wash., D.C., 42 p.

WooDWELL, G. M., P. P. Craig, and H. A. Johnson.

1971. DDT in the biosphere: Where does it go? Science
(Wash., D.C.) 174:1101-1107.

AN ENERGETICS MODEL FOR THE EXPLOITED YELLOWFIN TUNA,

THUNNUS ALB AC ARES, POPULATION IN
THE EASTERN PACIFIC OCEAN

Gary D. Sharp and Robert C. Francis^

ABSTRACT

An energetics model (ENSIM) for the exploited yellowfin tuna, Thunnus albacares, population in the
eastern Pacific Ocean is developed. Hydrodynamic properties and respiration-swimming work theory
are combined to describe the energy expenditure due to swimming as a function of length for tunas.
Growth and maintenance energetics are estimated and incorporated into a simplistic three process
model. This model is interfaced with a population simulator (TUNP0P) and minimal energy
requirements for the exploited yellowfin tuna population are derived for the simulated fishing years
1964-72. A theoretical unexploited population simulation is made, and the energy requirements by
this population are compared with primary productivity rates and minimum micronekton (forage)
standing stock availability. No obvious food limitation is indicated for yellowfin tunas greater than
40 cm, particularly since the exploited population is at a level of, at most, 50% of the unexploited
biomass estimates. Population limitation processes are examined and indications that the recruit-
ment rates are independent of exploited biomass are discussed.

The intent of studies of the population dynamics
of exploited populations is the determination of
the numbers, biomass, age structure, and poten-
tial yield from a population in order that rational
management decisions can be made about the
manner and rate of exploitation in order to insure
efficient utilization of the resource. The validity
of the resulting estimates of numbers, biomass,
and potential yield is of concern to all those
involved with the resource. Underestimations
generally result in conservative efforts which are
"safe" but not necessarily efficient. Overestima-
tions can result in reduced profit margins or, in
the extreme case, decimation of the resource.

Since the implementation of the program for
conservation of yellowfin tuna, Thunnus alba-
cares, in the eastern tropical Pacific in 1966, a
series of complex changes in the fishery have
occurred which make production model results
less and less comparable between years (Inter-
American Tropical Tuna Commission Annual Re-
ports). Attempts to account for multiple changes
in the effort variables and corresponding but
independent changes in the exploited population
have resulted in serious interpretation problems
as to the relative status of the exploited stock.

'Inter-American Tropical Tuna Commission, c/o Scripps In-
stitution of Oceanography, La Jolla, CA 92037.

The economic and temporal problems inherent
in the collection and analysis of biological data
and the difficulties in representation of the
biological processes in a useful mathematical
manner has served to hinder utilization in the
management procedures of what sparse physi-
ological and ecological information is available.

In this report, an energy budget model is de-
veloped for the exploited yellowfin tuna popula-
tion in the eastern Pacific Ocean within the
Inter- American Tropical Tuna Commission's Yel-
lowfin Regulatory Area (CYRA). The model will
be used to assess the energy flow through the
exploited yellowfin tuna population and also to
compare the estimated utilization of energy by
yellowfin tuna with the estimated primary pro-
ductivity in the CYRA. Comparisons will be
made using simulations of the population under
both exploited and unexploited conditions.

The energy budget estimates are interfaced
with an age dependent population simulation
model (TUNP0P) (Francis 1974) resulting in a
model of the energy utilization by semiannual
recruitment cohorts. This model is referred to as
ENSIM. The model incorporates the population
parameter estimates and variables of TUNP0P
and the empirical and estimated size dependent
relationships for the major energy consuming
processes, resulting in estimates of energy utili-
zation rates. The development of the empirical

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

relationships and the resulting formulations are
presented so as to encourage research in the area
so that improvements on this crude model can be
made in the future.

THE MODEL
Population Dynamics

In an attempt to produce a new, more detailed
method for evaluating the population or stock
status it was decided that the development of
TUNP0P, a biologically oriented population
simulator, would be appropriate. The only avail-
able population data which are collected on a
routine basis from within the fishery are length-
frequency information from commercial catches.
These data are collected according to criteria
which require that the several time-area strata
be sampled regularly and multiply, whenever
possible (Hennemuth 1961). Data from the period
1963-72 have been analyzed and processed in the
following manner.

The 12 existing sampling areas in the CYRA
were reassembled into three major areas: N —
North of lat. 10°N except east of long. 95°W;
5 — North of lat. 5°N to the boundary of area N;
S — all the CYRA south of the boundary of area 5
(see Figure 1). The areas N and S tend to have
separable length-frequency distributions during
any given time interval. Area 5 tends to have
unique components as compared to N and S, but
also contributions from both the other areas can
be observed in the data from area 5. (This phe-
nomenon is tjqjically nonseasonal or noncyclic
with respect to the fishing year and is probably
related to population and environmental pres-
sures within the separate areas.) In all three
areas, recruitment components of a semestral
nature are evidenced. The apparent relative
abundance of these components within the areas
changes seasonally and also between years (Table
1). Analysis of this phenomenon has made the
separation of the semestral cohorts seem the first
logical step when the available genetic, mor-
phometric, and length-frequency data are con-
sidered.

The catch data associated with each length-
frequency sample were obtained. The individual
sample sets were then given relative values pro-
portional to the contributions of the catches (in
weight) from which they were drawn. From this
basic processing of all the length-frequency data,

JO* 1^0° iiO"

00° 90° 80"

"-"tPN

- AU-.

V*k

-ssn?

-4- iiistT*!^

-a-X j: 5^j\

I 1

it: i ± ... S

J—X it -- -I

: i: qii :"

I M

^ ^^ J

*fc^ ;

-»^ 1

TV /

3" V /

- - ;^x j^V /

-- K + ±Js.V^--^ /

- -II ? """"w^S^

S "d>

- - -

--\^-- - .r

iit- --- - - - - " A

- W- t

!i >»

»«- -^ - ^

± — J

X - ^

-1 -1

± „.:..:...r+ -.^^

Figure l. — The study area CYRA (Commission Yellowfin
Regulatory Area) used in the simulations is enclosed in the
dark outline. Three subareas were used in the preliminary
population dynamics work in estimating cohort strength from
the length-frequency and catch and effort data appropriate to
these areas. N = North of lat. 10°N except inside of long.
95°W; 5 = North of lat. 5°N to boundary of N; S = all CYRA
South of boundary of subarea 5.

estimates of the catch composition with respect to
size-age for each fishing area were made and a
growth curve was determined for each of two
semestral cohorts. The two curves were essen-
tially identical and warrant no further discussion
here other than to say that from 40 to 145 cm fork
length it is possible to give relative monthly ages
to all individuals, given a length and correspond-
ing date of capture. The labeling problem was
handled such that any fish that was 40 cm from 1
January to 30 June is labeled S^ and correspond-
ingly 40-cm recruits from 1 July to 31 December
are labeled 83. The cohorts are identified in
relation to their recruitment year when they are
40 cm, not their spawned year. For example, a
40-cm fish caught in February 1969 is attributed
to the cohort labeled S^, 1969; and a 40-cm fish
caught in October 1968 is attributed to the
semester cohort labeled Sg, 1968. The two
semestral groups can be treated as independent
units in the population and provide a biological
basis in assessment of population size with re-
spect to size-age classes within the fishing year.
The annual growth increment in the most often
encountered cohort classes (40-140 cm) in the
fishery appears to be about 32 cm/yr; therefore,

FISHERY BULLETIN: VOL. 74, NO. 1

Table l. — For the years 1964-71 the data are presented for the catch in short tons by semestral cohort in the three areas (N, 5, S)
within the CYRA. Also given are the percent of the total catch (Sp^ + Sg + Big) by cohort within the areas. The category, Big,
represents the fish of length / greater than 145 cm which we feel are not ageable under the present system. The percent of the
individual semestral cohorts (S^ or Sg) caught in the three areas is also given. Note the erratic shifting of the cohort dominance
(S^ or Sg) in the catch as well as the distribution of the cohorts between areas.

Year

North
A

5
A

South
A

Total
A

North
B

5
B

South
B

Total
B

Total
A + B

Big

1964
% total A + B
% total A or B

1965

1966

1967

1968

1969

1970

1971

27,452

9,401

5,209

42,062

33,561

5,881

17,515

56,957

26.9

9.2

5.1

41.2

32.9

5.8

17.2

55.9

65.3

22.4

12.4

58.9

10.3

30.8

18,967

13,512

6,406

38,885

24,064

14,164

8,386

46,614

21.1

15.0

7.1

43.2

26.7

15.7

9.3

51.8

48.8

34.7

16.5

51.6

30.4

18.0

7,769

23,128

20,176

51,073

10,292

11,394

14,771

36,457

8.5

25.4

22.1

56.0

11.3

12.5

16.2

40.0

15.2

45.3

39.5

28.2

31.3

40.5

20,699

9,564

7,664

37,927

29,482

8,572

11,867

49,921

23.1

10.7

8.5

42.3

32.9

9.6

13.2

55.7

54.6

25.2

20.2

59.1

17.2

23.8

16,361

23,921

13,552

53,834

33,917

22,132

3,128

59,177

14.3

20.9

11.8

47.0

29.6

19.3

2.7

51.6

30.4

44.4

25.2

57.3

37.4

5.3

22,437

20,034

9,030

51,501

34,887

29,587

5,648

70,122

17.7

15.8

7.1

40.7

27.6

23.4

4.5

55.4

43.6

38.9

17.5

49.8

42.2

8.1

39,197

15,942

10,529

65,668

43,476

13,257

11,125

67,858

27.5

11.2

7.3

46.0

30.5

9.3

7.8

47.6

59.7

24.3

16.0

64.1

19.5

16.4

12,372

18,719

14,453

45,544

17,357

25,283

15,712

58,352

10.9

16.5

12.8

40.2

15.3

22.3

13.9

51.6

27.2

41.1

31.7

29.7

43.3

26.9

99,019

85,499

87,530

87,848

113,011

121,623

133,526

103,896

2,921
2.9

4,543
5.0

3,626
4.0

1,802
2.0

1,602

1.4

4,888
3.9

9,176
6.4

9,277
8.2

the mean lengths and modes of the two semes-
teral cohorts are separated by approximately 16
cm (Tomlinson and Sharp work in progress). A
significant number of animals may shift from the
leading edge of one labeled distribution into the
trailing edge of the other, but we are assuming
that countershifts are equally as probable and
both are irreversible. An effect of shortening the
sampling "season," since the implementation of
regulations, has been to distort the apparent
abundance of the two groups and merge the
modal distributions into a single amorphous dis-
tribution (Figure 2).

The cohorts are treated independently by the
model. Each cohort is considered to have a unique
effect in the analysis of the net biomass and
numbers estimates for a given fishing year. Dif-
ferential exploitation of these cohorts can be
determined from the catch-effort length-
frequency data and as such warrants this disin-
tegration technique as opposed to treating the
year class as a single unit. We have, however,
decided not to present in this report the area
breakdown results in the simulations. When the
cohorts are separated, it is possible to construct a
catch table for each from the length-frequency
sample data from the fishery. With this catch
table and the catch data (yield) it is possible to
determine the relative mortality (F) attributable

to fishing, by assuming a constant natural mor-
tality (M), a necessary, but perhaps poor assump-
tion in the case of tunas due to the inherent rapid
changes in ecological status as they grow. The
Murphy cohort analysis procedure (Murphy 1965;
Tomlinson 1970) was used for estimation of re-
cruitment at first availability to the fishery (A'40).
Using this approach we have generated the un-
derlying population structure for the historical
series we wish to represent.

Energetics

The energetics parameters for free-swimming
predatory species such as the tunas must be
size-related functions due to the broad range of
sizes commonly encountered in the fishery; 1.3 kg
to greater than 62 kg, or 40 cm to greater than
145 cm. In no case for fish has anyone measured
physiological parameters from such a range of
sizes.

Magnuson (1973) discussed the effect of gas
bladders and lift surfaces on the velocity of ob-
ligatory swimmers such as the tunas. He deter-
mined the relationships between size and mini-
mum velocity for maintenance of hydrostatic
equilibrium for several scombrid species, includ-
ing skipjack tuna, Katsuwonus pelamis, and
Thunnus albacares. This work has provided a

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

RE
(AS OF

20 30 40 50

LATIVE MONTHS OF AGE
JANUARY OF INDICATED YEAR)

basis for determining some of the relative energy
expenditures in the population simulation study.

The energy utilization which is simulated in
ENSIM is that attributable to 1) maintenance of
metabolic stasis, 2) growth, and 3) swimming.
Each is calculated independently and summed
with the others to give an estimate of the total
minimum energy utilized on a daily basis. No
attempt has been made to evaluate energy ex-
penditures due to gonad maturation or migratory
behavior beyond the daily forage or flight be-
havior levels because of the inherent void in our
knowledge of these processes in tunas.

Metabolic maintenance of stasis energy re-
quirements (£',„) are difficult to assess under
optimum conditions and are typically derived
from extrapolation of O2 consumption versus
activity relationships to a zero activity level. The
magnitude varies greatly between species and in
general is a tenuous function of size and physio-
logical state. It is essentially impossible to di-
rectly measure the stasis energy requirements
of tunas due to their continuous swimming be-
havior. Estimates of E,,, should not include the
energy expenditures due to even minimum swim-
ming activity if it is to be useful in the deter-
mination of energy expenditures due solely to
swimming work.

The respiration rate attributable to tissue stasis
can be estimated from the metabolic weight
(Wn^gt) of fish of length / from the equation:

£„, = 24 ^ W^et (modified from Winberg 1960)

where W^^^ = iMf)''-^

and

Mf = 1.858 X 10-2 (/)3.o2i (grams) (Chatwin 1959)

and where k is estimated to equal 1 cal/g h from
data and estimates for other highly active fishes
(Fry 1957; Winberg 1960). Therefore

E^ = 4.46 X 10-1 (/)302i cal/day.

Figure 2.— The numbers offish caught in the fishing years 1966,
1968, 1970, as a function of their recruitment month, and age,
relative to the fish of the year are graphically represented. Se-
mestral (A, B) and annual cohort labeling is as indicated. Note the
central tendency of the peaks within the semestral limits in the
years 1966 and 1968. In these years the fishing "season" was quite
long (>6 mo) as compared to 1970 (<3 mo), which combined with
cyclic migratory behavior and subsequent availability of cohorts
probably results in the drastic change from multimodality to the
amorphous distribution seen in the 1970 data.

FISHERY BULLETIN: VOL. 74, NO. 1

When estimates of the true stasis energy rela-
tions are finally available, they can be easily
incorporated into the model.

Probably the most difficult process to define,
estimate, and measure is that of growth. The
energy requisite to growth (Eg) can be esti-
mated minimally as the biomass gain per time
period as converted to calories. This is a highly
unsatisfactory method because of the many
energy requiring steps between ingestion of a
food organism and the consequential deposition
of the materials assimilated into the living bio-
mass of the growing organism (Phillips 1969).

One slight change in the accepted method-
ology of bioenergetic accounting which we will
make is in our definition of specific dynamic
action (SDA). If one is willing to accept that the
SDA contributed little other than heat to the feed-
ing organism, then it can be defined as the loss
of energy due to the inefficiency of the digestive
processes, including cost of transport, deamina-
tion, biosynthesis, and related processes. The
rate of inefficiency (percent of SDA energy with
respect to total ingested energy) is variable in
most animals studied as a function of feeding
level (Warren and Davis 1967) and environmental
conditions (Warren 1971). In our definition of
SDA we do not include the unavailable portion
of foodstuffs.

For our purposes we will assume that growth
of yellowfin tuna in the CYRA is relatively con-
tinuous with respect to season or environmental
state. There are several assumptions involved in
this basic tenet which require some discussion.
Tunas are highly endothermic animals, and
Carey and Teal (1966) have shown the presence
of a relatively high efficiency heat exchange
(conservation) mechanism in tunas. This sug-
gests that tunas are likely to be somewhat inde-
pendent of ambient temperatures in that the
temperature variability encountered within the
core of these fishes is likely less than the ambient
variability. Their large mass (>1 kg) would con-
tribute to thermal stability over a wide ambient
change (Neill and Stevens 1974).

Observations of temperature dependent activ-
ity indicate a lower activity as temperature de-
creases in small yellowfin tuna (<50 cm, <2.5 kg)
at a Qio of near 2 (Neill, pers. commun.). This size
of yellowfin tuna is rarely encountered in the
CYRA at temperatures below 23°C and is found
aggregated on the warm side of the north-south
surface temperature cline including this tempera-

ture, indicating some preference for tempera-
tures near 23°C. Preliminary studies of effects of
the environmental characteristics on the abun-
dance and availability of 40- to 70-cm yellowfin
tuna in the CYRA indicate a direct relationship
between the 23°C isotherm depth of the av-
erage number of fish per school, and the overall
availability of these fish to surface fishing gear
(Inter-American Tropical Tuna Commission
1975).

All this is emphasized to indicate the limited
range of temperatures likely to be affecting the
metabolic rates of yellowfin tuna as compared
to that affecting smaller species without the
complex stabilization mechanisms (heat ex-
changers, etc.) as is the typical situation in fishes.

The relative activity, mobility, and distribution
with respect to temperature of yellowfin tuna
can be used as supportive background for as-
suming a relatively stable growth energy avail-
ability as they developed, bringing us to the con-
clusion that a first approximation of the SDA
can be made with respect to the energy equiva-
lent to the biomass change on a daily basis.
From studies discussed by Paloheimo and Dickie
(1966) and Warren and Davis (1967) on several
species and estimates by Kitchel et al.^ for K.
pelamis, it appears that SDA probably accounts
for 30-40% of the total consumed calories which
could be part of the growth process. We have,
therefore, assumed that Eg is going to equal the
equivalent caloric value of the tissues plus the
SDA which will be given by the relation

SDA

(Biomass change in grams per day)

where, if 1 g is calorically equivalent to 1.46 kcal
(Kitchell et al. see footnote 2) then

3
Eg = — Biomass change (grams) (1.46 kcal/g)

= 2,190 kcal /kg growth.

Smit (1965) has provided the mathematical
basis for our determinations of energy output
and caloric requirements due to swimming. He
shows that:

(Meg S) (143 X 103) gcm2 (1)

Power

3,600

^Kitchell, J. F., W. H. Neill, and J. J. Magnuson. Bio-
energetics of skipjack tuna, Katsuwonus pelamis. Manuscr.

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

where Mg is the efficiency of the muscle tissue
when converting chemical energy to mechanical
work; S is the respiration due to activity in
mg 02/h; and^ is the acceleration due to gravity
(981 cm s'2). The propulsion efficiency is as-
sumed to be 0.90 (Lighthill 1970) and is included
in the resulting muscle efficiency figure.

For our purposes we assume M^ to be 0.18.
Therefore from Equation (1)

(Power) (3,600 s/h)

S =

(0.18) (143 X 103 g cm) (981 cm/s^)

mg Oa/h.

(lA)
From the hydrodynamics theory (Streeter 1962)

Power =-g-Ay3Crf^^

where p = the density of seawater (1.025 g/cm^)
A= 0.4(Z)2 from Bainbridge (1961) (cm^)
V'= is derived from Magnuson's empirical
relationships between / and species
velocity V (cm/s)
Cd = the coefficient of total drag of the fish,
which is derived from an empirical re-
lation including the results of studies
by Pyatetskiy (1971).

We can therefore rewrite the equation so that
respiration due to swimming is equal to

pAV^Cd
' 2 (7,017.66)

2.59 X 10-5 (/)2 (y)3 Crfmg Oz/h.

(2)

We now have an Equation (2) of three elements
for which we have solutions for two (V and Cd)
as functions of the third (I) given below.

V Determination

From Magnuson (1970), the relation for the
minimum velocity (Vioo) for sustained hydro-
static equilibrium by tunas is given as

100

1/2

(3)

where Cj^ = the coefficient of lift for the pectoral
fins
Af^ = the total lifting area of the pectoral
fins (cm2), log Af^ = -1.2154 +
1.87 log J

Ci = the coefficient of lift of the keel

A^ = the lifting area of the keel (cm^),
logA* = -2.7033 + 2.26 log /(cm2)

Lf = the total weight of the fish in sea-
water (dynes). (L, values are ob-
tained by multiplying Mf values
by appropriate constants as pro-
vided by Magnuson (1973) by
species and weight class.)

Mf= mass of the fish = 1.858 x lO'^
(/)3.02i (grams).

Determination of the Coefficient of
Total Drag C^

The relation between the total drag coefficient

iCd) and the Reynolds number (Re) for Atlantic

bonito, Sarda sarda , reported by Pyatetskiy (1971)

is taken to be representative in form for scombri-

/ V
form fishes. Re = , where v is the kinematic

viscosity of seawater or 0.01 cm^/s; / is the fish
fork length in centimeters; and V is the fish
velocity in centimeters per second.

An analytical expression was derived for esti-
mating the Cd values in the following manner:
R. Gooding (Gooding et al. 1973) of the National
Marine Fisheries Service Honolulu Laboratory,
Honolulu, Hawaii reported respiration rates for
unfed K. pelamis from 32 to 36 cm fork length,
swimming at or near minimum velocities (Vioo)-
From these data it was possible to calculate Cd
given the observed respiration rate (St<,tai) was
431.5 mg 02/kg h and / = 35 cm. The minimum
velocity (Vioo) = 59.1 cm/s and Re = 2.07 x 10^
at this velocity.

For skipjack tuna of Z = 35 cm, W^et = 200.5
g, so that

Sm = 60.0 mg Oa/h
Stotai - S^ = S, = 371.5 mg Oa/h.

From Equation (2) it is now possible to deter-
mine that

Cd-

371.5

2.59 X 10-5 (35)2 (59.1)3

= 0.057.

This value of Cd was related to the values
graphically displayed by Pyatetskiy (1971) and
what was assumed to be a good approximation

FISHERY BULLETIN: VOL. 74, NO. 1

of the total drag on the test animals was derived
relative to his graphed observations as a function
of Re. From Re, one can determine the approxi-
mate coefficient of total drag (C^) from the
relation:

Cd= 0.262 e -4-805 X 10- i?.

(4)

Gooding also reported respiration data for
skipjack, ranging from 45 to 53 cm, swimming at
or near Vioo where S^otai = 1.403 mg Oa/h. These
test animals had also been deprived of food for
24 h. Assuming / = 50 cm:

Wmet= 523.5 g; Vioo = 70.5 cm/s;
Re = 3.525 x 10^
S„ = 156 mg Oa/h;

Cd = 0.262 e-^-^^^ " "^'^ '^-^^^ ^ '°' = 0.048.
/.S, = 2.59 X 10-5 (50)2(70.5)3(0.048)
= 1,233 mg Oa/h.
Ss+Sn, -S,,,,i = {1,233 + 156} mg O2

= 1,389 mg Oa/h, (expected)
where S total = 1,403 mg 02/h, (observed)
leaving 14 mg 02/h, (difference).

The Relation (4) we have used for Q as a function
of Re appears to be adequate for our purposes.

Within the factors M^ and C^ there are an in-
separable pair of modifying effects which must
be accounted for, but which are essentially in-
determinate at the present state of the art. One
is the mechanical propulsion efficiency, and the
other is the effect of the short-term flux of the
rates of acceleration due to caudal fin position
and velocity wdthin a single tail beat cycle on
the "average" calculations of M^ and C^. The Me
and Q values are continuous variables within
the tail beat cycle and are inextricably bound
together. Where in the integration and estima-
tion of these two values the trade off is made is
inconsequential due to the equal and direct
effect of the estimate of one on the other value.
Until either value is measured and fixed, the
other coefficient is relative and therefore not
necessarily realistic.

The effect of velocity on propulsion efficiency
is probably great in tunas (and other large
organisms) due to several processes, including
local heating phenomena and subsequent con-
traction rate increases of the muscle fibers

(Walters 1962; Sharp and Vlymen^). The graded
increase in utilization of white muscle fibers as
velocity is increased should result in generalized
heating and increased overall efficiency of the
energy conversion processes in the muscles. This
and other effects may indeed account for the con-
siderable efficiency changes in work done as com-
pared to respiration rate when extended periods of
white muscle utilization are monitored (Kutty
1968).

The higher scombrids {Auxis, Euthynnus,
Katsuwonus, and Thunnus) have incorporated,
in various designs, a subcutaneous vascular
system which is the distribution mechanism for
transport of arterial and venous blood to and
from the warm swimming musculature (Kishi-
nouye 1923). The direct transport of "warm"
venous blood to the fish's surface probably
affects the hydrodynamics of the fish and con-
tributes to the dynamic flux of the Cd value. Since
no data are available for these phenomena, they
have to be ignored in this treatment of the swim-
ming energetics, but future laboratory studies
should not ignore or delete these potential
effectors.

Considering the range of possible error in
estimating both muscle efficiency and/or the co-
efficient of total drag, the close agreement be-
tween observed and expected respiration rates
indicates that we have useful estimates of energy
requirements.

The only available respiration-activity data
from tunas is for K. pelamis. Assuming that
Magnuson's (1973) empirical relations and
density multipliers are representative of the
relative hydrodynamic status of the several
species, these relations should give a similarly
good approximation of energy consuming proc-
esses in T. albacares as they appear to give for
K. pelamis.

The three continuous energy consuming pro-
cesses are, therefore, roughly accountable using
the previously described relations. The conver-
sion of oxygen consumption to caloric utiliza-
tion is made on the basis that 3.359 cal are avail-
able from 1 mg O2. Apparently the major energy
consumption process is swimming, including
feeding and flight behavior. The_energy ex-
pended is a function of the velocity Vjyp which is

^Sharp, G. D., and W. J. Vlymen III. The relation between
heat generation, conservation and the swimming energetics
of tunas. Manuscr.

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

in turn a function of the length of the individuals
(see Figure 3). In Magnuson's (1973) relation-
ships the variables necessary for a solution for
the minimum velocity are I and the density of the
fish. Magnuson (1973) provided data for fish
density (in the form of empirically derived multi-
pliers) by weight class for several species in-
cluding yellowfin tuna. We have extrapolated
his data to fit our size distribution with an asymp-
totic lower limit of fish density at 1.06 g/cm^
reached by 120-cm fish.

We are assuming that the animals have their
pectoral fins 75% extended all of the time that
they are in nonfeeding-flight behavior, hence
Ci = 0.75, and that the keel surface is 85% ef-
fective so that Cj , = 0.85. This results in a fish

that is swimming somewhat faster on the average
than its Vjoo or minimum velocity. These values
are "best guess" estimates and as such, repre-
sent only minor changes in the appropriate di-
rection as opposed to using absolute minimum
energy utilization in the population simulation.
Magnuson's Vioo for a 50-cm yellowfin tuna is
50.91 cm/s. Solving for the "typical" velocity under
our "best guess" conditions results in a V^y^ of
58.29 cm/s.

We have set a "typical" feeding-flight speed
at 3 m/s. This is an integrated average that in-
cludes all velocities above V^^^ and includes the
burst speed forays. Since the energy required
for different speeds is proportional to a cubic
function of the velocities, it should be noted
that the most probable velocity is less than 2 m/s,
since the energy requirements for a few short
bursts of up to 10 body lengths/s rapidly increase
the overall energy utilization. With this in mind.

FORK LENGTH ICMI

Figure 3. — The energy utilization (in kcal/day) for growth
(Eg), maintenance (£„,), and the total (Eg + E„ + E^ = ^total*
energy utilization are portrayed as functions of length 7.

we have at^ibuted 95% of the day or 22.8 h of
the day to V^yp requirements and 5% or 1.2 h to
Vfeed behavior. This is not to say that the fish are
limited to 1.2 h/day of feeding but that on the
average the increased velocity due to external
stimuli are exhibited for this period. One sus-
pects that the feeding of large and small tuna is
entirely different in nature, but for simplicity and
since no data are available, it is not unreason-
able to assume that the relative effectiveness of
feeding is somewhat similar over the life history
of the animals. Based on these estimates we
hope to have contrived a "reasonable" fiction for
use in our model. The need for better estimates
is obvious.

MODELING RESULTS

The model ENSIM computes the caloric re-
quirement of each semestral cohort in the ex-
ploited population, by quarter of the fishing year.
Summary data are listed after each quarterly out-
put which differentiate the semester A cohort
caloric expenditure from that of the semester B
cohort, and a composite total expenditure is
listed (see Table 2). An annual summary for 1972
is also generated and an example is presented
in Table 3.

Initial biomass and numbers, yield in weight
and numbers, gross growth, and average bio-
mass are tabulated for each quarter, and sum-
mary tables are generated for the individual
semestral cohorts as well as composite values.
The biomass of food ingested per day is gen-
erated for each cohort, assuming 1.00 kcal
(Paloheimo and Dickie 1966) are available per
gram food ingested. The minimum percent bio-
mass ingested per day with respect to the cohort
biomass is also calculated for each cohort (see
Figure 4). The caloric requirements for mainte-
nance, swimming (at V^yp, Vfeed); ^^^ growth are
tabulated by size of the average animal in each
cohort in the simulation by quarter (see Table 4).

We have simulated the fishing years 1964-72
and included the best available estimates for
cohort strength, fishing effort, and availability
parameters. We have also simulated a nonex-
ploited population which was recruited at the
average level for the data from the last 5 yr which
includes all the population indicated or expected
from inside our study area (see Figure 5). From
Figure 5, the plot of the average annual biomass
estimate, one can readily see the effect of fishery

FISHERY BULLETIN: VOL. 74, NO. 1

Table 2.— ENSIM output for quarter three of the 1972 simulation is presented. The calculated kilocalories expended by each
cohort (age-class) in the exploited population is given. The appropriate averages (N/-, weight (kg) and I) are also listed for each
cohort. Summary data are given by cohort and for both cohorts summed together.

Age

Maintenance

Swimming V^p

Swimming V^^^^

^ total

N-l

Weight (kg)

.634079E+11

.640848E+11

.280680E+12

.377884E+11

.445961 E+ 12

.186707E + 08

.176060E+01

.444182E+02

.889373E+11

.776321E+11

.354145E+12

.431443E+11

.563858E+12

.141776E + 08

.379126E + 01

.572565E + 02

.501222E+11

.307991 E + 11

.185411E+12

.264195E+11

.292751E+12

.521301E + 07

.646561E+01

.683217E + 02

.463850E+11

.196178E+11

.157393E+12

.237458E+11

.247142E+12

.292458E + 07

.120872E + 02

.840421 E+02

.352487E+11

.120881E+11

.111902E+12

.138239E+11

.173063E + 12

.151084E + 07

.195812E + 02

.985929E + 02

.176499E+11

.514907E+10

.528273E+11

.601945E+10

.816457E+11

.537688E + 06

.300053E + 02

.113553E + 03

.928307E+10

.302828E+10

.263712E+11

.297797E+10

.416605E+11

.208937E+06

.438055E + 02

128704E + 03

.465814E + 09

.166772E + 09

.128095E+10

.796327E + 08

.199317E+10

.868343E + 04

.55441 4E+ 02

.139141E + 03

.695863E + 09

.256764E + 09

.189348E + 10

.249005E + 08

.287100E+10

.122020E + 05

.598477E + 02

.142709E + 03

Total A

.153438E+12

.102566E+12

.565646E + 12

.729892E+11

.894639E+12

Total B

.158758E+12

.110257E+12

.606257E+12

.810347E+11

.956307E+12

Total

.312196E+12

.212823E+12

.117190E+13

.154024E+12

.185095E+13

Table 3. — The 1972 annual summary data are hsted which
give the yield in number and weight for each of the semestral
cohorts as well as the kilocalories utilized in the year by the
cohorts and the combined sum.

Yield weight Kilocalories

Yield numbers (metric ton) utilized

Total S^
Total S
Total S/i

0.491282E + 7
0,534418 E + 7
0. 1 02570 E + 8

0.653748 E + 5
0,640257 E + 5
0.129427 E + 6

3.85 E + 12
2.28 E + 12
7.13 E + 12

;,.o

1 1 1 1 1 1 1 1 1

2 w

^^^-^^^

- so

^"^^^^,_^

— ■ _______^

1 M

~ — -

^•

FORK LENGTH (CM )

Figure 4. — The amount of food required per day is given in
percent body weight of the individual yellowfin tuna of length /.

Table 4. — The estimates of the daily energy utilization (in kcal/day) for maintenance, swimming at V^yp and Vfggj, growth, and
the total daily energy utilized due to all these activities is provided for the average individual of length L and weight W for each
cohort in the population during each quarterly time period. The average number of individuals present in each cohort is given in
the column headed N. The semestral cohorts are separated (Total A or Total B) and the energy utilization estimates summed and
listed for each. The composite estimates (S^ + Sg) are also listed (Total).

Age

Maintenance

Swimming Vjyp

Swimming V(gg|j

^total

.377346E + 02

.381374E + 02

.167035E + 03

.224882E + 02

.265395E + 03

.192145E + 08

.176060E + 01

.444182E + 02

.667851E + 02

.604781 E + 02

.267903E + 03

.225852E + 02

.417752E-^03

.125166E + 08

.359406E + 01

.562530E + 02

.116280E + 03

.663364E + 02

.423896E + 03

.451705E^02

.651683E + 03

.863385E + 07

.718813E+01

.707599E + 02

.181619E + 03

.751903E + 02

.613072E + 03

.902863E + 02

.960167E + 03

. 562388 E + 07

.125513E + 02

.850967E + 02

.271052E + 03

.911480E + 02

.853897E + 03

,112988E + 03

.132909E + 04

.387916E + 07

.207038E + 02

,100429E + 03

.381489E + 03

.112386E + 03

,113300E + 04

.169496E + 03

.179638E + 04

.252689E+07

.317387E+02

.115684E + 03

.497444E + 03

.162886E + 03

,141127E + 04

,180946E + 03

.225255E + 04

.174290E + 07

.442248E + 02

.129111E + 03

.596227E + 03

.213496E + 03

.163949E + 04

.101935E + 03

.255115E + 04

.113565E + 07

.554626E + 02

,139159E + 03

.639329E + 03

.236943E + 03

.173697E + 04

.340073E + 02

.264725E + 04

.783083E + 06

.605190E + 02

.143236E + 03

Total A

.156184E + 04

.595451E + 03

.459307E + 04

.395600E + 03

.714596E + 04

Total B

,122612E + 04

.461550E + 03

.365347E + 04

.384302E + 03

.572544E + 04

Total

.278796E + 04

.105700E + 04

.824654E+04

.779903E + 03

.128714E + 05

SHARP and FRANCIS; ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

Figure 5. — The average biomass estimate of the exploited
yellowfin tuna population in the CYRA is shown. The historical
fishery label indicates the coastal fishery which operated prior
to 1965; the expanded fishery indicates the process of seaward
areal expansion which dramatically changed the estimates
of exploited biomass from 1966 until approximately 1968.
Fishery regulation was implemented in September 1966. The
simulation of the unexploited populations yielded estimates
of the average biomass for the two cohorts to be S;^ = 282,400
metric tons; Sg = 272,700 metric tons; S^ + Sg = 555,100
metric tons. Recruitment was assumed to be consistent with
recent levels.

NONREGULATED FISHERY

REGULATED FISHERY

HISTORICfiL FISHERY ^

EXPftNOEO FISHERY

G \ / '

/ 0

^O^

• SEMESTER A COHORTS

' '

X SEMESTER B COHORTS

O TOTAL [SEMESTER A +
SEMESTER B)

1964 1965 1966 1967 1968 1969 1970 1971 1972

Figure 6. — The catch in metric tons of yellowfin tuna from
the CYRA is shown for the study period. The cohorts and tottd
catch are indicated by symbols as in Figure 5.

-• SEMESTER fl COMOfiTS
-X SEMESTER B COHORTS
-o TOTAL (SEMESTER A * SEMESTER B)

FIGURE 7. — Estimates from ENSIM of the kilocalories used per
year by yellowfin tuna in the exploited CYRA population for
the 1964-72 period.

growth (areal expansion) on population size esti-
mates. From Figures 6 and 7 it is obvious that the
catch has great fluctuations (e.g., 1971) but the
energy flow seems to have stabilized in the ex-
ploited population estimates. This may be artifac-
tual but we think it may be significant to attempt
interpretation.

The ratio of yield in weight to gross growth is
another interesting indicator (Figure 8). Note the
differential rate of exploitation of the semestral
cohorts through time prior to 1967. The S^ and
Sg cohorts became approximately equally ex-
ploited in this respect about 1967 or at about the
end of the changes in fishery strategy and when

Rtg

loted

F<she

H>iloricol Fishery

X--

X Semes'*' B coho"s

-^.

Semes'er B )

"^^ ^^^q\

A 1

X Z®

\^^

1967 (968

YEAR

Figure 8. — The ratio of the yield in weight (catch) to gross
growth for the years 1964-72. Note the relative similarity of the
levels of the cohorts respective ratios in the regulated years as
compared to the preregulated years.

FISHERY BULLETIN: VOL. 74, NO. 1

regulation occurred. The indication is that since
approximately 1969, the biomass and exploita-
tion levels on the semestral cohorts have some-
how paralleled a somewhat uniform energy utili-
zation by the two cohorts, whereas from 1966
until 1969 a larger semester A biomass was
under exploitation compared to the semester
B cohort. The large discrepancies in biomass
caught as compared to gross growth in the early
data (1964-65) compared to the recent data
(1969-72) may be an indicator of the relative
health of the stocks under exploitation in recent
years in contrast to the preregulatory years.

SPECULATIONS

The utility of simulation studies lies in the
process of linking together observations, using
generalized principles where possible, to gen-
erate testable hypotheses which ultimately lead
to resolution of cause and effect relationships.
As examples, from the results of the simulation
model ENSIM, hypotheses were conceived con-
cerning the relative importance of forage or-
ganisms, primary productivity and the size of the
animals with respect to recruitment limitations.

Food as a Population Regulator

The availability of food is classically attributed
the role of limiting population size. We do not
intend to assail this premise, but intend only to
show that the most probable source of limitations
is at very early ages in tunas (<40 cm), and not
on the late juvenile or adult population.

Forage for tunas is generally considered to be
in the micronekton size range (1-10 cm). It
probably extends upwards to 30 cm or more in
length for larger sizes of tunas (Magnuson and
Heitz 1971; Perrin et al. 1973). Tunas eat largely
crustaceans, fishes, and cephalopods in most
regions (Alverson 1963; Magnuson and Heitz
1971; Perrin et al. 1973). These organisms are
poorly sampled by micronekton sampling devices.

The EASTROPAC cruises sampled from our
study area over the year 1967 and early 1968.
Productivity, micronekton, and most physical
and chemical properties which are linked to
biological productivity were sampled. EASTRO-
PAC data (Blackburn et al. 1970) indicate that
the average minimum micronekton night haul
contained 5 ml of micronekton per 10^ m^ of

water sampled. The samples represent a 200-m
water column.

The surface area of the CYRA is estimated to
be 5,012,643 sq nautical miles or 1.696 x lO^^ m2.
The minimum available forage is therefore

(1.696 X 1013 m2) (200 m)

5 ml forage

103 m^

)

= 1.696 X 10^3 cc.

If 1 cm^ forage has approximately 1 g or 1.25
kcal caloric equivalency, then one should expect
that there is a minimum forage availability of 1.25
kcal/m^ or assuming 80% utilization efficiency
of these calories by predators (Winberg 1960),
1.0 kcal/m^ are present for metabolic utilization.

Owen and Zeitzschell (1970) in their analysis
of EASTROPAC data also show that the primary
productivity averages 169 mg carbon m'^ day^
over long. 119°- 112° W, 219 mg carbon mr^ day^
at long. 105°W, and 282 mg carbon m'^ day^
along long. 98°W. They also indicate coastal
effects as being the probable cause of the east-
ward increase in productivity. The average pro-
ductivity over the entire study area was 205 mg
carbon m'^ day^.

The energetic equivalent value for 1 mg carbon
fixation is 11.4 cal (Piatt and Erwin 1973), so that
the average caloric productivity is 2,340 cal/m^
day (or 2.34 kcal/m^ day).

We have seen that the minimum estimate of
the micronekton standing stocks caloric value is
1,250 cal/m^, indicating that the probable daily
turnover rate is less than 125 cal/m^ so that
maintenance of this stock is not unreasonable if
the primary production is 2,340 cal/m^ day.

The yellowfin tuna population simulation pro-
cedure based on average Murphy recruitment
estimates of the 1966-71 S^ and Sg cohorts indi-
cates that an unfished population (exhibiting a
stable age structure) would have the biomass of
600,000 metric tons (6.0 x lO^^ g). Assuming
that the yellowrfin tuna (YF) are distributed pro-
portionally over the forage:

6.0 X 1011 g YF

1.696 X 1013 ni

= 3.54 X 10-2 g YF/m2

= 35.4 mg YF/m2;

35.4 mg YF/m2 x 1.2 cal/mg YF = 42.5 cal/m^.

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

Assuming the average caloric consumption by
the yellowfin tuna population per day to be 10%
of its caloric biomass, a somewhat higher than
realistic estimate, daily utilization in calories
would be 4.25 cal/m^ day. The results of the
ENSIM estimates of the total calories utilized per
year for the unexploited population was 14.96 x
10^^ cal/annum, so that the resulting utilization
per square meter day is given by:

14.96 X 10^5 cal/annum _ 2.5 cal

(365 day /annum) (1.696 x lO^^ m2) ~ m^ day

The results of the simulations of the exploited
fishery for the years 1964-72 yield estimates of
less than 50% of this figure as the energy utili-
zation by the yellowfin tuna population. One
would expect the true values of caloric utiliza-
tion to lie somewhere in the range from approxi-
mately 1.5 cal/m^ day to the upper value of 4.25
cal/m^ day.

With the primary productivity estimated to be
at an average level of 2.34 kcal/m^ day and forage
standing stock utilizable caloric values averaging
at a minimum of 1.00 kcal/m^, it seems hardly
likely that yellowfin tuna are food limited from the
40-cm recruitment size.

This brings up the problem of how the east-
ern tropical Pacific yellowfin tuna population
is limited. This, of course, is best taken in per-
spective. Population limitation examples are
typically taken from terrestrial populations and
extrapolations made to ecosimilar strategies in
closed systems such as lakes and estuaries
where primary productivity is greatly affected
by season, and indeed can be determined to
be the limiting factor in population numbers
and biomass.

In those marine animals where density de-
pendent growth functions are evidenced there is
generally a two-dimensional limitation imposed
such that crowding is likely to affect each indi-
vidual. For filter-feeding organisms, such as
herring and menhaden, the density dependent
function is easily conceptualized.

One needs only to examine the relative abun-
dance of food available to highly mobile preda-
tory species which feed opportunistically on
organisms ranging in size from 1 to 30 cm, which
are available on a relatively continuous basis
in a tropical system, to see that dogma general
to terrestrial, estuarine, limnetic, two-dimen-
sional substrate tied, or filter-feeding animal

ecology does not generally apply to the 40- to
140-cm yellowfin tuna.

There are, however, several possibilities con-
cerning the survival of yellowfin tuna from larvae
to 40 cm which would certainly fit into the
schemes which typically limit species. Since they
are probably particulate feeders (e.g., do not
undergo ecometamorphoses at early ages from
filter feeders to predators), it can easily be seen
that they are victims of the availability of con-
centrations of food at smaller sizes because of
their relative lack of mobility. If a 40-cm tuna
requires 10-20% of its body weight per day to
maintain, as compared to 3-5% in large yellowfin
tuna, then one can hypothesize that the smaller
predators must consume even greater amounts
due to the pressures of very rapid growth, feed-
ing activity, and competition with peers, indicat-
ing that they are more likely severely affected
by density of both conspecifics and food than are
the larger sized fish.

Another consideration is the size distribution
of the forage organisms. It is obvious that there
are considerably larger amounts of the smaller
food organisms than the bigger sizes, which
would perhaps indicate that the real density
competition pressures are on the intermediate
sizes (vis. 10-40 cm) as compared to the post-
larval sizes. This brings us to the next important
process, larval survival.

Spawning Survival Versus
Population Biomass

For our hypothesized unexploited population
of 600,000 metric tons of individuals from 40 to
140 cm fork length, we can calculate the requisite
number of postlarval survivors which must be
generated each year to maintain this stock at
equilibrium. Assuming 40-cm yellowfin tuna are
approximately 7 mo of age and that the survival
rate is constant for all ages after postlarval trans-
formation and is approximately equal to e"-^ on
an annual basis (Hennemuth 1961), the number
of postlarval survivors each year is given by the
relation

A^, = A^4oe

0.8(1)

If A^4o is approximately 2.12 x 10'^ individuals
per year in cohort S^, and 2.06 x 10^ in cohort Sq,
then there are approximately 6.67 x 10'' sur-
vivors/yr. If we assume that they are aggregated

FISHERY BULLETIN: VOL. 74, NO. 1

spatially but not temporally (there are two co-
horts of 3.33 X 10'' postlarvae spread approxi-
mately evenly over the year), approximately 9.13
X 10"* postlarvae enter the system daily. (This
is the equivalent of nearly 1% reproductive suc-
cess of either one 155-cm female or five 87-
cm females.)

The relative fecundity of yellowfin tuna is
given by Joseph (1963) to the following:

Number of eggs = 8.955 x lO'^ Z2.791
where / is the fork length of the fish in mm.

If we assume the average spawning female to
weigh 25 kg and we estimate the presence of
175,000 metric tons of females of reproductive
age in our unexploited population, then the equiv-
alent number of reproductive females is ap-
proximately equal to 7 x 10^. These females
would be an average of 107 cm in length and
therefore:

(8.955 X IQrH (l,0702-''9i) (7 x 10« females)
= 1.79 X 10^^ eggs produced.

So if 6.67 X 10"'' postlarvae start the process we
need invoke only 3.72 postlarval survivors per
million eggs spawned. This estimate is conserva-
tive due to the assumption that females only
spawTi once per year, whereas they could spawn
more often. (No evidence for or against multiple
spawnings is in existence for yellowfin tuna.)
It does, however, seem likely that spawning suc-
cess (survival to postlarvae) is greater than 3.72
individuals per million eggs produced (Sette
1943; Farris 1961). It is also important to mention
that all attempts at relating spawning biomass to
recruitment estimates for yellowfin tuna in the
CYRA have been futile. This could be due to error
in either, or both, estimates of spawning biomass
and recruitment and/or the possibility that
environmental conditions indeed override any
obvious relationships.

These comments are presented to point up the
likelihood that the density dependent factors for
limiting yellowfin tuna abundance are probably
more effective on the egg to larvae to juvenile
stages than at 40 cm or more. The larvae to 40-cm
fish are likely very narrowly distributed in the
water column (approximating a two-dimensional
distribution) due to thermal and energetic re-
quirements. The recruitment at 40 cm in the

highly productive regions such as the periphery
of the Costa Rica Dome and the Panama Bight-
Ecuador coastal regions can perhaps be best
explained by the high productivity levels in these
regions which ranges from 500 to 700 mg carbon
m'2 day"^ as compared to the 205 mg carbon m"^
day^ average CYRA carbon fixation rate, in con-
junction with the relatively shallow oxygen mini-
mum and thermal optima which probably act to
compress the available habitat toward the sur-
face. If one could invoke the ability of yellowfin
tuna to climb a food gradient, a simple volume
change in the preferred thermal-oxygen regime
combined with a negatively correlated food
gradient could result in the observed coastal
"emergence" of recruits, which "grow out" of
their previous thermal-oxygen limitations as they
develop, and exploit a significantly wider niche
than they could as relatively poikilothermal enti-
ties at sizes below 40 cm.

To summarize, larval tunas are relatively im-
mobile and for survival are probably dependent
on aggregations of food resources. The ability
of tunas, particularly postlarval sizes, to detect
food gradients is unknown, but may indeed ac-
count for the easterly trend in abundance of
recruits. The wider distributions of larger fish
(postrecruits) probably is a response to competi-
tive feeding problems and changing physiologi-
cal capabilities. These larger fish are increasing
their daily demands but are gaining in adaptive
physiological and morphological characteristics
which widen their niche as compared to smaller
sizes. Their mass and mobility insure their ability
to move rapidly from low to high availabilities
of food resources, in response to seasonal and
areal fluctuations in productivity, perhaps ac-
counting for the cyclic migratory behavior ob-
served in their first few years in the fishery. The
relative offshore surface distribution of the larger
fish (>40 cm) may be roughly correlated vdth the
depth distribution of the 22°-23°C isotherms, a
relationship which we are now starting to study.
As the larger fish grow in mass, they can afford
deeper and longer forays into colder than optimal
zones with low O2 availability to obtain larger and
more calorific food sources; and by thus increas-
ing the maximum excursion depth, competition
is likely to be less severe. The disaggregation of
larger sized fish into smaller schools (number
of individuals) may be accounted for by these
effects. The large yellowfin tuna in the offshore
areas are certainly concentrated at the surface

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

over highly productive regions where their main
sources of competition are probably porpoise
and bigeye tuna, Thunnus obesus. The porpoise-
tuna composite likely indicates the optimum
availability offish and squid in the eastern tropi-
cal Pacific. It is obvious from the Perrin et al.
(1973) studies that the two Stenella species and
tunas coexist but tend to feed differentially.
The tuna diet shares most of the organisms
found in both species indicating that they are
less selective and/or feed throughout the water
column.

No data support the concept of food limitation
for population size in yellowfin tuna in post-
recruit sizes and in most cases the arguments
tend toward the opposite conclusion. Since no
stable relationship can be found to exist be-
tween recruitment and spawning biomass, it is
unlikely that reproductive success is affected by
spawning biomass at the population levels we are
experiencing. More probable is that the environ-
mental parameters are more important in regulat-
ing the absolute numbers of surviving larval or
juvenile yellowfin tuna which are recruited to
the fishery.

In the future, we plan to incorporate the avail-
able productivity and environmental data (tem-
perature, oxygen, etc.) with a more complete
version of this model. We hope to determine the
environmental correlates with the fluctuations in
the catch, effort, and length-frequency data
generated from the fishery on yellowfin tuna. Pre-
liminary studies have been encouraging (Inter-
American Tropical Tuna Commission 1975) and
point up the need for data on the thermal pref-
erences (perhaps indicating energetic optima)
and the levels of environmental variability which
can be sensed and therefore compensated for by
the several tuna species at the various develop-
mental stages in their life cycles. Also obvious is
the need to work with smaller areas and corre-
sponding population segments rather than as-
suming "average" conditions in environmental
and population parameters. The ultimate goal of
these studies is the development of predictive
tools for use in assessing likely catch conditions
as well as the basic distributional properties of
the tunas. The use of unsupported guesses based
on overviews which integrate vast areas with sig-
nificant oceanographic and population structure
differences may do little more than obscure the
existing relationships which are important to
this goal. The application of the crude model we

have described in this study will depend upon
the development of better estimates of the
physiological parameters and appropriate use
of the areal breakdown in the population simu-
lator. Studies of trophic dynamics and competi-
tion interactions would help complete the pic-
ture necessary to "efficiently" manage a dynamic
resource. We hope to generalize, where possible,
the relationships which arise fi-om these analyses
in order to provide a useful descriptive tool as
well as a hypothesis testing device for studying
the occurrence, abundance, and availability of
tunas in the world ocean.

LITERATURE CITED

Alverson, F. G.

1963. The food of yellowfin and skipjack tunas in the
eastern tropical Pacific Ocean. [In Engl, and Span.]
Inter-Am. Trop. Tuna Comm. Bull. 7:293-396.
BAINBRIDGE, R.

1961. Problems of fish locomotion. Symp. Zool. Soc.
Lond. 5:13-32.
BLACKBURN, M., R. M. LAURS, R. W. OWEN, AND B.
ZEITZSCHEL.

1970. Seasonal and areal changes in standing stocks of
phytoplankton, zooplankton and micronekton in the
eastern tropical Pacific. Mar. Biol. (Berl.) 7:14-31.

Carey, F. G., and J. M. Teal.

1966. Heat conservation in tuna fish muscle. Zoology
56:1464-1469.
CHATWIN, B. M.

1959. The relationships between length and weight of
yellowfin tuna (Neothunnus macropterus) and skipjack
tuna (Katsuwonus pelamis) fi-om the Eastern Tropical
Pacific Ocean. [In Engl, and Span.] Inter-Am. Trop.
Tuna Comm. Bull. 3:305-352.

farris, d. a.

1961. Abundance and distribution of eggs and larvae
and survival of larvae of jack mackerel (Trachurus sym-
metricus). U.S. Fish Wildl. Serv., Fish. Bull. 61:247-279.
Francis, R. C.

1974. TUNP0P, a computer simulation model of the
yellowfin tuna population and the surface tuna fishery
of the eastern Pacific Ocean. [In Engl, and Span.]
Inter-Am. Trop. Tuna Comm. Bull. 16:235-279.

FRY, F. E. J.

1957. The aquatic respiration of fish. In M. E. Brown
(editor), The physiology of fishes, Vol. 1, p. 1-63. Aca-
demic Press Inc., N.Y.
Gooding, R., E. Poe, and C. Nagamine.

1973. Tuna newsletter No. 9 July 1973. Natl. Mar. Fish.
Serv., Southwest Fisheries Center, La Jolla, Calif

hennemuth, r. c.

1961. Size and year class composition of catch, age and
growth of yellowfin tuna in the eastern tropical Pacific
Ocean for the years 1954-1958. [In Engl, and Span.]
Inter-Am. Trop. Tuna Comm. Bull. 5:1-112.

Inter-American Tropical Tuna commission.

1975. Aimual Report of the Inter-American Tropical Tuna
Commission, 1974. [in Engl, and Span.] 169 p.

FISHERY BULLETIN: VOL. 74, NO. 1

JOSEPH, J.

1963. Fecundity of yellowfin tuna (Thunnus albacares)
and skipjack {Katsuwonus pelamis) from the Pacific
Ocean. [In EngL and Span.] Inter-Am. Trop. Tuna
Comm. Bull. 7:257-292.
KISHINOUYE, K.

1923. Contributions to the comparative study of the so-
called scombroid fishes. J. Coll. Agric, Imp. Univ.
Tokyo 8:293-475.
KUTTY, M. N.

1968. Respiratory quotients in goldfish and rainbow
trout. J. Fish. Res. Board Can. 25:1689-1728.
LIGHTHILL, M. J.

1970. Aquatic animal propulsion of high hydromechani-
cal efficiency. J. Fluid Mech. 44:265-301.
MAGNUSON, J. J.

1966. Continuous locomotion in scombroid fishes. (Abstr.)
Am. Zool. 6:503-504.

1970. Hydrostatic equilibrium of Euthynnus affinis, a
pelagic teleost without a gas bladder. Copeia
1970:56-85.

1973. Comparative study of adaptations for continuous
swimming and hydrostatic equilibrium of scombroid and
xiphoid fishes. Fish. Bull., U.S. 71:337-356.

MAGNUSON, J. J., AND J. G. HEITZ.

1971. Gill raker apparatus and food selectivity among
mackerels, tunas, and dolphins. Fish. Bull., U.S.
69:361-370.

Murphy, G. I.

1965. A solution of the catch equation. J. Fish. Res.
Board Can. 22:191-202.

NEILL, W. H., AND E. D. STEVENS.

1974. Thermal inertia versus thermoregulation in "Warm"
turtles and tunas. Science (Wash., D.C.) 184:1008-1010.

OWEN, R. W., AND B. ZEITZSCHELL.

1970. Phytoplankton production: Seasonal change in the
oceanic eastern tropical Pacific. Mar. Biol. (Berl.)
7:32-36.
PALOHEIMO, J. E., AND L. M. DICKIE.

1966. Food and growth of fishes. 11. Effects of food and
temperature on the relation between metabolism and
body weight. J. Fish. Res. Board Can. 23:869-908.

Perrin, W. F., R. R. Warner, C. H. Fiscus, and D. B. Holts.

1973. Stomach contents of porpoise, Stenella spp., and
yellowfin tuna, Thunnus albacares, in mixed-species
aggregations. Fish. Bull., U.S. 71:1077-1092.

PHILLIPS, A. M., JR.

1969. Nutrition, digestion, and energy utilization. In
W. S. Hoar and D. J. Randall (editors). Fish physiology.
Vol. 1, p. 391-432. Academic Press, N.Y.

PLATT, T., AND B. ERWIN.

1973. Caloric content of phytoplankton. Limnol.
Oceanogr. 18:306-310.

Pyatetskiy, V. Ye.

1971. Hydrodynamic swimming characteristics of some
fast marine fish, from translation of monograph; Kiev,
Bionika, Russian, No. 4; 1970, p. 3-120. Hydrodynamic
Problems of Bionics. JPRS 52605, p. 24-31.

Sette, O. E.

1943. Biology of the Atlantic mackerel (Scomber scom-
brus) of North America. Part I: Early life history, includ-
ing the growth, drift, and mortality of the egg and lar-
val populations. U.S. Fish Wildl. Serv., Fish. Bull.
50:149-237.

SMIT, H.

1965. Some experiments on the oxygen consumption of
goldfish iCarassius auratus L.) in relation to swimming
speed. Can. J. Zool. 43:623-633.

STREETER, V. L.

1962. Fluid mechanics. 3rd ed. McGraw Hill Book
Co., N.Y., 555 p.
TOMLINSON, P. K.

1970. A generalization of the Murphy catch equation.
J. Fish. Res. Board Can. 27:821-825.

Walters, V.

1962. Body form and swimming performance in the
scombroid fishes. Am. Zool. 2:143-149.

1966. The "problematic" hydrodynamic performance of
Gero's great barracuda. Nature (Lond.) 212:215-216.

WARREN, C. E.

1971. Biology and water pollution control. W. B.
Saunders Co., Philadelphia, 434 p.

WARREN, C, E., AND G. E. DAVIS.

1967. Laboratory studies on the feeding, bioenergetics,
and growth of fish. In S. D. Gerking (editor). The
biological basis for freshwater fish production, p. 175-
214. John Wiley & Sons Inc., N.Y.

WINBERG, G. G.

1956. [Rate of metabolism and food requirements of
fishes.] Nauchnye Tr. Belorussk. Gos. Univ. Minsk,
253 p. (Transl. 1960. Fish. Res. Board Can., Transl.
Ser. 194.)

SHARP and FRANCIS: ENERGETICS MODEL FOR YELLOWFIN TUNA POPULATION

APPENDIX.— GLOSSARY OF TERMS

E„

E^ =

E. =

F =

g
k

L, =

M =
M, =

Mf =

N. =

N^ =
Re =

wetted surface area of the fish.

the total lifting area of the pectoral fins.

the total lifting area of the keel.

the coefficient of lift of the pectoral fins.

the coefficient of lift of the keel.

coefficient of total drag of fish of length
J which includes an inseparable effi-
ciency term involving acceleration pro-
cesses during continuous swimming.

the daily caloric expenditure of fish of
length J attributable to growth in the
form of positive changes in mass.

the daily caloric expenditure of fish of
length 7 to maintain metabolic stasis.

the daily caloric energy expenditure of
fish of length I utilized by swimming
work, a function of swimming velocity

(V.eal)-

the instantaneous mortality rate due to
fishing.

acceleration due to the force of gravity.

the rate of oxygen consumption due to met-
abolic stasis of 1 g of respiring tissue,
not doing external work.

the length of a fish from snout to fork of
tail in millimeters.

the fork length of a fish in centimeters.

the total weight of a fish in seawater of
density p, in dynes.

the instantaneous natural mortality rate.

the efficiency of muscle when converting
chemical energy to mechanical work.

mass of the fish in grams where for yellow-

fin tuna: M

1.858 X 10-2 (/) 3.021

(Chatwin 1959).
the estimated number of individuals of
length I.

the number of postlarval survivors from

a spawning,
the number of recruits at 40 cm.
the Reynolds number.

'total

= the density of seawater, in this work p =
1.025 g/cm2.

= the rate of oxygen consumption due to
swimming activity, from the power
equation of Smit (1965).

= recruitment cohort label for all individuals
that attain 40 cm fork length from 1
January to 30 June of each year.

= recruitment cohort label for all individuals
that attain 40 cm fork length from 1
July to 31 December of each year.

= the oxygen consumption rate of fish of
length J attributable to metabolic stasis.

= the oxygen consumption rate of a fish of
length 7 attributable to swimming en-
ergy expenditures.

S + S

= respiration rate attributable

Mf X

10-3

V
V

100

to swimming and metabolic stasis en-
ergy expenditures.

= the kinematic viscosity of seawater.

= the constant velocity of a fish, in centi-
meters per second.

= the estimated integrated velocity of a fish
of length 7 used in determining Re and
C<f, and in the estimation of S.

= the minimum swimming speed of a fish of
given species and 7 for maintenance
of hydrostatic equilibrium (Magnuson
1973).

= the velocity which is "typical" of the
swimming speed of a fish of length 7.
Vfgej) = the velocity which is meant to integrate
all energy expenditures due to fish
swimming faster than V^yp, including
short bursts in feeding or flight be-
havior (assumed to be 3 m/s).
^reai " the average daily velocity of a fish of
length 7, = 0.95 V,^ + 0.5 Vfeed-
the metabolic weight of a fish, in grams
(Winberg 1960).

typ

met

EFFECTS OF TEMPERATURE AND SALINITY ON
THE SURVIVAL OF WINTER FLOUNDER EMBRYOS

Carolyn A. Rogers^

ABSTRACT

A series of experiments was performed to determine the optimum temperature and salinity for
incubating winter flounder, Pseudopleuronectes americanus, embryos. Eggs in lots of 50 were sub-
jected to a 0.5 to 45% salinity range and a 3° to 14°C temperature range in a total of 67 salinity-
temperature combinations. Highest proportion of viable hatches occurred at 3°C over a salinity range
of 15 to 35%. At temperatures above 3°C, the optimal range was 15 to 25%. Viable hatch decreased with
increasing temperature.

The winter flounder, Pseudopleuronectes
americanus (Walbaum), an important species in
local New England commercial and sport fishing
industries, occurs from Chesapeake Bay to the
northern shore of the Gulf of St. Lawrence
(Bigelow and Schroeder 1953). The adults dis-
perse into cooler offshore waters as temperatures
rise, but move back into embayments and es-
tuaries in the fall. Spawning occurs in shoal wa-
ters of these areas from February to mid-May with
the maximum in Rhode Island waters occurring
in March (Perlmutter 1947; Bigelow and
Schroeder 1953; Pearcy 1962). Winter flounder
spawn demersal eggs, which range from 0.74 to
0.85 mm in diameter when fertilized. Hatching
occurs in 15 to 18 days at 3° to 4°C, the tempera-
ture normally encountered in the natural envi-
ronment (Bigelow and Schroeder 1953).

This paper reports the optimum temperature
and salinity ranges for the development and sur-
vival of winter flounder embryos and larvae and
discusses the relationship between the two fac-
tors as it affects embryo development. An earlier
study (Scott 1929) indicated some of the effects of
temperature and salinity as separate factors on
the hatching of winter flounder eggs but pre-
sented no data on possible interaction of the two.
Forrester and Alderdice (1966) and Alderdice and
Forrester (1968, 1971a, b) working on the effects
of temperature and salinity on the embryonic de-
velopment of the English sole, Parophrys uetulus;
petrale sole, Eopsetta jordani; and Paciflc cod,
Gadus macrocephalus , respectively, indicated a

'Northeast Fisheries Center Narragansett Laboratory,
National Marine Fisheries Service, NOAA, Narragansett,
RI 02882.

relationship between the two factors, which
influenced early development, hatching time, and
viable hatch.

METHODS AND MATERIALS

Ripening adult winter flounder were captured
by trawl on 29 October 1970 at a depth of 23 to 30
m in Block Island Sound. Surface waters were
15°C, and a bottom temperature of 12°C was es-
timated for that area (Colton and Stoddard 1973).
The live fish were transported to the laboratory
where they were held in running water aquaria
until they were ripe in early February when am-
bient water temperature was 3°C. The fish were
fed clam worms, earthworms, and cut up clam
during the holding period. Eggs were stripped
into polyethylene dishpans, fertilized, and coated
with diatomaceous earth to prevent clumping, ac-
cording to the technique of Smigielski and Arnold
(1972). Fertilized eggs were transferred to incu-
bation baskets and held at 3°C in running seawa-
ter (32% salinity) for 24 h when normal develop-
ment could be distinguished. Day 1 embryos were
in the early blastoderm stage when the experi-
ments were started. Three separate experiments
were run at salinities ranging from 0.5 to 45%,
and at temperatures of 3° to 14°C. Each experi-
ment was run in duplicate.

To avoid bias, all salinities were prepared by
adding Instant Ocean ^ salts to normal seawater
(32%) to bring the salinity up to 50%. Experimen-
tal salinities were then made by diluting the
stock salinity with distilled water. Each salinity

^Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

Manuscript accepted March 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

ROGERS: EFFECTS OF TEMPERATURE AND SALINITY ON WINTER FLOUNDER

was checked with a refractometer to within
±0.15% of the test salinity. The test salinities
were cooled to the ambient seawater temperature
(3°C) at which the eggs were incubated for the
first 24 h.

Eggs in lots of 50 were counted into 100-ml
polyethylene beakers filled with the test
salinities. The beakers were covered with fitted
50-mm plastic disposable culture dish bottoms to
eliminate evaporation and placed in thermostati-
cally controlled water baths at the experimental
temperatures. Dead eggs or larvae were removed
daily and examined for stage of development.

Daily observations were made on the develop-
ment of embryos. The time of hatching and the
duration of the hatching interval were noted so
that mean hatching time (time from fertilization
to 50% hatch) could be calculated. Abnormal lar-
vae (those with curvature of the spine, abnormal
yolk sacs, or enlarged fin folds) were noted and
counted as nonviable since their chance of con-
tinued survival was considered to be small. Pre-
maturely hatched or aborted larvae were also
considered nonviable in calculations. Such larvae
were easily recognized since they were short,
thickened, often curled, and in no way resembled
a normal healthy larva.

Each experiment was terminated when all eggs
had either hatched or died, and when the larvae
could be judged normal or abnormal. From this
information, total percentage hatch (percentages
of eggs producing live larvae) as well as percen-
tage viable hatch (percentage producing viable or
normal larvae) was calculated. Salinities were
checked at the end of each experiment.

The experiment was set up as a factorial de-
sign. However, replications at different factor
combinations were unequal and there were mis-
sing data at 3°C due to equipment malfunction.
In view of this, a mean value of the replicates was
computed for each factor combination and values
for the missing data at 3°C were predicted from
the hyperbolic equation describing the actual
data at 3°C. The resultant design was a 2 factor, 6
X 12 (6 levels of temperature and 12 levels of
salinity) factorial design with no replicates. Dun-
can's multiple range test (Steel and Torrie 1960)
was used to compare the mean survivals for each
temperature and salinity condition.

RESULTS

The results of these experiments indicate that

winter flounder embryos are euryhaline, with
best survival occurring between 10 and 30% but
with some survival from 5 to 40%. Hatching oc-
curred at all temperatures tested, but the lower
temperatures produced the highest survival. In-
cubation time and hatching interval were de-
creased by increased temperatures and higher
salinities. Abnormal development occurred par-
ticularly at extremes of salinity but was also
influenced by temperature.

Effects of Salinity and Temperature
on Viable Hatch

Results of the temperature-salinity experi-
ments (Table 1) indicated an optimal salinity
range between 15 and 25% for temperatures
above 3°C and between 15 and 35% for 3°C (Fig-
ure 1, Table 2). Viable hatch was highest at 3°C
and lowest at 14°C with similar survival rates at
5, 7, and 12°C for all salinities. Percentage survi-
val at 10°C follows a similar curve at salinities of
25% and above, but was between 15 and 30%
lower than that of other temperatures at 20% and
below. At 3°C, high survival (>78%) occurred
from 15 to 35%>, but survival decreased sharply at
all other temperatures for salinities above 25%.

Table l. — Number of winter flounder eggs at each of 67
temperature-salinity combinations. Number of replicates
shown in parentheses

Salinity

Temperature (°C)

C^)

0.5

100(2)

100 (2)

100(2)

5.0

100(2)

300(6)

400(8)

300(6)

300 (6)

7.5

100 (2)

100(2)

50(1)

100 (2)

10.0

100 (2)

300(6)

400(8)

300(6)

15.0

300(6)

400(8)

200(4)

300(6)

20.0

100(2)

300(6)

400(8)

300(6)

300 (6)

300(6)

25.0

200(4)

300(6)

200(4)

30.0

100(2)

300(6)

400(8)

300(6)

35.0

100(2)

300(6)

400(8)

300 (6)

300(6)

37.5

100 (2)

100(2)

100 (2)

40.0

200(4)

300 (6)

200(4)

200 (4)

45.0

100 (2)

100(2)

100 (2)

100(2)

Influence of Temperature and Salinity
on Total and Viable Hatch

The influence of temperature and salinity is
shown in the percentages of mean total hatch and
mean viable hatch (Table 3). There is a sharp de-
crease in mean total hatch and mean viable hatch
at temperatures over 3°C, while these means ap-
proximate a normal distribution at the salinities
tested. The mean percentage of abnormal larvae
calculated from total and viable hatch data shows

FISHERY BULLETIN: VOL. 74, NO. 1

0.5

5 7.6 10

15 20 25

SALINITY (%.)

30 35

Figure l. — The effects of temperature and salinity on the percent viable hatch of winter flounder embryos.

Table 2. — Mean percent total and viable ( ) hatch at the various temperature-
salinity combinations.

Salinity

Temperature (°C)

(y„)

0.5

0(0)

0.0 (0)

5.0

26(0)

6.3 (0)

14.5 (6.5)

23.0 (0)

0.0 (0)

7.5

—

58.0 (26.0)

48.0 (26.0)

49.0 (21.0)

46.0 (17.0)

23.0 (7.0)

10.0

88 (61)

79,7 (65.7)

71.5 (57.8)

59.0 (32.0)

82.7 (65,3)

55.3 (32.7)

15.0

92 (84)

79.3(75.7)

76.5 (71.0)

71.0 (57.0)

77.0 (69.0)

69.3 (57.3)

20.0

100(99)

82.3 (79.3)

83.8 (82.0)

70.3(61.0)

78.3 (68.0)

61.3 (48.7)

25.0

—

75.5 (74,0)

69.3 (66.7)

74.0 (66,5)

62.0 (56 5)

57 5 (42.0)

30.0

74 (64)

54.7 (45.7)

59.8 (51.3)

50.0 (43.7)

63.0 (54.7)

48.7 (32.3)

35.0

84 (67)

31.3 (27.3)

42.8 (37.5)

31.0 (24.0)

47,0 (34,5)

21.0 (7.7)

37.5

—

40.0 (37.0)

86.0 (78.0)

57.0 (52.0)

38.0 (16.0)

5.0 (0)

40.0

—

19.0 (9.5)

63.0 (15.7)

34.0 (17.0)

26.0 (3.5)

0.0 (0)

45.0

0(0)

0.0 (0)

Table 3. — Means and ranges for percent total and viable
hatches and mean abnormal hatches for each salinity at all
temperatures, and each temperature at all salinities.

Mean % total hatch

Mean % viable hatch

Mean abnormal

Item

(Range)

hatch' (%)

0.5%

No hatch

5.07oo

12,8(2.3-26.0)

1.6(0-6.5)

11.2

7.57..

44.8(23.0-58.0)

19.4(7.0-26.0)

25.4

io.oy„

72.7(55.3-88.0)

52.4(32.0-65.7)

20.3

15.07..

77.5(69.3-92.0)

72.3(57.0-84.0)

5.2

20.07.O

79.3(61.3-100)

73.0(48.7-99.0)

6.3

25.07„

67.7(57,5-74.0)

61.1(42.0-74.0)

6.6

30.07„

58.4(48.7-74.0)

48.6(32.3-64.0)

9.8

35.07„

42.9(21,0-84.0)

33.0(7.7-67.0)

9.9

37.57.0

45.2(5.0-86.0)

36.6(0-78.0)

8.6

40.07..

15.1(9.5-21.0)

11.4(3.5-17.0)

3.7

45.07„

No hatch

3°C

77.3(26.0-100)

62.5(0-99.0)

14.8

5°C

51.7(6.3-82.3)

44.0(9.5-79.3)

7.7

57,3(14,5-86.0)

49.3(6.5-82.0)

8.0

10°C

48.1(2.3-74.0)

37.4(0-66.5)

10.7

12''C

56.3(13.0-82.7)

42.7(3.5-69.0)

13.6

14°C

42.6(5.0-69.3)

28.5(0-57.3)

14.1

'Mean abnormal hatch
viable hatches.

mean percent total hatches - mean percent

no trend with temperature, but a high percentage
of abnormal larvae for salinities of lO'L and be-
low. Lowest percentages for abnormal larvae
were for salinities between 15 and 35%. The low
percentage for 40.0% reflects low hatching rates
and mortality during embryonic stages and does
not reflect values which can be compared with
salinities of 37.5%o and below.

Analysis of variance performed on the survival
data indicate that salinity and temperature are
both significant factors (Table 4). Because of miss-
ing data (Table 1), it was not possible to test for
interaction between the two factors; however, by
examining the data, especially as it is expressed
in Figure 1, it is reasonable to conclude that an
interaction does occur. The multiple comparison
of means indicates significant differences be-
tween hatch means at various temperatures and

ROGERS: EFFECTS OF TEMPERATURE AND SALINITY ON WINTER FLOUNDER

Table 4. — Analysis of variance for the effects of tempera-
ture and salinity on the survival and hatching of winter
flounder embryos.

Source of
variation

Total

Salinity

Temperature

Residual

"significant at P = 0.005.

Table 5. — Duncan's multiple comparison of means for
temperature-salinity studies of winter flounder embryos.
(Means with similar symbols denote similar mean survi-
val percentages.)'

Degrees of

Sum of

Mean

freedom

squares

square

69.248.75

51,935.36

4,721.39

31.5

9,078.78

1,815.76

12.2

8,234.61

149.72

Temperature

Mean survival

(°C)

(%)

56. IV

36.2*

41. rV

31.4*x

32.4-^

18.9^

Salinities

Mean survival
(%)

0.5
5.0
7.5
10.0
15.0
20.0
25.0
30.0
35.0
37.5
40.0
45.0

0.0 V
1 . 1 v
21.6"
53.7°
69.9t
74.3t
67.4t
52.6°
35.6^
40.3x
15.3*
0.0 V

'P = 0.05.

salinities and allows a grouping of each in order
of its significance (Table 5). The grouping of the
hatch means for variations in both temperature
and salinity coincides closely with viable hatch
curves illustrated in Figure 1.

Incubation Time and Duration of
Hatching Interval

The time to 50% hatch and the total range of
hatching time for each temperature and salinity
combination are recorded in Table 6. Figure 2 il-
lustrates the time to 50% hatch and the mean
incubation time for each temperature and salin-
ity respectively. The mean hatching interval

> 22

o 18

2 -

5. 00

4.0 C

20 25 30

SALINITY (%.)

FIGURE 2.

-The effects of salinity on the time to 50% hatch of
winter flounder embryos.

ranges from 25 days at 3°C (10%) to 7 days at 12°
and 14°C (37.5 and 35% respectively). Individual
eggs hatched in as few as 5 days in most salinities
at 12° and 14°C, but took as long as 31 days at
3°C (10%). An inverse relationship for tempera-
ture with respect to the duration of hatching time
is evident.

There is also a trend toward the same inverse
relationship with respect to salinity as can be
seen in Figure 2 where the time to mean 50%
hatch at all temperatures decreased slightly with
increasing salinities. This phenomenon of greater
hatching time at low salinities was noted in
Pacific cod eggs by Forrester and Alderdice
(1966). When salinity means versus incubation
time is considered by least squares regression,
there is a low correlation coefficient and a regres-
sion relationship is not applicable (Figure 3).
However, temperature means have a high corre-
lation coefficient and there is a strong regression
relationship present.

Table 6. — Time in days to 50% hatch. Range of hatching interval in days shown in
parentheses. NH denotes no hatch.

Temperature

Salinity (%)

(°C)

5.0

7.5

10.0

15.0

20.0

25.0

30.0

35.0

37.5

40.0

(19-31)

(19-27)

(19-25)

(17-25)

19
(16-25)

(16-20)

(17-29)

(17-25)

(17-29)

(22-24)

(13-25)

(11-25)

(14-16)

(10-16)

(12-23)

(12-25)

(8-17)

(8-21)

(11-19)

(12-14)

(8-14)

(10-14)

(7-15)

(9-14)

(7-16)

(7-17)

(5-13)

(8-10)

(7-10)

(7-12)

(5-10)

(7-12)

(7-10)

(5-10)

(6-10)

(5-10)

FISHERY BULLETIN: VOL. 74, NO. 1

SALINITY %.
15 20 25 30

T 1 1 r

o SALINITY

•

y 16.0903-0.06824,

/• = -0.54509

f- 2.959

• TEMPERATURE

y- 28.7585-1.5441..

•S.

r -- 0.96095

/^= 48.342

O >s.

_ _ • \v

O 0 o

.^ ° ° SALINITY

N. o 0

• \.

^V •

TEMPERATURE

1 J 1 1

6 8 10 12 14 16 16

TEMPERATURE "C

Figure 3. — The mean hatching time of winter flounder
embryos for each temperature and sahnity.

Effects of Temperature and Salinity
on Embryonic Development

In each of the three experiments, general ob-
servations were made on the eggs, embryos, and
larvae (Figure 4). No development occurred in a
salinity of 0.5%; however, the eggs swelled ap-
proximately 20% before death occurred. A diame-
ter increase of 8 to 10% was also observed in eggs
held at 5%. Below 10°C, embryos held in 5% ap-
peared to develop normally, then died just prior to
hatching. At 10°C and above, most of the embryos
died during gastrulation. Embryos held in a sa-
linity of 10% had the highest mortalities just
prior to hatching and at hatching; many larvae
were observed dead partly emerged from the
chorion. Mortality occurred throughout develop-
ment at 12° and 14°C.

In salinities between 15 and 30%, most mor-
talities occurred just prior to hatching, although

MO bEVELOPMEWr

12-

<
UJ

CO L.\_AP5 1

£a^ftfcK30KJfV\.(VU bfeVELOPMEKir

' I

OF EMMlyO

)

—I 1 1 1 1 1 1 1 1 1 1 '_

0.5 5 7.5 10 15 20 25 30 35 37.5 40 45

SALINITY (%•)

Figure 4.— The qualitative effects of temperature and salinity on the development and hatching of wdnter flounder embryos.

ROGERS: EFFECTS OF TEMPERATURE AND SALINITY ON WINTER FLOUNDER

at temperatures of 10°C and above some mor-
talities usually occurred during gastrulation. At
salinities of 35 to 40%, abnormal development of
the embryos was observed. The embryos were
shorter and thicker than normal and died just
prior to hatching. Collapsing eggs were noted at
37.5'Ii and above. Embryos incubated at 40% died
during gastrulation and throughout development
at all temperatures while all embryos held at 45%
died during gastrulation. At both 40 and 45% em-
bryos exhibited shrinkage and often collapsed.

DISCUSSION

The results indicate that although temperature
and salinity are both significant, the major effect
of increased temperature is to decrease the incu-
bation period, whereas salinity is the factor
which has more effect on the successful hatching
and survival of winter flounder embryos and lar-
vae (Figure 4, Table 4). It is apparent however,
that an interaction between the two does occur
since, at the optimum experimental temperature
(3°C), the salinity range over which high percen-
tages of viable hatches occurred was extended by
10% (Figure 1). At higher than optimal experi-
mental temperatures, the survival curves appear
to be dictated primarily by salinity; however,
survival occurs over a broad enough range that
the embryos and larvae can be described as
euryhaline with regard to the natural environ-
ment in which they are normally spawned. At all
temperatures tested, there was a decrease in in-
cubation time at higher salinities, a phenomenon
which was also reported in studies done on
Clupea harengus (Holliday and Blaxter 1960) and
Pacific cod (Forrester and Alderdice 1966). Those
authors speculated that the relationships of
temperature and salinity with hatching are de-
pendent on conditions that minimize the energy
required of the embryos in maintaining osmotic
equilibrium with their environment. Salinity also
appears to influence the time of embryo mortal-
ity. Observations on eggs indicated that mortality
usually occurred either at gastrulation, in
salinities of 40 and 45% at all temperatures, or
just prior to hatching in the lower salinities. Bat-
tle (1930) noted increased mortality of the four
bearded rockling, Enchelyopus cimbrius, at
hatching in low salinities and she attributed this
to poorly developed tail musculature. McMynn
and Hoar (1953), working with embryos of the
Pacific herring, Clupea harengus pallasi, ob-

served that with the closing of the blastopore at
the end of grastrulation, embryos had a greater
ability to tolerate low salinities. However, many
embryos died just prior to hatching or when
partly emerged. Holliday (1965, 1969) observed a
similar occurrence in cod, Gadus callarius, and
plaice, Pleuronectes platessa. He felt that the low
specific gravity of such salinities made it difficult
for larvae to free themselves from the chorion so
that they died partly emerged. He also main-
tained that chorions did not rupture as easily at
low salinities. This phenomenon is also clearly
demonstrated for winter flounder in Table 3. The
highest percentages of abnormalities which were
aborted or partially hatched occurred at salinities
below 15%.

Results of these laboratory experiments indi-
cate that successful incubation of embryos oc-
curred over a temperature range which exceeded
normal spawning season temperatures by as
much as 10°C, but coincide quite closely with
natural observations for salinity, although there
is a shift in survival toward slightly higher
salinities than would have been expected. It is
possible that the adults, while being held in the
laboratory, were conditioned to slightly higher
salinities than would have been encountered in a
spawning migration into estuaries. This might
explain the differences between natural popula-
tions and results of laboratory experiments.

Most winter flounder populations move to in-
shore and estuarine waters to spawn (Perlmutter
1947; Bigelow and Schroeder 1953; Saila 1961),
but there are also spawning populations that re-
main in offshore shoals (Bigelow and Schroeder
1953; Marak et al. 1962). Field observations in
two estuaries of Narragansett Bay and in the Bay
itself indicate that spawning occurs at salinities
ranging from 11 to 32%. Plankton tows taken in
upper Chesapeake Bay produced one egg in 20%
with maximum numbers of larvae occurring be-
tween 6 and 14% (Dovel 1971). Salinities in sus-
pected offshore shoal spawning areas range from
32 to 35.5%, at the bottom (Bumpus 1973), so an
overall spawning range from 5 or 6 to 35.5% is
indicated for natural populations. The normal
temperature range for spawning is 0° to 3.3°C
with maximum temperatures for any appreciable
egg production and spawning being 4.2° to 5.6°C
(Bigelow and Schroeder 1953). Since the eggs are
demersal and adhesive, they are not subject to
transport into areas of unsuitable temperatures;
being estuarine, they are subjected instead to

FISHERY BULLETIN: VOL. 74, NO, 1

changes in salinity. However, the euryhaline
properties of the eggs insure successful incuba-
tion and larval development in a constantly vary-
ing salinity environment.

ACKNOWLEDGMENTS

I thank John Green and Geoffrey C. Laurence
for their review of the manuscript, and Geoffrey
C. Laurence for his assistance on the statistical
analyses. Technical assistance given by Thomas
Halavik and Alphonse Smigielski is gratefully
acknowledged. The illustrations were prepared
by Merrie Marsh and Lianne Armstrong.

LITERATURE CITED

Alderdice, d. F., and C. R. Forrester.

1968. Some effects of salinity and temperature on early
development and survival of the English sole iParo-
phrys vetulus). J. Fish. Res. Board Can. 25:495-521.

1971a. Effects of salinity and temperature on embryonic
development of the petrale sole Eopsetta jordani. J.
Fish. Res. Board Can. 28:727-744.

1971b. Effects of salinity, temperature, and dissolved
oxygen on early development of the Pacific cod (Gadus
macrocephalus). J. Fish. Res. Board Can. 28:883-902.

Battle, H. I.

1930. Effects of extreme temperatures and salinities on
the development of Enchelyopus cimbrius (L.). Con-
trib. Can. Biol. Fish., New Ser., 5:107-192.
BIGELOW, H. B., AND W. C. SCHROEDER

1953. Fishes of the Gulf of Maine. U.S. Fish Wildl.
Serv., Fish. Bull. 53, 577 p.
BUMPUS, D. F.

1973. Continental Shelf, arms of the sea. In Coastal
and offshore environmental inventory. Cape Hatteras
to Nantucket Shoals. Saul B. Saila, co-ordinator. Univ.
R.I., Mar Publ. Ser. 2, p. 1-1 - 1-46.
COLTON, J. B., Jr., and R. R. STODDARD.

1973. Bottom-water temperatures on the Continental
Shelf, Nova Scotia to New Jersey. U.S. Dep. Commer.,
NOAA Tech. Rep. NMFS CIRC-376, 55 p.

DOVEL, W. L.

1971. Fish eggs and larvae of the upper Chesapeake
Bay. Univ. Md. Nat. Res. Inst., Spec. Rep. 4, 71 p.

FORRESTER, C. R., AND D. F. ALDERDICE.

1966. Effects of salinity and temperature on embryonic
development of the Pacific cod (Gadus macrocephalus).
J. Fish. Res. Board Can. 23:319-340.
HOLLIDAY, F. G. T.

1965. Osmoregulation in marine teleost eggs and larvae.

Calif Coop. Oceanic Fish. Invest. Rep. 10:89-95.
1969. The effects of salinity on the eggs and larvae of
teleosts. In W. S. Hoar and D. J. Randall (editors),
Fish physiology. Vol. l,p. 293-311. Academic Press, N.Y.
HOLLIDAY, F. G. T., AND J. H. S. BLAXTER.

1960. The effects of salinity on the developing eggs and
larvae of the herring. Mar. Biol. Assoc. U.K. 39:591-603.

MARAK, R. R., J. B. COLTON, JR., AND D. B. FOSTER.

1962. Distribution of fish eggs and larvae, temperature,
and salinity in the Georges Bank-Gulf of Maine area,
1955. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish.
411, 66 p.
MCMYNN, R. G., AND W. H. HOAR

1953. Effects of salinity on the development of the Pacific
herring. Can. J. Zool. 31:417-432.
PEARCY, W. G.

1962. Ecology of an estuarine population of winter
flounder Pseudopleuronectes americanus (Walbaum).
Bull. Bingham Oceanogr. Collect., Yale Univ. 18(l):5-78.

perlmutter, a.

1947. The blackback flounder and its fishery in New
England and New York. Bull. Bingham Oceanogr.
Collect., Yale Univ. ll(2):l-92.
Saila, S. B.

1961. The contribution of estuaries to the offshore winter
flounder fishery in Rhode Island. Proc. Gulf Caribb.
Inst., 14th Annu. Sess., p. 95-109.

Scott, w. c. m.

1929. A note on the effect of temperature and salinity on
the hatching of eggs of the winter flounder (Pseudo-
pleuronectes americanus, Walbaum). Contrib. Can.
BioL 4(11):137-141.

Smigielski, A. S., and C. R. Arnold.

1972. Separating and incubating winter flounder eggs.
Prog. Fish-Cult. 34:113.

Steel, R. G. D., and J. H. Torrie.

I960. Principles and procedures of statistics with special
reference to the biological sciences. McGraw-Hill Co.,
N.Y., 481 p.

REEVALUATION OF FISHING EFFORT AND APPARENT ABUNDANCE

IN THE HAWAIIAN FISHERY FOR SKIPJACK TUNA,

KATSUWONUS PELAMIS, 1948-70

Richard N. Uchidai

ABSTRACT

Catch per effective trip, used in 1948-64 as an index of apparent abundance of skipjack tuna, Kat-
suwonus pelamis , in Hawaiian waters, is biased because effective trip, defined as one on which fish were
caught, underestimates effort. Catch per day fished, calculated from data collected in 1965-70, is a
refined index because effort includes days with or without catches. This paper describes the existence of
a linear relationship between catch per effective trip and catch per day fished in 1965-70, and a method
of estimating the latter from the former in 1948-64 based on this relationship. Fishing intensity, which
was measured by standard effective trips in past studies, is calculated in standard days fished. Changes
in catch per standard day fished are not associated with changes in relative fishing intensity. Skipjack
tuna abundance in Hawaiian waters, therefore, is fishery independent and is probably influenced by
availability and strength of year classes.

In the study of the dynamics of any exploited fish
population, data on commercial catch and fishing
effort can be interpreted in a number of ways,
giving various estimates of apparent abundance.
The ultimate objective, however, is to obtain the
best possible estimate of apparent abundance.

Prior to 1965, studies on catch and effort statis-
tics in the Hawaiian pole-and-line fishery for
skipjack tuna, Katsuwonus pelamis, defined
fishing effort as a "productive" or "effective" trip,
that is, one in which skipjack tuna were caught
(Yamashita 1958; Shippen 1961; Uchida 1966,
1967). Effective trip underestimated the actual
amount of fishing pressure, but it was used be-
cause catch report forms used by the fishermen in
1948-65 provided no spaces for recording zero-
catch trips.

Zero-catch trips should be considered as effort
expended to catch fish because they include time
spent searching for schools of fish. But the rela-
tive importance of search and fishing time de-
pends on type of gear used. Gulland (1969) used
whaling as an example of a fishery where the im-
portant measure was time spent searching, the
gear being operational only for a few minutes.
The other extreme was bottom trawling, where
the important measure was time spent catching
fish with the gear on the bottom and searching

^Southwest Fisheries Center Honolulu Laboratory, National
Marine Fisheries Service, NOAA, Honolulu, HI 96812.

Manuscript accepted May 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

was minimal. Beverton and Parrish (1956)
suggested that where searching time is impor-
tant, the gear may have to be regarded as being
engaged in searching for fish but giving no catch
until a school is encountered. For pole-and-line
fishing, where much time is devoted to searching
for schools of fish, Shimada and Schaefer (1956)
used the day spent on the grounds as the basic
unit of fishing time.

Catch reports of 1965-70 were used to obtain
two indices of skipjack tuna apparent abundance:
catch per effective trip (C/ET), calculated from
data on trips with catches, and catch per day
fished (C/DF), calculated from total days fished
including zero-catch fishing days. The purpose of
this study is to determine whether a relationship
exists between C/ET and C/DF. The importance
of the relationship is that it affords a means of
converting C/ET to C/DF for 1948-64, those years
for which no data on C/DF exist but for which
good C/ET information is available. A corrected
measure of apparent abundance, derived from
standard days fished instead of standard effective
trip, is used to estimate the relative fishing inten-
sity in 1948-70.

COLLECTION OF DATA

Data on skipjack tuna catch and fishing effort
were obtained from the Hawaii State Division of
Fish and Game, which collects fish catch statis-
tics in the Hawaiian Islands. In addition, catch

FISHERY BULLETIN: VOL. 74, NO. 1

and effort data were also collected routinely at
the cannery by personnel of the Honolulu
Laboratory, National Marine Fisheries Service.
The cannery records, however, were deficient in
that they did not provide information on vessels
not returning to Kewalo Basin, where the can-
nery is located, on vessels based on neighboring
islands, or on the area of operation.

Catch Reports of 1948-64

The forms for reporting skipjack tuna catch
have been revised several times over the years.
Essentially, all the different versions used in
1948-64 had spaces for recording the date of land-
ings, the amount of skipjack tuna landed, and the
area fished. The date of landing represented an
effective trip that may have lasted from one to
several days. Because Hawaiian vessels have
limited cruising range, a trip usually lasts 1 day.
Studies of interview data collected in 1960
showed that of 329 effective trips, 315 or 96%
lasted 1 day (Uchida 1967).

Catch Reports of 1965-70

The catch report forms of 1965-70 provided
spaces for recording not only the amount of skip-
jack tuna caught and the area fished, but also the
date of each day spent on the fishing ground, a
zero catch when no fish was caught, and the
number of men aboard per trip. Each entry repre-
sented 1 day's fishing. In using data for these
years, therefore, days with catches were assumed
to be equivalent to effective trips. The sum of
days with and without catches was taken as the
total number of days fished.

Reporting of Zero-Catch Trips

Review of catch reports and cannery records for
1965-70 showed that some vessels occasionally
failed to report zero-catch fishing days. When the
number of zero-catch trips recorded in the can-
nery records exceeded that reported in the catch
reports, the difference was assumed to be the
number of unreported zero catches. Most vessels
reported more zero catches in the catch reports
than were recorded in the cannery records; pre-
sumably, trips were not recorded at the cannery
when a vessel did not return to home port. These
catch reports were assumed to be accurate.

Not all unreported zero-catch days were ac-

counted for. In a few cases, vessels failed to indi-
cate a zero catch in the catch report after an un-
successful day of fishing and also failed to return
to Kewalo Basin, site of the cannery and home
port of the Honolulu-based fleet. Then, neither
the catch report nor the cannery record showed
the effort expended.

For Honolulu-based vessels, unreported zero-
catch days in 1965-70 varied between 0.5 and
3.8% of the estimated annual number of days
fished (Table 1). Differences between reported
and estimated number of days fished were not
significant it = 1.020; df = 5;P = 0.36); therefore
the few zero-catch days that went unreported
should not seriously affect the data in this study.

Table l. — Total days fished as reported, estimated number
and percentage of zero-catch days not reported, and esti-
mated total days fished by Honolulu-based Hawaiian skipjack
tuna fishing vessels, 1965-70.

Total days fished

Estimated

zero-catch

Estimated total

as reported

days not reponea

days fished

Year

(Number)

Number

Percent

(Number)

1965

1,938

0.5

1.948

1966

1,773

2.2

1,812

1967

1,678

3.8

1,745

1968

1.923

2.1

1,965

1969

1,469

3.5

1,523

1970

1,605

3.1

1,656

SOURCES OF VARIABILITY IN
FISHING POWER AMONG VESSELS

Fishing power is usually calculated on the
basis of a physical feature of the vessel such as
gross tonnage or engine horsepower. Differences
in fishing power, however, are certainly more
complicated than a comparison of these physical
attributes. Rothschild (1972) stated that "A con-
siderable portion of the variability in fishing
power among fishing units can be attributed to
variability in skill of the fishing skipper." Fishing
skill cannot be measured easily, but its influ-
ence on the fishing power of the vessels should
be understood.

Variability in crew size from trip to trip also
complicates the comparison of fishing power
among the vessels. For example, catch reports
showed that crew size in 1970 varied between 5
and 11 men per trip. Frequently, small vessels
were fully crewed while large vessels operated
shorthanded. The result was that some of the
small vessels were outperforming the larger ones
in some years.

UCHIDA: REEVALUATION OF FISfflNG EFFORT

ANALYTICAL PROCEDURES

In the sections that follow, the procedures used
in grouping vessels and fishing areas and in
treating the data are discussed.

Classes of Vessels

The difficulties that arise from differences in
fishing power among the vessels may be reduced
by separating them into relatively homogeneous
classes, using physical features such as gross
tonnage. It is convenient, therefore, to determine
which of the physical features of the vessels is, on
the average, proportional to fishing power, and to
use it to group the vessels into classes.

In a study covering the period 1952-62, the ves-
sels were grouped into two size classes according
to their bait-carrying capacities. Class 1 vessels
had capacities up to 3,000 liters per baitwell
whereas class 2 vessels had capacities greater
than that (Uchida 1967). But the ability of class 2
vessels to catch more fish than class 1 vessels is
not necessarily a permanent characteristic. Al-
though baitwell capacity was a good measure of
fishing power in the 1952-62 study, it did not
reflect fishing power of the vessels satisfactorily
after 1962. In 1963-70, some vessels with small
bait capacities had catch rates as high as or
higher than those with larger capacities.
Reevaluation of the data showed that gross ton-
nage provided a better approximation of vessel
performance. CIET and bait capacity were corre-
lated significantly in 8 out of 11 yr in 1952-62, but
only in 2 out of 8 yr in 1963-70 (Table 2). Correla-
tion between CIET and gross tonnage, on the
other hand, was significant not only in 8 yr in
1952-62, but also in 6 yr in 1963-70. For this
study, therefore, vessels of 27 to 44 gross tons
were called class 1 and those of 45 to 77 gross tons
were called class 2. The selection of the division
point between class 1 and class 2 vessels was
based on the tendency of CIET, when plotted
against gross tonnage, to be closely grouped
among class 1 vessels for almost all the years
examined. In contrast, CIET of class 2 vessels
varied widely in most years.

The relationship of fishing power to vessel age
and to bait usage cannot be overlooked. Among 8
class 1 vessels fishing in 1963-70, only 1 was built
after World War II whereas 9 out of 12 class 2
vessels fishing in 1963-70 were built after the
war. The relative comfort and reliability of most

Table 2. — Correlation coefficients of CIET on baitwell capacity
and on gross tonnage of Hawaiian skipjack tuna fishing
vessels, 1952-70. A single asterisk denotes probabilities be-
tween 0.05 and 0.01; two asterisks denote probabilities equal
to or less than 0.01.

Correlation coefficient of

Correlation coefficient

Year

CIET on baitwell capacity

of CIET on gross tonnage

1952

0.326

0.387

1953

0,306

0.275

1954

0.602"

0.463*

1955

0.498"

0.490-

1956

0.390*

0.318

1957

0.461-

0.457*

1958

0.625"

0.678"

1959

0.721"

0.669"

1960

0.477-

0.464*

1961

0.462-

0.499*

1962

0.356

0.528*

1963

0.703"

0.757**

1964

0.403

0.596**

1965

0.368

0.327

1966

0.400

0.531*

1967

0.593-

0.521*

1968

0.434

0.529*

1969

0.382

0.516*

1970

0.510

0.447

class 2 vessels undoubtedly accentuated the rela-
tion between fishing power and tonnage by at-
tracting better captains and fishermen. Also, the
difference between vessel classes in the amount of
bait used was pronounced. Whereas class 1 ves-
sels used an average of 8.3 buckets of bait per day
fished, class 2 vessels averaged 12.3 buckets.

Each year in the Hawaiian fishery the same
few vessel captains vie for the distinction of being
captain of the "top boat." Variability in skill
among captains, therefore, complicated the com-
parison of fishing power among vessels. Further-
more, captains and crew frequently shifted from
one vessel to another, taking their fishing skills
with them. In 1965-70, for example, a minimum
of nine vessels changed captains and the transfer
of a highly regarded captain usually involved the
transfer of part of his former crew. The shifting of
personnel caused some high-producing vessels to
become low- or marginal-producers.

Fishing Areas

After the establishment of the vessel classes,
the data within each size class were then grouped
into inshore and offshore fishing areas. In the
Hawaiian fishery, the deployment of fishing effort
and the resulting catches are recorded according
to a statistical area system that was established
for Hawaiian waters by the Hawaii State Divi-
sion of Fish and Game in 1947 (Uchida 1970).
Basically, three general areas are recognized. The

FISHERY BULLETIN: VOL. 74, NO. 1

first extends from the coastline to just outside the
reef, a distance of about 4 km, and the second
extends from 4 to 37 km. Combined and called
inshore for this study, these two areas are made
up of relatively small statistical areas of unequal
sizes. It has been estimated that about 80% of the
effort and 75% of the skipjack tuna catch are con-
centrated within these areas (Uchida 1967).
Beyond 37 km is the third area, called offshore
here; the statistical divisions within it are large
and nearly equal in size.

The inshore fishing ground, restricted to waters
within 37 km of the coastline, covered roughly
69,000 km^. The offshore ground, on the other
hand, was restricted only by the range of the ves-
sels, and varied from year to year. In 1948-65, the
vessels covered 111,000 km^ in their offshore
fishing, but many distant offshore areas were vis-
ited in only 1 or 2 yr over this period. The offshore
areas visited most frequently totaled roughly
69,600 km2.

Comparison of Catch Per Effective Trip
and Catch Per Day Fished

The monthly catches of skipjack tuna in 1965-
70, separated into inshore and offshore areas
within each vessel size class, were divided by two
different units of effort. One was the number of
days with catches, which was assumed to be equiv-
alent to effective trips; and the index derived
was CIET. The other v/as the total number of
days fished, which included days of fishing with
and without catches; and the index was CIDF.
The assumption that days with catches was equiv-
alent to effective trips appears justified; Uchida
(1967) showed that 96% of the effective trips
lasted 1 day.

Figure 1 illustrates the relationship of the
monthly CIDF (Y) against CIET (X) calculated
for class 1 and class 2 vessels fishing the inshore
and offshore areas in 1965. The least squares re-
gression of y on X resulted in a close linear fit
with the regression line having an angle of 45°.

A good fit between CIET and CIDF can be ex-
pected because both indexes are small when
fishing is poor and large when fishing is good. In
Hawaiian waters, periods of high tuna apparent
abundance are characterized by the presence of
larger schools and more frequent encounters be-
tween vessels and fish schools (Uchida and
Sumida 1971).

o
a.

^ 3

5 2

: 'o™o're CLASS , VESSELS
:;3Tp"s°ho'reC^*SS 2 VESSELS

1 + 0.994 X

/o *

2 3 4 5 6

CATCH /EFFECTIVE TRIP ( METRIC TONS )

Figure l. — Relationship between catch per effective trip and
catch per day fished of Hawaiian skipjack tuna vessels, by
areas fished, January-December 1965.

Homogeneity of Data

At the outset of the study, it was decided that
one regression equation should be calculated for
each area within the size classes. The resulting
equations could then be used to estimate CIDF
from CIET for 1948-64. The decision to calculate
one equation for each area by pooling the data for
1965-70 is appropriate, because the data included
those years for which skipjack tuna catches from
Hawaiian waters were the lowest (1969) and
highest (1965) on record. Including data from
these 2 yr should provide sufficient low and high
values to determine accurately the slope and
level of each regression line.

Pooling is appropriate when the samples are
homogeneous; therefore, it was necessary to test
the hypothesis of homogeneity. Statistical testing
of the data, discussed in the following sections,
was confined to only one index, CIET, because of
the close association between CIET and CIDF.

The tests for homogeneity showed that yearly
variances of inshore CIET among class 2 vessels
differed significantly (x^ = 11.92; df = 5;P<0.05).
A plot of the yearly means and standard devia-
tions, shown in Figure 2A, indicated that they
were significantly correlated (r = 0.883; df = 22;
P<0.01). Furthermore, the distribution of CIET
was skewed because of many low and few high

UCHIDA: REEVALUATION OF FISHING EFFORT
2.0 1 : '

z
o

1.5

O 1.0

(A) UNTRANSFORMED C/ET

r = 0883 1 at =22. c<O.OI

0.5 1.0 1.5 2.0 2.5

MEAN (METRIC TON)

3.5

0.6

0.3

Q 01
K

o INSHORE
• OFFSHORE

i INSHORE
» OFFSHORE

CLASS I VESSELS

CLASS 2 VESSELS

o g

(B) LOG-TRANSFORMED DATA BEFORE ELIMINATION

r ! -0 458 . dt = 22, p<0 05

0.2

0.1

-0.1

•

A
A

•

fig
° ° o

. '

f ^ 0 085. d( = 22 , p>0 05

0.1

0.2
LOG MEAN

Figure 2. — Relationship between mean and standard devia-
tion of catch per effective trip, before and after logarithmic
transformation and elimination, by vessel size classes and
areas, 1965-70.

values. Because the application of routine statis-
tical procedures requires a normal distribution
and independence of the mean and standard de-
viation, a transformation of the data was re-
quired. A logarithmic transformation was
selected because the standard deviations tended
to be proportional to their means (Figure 2A).

Transformation of the Data

A logarithmic transformation has several
theoretical advantages in analyzing catch data
(Murphy and ElHott 1954; Gulland 1956). Usually
the transformation tends to stabilize the var-
iances and make them independent of the mean.
Furthermore, the random components tend to be
independently and normally distributed about
zero mean and with a common variance.

After the transformation, the means and stan-
dard deviations continued to be significantly but
negatively correlated (r = -0.458; df = 22;
P<0.05). Examination of the transformed data
revealed that there were two points (Figure 2B)
that were aberrant and diverged from the cluster
of other points. These points represented data for
class 1 vessels fishing offshore in 1969 and in-
shore in 1970. The original monthly catch data
showed that the catch rates were affected by very
low C/ET, all of which were 0.15 MT (metric ton)
or less. These catch rates fell close to or beyond
IJL±3cr and their elimination from subsequent
analysis reduced the correlation between the
means and standard deviation (Figure 2C) and
stabilized the variances (r = 0.058; df = 22;
P>0.05). Tests for homogeneity of variances also
indicated that the transformed data for all years
could now be grouped by areas within size classes.

Figure 3 shows the frequency distribution and
fitted normal curve of the deviations from the
mean of log C/ET for each area within the size
classes. None of the histograms departed sig-
nificantly from normality when chi-square tests
were applied. Therefore, the fit of the normal
curve is as good as can be expected (x^ ranged
from 2.18 to 7.59; P<0.05).

20-

INSHORE

— ' — • — I — I — t — r

OFFSHORE

X'= = 2I8
DF=3
P = 054

1 1 — I — r

CLASS I

-0.8 -04 0 +04 +08

>-
u

UJ
IT

CLASS 2

0 +04 +0 8 -08 -04 0 +04 +08

DEVIATION FROM MEAN LOG C/ET

Figure 3. — Frequency distribution and fitted normal curve
of the deviations from the mean of log C/ET.

FISHERY BULLETIN: VOL. 74, NO. 1

Differences in Log Catch Per Effective

Trip Between Vessel Classes,

Between Areas, and Among Years

A factorial analysis of variance in a ran-
domized complete-block design was used to test
whether significant differences occurred in log
CIET between vessel classes (blocks), and be-
tween areas and among years (main treatment
effects). The analysis showed that log CIET with
respect to the two vessel classes differed sig-
nificantly {F = 12.34; df = 1 and 265; P<0.01).
Significant differences in log CIET also occurred
with respect to inshore and offshore areas fished
{F = 9.38; df = 1 and 5;P<0.05). Furthermore, the
results showed significant differences occurred
among years fished {F = 9.45; df = 5 and 5;
P<0.05). A Duncan multiple-range test (Steel
and Torrie 1960), wath Kramer's (1956) extension
of the test, determined that a significant differ-
ence in the means occurred primarily between
1965 and 1969, years in which there were consid-
erable differences in fishing conditions.

Relation Between Log Catch Per Day
Fished and Log Catch Per Effective Trip

Log CIDF increased linearly with log CIET in
each of the areas within the size classes. Regres-
sion lines, fitted to the data pooled for 1965-70,
showed that the scatter about the regression lines
was relatively narrow; there were, however, a few
observations in each set of data that appeared to
have large residuals. To assess the validity or ap-
propriateness of the least-squares fitting of log
CIDF on log CIET, these residuals were analyzed.

Figure 4 shows the scatter diagrams in which
the residuals were plotted against log CIET for
the four sets of data. With the exception of a few
outliers which can be seen as isolated points with
extreme negative ordinates, there were no
noticeable peculiarities in the distribution of the
residuals. The outliers were rejected at a multiple
of the standard deviation using a premium of
2.5% (see Anscombe and Tukey 1963). The
overall distribution of the residuals after the
rejection procedure appeared in the form of a
horizontal band, which indicated that the least-
squares analysis of the log transformed data was
satisfactory.

After the rejection of large residuals, regres-
sion lines were fitted to the data as shown in Fig-
ure 5. The dashed lines on either side of the re-

+0 3

+0.2

+0.1

-0.1

-02

-03
+0 3

+0 2

+0 1

3 -0 2
I

CLASS I (INSHORE)

REJECTED

I I I I I I I L

I 1 1 I

e
o

o o

« s

1 ) I I 1 I I I
CLASS 1 (OFFSHORE)

o
o

"■rejected

111

5 ^^y'°'

1 1 [ 1 1 1 1 1

1 I 1

1 I 1 I 1 I I I

CLASS 2 (INSHORE)

o •

°<>°oo

f «^°*o °%°a° r. o o

o o ao

ft « **

1 *^R?JECTED

I 1 1

1 1

CLASS 2 (OFFSHORE)

o o

°8

° °°°»>i;V;*°% „= „

rf. o"*" „ ° 0 - o °

oo o o

»■>

1 1 1

.•-REJECTED 1,1,

-03

-0 4
g +03

- +0 2

< +0.1
o

"= -0 I h

-02

-03
+03

+0 2

+0 I

-0.1

-0.2

-0.3

-QA

-0.4 -02 0 +0 2 +0 4 +0 6 +0 8

LOG C/ET

FIGURE 4.— Plots of residuals (log C/DF - log CrDF) against
log C/ET for class 1 and class 2 vessels fishing inshore and
offshore in 1965-70.

gression lines indicate the 95% confidence limits
for the estimates of log CIDF. The values of the
regression equation and correlation coefficient of
log CIDF on log CIET are given in Table 3.

Substitution of values of log a and b into the
logarithmic equation logioC/DF = logioa +
blogioC lET and solution of the equation provided
estimates of CIDF from CIET, by month, for

Table 3. — Data on the regression and correlation of
\ogioC/DF on logioC/£r in the Hawaiian skipjack tuna fishery,
by vessel size classes and areas, 1965-70. Two asterisks
denote probabilities equal to or less than 0.01.

Vessel
size
class

Area

Log „a

Inshore

-0.11566

1.13915

0.963"

Offshore

-0.12549

1 .08370

0.954"

Inshore

-0.10342

1.13340

0.976"

Offshore

-0.12268

1.13120

0.968"

UCHIDA: REEVALUATION OF FISHING EFFORT

INSHORE

OFFSHORE

08-

07-

0.6-

0.5-

04-

03-

02-

u. 0 1-

■V

<-> 0-

-I -0 I -
-0.2-
-0.3-
-0.4-
-0.5-
-0 6-
-0.7-

T 1 1 r

CLASS I

NOT INCLUDED
IN REGRESSION

-08
09

0.8

0.7

0.6

0.4

0.3

0.2

u. 0 1

o 0

-" -0.1
-0 2
-0.3-
-0.4
-0.5-
-0 6
-0.7-
-0 8

J L

^NOT INCLUDED
IN REGRESSION

-0.4 -0.3 -02 -0 1 0 0 1 02 03 04 05 06 07 08 09 -0.4 -0.3 -02 -0 1 0 0 1 0.2 0 3 04 0.5 0.6 07 0.8 09

CLASS 2

NOT INCLUDED
'IN REGRESSION

J 1 L

NOT INCLUDED
IN REGRESSION

J I I I I I I L-

I I

-0 4 -0 3 -0.2 -0 1 0 0 1 0 2 0.3 0.4 0.5 0.6 0.7 0 8 0.9 -0.4 -0.3 -0.2 -0.1 0 0 1 0 2 0 3 04 0 5 0 6 07 0.8

LOG C/ET LOG C/ET

Figure 5. — Regression of log C/DF on log C/ET for class 1 and class 2 vessels fishing inshore and oflfshore in 1965-70.

0.9

FISHERY BULLETIN: VOL. 74, NO. 1

Table 4. — Estimating the number of days fished among class 1 vessels fishing in the

inshore area, January-December 1948.

Effective

Calculated

Estimated

Catch

trips

CIET

CIDF

days fished

Month

(MT)

(No.)

(MT)

Log,oC/£7

Log,oC/Df

(MT)

(No.)

January

205.48

2.66857

0.42627

0.36993

2.34388

February

108.87

1.49137

0.17358

0.08207

1.20803

March

59.33

0.82403

-0.08405

-0.21141

0.61458

April

76.91

0,77687

-0.10965

-0.24057

0.57468

134

May

133.94

119

1.12555

0.05136

-0.05714

0,87669

153

June

285.80

154

1.85584

0.26854

0.19024

1.54970

184

July

352.30

147

2.39660

0.37959

0.31675

2.07374

170

August

239.72

120

1 99767

0.30052

0.22668

1.68531

142

September

191.07

104

1.83721

0.26415

0.18525

1.53199

125

October

101.31

1.25074

0.09716

-0.00497

0.98861

102

November

49 59

1 1 2704

0.05194

-0,05649

0,87802

December

19 26

0.77040

-0.11328

-0.24470

0.56923

Total

1,823.58

1,115

1,374

1948-64. For example, Table 4 shows the data
used in the computations and the results obtained
among class 1 vessels fishing the inshore area in
1948. CIET was derived from the equation,

CIET (col. 3)

Monthly catch (col. 1)
Number of effective trips (col. 2)

and converted to logarithms (col. 4). Log CIDF
(col. 5) was derived from the equation,

log CIDF = log a + 6 log CIET

and converted to CIDF (col. 6). Days fished were
estimated from the equation.

Days fished (col. 7) =

Monthly catch (col. 1)
CIDF (col. 6)

Standardization of Catch Per Day Fished

A method of standardizing effort of different
size classes of vessel has been discussed by
Shimada and Schaefer (1956) for the eastern
Pacific yellowfin and skipjack tuna fishery. I used
a similar method to estimate relative fishing
power of class 1 vessels in the Hawaiian fishery so
that their unit of effort was comparable to that of
class 2 vessels, which were selected as the stan-
dard size class (Uchida 1966, 1967). Briefly, the
method involves the use of correction or efficiency
factors that are calculated from CIDF of the ves-
sel size classes. Efficiency factors adjust the
fishing effort of one size class to that of a standard
class. For example, under conditions of equal
abundance, the class 1 vessels can be expected to
produce a smaller catch than the class 2 vessels.
From the catches of the two classes, the fishing
power of class 1 vessels can be determined rela-

tive to class 2, the standard class, for a given
fishing area.

To illustrate the calculation of efficiency factors
and the standard unit of effort, the annual CIDF
given in Table 5 by vessel size classes and areas
were used. In 1948, the efficiency factor for class 1
vessels fishing inshore was 1.33/1.78 = 0.747 and
for offshore was 2.07/3.46 = 0.598. The efficiency
factors for class 2 vessels were fixed at 1.000 for
all years. The mean efficiency factor, 0.668, is the
geometric mean of the inshore and offshore val-
ues. The geometric mean is appropriate for av-
eraging ratios.

Varying from 0.59 to 0.82 (rounded) and av-
eraging 0.71 in 1948-70, the efficiency factors
demonstrated not only the greater capability of
class 2 vessels, but also the wide variability of the
factors from year to year. There was no evidence
that the efficiency of class 1 vessels increased or
decreased relative to class 2 vessels. Therefore,
neither the efficiency of the standard class nor
that of class 1 vessels has been altered by the loss
of the less efficient or marginal vessels.

MEASURES OF APPARENT

ABUNDANCE
AND FISHING INTENSITY

Estimate of the apparent abundance of skipjack
tuna on the fishing grounds, expressed as catch
per standard day fished (CISDF), can be calcu-
lated from efficiency factors and the total number
of days fished for each of the two classes of ves-
sels. For example, in 1948 there were an esti-
mated 1,444 fishing days among class 1 vessels
and 829 days among class 2 vessels. The standard
days fished is the sum of the products of the mean
efficiency factor and the total number of fishing
days of the size classes. CISDF is found by,

UCHIDA: REEVALUATION OF FISHING EFFORT

Table 5. — Catch per day fished inshore and offshore among class 1 and class 2
vessels, class 1 efficiency factors, and their geometric mean, 1948-70.

Inshore

Offsfiore

Efficiency

Efficiency Geometric

Year

Class 1

Class 2

factors

Class 1

Class 2

factors

mean

1948

1.33

1.78

0.747

2.07

3.46

0.598

0.668

1949

1.56

2.24

0.696

2.54

4.12

0.616

0.655

1950

1.34

1.74

0.770

2.10

3.38

0.621

0.692

1951

1.64

2.59

0.633

2.60

3.58

0.726

0.678

1952

1.31

1.66

0.789

1.31

2 19

0.598

0.687

1953

1.53

1.98

0.773

2.37

2.69

0.881

0.825

1954

1.36

2.54

0.535

2.89

3.80

0.760

0.638

1955

1.39

1,99

0.698

2.08

2.32

0.896

0.791

1956

1.90

2.36

0.805

2.30

3.27

0.703

0.752

1957

1.18

1 63

0.724

1.28

1.61

0.795

0.759

1958

1.17

1.87

0.626

1.79

2.36

0.758

0.689

1959

1.97

3.03

0.650

2.37

2.91

0.814

0.728

1960

1.32

2.02

0.653

1.94

2.40

0.803

0.727

1961

1.82

2.37

0.768

2.42

4.05

0.598

0.677

1962

1.49

2,45

0.608

2.22

3.43

0.647

0.627

1963

1.17

1.77

0.661

1.87

3.55

0.527

0.590

1964

1.40

1.69

0 828

2.07

2.90

0.714

0.769

1965

2.39

2.90

0.824

3.32

4.01

0.828

0.826

1966

1.54

1.82

0.846

1.93

2.91

0.663

0.749

1967

1.47

1.84

0.799

1.65

2.31

0.714

0.755

1968

1.57

1.68

0.934

2.04

2.93

0.696

0.807

1969

1.12

1.43

0.783

1.58

2.26

0.699

0.740

1970

1.32

1.74

0.759

1.30

2.36

0.551

0.646

r'/Qnj?

TCi + TC

fished

an av(

erage

of 86.1 da

^'^^^ - (EF) iDF,) +

DF2

1948-58 when their numbers d(

where TCi = total catch of class 1 vessels,
TC2 = total catch of class 2 vessels,
EF = efficiency factor,
DFi = days fished among class 1 vessels,

and
DF2 = days fished among class 2 vessels.

In 1948-70, C/SDF of skipjack tuna in Ha-
waiian waters ranged from a low of 1.61 MT in
1957 to a high of 3.29 MT in 1965, but no trend
with time was discernible (Table 6; Figure 6).

Relative fishing intensity is estimated from
C/SDF and the total state catch, which includes
catches of part-time as well as full-time vessels:

Relative fishing intensity —

C/SDF

where TC^ = total state catch.

When examined over the 23-yr period, fishing
intensity did not decrease appreciably despite a
gradual decrease in the number of vessels fishing
from a maximum of 28 in 1951 to 15 in 1970.
With a reduction in the fleet, which occurred
primarily among the older class 1 vessels, fishing
intensity would be expected to decline, but it did
not. The reason was that the average days fished
per vessel per year increased. Class 1 vessels

10 vessels and 121.2 days in 1959-70 when their
numbers further decreased from 8 to 4 vessels
(Figure 7). Class 2 vessels have not decreased in
number drastically, declining from 14 in 1955 to

11 in 1970. Averaging 86.9 days fished prior to
1964, class 2 vessels subsequently averaged 119.8
days per year.

INTERRELATION OF TOTAL
CATCH, FISHING INTENSITY,
AND APPARENT ABUNDANCE

The total catch of skipjack tuna, given in Table
6 and shown in Figure 6, fluctuated with C/SDF
in a similar fashion in 1948-70 (r = 0.902; df =
21; P<0.01). For the years studied, then, total
catch may be satisfactory as a gross index of
changing apparent abundance but may not be
suitable in future years because it is obviously
sensitive to changes in demand or fishing effort,
competition from other fisheries, and economic
constraints upon the fishery.

Changes in C/SDF are not associated with
changes in fishing intensity (r = 0.302; df = 21;
P>0.05); therefore, the apparent abundance of
skipjack tuna in Hawaiian waters is not
influenced by changes in the amount of fishing
effort expended, but by fishery-independent fac-
tors such as variations in availability, which in
turn is related to changes in the fishes' habits or

FISHERY BULLETIN: VOL. 74, NO. 1

Table 6. — Total landings in metric tons (MT) of skipjack tuna in Hawaii, catch per stan-
dard day fished, relative fishing intensity, catch per standard effective trip, and relative
effective fishing intensity, 1948-70.

Catch per

Relative

Catch per

Relative

standard

fishing

standard

effective fishing

Total catch

day fished

intensity

effective trip

intensity

Year

(MT)

(Class 2 days)

(MT)

(Class 2 trips)

1948

3,802.96

2.01

1,891

2.30

1,653

1949

4.488.23

2.53

1,773

2.85

1.575

1950

4,314.38

1.99

2,161

2.31

1,868

1951

5,863.37

2.93

2,001

3.28

1,788

1952

3,307.58

1.83

1,806

2.15

1,538

1953

5.470.15

2.14

2,552

2.46

2,224

1954

6,360.13

2.81

2,256

3.16

2.013

1955

4,397,43

1.95

2,248

2.26

1,946

1956

5,049.58

2.59

1,946

2.91

1,735

1957

2,780.66

1.61

1,726

1.90

1,464

1958

3,100.15

1.87

1,652

2.18

1,422

1959

5,630 65

2.93

1,919

3.26

1,727

1960

3,338.46

1.99

1.673

2.30

1,452

1961

4,941.66

2.69

1,835

3.01

1,642

1962

4,270.81

2.56

1,665

2.88

1,483

1963

3,67386

2.15

1,712

2.48

1,481

1964

4,093.10

1.98

2,065

2.29

1,787

1965

7,328.96

3.29

2,221

3.54

2,070

1966

4,256.82

2.24

1,896

2.52

1,689

1967

3,646.80

1.99

1,832

2.30

1,586

1968

4,227.41

2.04

2,067

2.32

1,822

1969

2,704,94

1.63

1,658

2.02

1,339

1970

3.334.46

1.89

1,760

2.19

1,523

z
o

2 t-

o £

to Q.

in Q

=) ir

O o

2 2

-^ — I — I — I — i — I — 1 — I — I — 1 — I — \ — r

TOTAL CATCH

CATCH/STANDARD DAY FISHED

RELATIVE FISHING INTENSITY

2 6

10
Q.

z
<
<o

z>
o

22 f
W

I-
z

-20

I 8

J I 1 1 L.

1950

1955

I960
YEAR

1965

1970

z
I

Figure 6. — Total catch, catch per standard day fished, and the
relative fishing intensity for skipjack tuna in Hawaii, 1948-70.

in the environment, and to the strength of the
year classes.

Catch per standard effective trip {CI SET) and
relative effective fishing intensity, the two indi-
ces used in previous studies (Uchida 1966, 1967,
1970), are also given in Table 6. As expected, both
CISDF and CISET fluctuated similarly in 1948-
70 (r = 0.998; df = 21; P<0.01). Likewise the

correlation between relative fishing intensity and
relative effective fishing intensity was sig-
nificant, indicating that changes in one paral-
leled changes in the other (r = 0.982; df = 21;
P<0.01). It can be concluded that although the
use of effective trips in previous studies produced
biased results, which deviated from more precise
estimates calculated from days fished, its use did

160

n — \ — \ — \ — I — I — \ — r

CLASS 2 VESSELS

_J \ 1 L_

1950 1952 1954 1956 1958 I960 1962 1964 1966 1968 1970
YEAR

Figure 7. — Average number of days fished per vessel per year
among class 1 and class 2 Hawaiian skipjack tuna vessels,
1948-70.

UCHIDA: REEVALUATION OF FISHING EFFORT

not lead to faulty conclusions about the status of
the Hawaiian skipjack tuna fishery. The only
serious bias appears to be that fluctuations in the
CI SET were slightly exaggerated and those in ef-
fective fishing intensity were dampened.

SUMMARY

The existence of a linear relationship between
catch per effective trip and catch per day fished in
1965-70 was described. Based on this relation-
ship, catch per day fished was estimated from
catch per effective trip for 1948-64.

Efficiency factors were used to standardize
fishing effort of class 1 vessels to that of class 2.
The data showed that in 1948-70, efficiency fac-
tors for class 1 vessels remained constant relative
to class 2 vessels. Fishing intensity, calculated in
standard days fished, did not decline over the
23-yr period despite the gradual decrease in the
number of vessels fishing. Data from the catch
reports showed that in the face of this decline in
fleet size, the remaining vessels increased effort
by fishing more frequently.

Total catch correlated significantly with
C/SDF; therefore, it was a good gross indicator of
skipjack tuna apparent abundance. Evidence
supported the conclusion that in Hawaiian wa-
ters, skipjack tuna apparent abundance was not
influenced by changes in the amount of fishing
effort expended but by fishery-independent fac-
tors. And although effective trips as a measure of
fishing pressure in previous studies underesti-
mated effort and, therefore, provided a biased
estimate of skipjack tuna apparent abundance
in the Hawaiian fishery, its use did not lead to
faulty conclusions.

ACKNOWLEDGMENTS

I am indebted to Kenji Ego and Tamotsu
Shimizu of the Hawaii State Division of Fish and
Game for their time and effort in designing and
issuing the revised catch report forms of 1964
from which the basic data for this study were ob-
tained. Thanks go also to William H. Lenarz,
Gene R. Huntsman, and William Nicholson for
reading the manuscript and offering valuable
suggestions for its improvement.

LITERATURE CITED

ANSCOMBE, F. J., AND J. W. TUKEY.

1963. The examination and analysis of residuals. Tech-
nometrics 5:141-160.
BEVERTON, R. J. H., AND B. B. PARRISH.

1956. Commercial statistics in fish population studies.
Rapp. P.-V. Reun. Cons. Perm. Int. Explor. Mer 140
(Part I): 58-66.
GULLAND, J. A.

1956. On the fishing effort in English demersal fisheries.
Fish. Invest. Minist. Agric. Fish. Food (G.B.), Ser. II,
20(5), 41 p.

1969. Manual of methods for fish stock assessment.
Part 1. Fish population analysis. FAO (Food
Agric. Organ. U.N.) Man. Fish. Sci. 4, 154 p.

KRAMER, C. Y.

1956. Extension of multiple range tests to group means
with unequal numbers of replications. Biometrics
12:307-310.
MURPHY, G. I., AND K. C. ELLIOTT.

1954. Variability of longline catches of yellowfin tuna.
U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 119, 30 p.
Rothschild, B. J.

1972. An exposition on the definition of fishing effort.
Fish. Bull., U.S. 70:671-679.
Shimada, b. M., and M. B. SCHAEFER

1956. A study of changes in fishing effort, abundance,
and yield for yellowfin and skipjack tuna in the eastern
tropical Pacific Ocean. Inter-Am. Trop. Tima Comm.
Bull. 1:351-469.
Shippen, H. H.

1961. Distribution and abundance of skipjack in the
Hawaiian fishery, 1952-53. U.S. Fish Wildl. Serv., Fish.
Bull. 61:281-300.
STEEL, R. G. D., and J. H. TORRIE.

1960. Principles and procedures of statistics: With spe-
cial reference to the biological sciences. McGraw-Hill,
N.Y., 481 p.
UCHIDA, R. N.

1966. The skipjack tuna fishery in Hawaii. In T. A.
Manar (editor). Proceedings, Governor's Conference
on Central Pacific Fishery Resources, State of Hawaii,
p. 147-159.

1967. Catch and estimates of fishing effort and apparent
abundance in the fishery for skipjack tuna (Katsuwonus
pelamis) in Hawaiian waters, 1952-62. U.S. Fish Wildl.
Serv., Fish. Bull. 66:181-194.

1970. Distribution of fishing effort and catches of skip-
jack tuna, Katsuwonus pelamis, in Hawaiian waters, by
quarters of the year, 1948-65. U.S. Fish Wildl. Serv.,
Spec. Sci. Rep. Fish. 615, 37 p.

UCHIDA, R. N., AND R. F. SUMIDA.

1971. Analysis of the operations of seven Hawaiian
skipjack tuna fishing vessels, June-August 1967. U.S.
Dep. Commer., Natl. Mar. Fish. Serv., Spec. Sci. Rep.
Fish. 629, 25 p.

Yamashita, D. T.

1958. Analysis of catch statistics of the Hawaiian skip-
jack fishery. U.S. Fish Wildl. Serv., Fish. Bull.
58:253-278.

SEASONAL AND INSHORE-OFFSHORE VARIATIONS IN THE

STANDING STOCKS OF MICRONEKTON AND

MACROZOOPLANKTON OFF OREGON

William G. Pearcy^

ABSTRACT

Dry weights of pelagic animals captured along an inshore-offshore station line with Isaacs-Kidd
mid-water trawls and 1-m diameter plankton nets during a 5-yr period provided evidence for seasonal
changes in the standing stocks of carnivores. Micronekton catches (fishes, shrimps, and squids) were
largest inshore (28 and 46 km offshore) in the winter (November-April), and offshore (84 and 120 km)
during the summer (May-October), the season of coastal upwelling. No seasonal difference was
detected in the biomass of herbivores, or in its primary components, the copepods and euphausiids.
Increased biomass of medusae during the summer resulted in significant seasonal differences in the
planktonic carnivores at the inshore stations.

The average biomass (grams per square meter) of small nektonic and planktonic carnivores,
averaged over the year, peaked at the 84-km station. The biomass of fishes was greater than shrimps
and the biomass of shrimps was greater than that of squids at all stations, except 46 km where shrimps
predominated. Herbivore biomass was maximal at 46 km, over the inner continental slope, largely
because of the high catches of euphausiids at this station. The occurrence of largest average catches at
intermediate distances from shore, and inshore-offshore shifts in peak biomass with seasons, may
result from seasonal changes in upwelling and downwelling and exclusion of vertical migrants
from shoal waters on the shelf

Herbivore: carnivore biomass ratios differed significantly between inshore and offshore stations.
Standing stocks of herbivores were several times larger than those of carnivores in nearshore waters,
but the ratio was about 1.0 in offshore waters. Coefficients of variation (s/x) of herbivore and plank-
tonic carnivore stocks for the entire sampling period were highest inshore, indicating high variabil-
ity, and decreased markedly in offshore waters. These trends suggest that, compared to offshore or
oceanic communities, the pelagic inshore-upwelling ecosystem may be less predictable and have a
lower ecological efficiency.

This research was designed to answer two ecologi-
cal questions about intermediate consumers in the
pelagic food chain off Oregon:

(1) Are seasonal variations obvious in the
standing stocks of small nekton and macrozoo-
plankton off Oregon, perhaps in response to up-
welling along the coast during the summer?

(2) Are there trends in the standing stocks of
these animals from oceanic waters into neritic
waters and, if so, do they reflect basic ecological
differences in these pelagic communities?

Pelagic animals such as fishes, squids, shrimps,
and euphausiids are ubiquitous in the open oceans
and are important intermediates in the food chain
between small plankton and large pelagic carni-
vores. Yet little is known about their seasonal
variations, inshore-offshore distributions, or gen-

'School of Oceanography, Oregon State University, Corval-
lis, OR 97331

eral ecology. The life span and generation time of
many of these intermediate consumers are 1
yr or greater, limiting short-term changes in
population sizes. Moreover, many of these ani-
mals reside below the depth of seasonal tempera-
ture change much of the time. They may under-
take diel vertical migrations, and some species
may migrate through the thermocline at night. In
any event, seasonal changes in physical environ-
ment are expected to be less pronounced than
those experienced by inhabitants of surface
waters. Thus, seasonal variations in population
size of these animals are expected to be less
than those of small planktonic organisms.

Movements of water may also affect seasonal
changes in the abundance of animals at one
locality, or spatial distributions within a general
region. In areas where water masses and as-
sociated pelagic fauna overlap and mix, species
structure may be complicated, primarily a result

Manuscript accepted April 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

70 y-^ - ^^

PEARCY: MICRONEKTON AND MACROZOOPLANKTON OFF OREGON

of advective processes rather than biological in-
teractions (McGowan 1971). In the headwater
region of the California Current off Oregon, how-
ever, the water type is predominantly Subarctic
and common species of some taxonomic groups
of pelagic animals are the same within and
among years (Pearcy 1972).

In addition to in situ population changes and
changes affected by advection, small nektonic
animals may be able to swim or to migrate
horizontally. Though migrations of large nektonic
animals such as tuna, salmon, hake, etc., are
known to result in large seasonal changes in the
abundance of these animals off Oregon, little
evidence exists for horizontal movements of mi-
cronekton, even on a reduced scale. This is
another reason to expect temporal stability of
their populations.

Basic differences in the structure and energy
pathways of neritic and oceanic ecosystems in the
northeastern Pacific have been inferred by differ-
ences in the seasonal production cycle, seasonal
variations in chlorophyll a concentrations, and
the size of individual phytoplankton and mi-
crozooplankton (McAllister et al. 1960; Anderson
1965; Parsons and LeBrasseur 1970; LeBrasseur
and Kennedy 1972). Inshore-offshore differences
in the standing stocks of pelagic herbivores and
carnivores, which have not been studied, are
therefore to be expected.

METHODS

Micronekton and macrozooplankton were col-
lected at night with 1.8-m Isaacs-Kidd mid-water
trawls (IKMT) and with 1-m diameter plankton
nets (MN) along stations west of Newport, Oreg.
(lat. 44°39.1'N). The stations were located 28, 46,
84, 120, and >120 km, respectively, offshore (Fig-
ure 1). Collections, made about every month,
totalled 243 IKMT tows between August 1962
and July 1967, and 179 MN collections between
June 1963 and July 1967.

The IKMT had a 5-mm (bar measure) nylon
liner throughout. Oblique tows were made to a
depth of approximately 200 m, except at inshore
stations where about one-half the depth of the
water column was sampled (40 m and 130 m at
the 28- and 46-km stations, respectively). Tow
speed was 6 knots. The trawl was lowered at 50 m
wire/min until a 4:1 scope was attained. The
trawl was then retrieved at 30 m wire/min to the
surface. Volume of water filtered and depth of

Figure l. — Location of the sampling stations off Newport,
Oreg. Stations are designed in kilometers from the coast.
Depth contours are in fathoms (100 fathoms = 183 m, 500
fathoms = 914 m, 1,000 fathoms = 1,829 m, 1,500 fathoms =
2,743 m).

trawling was estimated from TSK^ depth-
distance recorders and flowmeters.

The meter nets, which were made of 0.571-mm
Nitex, were towed immediately before or after
each IKMT tow. From June to November 1963
oblique tows were made to approximately the
same depths as the IKMT tows, but because of
difficulties resulting from preferential sampling
of near-surface waters, oblique tows were aban-
doned in favor of vertical tows in December 1963.
Vertical tows were from 200 m to the surface, or
from 60 or 150 m to the surface at the two inshore
stations. After a vertical wire angle was obtained,
they were retrieved at 50 m wire/min. Flow-
meters mounted in the mouth of MN's provided
estimates of volumes filtered. In a few instances
flowmeters malfunctioned. Volumes were then
estimated from the distance towed and 85%
IKMT filtration efficiency (Pearcy and Laurs
1966) or from the average volume of other MN
tows to the same depth.

^Tsurumi-Seiki Kosakusho Co. Reference to trade names
does not imply endorsement by the National Marine Fisheries
Service, NOAA.

FISHERY BULLETIN: VOL. 74, NO. 1

Samples were preserved with Formalin at sea
and sorted into taxonomic groups ashore. Wet
(drained) weights were obtained for micronekton
(fishes, shrimps, and squids). Micronekton from 32
different IKMT collections were dried to a con-
stant weight in a drying oven at 65°C. The mean
dry weight: wet weight ratios were then used to
convert wet weights of other collections to dry
weights. The means and standard deviations of
the dry:wet weight ratios were 0.23 ± 0.06 for
fishes, 0. 15 ± 0.02 for shrimps, and 0. 1 1 ± 0.04 for
squids.

Dry weights were obtained for all major taxa
sorted from MN samples: euphausiids, copepods,
chaetognaths, medusae, amphipods, salps-dolio-
lids, and shrimps. These taxa generally com-
prised over 95% of the total collection weights. The
remainder usually consisted of annelids,
pteropods, and heteropods. Ctenophores usually
disintegrated in the samples, but when fragments
were identifiable they were weighed with the
medusae. In this paper, dry weights are used as a
measure of standing stock, which is considered to
be synonymous with biomass.

Sampling Variability

Several series of IKMT's at a single station
during a single night were taken to assess sam-
pling variability. The variability of total micro-
nektonic dry weight per 1,000 m^ (Table 1) indi-
cates that the variance for these series was ap-
preciably less than the mean. These data on total
biomass of micronekton, which are not in dis-
agreement with the high variability encountered
for individual species of micronekton captured
in repeated tows at one station (e.g., Pearcy 1964;
Ebeling et al. 1970), suggest that most of the
temporal fluctuations of biomass illustrated in

Table l. — Sampling variability oftotal biomass of micronekton
and macroplankton (grams dry weight per 1,000 m^) collected
during repeated tows during separate nights.

Distance

offshore

Average

Variance

Gear

Date

(km)

tows

(s')

Mid-water

Dec 1964

2.7

0.6

trawl

Nov. 1966

120

4.7

0.9

Feb. 1967

120

1.8

0.2

Feb. 1967

120

1.5

0.02

June 1967

306

1.9

0.01

June 1967

120

2.2

0.4

Meter net

June 1964

5.0

3.1

June 1966

20.3

99.0

Nov. 1966

111

9.6

2.4

Feb. 1967

4.6

1.1

Mar. 1967

787

10.0

101.8

Figure 2 are independent of short-term sampling
variability.

Variances of macrozooplankton biomass from
repeated MN tows, on the other hand, were much
larger than those for the IKMT (Table 1). In two
out of the five series, variance surpassed the mean.
Hence, a larger portion of the temporal variability
of zooplankton can be ascribed to sampling varia-
bility.

RESULTS

Micronekton

Variations of the dry weights of micronekton
(fishes, shrimps, and squids) captured per 1,000 m^
are shown in Figure 2 for four stations, 1962-67.
Several trends are apparent. Seasonal peaks in
the biomass occur inshore at the 28- and 46-km
stations during the winter months, with very low
values during intervening months. A reversed
trend, though less pronounced, is found offshore at
the 84- and 120-km stations where maximum
catches generally were made during the summer
or fall months. Average biomass values appear to
be lowest inshore, highest at 84 km, and lower
again at 120 km where total variability is the
lowest.

The spatial peak of micronekton biomass at 84
km is more obvious in Figure 3, where dry weight
is plotted per square meter instead of per cubic
meter (to compensate for different depths of sam-
pling at inshore stations). The standing stocks of
fishes were greater than shrimps, and shrimp
stocks were greater than squids at all stations
except at 46 km where shrimps predominated. The
neritic, benthopelagic shrimp, Pandalus jordani,
occasionally made up the bulk of the biomass of
collections at both 28 and 46 km (Pearcy 1970).
However, mesopelagic animals comprised most of
the nighttime IKMT catches: mainly the fishes
Stenobrachius leucopsarus, Diaphus theta, Tar-
letonbeania crenularis, and Tactostoma macropus
(Pearcy 1964, 1972; Pearcy and Laurs 1966;
Pearcy and Mesecar 1971); the shrimp Sergestes
similis (Pearcy and Forss 1966, 1969); and the
squids Gonatus spp. and Abraliopsis felis (Pearcy
1965, 1972).

Seasonal variations in the total biomass
(grams/10 m^) of micronekton are illustrated in
Figure 4 for two general seasons: May-October,
which includes the upwelling season; and
November-April, when surface currents are usu-

PEARCY: MICRONEKTON AND MACROZOOPLANKTON OFF OREGON

A SONOIJF MAMJ JASONDIJFMAMJ JAS ONOI

1962 1963 I lae^

JFMAMJJASONOIJFMAHJ JASONDIJFMAMJ J

1965 1966 I 196/

Figure 2. — Biomass of micronekton captured in Isaacks-Kidd mid-water trawl collections at four sta-
tions, 1962-1967. Each point represents one collection. Average depth of tows was 40 m for 28-km
station, 130 m for 46-km station, and 200 m for 84- and 120-km stations.

TOTAL

DISTANCE OFFSHORE (km)

>I20

Figure 3. — Inshore-offshore variations in the average total
micronekton biomass (grams per 10 m^ ± 1 SE) and in its
component fishes, shrimps, and squids.

ally reversed, downwelling occurs, and the David-
son Current is often present along the coast
(Wyatt et al. 1972; Bakun 1973). The means and
medians of the biomass of total micronekton per 10
m^, and of its constituents — fishes, shrimps, and
squids — are given in Table 2 for these two sea-
sons, along with the probabilities that the two

e
o

>-
<r
o

MAY-OCX.

84 120

DISTANCE OFFSHORE (km)

>I20

Figure 4. — inshore-offshore variations in the biomass of mi-
cronekton during two seasons, May-October and November-
April. Shaded areas included means ± 1 SE.

seasonal values are the same. Seasonal differences
of total biomass are significant (P<0.05) at 46

FISHERY BULLETIN: VOL. 74, NO. 1

Table 2. — The mean and median biomass (grams dry weight per 10 m^) for micronekton and macro-
plankton during summer (S = May-October) and winter (W = November- April) at five stations (28,
46, 84, 120 and >120 km) off the Oregon Coast. Probabilities resulting from Mann- Whitney U and t
tests of seasonal differences are given.

Stn

1. 28 km

Stn. 46 km

Stn. 84 km

Stn. 120 km

Stn. > 120 km

Item

S W

Total micronekton

Mean

0.19

0.32

0.51 4.30

8.20 5.24

6.20

3.26

350 4.30

Median

0.004

0.03

0.38 2.75

7.84 4.04

5.04

2.76

3.28 3.18

Probabilities

S<W" P<0.01

S>W P = 0.08

S>W P-

=0.04

S<WP<0.01

S>W P<0.05

S>W"P<0.01

Probabilities

U test

Fishes

S<W"P<0.01

S>W"P=0.01

S>W* P-

=0.02

Shrimps

S<W P-0.03

Squids

S>WP<0.01

Total macroplankton

Mean

24.9

19.3

31.3 38.7

37.0 15.6

27,4

26.6

11.8 15.7

Median

12.6

12.1

9.4 8.1

12.2 6.5

8.6

4.9 5.0

Probabilities

S>W P<0.04

S >W" P<0.01

Probabilities

ftest

Copepods

Euphausllds

Salps

Medusae

S -W

•• P-

;0.01

S>W"P<0.01

S>W P = 0.06

Chaetognaths

Amphipods

Shrimps

S W P<0.05

NS - not significant.

t - too many zeros

for valid tests.

and 120 km using the non-parametric Mann-
Whitney U test (Tate and Clelland 1957 ) and at 46,
84, and 120 km using the parametric t test. Mann-
Whitney U tests for the three taxa of micronekton
indicated significant seasonal differences for
standing stocks of fishes at 46, 84, and 120 km,
for shrimps at 46 km and for squids at 84 km.

Macrozooplankton

Values for the biomass of macrozooplankton
collected at four stations during 1963-67 are
shown in Figure 5 and Table 3. Inshore-offshore
and seasonal trends are less apparent than for
micronekton. The total MN biomass per 10 m^ is
lowest at the 28-km stations, greater at the
120-km stations, and highest at the 46-, 84-, and
120-km stations (Table 3).

Of the taxonomic groups composing the MN
samples, copepods were most important on an
average dry weight basis at all stations except at
46 km where euphausiids were very abundant
(Table 3). The standing stock of medusae ranked
second after copepods at all stations except at 46
km where it ranked third after copepods. Even
though the maximum biomass of all groups oc-
curred at 46, 84, or 120 km on a square meter
basis, the maximum weights of copepods and

Table 3. — Biomass of zooplankton per 10 m^ collected with
1-m diameter nets at the stations off Newport, Oreg.

Stn.

Item

28 km

46 km

84 km

120 km :

>120km

Total biomass

Mean

Median

8.0

No. collections

Ave. sampling depth

152

200

Copepods

Mean

11.9

12.2

7.9

11.7

4.3

Median

2.4

2.0

2 1

2.5

1.5

Euphausllds

Mean

2.6

20.0

6.2

2.4

2.5

Median

0.6

3.7

2.2

1.5

1.1

Salps

Mean

0.04

0.03

3.2

4.1

1.4

Median

0.002

Medusae

Mean

5.8

2.3

6.8

6.4

3.8

Median

1.2

1,1

3.2

2.5

2.0

Chaetognaths

Mean

0 5

0.9

1.6

1.7

0.9

Median

0.07

0.6

1.0

1.1

0.7

Amphipods

Mean

0.07

0.1

0.2

0.3

Median

0.02

0.06

0.2

Shrimps

Mean

0.02

0.6

0.5

0.6

0.8

Median

0.2

0.1

medusae on a cubic meter basis were found at 28
km, nearest the coast.

Differences in the biomass of macrozooplankton
between the two seasons were only significant at
one station, 84 km offshore (Table 2), although
distinct peaks occurred during the summers of 2 yr
at 120 km (Figure 5). Surprisingly, most of the
taxonomic groups of zooplankton, including
copepods and euphausiids, evidenced no seasonal
changes at any stations. The only significant

PEARCY: MICRONEKTON AND MACROZOOPLANKTON OFF OREGON

100

I I I I I 1 I I I I T 1 1 i I i I l; I I I rr 1 1 1 I ! I I I

. . I I I I 1 1 1 vy I •;• I I I I I I I I
jjasond'jf' mamj jasond

1963

1964

JFMAMJJASONDJFMAMJJASOND

1965 ' 1966

J F M A M J J

1967

Figure 5. — Biomass of macrozooplankton captured in 1-m diameter plankton nets at four
stations, 1963-1967. Each point represents one collection.

differences were for medusae, whose standing
stocks in the summer exceeded those in the winter
at 28 and 46 km (Mann-Whitney U, P<0.01) and
perhaps at 84 km (P = 0.06), and for shrimps at 84
km, where again biomass was larger during sum-
mer than winter (Table 2).

Trophic Groups

To estimate seasonal and inshore-offshore vari-
ations in the standing stocks of the lower trophic
levels of oceanic consumers, the dry weights of the
various taxa were combined. Herbivores were
assumed to include copepods, euphausiids, and
salps-doliolids. Planktonic carnivores included
chaetognaths, medusae, amphipods, and shrimps.
Nektonic carnivores included fishes, squids, and
shrimps. Although it is recognized that some
euphausiids and copepods may be carnivorous, the
main species captured off Oregon, Euphausia
pacifica, Thysanoessa spinifera, and Calanus spp.,
are considered to be largely herbivorous.

Inshore-offshore variations in standing stocks
are illustrated in Figure 6. On the average, the
biomass of herbivores was greater than planktonic

^"£ 30
O

^25

y 15

g 10

PLANKTONIC j

CARNIVORE§-.| J._ \

NEKTONIC
CARNIVORES

— O'-

46 84 120

DISTANCE OFFSHORE (km!

>I20

Figure 6. — Inshore-offshore variations in the average biomass
(± 1 SE) of herbivores, planktonic carnivores, and nektonic
carnivores at five stations.

carnivores, and the biomass of these organisms
was greater than that of micronektonic carni-

FISHERY BULLETIN: VOL. 74, NO. 1

vores at all stations. The high catches of
herbivores at 46 km were due to abundant
concentrations of euphausiids. Both groups of
carnivores, on the other hand, had lowest bio-
mass at the inshore stations and attained maxima
farther offshore.

Seasonal variations in the standing stocks of
herbivores and planktonic carnivores are illus-
trated in Figure 7. Mann-Whitney U tests of
differences between the two seasons were not
significant (all P>0.1) for any station, providing
no evidence for seasonal changes in the biomass of
herbivores. The biomass of planktonic carnivores
increased wdth distance offshore during the winter
and tended to decrease during summer. The
biomass at 28 km was higher in summer than
winter (P<0.01), largely due to high catches of
medusae during the summer. At 84 km,

E
Q 20

5 10

20-
15-
10-

PLANKTONIC
CARNIVORES

MAY-
OCT

■-^NOV-APR

>i^

28 46 84 120 >I20

DISTANCE OFFSHORE (km)

Figure 7. — Seasonal variations in the average biomass (± 1
SE) of herbivores (upper) and planktonic carnivores (lower).

^ 20
>

y 10-

PLANKTONIC

NEKTONIC ■•■.
CARNIVORES'^

28 46 84 120

DISTANCE OFFSHORE (km)

>I20

Figure 8. — Variability in the catches of herbivore, planktonic
carnivores, and nektonic carnivores vs. distance offshore.
Variability is expressed as coefficients of variation based on
dry weights per 1,000 m^.

planktonic carnivores also appeared to be more
abundant during the summer (P = 0.08), again
because of higher catches of medusae. No seasonal
differences were apparent at other stations
(P>0.1).

The ratio of herbivore:carnivore biomass, as
expected from the data shown in Figure 6, aver-
ages about 2.0 at 28 km and 4.0 at 46 km, but
only about 1.0 at the oceanic stations 84, 120,
and >120 km. These ratios were ranked among
stations for individual cruises. The sum of the
ranks for stations were significantly different
(P<0.01, Friedman two-way ANOVA by ranks,
Tate and Clelland 1957). Thus herbivores pre-
dominated over carnivores in inshore waters,
whereas the standing stocks of herbivores and
carnivores were about equal in oceanic waters 84
km offshore and beyond. No seasonal differences
in herbivore:carnivore ratios were found (P>
0.05, Mann- Whitney U tests).

As a measure of variability of the standing
stocks of trophic groups over the sampling period,
coefficients of variation (six) of the catches are
plotted for each station in Figure 8. A marked
decline in the variability of both herbivores and
carnivores takes place from inshore into offshore
waters.

DISCUSSION

Regional Comparisons of
Zooplankton Standing Stocks

Values for the standing stocks of zooplankton in

PEARCY: MICRONEKTON AND MACROZOOPLANKTON OFF OREGON

the upper 140 to 300 m are summarized by
Gushing (1971) for upwelling regions of the
world. The average biomass of zooplankton col-
lected within 120 km of the Oregon coast (Table
4) is within the range of values given by Gush-
ing, after conversion to displacement volume per
1,000 m^ and to grams carbon per square meter.

Zooplankton standing stocks off Oregon can also
be compared with those reported by the Galifornia
Gooperative Oceanic Fisheries Investigations
(GALCOFI) which used 0.25-0.55-mm mesh in
nets towed obliquely from 140 m to the surface.
Zooplankton displacement volumes near the Ore-
gon coast accord with values of Reid et al. (1958)
and Reid (1962) greater than 400 cm3/l,000 m^ for
July and August 1955 from Point Gonception,
Galif , to northern Washington, and with Thrail-
kill's (1956) values of 100-900 cm3/l,000 m^ for
1949 and 1950 off Oregon and northern Galifor-
nia. Smith's (1971) median displacement vol-
umes for pooled areas within 100 miles of shore
between Point Gonception and San Francisco
Bay, Galif., are 200-400 cm3/l,000 m^ during
April-July 1951-60, with decreased volumes
south of Point Gonception. Median displacement
volumes for Oregon (either on an annual or a
summer basis, Tables 2 and 4) are appreciably
lower than Smith's values for northern Galifor-
nia. This difference may be ascribed to differ-
ences between vertical and oblique tows, mesh
size, or annual differences in standing stocks. Or,
a real trend may exist for the nearshore zoo-
plankton standing stocks to increase in the
Galifornia Gurrent system between Oregon and
northern Galifornia, a trend that may be attrib-
uted to the more intense upwelling — and hence
higher productivity — that occurs off northern
Galifornia (Bakun 1973).

Zooplankton volumes within 120 km of Oregon
are several times those given by McAllister

Table 4. — Dry weight of Oregon zooplankton converted to dis-
placement volumes and grams carbon.

Stn.

stn.

Item

28 km

46 km

84 km

120 km

>120km

Mean cm^/i.OOOm^*

552

450

228

274

Median cm^/i.ooo m^

160

157

140

121

Mean gC/m^t

1.1

1.8

1-3

1.3

0.7

Mean gC/m^J

2.3

3.9

2.8

2,9

1.5

"Conversion based on data of Ahlstrom and Thrailkill (1963, Table 7): wet
weight plus interstitial water ("displacement volume) « 0.06 = dry weight.

tC was estimated to be 50% of the dry weight (see Omori 1969,
Table 5).

tCalculated using Gushing s (1971) conversion of 0.065 x displacement
volume = gC. This conversion assumes that displacement volumes do not in-
clude interstitial water, but according to the data of Ahlstrom and Thrailkill
(1963, Table 7) an average of 42% of the wet weight of mixed zooplankton is
interstitial water.

(1961) and LeBrasseur (1965) for oceanic areas of
the Gulf of Alaska (0-150 m vertical tows with a
0.45-cm diameter net, 0.35-mesh), even after
their catches are adjusted for the relatively low
catching power of their net (McAllister 1969;
LeBrasseur and Kennedy 1972). Average vol-
umes at weather station "P" (lat. 50'^N, long.
145°W) were more similar to those at the station
>120 km off the Oregon coast. Increased produc-
tivity associated with coastal upwelling along
Oregon, therefore, enhances the average zoo-
plankton standing stocks out to about 120 km
from shore several times above the stocks farther
offshore or upstream in the North Pacific Drift
(see also Reid 1962). The width of this zone of
high zooplankton standing stocks appears to be
considerably less than the 200-500 km reported
by Gushing (1971) for the region off northern
Galifornia.

Seasonality of Standing Stocks

Seasonality in the biomass of zooplankton,
with maxima in the summer and minima in the
winter, has been reported in the Galifornia Gur-
rent system off central Galifornia (Lasker 1970;
Smith 1971) and in waters off the Oregon- Wash-
ington coast (Peterson 1972). Yet there was lim-
ited evidence for differences in macrozooplankton
standing stocks between the two seasons in Ore-
gon waters. Thus seasonality of standing stocks
appeared to be more pronounced for micronekton
than macrozooplankton, or for carnivores than
herbivores. This may be because the high vari-
ability of macrozooplankton catches (Figure 8)
makes important seasonal changes difficult to
detect. Also the months selected for the two
seasons may not match the periodicity of natural
cycles. Another possible explanation is that the
seasonality in catches of common animals such as
Euphausia pacifica and Calanus spp. may be less
than that in small herbivores with shorter life
spans and generation times. Small copepods such
as Pseudocalanus, Oithona, and Acartia, which
were not sampled adequately with my nets, are
known to be abundant in Oregon-Washington
waters in the summer, especially in upwelled
waters along the coast (Frolander 1962; Gross
1964; Peterson 1972; Peterson and Miller 1975).

Inshore-Offshore Variations

Largest standing stocks of macrozooplankton
and micronekton (grams per square meter but not

FISHERY BULLETIN: VOL. 74, NO. 1

grams/1,000 m^, Tables 3 and 4, Figure 6) were
found intermediate distances off the Oregon
coast, namely over the continental slope at sta-
tions 46 and 84 km offshore. A trend for maxima
at intermediate distances offshore has been re-
ported for other regions. Standing stocks of zoo-
plankton were highest at the edge of the shelf or
over the inner slope off New York (Grice and Hart
1962), intermediate distances from shore off
California (Smith 1971), and near mid-shelf in
the Florida Current off Cape Hatteras, N.C. (St.
John 1958). Macrozooplankton and micronekton
collected with a 0.9-m IKMT off Vancouver Is-
land, Canada and Washington were maximal
over the outer edge of the shelf (Day 1971). The
reduced feeding activity of pink, chum, and sock-
eye salmon as they approach the coast is pur-
portedly explained by the low macroplankton
concentrations in neritic waters and higher con-
centrations in offshore waters of the northwest-
ern Pacific off Kamchatka (Andrievskaya 1957;
Mednikov 1958). All of these studies indicate that
small intermediate consumers may achieve
maximum importance in the pelagic food chain in
deep waters beyond the inner shelf (see also
Wilhams et al. 1968).

The reason why catches of micronekton and
macrozooplankton were higher offshore than
nearshore may be related to their vertical migra-
tions. Most of the species of micronekton and
euphausiids caught in upper waters at night
undertake diel vertical migrations (Pearcy and
Laurs 1966; Pearcy and Forss 1966; Brinton
1967; Pearcy and Mesecar 1971); hence they may
be most abundant in waters deep enough to
permit vertical movements but where productiv-
ity is enhanced near land (Pearcy 1964). If they
drift over the shelf, they may be eaten by large
benthic or pelagic predators (Isaacs and
Schwartzlose 1965; Pereyra et al. 1969).

The inshore-offshore changes in standing
stocks of micronekton for the two seasons (Figure
4) suggest that these distributions are interre-
lated. Movement of animals may be correlated
with seasonal oceanographic changes. During the
summer, when the biomass increases greatly
from 46 km to a peak at 84 km, large inshore-
offshore gradients also occur in physical proper-
ties because of upwelling, and there is an offshore
component of nearshore surface waters (Pillsbury
1972). During the winter, when biomass from
46 to >120 km is relatively uniform, inshore-
offshore gradients are weak, surface currents are

onshore, and downwelling occurs (Hebard 1966;
Laurs 1967). The significant increase in biomass
at 46 km in the winter may be caused by inshore
advection of surface water and animals and the
concentrating effect of shallow water near the
edge of the shelf on vertical migrants. The peak
at 84 km in the summer, though far from the
coast, may be related to upwelling. Sometimes
Laurs (1967) found maximum biomass of carni-
vores at 65-84 km and maxima of lower trophic
levels closer inshore off Brookings, Oreg., during
the summer, suggesting a succession of trophic
level maxima such as reported by Sette (1955),
King ( 1958), and Vinogradov and Voronina ( 1962)
in areas of oceanic upwelling in equatorial
waters.

Herbivore: Carnivore Ratios

Others have also found that the herbivore:
carnivore biomass ratios decrease from shal-
low, eutrophic waters to oceanic waters. Grice and
Hart (1962) reported that well over one-half of
the zooplankton by volume in shelf waters off
New York herbivorous, while in the Sargasso Sea
only about one-half belonged to this trophic level.
The percentage of herbivores in the zooplankton
catches decreased from inshore waters that were
affected by upwelling into offshore waters of the
California Current off Baja California (Longhurst
1967). Greze ( 1970) reported that the biomass and
production of herbivores and carnivores was a
larger percentage of that of primary producers in
the Equatorial Atlantic or Ionian Sea than the
shallow waters of the Black Sea or Sevastopol
Bay. These trends suggest that (a) a smaller
fraction of the herbivorous biomass is captured
in oceanic than neritic waters because of escape-
ment through coarse mesh or avoidance, (b) pro-
duction per unit biomass of herbivores is higher
relative to that of carnivores in offshore waters,
or (c) that ecological efficiences (food consumed by
tropic level n + 1 to food consumed by trophic
level n ) are higher in oceanic than neritic waters.

ACKNOWLEDGMENTS

This research was supported by the National
Science Foundation (Grant GB-1588) and the
Office of Naval Research (Contract NOOO 14-67-
A-0369-0007 under project NR 083-102). I am
grateful to Harriet Lorz, Henry Donaldson, Lyle
Hubbard and others who helped with the field

PEARCY: MICRONEKTON AND MACROZOOPLANKTON OFF OREGON

and laboratory work. Charles B. Miller and Law-
rence F. Small made helpful comments on the
manuscript.

LITERATURE CITED

AHLSTROM, E. H., AND J. R. THRAILKILL.

1963. Plankton volume loss with time of preservation.
Calif. Coop. Oceanic Fish. Invest. Rep. 9:57-73.

Anderson, G. C.

1965. Fractionation of phytoplankton communities off
the Washington and Oregon coasts. Limnol. Oceanogr.
10:477-480.

ANDRIEVSKAYA. L. D.

1957. The food of Pacific salmon in the northwestern
Pacific Ocean. Vses. Nauchno-Issled. Inst. Morsk.
Rybn. Khoz. Okeanogr. (VNIRO), Mosc, p. 64-75. (Fish.
Res. Board Can., Transl. Ser. 182, 12 p., 1958).

Bakun, a.

1973. Coastal upwelling indices, west coast of North
America, 1946-71. U.S. Dep. Commer., NOAA Tech. Rep.
NMFS SSRF-671, 103 p.
Brinton, E.

1967. Vertical migration and avoidance capability of
euphausiids in the California Current. Limnol. Ocean-
ogr. 12:451-483.

Cross, f. a.

1964. Seasonal and geographic distribution of pelagic
copepods in Oregon waters. M.S. Thesis, Oregon
State Univ., Corvallis, 71 p.

CUSHING, D. H.

1971. Upwelling and the production of fish. Adv. Mar.
Biol. 9:255-334.
DAY, D. S.

1971. Macrozooplankton and small nekton in the coastal
waters off Vancouver Island (Canada) and Washington,
spring and fall of 1963. U.S. Dep. Commer, Natl. Mar.
Fish. Serv., Spec. Sci. Rep. Fish. 619, 94 p.
EBELING. a. W, R. M. IBARA, R. J. LAVENBURG, and F. J. ROHLF
1970. Ecological groups of deep-sea animals off Southern
California. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 6,
43 p.
FROLANDER H. F.

1962. Quantitative estimations of temporal variations of
zooplankton off the coast of Washington and British
Columbia. J. Fish. Res. Board Can. 19:657-675.
GREZE, V. N.

1970. The biomass and production of different trophic
levels in the pelagic communities of south seas. In
J. H. Steele (editor). Marine food chains, p. 458-467.
Univ. Calif. Press, Berkley and Los Ang.

Grice, G. d., and a. d. Hart.

1962. The abundance, seasonal occurrence and distribu-
tion of epizooplankton between New York and Bermuda.
Ecol. Monogr. 32:287-309
HEBARD, J. F.

1966. Distribution of Euphausiacea and Copepoda off
Oregon in relation to oceanographic conditions. Ph.D.
Thesis, Oregon State Univ., Corvallis, 85 p.

ISAACS, J. D., AND R. A. SCHWARTZLOSE.

1965. Migrant sound scatterers: Interaction with the sea
floor Science (Wash., D.C.) 150:1810-1813.

KING, J. E.

1958. Variation in abundance of zooplankton and forage
organisms in the central Pacific in respect to the equa-
torial upwelling. Proc. Ninth Pac. Sci. Congr. 16:98-107.

LASKER, R.

1970. Utilization of zooplankton energy by a Pacific
sardine population in the California current. In J. H.
Steele (editor), Marine food chains, p. 265-284. Univ.
Calif. Press, Berkley and Los Ang.

Laurs, R. M.

1967. Coastal upwelling and the ecology of lower trophic
levels. Ph.D. Thesis, Oregon State Univ., Corvallis,
103 p.
LeBRASSEUR, R. J.

1965. Seasonal and annual variations of net zooplankton
at Ocean Station P. 1956-1964. Fish. Res. Board Can.,
Manuscr Rep. Ser. 202, 33 p.
LeBRASSEUR, R. J., AND O. D. KENNEDY

1972. Microzooplankton in coastal and oceanic areas on
the Pacific Subarctic Water mass: A preliminary report.
In A. Y. Takenouti (editor). Biological oceanography of
the northern North Pacific Ocean, p. 355-365. Idemitsu
Shoten, Tokyo.
LONGHURST, A. R.

1967. Diversity and trophic structure of zooplankton
communities in the California Current. Deep-Sea Res.
14:393-408.

McAllister, C. D.

1961. Zooplankton studies at ocean weather station "P"

in the northeast Pacific Ocean. J. Fish. Res. Board Can.

18:1-29.
1969. Aspects of estimating zooplankton production from

phytoplankton production. J. Fish. Res. Board Can.

26:199-220.

McAllister, C. D., T R. Parsons, and J. D. H. Strickland.

I960. Primary productivity and fertility at station "P" in
the Northeast Pacific Ocean. J. Cons. 25:240-259
McGowan, J. A.

1971. Oceanic biogeography of the Pacific. In B. M.
Funnell and W. R. Riedel (editors), The micropaleontol-
ogy of oceans, p. 3-74. Cambridge Univ. Press, Camb.

MEDNIKOV, B. M.

1958. On the plankton of the northwestern part of the

Pacific Ocean. Vses. Nauchno-Issled. Inst. Morsk. Rybn.

Khoz. Okeanogr (VNIRO), Mosc, p. 76-86. (Fish. Res.

Board Can., Transl. Ser 183, 7 p., 1958).
OMORI, M.

1969. Weight and chemical composition of some impor-
tant oceanic zooplankton in the North Pacific Ocean.
Mar Biol. (Berl.) 3:4-10.

Parsons. T. R., and R. J. LeBrasseur.

1970. The availability of food to different trophic levels in
the marine food chain. In J. H. Steele (editor), Marine
food chains, p. 325-343. Univ. Calif Press, Berkley and
Los Ang.

PEARCY W G.

1964. Some distributional features of mesopelagic fishes
off Oregon. J. Mar. Res. 22:83-102.

1965. Species composition and distribution of pelagic
cephalopods from the Pacific Ocean off Oregon. Pac.
Sci. 19:261-266.

1970. Vertical migration of the ocean shrimp, Pandalus
jordani: A feeding and dispersal mechanism. Calif.
Fish Game 56:125-129

FISHERY BULLETIN: VOL. 74, NO. 1

1972. Distribution and ecology of oceanic animals off
Oregon. In A. T. Pruter and D. L. Alverson (editors), The
Columbia River estuary and adjacent ocean waters, p.
351-377. Univ. Wash. Press, Seattle.

PEARCY, W. G., AND C. A. FORSS.

1966. Depth distribution of oceanic shrimps (Decapoda:
Natantia) off Oregon. J. Fish. Res. Board Can. 23:
1135-1143.

1969. The oceanic shrimp Sergestes similis off the Ore-
gon coast. Limnol. Oceanogr. 14:755-765.

PEARCY, W. G., AND R. M. LAURS.

1966. Vertical migration and distribution of mesopelagic
fishes off Oregon. Deep-Sea Res. 13:153-165.

PEARCY, W. G., AND R. S. MESECAR.

1971. Scattering layers and vertical distribution of
oceanic animals off Oregon. In G. B. Farquhar (editor),
Proceedings of an international symposium on biologi-
cal sound scattering in the ocean, p. 381-394. Maury
Cent. Ocean Sci., Wash,, D.C.

PEREYRA, W. T., W. G. PEARCY, AND F. E. CARVEY, JR.

1969. Sebastodes flavidus, a shelf rockfish feeding on
mesopelagic fauna, with consideration of the ecological
implications. J. Fish. Res. Board Can. 26:2211-2215.

PETERSON, W. K.

1972. Distribution of pelagic Copepoda off the coasts of
Washington and Oregon during 1961 and 1962. In A. T.
Pruter and D. L. Alverson (editors), The Columbia River
estuary and adjacent ocean waters, p. 313-343. Univ.
Wash. Press, Seattle.

PETERSON, W. T., AND C. B. MILLER.

1975. Year-to-year variations in the planktology of the
Oregon upwelling zone. Fish. Bull., U.S. 73:642-653.

PILLSBURY, R. D.

1972. A description of hydrography, winds, and currents
during the upwelling season near Newport, Oregon.
Ph.D. Thesis, Oregon State Univ., Corvallis, 163 p.

Reid, J. L., Jr.

1962. On circulation, phosphate-phosphorous content,
and zooplankton volumes in the upper part of the
Pacific Ocean. Limnol. Oceanogr 7:287-306.
REID, J. L., JR., G. I. RODEN, AND J. G. WYLLIE.

1958. Studies of the California Current system. Calif
Coop. Oceanic Fish. Invest. Progr. Rep. 1 July 1956 to 1
January 1958, p. 25-57.
SETTE, O. E.

1955. Consideration of midocean fish production as re-
lated to oceanic circulatory systems. J. Mar. Res.
14:398-414.

Smith, P. E.

1971. Distributional atlas of zooplankton volume in the
California Current region, 1951 through 1966. Calif.
Coop. Oceanic Fish. Invest. Atlas 13, 144 p.

St JOHN, P. A.

1958. A volumetric study of zooplankton distribution
in the Cape Hatteras area. Limnol. Oceanogr. 3:387-397.
Tate, M. W, and R. C. Clelland.

1957. Nonparametric and shortcut statistics in the social,
biological, and medical sciences. Interstate Printers
and Publishers, Inc., Danville, 111., 171 p.
THRAILKILL, J. R.

1956. Relative areal zooplankton abundance off the
Pacific Coast. U.S. Fish Wildl. Serv., Spec. Sci. Rep.
Fish. 188, 85 p.

Vinogradov, M., and N. Voronina.

1962. The distribution of different groups of plankton in
accordance with their trophic level in the Indian Equa-
torial Current area. Rapp. P.-V. Reun., Cons. Perm. Int.
Explor Mer 153:200-204.
WILLIAMS, R. B., M. B. MURDOCH, AND L. K. THOMAS.

1968. Standing crop and importance of zooplankton in
a system of shallow estuaries. Chesapeake Sci. 9:42-51.
Wyatt, B., W. V. Burt, and J. G. Pattullo.

1972. Surface currents off Oregon as determined fi-om
drift bottle returns. J. Phys. Oceanogr. 2:286-293.

CULTURE AND GROWTH OF NORTHERN ANCHOVY,
ENGRAULIS MORDAX, LARVAE

John R. Hunter^

ABSTRACT

Culture techniques used to rear larval anchovy through metamorphosis using laboratory cultured
foods are described. Anchovy larvae fed dinoflagellates Gymnodinium splendens, rotifers Brachionus
plicatilis, harpacticoid copepods Tisbe furcata, and brine shrimp nauplii Artemia salina, completed
metamorphosis (35 mm) in 74 days at 16°C with a minimum survival of 12.5*^. Growth in length and
weight were recorded over this interval and an excellent fit to the Laird-Gompertz growth equation was
obtained. Growth was comparable to that on a wild plankton diet. In a starvation experiment, most of
the fish that completed metamorphosis withstood a starvation period of 12- 15 days, whereas those that
had not completed metamorphosis did not.

Knowledge of the growth rate of northern an-
chovy, Engraulis mordax Girard, is essential for
estimating year class success or larval survival.
Another important element in estimating sur-
vival is the time fish or larvae can withstand
starvation. In this report I describe the growth
rate of larval anchovy to metamorphosis and
present data on the ability of newly metamor-
phosed juveniles to withstand starvation. Special
attention is also given to culture techniques be-
cause this is the first time northern anchovy have
been reared through metamorphosis entirely on
cultured foods.

Kramer and Zweifel (1970) recorded the growth
of anchovy larvae at 17° and 22°C for periods of
up to 34 days. In their experiments larvae at-
tained an average length of 17 mm but did not
reach metamorphosis, which is complete at about
35 mm standard length. Their larvae were fed
wild plankton supplemented by Artemia salina
nauplii. In the ensuing years, rearing techniques
using cultured foods have gradually been de-
veloped: Gymnodinium splendens for 3- to 5-day-
old larvae (Lasker et al. 1970), and Brachionus
plicatilis for 5- to 20-day-old larvae (Theilacker
and McMaster 1971). This paper describes the use
of the harpacticoid copepod Tisbe furcata which
are the proper size food for larvae older than 20
days (10 mm). All previous attempts to rear
anchovy larvae beyond 35 days on cultured foods
have failed. In all attempts Artemia nauplii were
used after 20 days.

'Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

METHODS

Five rearing experiments were done, four at
16°C and one at 17° to 18°C (Table 1). Eggs for all
experiments were obtained from a captive popu-
lation of anchovy which were maintained in
breeding condition continuously at the South-
west Fisheries Center La Jolla Laboratory
(Leong 1971).

Rearing tanks were cylindrical, black fiber-
glass, 122 cm diameter, 36 cm deep, covered
with a transparent acrylic plastic top, and im-
mersed in a water bath regulated by a refrigera-
tion unit. Temperature was maintained near
16°C in all but one experiment, and the salinity
was 35%. Fluorescent lamps suspended directly
over each tank provided about 2,000 Ix at the
water surface. The volume of water in the tanks
gradually increased from an initial volume of 200
liters of filtered seawater to 400 liters by about
20 days because of additions of seawater contain-
ing algae and food organisms. Thereafter, the
volume was maintained at about 400 liters by
siphoning water from the bottom from time to
time which also cleaned the tank.

Records were kept of the quantity of food or
algae added to tanks and on alternate days 16,
0.20-ml aliquots were taken to measure the den-
sity of Brachionus plicatilis, Gymnodinium
splendens, and Artemia salina nauplii in the
tanks. Concentrations of Tisbe furcata in the
tanks were not recorded because they were con-
centrated on or near the walls and bottom of the
tank, but records were kept of the numbers added
to the tank. Details regarding the feeding of

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

FISHERY BULLETIN: VOL. 74, NO. 1

o o

in o +1

CO c\j

8°

■* o +1
iri

CO 00
in

o o
o

CM O +1

00
CO

Cvj

'" "-i

' — '

co'r^

ir>

r~.

a, o-uj u, c

S 0)

_OT- -

X 3 <D ^ ^ 0)
LU2I- Q.

2? CD

E o

5 CD «

ra co^

52 ra

E Q)

CtS

^= c 1
2 5.£?

« 03

E 01

CD c 2 ~;
2 c ?

*^ T3 ■— C

E <D
CD

5 <1> ^

2"; != ?

H? CO

(D "O
E <D

3 n

5 CD i
CD CD 4

c
o

T3
CD

CO in I I

O) CD T h- CD
»- 't CD ^ T^
CO ■.-

O ^ O Tt t

CM I-- CM r^ t^

in T^ in -A CO

CM CM -a-

O -r- O 00

CM O CD CD

O CO O 00

Cvi CO C\J CO I

•>a- r^ 4 -A I

CM CM

CM CO I I I

CM CM CD r^
CO

O CM O CM
CM Tt CM Tt

in T^ in ^

CM CM

■.- CO
CM

r^ CO CD

■r- C) CD I

O CD O
CM ^ CM

^2 I I ^

CO in h- CO r)
o o 1- Tt o

o -^ O -'^ -<;J-

CVJ CD CM CO CO

LO -A IT) T^ Tt

CSJ C\J OJ

I I I

o o o

c> cb CM

CO O '-

CM CO

O 00 tT

CM ■<)• r-.

CD T^ CO

CM CD

2 111 III

o o

oo o

>- CM

O CD

CM CO I

CM -A I

■■- CM

I I I

I III

I ° III

I I

I I I

3
_ c;
o -9

o 2

S DD

O
,^ S _

3 .._
CD o

o -o

O CD "D "D IB

^^ ^ f— r- —

c:> p c c

m C "1 fc CD H) CD

c _

(D O

Q) (/I O en

" ~ 3

= 5 = ^ o ic

o o

CD _

C CD

o "

— "D

>--o

CD "D
"D CD
0) <D
CD CD

E 5

CD _.

ro E
^^

TD CJ

E'^

<D -^

in o

3 ^

O o
,£ en

o o

anchovy larvae on Tisbe will be given in a sepa-
rate section.

To obtain grovvi;h rates, 15 or more larvae were
removed every other day from each tank, then
measured, rinsed in distilled water, dried, and
weighed in groups of 15.

CULTURE

Not until the fifth experiment was the proce-
dure developed sufficiently to rear anchovy
through metamorphosis. The first four experi-
ments ended when it became obvious that it
would be impossible to rear them to metamor-
phosis because of slow growth and high mortality.
Data are included from the first four experiments
to provide the background information for the
final successful rearing procedure.

In all experiments, a single inoculation of 40
liters of Gymnodinium splendens (1,500-2,000
cells/ml) was given at age 0 days. This was
sufficient to provide a final density in the tank in
excess of 100 cells/ml for about 12 days. Gym-
nodinium was cultured using techniques de-
scribed by Thomas et al. (1973). If fed only
Gymnodinium, survival of anchovy larvae re-
mains high for at least 12 days (about 45% at 12
days) but growth is depressed (Hunter in prep.).

In all experiments Brachionus was added on
the 4th or 5th day in numbers calculated to yield
a density of 30 to 50/ml in the tank (Table 1).
Subsequent additions were made daily or on
alternate days until day 20 in all experiments.

Nannochloris sp. was used to culture the rotifer
Brachionus (Theilacker and McMaster 1971), and
as a consequence Nannochloris was added to
larval rearing tanks in all experiments to main-
tain a food supply for the rotifers. Four liters
(about 13,000 cells/ml) were added on days 4 and
5, and further additions were made on the basis of
water color. If water in the rearing tank was
faintly green none was added as I wished to avoid
creating a bloom in the tank because it is difficult
to see the larvae in a dense bloom. To avoid a
bloom usually required a reduction in the quan-
tity added after 20 days. Nannochloris sp. is too
small (about 7 /xm) to be directly fed upon by
larval anchovy although larvae might ingest cells
accidently.

In experiments 1 and 2 Artemia was added at
about 20 days and the level of Brachionus was
allowed to slowly decline thereafter. In experi-
ment 3, Brachionus was maintained at a high

HUNTER: CULTURE AND GROWTH OF ENGRAULIS MORDAX

level to the end and no Artemia was used. Al-
though high mortalities on the order of 30 to 300
larvae/day occurred in all three experiments be-
tween ages 20 to 30 days, the larvae in experi-
ment 3, those fed only Brachionus , grew faster
(Figure 1) and had a higher survival than in the
two groups fed Artemia . From these three exper-
iments I concluded that Artemia was an in-
adequate food for 20-day-old anchovy larvae and
that growth and survival could be increased by
continuing to add large quantities of Brachionus
after 20 days. Clearly, an adequate food larger
than Brachionus was needed for 20-day-old
larvae.

The food selected was the harpacticoid copepod
Tisbe furcata. Tisbe is a common contaminant in
the seawater system of the Southwest Fisheries
Center and can be easily reared on dried foods
(Johnson and Olson 1948) or algae (DeVauchelle
and Girin 1974). Copepods collected from cultures
ranged from SO-yum nauplii to 1,000-/Lim adult
females but the typical size was about 650 yum
and comparable in size to Artemia nauplii. The
first attempt to rear anchovy using Tisbe (exper-
iment 4) began as the other experiments except
that I began adding Tisbe at age 12 days at the
average rate of 180,000/day. At age 20 days the
rate was increased to 240,000/day and the
Brachionus was allowed to decline. The larvae fed
on Tisbe but growth was slow and survival low.
The low survival was attributed to an insufficient
number of Tisbe in the tank, failure to maintain
Brachionus at a high level after 20 days, as I had

t 12

o 10

0' ■ ' ' i i ^ I I ' i ' I ' I ' I ' ^ I I ' i ■ ^ ' I ' I I I I I I

0 4 8 12 16 20 24 28 32 36 40 44 48

AGE (days)

Figure l. — Laird-Gompertz growth curves for lengths of
anchovy larvae in five rearing experiments. Growth equation
given in text; parameters for equation in Table 2. (Foods used
in experiments 1-5 in Table 1.)

in experiment 3, and too high an egg stocking
density (6,000 eggs).

The first four experiments established the
guidelines needed for experiment 5, the final and
successful rearing experiment. Over the first 20
days Brachionus, Nannochloris, and Gym-
nodinium additions were managed in the same
way as in experiment 3. After 20 days, additions
of Brachionus were increased above that used in
experiment 3 and maintained at a high level
until the end of the experiment on day 74. Tisbe
additions were begun at age 6 days at an average
rate of 260,000/day and increased to 306,000/day
after 20 days. These additions were begun before
most larvae were capable of feeding upon them in
order to bring the copepod density in the tank to a
high level at the time feeding on Tisbe became
common (about age 12 days, anchovy length, 7-8
mm). This procedure is practical because survival
of Tisbe in the tank is high and consequently,
uneaten animals accumulate. Tisbe additions
ended at age 48 days (26 mm) because the quan-
tities needed exceeded the capacity of my cul-
tures. Although younger larvae did not survive
on a diet of Artemia nauplii it seemed possible
that larvae 26 mm long might survive because
they have a differentiated digestive tract, not
simply a straight tube as do younger larvae, and
they have a larger gut capacity (C. O'Connell,
Southwest Fisheries Center La Jolla Labora-
tory, pers. commun.) Rosenthal (1969) showed
that Artemia nauplii in the guts of herring larvae
were only partially digested whereas digestion of
copepods was nearly complete. From this he con-
cluded that poor survival of herring fed Artemia
could be attributed to digestive inefficiency. Past
experience in maintaining adult anchovy at the
Southwest Fisheries Center showed that they
survived on Artemia; thus, it seemed reasonable
that this might first occur when the digestive
tract became differentiated. For these reasons I
decided to change from a diet of Tisbe and
Brachionus to one of Artemia nauplii and Brachi-
onus at age 48 days. The change from copepods to
Artemia nauplii did not cause a noticeable mor-
tality nor a change in growth rate. Adult Artemia
were added at age 69 days as some of the fish
had metamorphosed and readily ingested adult
Artemia.

In this description of culture I have stressed
additions rather than density of food in the tank
because I felt they provided a more reliable out-
line of culture procedures. Density in the tank

FISHERY BULLETIN: VOL. 74, NO. 1

was measured before food was added and served
as a guide for the quantity of food to be added.
Where losses from ingestion or other sources of
mortality were high, the density measurements
tended to be lower than the level we attempted to
maintain. In experiment 5 we attempted to main-
tain the density of Brachionus between 50 and
100/ml and that ofArtemia nauplii at 2 to 3/ml.

In all experiments, 15 or more larvae were
removed on alternate days and consequently,
survival estimates include the effect of this sam-
pling. In experiments 1 to 4 no daily counts of
dead larvae were made until heavy mortalities
occurred after age 20 days. In experiment 5, daily
records of dead larvae were begun at age 54 and
continued to the end of the experiment (age 74
days). At age 54 days 20% of the larvae were alive
and at age 74 days, 374 larvae or 12.5% were
alive. If the tank had not been sampled survival
would probably have been greater because be-
tween 54 and 74 days the number of larvae
sampled, 151, exceeded the number that died in
the tank, 70. A total of 387 larvae were removed
during the experiment. A method exists for es-
timating mortality in rearing work independent
of the effect of sampling (Laurence 1974) but the
programming effort required seems unwarranted
for the objective of this paper. Collision with the
walls of the container was a frequent cause of
mortality over the last 3 weeks.

A survival of 12.5% at 74 days contrasts sharp-
ly with the other four experiments where nearly
all larvae died by 30 to 40 days. Prior to the study
described here, marked mortalities were common
after 20 days and in all of the attempts Artemia
was used as food. The pattern had become so
typical at this laboratory that we have called it
the 'Artemia syndrome" for some years. The
results of the current study suggest that the
cause of the Artemia syndrome may simply be an
inability of young clupeoid larvae with straight
tube digestive tracts to digest Artemia nauplii
but that Artemia may be used once the gut
becomes differentiated.

It is important to call attention to the fact that
6% of the anchovy larvae in experiment 3 were
able to survive for 42 days on a diet of only
Brachionus. Plaice, Pleuronectes platessa, larvae
have been reared through metamorphosis on only
Brachionus although growth was slower than
that on Artemia nauplii (Howell 1973). Howell
found that plaice larvae, immediately prior to
metamorphosis (12.7 mm), consumed 1,400 roti-

fers per day. In experiment 3, at age 42 days, the
mean length of the anchovy larvae was 21.6 mm
and dry weight was 5.5 mg. Assuming a digestive
efficiency of 100%, larvae of this weight would
have to ingest about 3,800 rotifers per day to
meet metabolic requirements (calculation based
on caloric value of Brachionus and anchovy respi-
ration data given by Hunter 1972). These results
illustrate the value of maintaining a high density
of rotifers in culture containers long after a larger
food has been added. They also suggest that some
fish larvae have the ability to ingest large quan-
tities of small prey and this could be of consider-
able benefit under natural conditions.

TISBE FURCATA AS A FOOD FOR
LARVAL FISH

The evidence for the use of Tisbe as a food for
rearing larval anchovy to metamorphosis is a
single rearing experiment. It would be preferable
to have additional experiments but none are
planned at present because current work is con-
cerned with only young stages and other species.
Two groups of Pacific mackerel. Scomber japon-
icus, have been reared to metamorphosis using
Brachionus and Tisbe as foods and this supports
the contention that Tisbe is a satisfactory food for
pelagic marine fish larvae. The work on Scomber
will be reported at a later date.

That larval anchovy ate Tisbe is supported by
records of stomach contents of larvae examined
during the course of the rearing work. Seventy-
four percent of the stomachs examined in experi-
ment 5 contained only Tisbe or Tisbe and
Brachionus and 26% contained only Brachionus
{N = 69, larval length = 8.6-18.8 mm). The num-
ber of Tisbe in stomachs of larvae increased
from 2.8 per larva (5.6-8.5 mm) to 18 per larva
(17.6-20.5 mm). (Data from experiments 4 and 5
combined — Table 2.) The average length of the

Table 2. — Number and mean length of Tisbe furcata in the
stomachs of anchovy larvae in experiments 4 and 5.

Larval at

ichovy

Number

Tisbe in stomachs

Length
class
(mm)

Total

Number

per
larva'

Mean

length

/im ± 2 SE

5.6- 8.5

2.8

506

± 57

8.6-11.5

3.6

681

± 28

11.6-14.5

102

6.4

714

± 28

14.6-17.5

10.9

758

± 32

17.6-20.5

18.0

734

± 43

'Includes only larvae that had either Tisbe and Brachionus or only Tisbe
In stomachs.

HUNTER: CULTURE AND GROWTH OF ENGRAULIS MORDAX

copepods ingested by larvae also increased with
larval length as expected (Arthur 1956).

Tisbe occurred throughout the rearing tank but
the greatest concentrations occurred on or near
the walls and on the bottom. Free swimming
copepods were plentiful near the walls of the tank
because Tisbe frequently leave the wall for short
periods. Anchovy larvae captured Tisbe that were
on the walls as well as free-swimming individu-
als. A pelagic copepod would be preferable to one
that prefers surfaces such as Tisbe furcata but I
have not been able to culture pelagic species in
sufficient quantities for rearing work.^

GROWTH

The length data from each of the five experi-
ments were fitted to the Laird-Gompertz growth
equation (Laird et al. 1965) using Marquardt's
Algorithm for fitting nonlinear models (Conway
et al. 1970). The equation for length was:

Loe

e-«<)

where L - standard length in millimeters
Lq = initial length at time 0

Aq = rate growth at time 0
a = rate of decay of growth.

A fit of the weight data from experiment 5 was
also made to the Laird-Gompertz equation:

The length-weight relationship for larvae in
experiment 5 was derived from the above two
equations by James Zweifel (Southwest Fisheries
Center) and had the form

In W = In Wo + K^^r

1 -

Kl - In (L/Lq)

/3/a

The Laird-Gompertz equation gave an excellent
fit to the growth in length and in weight and to
the length-weight relationship (Figures 2-4,
Table 3). The curvilinear nature of the length-
weight data evident in the log-log plot (Figure 4)
clearly indicates that a linear fit to log of length
and weight would lead to inaccurate estimates.

The growth of anchovy larvae in experiment 5
was about the same as that recorded by Kramer
and Zweifel (1970) for anchovy fed wild plankton
at 17°C. At age 34 days, the last day of their
experiment, the mean length of larvae was 17.4 ±
1.8 mm and that in experiment 5 at age 34 days
was 19.7 ±1.0 mm. Thus, over at least the first 34
days, growth on the cultured food diets was about
the same as that on wild plankton.

SURVIVAL AT METAMORPHOSIS

The object of this experiment was to determine
how long newly metamorphosed anchovy larvae
can survive without food. Most adult fishes and
presumably the anchovy can withstand prolonged
periods of starvation of weeks or months. On the

W = Woe

Kwn

.-^')

where W = dry weight in milligrams
Wo = initial weight at time 0

Bq = rate growth at time 0
fi = rate of decay of growth.

^At present our copepod culture system is composed of 10,
90-liter, glass, rectangular tanks maintained at 17° to 19°C.
The Tisbe are given green algae, either Tetraselmis or Nan-
nochloris, which is grown using commercial plant fertilizer (fish
emulsion). An inoculation of 50,000 to 100,000 copepodid-
adult stages yields on the average 500,000 copepods in these
stages in 2 weeks. A tank is drained, harvested, and reestab-
lished 5 days a week producing about 2.5 x 10® copepodid-adult
stages per week, which is sufficient to rear one group of anchovy
in the manner described. Occasional htirvests of over a million
in 2 weeks have been obtained suggesting that major improve-
ments in the technique are possible. Contamination by
Brachionus has been a problem because it increases the amount
of algae that must be added to the culture. A more detailed
description of this culture system would be premature but a
description of a similar method of mass culture exists (De-
Vauchelle and Girin 1974).

- 25

E
E

f 20

i±L

i±l

ill

1.^' ■!■

I ,

15 20 25 30 35 40 45 50 55 60 65 70 75
AGE (days)

Figure 2. — Laird-Gompertz growth curve for length of anchovy
larvae in experiment 5 and mean length ± 2 SE. Parameters
for equation given in Table 2.

FISHERY BULLETIN: VOL. 74, NO. 1

001 Li.

AGE (doys)

Figure 3. — Laird-Gompertz growth curve for dry weight of
anchovy larvae in experiment 5. Points are average weight
of larvae weighed in groups of 15-26 larvae each. Equation
for curve given in text; parameters for equation in Table 2.

other hand, larval anchovy, after they absorb
their yolk, survive only 1 to 2 days without food
(Lasker et al. 1970). The point at which this
extreme vulnerability to starvation ends is essen-
tial information for any model of anchovy ecology
and survival.

In this experiment fish reared to metamor-
phosis in experiment 5 were used. At age 74 days
a group of 53 fish (group 1) and one of 73 (group 2)
were placed into tanks containing only filtered
seawater and a sample of 29 fish was taken for
length and weight measurements. The tanks

lOOOr

50.0-

10.0

5 05

0.05

001

I I ' 111

5 10 20

LENGTH (mm)

100

Figure 4. — Length-weight relationship of anchovy larvae
reared in experiment 5. Equation for curve given in text;
parameters for equation in Table 2.

were the same as those described for the rearing
experiments and temperature was maintained at
16°C. Artemia nauplii were offered to group 1
after 12 days of starvation and to group 2 after 15
days; the experiment ended after 20 days. Daily
records were kept of water temperature and
lengths of dead fish; after 20 days all surviving
fish were measured. Total lipid content was also
monitored through the course of the experiment.

Table 3. — Parameters and 95% support plane' for Laird-Gompertz growth equation for length, experiments 1-5, and weight

for experiment 5. Symbols and equations are given in text.

Parameter

Support plane
Lower Upper

Parameter

Support plane

Parameter

Support plane

Number of

Experiment

Lower

Upper

Lower

Upper

observations

Length
1

2.4378

2.0856

2 7901

0 1349

0.1040

0.1658

0.06939

005366

0.08512

289

3.0600

2.5512

3.5688

0.0835

0.0614

0.1056

003936

0.02700

0,05171

358

28361

2.3894

3.2829

0.1088

0.0835

0.1342

004951

003696

006206

345

2,6711

2.1648

3,1774

0.1098

0.0792

0,1404

005185

0,03591

0,06779

345

24928

2.2219

2.7636

0.1167

0.1042

0.1292

004264

0.03865

0,04663

553

Weight
5

0.005758

0.003046

0.008470

0.2997

0.2552

0.3442

0.02725

0.02243

0.03206

'An approximation of the 95% confidence limits (Conway et al. 1970)

HUNTER: CULTURE AND GROWTH OF ENGRAULIS MORDAX

Fat was removed by Soxhlet extraction with
chloroform-methanol (Krvaric and Muzinic 1950)
from batches of 5 to 9 fish each. One such sample
was taken at the beginning of the experiment,
one from each group just before food was added,
and one from each group when the experiment
ended after 5 to 8 days of feeding.

A marked initial mortality occurred on the day
following the transfer of the two groups (Figure 5)
which was probably caused by handling. For this
reason the first day's mortality is excluded from
the analysis presented below, but the survival is
given for all days in the figure.

Table 4. — Lengths of fish in starvation groups and lengths
of fish that died during starvation.

= FOOD ADDED

Figure 5. — Percent survival of metamorphosed larvae reared
in experiment 5 during starvation periods of 12 and 15 days.
Arrow indicates end of starvation period.

After 12 days of starvation, 50% of the fish were
alive in group 1 (excluding the first day mortal-
ity) and 58% were alive in group 2 after 15 days of
starvation. One fish in group 1 died the day after
the first feeding. This was the only fish to die
after feeding began. Thus for fish averaging 35
mm in length, about 50% mortality is reached
after about 15 days of starvation and nearly all
surviving fish are able to recover from a starva-
tion period of that duration. Mortality during
starvation appeared to be dependent on size or
state of maturity, however. Metamorphosis is
completed in the northern anchovy when they
reach 35 mm standard length (E. H. Ahlstrom,
Southwest Fisheries Center La Jolla Laboratory,
pers. commun.). Eighty-three percent of the fish
that died were less than 35 mm whereas only 17%
of those longer than 35 mm died (Table 4). About
45% of the fish were less than 35 mm long at the
beginning of the experiment. These results are
similar to those obtained for herring larvae,
Clupea harengus. The number of days to irrevers-
ible starvation for herring larvae increased from

Number of fish

Percent

Length

r^ean length

Group

<35 mm

==35 mm

TotaP

<35 mm

mm ± 2 SE

Sample before

starvation

35.4 ± 1.8

All fisti^:

34.4 ± 1.5

36.0 ± 1.7

1 + 2

35.1 ± 1.1

Dead fish:

30.9 ± 1.4

31.8 ± 1.6

1 + 2

31.2 ± 1.1

'Fish that died on first day of starvation in groups 1 and 2 not included.
^Surviving fish measured at end of experiment after 5- to 8-day feed-
ing period.

6 days at the end of the yolk-sac stage to 15 days
at age 88 days (Blaxter and Ehrlich 1974).

Lipid content offish declined during the starva-
tion period from about 30% of dry weight to about
12% (Table 5). Recovery for the surviving fish was
rapid, as they returned to the 30% level after 5 to
8 days of feeding. Water content was inversely
related to fat as expected (lies and Wood 1965).
Fat content of muscle of adult anchovy is about 30
to 40% of dry weight during late summer and fall
when gonadal fat is low (Lasker, Southwest
Fisheries Center La Jolla Laboratory, unpubl.
data). Thus, fat levels of these newly metamor-
phosed larvae appeared to be about the same as
that of adult fish.

Table 5. — Total lipid and water content of anchovy at meta-
morphosis before, during, and after starvation.

Total

Elapsed

lipid

Dry

Mean

time

Water

dry wt

length

Treatment

(days)

(%)

(mg)

(mm)

Before starvation

78.1

30.6

74.6

35.0

End starvation:

Group 1

83.2

10.5

33,3

32.8

Group 2

82.9

13.4

54.2

36.9

End feeding:

Group 1

79.5

32.3

90.3

39.6

Group 2

79.2

32.3

83.7

39.8

Extreme vulnerability to starvation appears to
be characteristic of only the larval phase of the
northern anchovy and it is over by the time the
fish completes metamorphosis. There is a danger
in interpreting these data beyond these general
conclusions because reared fish may have more
fat than wild ones and this could alter the results
(Balbontin et al. 1973).

FISHERY BULLETIN: VOL. 74, NO. 1

ACKNOWLEDGMENTS

James Zweifel (Southwest Fisheries Center La
Jolla Laboratory) fit the Laird-Gompertz growth
equation to the data and derived the length-
weight relationship from the grovd;h equations.
Carol Sanchez (Southwest Fisheries Center La
Jolla Laboratory) assisted in all phases of this
work. Reuben Lasker and Gary Stauffer (South-
west Fisheries Center La Jolla Laboratory) re-
viewed the manuscript.

LITERATURE CITED

ARTHUR, D. K.

1956. The particulate food and the food resources of the
larvae of three pelagic fishes, especially the Pacific
sardine, Sardinops caerulea (Girard). Ph.D. Thesis.,
Univ. California, Scripps Inst. Oceanogr., 231 p.

Balbontin, F., S. S. De Silva, and K. F. EHRLICH.

1973. A comparative study of anatomical and chemical
characteristics of reared and wild herring. Aquaculture
2:217-240.

BLAXTER, J. H. S., AND K. F. EHRLICH.

1974. Changes in behaviour during starvation of herring
and plaice larvae. In J. H. S. Blaxter (editor). The early
life history of fish, p. 575-588. Springer- Verlag, Berl.

Conway, G. R., N. R. Glass, and J. C. Wilcox.

1970. Fitting nonlinear models to biological data by
Marquardt's algorithm. Ecology 51:503-507.
DeVauchelle, B., and M. GIRIN.

1974. Production du rotifere "Brachionus plicatilis O. F.
Muller" en elevage mixte avec le copepode "Tisbe
furcata" (Baird). Cent. Nat. Exploit. Oceans, Colloq.
Aquaculture, Ser.: Actes Colloq. 1:87-100.
HOWELL, B. R.

1973. Marine fish culture in Britain VIII. A marine roti-
fer, Brachionus plicatilis Muller, and the larvae of the
mussel, Mytilus edulis L., as foods for larval flatfish.
J. Cons. 35:1-6.

HUNTER, J. R.

1972. Swimming and feeding behavior of larval anchovy
Engraulis mordax. Fish. Bull., U.S. 70:821-838.

ILES, T. D., AND R. J. WOOD.

1965. The fat/water relationship in North Sea herring
(Clupea harengus), and its possible significance. J.
Mar. Biol. Assoc. U.K. 45:353-366.
JOHNSON, M. W., AND J. B. OLSON.

1948. The life history and biology of a marine harpacti-
coid copepod, Tisbe furcata (Baird). Biol. Bull. (Woods
Hole) 95:320-332.
KRAMER, D. AND J. R. ZWEIFEL.

1970. Growth of anchovy larvae (Engraulis mordax
Girard) in the laboratory as influenced by temperature.
Calif Coop. Oceanic Fish. Invest. Rep. 14:84-87.
KRVARIC, M., AND R. MUZINlC.

1950. Investigation into the fat content in the sar-
dine tissues (Clupea pilchardus Walb.). Acta Adriat.
4:289-314.
LAIRD, A. K., S. A. TYLER, AND A. D. BARTON.

1965. Dynamics of normal growth. Growth 29:233-248.
LASKER, R., H. M. FEDER, G. H. THEILACKER, AND R. C. MAY.

1970. Feeding, growth, and survival of Engraulis mordax
larvae reared in the laboratory. Mar. Biol. (Berl.)
5:345-353.

LAURENCE, G. C.

1974. Growth and survival of haddock (Melanogrammus
aeglefinus) larvae in relation to planktonic prey concen-
tration. J. Fish. Res. Board Can. 31:1415-1419.

LEONG, R.

1971. Induced spawning of the northern anchovy,
Engraulis mordax Girard. Fish. Bull., U.S. 69:357-360.

ROSENTHAL, H.

1969. Verdauungsgeschwindigkeit, Nahrungswahl und

Nahrungsbedarf bei den Larven des Herings, Clupea

harengus L. Ber. Dtsch. Wiss. Komm. Meeresforsch.

20:60-69.
THEILACKER, G. H., AND M. F. MCMASTER.

1971. Mass culture of the rotifer Brachionus plicatilis

and its evaluation as a food for larval anchovies.

Mar. Biol. (Berl.) 10:183-188.
Thomas, W. H., A. N. dodson, and C. A. Linden.

1973. Optimum light and temperature requirements for
Gymnodinium splendens, a larval fish food organism.
Fish. Bull., U.S. 71:599-601.

EFFECTS OF COOKING IN AIR OR IN NITROGEN ON THE

DEVELOPMENT OF FISHY FLAVOR IN THE BREAST MEAT

OF TURKEYS FED TUNA OIL WITH AND WITHOUT

a-TOCOPHEROL SUPPLEMENT OR INJECTION

L. Crawford and M. J. Kretsch^

ABSTRACT

The breast meat of turkeys which had been fed fish oil with and without a-tocopherol supplement or
injection were cooked in air or under nitrogen with a slight vacuum. Cooking under nitrogen
prevented the development of fishy flavor nearly as well as dietary a-tocopherol acetate supplementa-
tion. Some evidence is given which shows that fishy fiavor develops postmortem (during cooking) and
not in vivo.

Crawford et al. (1974) explored the effects of
feeding fish oil with and without a-tocopherol
acetate on the flavor of turkeys. This paper and
other work by Crawford et al. (1975) showed
that dietary a-tocopherol can be very effective in
preventing the development of fishy flavor Simi-
larly, a-tocopherol acetate had a profound effect
on the "elimination" of fishy flavor when it and
beef fat were substituted for fish oil in the rations
of turkeys that had been fed diets containing fish
oil for several weeks. Injection of a-tocopherol (a
few days before slaughter) into the thighs of
turkeys fed diets containing fish oil showed a
positive effect on the reduction of fishy flavor.

Consideration of these results and the finding
that poultry carcass stability is related to the
degree of lipid unsaturation and the tocopherol
content (Mecchi, Pool, Behman, Hamachi, and
Klose 1956; Mecchi, Pool, Nonaka, Klose, Mars-
den, and Lillie 1956; Webb, Brunson, and Yates
1972, 1973; Webb, Marion, and Hayse 1972)
led us to the reasoning that flshy flavor in poultry
may result from in vivo and/or postmortem oxida-
tion of lipids containing long chain a>-3 fatty
acids. Crawford et al. (1975) entertained the
possibility that such oxidation and subsequent
fishy flavor development occur mostly in vivo. At
first glance, the effects of dietary a-tocopherol
acetate on prevention of fishy flavor seem to
support this hypothesis. However, the effective-
ness of injecting a-tocopherol only a few days
before slaughter casts some doubt on this reason-

'Western Regional Research Laboratory, Agriculture Re-
search Service, U.S. Department of Agriculture, Berkeley, CA
94710.

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

ing since in vivo oxidation prior to injection
should have had ample time to occur. Whereas
this doubt does not call for total apostasy, it does
suggest that postmortem oxidation and sub-
sequent development of fishy flavor is indeed a
possibility and deserves consideration.

The exact nature and origin of flshy flavor in
turkeys is not known, but it is known that the
development of such flavor requires the uptake of
oi-S fatty acids from dietary oils rich in these fatty
acids. Most fish oils are rich sources of long
chained w-3 fatty acids which are readily taken
up into the carcass of turkeys when included in
their diet. Linseed oil contains more than 50%
linolenic acid and when incorporated into turkey
diets, the linolenic acid is taken up and elongated
to the longer chained homologues thereby caus-
ing fishy flavor to develop (-Klose et al. 1951;
Miller et al. 1967a, b; Crawford et al. 1974).

If postmortem oxidation plays a major role in
the development of fishy flavor, it is likely that
the development would occur largely during cook-
ing. Pippen and Nonaka (1963) found that the
amount of volatiles from raw chicken was small
and the aroma rather insipid when compared to
the relatively large amount of highly odoriferous
volatiles from cooked chicken. They also reported
that chicken boiled in air yielded a more complex
and larger volatile fraction than chicken boiled in
nitrogen. Crawford (1972) reported that replace-
ment of air in the headspace with nitrogen gave
some protection against scorch during the retort-
ing of 4-pound cans of tuna. This suggests that
less carbonyls (volatiles) were formed under ni-
trogen since volatile carbonyls, sugars, and

FISHERY BULLETIN: VOL. 74, NO. 1

amino compounds (Fujimoto et al. 1968) have
been implicated in such nonenzymatic browning
(Tarr 1954; Jones 1962).

It is clear that the development of the normal
aroma of poultry is time-temperature dependent
and that air or nitrogen cooking atmospheres
have profound effects on the development of this
aroma. Therefore, it is likely that control of the
cooking atmosphere may affect the development
of fishy flavor in poultry meat if this flavor
requires air and/or heat for its development.

This paper explores the effects of cooking in
different atmospheres on the flavor of breast meat
from turkeys fed diets containing tuna oil w^ith
and without dietary a-tocopherol acetate or
a-tocopherol injection. Diced breast meat was
cooked in air as well as under nitrogen with a
slight vacuum.

EXPERIMENTAL

Turkey Diets and Feeding

The turkeys used in this experiment were
taken from groups of turkeys raised experi-
mentally for other work. Their diets and feeding
are described in some detail by Crawford et al.
(1975). Briefly, there were 50 White Broad Breast
poults in experiment C that were divided into five
groups of 10 each and they were fed as follows:
chick starter (6.75% fish meal) was fed to 3 wk of
age, then a 50:50 mixture of chick starter and a
50% soybean meal basal diet for a few days,
followed by the 50% soybean meal diet supple-
mented with 2% soybean oil and 2% beef fat to 8
wk of age. At 8 wk of age, the following fat and oil
supplements replaced the previous ones and they
were fed from 8 to 14 wk of age:

Oil Supplement to Basal Diet^
4% BF

Group
1 C
2C
3 C
4C
5C

2% BF + 2% TO
2% BF + 2% TO
2% BF + 2% TO
2% BF + 2% TO

iBF = Beef fat; TO = Tuna fish oil.

At 14 wk of age, the above groups of turkeys
were fed a 30% soymeal basal diet plus the
following oil supplement to 16 wk of age:

Group Oil Supplement to Basal Diet^

1 C Keep on 4% BF

2 C Change to 4% BF

Group
1 B
2B
3 B

4B
5B

3 C Change to 4% BF + 100 mg Vit. E/kg

4 C Change to 4% BF + 200 mg Vit E/kg

5 C Keep on 2% BF + 2% TO

^BF = Beef fat; Vit. E = dl a-tocopherol acetate;
TO = Tuna fish oil

In experiment B, 50 poults were obtained and
handled as above. On day 3, they were fed a basal
diet plus 4% beef fat to 14 wk of age. From 14 to 16
wk of age, they were fed as follows:

Oil Supplement to Basal Diet
4% BF

2% BF + 2% TO

2% BF + 2% TO (+ injection of 170
mg a-tocopherol into thigh at 72,
48, 24 h before sacrifice)
2% BF + 2% TO + 100 mg Vit. E/kg
2% BF + 2% TO + 500 mg Vit. E/kg

Sampling, Canning, and Analysis

All turkeys were sacrificed at 16 wk of age then
handled and stored at -30°C as described by
Crawford et al. (1974). Two turkeys from each
group were randomly selected and thawed over-
night in a 2°C cold room. The breasts were excised
and diced in the cold after the skin had been
removed. Breast meat from turkeys of the same
group were mixed together and appropriately
identified. The diced breast meat was canned
immediately as follows: breast meat from each
group was hand packed into 307 x 113 cans (eight
cans per group) leaving a headspace of about V2
inch. All cans from each group were alternately
evacuated and flushed with nitrogen several
times. On the final nitrogen flush, the lids were
sealed when the vacuum dropped to 5 inches.
Four of the cans from each group were frozen at
-30°C until used and the other four cans were
cooked immediately at 116°C (15 psi) for 80 min to
an internal temperature of ca. 112°-115°C, cooled,
and stored at 2°C until used. The four uncooked
cans from each group were removed from -30°C
storage, thawed to about 2°C, opened, and the
contents cooked in aluminum trays (with loose
covers) at about 117°C for 30 min (internal tem-
perature ca. 70°C) before serving. Those cans that
were cooked at 116°C were warmed in boiling
water for 10 min before opening and serving.
Organoleptic analysis was performed by a panel
of eight judges using a balanced incomplete block
design (t = 5, r = 4). Only one panel per day was

CRAWFORD and KRETSCH: FISHY FLAVOR IN TURKEY

convened and the air and nitrogen packs w^ere
randomly offered from day to day. Duncan's mul-
tiple range test (a = 0.05) was used to compare
the adjusted mean of the taste panel scores. The
scoring was: 1 = no fishy flavor, 5 = very fishy
flavor

RESULTS AND DISCUSSION

The results reported in Table 2 are to be inter-
preted with some caution because of the low level
of fishiness in the meat from turkeys fed 2% fish
oil for only 2 wk. Therefore, only trends are
indicated for the results in Table 2 where statisti-
cal significance could not be achieved.

Tables 1 and 2 report Duncan's multiple range
test of the mean taste panel scores of breast meat
cooked in air or nitrogen from turkeys fed various
diets containing tuna oil and/or beef fat with and
without dietary a-tocopherol acetate or a-tocoph-
erol injection. All meats that contained
«-tocopherol gave taste panel scores that were
comparable to the scores for the control for all
methods of cooking. When breast meat is cooked
under nitrogen with a slight vacuum no appreci-
able difference in flavor is caused by any of the
diets. However, the breast meat from turkeys fed
diets containing 2% tuna oil (treatments 5C and
2B) did have slightly higher scores, although not
statistically different from the control (treat-
ments IC or IB, 47c beef fat). The breast meat
cooked in air from turkeys fed diets containing
2% tuna oil (treatments 5C and 2B) showed more
off flavor than those cooked in nitrogen when
each is compared to its control (treatments IC
or IB, 4% beef fat). Furthermore, the order and
rank of the scores for the air-cooked meat were
very similar to those of breast meat from whole
roasted turkeys previously reported by Crawford
et al. (1975). These turkeys were randomly
selected from the same groups of turkeys used in
this experiment and were roasted at 177°C to
center breast temperature of about 70°C.

From the results of this experiment, it may be
concluded that cooking breast meat of potentially
fishy flavored turkeys under nitrogen is nearly as
effective in preventing fishy flavor development
as feeding a-tocopherol acetate (in the diets with
the tuna oil) and roasting in the normal manner
This implies that fishy flavor develops postmor-
tem and requires air for its development. Alter-
nately, it could be concluded that cooking under
nitrogen per se had practically no effect in pre-

TABLE 1. — Duncan's multiple range test of mean' taste panel
scores^ for breast meat cooked in air or nitrogen from turkeys
fed various diets containing tuna oil and/or beef with and
without a-tocopherol acetate.

Cooked in nitrogen

Cooked

in air

Roasted no
5 Treatment^

rmally"

Treatment^ Scores

Treatment^

Scores

5C 2% TO 2.05

5C 2% TO

3.23

5C 2% TO

3.14

4C 4% BF + 1 .80

3C 4°o BF

+ 1.66

2C 4% BF

2.43

200 E

100 E

3C 4% BF + 1.77

2C 4% BF,

1.63

3C 4°o BF +

1.31

100 E

2C4%BF, 1.71

1 C 4°o BF

1.24

IC 4% BF

1.29

1C4%BF 1.65

4C 4°o BF
200 E

' 1.08

4C 4% BF +
200 E

0.99

'Mean taste panel scores connected by a common line are not signifi-
cantly different at tfie 0.05 probability level.

^Taste panel scoring: 1 = no fishy flavor, 5 = very fishy flavor. Abbrevia-
tions: TO = tuna oil; E = mg d/ a-tocopherol acetate per kilogram of diet;
BF = beef fat; BF, = beef fat substituted for 2% TO +2% BF.

^AII groups (except group IC, the control which was maintained on diet
with 4°o BF for all 1 6 wk) were fed a basal diet with 2% TO plus 2% BF from
8 to 14 wk of age and from 14 to 16 wk of age, they were fed a basal diet
with: group IC =4°b BF, group 2C = change to 4*^0 BF, group 3C = change
to 4°o BF -r 100 mg/kg a-tocopherol acetate, group 4C = change to 4''o BF
* 200 mg a-tocopherol acetate, group 5C = kept on 2% TO + 2% BF.

■■These results for the breast meat of normally roasted whole turkeys
were previously reported by Crawford et al. (1975).

Table 2. — Duncan's multiple range test of mean^ taste panel
scores^ for breast meat cooked in air or nitrogen from turkeys
fed various diets containing tuna oil and/or beef fat with and
without a-tocopherol acetate supplement or injection.

Cooked in n

itrogen
Scores

Cooked

n air

Roasted nc
Treatment^

rmally"

Treatment^

; Treatment^

Scores

2B 2% TO

1.82

2B 2% TO

2.16

2B 2% TO

2.23

48 2% TO +

1.71

5B 2% TO +

1.74

5B 2% TO +

2.18

100 E

500 E

5B 2% TO +

1.61

4B 2% TO +

1.41

4B 2% TO +

1.86

500 E

100 E

3B 2% TO -

n 1.59

3B 2% TO +

In 1.35

3B 2% TO +

In 1.32

IB 4°o BF

1.43

1 B 4% BF

1.22

1B 4% BF

1.19

'Mean taste panel scores connected by a common line are not signifi-
cantly different at the 0.05 probability level.

^Taste panel scoring: 1 ^ no fishy flavor, 5 = very fishy flavor. Abbrevia-
tions: TO = tuna oil: E = milligrams d/ a-tocopherol acetate per kilogram of
diet; In = inject <i-tocopherol; BF = beef fat.

^AII groups were fed a basal diet *4% BF to 14 wk of age and from 14 to
16 wk of age, they were fed a basal diet with: group IB = 4% BF, group 2B
= 2°o BF + 2°o TO, group 3B =2% BF + 2% TO (+ inject 170 mg of
a-tocopherol into thigh 72, 48. and 24 h before sacrifice), group 4B = 2%
BF + 2% TO + 100 mg a-tocopherol acetate per kilogram, group 5B = 2%
BF + 2% TO + 500 mg a-tocopherol acetate per kilogram.

"These results tor the breast meat of normally roasted whole turkeys
were previously reported by Crawford et al. (1975).

venting this development but that fishy flavor
had already developed in vivo and the heat of
cooking at 116°C for 80 min destroyed the compo-
nents which cause this flavor. Some observations
and recent work (Crawford unpubl. data) tend to
support the first conclusion.

We have observed that the odor of fresh raw
turkey was insipid regardless of the type of diet-
ary oil. However, after comminuting and storing
in the refrigerator overnight, the flesh from tur-
keys fed tuna oil smelled fishy while the odor of
beef fat-fed turkeys remained rather insipid.

FISHERY BULLETIN: VOL. 74, NO. 1

Fresh tuna fish also has very Httle odor but will
develop a characteristic odor during refrigerated
storage or cooking. These observations tend to
support the supposition that fishy flavor develops
during postmortem oxidation.

Additionally, volatiles were steam distilled
from the same tuna oil that was fed to the turkeys
in this experiment. These volatiles appeared to
have the same fishy aroma as turkeys judged to
have fishy flavor by the taste panel. The volatiles
were added to water (ca. 2 ix\/125 ml) in cans with
a nitrogen or air headspace plus a slight vacuum
and cooked at 116°C in the same fashion as the
breast meat. An odor panel revealed little, if any,
loss in character or intensity for the odor of the
volatiles cooked under nitrogen or in air. Al-
though this experiment with the volatiles offers
only deductive reasoning, it nonetheless lends
support to the argument that a heat-stable fishy
flavor develops during cooking in air and that
cooking under nitrogen prevents the development
of this flavor.

ACKNOWLEDGMENTS

Acknowledgment is given for the indispensable
assistance of Helen H. Palmer, D. W. Peterson, K.
E. Beery, A. W. Brant, Carol Hudson, E. P. Mec-
chi, Ko Ijichi, and Linda Eldridge. Further grat-
itude is extended to Hoffman La-Roche, Inc., Pa-
cific Vegetable Oil International, Inc., Star-Kist
Foods, and Van Camp Sea Food.

LITERATURE CITED

Crawford, L.

1972. The effect of premortem stress, holding tempera-
tures, and freezing on the biochemistry and quahty of
skipjack tuna. NOAA Ifech. Rep. NMFS SSRF-651, 23 p.

Crawford, L., D. W. Peterson, M. J. Kretsch, a. L.

LILYBLADE, AND H. S. OLCOTT.

1974. The effects of dietary a-tocopherol and tuna, saf-
flower, and linseed oils on the flavor of turkey. Fish.
Bull., U.S. 72:1032-1038.

Crawford, L., m. J. kretsch, D. W. Peterson, and a. L.

LILYBLADE.

1975. The remedial and preventative effect of dietary
«-tocopherol on the development of fishy flavor in turkey
meat. J. Food Sci. 40:751-755.
FUJIMOTO, K., M. MARUYAMA, AND T. KANEDA.

1968. Studies on the brown discoloration of fish
products — I. Factors affecting the discoloration. [In
Jap., Engl, abstr.] Bull. Jap. Soc. Sci. Fish. 34:519-523.
JONES, N. R.

1962. Browning reactions in dried fish products. In J.
Hawthorn and J. M. Leitch (editors). Recent advances in
food science, 2:74:80. Butterworths, Lond.

Klose, a. a., E. p. Mecchi, h. L. Hanson, and h.
lineweaver.

1951. The role of dietary fat in the quality of fresh and
frozen storage turkeys. J. Am. Oil Chem. Soc.
28:162-164.

Mecchi, e. p., M. F. Pool, G. a. behman, m. hamachi, and
A. A. KLOSE.

1956. The role of tocopherol content in the comparative
stability of chicken and turkey fat. Poult. Sci. 35:1238-
1246.

Mecchi, e. p., m. f. pool, m. nonaka, a. a. klose, s. J.
Marsden. and R. S. Lillie.

1956. Further studies on tocopherol content and stability
of carcass fat of chickens and turkeys. Poult. Sci.
35:1246-1251.

Miller, D., E. H. Gruger. Jr., k. C. Leong, and G. M.
Knobl, Jr.

1967a. Effect of refined menhaden oils on the flavor and

fatty acid composition of broiler flesh. J. Food Sci.

32:342-345.
1967b. Dietary effect of menhaden oil ethyl esters on the

fatty acid pattern of broiler muscle lipids. Poult. Sci.

46:438-444.

PippEN, E. L., AND M. Nonaka.

1963. Gas chromatography of chicken and turkey volatiles:
The effect of temperature, oxygen, and type of tissue
on composition of the volatile fraction. J. Food Sci. 28:
334-341.

TARR, H. L. a.

1954. The Maillard reaction in flesh foods. Food Tfechnol.
8:15-19.

Webb, J. E., C. C. Brunson, and J. D. Yates

1972. Effects of feeding antioxidants on rancidity de-
velopment in pre-cooked, frozen broiler parts. Poult. Sci.
51:1601-1605.

1973. Effects of feeding fish meal and tocopherol on the
flavor of precooked, frozen turkey meat. Poult. Sci.
52:1029-1034.

Webb, R. W., W. W. Marion, and P. L. Hayse.

1972. Tocopherol supplementation and lipid stability in
the turkey. J. Food Sci. 37:496.

BIOLOGY OF FIVE SPECIES OF SEAROBINS (PISCES,
TRIGLIDAE) FROM THE NORTHEASTERN GULF OF MEXICO

Thomas C. Lewis and Ralph W. Yerger*

ABSTRACT

Geographically, Gulf populations oi Prionotus alatus appear to be restricted almost exclusively to
the eastern portion of the Gulf of Mexico, while Be//ator militaris, P. martis, P. roseus, and P. stearnsi
occur over the entire Gulf Bathymetrically, P. martis is a shallow shelf species; B. militaris and P.
roseus, middle shelf species; P. alatus, middle to deep shelf species; P. stearnsi, deep shelf species.
The size (standard length) of fi. militaris, P. alatus, P. martis, and P. roseus showed a significant
positive correlation with increasing depth of capture. Bellator militaris showed a significant "prefer-
ence" for fine sandy silt, clay, or mud bottoms. Prionotus stearnsi was captured in significantly
greater numbers during daytime trawling and is postulated to swim actively in the water column at
night. It appears to spawn from late summer to fall or early winter, while the remaining species spawn
from fall to spring or early summer. Adult P stearnsi differed in food habits by consistently consum-
ing relatively large fishes, while juveniles of this species and all the age groups of the other four
species fed consistently on crustaceans.

Searobins of the family Triglidae are commonly
taken in shrimp trawls along the coast of the Gulf
of Mexico where they comprise an important ele-
ment of the benthic shelf ichthyofauna (Miles
1951; Hildebrand 1954; Springer and Bullis 1956;
Bullis and Thompson 1965; Roithmayr 1965;
Franks et al. 1972). They are not commercially
important in the Gulf of Mexico, but at least some
species are included among the bottomfishes
that are canned for pet food and reduced for fish
meal by commercial Gulf fisheries (Roithmayr
1965). Triglids also present a rich source of food
for the larger, commercially important fishes from
the Gulf Prionotus ophryas, P. roseus, and P.
stearnsi have been found in the stomachs of red
snapper, Lutjanus campechanus , taken off Pensa-
cola, Fla. (Jordan and Swain 1885; Jordan and
Evermann 1887). Prionotus roseus was reported
from the stomachs of red grouper, Epinephelus
morio, off Tampa, Fla. (Jordan and Evermann
1887). Hildebrand (1954) regarded P stearnsi as
one of the most important forage fishes in the
western Gulf where it was noted in the stomachs
of rock sea bass, Centropristis philadelphica; red
snapper; sand seatrout, Cynoscion arenarius; and
inshore lizardfish, Synodus foetens.

Despite their importance as forage fishes, few or
no data are available on the biology of the Gulf
species, particularly on those found in deeper

^Department of Biological Science, Florida State University,
Tallahassee, FL 32306.

water. What little is known appears widely scat-
tered in the literature, usually in faunal lists. The
only in-depth studies on the biology of western
North Atlantic triglids (Marshall 1946; McEach-
ran and Davis 1970) are on the two species (Pri-
onotus carolinus and P. evolans) that do not occur
in the Gulf.

Our study was undertaken to analyze the spe-
cies composition of the northeastern Gulf triglid
fauna on the continental shelf between 20 and
190 m, to determine the distribution and abun-
dance of this fauna, and to investigate aspects of
their biology. Thirteen species (Bellator brachy-
chir, B. egretta, B. militaris, Prionotus alatus, P.
martis, P. ophryas, P. paralatus, P. roseus, P. rubio,
P. salmonicolor, P. scitulus, P. stearnsi, P. tribulus)
were collected, but only five species (B. militaris,
P. alatus, P. martis, P. roseus, P. stearnsi) were
taken in sufficient numbers to report on their
biology

MATERIALS AND METHODS

Specimens were collected from July 1969 to
October 1971 aboard the RV Tursiops and the
USNS Lynch. Most cruises were conducted
aboard the Tursiops from October 1970 to Octo-
ber 1971 as part of the "Gulf Shelf Project"
conducted by the Edward Ball Marine Labora-
tory, Department of Oceanography, Florida State
University. Fishes were captured in a 16-foot
(4.9-m) try-net otter trawl with a %-inch (1.9-cm)

Manuscript accepted August 1975.
FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

FISHERY BULLETIN: VOL. 74, NO. 1

square mesh body and a Vs-inch (0.3-cm) square
mesh cod end Hner.

The study area extended along the northeast-
ern Gulf of Mexico from east of the Mississippi
River Passes, La., to the w^estern edge of Apala-
chee Bay, Fla., over a depth range of 20 to 190 m
(Figure 1). The easternmost stations (between
long. 84°37'W and 85°30'W) were visited, with
few exceptions, in October and December 1970,
and January, April, May, July, August, and
September 1971. The remaining stations were
visited only once during cruises conducted in one
of the following months: July, October, and De-
cember 1969; October and November 1970; Janu-
ary, February, April, July, and October 1971.
Station locations were determined through loran.
Station depth was recorded from fathometer
readings. Depths for a few stations were extrapo-
lated from soundings recorded for that location
on "1100 Series" U.S. Coast and Geodetic Survey
maps. (For complete station data and specimens
examined see Lewis 1973.) The principal investi-
gators of the Gulf Shelf Project determined the
sampling regime for each station. One trawl
sample was taken at each station. Trawling time
on the bottom ranged from 10 to 60 min. The time
duration for the majority of trawls at shallow
stations (i.e. less than 90 m) was 10 min; for the
deeper stations, 20 min. In order to standardize
these trawling efforts, catches were recorded as
number of fish collected per 10 min trawling

(catch per unit effort), and transformed [Y - log
{X + 1)] for analysis of the variance. Data for all
stations (when available) were used for analysis
regardless of whether or not the particular spe-
cies was present.

Bottom temperature was recorded for most
stations by bathythermograph and on a few
occasions either by expendable bathythermo-
graph or reversing thermometers. Bottom type
was determined by examination of samples taken
in a bucket dredge dragged over the trawl area.
Bottom type was divided into two major classes;
coarse sand overlain with shell hash (type I), and
fine sandy silt, clay, or mud (type ID. Data for
bottom type were not collected at some stations
and consequently fishes taken at these stations
were not used in analysis of bottom type. Night
was considered to be that interval of time be-
tween 1 h after sunset and 1 h before sunrise at
that time of the year, while day was considered to
be between 1 h after sunrise and 1 h before
sunset. Fish collected at dawn or dusk were not
used in the analysis of time of capture data.

The standard length (SL) of each fish was
measured to the nearest millimeter. Identifica-
tions were made following Ginsburg (1950),
Miller (1965), and Miller and Kent (1971). Speci-
mens were preserved in 10% Formalin^ origi-

^Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

Figure l.— Map of the study area sampled by the RV Tursiops and USNS Lynch between July 1969 and October 1971.

LEWIS and YERGER: BIOLOGY OF FIVE SEAROBINS

nally, transferred to 40*^ isopropyl alcohol and
deposited in the Florida State University
collection.

Gonads were examined from specimens taken
in October, November, and December 1970, and
January, February, April, May, July, August,
September, and October 1971. Size at sexual
maturity was determined by the first appearance
of ripe or developing ova in females and enlarged
testes in males. Females with numerous ripe ova
were judged to be ready to spawn at or very near
the date of capture. A ripe egg was determined to
be one that was transparent and filled with
numerous oil globules. Its size was measured to
the nearest 0.1 mm with an ocular micrometer.

Stomachs (including the posterior esophagus)

were removed and the contents analyzed for
identifiable remains. Food items were identified
at least to class, and where possible to order and
suborder. The importance of food taxa was judged
by their numerical abundance.

RESULTS

Bellator militaris (Goode and Bean)
Horned Searobin

Bellator militaris was collected widely at
depths of approximately 20 to 100 m (Figure 2a)
and temperatures of 15° to 28°C. Specimens
ranged in size from 24 to 111 mm SL. This species
showed the greatest density of all the species

FIGURE 2.— Distribution within the
study area of: A, Bellator militaris
and B, Prionotus alatus.

FISHERY BULLETIN: VOL. 74, NO. 1

caught, yielding 1.8 specimens per 10 min trawl-
ing within its depth range (Table 1).

This species was most abundant between 80
and 90 m (Figure 3a). There was a gradual
increase in abundance to this depth range fol-
lowed by a sharp decrease. There was a sig-
nificant (P<0.001) positive relationship between
increasing size and increasing depth of capture.

A statistically greater (P< 0.025) number of B.
militaris were taken over a fine sandy mud, silt,
or clay bottom (Table 2). There was no statistical
difference in catch for day versus night trawling
(Table 3).

Bellator militaris appeared to reach sexual
maturity at about 65 mm SL in both sexes. The
spawning season was protracted as indicated by
the presence of females with numerous ripe ova
(0.7 to 0.9 mm in diameter) from November 1970
to July 1971.

Bellator militaris fed primarily on crustaceans
(90 to 95% of the total stomach contents). Juve-
niles (Table 4) appeared to feed primarily on
amphipods and natantian decapods; adults (Ta-
ble 5), on natantian decapods, amphipods, and
mysids. Adults also fed to a lesser extent on very
small fishes (usually less than 15 mm SL),
polychaetes, bivalves, and gastropods.

Table l. — Number of specimens of five species of triglids
collected and the mean number of fish per 10 min trawling.

No. of

Mean no. per

Species

specimens

10 min trawling'

Bellator militaris

277

1.8

Prionotus alatus

162

1.0

P martis

109

1.2

P. roseus

162

1.2

P. stearnsi

113

0.7

s -

4 -

7 -

(

2 -

7 -

2 -

40 n

I I — ^j—^ — I — ^ — — I — ^ — — r

1 4

(

1 4

1 3

'For trawls within the depth and geographic range of the species.

1
2S

1
D

1 1
10S

EPTH (m

Iters)

I4i

1
US

Prionatus alatus Goode and Bean
Spiny Searobin

With one exception, all specimens of P. alatus
were collected east of the De Soto Canyon (Figure
2b). For this reason all analyses of this species
were based only on data from stations east of the
Canyon. Sizes ranged from 24 to 140 mm SL;
collection depth, from 40 to 190 m; temperature,
from 14° to 28°C. Prionotus alatus ranked fourth
in density over its depth and geographic ranges
(Table 1).

Prionotus alatus appeared to be most abundant
around the 80- to 90-m interval of its depth range
(Figure 3b). There was a rapid increase in catch to

Figure 3. — Relationship of depth of capture versus catch per
unit effort. A. Bellator militaris, B. Prionotus alatus, C. Prio-
notus roseus, D. Prionotus martis, and E. Prionotus stearnsi.
Number above each bar refers to the number of 10-min trawl-
ing intervals at that particular depth.

this point followed by a gradual decline. As in
B. militaris, there was a significant (P<0.001)
positive relationship between increasing size and
increasing depth of capture.

There were no statistical differences in catch
per unit efforts between bottom types (Table 2)
and between day and night (Table 3).

Prionotus alatus appeared to reach sexual ma-
turity at about 100 mm SL for both sexes. Females
with numerous ripe ova (0.8 to 1.0 mm in diam-

LEWIS and YERGER: BIOLOGY OF FIVE SEAROBINS

Table 2. — The relationship of bottom type to density for five species of trigHds.

Type P

Type in

Mean no^

per

Mean no.

per

Species

min trawling

Variance

2/V

mln trawling

Variance

2/V

Bella tor militaris

0.8

4.7

3.1

37.7

6.3*

Prion otus alatus

2.2

0.7

2.3

0.6

P. martis

1.7

27.5

0.6

1.3

0.3

P roseus

1.4

8.8

1.3

11.4

0.0

P. stearnsi

1.1

9.5

0.8

2.5

0.2

'Type I = coarse sand bottom overlain withi shell hash. Type II = fine sandy mud, silt or clay bottom.
^N = the number of 10-min trawling intervals within the depth and geographic range of the species
^For one factor analysis of the variance for data transformed to V = log(X + 1).
•Significant at P<;-0.025.

Table 3. — Comparison of day versus night trawling for five species of triglids.

Night

Day

Mean no.

per

an no.

per

Species

min trawling

Variance

10 min trawling

Variance

Bellator militaris

1.9

30.5

2.1

24.3

0.1

Prionotus alatus

1.4

14.9

0.8

3.4

0.3

P. martis

2.0

27.0

1.1

4.7

1.8

P roseus

1.0

5.6

0.9

9.4

0.4

P. stearnsi

0.2

1,7

1.6

7.9

16.5-

'A/ = the number of 10-min trawling intervals within the species' depth and geographic range
^For one factor analysis of the variance for data transformed to V = log(X + 1).
•Significant at P< 0,001.

Table 4. — Percent of total stomach contents for the juveniles
of five species of searobins (n = the number of stomachs that
contained identifiable remains).

Table 5. — Percent of total stomach contents for the adults of
five species of searobins ( n = the number of stomachs exam-
ined that contained identifiable remains).

Senator

Prionotus

Bellator

Prionotus

militaris

alatus

martis

roseus

stearnsi

militaris

alatus

martis

roseus

stearnsi

Taxa

n = 15

n = 24

n = 5

n = 14

n = 10

Taxa

n = 59

n = 30

n = 25

n = 54

n = 14

Crustacea:

Ostracoda

1.1

3.0

—

Ostracoda

1.2

0.3

—

0.1

—

Copepoda

7.9

—

2.2

Copepoda

3.9

—

Stomatopoda

—

10.4

—

Stomatopoda

2.4

4.3

—

1.5

—

Amphipoda

41.6

16.4

38,4

102

86.8

Amphipoda

30.1

2.5

21.1

3.0

4.5

Isopoda

—

3.0

7,7

—

2.2

Isopoda

1.7

1.8

—

0.5

—

Mysidacea

4.5

13.4

—

13.3

—

Mysidacea

18.6

5.0

6.4

4.7

—

Decapoda:

Natantia

25.8

31.3

7.7

71.6

2.2

Natantia

30.9

71.3

39.8

82.4

9.1

Reptantia

7.9

9.0

23.1

1.8

—

Reptantia

5.2

11.0

10.5

4.5

9.1

Megalops

1.1

4.5

—

4.4

Megalops

0.9

1.0

1.2

—

9.1

Zoea

—

Zoea

0.5

—

Annelida:

Polychaeta

1.5

23 1

1.3

—

Polychaeta

0.8

—

4.6

2.0

—

Mollusca:

Bivalvia

3.4

1,5

—

1.8

—

Bivalvia

1.5

0.3

0.6

0.2

—

Chordata:

Gastropoda

0.3

—

0.1

—

Verlebrata

Cephalopoda

—

4.5

Osteichthyi

3S —

6,0

—

2.2

Echinodermata:

Ophiuroidea

—

1.2

0.2

—

Chordata:

Cephalochordata

11.7

—

eter) were

collected from November 1970 to April

Vertebrata:

1971. No i

females we

!re collei

cted in

Mav n

r June

Osteichthyes

2.0

2.5

2.9

0.8

63.7

and those collected in July 1971 were not ripe,
indicating that spawning ceased somewhere dur-
ing this interval.

Prionotus alatus fed primarily on crustaceans
(91 to 97% of total stomach contents). Juveniles
(Table 4) fed on decapods, amphipods, mysids, and
stomatopods; adults (Table 5), chiefly on decapods.
Small fishes (usually less than 15 mm SL) made
up the only substantive non-crustacean food item
in both adults and juveniles.

Prionotus roseus Jordan and Evermann
Bluespotted Searobin

Specimens of P. roseus ranging in size from 240
to 170 mm SL were collected throughout the study
area at depths of 20 to 90 m (Figure 4a) and bottom
temperatures of 16° to 28°C. It ranked, with P
martis, second in density within its depth range
(Table 1).

FISHERY BULLETIN: VOL. 74, NO. 1

Figure 4. — Distribution within the
study area of: A, Prionotus roseus
and B, Prionotus mortis (•) and
Prionotus stearnsi (A).

This species was most abundant between 60
and 70 m. As with the previous two species, P.
roseus showed a significant (P< 0.001) positive
relationship between increasing size and increas-
ing depth of capture.

There were no statistical differences in catches
between bottom types (Table 2) or between night
and day collections (Table 3).

Prionotus roseus appeared to reach sexual ma-
turity at 100 mm SL for both sexes. Spawning
period was protracted. Females with numerous
ripe ova (0.7 to 0.8 mm in diameter) were col-
lected from December to May 1971.

Prionotus roseus also fed primarily on crusta-
ceans (97% of the total stomach contents). Juve-
niles (Table 4) fed chiefly on decapod shrimp,

mysids, and amphipods; adults (Table 5) even
more exclusively on decapods.

Prionotus martis Ginsburg
Barred Searobin

Prionotus martis was collected widely except at
the western edge of the study area at depths of
approximately 20 to 45 m (Figure 4b). Sizes of
specimens ranged from 51 to 159 mm SL and
bottom temperature fi-om 17° to 28°C. Prionotus
martis ranked, with P. roseus, second for density
within its depth range (Table 1).

Prionotus martis was most abundant at the
20- to 30-m interval of its depth range (Figure
3d). As was the case forB. militaris, P. alatus, and

LEWIS and YERGER: BIOLOGY OF FIVE SEAROBINS

P. roseus, this species showed a significant
(P<0.001) positive relationship between increas-
ing size and increasing depth.

No statistical differences in catch per unit effort
between bottom types (Table 2) or night and day
collections (Table 3) were observed.

Determination of size at sexual maturity in P.
martis was inexact due to a paucity of specimens
less than 100 mm SL. Individuals of both sexes at
100 mm SL were mature, while nine specimens
below this size were immature. Consequently 100
mm SL was tentatively given as the size at sexual
maturity for both sexes. Likewise, the exact
spawning season for this species was difficult to
determine. Females with numerous ripe ova (0.6
mm in diameter) were collected from October to
December 1970. A large sample of females in
January 1971 contained no ripe individuals,
while a sample from April 197 1 contained one ripe
female.

Prionotus martis fed primarily on crustaceans
but not as extensively as the previous three
species (around 80% of the total stomach con-
tents). Juveniles (Table 4) appeared to feed heav-
ily on amphipods, polychaetes, and decapod crabs;
adults (Table 5) on decapod crabs and shrimp,
amphipods, and cephalochordates. The only
other important food items for adults were poly-
chaetes and very small fishes (usually less than
15 mm SL).

Prionotus stearnsi Jordan and Swain
Shortwing Searobin

Prionotus stearnsi was collected widely at
depths of approximately 60 to 185 m (Figure 4b)
and temperatures from 14° to 21°C. Specimens
ranged in size from 11 to 117 mm SL. It ranked
fifth in density within its depth range.

Prionotus stearnsi was fairly evenly distributed
within its depth range, but was slightly more
abundant at shallower depths (Figure 3e). Unlike
the previous four species, there was no significant
relationship between increasing size and increas-
ing depth of capture.

There was no significant difference in catch be-
tween bottom types (Table 2). There was, however,
a significantly (P<0.001) greater catch during
daytime trawling (Table 3).

Prionotus stearnsi appeared to reach sexual
maturity at about 60 mm SL in both sexes. No ripe
females were collected during the 1970-71 season
and only one female in October and two in

December of 1969 contained numerous ripe ova
(0.6 mm in diameter).

This species appeared to have different feeding
habits between adults and juveniles. The latter
(Table 4) fed primarily on small crustaceans (98%
of the number of food organisms), the former
(Table 5) chiefly on relatively large fishes (usu-
ally larger than 25 mm SL; 64% of the number of
food organisms). The only other important food
among adults was decapod crustaceans.

DISCUSSION

Geographic Distribution

Four of the five species {B. militaris, P. martis,
P. roseus, P. stearnsi) have been previously re-
corded over the entire northern Gulf of Mexico
(Ginsburg 1950; Springer and BuUis 1956; Bullis
and Thompson 1965; Burns 1970; Franks et al.
1972). Prionotus alatus has been reported almost
exclusively from east of the De Soto Canyon, but
Ginsburg (1950), Burns (1970), Miller and Kent
(1971), and Franks et al. ( 1972, based on the same
two specimens examined by Burns) reported
small numbers west of the Canyon. Our study
confirms this distribution and we conclude that
P. alatus is quite rare in the western portion of
the northeastern Gulf, where it is replaced by
P. paralatus.

Depth Distribution

The triglids collected in this study fit into four
bathymetric categories: 1) shallow shelf and in-
shore species, 2) shallow shelf to midshelf species,
3) shallow to deep shelf species, and 4) midshelf to
deep shelf species.

Prionotus martis is a shallow shelf and inshore
species. Springer and Bullis (1956) reported it
from 200 fathoms (366 m) but we feel that this
record is based on either a misidentification or
incorrect station data. All other specimens in
their paper came from 25 fathoms (46 m) or less.
The maximum depth for our study, 44 m (24
fathoms), is probably the maximum depth
reached by this species. It also enters shallow
water, being reported from 6 fathoms or less by
Reid (1954), Bulhs and Thompson (1965), Rich-
mond (1968), and Hastings (1972).

Bellator militaris and P. roseus fall into the
second category; the maximum depth for both
species was about 90 to 100 m. However, B.

FISHERY BULLETIN: VOL. 74, NO. 1

militaris appears to reach 100 fathoms (183 m)
off southwestern Florida (Longley and Hilde-
brand 1941; Springer and Bullis 1956; Moe and
Martin 1965) as does P. roseus (Springer and
Bullis 1956). Bellator militaris has been recorded
by Bullis and Struhsaker (1970) from the 100- to
150-fathom (180- to 270-m) interval in their Ca-
ribbean study, and at 100 and 1,175 fathoms ( 180
and 2,150 m) in the northern Gulf by Springer and
Bullis (1956). The latter figure is likely wrong.
Since neither species was collected at 100 fathoms
(183 m) in the present study despite intensive
collecting at this depth, we conclude they rarely if
ever reach this depth in the northeastern Gulf
Both are seldom recorded from less than 20 m.
Moe and Martin (1965) recorded B. militaris in
less than 3 fathoms (5.5 m) and P. roseus from
approximately 6 fathoms (11 m) off Tampa, Fla.

Miller and Kent (1971) gave the depth range for
P. alatus as 30 to 250 fathoms (55 to 457 m)
which would place it in the shallow to deep shelf
category. Our study reveals that this species
occasionally enters water shallower than 30 fath-
oms (55 m); two specimens were collected in 44 m
of water.

Our study indicates that P. stearnsi is a mid-
shelf to deep shelf species. Like that of P. alatus,
its depth range extends to deeper waters than
those found in our study area. Excluding the
armored searobins (which are often placed in
Triglidae), it is one of the deepest dwelling west-
ern North Atlantic triglids. Ginsburg (1950) ex-
amined specimens from 169 fathoms (309 m).
Bullis and Struhsaker ( 1970) reported it from the
150- to 200-fathom (274- to 366-m) interval.
Springer and Bullis (1956) reported P. stearnsi
from as deep as 250 fathoms (457 m, excluding the
same erroneous 1,175-fathom station reported for
B. militaris). Prionotus stearnsi has also been
recorded from shallower waters. Ginsburg ( 1950)
listed specimens from 13 fathoms (24 m), Hilde-
brand (1954) from 12 fathoms (22 m), and
Springer and Bullis (1956) from 5.5 fathoms (10
m), though this last figure is based on a field
identification and is subject to error. We never
collected P. stearnsi at depths less than 60 m
despite intensive collecting and conclude that it
rarely enters shallower waters in the northeast-
ern Gulf

Size-Depth Relationship

In their study in Gulf waters off Pinellas
100

County, Fla., Moe and Martin ( 1965) reported that
larger specimens of various fishes consistently
occurred at deeper depths. They pointed out that
this phenomenon had been noted before and was
correlated with increasing salinity (e.g. Gunter
1945). However, they were unable to draw such a
correlation, since salinity changed so little over
their study area. Topp and Hoff (1972) showed
statistically significant increases in the mean
size of Syacium papillosum (a bothid) collected
between 18 and 37 m and between 37 and 55 m
off southwestern Florida. Our results point to
similar conclusions. We found a highly signifi-
cant (P<0.001) positive relationship between in-
creasing size and increasing depth of capture for
all species except P. stearnsi. We concur with Moe
and Martin (1965) that this is not correlated
with salinity changes (which are small in our
study area).

Temperature

The four species in the first three bathymetric
categories occurred over a wide range of tempera-
tures. The only species that could in any way be
restricted by the temperature of its environment
is P. stearnsi, the deep shelf species, which was
taken over a limited range from 14° to 21°C.

Bottom Type

Bellator militaris was the only species which
showed any significant bottom type preference; it
was found in greater abundance over fine sandy
mud, silt, or clay bottoms. We conclude that
bottom type, at least as categorized in this study,
does not play a very important part in the dis-
tribution of four of the five species studied.

Time of Capture

Only one species, P. stearnsi, showed a sig-
nificant difference in the catch per unit effort
between day and night trawls; it was more abun-
dant in daytime trawls. Bellator militaris and P.
roseus were equally abundant in both day and
night trawls, while f! alatus andf! martis tended,
though not conclusively so, to be caught in
greater numbers at night. Hoese et al. (1968)
noted that P. tribulus crassiceps as well as other
unidentified triglids tended to be caught more
frequently at night, though not significantly so.

LEWIS and YERGER: BIOLOGY OF FIVE SEAROBINS

The occurrence of P. stearnsi in such greater
numbers during the day is difficult to explain.
Two opposing hypotheses can be postulated.
First, P. stearnsi may be a diurnal species, active
over the bottom during the day, and perhaps
burrowing during the night and thus eluding
capture. Or second, P. stearnsi may be nocturnal;
during the day it may rest on the bottom exposed
to daytime trawls, while at night it may ascend
into the water column to feed beyond the reach of
the trawl. We favor the second possibility be-
cause of the general physiognomy of this species.
Food habits, as will be discussed, also suggest a
more actively swimming existence compared
with other triglids.

Reproduction

Sexual Maturity

Bellator militaris and P. stearnsi are rather
small triglids maturing at 65 and 60 mm SL and
reaching a maximum size around 120 and 135
mm, respectively (Ginsburg 1950). Prionotus
alatus, P. martis, and P. roseus mature at about
100 mm SL and attain at least 189 mm (Ginsburg
1950), 166 mm (Reid 1954), and 225 mm (Gins-
burg 1950), respectively. Marshall (1946) found
that P. carolinus and P. euolans mature at about
140 and 180 mm SL, respectively, and attain a
much larger size than any Gulf species. It ap-
pears that the size at sexual maturity is largely

a function of the size attained by the particu-
lar species.

Spawning Season

Spawning seasons for the triglids collected in
this study can be separated into two ill-defined
categories: 1) Late summer to fall or early win-
ter and 2) late fall to spring or summer (see
Figure 5).

Prionotus stearnsi appears to fit into the first
category. In our study ripe females were collected
only in October and December. Longley and
Hildebrand (1941) reported collecting a ripe fe-
male in August off the Tortugas. These limited
data and a large number of very small specimens
in collections from October, December, and Janu-
ary indicate that P. stearnsi probably spawns
from late summer to late fall or early winter. The
paucity of ripe females suggests that this species
may spawn at greater depths than those sampled
in this study.

Three of the remaining four species {B. mili-
taris, P. alatus, P. roseus) had obviously pro-
tracted spawning seasons from fall to late spring
or summer. The presence of a number of small
individuals collected throughout the year further
corroborated the length of the reproductive
period.

Prionotus martis was in spawning condition in
October, December, and April. The presence of
only a few juveniles in this study leads us to

Figure 5. — Spawning seasons for
five species of searobins.

101

FISHERY BULLETIN: VOL. 74, NO. 1

believe that the young develop in shallow^er
water. The bulk of spawning appears to take
place from late fall to late winter or early spring
since all specimens less than 45 mm SL that we
have examined came from March and April
collections from water less than 20 m deep. Also,
Hastings (1972) collected small specimens of
P. martis during February to April only (great-
est abundance in April) during his seasonal stud-
ies of the jetty fauna at Destin and Panama
City, Fla.

Food Habits

Rapid retrieval of the trawl from the bottom
often resulted in eversion of stomachs, especially
in the deeper water species. Hence, analysis of
food habits was impeded by small sample sizes.
Also, the use of numerical abundance of taxa to
determine dietary preferences presents an obvi-
ous bias. Large numbers of small individuals
would appear dominant when, in fact, they might
make only a small percentage of the volume of
food consumed. This was the case in the domi-
nance of amphipods in the stomachs of juvenile fl
stearnsi. In general, however, individuals of the
numerically dominant taxa tended to be domi-
nant in size also.

On the basis of these limited data, four of the
five species (B. militaris, P. alatus, P. martis, P.
roseus) and the juveniles of the fifth {P. stearnsi)
appear to feed primarily on benthic crustaceans
and other benthic organisms. Reid (1954) and
Springer and Woodburn (1960) examined P sci-
tulus latifrons and P. tribulus crassiceps from the
northeastern Gulf and also found that both spe-
cies fed primarily on crustaceans. Likewise, Mar-
shall (1946) found the same to be true for P.
carolinus and P. evolans from the Atlantic coast.

In contrast, the adults of P stearnsi appear to
consume primarily other fishes. The food habits
of the adults of this species are different from all
other western North Atlantic triglids examined.
Its piscivorous habit lends support to our earlier
contention that this species is more mobile than
its congeners. This type of diet would imply an
active pursuit of their prey.

The fusiform shape of this species also implies
an active mode of existence. The head of P.
stearnsi with its terminal mouth does not appear
to be adapted for bottom feeding. The free rays of
the pectoral fins are more slender and less de-
veloped; they likely are not used extensively as

tools for searching along the bottom as in other
triglids.

ACKNOWLEDGMENTS

We thank Patrick M. McCaffrey for the sugges-
tion that initiated this study and for supplying us
with and helping to collect most of our specimens.
We express our gratitude to George C. Miller,
who confirmed some of our identifications early
in this study, Robert W. Hastings, Christopher C.
Koenig, and Robert L. Shipp, who provided us
with their assistance and encouragement. Ship-
time aboard the RV Tursiops for the Gulf Shelf
Project cruises was funded through National
Science Foundation Contract No. GD-28174, Pat-
rick M. McCaffrey, principal investigator. The
senior author received support through a Na-
tional Science Foundation Trainee Fellowship
during the study.

LITERATURE CITED

BULLIS, H. R., JR, AND P. J. STRUHSAKER.

1970. Fish fauna of the western Caribbean upper slope.
Q. J. Fla. Acad. Sci. 33:43-76.
BULLIS, H. R., Jr., and J. R. THOMPSON.

1965. Collections by the exploratory fishing vessels
Oregon, Silver Bay, Combat, and Pelican made during
1956 to 1960 in the southwestern North Atlantic. U.S.
Fish Wildl. Serv., Spec. Sci. Rep. Fish. 510, 130 p.
BURNS, C.

1970. Fishes rarely caught in shrimp trawl. Gulf Res.
Rep. 3:110-130.

franks, j. s., j. y. christmas, w. l. siler, r. combs, r.
Waller, and C. Burns.

1972. A study of nektonic and benthic fauna of the
shallow Gulf of Mexico off the State of Mississippi as
related to some physical, chemical, and geological
factor. Gulf Res. Rep. 4, 148 p.
GINSBURG, I.

1950. Review of the western Atlantic Triglidae (fishes).
Tex. J. Sci. 2:489-527.
GUNTER, G.

1945. Studies on marine fishes of Texas. Publ. Inst.
Mar. Sci., Univ. Tex. 1, 190 p.
HASTINGS, R. W.

1972. The origin and seasonality of the fish fauna on a
new jetty in the northeastern Gulf of Mexico. Ph.D.
Thesis, Florida State Univ., Tallahassee, 555 p.
HILDEBRAND, H. H.

1954. A study of the fauna of the brown shrimp iPenaeus
aztecus Ives) grounds in the western Gulf of Mexico.
Publ. Inst. Mar. Sci., Univ. Tex. 3:233-366.

Hoese, h. d., b. J. copeland, F. n. moseley, and e. d. Lane.

1968. Fauna of the Aransas Pass Inlet, Texas. III. Diel
and seasonal variations in trawlable organisms of the
adjacent area. Tex. J. Sci. 20:33-60.

102

LEWIS and YERGER: BIOLOGY OF FIVE SEAROBINS

JORDAN, D. S., AND B. W. EVERMANN.

1887. Description of six new species of fishes from the
Gulf of Mexico, with notes on other species. Proc. U.S.
Natl. Mus. 9:466-476.
JORDAN, D. S., AND J. SWAIN.

1885. Description of three new species of fishes (Prio-
notus stearnsi, Prionotus ophryas and Anthias vivans)
collected at Pensacola, Florida, by Mr. Silas Stearns.
Proc. U.S. Natl. Mus. 7:541-545.

Lewis, T. C.

1973. Biology of searobins (Pisces: Triglidae) from the
northern Gulf of Mexico. M.S. Thesis, Florida State
Univ., Tallahassee, 84 p.
LONGLEY, W. H., AND S. F. HILDEBRAND.

1941. Systematic catalogue of the fishes of Tortugas,
with observations on color, habits, and local distribution.
Carnegie Inst. Wash. Publ. 535:1-331. (Tortugas Lab.
Pap. 34.)
MARSHALL, N.

1946. Observations of the comparative ecology and life
history of two sea robins, Prionotus carolinus and Prio-
notus evolans strigatus. Copeia 1946:118-144.
MCEACHRAN, J. D., AND J. DAVIS.

1970. Age and growth of the striped searobin. Trans
Am. Fish. Sec. 99:343-352.
MILES, R. M.

1951. An analysis of the "trash fish" of shrimp trawlers
operating in Apalachicola Bay and the adjacent Gulf
of Mexico. M.S. Thesis, Florida State Univ., Talla-
hassee, 46 p.
MILLER, G. C.

1965. A new species of searobin (Triglidae). Q. J. Fla.
Acad. Sci. 19:259-266.

MILLER, G. C., AND D. M. KENT.

1971. A redescription of Prionotus beani (Pisces, Trig-
lidae). Q. J. Fla. Acad. Sci. 34:223-242.

MOE, M. A., JR. AND G. T. MARTIN.

1965. Fishes taken in monthly trawl samples offshore of
Pinellas County, Florida, with new additions to the fish
fauna of the Tampa Bay area. Tulane Stud. Zool.
12:129-151.
REID, G. K., Jr.

1954. An ecological study of the Gulf of Mexico fishes,
in the vicinity of Cedar Key, Florida. Bull. Mar. Sci.
Gulf Caribb. 4, 94 p.

RICHMOND, E. A.

1968. A supplement to the fauna and flora of Horn Island,
Mississippi. Gulf Res. Rep. 2:213-254.

ROITHMAYR, C. M.

1965. Industrial bottomfish fishery of the northern Gulf
of Mexico, 1959-63. U.S. Fish Wildl. Serv., Spec. Sci.
Rep. Fish. 518, 23 p.

SPRINGER, S., AND H. R. BULLIS, jR.

1956. Collections by the Oregon in the Gulf of Mexico.
U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 196,
134 p.

SPRINGER, V. G., AND K. D. WOODBURN.

1960. An ecological study of the fishes of the Tampa
Bay area. Fla. State Board Conserv., Mar. Lab., Prof
Pap. 1, 104 p.

TOPP, R. W., AND F. H. HOFF, JR.

1972. Flatfishes (Pleuronectiformes). Memoirs of the
Hourglass Cruises. Vol. IV. Part II. Fla. Dep. Nat. Re-
sour., Mar. Res. Lab., 135 p.

103

AN ACOUSTIC METHOD FOR THE HIGH-SEAS ASSESSMENT OF

MIGRATING SALMON ^

Gary Lord,^ William C. Acker,^ Allan C. Hartt,^ and Brian J. Rothschild"

ABSTRACT

A system of free-floating acoustic buoys with upward-looking transducers has been developed for use in
assessing high-seas salmon stocks. The transducers, operating at 120 kHz, are suspended 46 m below
the surface. The fish counts and the range to each fish are obtained in digital form, and the data are
radioed from each buoy to the tending vessel where the data are decoded and recorded on magnetic tape.
The present system consists of four buoys although the receiver-decoder system can accommodate up to
10 buoys operating synchronously.

The assessment of fish stocks is of obvious impor-
tance to all segments of the fishing industry in
planning their respective operations. A problem
of particular interest to the United States Section
of the International North Pacific Fisheries
Commission has been the assessment of imma-
ture sockeye salmon, Oncorhynchus nerka, which
occur in abundance each summer south of the
central Aleutian Islands in the North Pacific
Ocean. It has been found (Hartt 1962, 1966) that
immature sockeye salmon, mainly of Bristol Bay
origin, migrate westward through this area in
summer and that their relative abundance is re-
lated to the number of mature fish returning to
Bristol Bay the following year (Fisheries Re-
search Institute Staff 1960; Rogers 1972, 1973,
1974). This information has been used since 1960
as a means of forecasting the Bristol Bay run.
Because the size of the run may vary by a factor of
10, an accurate forecast with a lead time of nearly
a year is of obvious importance to the fishing and
canning segments of the industry. Mathews
(1966) has shown, by means of a comprehensive
model simulating the cannery portion of the
fishery, the relative value of run forecasts of vary-
ing precision. Run forecasts are also of value to
the fishery management agencies in setting pre-
liminary escapement goals and in planning their

'Contribution No. 438, College of Fisheries, University of
Washington, Seattle, WA.

^Fisheries Research Institute, University of Washington,
Seattle, WA 98195.

^Applied Physics Laboratory, University of Washington,
Seattle, WA 98195.

"Southwest Fishery Center, National Marine Fisheries Ser-
vice, NOAA, P.O. Box 271, La Jolla, CA 92038.

early season strategies to meet these anticipated
goals.

The assessment of immature fish at Adak Is-
land has been done by the Fisheries Research In-
stitute using a fine-mesh purse seine 400 fathoms
(730 m) long at a series of stations from 5 to 50
nautical miles off the southern shore of Adak Is-
land. From 1956 through 1967, no station pattern
was followed — purse seine sets were made ran-
domly, mainly in an area within 20 nautical miles
of shore. Since 1968, the fishing has been con-
ducted uniformly at five stations spaced at ap-
proximately 10 nautical mile intervals between 5
and 50 nautical miles offshore.

Although the purse seine is a useful tool for
providing information on abundance, species
composition, and age composition of the stocks
present, it suffers from several disadvantages as
a research tool. Its use is limited to periods of
moderate sea conditions resulting in significant
gaps in the time-space coverage in this particu-
larly stormy region. A maximum of five sets can
be made in a day under ideal conditions which
yields only 2y2 h of actual fishing. Also, seines
give no direct information on depth stratification
or schooling of the fish, and in areas where the
direction of migration of the fish is not uniform,
multiple sets are required to sample all of the
stocks present. Variability in direction of migra-
tion is not a serious problem in the area south of
the central Aleutian Islands because the direc-
tion of migration of immature salmon is uni-
formly westward (Hartt in press). In an effort to
overcome the sampling limitations of the purse
seine, the Fisheries Research Institute and the
Applied Physics Laboratory jointly developed an

Manuscript accepted August 1975.
FISHERY BULLETIN; VOL. 74, NO. 1, 1976.

104

LORD ET AL.: ACOUSTIC ASSESSMENT OF MIGRATING SALMON

acoustic assessment system that can be used
alone or in conjunction with the purse seine.
Some of the anticipated advantages of such a
system were that it could obtain abundance
estimates and swimming depths with around-
the-clock operation in a wide range of weather
conditions.

PRELIMINARY CONSIDERATIONS

The final configuration of the system was
determined, to a large extent, by consideration of
problems related to obtaining adequate numbers
of representative samples. The preliminary indi-
cations were that most of the fish of interest were
concentrated near the surface at a depth of 10 m
or less. This conclusion was based on the results
of experiments in which longline and gill net gear
were fished at various depths (Manzer 1964;
Machidori 1966; French et al. 1967) and also by
direct visual observation of salmon in the purse
seines. This concentration offish near the surface
precluded the use of a hull-mounted device since
such a system would necessarily exclude the top 3
or 4 m of the water column. This led to considera-
tion of a transducer suspended in some manner
below the main body of fish.

A transducer mounted on a towed platform
with a coaxial cable to the towing vessel was con-
sidered initially but was abandoned because of
the anticipated difficulty of developing a platform
that could maintain depth and attitude stability
while maintaining position to the side of the ves-
sel. Since the extreme water depths precluded an
anchored system the approach eventually
adopted was to suspend the transducers from
free-floating surface buoys with self-contained
electronics. Consideration of the anticipated
sampling statistics indicated the need for a
multibuoy system which in turn suggested radio
telemetry of the data from the buoys to a central
shipboard receiver and recorder This is the type
of system that was eventually constructed.

The sampling statistics of particular relevance
to the design and operation of the buoy system
concerned: a) the level of effort required to obtain
a specified precision in the estimation of the
density of the fish and b) the purse seine effort
required to obtain comparable precision in the
estimation of the species composition of the
population.

Extensive purse seining over a period of several
years indicated that the salmon were relatively

sparse and probably did not school or otherwise
interact to a significant degree. Under these con-
ditions the echo counts may be assumed to have a
Poisson distribution with a parameter, /jl, that is
proportional to the number density^ of the fish.
Thus we have.

M = PoVs

(1)

where V, is the sampling volume of a single
counter and po is the average fish density defined
so that fjL is the expected number of echoes per
acoustic pulse. If we assume large sample theory,
the minimum number of acoustic pulses required
to be 100a% confident that the relative error of
the estimate of po does not exceed e is given by,

M„

e'PoVs

(2)

where c?^, is the lOOa^f point (two-sided) of ^^(0,1).
The crucial feature of Equation (2) is that the
sampling effort must be increased as either po or
Vg decrease. Preliminary estimates of po based on
purse seine data, while quite crude, indicated
that V, should be as large as possible subject only
to the tradeoffs necessary to obtain an adequate
signal to noise ratio. Also, the need for multiple
buoys sampling mutually disjoint volumes was
indicated.

The high-seas salmon population generally
consists of a mixture of species so that it is neces-
sary to determine species composition by some
means. In the area south of the central Aleutians
significant numbers of chum salmon, O. keta,
occur mixed with immature sockeye salmon dur-
ing the sampling period, and occasionally pink,
O. gorbuscha; coho, O. kisutch; and chinook
salmon, O. tshawytscha, are present in small
numbers. The only nonsalmonid species gener-
ally found in this area, at the depths being sam-
pled, is the Atka mackerel, Pleurogrammus
monopterygius. This species generally occurs in
small numbers relative to the salmon so that, if
counted, it will not seriously affect the estimates
of the density of the salmon. Further, this species
does not have a swim bladder so that by the
proper choice of the detection threshold level
these fish will not be detected by the sonar.

^For echo counting the number density is the quantity of
interest. If acoustic echo integration is utilized, the density on
a mass basis is appropriate.

105

FISHERY BULLETIN: VOL. 74, NO. 1

Species and age discrimination by acoustic means
is not currently possible so that it is necessary to
obtain samples by purse seine in order to deter-
mine the species and age composition of the popu-
lation.

Since sockeye salmon is the only species of in-
terest, we may treat the purse seine samples as
binomial events in which the parameter of in-
terest is the proportion, p, of sockeye salmon
present in the population. If asymptotic normal-
ity is again assumed, it is found that the sample
size required to be 100a% confident that the
relative error of the estimate of p will not exceed
8 is given by,

n =

(3)

where d^ is that defined for Equation (2).

Equations (2) and (3), as indicated, provide
information on the sampling necessary to achieve
prescribed levels of precision in the population
estimates. A direct but somewhat crude compari-
son of the purse seine and the acoustic buoys may
be made on an area basis. The purse seine has a
nominal length of 400 fathoms or about 732 m.
The area swept out in a round haul^ is about
42,600 m^. For a transducer having a beam width
of 28° to the 3 dB points and suspended 46 m
below the surface, the area ensonified is approxi-
mately 390 m^. Thus the purse seine sweeps out
an area about 110 times as great as the area
ensonified by a single acoustic pulse. The pulse
interval is approximately 10 s. However, it has
been found that the individual fish remain in the
pattern for longer periods of time, typically about
30 s, although a precise estimate is not available
at this time. Thus, a single buoy would have to
operate for at least 1 h to obtain coverage equiva-
lent to a single round haul. The additional cover-
age obtained using 30-min tow hauls is not
known precisely but the limited data available
indicate a factor of two or three over the round
hauls. Thus, to provide coverage comparable to
that of the purse seine a single buoy would have
to operate for a minimum of 3 h. A comparable
sampling time is obtained using Equation (2)
with poV,. estimated using a typical seine haul of
150 fish. The seine hauls may vary from zero to

well over 1,000 fish from which it follows that the
time required for adequate acoustic samples may
vary, inversely, by corresponding amounts. The
sampling considerations just outlined played a
significant part in the choice of the hardware
configuration and the decision to utilize multiple
buoys.

SYSTEM DESIGN
AND CHARACTERISTICS

Figure 1 is a schematic illustration of the
high-seas assessment system showing only a
single buoy. In operation up to 10 buoys can be
deployed, each sending information to the ship-
board decoding and recording system. A four-
buoy system has been used at Adak to help assess
the migrating salmon population. A simplified
block diagram of the buoy system is given in Fig-
ure 2, and a photograph of the buoy is shown in
Figure 3. The buoy and shipboard system are dis-
cussed below.

The buoy contains an acoustic system which
gathers fish count and depth distribution data, a
logic system which processes and provides tempo-
rary storage for these data, and a telemetry sys-
tem which sends data to the monitoring ship. The
acoustic system operates at 120 kHz and samples
the population every 10 s. Sample rates can be
changed to 5-s or 2.5-s intervals if desired. The
system transmits a 200-/xs pulse (24 cycles at 120
kHz) at a source level of + 106 dB. Target returns
must be greater than a preset threshold (approx-
imately 2V) for at least 100 fjLS before they are
validated. This technique, and an adequate
source level to give a worst case'' signal-to-noise
ratio of 10 dB, minimizes false target counts.

Pulse elongation and amplitude testing tech-
niques are used to automatically adjust "end-of-
sample" so that surface returns and near surface
bubbles are not counted.

Measurements at the University of Washing-
ton and at Adak during summer operation have
shown that the "average" target size of the mi-
grating salmon is about - 30 dB within the aspect
angles encountered in the sample volume.

A typical plot of signal return versus aspect
angle from a single fish is shown in Figure 4. This
polar diagram shows target strength from the

*Purse seining is normally done in a standard manner using
tow hauls in which the seine is held open in a semicircle for
30 min before closing and pursing. In a round haul the seine
is set in a circle and pursed immediately after closing the circle.

'The worst-case condition exists for minimum target strength
( -45 dB) at maximum range (46 m) at the -3 dB point in the
transducer beam pattern.

106

LORD ET AL.: ACOUSTIC ASSESSMENT OF MIGRATING SALMON

RADIO LINK

BUOY WITH
TRANSMITTER

22.1-m SURFACE
SEARCH DIAMETER

TRANSDUCER

ventral, head, dorsal, and tail aspects. Inspection
of this figure shows that the target strength de-
creases rapidly for head or tail views but is fairly
constant at -30 dB over ±30° when viewed from
the dorsal or ventral aspect. This severe depen-
dence of target strength on aspect angle was the
limiting factor in the choice of transducer beam
width. For the high-seas system, a 28° conical
beam is used. This gives an adequate sample vol-
ume and minimizes target-size fluctuations to a
manageable level. A time-variable-gain (TVG)
receiver, adjusted so that its output for a particu-
lar target is independent of target range, is used
to limit signal dynamic range at the detector.
This technique, and proper adjustment of abso-
lute sensitivity, keeps the search volume rela-
tively constant over a fairly wide range of target
strength (±15 dB).

Estimates of the fish density in the Adak area
indicate that the average count per sample will
be less than one fish. Schooling habits of these
salmon also reveal that only rarely will more
than a few salmon be included in any one sample.

RECEIVING
ANTENNA

SHIPBOARD RECEIVER
a DIGITAL RECORDER

Figure l. — Schematic illustration of prototype buoy used at
Adak in 1974.

COAJtIAl.
CABLE —

TRANSDUCER

UNDERWATER
SYSTEM

TRANSMITTER

TRANSMITTER-
CONTROL
(REP RATE PULSE
LENGTH, ate)

FISH
DETECTOR

TELEMETRY
CONTROL a

TRANSMITTER

MASTER TIMING

AND DATA

PROCESSING

LOGIC

SERIAL

DATA

STORAGE

PREAMPLIFIER
AND TVG
AMPLIFIER

SURFACE

DETECTOR

BUOY ELECTRONICS

RADIO
RECEIVER

DATA

DETECTOR ft

SYNCHRON IZER

BUFFER
STORAGE

TIMING ft

CONTROL
LOGIC

Digital

TAPE

RECORDER

SHIPBOARD SYSTEM

Figure 2. — Simplified block diagram of the prototype high-
seas system.

107

FISHERY BULLETIN: VOL. 74, NO. 1

Figure 3. — Buoy deployed with detection and recovery gear.

180°

Figure 4. — Typical signal level as function of aspect angle
for a single fish.

Storage was therefore limited to include data
from a maximum of seven fish. The number offish
counted per sample and depth of each fish (to the
nearest meter) are stored in a serial shift register
memory, as are data on buoy identification and a
data synchronization code. These data are stored
in a format which makes telemetry noise and
false counts easily recognizable and therefore
easy to eliminate.

After all data from one acoustic pulse are
gathered and stored they are automatically

shifted through the telemetry system for trans-
mission to the monitoring ship. Frequency-shift-
keying (FSK) through the audio inputs of com-
mercially available transceivers is currently
used. The frequency response of the audio chan-
nels limits the bit rate to 100 Hz. A reliable te-
lemetry range of 15 km in fairly rough seas has
been achieved using this technique. A 6-MHz
telemetry system has been developed which will
increase the useful range to 100 km and will
allow the use of radio direction finding equipment
found aboard most seagoing vessels for buoy
recovery.

In the current configuration, the buoys will
operate continuously for 5 days before battery
recharging is necessary. The acoustic and logic
systems were carefully designed to minimize
average power drain. COSMOS elements were
used in logic design, and transmitter and receiver
standby current is very low. Battery life is there-
fore limited by the telemetry system. The rela-
tively low data rate requires that the transmitter
be on for 0.6 s/sample. However, the redesigned
telemetry system can increase data rate by an
order of magnitude which will increase buoy life
between charges to more than 6 wk.

The shipboard system consists of a telemetry
receiver, a data synchronizer with buffer storage,
a printer, and a digital tape recorder. Data from
up to 10 buoys can be received and processed at
the monitoring ship. Real time readout is pro-
vided by the printer. The digital tape recorder
provides data storage for later computer analysis.

FIELD OPERATIONS AND RESULTS

The acoustic buoy system has been operated in
the Adak area during the summers of 1972, 1973,
and 1974. The 1972 operation suggested sig-
nificant design changes in the electroacoustic
portion of the system. These modifications were
accomplished during the winter of 1972-73. The
results of the 1973 operation indicated that spe-
cial attention had to be given to system sensitiv-
ity and field calibration which was done prior to
the start of the 1974 field season. The present
configuration represents an essentially final de-
sign with only minor modifications to be made in
the future.

Whenever feasible, the acoustic buoys have
been operated at the same station and at the
same time as the purse seine in order to obtain
comparable data. This was not always possible.

108

LORD ET AL.: ACOUSTIC ASSESSMENT OF MIGRATING SALMON

however, since it was more convenient to operate
the buoys continuously for several hours whereas
the seine vessel required only 2 h for a set after
which it proceeded to the next station. Occasion-
ally the buoys were operated at a station which
had been fished by the seine on the same day but
not at the same time. Also, even at the same sta-
tion, it was not feasible to set the seine directly
around the buoys so it cannot be said that the two
gears sampled precisely the same water. This is
of some significance in any gear comparison
since there was considerable set-to-set variation
in purse seine hauls made at the same station.

Buoy launch and recovery presented no diffi-
culty in any weather conditions in which buoy
operation was attempted. Buoy operation is usu-
ally limited by the presence of heavy breaking
seas with whitecaps in which case the entrained
air causes ambiguous echo counts. In the Adak
area the limiting weather conditions for opera-
tion of either the purse seine or the acoustic buoys
depend strongly on the wind direction. Generally
the purse seine can be operated in winds up to a
maximum of about 20 knots. The acoustic buoys
have been operated in higher winds with no seri-
ous difficulty in launch or recovery. However, the
aforementioned problem of entrained air usually
limits buoy operation to winds of less than 25
knots. The buoys, however, can be operated con-
tinuously for longer periods of time since, once
deployed, no further human activity is required
except to monitor the digital printout.

The buoys operate synchronously so that the
data for each acoustic pulse may be radioed to the
tending vessel as soon as it is obtained. The echo
count data are in digital form in which all of the
data from each acoustic pulse is coded into a
single 60-bit word for telemetry to the shipboard
receiver. Each of these 60-bit words contains:
a) buoy identification number, b) the number of
echo counts up to a maximum of seven, and c) the
range from the transducer to each of the targets.
The data system requires that the indicated
number of targets agrees with the number of
ranges actually recorded and that the target
ranges must form a nondecreasing sequence. This
redundancy permits the detection and rejection of
spurious or noise contaminated data. The binary
coded 60-bit words are formatted to be compatible
with the CDC -6400 computer^ used for the data

reduction. The tape reading and data processing
can be accomplished using only FORTRAN and
certain FORTRAN callable subroutines thus
avoiding the necessity of machine language pro-
gramming.

The range discrimination of the acoustic sys-
tem is 25 cm, i.e., two fish separated in range
from the transducer by 25 cm or more will be
detected as individual fish. Six binary bits are
allowed for each of the seven possible ranges.
This presently corresponds to a range resolution
of 1 m, i.e., more than one fish may be detected
and recorded in a single 1-m range increment if
they are physically separated in range by at least
25 cm. Target coincidence is a possibility, particu-
larly if the fish are dense or tend to school. This
has not been a problem in high-seas use since the
average number of echoes per pulse has been of
the order of one.

Figure 5 shows typical depth distribution his-
tograms corrected for the effect of a conical sam-
pling volume. The most striking feature is the
shallow depth at which most of the fish are found,
usually 5 m or less. This had been anticipated and
illustrates the need for an upward-looking device.

There is the possibility of ambiguity in the in-
terpretation of echoes originating very near the
surface. Indeed this usually proves to be the
limiting condition in the operation of the buoys.
This situation manifests itself by the consistent
presence of targets in two or more successive
range increments below the surface. More detail

DENSITY OF FiSH (ARBITRARY UNITS)
0 01 0.2 03 04 0 Ol 0.2 03 04 05 06 Q7

I 6

t-
Ol
bi

12-

DATE 7/11/74
STATION 2
BUOY NO 4

DATE 7/17/74
STATION I
BUOY NO I

12-1

^Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

Figure 5. — Typical depth distribution histogram.

109

FISHERY BULLETIN: VOL. 74, NO. 1

near the surface can be obtained by increasing
the depth resolution to 0.5 m from the present
1 m. This increased depth resolution would
necessitate the elimination of the first 14 m of the
46-m water column above the transducer since
only six bits are available for each range word.
This is a desirable tradeoff, however, in view of
the concentration of the fish near the surface.

Figure 6 is a series of plots of the computed
areal densities of the salmon obtained by inte-
grating the depth distribution histograms over
the depth. Also plotted are the purse seine
catches which were obtained in reasonable time
and space proximity to the buoys. The data are
reasonably consistent although significant
departures occasionally occur. There are several
possible sources for the observed discrepancies:
a) set-to-set variations in seine hauls, b) similar
variations in the sonar counts, c) the inability
of the purse seine and the acoustic buoys to
sample precisely the same volumes of water,
and d) possible attraction or avoidance of the
acoustic gear by the fish. The variations within
gear types can be explained by the "patchiness"
of the salmon. The digital printouts tend to show
small groups offish, rarely giving more than three
echo counts, occurring with widely varying inter-
arrival times. This observation indicates the exis-
tence of relatively large areas that are nearly
devoid of fish thus explaining the occasional
twofold variations in successive seine hauls
made at the same station.

Sonar gear avoidance or attraction by the fish
is a potentially serious problem, the magnitude of

" /\

Stotion 1

station 2

5 8-

UJ N

si /

\ /^^/X""

l§-

\/ \\

\ /^~

\^" ^^-/

\ / \/

— -^

^^V^ ^ \

1 —

1 1 1 -^ —

30 5

JUNE

10 15 20 2530 5 10 I

JULY JUNE JULY

20 25

Slolion 3

Stotion 4

Figure 6. — Plots of computed relative areal densities ( )

and purse seine catches ( — ) by date and station for 1974.

which is not yet known. Occasional sea lions have
been observed around the buoys but they usually
departed after several minutes. Also, there is lit-
tle evidence to indicate that the fish are attracted
to the sonar gear since none of the observed
targets remains in the ensonified region for more
than a few pulses. Sonar gear avoidance is a more
likely prospect. The Stellar sea lion is common in
the area being sampled and it is a known pred-
ator of salmon. The sonar buoy is similar in size
to that of a sea lion so that avoidance is a distinct
possibility. Secchi disc readings of 15 m are typi-
cal so that the buoy or cable may be detected at
significant distances by the fish. Day versus night
data differ slightly but as yet there are too few
data on which to base a conclusion concerning
gear avoidance.

All of the acoustic data from which Figure 6
was obtained were pooled and the sample correla-
tion coefficient for the buoy-purse seine was com-
puted. A value of 0.547 was obtained which,
under the assumption of normality, is significant
at approximately the 0.5% level {t distributed
with n - 2 = 19 df). The results indicate that the
acoustic buoys can obtain statistically significant
population information as well as such ancillary
information as depth distribution and density
during both day and night. Additionally, indirect
information on schooling is available by observ-
ing the interarrival tirnes of the fish although
this has not been investigated in detail.

The design of the acoustic buoy system is essen-
tially fixed although modifications for use in
other situations are possible. For example, a bot-
tom anchored version for use in water depths of
about 100 m has been designed but fabrication
has not begun. Another possible design change is
in the radio telemetry system. The present sys-
tem, while reliable, is inefficient. An improved
system has been designed and will be fabricated
upon allocation of a suitable frequency by the
Federal Communications Commission.

The current approximate unit cost per buoy, in-
cluding the radio transmitter, is $6,000. The
shipboard receiver-decoder cost is approximately
$2,000. The tape recorder currently used is a
Kennedy Model 1400 digital incremental which
records at 556 bits/inch on Va-inch magnetic tape
on 10-inch reels. It is "off the shelf but is inter-
faced to the receiver-decoder. The interfacing cost
is approximately $1,000 which should remain
constant for interfacing to any digital incremen-
tal recorder.

110

LORD ET AL.: ACOUSTIC ASSESSMENT OF MIGRATING SALMON

LITERATURE CITED

Fisheries research Institute Staff.

I960. Collected material on forecast of Bristol Bay red
salmon run in 1960. Univ. Wash., Fish. Res. Inst. Circ.
122. [21 p.]

French, R. R., d. r. Craddock, and J. R. Dunn.

1967. Distribution and abundance of salmon. Int. North
Pac. Fish. Comm. Annu. Rep. 1965, p. 82-94.
Hartt, a. C.

1962. Movement of salmon in the North Pacific Ocean
and Bering Sea as determined by tagging, 1956-1958.
Int. North Pac. Fish. Comm. Bull. 6, 157 p.

1966. Migrations of salmon in the North Pacific Ocean
and Bering Sea as determined by seining and tagging,
1959-1960. Int. North Pac. Fish. Comm. Bull. 19, 141 p.

In press. Problems in sampling Pacific salmon at sea.
In Symposium on the evaluation of methods of estimat-
ing the abundance and biological attributes of salmon
on the high seas, p. 128-197. Int. North Pac. Fish.
Comm. Bull. 32.
MACHIDORI, S.

1966. Vertical distribution of salmon (Genus Oncorhyn-

chus) in the North-western Pacific. I. [In Jap., Engl.
Summ.] Hokkaido Reg. Fish. Res. Lab. Bull. 31:11-17.
MANZER, J. I.

1964. Preliminary observations on the vertical distribu-
tion of Pacific salmon (genus Oncorhynchus) in the
Gulf of Alaska. J. Fish. Res. Board Can. 21:891-903.

Mathews, S. B.

1966. The economic consequences of forecasting sock-
eye salmon {Oncorhynchus nerka Walbaum) runs to
Bristol Bay, Alaska: A computer simulation study of the
potential benefits to a salmon canning industry from
accurate forecasts of the runs. Ph.D. Thesis, Univ.
Washington, Seattle, 238 p.
ROGERS, D. E.

1972. Forecast of the sockeye salmon run to Bristol
Bay in 1972, based on purse seine catches of im-
mature sockeye salmon south of Adak. Univ. Wash.,
Fish. Res. Inst. Circ. 72-3, 32 p.

1973. Forecast of the sockeye salmon run to Bristol Bay
in 1973. Univ. Wash., Fish. Res. Inst. Circ. 73-1, 33 p.

1974. Forecast of the sockeye salmon run to Bristol
Bay in 1974. Univ. Wash., Fish. Res. Inst. Circ. 74-1,
30 p.

Ill

ANALYSIS OF RETURNS OF TAGGED GULF MENHADEN

Paul J. Pristas,i Eldon J. Levi,^ and Robert L. Dryfoos^

ABSTRACT

From 1969 to 1971 nearly 76,000 adult Gulf menhaden, Brevoortia patronus, were tagged in the
northern Gulf of Mexico with internal metallic tags. From an estimated 28,000 recaptures it was con-
cluded that there is little east- west movement of adult Gulf menhaden during the fishing season from
April to October, and that there is little mixing of menhaden from different areas when fish move off-
shore during the winter. Total mortality appears to be high, but could not be estimated from the re-
turns. Few Gulf menhaden survive more than 3 yr.

Menhaden are industrial fish that are processed
into meal, oil, and solubles. From 1964 to 1973,
the annual purse seine catch of Gulf menhaden,
Brevoortia patronus, which support the largest
fishery in the United States, ranged from 316,000
to 728,000 metric tons. Scientists at the Atlan-
tic Estuarine Fisheries Center, National Marine
Fisheries Service, NOAA, Beaufort, N.C. have
been studying the fishery since 1964.

A scientifically interesting question, as well as
one of practical importance from the standpoint
of resource management, is whether Gulf men-
haden make extensive coastal movements dur-
ing or between fishing seasons. To determine
their movements in the area 75,673 adults were
tagged from 1969 to 1971. In this paper we
analyze recoveries from these fish through the
1973 fishing season.

FISHING AREAS

Although Gulf menhaden range from southern
Florida to Veracruz, Mexico (Reintjes 1969), the
purse seine fishery extends only from western
Florida to extreme eastern Texas, with most
fishing effort being expended in inshore waters
from Mississippi to western Louisiana. The fish-
ing season lasts from about early April until early
October, but some plants may begin operations
in late March while others may not begin until

^Atlantic Estuarine Fisheries Center, National Marine Fish-
eries Service, NOAA, Beaufort, NC 28516; present address:
Southeast Fisheries Center Panama City Laboratory, NMFS,
NOAA, Panama City, FL 32401.

^Atlantic Estuarine Fisheries Center, NMFS, NOAA, Beaufort,
N.C; present address: Gulf Breeze Field Station, NMFS, NOAA,
Gulf Breeze, FL 32516.

'Deceased.

nearly May. For this study, we arbitrarily divided
the fishery into three areas (Figure 1).

1. Western: waters and plants west of long.
92°W.

2. Central: waters and plants west of the
mouth of the Mississippi River to long.
92°W.

3. Eastern: waters and plants east of the
mouth of the Mississippi River to long.
86°W.

Plants were located at Moss Point, Miss, (three
plants); in Louisiana — Empire (two plants), Dulac
(two plants), Morgan City (one plant), Intra-
coastal City (one plant), and Cameron (three
plants); and Sabine Pass, Tex. (one plant). The
plants at Empire were considered to be in the
central area.

Because refrigerated carrier vessels may re-
main at sea up to 6 days and fish over a wide
area, we could not tell where their tagged fish
were caught but only where they were processed.
Two exceptions are one plant whose vessels
fished exclusively in the eastern area and another
plant whose vessels fished exclusively in the
western area. For tags recovered at these plants,
the area of capture was known. Although vessels
are far ranging and often travel long distances
to reported concentrations of fish, they tend to
fish most of the time within a restricted radius of
their plant. Most tagged fish, therefore, probably
were caught in the vicinity of the plant where the
tags were recovered.

METHODS OF TAGGING

Gulf menhaden, which spawn from about No-

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

112

PRISTAS ET AL.: RETURNS OF TAGGED GULF MENHADEN

91f ?«• ?V» 92'' 90

30'

28« -

26"
98

TEXAS

Figure l. — Three areas in which adult Gulf menhaden, Brevoortia patronus, were tagged, 1969-71.

vember to March, may arbitrarily be divided
into two broad age-classes, juveniles and adults.
Juveniles are less than a year old, inhabit the
estuaries and rivers during the summer, and
move into the open waters of the Gulf in autumn
when they are about 65 to 130 mm in fork length.
Except in late summer and autumn when some
of the larger fish become available, they are not
vulnerable to the purse seine fishery. Adults are
more than a year old (age 1 or older), inhabit the
larger sounds and inshore areas of the Gulf, and
are vulnerable to the purse seine fishery.

Techniques for tagging adult Gulf menhaden
followed those developed for tagging adult At-
lantic menhaden (Pristas and Willis 1973). A
numbered internal ferromagnetic tag (14.0 x 3.0
X 0.5 mm) was injected into the body cavity with a
tagging gun developed by Bergen-Nautik,'* a Nor-
wegian firm. Fish were obtained from com-
mercial purse seine catches and were tagged
aboard the carrier vessels.

Five percent of the fish tagged in 1969 and
10% of the fish tagged in 1970 were measured.
Because measuring fish reduced the number that
could be tagged, it was not done in 1971. Mean
lengths of fish released in the spring of 1969
ranged from 118 to 130 mm; means of those re-
leased in the spring of 1970 ranged from 157 to
171 mm; and means of those released in autumn
1969 ranged from 148 to 164 mm.

Individual fish were not aged. On the basis of

■•Mention of commercial firm does not imply endorsement
of product by National Marine Fisheries Service, NOAA.

length frequencies, nearly all the fish tagged
were judged to be either age 1 or age 2. Most of
those tagged in spring 1969, probably were age
1. Since the mean lengths were greater in 1970
than in 1969, a greater proportion in 1970 prob-
ably were age 2. Nearly all of those tagged in
autumn 1969 were age 1.

METHODS OF RECOVERING TAGS

Magnets, installed in reduction plants to re-
cover tags moving along the conveyer system
with the fish scrap and meal (Parker 1973), are
classified as either primary or secondary, de-
pending on their location. They were cleaned
about once a week to remove tags and other
scrap metal. Primary magnets are located be-
tween the fish scrap dryers and the storage areas.
Since newly processed fish scrap moves across
the primary magnets, the date tagged fish were
caught can be estimated. Tags recovered on
these magnets are referred to as primary recov-
eries. Secondary magnets are usually located in
the storage, transfer, or loading areas for scrap
and meal. Since fish scrap or meal that moves
across the secondary magnets may have been
in storage for several months or may have been
moved from one plant to another, the date
tagged fish were caught cannot be estimated,
and the plant at which the tags were recovered
cannot always be determined. Tags recovered
on these magnets are referred to as secondary
recoveries. In this paper we combine both types,
since we are interested only in the fishing sea-
son a tag was recovered.

113

FISHERY BULLETIN: VOL. 74, NO. 1

Because many tags that entered a plant be-
came lodged in machinery or passed over mag-
nets w^ithout being captured, the total number of
tags that entered a plant could only be estimated.
The estimates were based on the actual number
of tags recovered and the collective efficiency of
the magnets that recovered them. The efficiency
for each plant w^as estimated by adding 100
tagged fish to catches at regular intervals and
then determining the number of these test tags
that were recovered during the fishing season on
both primary and secondary magnets. The effi-
ciency for each plant, expressed as a percentage,
was the ratio of the number of test tags recov-
ered each fishing season to the number applied.

The number of tests varied from year to year
and plant to plant. In 1969 the number at each
plant ranged from 1 to 8 (100-800 tags); in 1970,
2 to 20 (200-2,000 tags); in 1971, 2 to 16 (200-
1,600 tags); in 1972, 3 to 17 (300-1,700 tags); in
1973, 3 to 16 (300-1,600 tags).

The percentage of tags recovered from each
series of 100 test tags varied from 10 to 907c . The
mean seasonal efficiency varied from 13^^ for the
least efficient plant to 73% for the most efficient.
It also varied from year to year for each plant.

For this study, the estimated total number of
field tags entering a plant was based on the
actual number of field tags recovered on both pri-
mary and secondary magnets. The total number
of field tags entering a plant each month was
estimated by dividing the actual number of tags
recovered by the mean annual plant efficiency.
Tags recovered in spring before fishing began
were added to recoveries from the previous year.

Tags remaining in various parts of a plant for
up to 2 yr before being recovered caused errors in
the recovery data. Nearly 1% of the test tags were
recovered in the second or third year (Table 1),
but the percentages varied from plant to plant.
Test tags introduced late in the season were re-
covered in subsequent years in greater numbers
than tags introduced early in the season. When a
field tag that actually had entered a plant in a
previous season was recovered, it would in effect

Table l. — Number and percentage of test tags recovered
during the year applied and after 1 and 2 yr

Test

No. of
test

Years

applied

After 1 yr

After 2 yr

year

tags

°o

No.

°0

No.

1969

5,600

1,964

35.1

0.50

7 0.13

1970

14,000

7,510

53.6

0 66

15 0.11

1971

11,900

5,317

44.7

0.55

9 0.76

be counted twice and expanded by the efficiency
factor two or more times. For example, if 100 tags
entered a plant whose efficiency was 0.50, the
number recovered would be 50. If 1 of the 50 un-
recovered tags were to be recovered the following
year and the recovery efficiency of the plant had
dropped to 0.25, the estimated number recovered
would be 4 ( 1/0.25 = 4). The estimated number of
tags recovered would be 104 instead of 100, an
error of about 4%, and 4 tags would be assigned to
the wrong year.

SPRING RELEASES AND
RECOVERIES

We tagged 26,995 fish in 1969, 17,775 in 1970,
and 22,800 in 1971. Of the number offish tagged,
the estimated percentages recovered through
1973 were 30.2, 51.5, and 32.5%, for 1969, 1970,
and 1971, respectively. Of the total number of
tags recovered, the largest percentages were in
the first year: 70.9% (1969); 84.3% (1970); 84.6%
(1971). Returns in the second or following year,
accounted for most of the remainder; 26.7%
(1969); 15.0% (1970); 14.3% (1971). Returns after
the second year ranged from 0.7 to 2.4% (Ta-
bles 2-4).

The actual numbers of field tags recovered
after the second year probably were much
smaller than the numbers reported. The percent-

TABLE 2. — Numbers of adult Gulf menhaden tagged in the
spring of 1969 and the estimated number recaptured in subse-
quent fishing seasons, by area.

Release

No. of Year

fisfi of

Area of recovery

area

tagged

recapture

Western

Central

Eastern

Total

Western

10,298

1969

1,839

273

2,164

1970

316

249

585

1971

1972

1973

Total

2,169

575

2,817

Central

3,699

1969

114

1,238

1,414

1970

172

255

1971

1972

1973

Total

187

1,430

1,693

Eastern

12,998

1969

2,188

2,195

1970

519

775

1,333

1971

1972

1973

Total

565

3,026

3,634

Combined

26,995

1969

1,953

1,518

2,302

5,773

1970

425

940

808

2,173

1971

100

168

1972

1973

Total

2,399

2,570

3,175

8,144

114

PRISTAS ET AL.: RETURNS OF TAGGED GULF MENHADEN

Table 3. — Numbers of adult Gulf menhaden tagged in the
spring of 1970 and the estimated number recaptured in subse-
quent fishing seasons, by area.

Release

No. of

fish
tagged

Year

recapture

Area of recovery

area

Western

Central

Eastern

Total

Western

9,100

1970

2.507

1,268

101

3,876

1971

286

479

814

1972

1973

Total

2.797

1.754

151

4.702

Central

5.100

1970

969

1.339

142

2.450

1971

273

367

1972

1973

Total

1.056

1.620

154

2,830

Eastern

3,575

1970

1,348

1.396

1971

160

192

1972

1973

Total

1.531

1.626

Combined

17,775

1970

3,476

2,655

1,591

7,722

1971

369

784

220

1.373

1972

1973

Total

3,853

3.469

1.836

9.158

Table 4. — Numbers of adult Gulf menhaden tagged in the
spring of 1971 and the estimated number recaptured in subse-
quent fishing seasons, by area.

Release

No. of

fish

tagged

Year

recapture

Area of recovery

area

Western

Central

Eastern

Total

Western

7,400

1971

1,711

843

2,602

1972

143

225

1973

Total

1,794

991

2.835

Central

5,200

1971

642

904

1.603

1972

1973

Total

669

966

1,694

Eastern

10,200

1971

2,008

2,066

1972

157

589

749

1973

Total

251

2.630

2.885

Combined

22,800

1971

2.353

1,805

2,113

6.271

1972

110

356

593

1,059

1973

Total

2,467

2,208

2,739

7,414

ages of test tags recovered after 1 yr (0.5%) and
2 yr (0.1%) probably underestimated the per-
centage of field tags that remained in a plant
and were recovered after 1 or 2 yr, since a greater
number of test tags were applied early rather
than late in the season and therefoi^e had a
greater chance of being recovered in the year
they were applied. The tendency of field tags that
had been out more than 2 yr to be recovered
early, rather than late, in the fishing season sug-
gests that some, at least, had remained in plants
over the winter. At plants where recovery effi-
cencies were relatively low, mainly plants in the
eastern and central areas, a greater percentage
of field tags were returned after 2 yr than at plants
where efficiencies were relatively high. Field tag

recoveries after 2 yr were highest at those plants
where test tag recoveries after 1 yr were highest.
The plant for which no field tag recoveries were
reported after 2 yr had the lowest percentage of
test tag recoveries after 1 yr — less than 0.1%.

Eastern Releases

Nearly all first year recoveries (tags recovered
the same year they were applied) were at plants
in the eastern area (99.7% in 1969; 96.5% in 1970;
and 97.2% in 1971), and no tags were recovered
in the western area. The only tags recovered in
the central area were at plants whose vessels
also fished in the eastern area. Second year re-
coveries (tags recovered the year after they were
applied) followed the same pattern as first year
recoveries, although a greater proportion of tags
were recovered in the central area. For 1969 re-
leases, no tags were recovered the second year
at the plant in the western area whose vessels
fished only in that area. For 1970 releases, no
tags were recovered the second year in the west-
ern area. For 1971 releases, only three tags were
recovered in the western area, all at a plant whose
vessels fished in all areas.

Central Releases

Although tags were recovered the first year at
plants in all areas, the highest percentages were
from plants in the central area (87.6% in 1969;
54.7% in 1970; 56.4% in 1971). The lowest per-
centages were at plants in the eastern area, as
might be expected, since the western and central
areas are continuous with each other but are
separated from the eastern area by the Missis-
sippi Delta. In 1969 and 1970 no tags were recov-
ered at the plant in the eastern area whose
vessels fished only in that area. Some tags were
recovered in the western area at the plant whose
vessels fished only in that area. The majority of
second year recoveries also was at plants in the
central area (71% for all release years com-
bined); the fewest were at plants in the eastern
area (4% for all release years combined).

Western Releases

Most of the first year recoveries were at plants
in the western area (85.0% in 1969; 64.7% in
1970; 65.7% in 1971), and the fewest were at
plants in the eastern area (2% in 1969 and 1971;

115

FISHERY BULLETIN: VOL. 74, NO. 1

3% in 1970). No tags were recovered at the plant
in the eastern area whose vessels fished only in
that area. Fewest second year recoveries were at
plants in the eastern area (3% in 1969; 6% in
1970; 1% in 1971). Most second year recoveries
were at plants in the western area for fish tagged
in 1969 and in the central area for fish tagged in
1970 and 1971.

AUTUMN RELEASES AND
RECOVERIES

Fish were tagged in autumn (September) only
in 1969, when 900 were tagged in the western
area, 2,100 in the central area, and 5,103 in the
eastern area (Table 5). By the end of the fishing
season in October, 6% had been recaptured. In
the following year 33% were recovered. For all
years combined 42% were recovered.

As with tags of fish released in spring, tags of
fish released in autumn were recovered mainly at
plants in the area of release in both the first
and second year Few fish tagged in the western
area were recovered in the eastern area and few
fish tagged in the eastern area were recovered
in the western area. No fish tagged in the west-
ern area were recaptured at the plant in the
eastern area whose vessels fished only in that
area. Approximately 90% of the tags of fish re-
leased in the eastern area and recovered in the

Table 5. — Numbers of adult Gulf menhaden tagged in autumn
of 1969 and the estimated nxmibers recaptured in subsequent
fishing seasons, by area.

Release

No. of
fish

Year
of

Area of recovery

area

tagged

recapture

Western

Central

Eastern

Total

Western

900

1969

1970

141

1971

1972

1973

Total

106

197

Central

2,100

1969

166

177

1970

277

305

615

1971

1972

1973

Total

460

357

854

Eastern

5,103

1969

251

1970

617

1,300

1,938

1971

109

1972

1973

Total

682

1,628

2.331

Combined

8,103

1969

195

252

460

1970

371

988

1,335

2,694

1971

106

195

1972

1973

Total

587

1,128

1,667

3,382

central area were at plants whose vessels fished
up to 25% of the time in the eastern area.

CONCLUSIONS

The pattern of first year tag recoveries shows
clearly that adult Gulf menhaden make no exten-
sive east-west movement along the coast during
the fishing season from April to November.
Nearly all tags were recovered at plants located
in the same area in which the fish were tagged.
Some fish that were released in one area but
whose tags were recovered at a plant in another
probably were caught in the release area, since
vessels at most plants, though fishing mostly
within their own area, also were far-ranging. No
fish tagged in the eastern area were recovered at
plants in the western area; few fish tagged in the
western area were recovered at plants in the east-
ern area. At plants whose vessels fished exclu-
sively in either the eastern or western area, no
tags were recovered except those from fish re-
leased in the same or adjacent area.

Second year recoveries also point to little or no
mixing of fish from different areas during the
winter. Gulf menhaden apparently move offshore
during autumn and return again in spring to the
same general area they previously occupied.
Since the boundary between the western and
central areas is arbitrary and since we do not
exactly know where fish were recovered, the
greater number of second year returns in the cen-
tral, rather than western area of fish tagged in
the western area for 1970 and 1971 does not
necessarily indicate any significant shift of fish
from the western to the central area.

Because there were no estimates of tag losses
due to shedding or deaths caused by tagging,
and because the variability in recovery efficien-
cies was large and some tags tended to remain
in plants for long periods, calculation of fishing
and total mortality rates would be no more than a
mathematical exercise. We can estimate from the
data, however, whether fishing mortality and ex-
ploitation rates are high or low.

Both fishing mortality and exploitation rates
appear to be high. First year recoveries of spring
releases ranged from 21 to 43% of the number of
fish tagged. The total number of tags recovered
ranged from 30 to 51% for spring releases and
was 42% for the autumn releases. High tagging
mortality may account for the relatively low re-
turns for the 1969 and 1971 spring releases (30%

116

PRISTAS ET AL.: RETURNS OF TAGGED GULF MENHADEN

and 32%), since tagging mortality tends to be
greater for small Atlantic menhaden than for
large ones (Kroger and Dryfoos 1972), and the
fish tagged in 1969 were generally smaller than
those tagged in spring of 1970 or autumn of 1969.
It is unlikely that more than a small percentage
of any year class survive more than 3 yr Less than
2% of the estimated returns of fish tagged in
spring, and 7% of the returns of fish tagged in
autumn were recovered after the second year
Because of the tendency of tags to hang up in
plants, the majority of tags recovered after the
second year probably had come from fish caught
in the first or second season after being tagged.
If tags that hung up for 1 yr averaged 1.5% and
for 2 yr or more 0.2%, and if recovery efficiencies
averaged 50%, hung up tags could account for
nearly all tags reportedly recovered after 2 yr.
Since the majority of fish tagged were in the size
class of age- 1 fish, the percentage of returns after
2 yr should have been higher than it was if any
significant number survived more than 3 yr.

ACKNOWLEDGMENTS

Only a project report based on returns through

July 1971 had been prepared before the authors
transferred to other laboratories and work on the
manuscript was temporarily suspended. Revi-
sion had just begun when Robert L. Dryfoos died
suddenly in January 1974. William R. Nicholson
and Robert M. Lewis, Atlantic Estuarine Fisheries
Center, Beaufort, N.C., prepared the 1971-73
returns for the computer programs, incorporated
them into the previous data, and assisted in re-
vising and editing the final manuscript.

LITERATURE CITED

KROGER, R. L., AND R. L. DRYFOOS.

1972. Tagging and tag-recovery experiments with Atlan-
tic menhaden, Brevoortia tyrannus. U.S. Dep. Commer,
NOAA Tech. Rep. NMFS SSRF-664, 11 p.

Parker, R. O., Jr.

1973. Menhaden tagging and recovery: Part II — Re-
covery of internal ferromagnetic tags used to mark men-
haden, genns Brevoortia . Mar. Fish Rev. 35 (5-6):36-39.

PRISTAS, P. J., AND T D. WILLIS.

1973. Menhaden tagging and recovery: Part I — Field
methods for tagging menhaden, genus Brevoortia.
Mar. Fish Rev 35(5-6):31-35.
REINTJES, J. W.

1969. The Gulf menhaden and our changing estuaries.
Proc. Gulf Caribb. Fish. Inst., 22nd Annu. Sess., p. 87-90.

117

DEVELOPMENT AND EXAMPLE APPLICATION OF A SIMULATION
MODEL OF THE NORTHERN ANCHOVY FISHERY

Michael F. Tillman^ and Donald Stadelman^

ABSTRACT

A computer simulation model of the reduction fishery for northern anchovy, Engraulis mordax, is
described. The biological subroutine of this model is an age- structured paradigm which is modified to
account for age-dependent exploitation and variable recruitment. To demonstrate the model's utility,
two example applications are presented which provide insight into the problems of evaluating alterna-
tive regulations while lacking perfect knowledge of economic or biological behavior. The model's
current value lies in its use as a tool to identify research needs.

Based upon the systems analyses of Tillman
(1972) and Stadelman (1974), it appears that the
northern anchovy, Engraulis mordax Girard, con-
stitutes one of the largest latent fishery resources
available to American flag vessels. Relative to its
estimated biomass, only a minute fraction of this
species is harvested when compared, for example,
to catches taken by the fishery for Peruvian an-
choveta, E. ringens. The present northern an-
chovy fleet consists of only a small number of rel-
atively old vessels, and the processing capacity of
the fish meal plants servicing this fleet is quite
inadequate. Thus, unlike many major fisheries of
the United States which are marked by overex-
pansion and overcapitalization, the northern an-
chovy fishery is still underdeveloped.

According to the above authors, this lack of
development can be attributed to a variety of
natural and artificial barriers. The natural bar-
riers comprise those constraints over which man
has little or no control, including lack of predic-
tive ability concerning the short-term behavior of
the market for fish meal. Moreover, there pres-
ently is lacking definitive biological knowledge
concerning the inherent variation in size and
availability of the northern anchovy population,
its dynamic stock-recruit feedback mechanisms,
and its natural mortality processes. These gaps
provide the context of a dynamic and variable en-
vironment within which this fishery system oper-
ates and with which its managers must contend.

The artificial barriers, on the other hand, are

^Northwest Fisheries Center, National Marine Fisheries Ser-
vice, NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112.

^Institute of Governmental Research, University of Washing-
ton, 3935 University Way N.E., Seattle, WA 98195.

institutional constraints which man has imposed
upon the system. While the intent of these rules
or regulations may be to govern the activities of
fishery participants, their overall effect, in the
opinion of Tillman (1972) and Stadelman (1974),
has been to thwart economic development of the
fishery. For example, small quotas for reduction
purposes are intended to prevent overcapitaliza-
tion of the fishery but have also acted to hinder
the much needed replacement and renovation of
antiquated reduction equipment. Other artificial
barriers and their apparent effects, as perceived
by the foregoing authors, include the following:
areal and temporal closures to protect stocks, but
which act instead to reduce harvest efficiency;
union rules to maintain employment levels, but
which in fact work to prevent use of technological
innovations that would reduce harvesting costs or
increase efficiency; landing taxes of $2 per ton to
pay for research and management, but which in
fact act to reduce substantially the returns ob-
tained by private interests.

If an appropriate goal for decision makers is to
foster economic development of the northern an-
chovy fishery, then the above institutional bar-
riers would seem to present opportunities for
achieving that goal. Consequently, a computer
simulation model has been developed which pro-
vides the means for evaluating the biological and
economic consequences of changing various regu-
lations governing this fishery. The purpose of this
study is to briefly describe this simulation model
and to present two examples of its application
which demonstrate some of its utility. These ap-
plications focus on the evaluation of alternative
regulations when given imperfect knowledge of
biological or economic behavior. Finally, the

Manuscript accepted May 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

118

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

value of modelling this system is discussed, tak-
ing into account some of the present model's limi-
tations and shortcomings.

DEVELOPMENT OF
THE SIMULATION MODEL

General Description

The basic model of the northern anchovy
fishery is formulated in terms of GAMES, the
general-purpose simulator of resource use sys-
tems developed by Gales (1972). This Fortran IV
program has been designed to simulate the ac-
tivities of major sectors involved in the harvest-
ing and marketing of renewable resources. The
sectors modelled by GAMES include locations,
stocks, harvesters, processors, regulators, prod-
ucts, and markets.

A specific system such as the anchovy fishery
(Figure 1) is modelled by indicating, through ap-
propriate inputs, the number of entities in each

Market -
Products

Processors

Regulators

Los Angeles
Processors

Harvesters

Col. Fish
and Game

Small

Vessel

Fleet

Stocks

Lorge

Vessel

Fleet

Northern
Anchovy

Location

Southern
California

Figure l. — Graphic representation of logical relations be-
tween sectors of the present northern anchovy fishery. From
Tillman ( 1972).

sector and their logical linkages. The user must
also provide the values of parameters which
define system processes and structures and the
initial values of variables which describe systems
behavior. Tillman (1972) provides a detailed list-
ing of the values required for the northern an-
chovy model. Through appropriate control values,
the user also specifies that certain built-in deci-
sion routines be used or else provides algorithms
of his own design by adding subroutines to
GAMES or by modifying existing ones. The user
must also provide an appropriate biological model
of the stocks being exploited by the harvester-
processor sectors.

The main GAMES program resembles the par-
tial listing given in Figure 2. The "Labelled
COMMON Blocks" reserves sections of memory
for storage of the values of parameters and vari-
ables used in common by the 11 subroutines. Sub-
routine TAPEIN is called first and reads in the
initial values of these parameters and variables,
including the starting and ending years of simu-

PROGRAM MAIN
[Labelled COMMON Blocks]
CALL TAPEIN

DO 110 YEAR=NYEAR1, NYEAR2
DO 10 0 MONTH = 1,12

CALL

PROCS

CALL

HARVS

CALL

REGLS

CALL

STOCKS

CALL

HRVST

CALL

RMARKT

CALL

PRCES

CALL

CMARKT

CALL

STATS

CALL

SMS TAT

100 CONTINUE

110 CONTINUE

[Codir

g for Subroutines ]

Figure 2.— Partial listing of the main GAMES program.

119

FISHERY BULLETIN: VOL. 74, NO. 1

lation. The succeeding 10 subroutine call state-
ments are imbedded within a double "do-loop"
which is indexed by month and year. This double
loop is the principal timing mechanism of the
program. Hence, each of these 10 subroutines is
executed once a month in the order indicated and
either simulates a component of the system, their
interactions, or else produces output.

Subroutines PROCS, HARVS, and REGLS
make programmed monthly decisions for the sys-
tem's respective processors, harvesters, and regu-
lators. PROCS and HARVS simulate monthly
decisions concerning the processing capacity
committed, the number of days spent harvesting,
the number of harvesting units committed, and
the gear efficiency per unit. Moreover, since pro-
cessors have only limited storage capacity for raw
materials, HARVS adjusts allowable vessel ca-
pacities as if processors were establishing boat
quotas (a situation presently occurring in the re-
duction fishery); this prevents overfishing and the
consequent dumping of excess catches. REGLS
compares these decisions to standards (regula-
tions) supplied by the user or determined by the
subroutine. If regulations are "broken," the sub-
routine makes appropriate adjustments to the
values of those parameters associated with im-
proper decisions.

STOCKS is a user supplied subroutine which
simulates the biomass dynamics of the exploited
resource on a monthly basis. The northern an-
chovy subroutine is an age-structured model
which accounts for the processes of growth, mor-
tality, graduation, and reproduction for each of
the seven age-groups (ages 0-6) comprising the
population. The basic mathematical theory for
age-structured models is treated by Ricker (1958)
and Beverton and Holt (1957). This basic theory
has been modified to account for age-dependent
exploitation and variable recruitment processes
in the northern anchovy population. Similar
age-structured models have been developed in re-
cent years for other species by Tillman (1968),
Walters (1969), Fox (1973), and Francis (1974).

Described further in an ensuing section,
STOCKS feeds catch values to HRVST, the sub-
routine which then simulates the monthly har-
vesting process. HRVST determines the catch of
each stock by a harvester, his harvest propor-
tional costs, and the cumulative catch taken from
each stock.

RMARKT then simulates the sale of the har-
vesters' catches to the processors, and PRCES

transforms these newly purchased raw materials
into finished goods which are added to the proces-
sors' inventories. Subroutine CMARKT then
simulates the sale of these products on the open
market to final consumers. The quantities de-
manded are determined from a user supplied de-
mand curve and a sales price set by the processor.
STATS then computes and outputs financial
statements for the processors and harvesters. It
also provides physical reports describing through
key variables the activities of the harvester, pro-
cessor, stock, and market sectors. Subroutine
SMSTAT then provides user desired cumulative
physical reports. Although all reports may be
provided at monthly intervals, printout typically
is suppressed until the year's end.

The Biological Sector

Some Important Assumptions

Development of the biological model for north-
ern anchovy depends critically upon two assump-
tions. One concerns the stock structure of this
population and the other, its stock-recruit be-
havior. The following discussion briefly exam-
ines how reasonable these assumptions are and
hopefully provides some justification for their
application.

Mais (1974) and Tillman (1975) review the evi-
dence which generally supports the hypothesis
that three distinct stocks exist within the north-
ern anchovy's total geographic range. The
simplifying assumption has been made that the
reduction fleet fishes exclusively upon that stock
which resides in the southern California-
northern Baja California region of the California
Current system. Results of tagging studies indi-
cate that some mixing of adult members of adja-
cent stocks might conceivably occur due to sea-
sonal north-south migrations (Haugen et al.
1969). However, Mais (1974) cites evidence from
comparisons of length-frequency and age-length
distributions which, in his opinion, indicates that
very little, if any, mixing occurs. Moreover, he
concludes that anchovies in this region should be
treated as a single biological unit for manage-
ment (and therefore modelling) purposes.

Several studies (Cushing 1971; Tillman and
Paulik 1971; Murphy 1973) suggest that recruit-
ment in clupeid and engraulid populations is a
density-dependent process. Moreover, these
authors imply that the asymptotic stock-recruit

120

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

relationship of Beverton and Holt ( 1957) is gen-
erally applicable to populations which have an
extended spawning season, whose adults are
cannibalistic upon their own young, and whose
annual recruitment variations are relatively
small. Results from surveys for pelagic eggs and
larvae conducted off California indicate that the
northern anchovy spawns over virtually the entire
year (Ahlstrom 1966). Baxter (1967) stated that
this species is a filtering and biting feeder which
consumes its own eggs and larvae. Moreover,
Murphy (1966) noted that this species has never
had spectacularly good nor spectacularly bad
year classes and that this may have been a factor
in the relatively slow replacement of the Pacific
sardine, Sardinops sagax, by anchovies following
the collapse of the sardine fishery. Conse-
quently, since the northern anchovy apparently
fits the required life-style, an asymptotic stock-
recruit model does not seem too unreasonable
an assumption, although it is an admittedly cir-
cumstantial and speculative one at this time.

General Description of STOCKS

STOCKS' main job is to solve the catch equa-
tion and pass the result to subroutine HRVST.
The following description briefly summarizes the
sequence of operations which occur each month
and some of the parameter values required to
determine the catch in weight for each age group.
The details of parameter estimation are given
by Tillman (1972).

Following the combined adjustments of
PROCS, HARVS, and REGLS, STOCKS first re-
ceives the allowed values of the following vari-
ables: level of fishing effort (number of vessels),
vessel capacity (metric tons (MT)/boatday),
fraction of the month fished, and fishing power of
a vessel (Table 1 gives values of relative fishing
power for various-sized vessels for which eco-
nomic performance data are available). These
four variables are used to calculate equivalent

Table l. — Efficiencies and relative fishing powers of hypothet-
ical vessels operating on northern anchovy. From Till-
man (1972).

Vessel capacity

Calculated

Relative

Tons MT

efficiency

efficiency'

66 60

0.536

0.681

110 100

0.787

1 000

155 140

1.038

1.319

210 191

1.358

1.726

265 240

1.518

1.929

standard effort, in terms of boats fishing the
entire month instead of a fraction of it, and the
total harvesting capacity of the reduction fleet.

Next the age structure is updated by account-
ing for the process of graduation. Since the
great bulk of spawning activity occurs during
January-May, most anchovies have their birth
dates during these 5 mo. Table 2 gives the pro-
portion of each age-group that is expected to
graduate at the start of the months indicated.
Recruits due to enter in the current month are
added to the first age-group, and fish leaving the
last age-group disappear. Within each age-
group, size of the individual is computed as a
weighted average of the sizes of newly entered
and residual fish. From these adjusted weights
and numbers at age, the biomass of the popula-
tion is computed.

Contribution to spawning then is calculated
for the current month. The number of females
eligible to spawn is determined by the propor-
tion of females in the population (Table 3), by a
maturity at age schedule (Table 4), and by a
schedule of the incidence of monthly spawning
activity (Table 5). The egg production of these
spawning females is computed by a fecundity at
age schedule (Table 6). The results of this proce-
dure are additions to the number of eggs de-
posited on the stock's spawning ground.

Instantaneous total mortality rates then are

Table 2. — Probabilities of graduating from one age group into
the next for northern anchovy. From Tillman (1972).

Birtfi date

Proporl

on graduating

Cumulative proporlion

January

0.17

February

0.18

0.35

Marcti

0.25

0.60

April

0.25

0.85

tviay

0.15

1.00

Table 3. — Estimates of fraction of females by number in the
tottd northern anchovy population.

Source

Estimate

Source

Estimate

Clark and Phillips (1952) 0.57

(Vliller et al. (1955) 0.56

Miller and Wolf (1958) 0.52

l^acGregor (1968) 0.56

Collins (1969)
Collins (1971)
Average

0.60
0.58
0.56

Table 4. — Maturity at age schedule of northern anchovy. From
Tillman (1972).

MOO-MT (metric ton) vessel is standard.

Age-group

Fraction mature

Age-group

Fraction m£

0.10

1.00

0.40

1.00

0.80

1.00

0.95

121

FISHERY BULLETIN: VOL. 74, NO. 1

Table 5. — Incidence of monthly spawning activity by northern
anchovy as determined from larval counts. From Till-
man (1972).

Month

Fractional occurrence

Adjusted occurrence'

June

0.10

July

0.05

August

0.03

September

0.01

Octotier

0.02

November

0.03

December

0.03

January

0.11

February

0.20

Marcti

0.17

April

0.17

May

0.08

0.20
0.10
0.06
0.02
0.04
0.06
0.06
0.22
0.40
0.34
0.34
0.16

'Adjusted to insure two spawnings per year

Table 6. — Fecundity at age of northern anchovy, assuming
574 eggs/g body weight. From Tillman (1972).

'Average weight in month 10, March, the midpoint of the major spawning
period.

computed for each age group, which may be sub-
jected to a different total mortality, Z (A,M),
depending on natural mortality rate, catchability
coefficient, seasonal availability factor, and the
total units of standard effort operating upon
the stock during the month:

Z(A,M) ^NM + F(AM)

where NM = constant natural mortality rate
F(A,M) = age specific fishing mortality
rate
= Q(A) ■ AV(M) ■ FF(M)
where Q(A) = age specific catchability
coefficient
AV(M) = monthly availability of the stock
FF(M) = standardized level of effort.

According to Schaefer (1967), NM = 1.10 and is
a constant parameter. Table 7 shows how catch-
ability decreases for ages which are not fully re-
cruited. Figure 3 indicates how availability varies
throughout the year, based upon extrapolations
of Messersmith's (1969) catch-per-unit-effort
(tons/hour) data for two seasons; this seasonal
pattern likely is associated with the spawning be-
havior of adults (Tillman 1972).

Given these mortality rates, the catch of an-

TABLE 7. — Age specific catchability coefficients for northern
anchovy given different areal restrictions and assuming full
recruitment occurs at age 2. From Tillman (1972).

Age

Coefficient when
inshore closed (10"^)

Coefficient when
inshore open (10'^]

2-6

0.24
2.78
9.04

0.38
4.10
9.04

Age-group

Average weight'

(g)

Fecundity
(eggs/spawning)

9.1

5.200

14.9

8,600

.C)

20.4

1 1 .700

25.1

14.400

28.9

16,600

31.9

18,300

34.2

19,600

.5-

-I 1 1 1-

-1 r 1 1 1

Jon Feb Mar Apr May Jun Jul Aug Sepf Oct Nov Dec Jan

Figure 3. — Average monthly availability of northern anchovy
in the southern California area. From Tillman (1972).

chovy is then computed for the month subject
to the constraint that it may not exceed the reduc-
tion fleet's total or assigned harvesting capacity.
The fleet and natural mortality at first compete
exponentially to determine the number of fish
each would take if harvesting capacity were un-
limited. The temporary catch in numbers is cal-
culated as:

CN(A) =

F(A,M)
Z(A,M)

N(A) ■ EXP

where

-ZiA.Mi iDT/NCYCLi

EXP - 1 - e"
N(A) = size in numbers of age group
DT = 1 mo

and F(A,M) and Z(A,M) are defined as above.
NCYCL is a parameter which determines the
accuracy of the solution and typically is set at 4,
yielding an effective DT of 1 wk.

122

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

The fleet's catch in weight then is temporarily '^

computed as the sum

CW(M) = ^CN(A) ■ WT(A,M)

where WT(A,M) is current weight at age. If
CW(M) exceeds the allowed harvesting capacity
of the fleet, CAPAC(M), the catch in weight is
adjusted downward:

RC = CAPAC(M)/CW(M)
CW(M)' = 2RC ■ CN(A) ■ WT(A,M).

Also, the fleet is rendered inactive for the re-
mainder of the week.

Fish credited to the harvester in excess of ca-
pacity are subjected to natural mortality and then
returned to the population. Once the catch cycle
has been completed, the number of fish remain-
ing in an age-group is determined by subtracting
the numbers caught and the numbers taken by
natural mortality.

Grov\i;h in length which occurred during the
month then is computed utilizing a von Berta-
lanffy equation (Beverton and Holt 1957). Figure
4 shows the growth in length curve for the follow-
ing parameter values: Loc = 15.91 cm, /^ = 0.32,
^0 = -2.08. New individual weights at age are
then computed from a cubic weight-length
relation.

Finally, future recruitment is calculated from
the number of eggs deposited on the stock's
spawning ground and an egg to recruit survival
rate:

RECRT(M) ^EGGS(M) ■ SER ■ SMULT( RATIO)

where SER is the equilibrium egg to recruit sur-
vival rate and SMULT( RATIO) is a multiplier
which adjusts SER in a density-dependent man-
ner Given Vrooman and Smith's (1971) estimate
of equilibrium spawning stock size (SEQ =
4.55 X 10« MT), Tillman (1972) estimated equi-
librium recruitment (REQ = 420 x 10^ fish)
and equilibrium numbers of eggs {EEQ = 2 x
1015 eggs) to obtain SER = 0.00021. The num-
ber of new recruits created during the current
month will subsequently enter the fishable stock
after a prerecruit period of 6 mo.

The appropriate value of SMULT( RATIO)
is determined from

14-

10-

o Observed
* Calculated

I 2 3 4 5 6 7

Time (years)

Figure 4. — Asymptotic growth in length of northern anchovy.
From Tillman ( 1972).

SMULT(RATIO)

A +B ■ RATIO

where RATIO provides a measure of the current
spawning stock size, SP(M), relative to its
equilibrium level, SEQ:

RATIO =SP(M)/SEQ.

This formulation insures that the stock-recruit
process behaves in an asymptotic manner, as has
been assumed.

Although data are lacking to estimate specific
values for stock-recruit parameters A andB, sets
of arbitrary values can be determined by defining
a family of curves which pass through the same
equilibrium point (SEQ, REQ). Following Till-
man (1972), a unique curve in this family is
distinguished by its asymptotic level of recruit-
ment, RMAX, which can be defined as some mul-
tiple of the equilibrium level of recruitment:

RMAX = MULT ■ REQ.

A particular set of stock-recruit parameters can
then be determined as

B = IIMULT

123

FISHERY BULLETIN: VOL. 74, NO. 1

Vrooman and Smith's (1971) larval data provide a
rough measure of variation in recruitment during
1962-66, a recent period of population stability.
Comparison of their largest index of larval abun-
dance (63 X 10^2) with the mean value during this
period (48 x lO^^) indicates that values of MULT
apparently should not exceed 1.30. Table 8 lists
some representative values of SMULT(RATIO),
given MULT values in the range 1.05-1.20.

Table 8. — Egg to recruit survival multipliers (SMULT) for a
family of three stock-recruit curves passing through the same
equilibrium point. A unique curve depends on the value of
MULT which defines parameters A and B. Each multiplier
corresponds to given ratio between present and equilibrium
biomass of the spawning stock.

Table 9. — Costs and prices for the northern anchovy model
as adapted from Stadelman (1974).

RATIO

Curve
MULT

1.05
0.04762
0.95238

1.10
009091
090909

1 20
0.16667
0.83333

0.10
0.20
0.30
0.50
0.75
1.00
2.00
3.00

7,00
4.20
3.00
1 91
1.31
1.00
0.51
0.34

5.50
3.67
2.75
1.83
1.29
1 00
0.52
0.35

4.00
3.00
2.40
1.71
1.26
1.00
0.55
0.38

Some Economic Content

Costs and prices used in this study (Table 9)
have been adopted from among those estimated
by Stadelman (1974). While these values are
dated, particularly with respect to the price in-
crease experienced in 1974, they still serve to il-
lustrate our example applications. Following his
suggestion, it is assumed that landing taxes have
been removed, that the union has allowed fisher-
men to receive a guaranteed wage (rather than a
share), and that it also has permitted crew size to
be reduced on vessels equipped with power
drums. Such changes conceivably would permit
the fishery to take advantage of new technology
that would provide the impetus for its immediate
economic expansion. Moreover, it is assumed that
quotas have been removed. In their stead, deci-
sion makers allow the fishery to expand to its
economically optimal level, insuring however
that only that fleet size is used and that catch is
taken which supplies the optimal level of process-
ing capacity in the system.

These assumptions, particularly the ones per-
taining to crew wages and to quotas, may not be
very realistic, but they do provide the basis for
some interesting modelling applications. Their
use infers that the harvesting-processing configu-

Item

Without
power drum

With
power drum

Harvesting costs:

Annual fixed cost/vessel
(Depreciation, moorage,
property taxes, office and
shore expenses, insurance)

Return on investment (15%)

Guaranteed wages
(Crew and captain)

Drum cost (Depreciation
and return on investment)

Fixed cost/year
Fixed cosfday fished

(Fuel and maintenance)
Cost/t\/IT anchovy caught

(Net repair)

$30,126

24,779
132,000 (11)

186,90500
77.75

2.20

$30,126

24,779
84,000 (7)

6,900

145,805.00
77.75

2.20

Processing costs:
Annual fixed cost/plant
(Overhead, 1 5% return on
investment)
Purchase price of anchovy/MT
Processing cost of anchovy/IVIT
tVlarket pnces:
Fish meal/MT
Fish oil/N/TT

$150,000.00

25.00
5.50

250.00
110.00

rations of this study fulfill three criteria: 1) they
maximize net economic yields; 2) they allow for
payment of opportunity wages to crew members
and of opportunity returns^ to capital invested in
the system; 3) they utilize state of the art tech-
nology. Opportunity wages are set at a guaran-
teed salary of $12,000/man. Also, a 15% rate of
return is used to compensate an investor for his
loss of alternative uses of capital, for his risk, and
for his managerial skill.

State of the art technology implies the use of
new plants and new vessels. According to the
above study, a new plant has only limited storage
capacity for raw materials, a processing capacity
of 20 tons/h, and conversion factors of 0.20 for
meal and of 0.01 for oil. By working 20 h/day, 252
days/yr, such a plant could process 92,000 MT of
anchovy annually. The above study also found
that a 210-ton (191-MT) purse seiner was the most
economically efficient harvesting unit. A new
vessel of this size could be equipped with a power
drum, which would lead to a reduction in crew
size (from 10 to 6 men) but not necessarily to an
increase in harvesting efficiency.

Stadelman (1974) indicated that prices of fish

^One who invests labor or capital in a particular economic
opportunity should at least earn that amount which might be
returned by his next best investment alternative. The amounts
that could have been earned from this second choice are
termed opportunity returns; i.e., opportunity wages should be
earned by labor and opportunity returns by capital.

124

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

meal and oil in the United States are established
primarily by the world market for these products.
Consequently we have assumed that northern
anchovy processors can only accept the prices of-
fered for their meal and oil, rather than being
able to affect the world market through their own
efforts. In this case, demand curves for their
products are nonexistent, and the fixed prices
given in Table 9 hold throughout a given simula-
tion experiment.

APPLICATIONS OF THE MODEL

Analytical Technique

Nature of Results

Due to the rough nature of many of the esti-
mates utilized by the model, little credence has
been attached to the absolute values of economic
return, catch in weight, or population size ob-
tained in the following simulation experiments.
These results are at best only informed extrapola-
tions, and, even though their values are of the
proper orders of magnitude, it is not the intent of
the following applications to accurately predict
future returns, yields, or sizes. Of greater impor-
tance are the relations between values obtained
in different experiments. Consequently, the re-
sults have been analyzed on a comparative rather
than an absolute basis.

Criteria for Comparisons

The primary results obtained from each exper-
iment include the net economic return (before in-
come tax) generated annually by the entire sys-
tem, the number of days fished each season, the
annual catch in weight, and population size in
terms of annual average biomass. In most exper-
iments, these four variables satisfactorily mea-
sure the economic and biological performance
achieved during an experiment. In preliminary
long run equilibrium experiments, values of
these variables stabilized within a 10-yr period.
Thus, 10 yr has been chosen as the length of all
experiments.

Differences between various experiments are
measured primarily in terms of the differences
between respective net economic returns. Net
economic return is obtained by subtracting amal-
gamated harvester-processor costs from amalga-
mated gross revenues at the end of each year of

simulation. Amalgamated costs include the an-
nual opportunity costs of labor and capital.

Alternative Regulations and
Stock- Recruit Sensitivity

Recalling the spectacular decline of the sar-
dine fishery during the 1950's and fearing a
similar debacle over another forage species,
sportsmen and bait fishermen have become
allied in sponsoring state legislation to limit com-
mercial development of the northern anchovy. As
a consequence of their efforts, the reduction
fishery has been plagued by low quotas and cur-
rently cannot fish during the summer (15 May-
15 September) nor within 3 miles (4.8 km) of
shore. These two specific exclusions define areas
wherein tradeoffs might be made to gain conces-
sions from the sport and bait fisheries. Decision
makers might retain the summer or inshore clo-
sures intact to placate the nonindustrial groups
and receive in trade the concession of larger
quotas for industrial use of anchovy. Some idea
of what is lost by such trades might be obtained
by contrasting these closures to others wherein
more lenient measures were enforced.

Some evidence exists which indicates that con-
siderable gains in harvesting efficiency might be
achieved by lengthening the season to a year or
by opening the inshore area. In Figure 3, the
pattern of availability extrapolated for May-
September indicates that an improving trend is
expected during the summer. Also, Tillman's
(1972) analysis of age-specific catchability re-
vealed that age-groups 0 and 1 tend to be more
available in the inshore area than in the offshore
commercial fishery area; he subsequently calcu-
lated catchability coefficients reflecting this ap-
parent areal difference (results given in our
Table 7).

Using these catchability coefficients implicitly
assumes that older anchovies (ages 2-6) are
equally available in the inshore and offshore
areas. As indicated in Figure 3 we have, of course,
attempted to account for the seasonal availability
of older anchovies as related to their spawning
behavior, but the net result of spawning move-
ments might also tend to distribute older fish
farther offshore than younger ones. This cir-
cumstance would effectively reduce the inshore
catchability coefficient for older fish.

Unfortunately, data on the areal distribution of
age-groups, such as the age compositions of

125

FISHERY BULLETIN: VOL 74, NO. 1

catches taken at varying distances from shore,
were not available to examine this possibility in
detail. However, Messersmith et al. (1969) re-
ported that, during summer and fall echo-sounder
surveys, all sizes of anchovies were found concen-
trated close inshore. Since all sizes were encoun-
tered, we speculated that, if fishing were allowed
inside of 3 miles (4.8 km), the catchability coef-
ficient for older fish would become reduced only if
effort concentrated on or very near nursery
grounds, which occur on shallows and flats inside
of 50 fathoms. Although lower fuel costs might
dictate such a concentration, we further specu-
lated that enforcement of the current minimum
size limit of 10.8 cm would make fishing this far
inshore unattractive and thus curtail it.

Given these speculations, simulation experi-
ments were conducted in our first application to
examine the biological and economic conse-
quences of opening the inshore area to commer-
cial fishing and of allowing a 12-mo fishing
season. These were contrasted to a "present" sit-
uation consisting of a closed inshore area and an
8-mo season (15 September-15 May). Moreover,
sensitivity of the model to changes in the stock-
recruit relationship was examined given alterna-
tive areal-seasonal restrictions. Stock-recruit
curve 2 (Table 8) was arbitrarily chosen as the
standard for comparison in these experiments.
Each experiment thus determined how an opti-
mal harvesting-processing configuration (num-
bers of vessels and plants) defined for curve 2 per-
formed when stock-recruit curve 1 or 3 were in
effect. Essentially, then, each experiment simu-
lated the decision-making problem wherein a
manager assumes that a given biological situa-
tion is "true" and plans to meet it but then en-
counters a completely different situation.

The results of this first group of sensitivity ex-
periments are indicated in Table 10. The main
criteria for comparing performances under differ-
ent stock-recruit curves are the absolute and per-
centage differences in net economic returns indi-
cated in the last two columns of this table. In all
cases, relative to curve 2, harvesting-processing
systems performed better under curve 1 and
worse under curve 3. As seen from the larger re-
turns, catches, and biomasses generated and from
the fewer days of fishing required, curve 1 defined
a more productive biological regime relative to
curve 2. Likewise, from the smaller returns,
catches, and biomasses and from the generally
greater number of days of fishing required, curve
3 defined a less productive biological regime.

The economic consequences of imposing differ-
ent regulatory schemes can also be determined
from Table 10. Opening the inshore area would
generate about a 30^ improvement in net return.
Given our assumptions, such an increase is likely
due to the increased availability of O's and I's
which in turn leads to greater catches for the
same level of effort. On the other hand, a change
in season length would generate an improvement
in returns of 120-130%. Quite obviously, from
an economic viewpoint, the model indicates that
the preferable management scheme would be
a change to the 12-mo season. Barring that,
the next best scheme would be to open the in-
shore area.

However, these economic findings should be
tempered somewhat by sensitivity considera-
tions. Comparison of areas within seasons (Table
10) reveals that an open inshore area is less sen-
sitive to changes in stock-recruit relations than is
a closed inshore area. That is, the percentage
change in net returns is less for both curves 1 and

Table lO. — Sensitivity of optimal configurations to changes in stock-recruit curves and areal
restrictions, given M = 1.10 and deterministic availability.

Length

Stock-

Average

Diflerenrp

recruit

Rshing
time

biomass

Catch

Net return

^^ 111^^1^^'

season

Area

curve

(10^ MT)

(103 MT)

(106 dollars)

Absolute

8 mo

Inshore

144

3.92

491.4

6.010

—

closed

144

4.00

501.6

6.456

0.446

7.42

144

3.81

477.5

5.408

-0.602

-10.02

Inshore

141

3.87

537.1

8.014

—

open

140

3.96

547.1

8.454

0.440

5.49

142

3.75

523.9

7.43 k:

-0582

- 7.26

12 mo

Inshore

216

3.47

831.9

13.660

—

closed

215

3.57

870.5

15.341

1.681

12.31

216

3.32

796.9

12.136

-1.524

-11.16

Inshore

212

3.43

920.6

17 545

—

open

209

3.63

941.2

18.466

0.921

5.25

214

3.26

886.1

16.024

-1.521

- 8.67

'Situations used as standards for comparative purposes.

126

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

3 when the inshore area is open, greater when it
is closed. Also, in three of four comparisons of
seasons within areas, an 8-mo season is less sen-
sitive to changes in stock-recruit relations than is
the 12-mo season.

The greater sensitivity of the 12-mo season is
probably due to the greater level of effort exerted
(e.g., compare days fished) which would tend to
drive stock size down into more critical regions of
the stock-recruit curve and give rise to density-
dependent responses greater than those observed
under the 8-mo season. From a sensitivity view-
point then, harvesting-processing operations
planned for the 12-mo season or closed inshore
area would tend to suffer most from the present
lack of knowledge about stock-recruit behavior;
the 8-mo season or open inshore area would tend
to suffer least.

Considering our premise that trade offs might
be made between quotas and areal-seasonal re-
strictions, the above model results imply that giv-
ing up (trading off) an increased season length
represents a considerable loss of potential
economic benefit. Such a trade off would therefore
seem to require substantial compensation in the
form of increased quotas. Trading off a change in
areal restrictions, on the other hand, would seem
to provide considerably less bargaining power.
Moreover, opening the inshore area appears to
offer distinct advantages, not only in terms of
moderately increased net returns, but also in the
form of somewhat decreased operating risk given
a lack of biological knowledge. Consequently, the
model indicates that trading off a change in sea-
son length appears to be the most advantageous
tactic for plant and fleet managers if they seek
increased quotas.

Technological Change and
Employment

In their study of the San Pedro wetfish"* fleet,
Perrin and Noetzel (1970) estimated that the
number of jobs on vessels had decreased from 381
in 1963 to 238 in 1968. The figures reflected a
reduction not only in the size of the fleet but also
in the size of crew as well. In 1963 the average
crew size was 10.29 compared to the 1968 average

of 9.52. With such a decline in employment, it is
not surprising that the union opposes the intro-
duction of technology which would replace more
men (Stadelman 1974).

According to Hester et al. (1972), the applica-
tion of a power drum to purse seining by the
wetfish fishery would significantly reduce the size
of the crew. Based upon the foregoing author's
experiment with a 100-ton (91-MT) capacity ves-
sel, Stadelman (1974) estimated that for a 210-ton
(191-MT) purse seiner the introduction of a power
drum would reduce the crew from 10 to 6. This
would result in significantly reduced vessel
operating costs (Table 9) which might allow fleet
expansion and a subsequent increase in the over-
all level of employment. Simulation experiments
were therefore conducted to see if a favorable out-
come resulted which might dissuade the union
from opposing such technological innovation.

Table 11 lists the results obtained for a 12-mo
season for both the normal and the power drum
methods of purse seining. Use of the drum in-
creased net yield by 80% and the optimal level of
fishing effort by 38%. However, the optimal total
labor force was reduced from 544 required to man
the fleet to an estimated 459. Consequently, the
added vessels did not make up for the reduction in
crew size.

However, it should be noted that even with the
use of the power drum the level of employment
would exceed its 1968 level of 238 men. It is also
apparent that the additional net yield associated
with the power drum, some $2.6 million, might be
negotiated into a wage above $12,000. On the as-
sumption that 459 men would be employed, each
could receive an additional $5,664/yr and the
fishery would still yield the same annual net re-
turn as before the innovation. Alternatively, the
increased net yield could supply income to employ
215 workers in other activities at the $12,000
wage, whereas prohibition of the power drum
would save only 85 jobs in the fishery. This is
the type of trade off that must be weighed in
determining policy to increase the level of
employment.

^Wetfish are defined by Perrin and Noetzel ( 1970) to include
northern anchovy for reduction; and Pacific sardine, jack
mackerel, Trachurus symmetricus, chub mackerel, Scomber
japonicus, and Pacific bonito, Sarda chiliensis, for canning and
the fresh-fish market.

Table

11.

—Effect of

power drum on
long season.

employment

for a year-

Power
drum

Net yield
(millions)

Level of effort Labor
(standard vessels) force

Total gross

wages paid

(millions)

Wthout
With

$2.9
5.5

49
68

544
459

$6.5
5.5

127

FISHERY BULLETIN; VOL. 74, NO. 1

The foregoing results assume that the physical
efficiency of harvesting is not increased by the
power drum. The study by Hester et al. (1972)
revealed that the use of a power drum and fish
pumps to unload the nets often enabled the ex-
perimental vessel to get in an extra set during the
brief time fish were available before dawn. This
circumstance depended on the size of catches
being made since use of the equipment actually
increased the set time for very small catches. No
data were presented, however, as to the average
number of sets or the frequency of catch size for
evaluation of efficiencies.

The above analysis points up the importance
of union work rules permitting the use of new
technology. The application of the power drum to
vessels apparently would improve the economic
viability of the fishery, permitting its operation
even with old hulls or at fish meal prices below
$250/MT. Although use of the drum reduces crew
size on an individual vessel, its general adoption
apparently would provide considerable economic
incentive for fleet expansion, leading to an in-
crease in overall employment beyond its 1968
level.

To make this inference, however, we have as-
sumed away the real problem, which is not the
adoption of new technology but the alteration of
traditional union share agreements which pay
the crew a percentage of net revenues. Unless
new technology resulted in increased gross reve-
nue as well as a reduction in crew size, the same
share of the net revenue would simply be divided
among fewer crewmen, and the investor would
gain nothing to compensate him for the addi-
tional costs of the technological change. Con-
sequently, the present system does not allow the
investor a sufficient return, and the fishery suf-
fers in terms of employment levels as well as with
respect to economic efficiency.

DISCUSSION

In discussing his model of the ecological bio-
energetics of isopods, Hubbell (1971) indicates
that there is a twofold utility in modelling a given
system. First, the model can be regarded as a tool
to guide and orient future research on that sys-
tem. Second, once the model exhibits satisfactory
performance, it can be put to predictive use,
answering hypothetical questions about the con-
sequences of different input conditions upon sys-
tem behavior. As demonstrated by the preceding

applications, we feel that the northern anchovy
model definitely has the potential for fulfilling
both of these purposes.

However, in its current state of development
the model is admittedly speculative in some of its
content. Several of its shortcomings have already
been discussed, but perhaps its greatest failing is
that its behavior has not yet been adequately val-
idated. To do so would currently require the circu-
lar logic of testing the model against the very
data from which its assumptions and estimates
derive. Consequently we have been forced to rely
upon our own subjective view of what constitutes
well-behavedness in the model and have applied
this criterion in evaluating its performance.

According to Patten (1972), we probably could
do little more to validate the model since there
currently exists no theoretical base for approach-
ing this fundamental modelling problem. In any
regard, the predictive use of this model should
therefore be treated in only the most general of
terms, i.e., with the aim of gaining insight into
the structure and behavior of the anchovy fishery.
In this sense, it presently is a conceptual rather
than an analytical model.

This leaves its use as a tool for guiding and
planning research as the model's primary reason
for being. To that end it has proven quite useful,
providing a systematic means by which extant
data might be organized and pinpointing areas
characterized by a glaring lack of data. For
example, our approach to modelling stock-recruit
behavior was necessitated by a lack of appropri-
ate indices measuring recent stock and recruit-
ment sizes.

Additionally, we feel that the model provides
the capability for identifying and ranking critical
research areas. Management decisions must be
timely and as correct as possible, yet the cost of
collecting and analyzing relevant data is very
high both in money and time. Given budgetary
constraints, all research needs cannot possibly be
satisfied. Therefore, decision makers should be
asking themselves whether the cost of better in-
formation will be justified by a better choice of
management policy.

The model could play an important role here by
allowing the decision maker to test the sensitiv-
ity of his information upon policy alternatives.
Some policy sets will not be affected by slight
changes in estimates resulting from fuller infor-
mation: a somewhat higher growth rate than ini-
tially believed, for example, may not occasion any

128

TILLMAN and STADELMAN: SIMULATION MODEL OF ANCHOVY FISHERY

revision in policy. The degree of sensitivity thus
determines which information is trivial and
which is critical. Parameters of the model which
prove to have little or no effect on the decision
then need not be refined by further research.

ACKNOWLEDGMENTS

The work was based on parts of two disserta-
tions in partial fulfillment of the requirements for
the Ph.D. degree at the University of Washing-
ton. Support for the project was provided by the
West Coast Fishing Industry Study, NMFS Con-
tract No. 14-17-0007-1118, with James A.
Crutchfield, University of Washington, as princi-
pal investigator.

We are indebted to the late G. J. Paulik, Uni-
versity of Washington, for his continuous encour-
agement and support throughout all aspects of
the research and analysis which led to the com-
pletion of this study.

LITERATURE CITED

AHLSTROM, E. H.

1966. Distribution and abundance of sardine and anchovy
larvae in the California Current Region off California
and Baja California, 1951-64: A summary. U.S. Fish
Wildl. Serv, Spec. Sci. Rep. Fish. 534, 71 p.

BAXTER, J. L.

1967. Summary of biological information on the north-
ern anchovy Engraulis mordax Girard. Calif. Coop.
Oceanic Fish. Invest., Rep, 11:110-116.

BEVERTON, R. J. H., AND S. J. HOLT.

1957. On the dynamics of exploited fish populations.
Fish. Invest., Minist. Agric, Fish. Food (G.B.), Ser. 2,
19, 533 p.
CLARK, F. N., AND J. B. PHILLIPS.

1952. The northern anchovy (Engraulis mordax mordax)
in the California fishery. Calif Fish Game 38:189-207.
COLLINS, R. a.

1969. Size and age composition of northern anchovies (En-
graulis mordax) in the California anchovy reduction
fishery for the 1965-66, 1966-67, and 1967-68 seasons.
Calif Dep. Fish Game, Fish. Bull. 147:56-74.
1971. Size and age composition of northern anchovies (En-
graulis mordax)m the California reduction and canning
fisheries, 1968-69 season. Calif. Fish Game 57:283-289.
CUSHING, D. H.

1971. The dependence of recruitment on parent stock in
different groups of fishes. J. Cons. 33:340-362.
FOX, W. W.

1973. A general life history exploited simulator with pan-
dalid shrimp as an example. Fish. Bull., U.S. 71:1019-
1028.

FRANCIS, R. C.

1974, TUNP0P, a computer simulation model of the yel-
lowfin tuna population and the surface tuna fishery of
the eastern Pacific Ocean, [in Engl, and Span.]
Inter-Am. Trop. Tuna Comm, Bull. 16:235-279.

Gales, L. E.

1972. NEW GAMES: A multi-purpose interactive re-
source management program. Univ. Wash,, Quant. Sci.
Pap. 32, 40 p.
HAUGEN, C. W,, J, D, MESSERSMITH, AND R, H. WICKWIRE.
1969. Progress report on anchovy tagging off California
and Baja California, March 1966 through May 1969. In
The northern anchovy (Engraulis mordax) and its fish-
ery 1965-1968, p. 75-89 Calif Dep. Fish Game, Fish.
Bull. 147,
HESTER, F. J,, D, A, AASTED, AND R. E. GREEN,

1972. Experimental drum seining for wetfish in Califor-
nia. Commer, Fish, Rev, 34(l-2):23-32,
HUBBELL, S. P.

1971. Of sowbugs and systems: The ecological bioenerget-
ics of a terrestrial isopod. In B, C, Patten, Systems
analysis and simulation in ecology, Vol, 1, p, 269-
324, Academic Press, N.Y.

MACGREGOR, J. S.

1968. Fecundity of the northern anchovy, Engraulis mor-
dax Girard, Calif Fish Game 54:281-288.

MAIS, K. F.

1974, Pelagic fish surveys in the California Current, Calif
Dep, Fish Game, Fish Bull, 162, 79 p,
MESSERSMITH, J. D.

1969. A review of the California anchovy fishery and re-
sults of the 1965-66 and 1966-67 reduction seasons.
Calif Dep, Fish Game, Fish Bull. 147:6-32.

MESSERSMITH, J. D,, J, L, BAXTER, AND P. M. ROEDEL.

1969, The anchovy resources of the California Current re-
gion off California and Baja California, Calif. Coop.
Oceanic Fish. Invest., Rep, 13:32-38,

MILLER, D, J,, A, E. DAUGHERTY. F, E, FELIN, AND J. MAC-

Gregor.

1955, Age and length composition of the northern anchovy
catch off the coast of California in 1952-53 and 1953-
54, Calif, Dep, Fish Game, Fish Bull. 101:36-66.
MILLER, D. J., AND R. S. WOLF.

1958. Age and length composition of the northern anchovy
catch off the coast of California in 1954-55, 1955-56, and
1956-57. Calif Dep, Fish Game, Fish Bull, 106:27-72.
MURPHY, G. I.

1966, Population biology of the Pacific sardine iSandinops
caerulea). Proc, Calif Acad, Sci,, Ser 4, 34:1-84.

1973. Clupeoid fishes under exploitation with special ref-
erence to the Peruvian anchovy. Univ. Hawaii, Hawaii
Inst, Mar Biol,, Tech, Rep, 30, 73 p.

PATTEN, B. C.

1972. Systems analysis and simulation in ecology. Vol,
II. Preface, p. XI-XIV. Academic Press, Inc, N,Y,

PERRIN, W. F., AND B. G. NOETZEL.

1970. Economicstudy of the San Pedro wetfish boats. U.S.
Fish Wildl. Serv., Fish. Ind. Res, 6:105-138.

RICHER, W, E,

1958, Handbook of computations for biological statistics of
fish populations. Fish. Res. Board Can., Bull. 119, 300 p.
SCHAEFER, M. B.

1967. Dynamics of the fishery for the anchoveta Engraulis
ringens, off Peru. [In Span, and Engl.] Inst. Mar Peru,
(Callao) Bol. 1:191-303.

STADELMAN, D.

1974. Optimal policy and sensitivity analysis of the north-
em anchovy fishery: A simulation study, Ph,D, Thesis,
Univ. Washington, Seattle, 101 p.

129

FISHERY BULLETIN: VOL. 74, NO. 1

Tillman, M. F.

1968. Tentative recommendations for management of the
coastal fishery for Pacific hake, Merluccius productus
(Ayres), based on a simulation study of the effects of
fishing upon a virgin population. M.S. Thesis, Univ.
Washington, Seattle, 197 p.

1972. The economic consequences of alternative systems; a
simulation study of the fishery for northern anchovy, En-
graulis mordax Girard. Ph.D. Thesis, Univ. Washington,
Seattle, 227 p.

1975. Additional evidence substantiating existence of

northern subpopulation of northern anchovy, Engraulis
mordax. Fish. Bull., U.S. 73:212-215.
TILLMAN, M. F., AND G. J. PAULIK.

1971. Biological analysis of the northern anchovy fishery
system. Univ. Wash., Quant. Sci. Pap. 28, 65 p.
VROOMAN, A. M., AND P. E. SMITH.

1971. Biomass of the subpopulations of northern anchovy
Engraulis mordax Girard. Calif Coop. Oceanic Fish. In-
vest., Rep. 15:49-51.

Walters, C. J.

1969. A generalized computer simulation model for fish
population studies. Trans. Am. Fish. Soc. 98:505-512.

130

POPULATIONS OF SYMPATRIC SCULPINS, COTTUS ALEUTICUS

AND COTTUS ASPER, IN FOUR ADJACENT SALMONPRODUCING

COASTAL STREAMS ON VANCOUVER ISLAND, B.C.

J. C. Masoni and S. Machidori^^

ABSTRACT

General life history, distribution and abundance, age structure, and growth and survival are
documented for sympatric populations of two cottid fishes. Stream obstructions may largely determine
the distributional limits for both cottids with Cottus aleuticus penetrating farthest upstream. Biomass
density and size of individual fish increased with distance upstream, largest individuals living at the
upstream borders of their species ranges. Both sculpins were numerically most abundant in their lower
ranges, reflecting the common estuarine origin of benthic young. From 69 to 74% of their combined
biomass in the upper estuaries were C. asper while 75-100% was C. aleuticus in the upper stream
zone. Cottus asper grew more rapidly and mortality rates were similar, but the oldest C aleuticus was
age 8 and 145 mm in length, compared with age 6 and 144 mm for C asper. The length-weight
relation was similar for both species. The community role of these sculpins is explored with primary
focus on possible competition with the stream-dwelling salmonids, and recommendations are made
which might lead to increased production of salmonid smolts to the sea.

As part of a general study of the fish community
of Lymn Creek, populations of the sympatric
sculpins, Cottus aleuticus and C. asper, were
examined during 1968 with regard to population
structure, annual growth and mortality, and gen-
eral distribution and abundance in the system. In
addition, three adjacent streams (Cabin, Chef,
and Waterloo) were sampled in the fall of 1968 to
provide a comparative basis for interpreting the
findings at Lymn Creek. The present communica-
tion deals primarily with population characteris-
tics of sculpins in relation to life history. Their
role in the community, including possible compe-
tition with salmonids, is examined with a view of
enhancing salmonid production.

THE STUDY AREA

The four streams studied are neighboring sys-
tems emptying into the Strait of Georgia on the
east coast of Vancouver Island. They are small
streams (drainage area <20 km^, minimum sum-
mer flow <7 m^/min. Table 1), having similar
gradients and streambed materials, but Cabin
Creek is considerably smaller than the others.
Their watersheds are forested at a similar stage

'Department of the Environment, Fisheries and Marine Ser-
vice, Research and Development Directorate, Pacific Biological
Station, Nanaimo, B.C. V9R 5K6, Canada.

^Fisheries Agency of Japan, Far Seas Research Laboratory,
1000 Orido, Shimizu 424, Japan.

of second-growth conifers, primarily Douglas fir.
Lymn and Waterloo creeks closely resemble each
other, although the latter stream has fewer major
obstructions (logjams) hindering the upstream
migration of salmon. Lymn Creek differs from
the other three streams in having a swampy
sloughlike area resulting from beaver activities
near the estuary. Both Lymn and Chef creeks
course through some 200 m of intertidal meadow,
but Cabin and Waterloo creeks empty directly
onto the open beach. Extensive intertidal zones in
all four streams result at low tide when nearly
the entire zone is exposed to freshwater flow.

Unlike the other systems. Chef Creek is subject
to flow extremes, rapid runoff during freshets
and, during the late summer and early fall,
intermittent flow and isolated pools in the lower
reaches.

Cutthroat trout, Salmo clarki; coho salmon,
Oncorhynchus kisutch; three-spined stickleback,
Gasterosteus aculeatus; coastrange sculpin, C.
aleuticus; and prickly sculpin, C. asper, reside in

Table l. — Some physical characteristics of the four study

streams.

Drainage

Average

Minimum summer

area

width'

gradient

discharge

Stream

(km^)

(m)

(%)

(m^/min)

Cabin

2,3

1.5

1.2

0.5

Lymn

9.3

2.5

1.0

3.4

Waterloo

105

2.5

1.2

2.6

Chef

18.3

7.5

0.9

6.8

'Within the sculpin zone.

Manuscript accepted April 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

131

FISHERY BULLETIN: VOL. 74, NO. 1

all four streams. Chef and Waterloo creeks also
contain steelhead trout, S. gairdneri, and chum
salmon, O. keta. Chum salmon occasionally
spawn intertidally in Lymn Creek.

METHODS AND MATERIALS

Sampling the Populations

In Lymn Creek, sculpins were collected inci-
dentally to salmonids from April to July 1968. A
sampling schedule for cottids was initiated in
August and terminated in December 1968. Chef,
Cabin, and Waterloo creeks were sampled during
September and October.

Fish were collected in the estuaries by seine at
low tide. In the streams proper, collections were
made with a 440- V DC fish shocker (Smith-Roote
Laboratories, Mark V^). In both environments,
discrete sections of stream, usually 15- to 30-m
sections, were sampled and all fish captured were
removed.

Specimens were preserved in 5% Formalin. In
the laboratory, total length was measured to the
nearest millimeter and body weight to the near-
est 10 mg. Otoliths were removed for age deter-
mination.

No attempts were made to quantify the relative
or absolute efficiencies of the two sampling
methods. The habitat seined lent itself to efficient
seining, and it is considered that any increased
capture efficiency or size-related sampling bias
usually associated with electrical fishing devices
was, at least in part, cancelled by the increased
complexity of habitat typical of the stream proper
and the concentration of the two youngest age-
groups in the lower stream, including the es-
tuaries. Increased stream flow and turbid water
following the first significant rains in the late fall
probably reduced the efficiency of both collecting
methods to a considerable but unknown extent.
Therefore, growth and survivorship estimates
were based on data collected prior to the onset of
the rainy season.

In the laboratory, breeding activity was fol-
lowed by keeping adults allopatrically in 150-
liter fiber glass tanks at ambient freshwater
temperature with flow-through conditions, a rub-
ble substrate, and normal photoperiod. Em-

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

bryological development and larval responses to
salinity, illumination, current, and food were
investigated. Egg masses of known age and their
resulting larvae were kept in 3-liter glass jars
filled with aerated fresh water or seawater; and
mortality and feeding responses of larvae to mi-
crozooplankton were observed. The responses of
larvae of known age and salinity history to over-
head illumination and water currents were inves-
tigated in a Perspex test chamber.

Drift nets were set at several stations in
Lymn Creek during the hatching period in the
spring to document the timing and extent of the
hatching period, upper limits of the spawning
ground and characteristics of the fry moving
seaward.

Population Estimates

Estimates of population size in Lymn, Cabin,
and Waterloo creeks were attempted in the fall
for both species of sculpin. Population estimates
for Chef Creek were precluded by the large size of
the stream, which prevented representative sam-
pling across the stream at most stations. In the
other three streams, catches from individual sta-
tions were assumed to be representative of that
stream section, and population was calculated as
follows:

N= ICD

where C = station catch (fish/meter of stream)
where each station is representative
of a larger stream section D
D - stream section (in meters).

The estimated populations were distributed
among the various age-classes so as to reflect the
age-class composition of the station catches. Ad-
mittedly, these estimates are rather crudely de-
rived yet they yielded fairly consistent trends in
annual mortality, particularly for the Lymn
Creek populations (see Results, Annual Growth,
Mortality, and Length-Weight Relations). At-
tempts to apply mark and recapture techniques to
the problem of population estimation proved
fruitless due to extensive behavior changes
in marked fish following their release. These
changes (movement downstream or into the
streambed) seriously affected their vulnerability
to recapture and led to large scale overestimates
of actual population size.

132

MASON and MACHIDORI: POPULATIONS OF SYMPATRIC SCULPINS

Age Determination

Following dissection, otoliths were dried for
several days and then immersed in a 50% solu-
tion of glycerin and water. Otolith structure was
not clear when examination immediately fol-
lowed removal of the otolith from the specimen.
Otoliths of specimens preserved for more than 1
mo were partly decomposed by the preservative.

Whole otoliths were examined under a dissect-
ing microscope by reflected light against a black
background. In both species, the otolith had an
opaque nucleus around which were arranged con-
centric, alternating hyaline and opaque bands
extending to the margin. The opaque band
reflected rapid summer growth and the hyaline
band constituted the annulus. The first hyaline
band around the nucleus was not considered an
annulus but is assumed to reflect initial post-
larval growth, perhaps prior to the onset of a
benthic existence. The newly forming annulus
was readily discernible in specimens collected in
October and December.

Length-frequency histograms were found use-
ful to identify the young of the year (age 0) and
yearlings (age I).

RESULTS

General Life History

Both species of sculpins in these short coastal
streams are "coastal" forms (McAllister and
Lindsey 1960) which spawn during April and
May. The prickly sculpin undergoes a down-
stream spawning migration in the early spring
(Mason 1974a) and spawns in the estuary as re-
ported previously by Krejsa (1967). The coast-
range sculpin has been reported to make down-
stream migrations coincident with C. asper
(Shapovalov and Taft 1954; Hunter 1959) but no
such migration was recorded in Lymn Creek
where C. aleuticus spawned in situ throughout its
range in the stream as found in Alaskan streams
by McLarney (1968).

The breeding males are territorial and court
one or more females which deposit clusters of
adhesive eggs on the underside of large rocks or
debris forming the nest site. Following spawning,
the females depart and the males guard the eggs
until hatching. The newly hatched and trans-
parent larvae begin swimming upon hatching
and assume a pelagic life for some 30 days, grow-

ing from 5 mm at hatching to 12 mm in length
before assuming a benthic existence.

In the laboratory at 10°-12°C, the eggs of both
species were eyed at 9-10 days; the larvae were
active at 15 days; and hatching occurred 19-20
days following fertilization. Hatching commenced
in Lymn Creek on or before 11 May when water
temperature reached 10° C. On this date, larvae
began appearing in the driftnet catches and were
taken for some 5 wk until 19 June.

From drift net catches of the larvae in Lymn
Creek, coupled with laboratory studies on the
reproduction of both species, we concluded that
the eggs and larvae are euryhaline but survival
and growth of cultured larvae are better in sea-
water. Feeding on microplankton commenced
some 6-10 days following hatching of cultured
larvae when the yolk was noticeably depleted and
when most stream larvae were either in the
estuary or, in the case of coastrange sculpin
larvae, in the lower stream near the estuary.
Since the average size of the latter larvae in drift
samples from four stations located along 1,150 m
of stream above the estuary equalled that of
6-day-old larvae in culture at similar tempera-
tures, these larvae probably spend several days in
the nest vicinity and in downstream transport
following hatching.

Within several hours of hatching, larvae of
both species swam to the water surface and main-
tained themselves vertically immediately be-
neath the surface film by steady swimming
movements. This behavior was sustained through
the 25 days of culture in both fresh water and
seawater. Tests on 5-day-old and older larvae
showed that they were positively rheotactic at
velocities greater than 1 cm/s and swam actively
against the current in short bouts of rapid
swimming.

Post-spawned C asper remained in the estuary
of Lymn Creek throughout the summer and early
fall. Their return to upstream areas may coincide
with the spawning runs of salmon that commence
in October (Mason 1974a). The offspring of both
species remain in the estuarine zone until the
early summer of the following year when they
proceed to invade upstream areas.

Distribution and Relative Abundance

Both sculpin populations were limited to the
lower reaches and estuaries of all four streams,
with coastrange sculpins distributed farthest up-

133

FISHERY BULLETIN: VOL. 74, NO. 1

stream. The prickly sculpin was not found more
than about 1 km upstream from high tide mark
where the stream gradient did not exceed 1.5%,
whereas the coastrange sculpin penetrated up-
stream some 1.6-2.7 km from high tide mark in a
range of stream gradients not exceeding 6%. In
Cabin Creek, the smallest stream, the same gen-
eral difference between the two species in lon-
gitudinal distribution prevailed, but the dis-
tances involved were reduced by a factor of 10.

The upstream distributional limits of both species
in all four streams are indicated in Figure 1.

Habitat segregation was evident in cohabitated
stream areas, large Casper occupying the
deepest locations in pools, under log jams and
undercut banks. Intermediate-sized C. asper and
large C. aleuticus were also found at these sites
but at shallower depths. Riffle and glide areas
were mainly occupied by small and medium-sized
C. aleuticus.

CABIN CREEK

■ Caleuticus

LB D C osper
I

35 cm foils

45cm fails

225
200

100

75
50
25

-3 0 3 6 9 12 15 18 21 24 27

DISTANCE FROM HIGH TIDE (0) IN TENS OF METERS

5
<

^ 60

CHEF CREEK

3m woterfall /
-impassable /

■

■ intermittent

^^^^-^^

1 _ i^L^am^ __

150

125

100

24
mpassable log

•s.

O 3 6 9 12 15 18 21 24

DISTANCE FROM HIGH TIDE MARK (0) IN HUNDREDS OF METERS

Figure l. — Autumnal distribution and abundance of Cottus aleuticus and C. asper in relation
to stream profile and streambed obstructions. High time mark (0) is a reference benchmark
determined by the highest spring tide.

134

MASON and MACHIDORI: POPULATIONS OF SYMPATRIC SCULPINS

In upstream areas devoid of C. asper, the large
coastrange sculpins were found in habitats which
were usually occupied downstream by large
prickly sculpins. Although the subyearlings of
both species were found in riffle habitats of the
intertidal zone, some habitat segregation was
evident since prickly sculpins tended to concen-
trate in riffle areas where water depth increased
and velocity lessened.

The upstream movement of both sculpins is
clearly hindered by minor obstructions in the
stream, and their respective upstream distribu-
tional limits are marked by similar but different
obstructions. These obstructions were usually
small log jams involving minor waterfalls al-
though in Chef Creek C. aleuticus was stopped by
a high waterfall (3-4 m) plunging over bedrock.
Obstructions resulting in differences in water
level greater than 30 cm were impassable for
C. asper while differences greater than 45 cm
were necessary to prevent upstream movement of

Figure 2. — Stream obstructions delimiting the upstream
distribution of sculpins in Lymn Creek. A. 45-cm waterfall
caused by a large cedar log which blocks the upstream move-
ment of Cottus aleuticus. B. 30-cm waterfall at the concrete
culvert under Trans-Canada Highway 1, which blocks the
upstream movement of C. asper.

C. aleuticus. The limiting structures in Lymn
Creek are shown in Figure 2.

The upstream limits of both species of sculpin
bore a general association with stream gradient,
since both stream gradient and frequency of log
jams increase with distance upstream, as do
streambed disjunctions causing higher falls (Fig-
ure 1).

Both species were distributed downstream into
the intertidal zone but to dissimilar extent. For
C. aleuticus, the downstream limit was the upper
edge of the barnacle zone (station 0 minus 250 m,
Figure 1) while C. asper was not collected below
the upper edge of the oyster zone (station 0 minus
400 m).

Both sculpins were most abundant in the lower
parts of their ranges (Figure 1) although the data
for C aleuticus in Chef Creek are inconclusive,
possibly due to upstream movement of fish from
the region of intermittent flow although such
movement was not observed. Skewed distribution
is most pronounced in populations of the two
smaller streams, Cabin and Lymn creeks, and in
large part is due to inequitable distribution of the
age-classes. The subyearling sculpins were found
to inhabit a narrow zone about the high tide
mark, within which the two species showed ex-
tensive overlap (Figure 3). The relative contribu-
tions of subyearlings to total catches were rather
low in Chef and Waterloo creeks, suggesting poor
reproductive success or poor recruitment in 1968.
This aspect will be dealt with again in a sub-
sequent section.

Neither species of sculpin undertook any obvi-
ous seasonal movements in Lymn Creek during
the period from August to December (Figure 4),
although the large catches of age 1+ prickly

C. aleuticus

Casper

enes of intermittent pools

during Aug.- Sept in —

Chet Creek

«.. LYMN CREEK
o- CHEF CREEK
i— WATERLOO CHEEK
a- CABIN CREEK

-4,5 -3 0 3 6 9

DISTANCE FROM HIGH TIDE MARK (0) IN HUNDREDS OF METERS

Figure 3. — Autumnal distribution and abundance of sub-
yearling sculpins.

135

FISHERY BULLETIN: VOL. 74, NO. 1

-3 0 3 6 9 12 15 18

DISTANCE FROM HIGH TIDE MARK (0) IN HUNDREDS OF METERS

Figure 4. — Relative distribution of subyearling and older
(1 + ) sculpins in Lymn Creek during the period August-
December.

sculpins made at the head of tide in December
suggest the return upstream of individuals which
made the downstream migration in the previous
spring.

In general, size of fish increased with distance
upstream, the largest individuals of both species
living at the upstream border of their respective
ranges (Figures 5-7); however, subyearling and
yearling sculpins of both species tended to be
larger both upstream and downstream from the
head of tide.

Age Structure

Age structure of populations of both species in
Lymn, Chef, and Waterloo creeks was determined
by reading the otoliths. Only the first two age-
classes could be identified from length frequency
histograms (Figures 5-7), and these modes agreed
with the otolith readings. The Lymn Creek popu-
lations were aged from three successive monthly
samples (August-October) that indicated similar
lengths within age-groups for this time interval
(Tables 2, 3). Slight length increases for a given
age-group reflected detectable growth.

oleul'Cus^78a
N
ospef =421

METERS
- 1212

-— ^■•'1 I

>^»fl<F^'

— -24010-303

I •■ ^T''

5 6 7 8 9

TOTAL LENGTH (cm)

Figure 5. — Length-frequency histograms for sculpin popu-
lations in Lymn Creek from collections made in September
and October. Sampling stations are identified as distances
upstream or downstream (-) from high tide mark (0) in meters.

5£.' 5

O 5
CO

(T 15

^.0
2 5

Z 0
5
0
5

oieuttcus:2G7
osper ' 140

• ^f""" — "i»w—

250
159

70
24

-ALEUTICUS

-38

5 6 7 8 9 10 M

TOTAL LENGTH (cm)

Figure 6. — Length-frequency histograms for sculpin popu-
lations in Cabin Creek from collections made in September
and October. Sampling stations are identified as distances
upstream or downstream (-) from high tide mark (0) in meters.

Both sculpins showed differences in age struc-
ture in the three streams (Tables 2, 3). There
were eight age-classes of C. aleuticus in Lymn
Creek but only five in Chef and Waterloo creeks.
For C. asper there were six age-classes in Lymn
and Waterloo creeks but only four in Chef Creek.

136

MASON and MACfflDORI: POPULATIONS OF SYMPATRIC SCULPINS

X
CO

u.
o

UJ
CD

s
z

^^^"

Ff^^—

■• ALEUTICUS

WATERLOO CREEK

. jIIMm L .-I. . — ,j ^

— . J»l J ».

■• ASPER

2242
1727
1242

939

€36

333

121

I J** >^ ■■ ' ■ " "■ I ■

5 6 7 8 9 10 II

TOTAL LENGTH (cm)

Lymn Creek contained older fish of both species
but C. aleuticus lived longer than did C. asper.

Distribution of Biomass

The autumnal distribution of biomass by
stream zone was derived from population esti-
mates and length-weight data for both species of
sculpin in Lymn, Waterloo, and Cabin creeks
(Table 4). Density of sculpin biomass (grams per
square meter) was lowest in the estuaries and
increased upstream. Cottus aleuticus showed the
greatest increase in biomass density with in-
creased distance upstream, particularly when
proceeding from the estuary upstream into the
lower stream zone. About 69-94% of sculpin
biomass in the estuaries was C. asper, whereas
about 60-100% of sculpin biomass were C. aleuti-
cus in the upper zones whose downstream
boundaries were marked by the first significant
streambed obstruction. Species biomass in the

Figure 7. — Length-frequency histograms for sculpin popu-
lations in Chef and Waterloo creeks from collections made
in September and October. Sampling stations are identified
as distances upstream or downstream (-) from a high tide
mark (0) in meters.

Table 2. — Age distributions of Cottus aleuticus in successive 5-mm intervals of total length, sexes combined. Number in parentheses
indicates total number of fish when not all fish in the length interval were aged.

14.5-19.4

19.5-244

24.5-294

29.5-344

34.5-39.4

39.5-44.4

44.5-494

49.5-54.4

54.5-594

59.5-64.4

64.5-69.4

69.5-74.4

74.5-79.4

79.5-84.4

84.5-894

89.5-944

94.5-99.4

99.5-104.4

104.5-109.4

109.5-114.4

114.5-119.4

119.5-124.4

124.5-129.4

129.5-134.4

134.5-139.4

139.5-144.4

144.5-149.4

Total
length

Lymn Creek

Chef Creek

Waterloo Creek

August

September

October

September-October

October

(mm)

1 II III IV

VII

0 1 II III

1 II III

0 1 II III IV

(7)
12(83)
5(73)
3(50)
8(15)
9

8
40
37
25

(4)
3(65)
6(75)

10(80)
6(40)

6(24)
10(154)
14(169)

3(50)

9
16

Total fish 240 116 25 16 4 5 11 288

25 12

6 426

2
2
9
9
12
3
3

(8)
3(137)
18(132)
9(64) 2(13)
3(6) 22(42)
39

5
10

9
12

5
13
17
13

40 19 7 2 347 140 40 14 1 52 60 17 12 4

137

FISHERY BULLETIN: VOL. 74, NO, 1

Table 3.-

-Age distributions ofCottus asper in successive 5-mm intervals of total length, sexes combined. Number in parentheses
indicates total number of fish when not all fish in the length interval were aged.

14.5-19.4

19.5-24.4

24.5-29.4

29.5-34.4

34.5-39 4

39.5-44.4

44.5-49.4

49.5-544

54.5-59.4

59.5-64.4

64.5-69.4

69.5-74.4

74.5-79.4

79.5-84.4

84.5-89.4

89 5-944

94.5-99.4

99.5-104.4

104.5-109.4

109.5-114.4

114.5-119.4

119.5-124.4

124.5-129 4

129.5-134.4

134.5-139.4

139.5-144.4

Total fish

Lymn Creek

Chef Creek

Waterloo Creek

Total
length
(mm)

August

September

October

September-October
0 1 II III

October

1 II III IV

0 1 II III

1 II III

1 II III IV

(3)

(25)
3(51)
6(55)
1(41)
5(33)
8(13)
1

1
5
8
8
11
8
2

1
7

15
8
2
5
1

(4)
(16)

1(39)
9(46)

18(49)

33(38)

1(3)

2(19)

8(30)

11(27)

10
3

(2)

1(3)

8(21)

7(17)

(8)
1(9)

(1)
7
2
1 1

2
6

3
2 1

1 1

222

43 39 7

1 229

116

13 22

4 14

Table 4. — The autumnal distribution of sculpin (Coitus)
biomass in three streams.

Stream Zone

Lymn

Estuary

Lower

Upper

Total area

Waterloo

Estuary

Lower

Upper

Total area

Cabin

Estuary

Lower

Upper

Total area

Sculpin biomass

(kg)

(g/m2)

2.727
4.345
4.772

0.98
1.72
3.38

11 844

0.310

9.052

16.454

1.76

0.69
3.49
3.45

25.816

0.037
1.493
0.508

3.31

033
4.87
2.85

C. aleuticus

C. asper

Table 5. — The autumnal distribution of sculpin iCottus)
biomass by age-class in three streams, expressed as a per-
centage of species biomass.

(g/m2)

(%)

(g/m^j

(%)

Stream

Ill

VII

0.31

31.6

0.67

68 8

Lymn Creek:

0.88

51.2

0.84

48.8

C. aleuticus

6.7

12.5

10.5

19.6

225

17.6

4.4

6.2

3.38

100.0

—

C asper
Waterloo CreSR-

17.2

14.0

35.2

19.9

10.8

2.7

1,17

663

0.75

39.0

C. aleuticus

286

20.8

32.3

15.3

0.04

5.8

0.65

94.2

C asper

4.0

47.2

249

18.5

2.95

84.5

0.54

15.5

Cabin Creek:

3.45

100.0

—

C. aleuticus

14.9

42.5

20.5

17.8

4.2

3.09

93.5
12.1

0.56
0.29

6.5
87.9

C asper

17.1

15.4

39.6

27.8

0.04

4.42

90.8

0.45

59.2

1.72

60.4

1.13

39.6

Annu

a\ (;

row

th.

Vfor

tahl

V. ai

2.038 3.41

2.79 81.8

0.62 18.2

lower stream zone was nearly equal in Lymn
Creek but was predominantly C. aleuticus (85%)
in Cabin and Waterloo creeks.

The two sculpins differed in relative distribu-
tion of biomass by age group within their popula-
tions (Table 5). Whereas C. asper in their third
growth season (age II) constituted 35-47% of
population biomass, the biomass of C. aleuticus
populations in Lymn and Waterloo creeks was
more evenly distributed in older age groups. The
contribution of age I to population biomass of
C. aleuticus was considerably higher in the two
smaller streams than in Lymn Creek and 3-5
times higher than for C. asper in these two
streams.

Length-Weight Relations

The annual growth of both sculpins showed a
consistent ranking in three streams. Growth was
most rapid in Lymn Creek, intermediate in
Waterloo Creek, and slowest in Chef Creek (Fig-
ure 8) although the growth of C. asper in Lymn
and Waterloo was not statistically different. Dis-
similarities in rate of grovvi:h were greatest for
C. aleuticus, possibly reflecting its greater re-
liance on the productivity of the freshwater
stream than in the case of C. asper, which spends
considerably more time in the estuary through-
out its life history.

C. asper grew more rapidly than did C. aleuti-
cus, the age-specific disparity in weight gain
increasing with age. Growth of the Lymn Creek

138

MASON and MACfflDORI: POPULATIONS OF SYMPATRIC SCULPINS

AGE-CLASS

Figure 8. — Annual growth rates (weight) of Coitus aleuticus
and C. asper in Lymn, Chef, and Waterloo creeks.

population, which was first sampled in early
April, was most rapid during the spring and early
summer and nearly completed by mid-August.
The largest coastrange sculpin captured was 145
mm in length and 8 yr old while the largest prickly
sculpin was 144 mm in length and 6 yr old.

Length-weight linear regressions based on
logged data were calculated for both species in
the three largest streams (Table 6) and compared
by analysis of variance. The length-weight rela-
tion was similar for both species in all three
streams except for the coastrange sculpin in
Chef Creek, which was considerably lighter per
unit length than in the other two systems (F2 2737
= 77.5). Slow annual growth and a lower slope
(6) may reflect poorer feeding conditions or as-
sociated population stress during the late sum-
mer when the flow in a 500- to 600-m section of
this stream becomes intermittent.

Estimates of average annual mortality for both
species of sculpins in Lymn, Cabin, and Chef
creeks ranged between 58 and 75%, the differ-
ences between species and streams depicted in
Figure 9 being statistically non-significant. Al-
though similar for both sculpins, mortality in
Waterloo Creek was considerably lower than in
the other three streams 38-40%. No estimate of

Table 6. — Length-weight regression parameters (log y =
a + bx) for Cottus aleuticus and C. asper in three streams
on Vancouver Island, B.C.

Param-

aleuticus

asper

eter

Lymn

Chef Waterloo

Lymn

Chef
73

Waterloo

1,565

767 397

1,225

-5.312

-5.001 -5.297

-5.268

-5.143

-5.363

3.237

3.041 3.224

3.203

3.122

3.259

0.993

0994 0,996

0.992

0.997

0,998

Syx

00096

00122 0.0136

00115

0.0308

0.0268

annual mortality was attempted for C asper in
Chef Creek due to the small population present.
Despite close agreement to the linear function
of the majority of point estimates, some points for
young and old age-classes deviated considerably
and are taken to indicate poor survival, low re-
cruitment of subyearlings from the estuary in
some years, or inadequate sampling. For exam-
ple, poor survival of age I of C. asper is indicated
for Lymn, Waterloo, and Cabin creeks (Figure 9).
Similarly, age 0 of both species were poorly repre-
sented in Waterloo Creek, as were age 0 in Chef
Creek, despite intensive sampling in the down-
stream areas in which they were distributed. In
Chef Creek, age IV of C. aleuticus was very poorly
represented, suggesting either a sudden exten-
sive mortality or inadequate sampling effort in
the larger pools upstream where these fish reside.

DISCUSSION

The ecological importance of cottid fishes in the
simple fish communities of these coastal streams
remains essentially unknown but the present
findings appear to be timely in view of the resurg-
ing interest in enhancing the natural production
of anadromous stream salmonids. Previous

10.000

1000

q: 100

<
jr 10

C.ALEUTICUS
0 \ 8
A ■••\ 0

y \ \ ^"'■'

°'^ \ V

\ '.ox
\ \

\ \

\«
\ \
\ \

•

C ASPER

■

\ • \

\
\
\

-•- LYMN CREEK
-0- CHEF CREEK
■••A- WATERLOO CREEK
--D- CABIN CREEK

•

1 1 1

y Vr yl 0
AGE-CLASS

Figure 9. — Declining numbers with increasing age within
sympatric populations of Cottus aleuticus and C. asper in
four streams. Straight lines describe least-square regressions
of best fit.

139

FISHERY BULLETIN: VOL. 74, NO. 1

studies on C. aleuticus and C asper, which are
widely distributed and commonly abundant in
coastal streams from California to Alaska, have
emphasized their potentially destructive role as
predators on the eggs and fry of salmon and trout
(Shapovalov and Taft 1954; Hunter 1959; McLar-
ney 1967). Conversely, it has been generally
shown that sculpins in streams of the North
Temperate Zone prey incidentally on salmon and
trout, but sculpins do share a common source of
food — the benthic invertebrate community.

The probable importance of interspecific com-
petition in general, and for food in particular, in
such streams where the several species of fishes
consume in common a wide variety of food or-
ganisms has been readily acknowledged (Hartley
1948; Maitland 1965; Mann and Orr 1969) but
continues to defy quantitative analysis. The over-
lapping summer foods of juvenile coho salmon,
cutthroat trout, and coastrange sculpins in Cabin
Creek (Table 7) clearly show the possibility of
competition for food in the present study streams.
Numerically, Ephemeroptera and Diptera were
important in all three diets but most important in
the coho salmon diet, while Trichoptera were
most important in the trout and sculpin diets.
The sculpins showed the least varied diet as
Ephemeroptera, Diptera, and Trichoptera com-
posed nearly 95% of the food items consumed.
Dietary differences can be related to behavioral
differences in feeding and habitat response. The

Table 7. — The percentage composition by frequency of
occurrence (0) and number (N) of the midsummer (June-July)
foods eaten by juvenile coho salmon, cutthroat trout, and
coastrange sculpins in Cabin Creek. Based on 30 fish of each
species collected simultaneously.

Coho

Troul

Sculp
<7cO

Food category

Oligochaeta

10.0

Diplopoda

—

40.0

6.2

—

Collembola

23.3

—

Ephemeroptera

46.7

17.4

30.0

9.2

60.0

30.3

Plecoptera

16.7

4.1

20.0

3.1

3.3

Hemiptera

20.0

3.3

10.0

1.5

—

Coleoptera

Adults

30.0

6.2

30,0

Larvae

13.3

1.7

—

Trichoptera

10.0

2.1

400

262

56.6

44.9

Lepidoptera'

3.3

50.0

6.7

2.8

Diptera

Adults

70.0

285

20,0

7.7

—

Larvae

26.7

20.7

43,4

154

33.3

18.3

Hymenoptera'

10.0

1.2

20,0

3.1

—

Araneida

36,7

6.6

10,0

1.5

—

Acarina

6.7

1.7

23.3

10.8

3.3

1.8

Gasteropoda

3.3

—

'Refers to adult stage, all categories of Insecta are larval stages unless
noted otherwise.

sculpins were abundant in all habitats but ate
few foods of surface origin, being crepuscular
grazers on the benthos. The trout were princi-
pally riffle-dwellers and grazed the benthos (both
trout and sculpins ate large numbers of Trichop-
tera larvae) but exploited the invertebrate drift to
a lesser extent than did the coho salmon, which
preferred the pool and glide habitats of low cur-
rent velocity. Despite this behavioral diversity,
niche differentiation remains poorly developed in
the Eltonian sense discussed by Weatherley
(1963) who proposed that the niche be defined as
"...the nutritional role of the animal in its
ecosystem... ."

Recent experiments have clearly illustrated
that populations of juvenile coho salmon in these
streams are limited by their food supply during
the summer months (Mason 1974b, 1974c). Rates
of growrth, survival and emigration were amena-
ble to manipulation by varying population den-
sity and food availability. Thus, in that young
coho salmon share a common food supply with
both trout and sculpins, the likelihood of food
competition is strongly suspected.

Since direct documentation of competition
among stream fishes in natural environments
continues to elude us, the inferential definition of
competition proposed by Maitland (1965) appears
to have greater utility than the modus operandi
definition of Larkin (1956), ". . .the demand, typi-
cally at the same time, of more than one or-
ganism, for the same resources of the environ-
ment in excess of immediate supply." Maitland
(1965) suggested that competition occurs "...
when the presence of more than one species
causes the average total biomass (standing crop)
of one of them to be less than it would be if that
species were existing alone — species which are
directly parasitic or predatory on one another
being excepted."

Fish biomass in small coastal streams of Van-
couver Island usually ranges between 7 and 10
g/m^ in midsummer (unpubl. data). Of this 3-6
g/m^ (50-80%) consists of sculpins (C. asper and
C. aleuticus) in the first several kilometers above
the estuarine zone. Studies by Brocksen et al.
(1968) have shown that, within the carrying
capacity of laboratory streams producing natural
drift foods, production of cutthroat trout was de-
termined by the biomass ratio of trout and scul-
pin, C. perplexus, at time of stocking, whereas
sculpin production remained independent of trout
biomass. These results were obtained over a

140

MASON and MACfflDORI: POPULATIONS OF SYMPATRIC SCULPINS

range of species biomass levels commensurate
with those encountered in nature and suggest
that the availability of drift foods for the trout
was determined by the intensity of grazing by
sculpins on the stream benthos.

From the present study, the restricted ability
of both species of sculpins to surmount obstacles
in the streambed, coupled with the life history
features of planktonic young and downstream
spawning migrations, lend themselves to the po-
tential development of a management strategy
for enhancing the production of salmonid smolts
to the sea. If the findings of Brocksen et al. (1968)
can be corroborated in stream simulator systems
more closely approximating the natural envi-
ronment, studies on the locomotory ability of
these sculpins relative to the performance of their
communal salmonids could provide the design
criteria for physical barriers to be located on test
streams at suitable sites above the influence of
high tide.

LITERATURE CITED

BROCKSEN, R. W., G. E. DAVIS, AND C. E. WARREN.

1968. Competition, food consumption, and production
of sculpins and trout in laboratory stream communi-
ties. J. Wildl. Manage. 32:51-75.
Hartley, p. H. T.

1948. Food and feeding relationships in a community
of fresh-water fishes. J. Anim. Ecol. 17:1-14.

Hunter, J. G.

1959. Survival and production of pink and chum salmon
in a coastal stream. J. Fish. Res. Board Can. 16:835-886.
KREJSA, R. J.

1967. The systematics of the prickly sculpin, Cottus
asper Richardson, a polytypic species. Part H. Studies
on the life history, with especial reference to migra-
tion. Pac. Sci. 21:414-422.

Larkin, p. a.

1956. Interspecific competition and population control
in freshwater fish. J. Fish. Res. Board Can. 13:327-342.
Maitland, p. S.

1965. The feeding relationships of salmon, trout, min-
nows, stone loach and three-spined sticklebacks in the
River Endrick, Scotland. J. Anim. Ecol. 34:109-133.
MANN, R. H. K., AND D. R. O. ORR.

1969. A preliminary study of the feeding relationships
of fish in a hard-water and a soft-water stream in
southern England. J. Fish. Biol. 1:31-44.

Mason, J. C.

1974a. Movements of fish populations in Lymn Creek,
Vancouver Island: A summary from weir operations
diu-ing 1971 and 1972, including comments on species
life histories. [Can.] Dep. Environ., Fish. Mar. Serv.
Tech. Rep. 483, 35 p.

1974b. A first appraisal of the response of juvenile coho
salmon (O. kisutch) to supplemental feeding in an
experimental rearing stream. [Can.] Dep. Environ.,
Fish. Mar. Serv. Tech. Rep. 469, 21 p.

1974c. A further appraisal of the response to supple-
mental feeding of juvenile coho (O. kisutch) in an
experimental stream. [Can.] Dep. Environ., Fish. Mar.
Serv. Tech. Rep. 470, 26 p.
McAllister, D. E., and C. C. Lindsey.

I960. Systematics of the freshwater sculpins (Cottus)
of British Columbia. Natl. Mus. Can. Contrib. Zool.,
Bull. 172:66-89.
McLARNEY, W. O.

1967. Intra-stream movement and food habits of a popu-
lation of coastrange sculpins, Cottus aleuticus, in rela-
tion to a spawning run of the pink salmon, Oncorhyn-
chus gorbuscha. Ph.D. Thesis, Univ. Michigan, Ann
Arbor, 154 p.

1968. Spawning habits and morphological variation in
the coastrange sculpin, Cottus aleuticus, and the prickly
sculpin, Cottus asper. Trans. Am. Fish. Soc. 97:46-48.

Shapovalov, l., and a. C. TAFT.

1954. The life histories of the steelhead rainbow trout
(Salmo gairdneri gairdneri) and silver salmon (Oncorhyn-
chus kisutch). Calif Dep. Fish Game, Fish Bull. 98:1-375.

Weatherley, a. H.

1963. Notions of niche and competition among animals,
with special reference to freshwater fish. Nature
(Lond.) 197:14-17.

141

REVIEW OF THE DEEP-SEA FISH GENUS SCOPELENGYS

(NEOSCOPELIDAE) WITH A DESCRIPTION OF A NEW SPECIES,

SCOPELENGYS CLARKEI, FROM THE CENTRAL PACIFIC

John L. Butler^ and Elbert H. Ahlstrom^

ABSTRACT

Scopelengys has been known previously from a few widely scattered collections. Recent collections by
the Scripps Institution of Oceanography in the Pacific, the RV Walther Herwig in the Atlantic, and
the International Indian Ocean Expedition have made possible a critical study of this genus. No
significant differences were found in either morphometric characters or meristic counts between
specimens of S. tristis Alcock from the eastern North Pacific (la 1. 16 ° to 33°N, long. 117° to 126°W) and
those from the eastern South Pacific (lat. 5° to 16°S, long. 77° to 90°W). When Pacific Ocean specimens
were compared with those from the Atlantic and Indian oceans, no significant differences were
found in morphometric characters, and although differences in average meristic counts were some-
what larger between oceans than among Pacific specimens, such differences exceed one for only one
meristic character (gill rakers), and the ranges for all counts from all oceans almost completely
overlapped.

Scopelengys clarkei is described from the central North Pacific. It differs from S. tristis mainly in
pectoral ray count (2.5 average difference), average counts of vertebrae (3.3 average difference),
deeper caudal peduncle, narrower maxillary, and in a differently pigmented larva.

In 1890, Alcock described a new genus and
species, Scopelengys tristis, from a single denuded
specimen collected in the Arabian Sea. Although
there was no evidence of photophores, Alcock
placed his new genus in the family Scopelidae
( = Myctophidae) allowing that the "exact position
among the Scopelidae cannot be accurately de-
fined at present." Garman (1899) described S.
dispar from two specimens collected in the Gulf of
Panama. Garman distinguished S. dispar from S.
tristis by its lower dorsal- and anal-fin ray counts.
Scopelengys dispar was considered a junior
synonym by Parr (1928), Bolin (1939), and Nor-
man (1939). Until 1963, Scopelengys was known
only from the Indian and Pacific oceans. Its dis-
covery in the Caribbean Sea by Mead (1963) re-
sulted in the description of a third species, S.
whoi Mead.

A recent survey of mid-water fishes conducted
by the California Cooperative Oceanic Fisheries
Investigations (CalCOFI) provided us with
specimens which indicated that two species of
Scopelengys were present in the Pacific Ocean.
Additional specimens made available to us by
Thomas A. Clarke of the Hawaiian Institute of

'Smithsonian Institution, Southwest Fisheries Center, Na-
tional Marine Fisheries Service, NOAA, La Jolla, CA 92038.

^Southwest Fisheries Center, National Marine Fisheries Ser-
vice, NOAA, La Jolla, CA 92038.

Marine Biology (see in this regard Clarke 1973),
confirmed that the second form was an unde-
scribed species. Study of Scopelengys from the At-
lantic, Pacific, and Indian oceans indicates thatS.
dispar andS. whoi Mead are synonyms ofS. tris-
tis Alcock.

METHODS AND MATERIALS

Measurements were made following Hubbs and
Lagler (1958). Measurements are given in percent
of standard length (SL), unless indicated other-
wise. Only lath-shaped gill rakers on the first gill
arch are included in gill raker counts. Vertebral
counts were determined from radiographs; the
urostyle was included as one vertebra.

Morphometric and meristic data were obtained
from 211 specimens from the Atlantic, Pacific, and
Indian oceans. Subsamples equal to the smallest
N (32 in the Atlantic) were randomly taken from
the Indian Ocean, the eastern North Pacific be-
tween lat. 16° and 33°N and long. 117° to 126°W,
and the eastern tropical Pacific between lat. 5°
and 16°S and long. 77° to 90° W. Morphometric
data were compared by analysis of covariance.
Meristic data were compared by Tukey's multiple
comparison procedure at the 5% level (Rothschild
1963).

Material was examined from the following col-
lections: Scripps Institution of Oceanography

Manuscript accepted June 1975.

FISHERY BULLETIN: VOL. 74, NO. 1. 1976.

142

BUTLER and AHLSTROM: NEW SPECIES, SCOPELENGYS CLARKEI

(SIO); University of Southern California (USC);
Institut fiir Seefischerei, Hamburg (ISH);
Museum of Comparative Zoology (MCZ); U.S. Na-
tional Museum (USNM); International Indian
Ocean Expedition (IIOE); and Field Museum of
Natural History (FMNH).

GENUS SCOPELENGYS ALCOCK 1890

Type-species Scopelengys tristis Alcock, by
monotypy.

Description. — Head and body laterally com-
pressed, eyes small, mouth large. Premaxillary,
dentary, and palatines with bands of villiform
teeth. Teeth absent at symphysis of upper and
lower jaw. Vomer indented at head with teeth in
two patches. Teeth on basihyal and on gill rakers.
Anterior gill rakers reduced to toothed knobs.
Maxillary extending past eye, expanded pos-
teriorly. Supramaxillary present. Head and body
covered with large deciduous, cycloid scales. Pec-
toral fins lateral, extending beyond bases of pel-
vic fins. Pelvic fins abdominal. Origin of dorsal
fin about over base of pelvic fin. Anal fin com-
pletely behind dorsal. Base of adipose fin over
posterior half of anal fin. No photophores. No
swim bladder in adults.

D 11-13; A 12-14; P 12-17; V 8; Br 8; C principal
19 (1 + 17 + 1); procurrent C 6-9 dorsal and 7-8
ventral, hypurals (including parhypural) 4-1-3;
epurals 3; uroneurals 2. Urostyle with two centra.
As in all myctophiform fishes retaining two ural
centra (personal observation reenforced by
Rosen and Patterson 1969), the anterior ural cen-
trum (labelled PUi -I- Ui in Rosen and Patterson)
supports both the parhypural and the 2 inferior
hypurals, whereas the posterior ural centrum
(U2 in Rosen and Patterson) is associated exclu-
sively with the 4 superior hypurals.

Scopelengys tristis Alcock

Scopelengys tristis Alcock 1890:302.

Scopelengys dispar Garman 1899:254, plate 54,
fig. 2-2d.

Scopelengys lugubris Garman 1899:400, (syn-
onym Scopelengys dispar).

Scopelengys whoi Mead 1963:255, fig. 1.

Description of Adult

Body moderately slender, maximum body depth

at nape, tapering to a narrow caudal peduncle
(Figures lA, 2A); body depth at dorsal origin
11.7-19.8 (15.4); least depth at caudal peduncle
5.6-8.3 (6.8). Dorsal profile of head slightly con-
cave; head length 24.4-33.9 (29.4); head depth
16.7-25.5 (20.2); eye small, orbit 3.1-4.2 (3.5);
snout 7.5-10.1 (8.8). Width of maxillary as per-
centage of its length 29.9-36.7 (32.2). Snout to:
dorsal fin origin 36.1-47.0 (41.9); anal fin origin
56.4-72.6 (66.4); ventral fin origin 34.7-48.0 (41.8).
Meristic Data.— D.11-13 (11.5); A 12-14 (13.0);
P 14-17 (15.4); vertebrae 29-32 (30.8); total gill
rakers 7-11 (8.5).

Larvae

Twenty-five specimens 3.5-10.3 mm were avail-
able from the eastern Pacific. Measurements and
counts were given for two eastern Pacific (EAS-
TROPAC) specimens (6.2 and 6.4 mm SL) by
Okiyama (1974) and the smaller specimen illus-
trated. The larvae have a small round eye with-
out choroid tissue, a snout as long proportionately
as in adults, a gut terminating just forward of the
anal fin, and a gas bladder, best seen on late
preflection and flexion specimens, becoming
obscured by overlying musculature in larger
postflexion specimens.

Rays form early in the pectoral fins; a 3.5-mm
specimen has large pectorals extending posteriad
to the anus; caudal fin forms and notochord flex-
ion occurs between ca. 5 and 7 mm; dorsal and
anal fins form during flexion; pelvic buds appear
between 6.5 and 7.0 mm; fin formation, including
procurrent caudal rays, complete by about 10.0
mm. Pigmentation is scanty; pigment develops on
dorsal margin of peritoneal cavity, spreading lat-
erally on prefiexion and flexion stage specimens
but becoming obscured on postflexion larvae;
preflexion larvae have a series of 6 or 7 small,
inconspicuous spots along the ventral margin of
the tail which are later obscured by the anal fin
formation and lacking on late postflexion larvae;
head pigment, best developed on postflexion
specimens, consists of a striking horizontal bar
extending from snout to eye and continuing be-
hind the eye onto the operculum (Figure 3A).

Distribution

Records are from the tropical Atlantic, Pacific,
and Indian oceans (Figure 4). The range is ex-
panded poleward in the eastern part of the Pacific

143

FISHERY BULLETIN: VOL. 74, NO. 1

Figure l. — A: Scopelengys tristis, 126 mm, Velero IV, cruise 1238, stn. 18762/10. B: S. clarkei, 176 mm. SIO 73-160, holotype.

Figure 2.— A: Scopelengys tristis, 126 mm, Velero IV, cruise 1238, stn. 18762/10. B: S. clarkei, 176 mm. SIO 73-160, holotype.
144

BUTLER and AHLSTROM: NEW SPECIES, SCOPELENGYS CLARKEl

Figure 3. — A: Scopelengys tristis, 13.9 mm, from the western Indian Ocean. B: S. clarkei, 15.4 mm, from off Hawaii.

—

^^ .^^

4-r^-^^

M \

V ^ Tyi/^/^rn-N-MQ^ \

1 lA

A W Hj77-/^uTrH~L^ ^ a:*^ <^ * \

UTT

XtUA

\ iw w/74sZyA7WT^;ai^«^ ^ V

XU-

VA\. .r^^Vi un-i~un~H4^

^ \

/^^'■^~ld//rT~r~l~L

TTTT

xti

IX^ (<tWA /%cZ7y/7S^/Xzr^^^:ffl

/ / / 1

a_3Sk x^^xr^^ HOJ^ m~H-U4T v ^s^r-.^. ^ Rr

■I'M

_LP».^^\T\ L'.Ttf+44~n/ W 2uV^ X^V N\

rrrr

*T\l

-PMOtTA rtfi-i-U-LLU ^ z£J_UX f+t^^

••

XUJ^^W^-

rNi-hH-M-JiJV

'=k::AXjLU

OrfHlYII

(- 4^-Li

L. :ii 1

^TC-

■i \

^IT 1

I "

..J

•

o*-

. ,1-

•

\r 1

■-^. „ . . yi 1 : i I

•0.

«•

M^^

■

•, 1

' ^

[

\>-

rm.

S ?^ 1

■

[ ^

i i

''^rvX'Vh\\.

1 1

-tft+J .-

I AT

-U-iu

Tttf+^

\\\t^ VTITr

•1

-rh+ruji vrr

rni_L

1 .^

jTTTT

~j~f~'~~-LJj

TTT

IlttH-UlN \ kWL^

44-4-

XWuJJJt

/~l--^~*~~i~[/

■1

Xl^^-^

JJfT

tf-wZ/yTTV- A AW\r\

TTLL

/njJTFf-hh

r-t-L^/"

-TT^

i -LL

1 / /X

;

rti~N~LrTf~U V WUrVr^rrr

XU-r

/^\ L

Liliihr-

-///"'"TV

■"^

TOX

mffSM^ \^^^

'\\\\

llLr

tj//7~/~/-/J'yy~TC/

JJTn

DXtt

1 1/ ,rT7

T/VVVtCV/Ts//

r\ 1

Vr^

-^m

T~r7~/~/yyry0//

^^^^^^^ \^^^^^^

-4ra

i~f~f~^~/~i/yyyL/y^

\X>cS<<"^\wvV\'^

LLL

Ji//77

/rr-Zj/yy^/y^//

N^^^^^vXIliL

W^^^^^^^ ^""^^^^^^

\II_LLZ7777

/Ty^y/yy^/y^/

V \l

///

WvmiLWZZ=^

'7%///yCy'?C^

^^^ ^^^^ __

^^"

^^•«*TS

UM/THM (Ouiu.

Mf * MK

UfC

■o

Figure 4. — Distribution of Scopelengys tristis (circles) and S. clarkei (triangles). Small symbols indicate larvae.

145

FISHERY BULLETIN: VOL. 74, NO. 1

and Atlantic oceans and is narrowed along the
equator to the west. In the western part of the
Pacific and Atlantic, the species appears to be
rare. Records of larvae from the Indian Ocean
were presented by Nellen (1973).

Geographic Variation

Most of the specimens studied from each area
were in poor condition, which added to the vari-
ability of body proportions (Table 1). No significant
difference was found in any morphometric
character between regions. Meristic characters of
32 specimens each from four areas are presented
in Table 2. Samples from the two eastern Pacific

areas showed no significant differences between
means of any meristic character Indian Ocean
specimens differed from Pacific material in mean
vertebral counts (30.4 vs. 30.9), pectoral-fin ray
counts (15.2 vs. 15.7), and in gill raker counts (9.1
vs. 7.9). Atlantic material differed from Pacific
material in dorsal-fin ray counts (12.0 vs. 11.4), in
anal-fin ray coimts (13.4 vs. 12.8), in gill raker
counts (9.2 vs. 7.9), and in pectoral-fin ray counts
(15.0 vs. 15.7). Atlantic material differed from In-
dian Ocean material in dorsal-fin ray counts (12.0
vs. 11.1), anal-fin ray counts (13.4 vs. 12.9), and
vertebral counts (31.1 vs. 30.4). Although these
differences are small, they are as marked be-
tween Indian and Atlantic ocean specimens as be-

TABLE 1. — Comparison of morphometric characters of Scopelengys tristis from four geographic areas {N = 32 for each area).

Eastern

North Pacific

tropi

cal Pacific

Indian Ocean

Atlantic Ocean

Character

Mean

Range

Mean

Range

Mean

Range

Mean

Range

Standard length (mm)

124.0

73.8-154.3

130.2

74.5-172.0

104.6

28.8-185.9

133.1

69.5-185.8

Head length

30.9

28.7- 33.9

28.5

28 2- 33.2

30.5

27.8- 33.5

27.6

24.4- 32.2

Head depth

20.2

18.6- 25.5

20.0

16.7- 22.5

20.9

17.2- 24.6

19.6

17.5- 21.7

Snout to origin of dorsal fin

42.7

39.6- 45.5

42.5

395- 46.2

41.9

36.1- 47.0

40.6

37.1- 43.6

Snout to ongin of pelvic fin

41.7

39.4- 45.3

428

39.4- 48.0

41.2

34.7- 46.1

41.6

37.8- 46.9

Snout to ongin of anal fin

663

63.0- 69.5

675

62.5- 72.1

65.2

56.4- 71.2

66.6

61.3- 72.6

Least depth of caudal peduncle

6.8

5.7- 8.1

7.0

6.1- 7.9

6.7

56- 8.3

6.9

5.9- 8.0

Body depth at origin of dorsal fin

16.0

13.3- 19.4

15.6

12.6- 19.3

14.5

11.7- 17.2

15.7

12.6- 19.8

Table 2.— Meristic data for Scopelengys tristis from the eastern North
Pacific, eastern tropical Pacific, the Indian Ocean, and the Atlantic Ocean.

Meristic character
Area

Numbers of character
and frequency

Mean

Overall
mean

Dorsal rays

32
32
32
32

64
64
64
64

32
32
32
32

63
63
63
63

11.44
11 34
11.12
11.97

12.75
12.81
12.94
13.41

15.59
15.88
15.19
1503

30 88
31.03
30.38
31.06

7,86
7.89
9.08
9.22

Eastern North Pacific
Eastern tropical Pacific
Indian Ocean
Atlantic Ocean

11.47

Anal rays

Eastern North Pacific
Eastern tropical Pacific
Indian Ocean
Atlantic Ocean

10
8
5

20
22
24

2
2
3

12.98

Pectoral rays

Eastern North Pacific
Eastern tropical Pacific
Indian Ocean
Atlantic Ocean

33
20
48
62

7
12

15.42

Vertebrae

Eastern North Paafic
Eastern tropical Pacific
Indian Ocean
Atlantic Ocean

1
1

5
3

20
3

4
2

30.84

Gill rakers

10 11

Eastern North Pacific
Eastern tropical Pacific
Indian Ocean
Atlantic Ocean

17
1

7 2
12 2

8.51

146

BUTLER and AHLSTROM: NEW SPECIES, SCOPELENGYS CLARKEI

tween specimens from these areas and from the
Pacific. Because there is no clinal pattern in the
variation and because of extensive overlap in all
counts, no taxonomic importance was placed on
the small meristic differences.

Garman distinguished S. dispar from S. tristis
by the lower dorsal and anal-fin ray counts: D 11
vs. 12 and A 12-11 vs. 13 (Garman 1899). The types
of S. dispar are in poor condition but the anal fins
appear to have 12 or 13 rays (Robert Schoknecht
pers. commun.). The counts of S. dispar are
within the range of S. tristis. Scopelengys dispar
has been correctly considered a junior synonym
by Parr (1928), Bohn (1939), and Norman (1939).
Scopelengys lugubris Garman 1899:400, the
specific name regarded as a lapsus calami by
Bolin (1939), is a synonym ofS. dispar, hence ofS.
tristis. Scopelengys whoi was described from the
Carribbean Sea (Mead 1963). The diagnosis was
based on a shorter head, higher number of anal
fin rays (14 vs. 12-13), and the insertion of the
pelvic fin in advance of the origin of the dorsal.
According to Mead (1963), however, the head
length is ". . .a poor measurement because of the
condition of the opercular flap." The anal-fin ray
count is within the range of S. tristis (Table 3).
The insertion of the pelvic fin is a variable
character in S. tristis. In most specimens the fin is
inserted below the origin of the dorsal fin but in-
sertion in advance of the dorsal is not uncommon.
Based on this study, we conclude that S. whoi is a
junior synonym of S. tristis.

Study Material

PACIFIC OCEAN ADULTS.— SIO 51-186 1
(134); SIO 64-21 6(78-148); SIO 65-243 2(122-134);
SIO 64-997 1(122); SIO 65-244 1(75); SIO 55-229
9(31-113); SIO 65-206 1(92); SIO 60-212 4(20-133);
SIO 52-309 2(36-56); SIO 73-170 1(49); SIO 73-171
1(30); SIO 55-265 1(54); SIO 65-620 1(139); SIO
65-606 4(92-151); SIO 65-220 5(14-138); SIO

65-611 17(85-176); SIO 51-84 3(74-123); SIO
69-497 6(92-170); SIO 72-186 8(73-179); SIO
65-215 1(121); SIO 54-124 1(147); SIO 52-367
1(145); SIO 60-232 1(168); SIO 65-213 3(88-158);
SIO 60-219 2(42-170); SIO 55-246 4(65-140); SIO
68-579 1(140); SIO 53-235 1(154); SIO 51-146
3(127-144); SIO 65-603 17(62-160); SIO 55-244
2(159-167); SIO 72-195 17(88-175); SIO 65-608
14(43-200); SIO 72-193 2(106-169); SIO 72-192
18(10.2-177); SIO 60-216 2(42-76); SIO 60-218
1(48); SIO 66-355 1(135); SIO 69-19 1(24); SIO
72-182 1(90); SIO 66-407 1(42); SIO 64-24 1(116);
SIO 60-234 1(69); SIO 64-13 1(113); SIO 52-409
1(65); SIO 59-202 1(83); SIO 52-90 1(113); SIO
64-15 1(85); SIO 63-444 1(103); SIO 60-243 4(18-
44); SIO 68-534 1(28); SIO 65-443 1(142); SIO
68-104 1(97); SIO 60-209 1(78); SIO 52-363 2(56-
115); SIO 64-28 3(95-144); SIO 57-43 1(126); SIO
65-237 1(128); SIO 61-32 2(105-106); SIO 63-42
1(109); SIO 66-30 1(113); SIO 51-45 1(132); SIO
60-215 7(19-94); SIO 52-32 1(150); SIO 50-270
2(110-115); SIO 51-77 1(110); SIO 51-189 1(120);
SIO 54-82 1(107); SIO 54-102 2(116-147); USC Vel-
ero IV, cruise 1238, stn. 18762/10; MCZ 41695
2(121-141); USNM 135842 1 (X-ray); MCZ 28058 1
(X-ray) (lectotype S. dispar Garman).

PACIFIC OCEAN LARVAE^.— Larvae taken
at 17 EASTROPAC stations and 2 CalCOFI sta-
tions as follows: EASTROPAC stations 11.282
1(4.8); 13.105 1(5.5); 13.172 2(6.4, 6.8); 20.018
1(5.5); 30.114 2(4.0, 4.5); 45.032 1(8.1); 45.073
1(6.0); 45.078 1(10.3); 45.293 1(6.6); 45.316 1(6.9)
46.034 1(6.2); 46.096 2(6.7, 6.9); 47.001 1(5.2)
47.005 4(3.5-4.3); 47.035 1(7.0); 47.040 1(5.3)
47.065 1(9.2); CalCOFI 7205-20.127 1(5.0); 4907-
112 1(9.1).

ATLANTIC OCEAN.— MCZ 41638 l(X-ray)

^Station data in EASTROPAC Information Paper 6 and Ahl-
strom (1972).

Table 3. — Means and differences among means of meristic counts of Scopelengys tristis from four areas (eastern North
Pacific, ENP; eastern tropical Pacific, ETP; Indian Ocean, 10; and Atlantic Ocean, AO) and S. clarkei.

S. tristis

S. c/a

rkei

Difference in

counts between

S. clarkei and S. tristis

Average Least

Overall

Mean

Greatest

differences

among

regions

character

Range

mean

ENP

ETP

Range

Mean

difference

Dorsal rays

11-13

11.47

11.4

11.3

11.1

12.0

0.9

13.0

1.5

1 .0-AO

Anal rays

12-14

12.98

12.8

12.9

13.4

0.6

14.0

1.0

0.6-AO

Pectoral rays

14-17

15.42

15.6

15.9

15.2

15.0

0.9

12-13

12.9

2.5

2.1-AO

Vertebrae

29-32

30.84

30.9

31.0

30.4

31.1

0.7

34-35

34.1

3.3

30-AO

Gill raker

7-11

8.51

7.9

9.1

9.2

1.3

7-10

8.2

0.3

0.3-EP

147

FISHERY BULLETIN: VOL, 74, NO. 1

(type S. whoi Mead); USNM 20678, 5(152-164),
eastern tropical Atlantic, lat. 07°32'N, long.
20°54'W, 1813-2125, 12 April 1971, 1,300 m,
1,600-mesh Engels trawl, RV Walther Herwig;
ISH 623/68, 7(73-162), eastern tropical Atlantic,
lat. 12°07'N, long. 23°08'W, 30 January 1968,
2,000 m, 1,600-mesh Engels trawl, RV Walther
Herwig; ISH 2095/71, 1(167), eastern tropical At-
lantic, lat. 05°30'S, long. 16°28'W, 9 April 1971,
1,950 m, 1,600-mesh Engels trawl, RV Walther
Herwig; ISH 2447/71, 12(86-160), eastern tropical
Atlantic, lat. 04°38'N, long. 19°21'W, 13 April
1971, 756 m, 1,600-mesh Engels trawl, RV
Walther Herwig; ISH 3099/71, 5(132-160), eastern
tropical Atlantic, lat. 07°32'N, long. 20°54'W, 14
April 1971, 1,300 m, 1,600-mesh Engels trawl, RV
Walther Herwig; ISH 2942/71, 2(134-155), eastern
tropical Atlantic, lat. 23°47'N, long. 20°59"W, 19
April 1971, 2,100 m, 1,600-mesh Engels trawl, RV
Walther Herwig.

INDIAN OCEANl— IIOE 7001 Anton Bruun
III, 16 (25-94); IIOE 7004 Anton Bruun III, 7
(32-120); IIOE 7012 Anton Bruun III, 2 (23-25);
IIOE 7022 Anton Bruun III, 1 (113); IIOE 7027
Anton Bruun III, 1(138); IIOE 7037 Anton Bruun
III, 2 (40-87); IIOE 7046 Anton Bruun III, 3 (66-
179); IIOE 7143 Anton Bruun VI, 1 (131); IIOE
7147 Anton Bruun VI, 28 (28-142); IIOE 7153
Anton Bruun VI, 4 (42-161); IIOE 7154 Anton
Bruun VI, 12 (48-114); IIOE 7163 Anton Bruun VI,
12 (28-152); IIOE 7165 Anton Bruun VI, 3 (22-27);
IIOE 7206 Anton Bruun VI, 1 (27); IIOE 7277
Anton Bruun VI, 2 (40-87).

Scopelengys clarkei n.sp.

Holotype

SIO 73-160, female (176 mm), central Pacific,
lat. 29°56.0'N, long. 144°56.6'W, 0224-0556 h; 14
February 1973, 10-foot IKMT, 0-1,000 m, RV
Alexander Agassiz.

Paratypes

USNM 210707, male (160 mm), central Pacific,
lat. 21°20-30'N, long. 158°20-30'W, 1204-1637 h;
15 September 1970, 10-foot IKMT, 0-1,000 m, RV
El Pescadero I; USNM 210706, male (156 mm).

"Station data in Nafpaktitis and Nafpaktitis (1969).

148

central Pacific, lat. 24°N, long. 139°W, 0049-0149
h; 29 November 1972, 50-foot Universal trawl,
0-494 m, RV David Starr Jordan; FMNH 76366,
female (154 mm), central Pacific, lat. 22°N, long.
158°W, 1240-1645 h; 13 November 1969, 10-foot
IKMT, 0-800 m.

Other Materials Studied

SIO 51-76, female (109 mm), southeast of
Guadalupe Island, 17 March 1951, 10-foot IKMT,
0-549 m; FMNH 76367, juvenile (65 mm), central
Pacific, lat. 21°20-30'N, long. 158°20-30'W, 0421-
0600 h; 27 February 1971, % Cobb trawl, 0-150 m,
RV Townsend Cromwell; FMNH 76368, juvenile
(42 mm), central Pacific, lat. 21"20-30'N, long.
158°20-30'W, 2236-0105 h; 16-17 November 1969,
10-foot IKMT, 0-250 m, RV Teritu; T. Clarke,
71-3-9, larva (15 mm), central Pacific, lat. 21°20-
30 'N, long. 158°20-30'W, 1252-1645 h; 2 March
1971, 10-foot IKMT, 800-900 m, Y(N El Pescadero I,
retained at the Southwest Fisheries Center.

Adult Morphology

Body proportions of the holotype are given first,
followed, in parentheses, by range of values for
holotype and three paratypes. Body slender;
greatest body depth at origin of dorsal fin, 19.0
(18.4-19.0), tapering to a moderately deep caudal
peduncle (Figures IB, 2B), less than three in
length of head, 9.4 (9.4-10.2). Head slightly con-
cave in dorsal profile, head length 25.4 (24.5-
26.4); head depth 17.6 (16.7-17.9); eye small, orbit
3.0 (2.9-3.6); interorbital width 8.7 (7.6-8.7);
snout about one-third of head length, 8.3 (7.7-
8.8); length of maxillary 11.3 (11.3-12.6), greatest
width of maxillary 3.1 (2.8-3.6). Snout to: dorsal
fin origin 43.5 (39.0-43.5); anal fin origin 68.6
(65.1-69.6); pelvic fin origin 40.2 (40.2-43.4).
Length of dorsal fin base 17.3 (17.0-19.4); length of
anal fin base 16.0 (16.0-17.9). Color dark brown,
preserved in alcohol.

Meristic Data

Counts are based on all seven specimens. D 13
(7); A 14 (6), ? (1); P 13/13 (6), 13/12 (1); V 8/8 (7);
principal C 10 + 9 (7), procurrent C 7-8/6-9; bran-
chiostegal rays 8/8 (7); vertebrae 15 + 19 = 34 (6),
15 + 20 = 35 (1); gill rakers 1-2 + (6-8) = 7-10
(mean 8.2).

BUTLER and AHLSTROM: NEW SPECIES, SCOPELENGYS CLARKEl

Larvae

A single specimen was available, 15.4 mm SL
(Figure 3B). Body shape similar to that of adults
but with a relatively larger head — length 35.7 and
depth 25.0; eye 5.5; snout 12.8; body depth 23.5;
least depth of caudal peduncle, 14.6. Fin origins
farther back on body than in adults. Snout to:
dorsal fin origin 50.0; anal fin origin 72.8; pelvic
base 53.6. Pigment confined to head and nape,
extensively developed on the operculum and
lower jaw; a small pigment patch on upper jaw
behind eye; several melanophores on mid-brain;
body pigment confined to nape and to a patch an-
terior to pectoral base.

Name

This species from the central North Pacific is
named in honor of Dr. Thomas A. Clarke of the
Hawaii Institute of Marine Biology.

COMPARISON OF SCOPELENGYS

CLARKEl AND SCOPELENGYS

TRISTIS

Scopelengys clarkei differs from S. tristis in
meristic counts, in some morphometric charac-
ters, and in larval pigmentation.

For differences in meristic characters, refer to
Table 3. Most marked differences are in average
number of vertebrae — 34.1 (S. clarkei) vs. 30.8
= 3.3; average pectoral-fin ray count — 12.9 {S.
clarkei) vs. 15.4 =2.5; and average dorsal-fin ray
count — 13.0 vs. 11.5 =1.5. As regards morphomet-
ric characters, S. clarkei has a deeper caudal
peduncle, a narrower maxillary, and a more
fusiform body. Several distinctive adult charac-
ters also can be recognized in larger larvae of the
two species, i.e., differences in meristic characters
and depth of caudal peduncle. The most striking
differences between larvae of the two species are
found in the head pigment which is restricted to
an eye-bar in S. tristis, as compared with the scat-
tered pigment on the operculum, lower jaw, etc. of
S. clarkei.

The two species are similar in general body
shape, head size, eye size, length of snout, and
position of fins on the body. Scopelengys clarkei
has its greatest body depth at the dorsal origin,
whereas S. tristis has its greatest body depth at
the nape.

When an analysis of covariance was performed
on the morphometric characters of 7 S. clarkei
and 32 S. tristis from the eastern North Pacific,
eastern tropical Pacific, Indian, and Atlantic
oceans, only the least depth of caudal peduncle

E
6

o
liJ

<
O

20p

18-
16-

14
12
10

Scopelengys clorkei

Scopelengys tristis

INDIAN OCEAN

EASTERN TROR PAG.
EASTERN NORTH PAG
EASTERN TROP. ATL.

20 30

50 60

70 80 90 100 no 120
STANDARD LENGTH (mm)

140 150

160

Figure 5.— Regression of least depth of caudal peduncle on standard length of Scopelengys tristis and S. clarkei.

149

FISHERY BULLETIN: VOL. 74, NO. 1

showed a significant difference at the 1% level,
F = 3.72, between the two species.

The Atlantic specimens of S. tristis had counts
for four characters that were closer to those of S.
clarkei than were counts of these characters from
other geographic areas. These differences in
counts between Atlantic S. tristis and S. clarkei
were as follows: dorsal fin rays 1.0 (12.0 vs. 13.0),
anal fin rays 0.6 (13.4 vs. 14.0), pectoral fin rays
2.1 (15.0 vs. 12.9), and vertebrae 3.1 (31.1 vs. 34.2).
Differences of two in pectoral-fin ray counts and
three for vertebrae are much greater than the
regional variability found among specimens of
S. tristis.

ACKNOWLEDGMENTS

We are grateful to the following individuals
and institutions for the loan of specimens: T.
Clarke, Hawaiian Institute of Marine Biology,
Kaneohe, Hawaii; G. Kreftt, Institut fiir
Seefischerei, Hamburg, Germany; R. H. Rosen-
blatt, M. A. Barnett, J. Copp, and D. Dockins,
Scripps Institution of Oceanography, La Jolla,
Calif; B. G. Nafpaktitis, University of South-
ern California, Los Angeles, Calif; R. K. Johnson,
Field Museum of Natural History, Chicago,
111.; W. Nellen, Institut fiir Meereskunde,
Kiel, Germany; R. Schoknecht and M. M. Dick,
Museum of Comparative Zoology, Cambridge,
Mass.; R. H. Gibbs and W. R. Taylor, Division
of Fishes, National Museum of Natural His-
tory, Washington, D.C. We give special thanks to
H. G. Moser, Southwest Fisheries Center, Na-
tional Marine Fisheries Service, La Jolla, and to
C. L. Hubbs, Scripps Institution of Oceanography,
for their advice and criticisms and to R. Schok-
necht for examining the types of S. whoi Mead
and S. dispar Garman.

LITERATURE CITED

Ahlstrom. E. H.

1972. Kinds and abundance of fish larvae in the eastern
tropical Pacific on the second multivessel EASTROPAC
survey, and observations on the annual cycle of larval
abundance. Fish. Bull, U.S. 70:U53-1242.

Alcock, a.

1890. On the bathybial fishes of the Arabian Sea, obtained

during the season 1889-90. Ann. Mag. Nat. Hist., Ser. 6,

6:295-311.
BOLIN, R. L.

1939. A review of the myctophid fishes of the Pacific coast

of the United States and of lower California. Stanford

Ichthyol. Bull. 1:89-156.
CLARKE, T. A.

1973. Some aspects of the ecology of lanternfishes (Myc-

tophidae) in the Pacific Ocean near Hawaii. Fish. Bull.,

U.S. 71:401-434.

Garman, S.

1899. Reports on an exploration off the west coasts of
Mexico, Central and South America, and off the Gala-
pagos Islands, in charge of Alexander Agassiz, by the
U.S. Fish Commission steamer "Albatross," during 1891,
Lieut. Commander Z. L. Tanner, U. S. N., commanding.
XXVI. The fishes. Mem. Mus. Comp. Zool., Harvard
Coll. 24:1-431.

HUBBS, C. L., AND K. F. LAGLER.

1958. Fishes of the Great Lakes region. Revised ed. Cran-
brook Inst. Sci. Bull. 26, 213 p.

Mead, G. W.

1963. Observations on fishes caught over the anoxic waters
of the Cariaco Trench, Venezuela. Deep-Sea Res.
10:251-257.
Nafpaktitis, B. G., and M. Nafpaktitis

1969. Lanternfishes (family Myctophidae) collected during
cruises 3 and 6 of the RA^ Anton Bruun in the Indian
Ocean. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 5, 79 p.

Nellen, W.

1973. Fischlarven des Indischen Ozeans. "Meteor"
Forsh-Ergebnisse, Ser. D, 14:1-66.

Norman, J. R.

1939. Fishes. John Murray Exped. 1933-1934, Sci. Rep.
2(1):1-116.
OKIYAMA, M.

1974. The larval taxonomy of the primitive Myctophiform
fishes. In J. H. S. Blaxter (editor). The early life history
of fish, p. 609-621. Proc. Int. Symp. Dunstaffnage Mar.
Res. Lab., Scott. Mar. Biol. Assoc, Oban, Scotl. May
17-23, 1973. Springer- Verlag, N.Y.

Parr, a. e.

1928. Deepsea fishes of the order Iniomi from the waters
around the Bahama and Bermuda islands. With anno-
tated keys to the Sudidae, Myctophidae, Scopelarchidae,
Evermannellidae, Omosudidae, Cetomimidae and Ron-
deletidae of the world. Bull. Bingham Oceanogr. Col-
lect., Yale Univ. 3(3):1-193.

Rosen, D. E., and C. Patterson.

1969. The structure and relationships of the Paracanthop-
terygian fishes. Bull. Am. Mus. Nat. Hist. 141(3 1:357-474.

Rothschild, B. j.

1963. Graphic comparisons of meristic data. Copeia
1963:601-603.

150

WEIGHT LOSS, MORTALITY, FEEDING, AND DURATION OF

RESIDENCE OF ADULT AMERICAN SHAD,

ALOSA SAPIDISSIMA, IN FRESH WATERS

Mark E. Chittenden, Jr.^

ABSTRACT

Linear regression equations are given for each sex for the regressions of total weight, somatic weight,
and gonad weight on length prior to spawning, and for total weight on length after prolonged stay in
fresh water

Most shad began to return seaward by late June and probably had spent a maximum of about 2 mo
in fresh water Many fish, however, remained near the spawning grounds well into summer; and
many died near the spawning grounds, probably from starvation. Opportunistic feeding oc-
curred on "planktonic" items, but adult shad do not regularly obtain energy sufficient to maintain
their weight in fresh water. Weight loss was related to sex and increased with increasing size. Mean
length males and females averaged 45 and ^T7c total weight loss, respectively. Daily somatic
weight loss was at least 5.75 g for males of average size and 12.47 g for females.

The anadromous American shad, Alosa sapidls-
sima, an important commercial and sport fish,
ranges widely on the Atlantic and Pacific coasts of
North America. There is much literature on this
fish, but little of it pertains to adults in fresh wa-
ter, except for aspects of their spawning and popu-
lation dynamics. In the course of other studies on
the Delaware River from 1960 to 1968, I made
many opportunistic observations on weight loss,
mortality, feeding behavior, and duration of resi-
dence of adult shad on their spawning grounds in
fi'esh water. This paper summarizes those obser-
vations and presents data on total-fork length
conversion, regressions of total weight, somatic
weight and gonad weight on length prior to
spawning, and regressions of total weight on
length after spawning.

MATERIALS AND METHODS

Adult shad were collected during their spawn-
ing runs at Lambertville, N.J., 22.5 km above
tidal water (but far downstream of the present-
day spawning grounds) using a 76-mm stretch-
mesh, 107 m long and 3.6 m deep haul seine that
was paid out from a boat and landed about 400 m
downstream. Sampling occurred at 3- or 4-day in-
tervals from 5 April to 19 May 1963, from 20

'Based on part of a dissertation submitted in partial fulfill-
ment of the requirements for a Ph.D. degree, Rutgers Univer-
sity, New Brunswick, N.J.

^Department of Wildlife and Fisheries Sciences, Texas A&M
University, College Station, TX 77840.

March to 18 May 1964, from 26 March to 7 May
1965, and from 27 March to 19 May 1966. Data for
the period 1959-62 were obtained from rotenone
surveys (hereinafter referred to as the Tri-State
Surveys) during July and August by the States of
New Jersey, New York, and Pennsylvania in
cooperation with the U.S. Fish and Wildlife
Service.

I examined grossly the stomach contents of
many adults captured during the Tri-State Sur-
veys in mid-July 1961, most of the 526 fish col-
lected at Lambertville and many fish captured on
the spawning grounds after 1962.

Length and total weight were determined on
most fish in 1961 and 1962 and on all fish thereaf-
ter. Gonad weight was measured after 1962.
Length, always taken in inches, was measured as
fork length during 1961 and 1962 and as total
length thereafter. To develop conversion factors,
both measurements were taken on 490 adults col-
lected at Lambertville during 1963 and 1964 and
on 100 young captured in summer 1966. Total
weight was measured in pounds (to the closest 0.1
lb) during 1961-63 but in grams thereafter Gonad
weights were always taken in grams (to the
closest 0.1 g). All measurements were converted
to grams, millimeters, and fork lengths for pre-
sentation herein.

Regression analyses and related statistics were
calculated using computer program BMD-03R
(Dixon 1967). All regressions presented herein
were significant at a = 0.005. The coefficient of

Manuscript accepted July 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

151

FISHERY BULLETIN: VOL. 74, NO. 1

determination (Steel and Torrie 1960) was used to
estimate the amount of variation in y associated
with variation in x. Residuals were used to
examine the data for differences due to categories
of classification such as year of collection. Size
ranges are given within which regressions were
linear.

When first referred to, locations are followed in
parentheses by their approximate distances in
kilometers upstream from Marcus Hook, Pa.,
which is situated about 90 km downstream from
the fall line at Trenton, N.J. and near the transi-
tion between brackish and fresh water.

RESULTS AND DISCUSSION

Total and Fork Length Conversion

The relationship between total length (TL) and
fork length (FL) for 590 young and adult fish was
linear, and 99.96*7^ of the variation in one mea-
surement was explained by variation in the other.
Regression equations were FL = 1.28 + 0.88 TL
and TL = 1.00 + 1.13 FL. Extreme deviations
from regression were about ±7.6 mm for adults
and less for young. The slope of the regression of
fork length on total length coincides with La
Pointe's (1958) factor of 0.894 to convert total
length to fork length.

Total Weight- Length Relationships
Prior to Spawning

The relationships between total weight (TW)
and length determined for fish captured at Lam-
bertville were TW = 1,106.77 + 8.09 (FL - 427.98)
for 268 males and TW = 1,737.26 + 11.54 (FL -
476.71) for 244 females. About 81^f (males) and
78% (females) of the variation in total weight
was associated with variation in length. Valid
ranges for linear interpolation were about 330-
520 mm for males and 410-550 mm for females.

The observed arithmetic mean weights with
95% confidence limits were 1,107 ± 36 g for males
and 1,737 ± 45 g for females. The smallest males
were 272 and 680 g and the smallest female was
1,089 g. The heaviest male and female fish were
1,905 and 2,585 g, respectively.

Somatic Weight-Length Relationships
Prior to Spawning

The relationships between log somatic weight
152

(SW) and length determined for 85 males and 130
females captured at Lambertville in 1964 and
1965 were logio SW - 3.0047 + 0.0036 (FL -
428.20) for males and logio SW = 3.1807 + 0.0029
(FL - 480.73) for females. About 91% (males) and
81% (females) of the variation in log somatic
weight was associated with variation in length.
Valid ranges for linear interpolation were about
360-500 mm for males and 410-540 mm for
females. Mean somatic weights with 95%
confidence limits were 1,011 ± 56 g for males and
1,516 ± 38 g for females.

Gonad Weight- Length Relationships
Prior to Spawning

The relationships between log total gonad
weight (TGW) and length determined for 267
males and 244 females captured at Lambertville
were logio TGW = 1.8633 + 0.0033 (FL - 428.43)
for males and logio TGW = 2.3892 + 0.0024 (FL
- 476.93) for females. Valid ranges for linear in-
terpolation were about 330-520 mm for males and
410-550 mm for females. About 45% (males) and
26% (females) of the variation in log total gonad
weight was associated with length variation.
Much variation in gonad weight, especially for
females, is not explained by the regression equa-
tions. Much gonad development occurs during the
spawning run (Chittenden 1969), and residual
plots suggested that gonad weights were heavier
in 1963 than in 1964. These factors account for
some unexplained variation in gonad weight.

Mean total gonad weights with 95% confidence
limits were 73 ± 7 g for males and 245 ± 22 g for
females.

Duration of the Freshwater Residence

Most fish begin to return seaward by about late
June. I observed hundreds of adults near Han-
cock, N.Y. (403) until 17 June 1964, but very few
were present on 14 July. Most fish had died or
migrated seawards during the interim period. De-
laware River shad runs begin in early April at
Lambertville and the peak occurs about 1 May,
depending upon the degree of pollution near
Philadelphia (34) (Chittenden 1969). This
suggests most fish probably spend a maximum of
2 mo in fi^esh water before returning seaward, in
agreement wdth Bean's (1892, 1903) observations.

Many fish remain near the spawning grounds
well into summer. The Tri-State Surveys cap-

CHITTENDEN: ADULT AMERICAN SHAD IN FRESH WATER

tured many adults during midsummer between
Skinners Falls, N.Y. (348) and Minisink Island,
N.J. (266): 538 fish were captured at three sta-
tions in mid-July 1961; 237 fish were captured at
two stations in mid-July 1962; 30 adults were
captured near Milford, Pa. (269) on 7 August 1959,
and 13 were captured there on 1 August 1961.

Upstream Mortality

There was a large mortality of shad upstream
near the spawning grounds about the end of the
spawning period. In 1963, I observed many dead
fish along the banks or in shallow water on 5
July; and a surface gill net set overnight at Mil-
ford, Pa. on 22 June captured 15 fish that ap-
peared to have been dead for several days. In
1964, dead shad first appeared in the East Branch
near Hancock about 14 June; on that date, I
walked the bank for about 0.8 km and observed
26 dead fish within 10 m of the shoreline. I ob-
served hundreds of dead shad on 8 July 1964 dur-
ing a 19-km float from Matamoras, Pa. (274) to
Dingmans Ferry, Pa. (258). I frequently saw dead
fish in shallow water during August.

Shad may die before being completely spent.
Some dead fish examined near Hancock had
ovaries about a fourth the size of those in fish
captured at Lambertville. The ovaries of these
dead fish contained many translucent eggs, a
criterion (Milner 1874; Brice 1898; Leach 1925)
indicating that the fish is ripe.

shad, 2) 6 darters and 17 shad, 3) 46 shad, and 4) 15
shad. Young shad were the first fish to react to
rotenone, and the adults probably foraged on dis-
tressed and dying young.

Weight Loss in Fresh Water

Much weight was lost while the adult shad
were in fresh water Fish captured near Hancock
had noticeably lost weight by late May, and they
became more emaciated the longer they remained
in fresh water. Tri-State Survey data obtained
10-13 July 1961 from Belvidere, N.J. (197) to Han-
cock, N.Y. and 16-17 July 1962 at Minisink Island
and Skinners Falls were used to estimate the
changed weight-length relationship for each sex.
The relationships between total weight and
length of these fish were TW = 536.34 + 3.24(FL -
407.34) for 296 males and TW = 661.29 + 3.01(FL
- 451.18) for 19 females. Valid ranges for linear
interpolation were about 265-450 mm for males
and 340-475 mm for females. About 66% (males)
and 63% (females) of the variation in total weight
was associated with variation in length. These
regressions explain less variation in total weight
than the 80% explained for fish taken at Lam-
bertville.

The average percentages of total weight loss in
fresh water were estimated by comparing Lam-
bertville and Tri-State Survey regression means
at different lengths for each sex (Figure 1). The

Feeding Behavior in Fresh Water

Feeding did occur in freshwater, at least near
the upstream spawning grounds. The stomachs of
most shad captured at Lambertville were empty,
but a few contained a slight amount of amorphous
material. Stomachs of fish collected upstream
from Port Jervis, N.Y. (295) in late May and June
frequently contained a few insects. I observed a
large mayfly hatch in late May 1964 near Han-
cock: hundreds of adult shad were rising to the
surface, apparently to feed, and the stomachs of
many fish (about 50) captured by angling were
packed with mayflies. Similar surface feeding be-
havior was observed on several other occasions,
although fish were not collected to confirm feed-
ing. Many adults captured during the Tri-State
Surveys contained recently eaten young shad and
shield darters, Percina peltata.For example, four
stomachs contained: 1) 2 darters and 9 young

en
(/)
O

300

400 500

FORK LENGTH (mm)

600

FIGURE 1. — Minimum average total weight loss of American
shad in fresh water.

153

FISHERY BULLETIN: VOL. 74, NO, 1

average percent weight loss depended upon
length. Large fish lost a greater percentage than
small fish. Average total weight loss was from 30
to 50% for 359-493 mm FL males and from 48 to
62% for 421-531 mm FL females, sizes which
closely approximate the observed size range of
fish in the 1963 and 1964 runs (Chittenden 1969).
The observed mean fork lengths of fish captured
at Lambertville were 428 mm for males and 477
mm for females, based upon the regression equa-
tions, and these sizes averaged 45 and 57% total
weight loss, respectively.

Somatic weight loss, a better measure of the
toll taken by the spawning migration, was esti-
mated by subtracting the predicted total gonad
weight from the predicted total weight at Lam-
bertville before making a comparison with the
Tri-State Survey total weight regressions. No
correction was made for the gonads of fish cap-
tured during the Tri-State Surveys; however,
these were a negligible fraction of the total
weight. The total testes weights of 15 males col-
lected near Hancock on 14 July 1964 and on 21, 24
June and 1 July 1965 ranged from 3.7 to 27 g and
averaged 15.9 g while the total ovary weights of 3
females collected then varied from 18.2 to 35 g
and averaged 27.1 g. The average percentage of
somatic weight loss in males was 24% at 359 mm,
46% at 493 mm, and 42% for the mean-sized male
of 428 mm. For females, somatic weight loss was
38% at 421 mm, 56% at 531 mm, and 50% for the
mean-sized female of 477 mm.

Absolute daily weight loss was estimated from
the duration of the freshwater residency. Fish
captured during the Tri-State Surveys had prob-
ably been upstream about 75 days. This approxi-
mates their maximum stay in fresh water because
the peak of the run at Lambertville is about 1
May (Chittenden 1969), and most fish move sea-
ward from the Hancock area by late June. There-
fore, the average daily loss in somatic weight of
males was 1.63 g at 359 mm, 9.37 g at 493 mm,
and 5.75 g for mean-sized males of 428 mm. For
females the average daily loss in somatic weight
was 5.75 g at 421 mm, 18.87 g at 531 mm, and
12.47 g for mean-sized females of 477 mm.

Daily weight loss can be used to suggest how
long fish of different sizes can remain in freshwa-
ter before death. The amount of weight loss which
results in death of shad is not known, but death
occurs in many animals when weight loss exceeds
40% (Curtis 1949). Assume 50% for simplicity in
calculation, this may not be quite correct, but it

may be conservative and the size pattern, at
least, remains the same if the percentage is a con-
stant. From this, males could remain 154 days at
359 mm, 81 days at 493 mm, and the average
sized male (428 mm) could remain 90 days.
Females could remain 100 days at 421 mm but
only 68 days at 531 mm, and the mean-sized
female of 477 mm could remain 75 days. There is
apparently little difference in the amount of time
an average to maximum-sized fish can spend in
fresh water before death, but small fish can sur-
vive much longer.

GENERAL DISCUSSION

Weight loss data presented herein agrees
reasonably with those of Leggett (1972) who
noted that his figures were probably underesti-
mates. The present figures ignore weight loss in
the 100-km migration between Marcus Hook and
Lambertville and may be based on a longer than
average stay in fresh water. Both factors tend to
underestimate weight loss which affects related
estimates.

Many shad apparently remain upstream near
the spawning grounds well into the summer.
However, the percentage they comprise of the run
is unknown. A few fish remain far upstream until
late fall. Bishop (1936) captured emaciated indi-
viduals 305-330 mm long near Hancock in
November These fish must have migrated up-
stream during the previous spring, because low
dissolved oxygen water near Philadelphia pre-
sents a virtually impassable barrier through
summer and fall (Ellis et al. 1947; Sykes and
Lehman 1957; Chittenden 1969). Nichols (1959)
captured an emaciated male during October in
the Connecticut River and estimated it had been
in freshwater at least 120 days. I captured an
emaciated male (287 mm FL, 194 g) in fresh water
in the James River, Va. on 7 October 1969.

The finding of little or no food in adults col-
lected at Lambertville is similar to the reports of
Bean (1903), Leim (1924), Leach (1925), Hilde-
brand and Schroeder (1928), Moss (1946), and
Hildebrand (1963) that adults take little or no
food while ascending rivers. My observations of
instances of intensive feeding while upstream are
exceptional, although Atkinson (1951) reported
an artificial instance of feeding in freshwater
ponds. Adult shad at sea feed largely on
planktonic forms such as copepods and mysids
(Leim 1924; Hildebrand and Schroeder 1928;

154

CHITTENDEN: ADULT AMERICAN SHAD IN FRESH WATER

Bigelow and Schroeder 1953; Hildebrand 1963;
Leim and Scott 1966), although Holland and Yel-
verton (1973) reported that they occasionally take
large amounts offish. Atkinson (1951) attributed
the general absence of food in the stomachs of
adults to their planktonic feeding habit and the
absence of suitably large plankton in fresh water.
My observations suggest that adult shad would
opportunistically feed in freshwater if suitably
large "planktonic" forms were readily available.

Although adults feed opportunistically in
fresh water, they do not regularly obtain energy
sufficient to maintain their weight and must use
energy reserves accumulated during their life at
sea to support migration in fresh water, final de-
velopment of the gonads, and spawning. Adults
use up their somatic substance at a size and sex
dependent rate of at least about 1.6-18.9 g/day.
Their physical activity deteriorates greatly as
Fowler (1908) and Walburg (1960) noted. Death
by starvation may occur when weight loss ex-
ceeds 40% (Curtis 1949), and this is probably the
main cause of the mortality I observed on the
spawning grounds. Further work is needed to
quantitatively describe upstream mortality, but
its magnitude would appear large as Bean (1892,
1903) and Anonymous (1902) also observed in the
Delaware River and Walburg (1960) observed in
the St. Johns River, Fla.

Weight loss was related to sex and size in
agreement with Leggett (1972). The apparent re-
lationship between weight loss and sex, however,
may not be direct. Metabolic rate, in general, in-
creases with temperature within limits. Leggett
(1972) noted that females tend to migrate later
and at a higher temperature than males and
suggested that temperature was responsible for
the apparent sex difference in weight loss. The
relationship between size and total metabolism in
a wide variety of organisms can be expressed as:

log M = \oga + b \ogW

where M is total metabolism and W is weight
(Paloheimo and Dickie 1966; Prosser 1973). The
relationship between metabolic rate and size can
be expressed (Prosser 1973) as:

log M/W = log a + (6 - 1) log W.

From the latter expression it follows that a b
value less than 1.0 implies that the metabolic rate
decreases with increasing size, while a b value

greater than 1.0 indicates that the metabolic rate
increases with size. The value generally found for
b is about 0.8 (Paloheimo and Dickie 1966; Pros-
ser 1973), although Fry (1971) cautions that this
value should not yet be accepted as dogma. Pres-
ent findings on the relationship between size and
weight loss in shad on their spawning migration
are consistent with a b value greater than 1.0.
Calculations made herein obviously assume that
adult fish of all sizes are in fresh water the same
length of time. If 6 is not greater than 1.0, we
must conclude that: 1) small adults enter fresh
water later than large fish and thus are in fresh
water for a shorter period of time, or 2) small fish
make better use of available freshwater food
resources.

Estimates of the time that adults can remain in
fresh water suggest that only small fish can sur-
vive upstream into the fall. The small fish I cap-
tured in the James River in October apparently
had lost only about 33-39% of its weight in com-
parison with the Delaware River somatic weight
regression at Lambertville and an unusually
small fish (285 mm FL, 288 g) captured at Lam-
bertville. It is noteworthy that, except for
Nichols' (1959) report of a 430 mm FL male, the
adult shad reported in fresh water during the fall
have all been males about 305 mm long. Fish this
small, however, are rare in the age compositions
reported from many Atlantic Coast rivers (Talbot
1954; Fredin 1954; Walburg 1956, 1957, 1960,
1961; Sykes 1956; Sykes and Lehman 1957; Wal-
burg and Sykes 1957; La Pointe 1958; Nichols and
Tagatz 1960; Nichols and Massmann 1963; God-
win 1968; Leggett 1969; Chittenden 1975).

ACKNOWLEDGMENTS

For assisting in field collections, I am deeply
grateful to J. Westman, J. Hoff, J. Harakal, D.
Riemer, J. Barker, F. Bolton, R. Coluntuno, K.
Compton, R. Gross, C. Masser, R. Stewart, J.
Miletich, S. Hoyt, L. Schulman, H. Dinje, H.
Buckley, J. Musick, M. Bender, J. Gift, C.
Townsend, R. Bogaczk, and K. Marcellus of or
formerly of Rutgers University, Harvard Univer-
sity, New Jersey Division of Fish and Game, and
New York Department of Environmental Con-
servation.

Fred and William Lewis, Jr. generously gave
permission to collect shad at their fishery at
Lambertville and frequently assisted in seining.
J. D. McEachran and W H. Neill of Texas A&M

155

FISHERY BULLETIN: VOL. 74, NO. 1

University reviewed the manuscript. The U.S.
Bureau of Sport Fisheries and Wildhfe, New Jer-
sey Division of Fish and Game, Pennsylvania
Fish Commission, and New York Department of
Environmental Conservation permitted use of
data collected during the Tri-State Surveys of the
Delaware River. Financial support was provided,
in part, by Rutgers University, The Sport Fishing
Institute, Delaware River Basin Commission,
and U.S. Public Health Service. One collection
was made while the author was employed at the
Virginia Institute of Marine Science.

LITERATURE CITED

ANONYMOUS.

1902. Rep. Pa. State Comm. Fish. 1902.
ATKINSON. C. E.

1951. Feeding habits of adult shad [Alosa sapidissima) in
fresh water. Ecology 32:556-557.

Bean, T. H.

1892. The fishes of Pennsylvania. Rep. Pa. Comm. Fish.
1889-90-91, p. 1-149

1903. Catalogue of the fishes of New York. N.Y. State
Mus. Bull. 60, 784 p.

BIGELOW, H. B., AND W. C. SCHROEDER.

1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv.,
Fish. Bull. 53, 577 p.

BISHOP, S.C.

1936. Fisheries investigations in the Delaware and Sus-
quehanna Rivers. In A biological survey of the Delaware
and Susquehanna watersheds, p. 122-139. N.Y. Conserv.
Dep. Suppl. 25th Annu. Rep., 1935.
BRICE, J. J.

1898. A manual of fish-culture, based on the methods of
the United States Commission of Fish and Fisher-
ies. U.S. Comm. Fish Fish. Part 23, Rep. Comm. 1897,
append., p. 1-261.
Chittenden, M. E., Jr.

1969. Life history and ecology of the American shad, Alosa
sapidissima, in the Delaware River. Ph.D. Thesis, Rut-
gers Univ., New Brunswick, N.J., 458 p.

1975. Dynamics of American shad, Alosa sapidissima,
runs in the Delaware River. Fish. Bull., U.S. 73:487-494.

Curtis, B.

1949. The life story of the fish. Scientific Book Club,
Lond., 276 p.
DiXON, W. J. (editor).

1967. Biomedical computer programs. 2nd ed. Univ. Calif.
Press, Berkeley, 600 p.
ELLIS, M. M., B. A. WESTFALL, D. K. MEYER, AND W. S.
PLATNER.

1947. Water quality studies of the Delaware River with
reference to shad migration. U.S. Fish Wildl. Serv., Spec.
Sci. Rep. 38, 19 p.
FOWLER, H. W

1908. The king crab fisheries in Delaware Bay, and further
note on New Jersey fishes, amphibians and rep-
tiles. Annu. Rep. N.J. State Mus. 1907 Part 3, p. 349-432.
FREDIN, R. A.

1954. Causes of fluctuations in abundance of Connecticut

River shad. U.S. Fish Wildl. Serv., Fish. Bull. 54:247-
259.
FRY, F. E. J.

1971. The effect of environmental factors on the physiology
of fish. In W. S. Hoar and D. J. Randall (editors), Fish
Physiology, Vol. 6, p. 1-98. Academic Press, N.Y.

GODWIN, W. F.

1968. The shad fishery of the Altamaha River,
Georgia. Ga. Game Fish Comm., Mar. Fish. Div., Con-
trib. Ser., 8, 39 p.

HILDEBRAND, S. F.

1963. Family Clupeidae. In Fishes of the western North
Atlantic, Part 3, p. 257-454. Mem. Sears Found. Mar.
Res. Yale Univ. 1.
HILDEBRAND, S. F., AND W. C. SCHROEDER.

1928. Fishes of Chesapeake Bay. U.S. Bur Fish., Bull. 43,
Part 1, 366 p.
HOLLAND, B. F., JR., AND G. F. YELVERTON.

1973. Distribution and biological studies of anadromous
fishes offshore North Carolina. N.C. Dep. Nat. Econ. Re-
sour., Div. Commer. Sport Fish., Spec. Sci. Rep. 24, 132 p.
LaPOINTE, D. F.

1958. Age and growth of the American shad, from three
Atlantic Coast Rivers. Trans. Am. Fish See. 87:139-150.

LEACH. G. C.

1925. Artificial propagation of shad. Rep. U.S. Comm.
Fish. 1924, append. 8, p. 459-486. Doc. 981.
LEGGETT, W. C.

1969. A study of the reproductive potential of the Ameri-
can shad (Alosa sapidissima) in the Connecticut River,
and of the possible effects of natural or man induced
changes in the population structure of the species on
its reproductive success. Essex Mar. Lab., Essex, Conn.,
72 p.

1972. Weight loss in American shad (Alosa sapidissima,
Wilson) during the freshwater migration. Trans. Am.
Fish. Soc. 101:549-552.

leim, a. H.

1924. The life history of the shad (Alosa sapidissima (Wil-
son) ) with special reference to the factors limiting its
abundance. Contrib. Can. Biol., New Ser. 2:163-284.
LEIM, A. H., and W. B. Scott.

1966. Fishes of the Atlantic Coast of Canada. Fish. Res.
Board Can., Bull. 155, 485 p.
MILNER, J. W

1874. Report on the propagation of the shad (Alosa
sapidissima) and its introduction into new waters by the
U.S. Fish Commissioner in 1873. U.S. Comm. Fish Fish.,
Part 2, Rep. Comm. 1872-73, p. 419-451.
MOSS, D. D.

1946. Preliminary studies of the shad (Alosa sapidissima)
catch in the lower Connecticut River, 1944. Trans. North
Am. Wildl. Conf. 11:230-239.
NICHOLS, P. R.

1959. Extreme loss in body weight of an American shad
(Alosa sapidissima). Copeia 1959:343-344.

NICHOLS, P. R., AND W. H. MASSMANN.

1963. Abundance, age, and fecundity of shad, York River,
Va., 1953-59. U.S. Fish Wildl. Serv., Fish. Bull. 63:179-
187.
NICHOLS, P. R., AND M. E. TAGATZ.

1960. Creel census Connecticut River shad sport fishery,
1957-58, and estimate of catch, 1941-56. U.S. Fish Wildl.
Serv., Spec. Sci. Rep. Fish. 351, 12 p.

156

CHITTENDEN: ADULT AMERICAN SHAD IN FRESH WATER

PALOHEIMO, J. E., AND L. M. DICKIE.

1966. Food and growth of fishes. II. Effects of food and
temperature on the relation between metabolism and
body weight. J. Fish. Res. Board Can. 23:869-908.
PROSSER. C. L.

1973. Oxygen: respiration and metabolism. In C. L. Pros-
ser (editor). Comparative animal physiology, 3rd ed., p.
165-211. Saunders, Phila.

Steel, R. G. D., and J. H. Torrie.

I960. Principles and procedures of statistics with special
reference to the biological sciences. McGraw Hill, N.Y.,
481 p.
SYKES, J. E.

1965. Shad fishery of the Ogeechee River, Georgia, in
1954. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 191,
lip.
Sykes, J. E., AND B. A. Lehman.

1957. Past and present Delaware River shad fishery and
considerations for its future. U.S. Fish Wildl. Serv.,
Res. Rep. 46, 25 p.

TALBOT. G. B.

1954. Factors associated wdth fluctuations in abundance of
Hudson River Shad. U.S. Fish Wildl. Serv., Fish. Bull.
56:373-413.

WALBURG, C. H.

1956. Commercial and sport shad fisheries of the Edisto
River South Carolina, 1955. U.S. Fish Wildl. Serv., Spec.
Sci. Rep. Fish. 187, 9 p.

1957. Neuse River shad investigations, 1953. U.S. Fish
Wildl. Serv, Spec. Sci. Rep. Fish. 206, 13 p.

1960. Abundance and life history of shad, St. Johns River,
Florida. U.S. Fish Wildl. Serv., Fish. Bull. 60:487-501.

1961. Natural mortality of American shad. Trans. Am.
Fish. Soc. 90:228-230.

WALBURG, C. H., and J. E. SYKES.

1957. Shad fishery of Chesapeake Bay with special em-
phasis on the fishery of Virginia. U.S. Fish Wildl. Serv.,
Res. Rep. 48, 26 p.

157

DISTRIBUTION, ABUNDANCE, AND SIZE OF PENAEID SHRIMPS
IN THE ST. ANDREW BAY SYSTEM, FLORIDA

Harold A. Brusher and Larry H. Ogren^

ABSTRACT

Shrimp collections were made every 2 weeks at 12 stations in varying depths (1.5-12.2 m) of the St.
Andrew Bay system, Fla., from September 1972 through August 1973. The eight species of penaeid
shrimps caught in 312 trawl hauls were, in decreasing order of abundance: pink shrimp, Penaeus
duorarum; broken-neck shrimp, Trachypenaeus similis; rock shrimp, Sicyonia brevirostris; rock
shrimp, S. dorsalis; broken-neck shrimp, T. constrictus; brown shrimp, P. aztecus; white shrimp, P.
setiferus; and rock shrimp, S. typica. Of the total catch of penaeids, 57.7% were of the genus Penaeas,
22.6% oi Sicyonia , and 19.7% oi Trachypenaeus . Penaeids were more abundant in the sections of the
bay system close to the Gulf of Mexico. Seasonal abundance varied for each species. Shrimps of the
genus Penaeus were larger in deeper sections of the bay. The hydrological characteristics of the St.
Andrew Bay system are much more similair to the waters of the Gulf of Mexico than are those of other
estuaries of the northern gulf This similarity probably accounts for the relatively high abundance of
shrimps of the genera Trachypenaeus and Sicyonia in the bay system. Also, this similarity probably
delays the gulfward migration of shrimps of the genus Penaeus and accounts for their large sizes in
the system.

Personal observations made on exploratory col-
lecting trips and on cruises aboard shrimp trawl-
ers within the St. Andrew Bay system in north-
west Florida had led us to believe that some
species of marine organisms normally found in
offshore waters of the Gulf of Mexico occurred
commonly within the system. For example,
penaeid shrimps of the genera Trachypenaeus
and Sicyonia, which are rare in bay systems of
the northern gulf, were observed frequently. Also,
shrimps of the genus Penaeus appeared to be
much larger within the St. Andrew Bay system
than other estuarine areas. It thus appeared to us
that the penaeid shrimps of the St. Andrew Bay
system were unusual in terms of species composi-
tion and size.

Although utilization of estuarine waters by
populations of shrimps of the genus Penaeus is
well known (Lindner and Cook 1970; Cook and
Lindner 1970; Costello and Allen 1970), the
abundance, distribution, and size are not com-
pletely described for all penaeid species within
many estuarine waters. This information is espe-
cially lacking along the northwest Florida coast.
The objectives of our study were to estimate these
parameters for penaeid shrimps in the St. An-
drew Bay system.

'Gulf Coastal Fisheries Center, Panama City Laboratory, Na-
tional Marine Fisheries Service NOAA, Panama City, FL
32401.

STUDY AREA

The St. Andrew Bay estuarine system is lo-
cated on the northwest coast of Florida between
lat. 30°00' and 30°20'N and long. 85°23' and
85°53'W. The system consists of four bays —
North, West, East, and St. Andrew (Figure 1) —
with mean depths of 1.8, 2.1, 2.1, and 5.2 m, re-
spectively, and covers an area of 280 km^
(McNulty et al. 1972). Various aspects of the
physical and biological characteristics of the St.
Andrew Bay system have been presented by
Ichiye and Jones (1961), Waller (1961), Vick
(1964), Hopkins (1966), Salsman et al. (1966),
Cosper (1972), and McNulty et al. (1972).

GUIF OF MEXICO

Figure l. — Location of sampling stations in the St. Andrew
Bay System, Fla.

Manuscript accepted July 1975.
FISHERY BULLETIN: VOL. 74, NO. 1,

158 /S^^-/4(a.

1976.

BRUSHER and OGREN: PENAEID SHRIMPS IN ST. ANDREW BAY SYSTEM

Waters in the St. Andrew Bay system are rela-
tively high in transparency. This high transpar-
ency results in part from the porosity of the soils
of the watershed, the low freshwater inflow, and
the proximity of the system to the clear waters of
the northeastern Gulf of Mexico. In terms of ex-
tinction coefficients, the transparency of gulf
waters adjacent to St. Andrew Bay are typical of
clear oceanic waters (Tolbert and Austin 1959).

The bottom of the bay system is composed of
distinct sediment regimes. The sand regime
(>80% sand) is generally restricted to areas near
the passes and in depths less than 6 m. The silt-
clay regime (>50% clay, <50% silt, and <20%
sand) is located in the deeper waters of the sys-
tem, but not in the passes (Waller 1961).

The bay system also contains areas covered by
rooted submerged vegetation. The submerged
vegetation includes turtle grass, Thalassia tes-
tudinum; manatee grass, Syringodium filiforme;
and shoal grass, Diplanthera wrightii. These
grasses cover an area of about 3,200 hectares.

METHODS

Sampling was conducted every 2 wk from 6
September 1972 through 21 August 1973 at 12
stations (Figure 1, Table 1). Two consecutive
nights were required to sample at all stations
with samples taken between sunset and 0200 h.
On 23-24 August 1973 additional sampling was
conducted between 1000 and 1400 h at the 12 sta-
tions to compare day catches with the night
catches of 20-21 August 1973.

Biological samples were obtained at each sta-
tion with an 11.5-m wing trawl with stretched
meshes of 7.6 cm in the wings, 3.8 cm in the body,
and 2.5 cm in the cod end. The trawl was towed at
about 3.5 knots for 10 min. The entire catch at
each station was placed on ice and transported to

Table l. — Locations and depth ranges of sampling stations in
the St. Andrew Bay system, Fla.

Identifying

Depth range

station

Lat.'

Long.'

landmark

(m)

30050N

85°31.0'W

Goose Point

4.6- 6.1

30 06,3N

8535.0W

Shoal Point

7.6- 9.1

30=07.6N

85='37.7'W

Palmetto Point

7.6- 9.1

30 09.0'N

85=40. 8W

Redflsh Point

10.7-12.2

30'09,5'N

85'41.6W

Baker Bayou

6.1- 7.6

30'06.2N

85°41.3'W

Stiell Island

6.1- 7.6

3009.4'N

85°42.8'W

Courtney Point

76- 9.1

30M0.4'N

85=43. 0'W

Lake Huntington

6.1- 7.6

30'10,5'N

85°44.2'W

Dyers Point

10.7-12.2

30=1 4. VN

85'44.3'W

Shell Point

6.1- 7.6

30°15.7'N

85=46. 6'W

Breakfast Point

3.1- 4.6

30^'15.4'N

85=40.0'W

Haven Point

1.5- 3.1

the laboratory and frozen. Catches were thawed
and processed usually within 1 wk of collection.
Penaeid shrimps from each sample were enumer-
ated by species, and 30 individuals, or all if less
than 30, were measured to the nearest 0.5 cm
total length (tip of rostrum to tip of telson).

Environmental data were also obtained at each
station. A water sample for determining dis-
solved oxygen and turbidity was taken 0.5 m
above the bottom at each station with a 3-liter
water sampler Salinity and temperature were
determined in situ with a Beckman^ RS5-3 por-
table salinometer (accuracy ±0.5°C and ±0.31,)
at the above mentioned depth. Turbidity was
determined with a Hach turbidimeter (Formazin
turbidity units — accuracy ±0.02 FTU), and dis-
solved oxygen determined by the modified Wink-
ler method (accuracy ±0.05 ml/liter).

For each species, differences in catch per unit
effort (average catch per tow), and in size (aver-
age length by date) between subareas were tested
statistically with Tukey's a;-procedure (Steel and
Torrie 1960). For length comparisons, data were
used for only those dates when shrimps of a
species were caught in all subareas. For compari-
sons of distribution and abundance, the data were
grouped into the following subareas: East Bay
(stations 1, 2); North Bay (station 12); West Bay
(stations 10, 11); St. Andrew Bay (stations 3-5,
7-9); and East Pass (station 6).

Mean catches per tow and mean total lengths
were also compared between upper and lower bay
areas. The upper area included all stations in
East Bay, North Bay, and West Bay, and the lower
area included all stations in St. Andrew Bay and
East Pass.

ENVIRONMENTAL FACTORS

Mean values of environmental factors near the
bottom were determined for subareas. Salinities
and dissolved oxygen were higher in St. Andrew
Bay and East Pass than in the other subareas
(Table 2). Turbidities in North Bay, East Bay, and
West Bay were greater than in St. Andrew Bay
and East Pass. Bottom temperatures, however,
were similar among subareas.

When subarea data were combined into the re-
spective upper and lower areas, the average val-
ues were: salinity— 29.2, 33.2%; turbidity— 3.0,

'United States Department of Commerce, Nautical Chart 868-SC.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

159

FISHERY BULLETIN: VOL. 74, NO. 1

Table 2. — Annual means and ranges of environmental factors measured in
1972-73 in five subareas of the St. Andrew Bay system, Fla.

Environmental
factor

North
Bay

West
Bay

East
Bay

East
Pass

St. Andrew
Bay

Salinity (%.)
Mean

27.20

29.08

30.34

32.97

33.27

Range

13.1-32.5

20.5-34.1

25.3-33.9

30.3-35.2

30.6-35.6

Turbidity (FTU)'
Mean

2.69

3.40

2.63

1.09

1.75

Range

0.50-13.00

1 53-7.55

1.50-5.20

0.60-2.15

0.87-4.09

Temp (°C)
Mean

21.74

21.82

21.79

22 13

21.74

Range

13.1-31.1

13.6-30.2

13.8-29.9

13.0-30.2

13.2-30.0

Dissolved 02 (ml/liter)
Mean

3.87

3.77

3.27

4.43

4.01

Range

1.33-5.37

2.06-4.70

1.64-5.58

3.47-5.13

3.13-4.80

No. of samples

182

'Formazin turbidity units.

1.7 FTU; temperature— 21.8°, 21.8°C; dissolved
oxygen — 3.6, 4.1 ml/liter. Generally, salinity and
dissolved oxygen values were higher in the lower

area, turbidity values were higher in the upper
area, and temperatures were similar between
areas (Figure 2). The only noteworthy variation

Figure 2. — Mean values of salinity, turbid-
ity, temperature, and dissolved oxygen in the
upper and lower areas of the St. Andrew Bay
system, Fla., 1972-73.

"4 i« 3 17 30 U 27 13 26 8 22 5 20 5 19

Stp OCT. NOV. DEC. JAN. FEB MAR,

1972

3 11 30 14 29 13 2S 9 23

APR MAT JUN. nil.

1973

7 20
AUG.

160

BRUSHER and OGREN: PENAEID SHRIMPS IN ST. ANDREW BAY SYSTEM

in these values occurred in the salinity of the
upper area where heavy spring rains accounted
for an exceptional drop in salinity in early April.
Compared to hydrological data from other
northern gulf estuaries (Gunter 1950; Swingle
1971; Dunham 1972; Stokes 1974), the values
within the St. Andrew Bay system are more
oceanic than estuarine (Waller 1961; Hopkins
1966).

CATCHES

Eight species of penaeids were taken during
the study: three species ofPenaeus (P. duorarum,
P. aztecus, and P. setiferus), two species of
Trachypenaeus (T. similis and T. constrictus) , and
three species of Sicyonia iS. breuirostris, S. dor-
salis, and S. typica). Catches of each species at
each of the 12 stations are shown in Table 3. The
greatest number of individual shrimps (species
combined) was taken at station 4 (St. Andrew
Bay), the least at station 2 (East Bay). Penaeus
duorarum was the most abundant species, S.
typica the least. Since only 25 S. typica (ranging
in size from 3.5 to 5.5 cm) were caught, this
species will not be discussed in the following
sections.

Although methods were similar, a striking dif-
ference was apparent between our catches and
those from other estuarine systems in the north-
ern Gulf of Mexico. In our study, 57.6% of the
total penaeid catch consisted of members of the
genus Penaeus, 22.6% of the genus Sicyonia, and
19.7% of Trachypenaeus. In contrast, studies in
other estuarine systems in Alabama (Swingle
1971), Louisiana (Dunham 1972), and Texas
(Gunter 1950; Moffett 1968; Stokes 1974) showed

that the genus Penaeus represented 99 to 100% of
the total trawl catch of penaeids.

DISTRIBUTION AND
ABUNDANCE

To determine where shrimp were more abun-
dant in the St. Andrew Bay system, relative
abundances were compared by subarea (Table 4).
Significant differences were found for four of the
seven species: T. similis, S. breuirostris, S. dor-
salis, and T. constrictus. Either St. Andrew Bay or
East Pass or both had significantly greater abun-
dance of these species than the other subareas.

When subarea data were combined for each
species and apportioned into upper and lower
areas, the relative abundances were greater in
the upper area for P. aztecus and P. setiferus and
were greater in the lower area for the other
penaeids. Average catches per tow for the upper
and lower areas, respectively, were: P. duorarum,
110.8, 129.3; T. similis, 12.8, 49.4; S. breuirostris,
6.0, 51.3; S. dorsalis, 2.9, 32.9; T. constrictus, 3.1,
14.8; P. aztecus, 10.1, 4.6; P. setiferus, 2.7, 0.3.

To determine seasonal distribution and abun-
dance, the catches per tow were calculated by
area and by date for each species. The results,
shown in Figure 3, indicate summer and fall
abundances for the three species of Penaeus, al-
though not necessarily in both areas. For Trachy-
penaeus and Sicyonia, seasonal abundances were
evident only in the lower area, with T similis and
S. dorsalis more abundant during spring and
summer, S. breuirostris more abundant during
winter and early spring, and T constrictus during
spring.

Table 3. — Total numbers of penaeid shrimps caught in 312 trawl hauls within the St. Andrew Bay system, Fla., from September 1972

through August 1973.

Station

Species

Total

Pink shrimp,

Penaeus duorarum

3,485

1,613

2,724

1,879

5,097

3,115

3,348

4,767

2,382

3,062

3,371

2,737

37,580

Broken-neck shrimp,

Trachypenaeus similis

1,140

1,553

2,724

101

418

1,095

1,218

1,878

383

10,599

Rock shrimp.

Sicyonia brevirostris

147

984

1,758

3,812

1,552

717

198

9,234

Rock shrimp,

Sicyonia dorsalis

273

632

3,433

247

434

226

993

6,387

Broken-neck shrimp,

Trachypenaeus constnctus

150

207

704

907

275

248

208

122

3,064

Brown shrimp,

Penaeus aztecus

125

144

119

146

187

165

197

342

279

1,889

White shrimp.

Penaeus setiferus

166

424

Rock shnmp.

Sicyonia typica

Total

3,802

3,201

5,368

9,355

7,749

8,657

6,905

7,356

5,768

3,832

3,988

3,221

69,202

Rank

161

FISHERY BULLETIN: VOL. 74, NO. 1

Table 4. — Comparisons of mean catch per tow of penaeid shrimps between
subareas (Tukey's w-procedure with 125 df) of the St. Andrew Bay system, Fla.,
from September 1972 through August 1973.

Species

Subarea, mean catch in parentheses and significance lines'

Penaeus duorarum

Trachypenaeus similis

Sicyonia brevirostris

East

North

East

West

St. Andrew

Bay

Pass

Bay

(100.4)

(105.3)

(119.8)

(124.4)

(130.9)

North

West

East

St. Andrew

Bay

Pass

Bay

(0.1)

(7.7)

(15.9)

(24.3)

(55.0)

North

East

West

St. Andrew

East

Bay

Pass

(0.4)

(0.6)

(0.7)

(35.4)

(146.6)

North

West

East

St. Andrew

Bay

Pass

Bay

dorsalis

(0.0)

(1.8)

(5.5)

(9.5)

(36.8)

East

West

North

St. Andrew

East

Bay

Pass

constrictus

(2.2)

(3.1)

(4.7)

(11.4)

(34.9)

East

St. Andrew

East

West

North

Bay

Pass

Bay

aztecus

(4.0)

(4.4)

(5.6)

(10.5)

(10.7)

East

St. Andrew

East

North

West

Pass

Bay

setiferus

(0.0)

(0.4)

(1.4)

(2.6)

(4.1)

'Any two means not underscored by the same line are significantly different at the 5% level.

2
o

>
a
z

Z
<

Figure 3. — Mean catch per tow of seven
penaeid shrimp species in the upper and
lower areas of the St. Andrew Bay system,
Fla., 1972-73.

6 IS 3 17 30 14 27 13 26 8 23 5 20 5 19

SEP OCT NOV DEC JAN FEB MAR

1972

3 18 30 14 29 13 25 9 23 7 20

APR MAY JUN JUL AUG

1973

162

BRUSHER and OGREN: PENAEID SHRIMPS IN ST. ANDREW BAY SYSTEM

Penaeid shrimps taken from the St. Andrew
Bay system showed definite habitat preference by
genera when abundance was related to depth. As
shown in Table 5, the higher mean catches per
tow for Penaeus occurred in the shallower waters,
while those for Trachypenaeus and Sicyonia oc-
curred in the intermediate and deeper waters of
the sampled area. Ninety -two percent of all Tra-
chypenaeus and Sicyonia were taken from the
lower area where the average station depth was
8.6 m.

Day and night comparisons showed mean catch
per tow to be greater at night for all seven species
(Table 6).

Table 5. — Comparisons of mean catch per tow and mean
length (cm) of penaeid shrimps in relation to depth and species
within the St. Andrew Bay system, Fla., from September 1972
through August 1973.

1.5-4.6 m

4.6-7.6 m

7.6-12.2 m

Species

Stn. 11, 12

1, 5, 6, 8, 10

2. 3, 4, 7, 9

Penaeus duorarum

117.5

150,2

91,9

(9.1)

(9.5)

(10.0)

Trachypenaeus similis

0.2

16.9

64.6

(6.3)

(6.8)

Sicyonia brevirostris

0.3

48.6

22.3

(6.1)

(6.0)

(6.2)

S. dorsalis

0.0

4.8

38.8

(-)

(5.3)

(5.5)

T. constrictus

4.2

15.1

6.9

(4.7)

(4.8)

(4.9)

P. aztecus

12.0

5.0

4.7

(11.1)

(12.4)

(12.7)

P. setiferus

4.6

0.8

0.6

(11.5)

(12.9)

(14.1)

Table 6. — Comparisons of mean catch per tow and mean total
length (cm) between day and night catches of penaeid shrimps
taken from the St. Andrews Bay system, Fla., in August 1973.

Species

Day

Night

Penaeus duorarum
Tracliypenaeus similis
Sicyonia brevirostns
S. dorsalis
T. constrictus
P. aztecus
P. setiferus

No of tows

34.8

172.4

(8.2)

(8,4)

0.3

8.0

(6.4)

(6.2)

1.7

(-)

(7.6)

1.2

5.2

(5.6)

(5.4)

0.3

(-)

(4.1)

3.0

9.5

(13.1)

(13.2)

0.8

1.5

(11.1)

(10.2)

SIZE

Shrimps of the genus Penaeus were larger than
shrimps of the other two genera. Penaeus seti-
ferus had the largest mean length, while S. dor-
salis had the smallest. Mean total lengths in cen-
timeters and length ranges in centimeters for

each species in the St. Andrew Bay system were:
P. duorarum, 9.5, 4.0-18.5; T. similis,6.6, 3.0-10.0;
S. brevirostris, 5.7, 2.8-9.5; S. dorsalis, 5.5, 2.0-
7.8; T. constrictus, 4.5, 2.5-8.0; P. aztecus, 12.4,
4.5-18.5; and P. setiferus, 13.3, 7.0-16.0.

Differences in lengths of shrimps associated
with water depth were examined (Table 5); nota-
ble differences were discernible only for the genus
Penaeus, the larger specimens of which generally
were found in deeper waters. This relation has
also been reported by others (Lindner and Cook
1970; Cook and Lindner 1970; Costello and Allen
1970). Species of Trachypenaeus and Sicyonia
showed little difference in mean lengths with
water depths, although the largest mean sizes
were found in the deeper zone.

Examination for differences in lengths associ-
ated with sampling at night and during the day
revealed clearly that hour of sampling had no ef-
fect on size of captured shrimps (Table 6).

Comparisons of mean total lengths for the
seven species between those subareas from which
sufficient data were available showed that the
largest shrimps were in either St. Andrew Bay or
East Pass (Table 7). However, statistically sig-
nificant differences were found for only three
species: P duorarum, T. similis, and P setiferus.

For five of the seven species, larger specimens
were caught in the lower area more often than in
the upper area. The situation was reversed for S.
brevirostris, whereas, for T constrictus the mean
sizes for the two areas were the same. Mean
lengths in centimeters by species between upper
and lower bay areas, respectively, were: P.
duorarum, 9.1, 9.9; T. similis, 6.4, 6.7; S. bre-
virostris, 6.3, 5.7; S. dorsalis, 5.4, 5.6; T. constric-
tus, 4.5, 4.5; P. aztecus, 11.9, 12.8; and P. setiferus,
11.7, 14.7.

Shrimps of the genus Penaeus were almost con-
sistently larger in the lower area throughout the
year (Figure 4). As shrimps of this genus grow
larger, they tend to move into deeper, more saline,
and less turbid waters.

When present in both areas at the same time,
the two species of Trachypenaeus were larger in
the lower area more often than in the upper,
whereas the reverse was true of the two species of
Sicyonia.

DISCUSSION AND CONCLUSIONS

In general, water depths and salinities are
greater, and turbidities, temperature fluctua-

163

FISHERY BULLETIN: VOL. 74, NO. 1

Table 7. — Comparisons of mean total length (cm) of penaeid shrimps between subareas
(Tukey's u;-procedure) of the St. Andrew Bay system, Fla., from September 1972 through
August 1973.

Species

Subareas, mean total length in parentheses, and significance lines'

Penaeus duorarum

North
Bay
(8.89)

West
Bay

(9.12)

East
Bay
(9.19)

East
Pass
(9.77)

St. Andrew
Bay
(682)

East
Pass
(6.30)

North
Bay
(4.77)

St. Andrew
Bay
(9.81)

East
Pass
(4.90)

120

Trachypenaeus similis

East
Pass
(5.86)

West
Bay
(6.20)

East

Bay
(6.67)

Sicyonia brevirostns
S. dorsalis
T. constrictus

St. Andrew
Bay

(5.66)

East
Bay
(5.37)

East
Bay

(4.23)

East
Pass
(5.81)

West
Bay
(5.44)

St. Andrew
Bay
(4.43)

St. Andrew
Bay
(5.50)

West
Bay
(4.67)

36
10

P. aztecus

North
Bay
(11.41)

West
Bay

(11.53)

East
Bay
(12.50)

St. Andrew
Bay
(12.79)

East
Pass
(12.96)

P. setiferus

East
Bay

(11.03)

West
Bay

(11.68)

North
Bay
(12.90)

St Andrew
Bay
(14.68)

'Any two means not underscored by the same line are significantly different at the 5% level.

tions, and river discharges are lower in the St.
Andrew Bay system than in other northern gulf
estuaries (Apalachicola Bay to the Rio Grande
River). The dominant group of spermatophytes in
the lower area are the submerged sea grasses,
whereas in most other northern gulf estuaries the
dominant groups are the emergent grasses in the
intertidal zone (Kutkuhn 1966). This unusual es-
tuarine environment in the St. Andrew Bay sys-
tem may induce shrimps of the genus Penaeus to
remain within the system for longer periods of
time, especially in the lower areas where oceanic
conditions often prevail.

Such environmental differences probably ac-
count for the differences observed in composition,
abundance, and size of penaeid shrimps between
the St. Andrew Bay system and other estuarine
systems in the northern Gulf of Mexico. For
example: 1) large adult (total length ranges of
16.5 to 18.5 cm) P. duorarum and P. aztecus usu-
ally occur only in offshore waters, but we caught
many of these large specimens throughout the St.
Andrew Bay system; 2) in low salinity waters
characteristic of other bay systems subadult P.
setiferus and P. aztecus are more abundant than
P. duorarum, whereas in the St. Andrew Bay
system we found subadult P. duorarum more
abundant than P. setiferus and P. aztecus; and 3)
previous reports indicated that T. similis, S.

brevirostris, and .S. dorsalis do not ordinarily
enter estuaries (Eldred 1959; Joyce 1965; Kut-
kuhn 1966; Cobb et al. 1973), but we caught many
individuals of these species within the St. Andrew
Bay system.

The abundance of shrimps of Trachypenaeus
and Sicyonia in the St. Andrew Bay system con-
trasts sharply with those reported from other
estuarine areas of the Gulf of Mexico. Other in-
vestigators have included catches made adjacent
to barrier islands or tidal passes and reported
abundances of less than 1 shrimp per tow. (Dun-
ham 1972; Gunter 1950; Saloman 1964, 1965;
Swingle 1971). In our study, average catch per tow
(excluding Station 6, which is adjacent to a bar-
rier island) for each species was: T. similis, 36; T.
constrictus, 8; S. brevirostris, 19; S. dorsalis, 21.

Periods of greatest abundance of S. brevirostris
in offshore waters of the northwestern and south-
eastern gulf occur in summer and early fall
(Brusher et al. 1972; Cobb et al. 1973). In the St.
Andrew Bay system, this species was almost ab-
sent during this period. We believe that this
shrimp migrates from inshore to offshore gulf
waters during spring months.

Means and ranges of total lengths of species of
Trachypenaeus or Sicyonia taken in other es-
tuarine areas were usually less (Swingle 1971;
Dunham 1972) than those taken in offshore areas

164

BRUSHER and OGREN: PENAEID SHRIMPS IN ST. ANDREW BAY SYSTEM

>
a

< '

O 4

^•

< \6

S u

16
14
13
10
8

Penaeus duorarum

Trachypenaeus iimilis

Sicyonia brevirosfris

Sicyonio dorsalis

Trachypenaeus constrictui

LOWER AREA
UPPER AREA

Penaeus aziecus

Penaeus setilerus

''''''

'''''''

Figure 4. — Mean total lengths of seven
penaeid shrimp species in the upper and
lower areas of the St. Andrew Bay system,
Fla., 1972-73.

6 IS

3 17 30 14 27

13 26

8 22

5 20

5 19

18 30

14 29

13 25

9 23

7 20

SEP

OCI NOV

1972

DEC

JAN

FEB

MAR

APR

1973

MAY

JUN

JUL.

AUG

of the Gulf of Mexico (Brusher et al. 1972). The
mean total lengths of the penaeids with the ex-
ception of T. constrictus (Table 7) were similar to
those reported by Brusher et al. (1972) for speci-
mens caught in the Gulf of Mexico. We believe
that species of Trachypenaeus and Sicyonia
utilize St. Andrew Bay as a nursery area owing to
the similarity of the bay to offshore oceanic
habitats.

Of the three species oi Penaeus caught in this
study, P. duorarum was the most abundant. High
abundance of P. duorarum was expected, because
the highest concentration of this species in the
Gulf of Mexico occurs in the eastern areas (Cos-
tello and Allen 1970). Costello and Allen as-
sociated P. duorarum with grass beds; grass beds
are abundant in St. Andrew Bay. Low abundance
of P. aztecus and P. setiferus was expected also, as
these are found most abundantly in the north-

western (Texas coast) and north central (Louisi-
ana coast) portions of the Gulf of Mexico, respec-
tively (Cook and Lindner 1970; Lindner and Cook
1970).

Although similar gear and trawling methods
were used, mean total lengths and length ranges
of P. aztecus and P. duorarum caught in the St.
Andrew Bay system differed greatly from those
caught in other gulf estuaries (Saloman 1965;
Trent et al. 1969; Dimham 1972). Our catches in-
cluded many specimens over 13.0 cm total length
which, according to Joyce (1965), is well above the
size at which shrimps of the genus Penaeus are
believed to leave estuarine areas. Shrimps of this
genus greater than 10 cm total length are usually
found in offshore waters (Lindner and Cook 1970;
Cook and Lindner 1970; Costello and Allen 1970).

We conclude that the St. Andrew Bay system is
unusual among estuaries of the northern Gulf of

165

FISHERY BULLETIN: VOL. 74, NO. 1

Mexico; its environmental qualities which are
much more similar to those in the gulf account for
the common occurrence in the bay of penaeid
shrimps of the genera Trachypenaeus and Si-
cyonia normally found in the offshore waters of
the open gulf; the unusual environmental factors
within the system also delay the migration of
penaeid shrimps of the genus Penaeus into the
open gulf, thereby allowing them to grow larger
within the St. Andrew Bay system.

ACKNOWLEDGMENTS

We thank Maxwell Miller and Leslie Touart for
their help in collecting and processing the sam-
ples and David Muenzel, Captain of the RV
Rachel Carson, for his assistance in keeping us on
schedule. We gratefully acknowledge the critical
reviews of this manuscript by David Aldrich
(Texas A&M University), Donald Allen (National
Marine Fisheries Service, NOAA), and William
Lyons (Florida Department of Natural Re-
sources).

LITERATURE CITED

brusher, h. a., w. c. renfro, and r. a. neal.

1972. Notes on distribution, size, and ovarian development
of some penaeid shrimps in the northwestern Gulf of
Mexico, 1961-62. Contrib. Mar Sci. 16:75-87.

Cobb, S. p., C. R. Futch, and D. K. Camp.

1973. The rock shrimp, Sicyonia brevirostris Stimpson,
1871 (Decapoda, Penaeidae). Mem. Hourglass Cruises
3(l):l-38.

Cook, H. L., and M. J. Lindner.

1970. Synopsis of biological data on the browTi shrimp
Penaeus aztecus aztecus Ives, 1891. FAO (Food Agric.
Organ. U.N.) Fish. Rep. 57:1471-1497.
COSPER, T. C.

1972. The identification of tintinnids (Protozoa: Ciliata:
Tintinnida) of the St. Andrew Bay system, Florida. Bull.
Mar Sci. 22:391-418.
COSTELLO, T J., AND D. M. ALLEN.

1970. Synopsis of biological data on the pink shrimp
Penaeus duorarum duorarum Burkenroad, 1939. FAO
(Food Agric. Organ. U.N.) Fish. Rep. 57:1499-1537.
DUNHAM, F.

1972. A study of commercially important estuarine-
dependent industrial fishes. La. Wildl. Fish. Comm.,
Tech. Bull. 4, 63 p.
ELDRED, B.

1959. A report on the shrimps (Penaeidae) collected from
the Tortugas controlled area. Fla. State Board Conserv.
Mar. Lab., Spec. Sci. Rep. 2, 6 p.
GUNTER, G.

1950. Seasonal population changes and distributions as re-
lated to salinity, of certain invertebrates of the Texas
coast, including the commercial shrimp. Publ. Inst. Mar
Sci., Univ. Tex. 1(2):7-51.

Hopkins, T. L.

1966. The plankton of the St. Andrew Bay system, Flor-
ida. Publ. Inst. Mar Sci., Univ. Tex. 11:12-64.

ICHIYE, T, AND M. L. Jones.

1961. On the hydrography of the St. Andrew Bay system,
Florida. Limnol. Oceanogr. 6:302-311.
JOYCE, E. A., JR.

1965. The commercial shrimps of the northeast coast of
Florida. Fla. State Board Conserv. Mar. Lab., Prof Pap.
Ser. 6, 224 p.

KUTKUHN, J. H.

1966. The role of estuaries in the development and per-
petuation of commercial shrimp resources. Am. Fish.
Soc, Spec. Publ. 3:16-36.

LINDNER, M. J., AND H. L. COOK.

1970. Synopsis of biological data on the white shrimp
Penaeus setiferus (Linnaeus) 1797. FAO (Food Agric.
Organ. U.N.) Fish. Rep. 57:1439-1469.

MCNULTY, J. K., W. N. LINDALL, JR.. AND J. E. SYKES.

1972. Cooperative Gulf of Mexico estuarine inventory and

study, Florida: Phase 1, area description. U.S. Dep.

Commer, NOAA Tech. Rep. NMFS CIRC-368, 126 p.
MOFFETT, A.

1968. A study of Texas shrimp populations, 1968. Tex.
Parks Wildl. Dep., Coastal Fish. Proj. Rep., p. 67-93.

SALOMAN. C. H.

1964. The shrimp Trachypeneus similis in Tampa Bay. Q.
J. Fla. Acad. Sci. 27:160-164.

1965. Bait shrimp {Penaeus duorarum) in Tampa Bay,
Florida — biology, fishery economics, and changing
habitat. U.S. Fish Wildl. Serv, Spec. Sci. Rep. Fish. 520,
16 p.

SALSMAN, G. G., W. H. TOLBERT, AND R. G. VILLARS.

1966. Sand-ridge migration in St. Andrew Bay, Flori-
da. Mar. Geol. 4:11-19.

Steel. R. G. D., and J. H. Torrie.

1960. Principles and procedures of statistics, with special
reference to the biological sciences. McGraw-Hill Book
Co., N.Y., 481 p.

Stokes, G. M.

1974. The distribution and abundance of penaeid shrimp in
the lower Laguna Madre of Texas, with a description of
the live bait shrimp fishery. Tex. Parks Wildl. Dep.,
Tech. Ser. 15, 32 p.
SWINGLE, H. A.

1971. Biology of Alabama estuarine areas — cooperative
Gulf of Mexico estuarine inventory. Ala. Mar Resour.
Bull. 5, 123 p.

TOLBERT, W. H., AND G. B. AUSTIN

1959. On the nearshore marine environment of the Gulf of

Mexico at Panama City, Florida. U.S. Navy Mine Def

Lab., Tech. Pap. TP-161, 104 p.
TRENT, W. L., E. J. PULLEN, C. R. MOCK, AND D. MOORE.

1969. Ecology of western Gulf estuaries. In Report of the
Bureau of Commercial Fisheries Biological Laboratory,
Galveston, Texas, fiscal year 1968, p. 18-24. U.S. Fish
Wildl. Serv., Circ. 325.

VICK, N. G.

1964. The marine ichthyofauna of St. Andrew Bay,
Florida, and nearshore habitats of the northeastern Gulf
of Mexico. Tex. A&M Univ. Dep. Oceanogr. Meteorol.,
Proj. 286-D, Ref 64-19T, 77 p.

WALLER, R. A.

1961. Ostracods of the St. Andrew Bay system. M.S.
Thesis, Florida State Univ., Tallahassee, 46 p.

166

SOME FEATURES OF COHO SALMON, ONCORHYNCHUS KISUTCH,

FRY EMERGING FROM SIMULATED REDDS

AND CONCURRENT CHANGES IN PHOTOBEHAVIOR

J. C. Mason^

ABSTRACT

The emergence of sibling coho fry from simulated redds lasted 20-23 days during which 97-98% of the
fry emerged. Average size of emerging fry increased with time but the largest fry emerged during the
peak of emergence. No clear preference was shown for nocturnal or daylight emergence but the latter
increased with time. Fry showed a positive current response, 69-82% moving upstream following
emergence. Most fry emerged when yolk reserve was reduced to less than 10% of total dry weight.
Later-emerging fry did not have lower yolk reserves, but fry moving downstream had slightly more
yolk reserve than did fry moving upstream. Fry which were captured shortly after emergence had fed
actively but had not yet filled their air bladders. Chironomids composed 70% of their diet.

Photoresponse of sibling fry denied the redd experience was studied in light-dark choice boxes with
reference to the timing of emergence of fry from the simulated redds. The pronounced photonegative
behavior of the denied fry was suddenly lessened at time of emergence but remained photonegative.
Weakening of the negative photoresponse was not the outcome of starvation or recent light experi-
ence, and was not modified by repeated testing. Retention of the photonegative response is referred to
hiding behavior and use of the gravel bed as a refuge.

The anadromous female Pacific salmon, On-
corhynchus, usually buries her eggs in several
adjacent pockets in streambed or lakeshore ma-
terials and these egg pockets collectively consti-
tute a redd. The eggs hatch after several months
and the larvae may spend several weeks or
months using up their extensive yolk stores prior
to emerging from the redd area into open water.

Mortality during this extended period of sub-
terranean life may be considerable (Royce 1959)
and probably routinely exceeds 70% for most
species of salmonids in natural habitats. Adap-
tion to suboptimal conditions includes physiologic
and behavioral responses in the embryo and larva
which were reviewed, especially for sockeye
salmon, O. nerka, by Bams (1969).

Because destructive influences on the egg and
alevin stages are amenable to amelioration
through manipulation of substrate structure and
flow regime, spawning channels pioneered by
Wickett (1952) at Nile Creek have become a
major component of salmon enhancement
strategy. Despite these advances, we have yet to
define optimal redd conditions, biotic and abiotic,
which maximize preemergence survival of any

'Department of the Environment, Fisheries and Marine Ser-
vice, Research and Development Directorate, Biological Sta-
tion, Nanaimo, B.C. V9R 5K6, Canada.

salmonid. Furthermore, fry surviving to
emergence may face extended ecological conse-
quences of suboptimal conditions in the redd
which alter timing of, or size at, emergence (Ma-
son and Chapman 1965; Mason 1969). Neither
can we yet define for the emerging fry physiologic
and behavioral states which optimize survival in
open waters. Thus, premature emergence, imply-
ing underdevelopment and reduced ability to re-
spond adaptively is not referrable to a defined
state of normality.

Alevins of Oncorhynchus , as are those ofSalmo
and Saluelinus (White 1915; Stuart 1953;
Woodhead 1957), are initially negatively photo-
tactic and respond to light by hiding (Hoar 1958).
They become positively phototactic and rheotac-
tic as emerged fry, orientation to current preced-
ing the shift from negative to positive phototaxis
(Dill 1969) as in Salmo (Grey 1929a; Stuart 1953)
but the timing of this photobehavioral change in
relation to emergence and remaining yolk re-
serve remains unknown in Oncorhynchus and
disagreement has arisen as to its timing in Salmo
(Woodhead 1957). Histophysiological studies by
Ali (1959) showed that only emerged fry and
older stages of Oncorhynchus are capable of full
retinomotor responses; however, partially devel-
oped responses have obvious survival value.

In this paper, some features of sibling coho fry

Manuscript accepted September 1975.
FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

167

FISHERY BULLETIN: VOL. 74, NO. 1

emerging from simulated stream redds are de-
scribed. Light and current responses; length,
weight, and condition; remaining yolk reserves at
emergence; and changes in photoresponse were
investigated. The possible effects on photore-
sponse of repeated testing, previous exposure to
light, and feeding experience were also examined.

MATERIALS AND METHODS
Emergence from Simulated Redds

The emergence of coho salmon fry of known
parentage (two males x one female) from four
simulated redds was investigated in two pairs of
wooden channels (Figure 1) located outdoors.
Each channel was divided into three equal-sized
compartments, and to simulate a redd, each
center compartment was filled to a depth of 27 cm
with stream pebbles 2-5 cm in diameter. A stand-
pipe terminating at its lower end in a 10 cm x 10
cm platform on 10 cm stilts so as to enclose a
chamber of 100 cm^ volume was buried in each
redd at this time. In each redd the gravel surface
was entirely underwater, but a shallow median
depression served to concentrate the surface flow
issuing through the V-notch openings.

The frames of the inner partitions were covered
with a double layer of fine plastic screen to allow
for circulation through the redds. Water flow
through each channel was 12 liters/min, about
30% of which passed through the redds.

Ten days after hatching, 150 alevins from eggs
incubated and hatched in standard baskets and
previously unexposed to light were introduced
into each redd at night via its standpipe and al-
lowed to emerge spontaneously. Each standpipe
was cleared of fry 1 h after stocking the redd by

inserting a wire rod capped with rubber stoppers
at either end and leaving the rod in place. Emerg-
ed fry could enter either the upstream or down-
stream compartments by way of the V-notch
openings and were collected there daily at dawn
and dusk.

Emerging fry were anesthetized with MS-222,2
fork length was measured to the nearest 0.1 mm
using a dissecting microscope, weight determined
to the nearest 0. 1 mg on a Mettler Grammamatic
balance after blotting, and the fry then preserved
in 5% Formalin. For each redd, samples of 20 fry
were extracted from each quartile of the emerg-
ing population (total of 80 fry per redd) divided
between fry moving upstream or downstream fol-
lowing emergence. Yolk reserve at emergence
was determined by dissecting out the yolk ma-
terial, drying both yolk and fry to constant
weight at 80°C, and expressing yolk reserve as a
percentage of total dry weight.

The resulting data were processed by regres-
sion and analysis of variance techniques to ex-
pose possible correlations between length,
weight, condition {K) and yolk reserve with time,
directional movement in current, and emergence
during the daylight or darkness.

Photoresponse Tests

Ten days after hatching, sibling alevins from
the same experimental stock as those used for the
emergence study but denied the redd experience
were separated into five groups of 50 fish each
and held indoors in wire baskets except during
testing. Two groups were held in complete dark-
ness. One of these groups was tested frequently
(dark experimental, DE); the other was tested
once then not retested until 15 days later (dark
control, DC). The three remaining groups were
held in baskets partly exposed to daylight of
about 200 ft-candles peak intensity from an ad-
jacent window and were given three different
treatments. One group was tested frequently
(light experimental, LE); one was tested once
then not retested until 15 days later (light con-
trol, LC). The remaining group was not tested
until the 18th day and, in contrast to the other
groups, was fed frozen ground beef liver three
times daily from day 9 onward (light control
plus food).

Figure l. — Compartmentalized wooden channels. Center
compartments contained the simulated redds. Dotted areas
signify screens.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

168

MASON: FEATURES OF EMERGING COHO SALMON FRY

Photoresponse tests were conducted in four
choice boxes placed in an uncompartmentalized
replicate of the emergence channels and located
adjacent to them. The choice boxes were con-
structed of fine plastic screen on a wire frame-
work (Figure 2) and divided equally into two
compartments by a vertical partition that al-
lowed a passage height of 1.5 cm beneath it. Both
hinged top and the partition were made of black
polyethylene sheeting. The wooden channel was
covered with the same material, except in the
areas taken up by the boxes, so that the com-
partments not covered by the hinged tops re-
ceived most of the illumination in the boxes. Each
box presented a choice between sharply contrast-
ing light conditions rather than between "light"
and "no light," because some light leaked under
the partitions. A series of mirrors was mounted 1
m above the water surface, allowing observation
from a blind.

Water flow in the channel was 10 liters/min
and velocity less than 10 cm/min. Water depth in
the choice boxes was 10 cm providing an air space
of 3 cm between the water surface and the ceiling
of the covered compartment. Average fish density
was set so as to allow about twice as much water
volume and 2.4 times as much bottom area as in
the holding baskets. Temperature of the water
supply (stream) ranged from 7.8° to 11.7°C during

Figure 2. — Light-dark choice box showing the reversible

opaque lid.

the experimental period. Light intensities at the
exposed water surface ranged from 700 to 4,000
ft-candles during photoresponse tests.

The procedure for a photoresponse test was as
follows. The appropriate group of fry was trans-
ferred to the test site in a covered pail, 40 fry were
netted out and 10 fry put in each of the four choice
boxes with the lids in an upright position. The
lids were then closed in a common direction, and
the remaining fry were returned to their holding
basket. In the choice boxes, all fry swam into the
dark compartments when the lids were dropped.
After 30 min, the number of fry observed in the
light compartments were recorded every 10 min
for 40 min (5 observations in each of 4 compart-
ments = 20 observations). A fish was considered
to be in a light compartment when its head was
visible. The lids were then reversed and, after 10
min, five additional observations were made at
10-min intervals. Thus, for each test, 40 counts
were recorded on 40 fry, which spent 2 h in the
choice boxes per test and about 10 min in the
transfer process. The photoresponse tests were
initiated 1 day before fry began emerging from
the simulated redds and continued until the 22nd
day of emergence. Length and weight measure-
ments were taken for all fry groups on the follow-
ing day. Data were tested for homogeneity using
chi-square. There was no significant difference
(P<0.01) between the first and second runs of five
observations each made in individual choice
boxes, x^ values ranging from 0.0 to 2.8 in 132
pairs of runs. Similarly, the data from individual
choice boxes proved homogeneous within each
test in 29 of the 33 tests performed (P<0.01) x^
values ranging from 1.1 to 7.8 with 3 degrees of
freedom. The remaining four tests contained
heterogeneous data, x^ values ranging from 14.7
to 36.5 and were excluded from further analysis.
With homogeneity assured within most tests, the
data within tests were pooled and processed.

RESULTS

Emergence from the Simulated Redds

Fry began emerging 25 days after hatching and
15 days following introduction to the redds.
Emergence proceeded for 20-23 days during
which 97-98% of the fry emerged. All four redds
showed a similar pattern of emergence, peaking
at the same time, 74 to 94% of the fry emerging
during the median 10 days (Figure 3).

169

FISHERY BULLETIN: VOL. 74, NO. 1

®J

1 1^

Jir"

nnl

\ J

A^ J^

DAY OF EMERGENCE

Figure 3. — Timing of coho salmon fry emergence from the
simulated redds.

The average size of fry increased significantly
(P<0.01) as emergence proceeded but the largest
fry emerged during the peak of emergence from
day 10 to day 15 (Figure 4).

More fry emerged at night than during the day
in redds 2 and 3 (57% and 60%, respectively), but
more fry emerged during the day in redd 1 (Table
1). No preference was shown by fry in redd 4
which emerged in equal numbers. Dividing the
data into two time intervals, days 1 through 11,
and days 12 through 24, revealed that emergence
during the day increased some 30% in all four
redds in the latter period.

Emerging fry showed a strong positive current
response, the majority (69-82%) moving upstream
subsequent to emergence, upstream movement
increasing but slightly as emergence proceeded.

There were no significant differences in aver-
age length and weight (P<0.01) of fry moving up-
stream or downstream following emergence, but
fry emerging during the day were, on the aver-
age, larger than those emerging at night (Table
1), significantly so in two redds (redd 3, P<0.05;

I
I-

* ■

35^

••:?

^^■.

12 15

21 24

6 9 12 15

DAY OF EMERGENCE
Figure 4. — Size of coho salmon fry at emergence including the regression lines.

170

MASON: FEATURES OF EMERGING COHO SALMON FRY

Table l. — Average lengths of sibling coho fry emerging from four simulated redds, stratified as to night
and day timing and direction of movement. Values in parentheses are percentages.

Upstream

Downstream

Nigfit

Day

Redd

movement

emergence

Number of fry

94(69.1)

42(30.9)

53(39.0)

83(61.0)

Mean fork lengtfi (mm)

± SE

38.39 ± 0.11

38 12 ± 0.18

38.13 ± 0 18

0.15 ± 0.15

Number of fry

115(80.4)

28(19.6)

82(57,3)

61(42.7)

Mean fork lengtfi (mm)

± SE

38,08 ± 0.11

37.91 ± 0.21

38.00 ± 0.12

38,11 ± 0,15

Number of fry

113(79.0)

30(21.0)

86(60,1)

57(39,9)

Mean fork lengtfi (mm)

± SE

38.25 ± 0.11

38.48 ± 0.18

38.14 ± 0.11

38,54 ±0,15

Number of fry

116(82,3)

25(17.7)

70(49.6)

71(50.4)

Mean fork length (mm)

± SE

38.10 ± 0,11

38.00 ± 0.38

37.68 ± 0.15

38.48 ± 0 13

Total emerging fry

438(77.8)

125(22.2)

291(51.7)

272(48,3)

Pooled

mean fork lengtfi (mm)

38.19

38.14

38.12

38.38

redd 4, P<0.01). This is the outcome of the ten-
dencies for both increased emergence during
the day and increased size at emergence as time
progressed.

Of the 584 fry that emerged from the simulated
redds, 14 rather small fry emerged 5 or more days
prior to the onset of general emergence. Twelve of
these fry emerged at night and went downstream.
They, and seven additional fry which also moved
downstream and were designated as cripples due
to truncated vertebral columns, were deleted from
the analyses.

Most fry emerged when their yolk reserve was
reduced to less than 10% of total dry weight (Fig-
ure 5), average reserve being 5-7% of total dry
weight. The three rather high points (days 9-10)
for fry moving downstream represent small sam-
ples whose means were inflated by premature fry.
Yolk reserve in these samples was either less
than 8% or ranged between 26 and 60% for indi-
vidual fry. The large standard errors shown in
Figure 5 are all associated with mean values
inflated by premature fry. Yolk reserve did not
diminish with time, indicating that the majority
of fry were in a similar nutritional state at
emergence. Although there were no significant
differences in length and weight between fry
moving up or downstream, fry moving down-
stream had more yolk reserve (9.2%) than did fry
moving upstream (7.4%), this difference being
significant at the 1% level. Similarly, in 13 of 16
possible pairs of samples from the four redds, the
downstream fry contained more yolk reserve.

To discover if fry were feeding within a short
time of emergence, the digestive tracts of 75 fry
emerging from redds 1 and 2 were examined.
These fry were representative with regard to
night or day emergence and upstream or down-
stream movement following emergence, through-
out the period of emergence. No dietary differ-

ences were found in the 74 fry that had fed
shortly before their capture. As both stomach and
intestine contained food particles, fry emerging
at night probably fed that night or during the
preceding hours of daylight while in the redd.
Chironomids constituted nearly 65% of their diet
(Table 2), mites and Collembola made up 17%,
and together these three items were consumed by
83% of the fry.

At time of capture in the upstream and down-
stream compartments, no fry had yet reached
neutral buoyancy although some had partially
filled air bladders.

O
O

40 •

o 20

^ in

O redd 1

A redd 2

D redd 3
D redd 4

» J

■

■ b

. \

I '1

■

D
I

4 6 8 10 12 14 16 IB

DAY OF EMERGENCE

Figure 5. — Yolk reserve of coho salmon fry at emergence.
Solid symbols indicate downstream movement following
emergence; open symbols indicate upstream movement. Ver-
tical bars indicate ±2 SE. For the remaining points, the range
in SE was 0.1-1.0, and 90% ranged from 0.1 to 0.6.

171

FISHERY BULLETIN: VOL. 74, NO. 1

Table 2. — Diet of 75 coho salmon fry emerging from two of four
simulated redds supplied with river water.

Number of

% of total

Food item

items

incidence

Chironomidae;

Larvae

34.4

32.0

Pupae

16.0

21 4

Imagines

12.8

14.7

Total

63,2

68.1

Hydracarina

13.6

2.7

Collembola

10.4

12.0

Ephemeroptera nymphs

4.0

6.7

Arachnida

3.2

5.3

Trichoptera larvae

1.6

2.7

Plecoptera nymphs

0.8

1.3

Coleoptera imagines

0.8

1.3

Hymenoptera

1.3

Plant fragments

1.6

2.7

Concurrent Changes in Photoresponse

Photoresponse testing of fry denied the redd
experience began on day 1, 1 day before their
counterparts in the simulated redds began
emerging. Their photoresponse remained essen-
tially negative throughout the time period when,
normally, they would have emerged. Until the
eighth day of emergence (day 9), less than 3% of
the denied fry were seen in the light compart-
ments (Figure 6) and they remained strongly
photonegative although nearly 13% of their sibs
had emerged from the redds. By day 12, the col-
lective negative photoresponse had weakened
considerably, and nearly 15% of the denied fry
were recorded then in the light compartments. By
the 16th day of emergence, when 90% of their sibs
had emerged, the percent of the fry recorded in
the light compartments reached a plateau. From
day 17 onward, 20-30% of the fry were seen in the
light compartments (15 of 19 tests), but the re-
sponse was more variable during the last day of

testing, two of the tests (LC and DC) providing
heterogeneous data. Interaction stemming from
territorial behavior was the most likely source of
variability, the light compartments being sporad-
ically defended by single fry attempting to drive
the others away.

Despite a decidedly negative photoresponse
during the first 10 tests (Figure 6, Table 3), in 8 of
these tests more fry held in darkness between
tests were recorded in the light compartments
than were those exposed to illumination between
tests (P<0.01). Because there was no significant
difference attributable to light history in sub-
sequent tests, novelty due to limited light experi-
ence may have stimulated exploratory behavior
during testing in fry held in darkness between
tests. When tested on days 15 and 17, the control

DAY OF EMERGENCE

Figure 6. — Change in photoresponse of coho salmon fry
held in baskets between tests. The histogram depicts the
concurrent rate of emergence of 584 sibling fry from the four
simulated redds.

TABLE 3. — Fry sightings in the light compartment of each of four choice
boxes containing 10 fry during photoresponse tests, expressed as a percen-
tage of possible sightings (400/test). Bracketed values are standard errors.

Light

Dark

Light

Dark

Light control

experimental

control

plus food

Day

(LE)

(DE)

(LC)

(DC)

(LC+F)

0.3(0.3)

2.5(0.6)

0.5(0.3)

1.8(0.5)

03(0.6)

2.0(0 8)

0.3(0.1)

0,5(0.3)

0.0

3.0(0.8)

13.3(1.6)

13.8(1.5)

32.2(2.4)

28.0(2.3)

•17

26.3(1.8)

22.3(1.9)

36.3(2.9)

30.0(2.3)

26.8(2.8)

'29.3(2.9)

26.5(2.3)

21.5(1,8)

23.3(1.9)

24.3(2.0)

24.8(1.4)

220

25.0(1.6)

23.8(1.9)

19.8(1.6)

'16.3(1.9)

27.0(1.8)

27.5(2.3)

'41.0(3.6)

'10.8(1.3)

'Heterogeneous data.

2Fed

In previous evening and 1 h prior to testing.

172

MASON: FEATURES OF EMERGING COHO SALMON FRY

groups LC and DC showed higher counts
(P<0.01) than did their experimental counter-
parts LE and DE tested on day 17 (Table 3).
Nonsignificant differences in subsequent tests
suggested that frequency of testing may have de-
pressed the magnitude of photoresponse change.

Fry receiving supplemental food (LC-l-F) made
scores similar to DC and LE groups (P<0.01)
when tested on day 18, but lack of homogeneity in
the data from one of the three tests performed
precluded further evaluation.

Light history and recent feeding did not sig-
nificantly affect response level when the four pre-
viously unfed groups were tested on day 20 {t =
1.3 with 158 df, P<0.020).

Differences in average length among the four
unfed groups of fry 1 day after the last tests were
not significant (Table 4, F = 0.33 with 3, 96 df)
but fish in the LE and DE groups weighed sig-
nificantly more and therefore had higher K val-
ues. Their heavier weight is attributed to feeding
on natural drift foods available only in the choice
boxes. The control group given supplemental food
from day 9 onward were significantly longer than
the other four groups of fry in average length {F
= 11.4 with 4, 122 df, P<0.01) and weighed con-
siderably more.

Table 4. — Average lengths, weights, and condition factors {K)
of samples of 25 coho fry used in the photoresponse experiment,
measured 1 day after final testing.

Fork length

Live weight

Treatment

(mm) SE

(mg)

Light experimental

38.38 ± 0.23

442.2

0.783

Dark experimental

38.29 ± 0.21

432.4

0.771

Light control

38.33 ± 0.19

391.6

0.695

Darl< control

38.17 ± 0.23

3996

0.719

Light control with

food supplement

39.60 ± 0.24

473.6

0.763

'K = W X 10^//-^ where W is weight in milligrams and/, is length in millimeters.

The average length of fry emerging from the
redds (Table 1) did not differ significantly from
that of the unfed siblings used in the photore-
sponse tests (Table 4). However, the emerging fry
weighed somewhat less than fry of the experi-
mental groups but more than those of the control
groups (X = 425.7 mg) and were in similar condi-
tion to the experimental groups (K = 0.766).

DISCUSSION

Emergence from these simulated redds in-
volved several differences from that reported by
Koski (1966) for natural redds of coho salmon.

Fry from individual natural redds took from 10 to
47 days {X = 35 days) to complete emergence
which peaked 8-10 days after first emergence,
and size of fry decreased as emergence proceeded.
In the simulated redds, duration of emergence
was 20-23 days peaking at 12-13 days and size
increased with time although yolk reserve re-
mained nearly constant. The physical structure of
the natural redd, particularly the proportion of
smaller particle sizes, restricted permeability
and impeded emergence. Low permeability re-
duced size of fry and increased mortality, later-
emerging fry and those failing to emerge that
were excavated from redds were emaciated,
weight loss indicating exhaustion of yolk prior to
emergence.

As yolk reserves remained fairly constant
throughout emergence from the simulated redds,
the larger, later-emerging fry probably developed
from larger eggs. Koski (1966) found that large
female spawners produced large fry at emer-
gence, but large size of progeny did not alleviate
physical hindrance to emergence, typifying the
majority of redds, leading to decreasing size of fry
as emergence progressed.

The strong upstream response shown by fry
emerging from the simulated redds is charac-
teristic of coho fry emerging in natural streams.
Apart from counteracting downstream transport,
upstream movement provides for the seeding of
upstream rearing areas unavailable to, or not
used by, spawners. The small but significant dif-
ference in yolk reserve between fry moving up-
stream or downstream may reflect, rather than a
minor difference in swimming ability, behavioral
differences associated with rising aggression,
onset of territoriality, and commencement of
feeding on the invertebrate drift.

The lack of preference for nocturnal emergence
is in contrast to findings for sockeye salmon; pink
salmon, O. gorbuscha; and chum salmon, O. keta,
fry which emerge primarily at night (Neave 1955;
Heard 1964). But like these other species, the
coho salmon fry retained a photonegative re-
sponse at emergence of potential survival value,
e.g. escape from predators. Stuart (1953) also re-
ported that fry of brown trout, Salmo trutta, re-
mained photonegative during their ascent in
simulated redds, even upon reaching positions
only 1 or 2 inches from the gravel surface. For
several days after emerging, fry of coho salmon
and cutthroat trout, S. clarki, will bolt back into
the gravel bed when disturbed (pers. obs.) and

173

FISHERY BULLETIN: VOL. 74, NO. 1

similar observations led Neave (1955) to com-
ment that migrating chum and pink salmon fry,
failing to reach the ocean in a single night, hide
during the day and resume migration at night-
fall. Hiding behavior disappears in coho salmon
fry at time of complete yolk absorption but is re-
tained for several days at high light intensities
(Hoar 1958); this suggests a threshold intensity
for the avoidance response which increases as the
alevin stage proceeds.

Concurrence between change in numbers of fry
observed in the choice chambers, a collective re-
sponse, and the accumulated number of emergent
sibs could reflect either a sudden shift in photo-
response of individual fry or gradual erosion of
the negative response occurring simultaneously
in all fry. The sudden shift alternative is best
supported by three patterns of behavior noted in
the choice chambers. Individual fry were ob-
served to spend considerable time in the light
compartment upon entering it, alternately
swimming about slowly and remaining locally
quiescent. Positions were commonly adopted with
the head projecting into the light compartment
(Figure 1), or entrance, and departure was rapid,
irrespective of the presence or absence there of
other fry until the last few days of testing when
aggression was observed (Figure 6).

Despite near depletion of vitellus at time of
emergence, the shift in photoresponse did not ap-
pear to be due to starvation because the response
was not altered significantly by feeding. This is of
interest as Smith (1952) reported marked meta-
bolic changes in rainbow trout, S. gairdneri, ale-
vins a few days prior to emergence, suggesting
that these physiological events signified the onset
of starvation. The change in photobehavior ap-
pears to be an ontogenetic behavioral change
normally associated with emergence from the
redd rather than one instigated by nutritional
deficiency, premature feeding, or light experi-
ence. It remains unclear as to whether or under
what conditions such stimuli can modify this
change significantly; however, under hatchery
conditions, Harvey (1966) found that sockeye
salmon fry took food 2 wk after hatching but that
emergence of fry from a simulated redd coincided
with complete yolk absorption some 3 wk later.
Heard (1964) noted that most emerging sockeye
salmon fry trapped from natural redds in an
Alaskan stream contained little or no yolk, re-
mained photonegative, and emerged primarily
during hours of darkness.

The timing of the photoresponse change rela-
tive to emergence and yolk reserves may vary
within common limits for most stream salmonids
and differences may reflect species-specific adap-
tions of value to fishery biologists. As in the fry
emerging from the simulated redds, the yolk re-
serve of coho salmon fry emerging from natural
redds averaged 7% (unpubl. data). Stuart (1953)
observed a definite change in photoresponse of S.
trutta when yolk neared depletion, and the photo-
response change was employed by Gray (1929b)
to denote the conclusion of incubation when
measuring the effect of temperature on alevin
size at time of yolk depletion. Woodhead (1957)
disagreed with Stuart as to the timing of the
photoresponse change in S. trutta, and asserted
that it occurred coincident with maximum activ-
ity of the alevin 15 days after hatching when yolk
reserve constituted 70% of the dry weight of the
fry. This considerable difference in timing re-
mains unresolved.

Denying the photoresponse fry streambed ex-
perience during the last few weeks of the alevin
stage had no apparent effect on the final size of
the fry, probably due to their advanced stage of
development prior to application of treatment dif-
ferences. Marr (1963, 1965) has shown that de-
velopmental efficiency is reduced by exposure to
natural light or lack of substrate contour which
stimulate locomotor activity at the expense of
growth. However, marked effects on locomotor ac-
tivity were only measurable until development
was 75-80% complete. The weight disparity be-
tween experimental and control groups of fry
(Table 4) which was the outcome of weight loss or
reduced weight gain is presumed to be an out-
come of reduced feeding opportunity.

In summary, the present results show that coho
salmon fry underwent a definite shift (sudden or
otherwise) from a strong to a weak negative
photoresponse. This shift was accompanied by a
positive response to water current leading to pre-
ferred movement upstream. The emerging fry
was an actively feeding animal yet to fill, or in
the process of filling, its air bladder, fed in the
gravel prior to emergence, and emerged when av-
erage yolk reserves declined to 7% of total dry
weight. In contrast to fry emerging from natural
redds (Koski 1966), later-emerging fry were
larger than those emerging earlier and may have
derived from larger eggs. Because first-emerging
fry held ecological advantage over later-emerging
fry in stream aquaria (Mason and Chapman

174

MASON: FEATURES OF EMERGING COHO SALMON FRY

1965), the timing of emergence and environmen-
tal conditions which modify it and the ecological
state of fry at emergence should be fruitful con-
siderations in future research.

ACKNOWLEDGMENTS

I am grateful to R. A. Bams and W. Percy
Wickett (Pacific Biological Station) for con-
structive criticism, and to D. W. Rimmer for tech-
nical assistance.

LITERATURE CITED

Ali, m. a.

1959. The ocular structure, retinomotor and photobe-
havioral responses of juvenile Pacific salmon. Can. J.
Zool. 37:965-996.
BAMS, R. A.

1969. Adaptations of sockeye salmon associated with
incubation in stream gravels. In T. G. Northcote (editor),
Salmon and trout in streams, p. 71-87. H. R. MacMillan
Lectures in Fisheries. Univ. British Columbia, Vancou-
ver, B.C.
DILL, L. M.

1969. The sub-gravel behavior of Pacific salmon larvae.
In T. G. Northcote (editor), Salmon and trout in streams,
p. 89-99. H. R. MacMillan Lectures in Fisheries. Univ.
British Columbia, Vancouver, B.C.

Gray, J.

1929a. The growth of fish. IL The growth-rate of the

embryo oi Salmo fario. J. Exp. Biol. 6:110-124.
1929b. The growth of fish. HI. The effect of temperature

on the development of the eggs of Salmo fario. J. Exp.

Biol. 6:125-130.

Harvey, H. H.

1966. Commencement of feeding in the sockeye salmon
iOncorhynchus nerka). Verb. Int. Ver. Theor. Angew.
Limnol. 16:1044-1055.

Heard, W. R.

1964. Phototactic behaviour of emerging sockeye salmon
fry. Anim. Behav. 12:382-388.
HOAR, W. S.

1958. The evolution of migratory behaviour among juve-

nile salmon of the genus Oncorhynchus. J. Fish. Res.

Board Can. 15:391-428.
KOSKI, K. V.

1966. The survival of coho salmon (Oncorhynchus kisutch)

from egg deposition to emergence in three Oregon

coastal streams. M.S. Thesis, Oregon State Univ.,

Corvallis, 84 p.
MARR, D. h. a.

1963. The influence of surface contour on the behaviour

of trout alevins S. trutta L. Anim. Behav. 11:412.
1965. Factors affecting the growth of salmon alevins and

their survival and growth during the fry stage. Assoc.

River Auth. Yearb., 1965:1-9.

Mason, J. C.

1969. Hypoxial stress prior to emergence and competi-
tion among coho salmon Iry. J. Fish. Res. Board Can.
26:63-91.

Mason, J. C, and D. W. Chapman.

1965. Significance of early emergence, environmental
rearing capacity, and behavioral ecology of juvenile
coho salmon in stream channels. J. Fish. Res. Board
Can. 22:173-190.
NEAVE, F.

1955. Notes on the seaward migration of pink and chum
salmon fiy. J. Fish. Res. Board Can. 12:369-374.
ROYCE, W. F.

1959. On the possibilities of improving salmon spawning
areas. Trans. North Am. Wildl. Conf 24:356-366.

Smith, S.

1952. Studies in the development of the rainbow trout
(Salmo irrideus). H. The metabolism of carbohydrates
and fats. J. Exp. Biol. 29:650-666.

STUART, T. A.

1953. Spawning migration, reproduction and younger
stages of loch trout (Salmo trutta L.). Freshwater
Salmon Fish. Res. 5:1-39.

WHITE, G. M.

1915. The behaviour of brook trout embryos from the time
of hatching to the absorption of the yolk sac. J. Anim.
Behav. 5:44-60.
WICKETT, W. P.

1952. Production of chum and pink salmon in a controlled
stream. Prog. Rep. Pac. Coast Stn., Fish. Res. Board
Can. 93:7-9.
WOODHEAD, P. M. J.

1957. Reactions of salmonid larvae to light. J. Exp. Biol.
34:402-416.

175

FEEDING BEHAVIOR, FOOD CONSUMPTION, GROWTH, AND
RESPIRATION OF THE SQUID LOLIGO OPALESCENS RAISED

IN THE LABORATORY

Ann C. Hurley^

ABSTRACT

The squid Loligo opalescens was raised in the laboratory to a maximum age of 100 days on a diet of
Artemia nauplii and adults. Newly hatched squid (2.7 mm mantle length) readily attacked Artemia
nauplii (length 0.7 mm), Artemia adults (length 5 mm), copepods (length 1 mm), and larval fish
(length 4 mm). Feeding rates varied between 35 and 80% of squid body weight per day. Growth rate
was highly variable in different individuals, ranging from 0.5 to nearly 4.5 mm mantle length per
month. Respiration rates were obtained at 15°C for squid of three different ages and at 10°, 15°, and
20°C for 1-day-old squid.

The squid Loligo opalescens Berry is a common
pelagic predator off the west coast of North
America from British Columbia to Baja Califor-
nia. Because a fishery exists for this species, con-
siderable information is available concerning
adults in the spawning schools (Fields 1965), but
little is known about the early life stages. In a
paper on larval squid abundance off California,
Okutani and McGowan (1969) found few L.
opalescens in their samples; and McGowan (1954)
reported that despite considerable effort he could
not catch newly hatched L. opalescens over the
spawning grounds.

To obtain information on the early life history, I
reared L. opalescens in the laboratory. Several
workers have succeeded in rearing decapod
cephalopods, but all of the species they used tend
to be closely associated with the bottom (Choe
1966, three species of Sepia, the squid Sepioteu-
this lessoniana, the sepiolid Euprymna berry i;
LaRoe 1971, S. sepioidea; Boletzky et al. 1971,
four species ofSepiola and two species of Sepietta;
Arnold et al. 1972, the sepiolid E. scolopes). At-
tempts to raise pelagic species such as Loligo
opalescens have met with little success (Fields
1965; Arnold et al. 1974). Workers have attrib-
uted their failure to lack of food and to infec-
tions. I describe here a simple technique for rear-
ing early stages of L. opalescens and present data
on the growth, respiration, and food requirements
of L. opalescens reared for 100 days in the
laboratory.

'Scripps Institution of Oceanography, University of Califor-
nia, La Jolla, CA 92093.

MATERIALS AND METHODS

Five groups (referred to as groups 1 through 5)
of squid have been reared, three ( 1 through 3) of
which will be described in detail in this report.
Eggs were collected from a water depth of 20 m off
La Jolla, Calif, and were maintained in circulat-
ing seawater at about 13°C. The young squid
were transferred to the rearing tanks after they
had hatched. Fields (1965) and McGowan (1954)
have described the methods of egg deposition and
structure of the egg masses in detail.

The rearing tanks were cylindrical (122 cm
diameter, 36 cm deep) and made of black fiber
glass. Tanks were illuminated by fluorescent
lights which had a cycle of 18 h light, 6 h dark.
During the dark period, lights in other rooms of
the aquarium building provided a source of dim
light. The tanks were immersed in water baths
which kept the temperature within the tanks be-
tween 15° and 17°C. Squid were transferred to the
rearing tanks with a beaker. Squid in groups 2
and 3 were counted during transfer. In group 1,
the number of squid was estimated after the
squid were in the tank. Groups 1 and 2 began
with 300 squid; group 3 began with 250. The
water in the tanks was noncirculating. Each tank
was aerated by a gently bubbling air supply. The
squid in group 1 were transferred to a holding
tank on day 62 and on day 76, and on each day
their tank was drained, cleaned, and refilled.
Tanks 2 and 3 were both similarly cleaned on day
49. Dead food was removed from the bottom of all
tanks with a siphon, and small amounts of seawa-
ter were added to maintain a constant volume.

Manuscript accepted September 1975.
FISHERY BULLETIN: VOL. 74. NO. 1, 1976.

176

HURLEY: LOLIGO OPALESCENS RAISED IN THE LABORATORY

During the first 4 wk the squid (groups 1
through 3) were fed newly hatched brine shrimp,
Artemia salina, nauplii which were kept at densi-
ties ranging from 1 to 20 nauplii/ml. After this
time, small adult brine shrimp were added (aver-
age length 5.4 mm; range 2.5 to 8.0 mm) and were
the major source of nourishment for the remainder
of the rearing period. In groups 4 and 5, small
adult Artemm as well as nauplii were used as food
during the first 4 wk.

Squid were measured using an optical microm-
eter on a dissecting microscope. Measurements
are of dorsal mantle length (measured dorsally
from the tip of the tail to the farthest anter-
ior point on the mantle). Mantle length is less
variable than a measurement of total length,
which depends upon the degree of stretch of the
arms and tentacles. To make possible conversions
to total length, measurements were made of both
dorsal mantle length (ML) and total length (to
tips of arms, not tentacles) (TL) on 35 juvenile
animals, and the average ratio ML/TL was 0.62
± 0.014 ( ±2 SE). Measurements are all on freshly
dead unpreserved animals. For weight measure-
ments, squid were rinsed in distilled water and
oven dried at 60°C to a constant weight.

Respiration measurements were made using a
Warburg constant volume respirometer with
respiration vessels kept at constant temperature
in a water bath. The respiration vessels contained
from 2 to 30 squid and were kept in constant mo-
tion by gentle shaking.

Estimates were made of the number of squid
surviving at intervals throughout the study. The
number of squid alive on any day was the average
of three counts taken of live animals in the tank.

Daily observations were made of the feeding
behavior of the squid. At various times through-
out the day, a squid was selected and observed for
about 5 min. The number of feeding attempts and
successful captures of prey were recorded.

RESULTS

Survival

Mortality in all of the tanks was initially high
(Figure 1). This is similar to what LaRoe (1971)
found in rearing Sepioteuthis sepioidea. LaRoe
speculated that the high initial mortality was due
to insufficient quantities of food. This probably
was not the case in my studies, as a large amount

lOOi

^ 60

ff 50

^ 40

30-

20-

D GROUP I
A GROUP 2
O GROUP 3

V& ''*^'^^'** ®-o-ocp-oo^

^-.

2°=^

30 40

50
DAYS

100

Figure l. — Estimated percent survival of Loligo opalescens in
the rearing tanks. Group 1 started with 300 squid; group 2, with
300 squid; and group 3, with 250 squid.

of food was continually available at this stage.
Some of this mortality could have been caused by
squid which did not initiate feeding. Fields (1965)
found that L. opalescens which did not appear to
be feeding lived up to 10 days and still had some
internal yolk reserves left at the end of this time.
From 30 to 60 days mortality was low, but after
60 to 70 days mortality again increased. It is pos-
sible that the brine shrimp did not provide an
adequate diet for squid older than 60 days.

Feeding Behavior

Attack

The attack of a young L. opalescens is similar to
that described for adult Loligo (Fields 1965),
Sepioteuthis (LaRoe 1970), and Sepia (Messenger
1968). Messenger divided the Sepia attack into
three motor patterns: attention, positioning, and
seizure. These three patterns may also be used to
describe the attack of yoimg L. opalescens. Dur-
ing attention, the squid orients toward a particu-
lar prey. The arms and tentacles are extended in
front of the squid and form a tight cone which is
pointed toward the prey. Color changes such as
those noted for Sepioteuthis (LaRoe 1970) and
Sepia (Messenger 1968) were not observed.

177

FISHERY BULLETIN: VOL. 74, NO. 1

After the squid oriented toward a particular
prey, it approached the prey until it was within
attacking distance (positioning). This distance
was not constant. At times there was no clear
separation between the attention and positioning
patterns. LaRoe (1970) suggested that the posi-
tioning approach is an example of an aggression-
fear conflict. This appears to be the case in Loligo.
The young squid would sometimes flee rapidly
after closely approaching a large prey.

The prey was usually captured with the tenta-
cles (seizure), although occasionally the arms
alone were used. The arms were used to maneu-
ver the food toward the mouth. At times a new
attack began while the squid was holding other
prey in the arms.

LaRoe (1970) reported that for Sepioteuthis
sepioidea physical fights over food were rare. This
was not true for young L. opalescens. Fighting
between squid was never observed when prey was
small (brine shrimp nauplii), but if the prey was
large and could not be completely enclosed within
the arms, other squid would often chase the one
which caught the food and try to take the food
away from it. Often several (in one case, four)
squid held on to the captured prey and all fed on
it. The prey would be tugged about until one
squid pulled it away from the others. This be-
havior occurred even when there was an abun-
dance of prey in the tank. This attack on captured
prey at times allowed small squid to eat larger
prey organisms than they could normally subdue
alone.

Prey Selection

Unlike Sepioteuthis (LaRoe 1971), young L.
opalescens were not extremely selective as to the
type and size of prey they would attack. Within a
few days after hatching, the young Loligo (2.7
mm ML) readily attacked Artemia nauplii (0.7
mm long), Artemia adults (5 mm long), copepods
(1 mm long), and larval fish (4 mm long). Occa-
sionally, squid attacked and ate dead prey (e.g.,
6.ea.6. Artemia dropped into the tank), but usually
the food had to move before it was attacked. An
exception to this was that the squid attacked fish
larvae which appeared to be motionless in the
water.

When the squid were 17 days old, nine squid
from group 2 were placed in a small cylindrical
container (8 liters of water) to determine whether
a food size preference existed in Loligo. The food

used was Artemia nauplii (0.6 to 0.8 mm long) at
10/ml and small adult Artemia (2 to 4 mm long)
at 0.2/ml. After the squid were added, I recorded
the number of attacks until a prey was captured
and the type of prey being attacked. If no prey
was captured in 20 min, I selected another squid.
At this age, the squid attacked both large and
small prey. During the 164 min of observation, 23
nauplii were attacked (9 actually captured) and
30 adults were attacked (8 actually captured).
These results are different from those given for
Sepioteuthis sepioides (LaRoe 1971). That squid
only attacked food species in a very limited size
range. Within several days, Sepioteuthis would
cease to attack the prey it had previously eaten
and would only attack larger prey. This seemed to
occur when the squid were 1 to IV2 times as large
as their prey. Although Loligo captured both
large and small prey with about equal frequency,
a preference may exist for larger prey as their
density in the container was much lower.

An experiment was run with group 1 when the
L. opalescens were 49 days old. In this case the
choice was between two different prey species of
approximately the same size. Two thousand
2-day-old chub mackerel, iScom6erJaponicus, lar-
vae were added to one of the rearing tanks where
the squid had been feeding on Artemia adults.
There were approximately 2,000 Artemia in the
tank. The same method was used to record feeding
as in the previous experiment. Observation time
in this case was 69 min. The squid showed a high
incidence of attacks on fish larvae (52 attacks, 6
captures) even though the success rate was much
lower than when attacking Artemia (4 attacks, 3
captures). This may indicate a preference for fish
larvae, but without further experiments it is im-
possible to say whether this is true.

Feeding Success

The ability of the squid to successfully complete
an attack sequence depended on the size and
species of prey and the age and experience of the
young squid. Figure 2 is a record of the percent of
successful attacks on Artemia nauplii as a func-
tion of the age of the squid. Each point is an aver-
age from the squid observed during that day. The
number of squid observed per day ranged from 5
to 11, with the total daily observation time rang-
ing from 25 to 55 min. The attack efficiency in-
creased with the age of the squid, but a number of
prey were still being lost even after 3 wk. LaRoe

178

HURLEY: LOLIGO OPALESCENS RAISED IN THE LABORATORY
100.

90-

40-

20-

• •

— I 1

10 IB

DAYS AFTER HATCHING

—I
30

Figure 2. — Percent of attacks on Artemia nauplii which were
successful as a function of the age of the squid.

(1970) found that for Sepioteuthis , the majority of
the prey were lost because the squid were unable
to judge the attack distance. In my experiments,
most unsuccessful attacks occurred because the
prey managed to escape after being initially
struck. Some of the variability in success rates
may have been due to different motivational
states of the squid.

Feeding Rates

Several methods were used to determine the
food ration of the developing squid. When the
squid fed on nauplii, feeding rates were deter-
mined at irregular intervals by choosing a squid
and watching it for 5 min to determine the
number oi^ Artemia nauplii consumed during this
period. All of the observations accumulated dur-
ing a given week were combined. For each week, I
calculated the food eaten over a 24-h and 18-h
feeding period. The squid captured prey when the

Table l. — Estimated feeding rates (percent body weight eaten
{)er day) of squid in rearing tanks. Each value is average for all
values for a given week. Values through week 4 are based upon
observed short-term feeding rates on Artemia nauplii and are
given for assumed 18- and 24-h feeding periods. Subsequent
values are based on counts of Artemia adults consumed in tanks
1 and 2.

jplil

Adults

Week

18 h

24 h

Tank 1

Tank 2

—

4
5
6

—

overhead lights were off, but it was not possible to
establish how much was eaten. When adult Ar-
temia was the primary source of nourishment,
record was kept of the approximate number of
food organisms introduced to the tank and their
average weight. There is some error introduced
here because some of the brine shrimp died and
were not consumed. The average weight of the
squid during each week was obtained from the
growth data and length-weight relationships pre-
sented in the next section. Average weight of
Artemia adults was 0.3 mg (obtained from six
random samples of 10 to 20 individuals each) and
average weight of nauplii was 0.002 mg (John R.
Hunter pers. commun.). Food consumption is
shown in Table 1.

One short-term experiment was performed to
examine the feeding rate of 36-day-old squid on
yolk-sac larval anchovies. Five squid were placed
with 100 anchovy larvae in 8 liters of water and
were left for 285 min. At the end of this period 58
larvae had been eaten. This gives a feeding rate
of 2.4 larvae/squid • hour Theilacker and Lasker
(1974) gave the average weight of a larva of this
size as 0.022 mg. Using this information and the
average weight of the squid, a feeding rate of
0.028 mg anchovy/mg squid h is obtained.

Growth

Since the number of squid being reared was
small, specimens were not sacrificed for growth
measurements alone. Every time a squid died, it
was immediately measured. These measure-
ments constitute the majority of the points on the
growth curve shown in Figure 3. The points indi-
cated by the x's are measurements which were
made on squid that had been selected while alive

179

FISHERY BULLETIN: VOL. 74, NO. 1

Figure 3. — Size data ior Loligo opalescens. A dot denotes mea-
surement made on squid which had died, and x denotes mea-
surement made on squid that had been selected while alive to
give an indication of the size range of individuals in the tanks.
For days 1, 17, and 22, the numbers of squid measured, means,
and ranges are given. The upper solid line gives a constant
growth rate of 4.5 mm/mo. The lower one gives a rate of 0.5
mm/mo.

to give an indication of the full size range of squid
in the tank. Since the squid were not randomly
sampled during this time, Figure 3 cannot be
taken to give an average growth rate for the
population, but it does give an indication of the
range of growth rates. There was a large differ-
ence in the rates of growth of individuals. Maxi-
mum growth rates were nearly 4.5 mm/mo (upper
line in Figure 3). Minimum growth rates were 0.5
mnVmo (lower line in Figure 3).

The linear regression equation for the log
length-log weight relationship for the developing
squid is log weight (mg) = -1.22 + 2.37 log
length (mm) with little scatter around the re-
gression line.

Respiration

Measurements were taken of the oxygen con-
sumption of young L. opalescens using a Warburg
respirometer and a constant temperature water
bath. Measurements were taken at 15°C for squid

of three different ages and at 10°, 15°, and 20°C
for 1-day-old squid (Table 2). Average oxygen
consumption values are as follows: 1 day, 10°C,
1.5 /ul 02/mg squid h; 1 day, 15°C, 2.5 /aI 02/mg
squid -h; 1 day, 20°C, 3.5 /ul 02/mg squid h; 3 wk,
15°C, 3.5 )ul 02/mg squid h; 8 wk, 15°C, 3.7 /xl
02/mg squid h. These measurements may be ar-
tificially high because of the crowding which oc-
curred in the small respiration vessels. It was ob-
served, however, that the oxygen consumption
tended to decrease (at a given temperature) with
increasing number of animals present in the
same vessel. It is possible that these lower rates
occurred because some of the animals became
moribund in the crowded conditions. But this is
not likely, since the respiration rates remained
constant over the course of the 2-h experiments.

To compare these measurements to those made
by other investigators, conversion factors had to
be obtained to transform dry weight to wet
weight. The ratio wet/dry was calculated for nine
juvenile squid and gave a mean of 5.4 ± 0.21 ( ±2
SE). Wet weights were calculated by placing the
squid on the weighing pan, blotting it with filter
paper, weighing it at measured time intervals,
and extrapolating the line obtained to zero time.

The previous rates expressed in terms of wet
weight are: 0.28, 0.46, 0.65, 0.65, and 0.69 ix\
02/mg squid h. These values are similar to those
obtained by LaRoe (1971) for 2- and 6-day-old
Sepioteuthis sepioidea (0.64 ixVmgh at 23°C) and
with the figure of 0.60 fiVmgh for adult L. pealei,
calculated from data in Redfield and Goodkind
(1929).

Table 2. — Oxygen consumption rates for Loligo opalescens.
Respiration vessels had a volume of 18 ml and contained approx-
imately 5 ml seawater. The duration of the experiments was 2 h.

Temp.
(°C)

Age of
squid
(days)

Number of
squid/vessel

Range of oxygen consumption
(/jl Oj/mg squid (dry wt) h)

10
15
20
15
15

3
3
3
1
2

10-30
8-25

10-21

10
2-3

1.4-1.6

2.1-3.6

3.2-3.8

3.5

3.5-3.9

DISCUSSION

It is extremely difficult to assess the role which
an animal such as L. opalescens plays in the
California Current ecosystem. Estimates of popu-
lation size of adults are very poor because of the
difficulties involved in sampling large active

180

HURLEY: LOLIGO OPALESCENS RAISED IN THE LABORATORY

animals. Fisheries statistics are not particularly
helpful because the catches come mainly from a
few locations. It has been possible to get some
field information on the diet of the adult squid
(Fields 1965) but these data are completely
lacking on such necessary information as feed-
ing rates.

It appears to be equally difficult to obtain in-
formation on young L. opalescens from field sam-
ples. The young squid have well-developed eyes
and are very sensitive to vibrations. Therefore,
even the young are likely to be able to avoid
many nets. Okutani and McGowan (1969) pub-
lished data on the abundance of young L. opales-
cens (size range 3.5 to 7 mm dorsal ML) taken
in net tows during the California Cooperative
Oceanic Fisheries Investigations cruises in 1954
to 1957. In their report, however, they emphasized
the problems involved in sampling the young
squid and stressed that the abundances given
probably should only be used to compare relative
abundances of different species. They found that
L. opalescens was the third most abundant
species of larval squid present in their samples,
but that its abundance was quite low when com-
pared to the most common fish larvae present
(e.g., 0.008 times the abundance of northern an-
chovy, ^n^raw/is mordax).

If the role of a young L. opalescens as a predator
is to be evaluated, it is necessary to know the type
of prey which it eats. Fields (1965) has deter-
mined the diet of the adult squid from an exami-
nation of stomach contents, but to my knowledge
no one has done a similar study on the very small
squid. From the laboratory results presented in
this paper, it appears that young L. opalescens
must be considered as predators on a wide range
of prey types and prey sizes. They are capable of
preying on species ranging in size from 0.7 to 7
mm and they readily attack prey species ranging
from brine shrimp adults and nauplii to copepods
and larval fish. McGowan (pers. commun.) has
found that they also successfully attack the mysid
Metamysidopsis elongata.

It is also possible to use the data presented here
to estimate a feeding rate for the young squid.
The respiration data can be used to calculate the
amount of food a young squid would need to sus-
tain itself. The respiration rate of the squid in the
rearing tanks can be taken as 3 ix\ 02/mg dry
wth. An average value for the caloric value of
oxygen consumed is 5 x 10"^ cal//il of O2. There-

fore, a newly hatched squid (2.7 mm ML, weigh-
ing 0.625 mg) would use 0.22 cal for respiration
alone in 24 h.

It is possible to determine how many prey items
of different types of prey would satisfy this re-
quirement. A newly hatched Ar^emia nauplius is
the equivalent of 0.0096 cal (John Hunter pers.
commun.). Therefore, a newly hatched squid
would need 23 Artemia nauplii per day. If the
squid were instead feeding on newly hatched
northern anchovies, it would need a total of 2 an-
chovy larvae per day (using a value of 5 cal/mg,
weight of larva = 0.022 mg; Theilacker and
Lasker 1974). Similar calculations can be made
for older squid. A squid 7 mm ML (~2 mo old, 6
mg) would consume 225 nauplii or 20 anchovy
larvae simply to meet its metabolic needs. The
actual amount of food consumed per day was ap-
preciably more than this, averaging about 50% of
body weight per day. At this rate, a newly
hatched squid would consume 150 nauplii or 14
anchovy larvae per day, while a 7-mm squid
would consume 1,500 nauplii or 135 anchovy lar-
vae per day.

Data on feeding rates and abundance could be
used to calculate the impact that young squid
might have on populations of potential prey
items, but before such calculations can be mean-
ingful, more information must be known about
the ability of the squid to locate sources of food.
Loligo opalescens was only one hundredth as
abundant as the most common fish larvae (Oku-
tani and McGowan 1969). But with feeding rates
of 15 to 135 larvae per day, young squid could
potentially have a large impact on such popula-
tions if they concentrate on this type of food and if
they have effective means of finding such prey.
Laboratory observations indicate that larval fish
may be a preferred food, and the squid do occur in
areas where larval fish are common. Okutani and
McGowan found that L. opalescens was most
common in the upper 40 m, and this is the
stratum where the highest abundance of north-
ern anchovy larvae occur (Ahlstrom 1959).

ACKNOWLEDGMENTS

I thank J. Hunter, R. Lasker, and D. Lange for
help during this work. This study was done while
I was on a NOAA Associateship at the Southwest
Fisheries Center, National Marine Fisheries Ser-
vice, NOAA, La Jolla, Calif.

181

LITERATURE CITED

Ahlstrom, E. H.

1959. Vertical distribution of pelagic fish eggs and larvae
off California and Baja California. U.S. Fish. Wildl.
Serv., Fish. Bull. 60:107-146.

Arnold, J., W. Summers, D. Gilbert, R. Manalis, N. daw,
AND R. Lasek.

1974. A guide to laboratory use of the squid Loligo
pealei. Mar Biol. Lab., Woods Hole, 74 p.
Arnold, J. M., C. T. Singley, and L. D. Williams-Arnold.

1972. Embryonic development and post-hatching survival
of the sepiolid squid Euprymna scolopes under laboratory
conditions. Veliger 14:361-364.
BOLETZKY, S. VON, M. V. VON BOLETZKY, D. FROSCH, AND V.
GATZL

1971. Laboratory rearing of Sepiolinae (Mollusca:
Cephalopoda). Mar Biol. (Berl.) 8:82-87.
CHOE, S.

1966. On the eggs, rearing, habits of the fry, and growth of
some Cephalopoda. Bull. Mar. Sci. 16:330-348.
FIELDS, W. G.

1965. The structure, development, food relations, repro-
duction, and life history of the squid Loligo opalescens
Berry. Calif. Dep. Fish Game, Fish. Bull. 131, 108 p.
LaROE, E. T.

1970. The rearing and maintenance of squid in confinement

FISHERY BULLETIN: VOL. 74, NO. 1

with observations on their behavior in the laboratory.
Ph.D. Thesis, Univ. Miami, Coral Gables, Fla., 136 p.
1971. The culture and maintenance of the loliginid squids
Sepioteuthis sepioidea and Doryteuthis plei. Mar Biol.
(Berl.) 9:9-25.

MCGOWAN, J. A.

1954. Observations on the sexual behavior and spawning
of the squid, Loligo opalescens, at La Jolla, Califor-
nia. Calif. Fish Game 40:47-54.

Messenger, J. B.

1968. The visual attack of the cuttlefish, Sepia officinalis.
Anim. Behav. 16:342-357.

OKUTANI, T., AND J. A. MCGOWAN.

1969. Systematics, distribution, and abundance of the
epiplanktonic squid (Cephalopoda, Decapoda) larvae of
the California current, April, 1954-March, 1957. Bull.
Scripps Inst. Oceanogr., Univ. Calif. 14, 90 p.

Redfield, a. C, and R. GOODKIND.

1929. The significance of the Bohr effect in the respiration
and asphyxiation of the squid, Loligo pealei. J. Exp. Biol.
6:340-349.

Theilacker, G. H., and R. Lasker.

1974. Laboratory studies of predation by euphausiid
shrimps on fish larvae. In J. H. S. Blaxter (editor), The
early life history of fish, p. 287-299. Springer- Verlag,

N.Y.

182

CONTRIBUTION OF THE NET PLANKTON AND NANNOPLANKTON

TO THE STANDING STOCKS AND PRIMARY PRODUCTIVITY IN
MONTEREY BAY, CALIFORNIA DURING THE UPWELLING SEASON

David L. Garrison^

ABSTRACT

Net plankton and nannoplankton standing stocks and primary production were measured in Mon-
terey Bay, Calif, from January through August 1972. Throughout the period of seasonal upwelling,
the phytoplankton stocks were dominated by net plankton. Both fractions showed seasonal changes:
the net plankton concentrations increased dramatically during upwelling, while nannoplankton
concentrations were decreased. Nannoplankton growth rates exceeded net plankton rates at in-
cubator light levels; however, at higher in situ light levels near the surface, this relationship ap-
peared to be reversed.

Nannoplankton decreases may have been related to their selective removal from the area of up-
welling by horizontal advection or selective grazing on the nannoplankton fraction. Net plankton
dominance during upwelling has been related to their higher growth rates when populations are
retained in shallow nutrient-rich nearshore waters.

Frequently, phytoplankton are divided into two
size classes, depending on whether they are re-
tained by fine mesh nets (net plankton) or pass
through the mesh (nannoplankton). The inade-
quacy of net collections for estimating standing
stocks or production is clear. The standing stocks
of the two fractions and their relative contribu-
tions to primary productivity, however, are less
well-known. The size distribution, which may be
environmentally controlled (Semina 1972; Par-
sons and Takahashi 1973), is an important fea-
ture of the phytoplankton populations because
the size of the primary producers may affect the
length and efficiency of pelagic food chains
(Ryther 1969; Parsons and LeBrasseur 1970). The
purpose of this study was to determine the rela-
tive importance of the two fractions during the
upwelling season in Monterey Bay, a neritic envi-
ronment of the California Current system.

Most previous studies reported that the nan-
noplankton fraction usually exceeds the net
plankton fraction, often accounting for 80 to
100% of the standing stocks and primary produc-
tion (e.g., Steeman Nielsen and Jensen 1957;
Holmes 1958; Yentsch and Ryther 1959; Kawa-
mura 1961; Holmes and Anderson 1963; Teixeira
1963; Gilmartin 1964; Saijo 1964; Anderson 1965;

'Moss Landing Marine Laboratories, Moss Landing, CA
95039; present address: Coastal Marine Laboratory, University
of California, Santa Cruz, CA 95064.

Manuscript accepted September 1975.
FISHERY BULLETIN: VOL. 74. NO. 1. 1976.

Saijo and Takesue 1965; Malone 1971a, c; Parsons
1972; McCarthy et al. 1974). Only a few authors
reported net plankton dominated communities
(Digby 1953; Subrahmanyan and Sarma 1965). It
is difficult to compare these studies, however, be-
cause mesh sizes of 22 to 110 ^im have been vari-
ously used to separate the net plankton and nan-
noplankton fractions.

The nannoplankton fraction may show little
seasonal fluctuation, while the net plankton
shows pronounced seasonal trends with periods of
abundance corresponding to increased water
temperatures (Yentsch and Ryther 1959), peak
periods of primary production (Subrahmanyan
and Sarma 1965), or seasonal upwelling (Malone
1971c). Malone (1971a) reported higher net:
nanno ratios for standing stocks and production
in neritic environments as compared with oceanic
areas and pronounced onshore to offshore lower-
ing of the ratio in the California Current region
during upwelling (Malone 1971c). The growth
rate (as indicated by the assimilation ratio
= mg C mg Chi a"^ h"^) of the nannoplank-
ton fraction is greater than that of the net plankton
fraction (Yentsch and Ryther 1959; Saijo and
Takesue 1965; Malone 1971a, c).

Arguments presented for the predominance of
net plankton or nannoplankton in a given envi-
ronment relate cell area to volume ratios (Malone
1971a, c; Eppley 1972; Parsons and Takahashi
1973). There is a general relationship between

183

FISHERY BULLETIN: VOL. 74, NO. 1

cell size and the ability to take up nutrients
(Dugdale 1967; Eppley et al. 1969; Eppley and
Thomas 1969). Large species generally have
higher half saturation constants (Kg) and may
have higher maximum uptake rates (Vmax)'
whereas small species have low^er Kg and V^ax
(Dugdale 1967). Maximum net plankton grovi'th
rates are favored at higher ambient nutrient
concentrations while nannoplankton reach their
maximum growth rates at lower ambient nutrient
levels. There is also a direct relationship of in-
creasing cell size (or chain length) with increas-
ing sinking rates (Smayda 1970), and larger cells
and chain formers tend to be aggregated in
areas of upward advection, while motile or posi-
tively buoyant cells tend to be concentrated in
areas of downward advection (Stommel 1949).
Net plankton will have a longer residence time
in the euphotic zone and concentrate in areas
of upwelling, while the nannoplankton (if the
population is primarily motile flagellates) will be
concentrated in areas of downwelling.

Parsons and Takahashi (1973) related the
growth rate (/u,) to physiological characteristics of
the cell (maximum grovvi:h rate, half saturation
constants for nutrients and light, and sinking
rates) and environmental conditions (incident
radiation, extinction coefficients, mixed layer
depth, and upwelling rates) and used the rela-
tionship to explain characteristic phytoplankton
cell size in a number of environments. Recently,
Laws (1975) expanded the Parsons and Taka-
hashi model and showed that under certain light
conditions the decreasing respiration rate with
increasing cell size may regulate the growth rate
of large versus small cells.

The effect of grazing on the net:nanno ratios
and, conversely, the size of the primary producers
on food chains have not been well documented.
Grazing may ultimately control net plankton
stocks (Malone 1971c; Ryther et al. 1971) and de-
termine the lower net:nanno standing stock ra-
tios in oceanic as opposed to neritic areas (Malone
1971a). Grazing has been suggested as the pri-
mary cause for failure of phytoplankton stocks
to develop in otherwise favorable waters (Mc
Allister et al. 1960; Strickland et al. 1969).
Shorter food chains have been shown for some
clupeid fishes which feed directly on the large
phytoplankton species (e.g., Bayliff 1963; Rojas
de Mendiola 1969; Dhulkhed 1972) and for her-
bivorous euphausids in the diatom-rich antarctic
region (Marr 1962). The general argument for

larger phytoplankton cells resulting in shorter,
more efficient food chains may not always apply
to the smaller grazers, as Parsons and LeBras-
seur (1970) have reported on selective feeding re-
lated to cell shape.

Previous studies have been made on the hydro-
graphic seasons in Monterey Bay and their rela-
tionship to the seasonal phytoplankton blooms
(Bolin and Abbott 1963; Abbott and Albee 1967).
Malone (1971c) reported the seasonal variability
of the net plankton and nannoplankton in the
California Current, which included one deep sta-
tion on the edge of Monterey Bay. The present
study was part of a monthly sampling program
conducted by Moss Landing Marine Laboratories
to provide information on the hydrographic con-
ditions and plankton populations in Monterey
Bay, particularly from the extensive shallow
areas of the bay. Although it was not possible to
carry this study through a complete seasonal cy-
cle, information is presented for the upwelling
period, when seasonal blooms of phytoplankton
appear in Monterey Bay.

MATERIALS AND METHODS

Measurements of primary productivity and
phytoplankton standing stocks were made at sta-
tions 3 and 8 for the period January through Au-
gust 1972 and at station 15 for the period June
through August 1972 (Figure 1). The stations
were located over the Monterey Submarine Can-
yon at depths of 110, 240, and 718 m, respectively.
Samples were taken monthly during hydrograph-
ic and plankton cruises conducted by Moss Land-
ing Marine Laboratories and, occasionally, be-
tween these periods on instructional cruises.
Sampling times varied between cruises but fell
between 0700 and 1100 h.

Samples were collected with 5-liter Niskin
water sampling bottles from depths correspond-
ing to 100, 50, 25, 10, and 1% light penetration
levels as measured with a submarine photometer
or calculated using the relationship: depth of 1%
light = 3.5 X Secchi disk (Silver and Hansen
1971a). Hydrographic parameters (salinity, °L;
temperature, °C; O2) and nutrients (PO4, NO3,
NO2, NH3, Si02) were samples at standard depths
(Broenkow and Benz 1973).

Primary productivity was measured using the
carbon-14 method (Steeman Nielsen 1952). For
each depth two light and one dark bottles were
innoculated with 5 or 10 fxCi of Naa^^COg. The

184

GARRISON: NET PLANKTON AND NANNOPLANKTON IN MONTEREY BAY

MONTEREY
BAY

37* N

40'

36'3r

Figure l. — Location of stations in Monterey Bay. Broken
lines indicate the position of the 100-fathom (183-m) contour
line.

samples were incubated immediately after collec-
tion for 3 to 4 h in a shipboard incubator (Doty
and Oguri 1958) using Luxor Magnalux fluores-
cent lamps^ (approx. 0.06 langley min"^). Neutral
density filters of 50, 25, 10, and 1% transmittance
were used on subsurface samples.

The net plankton and nannoplankton fractions
were separated by passing the samples through a
22-/xm Nitex-net filter (net plankton) and then a
Gelman, type A glass-fiber filter having 0.3-/>tm
pore size (nannoplankton). Both filters were
washed with approximately 20 ml of freshly
filtered seawater and placed directly in scintilla-
tion fluor for counting at a later time.

All samples were counted for at least 10 min
with a Nuclear Chicago (Unilux II) scintillation
counter. Carbon uptake was calculated as out-
lined in Strickland and Parsons (1968). Since
Malone (1971b) reported no diurnal periodicity in
assimilation ratios in the California Current re-
gions, daily production was estimated by using

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

the sunrise to sunset interval as the day length
and multiplying by the hourly production rates
that were determined during the first part of the
day.

Phytoplankton standing stocks were measured
as chlorophyll a by using the fluorometric method
of Holm-Hansen et al. (1965). The Turner fluoro-
meter (model 111) was calibrated using the spectro-
photometric method for chlorophyll a as outlined
by Strickland and Parsons (1968). The two size
fractions were separated by taking two replicate
samples from each depth and passing one
through a Gelman glass-fiber filter (total chloro-
phyll) while the other sample was filtered through
22 ;um Nitex-net filter and then a glass-fiber filter
(nannoplankton). Both filters were immediately
frozen, stored in the dark, and analyzed within a
month after collection. Net plankton was calcu-
lated as the difference between total chlorophyll
and nannoplankton chlorophyll.

Productivity and chlorophyll a values deter-
mined for the discrete samples were integrated to
the depth of the 1% light level by trapezoidal ap-
proximation. Carbonxhlorophyll a ratios vary
widely and depend on light and nutrient condi-
tions. For most of the study, nutrient levels were
high and a C:Chl a ratio of 40 was used to convert
chlorophyll a to carbon biomass (Lorenzen 1968;
Eppley et al. 1970; Eppley et al. 1971). Phyto-
plankton growth rate and standing stock dou-
bling time were calculated using exponen-
tial growth expression.

RESULTS

In January, the weak thermal gradient in the
upper 50 m (Figure 2) is indicative of the David-
son Current period, when the subsurface counter-
current extends to the surface and flows north-
westward on the inshore side of the California
Current (Reid et al. 1958; Bolin and Abbott 1963;
Smethie 1973). Rising isotherms and nitrate iso-
pleths from February through May indicate up-
welling over the Monterey Submarine Canyon.
After May there was a slacking or an end to up-
welling, and the isotherms and isopleths are
found progressively deeper as denser upwelling
waters subside. In July and August, conditions of
the oceanic period were evident with low nutrient
levels, higher surface temperatures, and lower
salinities; however, upward movement of the
isotherms and isopleths in August may indicate a
developing upwelling pulse.

185

FISHERY BULLETIN: VOL. 74, NO, 1

300

Jan. Feb. Mar. Apr. May June July Aug.

<v
Q

100 -

200 -

300

Jon. Feb. Mar. Apr. May June July Aug.

Figure 2. — Average depth of isotherms and nitrate isopleths
for hydrographic stations samples over Monterey Submarine
Canyon, January through Augxast 1972 (data from Broenkow
and Benz 1973).

Standing Stocks and Primary
Production

In January, at the end of the Davidson Current
period, standing stocks were near their lowest
levels and nannoplankton dominated (Table 1,
Figure 3). Throughout the period from February
through July, however, the net plankton fraction
exceeded the nannoplankton. In August, the
standing stocks were again predominantly nan-
noplankton. Estimated primary production fol-
lowed the general trend shown by the standing
stocks (Figure 4), but lower production per unit
chlorophyll for the net plankton fraction in Jan-
uary and July is apparent. The highest standing
stock was measured in April at the time the
isotherms and nutrient isopleths reached their
highest positions (see Figure 2). At this peak, the
stocks were 97% net plankton, and net plankton
concentrations in the euphotic zone ranged from
4.63 to 6.88 mg Chi a m"^. Concentrations of net

80
w

§ 60

g 40

NET ji.
1 NANNO

8
3

STANDIN
8

'kn

il i

Jon Feb. Mor Apr Moy June July Aug.

FlGURE 3. — Phytoplankton standing stocks in the euphotic
zone, January through August 1972. Numbers over histogram
bars refer to stations.

Moy

July

Aug-

FIGURE 4. — Estimated primary production in the euphotic
zone, January through August 1972. Numbers over histogram
bars refer to stations.

plankton as high as 9.26 mg Chi a m'^ were re-
corded in June. During the April peak, the cor-
responding total productivity was approximately
1.1 g C m"2 day"^ It is difficult to equate incubator
productivity to in situ productivity; however,
these values are similar to productivity estimates
calculated from nutrient uptake and oxygen pro-
duction in the water column (Smethie 1973).

The changes in the ratios of the two fractions
were largely a result of changes in the biomass of
the net plankton fraction. The net plankton frac-
tion experienced large seasonal changes in con-
centrations, and occasionally there was sig-
nificant vertical stratification within the water
column; however, nannoplankton fluctuations fell
within a much narrower range (Figure 5). There
were significant differences in the average con-
centrations in the euphotic zone of the two frac-
tions in all three hydrographic seasons, and both
fractions showed significant differences between
seasons (Mann Whitney U test; P = 0.01). The

186

GARRISON: NET PLANKTON AND NANNOPLANKTON IN MONTEREY BAY

Table i. -

— Standing stock,

primary production, and growth rate ( fi

) of the net plankton and nannoplankton

in Monterey Bay for

the period January through August 1972.

Euphotic

Phytoplankton

Primary

standing

stock

production

Growth rate,

Station

depth
(m)

(mg Chi a m-^)

(g C m-2 day-')

(doubling day-')

Date

Net

Nanno

Net

Nanno

Net

Nanno

20 Jan.

3
8

20
28

2.2
5.8

10.2
18.5

27 Jan.

4.6

7.8

0.017

0.076

• D^

0.1

0.3

6.0

12.0

0.012

0.113

0.1

0.3

15 Feb.

18.2

2.4

0.142

0.084

0.3

0.9

36.4

3.0

0.437

0.137

0.4

1.1

1 Mar.

3
8

23.2
10.0

2.8
4.4

8 Mar.

3
8

10
40

17.6
61.0

3.2
7.6

23 Mar.

3
8

11
15

0.262
0.706

0.116
0.126

18 Apr.

29.8

3.0

0.491

0.075

0.5

0.7

95.4

2.6

0.955

0.153

0.5

1.3

16 May

11.6

5.8

0.058

0.160

0.2

0.8

34.0

9.6

0.558

0.418

0.5

1.1

20 June

73.0

3.4

0.612

0.067

0.3

0.6

20.8

3.4

0.401

0.115

0.6

0.9

2.0

5.0

0.011

0.093

0.2

0.5

20 July

31.6

8.2

0 172

0.276

0.2

0.9

45.0

8.6

0.085

0.286

0.1

0.9

7.6

9.6

0.010

0.207

>0.1

0.6

29 Aug.

M.4

20.2

0.094

0.507

1.4

0.7

14.6

32.4

0309

0.612

0.6

'3.2

18.7

0.488

0.252

2.3

0.4

'Value appears low, corresponding growth rate [fJi) may be too high.

seasonal effect during upwelling seems to be a
reduction of the average concentration of nan-
noplankton and an increase in the average con-
centration of net plankton.

ro
I

E
oi

o
o> 2

9.26

mg Chi a M''

6.88

1^ .

Jan. Feb Mar. Apr. May June July Aug.

Figure 5. — Seasonal changes in the concentration of net
plankton chlorophyll (heavy line) and nannoplankton chloro-
phyll (thin line). (Davidson Current period — January; upwell-
ing period — February through June; oceanic period — July,
August.) Average and range of concentrations in the euphotic
zone are shown. The number of samples for each month is
given in Table 1.

Standing Stock Growth Rate

The growth rate, /x (doublings day"^), and as-
similation ratio (mg C mg Chi a'^ h"^), of the nan-
noplankton fraction was greater than the corre-
sponding value for the net plankton during all
three seasons, and both fractions showed their
highest growth rate during the upwelling period;
however, assimilation ratios of the surface sam-
ples for both fractions were higher in the oceanic
period than during upwelling (Table 2). There is
no correlation {P > 0.10) between the growth
rates of either phytoplankton fraction and aver-
age nutrients (NO3, Si02) in the upper 10 m on
individual sampling days for the three hydro-
graphic periods.

Net plankton growth rates exceeded nanno-
plankton growth rates in only two of the samples;

Table 2. — Growth rates of the standing stocks in the euphotic
zone and assimilation ratios of surface samples.'

Hydro-
graphic
period

Growth rate, fj.
(doublings day-')

Assimilation ratio
(mg C mg Chi a-' h-')

Net Nanno

Davidson
Current
Upwelling
Oceanic

0.1
0.4
0.2

± 0.0(2) 0.3 ± 0.1(2)
±0.1(9) 0.9 ± 0.2(9)
± 0.3(4) 0.7 ± 0.2(6)

0.4 ± 0.2(2) 2.2 ± 0.5(2)
2.7 ± 1.5(9) 5.2 ± 2.2(9)
3.0 ± 1.6(3) 10.3 ± 1.2(4)

'Growth rates were calculated from daily productivity and standing stock
estimates integrated to the depth of 1% light penetration, while assimilation
ratios are for surface samples incubated at 0.06 langley min*'. 7 ± SD(W);
questionable data indicated in Table 1 have been excluded.

187

FISHERY BULLETIN: VOL. 74, NO. 1

however, the growth rates were determined at in-
cubator light levels which were not representa-
tive of in situ conditions. The regression of light
level on the ratio of the growth rates (/u net:/u,
nanno) is significant (P < 0.01) during the up-
welling months (Figure 6). Light levels approxi-
mately equivalent to full incubator light are
found at depths of 8 to 15 m during the upwelling
period, and the upper one-fourth to one-third of
the euphotic zone receives light which is in excess
of incubator light levels.

c
c
o

.06 0.6

lO'^langley min'

Figure 6. — Regression of incubator light levels on the
net'.nanno growth rates.

Distribution in the Water Column

Since nannoplankton concentrations were rela-
tively homogeneous in the water column, max-
ima were often not well defined. Net plankton
maxima, however, were usually apparent and cor-
responded to the depth of the seasonal pycnocline.
There was no regularly observed depth relation-
ship between nannoplankton and net plankton
maxima, and they often were at the same depth.
Phaeophytin peaks appeared at the surface and in
conjunction with, or just below, the chlorophyll
maxima. High NH3 concentrations in the deeper
phaeophytin maxima may be indicative of grazing
on the phytoplankton stocks in the chlorophyll
maxima (see Figures 7-10).

During the Davidson Current period there is
little vertical stability in the water column, and
the net plankton stocks are poorly developed
(Figure 7). With the onset of upwelling net
plankton stocks develop above the strong, shal-
low pycnocline (Figures 8, 9) and the nanno-

NOj ;ug atomt liter''

NH, IO"'>jg atoms liter"'
10 20

100

25.00

26.0O

«^

Figure 7. — Vertical distribution of phytoplankton standing
stocks, phaeophytin, and hydrographic parameters during
the Davidson Current period.

NOj^jg atoms liter''
NH3 IO''*jg atoms liter''
10 20

x: 50

o.
Q

100

25.00

26.00

O-f

Figure 8. — Vertical distribution of phytoplankton standing
stocks, phaeophytin, and hydrographic parameters during
upwelling period. Station was sampled during a flowing tide.

plankton stocks decline. With strong or persistent
upwelling, the pycnocline may intersect the sur-
face and the phytoplankton stocks are concen-
trated in a relatively shallow layer (Figure 9).

After a slacking of upwelling the denser waters
subside and the pycnocline depths become pro-
gressively deeper. The surface layer can be
strongly stratified by the onshore movement of
warmer, low salinity oceanic water, and nutrient
concentrations in the near surface waters are low
during the oceanic period. The net plankton

188

GARRISON: NET PLANKTON AND NANNOPLANKTON IN MONTEREY BAY

NO3 xig atoms lifer''
NHj IO-'>ug atoms liter"'

10 20

-c 50

a

100

Figure 9. — Vertical distribution of phytoplankton standing
stocks, phaeophytin, and hydrographic parameters during
upwelling period. Station was sampled during an ebbing tide.

NOj iug atoms liter''
NH3 IO''4Jg atoms liter''

jz 50

Q.
(U
Q

100-

STATION 8
18 JUL. 72

25.00

26.00

Figure 10. — Vertical distribution of phytoplankton standing
stocks, phaeophytin, and hydrographic parameters during
the oceanic period.

maximum remains associated with the sinking
pycnocline and, although nutrients do not reach
limiting concentrations in the pycnocline, light
levels are below optimal intensity for maximum
growth rates (Figure 10).

Broenkow and McKain (1972) demonstrated
that tidal effects have a marked influence on the
distribution of hydrographic parameters over the
canyon: during a flow tide there is a down-canyon
current and isotherms and isopleths over the
canyon are depressed; conversely, during an ebb

tide the flow is up the canyon and isotherms and
isopleths are nearer the surface. The source wa-
ters for the down-canyon flow are subsurface wa-
ters from the shallow areas adjacent to the can-
yon. These tidal effects can be identified in the
distribution of the phytoplankton stocks (Silver
and Hansen 1971b), but their importance is un-
known. The chlorophyll a maximum at station 8
(in Figure 8) appears to be an intrusion of stocks
developed in shallower areas and carried to depth
by the down canyon flow during the flow tide.
Station 3 was sampled earlier during an ebb tide,
and the sigma-t surface at 50 m (crt = 26.14) was
found deeper than 100 m at station 8 (see Figure
8). At a full ebb tide the pycnocline and the stand-
ing stocks may be located very near the surface
(Figure 9).

DISCUSSION

The net plankton-dominated blooms that de-
veloped during this study were similar to those
described by Bolin and Abbott (1963) and Abbott
and Albee (1967) in their close association with
seasonal upwelling and in their composition (i.e.,
the net plankton was dominated by colonial
diatoms — M. Silver unpubl. data^). Malone
(1971c) noted an increase in net plankton fraction
during the upwelling season; however, he re-
ported net plankton dominated stocks only dur-
ing strong upwelling pulses. Malone also reported
a marked decrease in net plankton chlorophyll
and productivity between inshore and offshore
stations near the end of the upwelling season. Al-
though these studies cannot be directly com-
pared, they suggest phytoplankton blooms which
develop during upwelling are mostly net plank-
ton forms, and higher standing stocks may develop
inshore.

There seems to be a fundamental contradiction
in the measured growth rates of the two fractions
and the observed standing stocks. The growth
rates of the nannoplankton were consistently
higher than those of the net plankton, whereas
the standing stocks of nannoplankton decrease
and the stocks of net plankton increase during
the upwelling season. The observed development
of the stocks could result theoretically from one
or a combination of the following conditions: 1)

^The unpublished data supplied by M. Silver can be found in a
data report filed in 1971-72 at Oceanographic Services, Inc., 135
East Ortega Street, Santa Barbara, CA 93101.

189

FISHERY BULLETIN: VOL. 74, NO. 1

the nannoplankton fraction may be selectively
removed from the area by horizontal advection
because of their low sinking rates; 2) nanno-
plankton may be selectively grazed; 3) environ-
mental conditions may favor higher net plankton
growth rates.

Malone (1971c) discussed the argument for
selective removal of nannoplankton from upwell-
ing areas by horizontal advection. Briefly re-
stated, nannoplankton cells tend to have slower
sinking rates than net plankton cells (or they are
motile) and in convection cells they will tend to be
removed from the areas of upward movement and
concentrated in areas of downward movement
(Stommel 1949). In upwelling areas then, nan-
noplankton may be selectively removed by mass
transport of surface waters offshore. There is lit-
tle direct evidence to show that this takes place;
however, the advection hj^othesis is supported
by the observed decrease in nannoplankton
stocks between the Davidson Current period and
the upwelling period. During the Davidson Cur-
rent period there is a general onshore movement
of surface waters with water sinking along the
coast, while during the upwelling period the cir-
culation is reversed and water moved upward
along the coast, and the surface waters are trans-
ported offshore (Skogsberg 1936; Bolin and Ab-
bott 1963). Malone (1971c) found the level of the
nannoplankton stocks remained relatively con-
stant throughout the year; however, he reported
that during periods of onshore water movement
there was an enhancement which could be attrib-
uted to concentrating the nannoplankton in an
area of downward water movement.

The decrease in nannoplankton stocks reported
in the present study may have been the result of
selective grazing by microzooplankton and
planktotrophic larvae (Thorsen 1950; Beers and
Stewart 1969; Parsons and LeBrasseur 1970). In
this area many of the benthic invertebrates have
their reproductive season during the spring (M.
Houk pers. commun.)"*; increased grazing pres-
sure by these larvae may have caused the de-
crease in nannoplankton stocks. However, the
extent of grazing on either fraction of the phy-
toplankton in Monterey Bay is not known.
Zooplankton samples were collected as part of
the routine sampling program, but gelatinous

■•M. Houk, Department of Natural Science, University of
California, Santa Cruz, CA 95064.

and colonial phytoplankton could not be sepa-
rated from the zooplankton for biomass estimates.

Throughout the period of upwelling, nitrate
levels in the upper 10 m remained high (> 5 ^ig
atoms liter"^) and the chlorophyll maximum was
frequently located near the surface. At these
shallow depths light levels were in excess of in-
cubator light levels (0.06 langley min"M. Eppley
et al. (1969) have shown that the diatoms
Skeletonema costatum and Ditylum brightwellii
grow faster than Coccolithus huxleyi at high light
levels (0.1 langley min"^) when nitrate levels are
in excess of 0.8 /xg atoms liter"^, while at lower
light levels (0.02 langley min"^), the situation is
reversed and C. huxleyi will grow faster at any
nitrate concentration. In situ nutrient and light
conditions near the surface during the upwelling
period should favor net plankton growth.

In the present study and in that of Malone
(1971c), growth rates of the net plankton were
lower than the growth rates of the nannoplank-
ton; however, the two fractions responded differ-
ently to increasing light as showai by the ratio of
the growth rates (/x net.fx nanno) increasing with
higher light levels (Figure 6). The regression pre-
dicts that net plankton growth rates would ex-
ceed the nannoplankton growth rates at light
levels similar to those where Eppley et al. (1969)
showed a reversal of growth rate relationships.
Estimated light levels in the upper part of the
euphotic zone are higher than the incubator light
levels which have been used in this study and
that of Malone. Since the net plankton growth
rates show greater enhancement with increasing
light than the nannoplankton, light levels in the
upper water column may favor the growth of the
net plankton fraction and lead to net plankton
domination of the standing stocks.

Laws (1975) suggested that, under certain en-
vironmental conditions, large cells may realize a
higher net growth rate because of a decreasing
respiration rate with increasing cell size. In
Laws' model, when surface light levels are low or
the product of the attenuation coefficient and
mixed layer depth is large, integral productivity
efficiency is low and respiration losses become
more important. During the present study, how-
ever, under low light levels, the net growth rates
of the smaller cells (nannoplankton) exceeded
larger cells, and the phytoplankton populations
were net plankton dominated at a time when the
mixed layer was extremely shallow.

Notwithstanding the possible effects of selec-

190

GARRISON: NET PLANKTON AND NANNOPLANKTON IN MONTEREY BAY

tive grazing on the nannoplankton or their selec-
tive removal by horizontal advection, the de-
velopment of the upwelling bloom in Monterey
Bay is largely a result of the increase in the net
plankton fraction and may be explained in terms
of conditions which are favorable for net plankton
growth. High nutrient concentrations can be
maintained in the euphotic zone by downward
mixing from the surface which extends below the
pycnocline or by a continual input of nutrients to
the surface waters by upwelling. Optimal light
levels, however, are found only in the upper part
of the euphotic zone. The combination of these
conditions that constitute optimal growth condi-
tions for the net plankton fraction occur when the
phytoplankton stocks are restricted to a shallow
mixed layer above the pycnocline which has been
"pushed up" by upwelling water. Optimal growth
conditions vary spatially and seasonally and may
be primarily responsible for the net plankton and
nannoplankton relationship observed in Mon-
terey Bay.

Nutrients do not appear to limit the growth
rates of either fraction as correlation coefficients
of nutrient levels with growth rates were not sig-
nificant and, although nutrient levels change
seasonally, Malone (1971c) reported little sea-
sonal variation in assimilation rates. Light
levels, however, are potentially limiting a short
distance from the surface and can influence the
ratio of net:nanno growth rates.

An increase in the depth of the mixed layer
results in a decrease in the average light expo-
sure for phytoplankton cells in the mixed layer
(Parsons and Takahashi 1973). The net plankton
fraction will be more strongly influenced than the
nannoplankton because their optimal growth
rates occur at light levels near the surface, and
their vertical distribution is strongly controlled
by water movement. Upwelling water move-
ments result in a shallow pycnocline and shallow
mixed layer; with a slack in the upwelling rate,
the pycnocline sinks and there is a deeper mixed
layer. In the present study, net plankton maxima
were concentrated above the pycnocline, whereas
no particularly strong relationship between the
nannoplankton maxima were observed (the nan-
noplankton maxima were often not well defined).
Malone (1971c) showed that the net plankton
maxima were located below the nannoplankton
maxima during periods when upwelling was
slack or that both were located at the surface dur-
ing periods of upwelling, and he emphasized the

role of upward movement in controlling the verti-
cal distribution of the net plankton fraction.

Malone (1971c) showed an onshore to offshore
decrease in the ratio of net:nanno standing
stocks. Yoshida (1967) showed the potential for a
narrow zone of stronger upwelling associated
with the edge of the continental shelf where the
effects of upwelling are maximal at the edge of
the shelf and decrease exponentially shoreward
and seaward. A decrease in the upwelling rate
away from the continental shelf would result in
reduced suspension of sinking cells, a deeper
mixed layer, and lower average light levels for
phytoplankton cells in the mixed layer and could
reduce the net:nanno growth rate ratio. Malone's
data showed shallow mixed layers during periods
of strong upwelling at inshore stations and a
trend for an increasing mixed layer depth
offshore. In Monterey Bay during the upwelling
season, the mixed layer is frequently shallow or
the pycnocline intersects the surface. There are
considerable amounts of hydrographic data which
show this characteristic distribution (Broenkow
and Benz 1973) and corresponding phytoplankton
standing stock data which show significant strat-
ification of the phytoplankton standing stocks
above the shallow pycnocline (Silver see footnote
3).

The depth of the pycnocline and mixed layer
vary seasonally in response to the upward move-
ment of isotherms during upwelling and the sink-
ing of isotherms when upwelling ceases. Upwell-
ing, however, is not a continuous process and
may be particularly sporadic near the end of the
upwelling season (Bolin and Abbott 1963;
Smethie 1973). Malone (1971c) reported net
plankton dominated stocks only during periods of
strong upwelling, which suggests that in deep
water continual upwelling is necessary to main-
tain optimal growth conditions for the net
plankton fraction. During the present study the
net plankton fraction dominated the phytoplank-
ton populations in shallow water throughout the
upwelling season. This evidence and previous
evidence for an offshore decrease in the netrnanno
ratios (Malone 1971c) suggest that physical pro-
cesses in shallow water are sufficient to maintain
net plankton populations and mitigate the lack of
continual upwelling.

The physical processes in shallow water that
could serve to maintain favorable growth condi-
tions for the net plankton fraction or maintain
the population between periods of favorable con-

191

FISHERY BULLETIN: VOL. 74, NO, 1

ditions are poorly known. Tidal mixing and in-
creased turbulence in shallow water could facili-
tate cell suspension of sinking populations or
resting spores, and increase nutrient input to the
surface waters. Over Monterey Canyon and, to a
lesser extent, in the shallow areas of the bay, the
vertical distribution of nutrients (Broenkow and
McKain 1972; Smethie 1973) and phytoplankton
stocks (Silver and Hansen 1971b; Silver see foot-
note 3) are strongly influenced by tidal effects.
Turbulence and mixing in deep water results in a
decrease in the average amount of light to which
a phytoplankton cell is exposed; however, in shal-
low water the depth of mixing is limited by the
bottom and mixing here may result in resuspen-
sion of sinking cells. Many of the neritic diatoms
form resting spores which sink to the bottom and
may be an important source of innoculum to ini-
tiate blooms if they are resuspended by turbu-
lence during favorable growth conditions.

The decline in the net plankton populations
during this study corresponded to the influx of
oceanic waters in July. The end of net plankton
domination of the population appears to have
been the result of the low nutrient concentrations
in the oceanic surface waters and subsidence of
previously upwelled waters and its entrained net
plankton populations. During oceanic conditions,
nutrient levels in the surface waters favor the
growth of nannoplankton and the light levels in
the sinking net plankton maxima are not optimal
for growth. Malone (1971c) suggested, however,
that the net plankton are ultimately limited by
grazers as the grazing index (phaeo:Chl a) in-
creased and the netplankton concentrations de-
creased even before the end of the upwelling
period. Direct evidence for the extent of grazing
in Monterey Bay is not available; however, when
upwelling becomes sporadic and periodic influxes
of oceanic water occur, the stage is set for a de-
cline in the net plankton fraction without the
need for an increase in grazing pressure.

ACKNOWLEDGMENTS

I am grateful for the help of David Seielstad,
Sara Tanner, and many others who participated
in the sampling cruises. I thank W. W. Broenkow,
Scott McKain, and Sandra Benz for providing the
hydrographic data. I am particularly indebted to
Mary Silver for her encouragement, support, and
advice throughout the study and during the prep-
aration of this manuscript. Greg Cailliet re-

viewed the manuscript and offered suggestions
for its improvement.

This research was supported by Grant 2-35137
from the office of Sea Grant Programs, National
Oceanographic and Atmospheric Administration,
Department of Commerce; the Association of
Monterey Bay Area Governments; and the Soci-
ety of the Sigma Xi and was based on a thesis
submitted as a partial requirement for a M.A.
degree at San Francisco State University, Calif

LITERATURE CITED

ABBOTT, D. P., AND R. ALBEE.

1967. Summary of thermal conditions and phytoplankton
volumes measured in Monterey Bay, California 1961-
1966. Calif Coop. Oceanic Fish. Invest. Rep. 11:155-156.
ANDERSON, G. C.

1965. Fractionation of phytoplankton communities off
the Washington and Oregon coasts. Limnol. Oceanogr.
10:477-480.
BAYLIFF, W. H.

1963. The food and feeding habits of the anchoveta,
Centengraulis mysticetus, in the Gulf of Panama. Inter-
Am. Trop. Tuna Comm. Bull. 7:399-432.
BEERS, J. R., AND G. L. STEWART.

1969. The vertical distribution of microzooplankton and
some ecological observations. J. Cons. 33:30-44.
BOLIN, R. L., AND D. P. ABBOTT.

1963. Studies on the marine climate and phytoplankton
of the central coastal area of California, 1954-1960.
Calif Coop. Oceanic Fish. Invest. Rep. 9:23-45.
BROENKOW, W. W., AND S. BENZ.

1973. Oceanographic observations in Monterey Bay,
California January 1972 to April 1973. Moss Landing
Mar. Lab., TP73-3, 336 p.
BROENKOW, W. W., AND S. J. MCKAIN.

1972. Tidal oscillations at the head of Monterey Sub-
marine Canyon and their relation to oceanographic
sampling and circulation of water in Monterey Bay.
Moss Landing Mar. Lab., TP72-5, 42 p.
Dhulkhed, M. H.

1972. Food of the oil sardine taken by bottom nets and
surface gill net in the Mungalore area. Indian J. Fish.
19:163-166,
DiGBY, P. S. B.

1953. Plankton production in Scoresby Sound, East
Greenland. J. Anim. Ecol. 22:289-322.
DOTY, M. S., AND M. OGURI.

1958. Selected features of the isotopic carbon primary
productivity technique. Rapp. P.-V. Reun., Cons. Perm.
Int. Explor. Mer 144:47-55.
DUGDALE. R. C.

1967. Nutrient limitation in the sea: dynamics, identifica-
tion, and significance. Limnol. Oceanogr. 12:685-695.
Eppley, R. W.

1972. Temperature and phytoplankton growth in the sea.
Fish. Bull., U.S. 70:1063-1085.
Eppley, R. W., a. F. Carlucci, O. holm-Hansen, D. Keifer,
J. J. McCarthy, E. Venrick, and P. M. Williams.

1971. Phytoplankton growth and composition in ship-
board cultures supplied with nitrate, ammonium, or

192

GARRISON: NET PLANKTON AND NANNOPLANKTON IN MONTEREY BAY

urea as the nitrogen source. Limnol. Oceanogr.
16:741-751.

EPPLEY, R. W,, F. M. H, REiD, AND J. D. H. Strickland.

1970. The ecology of the plankton off La Jolla, California,
in the period April through September, 1967. Part III.
Estimates of phytoplankton crop size, growth rate, and
primary production. Bull. Scripps Inst. Oceanogr., Univ.
Calif 17:33-42.
EPPLEY, R. W., J. N. ROGERS, AND J. J. MCCARTHY.

1969. Half-saturation constants for uptake of nitrate and
ammonium by marine phytoplankton. Limnol.
Oceanogr. 14:912-920.

EPPLEY, R. W., AND W. H. Thomas.

1969. Comparison of half-saturation constants for growth
and nitrate uptake of marine phytoplankton. J. Phycol.
5:375-379.
GILMARTIN, M.

1964. The primary production of a British Columbia
fjord. J. Fish. Res. Board Can. 21:505-538.

Holmes, R. W.

1958. Size fractionation of photosynthesizing phytoplank-
ton. In Physical, chemical, and biological oceano-
graphic observations obtained on expedition SCOPE
in the Eastern Tropical Pacific, November - December
1956, p. 69-71. U.S. Fish Wildl. Serv., Spec. Sci. Rep.
Fish. 279, 117 p.
Holmes, R. W., and G. C. Anderson.

1963. Size fractionation of C'^-labeled natural phyto-
plankton communities. In C. H. Oppenheimer (editor).
Symposium on marine microbiology, p. 241-250.
Charles C. Thomas Publisher, Springfield, 111.
HOLM-HANSEN, O., C. J. LORENZEN, R. W. HOLMES, AND J. D. H.

Strickland.

1965. Fluorometric determination of chlorophyll. J. Cons.
30:3-15.

KAWAMURA, T.

1961. Data on phytoplankton photosynthetic activity.
Data Rec. Oceanogr. Obs. Expl. Fish (Hokkaido)
5:142-165.

LAWS, E. A.

1975. The importance of respiration losses in controlling
the size distribution of marine phytoplankton. Ecology
56:419-426.
LORENZEN, C. J.

1968. Carbon/chlorophyll relationships in an upwelling
area. Limnol. Oceanogr. 13:202-204.
Malone, T. C.

1971a. The relative importance of nannoplankton and net
plankton as primary producers in tropical oceanic and
neritic phytoplankton communities. Limnol. Oceanogr.
16:633-639.
1971b. Diurnal rhythms in net plankton and nannoplank-
ton assimilation ratios. Mar. Biol. (Berl.) 10:285-289.
1791c. The relative importance of nannoplankton and
net plankton as primary producers in the California
Current system. Fish. Bull., U.S. 69:799-820.
MARR, J.

1962. The natural history and geography of the Antarctic
krill (Euphausia superba Dana). Discovery Rep.
32:33-463.

McAllister, C. D., T R. Parsons, and J. D. H. Strickland.

I960. Primary productivity and fertility at station "P" in
the north-east Pacific Ocean. J. Cons. 25:240-259.

McCarthy, J. J., w. R. Taylor, and M. E. loftus.

1974. Significance of nannoplankton in the Chesapeake

Bay estuary and problems associated with the mea-
surement of nannoplankton productivity. Mar. Biol.
(Berl.) 24:7-16.

Parsons, T R.

1972. Size fractionation of primary producers in the Sub-
arctic Pacific Ocean. In A. Y. Takenouti (editor), Bio-
logical oceanography of the northern North Pacific
Ocean, p. 275-278. Idemitsu Shoten, Tokyo.

Parsons, T. R., and R. J. LeBrasseur.

1970. The availability of food to different trophic levels
in the marine food chain. In J. H. Steele (eidtor). Marine
food chains, p. 325-343. Oliver & Boyd, Edinb., Scotl.

Parsons, T. R., and M. Takahashi.

1973. Environmental control of phytoplankton cell size.
Limnol. Oceanogr. 18:511-515.

Reid, J. L., JR., G. I. Roden, and J. G. Wyllie.

1958. Studies of the California Ciu-rent system. Calif.
Coop. Oceanic Fish. Invest., Prog. Rep., 1 July 1956 to
1 Jan. 1958, p. 27-56.
ROJAS DE MENDIOLA, B.

1969. The food on the Peruvian anchovy. J. Cons.
32:433-434.
RYTHER, J. H.

1969. Photosynthesis and fish production in the sea.
Science (Wash., D.C.) 166:72-76.

RYTHER, J. H., D. W. MENZEL, E. M. HULBURT, C. J. LORENZEN,
AND N. CORWIN.

1971. The production and utilization of organic matter
in the Peru coastal current. Invest. Pesq. 35:43-59.

SAIJO, Y.

1964. Size distribution of photosynthesizing phytoplank-
ton in the Indian Ocean. J. Oceanogr. Soc. Jap.
19:187-189.

SAIJO, Y., AND K. TAKESUE.

1965. Further studies on the size of photosynthesizing
photoplankton in the Indian Ocean. J. Oceanogr. Soc.
Jap. 20:264-271.

SEMINA, H. J.

1972. The size of phytoplankton cells in the Pacific
Ocean. Int. Rev. Gesamten Hydrobiol. 57:177-205.

Silver, M., and J. Hansen.

1971a. Plankton studies. In Environmental studies of
Monterey Bay and the central California coastal zone.
Annual Report July 1971, p. 116-131. Moss landing Mar.
Lab., Moss Landing, Calif

1971b. Plankton studies in nearshore Monterey Bay. In
Abstract volume, Second National Coastal and Shallow
Water Research Conference. Geog. Programs, Off
Nav. Res.
Skogsberg, T.

1936. Hydrography of Monterey Bay, California. Thermal
conditions, 1929-1933. Trans. Am. Philos. Soc, New
Ser. 29:1-152.
SMAYDA, T. J.

1970. The suspension and sinking of phytoplankton in
the sea. Oceanogr. Mar. Biol., Annu. Rev. 8:353-414.

SMETHIE, W. M.

1973. Some aspects of the temperature, oxygen and
nutrient distributions in Monterey Bay, CaUfomia. Moss
Landing Mar, Lab. TP73-1, 97 p.

Steeman Nielsen, E.

1952. The use of radio-active carbon (C") for measuring
organic production in the sea. J. Cons. 18:117-140.

Steeman Nielsen, E., and E. a. Jensen.

1957. Primary oceanic production. The autotrophic pro-

193

FISHERY BULLETIN: VOL. 74, NO. 1

duction of organic matter in the ocean. Galathea Rep.
1:49-136.
STOMMEL, H.

1949. Trajectories of small bodies sinking slowly through
convection cells. J. Mar. Res. 8:24-29.

Strickland, J. D. H., R. w. eppley, and B. Rojas de

Mendiola.

1969. Phytoplankton populations, nutrients, and photo-
synthesis in Peruvian coastal waters. Inst. Mar. Peru
(Callao), Bol. 2:4-45.

Strickland, J. D. H., and T. R. Parsons.

1968. A practical handbook of seawater analysis. Fish.
Res. Board Can., Bull. 167, 311 p.
Subrahmanyan, R., and a. H. Viswanatha Sarma.

1965. Studies on the phytoplankton of the west coast
of India. Part IV. Magnitude of the standing crop for
1955-1962, vdth observations on nanoplankton and its

significance to fisheries. J. Mar. Biol. Assoc. India
7:406-419.
TEIXEIRA, C.

1963. Relative rates of photosynthesis and standing stock
of net phytoplankton and nannoplankton. Bol. Inst.
Oceanogr, (Sao Paulo) 13:53-60.
THORSEN. G.

1950. Reproductive and larval ecology of marine bottom
invertebrates. Biol. R«v. (Camb.) 25:1-45.
YENTSCH, C. S., AND J. H. RYTHER.

1959. Relative significance of the net phytoplankton and
nannoplankton in the waters of Vineyard Sound. J.
Cons. 24:231:238.
YOSfflDA, K.

1967. Circulation in the eastern tropical oceans with
special reference to upwelling and undercurrents. Jap.
J. Geophys. 4(2):l-75.

194

ABUNDANCE OF MACROCRUSTACEANS IN A NATURAL MARSH AND
A MARSH ALTERED BY DREDGING, BULKHEADING, AND FILLING^

Lee Trent,2 Edward J. Pullen,^ and Raphael Proctor^

ABSTRACT

Indices of abundance of macrocrustaceans during March-October 1969 in West Bay, Tex., were
determined for day and night and statistically compared between 1) a natural marsh area, 2) upland
and bayward canal areas of a housing development, and 3 ) an open bay area. Significance levels of 5% or
1% were used in the statistical comparisons. Catches of brown shrimp, Penaeus aztecus; white shrimp,
P. setiferus;h\ue crab, Callinectes sapidus;and pink shrimp, P. duorarum, were significantly greater at
night than during the day at one or more stations in the marsh. More grass shrimp, Palaemonetes sp.,
were caught at night than during the day, but the differences were not statistically significant.
Individuals of each species appeared to migrate into the more shallow areas of the marsh at night. At
night, brown shrimp and blue crabs were significantly more abundant in the marsh and bayward canal
areas than in the upland canal and bay areas, white shrimp were significantly more abundant in the
marsh area than in the other three areas, and pink shrimp were significantly more abundant in the
marsh than in the upland and bayward canal areas. During the day, brown shrimp were significantly
more abundant in the bayward canal area than in the upland canal and bay areas, while pink shrimp
were significantly more abundant in the marsh area than in the upland canal area. The generally lower
catches of each species in the open bay and upland canal areas when compared with the marsh and
bayward canal areas were attributed to: 1 ) permanent loss of intertidal vegetation in the housing
development; 2) low abundance of detrital material and benthic macroinvertebrates in the open bay
and upland canal areas; and 3) eutrophic conditions in the upland canal area.

Development of bayshore property into housing
sites by dredging, bulkheading, and filling is oc-
curring in many estuaries. When this property is
developed, shallow bay and tidal marsh areas are
often dredged or filled with spoil, thus changing
the environment for marine organisms. Informa-
tion is available on some of the environmental
changes that are critical, but the effects of these
changes on the abundance of macrocrustaceans in
Gulf coast estuaries are poorly known.

Ecological studies conducted by personnel of the
National Marine Fisheries Service in the Jamaica
Beach housing development in West Bay, Tex.,
during 1969 were reported by Trent et al. (1972).
That report described the study area and included
summary information on phytoplankton produc-
tion, oyster production, benthic organisms, sedi-
ments, hydrology, and the abundance of macro-
crustaceans and fishes. Detailed analyses were
reported by Corliss and Trent (1971) on phyto-

iContribution No. 398, Gulf Coastal Fisheries Center, Na-
tional Marine Fisheries Service, NOAA, Galveston, TX 77550.

^Present Address: Gulf Coastal Fisheries Center Panama City
Beach Laboratory, NMFS, NOAA, Panama City, FL 32401.

^Present address: Department of the Army, U.S. Corps of En-
gineers, Galveston, TX 77550.

Manuscript accepted April 1975.

FISHERY BULLETIN: VOL. 74, NO. 1, 1976.

plankton production, Moore and Trent (1971) on
oyster production, and Gilmore and Trent (1974)
on benthic organisms and sediments.

Mock (1966) studied the abundance of brown
shrimp, Penaeus aztecus, and white shrimp, P.
setiferus, in Galveston Bay, Tex., after the
bayshore was altered by bulkheading. He stated
that catches of brown shrimp were 2.5 times
greater, and catches of white shrimp were 14
times greater in a natural habitat than in a bulk-
head area.

The objectives of this study in the Jamaica
Beach area during 1969 were to evaluate relative
abundance of selected macrocrustaceans be-
tween: 1) day and night; 2) housing development
canals, natural marsh areas, and open bay areas;
and 3) areas with different concentrations of dis-
solved oxygen.

STUDY AREA AND METHODS

The study area, located in West Bay, included a
natural marsh area, an open bay area, and a
canal area. The canal area was similar to the
natural marsh before it was altered by channeli-
zation, bulkheading, and filling (Figure 1). The
altered area, which included, prior to alteration,

195

FISHERY BULLETIN: VOL. 74, NO. 1

Figure l. — Study area and sampling locations in the Jamaica
Beach area of West Bay, Tex.

Winkler method (Carritt and Carpenter 1966).
Crustaceans were collected in a trawl that had a
mouth opening of 0.6 m by 3.0 m and a stretched
mesh of 28.0 mm in the body and 2.5 mm in the
cod end. At each station the trawl was towed 200
m at about 2 knots. "Abundance" and "catch" are
used synonymously in this report as our index of
relative abundance. These terms refer to either
the number or average number of animals caught
per 200-m tow with the trawl.

Data were treated differently than those re-
ported by Trent et al. (1972) in that stations 1-5
in the altered area were subclassified into upland
canal area (stations 1-3) and bay ward canal area
(stations 4, 5); classification of stations 6-9 in the
marsh and station 10 in the bay remained the
same.

The data were treated statistically as follows
for the five species caught in greatest abundance
(Table 1): differences in catches between day and
night were tested with a paired-comparison ^test
using individual catches at a station as observa-
tions; differences between areas were tested with
Tukey's a;-procedure (Steel and Torrie 1960)
using the average catch by area, date, and time of
day as observations.

COMPARISONS OF CATCH
BETWEEN DAY AND NIGHT

Eight genera and at least 11 species were rep-
resented in the catches (Table 1). Four species
and members of the genus Palaemonetes were

about 45 hectares of emergent marsh vegetation
(predominantly Spartina alterniftora) , intertidal
mud flats, and subtidal water area was reduced to
about 32 hectares of subtidal water area by
dredging and filling; water volume (mean low
tide level) was increased from about 184,000 m^
to about 394,000 m^. Ten sampling stations were
established in the study area. Average water
depths (mean low tide level) at stations 1 through
10 were 1.6, 2.6, 2.2, 1.4, 1.3, 0.5, 0.2, 0.4, 0.5, and
1.0 m, respectively.

Samples of water and crustaceans were col-
lected during the day between 1000 and 1400 h
and at night between 2200 and 0200 h at 2-wk
intervals from 25 March to 21 October 1969 at
each station. Water samples for determining dis-
solved oxygen were taken 30 cm above the bot-
tom. Oxygen was measured using a modified

Table l. — Species or genera and total numbers of crustaceans
caught by area during the study.

Species

Brown shrimp,

Penaeus aztecus
White shrimp,

P. setiferus
Grass shnmp,

Palaemonetes sp.
Blue crab.

Callinectes sapidus
Pinl< shrimp.

Penaeus duorarum
Mantis shrimp,

Squilla sp.
Brokenback shrimp,

Trachypenaeus sp.
Stone crab,

Menippe mercenaha
Mud crab,

Eurypanopeus sp.
Swimming crab,

Callinectes similis
Pistol shrimp,

Alpheus sp.

Upland Bayward
canal canal Marsh

Bay

6.112

16,195

27,063

2,505

1,150

2,738

10,961

172

8,336

181

583

1,149

636

196

TRENT ET AL.: ABUNDANCE OF MACROCRUSTACEANS IN MARSHES

caught in sufficient numbers for detailed
analyses.

Brown shrimp was caught in greater numbers
during the day in the canal and bay areas and in
greater numbers at night in the marsh area ex-
cept at station 6 (Table 2). In the canals, day
catches were much greater than night catches at
the upland canal stations but were only slightly
greater than night catches at the bayward canal
stations. In the marsh, night catches were sig-
nificantly greater than day catches at stations 8
and 9, slightly greater than day catches at station
7, and less than day catches at station 6.

White shrimp was caught in greater numbers
at night than during the day at all stations except
station 5. The differences were statistically sig-
nificant at stations 7-9.

Grass shrimp, Palaemonetes sp., was caught in
greater numbers during the day at two of the
canal stations and in greater numbers at night at
the remaining stations; the differences were not
statistically significant, however.

Blue crab, Callinectes sapidus, was caught in
greater numbers during the day at the upland
canal stations (significant at station 3) and in
greater numbers during the night at the remain-
ing stations (statistically significant at stations
5-8).

Pink shrimp, Penae us duorarum, was caught in
greater numbers at night than during the day at
all stations except station 6. Differences were
statistically significant at stations 5 and 8.

COMPARISONS OF
CATCH BETWEEN AREAS

Statistically significant differences in night
catches between areas were observed for four of
the five species; day catches were significantly
different between areas only for brown and pink
shrimps (Table 3). Abundance of brown shrimp
during the day was significantly greater in the
bayward canal area than in the upland canal and
bay areas, whereas at night, brown shrimp were
significantly more abundant in the marsh and
bayward canal areas than in the other two areas.
Catches of white shrimp at night were sig-
nificantly greater in the marsh area than in the
other three areas. Blue crabs were significantly
more abundant at night in the marsh and bay-
ward canal than in the bay and upland canal
areas. Catches of pink shrimp were significantly
greater in the marsh than in the upland canal
area during the day and significantly greater in
the marsh than in both canal areas at night.

CATCH RELATED TO
DISSOLVED OXYGEN

Mean dissolved oxygen values and mean catch
of each species by date and area are shown in
Figure 2. Mean oxygen values in the bajrward
canal, marsh, and bay areas were above 3.0 ml/
liter throughout the study except on 1 July in the
bayward canal and on 23 September in the

Table 2. — Comparisons between day and night catches (mean number caught per tow) by species and station

(paired comparison f-test with 15 df ).

Upland

Bayward

Bay

Species and

canal stations

canal stations
4 5

Marsh stations

station

time of day

Brown shrimp;

Day

194.4

27.6

77.3

222 9

298.1

210.7

137.1

212.5

93.5

81.4

Night

47.1

14.5

21.1

203.4

287.8

167.6

177.3

481.9

210.8

75.2

f-value

-1.90

-1.02

-1.24

-0.42

-0.18

-1.04

0.98

3.29"

4.43"

-0.20

White shrimp:

Day

5.8

1.7

12.5

30.1

73.4

76.0

16.1

4.4

8.1

2.9

Night

11.9

3.3

36.7

35.0

32.6

127.6

188.6

178.4

85.8

7.9

f-value

0.79

1.42

1.23

0.75

-0.89

1.18

3.25"

2.93-

2.55-

2.00

Grass shnmp:

Day

0.1

1.5

0.0

0.4

31.8

37.7

22.4

2.2

0.4

Night

1.0

0.4

0.1

0.6

320.4

43.0

61.0

2.5

0.9

f-value

1.45

-0.94

1.60

-1.23

1.00

1.03

0.21

1.40

0.18

1.09

Blue crab:

Day

3.9

1.5

2.4

6.6

7.6

8.8

2.3

8.0

2.6

1.3

Night

1.6

0.8

1.1

7.4

14.8

18.8

10.7

16.2

4.4

2.4

f-value

-1.61

-1.74

-2.77*

0.46

2.74-

1.93

2.87*

2.85-

1.04

1.28

Pink shrimp:

Day

0.1

0.5

0.2

4.8

2.1

1.1

0.2

1.0

Night

1.6

1.9

1.1

1.2

3.2

0.6

4.2

12.2

14.6

2.8

f-value

1,67

1.24

1.52

0.90

2.35-

-1.79

1.20

2.12-

2.04

1.93

'Significant at 5% level.
"Significant at 1% level.

197

FISHERY BULLETIN: VOL. 74, NO. 1

Table 3. — Comparisons of catches between areas (bay; bay-
ward canal, BC; marsh; upland canal, UC) by species and time
of day (Tukey's w-procedure with 60 df).

Species and
time of day

Area, mean catch ( ), and significance lines'

Brown shrimp:
Day

Night

White shrimp:
Day

Night

Grass shrimp:
Day

Night

Blue crabs:
Day

Night

Pink shrimp:
Day

Night

Bay

(81.4)

(99.7)

Marsh
(163.4)

(260.5)

UC
(27.6)

Bay
(2.9)

Bay
(75.2)

(6.7)

(245.7)

Marsh
(26.1)

Marsh
(259.4)

(51.7)

Bay
(7.9)

UC
(17,3)

(33.8)

Marsh
(145.1)

Bay
(0.4)

BC
(0.4)

UC
(0.5)

Marsh
(23.5)

(0.3)

UC
(0.6)

Bay
(0.9)

Marsh
(106.7)

Bay
(1.3)

(2.8)

Marsh
(5.4)

(7.1)

(1.2)

Bay
(2.4)

(11.2)

Marsh
(12.5)

(0.1)

(0.3)

Bay
(1.0)

Marsh
(2.0)

(1.5)

BC
(2.2)

Bay
(2.8)

Marsh
(7.9)

'Any two means not underscored by the same line are significantly different
at the 5% level.

marsh. In contrast, mean oxygen values observed
in the upland canal area remained below 3.0 ml/
liter from 20 May to 12 August and were below
2.0 ml/liter on three occasions. From 20 May to 12
August, about 24% of the individual observations
of oxygen values from the upland canal stations
were below 1.0 ml/liter, whereas all individual
observations from the other three areas were
above 1.5 ml/liter.

The normal patterns of seasonal abundance
were reflected for brown shrimp, white shrimp,
and blue crabs by catches in the bayward canal,
marsh, and bay areas (Figure 2). Immigration
and emigration in Galveston Bay by brown and
white shrimps occur during different seasons
(Baxter and Renfro 1966; Trent 1967; Pullen and
Trent 1969). Brown shrimp postlarvae immigrate
in late winter and early spring and most of the
juveniles emigrate in late spring and early sum-
mer. White shrimp postlarvae immigrate in the
summer, and the juveniles emigrate in the fall or
early winter depending on water temperature.
Blue crabs are abundant throughout the year in
Galveston Bay (Chapman 1965).

N. 3

BROWN SHRIMP

■*'''''"'"~'>^f,,„X«t""""*'''''i^

PINK SHRIMP

^Hfe»«,, „.„,„„

25 8 22 6 20 3 17 I IS 29 12 26 9 23 7 21
MAR APR MAY JUNE JULY AUG SEPT OCT

Figure 2. — Mean dissolved oxygen values, and mean catch of
each species by area and time of year.

Patterns of seasonal abundance for grass and
pink shrimps are not documented for the Galves-
ton Bay system. In Redfish Bay, Tex. (about 150
miles southwest of our study area), Hoese and

198

TRENT ET AL.: ABUNDANCE OF MACROCRUSTACEANS IN MARSHES

Jones (1963) caught grass shrimp in greatest
numbers during late winter and early spring and
pink shrimp in greatest numbers during spring
and early fall. Seasonal abundance patterns
reflected by catches in this study were similar to
those reported in Redfish Bay: for grass shrimp in
the marsh area; and for pink shrimp in the bay-
ward canal, marsh and bay areas during late
summer and early fall.

Seasonal abundance of brown shrimp, white
shrimp, blue crabs, and pink shrimp deviated
from what we expected in the upland canal area.
These deviations were probably caused by low dis-
solved oxygen. During the period of low dissolved
oxygen (below 3.0 ml/liter; from 20 May to 12
August) in the upland canal area, mean catches of
brown shrimp dropped and remained below the
mean catches of brown shrimp in the other three
areas; mean catches of white shrimp and blue
crabs remained below mean catches of white
shrimp and blue crabs in the bayward canal and
marsh areas after 3 June. The abundance of pink
shrimp increased on 29 July in all areas except the
upland canal area and remained higher than in
the upland canal area until 7 September Grass
shrimp were not caught in large numbers in any
area except the marsh and therefore were not used
to evaluate the effects of low dissolved oxygen.

DISCUSSION AND SUMMARY

Indices of abundance revealed differences in
day-night distribution of brown shrimp, white
shrimp, blue crabs, and pink shrimp in the study
area. Assuming that our catch per unit effort data
provided an index which unbiasedly represented
density, migration of individuals of all four
species into the more shallow areas of the marsh
at night best explains these distributional differ-
ences. Inherent in the assumption that catch per
unit effort unbiasedly estimates density is the
equal vulnerability of the animals to capture dur-
ing both day and night. Factors which could make
this assumption invalid include: 1) burrowing or
swimming above the trawl by the animals during
one but not the other time period, and 2) avoid-
ance of the trawl during the day or night. Re-
gardless of the correctness of our assumption, the
importance of sampling during both day and
night to determine differences in abundance be-
tween areas was clearly shown.

All five species were more abundant in the

marsh than in the upland canal area during both
day and night. Brown shrimp, white shrimp, blue
crabs, and pink shrimp were more abundant in
the bayward canal area than in the upland canal
area. The distributional patterns of pink shrimp
and blue crabs in this study were similar to those
reported by Lindall et al. (1975), who provided
data showing that catches of blue crabs and pink
shrimp were highest in the bayward portion of an
upland canal in a housing development in Tampa
Bay, Fla.

Four factors probably account for most of the
differences observed in abundance of shrimps be-
tween areas. Intertidal vegetation was perma-
nently eliminated by alteration of the natural
area for the housing development. Detrital mate-
rials and abundance of benthic macroinverte-
brates were lowest in the open bay area, low in
the upland canal area, and highest in the bay-
ward canal and marsh areas (Gilmore and Trent
1974). Eutrophic conditions observed represent
the fourth factor.

Eutrophic conditions, indicated by the observed
low values of dissolved oxygen in the upland ca-
nals of the housing development during the
summer, probably account for the comparatively
low catches of brown shrimp, white shrimp, pink
shrimp, and blue crabs during that period.
Further evidence of eutrophication in this area
was provided by studies on: the American oyster,
Crassostrea virginica, in which setting, survival,
and growth rates were less in the upland canal
area than in the marsh area (Moore and Trent
1971); phytoplankton in which production was
higher in the upland canal area than in the
marsh or bay areas (Corliss and Trent 1971); and
benthic macroinvertebrates in which the abun-
dance of the organisms declined drastically dur-
ing the summer months in the upland canal area
(Gilmore and Trent 1974). Symptoms of eutrophic
conditions in the upland canals of the housing
development include inadequate water exchange
and high nutrient levels. These factors were dis-
cussed in detail by Moore and Trent (1971).

Alteration of estuaries by dredging and filling
for housing developments and boat basins results
in an environment highly susceptible to recur-
rent low dissolved oxygen levels. This probelm of
low oxygen has been shown also in Forida (Taylor
and Saloman 1968; Lindall et al. 1973) and
California (Reish 1961). Stresses resulting from
low dissolved oxygen reduce the abundance of
crustaceans and other animals in the stressed

199

FISHERY BULLETIN: VOL. 74, NO. 1

areas and may produce mass mortalities. Flow
dynamics and sedimentation patterns should be
carefully evaluated when new developments in
estuaries are being considered in order to prevent
areas of stagnant water from being created.

ACKNOWLEDGMENTS

Sincere appreciation is extended to Edwin A.
Joyce, Jr., and his staff, Florida Department of
Natural Resources, for reviewing this manuscript
and for helpful suggestions.

LITERATURE CITED

Baxter, K. N., and W. C. Renfro.

1966. Seasonal occurrence and size distribution of postlar-
val brown and white shrimp near Galveston, Texas, with
notes on species identification. U.S. Fish Wildl. Serv.,
Fish. Bull. 66:149-158.
CARRITT, D. E., AND J. H. CARPENTER.

1966. Comparison and evaluation of currently employed
modifications of the Winkler method for determining dis-
solved oxygen in seawater; a NASCO report. J. Mar. Res.
24:286-318.
CHAPMAN, C. R.

1965. Estuarine program. In Biological Laboratory, Gal-
veston, Tex. fishery research for the year ending June 30,
1964, p. 60-75. U.S. Fish Wildl. Serv., Circ. 230.

Corliss, J., and L. Trent.

1971. Comparison of phytoplankton production between
natural and altered areas in West Bay, Texas. Fish. Bull.,
U.S. 69:829-832.
Gilmore, G., and L. Trent.

1974. Abundance of benthic macroinvertebrates in natural
and altered estuarine areas. U.S. Dep. Commer., NCAA
Tech. Rep. NMFS SSRF-677, 13 p.

HOESE, H. D., AND R. S. JONES.

1963. Seasonality of larger animals in a Texas turtle grass
community. Publ. Inst. Mar. Sci., Univ. Tex. 9:37-47.
LINDALL, W. N., JR., W. A. FABLE, JR., AND L. A. COLLINS.
1975. Additional studies of the fishes, macroinvertebrates,
and hydrological conditions of upland canals in Tampa
Bay, Florida. Fish Bull., U.S. 73:81-85.
LINDALL, W. N., Jr., J. R. Hall, and C. H. Saloman.

1973. Fishes, macroinvertebrates, and hydrological condi-
tions of upland canals in Tampa Bay, Florida. Fish. Bull.,
U.S. 71:155-163.

Mock, C. R.

1966. Natural and altered estuarine habitats of penaeid
shrimp. Proc. Gulf Caribb. Fish. Inst., 19th Annu. Sess.,
p. 86-98.

Moore, D., and L. Trent.

1971. Setting, growth, and mortality of Crassostrea vir-
ginica in a natural miarsh and a marsh altered by a hous-
ing development. Proc. Natl. Shellfish. Assoc. 61:51-58.

Pullen, E. J., AND W. L. Trent.

1969. White shrimp emigration in relation to size, sex,
temperature and salinity. FAQ Fish. Rep. 57:1001-1014.
REISH, D. J.

1961. A study of benthic fauna in a recently constructed
boat harbor in southern California. Ecology 42:84-91.

Steel, r. g. d., and j. r. torrie.

I960. Principles and procedures of statistics, with special
reference to the biological sciences. McGraw-Hill Book
Co.,N.Y., 481 p.
Taylor, J. L., and C. H. Saloman.

1968. Some effects of hydraulic dredging and coastal de-
velopment in Boca Ciega Bay, Florida. U.S. Fish Wildl.
Serv., Fish. Bull. 67:213-242.
TRENT, W. L.

1967. Size of brown shrimp and time of emigration from the
Galveston Bay system, Texas. Proc. Gulf Caribb. Fish.
Inst., 19th Annu. Sess., p. 7-16.

Trent, w. l., e. j. pullen, and d. moore.

1972. Waterfront housing developments: Their effect on the
ecology of a Texas estuarine area. In M. Ruivo (editor).
Marine pollution and sea life, p. 411-417. Fishing News
(Books) Ltd., West Byfleet, Surrey.

200

NOTES

MORTALITIES AND EPIBIOTIC

FOULING OF EGGS FROM

WILD POPULATIONS OF THE

DUNGENESS CRAB, CANCER MAGISTER'''

Cultured crustaceans have been found to be sus-
ceptible to fouling by a variety of epibionts. Nilson
et al. (1975) recently described mortalities attrib-
uted to epibiotic fouling in the eggs and larvae of
the American lobster, Homarus americanus, the
larvae of the prawn, Pandalus platyceros, and lar-
vae of the Dungeness crab, Cancer magister Dana.
This same type of fouling has also been found on
juveniles of Penaeid shrimp, where it causes death
in rearing ponds with low oxygen content by in-
habiting the gill filaments and suffocating the
animal (Johnson et al. 1974; Lightner et al. 1975).
The organisms most commonly encountered have
been filamentous bacteria and algae.

Work on the larval cultivation of the Dungeness
crab at the Bodega Marine Laboratory, Bodega
Bay, Calif., revealed heavy fouling on the eggs of
oviposited female crabs held in rearing tanks.
Further investigation showed that the condition
also existed on eggs of crabs obtained from local
fishermen. Egg masses with extensive fouling also
showed a large number of empty egg cases, al-
though eyespot development on the remaining
embryos showed the time until hatching to be dis-
tant. Similar fouling of the eggs of wild caught
Atlantic blue crabs, Callinectes sapidus, has been
observed and well documented (Sandoz et al. 1944;
Rogers-Talbert 1948). With Callinectes, however,
the predominant fouling organism appears to be
the fungus Lagenidium callinecti.

These observations of fouling and mortality in
the natural population suggest a possible explana-
tion for the decline in Dungeness crab catches
recorded in the San Francisco Bay region since
1960 (Biostatistical Section 1961, 1963, 1964, 1965;
Greenhood and Mackett 1965, 1967; Heimann and
Frey 1968a,b; Heimann and Carlisle 1970; Pinkas
1970; Bell 1971; Oliphant 1973). In order to inves-
tigate this possibility, a distributional study was
undertaken, comparing mortalities and epibiotic

fouling of crab eggs from various locations along
the coast of northern California.

Materials and Methods

Egg samples of C. magister were obtained from
fishermen along the northern California coast
during the period from 27 November 1974 to 30
January 1975. A total of 105 samples of eggs from
individual crabs were obtained from six regions
which included the following localities (Figure 1):
region I — Pacifica (4 samples); region II —
Drake's Bay (18 samples); region III — Point Reyes
(39 samples); region IV — Bodega Bay, Russian
River, and Gualala (10 samples); region V — Fort
Bragg (20 samples); region VI — Eureka (14 sam-
ples).

In the field, a portion of eggs were removed from
the Dungeness crab egg masses and placed in vials

Froncisco Boy

'This work was supported by California State Legislature
Aquaculture funds.

^This work was done at the University of California, Bodega
Marine Laboratory at Bodega Bay, CA 94923.

Figure l.— The coast of northern California showing the
Dungeness crab collection sites: I - Pacifica; II - Drake's Bay; III
- Point Reyes; IV - Bodega Bay, Russian River, and Gualala; V -
Fort Bragg; VI - Eureka.

201

containing 10% Formalin^ in seawater. The sam-
ple size was variable — all exceeded 100 eggs, usu-
ally several hundred. The vials were then shipped
to the laboratory for examination with the aid of a
dissecting microscope. The epibiotic organisms
were clearly visible using transmitted light for
illumination (Fisher et al. 1975). Closer examina-
tion of the egg cases was carried out with a phase
microscope to aid in the characterization of the
fouling organisms. Portions of the samples were
categorized as to the comparative developmental
state of the eggs, extent of epibiotic fouling, and
egg mortality by the following methods:

1. The following observations of the eyespots
which develop as the embryos develop were used to
give a comparative estimate of the time the eggs
had been carried externally on the female:

Dl. No visible eyespot.
D2. Emerging eyespot.
D3. Full eyespot.

Any samples which showed evidence of hatch-
ing were not used. Occasionally, there was varia-
tion in the degree of development of the eggs from
a single sample, in which case the eggs that had
developed furthest were used for observation.

2. The extent of epibiotic fouling was deter-
mined by the following observations of the exter-
nal egg membrane:

Fl. None — no evidence of epibionts at 100 x
(Figure 2A).

F2. Light — occasional short filaments.

F3. Moderate — the majority of the surface
covered with short filaments and occa-
sional long filaments (Figure 2B).

F4. Heavy — the surface extensively covered
with short and long filaments (Figure
2C).

F5. Very heavy — the surface extensively
covered with short filaments, long fila-
ments, and detrital material.

3. The number of empty egg cases was used as
an estimate of mortality.

Ml. <10% mortality.
M2. 10-25% mortality.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

M3. 26-50% mortality.
M4. 51-75% mortality.
M5. 76-100% mortality.

Only empty egg cases (Figure 3) were consid-
ered mortalities. Other abnormal conditions,
such as discolored eggs which might have eventu-
ally led to mortalities, were observed but not used
in the estimates. All developmental stage D3
samples were checked for emerging embryos to
ensure that the empty egg cases were not due to
hatching.

In addition to the field samples, seven ovigerous
females from the Point Reyes area were examined
before being placed into flow-through seawater
tanks at the laboratory. After 25 days the eggs
were reexamined to determine the progress of the
infestation. In addition, one complete egg mass
from an ovigerous female was examined to deter-
mine the homogeneity of the fouling condition
throughout the egg mass.

Results

Observation of eyespot development placed
10.5% of the samples into category Dl, 35.2% into
D2, and 54.3% into D3. Fouling was observed in all
developmental categories, but mortalities were
generally higher in the more developed eggs. The
histograms presented in Figure 4 show the percent
of samples from each region placed in each mortal-
ity category (M1-M5) and fouling category (F1-F5)
after combining the developmental categories.

The eggs of the seven females held in the labo-
ratory for 25 days showed an average increase
in their development, fouling and mortality of
one level in each category. The greatest observed
change was on an egg mass in developmental
stage 2 which originally showed light fouling (cat-
egory 2) and were in mortality category M2.
After 25 days it was in developmental stage 3
and showed very heavy fouling (category 5) and
had advanced to mortality category, M5. Another
showed no increase in fouling as it matured
from developmental stages 1 to 3, but the egg
mortality category advanced from Ml to M3.

Examination of the entire egg mass of one
specimen showed that the extent of the fouling
was variable and concentrated mostly on the
periphery of the mass and on the inner eggs near
the fold of the abdomen. This raises the possibility
of sampling error; however, it would probably be
insignificant since the field samples came primar-
ily from the exterior of the egg masses.

202

n**

■'■f

^•^%lfc^ ^.

•'4:^

-'■ r'

Figure 2. — Dungeness crab egg samples showing (A) no epibio-
tic folding on the egg membrane (78x), (B) moderate epibiotic
fouling on the egg membrane (96x ), and (C) very heavy epibiotic
fouling on the egg membrane (57 x).

The epibiotic fouling organisms found were
similar to those noted on other crustaceans by
Nilson et al. ( 1975). Particularly prominent were
the long filamentous cyanophytes which resem-
bled Oscillatoria and bacterial filaments similar to
Leucothrix. In heavily fouled samples stalked pro-
tozoans (vorticellids) were also observed. These
and the filamentous organisms trapped detrital
material, which added to the overall contamina-
tion of the eggs. Fouling on the egg stalk was often
more extensive than fouling on the egg membrane
proper. Empty egg cases also showed heavier foul-
ing than those containing embryos. In many cases
where fouling was observed, worms were found,
and the population of worms was generally larger
on egg samples with heavier fouling. The worms
were identified as the nemertean egg predator
Carcinonemertes epialti as described by Kuris
(1973).

203

Figure 3. — Dungeness crab egg sample showing empty egg cases representing egg mortalities alongside

viable eggs from the same egg mass (24. 5x).

Discussion

Various workers have attributed mortalities
(Johnson et al. 1974; Lightner et al. 1975; Nilson
et al. 1975; Fisher et al. 1975) in cultured crusta-
ceans to epibiotic fouling. These reports suggest
that death may be caused either by mechanical
interference in larval molting or restriction of
gaseous exchange across the egg or gill mem-
brane. The fouling organisms may also consume a
great deal of the available oxygen from the envi-
ronment. The dramatic effect of this condition may
be seen in Figure 5 where the moderately fouled
egg case is entirely intact, yet the embryo is at-
rophied and nonviable.

Infestation with fouling organisms presumably
does not begin until the eggs are oviposited. Al-
though heavy fouling may occur, few mortalities
are observed in the early developmental periods.
Fouling on the eggs held in rearing tanks pro-
gressed as the eggs developed. The progression
was an increase in the number or filament length
of any one type of the organisms or the addition of
other types of organisms. By the second and third
developmental categories, mortalities were regu-
larly encountered where fouling occurred.

The samples obtained from regions II and III
showed the heaviest epibiotic fouling, as well as

the highest levels of mortality. In comparison, re-
gion V showed the least extensive fouling and the
fewest mortalities. This suggests that there is a
relationship between epibiotic fouling and egg
mortality.

Closer examination of the histograms in Figure
4 reveals a possible trend of mortalities and foul-
ing progressively decreasing from region II to re-
gion V. Although the number of samples obtained
from region I may not be conclusive evidence, they
suggest that the trend may not continue south of
San Francisco Bay. The region VI data show a
slight reversal of the trend although mortalities
and fouling are still comparatively low.

The mortalities observed in regions II and III
are particularly relevant when the coastal crab
catch over the last 25 yr is considered. Figure 6
shows a general coast-wide decline in Dungeness
crab catch commencing in 1958. In 1965, the
northern fishery areas began a strong recovery,
whereas the San Francisco area remained at low
level. During this decline, the catch of the San
Francisco fishery dropped from 8y2 million pounds
to less than 1 million pounds where it has re-
mained.

Several studies have investigated the potential
impact of overfishing on the Dungeness crab popu-
lation. Poole (1962) and Cordier (1966) showed

204

Ml I 2 ' 3 ' 3 I 5 '

Morfolity levels
»

Fouling levels

Figure 4. — Histograms representing the percent of Dungeness
crab samples from each region found in the mortality and fouling
categories. The arrows represent increasing mortalities and in-
creasing fouling. Sample sizes from each region are shown.

that 99% and 98%, respectively, of the adult
female population had been inseminated, indicat-
ing that the fishing industry (which only legally
catches males greater than 6^^ inches across the
carapace) is not significantly reducing the repro-
ductive capabilities of the crab population. Also,
tagging studies have shown that an estimated 90
to 100% of the legal-size males in fishing areas of
the California coast have been caught each year
since 1929 (Pacific Marine Fisheries Commission
1965). Cleaver (1949) and Peterson (1973) stated
that the fishing pressure has been similar in
Washington and Oregon. It therefore appears that
fisheries along the coast are capable of maintain-
ing production despite the virtually maximum
fishing pressures. Poole and Gotshall ( 1965) con-
cluded that the fishing regulations at that time
were sufficient to protect the crab from depletion
through overfishing.

Physical factors may be responsible for periodic
fluctuations in crab abundance. The Pacific
Marine Fisheries Commission (1965) suggested
that shifting currents played a role in these fluc-

FlGURE 5. — A single Dungeness crab egg showing an intact
membrane, an atrophied and nonviable embryo (168 x).

- Eureka- Ft Bragg
-Oregon
San Frartciico

Crab 5«oson Ycori (1948-19/21

Figure 6. — A graph comparing the Dungeness crab catches
reported from 1948 to 1972 in three areas. Note that the San
Francisco crab catch did not increase from the 1961-62 level.

tuations by disturbing larval settlement. Lough
(1974) found a correlation between rainfall during
salinity-sensitive larval stages and crab catch 4 yr
later when those larvae were to enter the fishery.
Peterson (1973) and Botsford and Wickham (1975)
have found a positive correlation between upwel-
ling intensity and crab catch.

205

Our observations indicate that disease is a fac-
tor to be considered in evaluating the decline of
the San Francisco area crab population. The re-
productive capacity of the population must be af-
fected by this epibiotic fouling condition especially
if it can also infest the larval stages as indicated by
the studies on other crustaceans (Fisher et al.
1975).

The variety of fouling organisms and the geo-
graphical trends observed in this disease situation
suggest a complex relationship with external en-
vironmental factors. In view of the saprophytic
nature of the fouling organisms, their major
source of nutrients is probably external. As such,
the growth of the contaminants are affected by the
nutrient level in the seawater.

It appears that the external factors involved
may originate in the San Francisco Bay effluent.
This is suggested by the decreasing trend of mor-
talities and fouling heading north from this area,
presumably reflecting the dilution of the effluent
waters. The normal water currents in this area
flow in a southerly direction; however, during the
period from November through February, the pre-
vailing inshore flow is the northerly Davidson
Current (Reid et al. 1958). During the egg-bearing
season, the effluent from San Francisco Bay is
carried northward.

The observations of this study were limited by
the collection of samples during only the 1974-75
crab season. Because of the potential relationship
of these findings to a valuable natural resource,
we felt that it was important to communicate the
available information. It is clear that further stud-
ies during the next season will enhance our under-
standing of the situation.

Acknowledgments

We thank Robert Shleser for his support under
California State Legislature funds and in the
preparation of the manuscript; Edgar Nilson for
his assistance and suggestions; Louis Cavellini,
Earl Carpenter, Tom Burke, and the crab fisher-
men of northern California who obtained egg
samples.

Literature Cited

Bell, R. R.

1971. California marine fish landings 1970. Calif. Fish
Game, Fish Bull. 154, 50 p.
BIOSTATISTICAL SECTION, MARINE RESOURCES OPERATIONS.
1961. The marine catch of California for the year
1960. Calif Fish Game, Fish Bull. 117, 45 p.

1963. The California marine fish catch for 1961. Calif.
Fish Game, Fish Bull. 121: 1-47.

1964. The California marine fish catch for 1962. Calif.
Fish Game, Fish Bull. 125, 45 p.

1965. The California marine fish catch for 1963. Calif.
Fish Game, Fish Bull. 129, 45 p.

BOTSFORD, L. W., AND D. E. WiCKHAM.

1975. Correlation of upwelling index and Dungeness
crab catch. Fish. Bull., U.S. 73:901-907.

Cleaver, F. C.

1949. Preliminary results of the coastal crab (Cancer
magister) investigation. Wash. Dep. Fish., Biol. Rep.
49A:47-82.
CORDIER, P.

1966. Cruise report 66-N-9 crab. Calif. Dep. Fish Game,
Mar. Resour. Oper.

FISHER, W. S., E. H. NiLSON, AND R. A. SHLESER.

1975. Diagnostic procedures for diseases found in egg, lar-
val, and juvenile cultured American lobsters [Homarus
americanus). Proc. 6th Annu. Workshop World
Maricult. Soc, Seattle, Wash., 1975.
Greenhood, E. C, and D. J. MACKETT.

1965. The California marine fish catch for 1964. Calif. Fish
Game, Fish Bull. 132, 45 p.

1967. The California marine fish catch for 1965. Calif.
Fish Game, Fish Bull. 135:1-42.

Heimann, R. F. G., and H. W. Frey.

1968a. The California marine fish catch for 1966. Calif.

Fish Game, Fish Bull. 138:1-48.
1968b. The California marine fish catch for 1967. Calif

Fish Game, Fish Bull. 144, 47 p.

Heimann, R. F. G., and J. G. Carlisle, jr.

1970. The California marine fish catch for 1968 and histori-
cal review 1916-1968. Calif. Fish Game, Fish Bull. 149,
70 p.

JOHNSON, S. K., J. C. Parker, and H. Holcomb.

1974. Control of Zoothamnium sp. on penaeid
shrimp. Proc. 4th Annu. Workshop World Maricult. Soc,
Monterrey, Mexico, 1973.

KURIS, A.

1973. Population interactions between a shore crab and two
symbionts. Ph.D. Thesis, Univ. California, Berkeley.
LIGHTNER. D. v., C. T. FONTAINE, AND K. HANKS.

1975. Some forms of gill disease in penaeid shrimp. Proc.
6th Annu. Workshop World Maricult. Soc, Seattle, Wash.,
1975.

LOUGH, R. G.

1975. Dynamics of crab larvae (Anomura, Brachyura) off
the central Oregon coast, 1969-1971. Ph.D. Thesis, Ore-
gon State Univ., Corvallis, 229 p.
NILSON, E. H., W. S. FISHER, AND R. A. SHLESER.

1975. Filamentous infestations observed on eggs and lar-
vae of cultured crustaceans. Proc. 6th Annu. Workshop
World Maricult. Soc, Seattle, Wash., 1975.
OLIPHANT, M. S.

1973. California marine fish landings for 1971. Calif Fish
Game, Fish Bull. 159, 49 p.

Pacific Marine fisheries Commission.

1965. Discussion that followed the report on Dungeness
crabs. 16th and 17th Annu. Rep. Pac. Mar. Fish. Comm.,
p. 38-39.

Peterson, W. T.

1973. Upwelling indices and annual catches of Dungeness
crab, Cancer magister, along the west coast of the United
States. Fish. Bull., U.S. 71:902-910.

206

PINKAS, L.

1970. The California marine fish catch for 1969. Calif Fish
Game, Fish Bull. 153, 47 p.
POOLE, R.

1962. Cruise report 62-N-2g, h, i and 1 crab. Calif. Dep.
Fish Game, Mar. Resour. Oper.
POOLE, R., AND D. GOTSHALL.

1965. Regulations and the market crab fishery. Outdoor
Calif. 26(9):7-8.
REID, J. L., JR., G. I. RODEN, AND J. G. WYLLIE.

1958. Studies of the California Current System. Calif.
Coop. Oceanic Fish. Invest. Rep. 1 July 1956 - 1 January
1958, p. 27-56.
ROGERS-TALBERT, R.

1948. The fungus Lagenidium callinectes Couch (1942) on
eggs of the blue crab in Chesapeake Bay. Biol. Bull.
(Woods Hole) 94:214-228.

Sandoz, M. D., R. Rogers, and C. L. Newcombe.

1944 Fungus infection of eggs of the blue crab Callinectes
sapidus Rathbun. Science (Wash., D.C.) 99:124-125.

WILLIAM S. FISHER

Department of Food Science and Technology
University of California
Davis, CA 95616

Daniel E. Wickham

Department of Zoology
University of California
Berkeley, CA 94620

SECOND RECORD OF BLACK SKIPJACK,

EUTHYNNUS LINEATUS,

FROM THE HAWAIIAN ISLANDS

Matsumoto and Kang (1967) reported the first
capture of the black skipiack, Euthynnus lineatus
Kishinouye, in the Hawaiian Islands. Recently
(14 July 1975), a second black skipjack was taken
in these waters by a Hawaiian pole-and-line skip-
jack tuna fishing vessel, the Mar/m, skippered by
Walter Asari. The fish was noticed by a fish re-
ceiver at Hawaiian Tuna Packers, Richard How-
ell, who contacted Robert T. B. Iversen, South-
west Region Representative stationed at the
Southwest Fisheries Center Honolulu Laboratory.
Iversen brought the fish to me for identification.

The specimen, 454 mm fork length, and weigh-
ing 1.53 kg, was caught from a school of small
skipjack tuna, Katsuwonus pelamis, at the ex-
treme tip of Penguin Banks, about 40 km south of
the eastern end of Oahu. The specimen is de-
posited in the U.S. National Museum collection
(USNM 214683).

Measurements in millimeters taken according
to the methods described by Grodsil and Byers

(1944) are as follows: Fork length - 454; head
length - 126; 1st dorsal insertion - 144; 2d dorsal
insertion - 271; anal fin insertion - 306; ventral
fin insertion - 144; greatest body depth - 112;
greatest body width - 73; dorsal-ventral distance -
108; dorsal-anal distance - 188; ventral insertion
to vent - 160; length 1st dorsal base - 130; length
2d dorsal base - 29; length anal base - 25; length
pectoral - 70; height 1st dorsal - 61; height 2d
dorsal - 28; height anal - 28; diameter of iris - 19;
maxillary length - 50; snout to posterior margin
of eye - 54.

Counts: 1st dorsal spines - 14, plus 1 imbedded;
2d dorsal rays - 12; dorsal finlets - 8; anal
rays - 12; anal finlets - 7; pectoral rays - 26; gill
rakers - left side 9 -H 1 -h 24 = 34, right side 9 +
1 + 25 = 35.

The external characters agree with that of the
previous capture (Matsumoto and Kang 1967)
and with Godsil's (1954) description of the
species. Five black unbranched stripes run paral-
lel to the longitudinal axis of the body on the back
fi'om the corselet to the caudal fin, and five or six
faint unbranched stripes run horizontally on the
belly. Two black thoracic spots are located on each
side at the indentation of the corselet near the
ventral margin of the body.

The vertebral count is 20 + 17 = 37. As in the
previous capture, four large protuberances are
present on the 31st vertebra, a characteristic of
this species (Godsil 1954).

Although this is only the second specimen re-
corded, an interview with the skipper of the ves-
sel disclosed that fish similar to this are often
caught but are not reported. The question posed
in 1967 as to whether this is a chance migrant
from the eastern Pacific Ocean still stands.

Literature Cited

GODSIL, H. C.

1954. A descriptive study of certain tuna-like fishes. Calif.
Dep. Fish Game, Fish Bull. 97, 185 p.

Godsil, H. C, and R. D. Byers.

1944. A systematic study of the Pacific tunas. Calif Dep.
Fish Game, Fish Bull. 60, 131 p.

Matsumoto, W. M., and T. Kang.

1967. The first record of black skipjack, Euthynnus
lineatus, from the Hawaiian Islands. Copeia 1967:837-
838.

Walter M. matsumoto

Southwest Fisheries Center Honolulu Laboratory
National Marine Fisheries Service, NOAA
Honolulu, HI 96812

207

OPTICAL MALFORMATIONS INDUCED BY

INSECTICIDES IN EMBRYOS OF THE
ATLANTIC SILVERSIDE, MENIDIA MENIDIA

Since the banning of DDT from use in the United
States, other insecticides such as malathion,
parathion, and Sevin^ (carbaryl) have come into
greater use. Though not persistent like DDT,
these insecticides, Uke DDT, find their way into
aquatic ecosystems and thus into the spawning
grounds of aquatic organisms. Various insec-
ticides have been shown to cause developmental
abnormalities. Malathion, for example, has been
shown to cause skeletal malformations in birds
(McLaughlin et al. 1963; Walker 1967; Greenberg
and LaHam 1969), mammals (Tanimura et al.
1967), and reptiles (Mitchell and Yntema 1973).
The experiments described herein were de-
signed to study the effects of DDT, malathion, and
Sevin on the development of the Atlantic silver-
side, Menidia menidia. Since previous studies had
all indicated that sensitivity decreases with em-
bryonic age, we initiated our treatment early in
development.

Materials and Methods

Adult M. menidia, from the vicinity of Mon-
tauk, N.Y., were collected by a seine during June
and July. Eggs and sperm were obtained by
stripping the fish, as described by Costello et al.
(1957:228-233). The fertilized eggs were sep-
arated into small clumps and, after being washed,
were placed randomly in glass finger bowls in 100
ml of Millipore-filtered seawater (salinity 30%)
and incubated at 20°C. The insecticides malath-
ion (95% analytical reagent, Supelco Inc., Belle-
fonte, Pa.), DDT (p,p'-DDT, 729c technical grade,
Montrose Chemical Co., Torrance, Calif., recrys-
tallized from ethanol to yield 98% p,p'-DDT), and
Sevin (99.2% carbaryl. Union Carbide Corp., New
York, N.Y.) were introduced as acetone solutions
into experimental dishes during either early
cleavage (2-4 cell stage) or late cleavage (about
100 cells— see Costello et al. 1957, fig. 104), at
concentrations of 10 to 500 parts per billion (ppb).
Control dishes received an equivalent amount of
acetone (10 /xl). The solutions were not changed;
thus we were studying the effect of a single appli-
cation of the chemicals (the concentration of

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

which undoubtedly decreased over time due to
adsorption). Development was followed with ref-
erence to the descriptions of Costello et al. (1957).
At appropriate times, eggs were examined to see
the percentage which had successfully completed
gastrulation and, later, the percentage which had
successfully initiated heartbeat. In the first two
experiments hatching rates were noted and only
the newly hatched fry were examined for mal-
formations. Since they appeared normal, in the
subsequent experiments embryos were examined
for malformations with considerably more suc-
cess. Some embryos were preserved in glutaral-
dehyde, dechorionated, sectioned, and stained
with hematoxylin and eosin.

A repeat experiment was performed in the fol-
lowing summer using the same procedures.

Results

In the first experiment, eggs were treated at
the late cleavage stage with malathion at 10 and
100 ppb and Sevin at 25 and 100 ppb. There were
over 200 eggs in each dish. Percents of successful
axis formation and heartbeat initiation were
lower than controls in most treated groups (Table
1) but did not always show a dose-related effect.
Hatching commenced 14 days after fertilization
and continued for 6 days, at which time the ex-
periment was terminated. No difference was
noted in hatching times in the various groups and
no abnormalities were observed in the fry, al-
though some dead ones were seen in each group.

In the second experiment, eggs at the 2-4 cell
stage were exposed to DDT at 25 and 100 ppb and
to malathion at 10 and 100 ppb. There were again
about 200 eggs in each dish. As in the previous
experiment (Table 1) treated groups had lower
rates of axis formation and of heartbeat initiation
than controls. Hatching commenced 14 days after
fertilization and continued for 6 days, at which
time the experiment was terminated. No differ-
ence was noted in hatching times in the various
groups and no abnormalities were noted in the
fry, although, as before, some dead ones were
noted in each group.

In the third experiment, eggs at the late cleav-
age stage were exposed to DDT at 10, 25, and 100
ppb, malathion at 10, 100, and 500 ppb, and Sevin
at 25, 100, and 500 ppb. There were about 50 eggs
in each dish. When checked for axis formation
and heartbeat initiation, the treated eggs were
again lower than controls. Embryos were care-

208

Table l. — Insecticide effects on percentage of axis formation, heartbeat, optic abnormalities, and hatching.

Concentrations in parts per billion (ppb).

Control

DDT (ppb)
25 100

Malathi

ion (ppb)

Sevin (ppb)

Item

100

500

1,000

2,500

100

500

Experiment 1

(late cleavage):

Axis formation

Heartbeat

Hatch

Experiment 2

(2-4 cell stage):

Axis formation

Heartbeat

Hatch

Experiment 3

(late cleavage):

Axis formation

Heartbeat

Optic anomalies

Experiment 4

(late cleavage):

Axis formation

Heartbeat

Optic anomalies

Experiment 5

(late cleavage):

Axis formation

Heartbeat

Optic anomalies

fully examined for developmental abnormalities,
and various optic malformations were discovered
in the insecticide-treated embryos. These took the
form of unilateral and bilateral microphthalmia
(reduced size of eyes), unilateral and bilateral
anophthalmia (absence of eyes), and cyclopia (a
single median eye) (Figure 1). Severely retarded
embryos were also noted. Percentages of those
with successful axis formation which showed op-
tical abnormalities were quite high in all treated
groups, while none were observed in the control
group. Abnormal embryos were fixed prior to

hatching. (It was assumed that they would die
prior to hatching since no abnormal fry had been
found in the previous experiments.) At hatching,
which commenced 15 days after fertilization and
continued for 7 days, one fish with scoliosis was
noted in 10 ppb malathion.

In the fourth experiment, eggs were again ex-
posed at the late cleavage stage to DDT at 10, 25,
and 100 ppb, malathion at 10, 25, and 100 ppb,
and Sevin at 10 and 25 ppb. There were about 200
eggs in each dish. When checked for axis forma-
tion and heartbeat, treated groups were lower

FIGURE 1.— Photomicrographs of whole, fixed, 2-wk-o\d Menidia menidia embryos at approximately 20 x. A is a control embryo,
while B is a 10 ppb Sevin-treated embryo with unilateral anophthalmia (the site of the undeveloped eye is marked by X), and C is
a cyclopic embryo from a 10 ppb malathion-treated batch (transmitted light illuminates the single lens at L).

209

than controls. At this time, and for several days
after, abnormal embryos were noted. These in-
cluded the severely retarded embryos and the
optical abnormalities noted earlier. Only one
control embryo showed slight microphthalmia.
Hatching commenced after 11 days and continued
for 9 days, at which time the experiment was
terminated. After hatching, lordotic fry were seen
in the 10 ppb malathion, 10 ppb Sevin, and 25 ppb
DDT groups. These skeletal abnormalities were
quite rare, however.

Eye diameters of hatched fry were measured
with an ocular micrometer to see if there were
slight reductions in optic size in the apparently
normal specimens, but no difference between ex-
perimental and control fry was seen.

The fifth experiment was performed the follow-
ing summer using about 100 eggs per dish. Eggs
were exposed at late cleavage to DDT at 10, 25,
and 100 ppb, Sevin at 10, 25, and 100 ppb, and
malathion at 1 and 2.5 ppm. Treated groups were
again lower than controls in rate of axis forma-
tion and heartbeat initiation. Abnormal embryos
were seen in most treated groups (Table 1) and all
embryos which exhibited optic malformations
also showed retardation, stunting of growth,
sparse body pigment, and abnormal cardiac de-
velopment in which the heart remained a very
thin, feebly beating tube without differentiation
of the chambers. There were also embryos with
this syndrome in which the eyes appeared nor-
mal. Hatching commenced after 12 days, and sev-
eral fry with scoliosis were seen in the mala-
thion dishes.

Discussion

The three insecticides reduced survival of
Menidia embryos, although this reduction was
not always correlated with the dose and varied in
different batches of eggs. The main embryotoxic
effect was at early stages, preventing successful
axis formation. Of those which formed axes, most
went on to establishment of heartbeat.

Notable optic malformations were observed in
embryos exposed to DDT, malathion, and Sevin.
These three insecticides are quite different from
each other chemically, and the fact that they all
produced similar malformations may indicate
that this species has a propensity toward this
type of malformation and various agents can in-
voke them. This propensity is supported by the
presence of one control embryo with slight mi-

crophthalmia in one eye. McEwan et al. (1949)
likewise concluded that the jewelfish, Hemi-
chromis bimaculata, had a tendency to vary ab-
normally in certain directions and that an ab-
normal environment accentuated this tendency.
The most common optic abnormalities seen in our
fish were unilateral anophthalmia and microph-
thalmia. True cases of cyclopia were rare, though
several embryos showed partial convergence of
the eye cups, with optic cups directed somewhat
ventrally rather than laterally.

Stockard (1907) produced cyclopia in Fundulus
embryos by treatment with MgCl2. In another
study (1910) he produced cyclopean, anophthal-
mic, and monophthalmic Fundulus embryos after
treatment with alcohol, results similar to those in
the present study.

Histological examination of our material re-
vealed a case in which the optic cup had partly
formed, but appeared to be facing inward rather
than outward and had lost its connection to the
brain. No lens was present in this specimen.
Smithberg (1962) found that tolbutamide caused
eye malformations in the medaka, Oryzias
latipes. However, these malformations involved
degeneration of the eye cup after the lens had
been formed, and lenses were present in all the
abnormal embryos. These malformations were
accompanied by circulatory defects, which he
considered responsible for the eye defects.

Retardation of development was seen by Battle
and Hisaoka ( 1952) in their studies of effects of
ethyl carbamate (urethan) on embryos of the
zebrafish, Brachydanio rerio. Some of their em-
bryos also exhibited optical malformations in-
cluding anophthalmia, microphthalmia, and
cyclopia. In Hisaoka's subsequent study (1958) of
2-acetylaminofluorene on zebrafish embryos,
microphthalmia was one abnormality produced
by this carcinogen. The antibiotic chloram-
phenicol was found by Anderson and Battle
(1967) to cause a variety of teratogenic effects in
zebrafish, including cyclopia and intermediate
stages leading to this condition. Colchicine was
likewise found by Waterman (1940) to cause a
variety of anomalies in the medaka, including
cyclopia.

Aside from general retardation, the optic mal-
formations were the major teratological effect of
the insecticides on Menidia the first year.
Skeletal malformations were also noted but they
were relatively rare. In the following year, a vari-
ety of malformations in addition to the optic ab-

210

normalities were produced. This difference is
perplexing, and is probably due to genetic dif-
ferences among individuals of this species in
susceptibility to the chemicals. This is more
understandable when it is realized that rela-
tively few females can supply all the eggs needed
for an entire experiment. Such variability in
response makes this species a poor one to use in
teratological studies.

Effects were seen at dosages as low as 10 ppb.
These are levels far lower than those which pro-
duced noticeable effects in Fundulus heteroclitus
embryos (Weis and Weis 1974) in which it was
necessary to increase the dosage to parts per mil-
lion. This may be due to differential permeability
of the chorions of the two species and/or to a
higher general resistance of Fundulus which is
generally considered a hardier fish than Menidia.
The dose levels which affected Menidia are levels
near those which have been found temporarily, in
solution or suspension, in natural areas (Finley et
al. 1970; Kennedy and Walsh 1970). Therefore
these adverse effects could occur during the de-
velopment of fish embryos in nature.

Acknowledgments

We thank John C. Baiardi, Director of the New
York Ocean Science Laboratory for making
facilities available to us. We are especially grate-
ful to Eugene Premuzic and John P. Wourms for
sharing their laboratory and advice. Thanks are
also extended to S. H. Gilani for reviewing our
manuscript.

Literature Cited

Anderson, p. d., and h. I. Battle.

1967. Effects of chloramphenicol on the development of the
zebrafish, Brachydanio rerio. Can. J. Zool. 45:191-205.
BATTLE, H. I., AND K. K. HISAOKA.

1952. Effects of ethyl carbamate (urethan) on the early
development of the teleost Brachydanio rerio. Cancer
Res. 12:334-340.
COSTELLO, D. P., M. E. DAVIDSON, A. EGGERS, M. H. FOX, AND
C. HENLEY.

1957. Methods for obtaining and handling marine eggs
and embryos. Mar. Biol. Lab., Woods Hole, Mass., 247 p.
FINLEY, M. T., D. E. FERGUSON, AND J. L. LUDKE.

1970. Possible selective mechanisms in the development of
insecticide-resistant fish. Pestic. Monit. J. 3:212-218.
GREENBERG, J., AND Q. N. LAHAM.

1969. Malathion-induced teratisms in the developing
chick. Can. J. Zool. 47:539-542.

HISAOKA, K. K.

1958. The effects of 2-acetylaminofluorene on the em-
bryonic development of the zebrafish. I. Morphological
studies. Cancer Res. 18:527-535.
KENNEDY, H. D., AND D. F. WALSH.

1970. Effects of malathion on twfo warmwater fishes and
aquatic invertebrates in ponds. U.S. Bur. Sport Fish.
Wild!., Tech. Pap. 55, 13 p.
MCEWAN, R. S., J. B. BRIGGS, AND M. S. GILBERT.

1949. The effects of various abnormal agents applied to

early developmental stages of Hemichromis bimaculata,

and their theoretical significance. Anat. Rec. 105:491

(abstr.)

MCLAUGHLIN, J., Jr., J.-P MARLIAC, M. J. VERRETT, M. K.

MUTCHLER, AND O. G. FiTZHUGH.

1963. The injection of chemicals into the yolk sac of fertile
eggs prior to incubation as a toxicity test. Toxicol. Appl.
Pharmacol. 5:760-771.
MITCHELL, J. T, AND C. L. YNTEMA.

1973. Teratogenic effect of malathion and captan in the
embryo of the common snapping turtle, Chelydra serpen-
tina. Anat. Rec. 175:390.
Smithberg, M.

1962. Teratogenic effects of tolbutamide on the early de-
velopment of the fish, Oryzias latipes. Am. J. Anat.
111:205-213.
STOCKARD, C. R.

1907. The artificial production of a single median Cyclo-
pean eye in the fish embryo by means of sea water solu-
tions of magnesium chloride. Arch. Entwicklungsmech.
Org. (Wilhelm Roux) 23:249-258.

1910. The influence of alcohol and other anaesthetics on
embryonic development. Am. J. Anat. 10:369-392.
TANIMURA, T, T KATSUYA, and H. NISHIMURA.

1967. Embryotoxicity of acute exposure to methyl para-
thion in rats and mice. Arch. Environ. Health
15:609-613.
WALKER, N. E.

1967. Distribution of chemicals injected into fertile eggs
and its effect upon apparent toxicity. Toxicol. Appl.
Pharmacol. 10:290-299

Waterman, A. J.

1940. Effect of colchicine on the development of the fish
embryo, Oryzias latipes. Biol. Bull. (Woods Hole)
78:29-34.
WEIS, P., AND J. S. WEIS.

1974. Cardiac malformations and other effects due to in-
secticides in embryos of the killifish, Fundulus hetero-
clitus. Teratology 10:263-267.

JUDITH S. WEIS

Department of Zoology and Physiology
Rutgers University
Newark, NJ 07102

peddrick Weis

Department of Anatomy

College of Medicine and Dentistry of New Jersey

Newark, NJ 07103

211

GOOSE BARNACLES

(CIRRIPEDIA: THORACICA)

ON FLOTSAM BEACHED AT

LA JOLLA, CALIFORNIA

The macroscopic floating biota of the ocean
surface — the pleuston — has been comparatively
httle studied (see review by Cheng 1975). It com-
prises a few species of insects, which skim over the
surface; a few species of siphonophores equipped
with floats; a few species of barnacles; etc. These
organisms can be collected by the use of special
nets towed at the level of the ocean surface, but the
numbers of such tows made on oceanographic ex-
peditions have been comparatively few compared
with the much larger numbers of plankton tows
made below the sea surface. Under exceptional
circumstances, when an onshore wind blows for an
extended period, pleustonic organisms are cast
ashore in appreciable numbers, presenting un-
usual opportunities to study numbers of individu-
als of this little known community. Such mass
beachings of the siphonophores Physalia and Ve-
lella have been reported in several parts of the
world (Bingham and Albertson 1974; Cheng
1975). This paper presents some data on a mass
beaching of pleustonic goose barnacles, mostly at-
tached to floating objects and mostly still living,
found washed ashore between 5 and 9 July 1974,

in front of the Scripps Institution of Oceanog-
raphy, La Jolla, Calif.

Methods

A stretch of beach approximately 1 km long and
5 m wide was searched systematically for five
successive days, around the time of the low tide in
daylight, and every barnacle or piece of flotsam
bearing barnacles was collected, taken to the
laboratory in plastic bags, and there kept in tanks
with running seawater. Some observations were
made on the living animals, which remained alive,
feeding actively, for several days, and specimens
were photographed (Figure lA-F). They were
sorted according to substrate, the species were
identified, and the lengths of the capitula were
measured from base of scutum to apex of tergum
(peduncle lengths being variable).

Observations

In all, some 329 substrate objects were collected
and examined; they bore a total of 2,555 individual
barnacles. The data, for all collections, are sum-
marized in Tables 1 and 2, and the size distribu-
tions of each species on each of the major substrate
types are shown in Figure 2A-L. The following
generalizations were made on the basis of this
material.

Table l. — Numbers and percentages of substrates bearing barnacles: Lepas
(Dosima) fascicularis and Lepas (Lepas) pad fica.

Total

Number

Dosima +

Lepas

Dosima
No. %

Lepas

% of total
Dosima

specimens

Substrates

No.

Lepas

Feathers

878

657

221

Sea grass leaves:
Phyllospadix
Zostera

537
373

Subtotal

910

835

Brown algae;
Macrocystis

202

117

Colpomenia
Egregia
Hall dry s
Sargassum
Scytosiphon

18
3

55
6
2

Subtotal

Terrestrial debris:

Wood

Peanut shells

Plastic straws

Cigarette filters

Subtotal

Tar lumps
None

322

113
61

35
82

209
13

65
18

6
3

33
2

Total

2,555

1,913

642

100

212

wmmmt^

>',:o-

-tiSl, .B

Figure l. — Lcpas (Dosima) fascicularis and Lepas (Lepas) pacifica, living specimens photographed in aquarium. A. Specimens of
Dosima supported by their own floats at the water surface (note young barnacles attached to specimen on left); B. Right-hand specimen,
from Figure lA, showing cirri withdrawn; C. Dosima on detached float of Macrocystis; D. Dosima on feather; E. Goose barnacles,
mostly Lepas, on piece of Macrocystis; F. Small specimens of Lepas on flat lump of tar.

213

Table 2. — Numbers and percentages of barnacles [Lepas (Dosima) fascicularis and Lepas

(Lepas) pacifica] on various substrates.

Dosima only

Lepas

only

Dosima

+ Lepas

Total of substrates

Substrates

No.

Feathers

45.5

46.5

Sea grass leaves:

Phyllospadix

Zostera

Subtotal

122

Brown algae:

Macrocystis

Colpomenia

Egregia

Halidrys

Sargassum

Scytosiphon

Subtotal

Terrestrial debris:

Wood

Peanut shells

Plastic straws

Cigarette filters

Subtotal

Tar lumps

Total

193

329

100

The most common barnacle-bearing substrate
was foimd to be bird feathers (90 items). The next
most common were leaves of the surfgrass Phyl-
lospadix (86 pieces) and tar (57 lumps). Other sub-
strates included bits of brown algae Colpomenia,
Egregia, Halidrys, Macrocystis, Sargassum,
Scytosiphon; leaves of the sea grass Zostera; pieces
of wood; cigarette filters; peanut shells; and plastic
drinking straws. (Pieces of other debris without
barnacles, such as polystyrene cups and plastic
bottles and caps — many clearly of local origin —
were not collected and are not further discussed
here.)

Most of the barnacles belonged to two species:
Lepas (Dosima) fascicularis Ellis and Solander,
the soft blue barnacle (about 75% of the individu-
als); and Lepas (Lepas) pacifica Henry, a common
Pacific goose barnacle (about 25%). Two other
species of barnacle were also found: three speci-
mens of Tetraclita squamosa on pieces of Mac-
rocystis stipe, and one young specimen oi Lepas
(Lepas) anatifera on a piece of tar. These have not
been included in the data of Tables 1 and 2, and
will not be considered further.

Unattached (Figure lA, B)

An appreciable number of the Dosima speci-
mens (61) were found unattached to flotsam,
either occurring singly, each with its own float, or
else with several specimens sharing a communal

float. Whether these had previously been attached
to any substrate was not determined. The 13 unat-
tached Lepas specimens found in our collections
had probably become detached from substrates
after they were collected.

Feathers (Figures ID, 2 A, B)

The feathers bearing barnacles were mostly
large, more than 10 cm long, and were relatively
intact with both quill and vanes. Most were white
or grey; the species of seabirds from which they
originated were not identified. Though a few of the
barnacles were attached singly along the shaft,
most occurred in clusters, generally near the dis-
tal end of the feather. Such clusters comprised as
many as 20 individuals of different sizes, many or
all of which must have contributed to the com-
munal bubble floats which in some specimens
reached a diameter of almost 20 mm. The largest
Dosima specimen found on a feather was 20 mm
long; the largest Lepas, only 13 mm. About 50% of
the feathers bearing barnacles had only Dosima
specimens; only seven (7.8%) were found carrying
Lepas alone, and on all of these the barnacles were
rather small and few. On the feathers that carried
a mixture of both species, the majority of the ani-
mals were Dosima; in fact, some 18 of the Lepas
specimens (all less than 10 mm) were found at-
tached to the larger individuals of Dosima. The
highest cluster numbers found on single feathers

214

Dosima

FEATHERS

Lepos

° lOOn
a.

SEA GRASS LEAVES

MACROCYSTIS

Figure 2. — Size-frequency distributions of Lepas (Dosima)
fascicularis and Lepas (Lepas) pacifica on various substrates
as indicated.

were 34 for Dosima and 15 for Lepas. Thirty-six of
the Dosima clusters consisted of more than 10
individuals, whereas only six of the Lepas clumps
on feathers comprised more than 10 animals.
From these data it appears that on feathers
Dosima is much commoner than Lepas and can
occur more densely and in larger clumps, presum-
ably because of its ability to produce its own float.

Sea Grass Leaves (Figure 2C, D)

Many of the Phyllospadix and Zostera leaves
bearing barnacles had been completely bleached;

possibly they had become detached from the par-
ent plants and had drifted out to sea before being
colonized. The majority of the leaf sections col-
lected were found to carry one or more specimens
of Dosima. Almost 809^ carried only Dostma; only
3% bore Lepas alone; the rest had both. As in the
case of the feathers, the Dosima specimens at-
tached to leaves had produced their own floats, as
many as 23 individuals being found in one cluster.
The largest specimens of Dosima found on Phyl-
lospadix andZostera were 22 mm and 19 mm long,
respectively. In contrast, on these substrates the
Lepas individuals generally occurred either singly
or in pairs, and the majority of these animals did
not measure more than 5-6 mm in length, though a
few of those which occurred together with Z)ostma
exceeded 10 mm. Presumably, larger specimens of
Lepas cannot be supported by a floating leaf sec-
tion unless additional buoyancy is supplied by
floats of Dosima.

Brown Algae (Figures IC, E, 2E-H)

It is significant that the only algae found bear-
ing barnacles are parts of brown algae
(Phaeophyta), which either produce well-
differentiated gas-filled floats or, as in the cases of
Colpomenia and Scytosiphon, have hollow thalli
usually filled with air. The majority of the barna-
cles were found on float-bearing segments of Mac-
rocystis, and in Tables 1 and 2 the data for this alga,
which occurs in offshore waters, are presented
separately from those of other brown algae, which
are more or less intertidal. Since none of these
algae normally carry goose barnacles while grow-
ing in their natural habitats, it appears probable
that the pieces of thallus were colonized by barna-
cles after they had been detached. They must have
floated for some time, however, since the barnacles
had reached appreciable sizes: up to 21 mm in
length for Dosima and up to 12 mm in length for
Lepas. With the exception of one piece of Egregia
bearing a small 2-mm Lepas, the littoral brown
algae bore only Dosima (83 specimens in all),
whereas a large proportion of the Macrocystis
pieces bore mixed populations.

Terrestrial Debris (Figure 21, J)

The majority of the fragments grouped in this
category were pieces of wood, which may be con-
sidered a "natural" substrate since fallen branches
are a normal component of the flotsam carried

215

by rivers out to sea. So far, plastics — in pieces
sufficiently large and buoyant to support goose
barnacles — evidently constitute a substrate of
only minor importance for this kind of animal.

Tar (Figures IF, 2K, L)

The 57 pieces of barnacle-bearing tar, presuma-
bly originating from natural seepage or oil bun-
kers, were mostly flattened 2-3 mm thick, 10-60
mm in diameter. This substrate, unlike those de-
scribed hitherto, appeared to be preferred by
Lepas. More than 42% of the lumps collected bore
only this species, and many of the pieces had more
than 10 animals attached. About 65% of the bar-
nacles found on tar were of this species. Some were
more than 15 mm long. They were generally not
clumped, but occurred scattered over the surface of
the substrate, often on both upper and under sur-
faces, suggesting that the lump had repeatedly
turned over while afloat on the ocean. Compara-
tively fewer of the tar lumps bore only specimens
ofDosima, and only 10 of these had more than 10
animals each. Per unit of surface area, the indi-
viduals of Dosima appeared to be more sparsely
distributed on tar than on feathers or grass leaves.

Discussion

Lepas (Dosima) fascicularis is the most
specialized pleustonic goose barnacle, with an al-
most uncalcified shell and a gas-filled bubble float.
The larval stages were described on the basis of
material collected and reared during the Chal-
lenger Expedition (Willemoes-Suhm 1876). Since
there were several errors and omissions in that
paper, all the stages were redescribed by Bain-
bridge and Roskell (1966).

Boetius (1952-53) reported that all of the speci-
mens of Dosima, which he found on the Danish
North Sea coast in September 1952, had floats
roughly proportional in diameter to the length of
the animal. These barnacles are able to support
themselves in the adult stage by their own float,
but the cyprid larvae must settle on some sub-
strate before they can metamorphose. The larvae
of Dosima have been shown to settle preferen-
tially on small floating objects; only later do they
produce a bubble float which enables them to stay
at the sea surface even when detached from such a
support (Boetius 1952-53; Newman 1974). In our
collections, all of the Dosima specimens, but none
of the Lepas specimens, were attached to bubble

floats of their own making. Some 27 individuals of
Lepas (1-10 mm), the smaller of the two species,
were found attached to larger specimens of
Dosima, but, despite their larger absolute num-
bers, only 8 Dosima specimens (1-14 mm) were
found on other animals of this species. Evidently
floating barnacle colonies do not normally grow by
accretion in this way.

The blue pigment of Dosima was studied by Fox
et al. (1967), who reported that it is a conjugated
carotenoid. Although many of the blue barnacles
which they studied (washed ashore in the same
location) were found attached to the floats of Ve-
lella, and although we have found large numbers of
these siphonophores stranded at various other
times in recent years, we found no Velella floats
among the barnacle substrates in this study. In
fact, although hundreds of pleustonic barnacles
were stranded on our beach during the period
studied, we found no specimen of Physalia, Velella ,
or lanthina, which are all common components of
the pleuston community in the open ocean. We
found only one Glaucus (a pelagic nudibranch), a
few specimens of Fiona (another nudibranch,
normally associated with Macrocystis), and sev-
eral polychaete worms. This probably indicates
the relatively nearshore rather than oceanic ori-
gin of the barnacle colonies. Although, when
brought back to a laboratory aquarium and given
fresh running seawater, many of the specimens
remained alive and apparently healthy for more
than 1 wk, such stranded animals are normally
unable to return to the sea. When exposed to the
sun on the beach they would probably be eaten by
gulls or dry up within a few hours.

We have not attempted to study the gut contents
of our animals but assume that, like other barna-
cles in nature, they probably feed mainly on
microorganisms and small zooplankton (Howard
and Scott 1959; Crisp and Southward 1961). We
noted that in the laboratory, when supplied with a
suspension of the unicellular alga Platymonas ,
many individuals of Dosima extended their cirri,
apparently moving them towards the food source,
directing it towards the mouth.

Goose barnacles are hermaphrodites. Adults
develop both male and female organs at the same
time and can cross-fertilize each other. The eggs
are brooded in the mantle cavities, and hatch as
larvae which live in the plankton before settling.
They attach themselves to a solid substrate by an
adhesive secreted by the cement glands; the com-
position of the cement of Lepas fascicularis has

216

been analyzed by Barnes and Blackstock (1974).
We do not know how long it takes for them to reach
the adult stage after metamorphosis. Horn et al.
(1970), who collected 150 specimens ofLepas pec-
tinata, 2-8 mm long, attached to four lumps of tar
found floating on the sea surface, noted that, in the
laboratory, these animals increased in length by
about 1 mm per week. The larger specimens in our
collections (20 mm for Dosima, 15 mm for Lepas)
contained mature eggs. We have no information
on the numbers of generations in the year; our size
distribution data (Figure 2) show no evidence for
separate generations (which might have been in-
dicated by distinguishable size-class modes).
Lepas species are known to be widely distributed
from tropical to polar seas. Our specimens proba-
bly came from populations floating in the eastern
Pacific Ocean, which is the most likely area af-
fected by the anomalous meterological conditions
occuring during June and July 1974 (J. Namias,
pers. commun.).

Summary

A total of 1,913 specimens of Lepas (Dosima)
fascicularis and 642 specimens of L. (Lepas)
paciftca, many still alive, were collected on a
1,000-m stretch of beach at La Jolla between 5 and
9 July 1974. They were attached to various sub-
strates which had enabled them to float at the sea
surface before being cast ashore. The predominant
substrates were feathers (90 pieces, bearing 657
Dosima, 221 Lepas), sea grass leaves (122 pieces:
835 Dosima, 75 Lepas), brown algae (39 pieces:
200 Dosima, 86 Lepas), and tar (57 pieces: 113
Dosima, 209 Lepas). Dosima is the predominant
species on most of the substrates whereas tar
lumps appeared to be preferentially settled by
Lepas. The size distributions (Dosima, 1-22 mm;
Lepas, 1-16 mm) provided no indications of gen-
erational discontinuities. The beaching of
these normally pleustonic animals should be
considered in relation to preceding and prevail-
ing wind conditions.

Acknowledgments

We thank Connie L. Fey for her patient help in
sorting and measuring the barnacles, James E.
Rupert and James R. Lance for taking the photo-
graphs, William A. Newman for specific iden-
tification of the barnacles and valuable comments

on the manuscript, and Jerome Namias for discus-
sion of relevent meteorological data. Financial
support from the Marine Life Research Group of
Scripps Institution of Oceanography and from the
Foundation for Ocean Research to Lanna Cheng is
also grate fill ly acknowledged.

Literature Cited

BAINBRIDGE, V., AND J. ROSKELL.

1966. A re-description of the larvae of Lepas fascicularis
Ellis and Solander with observations on the distribution
of Lepas nauplii in the north-eastern Atlantic. In H.
Barnes (editor), Some contemporary studies in marine
science, p. 67-81. George Allen & Unwin Ltd., Lond.

Barnes, H., and J. Blackstock.

1974. The biochemical composition of the cement of a
pedunculate cirripede. J. Exp. Mar. Biol. Ecol. 16:87-91.

Bingham, F. O., and H. D. Albertson.

1974. Observations on beach strandings of the Physalia
(Portuguese-man-of-war) community. Veliger 17:220-
224.

BOETIUS, J.

1952-53. Some notes on the relation to the substratum of
Lepas anatifera L. and Lepas fascicularis E. et S. Oikos
4:112-117.

Cheng, l.

1975. Marine pleuston — animals at the sea-air interface.
Oceanogr. Mar. Biol., Annu. Rev. 13:181-212.

Crisp, D. J., and a. J. southward.

1961. Different types of cirral activity of barnacles. Philos.
Trans. R. Soc. Lond., Ser. B, Biol. Sci. 243:271-307.
Fox, D. L., V. E. SMITH, AND A. A. WOLFSON.

1967. Disposition of carotenoids in the blue goose barnacle
Lepas fascicularis. Experientia 23:965-967.

Horn, M. H., J. M. Teal, and R. H. Backus.

1970. Petroleimi lumps on the surface of the sea. Science
(Wash., D.C.) 168:245-246.

Howard, G. K., and H. C. Scott.

1959. FVedaceous feeding in two common gooseneck barna-
cles. Science (Wash., D.C.) 129:717-718.

Newman, W. a.

1974. Cirripedia. Encyclopaedia Britannica, 15th ed.

4:641-643.
WILLEMOES-SUHM, R. VON.

1876. On the development of Lepas fascicularis and the

Archizoea of Cirripedia. Philos. Trans. R. Soc. Lond., Ser.

B, Biol. Sci. 166:131-154.

LANNA CHENG
RALPH A. LEWIN

Scripps Institution of Oceanography
University of California
La Jolla, CA 92093

217

CALORIC VALUES OF SOME
NORTH ATLANTIC CALANOID COPEPODS

Evaluation of the dynamics of energy exchange of
a marine ecosystem necessitates a knowledge of
the caloric equivalents of its living constituents.
This information, in combination with informa-
tion on growth, metabolism, and assimilation
rates can lead to predictions of energy conver-
sion between trophic levels and estimates of pro-
duction.

Researchers have accumulated a considerable
quantity of data concerning the caloric value of
marine organisms (Cummins 1967; Thayer et al.
1973; Tyler 1973); however, values recorded for
marine, planktonic copepod species have been few
(Slobodkin and Richman 1961; Comita et al. 1966;
Cummins 1967). My research reports the caloric
values for seven species of marine copepods, six of
which apparently have not been previously re-
corded. These studies are part of an overall inves-
tigation of the bioenergetics of the early life stages
of some North Atlantic fish species.

Materials and Methods

Plankton samples were collected in July and
August 1972 off Narragansett Bay, R.I. except for
samples of Pseudocalanus minutus which were
collected in April 1971 off the coast of Delaware.
All samples were preserved in 5% Formalin^ and
were prepared and combusted in July and August
1972. Laboratory preparation included rinsing the
samples in distilled water for 1 h, sieving through
a coarse mesh screen to remove large detritus, and
hand sorting adults of the various copepod species
under a dissecting microscope. Pure copepod
species samples were dried for 24 h at 90°C and
desiccated in a silica gel desiccator after which
they were made into pellets for combustion. All
combustion was done in a Parr 1241 automatic,
adiabatic calorimeter adapted for a microbomb.
Combustion samples for each copepod species were
done in triplicate. Percent ash for each copepod
species was determined by ashing uncombusted
pellets in triplicate at 500°C for 4 h in a muffle
furnace.

Results

Mean values for the caloric determinations of

the seven species of copepods (Table 1) were as
follows: 5,251.9 cal/g dry weight, 5,626.3 cal/g
ash-free dry weight, and 6.70% ash. Statistical
analysis of the means of caloric values for each
species (Duncan's New Multiple Range Test, Steel
and Torrie 1960) indicated that Calanus ftninar-
chicus had significantly higher values of both
calories per gram dry weight and calories per
gram ash-free dry weight than all other species,
that Temora longicornis had significantly lower
values for calories per gram ash-free dry weight
than all species except Centropages hamatus, and
that the differences between Acartia tonsa, Tor-
tanus discaudatus, P. minutus, Centropages
typicus, and C. hamatus were minimal (Table 1).
Temora longicornis had the highest percent ash.
Acartia tonsa and P. minutus also had relatively
high ash values in comparison with the other
species, while Calanus finmarchicus was inter-
mediate and higher than the three remaining
species (Table 1).

Table l. — Caloric and ash values for some North Atlantic
copepods. Species are recorded in order from largest to smallest
mean value under each category. Those species side-scored have
similar means (Duncan's New Multiple Range Test, P = 0.05).

Standard
Species Mean deviation

cal/g dry weight
{Calanus finmarchicus 6,425.1 ±187.0

Tortanus discaudatus
Centropages typicus
Acartia tonsa
Pseudocalanus minutus
Centropages hamatus

5,398.3
5,244.7
5,160.0
5,070.9

4,998.6

±14.6
±183.3

±78.8
±181.7
±246.3

Temora longicornis

4,466.3

±92.8

cal/g
Calanus finmarchicus

ash-free dry weight

6,835.2

±191.2

Acartia tonsa
Tortanus discaudatus
Pseudocalanus minutus
Centropages typicus

5,664.1
5,642.0
5,541.9
5,503.4

±86.6

±15.3

±198.6

±192.3

Centropages hamatus
Temora longicornis

5,212.3
4,984.7

±256.9
±103.6

Temora longicornis

% ash

10.40

±0.16

Acartia tonsa
Pseudocalanus minutus

8.90
8.50

±0.16
±0.11

Calanus finmarchicus

6.00

±1.82

Centropages typicus
Tortanus discaudatus
Centropages hamatus

4.70
4.32
4.10

tO.28
b0.07
t0.13

^Reference to trade names does not imply endorsement by
National Marine Fisheries Service, NOAA.

Discussion

Since the species in this study were preserved in
Formalin for short periods of time and rinsed in
distilled water to remove the Formalin before pro-
cessing, the estimates of caloric and ash content

218

and dry weight may have been slightly affected
due to an unknown loss of chemical constituents.
Methods of preservation of animals before com-
busting or determining chemical composition and
weights have been a subject of debate. Omori
(1970) showed there was considerable variation
with no apparent trend of chemical composition
and weight ofCalanus cristatus that were frozen,
dried, or preserved in Formalin. Except for dry
weight, which was lowest in Formalin-preserved
specimens, he found no clear relationship between
percent ash, carbon, nitrogen, and hydrogen com-
position and the methods of preservation. Faustov
and Zotin ( 1965) determined that fixing by drying
or in 4% Formalin had no significant effect on the
caloric value of fish embryos and, consequently,
results obtained with fresh or fixed material could
be directly compared. In the present study, sam-
ples of fresh and preserved (5% Formalin) C.
finmarchicus were compared. Calories per gram
dry weight and percent ash were less for the pre-
served sample, however, the differences were min-
imal (274.8 cal/g dry weight and 3.78% ash which
corresponds to 275.0 cal/g ash-free dry weight) and
only slightly greater than one standard deviation
(Table 1).

In view of the apparent lack of specific effects of
preservation method on chemical composition,
weights, and caloric values reported in the litera-
ture and the results with C finmarchicus in this
research, it may be concluded that the values pre-
sented in this paper are only slightly underesti-
mated, if at all. Also, since all samples in this
study were treated the same way, relative com-
parisons between them should be valid.

Attempts to explain the differences in caloric
values on the basis of phylogeny proved in-
adequate. All species are calanoid copepods and,
although C. finmarchicus and P. minutus are
members of a different, more primitive taxonomic
subdivision under the Calanoida than the other
species (Sars 1903), the values for P. minutus
were statistically more similar to the lower values
for the other species than to C. finmarchicus.

There is a lack of information on the specific
chemical composition of the species tested in this
research with the exception of C finmarchicus.
Calanus finmarchicus is known to have a reasona-
bly high fat content. Comita et al. (1966) noted
that, upon fixation, globules of fat were extruded
from living specimens and that a layer of oil
formed on the surface of the fixed sample. They
determined the caloric value of the fat of C.

finmarchicus to be 9,500 cal/g. Fisher (1962) de-
termined the lipid content for a number of marine
Crustacea and found the concentrations in C
finmarchicus to be consistently among the higher
values recorded. Although there are no fat content
values for the six other species tested in this re-
search to compare with C. finmarchicus, the im-
plication is that the lipid content in C. finmarchi-
cus may be the cause of its higher caloric value.
The caloric determinations of C. finmarchicus
recorded in this research (Table 1) compare closely
with the results of other workers (Slobodkin 1962;
Comita and Schindler 1963; Comita et al. 1966). In
fact, the caloric values of C. finmarchicus have
been some of the highest recorded for copepods.

Temora longicornis had lower caloric values
than the other species and the highest percentage
of ash (Table 1). This may be the result of its
morphology which is somewhat different com-
pared to the other species. It has a proportionately
rounder and deeper cephalothorax that may con-
tribute to a higher percentage of inorganic exo-
skeleton.

The overall means for the caloric values of all
the species (5,251.9 cal/g dry weight and 5,626.3
cal/g ash-free dry weight) are similar to composite
sample caloric values recorded by other inves-
tigators. A calculation based on the data of Os-
tapenya et al. (1967) using their values of calories
per gram dry weight and percent organic matter
for Gulf of Mexico plankton samples, which were
predominantly copepods including Acartia sp.,
Centropages sp., and Temora sp. (separate values
for each of these genera were not reported), pro-
duced a mean value of 5,187 cal/g ash-free dry
weight. A similar confirming value of 5,016 cal/g
dry weight was obtained using the percent organic
matter in the dry material in my research (calcu-
lated by subtracting the mean percent ash, 6.70%,
from 100) and the regression relationship between
that and ash-fi-ee dry weight devised by Piatt et al.
(1969).

Seasonal changes in the caloric value of zoo-
plankton have been verified in several studies
(Comita et al. 1966; Conover 1968; Siefken and
Armitage 1968). The species in this study undoubt-
edly undergo seasonal variations also, and this is
a subject for future investigation. However, all the
species used in this research, with the exception of
P. minutus, were collected at approximately the
same time in the same general area and can be
used for a comparison of the potential energy avail-
able to predators at a particular time and place.

219

Examination of data on the abundance of adult
and nauplii stages in the Narragansett Bay and
Block Island Sound areas (Deevey 1952; Faber
1966) for the time of year samples for this research
were collected (July-August) showed that, al-
though all seven species were present, only A.
tonsa, T. longicornis, andC. hamatus were availa-
ble in sufficient quantity to be considered major
prey organisms. They represented 24.6, 10.8, and
10.4%, respectively, of the total copepods availa-
ble, while the other four species were less than 3%.
The results of this study in calories per gram ash-
free dry weight (Table 1) show that A. tonsa had
the second highest value while C. hamatus and T.
longicornis had the two lowest values. In fact, the
difference between A. tonsa and T. longicornis is
680 cal/g. This indicates, assuming equivalent as-
similation rates, that predators utilizing the
copepods like A. tonsa with higher caloric values
may have an advantage in acquiring energy for
growth and metabolic processes. Predators feed-
ing on copepods with lower values, especially T.
longicornis, would have to consume more prey or-
ganisms for an equivalent energy intake and,
given the same density of plankton, would spend
more energy searching for their prey.

Acknowledgments

I thank John B. Colton, Jr. for his critical re-
view of the manuscript and Stephen Hale for his
technical assistance.

Literature Cited

COMITA, G. W., AND D. W. SCHINDLER.

1963. Calorific values of microcrustacea. Science (Wash.,
D.C.) 140:1394-1396.
COMITA, G. W., S. M. MARSHALL, AND A. P. ORR.

1966. On the biology of Calanus finmarchicus . XIII. Sea-
sonal change in weight, calorific value and organic mat-
ter. J. Mar. Biol. Assoc. U.K. 46:1-17.

CONOVER, R. J.

1968. Zooplankton — Life in a nutritionally dilute environ-
ment. Am. Zool. 8:107-118.
Cummins, K. W.

1967. Calorific equivalents for studies in ecological ener-
getics. 2nd ed. Univ. Pittsburg, Pittsburg, 52 p.

DEEVY. G. B.

1952. Quantity and composition of the zooplankton of Block

Island Sound, 1949. Bull. Bingham Oceanogr. Collect.,

Yale Univ. 13:120-164.
Faber, D. J.

1966. Seasonal occurrence and abundance of free-swim-
ming copepod nauplii in Narragansett Bay. J. Fish. Res.
Board Can. 23:415-422.
FAUSTOV, V. S., AND A. I. ZOTIN.

1965. Changes in the heat of combustion of the eggs of fishes

and amphibians during development. Akad. Nauk SSSR

(Doklady) Biol. Sci. 162:965-968.
FISHER, L. R.

1962. The total lipid material in some species of marine

zooplankton. Rapp. P.-V. Reun., Cons. Perm. Int. Explor.

Mer 153:129-136.
OMORI, M.

1970. Variations of length, weight, respiratory rate and

chemical composition of Calanus cristatus in relation to

its food and feeding. In J. H. Steele (editor), Marine

food chains, p. 113-126. Univ. Calif. Press, Berkeley.
OSTAPENYA, A. P., L. M. SUSHCHENYA, AND N. N. KHMELEVA.

1967. Caloricity of plankton from the tropical zone of the
ocean. [In Russ., Engl, abstr.] Okeanol. Keanologiza
6:1100-1107.

PLATT, T., V. M. BRAWN, AND B. IRWIN.

1969. Caloric and carbon equivalents of zooplankton
biomass. J. Fish. Res. Board Can. 26:2345-2349.
SARS, G. O.

1903. An account of the Crustacea of Norway, Vol. 4, Cope-
poda Calanoida. Bergen Museum, Christiana, 171 p.
SIEFKEN, M., AND K. B. ARMITAGE.

1968. Seasonal variation in metabolism and organic nu-
trients in three Diaptomus (Crustacea: Copepoda). Comp.
Biochem. Physiol. 24:591-609.

SLOBODKIN, L. B.

1962. Energy in animal ecology. Adv. Ecol. Res. 1:69-101.
SLOBODKIN, L. B., and S. RICHMAN.

1961. Calories/gm in species of animals. Nature (Lond.)
191:299.
STEEL, R. G. D., and J. H. TORRIE.

1960. Principles and procedures of statistics. McGraw-Hill
Book Co., Inc., N.Y., 481 p.

thayer, g. w., w. e. schaaf, j. w. angelovic, and m. w.
LaCroix.

1973. Caloric measurements of some estuarine organisms.
Fish. Bull., U.S. 71:289-296.
TYLER, A. V.

1973. Caloric values of some North Atlantic invertebrates.
Mar. Biol. (Beri.) 19:258-261.

Geoffrey C. Laurence

Northeast Fisheries Center Narrangansett Laboratory
National Marine Fisheries Service, NOAA
Narragansett, RI 02882

METHOD FOR RESTRAINING
LIVING PLANKTONIC CRUSTACEANS*

Studies of the feeding and swimming mecha-
nisms of small, active planktonic crustaceans re-
quire restraining the organisms so that water
flow and limb movements can be observed under
the microscope. The usual technique is to place
the organism in a watch glass or cavity slide
(Cannon 1928; Gauld 1966) or to secure the dorsal
side of the animal to a drop of stopcock grease in

'Contribution No. 3488 from the Woods Hole Oceanographic
Institution. This work was supported by NSF Grant GA-41188.

220

some type of water chamber (McMahon and
Rigler 1963). For many studies, these methods
are undesirable because of the confinement of the
animal to a small volume of medium or because of
the solid boundaries nearby, both of which affect
the flow of water and possibly the movement of
limbs or other behavior by the animal (Lowndes
1935). Whenever the animal must be placed
within a relatively large volume of water, other
methods must be used. In a study of mate-seeking
behavior, Katona (1973) tethered female copepods
by means of fine stainless steel wires looped
about their bodies. While this method allows the
subsequent release of the animals unharmed, the
restraining wire can interfere with limb
movements.

I have found a relatively simple method for re-
straining small crustaceans in large volumes of
water for extended periods of microscopic exami-
nation. A short segment (1-2 cm) of nylon mono-
filament fishing line of small diameter relative to
the organism is mounted in a dissecting needle
holder or pin vise. The free tip of the mono-
filament is then cut off square with a razor blade.
The animal is placed dorsal side up in a small
drop of water on a microscope slide or watch
glass. The tip of the monofilament is dipped in a
fresh droplet of "instant" drying polymer glue
(such as Dixon Duradix)^ and quickly applied and
held to the center line of the dorsal surface of the
animal for about 5 s. The organism can then be
lifted from the slide and placed in the test vessel,
with the dissecting needle holder mounted in a
micromanipulator or other type of clamping de-
vice. The rapid filming over of the glue and its
tendency to spread when placed on the wet ani-
mal sometimes makes a neat attachment difficult
and several attempts may be needed before a
satisfactory mount is achieved.

Organisms restrained in this way appear to
carry out swimming movements in a natural
manner and live for several days on the mount.
Removal of the animal from the monofilament
usually results in its death. To make limb move-
ments easier to observe, organisms can be vitally
stained with neutral red prior to mounting (Dres-
sel et al. 1972).

I have since found a description of this mounting
technique given by Scourfield (1900) in which he
regrets that no satisfactory cement could be found.
The polymer glues appear to solve the problem.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Literature Cited

Cannon, H. G.

1928. On the feeding mechanism of the copepods, Calanus
finrrmrchicus and Diaptomus gracilis. Br. J. Exp. Biol.
6:131-144.
DRESSEL, D. M., D. R. HEINLE, AND M. C. GROTE.

1972. Vital staining to sort dead and live copepods. Chesa-
peake Sci. 13:156-159.

Gauld, D. T.

1966. The swimming and feeding of planktonic cope-
pods. In H. Barnes (editor), Some contemporary studies
in marine science, p. 313-334. George Allen and Unwin,
Ltd., Lond.

KATONA, S. K.

1973. Evidence for sex pheromones in planktonic cope-
pods. Limnol. Oceanogr. 18:574-583.

LOWNDES, A. G.

1935. The swimming and feeding of certain calanoid cope-
pods. Proc. Zool. Soc. Lond. 1935:687-715.
MCMAHON, J. W., AND F. H. RiGLER.

1963. Mechanisms regiilating the feeding rate of Daphnia
magna Straus. Can. J. Zool. 41:321-332.
SCOURFIELD. D. J.

1900. The swimming peculiarities of Daphnia and its al-
lies, with an account of a new method of examining
living Entomostraca and similar organisms. J. Quekett
Microsc. Club 7:395-404.

LOREN R. HAURY

Woods Hole Oceanographic Institution
Woods Hole, MA 02543

OBSERVATIONS ON
THE BIGEYE THRESHER SHARK,

ALOPIAS SUPERCILIOSUS, IN
THE WESTERN NORTH ATLANTIC

Thresher sharks of the genus Alopias are distrib-
uted throughout the tropical and warm temper-
ate zones of the world's oceans. Of the two species
reported from the western North Atlantic, the
thresher shark, A. vulpinus, is commonly found
in coastal waters of the middle Atlantic states
(Bigelow and Schroeder 1948). The second
member of the genus, the bigeye thresher, A.
superciliosus , is a little known offshore resident
of the continental slope and open sea.

Lowe first described the bigeye thresher in
1840 from a specimen taken off the island of
Madeira (Bigelow and Schroeder 1948). The
species was not reported again until 1941 when
Springer (1943) documented the occurrence of a
gravid female taken near Salerno, Fla. Records of
other bigeye threshers from the Atlantic include
a gravid female, two embryos, a juvenile male,
and an 18-foot specimen all taken from the north

221

coast of Cuba in the late 1940's (Bigelow and
Schroeder 1948); an adult female from Nassau in
1962 and an adult male from Cape Hatteras,
N.C., in 1963 (Fitch and Craig 1964). Bigelow and
Schroeder (1948) reported proportional mea-
surements from two individuals taken off Cuba;
Strasburg (1958) and Fitch and Craig (1964) re-
ported similar data from two Pacific specimens.
We report observations of A. superciliosus
taken on pelagic longlines aboard the commercial
fishing vessel Cap'n Bill III, in 1962, the RY Dol-
phin of the Sandy Hook Laboratory in 1966-69,
and the RV Gosnold of the Woods Hole Oceano-
graphic Institution in 1971. All previous evidence

suggests A. superciliosus is not abundant any-
where in its range. However, our data, together
with anecdotal information from experienced
commercial longliners, show that concentrations
of bigeye threshers occur during April-June off
Cape Hatteras. Other sharks and teleosts occur-
ring in the area with A. superciliosus included
blue shark, Prionace glauca; short fin mako shark,
Isurus oxyrinchus; scalloped hammerhead,
Sphyrna lewini; bignose shark, Carcharhinus al-
timus; night shark, Hypoprion signatus; dusky
shark, C. obscurus; and silky shark, C falciformis ,
along with swordfish, Xiphias gladius; and yel-
lowfin tuna, Thunnus albacares. Additional

40*

35*

30*

65«

Number taken at each station

Cruise

Sta.

Position

(Start Haul)

Depth

No.

Vessel

No.

Date

No.

Lat 1 tude
liClO'

LongI tude

Heters
110

Caught

Capt. Bill III

11-12-62

R/V Oolpfiin

0-66-'.

S-Oft-Sfi

35°30'

7'<°'.7'

91'.

R/V Dolphin

D-66-6

6-08-66

35°'<2'

7'<°36'

1829

R/V Dolphin

D-68-5

6-06-68

35°12'

7'.°56'

1280

R/V Dolphin

D-68-5

6-07-68

35°I8'

7'<°57'

220

R/V Dolphin

D-69-7

'(-03-69

27°I9'

63°00'

5'.86

R/V Dolphin

D-69-1 1

5-17-69

36°I2'

Tk'sr

R/V Dolphin

0-69-11

5-18-69

35°23'

7'.°5I'

768

R/V Dolphin

D-69-n

5-19-69

35°'.5'

7'.°'(6'

886

R/V Gosnold

175

it-lB-ZI

35°33'

7'.°36'

1829

R/V Gosnold

175

2-20-71

3'.°35'

75°25'

1829

z®

_l L.

40'

35'

-30'

80*

75*

70'

65'

Figure l. — Location o{Alopias superciliosus longline catches in the western North Altantic.

222

species taken occasionally, included sandbar
shark, C. milberti; oceanic white tip, C lon-
gimanus; and porbeagle, Lamna nasus; bluefin
tuna, Thunnus thynnus; white marlin, Tetrap-
turus alhidus; sailfish, Istiophorus platypterus;
dolphin, Coryphaena hippurus; and lancetfish,
Alepisaurus sp.

All longline sets resulting in catches of bigeye
threshers were made between 0000 and 0300
with gear retrieval beginning after 0700. The
depth at which the gear was fished ranged from
near surface to a maximum of 65 m and was
controlled by float lines of varying length. Tem-
perature-depth profiles obtained from bathyther-
mograph casts were routinely used to determine
the optimum depth for the gear. The best catches
of bigeye threshers were made in areas where the
water regime ranged from 16° to 25°C at the sur-
face to a minimum of 14°C at 75 m.

A total of 65 A. superciliosus were hooked at 11
longline stations (Figure 1); of these, 7 broke free
as they were being held alongside the vessel, 23
were tagged and released, and 35 ( 15 females and
20 males) were brought aboard for examination.
Length measurements and internal examination
of stomachs and reproductive organs were made
on all sharks brought aboard. Total lengths (TL)

for the 15 females ranged from 233 to 399 cm {x —
312 cm); the 20 males ranged from 155 to 352 cm
(x = 307 cm).

Morphometric measurements from eight males
and four females, summarized in Table 1 as per-
cents of fork length, were collected following the
methods of Bigelow and Schroeder (1948). Fork
length (FL) measurements were used as a pri-
mary growth parameter in the morphometric re-
lationships in order to discern more accurately
any changes occurring in body proportions with
increasing size. The same accuracy could not be
expected if total lengths were used because of the
difficulty in obtaining precise length measure-
ments due to the extreme size and shape of the
caudal fin.

Proportional data from Table 1 shows that al-
lometric growth is reflected in several characters.
The most obvious change associated with increas-
ing fork length is a proportionately shorter head
length resulting in a decrease in the ratios of
snout to: eye, nostrils, mouth, first gill, and pec-
toral fin. The relative size of the eye and mouth
also decrease as the body lengthens. Characters
that increase allometrically with growth include
height of first dorsal, length of claspers in males,
and interspaces between fins except in females

Table l. — Proportional dimensions of body parts in percent of fork length for 12 Alopias superciliosus.

Male

Female

Body part

Total length (cm)

155.0

307.0

315.0

331.0

332.0

342.4

351.7

339.0

257.5

340.0

355.0

399.0

Fork length (cm)

100.0

188.0

192.5

197,0

1970

207.0

212.5

217.0

167.0

207.0

210.0

221.0

% of total length

64.5

61.2

61.1

59,5

59.3

60.4

64.0

64.9

60.8

59.1

55.3

Distance from snout to:

eyes

9.0

6.4

7.4

6.3

7.5

6.7

7.4

6.7

6.8

7.2

nostrils

6.5

5.3

6.0

5.6

5.2

6.0

5.5

5.7

5.5

5.0

5.2

mouth

9.5

7.8

8.2

7.9

7.7

7.9

7.8

8.3

8.0

7.5

7.4

7.9

first gill (base)

25.5

21.2

23.6

23.4

20.6

22.7

22.3

226

22.4

22.9

22.1

21.7

pectoral

29.0

24.1

282

27.2

24.4

26.1

25.7

24.0

25.6

26.3

24.3

25.8

first dorsal

57.0

55.4

52.5

51 8

55.1

55.5

53.2

53.0

51.2

52.2

51.0

52.3

second dorsal

82.0

82.2

79.2

80.5

81.7

82.1

80.5

802

79.9

82.6

79.5

80.5

pelvic

66.0

65.2

66.2

66.0

64.5

66.9

65.2

64,1

65.3

64.4

64.3

66.0

anal

87.0

88.0

87.6

87.3

863

87.9

87.8

85.7

83.5

85.0

83.3

86.0

upper caudal pit

90.5

90.8

89.7

89.3

90.4

90.6

89.7

88.8

91.8

90.0

90.5

Interspace between:

1st & 2nd dorsal

16 8

17.8

16.4

17.0

198

16.6

17.2

18.0

18.9

18.4

17.9

18.1

2nd dorsal & caudal

7.2

7.8

8.8

8.1

8.3

9.2

7.7

7.9

8.2

pelvic & anal

9.5

12.2

13.0

12.9

11,3

11.8

13.2

12.9

7.2

8.2

7.4

8.1

anal & caudal

3.2

3.6

3.4

4.3

4.6

3.6

3.3

3.2

6.0

4.8

4.3

5.0

nostrils (proximal)

4.5

2.5

2.7

2.5

2.9

2.6

2.8

2.4

2.7

2.4

2.5

Height of:

first dorsal

10.0

11.5

13.0

11,5

11.9

11.7

11.6

12.4

11.8

13.5

12.8

14.0

free tip

1.5

1.6

1.7

2.0

1.9

1.8

1.9

1.3

second dorsal

1.1

1.3

1.0

1.8

free tip

2.1

2.5

2,5

2.4

2.6

2.8

2.7

3.4

2.5

4.3

Diameter of eye'

horizontal

3.5

2.5

2.6

2.4

2.9

2.5

3.5

2.8

3.2

vertical

4.0

4.2

4.5

4.4

3.8

Right clasper

3.0

12.4

13.0

12.9

11.9

12.4

10.8

12.0

Left clasper

3.1

12.4

11.4

12.9

11.7

12.1

11.6

12.0

Width of mouth

9.0

6.2

7.3

7.0

7.7

8.3

7.6

7.5

8.1

Height of mouth

5.0

4.5

4.7

4.8

4.4

4.3

5.0

3.6

4.7

4.3

4.5

Max length pectoral fin

32.3

31.2

33.6

32.0

32.1

31.6

31.8

32.3

35.5

32.4

33.5

'Orbit.

223

where the distance between anal and caudal fin
decreases.

The length-weight relationship for this species
(Figure 2) was derived using data from 5 females
and 11 males. To determine the regression line,
the equation, log Y = 11.1204 + 2.99269 log X
was calculated using the nonlinear least squares
method of Pienaar and Thomson (1969).

Clark and von Schmidt (1965) noted that adult
and juvenile males of several species of sharks
can be distinguished by the differences in the rel-
ative size and rigidity of the claspers. This
characteristic applies to A. superciliosus. Of the
males examined, the claspers of all but five indi-
viduals were large (10.8-13.0% of their FL), heav-
ily calcified, and quite obviously mature. Internal
examinations of the larger males revealed the
presence of sperm in the epididymis and sper-

200-

150-

100

SO-

<v/
♦ /

o MALES

20-

1 1

X FEMALES

15 ■

o/
o/

lO J

1 1 1 r

150 200 250 300

FORK LENGTH IN CENTIMETERS

Figure 2. — Length- weight relationship for Alopias
superciliosus.

matophores in the lower ductus deferens. The
smallest male positively identified as mature was
307 cm TL. A smaller individual however of 289
cm TL had testes in a relatively advanced state of
development. Female A. superciliosus apparently
mature at a larger size than males. Of the 13
females examined (233-355 cm TL) only the
largest was mature. Ovaries of immature indi-
viduals were 10-13 cm long and 3-5 cm wide and
contained thousands of white opaque follicles
from less than 1 to 5 mm diameter. The oviducts
were firm, ribbonlike tubes 0.5 to 2.5 cm in di-
ameter The 355-cm female differed in that the
ovary was 30 cm long and 10 cm wide and con-
tained yellow ova up to 10 mm in diameter. Also
the oviducts in this individual were considerably
larger (10 cm in diameter) and more flaccid and
similar in appearance to the post gravid condition
of other species we have seen. We suggest A.
superciliosus males may mature at 290-300 cm
TL, but females are not mature until they reach
350 cm.

Examination of the stomachs showed 17
(48.5%) were empty. Of the 18 that contained food
the most common items were squid (66%) and
scombrid remains (27%). One stomach contained
remains of 5 lancetfish; another, 30 small (5-10
cm) herringlike fishes; and a third had parts of a
small billfish, tentatively identified as an is-
tiophorid. The occurrence of two or more whole
longline baits in stomachs was not uncommon
and suggests they had been dislodged from hooks
elsewhere on the line. Alopias superciliosus may
utilize its tail to herd or stun its prey in the man-
ner described for A. uulpinus (Bigelow and
Schroeder 1948; Strasburg 1958). Several indi-
viduals including some of those lost at the rail
were foul hooked in the tail.

Acknowledgments

We are indebted to Frank Carey and John
Mason of the Woods Hole Oceanographic Institu-
tion who assisted during cruises aboard the RV
Gosnold and provided measurements on the
399-cm female; to Martin Bartlett for his assis-
tance aboard the Cap'n Bill III; to commercial
longliners Phil Rhule and Deba Larson and
James Beckett of the Canadian Fisheries Re-
search Board of Canada for their anecdotal in-
formation; and to Michael L. Dahlberg for help in
adapting the length-weight program for our
purposes.

224

Literature Cited

BIGELOW, H. B., AND W. C. SCHROEDER.

1948. Sharks. In A. E. Parr and Y. H. Olsen (editors),
Fishes of the western North Atlantic. Part One, p. 59-
546. Sears Found. Mar. Res., Yale Univ., Mem. 1.
CLARK, E., AND K. VON SCHMIDT.

1965. Sharks of the central Gulf Coast of Florida. Bull.
Mar. Sci. 15:13-83.

FITCH, J. E., AND W. L. Craig.

1964. First records for the bigeye thresher (Alopias super-
ciliosus) and slender tuna ( Allothunnus fallal) from
California, with notes on eastern Pacific scombrid oto-
liths. Calif Fish. Game 50:195-206.
PlENAAR, L. v., AND J. A. THOMSON.

1969. Allometric weight-length regression model. J. Fish.
Res. Board Can. 26:123-131.

Springer, S.

1943. A second species of thresher shark from Flor-
ida. Copeia 1943:54-55.
STRASBURG, D. W.

1958. Distribution, abundance, and habits of pelagic
sharks in the central Pacific Ocean. U.S. Fish Wildl.
Serv., Fish. Bull. 58:335-361.

Charles E. Still well
John G. Casey

Northeast Fisheries Center Narragansett Laboratory
National Marine Fisheries Service, NOAA
RR 7-A, Box 522-A
Narragansett, RI 02882

EPIZOITES ASSOCIATED WITH

BATHYNECTES SUPERBUS (DECAPODA:

PORT.UNIDAE)''2

The only known documentation of epizoites occur-
ring on Bathynectes superbus (Costa 1853) is that
of Capart (1951), who noted a stalked barnacle,
Scalpellum sp., on specimens from the South At-
lantic coast of Africa. This note describes epizoites
present on B. superbus from the western North
Atlantic Ocean.

Crabs were obtained from several cruises along
the eastern coast of North America (lat. 36°33'N-
39°38'N to long. 73°00'W-74°43'W): RV Columbus
Iselin (cruise 73-10) from 252 to 335 m; RV Dan
Moore (73-030) from 122 to 232 m; RV Albatross IV
(74-4) from 236 to 300 m; and RV Eastward (E-2-
74) from 280 to 350 m. Gills, branchial chambers,
and external surfaces of 172 crabs were examined.
Crabs often supported more than one epizoite.

'Contribution No. 740, Virginia Institute of Marine Science,
Gloucester Point, VA 23062.

^Research supported partly by National Oceanic and Atmo-
spheric Administration, Office of Sea Grant (No. 04-3-158-49).

Crabs were most heavily fouled (65%) with a
"'Perigonimus" -like hydroid. Quotations are pres-
ent around the name "Perigonimus" because the
genus is not valid and is a representative of a
poorly known group, the systematics of which
need revision (D. R. Calder, pers. commun.). The
"Perigonimus" -like hydroid was most frequently
found associated with setae along the ventral an-
terolateral border and on the ecdysial suture line.
Trilasmis (Poecilasma) kaempferi inaequilaterale
Pilsbry (Cirripedia: Scalpellidae) was found on
13% of thefi. superbus examined. It was present
on all exposed regions of the carapace, pereopods,
and abdomen. An eastern Atlantic specimen in the
U.S. National Museum collections {Geronimo-2-
203) had approximately 100 T. k. inaequilaterale
on the dorsal carapace, pereopods, eyes, and
mouthparts. Anomia aculeata (Pelecypoda) was
relatively abundant (14%) and frequently occurred
in indentations of the dorsal carapace and on
the carinae of pereopods. Other organisms on the
carapace were calcareous tubes of an unidentified
polychaete (<1%) and Stegopoma plicatile, a the-
cate hydroid (<1%). The latter were found along
the ventral anterolateral surface of the carapace.
No organisms were found within the branchial
chamber.

Figure 1 shows the occurrence of epizoites on 5.
superbus according to sex, size group, and molt
stage. Size groups of short carapace width (=s35
mm, 36-45 mm, 46-57 mm, 2=58 mm) are based on
arbitrarily chosen modes from a size-frequency
distribution (Lewis 1975).

Crabs were assigned to molt stages described by
Drach and TchernigovtzefF(1967): anecdysis (Ci-
C4), proecdysis (DrD4), postecdysis (A1-B2).

There is apparently no preference of epizoites
for male or female crabs, but there is an associa-
tion with molt stage and size. As expected, crabs in
anecdysis are more heavily fouled than those
which have recently molted (A1-B2). Larger crabs
(>46 mm) supported a variety of epizoites while
those ^35 mm were colonized by Perigonimus
only. This may be attributable to the greater sur-
face available for epizoite set on larger crabs and
the lower frequency of molt for these crabs.

The epizoites are inhabitants of the shelf-edge
upper slope habitat within the bathymetric range
of Bathynectes. Trilasmis (Poecilasma) has a
known range along the western Atlantic from
Martha's Vineyard, Mass. to Key West, Fla., hav-
ing been recorded at depths from 21.6 to 1,733 m,
chiefly on the carapace of the brachyurans Geryon

225

_J
<

Peridonimus

Anomia

Trilasmis
I I SteQopoma

Worm tubes
^ No fouling

Figure l. — Occurrence ofepizoiteson male and female Ba^/i^'/ieciessMperfcus atanecdysis
(C1-C4), proecdysis (D1-D4), and postecdysis (A2-B2) for four modal size (short carapace
width) groups («35 mm, 36-45 mm, 46-57 mm, s=58 mm).

quinquedens Smith (Pilsbry 1907) and Cancer
borealis Stimpson, collected from the same cruises
from yjhich Bathynectes were obtained. Trilasmis
(Poecilasma) was also observed on mature
lobsters, i/omarus americanus H. Milne-Edwards.
These decapods are bathymetric associates of 5.
superbus (Lewis 1975). Trilasmis (Poecilasma)
has also been found on Hyposophrys noar, a
brachyuran from the Straits of Florida (Williams
1974).

Anomia aculeata has been recorded from the
Arctic Ocean to Cape Hatteras, N.C. within a
bathymetric range of 1.8 to 144 m (Smith 1937).
The stations at which this pelecypod occurred on
Bathynectes were in depths greater than 200 m.

The hydroid, Stegopoma plicatile, is common
along the east coast of the United States from
Hudson Bay to Cape Hatteras with a bathymetric
range of 45 to 1,733 m (Fraser 1944).

Acknowledgments

I thank Frank Holland (North Carolina Divi-

sion of Commercial and Sports Fisheries) for col-
lection of specimens (RV Dan Moore) and Charles
Wenner (Virginia Institute of Marine Science; RV
Albatross IV). Ship time on RV Columbus Iselin
and RV Eastward was provided by J. A. Musick
through NSF grants GS-37561 and GS-27725, re-
spectively. Dale Calder (South Carolina Marine
Resources Research Institute) identified the hy-
droids found onB. superbus, and Mariana Doyle
(U.S. National Museum) confirmed identification
of Trilasmis (Poecilasma) kaempferi in-
aequilaterale . I also thank George Grant, P. A.
Haefner, Jr., Fred Jacobs, and W. A. Van Engel for
their criticism of the manuscript.

Literature Cited

Capart, a.

1951. Crustaces decapodes, Brachyures. Exped. oceanogr.

Beige dans les Eaux cotieres afr. Atl. Sud. (1948-1949)

3(l):ll-205.
Drach, p., and C. TCHERNIGOVTZEFF.

1967. Sur la methode de determination des stades d'inter-

mue et son application generale aux crustaces. Vie

Milieu 18:595-609.

226

FRASER, C. M.

1944. Hydroids of the Atlantic coast of North America.
Univ. Toronto Press, Toronto, 451 p.

Lewis, E. G.

1975. Contributions to the biology of Bathynectes superbus
(Costa) (Decapoda: Portunidae) from the Chesapeake
Bight of the western North Atlantic. M.A. Thesis, Col-
lege of William and Mary, Williamsburg.
PILSBRY, H. A.

1907. The barnacles (Cirripedia) contained in the collec-
tions of the U.S. National Museum. Smithson. Bull.
60:1-122.

Smith, M.

1937. East coast marine shells. Edwards Brothers, Inc.,

Ann Arbor, 308 p.
WILLIAMS, A. B.

1974. A new species oiHypsophrys (Decapoda: Homolidae)

from the Straits of Florida, with notes on related crabs.

Proc. Biol. Soc. Wash. 87:485-492.

ELIZABETH G. Lewis

Virginia Institute of Marine Science
Gloucester Point, VA 23062

227

INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN

Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the
following instructions. These are not absolute requirements, of course, but desiderata.

CONTENT OF MANUSCRIPT

The title page should give only the title of the
paper, the author's name, his affiliation, and
mailing address, including Zip code.

The abstract should not exceed one double-
spaced page.

In the text. Fishery Bulletin style, for the most
part, follows that of the Style Manual for Biologi-
cal Journals. Fish names follow the style of the
American Fisheries Society Special Publication
No. 6, A List of Common and Scientific Names of
Fishes from the United States and Canada, Third
Edition, 1970. The Merriam-Webster Third New
International Dictionary is used as the authority
for correct spelling and word division.

Text footnotes should be typed separately
fi'om the text.

Figures and tables, with their legends and
headings, should be self-explanatory, not requir-
ing reference to the text. Their placement should
be indicated in the right-hand margin of the
manuscript.

Preferably figures should be reduced by photo-
graphy to 5% inches (for single-column figures,
allowing for 50% reduction in printing), or to 12
inches (for double-column figures). The maximum
height, for either width, is 14 inches. Photo-
graphs should be printed on glossy paper.

Do not send original drawings to the Scientific
Editor; if they, rather than the photographic re-
ductions, are needed by the printer, the Scientific
Publications Staff will request them.

Each table should start on a separate page.
Consistency in headings and format is desirable.
Vertical rules should be avoided, as they make
the tables more expensive to print. Footnotes in
tables should be numbered sequentially in arable
numerals. To avoid confusion with powers, they
should be placed to the left of numerals.

Acknowledgments, if included, are placed at
the end of the text.

Literature is cited in the text as: Lynn and Reid
(1968) or (Lynn and Reid 1968). All papers re-
ferred to in the text should be listed alphabetically
by the senior author's surname under the heading

"Literature Cited." Only the author's surname
and initials are required in the literature cited.
The accuracy of the literature cited is the re-
sponsibility of the author. Abbreviations of
names of periodicals and serials should conform
to Biological Abstracts List of Serials with Title
Abbreviations. (Chemical Abstracts also uses this
system, which was developed by the American
Standards Association.)

Common abbreviations and symbols, such as
mm, m, g, ml, mg, °C (for Celsius), %, %» and so
forth, should be used. Abbreviate units of mea-
sure only when used with numerals. Periods are
only rarely used with abbreviations.

We prefer that measurements be given in
metric units; other equivalent units may be given
in parentheses.

FORM OF THE MANUSCRIPT

The original of the manuscript should be typed,
double-spaced, on white bond paper. Please triple
space above headings. We would rather receive
good duplicated copies of manuscripts than car-
bon copies. The sequence of the material should
be:

TITLE PAGE

ABSTRACT

TEXT

LITERATURE CITED

APPENDIX

TEXT FOOTNOTES

TABLES (Each table should be numbered with
an arable numeral and heading provided)

LIST OF FIGURES (entire figure legends)

FIGURES (Each figure should be numbered
with an arable numeral; legends are desired)

ADDITIONAL INFORMATION

Send the ribbon copy and two duplicated or
carbon copies of the manuscript to:

Dr Bruce B. Collette, Scientific Editor

Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Fifty separates will be supplied to an author
free of charge and 100 supplied to his organiza-
tion. No covers will be supplied.

Contents — continued

BRUSHER, HAROLD A., and LARRY H. OGREN. Distribution, abundance, and size of
penaeid shrimps in the St. Andrew Bay system, Florida 158

MASON, J. C. Some features of coho salmon, Oncorhynchus kisuich, fry emerging from simu-
lated redds and concurrent changes in photobehavior 167

HURLEY, ANN C. Feeding behavior, food consumption, growth, and respiration of the squid
Loligo opalescens raised in the laboratory 176

GARRISON, DAVID L. Contribution of the net plankton and nannoplankton to the standing
stocks and primary productivity in Monterey Bay, California during the upwelling season . 183

TRENT, LEE, EDWARD J. PULLEN, and RAPHAEL PROCTOR. Abundance of macrocrusta-

ceans in a natural marsh and a marsh altered by dredging, bulkheading, and filling 195

Notes

FISHER, WILLIAM S., and DANIEL W. WICKHAM. Mortalities and epibiotic fouling of eggs

from wild populations of the Dungeness crab. Cancer magister 201 ■^

MATSUMOTO, WALTER M. Second record of black skipjack, Euthynnus lineatus, from the

Hawaiian Islands 207

WEIS, JUDITH S., and PEDDRICK WEIS. Optical malformations induced by insecticides in

embryos of the Atlantic silverside, Menidia menidia 208

CHENG, LANNA, and RALPH A. LEWIN. Goose barnacles (Cirripedia: Thoracica) on flotsam

beached at La Jolla, California 212 ■*'

LAURENCE, GEOFFREY C. Caloric values of some North Atlantic calanoid copepods .... 218 —

HAURY, LOREN R. Method for restraining living planktonic crustaceans 220 "

STILLWELL, CHARLES E., and JOHN G. CASEY. Observation on the bigeye thresher shark,

Alopias superciliosus, in the western North Atlantic 221

LEWIS, ELIZABETH G. Epizoites associated with Bathynectes superbus (Decapoda:

Portunidae) 225

5 ^'^^ ^

AMERICAS
,> ^^ FIRST INDUSTRY

W rxv»r\ ftQC-O'a'a

Fishery Bulletin

^ National Oceanic and Atmospheric Administration • National Marine Fisheries Service

Mm Biological laiioralorii j

LIBRARY
AUG^ »7S

Vol. 74, No. 2 I Woods Hole, M^ss^ j ^p^H .,975

PERRIN, WILLIAM R, JAMES M. COE, and JAMES R. ZWEIFEL. Growth and
reproduction of the spotted porpoise, Stenella attenuata, in the offshore eastern
tropical Pacific 229

SAKAGAWA, GARY T., and MAKOTO KIMURA. Growth of laboratory-reared
northern anchovy, Engraulis mordax, from southern California 271

HEWITT, ROGER R, PAUL E. SMITH, and JOHN C. BROWN. Development and use
of sonar mapping for pelagic stock assessment in the California Current area . . . 281

GRIFFIN, WADE L., NEWTON J. WARDLAW, and JOHN R NICHOLS. Economic

and financial analysis of increasing costs in the Gulf shrimp fleet 301 V

LIVINGSTON, ROBERT J., GERARD J. KOBYLINSKI, FRANK G. LEWIS, III, and

PETER F. SHERIDAN. Long-term fluctuations of epibenthic fish and invertebrate l

populations in Apalachicola Bay, Florida 311^T

HAYNES, EVAN. Description of zoeae of coonstripe shrimp, Pandalus hypsinotus,
reared in the laboratory 323 V

CHITTENDEN, MARK E., JR. Present and historical spawning grounds and nurser-
ies of American shad, Alosa sapidissima, in the Delaware River 343

LOUGH, R. GREGORY. Larval dynamics of the Dungeness crab, Cancer magister, off
the central Oregon coast, 1970-71 353

HANSON, CHARLES H., and JONATHAN BELL. Subtidal and intertidal marine
fouling on artificial substrata in northern Puget Sound, Washington 377

HASTINGS, ROBERT W, LARRY H. OGREN, and MICHAEL T MABRY. Observa-
tions on the fish fauna associated with offshore platforms in the northeastern Gulf of
Mexico 387

CRADDOCK, DONOVAN R. Effects of increased water temperature on Daphnia
pulex 403 ^

ABLE, K. W., and J. A. MUSICK. Life history, ecology, and behavior oi Liparis
inquilinus (Pisces: Cyclopteridae) associated with the sea scallop, PZacopecien ma-
gelanicus 409

KJELSON, MARTIN A., and GEORGE N. JOHNSON. Further observations of the 1

feeding ecology of postlarval pinfish, Lagodon rhomboides, and spot, Leiostomus
xanthurus 423

BREWER, GARY D. Thermal tolerance and resistance of the northern anchovy, I

Engraulis mordax 433

(Continued on back cover)

Seattle, Washington

U.S. DEPARTMENT OF COMMERCE

Elliot L. Richardson, Secretary

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION

Robert M. White, Administrator

NATIONAL MARINE FISHERIES SERVICE

Robert W. Schoning, Director

Fishery Bulletin

The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and
economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in
1904 and the Fishery Bulletin of the Fish and WildUfe Service in 1941 . Separates were issued as documents through volume 46; the last
document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a
numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin
instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical,
issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office,
Washington, D.C. 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and
in exchange for other scientific publications.

EDITOR

Dr. Bruce B. Collette

Scientific Editor, Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Editorial Committee

Dr. Elbert H. Ahlstrom

National Marine Fisheries Service

Dr. WilHam H. Bayliff

Inter-American Tropical Tuna Commission

Dr. Roger F. Cressey, Jr.
U.S. National Museum

Mr. John E. Fitch

California Department of Fish and Game

Dr. William W. Fox, Jr.

National Marine Fisheries Service

Dr. Marvin D. Grosslein
National Marine Fisheries Service

Dr. Edward D. Houde
University of Miami

Dr. Merton C. Ingham

National Marine Fisheries Service

Dr. Reuben Lasker

National Marine Fisheries Service

Dr. Sally L. Richardson
Oregon State University

Dr. Paul J. Struhsaker

National Marine Fisheries Service

Dr. Austin Williams

National Marine Fisheries Service

Kiyoshi G. Fukano, Managing Editor

The Fishery Bulletin is published quarterly by Scientific Publications Staft National Marine Fisheries Service, NOAA,
Room 450, 1107 NE 45th Street, Seattle, WA 98105. Controlled circulation postage paid at Seattle, Wash.

Fishery Bulletin

CONTENTS

Vol. 74, No. 2 April 1976

PERRIN, WILLIAM R, JAMES M. COE, and JAMES R. ZWEIFEL. Growth and
reproduction of the spotted porpoise, Stenella attenuata, in the offshore eastern
tropical Pacific 229

SAKAGAWA, GARY T., and MAKOTO KIMURA. Growth of laboratory-reared
northern anchovy, Engraulis mordax, from southern California 271

HEWITT, ROGER R, RAUL E. SMITH, and JOHN C. BROWN. Development and use
of sonar mapping for pelagic stock assessment in the California Current area . . . 281

GRIFFIN, WADE L., NEWTON J. WARDLAW, and JOHN R NICHOLS. Economic
and financial analysis of increasing costs in the Gulf shrimp fleet 301

LIVINGSTON, ROBERT J., GERARD J. KOBYLINSKI, FRANK G. LEWIS, III, and
PETER F. SHERIDAN. Long-term fluctuations of epibenthic fish and invertebrate
populations in Apalachicola Bay, Florida 311

HAYNES, EVAN. Description of zoeae of coonstripe shrimp, Panc?aZr/s hypsinotus,
reared in the laboratory 323

CHITTENDEN, MARK E., JR. Present and historical spawning grounds and nurser-
ies of American shad, Alosa sapidissima , in the Delaware River 343

LOUGH, R. GREGORY. Larval dynamics of the Dungeness crab, Cancer magister, off"
the central Oregon coast, 1970-71 353

HANSON, CHARLES H., and JONATHAN BELL. Subtidal and intertidal marine
fouling on artificial substrata in northern Puget Sound, Washington 377

HASTINGS, ROBERT W., LARRY H. OGREN, and MICHAEL T. MABRY Observa-
tions on the fish fauna associated with offshore platforms in the northeastern Gulf of
Mexico 387

CRADDOCK, DONOVAN R. Effects of increased water temperature on Daphnia
pulex 403

ABLE, K. W., and J. A. MUSICK. Life history, ecology, and behavior of Liparis
inquilinus (Pisces: Cyclopteridae) associated with the sea scallop, P/acopecten ma-
gelanicus 409

KJELSON, MARTIN A., and GEORGE N. JOHNSON. Further observations of the
feeding ecology of postlarval pinfish, Lagodon rhomboides, and spot, Leiostomus
xanthurus 423

BREWER, GARY D. Thermal tolerance and resistance of the northern anchovy,
Engraulis mordax 433

(Continued on next page)

Seattle, Washington

For sale by the Superintendent of Documents, U.S. Government Prmtmg Office, Washing-
ton, D.C. 20402 — Subscnption price: $11.80 per year ($2.95 additional for foreign mail-
ing). Cost per single issue - $2.95.

Contents — continued

Notes

HARRELL, LEE W., ANTHONY J. NOVOTNY, MICHAEL H. SCHIEWE, and
HAROLD O. HODGINS. Isolation and description of two vibrios pathogenic to
Pacific salmon in Puget Sound, Washington 447

MAY, NELSON, LEE TRENT, and PAUL J. PRISTAS. Relation offish catches in gill
nets to frontal periods 449

LANSFORD, LAWRENCE M., CHARLES W. CAILLOUET, and KENNETH T
MARVIN. Phosphoglucomutase polymorphism in two penaeid shrimps, Penaeus
brasiliensis and Penaeus aztecus subtilis 453

PERRIN, WILLIAM F. First record of the melon-headed whale, Peponocephala electra,
in the eastern Pacific, with a summary of world distribution 457

CARLSON, H. RICHARD. Foods of juvenile sockeye salmon, Oncorhynchus nerka, in
the inshore coastal waters of Bristol Bay, Alaska, 1966-67 458

LAIRD, CHAE E. , ELIZABETH G. LEWIS, and PAUL A. HAEFNER, JR. Occurrence
of two galatheid crustaceans, Munida forceps and Munidopsis bermudezi, in the
Chesapeake Bight of the western North Atlantic Ocean 462

WEIS, JUDITH S. Effects of mercury, cadmium, and lead salts on regeneration and
ecdysis in the fiddler crab, Uca pugilator 464

FUIMAN, LEE A. Notes on the early development of the sea raven, Hemitripterus
americanus 467

Vol. 74, No. 1 was published on 8 April 1976.

GROWTH AND REPRODUCTION OF THE SPOTTED PORPOISE,
STENELLA ATTENUATA, IN THE OFFSHORE EASTERN

TROPICAL PACIFIC

William F. Perrin, James M. Coe, and James R. Zweifel^

ABSTRACT

This study is based on data from several thousand specimens of spotted porpoise, Stenella attenuata,
incidentally killed in the purse seine fishery for yellowfin tuna, Thunnus albacares. Average length
at birth is 82.5 cm. Gestation is 11.5 mo. Average length at 1 yr is 138 cm. Length-weight equations
are given for fetuses and postnatal males and females. Age was estimated from dentinal layers in
thin sections of teeth. A two-phase Laird-Gompertz growth model was fitted to the layer-length data.
Direct calibration of the dentinal layers beyond the first year (two layers) was not possible, and three
alternative hypotheses were considered: 1) two layers per year, until pulp cavity occluded, 2) two
layers per year in first year, and one per year thereafter, and 3 1 two layers per year until puberty, and
one per year thereafter The second alternative is most probably the correct one, but reproductive
parameters were estimated in terms of layers. Breeding is diffusely seasonal, with prolonged calving
seasons in spring and fall and a pronounced low in winter A third calving season may exist in the
summer. Average age at attainment of sexual maturity of males is approximately 12 layers (average
length about 195 cm and averjige weight about 75 kg). Females attain sexual maturity on the
average at about 9 layers and 181 cm. Ovarian changes in adult females are described. Apparently
postreproductive females were encountered in the samples. It is concluded that corpora albicantia of
ovulation and pregnancy persist indefinitely in the ovaries. It was not possible to distinguish between
the two types of corpora. Ovulation rate changes with age, from about four per layer in very young
adult females, to about one per layer in older females. The average calving interval is 26 mo long and
consists of 11.5 mo of pregnancy, 11.2 mo of lactation, and 3.3 mo of resting and/or estrus. About 9.6%
of lactating females are also pregnant. Pregnancy rate decreases with age, from about 0.6 per year at
8 to 10 layers, to about 0.3 at 16 layers. The overall sample contained 44.9% males and 55.1% females.
Sex ratio changes with age, from near parity at birth, indicating higher mortality rates for males.
Gross annual production of calves, based on age and sex structures of the sample and the estimated
pregnancy rate, is 14.4% of the papulation per year. No evidence was found of age or sex segregation
in schooling. The estimated parameters differ in a consistent way from those estimated for a
population of Stenella attenuata in the western Pacific, possibly reflecting the exploitation in the
eastern Pacific.

Porpoises of the genera Stenella and Delphinus
are killed incidentally in the tuna seine fishery in
the eastern tropical Pacific (Perrin 1969, 1970a;
National Oceanic and Atmospheric Admin-
istration^). Since 1968, the National Ma-
rine Fisheries Service (NMFS) has conducted
a program of research into the population biology
of the major porpoise species to assess the impact
of this fishing mortality on the porpoise stocks.
The purpose of this paper is to describe the life
history of the spotted porpoise, Stenella attenuata

^Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA
92039.

^National Oceanic and Atmospheric Administration. 1972.
Report of the NOAA Tuna-Porpoise Review Committee, Sep-
tember 8, 1972. Unpubl. rep. U.S. Dep. Commer, Wash.,
D.C., 63 p.

(Gray),^ the animal most frequently killed in the
fishery.

Little information on life history of the spotted
porpoise has been available until very recently.
Harrison et al. (1972) examined the gonads of 6
specimens from Japan (5 males and 1 female) and
45 specimens of S. attenuata fi-om the eastern
tropical Pacific (19 males and 26 females), but did
not separate their results and conclusions fi-om

Manuscript accepted December 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

^The taxonomy of the spotted porpoise has long been con-
fused. Recent morphological studies (Perrin in press) have
shown that the spotted porpoise in the tuna fishery is
conspecific with the spotted porpoise occurring around
Hawaii. The name S. attenuata (Gray 1846, holotype from un-
known locality) applied by True (1903) to the Hawaiian form is
used here for the eastern Pacific form, taking priority over S.
graffmani (Lonnberg 1934). This usage is strictly provisional,
pending the completion of current taxonomic studies, when a
different name, such as S. dubla (G. Cuvier 1812) or S. frontalis
(G. Cuvier 1829) may take priority.

229

FISHERY BULLETIN: VOL. 74, NO. 2

those for S. longirostris . Preliminary unpublished
results of our studies indicate that these two
species are probably disparate in such growth
parameters as length at birth, length at maturity,
and asymptotic length. Harrison et al. (1972)
stated that lengths of the fetuses examined indi-
cate that parturition occurs both in the spring
and in the autumn. They described in detail the
gross and microscopic histological appearances of
several pairs of ovaries. A maximum of nine cor-
pora albicantia were encountered. They con-
cluded that if all the corpora albicantia in ovaries
of specimens of this species do not represent past
pregnancies, either the fertility is very low or the
corpora are not permanent.

Nishiwaki et al. (1965) published length-
frequency distributions of 34 fetuses (up to 106 cm
long) and 194 postnatal animals (104 to 208 cm)
from a school driven ashore in Japan. They esti-
mated that gestation lasts 1 yr, length at birth is
about 105 cm, juveniles reach 150 cm in 6 mo, and
adult size (180 cm for females and 190 cm for
males) is reached in 1 yr. They concluded that
there are two seasons for mating and parturition,
in the spring and in the autumn, and that there
are fewer males than females among adults. On-
togenetic changes in coloration, external propor-
tions, organ weights, the skeleton, parasite load,
and feeding habits have been described (Perrin
1970b, in press; Perrin and Roberts 1972; Dailey
and Perrin 1973; Perrin et al. 1973).

Kasuya et al. (1974) recently published results
of a study of several hundred specimens caught in
the Japanese fishery for S. attenuata. Their re-
sults are discussed and compared with ours in the
body of this paper.

METHODS AND MATERIALS

Observer Program

Beginning in 1968, NMFS placed observers
aboard U.S. tuna seiners to collect information on
the incidental take of cetaceans in the eastern
tropical Pacific. Observers were placed on 1 cruise
in 1968, 5 in 1971, 12 in 1972, and 22 in 1973. Most
of the cruises were 30 to 60 days long. In addition,
biological data were collected during chartered
cruises of commercial seiners: one in 1971, one in
1972, and two in 1973.

The data collecting had to be carried out in
such a way as to not interfere with the fishing
operation. Hence, the amount of information col-

lected on the animals killed in a net set varied
vddely, depending on the amount of time that was
available before the next set was made. Following
is the hierarchy of types of data that were col-
lected (sample sizes were largest for the first and
smallest for the last):
Animals killed were:

1. Counted (estimates were made in cases
where counts were not possible), usually
on the deck or in the net,

2. Identified to species (and race when
possible),

3. (S. attenuata only) identified to develop-
mental color pattern phase (Perrin 1970b),
and sexed,

4. Measured (to nearest centimeter with 2-m
calipers), and

5. Dissected to collect information on repro-
ductive condition (for females, mammaries
were examined and reproductive tract col-
lected; for males, the right testis was col-
lected) and age (a section of the left lower
jaw at midlength was collected). The gonad-
al material and jaw sections were pre-
served in 10% Formalin.'* Small fetuses (^ 30
cm) were preserved in the uterus. Larger
fetuses were removed from the uterus and
frozen.

For each specimen that was at least measured
(step 4 above), a field serial number was assigned,
and a specimen data sheet was filled out. Data for
specimens that were not at least measured were
collected on a running tally.

The Study Area

One of us has described the distribution ofS.
attenuata in the eastern tropical Pacific (Perrin
1975). The known occurrence of mixed aggrega-
tions of cetaceans and tuna is strongly correlated
with certain oceanographic conditions peculiar to
that region. The porpoise-tuna association is
known only in the eastern tropical portion of the
Pacific. That area, which has been called the
North Pacific Equatorial water mass (Seckel
1972), has an unusual oxygen-salinity-
temperature structure. The reason for this is not

••Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

230

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

fully understood but certainly has to do with the
latitude of the area, its position relative to the
rest of the Pacific and to the American continents,
and the shapes of the adjacent land masses. These
factors interact with general global oceanic and
atmospheric circulation to produce a water mass
with relatively high surface temperature, low
surface salinity, a strongly developed, shallow
thermocline (usually within 100 m of the surface),
and a pronounced, thick oxygen minimum layer
just below the thermocline. The effect is to create
a very extensive but shallow warm habitat with a
sharp oxythermal floor. To the west, these condi-
tions tail off along a divergence centered on lat.
10°N (Wyrtki 1964). The conditions must be caus-
ally interrelated, but one of the more striking
correlations with the occurence of the mixed-
species aggregation is in the thickness of the oxy-
gen minimum layer (Figure 1).

The occurrence of the aggregation is not tightly
correlated with the geographic distributions of
the major prey species of the participating pred-
ators. Major shared prey items are the omma-
strephid squid Dosidicus gigas, an unidentified
ommastrephid (probably Symplectoteuthis sp.), a
scombrid fish Aj/x/s sp. (A. thazard or A. rochei),
and the exocoetid fish Oxyporhamphus microp-
terus (Perrin et al. 1973). Dosidicus gigas is
primarily equatorial but migrates sporadically as
far as California and southern Chile, far beyond
the limits of the distribution of the mixed-species
aggregation (Clarke 1966; Young 1972). Species of
Symplectoteuthis occur widely in the tropical
Pacific and Indian oceans (Clarke 1966). Auxis
thazard occurs in "tropical and subtropical wa-
ters of the Indo-Pacific and Atlantic oceans," and
A. rochei in "tropical and subtropical waters of
the Indo-Pacific and Atlantic oceans, including
the Mediterranean Sea" (Richards and Klawe
1972). The genus Oxyporhamphus is also pantropi-
cal (Bruun 1935). At least some of the several
myctophid fishes in the aggregate apparently are
a mainstay of the diet of the spinner porpoise in
mixed schools (Perrin et al. 1973) and are not re-
stricted to the tropics but occur also in temperate
waters of the eastern Pacific (Moser and Ahlstrom
1970) and elsewhere. These facts, combined with
the pantropical distributions of the cetaceans,
tunas, and birds, suggest that the multispecies
aggregation does not have its roots in the dis-
tribution of the component species or their prey
but rather in the peculiarities of the physical
oceanography of the region.

The Sample

In 1971 and early 1972, when more specimens
were decked than could be processed in the time
available (the limit per net set was usually about
35 to 40 specimens), adult females were selected
for measuring and dissection. The intention was
to insure that sample sizes would be large enough
to allow estimation of pregnancy rate with
adequate precision. The information on age struc-
ture of the catch for that period is limited to the
coloration phase data. The observer program sub-
sequently expanded, and beginning in October

1972 no selection was practiced in determining
which animals were to be dissected; the first 35 to
40 specimens of both sexes and all ages that came
to hand were set aside for measuring and dissec-
tion and the remainder discarded. The length
data for 1968 and for October 1972-December

1973 are presumably cross-sectional with respect
to the kill.

The sample of animals at least measured in-
cluded 3,504 postnatal animals and associated
fetuses fi'om known localities and 23 from impre-
cisely known localities (Figure 2). Coloration
phase and sex data were collected for another
6,150 specimens. In addition, some data were
available for 45 other specimens collected by
other research agencies, museums, and private
individuals. Because of the seasonal nature of the
tuna fishery, the sample is heavily biased toward
the early months of the year, with minimal cov-
erage of the latter part of the year and practically
no specimens from the summer months (Table 1).

Two races of S. attenuata exist in the eastern
tropical Pacific — a large coastal form and a
small offshore form (Perrin 1975, in press). This
paper deals only with the offshore form. The es-
timates of life history parameters cannot be as-
sumed to apply also to the coastal form.

Table L— Samples of postnatal spotted porpoise by month

for all years.

Month

Males

Females

Total

January

748

443

1,191

February

263

209

472

March

298

147

445

April

216

155

371

May

181

278

June

127

July

August

1 1

September

October

222

158

380

November

110

197

December

Total

2,144

1,383

3,527

231

FISHERY BULLETIN; VOL. 74, NO. 2

30'

20'

10'

0°-

160=

140"=

140*

120

100*

Figure l. — CompEirison of the known occurrence of spotted porpoise in the eastern Pacific (above)
with average thickness of the subsurface layer of water (contours in meters) in which the dissolved
oxygen is less than 0.25 ml/liter (below, after Knauss 1963). The entire layer lies above 1,000 m.

232

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

135° 130" 125" 120"

0° 105° 100° 95" 90° 85° 60"

123

Galopagos is c^

4 1 8 -■

135° (30" 125" 120" 115° 110° 105° 100" 95° 90° 85° 80°

Figure 2. — Samples of spotted por-
poise used in life history studies by
5° square. Does not include speci-
mens that were not at least
measured.

Because the field program is a continuing one,
the sample sizes for the various analyses were
different and depended on how much material
was available at the time each analysis com-
menced. Restrictions on sample size are set out in
the text below.

Laboratory Procedures

Fetuses were measured with dial calipers or
with calipers mounted on a 1-m stick. Postnatal
animals were weighed to the nearest pound on
platform scales. Fetuses were weighed to the
nearest gram on a triple beam balance. Testes
were weighed to the nearest gram on a platform
balance. A 1-cm^ cube from the center of each
testis^ and a similarly sized sample of the
epididymis from midlength of the testis were sec-
tioned and stained with hematoxylin and eosin.

*Some early samples were taken near the dorsal surface of
the testis. Tubule diameter in these was subsequently found
not to differ relative to length, weight, and age of the animal
from that in those taken at the center of the testis, and the lots
were therefore combined for analysis.

The mounted sections were subsequently
examined under a compound microscope.

Ovaries were weighed to the nearest 0.1 g on a
platform balance. They were then cut into trans-
verse sections approximately 1 mm thick with a
scalpel and the sections examined under a dissect-
ing microscope. The corpora albicantia in each
ovary were scored to eight categories based on
size, color, vascularization, and gross appearance
(categories described below). If a corpus luteum
was present, it was measured with dial calipers to
the nearest millimeter in its three largest dimen-
sions. The diameter of the largest follicle was
measured to the nearest 0.1 mm.

Age was estimated for 442 animals by exami-
nation of dentinal layers in the teeth. Three or
four teeth were extracted from the lower right
tooth row at approximately midlength and
mounted on wooden blocks in dental wax or plas-
tic resin. A longitudinal section 0.012 inch
(0.31 mm) thick was cut from each tooth with a
diamond saw. The sections were cleared for sev-
eral days in a 1:1 mixture of glycerine and 95%
ethanol, mounted under cover slips in balsam,
and examined with transmitted light under a

233

FISHERY BULLETIN: VOL. 74, NO. 2

compound microscope at approximately 30
diameters. One postnatal layer was considered to
consist of an opaque subunit and a translucent
subunit (Figure 3). The layers in most of the teeth
examined were not as well-defined or as regular
in thickness as those illustrated by Kasuya (1972)
for Stenella coeruleoalba or by Klevezal' and
Kleinenberg (1969) for Delphinus delphis. Teeth
from 39 of the 442 animals were completely un-
scorable, being heavily worn or showing no dis-
crete layers in the sections examined. All the
teeth were scored several times, over a period of
several months, without referring to specimen
numbers or to values obtained previously, until the
scorer felt confident of the results. The values
used in the analyses are those obtained in the
final round of scoring. The teeth were scored to
the nearest postnatal layer when possible, or a
range, e.g., "8 to 10 layers," was estimated. Aver-
age accuracy is estimated at ±1 layer for teeth
with 5 layers or less and ±2 layers for teeth with
5 to 12 layers. Convoluted secondary dentine was
present in most of the teeth vdth more than 12
layers, making counts very difficult and of dubi-
ous reliability. We feel that the counts for many of
these teeth are probably underestimates. Teeth in
which the pulp cavity was entirely closed in all
sections examined were scored as "occluded."

The NORMSEP computer program was used to
define modes in the length-frequency distri-
butions for fetuses. The program was written by
Hasselblad (1966) and modified by Patrick K.
Tomlinson, Inter-American Tropical Tuna Com-
mission. The program separates the mixture of
normal length distribution into its components,
assuming that the lengths of individuals within
age groups are normally distributed and that an
unbiased sample of the length distribution was
obtained.

GROWTH

Length at Birth

Average length at birth of 82.5 cm was obtained
from a linear regression line based on 3-cm group-
ings of fetuses and neonatals (Figure 4). The
largest fetus of the 461 examined was 904 mm
long. The smallest neonatal animal was 780 mm
long. Eighty-six calves and fetuses between 73 and
94 cm were measured in random samples. As-
sumptions inherent in the method used to arrive
at this estimate are that pregnant females and

calves are 1) equally vulnerable to capture in the
purse seine, 2) equally likely to die once captured,
and 3) equally represented in the sample of dead
animals measured. For example, if neonates were
less likely to be included in the samples than were
pregnant females, average length at birth would
be overestimated. Other potential sources of error
are differential rates of prenatal and postnatal
natural mortality and premature births caused by
stresses imposed by pursuit and by capture in the
purse seine.

Gestation Period and Fetal Growth

The most commonly used method for estimating
the gestation time of cetaceans is that of Huggett
and Widdas ( 195 1). They showed that for a variety
of mammals of widely different orders, a plot of the
cube root of fetal weight on age is linear except
during the first part of pregnancy, when growth is
exponential. Their model can be expressed in the
general formula W" = a(t - to), where W =
weight, t = age, a = the "Specific Fetal Growth
Velocity," andto = "the intercept where the linear
part of the plot, if produced backwards, cuts the
time axis." Laws (1959) applied the method of
Huggett and Widdas to fetal length/time data for
three odontocetes [Physeter catodon, Delphinap-
terus leucas , and Phocoena phocoena) and obtained
estimates of gestation periods (15, 14, and 11 mo,
respectively). He assumed that weight is propor-
tional to the cube of length and used the form L =
aitg - to),-whereL = length. This assumption is not
entirely correct (see length-weight results below),
but is a close enough approximation of the real
relationship between length and weight to allow
its use in estimating gestation period. Laws' esti-
mates corresponded closely with other estimates
obtained by more direct methods. Laws' version of
Huggett and Widdas' method is used here.

A gestation period of 11.5 mo was obtained from
an analysis based on 281 fetal and postpartum
specimens collected in January, February, March,
April, May, and October 1972 (Figure 5). The
January-May samples comprised all of the fetuses
of all of the females examined. The postpartum
samples in these months were not random and are
therefore not included. The October samples were
random over all age-classes in the catch; therefore,
all specimens less than 160 cm long, approxi-
mately the length at onset of puberty (Harrison et
al. 1972), are included in the plot. Obvious modes
are present in the length distributions (seasonal-

234

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

\r\

I) '*

:-?•<.

■v^.

Figure 3. — Longitudinal thin sections of teeth from two specimens of Stenella attenuata from the
offshore eastern tropical Pacific. (Left) field number CV300 male, 144 cm, with two postnatal
dentinal layers; (right) number LR55 female, 191 cm, with 13 layers.

235

I—
<

100

Average

length at
birtti

I I I I I

74 77 80 83 86 89 92

(14) (9) (17) (II) (II) (17) (7)

LENGTH (cm)

Figure 4. — Linear regression analysis of percent postnatality
on body length for 86 fetuses and calves of Stenella attenuata
from the offshore eastern Pacific grouped in 3-cm intervals.
Sample size for each 3-cm interval in parentheses.

ity is discussed below). Apparent progression of
the smaller mode in the January 1972 sample is
consistent with a gestation period of roughly 1 yr.
Sample sizes for the other apparent modes are not
large enough for similar analysis. Linear regres-
sion analysis of the modal lengths plotted on
month (Figure 6) yields an estimate of the slope to
use in Laws' equation:

L = 8.283 {t - to), or
length at birth = 8.283 {t^ - to)

with (tg -^(j) (using months of 30. 4 days) = 9.96 mo
or 303 days, where tg = total gestation period.

Laws (1959) proposed that to for length data is
slightly less than for weight data and assumed
^OLn = 0-9 ^0^^- Roughly interpolating between
Huggett and Widdas' values for tf^/tgOf 0.1 for
tg > 400 days and 0.2 for tg = 100 to 400 days
(using provisional tg = 330 days to enter the itera-
tion) (Figure 7) and applying Laws' correction,
^OLn of - 0.135 tg is obtained. This value yields an
estimate of gestation time of 11.5 mo (349 days).
The estimate of ^^ (47 days) is crude, but the true

FISHERY BULLETIN: VOL. 74, NO. 2

AVERAGE LENGTH AT BIRTH

ii*

-g 10

I I I I '~i p

JAN
1972

FEB.

0-*— ^

^ 20i

-FF| n I fp

MAR.

0
20

r-n

APR.

04
201

_a_

P , Fh n n

■T 1 \ r-

MAY

OCT.

0 20 40 60 80 100 120 140 160

LENGTH (cm )

Figure 5. — Length-frequency distributions by month for 281
fetal (open) and young (hatched) postnatal specimens of
Stenella attenuata captured by tuna seiners in the offshore
eastern tropical Pacific in 1972.

value probably lies between 0.12 tg and 0. 15 ^g . We
therefore estimate the gestation period to be 11.5
± 0.2 mo (interval between estimates using ^o =
0.12 tg and 0.15 tg), or 11.3 to 11.7 mo.

Postnatal Growth

The same cohort used for analysis of fetal
growth can be followed through the samples until
approximate length of 125 cm (Figure 8) at the age
of approximately 8.5 mo. In order to optimize res-
olution, 4-cm intervals were used, and the sam-
ples for April, May, and June were combined.
Modes were estimated with NORMSEP.

A linear regression line through the modal
lengths yields the postnatal growth equation

236

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

IbU

140

l?0

100

AVERAGE

LENGTH AT BIRTH

^^-^*

ill

;V^''

(')

^^9^

liJ

^ ^

1 1

1 2

•1971 I 1972-^

MONTH

Figure 6. — Linear regression analysis of modal lengths of fetal
and neonatal specimens of Stenella attenuata (from Figure 4).
Open circles are modes not included in the analysis. Modal
lengths defined with computer program NORMSEP (see
Materials and Methods).

CJ>

0.4

0.3

0.2

0.0

0 100 200 300 400 500

GESTATION PERIOD (days)-tg

Figure 7. — Ratio toltg interpolated between empirical esti-
mates of Huggett and Widdas (1951) — ". . .for gestation times up
to 50 days ^o ~ 0.4 x (gestation time), from 50-100 days
<o-0.3 X (gestation time), from 100-400 days <o- 0.2 x (gestation
time), over 400 days t^— 0.1 x (gestation time)."

L = 82.5 + 5.42 t,

where L = length in centimeters

t = postnatal age in months.

Analysis Based on Analogy with Other Cetaceans

Fetal growth in length of cetaceans, except for

OCTOBER 1972

^^^/'-

1 1 1 1

=3
C

"— '

<J)

III

if)

JANUARY 1973

[ — '

A./W^

1 1 1

FEBUARY 1973

MARCH 1973

-1

-^^J , LT^

oT , , , ,

80 96 112 128 144 160 176 192 208

LENGTH (cm)

Figure 8. — Length frequencies of fetuses and postnatal speci-
mens of Stenella attenuata between 64 and 204 cm long, of
both sexes, by month.

an initial slow phase {t^), is nearly linear. Post-
natal growth during at least a period equivalent in
length to the gestation period is also nearly linear,
but at a lower rate. The difference between the

237

150

4.5 cin/mo^,.''^^

100

1 1 1 1 1 i 1 1 1 1 1 1 1 1 1

\<<

1 11

-15 t-IO -5 /

CONCEPTION X

(-llmo) y^

5 10 15

-50

P. phocoena

/'i 1 cm/mo

Cf+(J

250

I ; . . i I , 1 1 1

200

1 1 1

' 4.0 cm/mo_,^''''''^

-15

-10 -5

-150 5 10 15

CONCEPTION
(-14.5 mo) /

Delphinapterus

-100

leucas

/ W 5 cm

/mo

-.0 9?

300

250

200

I I 1 uL

-J UJ I I 1— i L-

-20 t-15 -10
CONCEPTION
(-16 3 mo)

1 1 6 cm/mo

4 5 cm /mo

-150

100 Globicephalg
melaena

- 50

FISHERY BULLETIN; VOL. 74, NO. 2
550

78 cm/mo.

I I I I I I J I I

-I5| -10 -5

CONCEPTION
(-14.6 mo)

I I I I I

Physeter
catodon

27.8 cm /mo

-200

—J — 1 — I — I — I 1 1 I 1 1 1 1 1 1 i I I 1

15 20

200

150

1 1 I 1 1 1 1

5 1 cm/mo^^

^^\ 1 1 1 1 1 1 1 1 1 J 1 1 1

-15 ' -10

-5 /

5 10 15

CONCEPTION

(-12 mo)

-50

9 3 cm/mo

- S. coeruleoalba
-0 00

Figure 9. — Fetal growth and average postnatal growth during a period equal to the gestation period in five odontocete
cetaceans: Phocoena phocoena (gestation period and postnatal growth from M</)hl-Hansen 1954; to from Laws 1959; length at
birth from Fisher and Harrison 1970); Delphinapterus leucas (from Brodie 1971); Globicephala melaena (from Sergeant 1962);
Physeter catodon (from Best 1968, 1970); and Stenella coeruleoalba (from Kasuya 1972).

fetal rate and the average rate during a postnatal
period equal to the gestation period differs among
the five odontocete species for which sufficient
data exist (Figure 9) and is correlated with length
at birth (Figure 10). The least-squares line for log
of the difference between fetal and postnatal

growth rates (Y ) on log of length at birth (X ) yields
Y = -1.33 + 0.997X, from which a predicted Y of
3.75 cm/mo is estimated for S. attenuata and an
average growth rate in the first year of 4.66 cm/mo
is estimated. This average rate is close to those for
the other three delphinids (5.1 forS. coeruleoalba,

238

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

-o

O) '—

o. o

_ £

y Physeter

O '^

Log Y

= -133 +0997 Log X X

wi "

5 5

Delphinaplerus^/

/ ' Globlcepholo

2 c

o
a> —

Phocoeno /

/v4

r * 2 coeruleoQiba

»- trt

a> a;

5 "

1 2

a> —

■n o

O 3

*" CT

-— ai

Li.

20 30

40 60 100 160 200 300 400 600 800 1000

S attenuoto

Length at birth (cm )

Figure lO. — Relationship of difference between fetal growth
rate during linear phase and average growth rate during post-
natal period equal to gestation period to length at birth in five
odontocete cetaceans. Line is linear regression line of log
difference on log length. Data from Figure 11. Y is predicted
difference for Stenella attenuata from the offshore eastern
tropical Pacific.

4.5 for Globicephala, and 4. 5 for P. phocoe?ia)^; and
yields a predicted length at 1 yr of 138 cm.

Length Relative to Tooth Layers

Total length was plotted on number of postnatal
layers for 115 males and 306 females (Figure 11).
The teeth of five males and three females had
completely filled pulp cavities. These are included
in the plots in a separate category "occluded."

The plots of means for 2-layer intervals (the
points in Figure 12; the curves were fitted as ex-
plained below) very closely resemble the growth
curve obtained by Sergeant (1962) for
Globicephala. Asymptotic length (L ^) for females
is approximately 190 cm and for males approxi-
mately 200 cm. There appears to be a secondary
surge in growth at about 6 layers. With the restric-
tion that the curves must pass through birth
length of 82.5 cm and asymptotic lengths of 190
and 200 cm, it is not possible to fit any continuous
equation to the data satisfactorily. Continuous
curves that fit well at the upper and lower ranges
of layer count seriously underestimate length at 5

*Fisher and Harrison (1970) stated that their data suggest
that Phocoena in Canadian waters grows approximately 30 cm
during the first year of life, or at an average rate of about 2.5
cm/mo, as opposed to the 4.5 cm/mo hypothesized by M0hl-
Hansen (1954). However, they also suggested, and their figure 2
showed, an average rate of at least 5 cm/mo during the first
4 mo. It seems unlikely that the rate would drop to an average
of — 1.25 cm/mo in the remaining 8 mo of the first year.

to 7 layers. Kasuya (1972) also encountered
difficulty in attempting to fit a continuous model
to growth of a delphinid, S. coeruleoalba . Good fits
can be obtained, however, by assuming a dynamic
growth function. A two-phase version of Laird's
(1969) growth model was fitted to the 2-cm means
for all males and females, using an iterative
least-squares method. The occluded specimens
were assigned to the 16+ interval.
Laird's model is

Lit)

exp

exp (- at)

where

L(t) = length at time t

Lq = length at birth ( 82. 5 cm in this case)

t = time (layers in this case)

a = specific rate of exponential growth

a = rate ofdecay of exponential growth.

This model assumes that an organism's growth
pattern is determined at conception. The fitted
parameters a and a express the premise that
"growth is fundamentally exponential (implied by
the normal binary fission of cells), and it also un-
dergoes exponential retardation by some as yet
unknown physiological mechanism" (Laird 1969).

In the two-phase approach, separate equations
were simultaneously fitted to the upper and lower
range of means. The assumptions were made that
juvenile growth is the same for males and females
(supported by the data) and that the growth dis-
continuity comes at about the same age for males
and females. The only fixed point was 82.5 cm at 0
layers (birth). The convergence point (inflection in
the growth curve) was allowed to float to the posi-
tion that gave the best fit, with males and females
considered jointly for lesser ages. The equations
converged at 5.59 layers (rounded off to 6 below) at
which predicted length is 159.9 cm. The fit is excel-
lent for females (Figures 11, 12). Asymptotic
length is 190 cm at predicted age of 18 layers.
Average length of adult females (those with ovar-
ian scars) is 187.3 cm, based on a sample of 555
(Perrin 1975). The largest female of 2,138 mea-
sured was 220 cm long. The equation for juvenile
growth to less than 6 layers is

L = 82.5 exp

0.4817
0.7172

[l - exp (-0.7172^],

where L = length, in centimeters
t = age in layers.

239

FISHERY BULLETIN: VOL. 74. NO, 2

200

% 150
o

X
I-

<s>

21
LU

I-
O

100 -/

. ! • • • ••

N = 120

Average_lenqt^h_cit_bM^tlni ,

0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 OCCLUDED

POSTNATAL DENTINAL LAYERS (number)

—

• • • • •

200

—

• • I f jti^M* * 1 »■ ■ i

•
•
•

—

■^

150

— ^

• %/*

• :

LiJ

-J

N -309

100

. 1 1 1

1 1

Average lengt_h_at_b[rth

III 1 1 1 1

0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 OCCLUDED

POSTNATAL DENTINAL LAYERS (number)

Figure ll. — Scatterplots of body length on number of postnatal dentinal layers from males (top) and females (bottom)
otStenella attenuata from the offshore eastern Pacific. Lines are fit to the growth model (see text).

240

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

220
200

180

160

100-/

_^^i^^-— -

^^^

If i u

- /

Average length at birtti

- 8

23 8 7

N (Cf)

- 7

91 40 12

i 1 1 1 1 i

N(J)

0 1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18
POSTNATAL DENTINAL LAYERS (number)

Figure 12. — Fit of the double Laird growth model (see text) to
2-cm mean values of body length on number of postnatal
dentinal layers for males and females of Stenella attenuata
from the offshore eastern Pacific. For samples greater than
30, ± standard errors indicated as vertical line.

For females with 6 or more layers, the growth
equation is

' 0.0657

L = 159 exp

0.3707

[l - exp (- 0.3707U - 5.588))]

In this case, both the growth rate and the rate of
decay of growth are sharply lower than for
juveniles.

The fit for males is not as good (Figures 1 1, 12) as
it is for females, probably due to greater variabil-
ity and to inadequate sample sizes for the two
oldest strata (the tooth-reading effort was concen-
trated on females because of their importance in
population dynamics). Another possible explana-
tion for the relatively poorer fit for males is that
growth (real, or as inferred from tooth layers) in
adult males is more complex than in adult
females, and a model more complex than the Laird
model is called for. Inferred asymptotic length is
206 cm, achieved at predicted age of 26 layers.
Average length of adult males (defined as those
having testes weighing 200 g or more) is 200.7 cm,
based on a sample of 253 (Perrin 1975). The largest
male of 1,083 measured was 226 cm long. The
growth equation for males with 6 or more layers is

L = 159.5 exp

0.0524

, 0.2032
[l - exp (- 0.2032(^ - 5.588))]

The secondary growth rate (a, 0.0524) is very
slightly smaller than for females, but the rate of
decay {a, 0.2032) is sharply smaller, reflecting the
attainment of greater size in males. The equations
rearranged and reduced for estimating age (in
terms of layers) from length are

^(M and F <160 cm) = -1.394 In (7.531

- 1.48 In L)
= 5.588 - 2.698 In (29.606

- 5.64 In L)
= 5.588 -4.921 In (20.669

-3.878 In L).

^(F ^160 cm)
?(M^160cm)

Note: These equations should not be used to esti-
mate age from length except for grouped samples
of smaller animals (about 180 cm or less), for
which growth rate is still large compared to indi-
vidual variation in length.

The juvenile growth curve based on tooth layers
can be calibrated for the first year by comparison
with the growth curve derived from analysis of
modal progression (above) and by deduction from
what is known about juvenile growth of other
odontocetes (the fetal-postnatal growth argument
above). Estimated average length at 8 mo based on
analysis of modal progression is 125.5 cm. The
predicted number of layers at that length (Figure
12) is 1.53. If the average growth rate during the
first year is assumed to be the same as the average
during the first 8 mo, the predicted number of
layers at 1 yr (1.53 • 12 - 8) is 2.3. This extrapola-
tion, however, is a slight overestimate, because
while growth during the first year in delphinids is
approximately linear, there is some decay of rate.
The predicted number of tooth layers (using Fig-
ure 12) at 138 cm, the above-predicted length at 1
yr based on camparison with other odontocetes, is
2.0. It seems safe to assume that about 2 layers are
laid down during the first year of life.

Calibration of the remainder of the tooth-layer
curve is more difficult. Kasuya et al. (1974)
examined the innermost layer in teeth of S. at-
tenuata and related type and thickness of layer to
season of capture. They concluded that one layer
(one transparent plus one opaque subunit) repre-
sents 1 yr of growth. We found no correlation
between thickness of the innermost layer and
season of capture. Almost all of the samples for
which teeth were sectioned, however, were col-
lected in the first few months of the year. Lacking
such direct calibration, several alternative pos-
sibilities can be examined. The results, however.

241

must remain tentative and inconclusive until
growth has been monitored directly in one or more
captive or free-ranging, tagged individuals.
Some alternatives that can be considered are:

1. Two layers per year until the teeth are
occluded.

2. Two layers in the first year and one per year
thereafter until the teeth are occluded.

3. Two layers per year until puberty (about
nine layers in males and seven in females;
see section below on age at puberty), and one
per year thereafter.

This list of alternatives can be extended to great
length by making assumptions such as that layers
are laid down at irregular intervals, males and
females lay down layers at different rates, layers
disappear with age, etc., but the above are proba-
bly the main possibilities that should be con-
sidered. All references below to age are in terms of
layers, with the above alternative possibilities
considered or implied. None of the alternatives
can be eliminated with certainty. One tooth layer
deposited per year has been inferred for the west-
ern Pacific population of S . attenuata by Kasuya et
al. (1974). One layer per year has also been sug-
gested for other closely related delphinids, includ-
ing S. coeruleoalba (Kasuya 1972) and Tursiops
truncatus (Sergeant et al. 1973). Two tooth layers
per year have been found in Delphinapterus leucas
(Sergeant 1973), but this form is less closely re-
lated to Stenella. Thus, there is more support in
the literature for the one-layer-per-year model
(number 2 above) than for the others.

Length-Weight Relationships

Length-weight relationships were determined
for 218 fetuses, 66 postnatal males, and 33
nonpregnant, postnatal females by using linear
regressions of log weight on log length.

Fetuses

The fetuses ranged from 20 to 897 mm long and
weighed from 2 to 7,588 g. Ten fetuses less than 20
mm long were not included. The regression equa-
tion is

log W = 3.5532 + 2.501 logL,

where W = weight in grams

L = length in millimeters.

242

FISHERY BULLETIN: VOL. 74, NO. 2

In exponential form, the relationship is

W = 2.79 X 10-4 L2-501.

Females

The females ranged from 100 to 200 cm and
weighed from 12.0 to 69.1 kg. The regression equa-
tion is

log W = -4.1576 + 2.6120 log L,

where W = weight in kilograms

L = length in centimeters, or in exponen-
tial form, W = 6.95 x 10"^ L^^'^.

Males

The males ranged from 86 to 218 cm and
weighed from 6.8 to 90.0 kg. The regression equa-
tion is

log W = -4.7135 + 2.873 logL,

where W = weight in kilograms

L = length in centimeters, or in exponen-
tial form, W = 1.93 x lO'^ L^-^^s

The slopes of the regression equations are
statistically different (^-test ata = 0.05) for males
and females. Males are lighter for their length at
birth, and heavier for their length after about 135
cm has been attained.

Color Pattern

Perrin (1970b) has previously described the de-
velopment of the color pattern of S. attenuata in
the offshore eastern Pacific. The animal begins
life unspotted, develops dark spots ventrally that
later coalesce, as light spots develop dorsally. The
ontogenetic continuum can be divided into five
stages as defined below and as shown in Figures 13
and 14:

1. Newborn stage. Dark purplish-gray dorsal
surfaces and lateral blazes, with white ven-
tral surfaces and no spots; about 80 to 160 cm.

2. Two-tone stage. General two-tone pattern
with dark-gray surfaces above, lighter gray
lower surfaces, and a well-defined pattern in
varying shades of gray about the head and
flippers; no spots; about 95 to 175 cm.

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

The division between this and the pre-
vious category is somewhat subjective and
arbitrary.
3. Speckled stage. Same as two- tone but with
discrete, very dark-gray spots on the ven-
tral surfaces; discrete light-gray spots on the
upper, darker surfaces present on some
animals but lacking on others; about 140 to
190 cm.

120-
110-
100 -
90 -
80-
70-
60-
50-
40-
30-
20-
10-

1x1
O

FUSED
(604)

AVERAGE

0
30

20-

10 -

0 :

50-
40-
30-
20-

10-

0 ■

MOTTLED
(138)

AVERAGE

_tj_

SPECKLED
(255)

0
20

AVERAGE

' i

. i>-i .1:1

2- TONE
(288)

NEONATAL
(72)

1 I i

70 60 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230

TOTAL LENGTH (cm)

Figure 13. — Length-frequency distributions of males of
Stenella attenuata from the offshore eastern Pacific, by color
pattern phase.

Mottled stage. Ventral spots converging and
overlapping in places, but patches of the
lighter gray background still visible, yield-
ing a mottled effect; discrete or merging
light-gray spots present on the upper sur-
faces; about 155 to 210 cm.
Fused stage. Ventral spots completely
convergent, to give the effect of a uniform,
medium-gray to dark-gray surface; on close
inspection, the individual overlapping spots
still discernible; about 160 to 230 cm.

REPRODUCTION

Seasonality

Nishiwaki et al. (1965) suggested that S. at-
tenuata in Japanese waters breeds in the spring
and in the autumn. Harrison et al. (1972) stated
that lengths of fetuses indicate that parturition
in the eastern tropical Pacific (of S. graffmani =
S. attenuata ) also occurs both in the autumn and
in the spring. The postnatal length-frequency
data for large samples (Figures 15, 16; April 1968
and October 1972, for example) support the thesis
of major reproductive seasons in spring and au-
tumn but also suggest that there is a reproduc-
tive peak in summer as well. There is year-to-
year variation in the timing of reproductive
peaks, and there is some reproduction occurring
throughout most of the year. It is difficult to
define the reproductive seasons with precision be-
cause most of the sampling effort was in the early
(January- April) and late (October-December)
parts of the calendar year. The sampling inter-
sected obvious calving seasons in April 1968,
January 1972, October 1972, January 1973, and
June 1973 (Figures 15, 16). Calving peaks were
probably also present in some of the other sam-
pling months, but the samples were too small to
detect them or were biased in some fashion. A
summary of predicted birth dates for 373 fetuses
more than 15 cm long collected in 1971, 1972, and
1973, however, indicates that there may have been
three calving peaks in each of the 3 yr (Figure 17).
In each year there was a definite calving low in
winter. The synchrony was diffuse, and some
peaks were much sharper than others. The statis-
tical evidence for three annual peaks in calving is
weak, and when the data for all years are com-
bined, all that can be said with certainty is that
the calving season is prolonged, with a low point in
winter and a tendency for high points in spring
and fall.

The Male

Sexual development of the male was examined
under three criteria: 1) weight of testes, 2) aver-
age diameter of seminiferous tubules, and 3)
amoimt of sperm in the epididymis. Each of these
was examined relative to total length, weight,
and age (number of postnatal dentinal layers).

Weight of the testes (Figure 18) increases pre-
cipitously at body length of about 175 to 190 cm,

243

FISHERY BULLETIN: VOL. 74, NO. 2

3IOp
3CX)
290
280
270
260-
250-
240-
230-
220-
210-
200-
190-
180-
170-
160-
150-
140-
130-
120-
110-
100-
90-
80-
70-
60-
50-
40-
30
20-
I
0
60-
50-
40
30-
20
10

0
60

50-

40-

30-

20-

FUSED
(g83)

' I ' ' ■

I ■ I ■ I ■ I ■ I

AVERAQE

MOTTLED
(194)

'''!''■

. I

SPECKLED
(210)

m/BMBE

I ''■ ■ ' 1-^ I

Figure 14. — Length-frequency distributions of fe-
males of Stenella attenuata from the offshore east-
ern Pacific, by color pattern phases. On cruises be-
tween January 1971 and October 1972, adult
females (3^160 cm) were selected for measuring and
dissection. Earlier and later samples were non-
selective. Average lengths for neonatal, two-tone,
and fused are based on all the samples (no length
bias), and averages for speckled and mottled are
based on the nonselective samples (shaded). The
analyses of coloration transition are based on all
the samples for neonatal-to-two-tone and on the
nonselective samples for the remaining three
transitions. Size of sample used for calculation of
average is given in parentheses.

m/fjmt

. i i I t R-t-<

80 90 K)0 110 120 130 140 ISO 160 170 180 190 200 210 220 230

TOTAL LENGTH (cm)

244

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

AVERAGE LENGTH AT BIRTH

> 0

2 5
UJ

cr 5

0
5
0
5
0
5
0
5
0
15
10
5
0

NOVEMBER 1971

DECEMBER 1971

■ ' ' ' ^~i ''"' I

JAN 1972

'r^!;^^^^

MARCH 1972

APRIL 1972

-^ — .n

MAY 1972

r^l

I I

OCTOBER 1972

LiT_,

10
5
0
5
0
10
5
0
5
0
5
0
5
0
5-

j I I I i_i

,,n,

MAY 1973

OCTOBER 1973

r-l . . f—i I I 1 ) ) f— 1

-i 1 \ I I ■ ■ I

NOVEMBER 1973

i~i i r-t

' i i I 1 L.

DECEMBER 1973

-I l_l I L_i L.

I I I I I""" I I I I I

75 100 125 150 175 200 225 250 75

LENGTH (cm)

100 125 150 175 200 225 250

LENGTH (cm)

Figure 15. — Length-frequency distributions of postnatal male specimens ofStenella attenuata, 1968-73, by month.

but in animals larger than about 200 cm, there is
little correlation with length. The largest testes
encountered weighed 2,400 g and were possessed
by a male 196 cm long. However, some males
more than 210 cm long had testes weighing less
than 300 g. Testes weight begins to increase
sharply at 50- to 55-kg body weight and is
strongly correlated with weight in larger ani-
mals. Males in the sample that weighed more
than 70 kg (eight animals) had testes weighing
more than a kilogram. The male with the third
heaviest testes (2,017 g — heaviest testes for
which body weight also available) weighed 80 kg;

the heaviest male in the sample weighed 91 kg
and had testes weighing 1,348 g. A rapid increase
in testes weight (Figure 19) occurs at age 7 to 13
layers, with maximum size increasing until 12 to
16 layers. All animals with more than 14 layers
had testes weighing 500 g or more. Again, there
is wide variation in testes size relative to age.
Part of the variation is ascribable to the consider-
able error in the estimate of number of dentinal
layers (±2 layers for animals with more than 5 to
12 layers, more for older), but it must be con-
cluded that there is probably about a 5-layer
period during which the onset of puberty may

245

FISHERY BULLETIN: VOL. 74, NO. 2

bj 65

O 60
UJ

<E 55
U.

50
45
40
35
30
25
20
15
10

0
5
0

20
15

10
5
0
10
5
0
5
0

AVERAGE LENGTH AT BIRTH

APRIL 1968
(120

JANUARY 197!

I H~l

I I J i I I I ^

FEBRUARY 1971

_i I — ^ — L_^

MARCH 1971

MAY 1971

_i I J i I I I ' ' '

OCTOBER 1971

NOVEMBER 1971

' ' ' ' '

DECEMBER 1971

JANUARY 1972

FEBRUARY 1972

35
30
25

20
15
10
5
0
40
35
30
25
20

' ' i-i ' ■ ' ' ' ' 5

0
^ I I i '-I I i_i-j 35

J I 1 I 1 1—1 1 L_i 25

I ' I i-i-i-i 15

MARCH 1972

APRIL 1972 *

' ' I I I ■ ' I I ' I ' ' ' ' ' ' ' ' ' I

MAY 1972 *

^J^^

-^ 0

AVERAGE LENGTH AT BIRTH

OCTOBER 1972

FEBRUARY 1973

^^J""W^

MARCH 1973

J^— M^^U ^,,,

APRIL 1973

MAY 1973

m .— n

'''II'

!=!IL

JUNE 1973

it--

' ' I ' '

OCTOBER 1973

I I I ' ■ I ' ' '

JIl

NOVEMBER 1973

DECEMBER 1973

' I I ■ I ' '

n r-i n

Figure 16. — Length-frequency
distributions of postnatal female
specimens of Stenella attenuata,
1968-73, by month.

■'■5 100 125 150 175 200 225 75 100 125 150 175 200 225

LENGTH (cm) LENGTH (cm)

246

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

take place (age about 7 to 12 layers) and that
about 2 to 4 layers are required to attain "adult"

- 20

UJ
CO

It: 15

UJ
U-

z
o
I-
a:
o

o
<r
a.

rui

I I I I I I ' ' ' I

JFMAMJJASONDJFMAMJJASONDJFMAMJJiSOND

1971 1972 1973

(n = 34) (n = l52) (n = l87)

MONTH OF BIRTH

FIGURE 17. —Predicted month of birth for 373 fetuses oiStenella
attenuata, based on fetal growth curve.

testes size (500 to 2,000 g). The third largest
testes were possessed by an animal (202 cm long,
80 kg, discussed above) that had nonreadable
teeth that were worn to the gum in all four tooth
rows. Such tooth wear may be a correlate of rela-
tively great age.

The diameter of the seminiferous tubules be-
gins to increase rapidly at body length of about
155 to 170 cm (Figure 20), or at lengths about 15
cm shorter than those at which testes weight be-
gins to increase. Tubule diameter is definitely
correlated with body length until at least about
200 cm. The heaviest male (91 kg) had the largest
tubules. The plot of tubule diameter on layers (Fig-
ure 21) indicates that the tubules enter a rapid
development stage at 6 to 11 layers, before the
onset of a rapid increase in testes weight (Figure
19). Asymptotic diameter is about 170/xm and ap-
pears to be attained by 10 to 14 layers.

2400

2000

— 1,500

X
O

LlI

I-

C/)
UJ

1,000

500

v.. «•

M • • . • .

t • I..

I«

100

■f-

»«»*

m* * t , 1..

• ••! ' 1 , **

. -••• • . •

120

130

180

190

200

140 150 160 170

BODY LENGTH (cm)
Figure 18. — Scatterplot of testes weight on body length in Stenella attenuata

210

220

230

247

FISHERY BULLETIN: VOL. 74, NO. 2

1,800

1400 -

1200

tl 1000 -

800

400-

200-

• •

■ ^■^■i^-^-i» T>t!»-».*if»

0 1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19
LAYERS (number)

Figure 19. — Scatterplot of testes weight on number of postnatal
dentinal layers in Stenella attenuata.

~i.

Z50

•

• .

13
CD

200

• • •

150

••. *•► .• . •

• •• n •

••

•

100

• •

UJ
CO

•

• .

u.
o

•
•

.•

•
•
•

•• •

•
•

1 .. .• ^w.*

•

1 1

L-

1 1 1 1 1 1

100 120 140 160 ISO

BODY LENGTH (cm)

220

Figure 20. — Scatterplot of average diameter of seminiferous
tubules on body length in Stenella attenuata.

250

200

•

■

t-
to

•
•

•
•
t

•

• •

3
O

150

•>

•

• • •

•

Lj.

•

100

•

U-

•
•

•

•
•

•

t •

• ■T

•.•

• •

•
• •

•
•

•

•
•

1
t

1 1 1

J-

-L

-1.

1 1 1

1 1

Q 0 I 2 3 4 5 6 7 8 9 10 M 12 13 14 15 16 17 IB 19

LAYERS (number)

Figure 21. — Scatterplot of average diameter of seminiferous
tubules on number of postnatal dentinal layers in Stenella
attenuata.

248

Sperm in the epididymis were scored as "ab-
sent," "present in small numbers," or "copious"
(easily seen in the histological sample without
searching). The shortest individual with large
numbers of sperm in the epididymis was 179 cm
long. This animal weighed 62 kg. In animals
larger than 180 cm and heavier than 58 kg, pres-
ence or absence of sperm in the epididymis bears
little relationship to total length. Thirty-six large
adults (>200 cm) were equally distributed among
the three categories of no sperm, some sperm, and
copious sperm.

The smallest testes bearing epididymis with
sperm weighed 200 g, and the smallest testes
with copious sperm weighed about twice as much.
Some animals with testes heavier than 1.5 kg,
however, had no sperm in the epididymis. The
same pattern of wide variation is apparent in the
relationship between epididymis code and layers.
The youngest male with sperm in the epididymis
had 9 layers. The youngest animal with copious
sperm had 10 layers. After about 10 layers, there
appears to be no relationship between age and
presence or absence of large numbers of sperm.

In summary, the onset of puberty, as indicated
by a rapid increase in diameter of seminiferous
tubules and increase in testes weight, is at 7 to 12
layers (average ~-9 layers; an estimate of ages at
puberty) and at lengths of 155 to 170 cm and
weights of 40 to 50 kg. Sexual maturity is at-
tained about 2 to 4 layers later, at 10 to 14 layers,
^180 cm, and &58 kg. The midpoint of the range
of 10 to 14 layers, or 12 layers, may be taken as an
approximation of average age at attainment of
sexual maturity. Whether or not males at this
point are "socially mature" (sense of Best 1969)
can be determined only through behavioral
studies. Average length of males 12 layers old is
about 195 cm, and average weight is about 75 kg.

The Female

Attainment of Sexual Maturity

Harrison et al. (1972) described and figured the
ovaries of S. attenuata (as S. graffmani). The
ovaries weigh less than 0.5 g each at birth.
Weight increases gradually to about 1.5 g at
about age 6 to 8 layers (average ~7 layers; an
estimate of age at puberty), when there is a sud-
den increase in average ovary size and weight
due to presence of corpora of ovulation and/or
pregnancy (Figure 22). This change comes at an

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

16-

X 10
C2

LlJ

% ^

«. o°°

° ° o °

° ° ° o °

° 8

= • ° c^." 1 8

T° o ° •

8 °o8

too •

•

• o •

•

• • °

•

.'*?.i h:.*

•

•
•

••
•

•

• f. -th .

• • ,

• • .

► • 1

i , 1 , 1 , i

— 1 —

. 1 i i 1 i 1 1

0 2 4 6 8 10 12 14 16 18

LAYERS (no)

Figure 22. — Relationship between weight of ovaries and
number of postnatal dentinal layers in Stenella attenuata.
Open dots are females with a corpus luteum.

average total length of about 170 to 180 cm and
weight of 50 to 60 kg (Figure 23).

Analysis of lengths of females with and without
ovarian scars yields a more precise estimate of
length at attainment of sexual maturity. The
smallest of 1,410 specimens (160 cm long or
longer) that possessed scars were two that were
167 cm long (one with a corpus luteum only and
one wdth a corpus luteum and four corpora al-
bicantia). The largest female with no scars was
193 cm long. The length-maturity curve is
slightly asymmetrical, but a linear regression
line through the nearly linear central portion (M
= 5.76L - 960.95) estimates that average length
at which scarring is first evident is 175.4 cm. This
analysis probably underestimates length at at-
tainment of maturity, because some of the small
adults (170 to 180 cm) with many scars are those
that have stopped growing at a shorter-than-
average length. In other words, the left-hand por-
tion of the frequency distribution of physically
mature adults to an unknown extent artificially
elevates the central portion of the length-
maturity curve, making it asymmetrical.

An estimate of age and length at attainment of
sexual maturity can also be derived directly from
the smaller sample of females for which the
number of tooth layers was determined. The
youngest specimen exhibiting ovarian scarring

had 7.5 layers. The oldest with no scarring had 12
layers. The estimated age at which 50% have
scars is 9.14 layers (M = 19.5^ - 128.25). Pre-
dicted length at this age is 181.6 cm (based on
growth equation above). This estimate is less
biased than the others above but based on much
smaller samples, especially at the lower end of
the layer-maturity curve.

Another estimate of length and age at first ovu-
lation can be made by back extrapolation of a
relationship between body length and number of
corpora (including corpora lutea) in the ovaries
(Figure 24). Length increases with corpora count
until at least six to eight corpora have been ac-
cumulated, at about 183 to 190 cm. A fit of the
data to the Laird growth model (above) yields the
equation

L = 180.17 cm exp{0.0541[l - exp(-0.2815C)]},

where L = length in centimeters
C = number of corpora.

Back extrapolation of the curve to zero corpora
yields an estimate of 180.2 cm. Predicted age from
the growth equation is 8.74 layers.

An estimate of length at first conception can be
made by calculating the average length of preg-
nant females with a corpus luteum only (indicat-
ing first pregnancy) and subtracting the growth
that they can be assumed to have undergone dur-
ing pregnancy. Fifty-four primiparous females
averaged 181.7 cm in length (range 167 to 193 cm).
Predicted age at that length is 9.17 layers. The
average length of their fetuses was 372 mm. This
length is attained by about the beginning of the
sixth month of gestation. Using the growth equa-
tions above to predict growth during 6 mo for the
various tooth-layering models and substracting
the growth increment from 181.7 cm yields esti-
mates of length at first conception ranging from
177.7 to 180.0 cm (number 4 in Table 2). The
primiparous females in this sample, however, are
only those that became pregnant at the first ovu-
lation. This may cause the estimate to be an
underestimate, because many females ovulate
several times, and presumably continue to grow,
before becoming pregnant the first time (see
Ovarian Changes below).

The various methods of estimating age and
length at attainment of sexual maturity yield es-
timates of varying accuracy (Table 2). The esti-
mates based on tooth layers and length at first

249

20-

— 16
en

E 1^

^ 12

CO
LU

CC 10

<
>

2 -

FISHERY BULLETIN: VOL. 74, NO. 2
126

•I

90 100 110 120 130 140 150 160 170

BODY LENGTH (cm)

190

200 210

Figure 23. — Scatterplot of weight of both ovaries on body length in Stenella attenuata. Open dots represent females with a
corpus luteum. For animals 160 cm or longer, where sample for 5-cm interval is 10 or more, means (circled symbols) and ranges
are graphed and points are not plotted. Where the sample is &30, ± two standard errors are indicated by bars.

conception are the best of the four and probably
bracket the true values. Under method number 3,
age hypotheses numbers II and III are more prob-
ably correct than number I. Accordingly, we esti-
mate that sexual maturity is, on the average, at-
tained at 181 ± 1 cm and 9.0 (8.6 to 9.3) layers (5.1
to 8.3 yr, depending on the alternative layering
hypothesis used).

An increase in size of Graafian follicles is
another criterion of approaching sexual maturity.
Diameter of the largest follicle also shows a sharp
increase after 160 cm total length (Figure 25), con-

250

current wdth the increase in ovary weight (Figure
23). The largest follicle in immature females usu-
ally is less than 1 mm in diameter. The largest
follicles in most ovaries containing scars are be-
tween 1 and 8 mm in diameter, but a few follicles
(possibly cystic) as large as 10 to 16 mm in diame-
ter were encountered.

Ovarian Changes in Adults

The analyses of ovarian changes are based on
material collected through 1972. The corpus

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE
2IOr

2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18

CORPORA (number)

FIGURE 24. — Relationship between body length and number of
corpora in Stenella attenuata. Average (bar), ± two standard
errors (box), range (vertical line), and sample size shown.

Table 2. — Results of analyses of length and age at attainment
of sexual maturity in Stenella attenuata, with comments (in
parentheses) on pros and cons of the methods. Lengths and layer
counts predicted with the growth equations are in parentheses.

Analysis

Length
(cm)

Layers
(no.)

Age (yr)
under hypoth

1 II

esis
III

Length at which 50% have cor-
pora (probable underestimate).

175.4

(7.66)

6.7

Number ot tooth layers at which
50% have corpora (interpolation,
but small sample sizes).

(181.6)

9.14

4.6

8.1

5.6

Back-extrapolation of corpora-
length cun/e (large samples, but
extrapolation).

180.2

(8.74)

4.4

7.7

5.2

Length at first conception under
hypothesis: 1

II
III

(includes only those that become
pregnant at first ovulation; prob-
able underestimate).

177.7
180.0
180.0

(8.17)
(8.57)
(8.57)

(4.1)

7.6

5.1

luteum of pregnancy arises from the ruptured fol-
licle and has an important secretory function in
maintaining early pregnancy in all mammals
and full gestation in most (Amoroso and Finn
1962). The gross and microscopic structures of
corpora lutea in various delphinids, including S.
attenuata, have been described by Harrison et al.
(1972).

The corpus luteum decreases in size during ges-
tation (Figure 26). Of 242 females with corpora
lutea, 229 were pregnant. Eleven with fetuses
less than 20 mm long (range 1 to 20 mm) had

_J
o

LJ
O

Ll_
O

LJ
H
LlI

140 150 160 170 ISO 190

BODY LENGTH (cm]

200 210

Figure 25. — Relationship between body length and diameter
of the largest Graafian follicle in Stenella attenuata. Open dots
represent females with corpus luteum. For length ^160 and
n 5=10, means (circled symbols) and ranges shown. Forn 3=30, ±
two standard errors are shown. Not included are 27 "senile"
specimens with follicles <0, 1 nmi and five juveniles 88 to 122 cm
with 0- to 1-mm follicles.

corpora with diameters of 23 to 29 mm (average
26.0 mm, SD 2.90). The mean diameter dropped
sharply to 23.6 mm (range 21 to 27 mm, SD 2.27)
in 17 females with fetuses between 20 and 100
mm (using Student's t, means are significantly
different at a = 0.01). This amounts to about a
32% decrease in luteal volume. Size of the corpus
luteum continues to decrease at a slower rate, to
22.2 mm (range 19 to 28 mm, SD 1.79) in females
with fetuses 700 to 825 mm (average length at
birth is 825 mm) long, a further decrease in vol-
ume of about 15%. Luteal volume in females with
near-term fetuses is only about half of that
shortly after conception. Mean diameter in 10
females with fetuses longer than average birth
length (825 mm) was 24.0 mm (range 20 to 26 mm,
SD 2.21, greater than mean for 700 to 825 mm at
a = 0.01), a volume difference of about 38% more
than for fetuses 700 to 825 mm long. Delayed re-
gression (or re-enlargement) of the corpus luteum
is apparently correlated with greater-than-
average length at birth.

251

FISHERY BULLETIN: VOL. 74, NO. 2

1 1 1 1 1 1 1 1 1 1 1 1 1

!£2 Q in

1 1 1 1 1 1 1 1 1 1 1 1 1 t

COS

33
OQ-LlI

•"ogoSS^g^oSScxft^, o

(VJ

•

iijt/5

^^l-

OZ)

: • t

UJ<

liJ

o_i

0 oo o

o
o

II
0

I-

UTERI

NOT

DISTEN

Q
<rt UJ

=)0

® « ® ®

i ®

UTER
DISTEN

®» •

• • •

•

_,J. ^

—

O
O
CO

- n .

«>

!-5 •

—

. t .•••

• • *
•

1 ,

. • • •

—

o -p

O c

• • • • '

• ® ® • :

o =>

• •

—

o ^-

a.
o

• « 1 • •
• ' • •

lO UJ

t .

O O o
O 2

• t ® ®
•: i

^ X 1-

•® l-n.

z: <

O UJ > -J

—

O _l -! Q
ro 2 z

o <

. •!

1- t-

Z 2

•

r^ < <

9 2 Z

—

O o o

OJ LlI LiJ

^^ cr q:

Q. CL

® • •

. r-

II II

o • ®

—

• * t
® • :

® • t

r?^-«t-'

rTii|riii|iii

1 1 1 1 1 1 1 1 1 1 1 1 1 1

—^

C\J

(ujuj) sndaoo dO a3i3kMvia

o
G

o
u

.4-)

C
CO
u

0)
13

o
o

c
til

ca
E

T3
C
CO

M
C

tn
3
P.

!-•

5
■5

0)
CO

D
O

252

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

Nine obviously postpartum females had cor-
pora lutea 13 to 25 mm in diameter (Figure 26).
Four lactating females with uteri not obviously
distended had corpora lutea in the same size
range. Average luteal volume in these lactating
animals was less than half of that in animals at
parturition. Some of these 13 cases may represent
miscarriages.

The corpus luteum of pregnancy shrinks still
further during the suckling period, losing its
glandular appearance and becoming a corpus al-
bicans. Nine of 197 lactating females without cor-
pora lutea each had a single corpus albicans,
which must represent the regressed corpus
luteum of the first pregnancy. These corpora (Fig-
ure 26) were approximately spherical and 5.9 to
10.6 mm in diameter (average 8.5 mm). The lower
end of this range — about 6 mm — must approx-
imate the limit of regression during the suckling
period (about 11.2 mo; see below). The small
number of lactating females with corpora lutea
(13) compared to the number with only copora al-
bicantia (197) indicates that initial regression fol-
lowing parturition must be very rapid, perhaps
occurring in less than 15 days. Still further re-
gression in size and histological structure of the
corpus albicans of pregnancy probably occurs.
Many adult females have a large corpus albicans
(in most cases, one of several) between 3 and 6
mm in diameter (Figure 26) with greatly degen-
erated structure. Unless these corpora all repre-
sent ovarian events not resulting in pregnancy,
i.e., the females are all completely barren, the
corpus albicans of pregnancy must decrease in
diameter during a resting period following a
pregnancy, to possibly as little as 3 mm.

Multiple corpora lutea are uncommon in S. at-
tenuata. They were encountered in only 2 out of
258 females with corpora lutea. One of these was
pregnant with twin fetuses (males, 83 and 86
mm) in the left horn of the uterus. The left ovary
contained two corpora lutea of approximately
equal size, each possessing a surface scar of ovu-
lation, together with seven corpora albicantia vis-
ible on the surface. The right ovary was devoid of
scars. Another female with two corpora lutea had
a 592-mm fetus (male) in the left horn of the
uterus. The left ovary looked very much like that
of the specimen with twin fetuses, having two
corpora lutea of approximately equal size and
eight corpora albicantia on the surface. Neither
corpus luteum bore a discernible surface scar. The
right ovary was unscarred. There are two possible

explanations for the presence of two corpora lutea
in this specimen: 1) one of them was an accessory
corpus, or 2) one of a pair of twin fetuses was
aborted during early pregnancy. In any case, the
incidence of multiple corpora lutea is very low in
S. attenuata, less than 1% in the sample
examined. This is in sharp contrast to some other
cetaceans, in which rates of presence of accessory
corpora range to 15.6% iDelphinapterus leucas
— Brodie 1972). The contribution of double and
accessory corpora lutea to the accumulation of
corpora albicantia can be considered to be neg-
ligible in S. attenuata.

Corpora albicantia in S. attenuata represent
both regressed corpora lutea of pregnancy and re-
gressed corpora of ovulations that do not result in
pregnancy. This conclusion is based on the ac-
cumulation rate of corpora albicantia and on the
estimate of the mean length of the calving inter-
val (see below). We were not, however, able to
differentiate between small regressed corpora
lutea and regressed corpora of ovulation. This
impasse, also encountered by workers dealing
with other cetaceans (Harrison et al. 1972) is
caused by the wide and largely discordant varia-
tion in size, shape, surface texture, and internal
structure and color of the corpora albicantia. If
one looks at enough corpora, it is possible to find
corpora with these characters in almost any com-
bination of expressions.

Harrison et al. (1972) found no more than six
corpora albicantia in the ovaries of any Stenella
female. In the present sample, however, nearly
half (44%) of the females had more than six cor-
pora, including the corpus luteum. Fifty-five
females of 1,131 had 15 or more corpora; one had
28 (Figure 27). Three thousand five hundred and
two corpora from ovaries of 530 females were
scored to six categories. These categories are
somewhat arbitrary in view of the continuity of
regression and the wide variation discussed
above, but, nonetheless, they are useful in
analyzing the course of regression. The numbers
and proportion of total corpora complement rep-
resented by each of these categories varies with
the total number of corpora (Table 3, Figure 28).
The categories were defined as follows:

Type 1. Surface raised, smooth or slightly
wrinkled. Looks externally like a small corpus
luteum. Cortex white or yellow, with obvious
remnants of vascularization. Center solid or
loosely constructed, consisting mainly of white

253

FISHERY BULLETIN: VOL. 74, NO. 2

120 r-

100

>-
o

3
O

±1.

Th^ I

5 10 15 20

TOTAL CORPORA (number]

Figure 27. — Frequency distribution of corpora count in 1,131
females ofStenella attenuata.

connective tissue, 3.5 to 15.5 mm in diameter, av-
erage 7 mm. These corpora almost certainly are
nearly all regressed corpora lutea. Four hundred
fifty-six were encountered (13.29f ). Females w^ith
two or more corpora have, on the average, about
one Type 1 corpus (Figure 28), although as many
as five may be present (Table 3).

Table 3. — Types of corpora present in ovaries ofS. attenuata in
relation to total corpora. Averages in Figure 28.

Total number

Range of

number of each

tvoe of

of corpora

(including

corpora lutea)

Sample
size

corpus albica

ns— Type;

(no.)

0-1

0-2

0-1

0-3

0-2

0-1

0-3

0-4

0-3

0-1

0-2

0-5

0-4

0-5

0-2

0-5

0-4

0-6

0-3

0-1

0-5

0-6

0-3

0-2

0-1

0-5

0-6

0-7

0-2

0-4

0-2

0-5

2-9

0-4

0-3

0-1

10-11

0-4

0-5

3-10

0-2

0-1

(average 10,4)

12-14

0-2

0-5

3-14

0-3

0-4

0-3

(average 12.9)

15-27

0-4

0-5

7-19

0-2

0-9

0-1

(average 17.2)

Total

513

0-5

0-6

0-19

0-4

0-9

0-3

Type 2. Surface raised and wrinkled. Interior
white to yellow, often v^dth traces of luteal cortex
and vascularization. Center solid or loosely con-
structed, consisting mainly of white connective
tissue. Definitely less integrated in structure
than Type 1 (above). Diameter 3.0 to 12.0 mm,
average 6 mm. The evidence on accumulation
rate (below) suggests that these corpora are prob-
ably a mixture of regressed corpora lutea and
corpora of ovulation. We found 787 of this type
(22.5%). The number of Type 2 corpora is rela-
tively constant in females with three or more cor-
pora, at about one and one-half (Figure 28) vdth a

^ 12

c
<v
o

IjJ

Z)
Q.

tr
o
o

z
o

I-
cr
o

o
cr
a.

2 3 4 5 6 7 8 9 10 I I 12 13 14 15 16 17 '27
TOTAL CORPORA (number)

Figure 28. — Relationships between numbers of corpora of various types and total number of corpora i
ovaries of females of Stenella attenuata. Ranges and sample sizes in Table 4.

254

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

range of 0 to 6 (Table 3). This number is tightly
correlated with the number of Type 1 corpora
(Figure 28), indicating that there may be some
overlap in the classification criteria for these
categories.

Typed. Surface usually not raised; scar usually
smaller than Type 2 and heavily wrinkled. May
be pedunculate and flattened. May be flattened
against the surface or may run deep into the ova-
ry. Interior consists of white connective tissue.
May have yellow "stains" around the white
center. When many corpora are present, some of
this type may be present but not apparent at the
surface. Diameter 2.0 to 8.5 mm, average 3.5 mm.
This is a catch-all category for all small compact
corpora with surface scars and internal structure.
It probably includes both regressed corpora lutea
and corpora representing ovulation and other
events. We found 1,999 corpora of this type
(57.1% ). The number increases steadily with total
corpora number (Table 3, Figure 28), while the
numbers of Types 1 and 2 corpora remain con-
stant, indicating that Types 1 and 2 corpora re-
gress into and accumulate as Type 3 corpora. This
is assuming, of course, that total corpora count is
related to age (see below).

Type 4. Thin, flattened against the surface of a
new corpus luteum. Two to 15 mm in diameter.
These are Types 2 and 3 corpora that cannot be
allocated to those categories because of distortion
caused by the corpus luteum. One hundred were
encountered (2.9%).

Type 5. Surface trace very slight or apparently
absent. Interior deep yellow or orange, with no
concentrated connective tissue or apparent inter-
nal structure. Diameter 0.5 to 5.5 mm, average 2
mm. Harrison et al. (1972) have suggested that
this type of corpus is the end result of regression
of an atretic lutealized follicle. We encountered
149 (4.3%).

Type 6. A small surface scar with no discernible
internal structure. Two to 5 mm in diameter.
Only 11 corpora of this type were encountered
(0.3%). They may represent extremely regressed
corpora of other types or may originate from dif-
ferent ovarian events.

Types 1, 2, and 3 comprise a series of increasing
regression and/or decreasing complexity of origi-

nal structure, and it is probable that regressing
corpora lutea pass through these types or stages.
The shapes of the diameter frequency distribu-
tions (Figure 29) suggest that corpora albicantia
regress to an average size of about 3 mm in
diameter and then persist and accumulate at that
size for at least part of the remainder of the life of
the female. The skewness of the aggregate dis-
tribution (sum Types 1, 2, and 3 in Figure 29)
becomes even more significant when one consid-
ers that the volume of the corpus decreases as the
cube of the diameter. On a volume scale, the left

100

TYPE I
(n=456)

100

0
440

T 400
0)

c 350

300 -

Q.
(T

O 250

200

150 -

100

J\^

TYPE 2
(n=787)

TYPE 3
(n = 1,999)

5 10

DIAMETER (mm)

15 17

Figure 29. — Frequency distribution of diameter of Types 1,2,3,
and 5 corpora albicantia in Stenella attenuata.

255

FISHERY BULLETIN: VOL. 74, NO. 2

side of the curve would be steeper and the right
side less steep.

Consideration of the relative rates of deposition
of corpora in the left and right ovaries is impor-
tant to the question of persistence of corpora. The
distribution of corpora between left and right
ovaries is related to the number of corpora pres-
ent (Table 4). The first corpus occurs in the left
ovary about 94% of the time. Subsequent corpora
occur in the same ovary as preceding ones at
about the same rate (—95% ), causing a gradually
increasing percentage of animals with corpora in
both ovaries, until about 10 to 11 corpora have
been deposited, when emphasis switches sharply
to the opposite ovary (left or right). All females
with 15 or more corpora (27 specimens) had cor-
pora in both ovaries.

A group of 15 seemingly postreproductive

Table 4. — Location of corpora (corpora lutea and corpora
albicantia) in ovaries of 488 specimens of Stenella attenuata.

Sample

Location of corpora

Corpora

size

Left ovary

Rigtit ovary

Botti ovaries

(no.)

only (%)

(%)

93.6

—

85.0

7.5

86.5

4.5

9.0

88.7

1.9

9,4

788

2.1

19.1

75.0

4.2

20.8

73.4

2.2

24 4

61.0

2.4

36.6

70.6

2.9

26.5

10-11

46.8

2.1

51.1

12-14

6.5

0.0

93.5

15-27

0.0

100.0

Total

488

females was encountered. These specimens had
very small, obviously regressed ovaries with 10
to 15 Type 3 or smaller corpora albicantia (Figure
30). They had no corpora lutea or Type 1 corpora

•

•
•

•

•
•

•

•
•

•

-~" 7

_ •

•

CT>

•

•
•

•

••

•

=C 6

•

•o

•

•
•

••

•
h

•
•

•

•
•

•
•
•

•
••

•
•

•

• •

(f)

•

•
_ •

•

- :

•

•
••

••

•
•

•

•
•
•
•
•
•

•o

•

•
•
•

•

•
•

•

•
•
t

••

•

•
•

•

•
•

••
•»

•

•
o

•

•
o

•

•
•

•

••

•

•
o

o •

•

2
1
n

•

—L-

o
o

_L_

•

o
o

1 I 1

1 1

1 1 1

1 1 1 1

0 I 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

CORPORA (number)

Figure 30. — Scatterplot of ovaries weight on number of corpora in Stenella attenuata. Females with corpus luteum not included.

Open dots are females with no Type 1 or 2 corpora albicantia.

256

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE
I7r-

^ 12
E
E

•- 9

< 8

I-

UJ 6

<05

• X

•X

o
o

XX X

J I I I

o oo
J L

•o

• cX ••

J I

0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

CORPORA (number)

Figure 31. — Scatterplot of diameter of largest follicle on number of corpora inStenella attenuata. Females with corpus luteum in-
dicated with X . Open dots are females with nor corpus luteum or Type 1 or 2 corpora albicantia.

257

FISHERY BULLETIN: VOL. 74, NO. 2

albicantia. They also typically had very small
Graafian follicles (Figure 31). A consideration of
these females bears on the question of persistence
of corpora albicantia. Sergeant (1962) encoun-
tered similar females in Globicephala. They com-
prised about 5% of adult females. He called them
"senile" and concluded that the ovarian scars in
these animals represent some residual subset of
the maximum complement of corpora of preg-
nancy, ovulation, and other events. He implied
that they probably are the corpora of pregnancy,
since those corpora are larger at the outset and
presumably less likely to regress to the point of
macroscopic disappearance. The ovarian data for
S. attenuata do not support this hypothesis of dis-
appearance of some corpora in regressed ovaries.
The regressed ovaries have 10 to 15 corpora (Fig-
ures 30 and 31). The ovaries of other, still repro-
ductive females are larger and have 16 to 27 cor-
pora, although follicles are typically smaller than
in reproductive females w^ith fewer corpora (Fig-
ure 31). Three alternative hypotheses explain
this apparent dichotomy in females with 10 or
more corpora:

1. The usual maximum number of corpora is
about 21, and some corpora disappear in
postreproductive females, i.e., the "senile"
group in Figure 30 properly belongs at the
far right side of the plot at the end of a
downward trend in ovary weight (the hy-
pothesis of Sergeant 1962).

2. Corpora are laid down at about the same
rate in all individuals, but some become
postreproductive at about 10 to 15 corpora
while others continue to accumulate corpora
(16 to 27) until a greater age, i.e., the corpo-
ra scale in Figure 30 is effectively an age
scale. Under this hjrpothesis, corpora do not
disappear.

3. Corpora are accumulated at rates varying
widely among individuals, but the typical
maximum complement is 10 to 15 corpora,
i.e., the reproductive females with more
than 15 corpora in Figure 30 properly belong
in the body of the distribution in the left
two-thirds of the plot. A possible explana-
tion for widely varying rates of accumula-
tion is that some females are more fecund
and the senile period is reached with some
maximum number of pregnancies, so that
the varying ratios of corpora of pregnancy to
corpora of ovulation may produce the appar-

ent dichotomy. Sergeant (1973) found
greatly varying individual rates of ovulation
in the white whale, Delphinapterus leucas.

In order to examine these alternative hypoth-
eses, the females in Figure 30 and 31 with 10 or
more corpora were examined in three groups —
A, B, and C:

A. 10 to 15 corpora, reproductively active

(corpus luteum and/or Types 1 and 2
corpora albicantia).

B. 16 or more corpora, reproductively active.

C. 10 to 15 corpora, postreproductive (ovaries

regressed, no corpus luteum or Types
1 or 2 corpora albicantia).

The three groups were compared in terms of
corpora count, weight of ovaries, size of largest
follicle, number of dentinal layers, total length,
and relative corpora counts in left and right
ovaries (Table 5). Only nonpregnant females were
included in the sample for ovary weight. Follicle
size was examined separately for pregnant and
nonpregnant animals.

Ovary weight and follicle size for nonpregnant
animals decline progressively from A to C. This is

Table 5. — Characteristics of females of Stenella attenuata in
groups A, B, and C (see text).

Item

Corpora (no.)

Sample size

Average

11.2

189

12.9

Range

10-15

16-27

10-15

1.41

256

—

Ovary weight (g)

(nonpregnant)

Sample size

Average

4.4

3.6

2.2

Range

2.0-8.5

2.2-48

1,0-3,1

1.61

0.81

0,59

Largest follicle (mm)

(nonpregnant)

Sample size

<14

Average

2.9

1.5

<0.5

Range

<0.5-10.3

<0. 5-8.0

0.5-4.3

2.53

2.35

—

Layers (no)

Sample size

Average

13.1

13.2

Range

10.0-16.0

11.0-15,0

11.5-16.0

1.39

1.31

1.52

Length (cm)

Sample size

Average

190,1

190 3

187.0

Range

172-202

1 77-204

179-192

6.43

6,78

3,54

Left/right ovary

Sample size

Average in right (%)

Lett/nght (no./no.)

548/178

291/144

—

258

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

a requirement of hypothesis 1, above, but does not
eliminate hypotheses 2 and 3.

The three groups do not differ in average esti-
mated number of tooth layers. This may, in part,
be due to the difficulty of accurately counting the
innermost layers in teeth with more than 12
layers (the number of layers is probably undere-
stimated by as much as one-third in teeth with
large amounts of convoluted secondary dentine),
but careful comparison of the teeth of the three
groups in terms of other features presumably cor-
related with age, such as tip wear, degree of clo-
sure of the pulp cavity, and amount of secondary
dentine does not indicate that any group is older
than any other. This evidence is against
hypothesis 1, which requires that group C be
older than A, and hypothesis 2, which requires
that B be older than A and C.

Groups A and B have reached asymptotic
length (-190 cm). The animals in group C aver-
aged about 3 cm less. A statistical comparison of
A with B using Student's t indicates that the dif-
ference is significant at a = 0.05. These results do
not eliminate or support directly any of the
hypotheses. Since A, B, and C are about the same
age, the length data indicate that asymptotic
length may be less for females that become senile
with 10 to 15 corpora. This indirectly supports the
idea of considerable individual variation in life
history.

The most convincing evidence against hy-
pothesis 1 has to do with number of corpora in
right versus left ovaries. If the emphasis in cor-
pus deposition shifts from left to right at about 10
corpora, and if group C regresses from group B
(animals with about 20 corpora) losing about 6
corpora in the process, then group C should have
about equal numbers of corpora in the right and
left ovaries. If most corpora of ovulation come in
early reproductive life (as data analyzed below
indicate) and, as suggested by Sergeant (1962),
are more likely to disappear than corpora of preg-
nancy because of smaller initial size, then the
regressed group C should have, on the average,
more corpora on the right than on the left, be-
cause most of the corpora of preg-
nancy would be in the right ovary. Forty-one per-
cent of the corpora in 14 individuals having 18 to
22 corpora (average 19) were on the right. Only
2^c of the corpora in group C were on the right.
The difference between C and A (29 and 24% ) can
be accounted for simply by the difference in aver-
age total corpora count (12.9 and 11.2). These re-

sults eliminate the hypothesis (number 1 above)
of loss of corpora with regression of ovaries.

The various lines of evidence largely speak
against hypotheses 1 and 2 and support hy-
pothesis 3, that of great individual variation in
life history and of persistence of corpora albican-
tia. This is in line with findings by some other
workers in small cetaceans (Sergeant 1962, 1973;
Brodiel971).

The data on the relationship of percent occur-
rence of corpora lutea to number of corpora (Fig-
ure 28) also support the hypothesis of widely
varying rate of accumulation of corpora albican-
tia. After stabilization at about 50% at 3 to 4 cor-
pora, the rate declines after 8 to 9 corpora to 20%
at 13 corpora; but the rate for females with 17 to
27 corpora is again 50%. Assuming that fecundity
is inversely related to age, this pattern suggests
that the females in the 17 to 27 group are about
the same age as those in the 3 to 9 group.

Ovulation Rate

Even assuming that corpora albicantia persist
and represent various ovarian events, estimating
average rates of accumulation is difficult because
of 1) the above-mentioned unreliability of age
estimates based on more than 12 tooth layers, 2)
the evident individual variation in accumulation
rate, and 3) change in ovulation rate during the
reproductive span. All of these factors must con-
tribute to the scatter in a plot of corpora number
(including corpus luteum) on estimated age (Fig-
ure 32). Several workers have pointed out that
cetacean ovaries often contain two or more cor-
pora of the same size and same stage of regres-
sion. It has been suggested that these are the
result of multiple infertile ovulations or luteali-
zation of atretic follicles in newly mature ani-
mals (Harrison et al. 1972). Many in the present
series of ovaries had two or more corpora (of Type
1 or 2) that were very similar in size and struc-
ture and must have resulted from nearly contem-
poraneous events. One probable multiple ovula-
tion is apparent in Figure 32. This female, field
number CW0R8, possessed 7 or 8 well-defined
layers in its teeth. In spite of its extreme youth, it
had a small corpus luteum, three Type 1 corpora,
two Type 2 corpora, one Type 3 corpus, and one
Type 4 corpus. The uterus was empty, and there
was no milk in the mammaries. The animal could
not have been reproductively active for more than
about a year, but had already experienced eight

259

3
C

FISHERY BULLETIN; VOL. 74, NO. 2

15-

cr
o

Q-

O 10

im ®
®» «) • •

• ®M ••• • •

5 -

••• • A •• •

• •••••

-! \ r-

I 12 13 14

LAYERS (number)

T
16

OCCLUDED

Figure 32. — Scatterplot of number of corpora on number of postnatal dentinal layers in Stenella attenuata. Circled s3rmbols are
senile females [shriveled oveiries with no corpus luteum or Type 1 or 2 corpora albicantia].

apparently nonreproductive ovarian events that
resulted in corpora belonging to all of the t5^es
through which a corpus luteum must pass during
regression to a small corpus albicans.

Calculation of average ovulation rates from the
data in Figure 32 must take into account indi-
vidual variation in age at first ovulation. The fe-
males in Figure 32 were grouped into 10 one-
layer intervals beginning with 7.5 layers (Table 6).
The average reproductive age in interval p was
calculated as

A = [ 2 a,&< J ^ c,

where a, = % maturing in iih. interval (% mature
in i minus % mature in i - 1)
b, = average reproductive age in interval

p of females mature in i
c = % mature in interval p.

Average reproductive age in the ith interval of

260

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

TABLE 6. — Average reproductive ages and corpora counts of
females oiStenella attenuata used in estimating ovulation rate
based on corpora and tooth layers.

Average

Proportion

reprodu

ctive

Average

Layers

Sample

mature

age of mature

corpora

(no.)

(°o)

(laye

rs)

(no.

)

7.5- 8

46.2

0.50 1

4.50

8.5- 9

44.4

1.56 \

1.43

3.25 ■

4.21

9.5-10

79.2

1.67 )

4.53 )

10.5-11

84.0

2.56

6.42

11.5-12

94.2

3.25

8.35

12.5-13

100.0

4.05

8.92

13.5-14

100.0

5.06

8.71

14.5-15

100.0

6.07 )

10.87 1

15.5-16

100.0

7.08 \

6.60

9.86 \

10.45

>16

100.0

8.09 )

9.75 )

Total

224

Note: Teeth of all available females with more than 12 corpora were sec-
tioned, while only a nonselective subsample of females with fewer corpora were
included. The effect on estimate of average reproductive age is negligible, since
nearly all had 11 or more layers.

females maturing in i was set at 0.50 layers. Be-
cause of small sample sizes, the first three inter-
vals and the last three were pooled. The results
show an increase in average corpora count
(number of ovulations) with reproductive age
(Figure 33). A curvilinear fit to the interval
means, using a power model forced through the
origin, fits well and indicates that ovulation rate
is higher in animals of reproductive age 0-2
layers than in older animals. The breaking point
seems to come at about 12 layers, when about 6
corpora have been accumulated and rate appears
to become nearly constant. Average ovulation
rates estimated from the curve are about four
during the first layer, two during the second, and
about one per layer thereafter.

0 12 3 4 5 6 7

AVERAGE REPRODUCTIVE AGE (layers)

Figure 33. — Relationship between average number of corpora
and average reproductive age (in layers) in Stenella attenuata.

Calving Interval

The pattern of reproduction definable with the
methods used here consists of three phases: preg-
nancy, lactation, and a period of inactivity and/or
estrus called here "resting/estrus." The length of
pregnancy was estimated above as 11.5 ± 0.2 mo.
We estimated length of lactation in three ways,
based on 1) stomach contents of calves, 2) num-
bers of lactating females and calves, and 3) ratio
between numbers of lactating and pregnant
females.

The forestomachs of 45 calves less than 150 cm
long were opened and examined by eye for pres-
ence of milk. Twenty-one were empty. The
stomachs of four calves 120 to 130 cm long con-
tained both milk and solid food (fish and/or
squid). Stomachs of 8 smaller calves (80 to 115 cm)
contained only milk, and 12 of the larger calves
(130 to 150 cm) apparently contained only solid
food. About 130 cm appears to be the length at
which effective weaning occurs. The estimated
time required to grow to 130 cm is 9.4 mo (based
on growth curve above). This estimate is not very
reliable for two reasons: the sample is small, and
small amounts of milk could be present and
undetectable by eye, i.e., suckling could continue
at a low level after the effective shift to solid food.
The estimate can, however, be considered to be a
probable lower bound on length of lactation.

A second estimate is based on the assumptions
that 1) a suckling calf exists for each lactating
female and 2) the samples of specimens are un-
biased with respect to suckling calves and lactat-
ing females. Given these assumptions, the length
at which the cumulative frequency of calves in a
sample equals the number of lactating females
should be the average length at weaning. This
length in eight variously sized, 1-mo "random"
samples of calves and females ranged from 125 to
145 cm (Table 7). The aggregate estimate for the
eight samples pooled (320 lactating females) is
137 cm. Average age at 137 cm is estimated at
1.94 tooth layers, or (assuming two layers ac-
cumulated during first year) 11.6 mo. If calves
were overrepresented in the samples, this would
be an underestimate. If they were underrepre-
sented, it would be an overestimate. It would be
an overestimate if the assumption that the
number of lactating females equals the number of
nursing calves were not valid. The assumption is
not valid if the mortality of nursing calves is

261

FISHERY BULLETIN: VOL. 74, NO, 2

Table 7. — Length at which cumulative frequency of calves
equals the number of lactating females in eight 1-mo samples of
Stenella attenuata.

Lactating

Length (cm) at which cumulative

Sample

females

frequency of calves = no

(mo)

(no.)

lactating females

Oct. 1972

132

Jan. 1973

125

Feb. 1973

144

Mar. 1973

136

Apr. 1973

142

May 1973

142

June 1973

145

Nov. 1973

142

(Oct. 28-Dec. 1 1 )

Total

320

137

higher than that of lactating females and lacta-
tion continues after death of a nursing calf.

A third estimate of length of lactation was de-
rived from the ratio of lactating to pregnant
females. This analysis included all the material
from 1971 and 1972, when only adult females
were sampled, as well as the material included in
the calf-lactating female analysis above. Females
both lactating and pregnant were included in
both categories. The assumption is made that
samples were unbiased with respect to relative
representativeness for lactating and pregnant
females. The ratio was 0.95 in the 1971 sample (86
adult females), 1.00 in 1972 (455), 0.96 in 1973
(573), and 0.97 for the pooled samples {n = 1,114;
Table 8). The ratio of lactating to pregnant should
equal the ratio of the lactation period to the ges-
tation period. Gestation is 11.5 mo, therefore lac-
tation is by this method estimated at 0.97 times
11.5 mo, or 11.2 mo. Estimated length at this age
is 135.5 cm.

The three estimates of 9.6, 11.6, and 11.2 mo are
based on largely independent assumptions and
are close enough to each other to indicate that
length of lactation is almost certainly between 9
and 12 mo. Of the three, the central estimate, 11.2

mo, is best in terms of sample size and probable
validity of assumptions and is used below in es-
timating length of the calving interval.

The basic data used for estimating average
length of calving interval were the relative fre-
quencies of adult females in several reproductive
conditions (Table 8). Adult females were defined
as those wath at least one corpus luteum or corpus
albicans. Senile females were those with 10 or
more corpora albicantia, no corpus luteum or
Type 1 or 2 corpora albicantia and ovaries weigh-
ing less than 3.5 g. Resting/estrus females were
those nonsenile adults that were neither preg-
nant nor lactating. Many of these (16 to 31%) had
a corpus luteum. The corpus luteum may have
represented an undetected very early pregnancy,
a very recently aborted pregnancy, loss of a calf
shortly after birth (resulting in cessation of lacta-
tion), or may have been a corpus luteum of ovula-
tion. All of these alternatives may be represented
in the samples.

In calculating the proportions of females in the
three phases of pregnant, lactating, and resting
(Table 9), senile females were excluded. One-half
of the animals simultaneously pregnant and lac-
tating were assigned to the pregnant category
and one-half to the lactating category.

The average length of calving interval was es-
timated by two methods — 1) using the estimates
of gestation and lactation periods and 2) using the
percentage of females pregnant. The data for the
3 yr are comparable (Table 9), with the exception
of possible existence of a trend in proportion rest-
ing; therefore, length of calving interval was es-
timated from the pooled data. Eighty-four and
one-half percent of reproductive females were
pregnant or lactating. Pregnancy (11.5 mo) plus
lactation (11.2 mo) total 22.7 mo. If the proportion
in a phase is equal to the proportion of the total

Table 8. — Reproductive condition of 1,114 adult female specimens o{ Stenella

attenuata, collected 1973.'

1971

1972

1973

Total

No,

No.

Pregnant only (P)

36.0

180

39.7

233

40.7

444

39.6

Lactating only (L)

33.7

180

39.7

223

38.9

432

38.8

Pregnant and

lactating (PL)

15.1

3.5

3.0

4.1

Resting/estrus ( „ )
(R) ^^'

1) '-

«(

54 ;

14.1

-fi)

16.4

'«» (J)

15.2

Senile^

2.3

3.3

1,0

2.1

Total

100

455

100

573

100

1,114

100

'In the resting/estrus category, subcategones A and B (in parentheses) are specimens with and
without a corpus luteum, respectively.
^3^10 corpora, no Type 1 or 2 corpora, and ovaries s 3.5 g.

262

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

Table 9. — Proportions of 1,091 adult reproductive females of Stenella attenuata in
pregnant, lactating, and resting/estrus phases.

1971

1972

1973

Total

No.

Pregnant (P + V2PL

in Table 8)
Lactating (L+ V2PL)
"Resting" (R)

Total reproductive

females

37.5
35.5
11

44.6
42.3
13.1

100

188

440

42.8
42.8
14.4

100

241.5

231.5

567

42.6
40.8
16.6

100

467
455
169

1,091

42.8
41.7
15.5

100

calving interval spent in that phase, then total
length of the interval cycle is 22.7 mo divided by
0.845, or 26.9 mo.

A second estimate was obtained directly from
the proportion of females pregnant. In calculating
this proportion, all pregnant animals were in-
cluded (P + PL in Table 8): 490 of 1,091 reproduc-
tive females were pregnant, or 44.9% . Division by
length of pregnancy, 0.958 yr (11.5 mo), yields an
estimate of annual pregnancy rate, 0.469. The re-
ciprocal of pregnancy rate, 2.133 yr, or 25.6 mo, is
an estimate of average length of calving interval.

Both estimates of length of calving interval,
26.9 and 25.6 mo, are overestimates to the extent
that the "resting" females with corpora lutea rep-
resented uncounted pregnancies, but the effect
can be at most very minor For example, if all
these females represented undetected pregnan-
cies or pregnancies aborted during capture, the
unlikely extreme case, the estimates would be
25.7 and 24.7 mo respectively, an average differ-
ence of about 1 mo. Since the "resting" females
with corpora lutea probably represent a mixture
of causes and conditions, including nonfertile
ovulations, the probable effect on the estimates is
less than 1 mo. Considering this factor and the
closeness of the two estimates to each other, it
seems certain that the true length of the interval
is between 24 and 27 mo. The lower of the two
estimates, which is based on fewer assumptions
and calculations, was rounded off to 26 mo and is
used below in further analysis of life history. The
average pattern of events then, consists of 11.5 mo
of pregnancy, 11.2 mo of lactation, and 3.3 mo of
resting and/or estrus.

Overlapping Lactation and Pregnancy

About 9.6% of lactating females were also preg-
nant (Table 8). Most had fetuses less than 35 to
40 cm long (Figure 26), about halfway through
the gestation period. This suggests that overlap
when it occurs is usually about 5 to 6 mo long, i.e.,
conception occurs about halfway through the lac-

tation period of about 11 mo, making the calving
interval about 20 mo long instead of 26. The very
few lactating females with near-term fetuses may
have conceived during postpartum estrus or may
have begun to lactate shortly before parturition.
The data on Graafian follicles are consistent
with the theory that postpartum estrus occurs
during lactation (Figure 34). The largest follicle
in the ovaries of resting/estrus females (including
those presumably about to ovulate) is on the av-
erage 3 to 4 mm in diameter. After ovulation and
conception, the remaining large follicles regress
rapidly to about 2 mm (or become lutealized or
atretic). There is a further net decline during ges-
tation to about 1.5 mm, and during lactation the
main modal diameter is about 1.0 mm. During
both pregnancy and lactation, however, about
10% of the females (excluding senile individuals,
as defined above) have follicles that are within
the size range (^3.0 mm) of the presumably ripe
follicles present during the resting/estrus phase.
This is most clear-cut during lactation. Most of
the larger follicles during pregnancy occur in
females having fetuses 400 to 500 mm long, or
about halfway through the gestation period (Fig-
ure 34).

Decrease in Reproductive Rate with Age

Reproductive rate decreases with age. Age-
specific estimates of pregnancy rates and lactation
rate were calculated from a random sample of the
data for specimens for which teeth were sectioned
(stratified to insure representation of corpora-
number strata in about the proportions as in the
entire sample). The analysis shows decline of preg-
nancy rate from about 0.6 at 8 to 9 layers to about
0.3 at 16 to 17 layers (Figure 35). The weighted
rate for the pooled sample of 138 used in the calcu-
lation was 0.51, comparable to the rate of 0.47
obtained for 1,091 animals (above). The specimens
for which teeth were sectioned were about one-
third from 1971 and two-thirds from 1972, with a
few specimens from earlier years. Lactation rate

263

£
E

_l
O

_l
_l

o
li.

o
cr
<

o
cc

LlI

ll-

5 3

2 -

I -

PREGNANT

LACTATING
NON-PREGNANT

r-N-F

FISHERY BULLETIN; VOL. 74, NO. 2

-ill

RESTING/
ESTRUS

11%

♦aar ave

- 7

- 6

- 5

-AVE.

? 62 %

0 100 200 300 400 500 600 700 800 900

LENGTH OF FETUS (mm)
Figure 34. — Diameter of largest follicle in pregnant, lactating, and resting females of Stenella attenuata.

(Figure 35) increases from about 0.1 at 8 layers to
about 0.6 at 12 layers and then again decreases to
about 0.5.

The initial very low lactation rate compared to
pregnancy rate, of course, reflects the fact that a
very high percentage of the young females are
pregnant for the first time and thus cannot be
lactating. The lactation rate climbs rapidly to a
level about equal to the pregnancy rate (at about
12 layers) and behaves like the pregnancy rate
thereafter. The apparent decline of reproductive
rates in older females may be related to the
physiological or social mechanisms that cause the
appearance of postreproductive females in this age
group (see above; not included here).

Sex Ratios

The overall sex ratio was 44.9% males and
55.1% females (Figure 36). Many large samples
examined were predominantly female. Fourteen
of 32 single-school samples of 50 or more speci-
mens were more than 60% female, whereas none
was more than 60% male. The largest single-
school sample examined (342) was almost half and
half males and females.

Sex ratio changes with age (Table 10). This is, of
course, making the assumption that the samples
examined were representative of the population.
Neonates and two-tone animals were almost
equally divided between the sexes, but only about

264

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

PREGNflNCr RATE

LACTATION RATE

II 12 13

LAYERS (number)

Figure 35. — Age-related changes in pregnancy (solid line) and
lactation rates (dashed line) in Stenella attenuata, based on
tooth layer data. Postreproductive females not included.

PARITY

OVERALL RATIO

I I I I ■ I I I I ' I I I I I ' ' '

' I ' I ' I I

100 150 200 250

SAMPLE SIZE (no)

300

350

Figure 36. — Scatterplot of sex ratio (percent males) on sam-
ple size in single-school samples of five or more specimens of
Stenella attenuata. Overall ratio (dashed line) from Table 12.

Table lO. — Sex ratio, by color pattern stage, in 9,371 speci-
mens ot Stenella attenuata, 1971-73.

Color
pattern

Males

Females

stage

No.

Total

Neonate

Two-tone

Speckled

Mottled

Fused

Total

205
666
609
569
2,154

4,203

49.8
487
47.8
43.8
429

44.9

207
701
666
729
2,865

5,168

50.2
51.3
52.2
56.2
57.1

55.1

412
1,367
1,275
1,298
5,019

9,371

43% of the adults examined were males. The
greatest change in ratio, from 48.0 to 43.5% male,
comes about during the transition to mottled col-
oration between 7 and 8 layers of age. Assuming
random sampling of the population, male and

female mortality rates must diverge sharply at
this point.

Gross Annual Production

An estimate of average gross annual production
of calves for 1971 to 1973 was calculated based on
the estimate of annual pregnancy rate, the color
pattern phase data, and the proportions of mottled
and fused females found to be sexually mature
(Table 11).

Seven hundred and twenty-nine of 9,371 ani-
mals were mottled females (7.8%) and 2,865 were
fused females (30.6%). Of 127 mottled and 1,141
fused females, 47.4 and 88.4% were sexually
mature, respectively (Table 11). Average preg-
nancy rate was 0.469. Production = [(0.078 x
0.474) + (0.306 X 0.884)] 0.469 = 0.144 of the popu-
lation per year.

Table ll. — Sexual maturity (presence of ovarian corpora) in
mottled and fused females of Stenella attenuata, 1971-73.

Mottled

Fused^

Mat

J re

Meture
No. ' %

Year

No.

1971
1972
1973

Total

6
92

170

268

5
37
85

127

(- )
(40.2)
(50.0)

(47,4)

99
473
569

1,141

82 (82.8)
417 (88.2)
510 (89.6)

1,009 (88.4)

Schooling in Relation to Reproduction

Kasuya (1972) reported changes in structure
and size of schools of .S. coeruleoalba correlated
with breeding condition and breeding activities.
Kasuya et al. (1974) proposed a complex hypothet-
ical system of school formation and breakdown
determined by reproductive activities in the
Japanese population of S. attenuata. They
suggested that juveniles of S. attenuata in Japa-
nese waters leave breeding schools and school sep-
arately, rejoining the breeding schools at puberty.
There is nothing to indicate that this happens
in the eastern Pacific. We examined the coloration
structure (= age structure) of single-school sam-
ples. Of 324 single-school samples of seven or more
animals, only 1 (of 17 animals) contained no adults
(or neonatal calves, which would indicate presence
of adult lactating females in the school). This sam-
ple (8 two-tone, 2 speckled, and 7 mottled) was
from a school of about 600 spotted porpoise, S.
attenuata, congregated with about 600 spinner
porpoise, iS. longirostris. Given that about half the
animals examined were adults, the probability of

265

FISHERY BULLETIN: VOL. 74, NO. 2

a single-school sample of seven containing no
"fused" individuals is aboutO.Ol (= 0.5''). If schools
consisting only of juveniles were common, many
more all-juvenile samples would have been en-
countered. Conversely, juveniles (two-tone, speck-
led, and/or mottled) occurred in all but 3 of the 324
samples. It must be concluded that juveniles prob-
ably do not school separately in the eastern
Pacific. Another possibility, albeit unlikely, is that
all-juvenile schools exist but are not captured by
tuna seiners.

COMPARISON WITH
THE JAPANESE POPULATION

Many of the estimates of life history parameters
presented here differ from those published by
Kasuya et al. (1974) for the relatively unexploited
population of S. attenuata in Japanese waters (Ta-
ble 12). The differences could be caused by 1 ) differ-
ential procedures or analytical methods, 2) in-
trinsic racial differences between the populations,
or 3) differential population status, e.g., exploited
versus unexploited. The comparisons below of
similarly calculated average estimates, of course,
rest on the assumption that the overall samples in
both cases were not biased with respect to age or
sex. The major sampling differences between the
two studies is that the Japanese sample consisted
mostly of large samples from a few schools.

whereas our sample consisted mainly of aggre-
gated, small samples from many schools. Both
studies assume no sampling bias. Comparison of
large, single-month samples in the present study
with large, single-school samples in the Japanese
study (e.g., the October 1972 sample in Figures 15
and 16 with sample number 2 in Figure 2 of
Kasuya et al. 1974) indicate very similar length-
frequency distributions and support the idea that
the aggregated samples are probably not biased,
or, if biased, are biased in the same way. This
inference is, of course, based on the assumption
that the underlying population structures are
about the same in the two populations.

The estimate of Kasuya et al. ( 1974) of length at
birth was based on only 5 full-term fetuses and
newborn calves versus 86 in the present study.
Our estimate can, therefore, be considered more
reliable, although the possibility does exist that
length at birth is greater in the Japanese popula-
tion. The difference between the estimated
lengths at 1 yr for the two populations is about the
same as the difference between the estimates of
length at birth. Estimated length at attainment of
sexual maturity and maximum length (for males)
are also greater for the Japanese samples. The
estimate of length at maturity of males is greater
in spite of the fact that Kasuya et al. used a lower
testis-weight criterion than we did (68 versus 100
g). The average lengths of both adult males and

Table 12. — Comparison of estimates of average life history parameters of Stenella attenuata by Kasuya et al. (1974) and

in present paper.

Parameter (average)

Kasuya et al.

Perrin et al.

1 . Length at birth

89 cm

82.5 cm

2. Growth rate in 1st year

4.5 cm per mo

4.6 cm per mo

3. Length at 1 yr

143 cm

138 cm

4. Length at onset of sexual maturity:

Males

197 cm

-195 cm

Females

187 cm

181 ±1 cm

5. Age at onset of sexual maturity:

Males

10.3 layers (10 3 yr)

12 layers (6-11 yr)

Females

8.2 layers (8.2 yr)

9 layers (4.5-8 yr)

6. Average length of sexually mature adults:

Males

204-207 cm

200.7 cm

Females

192-195 cm

187.3 cm

7. Maximum length:

Males

234 cm

226 cm

Females

220 cm

8. Maximum number of consistently readable tooth layers

-13

12-13

9. Average ovulation rate (based on layers)

0.8 per layer

-1 per layer

(0.8 per yr)

(1 or 2 per yr) in fully
mature, more in younger

10. Pregnancy rate (overall)

0.27 per yr

0.47 per yr

11. Breeding seasons

3 per yr

multiple

12. Gestation

11.2 mo

11.5 ±0.2 mo

13. Lactation

29.3 mo

11.2 mo

14. Resting

9.8 mo

3.3 mo

15. Length of calving interval

4,19 yr

2.17 yr

16. Sex ratio:

Overall

0.76 male:1 female

0.81 male:1 female

At birth

1.3-1.5:1

1.00:1

Adults

0.58:1

0.75:1

266

PERRIN ET AL : GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

females are also greater in the Japanese popula-
tion, and in this case, all four of the estimates are
based on large and certainly adequate samples.
These differences all suggest that the Japanese
form is about 6 to 8 cm larger than the eastern
Pacific form.

The estimates of Kasuya et al. (1974) of age at
attainment of sexual maturity are based on their
conclusion that one tooth layer corresponds to 1 yr
of growth. It appears from comparisons of their
first-year growth curve with ours (note rate in first
year and length at 1 yr) that our first two layers
correspond to their first layer. Kasuya (1972) in
his paper on growth of S. coeruleoalba mentioned
observing "one or two faint translucent layers in
the thick opaque layer accumulated just after the
birth" that were "not used for age determination
because it was not expected to show the annual
accumulation cycle," and Kasuya et al. (1974)
stated that the "dentinal growth layers of this
species [S. attenuata] does not differ so much from
that of S. coeruleoalba.'' After the first year, our
hypothesis 2 corresponds to the assumption of
Kasuya et al . of one layer per year, e.g., nine layers
of Perrin et al. (1973) = eight layers of Kasuya et
al. = 8 yr.

The average length of calving interval in both
studies was estimated by several methods that
converged on the respective central estimates.
One minor difference between the two analyses is
that Kasuya et al. (1974) did not exclude postre-
productive females from the "resting/estrus" group.
Thus, their estimate of the average resting/estrus
period of 9.8 mo may be a slight overestimate. The
probable effect of this on the estimate of length of
total calving interval is very small, however, and
it therefore seems that the estimates are analyti-
cally comparable and that the difference between
them is real. Kasuya et al. estimated that indi-
vidual intervals in the Japanese population vary
from 23 to 60 mo, with modes at 28 to 30, 36 to
38, and 54 to 56 mo. The potential thus probably
exists for a shift in average length from 50 mo (4.17
yr) to 26 mo (2.17 yr) under exploitation.

Kasuya et al. (1974) used the same methods
used here to estimate length of the lactation period
and arrived at a "best" estimate of 29.3 mo, some
18 mo longer than our estimate of 11.2 mo. They
found that the major shift from milk to solid food
occurs at body length of about 133 cm, about the
same as in our sample, but that some suckling and
lactation of the mother continues for an average
additional 20 mo. The prolonged suckling is prob-

ably nutritionally a largely nonfunctional aspect
of general prolonged parental care. It has been
suggested on the basis of comparison of the life
histories and behavior of mysticetes and odonto-
cetes that this period in odontocetes may allow for
"sophisticated" communicational-navigational
training (Brodie 1969). Thus the apparent shorter
lactation period in the eastern Pacific, and the
concomitant shorter calving interval and higher
pregnancy rate, does not necessarily mean earlier
effective weaning, but may reflect a truncated pa-
rental care period.

The apparent overall sex ratios are almost the
same for the two populations, but the proportion of
males was higher at birth and lower at maturity in
the Japanese samples than in the eastern Pacific
samples. A lower proportion of males at birth
could be a response to exploitation. Kasuya et al.
(1974) suggested that the very low proportions of
males in mature age-classes in the Japanese
catches could be partially caused by segregation of
adult males or by differential catchability but are
largely due to differential mortality rates. If the
decrease in proportion of males with age is caused
by differential mortality, the apparent faster de-
crease in the Japanese population must mean that
the disparity in mortality rates between the sexes
is greater there than in the eastern Pacific.

In summary, the two sets of estimates differ in a
consistent way, and the differences are real. It
seems possible that the differences in some way
reflect exploitation in the eastern Pacific.

ACKNOWLEDGMENTS

This study would not have been possible without
the generous cooperation and assistance of the
owners, masters, and crews of the tuna seiners
Conte Bianco, Carol Virginia {now Carol S), Larry
Roe, Nautilus, Mary Antoinette, San Juan, Con-
cho, Kerri M, Queen Mary, Eastern Pacific, John F.
Kennedy, Sea Preme, Westport, Anne M, Pacific
Queen, J. M. Martinac, Lois Seaver, Marietta,
Independence, Sea Quest, Bold Contender, Jac-
queline A, Frances Ann, Elsie A, Sea Royal, Jac-
queline Marie, Trinidad, Mermaid, Bettie M, An-
tonina C, Day Island, Connie Jean, and Denise
Marie. Scientists and technicians who collected
data and specimens aboard the vessels include C.
E. Bowlby, R. W. Cunningham, W. E. Evans, R. S.
Garvie, J. M. Greene, D. B. Holts, J. La Grange, J.
S. Leatherwood, R. E. Loghry, R. L. McNeely, C. W.
Oliver, R. J. Olson, C. J. Orange, D. J. Otis, J. W.

267

FISHERY BULLETIN: VOL. 74, NO. 2

Ploeger, A. Poshkus, F. M. Ralston, S. B. Reiley, J.
M. Rosen, C. R. Ryan, K. D. Sexton, G. M. Treinen,
J. A. Young, and D. B. Zantiny. R. L. Brownell,
Jr., G. D. Fitzgerald, D. W. Rice, W. A. Walker, and
D. W. Waller contributed unpublished data. D. B.
Holts sectioned the teeth, and D. W. Rice assisted
with the readings. R. B. Miller processed and
examined the ovaries. T. D. Smith and N. K. Wiley
provided advice and assistance in data processing
and analysis. I. Barrett, P. J. H. van Bree, P. F.
Brodie, R. L. Brownell, Jr., W. Clark, W. E. Evans,

C. L. Hubbs, T. Kasuya, W. H. Lenarz, J. G. Mead,

D. W. Rice, D. E. Sergeant, T. D. Smith, and G.
Stauffer read the manuscript. We thank these per-
sons and others not mentioned for their invalu-
able assistance.

LITERATURE CITED

Amoroso, E. C, and C. a. Finn

1962. Ovarian activity during gestation, ovum transport
and implantation. In S. Zuckerman (editor), The ovary.
Vol. 1, p. 451-537. Academic Press, N.Y.
Best, p. B.

1968. The sperm whale (Physeter catodon) off the west
coast of South Africa. 2. Reproduction in the female. S.
Afr. Div. Sea Fish., Invest. Rep. 66, 32 p.

1969. The sperm whale (Physeter catodon) off the west
coast of South Africa. 3. Reproduction in the male. S.
Afr. Div. Sea Fish., Invest. Rep. 72, 20 p.

1970. The sperm whale (Physeter catodon) off the west
coast of South Africa. 5. Age, growth and mortality. S.
Afr. Div. Sea Fish., Invest. Rep. 79, 27 p.

Brodie, p. F.

1969. Duration of lactation in Cetacea: An indicator of
required learning? Am. Midi. Nat. 82:312-314.

1971. A reconsideration of aspects of growth, reproduc-
tion, and behavior of the white whale (Delphinapterus
leucas), with reference to the Cumberland Sound, Baffin
Island, population. J. Fish. Res. Board Can.
28:1309-1318.

1972. Significance of accessory corpora lutea in odonto-
cetes with reference to Delphinapterus leucas. J.
Mammal. 53:614-616.

Brunn, a.

1935. Flying-fishes (Exocoetidae) of the Atlantic. Sys-
tematic and biological studies. Dana Rep. 2(6), 106 p.

Clarke. M. R.

1966. A review of the systematics and ecology of oceanic
squids. Adv. Mar. Biol. 4:91-300.
BAILEY, M. D., AND W. F. PERRIN.

1973. Helminth parasites of porpoises of the genus
Stenella in the eastern tropical Pacific, with descrip-
tions of two new species: Mastigonema stenellae gen.
et sp. n. (Nematoda: Spiruroidea) and Zalophotrema
pacifkum sp. n. (Trematoda: Digenea). Fish. Bull., U.S.
71:455-471.

Fisher, H. d., and R. J. Harrison.

1970. Reproduction in the common porpose (Phocoena

phocoena) of the North Atlantic. J. Zool. (Lond.) 161:471-
486.

Harrison, r. J., R. L. Brownell, Jr., and r. C. Boice.

1972. Reproduction and gonadal appearances in some
odontocetes. In R. J. Harrison (editor). Functional anat-
omy of marine mammals. Vol. 1, p. 361-429. Academic
Press, N.Y.
Hasselblad, v.

1966. Estimation of parameters for a mixture of normal
distributions. Technometrics 8:431-444.
HUGGETT, A. St. G., and W. F. WIDDAS.

1951. The relationship between mammalian foetal
weight and conception age. J. Physiol. 114:306-317.

Kasuya. T.

1972. Growth and reproduction oi Stennella caeruleoalba
based on the age determination by means of dentinal
growth layers. Sci. Rep. Whales Res. Inst. (Tokyo)
24:57-79.
Kasuya,T., N. Miyazaki, and W. H. DAWBIN.

1974. Growth and reproduction of Stenella attenuata
in the Pacific coast of Japan. Sci. Rep. Whales Res.
Inst. (Tokyo) 26:157-226.

KLEVEZAL', G. A., AND S. E. KLEINENBERG.

1969. Age determination of mammals from annual layers
in teeth and bones. Izdatel'stvo "Nauka," Mosc, 1967.
(Translated by Israel Program Sci. Transl., Jerus., 128 p.)
KNAUSS, J. A.

1963. Equatorial current systems. In M. N. Hill (editor),
The sea. Vol. 2, p. 235-252. Interscience Publ., N.Y.
LAIRD, A. K.

1969. The dynamics of growth. Res./Dev. 20(8):28-31.
Laws, R. M.

1959. The foetal growth rates of whales with special
reference to the fin whale, Balaenoptera physalus Linn.
Discovery Rep. 29:281-307.
M0HL-HANSEN, U.

1954. Investigations on reproduction and growth of the
porpoise (Phocaena phocaena (L.)) from the Baltic.
Vidensk. Medd. Dan. Naturhist. Foren. Kbh.
116:369-396.
MOSER, H. G., AND E. H. AHLSTROM.

1970. Development of lanternfishes (family Myctophidae)
in the California Current. Part I. Species with narrow-
eyed larvae. Los Ang. Cty. Mus. Nat. Hist. Sci. 7, 145 p.

NISHIWAKI, M., M. NAKAJIMA, AND T KAMIYA.

1965. A rare species of dolphin (Stenella attenuata) ft-om
Arari, Japan. Sci. Rep. Whales Res. Inst. (Tokyo)
19:53-64.
PERRIN, W. F.

1969. Using porpoise to catch tuna. World Fishing 18(6):
42-45.

1970a. The problem of porpoise mortality in the U.S. tropi-
cal tuna fishery. Proc. 6th Annu. Conf. Biol. Sonar Diving
Mammals, Stanford Res. Inst., Menlo Park, Calif., p.
45-48.

1970b. Color pattern of the eastern Pacific spotted por-
poise Stenella graffmani Ldnnberg (Cetacea, Delphin-
idae). Zoologica (N.Y.) 54:135-149.

1975. Distribution and differentiation of populations of
dolphins of the genus Stenella in the eastern tropical
Pacific. J. Fish. Res. Board Can. 32:1059-1067.

In press. Variation of spotted and spinner porpoises
(genus Stenella) of the eastern tropical Pacific and
Hawaii. Bull. Scripps Inst. Oceanogr., Univ. Calif.

268

PERRIN ET AL.: GROWTH AND REPRODUCTION OF SPOTTED PORPOISE

PERRIN, W. F., AND E. L. ROBERTS.

1972. Organ weights of non-captive porpoise (Stenella
spp.). Bull. South. Calif. Acad. Sci. 71:19-32.

PERRIN, W. F., R. R. WARNER, C. H. FiSCUS, AND D. B. HOLTS.

1973. Stomach contents of porpoise, Stenella spp., and
yellowfin tuna, Thunnus albacares, in mixed-species
aggregations. Fish. Bull., U.S. 71:1077-1092.

RICHARDS, W. J., AND W. L. KLAWE.

1972. Indexed bibliography of the eggs and young of
tunas and other scombrids (Pisces, Scombridae), 1880-
1970. U.S. Dep. Commer., NOAA Tech. Rep., NMFS
SSRF-652, 107 p.
SECKEL, G. R.

1972. Hawaiian-caught skipjack tuna and their physical
environment. Fish. Bull., U.S. 70:763-787.
SERGEANT. D. E.

1962. The biology of the pilot or pothead whale Globi-
cephala melaena (Traill) in Newfoundland waters. Fish.
Res. Board Can. Bull. 132, 84 p.

1973. Biology of white whales (Delphinapterus leucas)
in western Hudson Bay. J. Fish. Res. Board Can.
30:1065-1090.
SERGEANT, D. E., D. K. CALDWELL, AND M. C. CALDWELL.
1973. Age, growth, and maturity of bottlenosed dolphin
(Tursiops truncatus) from northeast Florida. J. Fish.
Res. Board Can. 30:1009-1011.

TRUE, F. W.

1906. Notes on a porpoise of the genus Prodelphinus from
the Hawaiian Islands. Bull. U.S. Fish Comm.
23:811-815.

Wyrtki, K.

1964. The thermal structure of the eastern Pacific Ocean.
Erganzungsh. Dtsch. Hydrogr. Z. Reihe A 8(6), 84 p.

YOUNG, R. E.

1972. The systematics and areal distribution of pelagic
cephalopods from the seas off southern California.
Smithson. Contrib. Zool. 97, 159 p.

269

GROWTH OF LABORATORY-REARED NORTHERN ANCHOVY,
ENGRAULIS MORDAX, FROM SOUTHERN CALIFORNIA

Gary T. Sakagawa and Makoto Kimura^

ABSTRACT

The northern anchovy, Engraulis mordax, was experimentally reared in the laboratory at the South-
west Fisheries Center, La Jolla, Calif. Data from three experiments were used to empirically fit a
two-phase Gompertz growth model. The model describes growth from hatching to about 20 mo of age.
It was estimated that the average length of laboratory-reared anchovies is 102 mm at 1 yr old and 119
mm at 2 jT old. Growth of laboratory-reared anchovies was comparable to that of anchovies in
the wild.

Attempts to rear the northern anchovy, En-
graulis mordax, at the Southwest Fisheries
Center (SWFC), La Jolla, Calif., were begun in
1966 when G. O. Schumann collected anchovy
larvae in the ocean off La Jolla and successfully
reared them using wild plankton as food in the
laboratory (Bardach 1968). Schumann's success
was followed by other laboratory experiments in
which anchovies were reared from eggs, larvae,
and juveniles that were caught in the ocean (Ta-
ble 1). In 1970, Leong (1971) developed a method
for artificially inducing anchovies to spawn by
controlling the photoperiod and injecting hor-
mones. This technique is currently used at the
SWFC to produce eggs and to rear anchovies for
experimental purposes.

One of the purposes of the rearing experiments
at the SWFC has been to obtain physiological and
biochemical information needed for describing
the energy budget of the northern anchovy, and to
relate the results to the feeding dynamics of the
anchovy population in the California Current,
which consists of primarily young fish less than 3
yr old. Growth data are needed for analysis of the
budget, and various attempts have been made to
measure growth in the laboratory. Kramer and
Zweifel (1970) and Lasker et al. (1970) reported
growth rates of anchovy larvae. In this report we
extend their analyses to include growth from
hatching to about 20 mo old. We also present a
mathematical model that describes this growth
and compare our results with those of other
investigators.

^Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

SOURCES OF DATA

Data primarily from experiments of G. O.
Schumann (Schumann-I; Schumann-ID, G. O.
Schumann and A. Saraspe (Schumann-Ill), and
R. Leong (pers. commun., SWFC) were used in
our study (Table 1).

Schumann-II successfully reared larval an-
chovies for 22 days at about 22°C water tempera-
ture, which is higher than the temperature (15° to
16°C) at which anchovy larvae are frequently
found in large numbers in the California Current
(pers. commun., P. Smith, SWFC). The larvae
were fed wild plankton and samples were taken
for length measurement approximately daily.

Schumann-Ill reared anchovies from the egg
stage through the juvenile stage in aquaria for 83
days on a diet of wild plankton, Artemia salina,
and commercial trout food. The experiment was
conducted from March to June and the water
temperatures in the aquaria were not recorded.
However, during March to June the average
water temperature in rearing aquaria at the
SWFC is generally about 18° to 22°C.

Leong (pers. commun.) obtained juvenile an-
chovies from a live-bait dealer and reared the fish
to maturity in a 4.6-m diameter pool (13.2 kl)
with circulating seawater. The water tempera-
ture in the pool was a few degrees higher than the
prevailing water temperature oflFScripps Pier, La
Jolla, site of the water intake for the experimen-
tal pool (Lasker and Vlymen 1969). Leong fed the
fish a diet o( Artemia salina, ground squid and
anchovies, and commercial trout food. Once a
month about 25 fish were sacrificed for length and
weight measurements.

Manuscript accepted October 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

271

FISHERY BULLETIN: VOL. 74, NO. 2

Table l. — Laboratory experiments of rearing the northern anchovy at the Southwest Fisheries Center, La JoUa, Calif.

Source

Hunter (1976)

Kramer and Zweifel (1970)
Lasker et al. (1970)

Leong (unpubl. data)'

Paloma (see text footnote 3)

Schumann-I (G. O. Schumann unpubl. data)^

Sctiumann-ll (Kramer and Zweifel 1970)

Scfiumann-lll (G. O. Schumann and A. Saraspe unpubl. data)^

Theilacker and McMaster (1971)

Rearing

Average length
(mm^

Life stage
at start

start of

duration

rearing

(days)

Stan

Finish

Food

Eggs

April

4.0

35.0

Gymnodinium splendens,
Brachionus plicatilis,
Tisbe furcata, and
Anemia saline

Eggs

August and
September

3.2

17.4

Wild plankton and
A salina

Eggs

February

3.4

21.0

Bulla gouldiana.
G. splendens,
and A. salina

Juveniles

Apnl

474

88.3

117.7

Squid, anchovy,
A salina,
and trout food

Juveniles

November

624

75.0

106.2

Anemia salina and
trout food

Larvae

Inarch

18.0

81.9

Wild plankton

Eggs

March

2.9

16.2

Wild plankton

Eggs

March

3.5

67.1

Wild plankton, A.
salina, and trout
food

Eggs

12.0

Gymnodinium splendens,
B. plicatilis,
and A. salina

'Pers. commun.. Southwest Fisheries Center, La Jolla, Calif.
^Data are on file at the Southwest Fisheries Center, La Jolla, Calif.

In all of these experiments the fish were from
the southern California stock (Vrooman and
Smith 1971), reared at laboratory ambient water
temperature, and not subjected to experimental
treatment or excessive handling. All fish sampled
for measurements were sacrificed. The length
measurement is standard length.

TREATMENT OF DATA

The age of anchovies reared by Schumann-II
and Schumann-Ill were known because the an-
chovies were hatched from eggs at the start of the
rearing experiments. In Leong's {pers. commun.)
experiment, the exact age of his fish was not
known because juvenile fish of average length of
88.3 mm were used at the start of the experiment.
We estimated the age of Leong's fish from data
from Schumann-I in which anchovies were reared
for 97 days from an average length of 18.0 to 81.9
mm (Table 1), and data from Schumann-Ill which
indicated that an 18.0 mm fish, raised from eggs,
was about 30 days old. Our age estimate is 4 mo.

Several mathematical models describing
growth of organisms are available (e.g., Parker
and Larkin 1959; Richards 1959; Laird 1969).
The commonly used models in fisheries are the
exponential, the von Bertalanffy, and the Gom-
pertz models (Beverton and Holt 1957; Silliman
1969). The Gompertz model was selected for our
study because it was shown by Kramer and
Zweifel (1970) to be better than the exponential

model for describing growth of laboratory-reared
anchovy larvae and because it generally de-
scribes the growth of fishes fairly well. Also, pre-
liminary analysis of our data indicated that the
von Bertalanffy model poorly described the
growth of young fish.

The Laird version of the Gompertz growth
model (Laird 1969) describes an asymmetric sig-
moid curve of the form.

L, = L 0 exp {C [ 1 - exp i-at )] }

where Lq = length at zero age or hatching
C = a constant
a = rate of decay of exponential

growth
t = age in months.

This model was fitted to our data using an itera-
tive least squares procedure (Conway et al. 1970).
Our goal was to describe growth on a coarse
time scale, i.e., monthly rather than on a fine
time scale, i.e., daily.

GROWTH FROM HATCHING TO
JUVENILE STAGE

The Gompertz growth model and an exponen-
tial growth model were applied to data of Schu-
mann-II by Kramer and Zweifel (1970). Both
models described the data from Schumann-II
reasonably well, although the Gompertz model de-

272

SAKAGAWA and KIMURA: GROWTH OF LABORATORY-REARED ANCHOVY

scribed the data better. In the Kramer-Zweifel
analysis the length at zero age,Lo, was fixed at 2.5
mm, the average size at hatching. We also applied
the Gompertz grovd:h model to data of Schumann-
II. Kramer and Zw^eifel (1970) used data only for 17
days of growth. We used all of the data of Schu-
mann-II, which included sampling through 22
days of growth, and fitted the model first with Lq
fixed at 2.5 mm and again without this constraint,
i.e.,Lo was estimated. The results (Figure 1) indi-
cate that there is not much difference in the curves
withLo fixed or estimated within the range of the
data. Outside the range of the data, the curves
diverge considerably and there is a substantial
difference; the curve withLo estimated has a lower
asymptotic length (61 mm) than the curve withLo
fixed at 2.5 mm (asymptotic length of about 696
mm).

Zweifel and Lasker^ showed that a two-phase
Gompertz curve described the data from Schu-
mann-II better than a single-phase Gompertz
curve. The separation of the phases occurred at
about 6 days of age, the onset of feeding in an-
chovy larvae.

Schumann-Ill reared anchovies for a longer
period than Schumann-II. Fish reared by Schu-

^Zweifel, J. R. and R. Lasker. 1974. Prenatal and postnatal
growth of fishes — a general model. Unpubl. manuscr. Southwest
Fisheries Center, La JoUa, CA 92038.

mann-II, however, were larger than those reared
by Schumann-Ill at similar ages. For example, at
0.5 mo of age fish reared by Schumann-II aver-
aged 12.1 mm long and fish reared by Schu-
mann-Ill, 8.2 mm long. Although the sample size
is small, this difference is statistically significant
at the 1% probability level. Differences in rearing
procedures, i.e., diet and temperature of water,
probably produced the difference in growth
(Kramer and Zweifel 1970; Lasker et al. 1970).

The Gompertz growth model was applied to
data from Schumann-Ill first withLo fixed at 2.5
mm and then with Lo estimated (Figure 2). As in
the case with data from Schumann-II, this model
describes the grovd;h data reasonably well, and
the curve with Lo estimated has a lower asymp-
totic length (81 mm) than the curve with Lq
fixed (asymptotic length of about 93 mm).

GROWTH DURING JUVENILE
TO ADULT STAGE

Anchovies reared by Leong (pers. commun.)
were juveniles at the start of the experiment and
grew to an average size of 117.7 mm in 474 days
(Table 2). Growth was in steplike stages charac-
terized by rapid growth followed by a leveling off.
The first stage was between 4 and 12 mo of age
and the second was between 12 and about 20 mo
of age.

ccoQM ^-0.570t,
L» = 2.500e ^-^29(1-6 )

L,-2.140e 3-546 (l-e-'-279t)

0.0

Q2 03 0.4 0.5 0.6
AGE (months)

0.7

0.8

0.9

FIGURE 1.— Growth of anchovy larvae reared in the laboratory. The Gompertz
growth model of the form, L , = Loexp {C [ 1 - exp (-at)] } is used to describe the data.
Solid line is based on Lofixed at 2.500 mm and broken line is based onLoestimated,
2.140 mm. Data from Schumann-II (Kramer and Zweifel 1970). The mean (circle),
one standard deviation on each side of the mean, and sample size are shown.

273

FISHERY BULLETIN: VOL. 74, NO. 2

Lf = 2.500e ^^2' ^'"^ ^ 1-0

AGE (months)

Figure 2. — Growth of anchovy reared in the laboratory. The Gompertz growth
model of the formL, = Lq exp {C[l - exp (-an] } is used to describe the data. Solid
line is based on Lq fixed at 2.500 mm, and broken line is based on Lq estimated,
2.062 mm. Data from Schumann-Ill (unpubl. data. Southwest Fisheries Center,
La Jolla, Calif). The mean (circle), one standard deviation on each side of the
mean, and sample size are shown.

Table 2. — Estimated age and average standard length of
northern anchovies reared in the laboratory by R. Leong (pers.
commun., Southwest Fisheries Center, La Jolla, Calif.)

Age

Number

Average

length
(mm)

Standard

Days

Months

fish

deviation

120

4.00

88.30

6.34

153

5.10

92.35

4.90

189

6.30

94.63

5.85

231

7.70

97.68

6.40

270

9.00

96.48

6.19

301

10.03

99.30

5.14

351

11.70

101.52

4.59

385

12.83

105.72

5.26

413

13.77

109.16

5.75

444

14.80

109.23

6.07

471

15.70

110 58

6.82

503

16.77

114.56

6.57

533

17.77

117.38

6.11

562

18.73

116.32

7.07

594

19.80

117.68

6.69

The Gompertz growth model was applied, but
did not adequately fit the data. This is charac-
teristic of asymptotic models like the Gompertz
model when all the data points are for a segment
of the growth curve where growth is relatively
slow and the plot of the data exhibits little
curvature.

GROWTH FROM HATCHING
TO ADULT STAGE

Growth Curve

As indicated earlier, anchovies reared by
Shumann-III grew slightly faster than those of
Schumann-II, probably due to slight differences
in the rearing environment and procedures. Be-
cause our goal was to construct a general growth
curve and the differences in the data were rela-
tively slight, we elected to disregard the differ-
ence and pooled the data from the three experi-
ments (Schumann-II, Schumann-Ill, and Leong).
The Gompertz growth model was applied to the
pooled data. The results (Figure 3) indicate that
the model does not adequately describe the data.
For example, the model overestimates the sizes of
fish at about 4 to 12 mo old and underestimates
the sizes of fish older than about 13 mo. These
biases are caused by the steplike growth pattern
which produces plateaus at about 6 mo and 19 mo
of age.

To account for this steplike growth pattern, a

274

SAKAGAWA and KIMURA: GROWTH OF LABORATORY-REARED ANCHOVY
I30|-

8 10 12 14 16 18 20 22

AGE (months)

Figure 3. — Growth of northern anchovy reared in the laboratory. One- phase (broken Hne) and two-phase (soHd
hne) Gompertz growth models are used to describe the data. Data from Schumann-II (Kramer and Zweifel 1970),
Schumann-Ill, and Leong (unpubl. data. Southwest Fisheries Center, La Jolla, Calif.). The mean (circle), one
standard deviation on each side of the mean, and sample size are shown. Broken line is described by L , = 2.825 exp
{3.623 [l - exp (-2.877^]} and solid line by L, = 2.745 exp {3.563 [l - exp (-0.848t)]} for t «11 mo, andL,, .11,=
96.782 exp {0.213 [l -exp (-0.258 {t - ll})]} for Oil mo.

two-phase Gompertz model (Zweifel and Lasker
see footnote 2) was fitted to the data. The two-
phase model is essentially two separate Gompertz
equations that describe different segments of the
growth curve. The equations were fitted simul-
taneously and the convergence point of the equa-
tions was determined on the basis of least squares
analysis. Our best fit of the data was with the

equation, L, = 2.745 exp {3.563 [l - exp
(-0.848n] } for growth from hatching to 11 mo of
age and the equation, L(,.ii) = 96.782 exp {0.213
[1 - exp (-0.258 {t - 11 })] } for growth from 11 to
20 mo of age (Figure 3). From the equations, the
estimated average length of anchovies after 1 yr
of hfe is 101.6 mm and after 2 yr of life, 118.9 mm
(Table 3).

275

FISHERY BULLETIN: VOL. 74, NO. 2

Table 3. — Estimated growth for the first 24 mo of life of the
northern anchovy reared in the laboratory. Estimates are based
on a two-phase Gompertz growth curve (see text).

Standard

Age

length

Age

length

(mo)

(mm)

(mo)

(mm)

hatching

2.7

105.5

21.0

108.6

50.4

111.0

73.2

112.9

85.9

114.5

92.0

115.6

94.7

116.6

95.9

117.3

96.4

117.8

96.6

118.3

96.7

118.6

96.8

118.9

101.6

General Remarks

The steplike growth pattern is commonly found
in fishes. Gerking (1967) reviewed the literature
on this subject and noted that many temperate
species have seasonal, sigmoid growth curves.
Lockwood (1974) recognized this feature in the
growth of plaice and brown trout and applied a
multiphase von Bertalanffy growth model to de-
scribe the data mathematically. His results were
satisfactory but because the von Bertalanffy
growth equation does not describe a sigmoid
curve, his analysis was confined to growth for part
of the season only.

In this study we used the Gompertz growth equa-
tion to describe the sigmoid curve. The two-phase
model satisfactorily described our data for
laboratory-reared anchovy, and a cycle that occurs
at 12-mo intervals is evident in our results. This is
quite similar to the seasonal growth patterns de-
scribed by Gerking (1967), Mann (1971), Kroger et
al. (1974), and others. The cycle indicates that for
the northern anchovy, about 95% of the first year's
growth is completed by the 8th month of life and
about 9 1% of the second year's growth is completed
by the 20th month of life.

If this cyclic pattern in growth also occurs in
anchovies in the wild, then it may have a consid-
erable impact on yield models, such as yield-per-
recruit models, and on management decisions. It
might be that the best harvesting strategy in
terms of maximum yield-per-recruit is during the
period of the cycle when growth is relatively slow,
i.e., period of plateau. It seems important, there-
fore, that a multiphase growth function be
considered for use in yield models for northern
anchovy.

We point out the possibility that the cyclic pat-

tern could have been artificially created because
our data were from three cohorts that were reared
under different laboratory conditions during dif-
ferent periods of the year and the ages of fish
reared by Leong (pers. commun.) were estimated.
However, we discount that possibility because the
cyclic pattern persists even if our age estimates of
Leong's fish were off by 1 or 2 mo. Rearing condi-
tions, on the other hand, could have produced the
cyclic pattern if the pattern is influenced primar-
ily by environmental factors, e.g., temperature,
length of day, and food density and quality

WEIGHT-LENGTH RELATION

Weight-length relations for the northern an-
chovy were reported by several investigators (Ta-
ble 4). OnlyLaskeretal. (1970), however, reported
on estimates for laboratory-reared anchovies, and
their estimates were for anchovy larvae.

Length and weight data were collected by Leong
(pers. commun.) and Paloma^ from fish reared in
their experiments. We used their data from 757
fish to estimate the weight-length relation of
laboratory-reared anchovies of 70 to 131 mm long.
Data from Paloma were only from fish in their first
year of life, in which growth was somewhat simi-
lar to that of fish reared by Leong. Separate esti-
mates were made for males and females (Table 4),
and the results subjected to covariance analysis
(with log transformed data) to test whether the
relation could be represented by a single line. The
analysis indicated that the separate lines were
parallel and not significantly different from a
common line. The data were, therefore, pooled and
a weight-length relation estimated for the com-
bined (all sexes) data (Table 4).

Our estimates are compared with those of Col-
lins (1969) for anchovies from southern California
(Figure 4). Collins based his estimates on data
from anchovies caught in the reduction fishery off
southern California. For a given length, fish
examined by Collins were lighter than the
laboratory-reared fish. This phenomenon appears
common for fishes (Kramer 1969; Kimura and
Sakagawa 1972). Kimura and Sakagawa (1972)
mentioned that for Pacific sardines, differences in
diet and reduced amount of exercise because of
confinement were some possible causes for

^Paloma, P. 1971. Annulus formation in the scale of marked
anchovy Engraulis mordax Girard. Unpubl. manuscr. Southwest
Fisheries Center, La Jolla, CA 92038.

276

SAKAGAWA and KIMURA: GROWTH OF LABORATORY-REARED ANCHOVY

Table 4. — Coefficients of the weight-length relation for the northern anchovy as reported by various investigators. The coefficients are

for the equation, weight = a x length''.

Rearing

environment

Number
of

Unit of

measure

Range of length

Origin of sample

Source

Ocean

Laboratory

Sex

fish

Length

Weight

(mm)

Southern California

Clark and Phillips (1952)

Combined

sexes

3.453

27 X

10"

ounce

56-134

Collins (1969)

Male

926

3.049

8.1 X

io-«

gram

97-161

Collins (1969)

Female

1,513

2.984

1.1 A

10-5

gram

Lasker et al. (1970)

Combined

sexes

3324

1.5 X

10-"

3- 25

Present study^

Male

257

3.521

I.Ox

10-6

gram

73-126

Present study

Female

500

3.433

1.5x

10-6

gram

70-131

Present study

Combined

sexes

757

3.461

1.4 X

10-6

gram

70-131

Central California

Clark and Phillips (1952)

Combined

sexes

(')

3.252

7.2 X

10-6

ounce

114-160

Collins (1969)

Male

270

2.805

2.7 X

10-^

gram

80-171

Collins (1969)

Female

407

2.743

3.6 X

10-6

gram

'Clark and Phillips (1952) used data from 17 samples from southern California and 77 samples from central California but they did not specify the number of fish in
each sample.

^Unpublished data from R. Leong and P. Paloma (pers. commun). Southwest Fisheries Center, La Jolla, Calif. The coefficients with the more appropriate functional
regression (Ricker 1973) are: 1) male, a = 7.0 x 10-', b = 3.608; 2) female, a = 1.0 x la^, b = 3.518; 3) combined sexes, a = 9.2 x 10"^ b = 3.547

X
CD
liJ 20

COLLINS (1969) //

PRESENT STUDY//

100

125

150

175

STANDARD LENGTH [mm)

Figure 4. — Weight-length relation for northern anchovy from
southern California. Laboratory-reared fish were used in pres-
ent study, and fish caught in the California reduction fishery
were used by Collins (1969).

laboratory-reared fish being heavier than fish in
the wild. Zweifel and Lasker (see footnote 2) men-
tioned the possibility that the differences arise
when the curves are based on fish in different
phases of their growth cycle.

COMPARISONS OF GROWTH

Growth of anchovies reared in the laboratory

was studied by Kramer and Zweifel (1970),
Lasker et al. (1970), Theilacker and McMaster
(1971), Hunter (1976), and Paloma (see footnote
3). Kramer and Zweifel and Lasker et al. studied
the effects of diet and water temperature on
growth of anchovy larvae. They concluded that
larval growth was best at 22°C with wild
plankton as a food source. The growth curve of
Figure 3 for the larval stage is for fish reared on
wild plankton and ArtemJa saZma at about 22°C.
It is the best so far attained in the laboratory.

Theilacker and McMaster (1971) and Hunter
(1976) reared anchovy larvae on cultured foods.
Results of their studies show that growth of an-
chovies on cultured food diets is about the same as
that on wild plankton.

Paloma (see footnote 3) obtained juvenile an-
chovies from a live-bait dealer and reared the fish
in the laboratory for 624 days (Table 1). He in-
jected oxytetracycline hydrochloride into the fish
at various times to label the body structures for
ageing. At 2-wk intervals, scales and data on body
measurements were collected. Fish reared by
Paloma started at a smaller average size (75 mm
long) than fish reared by Leong (pers. commun.)
(88 mm long) and grew at a much slower rate (21
mm in about 470 days versus 30 mm in about 470
days for Leong's fish) wdthout a noticeable step-
like pattern. Perhaps the frequent handling, in-
jection of tetracycline, and small size of the rear-
ing pool (2.74-m diameter) contributed to the slow
growth and eliminated the steplike pattern.

Clark and Phillips (1952), Miller et al. (1955),
Collins (1969), and Collins and Spratt (1969),
studied growth of anchovies caught in the Califor-
nia fisheries. They used scales and otoliths for
ageing fish to the nearest whole year. Clark and
Phillips reported their results for the combined

277

FISHERY BULLETIN: VOL. 74, NO. 2

southern and central California fisheries. Miller
et al., Collins, and Collins and Spratt, on the other
hand, reported their results separately for each
fishery.

To compare growth of anchovies in the wild with
that of laboratory-reared anchovies, we limited
the comparison to fish from southern California to
eliminate possible regional biases in grow^th. We
also adjusted Miller et al.'s (1955) data upward by
1 yr to make them comparable to those of Collins
(1969) and Collins and Spratt (1969) (Figure 5).
This was necessary because Miller et al. did not
correct their age readings for date of capture (Au-
gust to March) and growth on the margin of the
scale relative to the birthdate (April 1); hence,
they underestimated the age of their fish by ap-
proximately 1 yr.

The growth curves in Figure 5 indicate that
anchovies in the wild are 95 to 115 mm long at
about lyroldand 115 to 125 mm long at about 2 yr
old and possibly grow^th was slower in the 1960's
than in the 1950's owdng to the dramatic increase
in the northern anchovy population (Spratt 1975).
Our growth estimates for laboratory-reared an-
chovies are 102 mm for 1 yr olds and 1 19 mm for 2
yr olds; hence, growth of laboratory-reared fish
seems to be similar to that of anchovies in the wild.
However, we note that this direct comparison is
not entirely valid because inherent biases exist in
the grovvi;h curves in Figure 5. The biases exist
because: 1) larger fish are generally more avail-
able to the reduction fishery than the live-bait
fishery (Messersmith 1969) and thus are over-
represented in the data for the reduction fishery;
2) live-bait fishermen "consciously avoid taking
large anchovies, since they are less desirable for
bait than smaller anchovies" (MacCall 1973:5-6)
and thus large fish are underrepresented in the
data for the live-bait fishery; 3) the true birthdate
of anchovies aged by otolith or scale readings is not
known although it is known that the birth date
varies (Kramer and Smith 1971; Smith 1972), the
ages, therefore, are not exact ages; and 4) growth
of several year classes are averaged and con-
sequently, variability in growth is reduced.

Spratt (1975), who also studied growth of the
northern anchovy from otoliths, accounted for
some of these biases by using back-calculated
lengths and fish from the reduction fishery, live-
bait fishery, and catches of a research vessel. He
estimated that the mean standard length of an-
chovies in the wild is 92 and 112 mm at the end of
the first and second year of life, respectively. These

I60r

150

LLJ

<
Q

<
Ul

<
OH

140

130

120

110

100

90-

80-

70-

1953-54 CANNING-REDUCTION

FISHERY
1965-66 REDUCTION FISHERY
1966-67 REDUCTION FISHERY

1966 LIVE-BAIT FISHERY

1967 LIVE-BAIT FISHERY

n m

AGE GROUP

Figure 5. — Growth curves for northern anchovy caught off
southern California in the fisheries for reduction (Collins 1969),
live-bait (Collins and Spratt 1969), and canning-reduction (Mil-
ler et al. 1955).

estimates are somewhat less than ours for
laboratory-reared fish but they are close.

It appears that growth of anchovies in the wdld is
similar to that estimated, on an annual basis, from
our growth curve. We have not demonstrated,
however, whether there is a cyclic pattern in
growth of anchovies in the wild similar to that
revealed in our results for laboratory-reared fish.
On the other hand, studies on growth of other
temperate fishes have shown that a seasonal cycle
is common, which leads us to believe that a sea-
sonal cycle exists for anchovies in the wild. The
use of our growth curve for describing the feeding
djmamics of northern anchovies of at least 2 yr of
age in the California Current is therefore practical
until a seasonal growth curve is described for an-
chovies in the wild.

ACKNOWLEDGMENTS

We are indebted to the many individuals that
contributed to the development of rearing proce-
dures for the northern anchovy at the Southwest
Fisheries Center, La Jolla, Calif. We especially
acknowledge the contributions of Roderick Leong
and Pedro Paloma, who generously provided us
with unpublished data. David Kramer, Reuben
Lasker, William Lenarz, Alec MacCall, and James

278

SAKAGAWA and KIMURA: GROWTH OF LABORATORY-REARED ANCHOVY

Zweifel read the manuscript and offered valuable
comments and suggestions. Robson Collins and
members of his staff at the California Department
of Fish and Game read the manuscript and
brought to our attention related studies that were
in press.

LITERATURE CITED

BARDACH, J. E.

1968. Fish culture. In The status and potential of aquacul-
ture, particularly fish culture, Part III, Vol. 2, p.
1-193. American Institute of Biological Sciences. Availa-
ble Clgh. Fed. Sci. Tech. Inf., Springfield, VA 22151 as PB
177 768.

BEVERTON, R. J. H., AND S. J. HOLT

1957. On the dynamics of exploited fish populations. Fish.
Invest. Minist. Agric. Fish. Food (G.B.), Ser. II, 19, 533 p.

Clark, F. N., and J. B. Phillips.

1952. The northern anchovy (Engraulis mordax mordax) in
the California fishery Calif Fish Game 38:189-207.

Collins, R. a.

1969. Size and age composition of northern anchovies (En-
graulis mordax) in the California anchovy reduction
fishery for the 1965-66, 1966-67, and 1967-68, seasons. In
The northern anchovy (Engraulis mordax) and its fishery
1965-1968, p. 56-74. Calif. Div. Fish Game, Fish Bull. 147.

Collins, R. a., and J. D. Spratt.

1969. Age determination of northern anchovies, Engraulis
mordax, from otoliths. In The northern anchovy (£n-
^raw/tsmorc/ajc) and its fishery 1965-1968, p. 39-55. Calif.
Div. Fish Game, Fish Bull. 147.

CONWAY, G. R., N. R. GLASS, AND J. C. WiLCOX

1970. Fitting nonlinear models to biological data by Mar-
qusirdt's algorithm. Ecology 51:503-507.

GERKING, S. D.

1967. Annual growth cycle, grovrth potential, and growth
compensation in the bluegill sunfish in northern Indiana
lakes. J. Fish. Res. Board Can. 23:1923-1956.
HUNTER, J. R.

1976. Culture and growth of northern anchovy, Engraulis
mordax, larvae. Fish. Bull., U.S. 74: 81-88.
KiMURA. M., AND G. T. SAKAGAWA

1972. Observations on scale patterns and growth of the
Pacific sardine reared in the laboratory. Fish. Bull., U.S.
70:1043-1052.
KRAMER, D.

1969. Synopsis of the biological data on the Pacific mac-
kerel. Scomber japonicus Houttuyn (Northeast Pacific).
U.S. Fish Wildl. Serv., Circ. 302, 18 p.

KRAMER, D., AND P. E. SMITH.

1971. Seasonal and geographic characteristics of fishery
resources, California Current region — V. Northern an-
chovy Commer. Fish. Rev. 33(3):33-38.

KRAMER, D., AND J. R. ZWEIFEL.

1970. Growth of anchovy larvae (Engraulis mordax Girard)
in the laboratory as influenced by temperature. Calif
Coop. Oceanic Fish. Invest. Rep. 14:84-87.

KROGER, R. L., J. F. Guthrie, and M. H. Judy.

1974. Growth and first annulus formation of tagged and

untagged Atlantic menhaden. Trans. Am. Fish. Soc.

103:292-296.
LAIRD, A. K.

1969. The dynamics of growth. Res./Dev. 20(8):28-31.
LASKER, R., AND L. L. VLYMEN.

1969. Experimental sea-water aquariimi. Bureau of Com-
mercial Fisheries Fishery-Oceanography Center, La
Jolla, California. U.S. Fish Wildl. Serv., Circ. 334, 14 p.

Lasker, R., h. m. feder, G. H. Theilacker. and r. C. May.

1970. Feeding, growth, and survival oi Engraulis mordax
larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345-
353.

LEONG, R.

1971. Induced spawning of the northern anchovy, En-
graulis mordax Girard. Fish. Bull., U.S. 69:357-360.

LOCKWOOD, S. J.

1974. The use of the von Bertalanffy growth equation to
describe the seasonal growth offish. J. Cons. 35:175-179.

MacCall. a. D.

1973. The mortality rate of Engraulis mordax in southern
California. Calif Fish Game, Mar Res. Tech. Rep. 4, 23 p.

Mann, R. H. K.

1971. The populations, growth and production of fish infour
small streams in southern England. J. Anim. Ecol.
40:155-190.

MESSERSMITH, J. D.

1969. A review of the California anchovy fishery and re-
sults of the 1965-66 and 1966-67 reduction seasons. In
The northern anchovy (Engraulis mordax) and its fishery
1965-1968, p. 6-32. Calif Div. Fish Game, Fish Bull. 147.

Miller, d. J., A. E. daugherty, f. e. felin, and J. mac-
Gregor.

1955. Age and length composition of the northern anchovy
catch off the coast of California in 1952-53 and 1953-
54. In Age determination of the northern anchovy, En-
graulis mordax, p. 36-66. Calif Div. Fish Game, Fish
Bull. 101.

Parker. R. R., and p. a. Larkin.

1959. A concept of grow1;h in fishes. J. Fish. Res. Board
Can. 16:721-745.

Richards, f. j.

1959. A flexible growth function for empirical use. J. Exp.
Bot. 10:290-300.
RICKER, W. E.

1973. Linear regressions in fishery research. J. Fish. Res.
Board Can. 30:409-434.
SILLIMAN, R. P.

1969. Comparison between Gompertz and von Bertalanffy
curves for expressing growth in weight of fishes. J. Fish.
Res. Board Can. 26:161-165.

Smith, p. E.

1972. The increase in spawning biomass of northern an-
chovy, Engraulis mordax. Fish. Bull., U.S. 70:849-874.

Spratt, J. d.

1975. Growrth rate of the northern anchovy, Engraulis mor-
dax, in southern California waters, calculated from
otoliths. Calif Fish Game 61:116-126.

THEILACKER, G. H., AND M. F. MCMASTER.

1971. Mass culture of the rotifer Brachionus plicatilis and
its evaluation as a food for larval anchovies. Mar. Biol.
(Berl.) 10:183-188.

Vrooman, a. M., and p. E. Smith

1971. Biomass of the subpopulations of northern anchovy
Engraulis mordax Girard. Calif Coop. Oceanic Fish. In-
vest. Rep. 15:49-51.

279

DEVELOPMENT AND USE OF SONAR MAPPING FOR PELAGIC
STOCK ASSESSMENT IN THE CALIFORNIA CURRENT AREA^

Roger P. Hewitt, Paul E. Smith, and John C. Brown^

ABSTRACT

A method for pelagic fish stock assessment is presented which utilizes a fixed sonar beam for mapping
fish schools. Samples of the two major acoustic properties of fish schools are presented, i.e., acousti-
cally derived horizontal dimensions (representative of school volume) and target strengths (which
may be representative of school compaction). Sampling biases and sources of sampling variability in
the measurement of these properties are discussed. The results of two experiments, conducted to
determine the weight of a fish school as a function of its acoustic characteristics, are presented. In the
first experiment, an acoustically transparent trap was used to recreate an aggregation offish and in
the second, commercial fishing boats were chartered to capture whole schools. An automated sonar
data acquisition and processing system is described and test results presented. The results of paired
automated surveys of the Los Angeles (southern California) Bight are presented and discussed. The
paper reports development of the sonar-fish school mapping method first documented by P. E. Smith
in 1970.

Field investigations, conducted in cooperation with the Navy and the California Department of
Fish and Game, indicate a median school size of 30 m diameter, a mean fish density of 15 kg offish
biomass per square meter of horizontal school area, and a biomass estimate of 1.23 to 2.30 x 10^
metric tons for pelagic schooled targets in the Los Angeles Bight.

Fishermen have used hydroacoustic apparatus
for locating concentrations of fish for almost as
long as practical echo sounding devices have been
available, although quantification of the informa-
tion they provide has been attempted only in re-
cent years. Horizontal echo ranging (sonar) to
locate fish schools was first used off the coast of
California in 1946 (Smith 1947; Smith and
Ahlstrom 1948). The 1950 progress report of the
California Cooperative Sardine Research pro-
gram notes the use of sonar and echo sounders on
the RV Yellowfin for locating fish schools, and
cites the "considerable experimental value" of the
acoustic apparatus. A research sonar on the
RV David Starr Jordan has been used to count
fish schools in the eastern tropical Pacific (Mc-
Clendon 1968) and in the California Current
area (Smith 1970). For recent reviews of the
use of echo sounders and sonars for fishery re-
search, consult Forbes and Nakken (1972) and
Cushing (1973).

The work presented here is a method for quan-
tifying sonar records and further using these re-

^Conducted under a grant from the Marine Research Com-
mittee of the California Department of Fish and Game as part
of the California Cooperative Oceanic Fisheries Investigations,
and in cooperation with the United States Navy.

^Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

Manuscript accepted October 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

cords for estimating the size of pelagic fish stocks.
The paper is divided into four sections:

1. The section entitled "Sources of sampling
variability" describes the scale and variance
of measured acoustic parameters of fish
schools, i.e., horizontal fish school dimen-
sions and peak target strength or echo in-
tensity. It further discusses major biases af-
fecting the measurement of these values.

2. The estimation of fish biomass in an
aggregation involves the determination of a
conversion factor by which the detected
horizontal area of a fish school may be mul-
tiplied. Experiments to determine the
weight of the fish under a square meter of
school area are described in a section en-
titled "Horizontal school area to biomass
conversion factors."

3. An automated data acquisition system is de-
scribed in the third section.

4. The results of a paired sonar survey of the
Los Angeles Bight, utilizing the automated
system and a biomass factor determined
during the cruise, are presented and discus-
sed in the fourth section.

This report is the second in a series describing

281

FISHERY BULLETIN: VOL. 74, NO. 2

progress on a number of objectives established in
early 1968. In order to develop "sonar mapping"
as a stock assessment tool, it was decided that
such a system should be able to: 1) count the
number of schools per unit area in the upper
mixed layer from a ship proceeding at 12 knots, 2)
measure the horizontal size of each fish school, 3)
calculate the biomass of each school, 4) estimate
the size of individual fish within a school, and 5)
distinguish the northern anchovy from all other
schooling species.

Smith (1970) developed a technique for "map-
ping" fish schools in the area where the northern
anchovy, Engraulis mordax, is abundant off the
coast of southern California. Sonar mapping dif-
fers from echo sounding; with sonar, estimates
can be made of the number offish schools per unit
area, of their horizontal dimensions, and of the
degree of aggregation of fish schools. We do not
routinely estimate depth of the school in the
water column, nor thickness of the school in the
vertical plane. Hull-mounted echo sounders pro-
vide estimates of the number of schools per line
transect deeper than 4 m, measures of chords
across the horizontal dimension of the school in
the plane of ship travel, depth in the water col-
umn, and thickness or vertical height of the fish
school. Experience indicates that the process of
"sonar mapping" encounters one or two orders
of magnitude more fish-school targets per unit of
ship time as compared to echo sounding from
the same vessel. It is important to emphasize that
this technique was developed because fish
schools are frequently found in the upper mixed
layer of the ocean where echo sounders are rela-
tively ineffectual at counting or measuring them.

In the first report on this project. Smith (1970)
described a series of experiments designed to de-
termine the feasibility of the use of sonar to count
and measure the size of pelagic fish aggregations
(objectives 1 and 2). Optimum instrument set-
tings were determined for source level, receiver
gain, pulse length, transducer bearing, trans-
ducer directivity, and range. Methods were de-
veloped for correcting target width (dimension
measured on axis parallel to ship's track) for the
effect of the beam angle and for correcting target
count "edge biases." Since no target was counted
unless it lay entirely within a specified range, the
latter adjustment was made to compensate for the
narrowing possible interval of detection for larger
targets.

Holliday (1972, 1974) investigated the fre-

quency domain processing of fish school echoes
using experimental equipment brought aboard
the David Starr Jordan. By detecting and
measuring Doppler spread, Holliday was able to
calculate tail beat amplitudes of schooled fish
and, indirectly, their length (objective 4).

Holliday also examined the resonance struc-
ture of pulse returns from fish schools and was
able to detect the presence or absence of a swim
bladder in the school constituents. This informa-
tion, when supplemented by observations on
school behavior and free vehicle camera drops,
may be used to distinguish anchovy from other
pelagic schooling organisms in a sample taken
randomly from targets encountered during a sur-
vey (objective 5). The statistical base thus ob-
tained would be applied to the entire survey.

The California Department of Fish and Game
(CF«&G) has been engaged in sea surveys using
sonar methods since 1967 (Mais 1974). Its ap-
proach has been the collection of large amounts of
data and its interpretation, while the work at the
Southwest Fisheries Center (SWFC) has been in
the isolation of sampling errors and the develop-
ment of an automated hydroacoustic data acqui-
sition and processing system. As such, the two
groups complement each other with field experi-
ence and technological development.

SOURCES OF
SAMPLING VARIABILITY

We have made the assumption that quantita-
tive errors associated with system instrumenta-
tion are small in comparison to errors generated
by sampling an adult schooling population whose
behavior is little understood. For this reason, we
monitored our sonar system response when it was
operated in a variety of circumstances and
changed that system in answer to practical rather
than theoretical considerations. Using operating
techniques developed in 1968, school size fre-
quency distributions were generated and a lower
detectable size threshold defined; school target
strengths were calculated and compared with
similar work conducted by the Navy and the
CF&G; the relationship between the detected oc-
currence of pelagic fish schools and bottom topog-
raphy was investigated; and the variable range of
detection of schools due to internal waves was
studied (Smith^).

^Smith, P. E. 1973. The effects of internal waves on fish school

282

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

Based on Smith's (1970) work, sonar mapping
cruises aboard the David Starr Jordan were con-
ducted with a 30-kHz sonar unit directed 90° to
starboard and 3° down. The sampled range band
was 200 to 450 m from the transducer. The re-
ceivers were rebuilt using solid state circuitry
with the remaining system as described by Smith
(SIMRAD 580-10 Scientific Sonar and Sounder).^

Target Size

Frequency distributions of fish school sizes
were generated from data taken on several
cruises (April-May, November, December 1973;
and March-April 1974) using the maximum dif-
ference between the leading and trailing edge of
the echo envelope, corrected for pulse length, on
an axis perpendicular to the ship's track. The cal-
culation of target widths (measured on an axis
parallel to the ship's track) was discontinued due
to uncertainties in choosing the effective beam
width (see Smith 1970), fluctuations in the ship's
speed, and the inability to quantify other factors
which may affect apparent target width (i.e.,
target strength).

School size distributions (based on range differ-
ences) remained nearly constant during several
sampling periods and agreed well with a much
larger sample collected by the CF&G. A total of
4,355 sonar targets were counted and assigned to
size classes on three cruises approximately 6 mo
apart. Ten-meter class intervals were used and
frequencies were corrected for recording edge bias
employing the method described by Smith (1970).
This bias is encountered when one excludes
targets which do not entirely occur within the
observation band. Thus, frequencies of targets
other than point sources, are underestimated by
virtue of the fact that their physical size limits
the probability of their detection. To determine
unbiased relative proportions of target sizes, one
must correct observed target count (those targets
which lie entirely within the observation band) to
a count of targets whose centers lie within the
observation band.^

mapping. Presented at the ICES-ICNAF-FAO Symposium on
the Acoustic Methods in Fisheries Research, Bergen, Norway,
Contrib. No. 8, 13 p.

"•Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

^It is assumed that range-dependent, size-specific target los-
ses are a minimum for the observation band sampled (Smith
1970). A similar study expanded to include the effects of target
strength on detection ranges would be of value.

In developing a correction for recording edge
bias, a diagram may be useful. In Figure 1 a
school of diameter d is shown at the maximum
and minimum ranges of detection for an observer
on a ship sampling an observation band of k
units. The difference between the maximum and
minimum range of detection isk - d units.

Let A represent the event that a school of d
diameter has its center within an observation
band of ^ units. LetB represent the event that a
school of d diameter is not intersected by either
edge of the observation band. Then the probabil-
ity of event B occurring given that event A has
occurred may be expressed:

P[BIA] =^^.

Further, let A^^ represent the count of targets of
diameter d who lie entirely within the observa-
tion band. Let N'^ represent the count of targets
of diameter d whose centers lie within the obser-
vation band. Since A^'^ represents both edge inter-
sected and non-edge intersected targets of diame-
ter d, the portion of non-edge intersected targets
may be estimated by:

Nd = N'aP[BIA] =N'a

Ship

k-d

Observation band

Figure l. — Plan view of sonar mapping technique showing
maximum and minimum ranges of detection for a target of
diameter d within an observation band of k units.

283

FISHERY BULLETIN: VOL. 74, NO. 2

In actual practice Nj is tabulated. N\i is then
estimated by rearranging the above expression:

N', = N,

where Nj = observed class frequency

N'ci= edge corrected class frequency
k = extent of the observation window

in meters (usually 250 m)
d = mean class diameter in meters.

As an example, when using a 250-m observation
band, a 50-m target may be entirely detected over
200 m of that band, whereas a 100-m target must
occur within a band of only 150 m to be detected.
If one counts 10 50-m targets and 3 100-m
targets, the counts, when corrected for edge bias,
will be 10(250)/(250 - 50) = 12.5 and 3(250)/(250
- 100) = 5, respectively.

Horizontal school area is calculated by multi-
plying A?^' by the area of a circle whose diameter is
equal to the class mark. The calculation is based
on the assumption that with an increasing sam-
ple size the school dimension perpendicular to the
ship's track will approximate the diameter of a
circle whose area is equal to the area of a given
school, however irregularly shaped. This assump-
tion contains the condition that the orientation of
a sample of schools is random and in no way re-
lated to that of the survey ship.

The resulting cumulative frequency diagram
(Figure 2) would indicate that over 507c of the
schools are less than 30 m in diameter while 90%
of the horizontal school area is contributed by
schools larger than 30 m in diameter. Mais' (1974)
experience with over 23,000 schools (corrected for
edge bias) in the same survey area indicated a
similar distribution with a mode at 30 to 40 m
(Figures 2, 3).

Smaller schools (<20 m in diameter) were
likely to be undersampled by both the National
Marine Fisheries Service (NMFS) and CF&G as
the probability of their detection decreases faster
with range than larger schools. Even if an expo-
nential model of target size obtains in nature,
schools smaller than 20 m would contribute little
in amounts of horizontal school area.

The significance of a negative bias in the lower
end of the observed school size distribution may
be evaluated by fitting a power curve to that por-
tion of the distribution between 15 and 165 m.

lUU

_._-^-j-j^' -7

NUMBER OF .. ■~_^,.^ /

SCHOOLS ,.^ ' ^ /

(CDFSG) ,'" ^^ ^^

^ y^ ^

/ / NUMBER OF /
/ / SCHOOLS /
, / (NMFS) /

' / /

/ / /''horizontal

; / / SCHOOL AREA

// / (NMFS)

'/ ''
'/ /

/ / /

'/ /

/ / /

,7 /

// /

/ /

7 /

/ /
/ /
/ /

/ /

/
/

/
/
/
/

■r' \ \ \ \ \ \ \ \ \ II

0 20 40 60 80 100 120 140 160

Figure 2. — Cumulative frequencies of sonar-detected fish
schools by size and their contributing horizontal area (NMFS
data only). The two modes in the CF&G data curve, drawn
from a much larger sample (5x), might suggest either a sys-
tematic sampling error or optimum fish school sizes.

The equation, derived by a least squares fit, as-
sumes the following form:

y = ax .

Using the NMFS sample of 4,355 targets:
AT', = 428,864 {D mX^-^''^

where A''', = edge-bias corrected target fre-
quency within class i
(Dm)i = mean diameter of class / in meters.

-3 20

_)
I- o
2 O 10
UJ I
o o
cr lo

40 60 80 100 120

MEAN CLASS DIAMETER (meters)

Figure 3. — Percent of total school count by size class. NMFS
data are represented by the shaded bars; the
open bars are calculated from CF&G data.

284

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

The correlation coefficient (r) = -0.969. Table
1 summarizes horizontal school area contri-
butions by size class for observed frequencies
corrected for edge bias and for frequencies de-
rived from the exponential model. In both cases
more than 90% of the area was contributed by
schools larger than 20 m. The importance of hori-
zontal school area is that it is probably propor-
tional to the tonnage offish in schools and, in this
sense, decreases the significance of any bias in
the counts of small schools.

Table l. — Cumulative percent of total horizontal school
area contributed by size class for observed frequencies (cor-
rected for edge bias) and for frequencies derived from an
exponential model.

Cumulative

% A

Mean class
diameter

A/'

Model

Observed

Model

420

429

21018

0.09

4.24

1,247

1,347

2682

248

9.12

843

937

1030

7.17

14.32

556

647

548

1352

19.73

403

491

342

21.50

25.32

277

355

235

30.11

31.05

182

246

172

38.44

36.91

124

177

131

4633

42.88

130

104

53 98

48.98

6063

5509

105

67.79

61.39

115

74.07

67.72

125

79.58

74.03

135

85.61

80.42

145

90.02

86.89

155

9357

93.53

165

99.99

>165

12
4,355

Diurnal and Seasonal Effects

Time specific frequency distributions were
drawn for data collected on cruises in April-May
and in November 1973 for the purpose of discern-
ing variations in sizes and detection of schools
during various times of the day. While variations
were noticed, their pattern was neither pro-
nounced nor consistent from cruise to cruise. This
is not to say that daily changes in schooling be-
havior do not exist, but that our data base is in-
sufficient, at present, to delineate them. In the
evening, discrete, well-formed schools of anchovy
have been observed to disperse into a thin scat-
tered layer but no program of study on this prob-
lem has been undertaken.

The data base is insufficient to detail seasonal
changes in school size distributions, although,
from communication with Mais and several
commercial fishermen, we have reason to expect
somewhat larger schools in the fall and smaller,

scattered schools in the spring. Mid-spring is con-
sidered to be the main spawning season of the
northern anchovy.

Target Strength

Acoustic target strength is proportional to the
ability of an object or group of objects to reflect
sound waves. Acoustic reflections from schools of
fish are not presently well enough understood for
rigorous characterization of the biomass of a fish
school by the use of sonar. Nevertheless, we have
measured apparent fish school target strengths
with the objective of providing data which may
lead to the quantification of fish schools in terms
of total biomass.

Peak echo amplitudes were collected and cor-
rected for propagation and absorption losses by
employing the active sonar equation:

EL = SL - 2TL + TS

where EL
SL

echo level in decibels (dB)
source level in decibels, reference
1 yubar at 1 m
TL = transmission loss in decibels
TS = target strength in decibels.

Solving for target strength and using signal vol-
tage level as a measure of echo level:

TS = 20\ogV - k + 40 \ogR + 2 ozR

where V = peak echo signal amplitude in

volts
k = calibration coefficient which is
the algebraic sum of source
level, receiver sensitivity, and
system gain expressed in
decibels
40 log R +

2a:i? = range dependent transmission
loss (assuming spherical losses
as in a homogeneous fluid)
where R = midrange of target
(as an approximation of the (lo-
cation of peak echo amplitude),
and cc = absorption coefficient
expressed in decibels per meter.

Figure 4 illustrates five samples of peak target
strengths computed from data taken by the

285

FISHERY BULLETIN: VOL. 74, NO. 2

NMFS, U.S. Navy, and CF&G. Two of the distribu-
tions are "absolute" target strength in decibels
and three are relative measurements, i.e., the
calibration coefficient was not included in the cal-
culations. The range of peak target strengths ob-
served in any one sample varies from 28 to 34 dB.
The two distributions of absolute target strength
were obtained with the same sonar unit aboard the
David Starr Jordan. The value of the calibration
coefficient was recomputed after hydrophone

calibration between cruises and remained con-
stant. As such, the favorable comparison between
the samples may be deceptive. The CF&G data
were obtained and processed in a similar fashion
using a 38-kHz sounder.

The theoretical target strength of a fish school
has been discussed by Weston (1967) and Uretsky
(1963). Modeling a fish school as a two dimen-
sional array of bubbles in a liquid, both Weston
and Uretsky predicted a sharp drop in response

30-

NOVEMBER
n=l98
)J = -l.ldB
s= 8.8dB
range= 35 dB

JULY-AUGUST
n=l78
x = -9.lldB
s= 6.62 dB
range = 32 dB

' -13 -7-1 5 II 17
RELATIVE TARGET STRENGTH (dB)

-21 -17 -13 -9-5-1 3 7 1
ABSOLUTE TARGET STRENGTH (dB)

. c.

L-|

MARCH-APRIL
n=272 '^
x= -22.72
s=6.l7
range= 34dB

—J

't^ .

40-

30-

20-

10-

-40 -30 -20 -10 0

RELATIVE TARGET STRENGTH (dB)

I 6. FEB., OCT, a DEC.
^"'~ n=ll7

x= 14.2
s= 8.2
range= 36 dB

16-

DECEMBER
n=l09
x=-9.ll
s= 5.79

ranges 28 dB

J L.

I'll I i_

3
O
UJ
(T
U.

5 :f

-20 -16 -12 -8 -4 0

ABSOLUTE TARGET STRENGTH (dB)

10 Figure 4. — Distributions of five samples of peak school
target strengths; a and b are from NMFS data, c and d
are from Navy data, and e is from CF&G data.

-7 -1 5 II 17 23
RELATIVE TARGET STRENGTH (dB)

286

HEWITT ET AL.; DEVELOPMENT AND USE OF SONAR MAPPING

with increasing frequency above resonance. Using
this approach, the energy scattered by the bound-
ary of a fish school ensonified (irradiated acousti-
cally) with 30 kHz sound becomes negligible.

Weston (1967) further suggested that an inco-
herent addition of reflected energy from indi-
vidual fish may be expected as sound is trans-
mitted across the boundary of a fish school. At 30
kHz, this component of target response becomes
dominant and is reduced (or enhanced) by multi-
ple scattering and absorption within the school.

The target response due to sound scattering by
individual fish, assuming a mean wave phase in-
terference of zero, may be calculated by summing
the scattering cross sections of the fish comprising
the target. Expressed in target strength, TS:

TS = TS, + 10 log n (decibels)

where TS, = the average target strength of the
individual scatterer
n = the number of scatterers contribut-
ing to the total echo.

The number of scatterers contributing to the
measured echo, n, may be estimated by applying
observed and theoretical school densities (fish per
cubic meter) to the ensonified volume. The enson-
ified volume may be estimated from:

V = —{d)iD) (cubic meters)

(1)

where —

the range extent of the volume
sampled by a sound pulse t sec-
onds long and moving at a speed
of c meters per second

D = the vertical dimension of the school
in meters

d = the horizontal dimension of the
school.

School dimensions, D and d, are further limited
by beam geometry, i.e., a school may not be fully
ensonified if its dimensions exceed the effective
beam width at the range of detection. The effective
horizontal beam width may be estimated as that
between the half-power points or:

2/?tani3

where R = range of detection

(B = 5° for the 30-kHz transducer used
in this study.

Thus, d is the smaller of the measured horizontal
dimensions or 0.175i?. Vertical dimensions offish
schools are not readily measured with sonar. How-
ever, in studying echograms of thousands of
schools, Mais (1974) noted less variation in the
vertical school dimension than the horizontal di-
mension and reported a mean school thickness
of 12 m. The vertical effective beam width is esti-
mated to be 12° or 42 m at 200-m range. If D is
then assumed to be 12 m for all schools, there is
no limitation imposed by the vertical beam width
except that caused by vertical positioning of the
school.

Using a 10 ms pulse length and estimating the
speed of sound in a seawater medium at 1,500
m/s, Equation (1) becomes:

V = 90d

where d is the smaller of the measured horizontal
dimensions or 0.175 R.

Mais (1974) reported visual observations of an-
chovy schools and estimated average packing
density at 50 to 75 fish/m^. Graves^ analyzed in
situ photographs of three anchovy schools and re-
ported a mean density of 115 fish/m^ at a mean
spacing of 1.2 body lengths. Hewitt'' used an
idealized model of anchovy school compaction and
calculated school densities of 0.5, 1.4, 6.6, 217,
and 4,219 fish/m^ at interfish distances of 10, 7, 4,
1, and 0.2 body lengths, respectively.

The target strength of an individual scatterer,
TS, may be estimated from considerations of
acoustic theory and extensions of empirical mea-
surements. Weston (1967) had shown the acoustic
response of an ideal gas bubble to be essentially
independent of frequency above resonance and
proportional to the surface area of the bubble.
When predicting the response of a fish swim
bladder, Weston suggested an enhancement of

^Graves, J. 1974. A method for measuring the spacing and
density of pelagic fish schools at sea. SWFC Administrative
Report No. LJ-74-44. Southwest Fisheries Center, NMFS,
NOAA, La Jolla, CA 92038.

Hewitt, R. 1975. Sonar mapping in the California Current
area: A review of recent developments. Unpubl. manuscr.
Southwest Fisheries Center, NMFS, NOAA, La Jolla, CA
92038. The compaction model cited here used an anchovy of 12
cm standard length and computed the space required for the
fish and a surrounding volume expressed in body lengths. The
inverse of the resulting volume yields compaction in fish per
cubic meter for a school of fish uniformly distributed in space.

287

FISHERY BULLETIN; VOL. 74, NO. 2

75% due to shape distortion. Expressed in target
strength:

TS, = 20 log L - 25 (decibels)

(2)

where L is the fish length in meters. Swim blad-
der volume is assumed to be 4.1% of total fish
volume and the radius of a sphere of equal vol-
ume equal to 0.043 L (after Haslett 1965).

Using a standard length of 12 cm as typical of
anchovy school constituents detected by sonar
(Mais 1974), Equation (2) yields a TS, of -43.4
dB. It should be noted that Equation (2) makes no
provision for reflection, interference, or attenua-
tion of sound waves by fish tissue.®

McCartney and Stubbs (1970) measured
maximum dorsal aspect target strengths of six
fish species at varying frequencies and lengths.
They fit Equation (3) to their data and further
showed that the swim bladder can account for
practically all of the scattering over a wide band
of frequencies:

TS, = 24.5 logL - 4.5 log X - 26.4

(3)

where X = the wavelength of incident sound
defined as c(f)'^, where c is the speed of sound in a
saltwater medium = 1,500 m/s"^ and /"is the fre-
quency. For a 12 cm anchovy ensonified with 30
kHz sound. Equation (3) gives a TS, of -43.1 dB.
Love ( 197 1 ) reviewed maximum dorsal and side
aspect target strength measurements made by
several investigators. The data were obtained
using fish from eight different generic orders,
varying 100-fold in length, some with swim blad-
ders and some without, and ensonified over a fre-
quency range of 8 to 1,480 kHz. For dorsal aspect,
Love related maximum target strength, fish
length, and frequency by:

TS, = 19.4 logL + 0.6 log A - 24.9. (4)

For the anchovy described above, Equation (4)
predicts a TS, of -43.5 dB at dorsal aspect.

Love described the side aspect data with the
following equation:

TSi = 22.8 logL - 2.8 log \ - 22.9 (5)

or -40.2 dB for the anchovy described at side
aspect.

A similar regression on target strength mea-

®Holliday (1972) reported an average swim bladder volume of
2.8% of the total fish volume for a sample of 239 anchovy. The
use of this value predicts an anchovy swim bladder response of
-44.3 dB.

surements taken from dead fish in dorsal aspect
by six investigators and collated by Haslett
(1965) would describe a TS, of -49.8 dB for a
12-cm fish ensonified at 30 kHz (McCartney and
Stubbs 1970). An application of the equations
that Shibata (1970) used to describe his results
yielded values of -42.8 dB for maximum dorsal
aspect target strength and -40.0 dB for maxi-
mum side aspect target strength.

Several authors have noted that acoustic
equipment commonly used by the biologist oper-
ates at frequencies (10 to 200 kHz) which gener-
ate sound at wavelengths comparable with the
size of fish under study. Interferences will occur
among the scattering components of a fish (swim
bladder, flesh, skeleton, and organs) and may be
expected to be a function of species and aspect.
Further, our measurements are of peak school
target strength taken from several transmissions
along one tangential to the school and may not be
the maximum value which would be obtained
from interrogation at several angles.

Let us return now to the original calculations,
i.e., the incoherent summation of echoes from an
aggregation of fish which may now be expressed

as:

TS = TS, + 10 log [q (90 d)]

(6)

where TS, may vary from —50 to -40 dB, g is the
school density in fish per cubic meter and may
vary from 0.5 to 4,219, and d may vary from 5 m
(mean diameter of the minimum class size) to 79
m (0.175 R at R = 450 m, the maximum range
within the observation band). The expected range
of peak school target strengths (assuming inco-
herent addition and no interference or absorption
within the school) are listed below for foiu- as-
sumptions of fish target strength, TS,:

TS,

Minimum TS

Maximum TS

where q = 0.5 fish/m^

where q = 4,219 fish/m^

and d = 5m

and cf = 79 m

40 dB

-16dB

+35 dB

43 dB

-19dB

+32 dB

45 dB

-21 dB

+30 dB

50 dB

-26 dB

+25 dB

where r = 10 ms, jS = 5°, and D = 12 m.

Based on a framework of several assumptions,
we may expect a range of peak school target
strengths of about 50 dB whose position on the
decibel scale is determined from the value one
assumes to be the average target strength of the
individual scatterers comprising the school.

288

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

From the data presented so far (Figures 3, 4) we
may assume the most probable target strength
for all schools to be -9 dB. Further, assuming
that the "typical" school has a vertical dimension
of 12 m and that the measured target strength is
the summation of scattering strength of the indi-
vidual fish ensonified with no effects from multi-
ple scattering or attenuation, we may use Equa-
tion (6) to estimate q:

TSj

Spacing

40 dB

0.93 fish/m^

8.1 body lengths

43 dB

1.86

6.5

45 dB

2.95

5.5

50 dB

9.33

3.4

Bottom Topography

Fixed transect surveys require that the dis-
tribution of schools be independent of fixed geo-
graphic locales whose scale is smaller than tran-
sect spacing.

A cruise in March-April 1974, was designed to
test a postulated relationship between the oc-
currence of pelagic fish schools and bottom topog-
raphy. The area chosen was the Los Angeles
Bight and for the purposes of the experiment was
defined as that body of water bounded by the
southern California coast from Pt. Arguello to the
U.S. -Mexican border and seaward by a line ex-
tending south from Pt. Arguello to a point west of
San Miguel Island, thence southeast along an ex-
tension of the Santa Rosa-Cortez Ridge to a point
north of the east end of Cortez Bank, thence east
to the intersection of the shoreline and the
U.S. -Mexican border. The survey area, excluding
island masses, contains approximately 11.5 x 10^
square nautical miles of sea surface area.

The "Bight" was further divided into four
classes of bottom topography and transects de-
signed to distribute survey effort within these
zones as described below. The method used was to
delineate and compute the combined areas of the
first three categories and then assign the remain-
ing area to the fourth general zone.

Total area

%of

(nautical

survey

sampling

Bottom topography

milesV

area

effort

Banks and seamounts

547

4.8

14.4

Basins and troughs

2,946

25.9

27.4

Escarpments and canyons

467

4.1

24.1

Slopes

7,510

65.2

34.1

Combined seas and swells in excess of 7
feet prohibited sonar operations on 1 day out of
12 and somewhat altered the distribution of sur-
vey effort. A detailed breakdown of zones and ac-
tual survey effort is listed in Appendix Table 1.

Daylight sonar tracking was accomplished dur-
ing two time periods separated by 2 wk: 25-29
March, 1 April, and 15-19 April 1974. No differ-
ence in schooling behavior was detected between
the two periods and results are presented for the
total cruise time in Appendix Table 2. If an area
was surveyed and no targets were detected, a "0"
under "No. targets obs." so indicates; if an area
was not surveyed during one or both time periods
then no numbers are recorded in the appropriate
columns. "Linear nautical miles surveyed" is the
distance traversed while sonar tracking over the
designated area. The observation window (250 m
wide beginning at 200 m from the ship, and 90° to
starboard from the ship's track) is multiplied by
the linear distance traversed and divided into the
number of targets observed to obtain target den-
sity, expressed in units of targets per square
nautical mile.

The geographic names of various topographic
features are commonly accepted and can be lo-
cated on National Ocean Survey bathymetric
maps (numbers 1205N-15, 1206N-16, 1306N-19,
and 1306N-20) with the exception of the following
features informally named for the sake of con-
venience: Coronado Bank (lying immediately to
the east of Coronado Escarpment), San Diego Es-
carpment (along the west side of the San Diego
trough), Cortez Escarpment (east-northeast of
Cortez Bank), San Clemente Bank (a relatively
deep bank northeast of San Clemente Island),
Santa Rosa North and South Bank, San Nicolas
Escarpment (southeast of San Nicholas Island),
Santa Cruz Bank (south-southeast of Santa Rosa
Island), Santa Barbara Escarpment (west of
Santa Barbara Island at the southeast end of
Santa Cruz Basin), Santa Barbara Bank (north of
Santa Barbara Island), and Santa Monica Es-
carpment (along the southwest side of Santa
Monica Basin).

The data fail to support the notion that the oc-
currence of pelagic fish schools can be related to
bottom topography over which they are detected.
Mean target densities (number of targets ob-
served per square nautical mile) were calculated
for the four classes of bottom topography and al-
though these densities range from 2.98 (banks
and seamounts) to 8.23 (escarpments and can-

289

FISHERY BULLETIN: VOL. 74, NO. 2

Table 2. — A comparison of the variance in detected target densities within the cleisses
of bottom topography (zone) and between the zones. Probabihty <0.5 that there is an
other than random relationship between the four classes of bottom topography and
detected school occurrence rates (target densities).

Targets observed (no

)

Target

25 Mar.-
1 Apr.

density (targets/nmi^)

Zone

25 Mar.-
1 Apr.

15-19 Apr.

Total

15-19 Apr. Total

Banks and seamounts
Basins and troughs
Escarpments and canyons
Slopes

36
117

29
194

244

229

38
361
258
263

3.57
4.42
2.11
8.55

0.75 2.98
12.08 7.74
12.81 8.23

3.25 5.98

Sum of squares

Degrees

of freedom

Means of squares F

Within zone
Between zones

72.9765
2.9932

29
3

2.5164

0.9977 0.40

yens), an analysis of the variance would suggest
that there is no variance between the zones that
could not be explained by the existing variability
within the zones (Table 2).

HORIZONTAL SCHOOL AREA TO
BIOMASS CONVERSION FACTORS

Fish Trap Experiment

The first effort toward determining a horizontal
school area to biomass conversion factor was con-

9.1m

-70m

NETTING

l/2" STRETCH

0 104mm MONOFILAMENT

-BAMBOO

(JLEAD (0 9 kg)

Figure 5. — Diagram of an acoustically transparent trap for
ensonifying a group of fish of known size and weight.

ducted in 1970 and briefly described in the dis-
cussion following the presentation of Smith's
( 1970) paper and transcribed in the publication of
that paper.

An acoustically transparent trap (Figure 5)
was constructed and live northern anchovy en-
closed. Two groups of fish were ensonified and
their horizontal area measured. A 354-kg group
yielded a target strength within the range fre-
quently encountered while a 2,017-kg group's
target strength was well above that observed in
nature for schooling fish.

Ensonification of additional weight groups was
not possible due to the presence of predators and
attempts at visual observation of the fish aggre-
gation using a manned submersible eventually
destroyed the trap. A value of 31 kg of fish
biomass/m^ was derived from the 354-kg group
and judged to be our best estimate (Table 3). Mais
(pers. commun.) reports from his experience

Table 3. — Computation of a horizontal school area to bio-
mass conversion factor from data gathered during the fish
trap experiment (February 1970).

50-fish

sample

354-kg group

2,017-kg group

Weight

No.

% of

Total

No.

Total

PVC PIPE

class

sample

weight

(g)

fish

weight

fish

(g)

fish

(g)

33.8

11,925

119,652

68,175

681,746

31.7

7,481

112,218

42,626

639,389

25.4

4,496

89,916

25,616

512,318

3.5

496

12,390

2,824

70,595

5.6

496

19,824

2,824

112,952

Total 50

100.0

24,894 354,000

142,065 2,017,000

Surface area'

354-kg group
2Mt/m2 No./m2

2,017-kg group
Mt/m2 No./m2

11.39

0.031 2,190

0.177 12,473

'The fish are schooled in an ellipse with a major radius of 2.90 m and
a minor radius of 1.25 m (surface area 11.39 m^).
^Metric tons per square meter.

290

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

a representative anchovy school compaction
around 50 fish/m^ or a distance of two body
lengths between fish. Using a single fish weight
of 18 g and an average school thickness of 12 m
(Mais 1974), one obtains a horizontal school area
to biomass conversion factor of 8.4 kg/m^.

Charter Boat Experiment

A second experiment was designed and exe-
cuted in late summer 1974, to relate measured
school size, calculated target strength, and school
compaction. Purse seine boats were chartered to
make directed sets on fish schools first ensonified
by the acoustic system aboard the David Starr
Jordan.^ Target strength and school size were
calculated from the observation. The fishing boat
supplied information on the tonnage caught and
the portion of the school taken. Using these data,
a biomass conversion factor was calculated for
each school by dividing the total estimated school
tonnage by a circular area based on the difference
between its near and far ranges.

Fifty-two sets were judged to be the minimum
sample size necessary to distinguish between two
estimates of the portion of detectable pelagic
aggregations that are schools of northern an-
chovy. Squire (1972), using data from 6 yr of ob-
servations from several commercial air spot-
ters, reported that at least 50% of the surface
schools off southern California can be expected to
be anchovy. Mais (pers. commun.) estimates that
90% of the schools sampled by mid-water trawl
are anchovy.

Seventy-six sets were made landing 1,901 short
tons of anchovy; 63 were directed by the David
Starr Jordan and 13 directed by the State of
California's RV Alaska. Forty-nine positive data
points were tabulated from the David Starr Jor-
dan's work and eight from the Alaska.

Average target size was 119 m (as measured by
the difference between the near and far ranges on
a line perpendicular to the ship's head) with a
range from 31 to 305 m. Average peak target
strength was +5.18 dB (as calculated from peak

amplitude and range dependent losses) with a
range from -9 to +18 dB and a SD of 5.63 dB.

Practical considerations forced us to expend a
larger portion of effort on schools of larger than
average size and target strength. This circum-
stance accounts for the fourfold increase in
median target size and a 15-dB increase in mean
target strength over a sonar-generated data base
reported earlier. In addition, this sample was cho-
sen from a detected school population whose
acoustic dimensions were, in general, larger than
that experienced on previous cruises.

To facilitate the direction of sets, the observa-
tion window was increased fi:"om 250 to 500 m
wide and moved 100 m closer to the vessel. A
time-varied gain increase was also accomplished
in the receiver previous to signal display on an
oscilloscope. Either or both of these changes to
the sonar system configuration could produce cir-
cumstances under which similar data distribu-
tions would appear to be different. Point scat-
ters encountered when plotting target size versus
target strength, target strength versus horizontal
school area to biomass conversion factor, and
target size versus horizontal school area to
biomass conversion factor are too wide to detect a
relationship between these school parameters.

A distribution of horizontal school area to
biomass conversion factors is presented in Figure
6. The distribution is skewed right with an
arithmetic mean of 15.16 kg/m^. While no rela-
tionship is as yet demonstrated between indi-
vidual target strengths and horizontal school
area to biomass conversion factors, the data have
contributed to a refinement of a general conver-

®Contracts were let for a total of 104 sets assuming 50% suc-
cess rate for positive sets and a permit was secured from CF&G
to land 2,500 tons of anchovy during the experiment. A charter
agreement was written establishing criteria for the successful
bidders as minimum tonnage bid with the proceeds from any
excess tonnage, not to exceed the permit, to be given to the
State. In addition, each boat was guaranteed a fixed fee over
and above the proceeds from the landed fish.

'0.125 0.25 0.5 I 2 4 8 16 32 64 128

HORIZONTAL SCHOOL AREA TO BIOMASS CONVERSION FACTOR (kg/m^)

Figure 6. — Distribution of horizontal school area to biomass
conversion factors obtained from the charter boat experiment.

291

sion factor based previously on only one data
point.

Eight horizontal school area to biomass conver-
sion factors calculated from sets directed by the
Alaska have a range from 10.14 to 30.22 kg/m^
with a mean value of 18.42 kg/m^. The Alaska
participated in the experiment during the last
2 wk when only large schools were available in
shallow water.

AUTOMATED HYDROACOUSTIC

DATA ACQUISITION AND

PROCESSING SYSTEM

In an effort to reduce observer subjectivity in
the collection of large amounts of sonar data
necessary for the isolation of sampling errors and
biases, a decision was made to develop the capa-
bility to automatically count and measure the
horizontal dimensions of sonar targets. Peak echo
amplitude was also to be measured with the in-
tention of eventually relating it to school compac-
tion and depth.

A digital PDP8/I computer with an additional
16k memory, an analog-to-digital converter and a
teletype terminal were acquired on loan from the
Naval Undersea Center at San Diego. Using this
gear, a project was undertaken which would allow
us to do automatically what we were doing man-
ually but with the additional benefits of real-time
target strength calculation and rapid raw data
processing.

The raw data used for hand target collection is
in the form of a paper record containing a field of
parallel lines, each line being an incremental dis-
tance along the survey track. If the amplitude of
the signal is sampled during the recording of one
of these lines, at a sample rate of 750 samples/s
(velocity of sound/two-way path length), the
result is a record of the instantaneous echo
amplitude at 1-m increments along a line per-
pendicular to the survey track.

When several of these lines have been recorded,
the result is a data field which is a numerical
counterpart of the paper record. Once the word
"target" is defined numerically, the number of
targets in this field can be counted.

The numerical definitions used for this purpose
are:

Threshold (THS) = some signal amplitude
greater than the average reverberation or
noise level.

292

FISHERY BULLETIN: VOL. 74, NO. 2

Target line = at least five consecutive samples
greater than THS, preceded and followed by
five samples below THS.

Target block = two target lines which have at
least five coincident and consecutive samples
greater than THS.

Target = a target block + A^ additional coinci-
dent target lines, bounded by noise (signal
less than THS).

The threshold, for the initial program was a
predetermined constant. The five sample target
line is selected on the assumption that a 5-m
target may be the smallest significant unit. The
two line target block is selected since random or
asynchronous noise greater than THS can cause a
target line, but will rarely cause at least five
coincident samples on consecutive lines. Three
consecutive lines of data are stored in the mem-
ory of the PDP8/I computer. As each new line of
data is stored it is tested for the presence of target
lines. When a target line is found, the amplitude
of the samples is compared and the value of the
peak amplitude is stored in the first data point
location.

The newest data line is then compared with the
previous one and any occurrence of a target block
is recorded in the block register. The previous
data line is compared with the oldest data line
and, with the information in the target block reg-
ister, the following decisions are made:

1. Is the target block the beginning of a new
target? If so, assign it a number and record
its initial range, final range, and peak
amplitude in the temporary target storage
register.

2. Is the target block the entire target? If so,
store its information in the final target stor-
age field with the current time and the ship's
speed.

3. Is the target block part of a previous target?
If so, update the temporary storage infor-
mation.

4. Is the target block the end of a previous
target? If so, update the temporary informa-
tion and store in final storage.

Additional logic decisions are required if two or
more previously recorded individual targets later
merge to form a single target, or if the inverse
should occur.

There are four analog data input lines to the

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

system which are multiplexed and sampled at
appropriate times by the analog-to-digital con-
verter. These are:

The start pulse — the trigger pulse for the sonar
transmitter.

The sonar signal — the 1,000 cycle band width
detected video from the sonar receiver.

The ship's speed — a DC voltage from the ship's
log proportional to speed.

The hour mark — a pulse from the ship's preci-
sion simplex clock system occurring at the
end of each hour.

The start pulse initiates the program, which
then counts 200 sample times before recording
data. Two hundred fifty samples are then taken
between 200 and 450 m, to be operated on by the
program as previously described. A running
count of the number of start pulses occurring
after the beginning of each new hour is kept and
used as a time base for all events recorded during
that hour. During data reduction, this count is
divided into 60 min and used to provide absolute
time data.

The ship's speed is recorded with each target,
and may be used to calculate the area surveyed. It
is used in the data collection program to deter-
mine when a hydrographic and/or biological sta-
tion has been reached and to suspend data record-
ing while on station; start pulses continue to be
counted, however, thus the time at the beginning
and end of the station is recorded.

In shipboard operation, the system requires no
attendance. Prior to leaving the dock, the com-
puter is started, and the hour counter is preset to
the current time. The sonar system is then
started and may be left in operation 24 h a day or
turned off at night. In either case, the data collec-
tion program will begin sampling automatically
at 0800 each morning and continue until 1600
each afternoon, except while on station. There are
six memory storage fields in the PDP8/I of 4,096
words each. One field is used for programming
and temporary data storage. The other five fields
provide final storage for 3,300 targets, at six data
words per target. At the end of the day ( 1600 h)
the data collection program in field zero is re-
placed by a general computational program used
in the PDP8/I called FOCAL. This program
change is accomplished automatically from a pre-
recorded magnetic tape cartridge. With FOCAL
programming, the stored target data is now re-

duced, summarized, and dumped onto periph-
eral mass storage capable of holding the entire
cruise.

When the output is finished, the collection
program is reread into field zero, and the com-
puter waits for 0800 h the following morning to
again begin data recording.

Field testing of this system was conducted in
July 1974, by comparing computer listings of
events with the corresponding wet paper records.
The system proved to have a greater resolution
than was felt necessary and the criteria for a
target block changed to two coincident and con-
secutive samples above threshold. Ten samples
below threshold rather than five were judged
adequate to terminate a target on any given line.
A variable threshold based on an integrated
value of volume reverberation is being developed.

The system was field tested under a wide vari-
ety of conditions and judged satisfactory for our
requirements. Figure 7 describes a cumulative

20 40 60 80 100 120 140
MEAN CLASS DIAMETER (meters)

Figure 7. — Cumulative frequency diagram of school count
and horizontal school area from a sample taken during the
field test of an automated sonar system in July 1974.

frequency diagram for school count and horizon-
tal school area. A median school size of 30 m
agrees with data from previous cruises.

293

FISHERY BULLETIN; VOL. 74, NO. 2

AUTOMATED SONAR SURVEY

An automated sonar survey of the Los Angeles
Bight was accompUshed during the last 2 wk of
the charter boat cruise. A 721-nautical-mile track
(Figure 8) was transected two times providing a
3.4% areal sample of the ll,500-mile2 Bight.
Each track (1.7% sample) was processed as a
separate survey.

Appendix Table 3 lists target counts on tracks 1
and 2 by target size and mid-range. Target size
refers to the maximum dimension normal to the
ship's track and is calculated from the difference
between the leading and trailing edges of the
echo envelope corrected for the pulse length ( 15 m
at 10 ms pulse length). The first mode, common to
both tracks at a target diameter of 30 m, is consis-
tent with earlier data collected by NMFS (approx-
imately 4,500 targets) and CF&G (approximately
23,000 targets). A second mode occurring at a
school diameter of 250 m is also common to both
tracks. This mode has not been seen before or
during any season in any year since sonar ac-
tivities were initiated off southern California. An
explanation for the mode, other than the reflec-
tion of an optimum school size, is that it may be a
bottom reverberation mode particular to the ob-
servation window used on the survey.

Bottom reverberation, as logged by the system,
was collected for 2 h over water depths of ap-
proximately 100 m during the cruise. Distribu-
tions of target size, midrange, and target strength
are shown in Appendix Table 4. Notable are two

32°30'

size modes at 50 and 225 m, an optimum mid-
range of 450 m, and an average target strength of
+ 5dB.

Targets contributing to the 250-m size class
mode have a midrange mode of approximately
450 m for both tracks 1 and 2. Average target
strength was -h? dB for the subsample. This in-
formation reinforces the theory that the 250-m
size class mode is caused by false targets caused
in turn by bottom reverberation. Changes in the
sonar system operating parameters (i.e., the en-
largement of the observation window and the ad-
dition of a time gain circuit) are assumed to be
responsible for the variation in system response.
These changes were made to facilitate the fish
biomass work and will not be in effect during the
sonar surveys to be conducted on a series of
California Cooperative Oceanic Fisheries Inves-
tigations cruises beginning in November 1974.
Operating procedures will be the same as used for
the initial field of testing of the automated hydro-
acoustic data acquisition and processing system.

Since those targets which begin or end beyond
the observation band are not counted, an edge
bias exists which is a function of the target size
and the extent of the observation window. Fre-
quencies within target size class intervals were
corrected for edge bias by the following formula:

N'^ = N,

500

I20°30' I20°00 II9°30 IIS^OO 1I8°30 II8°00' II7°30' II7°00'

Figure 8. — Los Angeles Bight including a 721-mile sonar sur-
vey track transected twice, 17-26 September 1974.

where Nd= frequency of observation within a
given size class
N'ci= frequency corrected for edge bias
d = mean class diameter.

The largest school size corrected for edge bias
was 160 m (target size distributions fi^om previ-
ous cruises, CF&G and NMFS, indicate that 160
m includes the 99th percentile). Table 4 lists ob-
served frequencies, edge corrected frequencies,
and horizontal school area contributions for size
classes up to a maximum mean class diameter of
160 m.

The total detected school area was 2.6 x 10^ m^
for track 1 and 1.4 x 10^ m^ for track 2. Integrat-
ing over the entire survey area by simple propor-
tion, assuming no stratification, and using a con-
version factor of 15.16 kg/m^, biomass estimates
of pelagic schooling fish in the Los Angeles Bight
were calculated at 2.30 x 10^ metric tons and
1.23 x 10^ metric tons for tracks 1 and 2, respec-

294

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

Table 4. — Observed frequencies, edge corrected frequencies, and horizontal school
area contributions for size classes (metric tons, mt) up to a maximum of 160 m school
diameter.

Class

limits

Cum.

(mt)

Mark

(f)

%IA/'

%A/'

(mt)2

%1N'A

%N'A

Track 1

-5

35.000

5.171

0.000

75.510

11.156

16.327

5,930,557

0.229

89.583

13.236

29.563

28,143.434

1.091

1.321

94.680

13.989

43.553

66,925.949

2.594

3.916

73.913

10.920

54.473

92,881.869

3.601

7.517

52.222

7.715

62.189

102,538 093

3.975

11.493

40.909

6.044

68.234

115,667.729

4.484

15.978

34.883

5.154

73.388

134,248.290

5.205

21.183

25.000

3.693

77.081

125,663.706

4.872

26.056

25.609

3.783

80.865

162,922.228

6.317

32.373

105

100

22.500

3.324

84.190

176,714.586

6.851

39.225

105

115

110

15.384

2.273

86.463

146,204.888

5.668

44.894

115

125

120

19.736

2.916

89.379

223,218.425

8.655

53.549

125

135

130

17.567

2,595

91.975

233,178.346

9.041

62.590

135

145

140

16.666

2.462

94.437

256,563.400

9.947

72.538

145

155

150

20.000

2,955

97.392

353,429.173

13.703

86.242

155

165

160

12
603

17.647

2.607

99 999

354,815.170
2,579,045.851

13.757

99.999

Tota

676.815

Track
-5

33.000

7.902

0.000

46.938

11.240

19.143

3,686,562

0.267

59.375

14.218

33 362

18,653,206

1.353

1.621

53 191

12.738

46.100

37,598.848

2729

4.350

42391

10.151

56.252

53,270.484

3,866

8.217

43.333

10.377

66.629

85,084,801

6.175

14.393

27.272

6.531

73.160

77,111.819

5.597

19.990

27.906

6.683

79.843

107,398.632

7.795

27 786

9.523

2.280

82.124

47,871.888

3.474

31.261

9.756

2.336

84.461

62,065,610

4.505

35.766

105

100

10.000

2.394

86.855

78,539.816

5.700

41.467

105

115

110

10.256

2.456

89.312

97,469.925

7.074

48.542

115

125

120

7.894

1.890

91.202

89.287.370

6,481

55.023

125

135

130

6.756

1.618

92.820

89.683.979

6,509

61.533

135

145

140

12.500

2.993

95.814

192,422.550

13.967

75.500

145

155

150

5.714

1.368

97.182

100,979.763

7.329

82.830

155

165

160

1 1 .764

2.817

99.999

236.543.446

17.169

99.999

Total

376

417.576

1,377,668.706

tively. Identification of the fish is not yet possible
on a routine basis. However, it is assumed that
the majority of schoohng fish in the Los Angeles
Bight are northern anchovy (Smith 1972; Squire
1972; Mais 1974).

DISCUSSION

It is our impression that the ultimate value of
sonar mapping is its potential to reconstruct
geographic patterns of school distributions at a
moderate cost of time both in data collection and
data reduction. However, before this potential
can be fully realized, several problems must be
recognized, investigated, and placed in proper
perspective.

With regard to counting and sizing targets:

1. An edge bias has been described which will
be present with any sonar system designed

to count and size schools. The determination
of effective detection ranges establishes a
finite observation band. Larger schools tend
to be undersampled relative to smaller
schools; in terms of school area the bias may
be significant.

Increasing the observation band would tend
to reduce the effect of edge bias. However,
the effects of target size and target strength
on maximum ranges of detection should be
investigated before defining the observation
band. Undersampling small schools may be
acceptable when considering their area con-
tribution.

Effective detection ranges may also be lim-
ited by inhomogeneities in the medium
caused by short-period internal waves.
Smith (see footnote 3) investigated this
phenomenon and suggested the only prac-
tical solution is a statistical approach

295

FISHERY BULLETIN: VOL. 74, NO. 2

whereby the number of sound velocity
profiles taken in an area-time stratum
would be limited to the number of samples
necessary to reduce the standard error to a
uniform value for all strata. A probability of
detection diagram could then be constructed
from the ray trace analyses and target
counts corrected by range. We have not so
far considered these effects in our area of
operation, however, the implication of
undersampling should be investigated when
designing a serious stock assessment survey
using sonar.

4. Diurnal and seasonal variations in school
sizes can be expected. In order to properly
evaluate their affect on a stock assessment
scheme the period and amplitude of these
variations must be measured. The collection
of a data base sufficient in size to detail
these changes, as well as geographic dis-
tribution patterns by season, was the pri-
mary motivation in designing an automated
data collection system.

5. While it appears that influences of bottom
topography on school distribution may be
neglected, there is no reason to expect areal
distributions to be uniform. In fact, there is
evidence from aerial reconnaissance, sonar
transects obtained at long ranges (2,500 m),
and fishermen that fish schools may be dis-
tributed in a highly contagious fashion simi-
lar to the distributions of fish eggs and lar-
vae. In our opinion, this is a most important
consideration in arriving at an optimum
survey design. Smith^° and MacCall^^ have
approached the problem by direct measure-
ment and simulation modeling and suggest
a transect spacing of 15 miles as adequate to
reconstruct groups of anchovy schools off
southern California.

e
in

6. Holliday (1972, 1974) demonstrated th
feasibility of sizing individual fish within
schools and provided information which
would aid in species identification. A de-

'"Smith, P. E. 1975. Precision of sonar mapping for pelagic
fish assessment in the Cahfomia Current area. SWFC Adminis-
trative Report No. LJ-75-60. Southwest Fisheries Center,
NMFS, NOAA, La Jolla, CA 92038.

'^MacCall, A. 1975. Anchovy population survey simulation.
Contribution No. 4, CalCOFI Anchovy Workshop, July 1975.
Document on hand at the Southwest Fisheries Center, NMFS,
NOAA, La Jolla, CA 92038.

velopment of these techniques as practical
additions to a sonar survey system would
reduce a presently loosely quantified factor,
i.e., the percent of detected schools which
can be expected to be the target species of a
survey.

With regard to school target strength:

1. The target strength of an individual fish is
an essential element in interpreting the
measured target strength of a school. At the
frequencies commonly used for sonar map-
ping we can expect interference of energy
reflected fi'om the various scattering parts of
a fish. This makes the target strength of a
fish strongly aspect dependent. Unfortu-
nately there is presently no method of
acoustically determining the aspect of indi-
viduals in a school and hence their effective
target strength. As such, the maximum dor-
sal or side aspect target strength is gener-
ally an overestimate and the use of these
values in interpreting school target
strengths results in an underestimate of the
number of individual scatterers.

2. We may also expect multiple scattering,
shadowing, and attenuation within a school.
These effects may tend to reduce or enhance
the target strength of a school and cannot be
evaluated until we know the effective con-
tribution of the fish taken as individual
scatterers. Love (1971) stated that the
quantification of a fish school using its
target strength is possible because the
target strength of a school depends on the
average size, number, distribution, and as-
pect of the individuals in the school. If the
effects of the distribution offish in space and
their aspect can be removed, we may as-
sume an average size and estimate their
numbers.

3. We have assumed spherical spreading losses
which may only be expected in a three-di-
mensional homogeneous fluid. In fact, the
upper mixed layer, in which we operate our
sonar, is characteristically bounded by den-
sity discontinuities which reflect and re-
fract sound waves. The actual path of
transmitted and target-reflected sound
waves may not be direct as implied in the
use of spherical transmission losses.

296

HEWITT ET AL.: DEVELOPMENT AND USE OF SONAR MAPPING

Continuing development of acoustic stock as-
sessment techniques rests on the comparison of
measurements and the best available theoretical
models for target strength and fish school bio-
mass. Improved instrumentation, particularly
data logging and processing equipment will make
the comparison more timely and useful. The
existing system will be used seasonally over the
entire California Current survey area (about
200,000 nautical miles^) in 1975. It is intended
that the data base thus furnished will allow a
balanced approach to such biological problems as
migration and patchiness of fish schools in the
context of better theory and instrumentation.

ACKNOWLEDGMENTS

We express our gratitude to the Marine Re-
search Committee of California for their support
and encouragement. The charter boat experiment
could not have been conducted without the assis-
tance of Peter Fletcher, Chairman of the Califor-
nia Fish and Game Commission, and Ray Arnett,
Director of the California Department of Fish and
Game. We extend our thanks to John Zankich
and his able crew aboard the FV Southern Ex-
plorer and to Peter Lipanovich and crew on the
FV Rigel for their exemplary effort. Robert Vent,
Isaac (Ed) Davies, and William Batzler of the
Naval Undersea Center provided equipment and
advice on a problem of common interest, i.e., the
acoustic characteristics of fish schools. Kenneth
Mais and his colleagues with the California De-
partment of Fish and Game Sea Survey were a
valuable source of experience. John Graves, a
graduate student at the University of California,
San Diego, and Gayle Turner, a NCAA Junior
Fellow, assisted in the field work. We are in-
debted to D. V. Holliday for his considerable assis-
tance in editing the manuscript.

LITERATURE CITED

CUSHING, D. H.

1973. The detection offish. Pergamon Press Ltd., Oxford,
200 p.

Forbes, S. T., and O. Nakken.

1972. Manual of methods for fisheries resource survey and
appraisal, Part 2: The use of acoustic instruments for fish

detection and abundance estimation. FAO Manuals in
Fisheries Science, 138 p.
Haslett, R. W. G.

1965. Acoustic backscattering cross sections of fish at
three frequencies and their representation on a universal
graph. Br. J. Appl. Phys. 16:1143-1150.

Holliday, D. v.

1972. Resonance structure in echoes fi"om schooled pelagic

fish. J. Acoust. Soc. Am. 51:1322-1332.
1974. Doppler structure in echoes from schools of pelagic

fish. J. Acoust. Soc. Am. 55:1313-1322.
LOVE, R. H.

1971. Measurements of fish target strength; A review.
Fish. Bull,, U.S. 69:703-715.

MAIS, K. F.

1974. Pelagic fish surveys in the California Cur-
rent. Calif Fish Game, Fish Bull. 162, 79 p.
McCartney, B. S., and a. R. Stubbs.

1970. Measurements of the target strength offish in dorsal
aspect, including swimbladder resonance. In G. B. Far-
quhar (editor), Proc. International Symposium on Biolog-
ical Sound Scattering in the Ocean, p. 180-211. Maury
Center for Ocean Science, Dep. Navy, Washington, D.C.
MCCLENDON, R. L

1968. Detection of fish schools by sonar (eastern tropical
Pacific, July-November 1967). Commer. Fish. Rev.
30(4):26-29.
SHIBATA, K.

1970. Study on details of ultrasonic reflection from indi-
vidual fish. Bull. Fac. Fish. Nagasaki Univ. 29:1-82.
Smith, O. R.

1947. The location of sardine schools by super-sonic echo-
ranging. Commer. Fish. Rev. 9(1): 1-6.

Smith, o. R., and E. H. ahlstrom.

1948. Echo-ranging for fish schools and observations on
temperature and plankton in waters off central Califor-
nia in the spring of 1946. U.S. Fish Wildl. Serv., Spec.
Sci. Rep. 44, 43 p.

Smith, p. E.

1970. The horizontal dimensions and abundance of fish
schools in the upper mixed layer as measured by so-
nar. In G. B. Farquhar (editor), Proc. International
Symposium on Biological Sound Scattering in the Ocean,
p. 563-591. Maury Center for Ocean Science, Dep. Navy,
Washington, D.C.

1972. The increase in spawning biomass of northern an-
chovy, En^rauZis mordax. Fish. Bull., U.S. 70:849-874.

Squire, J. L., Jr.

1972. Apparent abundance of some pelagic marine fishes

off the southern and central California coast as surveyed

by an airborne monitoring program. Fish. Bull., U.S.

70:1005-1019.
URETSKY, J. L.

1963. The acoustical properties of compacted schools

of fish. SIO Ref. 63-21, Scripps Inst. Oceanogr., Univ.

Calif., 3 p.
WESTON, D. E.

1967. Sound propagation in the presence of bladder

fish. In V. M. Albers (editor), Underwater acoustics, Vol.

2, p. 55-88. Plenum Press, N.Y.

297

FISHERY BULLETIN: VOL. 74, NO. 2

Appendix Table l.— Topographic breakdown of Los Angeles Bight by four
classes of bottom configuration (zone) and distribution of design and actual sam-
pling effort.

Zone and name

Area

(nautical

mlle^)

Total
area

(%)

Sampling effort

Design
(%)

Actual
(%)

Banks:
Thirtymile Bank
Fortymile Bank
Tanner Bank
Osborn Bank
San Clemente Bank
San Nicolas Bank
Santa Rosa N. Bank
Santa Rosa S. Bank
Coronado Bank
Santa Barbara Bank
Santa Cruz Bank
Lasuen Seamount

Total

Basins:
San Clemente Basin
Catalina Basin
San Nicolas Basin
San Diego Trough
San Pedro Basin
Santa Monica Basin
Santa Cruz Basin
Santa Barbara Basin

Total

Escarpments and canyons:
Coronado Escarpment and Canyon
Catalina Escarpment
San Clemente Escarpment
San Diego Escarpment
San Pedro Escarpment and Redondo Canyon
San Nicolas Escarpment
Santa Cruz Escarpment and Canyon
Santa Monica Escarpment
Santa Barbara Escarpment
Cortez Escarpment

Total

Slopes

44.8
39.6
50.2
13.4
37.3
125.4
17.7
34.0
19.1
72.7
79.3
13.2

546.7

91.7

540.8

497.3

264.2

145.6

490.3

213.2

733.2

1,976.3

37.3

99.5

97.4

38.9

33.4

34.2

75.4

15.5

23.3

12.4

467.3

4.8

14.4

9.4

25.9

27.4

34.4

4.1
65.2

24.1
34.1

23.7
32.5

298

HEWITT ET AL.; DEVELOPMENT AND USE OF SONAR MAPPING

Appendix Table 2. — Detected targets and target densities for four classes of bottom
topography (zone) in the Los Angeles Bight.

Zone and name

No.

targets

obs.

Linear

nautical miles

surveyed

Target

density

(targets/nmi^)

Banks and seamounts:
Thirtymile Bank
Fortymile Bank
Tanner Bank
Osborn Bank
San Clemente Bank
San Nicolas Bank
Santa Rosa Bank
Santa Rosa S. Bank
Coronado Bank
Santa Barbara Bank
Santa Cruz Bank
Lasuen Seamount

Inx excluding zero values

S|nx

Basins and trougtis:
San Clemente Basin
Catalina Basin
San Nicolas Basin
San Diego Trough
San Pedro Basin
Santa Monica Basin
Santa Cruz Basin
Santa Barbara Basin

In X excluding zero values
•S|nx

In X excluding zero values

S|nx

Slopes

3
18

12.40
16.99

6.19

1.79
7.85

1.20

Inx excluding zero values
S|nx

8.0

0.75

19.8

21.59

9.59

10.81

0.9835
1.1758

22.60

14.09

83.99

7.41

135

44.19

22.63

58.19

11.84

21.40

1.38

31.59

0.14

23.4

37.4

1 .4325
1.9237

19,40

0.38

28.19

3.94

40.39

0.55

172

51.53

24.73

19.80

9.35

14.99

18.78

33.18

0.89

11,00

12.4

6.79

6.55

1.2431
1.6145

9.40

20.19

7.71

6.00

680

5,00

7.99

22,80

1.62

7.80

7.20

4.16

12.39

5.38

3069

15.69

16.59

20.54

9.00

18.93

6.20

25.09

15.20

3.41

4.00

37.04

3.60

3.20

55.00

0.13

5.40

11.20

14.55

27.20

17.20

0.86

15.51

8.12

1.7850
1.5365

299

FISHERY BULLETIN: VOL. 74, NO. 2

Appendix Table 3. — Target counts by size and midrange detected on an auto-
mated survey of the Los Angeles Bight (tracks 1 and 2) during September 1974.

Size

100-

150-

200-

250-

300-

350-

400-

450-

500-

550-

(m)

150

200

250

300

350

400

450

500

550

600

Total

6- 15

120

16- 25

143

26- 35

139

36- 45

107

46- 55

56- 65

66- 75

76- 85

86- 95

96-105

106-115

116-125

126-135

136-145

146-155

156-165

177-175

176-185

186-195

196-205

206-215

216-225

226-235

236-245

246-255

256-265

266-275

276-285

286-295

296-305

306-315

316-325

326-335

336-345

346-355

356-365

366-375

376-385

386-395

Total

103

171

169

402

270

223

1,499

Appendix Table 4.

-Bottom reverberation by detected size, midrange, and target strength from data collected during 2 h in 100

fathoms on 7 September 1974.

Item

Mark

Relative %

Item

Mark

Relative %

Size

Midrange

25 m

5.4

14.3

6.5

100

5.4

125

4.8

150

6.0

175

10.7

200

15.5

225

14.3

250

13.7

275

3.6

300

340 m

3.0

360

4.1

380

3.6

400

5.3

420

14.8

440

26.0

460

24.9

Midrange

480

8.9

500

4.1

520

4.1

540

0.6

560

0.6

Target strength

-2dB

0.6

-1

0.6

2.4

6.5

10.7

• 3

12.4

14.8

16.6

14.8

3.6

5.3

4.1

4.7

3.0

300

ECONOMIC AND FINANCIAL ANALYSIS OF
INCREASING COSTS IN THE GULF SHRIMP FLEET ^ 2

Wade L. Griffin, Newton J. Wardlaw, and John P. Nichols^

ABSTRACT

The 115 Gulf of Mexico shrimp vessels used in this study were grouped into classes I (larger vessels)
through V (smaller vessels) based on their type of construction, length of keel, and index of effort. In
1973, class 11 vessels were the only vessels able to register a positive return to owner's labor and
management, $560; the other four classes registered negative returns. The payback period occurred
during the eighth year due to the sale of the vessels in classes II, III, and V, whereas payback did not
occur for classes I and IV. A positive rate of retvu-n on investment was experienced by the vessels in
classes II, III, and V in the amount of 13.21, 2.65, and 2.63%, respectively. The internal rate of return
on investment was negative for vessels in classes I and IV.

Input prices increased some 20% from 1973 to 1974 whereas production remained approximately
constant and ex-vessel shrimp prices were lower. Thus none of the classes of vessels would have
experienced a break-even cash flow for 1974. Increasing input cost another 10% above the 1974 level,
and assuming normal production, the average vessel in class 11 seems to be operating at a better than
a break-even level in 1975 assuming ex-vessel shrimp prices remaining constant at 1973 levels.
Classes I, HI, IV, and V experienced less than break-even cash flows under the same conditions in
1975.

The U.S. economy has faced some strong buffet-
ing in recent years. In spite of temporary wage
and price controls and other efforts by the admin-
istration, inflation has continued to be a major
problem for most sectors. The percentage in-
creases in the wholesale price index (including
all commodities) were 4.2% from 1971 to 1972,
13.1% from 1972 to 1973, and approximately 20%
from 1973 to 1974 (Board of Governors of the
Federal Reserve System 1974). Since inflation
can occur at different rates for different products,
profit and loss positions in almost every sector or
industry in the economy have been affected. Of
particular interest to shrimp vessel owners are
changes in the price for basic inputs used in the
shrimp industry: the price index for fuel, which
accounted for approximately 25%of variable costs
of shrimp production in 1971 (excluding crew
shares) (Hayenga et al. 1974) increased 76% from
December 1971 to December 1973; and the price

'Technical Article No. 11534 of the Texas Agricultural
Experiment Station.

^The work upon which this publication is based was
supported by the U.S. Department of Commerce, NOAA,
National Marine Fisheries Service under contract number
03-4-042-18, and partially supported through Institutional
Grant 04-3-158-18 to Texas A&M University by the National
Oceanic and Atmospheric Administration's Office of Sea Grant,
U.S. Department of Commerce.

^Department of Agriculture Economics, Texas Agricultural
Experiment Station, Texas A&M University, College Station,
TX 77843.

Manuscript accepted December 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

index for lumber, metals, and machinery and
equipment (inputs used in the construction of
shrimp vessels) jumped 46.5, 19.2, and 7.9%,
respectively, during the same period (Board of
Governors of the Federal Reserve System 1973).

With regard to prices and production in the
Gulf States, in 1973 ex-vessel shrimp prices
increased 33% from the 1972 figures, but land-
ings were off from the 1972 levels by 21%
(United States Department of Commerce 1974).

Due in part to the economic climate, vessel
owners, managers, financial institutions, and
marine resource researchers have come to rely
heavily upon cost and return data in analyzing
investment, financing, and profitability alterna-
tives within the Gulf shrimp industry. But a
classification problem exists because of the wide
range of combinations of vessel size, construc-
tion, power, and fishing capability within the
Gulf shrimp fleet and the wide range of variable
costs, fixed costs, investment requirements, and
profitability associated with the various vessel
configurations. It is the purpose of this paper to
investigate, for different vessel classes, the
profitability of investing in and operating a
vessel in the Gulf shrimp fleet based on data
collected for the 1973 calendar year, and then
with the data adjusted to estimated 1974 and
1975 levels.

301

FISHERY BULLETIN: VOL. 74, NO. 2

METHODOLOGY

Standard techniques of cost and return, cash
flow, and break-even analysis were used in this
study. A budget-generating computer program
was established to assimilate and report the data
according to each of the desired vessel configura-
tions, in the form of total costs and returns
budgets, unit costs and returns budgets, and
projected cash-flow budgets.

The vessels were classified in terms of their
average costs per pound of shrimp landed. An
average cost equation was estimated using
regression analysis with construction, keel
length (U.S. Coast Guard registry), and effort
index'* as dummy variables. Vessels included in
the sample were constructed of either wood or
steel. Grouping of vessels according to keel
length and effort index for use as dummy
variables in the regression analysis was based on
a natural frequency distribution of the vessels in
the sample.

It must be stressed here that this method of
classification is simply a means to group the
vessels for the purpose of analysis and is not
necessarily a criterion for evaluation of the
performance of the different classes. Performance
or profit depends not only upon unit cost but also
upon unit price. Even though one class of vessels
may have a higher average cost curve for a given
level and type of shrimp produced, it may not
necessarily produce less profit. Therefore, while
the product produced may be homogeneous with
respect to cost of production, it may be hetero-
geneous with respect to price.

DATA DESCRIPTION

Data Collection and Vessel Description

The cost and return and financial data used in
this study were gathered by personal interview
with shrimp vessel owners and/or managers
operating fi-om ports in Florida, Mississippi, and

Texas. Additional financial information was ob-
tained from officials of various lending institu-
tions which engage in shrimp vessel financing.
All data were for the period covering the calen-
dar year 1973.

The original sample for this study consisted of
126 vessels. However, due to incomplete data,
only 115 vessels were used in the analysis.
Vessels in the sample were constructed of wood
and steel, with keel lengths of from 45 to 78 feet,
and from 104 to 777 horsepower. The ages of the
vessels ranged from 1 to 36 yr.

Costs and Returns and
Cash Flow Data

Variable cost items were separated into vari-
able costs not directly proportional to catch: ice;
fuel; nets, supplies, and groceries; repair and
maintenance; and variable costs directly propor-
tional to catch: crew shares, payroll taxes, and
packing charges. Actual variable cost data re-
ported by the vessel owoiers were used except for
crew shares, payroll taxes, and packing charges,
which were determined on the basis of reported
pounds landed and gross revenues. Vessel owners
paid their captains and/or crew on the basis of a
percentage of pounds landed. This percentage
ranged from 30% in the eastern Gulf to 40% in
the western Gulf. Thirty-five percent was the
average share paid and is used in the analysis.

Fixed cost items were separated into: insur-
ance, depreciation, overhead, interest, and oppor-
tunity cost (required return to equity capital).
Fixed charges for insurance and overhead are re-
ported data. Charges relating directly to invest-
ment— depreciation, interest, required return on
equity capital for costs and returns, and principal
and interest for cash flow budgets — were stan-
dardized in terms of 1973 dollars in order to make
valid comparisons. Since most of the vessels in-
cluded in the sample were purchased new, vessel
owners (some of which were shipbuilders) were
asked to estimate the replacement value of their

■The effort index is defined as the amount of fishing power
that a vessel can exert in a day fished relative to that of a
standard vessel. The value for the effort index for each vessel
was calculated using the formula:

EI, =

(HP), 01385 (LFR), 0-4064
(38)0.1385 (14.6)0.4064

where EI, = effort index for vessel i, (HP), = horsepower for
vessel /, (LFR), = sum of the lengths of the footropes
measured in yards for vessel (, (38) = average horsepower of
the smallest class of vessels operating in the Gulf from 1962 to
1971, and 14.6 = average net size measured in yards of
footrope used by the smallest class of vessels for the same
period [Griffin, W. L., M. L. Cross, R. D. Lacewell, and J. P.
Nichols. 1973. Effort index for vessels in the Gulf of Mexico
shrimp fleet. (Unpubl. rep. to NMFS, contract no. 03-3-042-19
with the Tex. Agric. Exp. Stn., Tex. A&M Univ.].

302

GRIFFIN ET AL.: ECONOMIC AND FINANCIAL ANALYSIS

vessels in 1973 prices. Depreciation charges were
calculated using the straight-line method, based
on the estimated 1973 replacement value for each
vessel, and using an 8-yr depreciable life with
35% book salvage value. For the amortization
schedule, the same 1973 equivalent new vessel
costs were used, with 67% of the cost financed at
a 9% interest rate, for 8 yr, and with 12 equally
amortized payments per year. These terms were
found to be representative for 1973 through inter-
views with officers of financial institutions which
engage in shrimp vessel financing. The specific
amount of interest reported in each costs and
returns budget is for the fifth year of the amorti-
zation schedule since the majority of the vessels
in the sample taken were from 3 to 6 yr old.

Required return to equity capital is economic
rather than financial in concept and is an
attempt to place a value on the opportunity cost
of the equity capital committed to an investment.
At the time an owner invests in a shrimp vessel
he has several alternative investments available
with various rates of return associated with each.
Theoretically these different rates of return are
representative of the relative risks associated
with each — that is, risk and return vary directly.
Because the alternative investment opportuni-
ties are different for each owner, in the interests
of standardization the rate of interest charged by
financial institutions for shrimp vessel financing
(9%) was assumed to be the highest alternative
rate available to the owners for an investment of
equivalent risk and can be adjusted by an indi-
vidual owner to reflect his own investment
alternatives.

A note of explanation is necessary concerning
the cash flow budgets and cash flow analysis.
Terminal vessel value (sale value) and holding
period were established by asking each vessel
owner to estimate, in 1973 dollars, what that
same vessel would be worth as a used vessel if he
had held it for the number of years that he
customarily fishes a new vessel. Respondents
indicated they fished a new vessel from 3 to 15
yr, with 8 yr being the most frequent response,
and that even in periods of relative price stability
an 8-yr-old shrimp vessel is worth approximately
65% of its original cost. Furthermore, that
difference between the 35% book value for
depreciation purposes and the 65% terminal
value is evidenced by the fi*equency of income
taxes levied on vessel owners for depreciation
recapture at the time of replacement. For those

reasons, an 8-yr holding period and a 65%
terminal value were used in the cash flow
budgets.

RESULTS

Classification of Vessels

Vessels were grouped according to construc-
tion, keel length, and effort index (Table 1). All
vessels in the sample were either wood or steel.
Vessels were divided into three keel length inter-
vals: 45-62 feet, 63-69 feet, and 78-80 feet. The
range of effort indices was divided into three
intervals: 1.64-1.89 units, 1.90-2.19 units, and
2.20-2.51 units. Using these groupings for clas-
sification, 12 combinations were possible and the
vessels in the sample fell into 9 of those possible
combinations (See Appendix).

Predicted average cost values for the 115
vessels were plotted and vessels were classed into
five general categories as shown in Table 1,
where class I is the highest cost curve and class V
is the lowest. Classes I and II, the two highest
cost curves, consist entirely of steel vessels
whereas classes III, IV, and V consist entirely of
wooden vessels. The position of the average cost
curves seem to be related to vessel length for
each type construction except for class IV which
includes two length intervals.

These results are not surprising. Previous
research by Nichols and Griffin (1974) indicated
that smaller, less powerful wooden vessels can
produce a given quantity of shrimp at a lower
cost than can a larger, more powerful steel vessel:
As a matter of fact, their research showed that a
50% reduction in total effort exerted by the
shrimp fleet would only reduce total catch by
about 10%.

For the 4-yr period, 1962-65, the average vessel
exerted about 1. 16 units of effort in a day fished

Table l. — Classification of Gulf of Mexico shrimp vessels,
based on construction, keel length, and effort index from a
sample of 115 vessels, 1973.

Vessel

Keel length

Effort index

class

Construction

(feet)

(units)

Steel

70-78

1.90-2.19
2.20-2.51

Steel

63-69

1.90-2.19
2.20-2.51

III

Wood

63-69

1.90-2.19
2.20-2.51

Wood

45-62

1.90-2.19

63-69

1.64-1.89

Wood

45-62

1.64-1.89

303

FISHERY BULLETIN: VOL. 74, NO. 2

and the annual landings per vessel were 31,700
pounds of shrimp (heads-off). However, in the 4-
yr period, 1970-73, the average vessel exerted
about 1.68 units of effort in a day fished and the
annual landings were only 28,900 pounds (heads-
off). The average length of the vessel operating in
the Gulf also increased over time (Nichols and
Griffin 1975). Thus, as additional effort has been
added to the Gulf of Mexico shrimp fishery — by
increasing the number of vessels and/or the
average size of the vessel — the total pounds
landed have been divided between more and
more units of effort.

From these figures it is apparent that the
average Gulf of Mexico shrimp vessel has been
increasing in size and relative fishing power and
the annual landings per vessel have declined.
Due to the lower investment levels and lower
operating costs of smaller, less powerful vessels it
follows that those smaller vessels could produce a
pound of shrimp at a lower unit cost than could a
larger, more powerful vessel if both were fishing
the same or equally abundant fishing grounds.
However, two distinctions and/or disadvantages
of the smaller vessels compared to larger vessels
must be noted here. First, as discussed earlier,
shrimp is not a homogeneous product, and larger
shrimp command higher ex- vessel prices than do
smaller shrimp. Because the larger shrimp are
usually associated with deeper waters, farther
out in the Gulf, a smaller vessel with less
capacity both for deepwater trawling and for
holding fuel and shrimp is at a disadvantage
compared with a larger vessel because of that
depth and distance from shore.

The second distinction, and associated with the
first, is the fact that larger vessels are better able
to operate in and cope with rougher seas and the
frequent storms in the Gulf than are smaller
vessels. Therefore, the smaller vessels would
either be forced to trawl closer to shore for
smaller, less valuable shrimp, or for a given
period of time in the deeper fishing grounds with
typical weather conditions, the smaller vessels
would not be able to realize as many actual
fishing hours as a larger vessel operating in the
same waters during the same period of time.

Comparison of Classes

Table 2 shows a summary of the costs and
returns, equity requirements, payback period,
and internal rate of return for the five classes of
shrimp vessels operating in the Gulf of Mexico in
1973 (a more detailed break down of cost is
available from the authors). Class I vessels
received the highest price per pound, $2.03, for
the shrimp landed but produced 5,500 pounds
less shrimp than the smaller class II vessels.
Class I vessels also had the highest levels of
variable costs not proportional to catch, $45,152,
the highest fixed costs, $31,906, and the highest
total costs, $108,291, of any of the general classes.
These cost relationships were to be expected
since the larger steel vessels should have the
highest initial investment requirements and
operating costs. Due to low production and high
cost, these vessels averaged the greatest loss for
the year, $20,704, and payback did not occur. The
internal rate of return on investment was
negative.

Table 2. — Summary of costs and returns information, net present value analysis, and pay
back period for five classes of shrimp vessels operating in the Gulf of Mexico in 1973.

Vessel class

Item

III

Number of vessels

Catch (pounds)

43,146

48,602

39,170

30,716

30,950

Gross revenue:

Per pound ($)

2.03

1.89

1.93

1.65

1.55

Total ($)

87,587

91,802

75,764

50,770

48,044

Cost:

Variable

Not proportional to catch ($)

42,152

31 ,694

28,134

22.835

16,784

Total ($)

77.195

68.600

58,543

43.444

36,385

Fixed ($)

31.096

22.642

22,231

18,550

15,296

Total ($)

108,291

91,242

80,774

61,994

51,681

Returns above variable cost ($)

10,392

23,202

17,221

7,326

11,659

Net revenue ($)

-20,704

560

-5,010

-11,224

-3,637

Equity requirement ($)

47,407

38,921

30,630

24,200

22,176

Payback period (yr)

{')

(')

{')

Internal rate of return (%)

13.21

2.65

(')

2.63

'Does not occur.

^Does not occur through operations — payback in the eighth year is due to sale of the vessel

^Less than 0%.

304

GRIFFIN ET AL.: ECONOMIC AND FINANCIAL ANALYSIS

Class II vessels had the highest landings,
48,602 pounds, of the five classes of vessels. They
also had the highest gross revenues even though
the average price per pound received was $0.14
less than class I vessels. They did experience
relatively high total costs, yet the variable costs
not proportional to catch, the "manageable"
variable costs, were $10,500 less than class I.
Class II vessels were able to register a positive
return to owner's labor and management of
$560 — the only one of the classes to achieve that.
Payback occurred only with the sale of the vessel
in the eighth year. The internal rate of return on
investment was 13.21%, which was the highest of
the five classes.

Class in was the most populous class. Gross
revenue was approximately $15,000 below and
total costs were about $10,500 below those of
class II vessels. The difference in the total costs
was due to costs directly proportional to catch — a
reflection of the fact that class III vessels caught
roughly 9,000 pounds less shrimp than did the
class II vessels. Class IE vessels had a negative
net return of $5,010. The internal rate of return
on investment was 2.65% and payback occurred
during the eighth year only with the sale of the
vessel.

Class rV vessel production was about 9,000
pounds less than class III vessels and the price
per pound was about $0.30 less, so that gross
revenue was $25,000 lower for the class IV
vessels. Variable costs not directly proportional
to catch were roughly $5,000 lower, and total cost
was $19,000 less for class IV vessels than for class
III vessels. Because of the low level of production
and gross revenues, class IV vessels had the sec-
ond greatest net loss, $11,224, of any of the five
classes, and payback did not occur. The internal
rate of return on investment was negative.

Class V vessels reached roughly the same level
of production as did class IV vessels, but at
$6,000 lower variable costs not directly propor-
tional to catch. Comparison of the returns above
variable costs shows class V vessels contributed
over $4,000 more towards fixed costs than did
class IV, while receiving some $2,000 less in
gross revenues. Net revenue was a negative
$3,637, but was still the second highest with
respect to the other four classes. Payback
occurred in year 8 only with the sale of the vessel
and the internal rate of return on investment
was 2.63%.

Financial Analysis with Cost Adjusted
to 1974 and 1975

Fishing for shrimp in the Gulf of Mexico in

1973 was definitely not an enterprise in which
profits could be achieved across the board. Figure
1 shows the break-even undiscounted cash flow
analysis for each of the five vessel classes, based
on 1973 costs and for costs updated to 1974 and
1975. Costs for 1974 were calculated by increas-
ing all cost items (fixed and variable) by 20%^
except fuel and new vessel cost. Because fuel
represents such a large portion of a vessel's
operating costs, it was treated separately and
increased from 18 to 32 cents per gallon. New
vessel cost was held constant at 1973 levels since
there has not been a significant number of
vessels entering the industry since 1973. Infla-
tion is expected to continue to increase at a rate
between 5 and 15%; therefore, 1975 costs were
increased by 10% over 1974 levels with the
exception of new vessel prices. For comparison
purposes the vertical dashed lines, labeled 1973,
indicate the 1973 average landings and the hori-
zontal dashed lines indicate the 1973 average ex-
vessel price received for each vessel class.

1974 Analysis

Input prices continued their upward trends in
1974. At the same time landings showed approxi-
mately a 2% improvement over the 1973 levels,
but shrimp prices fell by approximately 20%; the
combined effect was a 15% drop in the value of
shrimp produced in the Gulf of Mexico in 1974.^
Figure 1 explains graphically the ramifications of
such conditions on the undiscounted break-even
cash flows for each of the five vessel classes. As
the graphs show, none of the classes would have
experienced a break-even cash flow for 1974
given the 20% decrease in shrimp prices and
minimal increase in landings over the 1973
levels. This of course means that none would

^Based on the July 1974 wholesale price index including all
commodities (Board of Governors of the Federal Reserve
System 1974).

^Total Gulf of Mexico shrimp landings (heads-off) m 1973
were 114.8 million pounds, average ex- vessel price per pound re-
ceived was $1.50 and the value was $171.7 million (United
States Department of Commerce 1974). Landings for the same
period in 1974 were 116.9 million pounds, the average ex-
vessel price per pound was $1.18, and the value of the land-
ings was $1374 million (United States Department of Com-
merce 1974-75).

305

FISHERY BULLETIN: VOL. 74, NO. 2

$/lb.

1973 Normal
t I

$/lb.

k.OO -

3.00 ^

2.00 _

1.00

Normal

CLASS

$/lb.

4.00 -

3.00

2. 00 = _:

1.00

1973

Normal
1

1
1

V 1

1
1
1
1

1^^^

"T"^^^^^r^^

30 kO 50 60

CLASS II I

71*

LB.

(1000)

/lb.

If. 00

1973

1
1

Normal

1
1

3.00

1
1

2.00

1.00

1
il

^^^

^^.

75
74
73

CLASS IV

LB.

:iooo)

$/lb. »-

3.00 -

2.00

1.00 -

1973 Normal

LB.

:iooo)

Figure l. — Break-even undiscounted cash flow analysis
(0% rate of return on investment) based on 1973 costs
and returns data, and with costs inflated to 1974 and
1975 levels, for five classes of shrimp vessels operating in
the Gulf of Mexico.

CLASS V

306

GRIFFIN ET AL.: ECONOMIC AND FINANCIAL ANALYSIS

have registered a positive return on investment.
As a matter of fact, the class II vessels, which had
the highest rate of return in 1973, would have had
to receive approximately $2.20 per pound of
shrimp landed to achieve a break-even invest-
ment (0% internal return on investment) if an-
nual production of shrimp is held constant at the
1973 level of 48,602 pounds. Since they only re-
ceived $1.89 per pound in 1973 and prices declined
in 1974, investment in a class II vessel in 1974
would have yielded a negative rate of return on
investment.

1975 Analysis

If inflation continues at a 10% rate in 1975, and
production remains at approximately the 1973
level. Figure 1 indicates that ex-vessel prices
would have to increase to approximately $3.10,
$2.25, $2.65, $2.50, and $2.00 per pound of
shrimp landed for vessel classes I, II, III, IV, and
V, respectively, to achieve even a zero internal
rate of return on investment. Or, on the other
hand, with prices remaining constant at the 1973
level, production would have to increase to ap-
proximately 66,000, 57,000, 52,000, 49,000, and
40,000 pounds of shrimp landed per vessel,
respectively.

However, based on production functions esti-
mated by Nichols and Griffin (1974) for the Gulf
of Mexico shrimp fleet where catch is a function
of effort, 1973 production of shrimp from the Gulf
was below normal. Average annual landings for
the vessels in the sample were estimated in a
normal year to be approximately 53,000, 59,000,
49,000, 37,000, and 38,000 pounds of shrimp
landed per vessel for classes I, II, III, IV, and V,
in that order. The vertical dashed lines in Fig-
ure 1 labeled "normal" indicate the average
landings for each class of vessel for the normal
production year.

The average vessel in class II seems to be
operating at better than a break-even level in
1975 assuming normal production and 1973 ex-
vessel prices for shrimp. That is, given normal
production, class II vessels would have to receive
$1.85 per pound for shrimp landed in 1975 while
the 1973 average price for the class was $1.89
per pound. But, a new vessel cost of $130,000
would be just enough to set this cash flow at the
break-even level and the replacement of a class II
type vessel is estimated to be in excess of
$150,000 in 1975.

From the graphs in Figure 1 it is obvious that

none of the other classes (I, III, IV, V) are
operating at the break-even level assuming a
normal production year and 1973 average shrimp
prices and new vessel costs. In order to bring the
cash outflows down to the levels necessary to
achieve break even, class Ill-type vessel owners
could only invest approximately $30,000 in a new
vessel in 1975, and class V owners could invest no
more than $40,000. To reiterate, these break-
even levels represent a zero internal rate of
return on investment. Significantly, class I and
class IV-type vessel owners could not achieve the
breakeven level even with a zero investment
requirement.

DISCUSSION AND IMPLICATIONS

The resolution of problems facing the Gulf
shrimp industry may come about as a result of
changing economic conditions and/or changes in
specific policies which may or may not be initia-
ted or suggested by the industry. A number of
possible changes have been suggested which bear
consideration.

One suggestion has been a fuel subsidy for the
fishing industry. This would be a direct saving to
vessel owners on the largest single input cost
item. Assuming a normal production year, it
would take a subsidy of 35, 13, 48, and 15 cents
per gallon for classes I, III, IV, and V, respec-
tively, to break even with a zero return on invest-
ment assuming prices stayed constant at the
1973 level. Chances of obtaining any relief in this
area are very slim. At best, the extent of such
relief would likely be limited to future increases
related to oil import taxes. Current fuel expenses
would probably not be reduced.

Efforts to improve the efficiency of fishing
operations are also a priority consideration. The
operation of fishing vessels during periods of
marginal profitability required improved man-
agement and closer consideration of the effects of
the day-to-day decisions in running the vessel.

Import quotas and tariffs are one suggested
alternative to the current cost-price squeeze in
the industry. By controlling imports it is antici-
pated that supplies on the market can be reduced
thus preventing prices from being depressed be-
low the domestic producer's costs. The goals of
free trade and stabilized or lower consumer
prices may make approval of the necessary con-
trols through the political process difficult to
realize.

307

FISHERY BULLETIN: VOL. 74, NO. 2

Market expansion and development programs
have also been suggested as a means of shifting
demand and increasing prices. Market develop-
ment is a long term process and the industry
should commit itself to such a program. This
suggests a greater continuity of programs than
the occasional reaction to crisis situations which
are evident in the recent history of the industry.

A much larger question has been introduced in
this discussion of economic efficiency. Industry
sources have indicated a concern that the indus-
try has become overcapitalized in shrimp trawl-
ing vessels. One classic solution to this is a total
fisheries management scheme which includes a
limited entry concept. Other conditions assumed
equal, this would increase catch per unit of effort
and would result in lower costs per unit of shrimp
landed. This is not a short-run solution, however.
It is only now being experimented with in U.S.
fisheries. A great deal of planning and informa-
tion would be needed to design and implement
such a program.

Long-run problems of limited entry include the
possibility of creating a stagnant, protected in-
dustry which loses touch with both the consumer
market and the market for resources. In the long-
run this may be more detrimental than going
through periodic readjustments such as that
which the industry currently faces.

If it can be assumed that the relative positions
of the unit cost and revenue curves remain
constant in the future and assume normal pro-
duction years, then based on the sample size of
each vessel class, the percentage reduction in
vessels needed for break even can be calculated.
Using class I as an example, in a normal year, the
14 vessels in class I would have landed a total of
742,000 pounds of shrimp. To experience a break-
even rate of return, each vessel would have to
land 66,000 pounds of shrimp. Dividing 66,000
pounds per vessel into 742,000 pounds implies
that class I's total production of 742,000 pounds
could only support approximately 11 vessels or
79% of the vessels sampled.'^

CONCLUSIONS

The major conclusion from the analysis pre-

Tt is obvious that if the total Gulf shrimp fleet were reduced
to 79% of its current size, total production would also decrease.
That is, the estimated reduction in the fleet should be adjusted
with respect to the production function. However, calculations
using the production function made less than a 1% difference.

sented here is that investment in a shrimp trawler
is unprofitable assuming the environment exist-
ing in 1973 when these data were collected and for
which the average relationships were estimated.
The analysis permits tracing the effects of altered
assumptions regarding average prices and vessel
landings on profitability. Only class II vessels
showed profits under the 1973 conditions.

The shrimp industry is undergoing consider-
able economic stress. The underlying causes
relate to factors in the general economy beyond
industry control and the rapid expansion in po-
tential fishing effort which occurred during the
period since the late 1960's. Means of coping with
this stress include both improved management to
reduce costs and various forms of government
programs will be necessary to permit the imple-
mentation of some of these ideas.

Perhaps some would prefer to allow a period of
significant readjustment forcing the marginal
firms to leave the industry. The costs of this
readjustment, both economic and social, must be
considered by those who propose this solution.
Several things could happen which would pre-
vent a significant readjustment: landings could
increase dramatically, the economy could recover
quickly thus improving demand and prices, or
input costs could decline. However, these things
may not happen soon enough to avoid the
difficult readjustment problems.

LITERATURE CITED

BOARD OF Governors of the Federal Reserve System.

1973. Federal Reserve Bulletin. Div. Admin. Serv., Board
Gov. Fed. Res. Syst., Wash., D.C. 59:837-A119.

1974. Federal Reserve Bulletin. Div. Admin. Serv, Board
Gov Fed. Res. Syst, Wash., D.C. 60:683-A88.

Hayenga, W. A., R. D. Lacewell, and W L. Griffin.

1974. An economic and financial analysis of Gulf of Mexico
shrimp vessels. Tex. Agric. Exp. Stn., Misc. Publ. 1138,
14 p.

Nichols, J. P., and W. L. Griffin.

1974. Recent trends in catch and fishing effort in the Gulf
of Mexico shrimp industry and economic implications.
Dep. Agric. Econ. Inf. Rep. 74-5 SP-1. Texas A&M
Univ., College Station.

1975. Trends in catch-effort relationships with economic
implications : Gulf of Mexico shrimp fishery. Mar Fish.
Rev 37(2): 1-4.

United States Department of Commerce.

1974. Fisheries of the United States, 1973. Curr Fish.

Stat. 6400, 106 p.
1974-75. Gulf coast shrimp data. Curr. Fish. Stat. 6442,

6462, 6503, 6523, 6544, 6564, 6593, 6613, 6633, 6652,

6671.

308

GRIFFIN ET AL.: ECONOMIC AND FINANCIAL ANALYSIS

WARDLAW, N. J., ni, AND W. L. GRIFFIN.

1974. Economic analysis of costs and returns for Gulf of
Mexico shrimp vessels: 1973. Dep. Agric. Econ. Tech.
Rep. 74-3, Tex. Agric. Exp. Stn., Tex. A&M Univ., College
Station.

APPENDIX

Average cost equations were estimated using
ordinary least squares regression analysis for
each of the nine groups of vessels by the use of
linear, quadratic, and log linear functions. In
general, considering all nine equations, the log
linear model gave the best statistical results
whereas the quadratic gave the worse. This

implies that the average cost curves were ever
decreasing over the range of the data available.
Predicted values from the log linear model for the
nine equations were plotted by the computer on
one graph and compared. Because all nine plots
were relatively parallel, economies of scale did
not exist over the range of the sampled data.

Since the plotted predicted values were rela-
tively parallel, one average cost equation was
estimated using construction, length, and effort
as dummy variables. All three were statistically
significant variables at least at the 95% level of
confidence in explaining the average costs of
producing shrimp. For a more detailed discussion
see Wardlaw and Griffin (1974).

309

LONG-TERM FLUCTUATIONS OF
EPIBENTHIC FISH AND INVERTEBRATE POPULATIONS IN

APALACHICOLA BAY, FLORIDA

Robert J. Livingston, Gerard J. Kobylinski, Frank G. Lewis, HI, and Peter F. Sheridan^

ABSTRACT

A 3-yr study was made concerning seasonal changes in the biota of Apalachicola Bay. The
Apalachicola River causes a temporal progression of changes of various environmental parameters in
the bay such as salinity, turbidity, nutrients, and detritus levels. Fishes were more widespread
in their distribution throughout the bay than invertebrates. This was thought to be related to
trophic response and habitat preference. High levels of relative dominance prevailed for both groups
with the top three species of each group accoiuiting for more than 80% of the total number of
individuals taken.

Peak levels of monthly abundance of various dominant fish species tended not to overlap through a
given 12-mo period. Invertebrate species abundance usually reached peak levels during summer and
fall periods. The seasonal appearance and distribution of organisms in the Apalachicola Bay system
was comparable to that found in other estuaries in the northern Gulf of Mexico. The temporal and
spatial distribution of estuarine fishes and invertebrates was associated with species-specific
reproductive cycles, trophic relationships, and habitat preferences. The Apalachicola estuary was
viewed as a seasonally stable system, with regular temporal fluctuations of the biota through each
annual cycle.

There is a rapidly growing literature concerning
fluctuations of populations of epibenthic es-
tuarine organisms (Dahlberg and Odum 1970;
Bechtel and Copeland 1970; Copeland and
Bechtel 1971; McErlean et al. 1973; Oviatt and
Nixon 1973; Copeland and Bechtel 1974; Calla-
way and Strawn 1974; Livingston 1975). Haed-
rich and Haedrich (1974) noted that seasonal
changes of fish populations in a Massachusetts
estuary allow more species to utilize the estuary
than if there were constant direct competition.
Staggered reproductive cycles were postulated as
a partial explanation for this "dynamic situa-
tion." Trophic variability was also considered a
mechanism for reduced competition. Copeland
and Bechtel (1974) identified key environmental
requirements for six Gulf coast species, and con-
sidered such limits as potential criteria for es-
tuarine management programs. Oviatt and
Nixon (1973) noted that although fish biomass
remained constant throughout the year, indi-
vidual species abundance varied seasonally. They
found that biomass and numbers of individuals
could not be accounted for on the basis of physical

^Department of Biological Science, Florida State University,
Tallahassee, FL 32306.

parameters alone, and it was considered that
biological functions such as competition and pre-
dation could be more important determinants of
species distribution in estuarine systems.

The present study is part of a comprehensive
field program in Apalachicola Bay, Fla. (Liv-
ingston et al. 1974). This is a relatively un-
polluted, shallow coastal estuary bounded by
barrier islands. The bay is physically dominated
by the Apalachicola River (Estabrook 1973;
Livingston et al. 1974). This paper is concerned
with long-term, seasonal fluctuations of epiben-
thic fish and invertebrate populations, and the
possible interrelationships of the physicochemical
and biological elements of the Apalachicola Bay
system.

MATERIALS AND METHODS

Field Operations

A detailed description of the sampling meth-
odology is already available (Estabrook 1973;
Livingston et al. 1974). Physicochemical and
biological samples were taken monthly from
March 1972 to February 1975 at a series of sta-
tions in East Bay and Apalachicola Bay (Figure
1). Water samples were taken at the surface and

Manuscript accepted December 1975.
FISHERY BULLETIN: VOL, 74, NO. 2, 1976.

311

FISHERY BULLETIN: VOL. 74, NO. 2

Apaiach.col.

Figure l. — The Apalachicola Bay system with permanent sam-
pling stations for long-term studies concerning fluctuations
of populations of epibenthic fishes and invertebrates.

bottom with a 1-liter Kemmerer bottle. Tempera-
ture was measured with a stick thermometer
and/or a YSP dissolved oxygen meter. Salinity
was determined with a temperature-compensated
refractometer periodically calibrated with stan-
dard seawater. Color was measured with a (Hach)
American Public Health Association platinum-
cobalt standard test while turbidity was deter-
mined with a Hach model 2100A turbidimeter.
Light penetration readings were taken with a
standard Secchi disk. River flow data were pro-
vided by the U.S. Army Corps of Engineers
(Mobile, Ala.) while local climatological informa-
tion was provided by the Environmental Data
Service, NOAA, U.S. Department of Commerce.

Biological collections were made with 5-m (16-
foot) otter trawls (%-inch mesh wing and body;
%-inch mesh liner). Repetitive, 2-min trawl tows
were taken at each station at speeds of 2-3 knots.
Seven subsamples were taken at stations 1, 2, 4,
5, and 6 while two samples were taken at stations
lA, IB, IC, 3, and 5A. All organisms were pre-
served in 10% Formalin, sorted and identified to
species, measured and/or counted (standard
length for fishes; total length for shrimps;
carapace width for blue crab, Callinectes sa-
pidus). Stations 1 and 4 were also sampled at
night, approximately 1-2 h after sunset for the
first 2 yr of the study.

All statistical analysis was carried out using an
interactive computer program designed for the
study of extensive data collections. The extent of

interstation community similarity was tested
using the C\ index of overlap (Morisita 1959;
Horn 1966). This index determines the probabil-
ity that two randomly drawn samples from popu-
lations X and Y will be the same species relative
to the probability that two individuals of the
same species will be drawn from population X or
Y alone.

K =

1 = 1

\y =

s
2 ^ x,y,

1 = 1

ik, + ky) XY

i = l

^Yellow Springs Instrument Co. Reference to trade names
does not imply endorsement by the National Marine Fisheries
Service, NOAA.

where S = number of species

X, and J, = number of individuals of the
ith species in populations
XandF
X and Y = total number of individuals in

the two communities
k^ and X^ = measures of diversity (Simp-
son 1949) as modified for
sampling with replacement
(Horn 1966).

Values for this index range from 0 (no species in
common) to 1. A hierarchical (stepwise) multiple
regression analysis was carried out using
monthly population size as the dependent vari-
able. Various physicochemical and biological
parameters (temperature, salinity, chlorophyll a,
turbidity, color, Secchi disk depth, total depth,
local rainfall, wind speed and direction, tidal
stage, river flow, and dissolved oxygen) were used
as the independent variables. All such functions
were tested in the same month of collection and
with a 1-mo lag in the physicochemical parame-
ters. Due to the relatively high number of inde-
pendent variables, the stepwise regression was
used whereby one variable at a time was sys-
tematically introduced into the equation, and, at
each step, the variable added was the one giving
the greatest increase in the multiple correlation
coefficient. While not necessarily giving the
"best" equation, this method is computationally
feasible, and frequently gives results comparable
to methods that would determine all possible re-
gressions. Since the salinity, color, and turbidity

312

LIVINGSTON ET AL.: LONG-TERM FLUCTUATIONS OF POPULATIONS

data had skewed distributions, logarithmic trans-
formations were used for such variables to ap-
proximate normality.

RESULTS

Physicochemical Parameters

Depths of the various stations ranged from 1 to
2.5 m. With the exception of shallow areas, such
as station 6, which are characterized by periodi-
cally moderate concentrations of widgeon grass,
Ruppia maritima, East Bay has a silty-sand bot-
tom with little benthic macrophyte development.
Stations proximal to river drainage (stations 2-4)
are marked by strong currents and seasonally
high deposits of allochthonous detritus (leaf lit-
ter, branches, etc.). Except for shallow fringing
areas, Apalachiocola Bay (stations 1, lA-lC) has
little benthic macrophyte development; it is
dominated by silty-sand bottom with interspersed
oyster bars.

The Apalachicola River is a major determinant
of the physical environment of the bay system.
There is a seasonal fluctuation in flow with peak
levels occurring during winter and spring
months. Local rainfall, with peaks during late
summer and early fall, is out of phase with this
pattern. During the present period of study, river
flow determined salinity throughout the bay.
Mean salinity in East Bay was lower than that in
Apalachicola Bay; oligohaline areas (stations 5A,
6) were without measureable salinity from mid-
winter to early summer. Outer bay stations had
higher salinities; at station IB, the salinity did
not go below lb"L during the 3-yr study. During
periods of increased salinity, the shallow bay sys-
tem was vertically stratified (Estabrook 1973;
Livingston et al. 1974). However, there was little

horizontal or vertical variability in water tem-
perature at any given time. Low temperatures oc-
curred during the winter months. Turbidity
levels were relatively high throughout the bay,
and were directly related to river flow rates. Color
levels reflected both river flow and proximity to
land runoff, with elevated levels in East Bay
areas during the summer. Although there were
various complex physical changes in different
areas of the bay due to basin physiography, local
runoff, tidal currents, depth, etc., the major
habitat features of the Apalachicola Bay system
were determined by river conditions.

Distribution of Fishes
and Invertebrates

Similarity coefficients (cumulative, by station)
are shown in Table 1. Species such as bay an-
chovy, Anchoa mitchilli; Atlantic croaker, Mic-
ropogon undulatus; and sand seatrout, Cynoscion
arenarius, were dominant throughout the sam-
pling area. Others such as scaled sardine, Haren-
gula pensacolae, and Gulf menhaden, Brevoortia
patronus, were taken primarily in East Bay. High
interstation similarity of species assemblages of
fishes was noted, although grass bed areas such
as station 6 were characterized by higher num-
bers of species than other (mud-flat) stations.
There was increased spatial variability among
the invertebrate assemblages. Species such as the
blue crab and the penaeid shrimps (Penaeus
setiferus , P. duorarum ) were more evenly distri-
buted throughout the system than others. Grass
shrimps (Falaemonetes pugio, P. vulgaris, P. in-
termedius) were more frequently taken in the
grass beds of East Bay while the brief squid, Lo/-
liguncula brevis, was a dominant species in
Apalachicola Bay. High levels of species similar-

TabLE 1. — C\ values (by station) for invertebrates and fishes taken in the Apalachicola Bay system

(March 1972-February 1975).

Stati
Day

on 1
Night

Station

Day

Night

Fishes

1 day

0.96

0.94

079

0.95

0.97

0.80

092

0.99

094

0.95

0.93

night

063

0.76

0.52

0.95

0.87

0.58

0.77

0.95

0,79

0.83

0.85

0.54

093

0.84

0.95

0.93

0,96

069

0.96

0.97

0.80

0.36

0.29

0.22

065

0.84

0.94

088

0.73

0.98

0.87

0.91

0.57

0.46

0,86

091

094

0,67

0.84

0.95

0.89

0,90

0.78

0.82

0.34

0.58

022

0.23

0.85

094

0,96

0.95

0,97

0.94

0.68

0.53

0.50

0 15

031

049

0,95

0,78

0.95

0.91

0.84

4 day

0.96

0.54

0.73

0.20

038

092

068

0,91

0.99

0.98

0.92

night

0.87

0.90

0.69

0.24

0.69

0,62

0.69

0,82

0,93

0.94

0.93

0.74

0.25

0,53

0.10

0.19

0,98

0.41

0.85

0,50

0.98

0.92

0.84

0.36

0.58

0.09

0.58

0,97

0.54

0.94

0,64

0,97

0.96

0.23

0.20

0.16

0.06

0.09

0.16

0.79

0.28

025

0.17

0.21

313

FISHERY BULLETIN: VOL. 74, NO. 2

ity were noted among river-dominated and East
Bay stations (1, 2, 4, 5, 5A); outer bay stations
(lA, IB, IC) also w^ere somew^hat alike according
to the Cx similarity analysis. Station 6, as a
grass-bed area, differed from most of the other
collections. Station IB, v^ith consistently higher
salinity than the other stations, differed in terms
of invertebrate species composition. These data
indicate that fishes are more widespread in their
distribution throughout the bay system than the
invertebrates, which were more habitat-specific
with respect to substrate, salinity, etc.

Seasonal Fluctuations of
Dominant Species

Comparative dominance figures for the 10 most
numerous fish and invertebrate species are given
in Table 2. Relative dominance is high in both
groups with the top three species of fishes and
invertebrates constituting 77.0 and 80.7% of the
respective combined totals. Some species such
as H. pensacolae, B. patronus, and Atlantic
threadfin, Polydactylus octonemus, were found
during limited periods (April 1973, April 1974,
and May-August 1973, respectively). Seasonal
variations in the six dominant species are shown
in Figure 2. The most conspicuous species was A.
mitchilli, which was particularly abundant dur-
ing the first year of study. Peaks of numbers
usually occurred during fall or early winter
(October-January). With M. undulatus, peak
levels usually were noted during late winter or
early spring (February-March) whereas C.
arenarius reached abundance during late spring
and summer months (usually around August).
The sea catfish, Arius felis, usually peaked by
midsummer (July) while Atlantic bumper,
Chloroscombrus chrysurus, and southern king-
fish, Menticirrhus americanus, were prevalent

Table 2. — The lO dominant species of fishes and inverte-
brates taken in the Apalachicola Bay system from March 1972
to February 1975. Figures are expressed in percentages of
total numbers of individuals.

Fish

Invertebrate

Anchoa mitchilli

42.3

Penaeus setiferus

40.1

Micropogon undulatus

26.0

Palaemonetes pugio

20.4

Cynoscion arenarius

8.7

Callinectes sapidus

20.2

Leiostomus xanthurus

Penaeus duorarum

5.3

Harengula pensacolae

2.6

Lolliguncula brevis

4.3

Bairdiella chrysura

1.6

Penaeus aztecus

2.6

Chloroscombrus chrysurus

Neritina reclivata

1.5

Polydactylus octonemus

1.4

Portunus gibbesii

1.1

Arius felis

Palaemonetes vulgaris

0.8

Brevoortia patronus

1.2

Rhithropanopeus harrisii

0.5

3000-

D
Q

tr
lij

£
E

"I I I

MJ S D M J S DMJ S DM

MJ S DM J S DM J S DM

TIME-MONTHS

Figure 2. — Seasonal changes of numbers of individuals and
mean size of six dominant species of fishes taken in the
Apalachicola Bay system from March 1972 to February 1975.

during late summer or early fall (August-
October). The spot, Leiostomus xanthurus, usu-
ally peaked during winter and spring months;
silver perch, Bairdiella chrysura, had a variable
abundance curve. Overall, there was considerable
regularity in the appearance of the dominant bay
fishes even though there was often a marked
within-species variation in total numbers from
year to year.

Annual fluctuations of the dominant inverte-
brate species are shown in Figure 3. The white
shrimp, Penaeus setiferus, was prevalent from
August to November with autumn peaks of
abundance; the other penaeids usually reached
high numbers in the late spring (P. aztecus) or
late summer (P. duorarum). Palaemonetes pugio
was usually found in the bay during spring
months (March-May) while P. vulgaris reached
high numbers in November. The blue crab peaked
during summer and winter periods. Early sum-
mer and fall peaks were noted for Lolliguncula

314

LIVINGSTON ET AL.: LONG-TERM FLUCTUATIONS OF POPULATIONS

1600-

J 1200
<

Q 800

Penaeus setiferus

— r I'l'i I I' I I I t 'I —

s s s

Penaeus duorarum

I I'l r

s s s

Penaeus aztecus

Palaemonetes
pugio

-1— m— t-TTT

I I I

s s

Callinectes sapidus

i'"i' I ■»' I r

s s s

Lolliguncula brevis

MJ S DMJ S DMJ S DM

I I I I I I

MjSDMJSDMJSDM

210

1^8

234

E
E

TIME-MONTHS

Figure 3. — Seasonal changes of numbers of individuals and
mean size of six dominant species of invertebrates taken in
the Apalachicola Bay system from March 1972 to February
1975 (Palaemonetes pugio and Lolliguncula brevis were not
measured).

brevis. Unlike the fishes which usually reached
peak levels during different months of the year,
the invertebrates tended to increase in numbers
during spring and fall periods.

Annual peaks of abundance often coincided
with influxes of juvenile fishes and invertebrates.
A more detailed analysis of this is shown for two
representative species of fishes (Figure 4) and in-
vertebrates (Figure 5). The young stages of Mic-
ropogon undulatus entered the bay during the
winter at which time there was a continuous re-
cruitment for several months. Decreased num-
bers coincided with gradual increases in size dur-
ing spring and summer months. With Cynoscion
arenarius, recruitment of young occurred during
spring and summer, with subsequent increases in
size during fall and winter months. The blue crab
had peaks of young individuals during summer
and winter periods although an almost continu-
ous succession of young crabs entered the bay
during the year. Young stages of Penaeus

setiferus were found during the summer with
growth occurring through fall and winter. The
other penaeid shrimps had similar growth pat-
terns with recruitment of the young during sum-
mer and fall periods. The data indicate that vari-
ous patterns of recruitment and growth occur
among the different estuarine species, although
the inverse relationship of numbers and size ap-
pears to hold for most of the dominants.

Results of the regression analysis are shown in
Table 3. Factors such as chlorophyll a , Secchi disk
readings, and color repeatedly accounted for some
of the variability associated with fluctuations of
estuarine populations. Often such associations
were made with a 1-mo lag in the independent
variable. In most cases, the given independent
variables accounted for less than 50% of the var-
iability of the population data. There was a dis-
tinct correlation with factors related to trophic
phenomena such as chlorophyll a and Secchi disk
readings; this would indicate that biological func-
tions such as feeding behavior and reproduction
could play an important role in the determination
of population shifts in the Apalachicola Bay sys-
tem. These data indicate that no single set of forc-
ing functions can account for the population
changes of various estuarine species. Species
abundance is dependent on complexes of interac-
tions and possibly can be accounted for more
adequately by relating such processes to djmamic
changes in physical variables as well as impor-
tant biological parameters. It is obvious that re-
gression analysis cannot account for changes in

Table 3. — Results of the stepwise regression analysis of
various independent parameters and species (population)
occurrence in the Apalachicola Bay system from March 1972
to February 1975. Independent variables are listed by order
of importance with R ^ expressed as a cumulative function of
the given parameters.

Species

Independent variables

Anchoa mitchilli

Chloroptiyll a, Secchi

0.38

Micropogon undulatus

River flow (lag), Secchi (lag)

0.46

Cynoscion arenarius

Chlorophyll a, wind, Secchi (lag), temp

0.83

Polydactylus octonemus

Chlorophyll a (lag), salinity, Secchi

058

Arius tells

Temp, wind

0.30

Leioslomus xanthurus

Turbidity (lag), Secchi, salinity, temp

0.85

Chloroscombrus chrysurus Temp (lag), temp, salinity

0.44

Mentlclrrhus americanus

Temp (lag)

0.19

Symphurus plaglusa

Color (lag), color, Secchi

0.63

Bairdlella chrysura

Wind, temp, color

0.40

Penaeus setiferus

Wind, chlorophyll a, incoming tide, color

0.48

Palaemonetes pugio

Turbidity

0.36

Callinectes sapidus

Secchi, incoming tide

0.49

Penaeus duorarum

Chlorophyll a, Secchi

0.41

Lolliguncula brevis

Chlorophyll a (lag), temp

0.43

Portunus gibbesll

Chlorophyll a (lag), Secchi

0.39

Palaemonetes vulgaris

Turbidity

0.32

Rhithropanopeus harrisll

Wind

0.18

Callinectes simills

Chlorophyll a, temp

0.34

315

Micropogon undulatus Cynoscion arenarius

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 4.— Monthly size-fre-
quency distribution of two
species of fishes taken in the
Apalachicola estuary from
March 1972 through February
1975.

0 20 21 40 41 60 61 80 81 1000120121140 141 160 0 10 11 20 21 30 31 40 41 50 61 60 (.170

LENGTH (mm) LENGTH(mm)

Callinectes sapidus

Penaeus setiferus

TOTAL

Figure 5. — Monthly size-frequency
distribution of two species of inver-
tebrates taken in the Apalachicola es-
tuary from March 1972 through Feb-
ruary 1975.

0-15 WsO 3l'45 46-60 61 75 76 90 91-105 106-120 0-20 21-40 41-60 61-80 81 100101 120 121140

LENGTH (mm)

LENGTH(mm)

316

LIVINGSTON ET AL.: LONG-TERM FLUCTUATIONS OF POPULATIONS

the adaptive response of populations to the ex-
tremely complex environment of the estuary. The
data indicate that, in this case, temperature and
salinity might not be as critical in the determina-
tion of seasonal fluctuations of estuarine popula-
tions as biological functions such as trophic re-
sponse and possibly reproduction.

DISCUSSION

A review of the literature (Gunter 1945, 1950;
Daugherty 1952; Reid 1955; Van Engel 1958;
Gunter and Hall 1965; Williams 1965; Tagatz
1968; More 1969; Perez Farfante 1969; King
1971; Lyons et al. 1971; Swingle 1971; Perret and
Caillouet 1974; Stokes 1974; Swingle and Bland
1974) confirms that although minor variations
were evident (notably among the fishes), there
was a generally high level of conformity concern-
ing the time of appearance of various dominants
in the Apalachicola estuary with previously re-
corded data from other northern Gulf areas. Al-
though such timing was essentially stable from
year to year, there was considerable within-
species variability in annual abundance. For
example, the bay anchovy was particularly dom-
inant during the summer and fall of 1972, while
fewer individuals were taken during the succeed-
ing 2 yr. The Atlantic bumper, although not con-
sidered a common Gulf species (Perret and Cail-
louet 1974), was relatively common in the
Apalachicola estuary, especially during the first
year of collection. Some species reflected particu-
lar habitat preferences: Palaemonetes pugio was
located primarily in grass-bed areas of East Bay
during periods of low salinity while L. brevis was
found in outer bay areas during summer and fall
periods of increased salinity. Although gen-
eralized temperature and salinity preferences
have been shown for various estuarine species
(Copeland and Bechtel 1974), as a whole these
organisms show a wide tolerance for short-term
changes in these parameters. This could help to
explain the general lack of importance of temper-
ature and salinity as critical variables in the
multiple regression analysis; quite obviously,
other functions such as acclimatization would
tend to complicate such a direct approach to de-
termination of causative agents. The multiple re-
gression technique was limited in its application
to causal relationships since various biological
functions are probably involved in the determina-
tion of a given population curve.

It is possible that trophic relationships and re-
productive cycles are of critical importance in the
spatial and temporal distribution of estuarine
populations. As in other Gulf estuaries, the
Apalachicola Bay system is dominated by
juvenile stages of a small number of species. The
bay anchovy, abundant in a size range of 35-50
mm, is considered to be a generalized zooplankti-
vore at this stage, feeding in the water column on
copepods, amphipods, mysids, larval and juvenile
shrimps and fishes, etc. (Darnell 1958; Odum and
Heald 1972; Carr and Adams 1973). Various
studies (Roelofs 1954; Darnell 1958; Fontenot and
Rogillio 1970) indicate that M. undulatus
(juveniles, 10-50 mm) feeds primarily on zoo-
plankton (copepods and amphipods) while C.
arenarius (juveniles, 40-99 mm) consumes larger
zooplanktors such as mysids, shrimp, and larval
or juvenile fishes (Darnell 1958; Springer and
Woodburn 1960). Juvenile (up to 40 mm) spot also
feed on zooplankton; more mature fish of this
species (40-200 mm) become benthic omnivores
(Roelofs 1954; Darnell 1958; Springer and Wood-
burn 1960). Juvenile B. chrysura (16-160 mm)
feed on copepods, mysids, shrimp, and small
fishes (Darnell 1958; Carr and Adams 1973).
Thus, the dominant fishes in the Apalachicola
Bay system are primarily planktivorous although
possible differences could exist in vertical feeding
distribution and the size and species composition
of the prey organisms. Previous work has shown
that Anchoa mitchilli feeds on small crustaceans
and C. arenarius eats the larger, more motile
crustaceans. Both Leiostomus xanthurus and B.
chrysura feed on small mid-water planktors
(mainly copepods) as early juveniles, with later
stages becoming benthic omnivores feeding
largely on mysids and shrimp. Increased concen-
trations of zooplankton occur in Apalachicola Bay
during the spring and summer while palae-
monetid shrimp are abundant during winter
and early spring (H. L. Edmiston pers. commun.).
Thus, diversity in feeding behavior would con-
tribute to the observed vertical partitioning of
prey organisms among various planktivorous
species; such data are consistent with the ob-
served distribution of fishes in Apalachicola
Bay at any given period of time.

Of the six most prevalent invertebrates in the
Apalachicola estuary, five are benthic omnivores
and one is a probable planktivore. Juvenile blue
crabs consume detritus while larger individuals
(20-200 mm) are omnivorous, feeding on detritus

317

FISHERY BULLETIN: VOL. 74, NO. 2

and plant material, mollusks, polychaetes, crus-
taceans, and fishes (Darnell 1959; Tagatz 1968;
Odum and Heald 1972). Penaeid shrimp are also
omnivores, feeding on similar forms (Williams
1965; Darnell 1958; Eldred et al. 1961; Odum and
Heald 1972). Palaemonetes pugio feeds primarily
on detritus (Adams and Angelovic 1970; Oviatt
and Nixon 1973; Welch 1975). Qualitative obser-
vations indicate that Lolliguncula brevis is a
planktivore (Dragovitch and Kelly 1967). Thus,
most of the epibenthic invertebrates utilize de-
tritus and are more closely associated w^ith sedi-
ment type, benthic macrophyte distribution, and
placement of allochthonous forms of detritus than
the planktivorous fishes; this, together with cer-
tain (species-specific) temperature and salinity
tolerances, could provide a partial explanation for
the observed differences in the spatial distribu-
tion of the fishes and invertebrates.

Another important evolutionary mechanism
for the partitioning of the energy resources of an
estuary is the temporal succession of species over
an annual cycle. Abundance interrelationships
expressed as percentage of total catch are shown
in Figure 6. There was a certain regularity of
percent representation of dominant species of
fishes and invertebrates in the Apalachicola sys-
tem. For example, relative occurrence of P. pugio
was high during spring months while Penaeus
setiferus was dominant during late summer and
fall. The blue crab was abundant during winter
periods. Among the fishes, C. arenarius was dom-
inant during the spring and summer while A.
mitchilli (after the first year of sampling) pre-
dominated in the fall andM. undulatus prevailed
during the late winter and spring. When a com-
parison was made among the 10 most dominant
species of fishes for peaks of abundance, such in-
creases were evenly distributed over a 12-mo
period. However, of the top 10 species of inverte-
brates, most peaks of abundance occurred during
fall periods (September-November) with second-
ary concentrations of peaks during early summer
(May-June). Livingston (in press), describing pat-
terns of species richness and diversity in
Apalachicola Bay, noted that there was an an-
nual double peak in fish and invertebrate diver-
sity although there was far more seasonal varia-
bility mN (numbers of individuals) and S (num-
bers of species) among fishes than invertebrates.
These data would tend to corroborate and eluci-
date such findings. Thus, although the top domi-
nants in both groups showed distinct temporal

sequences in relative peak abundance, there was
a tendency for increased numbers of invertebrate
species during summer and fall periods whereas
peaks of A'^ and S for fishes were more contin-
uously distributed throughout the year. Major dom-
inants for both fishes and invertebrates thus
showed temporal partitioning through an annual
cycle. The noted differences in temporal distribu-
tional patterns of fishes and invertebrates could
be related to trophic response, with the plank-
tivorous fishes competing for a more limited
resource than the omnivorous (detritovore and om-
nivore) invertebrate species.

Several conclusions can be made with regard to
the biotic component in the Apalachicola estuary.
Various independent ecological factors operate to
determine the spatial and temporal distributions
of such organisms. Biological functions, as adap-
tive responses to the physical and trophic environ-
ment, determine such distributional patterns,
allowing a somewhat orderly temporal succession
of dominant forms within certain broad trophic
spectra. Patterns of reproduction of various dom-
inant estuarine species have evolved in such a
way as to permit such long-term partitioning of
the estuarine environment. Superimposed on this
are certain in situ mechanisms whereby further
resource division occurs due to vertical and hori-
zontal distribution of the component species. This
is largely determined by various microhabitat
phenomena such as salinity, bottom type, cur-
rents, availability of detritus, etc. In addition,
biological determinants such as intraspecific
competition and predation further modify the in-
dividual component populations. Thus, no single
parameter prevails in the determination of the
community structure of an estuary which under-
goes predictable seasonal changes even though it
is a physically forced system. Although there is
considerable short-term fluctuation in the num-
bers of individuals of various populations, the
system maintains a certain temporal constancy
which, according to a traditional view of such
phenomena, could be termed stability. This does
not mean that such a system is not in a constantly
transient state; on the contrary, through various
natural and unnatural mechanisms such as
habitat alteration and destruction, hurricanes,
etc., the various population equilibria can be
shifted so that the system is no longer charac-
terized by a stable temporal succession of energy
utilization. Each population fluctuates around a
certain point of equilibrium; such fluctuations are

318

LIVINGSTON ET AL.: LONG-TERM FLUCTUATIONS OF POPULATIONS

Tl ME-- MONTHS (3/72-3/75:

Figure 6. — Relative importance (% of total) of four dominant species of invertebrates and fishes taken in
the Apalachicola Bay system from March 1972 through February 1975. Such species represent 82.4 and
86.0% of the respective 3-yr totals.

determined by various natural and man-induced
phenomena such as overfishing and pollution.
The stability of the system depends on the
maintenance of various populations within cer-
tain limits of fluctuation. This has serious impli-
cations for any estuarine management program.
Holling (1973) pointed out that instability (in the
sense of large fluctuations) of individual popula-
tions may actually introduce a capacity for per-
sistence or resilience. Such resilience can be at-
tributed not only to component populations but to

the system as a whole. Stability thus is seen as
the "ability of a system to return to an equilib-
rium state after a temporary distiu-bance," (Hol-
ling 1973). Resilience, however, is a measure of
the ability of a given system to absorb changes of
primary forcing functions and still persist. By
this measure, an estuarine system such as
Apalachicola Bay comprises various populations
which undergo considerable annual fluctuations
but nevertheless are maintained within a rela-
tively stable temporal succession.

319

FISHERY BULLETIN: VOL. 74, NO. 2

ACKNOWLEDGMENTS

We thank Duwayne A. Meeter and Richard
Gurnee for their advice concerning the statistical
applications in this study. Thanks are also due to
Robert C. Harriss and his staff at the Edward
Ball Marine Laboratory (Florida State Univer-
sity) for the use of various support facilities. The
authors also acknowledge Glenn C. Woodsum and
J. Elton Jernigan for their help with the compu-
ter programs used in this study. Many students
have also helped in the field, and are acknowl-
edged for their interest in this project. Much of
the field collections was funded by grants from
NOAA Office of Sea Grant, U.S. Department of
Commerce (Grant Number 04-3-158-43), and the
Board of County Commissioners of Franklin
County, Fla. Data analysis was supported by EPA
Program Element #1 BA025 under Grant
Number R-803339.

LITERATURE CITED

Adams, S. M., and J. W. Angelovic.

1970. Assimilation of detritus and its associated bacteria
by three species of estuarine animals. Chesapeake Sci.
11:249-254.
BECHTEL, T. J., AND B. J. COPELAND.

1970. Fish species diversity indices as indicators of pollu-
tion in Galveston Bay, Texas. Contrib. Mar. Sci.
15:103-132.

Carr, W. E. S., and C. a. Adams.

1973. Food habits of juvenile marine fishes occupying
seagrass beds in the estuarine zone near Crystal River,
Florida. Trans. Am. Fish. Soc. 102:511-540.

COPELAND, B. J., AND T. J. BECHTEL.

197 1. Species diversity and water quality in Galveston Bay,
Texas. Water Air Soil Pollut. 1:89-105.

1974. Some environmental limits ofsix Gulf coast estuarine
organisms. Contrib. Mar. Sci. 18:169-204.

Dahlberg, M. D., and E. p. Odum.

1970. Annual cycles of species occurrence, abundance, and

diversity in Georgia estuarine fish populations. Am. Midi.

Nat. 83:382-392.
DARNELL, R. M.

1958. Food habits of fishes and larger invertebrates of Lake
Ponchartrain, Louisiana, an estuarine community. Publ.
Inst. Mar. Sci., Univ. Tex. 5:353-416.

1959. Studies of the life history of the blue crab (Callinectes
sapidus Rathbun) in Louisiana waters. Trans. Am. Fish.
Soc. 88:294-304.

daugherty, F. M., JR.

1952. The blue crab investigation, 1949-50. Tex. J. Sci.
4:77-84.

Dragovich, a., and J. A. Kelly, Jr.

1967. Occurrence of the squid, Lolliguncula brevis, in
some coastal waters of western Florida. Bull. Mar.
Sci. 17:840-844.

eldred, r. m. Ingle, k. d. Woodburn, r. f. hutton, and
H. Jones.

1961. Biological observations on the commercial shrimp,
Penaeus duorarum Burkenroad, in Florida waters. Fla.
State Board Conserv., Mar. Lab., Prof. Pap. Ser. 3, 139 p.
ESTABROOK, R. H.

1973. Phytoplankton ecology and hydrography of Apa-
lachicola Bay. M.S. Thesis, Florida State Univ.,
Tallahassee.

FONTENOT, B. J., Jr., AND H. E. ROGILLIO.

1970. A study of estuarine sport fishes in the Biloxi marsh
complex, Louisiana. F-8 Completion Rep., La. Wild Life
Fish. Comm., 172 p.

GALLAWAY, B. J., AND K. STRAWN.

1974. Seasonal abundance and distribution of marine
fishes at a hot-water discharge in Galveston Bay, Texas.
Contrib. Mar. Sci. 18:71-137.

GUNTER, G.

1945. Studies on the marine fishes of Texas. Publ. Inst.

Mar. Sci., Univ. Tex. 1:1-190.
1950. Distributions and abundance of fishes on the
Aransas National Wildlife Refuge, with life history notes.
Publ. Inst. Mar. Sci., Univ. Tex. 1(2):89-101.
GUNTER, G., AND G. E. HALL.

1965. A biological investigation of the Caloosahatchee
estuary of Florida. Gulf Res. Rep. 2, 71 p.

HAEDRICH, R. L., AND S. O. HAEDRICH.

1974. A seasonal survey of the fishes in the Mystic River,
a polluted estuary in downtown Boston, Massachusetts.
Estuarine Coastal Mar. Sci. 2:59-73.

HOLLING, C. S.

1973. Resilience and stability of ecological systems.
Annu. Rev. Ecol. Syst. 4:1-23.

Horn, H. S.

1966. Measurement of "overlap" in comparative ecologi-
cal studies. Am. Nat. 100:419-424.

KING, B. D., III.

1971. Study of migratory patterns of fish and shellfish
through a natural pass. Tex. Parks Wildl. Dep., Tech.
Ser. 9, 54 p.

Livingston, R. J.

1975. Impact of kraft pulp-mill effluents on estuarine
and coastal fishes in Apalachee Bay, Florida, USA.
Mar. Biol. (Berl.) 32:19-48.

In press. Diurnal and seasonal fluctuations of estuarine
organisms in a north Florida estuary. Estuarine Coastal
Mar. Sci.
Livingston, R. J., R. L. Iverson, R. H. Estabrook, v. e.
Keys, and J. Taylor, Jr.

1974. Major features of the Apalachiocola Bay system:
Physiography, biota, and resource management. Fla.
Sci. 37:245-271.

Lyons, W. G., S. p. Cobb, D. K. Camp, J. A. Mountain, T.
Savage, L. Lyons, and E. a. Joyce, Jr.

1971. Preliminary inventory of marine invertebrates col-
lected near the electrical generating plant. Crystal
River, Florida, in 1969. Fla. State Board Conserv.,
Mar. Lab., Prof Pap. Ser. 14, 45 p.

McErlean, a. J., S. G. O'Connor, J. A. Mihursky, and C. I.
Gibson.

1973. Abundance, diversity, and seasonal patterns of
estuarine fish populations. Estuarine Coastal Mar.
Sci. 1:19-36.

320

LIVINGSTON ET AL.: LONG-TERM FLUCTUATIONS OF POPULATIONS

MORE, W. R.

1969. A contribution to the biology of the blue crab
(Callinectes sapidus Rathbun) in Texas, with a descrip-
tion of the fishery. Tex. Parks Wildl. Dep., Tech. Ser.
1,31 p.

MORISITA, M.

1959. Measuring of interspecific association and simi-
larity between communities. Mem. Fac. Sci. Kyushu
Univ., Ser. E, Biol. 3:65-80.

ODUM, W. E., and E. J. Heald.

1972. Trophic analyses of an estuarine mangrove com-
munity. Bull. Mar. Sci. 22:671-738.

OviATT, C. A., AND S. W. Nixon.

1973. The demersal fish of Narragansett Bay: An analysis
of community structure, distribution and abundance.
Estuarine Coastal Mar. Sci. 1:361-378.

PERRET, W. S., AND C. W. CAILLOUET, JR.

1974. Abundance and size of fishes taken by trawling in
Vermilion Bay, Louisiana. Bull. Mar. Sci. 24:52-75.

PfiREZ FARFANTE, I.

1969. Western Atlantic shrimps of the genus Penaeus.
U.S. Fish Wildl. Serv., Fish. Bull. 67:461-591.

REID, G. K., JR.

1955. A summer study of the biology and ecology of
East Bay, Texas. Part I. Introduction, description of
area, methods, some aspects of the fish community,
the invertebrate fauna. Tex. J. Sci. 7:316-343.

ROELOFS, E. W.

1954. Food studies of young sciaenid fishes, Micropogon
and Leiostomus, from North Carolina. Copeia 1954:
151-153.

SIMPSON, E. H.

1949. Measurement of diversity. Nature (Lond.) 163:688.
SPRINGER, V. G., AND K. D. WOODBURN.

1960. An ecological study of the fishes of the Tampa
Bay area. Fla. State Board Conserv., Mar. Lab., Prof
Pap. Ser. 1, 104 p.

Stokes, G. M.

1974. The distribution and abundance of penaeid shrimp
in the Lower Laguna Madre of Texas, with a descrip-
tion of the live bait shrimp fishery. Tex. Parks Wildl.
Dep., Tech. Ser. 15, 32 p.

Swingle, H. a.

1971. Biology of Alabama estuarine areas — cooperative
Gulf of Mexico estuarine inventory. Ala. Mar. Res.
Bull. 5, 12 p.

Swingle, H. a., and D. G. Bland.

1974. A study of the fishes of the coastal watercourses
of Alabama. Ala. Mar. Res. Bull. 10:17-102.

Tagatz, M. E.

1968. Biology of the blue crab Callinectes sapidus Rathbun,

in the St. Johns River, Florida, U.S. Fish Wildl. Serv.,

Fish. Bull. 67:17-33.
Van Engel, W. A.

1958. The blue crab and its fishery in Chesapeake Bay.
Part 1 — Reproduction, early development, growth,
and migration. Commer. Fish. Rev. 20(6):6-17.
WELSH, B. L.

1975. The role of grass shrimp, Palaemonetes pugio in
a tidal marsh ecosystem. Ecology 56:513-530.

WILLIAMS, A. B.

1965. Marine decapod crustaceans of the Carolinas.
U.S. Fish Wildl. Serv., Fish. Bull 65:1-298.

321

DESCRIPTION OF ZOEAE OF COONSTRIPE SHRIMP,
PANDALUS HYPSINOTUS, REARED IN THE LABORATORY

Evan Haynes^

ABSTRACT

Zoeae oiPandalus hypsinotus from ovigerous females caught in Kachemak Bay, Alaska, were reared
in the laboratory. Each of the six zoeal stages is described and illustrated, and a brief description is
given for postzoeal Stages VII-IX. The descriptions are compared with descriptions of zoeal stages of P.
hypsinotus given by other authors.

Although pandahd shrimp form a major fishery
resource along the Pacific coast of North America,
little has been published on their early life his-
tory, especially on identification of the larval
stages. Berkeley (1930) described the zoeal stages
of five pandalid species from British Columbia,
Pandalus borealis Kr0yer, P. danae Stimpson,
P. hypsinotus Brandt, P. platyceros Brandt, and
Pandalopsis dispar Rathbun. The first zoeal stage
of each species was obtained in the laboratory, and
various remaining stages were obtained from the
plankton. Berkeley also mentioned briefly the
growth and distribution of the zoeae. Of 14 species
of pandalid shrimps known to occur along the Pa-
cific coast of North America, only two species,
Pandalus jordani Rathbun and P. platyceros, have
been reared through all their zoeal stages in
the laboratory (Modin and Cox 1967; Price and
Chew 1972).

In 1972, the National Marine Fisheries Service
began an intensive investigation at its field sta-
tion at Kasitsna Bay, Alaska, on the early life
history of pandalid shrimp in Alaskan waters.
The initial objective of the investigation was to
describe in detail laboratory-reared zoeae of each
pandalid species previously unverified. This re-
port describes and illustrates each of the six zoeal
stages of coonstripe shrimp, P. hypsinotus, and
compares the stages obtained from laboratory-
reared zoeae with stages obtained from the plank-
ton by other authors. Brief descriptions of post-
zoeal Stages VII through IX are also included.

'Northwest Fisheries Center Auke Bay Fisheries Laboratory,
National Marine Fisheries Service, NOAA, P.O. Box 155, Auke
Bay AK 99821.

MATERIALS AND METHODS

Owigeroxxs Pandalus hypsinotus were caught at
depths of 54 m (30 fathoms) in shrimp pots in late
April 1973. They were kept in plastic buckets
filled with seawater for about V2 h and then were
put in plastic glass hatching boxes similar to those
used by Price and Chew (1972) for rearing zoeae of
spot shrimp, P. platyceros. The hatching boxes
were kept in a biologically filtered recirculating
aquarium system containing 190 liters (50 gallons)
of refrigerated seawater, of which 19 liters (5
gallons) were exchanged for fresh seawater every
other day. Salinity was maintained between 32
and 34'Z, and temperature between 6° and 8°C.
The quality and quantity of light were not
controlled, but direct sunlight was avoided. Most
zoeae were released at night but some were
released during daytime whenever a female
shrimp was stimulated to flex her abdomen
rapidly. No predation of zoeae by female shrimp or
by the zoeae themselves was noted. No prezoeae
were seen.

About 50 zoeae were transferred by large-bore
pipette to each of 25 500-ml beakers containing
about 400 ml of aquarium seawater. In addition, a
zoea was placed in each of 50 25- by 50-mm num-
bered plastic vials held in compartmented trays.
The zoeae in the beakers provided both individual
specimens and cast skins of various stages for dis-
section, and the individual zoeae in the vials pro-
vided a continuous sequence of cast skins with a
known history. The beakers and vials were both
checked daily for exuviae. Seawater in the hold-
ing containers was changed every other day and
the zoeae were fed newly hatched nauplii of brine
shrimp, Artemf a salina, from San Francisco Bay.

Manuscript accepted December 1975.
FISHERY BULLETIN: VOL. 74. NO. 2, 1976.

323

FISHERY BULLETIN: VOL. 74, NO. 2

The density of nauplii was controlled only to the
extent that a few nauplii remained in the con-
tainer at the end of each feeding period. The origi-
nal beakers and vials were used throughout the
study because the zoeae also fed on the algae that
grew on the sides and bottoms.

All zoeae molted at night. Of the deaths noted,
most were caused by failure to complete the molt-
ing process; the posterior half was shed success-
fully, but the anterior half remained attached to
the mouth parts and pereopods. Survival was
about 90%.

Illustrations were drawn from unstained zoeae
and from exuviae stained red with Turtox CMC-S^
(acid fuchsin stain mountant). Stained exuviae
show segmentation and setation more clearly
than unstained. Zoeae and exuviae were dissected
with the aid of a binocular dissecting microscope.
The dissected material was mounted on a slide
and drawn to scale with the aid of a camera
lucida. Detail was checked with a compound mi-
croscope up to 430 X.

In the final illustrations (Figures 1-6), for clar-
ity, setules on the setae are usually omitted but
spinulose setae are shown. Because the numbers
of setae on the surface of the carapace and abdo-
men are highly variable, especially from Stage IV
onward, they are figured only when useful in iden-
tification of a stage. For each pair of appendages
the left member is figured except for the mandi-
bles, which are drawn in pairs and figured from
the right side. Whole zoeae are also figured from
the right side. The figures are in part schematic
and represent typical setal counts. The setation
formulas proceed from the distal to the proximal
ends of appendages. Gill development is men-
tioned in the text but usually not shown in the
figures. The terms are defined as follows:

spinose — bearing many spines
spinous — spinelike
setose — set with bristles (setae)
spinulose — set with little spines.

Total length was measured from the anterior tip of
the rostrum to the posterior tip of the telson with
the aid of a dissecting microscope; the number of
specimens used to determine total lengths is given
for each stage. A minimum of 10 exuviae of each
stage was used to verify segmentation and seta-
tion unless noted otherwise. The term "stage"

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

denotes the intermolt period. Nomenclature of
larval appendages and gills follows Pike and
Williamson (1964) and Berkeley (1930) respec-
tively.

STAGE I ZOEA

Total length of Stage I zoea (Figure lA) 5.8 mm
(range 5.5-6.2 mm; 50 specimens). Live specimens
brightly colored by numerous yellow chromato-
phores edged reddish browm. A conspicuous yel-
low chromatophore occurs dorsally on each eye-
stalk and at base of telson. Smaller but distinct
chromatophores occur on nearly all appendages,
especially maxillipeds and pereopods. Tips of an-
tennule and antennal scale are tinged reddish
brown. Chromatophore pattern of specimens
preserved in 5% solution of Formalin and sea-
water for several days identical to the pattern on
live specimens except that yellow color changes to
reddish brown after preservation. Rostrum slen-
der, spiniform, without teeth, about one-third
length of carapace, and projects horizontally or
slightly downward. Carapace with small, some-
what angular dorsal prominence at base of ros-
trum and a smaller rounded prominence near
posterior edge; prominences occur in all zoeal
stages. Antennal and pterygostomian spines
present, but both usually hidden by sessile eyes;
no supraorbital spine.

ANTENNULE (FIGURE IB).— Antennule (first
antenna) consists of a simple unsegmented tubu-
lar basal portion, distal conical base, distal conical
projection, and a heavily plumose seta on a small
conical base; distal conical projection bears four
aesthetascs — one long, one short, and two of inter-
mediate length.

ANTENNA (FIGURE IC).— Antenna consists
of inner flagellum (endopodite) and outer antennal
scale (exopodite). Flagellum two segmented and
about one-fourth longer than scale; distal segment
is styliform, tipped by a plumose seta and a spine.
Distal segment may be partially segmented proxi-
mally. Protopodite bears spinous seta at base of
flagellum and a spine at base of scale, both of
which persist throughout zoeal development. An-
tennal scale distally divided into six segments
(two proximal joints incomplete) and fringed with
10 heavily plumose setae along terminal and in-
ner margins. A small seta occurs on outer margin
near base of terminal segments.

324

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

MANDIBLES (FIGURE ID).— Mandibles with-
out palps. Incisor process of left mandible usually
bears four teeth in contrast to the distinctly
triserrate incisor process of right mandible. Left
mandible bears one premolar denticle and right
mandible bears two. Two subterminal processes
occur on truncated molar process of left mandible
but not on right mandible.

MAXILLULE (FIGURE IE).— Maxillule (first
maxilla) bears coxal and basial endites and an
endopod. Proximal lobe (coxopodite) bears a stout
seta near base and 12 spinulose setae terminally
along with a series of extremely fine hairs. Me-
dian lobe (basipodite) bears 11 spinulose spines in
two rows on terminal margin and several fine
hairs subterminally. Endopodite originates from
lateral margin of basipodite and bears three
terminal and two subterminal setae; three of the
five spines are sparsely plumose, the remaining
two spinulose. There is no evidence of an outer
seta (representing a vestigial exopodite) on
maxillule.

MAXILLA (FIGURE IF).— Maxilla bears plate-
like exopodite (scaphognathite) with 16 long,
approximately equal, evenly spaced plumose
setae along outer margin and one longer and
slightly thicker seta (at proximal end). Endopo-
dite has four partly fused segments and bears
nine large plumose setae. Basipodite bilobed;
each lobe bears eight setae. Bilobed coxopodite
bears 16 setae, 4 on distal lobe and 12 on proxi-
mal lobe.

FIRST MAXILLIPED (FIGURE IG).— First
maxilliped most heavily setose of natatory ap-
pendages. Protopodite partially segmented; bears
7 setae on proximal segment and 18 slightly
smaller setae on distal segment; most setae on
protopodite plumose but some simple or spinu-
lose. Endopodite distinctly four segmented; seta-
tion formula— 4, 2, 1, 3. Exopodite a long slender
ramus segmented at base; has four terminal and
five or six lateral natatory setae. Epipodite a
single lobe.

SECOND MAXILLIPED (FIGURE IH). —Pro-
topodite bisegmented; distal segment bears eight
sparsely plumose setae, and proximal segment
bears a simple seta. Endopodite distinctly five
segmented; fourth segment expanded laterally;

terminal segment has at least two spinulose
setae; remaining setae on endopodite usually
sparsely plumose; setation formula — 7, 2, 1, 1, 3.
Exopodite similar to exopodite of first maxilliped
but slightly larger; has 4 terminal setae, 11 or 12
lateral natatory setae. No epipodite.

THIRD MAXILLIPED (FIGURE II).— Proto-
podite bisegmented; distal segment bears four
setae. Endopodite distinctly five segmented and
nearly as long as exopodite, giving it more pedi-
form appearance than either of the two preceding
appendages; setation formula — 4, 8, 2, 2, 2. Ex-
opodite similar to second maxilliped but slightly
longer; has 3 or 4 terminal setae and 14 lateral
natatory setae. No epipodite.

FIRST PEREOPOD (FIGURE IJ).— Endopo-
dite functionally developed and similar in form to
third maxilliped but slightly smaller Endopodite
distinctly five segmented; ends in simple conical
dactylopodite; setation formula — 3, 7, 2, 2, 2. Ex-
opodite naked. Protopodite bisegmented; has four
setae. Neither this nor remaining pereopods of
this stage have any evidence of epipodite.

SECOND PEREOPOD (FIGURE IK).— Sec-
ond pereopod similar to first except that it has
fewer setae and fourth or propodal joint is
slightly extended to form beginning of chela.

THIRD, FOURTH, AND FIFTH PEREOPODS
(FIGURE IL-N). — These three pereopods essen-
tially identical to each other except that they
decrease slightly in size from third to fifth. No
exopodites.

PLEOPODS. — No pleopods evident, not even
as small buds.

TELSON (FIGURE 10).— Telson not segmen-
ted from sixth abdominal segment; slightly
emarginate distally; bears 14 densely plumose
setae. Minute spinules at base of each seta;
larger spinules along terminal margin between
bases of four inner pairs and on the four inner
pairs of setae themselves. Enclosed uropods
visible. No anal spine.

STAGE II ZOEA

Total length of Stage II zoea (Figure 2A) 6.1

325

FISHERY BULLETIN: VOL. 74, NO. 2

Figure l. — Stage I zoea ofPandalus hypsinotus: (A) whole animal, (B) antennule, (C) antenna, (D) mandibles (right and left), (E)
maxillule, (F) maxilla, (G) first maxilliped, (H) second maxilliped, (I) third maxilliped, (J) first pereopod, (K) second pereopod, (L)
third pereopod, (M) fourth pereopod, (N) fifth pereopod, (O) telson.

326

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

0.5 MM

327

FISHERY BULLETIN; VOL. 74, NO. 2

0.2 MM

Figure 2. — Stage II zoea otPandalus hypsinotus: (A) whole animal, (B) antennule, (C) antenna, (D) mandibles (right and left), (E)
maxillule, (F) maxilla, (G) third maxilbped, (H) first pereopod, (I) second pereopod, (J) third pereopod, (K) foiirth pereopod,
(L) telson.

328

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

0.5 MM

329

FISHERY BULLETIN: VOL. 74, NO. 2

mm (range 5.6-6.5 mm; 50 specimens). Chro-
matophore color and pattern essentially identical
to Stage I except ventral surface of abdomen now
greenish. Rostrum still without teeth; not curved
downward as strongly as in Stage I. Carapace
same as Stage I except now has prominent supra-
orbital spine; antennal and pterygostomian spines
clearly visible. Eyes, sessile in Stage I, now
stalked.

ANTENNULE (FIGURE 2B).— Antennule
shows considerable change from Stage I, now
three segmented. It bears on terminal margin a
large outer and a smaller inner flagellum, outer
flagellum bears four groups of three aesthetascs
each, one group terminally and three groups
along inner margin; inner flagellum bisegmented
and bears three setae terminally, one long and
two short; originating at base of these two fla-
gella is a dorsal budlike projection bearing four
simple setae (projection and setae not shown in
Figure 2B). Proximal segment of antennule
laterally expanded at base, with about 12 small
setae arranged laterally near expansion; 3
lateral plumose setae and about 14 dorsally pro-
jecting but smaller plumose setae ring terminal
margin; large spine projects downward from ven-
tral surface. Second segment has 4 lateral plu-
mose setae, 2 long and 2 short, and about 10 dor-
sal plumose setae ringing terminal margin.
Third segment has seven lateral plumose setae —
five originating ventrally and the remaining two
dorsally — and three simple setae — two dorsal
and one lateral.

ANTENNA (FIGURE 2C).— Inner flagellum
nine segmented, about twice as long as scale;
distal segment tipped by about six small setae.
Spine on basipodite at base of inner flagellum re-
duced in size. Antennal scale fringed with 28-30
long, thin plumose setae along terminal and
inner margins. Joints at distal tip reduced to
four, three of them incomplete. Distal outer seta
of scale a stout spine.

MANDIBLES (FIGURE 2D).— More massive
than in Stage I but still without palps. Both
mandibles bear pair of premolar serrated denti-
cles, and molar processes are more developed.
Truncated end of molar process of right mandible
formed into curved lip. Subterminal processes
still present on left; mandible.

330

MAXILLULE (FIGURE 2E).— Endopodite es-
sentially unchanged from previous stage. Basi-
podite bears 10 spinose spines in two rows and
flve spinous setae on terminal margin, but no fine
hairs. Coxopodite bears 12 setae terminally,
5 spinous and considerably longer than remain-
ing 7.

MAXILLA (FIGURE 2F).— Similar to Stage I
except exopodite larger and now bearing 21 or 22
marginal plumose setae in addition to plumose
seta at proximal end. Lobes of basipodite bear
nine setae each instead of eight as in Stage I.

FIRST, SECOND, AND THIRD (FIGURE 2G)
MAXILLIPEDS.— Maxillipeds essentially iden-
tical to each other and nearly identical to first
stage except for an increase in size and a slight
variation in numbers of setae.

FIRST PEREOPOD (FIGURE 2H).— First pere-
opod functionally developed and similar in form
to third maxilliped. Exopodites fringed wdth 15-17
plumose setae. Endopodite six segmented. Propo-
dite projected slightly distally. Setae more num-
erous than in Stage I, especially on last two seg-
ments. This pereopod and the remaining four
have a pleurobranchia bud at their base.

SECOND PEREOPOD (FIGURE 21).— Similar
to first pereopod except propodite projection
longer and ischiopodite not segmented.

THIRD, FOURTH, AND FIFTH PEREOPODS
(FIGURE 2J, K).— Third, fourth, and fifth
pereopods essentially identical except for slight
differences in size, fifth being smallest. Seven
functional segments including dactyl opodite. Dac-
tylopodite bears spine at tip and three spines lat-
erally. No exopodite.

PLEOPODS (FIGURE 2A).— Pleopods evident
only as slightly swollen areas on abdominal
segments.

TELSON (FIGURE 2L).— Telson distinct from
sixth abdominal segment; bears 16 densely plu-
mose setae along margin. Spinule arrangement
essentially same as Stage I. Dorsal surface bears
four small simple setae. Uropods still enclosed
but longer than in first stage. No anal spine.

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

STAGE III ZOEA

Total length of Stage III zoea (Figure 3A) 6.7
mm (range 6.2-7.7 mm, 25 specimens). Chro-
matophore pattern similar to first two stages but
less yellow color and more reddish brown. Ros-
trum pointing slightly upward with one or two
small teeth at base. Supraorbital, antennal, and
pterygostomian spines still present on carapace.

ANTENNULE (FIGURE 3B).— Outer flagel-
lum distinctly three segmented; first and second
segments have two groups of three aesthetascs
each; distal segment has four aesthetascs. Inner
flagellum still bisegmented but about twice as
long as in Stage II. Remainder of antennule simi-
lar to Stage II except it is larger and more setose,
and lateral projection on proximal segment is
more arcuate.

THIRD MAXILLIPED (FIGURE 3G).— Similar
in shape to third maxilliped at Stage II but larger
and more spinous and propodite bears two small
spinulose spines. Numbers of setae on endopodites
of maxillipeds and pereopods on this and succeed-
ing stages are so highly variable that a specific
description of them would not be an aid in
identification of stage or species.

FIRST PEREOPOD (FIGURE 3H).— Exopodite
still present, more setose than Stage II. Propodite
bears a small spinulose spine near base. Pleu-
robranchia at base of this appendage and remain-
ing four pereopods barely larger than in Stage II.

SECOND PEREOPOD (FIGURE 31).— Most
significant changes are presence of chela on
endopodite and an additional segment on base of
ischiopodite.

ANTENNA (FIGURE 3C).— Antennal scale
with 32-36 lateral plumose setae; no segmenta-
tion at tip in this or later stages. Lateral margin
near base now has four additional simple setae.
Flagellum about 3 times length of scale; has
several additional segments and setae near base.

THIRD (FIGURE 3J), FOURTH, AND FIFTH
PEREOPODS.— Essentially similar; fifth small-
est as usual. Greater development from Stage II
shown by well-formed dactylopodite and more
setae. An additional segment occurs at base of
ischiopodite.

MANDIBLES (FIGURE 3D).— Both mandi-
bles without palps. Right mandible bears three
premolar processes; projections along anterior
molar edge stronger and truncated end not
curved into lip as in Stage II. Left mandible
molar processes also stronger, and subterminal
processes present.

MAXILLULE (FIGURE 3E).— Endopodite un-
changed from Stage II except two setae particu-
larly spinulose. Basipodite bears an additional
plumose seta and a group of small fine hairs sub-
terminally. Coxopodite now bears 14 instead of 12
setae and has more fine hairs than Stage II.

MAXILLA (FIGURE 3F).— Exopodite longer
than in Stage II, slightly curved, and bears 27
marginal plumose setae in addition to plumose
seta at proximal end. Lobes of basipodite bear 10
setae instead of 9 as in Stage II.

FIRST AND SECOND MAXILLIPEDS.— Epi-
podite on first maxilliped has rudiment of second
lobe. Otherwise, first and second maxillipeds
same as Stage II but slightly larger.

PLEOPODS (FIGURE 3A).— Pleopods evident
as small buds.

TELSON (FIGURE 3K).— Uropods free; bear
plumose setae and small, randomly located setae
on dorsal surface. Telson broader at tip than at
base and still slightly emarginate; bears seven
pairs of spinous setae and two pairs of lateral
spines. Base of telson bears a pair of simple setae
that increase in number in later stages and persist
in adults. Anal spine appears at this stage.

STAGE IV ZOEA

Total length of Stage IV zoea (Figure 4A) 7.5
mm (range 7.3-8.1 mm, 10 specimens). Chromato-
phore pattern and color considerably different
from previous stages. In general, numerous small
wine-red chromatophores occur on carapace, pere-
opods, and ventral surface of abdomen; small yel-
low chromatophores occur on carapace, anten-
nules, antennal scale, uropods, telson, and third
abdominal segment. Rostrum beginning to ac-
quire adult shape; 11-13 dorsal spines, 2 or 3 small
ventral spines, and 1 dorsal spine that may be

331

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 3. — Stage in zoea ofPandalus hypsinotus: (A) whole animal, (B) antennule, (C) antenna, (D) mandibles (right and left), (E)
maxillule, (F) maxilla, (G) third maxilliped, (H) first pereopod, (I) second pereopod, (J) third pereopod, (K) telson.

332

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

0.5 MM

333

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 4. — stage rV zoea ofPandalus hypsinotus: (A) whole emimiil, (B) antennule, (C) antenna, (D) mandible (right sind left), (E)
maxilla, (F) first maxilliped, (G) second maxilliped, (H) third maxilliped, (1) first pereopod, (J) second pereopod, (K) telson.

334

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

0.5 MM

335

FISHERY BULLETIN: VOL. 74, NO. 2

faint or distinct near acute tip. No supraorbital
spine in this or remaining stages (Figure 4A).
Small setae and groups of minute hairs irregu-
larly located on carapace.

ANTENNULE (FIGURE 4B).— Outer flagel-
lum four segmented and longer than in Stage III
beginning to acquire slender terminal portion as
in adult; five groups of aesthetascs, two groups on
first and second segments each and one group on
third; groups composed of 3, 3, 3, 4, and 5 aes-
thetascs. Inner flagellum four segmented, nearly
as long as outer flagellum. Rest of antennule
similar in shape to Stage III but larger; bears
additional spines and setae; lateral projection on
proximal segment more pronounced, and ventral
spine on proximal segment noticeably smaller
than in Stage III.

ANTENNA (FIGURE 4C).— Antennal scale
with 32-39 lateral plumose setae and is assuming
narrow, slightly curved form of adult; scale bears a
few simple setae medially and usually a large seta
on inner margin near tip. Inner flagellum not
much longer than Stage III, about SVz times
length of scale.

MANDIBLES (FIGURE 4D).— Incisor and mo-
lar processes of both mandibles separated by deep
cleft, and each mandible has unsegmented palp
bearing two setae terminally. Curved lip of right
mandible considerably larger than in Stage III.

MAXILLULE.— Similar to Stage III except
number of setae somewhat variable. Endopodite
usually has one seta but may bear additional
small setae. Basipodite has 12 spines and 9-13
setae terminally, 2 or 3 setae subterminally. Cox-
opodite usually has 15 setae.

MAXILLA (FIGURE 4E).— Exopodite fringed,
has 32 plumose setae in addition to plumose seta
at proximal end; separated from protopodite by
cleft and bears 3 setae along inner margin. Num-
ber of setae on endopodite reduced to four. Basi-
podite bears 12 setae on each lobe; proximal lobe
bears additional seta subterminally. Distal lobe of
coxopodite reduced in size and bears two setae
instead of four as in Stage HI; proximal lobe of
coxopodite bears eight long and five short setae.

FIRST MAXILLIPED (FIGURE 4F).— Epipo-
dite distinctly bilobed. Protopodite clearly two
segmented and bears 6 setae on proximal seg-
ment, 26 smaller setae on distal segment. Endopo-
dite three segmented and bears one long seta on
first segment and one long and one short setae
terminally on third segment. Exopodite bears 6
long plumose setae along proximal outer margin
and 9 or 10 natatory setae.

SECOND MAXILLIPED (FIGURE 4G).— Sec-
ond maxilliped has undergone considerable change
from Stage III and now is similar in shape to adult.
Endopodite five segmented; terminal segment
flattened with many short spinous setae on lateral
margins. Epipodite arises from coxopodite and is
single lobed.

THIRD MAXILLIPED (FIGURE 4H).— Exopo-
dite considerably reduced. Endopodite heavily
setose and spinous. Meropodite slightly enlarged
medially; not distinctly segmented from ischiopo-
dite. Basipodite enlarged medially somewhat
more than meropodite. Bud of mastigobranchia
arises from coxopodite.

FIRST PEREOPOD (FIGURE 41).— Exopodite
reduced as in preceding appendage. Endopodite
ends in simple, heavily setose conical dactyl, as in
the third maxilliped; ischiopodite articulates
somewhat laterally with meropodite. Pereopods of
this stage, except fifth pair, bear bud of masti-
gobranchia. Each pleurobranchia adult in shape
and clearly lobulated.

SECOND PEREOPOD (FIGURE 4J).— Exopo-
dite reduced in size as in third maxilliped and first
pereopod. Joints appear on carpal segment for
first time, 10 or 11 on left and 5-7 on right. Left
pereopod slightly longer (about one-tenth) than
right pereopod.

THIRD, FOURTH, AND FIFTH PEREOPODS.
— Essentially similar to pereopods of Stage III.

PLEOPODS (FIGURE 4A).— Pleopods cleft
slightly and without joints or setae.

TELSON (FIGURE 4K).— Lateral margins
nearly parallel but spaced slightly wider posteri-
orly and bear two spines on each margin. Termi-
nal margin straight and bears three pairs of
feathered spines, the second pair longest; two

336

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

simple setae — one long, one short — occur be-
tween first and second pairs of spines. Two pairs of
simple setae (inner pair stouter) occur at base of
telson and project noticeably at nearly right
angles to telson surface (Figure 4A). Both pairs of
uropods nearly as long as telson and fully devel-
oped; both bear numerous small setae irregularly
located on dorsal and ventral surfaces of both
pairs in addition to setae figured. Beginning of
transverse hinge (diaeresis) of exopodite of uropod
faintly evident.

STAGE V ZOEA

40-44 plumose setae; proximal expansion of ex-
opodite and setae along its inner margin, espe-
cially proximal seta, considerably longer than in
previous stages. Endopodite shaped as adult;
bears three setae. Shape and setation of basipo-
dite and coxopodite similar to Stage IV except
distal lobe of basipodite bears 15 setae and
proximal lobe of coxopodite bears 7 long and 5
short setae.

FIRST AND SECOND MAXILLIPED.— Simi-
lar to Stage IV except endopodite of first maxil-
liped bears two setae on second segment and
three or four on proximal segment.

Total length of Stage V zoea (Figure 5A) 9.2 mm
(range 8.4-10.1 mm, 10 specimens). Numerous
small wine-red chromatophores occur primarily
on cephalothorax but also along surface of abdo-
men to base of telson and on dorsal hump of third
abdominal segment; large wine-red chromato-
phore on side of carapace especially pronounced;
yellow chromatophores few and minute; occur in
head region at base of antennae, on antennules,
and on dorsal surface of eyes. Rostrum similar in
shape to adult; 15-17 dorsal teeth, in addition to 1
(rarely 2) near acute tip; 4 or 5 ventral teeth. Still
no setae between dorsal rostral teeth (Figure 5A).

ANTENNULE AND ANTENNA.— Essen-
tially similar to Stage IV. Inner flagellum of an-
tenna approximately 4 times length of scale.

MANDIBLES.— Mandibles larger but mor-
phology unchanged from Stage IV; mandibular
palp row three segmented and bears three or four
setae terminally (Figure 5B).

MAXILLULE (FIGURE 5C).— Maxillule adult
in shape. Endopodite bears one long seta termi-
nally, sometimes an additional short seta. Basipo-
dite bears 13 spines in two rows along terminal
margins: 5 of the spines are relatively long and
the remaining 8 short. Seventeen setae of various
lengths are distributed terminally and along
lateral margin of basipodite. Coxopodite bears
five long spinulose setae terminally and a row of
five shorter sparsely plumose setae extending
proximally; row of fine hairs and a medial seta
occur ventrally.

MAXILLA (FIGURE 5D).— Maxilla more adult
in shape than previously. Exopodite fringed with

THIRD MAXILLIPED.— Similar to Stage IV
except for a few additional setae, and exopodite is
reduced to remnant. Mastigobranchia similar in
shape to adult. Arthrobranchia small bud.

FIRST PEREOPOD.— Appendage with few ad-
ditional setae and spines. Exopodite remnant,
distal joint of ischiopodite more pronounced than
in Stage IV (Figure 5E). Arthrobranchia minute
bud. Mastigobranchia on this and pereopods two
to four; adult in shape.

SECOND PEREOPOD.— Exopodite remnant,
carpal joints of left and right pereopods 14-16 and
7. No arthrobranchia on this or remaining per-
eopods.

THIRD, FOURTH, AND FIFTH PEREOPODS.
— Distal joints of carpal and basial segments
pointed (Figure 5F), no additional joint at basis.
Setation essentially as shown in Figure 3J except
carpopodite and meropodite each bear a spine.

PLEOPODS (FIGURE 5A).— Pleopods bilobed,
segmented, and without setae.

TELSON (FIGURE 5G).— Lateral margins
nearly parallel but slightly farther apart at
center and bear two spines on each margin.
Terminal margin straight; arrangement of spines
and setae on margin similar to Stage IV. The two
pairs of setae at base of telson noticeably longer
than in Stage IV. Transverse hinge of exopodite
of uropod complete; numerous small setae located
randomly on dorsal and ventral surfaces in
addition to those figured.

337

FISHERY BULLETIN: VOL. 74, NO. 2

0.5 MM

FIGURE 5. — Stage V zoea of Pandalus hypsinotus: (A) whole animal, (B) mandibular palp, (C) maxillule, (D) maxilla, (E) first pereo-

pod, (F) fifth pereopod (segmentation only), (G) telson.

338

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

STAGE VI ZOEA

Total length of Stage VI zoea (Figure 6A) 10.8
mm (range 10.0-11.8 mm, 10 specimens). General
color wine-red, particularly on carapace and per-
eopods and along ventral abdomen; remainder of
telson greenish hue. Most appendages of this
stage differ in shape only slightly from those of
Stage V and succeeding stages and are not
figured in detail. Rostrum with 15-19 dorsal teeth
in addition to 1 (usually) but sometimes 2 dorsal
teeth near acute tip; 4-7, usually 5, ventral teeth.
A seta may occur between two or three dorsal
teeth (Figure 6A).

ANTENNULE. — Inner flagellum six segmen-
ted (rarely five). Outer flagellum eight segmen-
ted; bears seven (rarely eight) groups of three
(usually) aesthetascs each.

ANTENNA. — Antennal scale fringed with 40-
45 plumose setae; flagellum about 6 times length
of scale.

MANDIBULAR PALP.— Three segmented;
number of setae variable; setation formula — 6-8,
2-3, and 1-3.

MAXILLULE. — Endopodite unchanged from
Stage V. Basipodite bears about 20 setae and 13
spines; coxopodite bears 18 setae.

MAXILLA.— Exopodite fringed with 61 or 62
plumose setae. Three setae on endopodite. Seta-
tion formula of lobes of basipodite and coxopodite
21-22, 17-19, 2, 11-12.

FIRST MAXILLIPED.— Exopodite has 10 or 11
setae along proximal margin. Setation formula of
endopodite 2, 4, 5. Number of setae on protopodite
variable — 38-61 on basipodite, 7-12 on coxopodite.

SECOND MAXILLIPED.— More setose than in
preceding stages; about 50 setae on terminal
segment. No podobranchia.

THIRD MAXILLIPED.— No exopodite. Arthro-
branchia as two minute rounded buds.

FIRST PEREOPOD.— No exopodite. Arthro-
branchia bud at base of each pereopod except fifth.

SECOND PEREOPOD.— No exopodite; carpal
joints of left and right pereopods 19 and 7 or 8
respectively. Left and right meropodites with
three or four and one or two joints respectively.

THIRD, FOURTH, AND FIFTH PEREOPODS.
— Meropodite bears 4-6 spines. Fifth pereopod
bears neither bud of arthrobranchia nor epipodite.

PLEOPODS (FIGURE 6B).— All five pairs
segmented, biramus, and tipped with setae but
nonfunctional. Appendix interna small bud on in-
ner lamella of second and third pleopods only.

TELSON (FIGURE 6C).— Telson shows, for
first time, narrow shape similar to adult and
bears three pairs of dorsolateral spines. Terminal
margin rounded slightly; bears three pairs of
feathered spines and a pair of large setae
dorsally. Three pairs of stiff setae at base of
telson instead of two as in Stage V.

POSTZOEAL STAGES VII-IX

Total length of Stage VII zoea 12.1 mm (range
11.5-12.8 mm, four specimens). Pleopods func-
tional and appendix interna distinct on all
pleopods except first pair. Because abdominal
propulsion is evident at this stage, it is consid-
ered the first postzoeal (megalopa) stage (Wil-
liamson 1969). Dorsal rostral spines 19 or 20, 1 or
2 at acute tip; 7 or 8 ventral spines. Seta (usually
1, rarely 2) occurs between each pair of rostral
spines. Bud of podobranchia distinct, arises at
base of epipodite of second maxilliped; buds of
arthrobranchiae on third maxilliped distinct,
pointed. Telson bears four pairs of spines along
lateral margin, rarely an additional small spine
on either margin. Left and right carpal joints of
second pereopods 24 or 25 and 10 respectively.

Stages VIII and IX differ only slightly from
VII. Total length of Stage VIII zoea 12.4 mm
(range 11.1-13.0 mm, four specimens). Gill buds
more fully developed in VIII than in VII but not
yet lobulated. Left and right carpal joints of
second pereopod 28 and 10 or 11 respectively.
Total length of Stage IX zoea 13.6 mm (range
13.4-13.8 mm, three specimens). Rostrum with
one to three setae between dorsal rostral spines
and one to five setae between ventral spines; seta
between the two spines at rostral tip. Buds of
both podobranchiae and arthrobranchiae nearly
lobulated.

339

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 6.— Stage VI zoea ofPandalus hypsinotus: (.A) whole animal, (B) pleopods (1, 2. and 3), (C) telson.

340

HAYNES: DESCRIPTION OF PANDALUS HYPSINOTUS ZOEAE

COMPARISON OF ZOEAL

STAGES WITH DESCRIPTIONS BY

OTHER AUTHORS

Berkeley Q930j described and figured the first
stage zoeae of P. hypsinotus that she reared in
the laboratory. She also obtained the probable
second and third stages from the plankton, but
these were not described. Stage I zoeae reared by
Berkeley differed in several respects from mine,
but mostly in segmentation and setation of ap-
pendages. For instance, Berkeley showed the tel-
son separated from the sixth abdominal segment
by a joint whereas I do not. She described the tip
of the antennal scale as unsegmented, but my
zoeae have the tip divided into six segments. The
endopodites of the first and second maxillipeds of
her zoeae are unsegmented, and the exopodites of
the maxillipeds and first and second pereopods
are unjointed at their bases. In my zoeae, the
endopodites of the first and second maxillipeds
are segmented, and the exopodites of the maxilli-
peds and first and second pereopods are jointed at
their bases. Segmentation of appendages, espe-
cially in the early zoeal stages, is most clearly
seen in exuviae. Because Berekeley was unable
to obtain exuviae from her laboratory-reared
specimens, she probably missed seeing the
segmentation of most appendages.

Kurata's Q964) description of Stage I zoeae of
P. hypsinotus was also based on specimens reared
in the laboratory; the remaining stages (II-V) he
described were obtained from the plankton. The
most important differences between Kurata's
description and mine are: Stage I — Kurata's
zoeae bear a chela on the second pereopod and the
antennal scale is four segmented. In my zoeae the
chela does not appear until Stage III and the an-
tennal scale is six segmented. Stage 11 — The tip
of the antennal scale is two segmented in Kur-
ata's zoeae but four segmented in mine. Stage
III — On Kurata's zoeae, the marginal spines of
the telson vary from one to three pairs, and the
inner flagellum of the antenna is twice as long as
the antennal scale and has 9 or 10 joints. My
Stage III zoeae always have two pairs of marginal
spines and the inner flagellum of antenna is 3
times the length of the antennal scale and has 19
joints. Stage IV — The telson of Kurata's zoeae
decreases in width posteriorly; the inner flagel-
lum of antennule is two segmented; the tip of the
first pereopod bears a small chela; and the cara-
pace bears a supraorbital spine. The telson of my

Stage IV zoeae increases in width posteriorly; the
inner flagellum of antennule is four segmented;
the tip of the first pereopod bears a simple dac-
tylopodite in all stages i^including adults;; and
the supraorbital spine occurs only in Stages II
and ni. Stage V — The telson of Kurata's zoeae
bears 6-6 spines terminally; the carpopodites of
the second pereopods and the pleopods are with-
out joints; and the ceu^apace still bears a supra-
orbital spine. In my specimens, the telson bears
3-t-3 spines terminally; the carpopodites of the
left and right second pereopods bear 14-16 and 7
joints respectively; the pleopods are jointed; and
the carapace does not bear a supraorbital spine.

The cause for the morphological differences
between Kurata's description of the morphology
of the zoeae and mine is unknown but apparently
is not a result of my zoeae being reared in the
laboratory. My zoeae showed no variation in
number of zoeal stages and only negligible
morphological variation between individuals of
the same stage. Also, there were no morphologi-
cal differences between my zoeae reared in the
laboratory and the zoeae of P. hypsinotus avail-
able from local plankton collections (^Stages I-III).
The morphological differences between Kurata's
zoeae and mine may be due to geographical vari-
ation. Berkeley Q930) has showTi that pandalid
zoeae from the northeast Pacific are further ad-
vanced on hatching than those from the Atlantic,
although she did not have enough information to
compare identical species from both areas. Unfor-
tunately, Kurata's descriptions from Stage II
onw£u-d were based on specimens from the plank-
ton. Verification of geographical variation in
zoeal morphology will be possible only when
identification is based upon zoeae of known
parentage and the magnitude of variation is
established for each stage.

Segmentation of the antennal scale was used
by Lebour Q940j as one criterion for classifying
the early stages of pandalid zoeae into two
groups. The first group includes pandalid species
described by various authors as possessing a seg-
mented scale (Dichelopandalus bonnieri (Caul-
leryj, Pandalus montagui Leach, and P. propin-
quus G. O. Sars). The second group includes
pandalid species described by Berkeley (1930) as
possessing an unsegmented scale (P. stenolepis
Rathbun, P. hypsinotus, P. danae, and P. platy-
ceros). Price and Chew fl972) showed Lebour's
grouping to be invalid for P. platyceros. Kurata
(1964) described zoeae of P. hypsinotus as hav-

341

FISHERY BULLETIN: VOL. 74, NO. 2

ing a segmented scale. Laboratory-reared Stage I
zoeae known by me to possess a segmented scale
are Pandalopsis dispar, Pandalus stenolepis, P.
goniurus, P. borealis, P. danae, P. hypsinotus, and
P. platyceros. Berkeley obviously failed to recog-
nize the segmented scales on her specimens.
Therefore, Price and Chew's (1972) suggestion
that Lebour's grouping for classifying the early
stages of pandalid zoeae using segmentation of
the antennal scale be disregarded is valid.

In most Decapoda, the development of func-
tional pleopods provides a convenient and clear
distinction between the zoeal and postzoeal
stages because it is accompanied by several other
abrupt changes in morphology, such as loss or re-
duction of some or all of the thoracic exopodites
and changes in shape and body proportions. In
the Pandalidae, however, there is not always an
abrupt metamorphosis at this molt. Pike and
Williamson (1964) discussed how in P. montagui
the pleopods may become fully functional before
the exopodites on the pereopods show any
reduction; in P. danae the exopodites on the
pereopods and the third maxilliped degenerate
before the pleopods become functional; and in P.
kessleri Czernaivski the exopodites on the per-
eopods never become functional. In my zoeae the
development of functional pleopods occurred at
Stage VII, but other morphological changes
normally associated with postzoeal metamorpho-
sis occurred earlier, especially at the molt to
Stage IV. Morphological changes that occurred at
the molt to Stage VI are reduction of thoracic
exopodites; loss of supraorbital spines; changes in
color; changes in shape of rostrum, mandibles,
and second maxilliped; and segmentation of
carpopodite of the second pereopod. Depending
upon one's definition of "megalopa," it may be
valid to consider Stage VII of P. hypsinotus as the
megalopa; or one may consider stages IV through
VII are all megalopal or the term "megalopa" is
not strictly applicable to P. hypsinotus.

In addition to the morphological changes noted
above, abbreviated development of zoeae of P.
hypsinotus is also indicated by the occurrence of
thoracic exopodites on pereopods 1 and 2. In con-
trast, most Pandalidae without abbreviated de-
velopment have thoracic exopodites on pereopods
1-3. A notable exception is zoeae of P. platyceros,

which have thoracic exopodites on pereopods 1-3
but only four zoeal stages and 8 + 8 telson setae in
Stage I rather than the usual 7 + 7. Another
feature of abbreviated development in P. hypsi-
notus is the proximal extension and occurrence of
17 setae on the exopodite of the maxilla in Stage
I. Usually the exopodite of the maxilla in Stage I
of the Caridea has no proximal extension and
only five setae, as in the protozoea of the
Peneidea and most British Pandalidae (Lebour
1940; Gurney 1942). The abbreviated develop-
ment of zoeae of P hypsinotus agrees with the
findings of Berkeley (1930), who noted that zoeae
of most Pandalidae of the northeast Pacific tend to
be more developed when they hatch than is normal
for Caridea.

ACKNOWLEDGMENTS

D. I. Williamson of the University of Liverpool,
England, and C. Nyblade of the University of
Washington, Seattle, read an earlier version of
this manuscript and offered suggestions for
improvement.

LITERATURE CITED

Berkeley, A. A.

1930. The post-embryonic development of the common
pandalids of British Columbia. Contrib. Can. Biol.
6:79-163.

Gurney, R.

1942. Larvae of decapod Crustacea. Ray Soc. Publ. 129,
306 p.
KURATA, H.

1964. Larvae of decapod Crustacea of Hokkaido. 3. Pan-
dalidae. Bull. Hokkaido Reg. Fish. Res. Lab. 28:23-34.
Lebour, M. V.

1940. The larvae of the Pandalidae. J. Mar Biol. Assoc.

U.K. 24:239-252.
MODIN, J. C, AND K. W. Cox.

1967. Post-embryonic development of laboratory-reared

ocean shrimp, Pandalus jordani Rathbun. Crustaceana

13:197-219.

Pike, r. b., and D. L Williamson.

1964. The larvae of some species of Pandalidae (Decapo-
da). Crustaceana 6:265-284.

Price, V. A., and K. K. Chew.

1972. Laboratory rearing of spot shrimp larvae (Pandalus
platyceros) and description of stages. J. Fish. Res. Board
Can. 29:413-422.

Williamson, D. I.

1969. Names of larvae in the Decapoda and Euphausia-
cea. Crustaceana 16:210-213.

342

PRESENT AND HISTORICAL SPAWNING GROUNDS AND
NURSERIES OF AMERICAN SHAD, ALOSA SAPIDISSIMA,

IN THE DELAWARE RIVERA

Mark E. Chittenden, Jr.^

ABSTRACT

Spawning occurs from late May into July but mainly in a 3-wk period from late May to mid-late June.
Spawning ends progressively later proceeding upstream. Light intensity seemed to regulate when
spawning began each day. Fish selected shallow riffle areas in preference to pool habitat for spawning.
Spawning behavior is described.

Except for the most grossly polluted tidal water, spawning and nursery areas now extend throughout
fresh water of the main Delaware and into the East and West branches. The most important spawning
grounds and nurseries are now located from Port Jervis, N.Y., to Hancock, N.Y., and extend into the
lower East Branch; this has probably been the case since 1910-20. There has been a fundamental
upstream shift in the chief spawning grounds and nurseries since the decline of the Delaware River
shad runs, because these historically extended downstream from about Delaware Water Gap, Pa., and
included tidal water. Reasons for this shift suggest intrastream homing.

Only a small proportion of the historical nursery now contributes to production. Nursery and
spawning areas now contribute to production of adults in proportion to their distance from Philadel-
phia, Pa. The extent of the spawning and nursery area since about 1910-20 has probably expanded and
contracted around a core area in the upper Delaware near Hancock. Future prospects of Delaware River
shad are discussed. They depend upon water quality in the tidal area and the proposed Tocks Island
dam. Extirpation of the remnant runs is a distinct possibility.

The Delaware River basin once supported larger
landings of American shad, Alosa sapidissima,
than any other river system (Stevenson 1899).
Annual landings near the turn of the century av-
eraged about 14-17 million pounds but have con-
sistently been much less than 0.5 million pounds
since 1920 (Sykes and Lehman 1957; Chittenden
1974). Gross pollution near Philadelphia, Pa.
(Figure 1), has been the chief reason for the low
abundance since at least 1920 (Ellis et al. 1947;
Sykes and Lehman 1957; Chittenden 1969). If pol-
lution were cleared up, shad runs could be largely
restored (Chittenden 1969).

Spawning and nursery areas of shad in the
Delaware River are not well known, although the
U.S. Army Corps of Engineers proposes to con-
struct a dam near Tocks Island, a few kilometers
upstream of Delaware Water Gap, Pa. If proposed
fishways are not successful, this dam would pre-
vent access to nearly half the 406 km of fresh
water between Marcus Hook, Pa., and Hancock,
N.Y. Sykes and Lehman (1957) concluded that the

'Based on part of a dissertation submitted in partial fiilfill-
ment of the requirements for a Ph .D. degree, Rutgers University,
New Brunswick, N.J.

^Department of Wildlife and Fisheries Sciences, Texas A&M
University, College Station, TX 77843.

chief spawning and nursery areas were located
upstream of Tocks Island. Their studies were
made in 1950-52 when shad runs were almost
nonexistent, however, and their conclusion was
necessarily based on extremely limited data.
Shad runs markedly resurged during the early
mid-1960's when I made extensive collections and
observations of adults and young. This paper de-
scribes the spawning period, behavior during the
spawning period, recent and historical spawning
and nursery grounds, and discusses the future
prospects of shad in the Delaware River.

MATERIALS AND METHODS

Locations referred to are indicated in Figure 1
or, when first mentioned, by their approximate
distances upstream from Marcus Hook, situated
about 90 km downstream from the fall line at
Trenton, N.J., and near the transition between
fresh and brackish water.

Adults (278 males and 250 females) were col-
lected during the spawning runs at Lambertville,
N.J., using a 76-mm stretch-mesh, 107-m long
and 3.6-m deep haul seine at 3- or 4-day intervals
from 5 April to 19 May 1963, 20 March to 18 May
1964, 26 March to 7 May 1965, and 27 March to 19

Manuscript accepted September 1975
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

343

FISHERY BULLETIN: VOL. 74, NO. 2

N.Y
PA.

_ .-T^ t'^>--^Oown%ville (445)

14031 HancocO
(3901 Equ.nunkJ

Figure l. — The Delaware River. Numbers in parentheses
represent distance in kilometers from Marcus Hook, Pa.

May 1966. Low dissolved oxygen near Philadel-
phia blocked upstream passage of part of the 1965
spawning run, and few fish were captured at
Lambertville (Chittenden 1969); however, 43
dead males and 147 females were collected 21
May-10 June during a fish kill near Paulsboro,
N.J. The gonads of all adults collected were
examined to assess their degree of maturation fol-
lowing criteria of Leach (1925).

Data on the abundance of adults in the period
1959-62 were obtained from surveys (hereinafter
referred to as the Tri-State Surveys) during July
and August by the states of New Jersey, New
York, and Pennsylvania in cooperation with the
U.S. Fish and Wildlife Service. Rotenone was
used to collect. After 1962 I made many observa-
tions on adult abundance and gonad condition
during irregular collections upstream from
Dingmans Ferry, Pa., especially during annual
float trips in late May between Hancock and Port

Jervis, N.Y. Observations on behavior during the
spawning period were made chiefly in the East
Branch near Hancock.

Young fish were collected in nontidal fresh
water from 1963 to 1966 using 12-mm stretch-
mesh seines. In 1963, most collections were made
from Milford, Pa., upstream into the East and
West branches using a 1.8-m deep, 6-m long net or
a similar 10.7-m long bag seine. Most seine hauls
in 1963 captured few or no young, but a few hauls
captured many fish. Quantitative comparisons of
abundance were considered unreliable because of
the extremely contagious fish distribution. There-
fore, techniques were greatly modified in 1964. A
22.9-m long, 1.8-m deep net was paid out from a
pram. Lights (900 W for 1 h) were used at night to
attract young shad to the shoreline for most col-
lections during 1964 and thereafter. Only one
seine haul was made at a station when lights
were used, and collection sites were near deep
water.

During 1966, night seining with lights was con-
ducted at 2-wk intervals at Lordville, N.Y, Tus-
ten, N.Y, Dingmans Ferry, Belvidere, N.J.,
Riegelsville, Pa., and Scudders Falls, N.J., from
1-4 August to 27-29 September and weekly there-
after until 14 November following an unreplicated
two-way (stations and collections periods) experi-
mental design in which collections were made at
each station until the young completely vacated
nontidal water. No F tests for significant differ-
ences in abundance were possible because of the
inherent nature of the study: collecting with lights
made catches reliable but replication impossible;
intensive seawgird movement of the young by
mid-late August caused a stations by collection
period interaction which negated tests for main
effects. Supplementary collections using lights
were made during 1966 in the East and West
branches and downstream from Dingmans Ferry
(Table 1).

Nurseries refer herein to areas the young oc-
cupy during July and August. Data for 1963,
1964, and 1966 (after August) are presented in
Chittenden (1969, tables 35, 36, 38, 39, 41).

SPAWNING PERIOD

Nearly all spawning apparently occurred
within a 3-wk period fi-om about late May to mid-
late June, although some spawning extended well
into July. No fish had any translucent eggs until
early May at Lambertville, and only one running

344

CHITTENDEN: PRESENT SPAWNING GROUNDS OF SHAD

Table 1. — Summary of catch size in) and total lengths (mm)
of young American shad collected during July and August 1966.

Date

Location

Mean

Min

Max

July

Dingmans Ferry'

129

37.8

5.30

Belvidere

40.4

7.09

Scudders Falls

41.0

10.28

Erwinna

—

Riegelsvllle

51 2

839

Scudders Falls

68.3

8.50

August

1-4

Lordville

208

37 1

4.57

1-4

Tusten

516

43.5

922

1-4

Dingmans Ferry

193

62 1

9.67

1-4

Belvidere

629

8 19

1-4

Riegelsvllle

629

10.86

1-4

Scudders Falls

—

East Brancti, Hancock'

406

—

Downsville

—

Fishs Eddy

44.0

4.24

West Branch, Hancock

—

15-17

Lordville

363

468

9.21

15-17

Tusten

367

508

7.51

15-17

Dingmans Ferry

1,282

67.5

879

15-17

Belvidere

177

659

855

15-17

Riegelsvllle

74.3

10.14

15-17

Scudders Falls

948

870

109

29-31

Lordville

526

534

8.64

29-31

Tusten

62.1

9.95

29-31

Dingmans Ferry

124

707

880

100

29-31

Belvidere

753

861

29-31

Riegelsvllle

55.0

—

29-31

Scudders Falls

—

'Thie listed n v*/as estimated as about tialfttie total catch; large amounts of
detritus were mixed with the East Branch, Hancock catch, small fish were
hard to find and measurements were not taken.

ripe female was captured as early as 15 May. The
gonads of some dead fish collected near Marcus
Hook on 21 and 23 May 1965 were nearly ripe.
Three females seined at Skinners Falls, N.Y., on
3 June 1964 had partially spawned. In the East
Branch near Hancock, I observed much spawning
from 10 to 17 June 1964; a few adults moved into
a spawning area there after dark on 1 July 1965,
suggesting that some spawning occurred then.
Most spawning probably occurred before late
June, however, because there was a great mortal-
ity of adults by then (Chittenden 1976).

Spawning ended at a later date upstream than
it did downstream based upon the minimum sizes
of young captured (Table 1). Assuming a month
between hatching and transformation at about 25
mm (Walburg and Nichols 1967), spawning in
1966 ended about 7 June near Scudders Falls and
about 25 June near Riegelsvllle. At Belvidere,
spawning occurred at least until early June and
at Dingmans Ferry until 1 July. Spawning ended
near Tusten from 1 to 15 July and at Lordville
from 15 July to 1 August. Length frequencies of
young in July and August 1966 (Chittenden 1969)
also show that spawning ended later upstream
than it did downstream. However, the spawning
period probably varies slightly between years and

at different locations depending upon spawning
stimuli.

The spawning period is apparently prolonged
for individual fish. The ovaries of females cap-
tured near Hancock during June 1964 varied in
size, many ovaries being about one-third or two-
thirds the size of those from prespawning fish cap-
tured at Lambertville. This suggests prolonged
spawning of individuals as Lehman (1953) con-
cluded from egg diameter measurements.

BEHAVIOR DURING
THE SPAWNING PERIOD

During the day, behavior depended upon the
habitat occupied. The nontidal Delaware consists
of a sequential arrangement of shallow swift
riffles and slow-moving deep pools. Shad prefer-
red pools but were frequently observed in riffles
about 0.3 m deep. Schools offish circled slowly in
the pools but often formed a V in riffles. The point
of the V headed upstream or in the direction of
travel and left a readily observed wake. When the
school was stationary and facing upstream, the
fish at the point of the V moved to the rear after
about 30 s. The fish immediately behind these
leaders then moved to the point. This behavior
spreads energy expenditure among all members
of the school and may conserve energy as would
the preference for pools. Both may be important
to survival. Weight loss during the spawning
migration is high (Leggett 1972; Chittenden
1976), and starvation causes a large mortality on
the Delaware River spawning grounds (Chitten-
den 1976).

Adults were observed after dark in the shal-
lows by using a pole to suspend a lantern high in
the air. The large schools typical of the day seem
to disperse during the evening spawning period,
because only one to three fish were usually ob-
served. Several times a behavior was observed
which may have been the spawning act: a smaller
fish (male?) lined up on either side of a larger fish
(female?) bringing their vents in close proximity
while swimming; a brief splashing coincident
with a rattling sound occurred at or near the sur-
face; and the fish separated after a few seconds.
Splashing and rattling noises were continually
heard outside the lighted area. This behavior was
only witnessed after dark, and it occurred in
water as shallow as about 150 mm. Plankton nets
were not available to collect fresh eggs to confirm

345

FISHERY BULLETIN: VOL. 74, NO. 2

this was the spawning act. However, the vigorous
splashing and noise is similar to the observations
of Goode (1888) and Leach (1925) and of Leim
(1924) who used plankton nets to collect newly
fertilized eggs.

Light intensity seemed to regulate when
spawning began each day, and the shad seemed to
prefer shallow riffle areas for this activity. Few
fish were observed during the day in a shallow
riffle spawning site near Hancock, but many fish
moved from the upstream pool to the riffle as
evening approached. Concentration near the
riffle occurred earlier on overcast days than on
sunny days. I observed spawning only at night in
general agreement with Pennsylvania (1875),
Goode (1888), Leim (1924), Leach (1925), Walburg
and Nichols (1967), and Marcy (1972). In contrast,
Massmann (1952) found spawning at all hours in
the Pamunkey River, Va., although possibly more
intensively fi:-om noon to midnight. Water turbid-
ity probably influences the effect of light in reg-
ulating the daily onset of spawning. Spawning
probably tends to occur at night in clear water
such as the upper Delaware, but seems to begin
later during the day or occurs all day long in tur-
bid water typical of tidal areas such as the
Pamunkey River. Overcast skies apparently per-
mit spawning to begin earlier in the day.

SPAWNING GROUNDS

Important spawning grounds apparently ex-
tend no farther downstream than the Belvidere
area. During the Tri-State Surveys, greatest
numbers of adults were captured from Minisink
Island to Skinners Falls, and none were captured
downstream from Manunka Chunk (Table 2).
Few adults were captured from Long Eddy, N.Y.,
upstream. However, these collections were made
10-21 July which is well after most adults move
seaward or die (Chittenden 1976). Therefore, the
chief spawning grounds may have been farther
upstream.

Extensive observations from 1962 to 1968 gen-
erally support the Tri-State Survey collections,
but in contrast they suggest that the area from
Skinners Falls to the lower East Branch was ex-
temely important. Many adults were observed 31
May-1 June 1962 from Milford to Delaware Water
Gap, and 30 May-5 June 1963 from Mongaup
River (km 296) to a few kilometers above Calli-
coon, N.Y. (km 360). In 1964, hundreds of adults
were observed near Hancock and the lower East

Branch 29 May-20 June and (J. Musick pers.
commun.) near Milford on 31 May. Fewer adults
were observed after 1964, but they consistently
appeared from Sparrowbush, N.Y. (km 286), to the
lower East Branch in late May and early June.

Table 2. — Numbers of adult American shad captured during
the Tri-State Surveys.

Distance from

Marcus Hook, Pa.

Station

(km)

1959

1960

1961

1962

East Branch, Hancock

403

—

West Branch, Hancock

403

—

Long Eddy

378

—

Skinners Falls

346

107

134

Mongaup Area

292

271

—

Minisink Island

263

160

103

Tocks Island

218

Manunka Chunk

197

—

Raubs Island

152

—

Marshalls Island

132

—

Scudders Falls

—

Trenton Falls

—

Some spawning occurs downstream of Philadel-
phia; however, few fish which pass Philadelphia
spawn as far downstream as Lambertville. I col-
lected a nearly spent male on 10 June 1965 at
Marcus Hook. This fish undoubtedly had spawned
nearby, because low dissolved oxygen would have
prevented movement past Philadelphia after
April (Chittenden 1969). The Lewis Fishery at
Lambertville captured about 6,300 fish from 1963
to 1968, but only 21 were taken after 15 May.

Spawning extends into the lower West and East
branches, especially the latter, but dams prevent
movement upstream of Stilesville, N.Y., and
Downsville, N.Y. Young shad (27 mm total length)
were captured in the West Branch at Hancock on 9
August 1963 (Chittenden 1969, table 26). This
suggests spawning there because net movement of
the young is downstream. Adults were collected in
the East Branch at Hancock during the 1961 Tri-
State Surveys. Many occurred at least as far up-
stream as East Branch, N.Y. (km 430), in the runs
of 1962-65 (W. Kelly pers. commun.; my observa-
tions). I observed spawning in the East Branch
near Hancock in 1964 and 1965.

The adults ascend some tributaries, but it is not
certain if they spawn there. A female was caught
on 16 May 1961 in Big Flat Brook (km 235) about
10 km upstream from the Delaware (Anonymous
1961). Adults ascended several kilometers up the
Mongaup River from 1962 to 1964 and 6 km up the
Beaverkill River, an East Branch tributary (W.
Kelly pers. commun.).

346

CHITTENDEN: PRESENT SPAWNING GROUNDS OF SHAD

NURSERIES

The chief nursery in 1966 was apparently lo-
cated upstream from Dingmans Ferry and was
especially centered near Tusten and Lordville
(Table 1). Areas downstream from Tusten gradu-
ally decreased in relative importance. The chief
nursery extended into the lower East Branch;
many young were captured near Hancock on 7
August, but none were taken at Downsville and
few were collected at Fishs Eddy, N. Y. No fish were
captured in the West Branch near Hancock on 8
August, suggesting that the lower West Branch
was an unimportant nursery in 1966.

Two seemingly aberrant catches affect interpre-
tation of relative abundance upstream from Bel-
videre. The catch was small at Tusten on 30
August and very large at Dingmans Ferry on 17
August. Hundreds of young were attracted to the
lights on 10 and 21 August at Tusten which agrees
with the magnitude of catches on 4 and 16 August.
The Tusten catch on 30 August probably reflects a
seaward exodus offish after 21 August. A plateau
in size formed at Tusten by August 30 (Chittenden
1969, figure 47) when mean total length was 62
mm (Table 1). A plateau represents seaward
movement of larger fish, and seaward movement
of the young is probable when they reach 64 mm
(Chittenden 1969:248). Mean size at Dingmans
Ferry was 62 mm on 4 August and 67 mm on 17
August, so that the very large catch at Dingmans
Ferry on 17 August probably reflects an influx of
seaward moving young from farther upstream.

The Delaware River downstream of Belvidere
appears to be a relatively unimportant nursery.
Catches during July and August 1966 at
Riegelsville and Scudders Falls were consistently
much smaller than at stations farther upstream,
and a catch at Erwinna, Pa., in July was also
small. The largest catch in these 10 collections was
16 young. This is much smaller than the smallest
catch in 14 collections at Belvidere, Dingmans
Ferry, Tusten, and Lordville.

My collections and observations in 1963-65 gen-
erally agree with the nursery patterns of 1966. In
1963, young shad were observed and captured
from Dingmans Ferry to the lower East and West
branches; many were repeatedly observed and col-
lected in the lower East and West branches at
Hancock, and hundreds were observed near
Matamoras, Pa., on 19 July and at Skinners Falls
on 30 August. In 1964, young were captured from
Erwinna upstream to Cochecton, N.Y. (km 354):

hundreds were observed or captured at Belvidere,
Delaware Water Gap, Worthington Tract (km
217), Flatbrookville (km 235), Dingmans Ferry,
Sparrowbush, Pond Eddy (km 301), and Cochec-
ton. No collections were made upstream from
Cochecton in 1964 except on 18 August when no
young were captured using lights in the West
Branch at Hancock. In 1965, young were observed
or captured from Belvidere upstream to Pond
Eddy; hundreds were observed and captured at
Delaware Water Gap on 8 July, at Belvidere on 15
July, and at Dingmans Ferry, Sparrowbush, and
Pond Eddy on 21 July. No trips were made up-
stream of Pond Eddy in 1965.

GENERAL DISCUSSION

Historical Spawning and Nursery Areas

Shad migrated 68 km up the East Branch to
Shavertown (Bishop 1936) and 24 km up the West
Branch to Deposit in the early 1800's (Gay 1892).
A dam constructed at Lackawaxen, Pa., however,
blocked access upstream after 1823 (Slack 1874;
Smiley 1884; Gay 1892). Spawning grounds then
extended downstream from Lackawaxen for
about 70 yr until a fishway permitted upstream
access in 1891 (Bean 1892, 1903).

Apparently the chief spawning grounds were
historically downstream from Lackawaxen. The
shad catch along the Atlantic coast is primarily
age IV or older fish (Walburg and Nichols 1967).
Few Delaware River shad migrate upstream until
age III, and most now first do so at ages IV and V
(Chittenden 1975). No records exist of size or age
composition in the late 1800's-early 1900's when
Delaware River landings reached their zenith,
except that average weights about 1896 were 3.75
and 3.50 pounds (Stevenson 1899), 3.75 pounds
(Townsend 1901), and 4.2 pounds based upon
Smith's (1898) report on the numbers and pounds
caught. These weights are reasonably similar to
the mean weights of males (1,107 g) and females
1,737 g) captured at Lambertville from 1963 to
1965 (Chittenden 1976), so that recent Delaware
River data probably closely represent the age
structures near the turn of the century. There-
fore, renewed access to spawning grounds up-
stream from Lackawaxen could not have fully af-
fected landings until 1895 or 1896. Except for
1892, annual landings were about 13-14.5 million
pounds in the period of 1889-95 and about 13.9-
16.8 million pounds from 1896 to 1901 (Chitten-

347

FISHERY BULLETIN: VOL. 74, NO. 2

den 1974). The catches in these two periods are so
similar that it would appear that the Lackawax-
en Dam had little effect on abundance. The chief
spawning grounds may have been located even
further downstream than Lackawaxen, however,
because Abbott (1868) stated that shad were sel-
dom plentiful upstream from Delaware Water
Gap, and this is supported by Smiley 's (1884)
statement that no shad were seen farther up-
stream than Milford for 25 yr prior to 1872. Shad
were abundant at that time (Slack 1874).

Spawning grounds could have extended
downstream to about Marcus Hook, because shad
spawn in fresh water (Prince 1907; Leach 1925;
Hildebrand and Schroeder 1928; Massmann
1952). Consideration of preferred spawning and
nursery habitat and Delaware River morphology
suggests that tidal water was historically impor-
tant: the existence of an extensive tidal nursery
(and spawning area) immediately downstream
from extensive excellent spawning grounds was
probably important to the former abundance of
Delaware River shad (Chittenden 1973b). How-
ever, the contemporary literature conflicts on the
importance of the tidal Delaware (Pennsylvania
1897; discussion session after Meehan 1907; New
Jersey 1916).

The potential importance of the tidal Delaware
can be judged by comparison with other rivers.
Hudson River runs are entirely produced in tidal
water, because a dam constructed in 1840 at Troy,
N.Y. (Cheney 1896), blocks passage of shad to
nontidal water. Annual Hudson River landings
were 2-4 million pounds from 1936 to 1949 and
catches of about 5 million pounds have been re-
ported (Talbot 1954). Migration of shad in the
Potomac River is blocked by Great Falls, 16 km
upstream from tidal water, so that most fish are
probably from tidal spawning. Spawning grounds
in several Virginia rivers are in tidal waters
(Massmann 1952). Therefore, it appears that tidal
spawning was once very important in the Dela-
ware River, in agreement with Walford [a 1951
memorandum cited by Mansueti and Kolb (1953)]
who stated that the principal spawning area once
was probably a short distance above Gloucester,
N.J. (km 30).

The area near Hancock apparently became an
increasingly important spawning area — but
eventually for reduced numbers of fish — as the
Delaware River shad runs declined. Many fish
again moved upstream into the East Branch after
installation of the Lackawaxen fishway in 1890

(Bean 1892, 1903). Landings from 1904 to 1913, in
general, were only about 3-5 million pounds and
consistently have been much less than 0.5 million
pounds since 1920 (Sykes and Lehman 1957;
Chittenden 1974). In spite of this great decline,
many shad (240-350/seine haul) were captured at
Hancock until 1915 (Bishop 1936). Catches near
Hancock gradually declined after 1915, and a
shad fishing club captured only 60-75 fish annu-
ally after 1920 and less than 12 in some years
(Greeley 1936; Bishop 1936).

Many tributaries, particularly in the tidal
area, may have been used for spawning and as
nurseries; but their historical importance is not
clear. Adults entered many tributaries near
Philadelphia (Meehan 1896; Stevenson 1899).
The Lehigh and Schuylkill rivers were once fa-
mous shad streams (Gay 1892; Meehan 1896), al-
though dams were constructed after 1820 and
prevented access to these streams.

Recent Spawning and Nursery Areas

With the probable exception of the most grossly
polluted tidal areas, recent spawning and nursery
areas have extended throughout fresh water of
the Delaware and into the East and West
branches. In general, nurseries must be at or
downstream of spawning grounds, because the
young begin to disperse downstream upon trans-
formation from the post-larval stage — if not
sooner (Chittenden 1969).

The chief spawning grounds and nurseries now
extend no farther downstream than Belvidere.
Gonad condition, the presence of few adults after
mid-May, and the location of the chief nurseries,
especially during early July, indicate that very
little spawning occurs as far downstream as
Lambertville. The Delaware between Belvidere
and Philadelphia probably now serves as a nur-
sery primarily due to downstream dispersal of the
young. The importance of spawning grounds and
nurseries now increases proceeding upstream
from Belvidere towards Hancock. The most im-
portant spawning grounds and nurseries are lo-
cated from about Port Jervis to Hancock and ex-
tend into the lower East Branch.

Tidal water near Philadelphia is no longer
suitable as a nursery and probably not for spawn-
ing. Although conditions vary slightly between
years, in general, the minimum daily dissolved
oxygen is at or near 0 mg/liter from about mid-
May through early December in the 66-km

348

CHITTENDEN: PRESENT SPAWNING GROUNDS OF SHAD

stretch from Torresdale, Pa., to the Delaware
Memorial Bridge, the most severely affected area
being from Chester, Pa., to the Benjamin
Franklin Bridge (Chittenden 1969). Minimum
daily dissolved oxygen levels of about 2.5-3.0 mg/
liter are needed to permit mere survival of shad,
and this is not a reasonably normal existence
(Chittenden 1973a).

Some spawning probably occurs in fresh water
seaward of Philadelphia when low oxygen pre-
vents upstream passage of part of the run. There-
fore, this area would be a nursery. The area is
limited in extent, however, and survival of fish
may be precarious because of daily dissolved oxy-
gen fluctuations due to photosynthesis or tidal
movement of polluted water, de Sylva et al. (1962)
collected larval shad, but no juveniles, in the Del-
aware River estuary shore zone even though the
euryhaline young can and do utilize brackish
nurseries (Chittenden 1973b). Production of shad
seaward of Philadelphia, at best, apparently is
small because landings in the Delaware Basin
have been low for more than 50 yr.

The West Branch is apparently no longer an
important nursery. Young shad were repeatedly
collected at Hancock in 1963, but none were cap-
tured in two collections with lights in 1964 and
1966. Cold water releases from Cannonsville Res-
ervoir, which began after summer 1963, may ac-
count for the apparent absence of young in the
West Branch thereafter (Chittenden 1972). If so,
the East Branch and possibly the Delaware below
Hancock may be of precarious suitability for
spawning and nursery purposes, because Pepac-
ton Reservoir on the East Branch is also designed
for water release from the hypolimnion.

Tributaries act as nurseries and possibly
spawning grounds but are probably not impor-
tant to production today in the Delaware River.
Compton (1963) captured 38 young on 23 July
1962 in Big Flat Brook, nearly 1.6 km from the
Delaware, and adults have been observed in sev-
eral tributaries. Tributaries in nontidal water are
too small to support many fish, however, except
for the Lehigh River (km 168) which is dammed
near its junction with the Delaware. Those in
tidal water near or upstream of the Philadelphia
area are dammed, affected by tidal movement of
low oxygen water, or the young produced therein
reach Philadelphia too early in summer or fall to
successfully pass seaward (Chittenden 1969).

The present findings on spawning and nursery
areas agree with Sykes and Lehman's (1957) ob-

servations and with their descriptions of unpub-
lished findings of Cable: plankton tows were
taken in May 1944 from Bordentown, N.J., to
Equinunk, Pa.; the greatest concentration of eggs
was above Lackawaxen and no eggs were found
below Lumberville, Pa. Therefore, it would ap-
pear that the chief spawning grounds and nur-
series have remained about the same for at least
the last 30 yr and probably since about 1910-20.

Areas Contributing to
Successful Production of Adults

It appears that there has been a fundamental
shift in the chief spawning grounds and nurseries
since the decline of the Delaware River shad
runs. Historically the chief spawning grounds
were downstream of Delaware Water Gap and in-
cluded the tidal area. These areas are now of little
importance; since the decline, the chief spawning
grounds have been upstream of Delaware Water
Gap. The most important spawning grounds and
nurseries for the last 60 yr or more have seem-
ingly been near the Hancock area.

Implications of the shift in spawning and nur-
sery areas include the existence of an intrastream
homing tendency which brings the fish back to
spawn in their general area of birth. Chittenden
(1969) discussed in detail causes of the decline in
abundance of Delaware River shad and why
abundance has remained low. I suggested
(1969:424) that the shift in spawning and nursery
areas occurred because pollution near Philadel-
phia has selected for an upstream-spawning stock
based upon the time when the young reach
the Philadelphia area; fish produced farthest
downstream have the greatest probability of
reaching Philadelphia before dissolved oxygen
improves sufficiently to permit successful sea-
ward passage. This implies intrastream homing.
Interstream homing exists in shad (Hammer
1942; Hollis 1948; Talbot and Sykes 1958; Nichols
1960), but direct evidence of intrastream homing
is desirable.

Spawning and nursery areas near Hancock are
apparently the key to maintenance of the rem-
nant Delaware River shad runs, because Chit-
tenden (1969) demonstrated that the last fish to
move seaward were, in general, those produced
farthest upstream. The extent of the spawning
and nursery area since about 1910-20 or earlier
has probably expanded and contracted depending
upon the size of the run and spawning success.

349

FISHERY BULLETIN: VOL. 74, NO. 2

Important spawning and nursery areas probably
extend farthest downstream when the run is
large and spawning is successful. The upper Del-
aware area near Hancock is probably the core
around which expansion and contraction occurs.
Downstream sections of the nursery usually
contribute little or nothing to production of adults
even if the nursery expands. Since 1925, larger
shad runs in the Delaware River have depended
upon one year class which successfully passed the
Philadelphia area (Chittenden 1975). Down-
stream nurseries contribute to production only
when water quality near Philadelphia per-
mits shad passage earlier than normal; there is
usually catastrophic destruction of the young as
they pass Philadelphia (Chittenden 1969). There-
fore, in general, it appears that nursery and
spawning areas contribute to production in pro-
portion to their distance from Philadelphia. Only
a small part of the historical nursery area now
contributes to production of adults.

Future Prospects

Future prospects of shad in the Delware River
depend primarily upon water quality in the tidal
area and upon a dam near Tocks Island (Chitten-
den 1969). The present remnant runs appear
based upon stocks that spawn far upstream in a
small part of their former spawning grounds and
whose progeny pass tidal water in late fall when
dissolved oxygen increases. A greater area would
contribute to successful production if dissolved
oxygen increased earlier, because fish spawned
farthest downstream pass tidal water first. There-
fore, the magnitude of future runs will reflect dis-
solved oxygen conditions, because the area con-
tributing to production will change accordingly. If
recent or typical water quality was maintained,
future runs would usually be small. Fortuitous
circumstances would occasionally produce larger
runs as in the early 1960's.

Construction of a dam near Tocks Island would
greatly affect shad. They probably would be ex-
tirpated from the Delaware if successful fishways
for both adults and young are not provided and
water quality in the tidal area is unchanged. Cold
water reservoir releases drastically and ad-
versely affect usage of downstream spawning and
nursery areas, if only due to avoidance (Chitten-
den 1972). Cold water releases from a Tocks Is-
land dam would shift spawning and nursery
areas far downstream, and spawning grounds

under any water release circumstances would be
downstream of the area that presently produces
adults successfully. Therefore, the young pro-
duced would reach tidal water too early to pass
seaward successfully. Great water quality im-
provement would be needed in the tidal area just
to maintain the present small runs. Water qual-
ity improvement by flow augmentation might be
self-defeating, because the young now move
downstream even during the summer; and in-
creased discharge and temperature decrease
would accelerate this. The potential would be
brighter if successful fishways were provided. The
reservoir might be an excellent nursery for the
young judging from their pelagic habits, their
preference for pool habitats, and the former im-
portance of tidal nurseries. This, combined with
nurseries upstream from the reservoir, might es-
tablish larger runs — if the young passed the dam
and tidal water successfully. However, much
larger runs would be achieved with less risk at
possibly less cost if Delaware River water quality
in the tidal area were restored and the dam was
not built. Then, the outstanding recreational po-
tential of a clean tidal area in a great population
center would be restored — and the outstanding
recreational opportunity of an unobstructed Del-
aware River would not be lost.

ACKNOWLEDGMENTS

For assisting in collections, I am deeply grate-
ful to J. Westman and J. Hoff, J. Harakal, D. Rie-
mer, J. Barker, F. Bolton, R. Coluntuno, K.
Compton, R. Gross, C. Masser, R. Stewart, J.
Miletich, S. Hoyt, L. Schulman, H. Dinje, H. Buck-
ley, J. Musick, M. Bender, J. Gift, C. Townsend,
R. Bogaczk, and K. Marcellus of or formerly of
Rutgers University, Harvard University, the New
Jersey Division of Fish and Game and/or the New
York Department of Environmental Conservation.

Fred and William Lewis, Jr. generously gave
permission to collect at their fishery at Lam-
bertville and frequently assisted in seining. W.
Kelly of the New York Department of Environ-
mental Conservation and J. Musick, then at Har-
vard University, provided observations. J.
McEachran and R. Noble of Texas A&M Univer-
sity reviewed the manuscript.

The U.S. Bureau of Sport Fisheries and
Wildlife, New Jersey Division of Fish and Game,
Pennsylvania Fish Commission, and New York
Department of Environmental Conservation

350

CHITTENDEN: PRESENT SPAWNING GROUNDS OF SHAD

kindly permitted use of data collected during the
Tri-State Surveys of the Delaware River. Finan-
cial support was provided, in part, by Rutgers
University, The Sport Fishing Institute, Dela-
ware River Basin Commission, and U.S Public
Health Service.

LITERATURE CITED

Abbot, C. C.

1868. Catalogue of vertebrate animals of New Jersey. In
Geology of New Jersey, Append. E, p. 751-830.

Anonymous.

1961. Big Flat Brook shad. N.J. Outdoors 12(2):28.

Bean, T. H.

1892. The fishes of Pennsylvania. Rep. Pa. State Comm.
Fish. 1889-90-91:1-149.

1903. Catalogue ofthe fishes of New York. N.Y. State Mus.
Bull. 60:1-784.
BISHOP, S. C.

1936. Fisheries investigations in the Delaware and Sus-
quehanna Rivers. N.Y. State Conserv. Dep., 25th Annu.
Rep. 1935, Suppl. 10:122-139.

Cheney, a. N.

1896. Shad ofthe Hudson River. Annu. Rep. Comm. Fish.
Game For. N.Y. 1895:125-134.

Chittenden, M. E., Jr.

1969. Life history and ecology ofthe American shad, A/osa
sapidissima, in the Delaware River. Ph.D. Thesis, Rut-
gers Univ., New Brunswick, N.J., 458 p.

1972. Responses of young American shad, Alosa sapidis-
sima, to low temperatures. Trans. Am. Fish. Soc.
101:680-685.

1973a. Effects of handling on oxygen requirements of
American shad (Alosa sapidissima). J. Fish. Res. Board
Can. 30:105-110.

1973b. Salinity tolerance of young American shad, Alosa
sapidissima. Chesapeake Sci. 14:207-210.

1974. Trends in the abundance of American shad, Alosa
sapidissima, in the Delaware River Basin. Chesapeake
Sci. 15:96-103.

1975. Djrnamics of American shad, A/osa sapidissima, runs
in the Delaware River. Fish. Bull., U.S. 73:487-494.

1976. Weight loss, mortality, feeding, and duration of resi-
dence of adult American shad, Alosa sapidissima, in
fresh water. Fish. Bull, U.S. 74:151-157.

COMPTON, K. R.

1963. Angler harvest comparisons on the fly-fishing only
and open fishing stretches ofthe Big Flatbrook. D. J. Job
Compl. Rep. State N.J. Proj. F-20-R-1, 37 p.

de Sylva, d. p., f. a. Kalber, Jr., and C. N. Shuster, Jr.

1962. Fishes and ecological conditions in the shore zone of
the Delaware River estuary, with notes on other species
collected in deeper waters. Univ. Del. Mar. Lab., Inf Ser.,
Publ. 5, 164 p.

ELLIS, M. M., B. A. WESTFALL, D. K. MEYER, AND W. S.

Platner.

1947. Water quality studies of the Delaware River with
reference to shad migration. U.S. Fish Wildl. Serv., Spec.
Sci. Rep. 38, 19 p.
Gay, J.

1892. The shad streams of Pennsylvania. Rep. Pa. State
Comm. Fish. 1889-90-91:151-187.

GOODE, G. B,

1888. American fishes. W. A. Houghton, N.Y., 496 p.

Greeley, J. R.

1936. Fishes of the area with annotated list. N.Y. State
Conserv. Dep., 25th Annu. Rep. 1935, Suppl. 10:45-88.

Hammer, R. C.

1942. The homing instinct ofthe Chesapeake shad, Alosa
sapidissima (Wilson), as revealed by a study of their
scales. M.S. Thesis, Univ. of Maryland, College Park,
45 p.
HILDEBRAND, S. F, AND W. C. SCHROEDER.

1928. Fishes of Chesapeake Bay. Bull. U.S. Bur Fish.
43(1), 368 p.
HOLLIS. E. H.

1948. Thehoming tendency of shad. Science (Wash., D.C.)
108:332-333.

Leach, G. C.

1925. Artificial propagation of shad. Rep. U.S. Comm.
Fish. 1924. Append. 8, p. 459-486. (Doc. 981.)
LEGGETT, W C.

1972. Weight loss in American shad {Alosa sapidissima,
Wilson) during the freshwater migration. Trans. Am.
Fish. Soc. 101:549-552.

Lehman, B. A.

1953. Fecundity of Hudson River shad. U.S. Fish Wildl.
Serv., Res. Rep. 33, 8 p.

leim, a. H.

1924. The life history ofthe shad (Alosa sapidissima (Wil-
son)) with special reference to the factors limiting its
abundance. Contrib. Can. Biol., New. Ser. 2:161-284.
Mansueti, R., and H. KOLB.

1953. A historical review of the shad fisheries of North
America. Chesapeake Biol. Lab. Publ. 97, 293 p.
MARCY, B. C, Jr.

1972. Spawning ofthe American shad, Alosa sapidissima
in the lower Connecticut River Chesapeake Sci. 13:116-
119.
MASSMANN, W. H.

1952. Characteristics of spawning areas of shad, Alosa
sapidissima (Wilson) in some Virginia streams. Trans.
Am. Fish. Soc. 81:78-93.
MEEHAN, W. E.

1896. Fish, fishing and fisheries of Pennsylvania. Rep. Pa.
State Comm. Fish. 1895:108-245.

1907. The shad work on the Delaware River in 1907 and its
lessons. Trans. Am. Fish. Soc. 36:105-118.
NEW JERSEY.

1916. Annu. Rep. Board Fish Game Comm. N.J. 1916, 26 p.
NICHOLS, P. R.

1960. Homing tendency of American shad, Alosa sapidis-
sima, in the York River, Virginia. Chesapeake Sci.
1:200-201.
PENNSYLVANIA.

1875. Rep. Pa. State Comm. Fish. 1874.

1897. Rep. Pa. State Comm. Fish. 1897.
PRINCE, E. E.

1907. The eggs and early life history of the herring, gas-
pereau, shad and other clupeoids. Contrib. Can. Biol.
1902-1905:95-110.

Slack, J. H.

1874. Notes on the shad, as observed in the Delaware
River. In Notes on the natural history of the shad and
alewife, p. 457-459. U.S. Comm. Fish and Fish. Part 2.
Rep. Comm. 1872 and 1873.

351

Smiley, C. W.

1884. Notes on the shad season of 1884, with references to
other species. Bull. U.S. Fish Comm. 4:337-341.

Smith, H. M.

1898. Report of the Division of Statistics and Methods of the
Fisheries. U.S. Comm. Fish and Fish. Part 23. Rep.
Comm. 1897:CXXV-CXLVI.

Stevenson, C. H.

1899. The shad fisheries of the Atlantic coast of the United
States. U.S. Comm. Fish and Fish. Part 24. Rep. Comm.
1898:101-269.

SYKES, J. E., AND B. A. LEHMAN.

1957. Past and present Delawsire River shad fishery and
considerations for its future. U.S. Fish Wildl. Serv., Res.
Rep. 46, 25 p.

fishery bulletin: vol. 74, no. 2
Talbot, G. B.

1954. Factors associated with fluctuations in abundance of
Hudson River shad. U.S. Fish Wildl. Serv., Fish. Bull.
56:373-413.

Talbot, G. B., and J. E. Sykes.

1958. Atlantic coast migrations of American shad. U.S.

Fish Wildl. Serv., Fish. Bull. 58:473-490.
TOWNSEND, C. H.

1901. Statistics of the fisheries of the Middle Atlantic

States. U.S. Comm. Fish and Fish. Part 26. Rep.

Comm. 1900:195-310.

Walburg, C. h., and p. R. Nichols.

1967. Biology and management of the American shad and
status of the fisheries, Atlantic coast of the United States,
1960. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 550,
105 p.

352

LARVAL DYNAMICS OF THE DUNGENESS CRAB, CANCER
MAGISTER, OFF THE CENTRAL OREGON COAST, 1970-71

R. Gregory Lough^

ABSTRACT

The larval dynamics of the economically important Dungeness crab, Cancer magister, were investi-
gated from plankton samples collected bimonthly during 1970 and 1971 along a trackline near
Newport, Oreg. Larvae appeared at maximum densities (8,000/1,000 m^) within 15 miles of the coast
in late January 1970 and remained in the plankton until late May for an approximate larval period of
130 days. The bulk of the larval population was retained in the nearshore area by the strong along-
shore and onshore components of the surface currents and to some extent by the behavior of larvae in
determining their position in the water column. During the 1971 season, larvae appeared initially
at about the same time and densities, but a mass mortality may have occurred in the early zoeal
stages coinciding with the unusually severe weather in February and March. A significant difference
between the 1970 and 1971 larval populations was suggested by analysis of covariance using sea
surface temperature and salinity as environmental variables. However, the effect of the low tempera-
ture and salinity values that occurred during the winter of 1971 were not clearly indicated by
multiple regression analyses of laboratory experimental data to be the prime factors directly affect-
ing larval survival. Neither did a gut-fullness study of planktonic larvae substantially explain the
1971 larval mortality. Therein various hypotheses are explored in view of the present knowledge of
processes affecting larval survival and recommendations are suggested for further research.

It is well known that many species of economi-
cally important marine resources fluctuate
greatly in number and location. These fluctua-
tions may be explained in part by changes occur-
ring in the larval populations. That the larval
stage is the most critical period for the majority of
marine animals was originally emphasized by
Hjort (1914, 1926) for fish larvae and by Thorson
(1946) for marine invertebrate larvae. Survival
through this period is usually considered the
major factor in determining the strength of the
year class. The causes or extent of larval mortal-
ity, however, are still relatively unknown.

Bimonthly plankton samples were collected
from 1969 through 1971 along a transect off the
central Oregon continental shelf to document the
species of crab larvae present, their seasonality
and abundance, and their onshore-offshore dis-
tribution in relation to seasonal changes in
oceanographic conditions (Lough 1975b). A major
effort was made to assess the larval population of
the Dungeness crab. Cancer magister Dana, as it
supports one of the most important fisheries in
the Pacific Northwest.

Cancer magister occurs along the Pacific coast

'School of Oceanography and Marine Science Center, Oregon
State University, Corvallis, OR 97331; present address: North-
east Fisheries Center, National Marine Fisheries Service,
NOAA, Woods Hole, MA 02543.

fi'om Unalaska to lower California and ranges
from mean low water to 50 fathoms (91 m)
(Schmitt 1921). Although it prefers sandy or
sandy-mud bottoms of the nearshore area, speci-
mens have been found on all bottom types within
estuaries and on the continental slope. Adult fe-
males generally reach maturity by their second
or third year and may produce three or four
broods during a life-span (MacKay 1942; Cleaver
1949; Butler 1960). Egg-carrying females are
found in Oregon waters from October to March
with essentially one brood produced per year
(Waldron 1958). Field observations (Waldron
1958) and laboratory rearings by Poole (1966)
and Reed (1969) indicate that larvae hatch off
northern California and Oregon from January
through March and are present in the nearshore
waters through July for a total estimated larval
life of 128 to 158 days. Cancer magister passes
through five zoeal stages and one megalops dur-
ing its larval development before settling out
of the water and metamorphosing to the benthic
juvenile.

HYDROGRAPHIC FEATURES
OF STUDY AREA

The surface waters along the U.S. west coast
are dominated by the California Current; a slow,

Manuscript accepted September 1975.
FISHERY BULLETIN; VOL. 74, NO. 2, 1976.

353

FISHERY BULLETIN: VOL. 74, NO. 2

broad, and shallow current flowing equatorward
(Wooster and Reid 1963). The nearshore currents
vary seasonally and are dependent upon wind di-
rection and strength. During the fall and winter
months when the winds are predominantly from
the southwest, a subsurface countercurrent flow-
ing northward along the coast develops into the
Davidson Current. Drift bottle studies by Wyatt
et al. (1972), Burt and Wyatt (1964), and
Schwartzlose (1964) indicate that the Davidson
Current first develops along the Oregon-Wash-
ington coast in September reaching maximum
speeds between 0.5 and 2 knots within 20 miles of
the coast during the month of November.

The major change in the surface currents from
northward to southward occurs in March and
April (Wyatt et al. 1972). The phenomenon of
coastal upwelling occurs when the northwesterly
winds intensify and sometimes persist from May
to September. As the surface waters are trans-
ported offshore and to the southwest, cold, high
salinity waters from below a permanent pycno-
cline (60-100 m) are brought to the surface (Smith
et al. 1966). This zone of active upwelling occurs
within 20 miles of the coast but its effects can be
observed to the edge of the slope.

The area within 5 miles of the coast has not
been studied in much detail but is believed to be
dominated by mixing processes (Mooers 1970).
The surface currents are generally well corre-
lated with the wind direction, but tidal currents
predominate when the wind is reduced. A very
strong alongshore current with an onshore com-
ponent is indicated within 3-5 miles of the coast
(Keene 1971; Wyatt et al. 1972; Holton and
Elliot 1973).

The dominant processes modifying surface
water properties off" the Oregon coast during the
winter are rainfall and river runoff; while during
the summer, the major processes are upwelling in
conjunction with heating and mixing with the
Columbia River plume water (Pattullo and Den-
ner 1965). Surface temperatures and salinities
taken on early life history cruises from June 1969
through August 1971 at stations NHOl-NHlO are
presented in Figures 1 and 2. Temperatures
range annually from about 7° to 17°C and are
highest from May through October, peaking in
September. More variability is evident during the
summer due to surface heating interrupted by
local upwelling of near 7°C bottom water. Surface
salinity values are generally low during the
winter and high in the summer reflecting sea-

15 —

10 —

1 — I — I — rn — I — I — r~i — i — r~i — r

"1 I I I I r

JJASONDU rMAMJ JASOND|J FMAMJJA

1969

1970

1971

Figure l.— Surface temperature (°C) at stations NHOl, NH03,
NH05, and NHIO from June 1969 through August 1971.

sonal precipitation and upwelling, respectively.
The annual range of salinity is from about 25 to
35"L. Low salinity values at stations NH03 and
NH05 from November through April are proba-
bly associated with the Yaquina Bay plume
which flows north along the coast during the
winter (Kulm and Byrne 1966).

METHODS

Sampling Program and Gear

This study was conducted primarily on a
trackline off Newport, Oreg. (lat. 44°39.1'N)
across the continental shelf and slope. The 12 sam-
pling stations are designated on the Newport
Hydrographic line (NH) in Table 1, which cor-
respond in distance to nautical miles from the

354

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

NHIO

J ' j'aIs'o'n'd|j'f'm'a'm'j'j'a's'o'n'd|j'fIm'a'm'j'jIa '
1969 1970 1971

FIGURE 2.— Surface salinity (Z.) at stations NHOl, NH03, NH05,
and NHIO from June 1969 through August 1971.

TABLE 1. — Location of plankton sampling stations and bottom
depths along the Newport Hydrographic line (NH) off Newport,
Oreg.

Station
(Lat. 44'39.1'N)

Long.

Depth
(m)

NHOl

12405.4W

NH03

124^086W

NH05

124 10.7W

NHIO

124=17.7'W

NH15

124 24.yW

NH20

124^31. 7'W

NH25

124=387W

NH30

124=45.7'W

NH35

124=52.7'W

NH40

124 59.7W

NH50

125°13.7'W

NH60

125°27.7'W

142

330

220

340

1,060

1,300

2,850

coast. Plankton samples initially were collected
at the four inshore stations (NHOI-NHIO) con-
stituting the main series of samples from June
1969 through August 1971. The sampling pro-

gram was extended offshore to NH60 by 5- or 10-
mile intervals beginning with the 3 February
1971 cruise.

A high-speed bongo net sampler (Posgay et al.
1968) with a 0.2-m mouth diameter was used ex-
clusively from 22 June 1969 through 20 October
1970. The two cylinder-cone nets, 1.8 m in length,
were constructed of 0.233- and 0.571-mm nylon
mesh and had an effective straining surface (pore
size area) to mouth area ratio of ca. 10 to 1. A
30-pound lead ball or a 15-pound V-fin depressor
was attached to the sampler line.

Starting with the 4 November 1970 cruise, a
0.7-m diameter bongo net sampler was used in
conjunction with the 0.2-m sampler to strain a
greater volume of water and to reduce avoidance
by the larger larvae. The 0.7-m bongo nets had a
net length of 5.1 m, were constructed of 0.571-mm
nylon mesh, and had an effective straining area
ratio of ca. 8 to 1. Both samplers were equipped
with TSK^ flowmeters mounted on brackets 18
cm from the rim of the inside frame. A multiplane
kite-otter wire depressor (ca. 80 pounds), modified
after Colton ( 1959), was used with the dual bongo
net array to produce a wire angle ratio of 2 to 1.
The sampling objective was to make a high speed,
oblique, plankton tow, sampling the water col-
umn in equal stepped intervals from 150 m depth,
or in shallower areas from bottom to surface.
Wire was let out and retrieved at 50-75 m/min
while the vessel was underway at 2-3 knots. Most
of the samples represent daylight (0600-1800)
tows ranging in duration from 10 to 25 min. The
longer tows were generally made on stations
beyond 5 miles. Plankton samples were im-
mediately preserved in 5-109'f Formalin and later
buffered with sodium borate.

A bathythermograph (BT) cast was made at
each station near bottom or to 150 m depth. Sur-
face bucket temperatures also were taken at each
station to calibrate the BT readings. Salinity
samples were collected on the surface and near
bottom or to 150 m depth by a Nansen bottle cast
and analyzed by an inductive salinometer. Salin-
ity, temperature, and depth (STD) data from a
real-time printout computer were available for
several cruises.

The Nekton Cruise of 11-12 April 1970 at sta-
tion NH45 was included in this study as it is one
of the few cruises that sampled the offshore

^Tsurumi-Seiki Kosakusho. Reference to trade names does
not imply endorsement by the National Marine Fisheries Ser-
vice, NOAA.

355

plankton during 1970. The objective of this cruise
was to identify those organisms associated with
sound scattering layers in the upper 150 m of
seawater and, if possible, to follow their day-
night migration patterns. Six successive inte-
grated tows of approximately 45 min each were
taken to a depth of 150 m (total time: 1852-2355).
A standard 6-foot (1.8-m) Isaacs-Kidd mid- water
trawl (IKMT) with a 2.9 m^ mouth opening [1%-
inch (3.8-cm) mesh with a V4-inch (0.6-cm) linear
nylon liner] was used for this series. The second
series of eight samples alternately sampled from
surface to 150 m and from 150 m to surface with
an eight-bar electronic multiple plankton sam-
pler (EMPS) attached to the IKMT (Pearcy and
Mesecar 1970) (total time: 0134-0514). The
cylinder-cone nets were approximately 2.9 m in
length with a mouth diameter of 0.4 m and made
of 0.57 1-mm nylon mesh. Another series also
used the EMPS to sample eight discrete layers
from the surface to 330 m depth covering three
bands of scatters (total time: 0640-1113). Scatter-
ing layers were located using 12 and 38.5 kHz
echo sounders. One automated STD cast was
made.

Processing of Plankton Samples

Samples from both mesh sizes of the 0.2-m
bongo nets were processed for the nearshore area,
stations NHOI-NHIO. Only one side of the 0.7-m
bongo net sampler was processed to examine the
offshore area, NH15-NH60. Generally, the entire
sample was sorted, however, many required sub-
sampling using an 8-cm diameter plankton split-
ter (Longhurst and Seiburt 1967). Approximately
22% of the 0.2-m bongo net samples and 39% of
the 0.7-m bongo net samples required subsam-
pling. Those samples which required splitting
were usually from stations NHOl and NH03.

All crab larvae were removed from the samples
and positive identification of C. magister larvae
was made from the descriptions given by Poole
(1966) and from preserved specimens reared by
Thomas F. Gaumer, Fish Commission of Oregon,
Marine Laboratory, Newport, Oreg. Catches of
larvae were first converted by computer to
number per 1,000 m^ of seawater and ordered in a
format. Graphs of stage density against time
were plotted for the 0.2-m bongo net samples,
0.57 1-mm mesh, with the aid of a CalComp plot-
ter using the Oregon State University CDC3300
computer.

FISHERY BULLETIN: VOL. 74, NO. 2

SAMPLING VARIABILITY

The detailed analyses of the various methods
by which sampling variability affected the esti-
mates of larval crab abundance are given by
Lough (1975b). Variability estimates and sam-
pler comparisons were made in this study on
other species of crab larvae than C. magister for
the most part, as limited ship time and weather
played an important role in determining the ob-
jectives and priorities of the sampling program.
Analysis of variance techniques were used to es-
timate the variance of a single observation in the
manner of Winsor and Clarke (1940). Confidence
limits for a single observation of either sampler
usually exceeded the 50-200% range reported by
Winsor and Clarke due to the relatively low den-
sities of crab larvae sampled during replicate
tows. A range of an order of magnitude was con-
sidered necessary to distinguish a real difference
between any two observations. There was no sig-
nificant difference between the total number and
kinds of crab larvae caught by the two sides of the
different sized samplers. The 0.7-m bongo net
sampler gave smaller confidence limits for larval
crab catches and was much superior in establish-
ing significant differences between stations than
the 0.2-m bongo net sampler.

Most of the nearshore samples (NHOl-NHlO)
were taken during daylight hours; only 8.6% of
the 0.2-m bongo net samples were taken at night
between 1800 and 0600 h. More (26.7%) of the
0.7-m bongo net samples sorted beyond NHIO
were collected at night. Most larvae were caught
more abundantly in night tows than day tows for
both sized samplers. Day-night differences in lar-
val abundance were greater for the 0.2-m sampler
than the 0.7-m sampler. There was a nearly equal
distribution in the number of kinds of crab larvae
caught between day and night samples using the
0.7-m sampler; however, using the 0.2-m sampler,
significantly more kinds of larvae were caught at
night.

The results of the Nekton Cruise showed that
the larvae of C. magister occur in relatively low
densities offshore as far as station NH45 during
early April 1970. They are most likely to occur in
the surface waters above 120 m, the depth of the
thermo- and halocline and are probably as-
sociated with the first sound scattering layer at
25 to 90 m depth. A Mann- Whitney two sample
rank test (Tate and Clelland 1957) retained the
null hypothesis that there was no significant dif-

356

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

ference between the number of C. magister
megalopae, or the total number of larvae caught
in the first two series of tows. In the second series
of samples more larvae were caught towing from
surface to 150 m than from 150 m to svuface and
the total number of larvae decreased with time
(0134-0514) for both alternate types of tows. The
coefficients of variation (standard deviation/
mean) for the total number of larvae were about
the same for the first two series of tows (1.25 and
1.31, respectively) indicating a somewhat patchy
distribution of the larvae in the upper 150 m of
water at night. Very few larvae were caught dur-
ing the third series of tows.

RESULTS

Distribution and Abundance of
Cancer magister Larvae

Two larval seasons were encompassed by the
sampling program (Figure 3). Zoea 1 larvae made
their first substantial appearance during the first
season on 29 January 1970 at stations NH03,
NH05, and NHIO with maximum densities rang-
ing from 1,000 to 3,000/1,000 m^. The subsequent
zoeal stages were found most abundantly at sta-
tions NH05 and NHIO. Few zoea 4 and no zoea 5
stages were found at any of the four inshore sta-
tions. In general, the number of larvae captured
decreased from zoea 1 through 5. However, large
numbers of megalopae were found at stations
NHOl, NH03, and NH05, suggesting a general
inshore transport of larvae during this season.
Maximum densities of the magalopae ranged
from 1,000 to 8,000/1,000 m^, densities compara-
ble to those of the zoea 1 stage found earlier in the
year. Few megalopae appeared in the water col-
umn after 22 May 1970 and none after 16 July
1970. This indicates that the length of the larval
period in the plankton is approximately 130 days
(89-143 days). The summer upwelling conditions
did not appear to have any effect on the larvae
since the bulk of the megalopae had settled before
the onset of intense upwelling.

The major appearance of zoea 1 larvae during
the second season occurred at about the same
time (18 January 1971) and stations (NH03,
NH05, NHIO), and at about the same densities
(1,000-2,000/1,000 m3). However, the density of
the larvae appeared to decrease more rapidly at
zoeal stages 2 and 3, and virtually no larvae of
any stage were found after zoea 3. The 30 March

1971 cruise was the last sampling period which
caught any significant number of larvae. Very few
megalopae were found at any station through-
out the summer in day or night samples.

Cancer magister was the most abundant crab
larvae caught at station NH45, 11-12 April 1970
(Nekton Cruise). Its megalopae had the highest
densities of any larval stage with 19/1,000 m^,
followed by zoea 5 at 12/1,000 m^. Fewer zoea 4
and 3 were present. Scattered occurrences of all
larval stages were present the following year,
1971, to 60 miles offshore in the 0.7-m bongo net
samples. Megalopae and zoea 3-5 predominated
offshore with densities usually much less than
200/1,000 m^, suggesting that these larvae had
originated nearshore and subsequently drifted
offshore. Larvae present at stations NH35 to
NH60 are under the influence of the Columbia
River plume as indicated by the warmer tempera-
tures and lower salinities measured at these sta-
tions during the sampling period.

All observations indicate a dramatic difference
in the abundance of megalopae between the 2 yr.
Sampling was much more intensive during the
1971 season from the standpoint of day-night re-
plicate tows using both size samplers in the in-
shore and offshore areas when the megalopae
were sparse.

Climate and Hydrography 1970-1971

The winter of 1971 along the Oregon coast was
generally more severe than that of 1970.
Climatological records (U.S. Environmental Data
Service 1970,1971) for Newport and other ports of
Oregon show monthly mean air temperatures for
February and March 1971 to be substantially
lower than the same months during 1970. Also,
total precipitation generally was greater during
the winter of 1971 but showed considerable var-
iability along the coast. Ocean surface tempera-
tures correspondingly were much colder during
this period in 1971 than 1970. Conor et al. (1970)
and Conor and Elvin (197 1)^ reported Agate
Beach, Oreg. mean surf temperatures and Wyatt
and Cilbert (1971, 1972) reported monthly mean
surface temperatures for various ports along the
Oregon coast to be as much as several degrees
lower during the later winter of 1971 than 1970.

^Gonor, J. J., and D. W. Elvin. 1971. Inshore sea surface
temperature and salinity conditions at Agate Beach, and
Yaquina Head, Oregon in 1971. Unpubl. data. School
Oceanogr. Oreg. State Univ.

357

FISHERY BULLETIN: VOL. 74, NO. 2

5
4

2
1

I-

5
4--

3-
2-
1 ■•

o
o
o 5

• 3-

I--

5 4-
to 3-

5 +
4
3

2-
1 ■■

MEGRLOPn

JJflSONDJ F M fl M J J fl 5 5 N D J T

"" ■ MflMJjR

ZOEfl 5

5t
4--
3-
2-
I-

4 +
3
2
1 +

MEGflLOPn

W »»»«»»»«» — »»!»«»« I «» »»»« «»«t I

ZOER 5

a»ww m » wo ♦««»<• i — m — » — g^ga^*^** — ■ ■*•• »»i — # — ea — <■•> •!

J J"R s' 0 N d"j'f">i'r m J j'fl's "o n"d"J ^mTm J^'J'fl

ZOEn 4

JJflSONOJFMRMJJfiSONDJFMRMJJR

ZOEfl 3

o 5

\ 4

• 3

5 2

JJflSONOJFMflMJJfiSONDJFMflM-JjR O

ZOEfl 2

o 3

5-
4-

2-
1 •■

u»euB» ♦■ ■♦- fl»<B w »e # »#» I — m p<»aaaa»aa»B — ■ •wawt — • — »a»a ti<» ai

JJflSONDJFMRMJJRSONOJFMflMJjR

ZOEfl 1

5-
4-
3-
2-
1-

J T R S 0 N D J F M fl M J J fl S " N 0 J F M fl M J j"

ZOEfl 4

■ mW »a»a o»» » ♦ a<» i a^ aip «>a»aaaaaa» — 0 awe »at „ — , , _ -^ »-,

JJRSONDJFHRMJJRSONDJFMRMJJR

ZOEfl 3

Kuuwu wa»aata>ia # — a '«a 1 a<a sm mB999«mB«» — 0 awa »at a# — »»*a<a»»»H

JJflSONDJFMRMJJRSONDJFMRMJJfl

ZOEfl 2

ail)— »a»a WW » — » '«a 1 at aw wa^ageaaaB — ■ aw' »af a» — »a»a aaaa ai

JJflSONDJFMRMJJRSONOJFMRMJJR

ZOEfl 1

JflSONDJFMRMJJflS^NDJFMflMJJfl

1969 1970 1971
CflNCER MflGISTER flT NHGl

»a — ataafc » — ^-a^

I a^ Bp aj e#

aqaaeaa B »»

♦ — >•#•<> aaai

JJRSONDJFMRMJJflSONDJFMflMJJfl

1969 1970 197:
CflNCER MflGISTER flT NH03

FIGURE 3.— Density of Cancer magister larvae at stations NHOl, NH03, NH05, and NHIO from June 1969 through August 1971,

collected with the 0.2-m bongo net sampler, 0.57 1-mm mesh.

Salinity values show^ed considerable variability
among stations and months such that a
generalized trend could not be observed between
the two seasons. The anomalous winter of 1971
was further substantiated by Bakun's (1973) in-
dices of coastal upwelling intensity for selected
locations along the west coast of North America
based on offshore Ekman surface wind transport
from monthly mean surface atmospheric pressure
data. Positive values indicate periods of coastal
upwelling whereas negative values indicate
downwelling. January and February of 1970 at
lat. 45°N, long. 125°W show significantly greater
negative indices (-98 and -71, respectively)
than the same period in 1971 (-32 and -16, re-
spectively). High negative values are indicative
of strong downwelling along the coast which

358

Bakun stated would accelerate the southward
flow. In either case, more offshore surface water
would be transported onshore. During the March
transition period, the 1970 index was normal
( + 1); however during 1971 an anomalously high
negative index (-49) occurred. This indicates
that downwelling and subsequent transport of
surface waters was more intense during March of
1971 than 1970. Downwelling also was more in-
tense during March 1971 than in the previous 2
mo of that year. Drift bottle data compiled by
Wyatt et al. (1971) reported a 14.7% return for
bottles released off Nevi^jort from 25 February to
3 March 1970. By contrast, a 28.6% return oc-
curred during 6-9 March 1971. The average per-
cent return of drift bottles on all stations west of
Newport, 1961-71, during both February and

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

ql MEGfiLOPfl

3-
2-

I--

5
4
31

2
1

^ 3

^ 2

o '

o 5

\ 4

• 3

5 2

o 3

-J 2

5
4
3
2
1

ZOEn 5

JJRSONDJFMRMJJRSONDJFMflMJJfl

ZOER 4

* '<• t — ^»0 a»e»ea»a»»» — ■ 9mm »*i —

JJflSONDJFMflMJJfiSONOJFMflMJJfl

ZOEn 3

° 5|
\ 4
• 3
§2
1 +

JJRSONDJFMflMJJflSONDJFMRMJJfl CD

ZOER 2

^Bta »e »afliaa>»# ^M' t ^»frepq»»ag«»gp — » awe ii ' ) — 1# — '»» ae a»a» ai

JJflSONOJFMflMJjRSONDJFMflMJJR

ZOER 1

o 3

^ 2

»a»ama»fl

JJRSONDJFMRMJJRSONOJFMRMJJfi

1969 1970 1971
CRNCER MRGISTER RT NH05

MEGRLOPfl

JJHSONDJFMFlMJJfi Si N 0 J f M R M J J |!

ZOER 5

j'Tp's' o'n'd' j7"m' b"m" j"r •R"s"rN'iS'rr n" fi'T'j'j"^'

ZOER 4

JJHSONDJFMRMJJflSON 0 J P M AM

J J fl

ZOER 3

■ iMu » • •»» ♦ ♦'4-' — I — oai aaaaaaaaiaaa — ta»aaat '< aaaaaiaaaa

JJflSONDJFMflMJJflSOND J F M R M J J fl

ZOER 2

«M — • ^ ' 1 ' — I — ••-•^ — — »■ •■••!>■■ — » •!'• » ' I ' H »f > l>»»

JJflSONDJFMRMJJfiSONDJFHflMJJR

+•— ( — f I «»■

ZOER 1

JJflSONDJFMflMJJR S" NDJFHRMJJfl

1969 1970 1971
CANCER MRGISTER RT NHIO

March was 18% (Wyatt et al. 1972). High percen-
tages of returns near 3(f/c were observed only
during February and March 1961, 1962, 1963,
and 1967.

April 1970 showed a high positive index value
(4-25) indicative of upwelling processes, whereas
downwelling was still in process during the same
month in 1971 (-2). However, by May 1971 the
upwelling intensity was twice the magnitude of
that in 1970 ( + 66 and +33, respectively). In all
regards the year 1971 can be considered the most
anomalous whereas 1970 can be considered the
least anomalous of the sampling period and the
most typical over a 20-yr span. Kukla and Kukla
(1974) reported large-scale global anomalies in
weather patterns developing early in 1971. Snow
cover in the northern hemisphere increased
dramatically for the months of February, March,
April, and September 1971.

Larval Population Analyses
Between 1970 and 1971

Despite the rather restricted data set, a rigor-
ous statistical analysis is attempted at this point
to explore the relative importance of some envi-
ronmental variables associated with the C. ma-
gister larval populations. An attempt is made to
examine potential causative factors underlying
the difference in larval abundance between 1970
and 1971 seasons. A basic assumption in the
analysis is that the larval data collected in a
single sampling transect are representative of a
much larger homogeneous area. Patches of larvae
may be quite localized so that differences in lar-
val abundance from year to year may be due to
dispersal and not mortality caused by an en-
vironmental variable per se. However, the dis-
tribution of adult breeding populations are

359

FISHERY BULLETIN: VOL. 74, NO. 2

confined to shallow waters less than 50 m depth
and appear fairly uniform along the Oregon coast
based on commercial landings of legal-sized
adults (Waldron 1958). This implies that the dis-
tribution of larvae along the entire Oregon near-
shore area would be relatively homogeneous from
year to year. Wind induced turbulence and mix-
ing would tend to increase the homogeneity of the
larval population despite any initial patchiness.

If we assume that the total number of C magis-
ter larvae combined over the four inshore stations
(NHOl, NH03, NH05, NHIO) is representative of
the total population on a local basis, then the
question may be asked whether there is a signif-
icant difference in the population means between
the 2 yr, 1970 and 1971, and can a difference
be explained using the concomitant observations
of time, temperature, and salinity?

An analysis of multiple covariance was used to
test this hypothesis on two sets of data for C.
magister larvae. The first set of data compares the
sampling period from 29 January 1970 to 29 July
1970 with that of 18 January 1971 to 21 July
1971. This period includes, for these 2 yr, the first
major larval release through the time at which no
megalopae were present in the water column.
Larval density estimates from both sizes of mesh
of the 0.2-m bongo net sampler were used in the
analyses. Surface temperatures and salinities
comprised the only complete data set for the two
larval seasons and the average values of the four
inshore stations were used for each sampling
period. Nevertheless, sea surface temperatures
and salinities are representative of nearshore
subsurface conditions during the winter period
from November through March-April as exten-
sive wind mixing occurs in the shallow areas pro-
ducing isothermal conditions (Renfro et al. 1971).
During the spring and summer, a weak thermo-
cline of less than 2°C exists in the nearshore area
(<20 m). Larval and environmental data used in
the analyses are given in Appendix Table 1.

The mathematical model used for the initial
analysis was of the form:

Y = b + bo(y) + bi it) + b^m + b:,(S)
+ b^{T^) + 65(52) + b^{T X S)

where, Y = logio(X -I- 1) number of larvae per
4,000 m^ of water, 6 = a mean effect, y = a year
effect, ^ = a time effect (days elapsed since 1
January), T = linear effect of sea surface temper-
ature (°C), S = linear effect of sea surface salinity

360

(%o), T"^ = quadratic effect of temperature, S^ —
quadratic effect of salinity, and T x S = interac-
tion effect between temperature and salinity.

The b's in the model were estimated from a
general linear hypothesis testing computer pro-
gram contained in the Oregon State University
Statistical Program Library. Various hypotheses
can be specified by the user to test the importance
of the individual parameters in the model.

A summary of the analysis on the initial run is
given in Table 2. A highly significant difference
{1% level) was found between j' means after being
adjusted for all the covariates in the model. How-
ever, only t was found to be highly significant in
explaining the yearly difference. That is, the ap-
pearance of larvae in the plankton was of shorter
duration in 1971 than in 1970. Subsequently, a
new model was generated using only f as a
covariate:

Y = b + bo(y) + 61U).

The importance of t was again found to be
highly significant in explaining the difference be-
tween y population means of C. magister larvae
(Table 3).

Table 2. — A comparison of the total number of Cancer magis-
ter larvae for 1970 and 1971 (January through July) by analysis
of multiple covariance (full model).

Source of
vanation

Degrees of
freedom

Sum of
squares

Mean
square

f-level

12.983

15.079

1.323

1.537

0.120

0.140

0.513

0.594

0.296

0.344

7 X S

1.303

1.513

y (adjusted)

9.074

10538'

Residual

37.887

0.861

"F 99(1.44) - 7.12

Fitted model: / = -11.313 +
0.043(72)

0.470(y) - 0.018(f) - 5.076(7) + 2.576(S) +
- 0.060(S2) + 0.127(7;«S).

Mean of covariates

Year Mean Y t

T S 72 S^ T < S

1970 219518 12421

1971 1.57263 10938

10.36 32.23 109.05 1,040.49 333 13
10.01 31.71 102.74 1,007.00 316.70

Table 3. — A comparison of the total number of Cancer magis-
ter larvae for 1970 and 1971 (January through July) by analysis
of multiple covariance (reduced model).

Source of
vanation

Degrees of
freedom

Sum of
squares

Mean
square

f-level

y (adjusted)

Residual

45.149

9.448

42.271

45.149
9.448
0.863

52.336"
11.218"

'^99(1 491

Fitted model:

= 7.17

y = 3.856 + 4

366(y) - 0.017(f).

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

The second data set compares the sampling
period 29 January-2 May 1970 with 18 January-
14 May 1971. The period selected compares the
larval period prior to summer upwelling,
eliminating the erratic surface temperature and
salinity fluctuations. Most of the C. magister lar-
vae are megalopae by early May.

The same full model was used in the initial run
for the second data set and is presented in Table
4. There was a significant difference (5% level)
between y means after being adjusted for all the
covariates in the model. The covariates, t, T, and
T X S were all significant.

Table 4. — A comparison of the total number of Cancer magis-
ter larvae for 1970 and 1971 (January to May) by analysis of
multiple covariance (full model).

Table 5. — A comparison of the total number of Cancer magis-
ter leirvae for 1970 and 1971 (January to May) by analysis
of multiple covariance (reduced model).

Source of
variation

Degrees of
freedom

Sum of
squares

Mean
square

F-level

t
T
S

T y S

y (adjusted)

Residual

6.807
5.277
0.156
0.018
0.012
5.183
6.260
25.023

6.807
5.277
0.156
0.018
0.012
5.183
6.260
0.962

7.073-

5.483-

0 162

0.019

0.012

5385-

6.504-

'F 95(1,26)

Fitted model

= 4.22

Y = 180.944 +
0.037(7"2)

0.7^2{y) - 0.024(() - 20.294(7-) -
- 0.022(S2) + 0.656(7" x S).

4.907(S) -

Mean of covariates

Year

=an V f

T S

7 X S

1970
1971

2.80494 90.56
2.09844 80.13

10.35 31.76
9.59 31.28

107.91
92.82

1,010.36
979.08

32808
299.81

The initial model was reduced to the following
form:

Y = b + boiy) + bjt) + bHT)
+ baiS) + b^(T X S)

which greatly increased the significance of the
parameters in the final model (Table 5). A highly
significant difference (1% level) was found be-
tween _y means after being adjusted for all the
covariates. In explaining the difference between j
means of C. magister larvae, the covariate t was
most significant (1% level) followed by T and S,
and T X S at the 5% level.

The foregoing analyses support the contention
that there was a significant difference between
the C. magister larval populations of 1970 and
1971. Fewer larvae appeared in 1971 and they
appeared in the plankton for a shorter period of
time suggesting widespread larval mortality.
This apparent larval mortality was associated by
these analyses with the colder surface tempera-

Source of

Degrees of Sum of

Mean

variation

freedom squares

square

F-level

1 7.629

7.629

8.530"

1 5.859

5.859

6.551-

1 5 230

5230

5.845-

T y S

1 5.774

5.774

6.456-

y (adjusted)

1 8.650

8.650

9.672-

Residual

28 25.043

0.894

'F 95(1.281

= 4.20; ••F53„je,= 7.64

Fitted model

Y = 201.891 + 0.705(yj - 0.023(f)
0.641(7 X S).

- 20.547(7) -

- 6.148(S) +

tures and lower salinities that occurred during
the winter of 1971. The direct effects of tempera-
ture and salinity on larval survival will be
explored in the next section.

Temperature-Salinity Tolerance of
Laboratory-Reared Larvae

A laboratory study by Reed (1969) determined
the effects of temperature and salinity on the lar-
val survival of C. magister. However, it was
necessary to assess more thoroughly the effects of
these factors on survival during development and
to extrapolate from Reed's data in order to derive
better estimates of larval survival at the low
temperatures that occurred during the 1971 sea-
son. The response surface technique used in the
analysis of his data is not only valuable in its
predictive role, but also visually represents any
change in response at various stages of develop-
ment. Details of this response surface technique
and its application to the study of marine ecology
are discussed by Alderdice (1972).

A multiple regression analysis was applied to
Reed's (1969) survival data of C. magister after
20, 30, 40, and 50 days of culture at experimental
conditions. The mathematical model used in the
analysis was of the form:

Y = bo + biiS) + b^iT) + 63(52)
+ 64 (T^) + 65(8 X T)

where, Y = percentage survival, 60 = a constant,
S = linear effect of salinity, T = linear effect of
temperature, S^ = quadratic effect of salinity, T^
= quadratic effect of temperature, and S x T =
interaction effect between salinity and tem-
perature.

The 6's in the model were estimated by a step-
wise multiple regression computer program.
Further details of the regression analysis are

361

given by Lough (1975a). The calculated regres-
sion coefficients from a particular equation are
fitted by computer to a full quadratic equation in
temperature and salinity in order to print a con-
tour diagram of the response surface. Tempera-
ture and salinity scales on all plots were set to
range beyond the experimental conditions in
order to facilitate response comparison and to
allow the overall form of the surface to be vi-
sualized. Contours extrapolated beyond the ex-
perimental data lie outside the dotted lines.

A summary of the multiple regression analyses
on survival after the various periods of rearing
and the response surfaces are given in Table 6
and Figure 4. The analyses indicated that after
20 days of rearing under the experimental condi-
tions S and S^ were the two most important vari-
ables in the model. T and S x T were of lesser
importance but still contributed significantly to
the model. Analyses of the later rearing periods of
C. magister emphasized the effect of temperature
and showed the decreasing importance of both S
and S^ and S x T. This trend is more evident
when one compares the response surface plots
from 20 through 50 days of rearing. After 20 days
of rearing, the response surface contours are
nearly circular, with a slight tilt to the main axis.

FISHERY BULLETIN: VOL. 74, NO. 2

indicating a small interaction effect. The axis of
the contours tilts progressively towards the
temperature axis until, at 50 days of rearing, the
contour axis is almost perpendicular to the tem-
perature axis. Also, the survival contours progres-
sively constrict about the temperature axis with
time showing the narrowing of the temperature
range tolerated by the larvae. Maximum survival
(80% contour) at 20 days is predicted to occur be-
tween 6.5° and 17.5°C and 21.5 and 35.01., while
at 50 days, maximum survival is predicted to
occur between 9.0° and 15.0°C and above 28.5'L.
The area of maximum survival (80% contour)
shifts somewhat during the 20- to 50-day period
from an initial low salinity-wide temperature
range to a high salinity-low temperature toler-
ance. However, when the 20- and 50-day survival
polynomials were tested by an analysis of
covariance (Ostle 1963:205), they were not found
significantly different in their response (Table 7).
In summary, salinity appears to exert an im-
mediate effect on C. magister larval survival,
while the effect of temperature becomes increas-
ingly important with time.

Survival at a given temperature, salinity, and
time can now be estimated using the fitted equa-
tions. All of the fitted equations for the four time

Table 6. — Multiple regression analyses of Cancer magister larval survival in 20 temperature and salinity

combinations.

Regression

Degrees of

Significance

step number

Vanable

fl2

F-value

freedom

level

Coefficients

f-value

level

20 days

0.505

18.378

(1,18)

29.4369

4069

0.591

3.723

(2.17)

-0.4720

3040

0.659

3.030

(3,16)

N.S.

-0.7068

4635

0.834

15.819

(4,15)

23.4636

4.559

S xT
Constant

0.865

3.272

(5,14)
30 days

-0.2277
-457.6092

1.809

NS.'

0.417

12.878

(1.18)

18.3726

2.026

N.S.

0.529

4.044

(2,17)

-0.6903

3.611

0.702

9 290

(3,16)

23.0272

3.569

S X 7

0.744

2.443

(4,15)

N.S.

-02503

1 586

N.S.

Constant

0.768

1.446

(5,14)
40 days

N.S.

-0,2340
-335.2887

1 202

N.S.

0.416

12.830

(1.18)

143243

1.602

N.S.

0.491

2.511

(2,17)

N.S.

-08113

4 305

0.744

15.824

(3,16)

25.5095

4011

S ' 7

0.768

1.509

(4,15)

N.S.

-0.1892

1.217

N.S.

S2
Constant

0779

0.713

(5,14)
50 days

N.S.

-0.1620
-3138493

0,844

NS.

0.373

10.717

(1,18)

13,4195

1 687

N.S.

0.432

1.756

(2,17)

N.S,

-0.8265

4,931

0.757

21.339

(3,16)

25.7928

4.559

S X 7

0.778

1.451

(4,15)

N.S.

-0.1662

1.201

NS.

Constant

0.791

0.901

(5,14)

N.S.

-0.1620
-305.2337

0949

N.S

'N.S. = Not

significant.

362

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

IT
Ld

2 10

5 •

20 25 30

SALINITY (%.)

1 —

( , . ■ . 1 . .— . . ( 1 ■ ■ ■

"~~'--

^ \ -~

I
\

60 1

\l\

\ \
40 \

20 \^

"^^^^^^

^0^^^^

^^=

15 20 25 30

SALIN I TY (%o)

UJ
(£

I-
<

20 25 30

SALINITY (%o)

20 25 30

SALINITY (%o)

Figure 4. — Response surface estimation of percent survival oiCancer magister larvae after (A) 20 days, (B) 30 days, (C) 40 days, and
(D) 50 days of development at 20 different temperature and salinity combinations.

periods explained a significant 77-87% of the var-
iance in the data. The lowest surface temperature
and salinity reported for any sampled station dur-
ing the 1971 season was 7.4°C and 25.17%. After
20 days at this combination, 76.8% survival is
predicted; after 50 days, 44.6% survival. The
monthly mean surface temperature and salinity
compiled at the Oregon State University Marine

Science Center dock, Newport, is reported by
Wyatt and Gilbert (1972) for March 1971 to be
8.81°C and 30.12'L. Survival of 92.3% is predicted
at this temperature and salinity combination
after 20 days, and 71.0% survival after 50 days.
The direct effect of these temperatures and
salinities found off the central Oregon coast on
the survival of C. magister larvae would appear to

363

FISHERY BULLETIN: VOL. 74, NO. 2

Table 7. — An analysis of covariance between the 20- and
50-day survival poljmomials of Cancer magister larvae. Null
hypothesis: no significant difference between 20- and 50-day
survival polynomials.

Sum of

Mean

Source of variation

squares

square F-value

Polynomial 1 ; 20-clay survival

4,217 780

Polynomial 2: 50-clay survival

5.096 684

Total; Polynomial 1 and 2

9,314.464

332.659

Polynomial 3: Combined 20-

and 50-day survival

13,287.052

Difference: Polynomial 3 and

total

3,972.588

662.098 M.99

'Not significant, F 95,5 25, =2.44.

be minimal. Forty-five percent survival would
still occur, even after an unrealistic period of 50
days at nonconservative temperatures and
salinities.

Gut-Fullness Analysis of
Planktonic Larvae

The physical appearance of C. magister larvae
was examined for clues to the difference in the
larval populations between the two seasons, 1970
and 1971. Whatever happened to the larvae oc-
curred early in their development during the
months of February and March 1971, as a marked
decrease in the total larval population was ob-
served by the second zoeal stage. Those larvae
examined from the 1971 season appeared more
flaccid wdth a soft exoskeleton, had less eye pig-
mentation, and were more transparent compared
to the larvae caught during the 1970 season.
However, these features of appearance could not
be readily quantified. Further examination indi-
cated a possible difference on a population basis
in the amount of food in their guts among stages,
stations, and years. Differences in larval gut-
fullness may indicate good versus poor food
availability, or possibly a dying larval population

weakened by some factor in their environment
other than food.

Food and/or feces in the guts could readily be
seen through the body wall up to the fourth or
fifth zoeal stage and a close estimation of the per-
centage fullness could be made by noting the
proportion of gut segments filled with food. The
larval body can be divided into eight equal seg-
ments; the thorax constituting twice the length of
an abdominal segment. The food or feces was con-
sidered to be of the same approximate diameter
and could be estimated to within 3% of the total
gut length. A sample size of 30 larvae was neces-
sary before any significant difference could be
considered.

The 0.2-m bongo net samples were used to com-
pare the 1970 and 1971 larval seasons at stations
NHOl, NH03, NH05, and NHIO. Samples were
combined with both meshes of the 0.2-m bongo
nets. Only whole larvae were used and usually
the entire sample was analyzed. Specimens from
the 0.7-m bongo net samples were used to com-
pare inshore-offshore larval gut-fullness between
the 12 stations, NHOl through NH60, for the
1971 season.

Zoea 1 larvae from the 1970 season showed
maximum mean percentage gut-fullness at sta-
tions NH03 and NHIO compared to those from
NHOl and NH05 (Table 8). A general decrease in
gut-fullness was observed with increasing stage
of development. Surprisingly, all zoeal stages of
larvae caught during the 1971 season showed an
increase in gut-fullness over those of the 1970
season. The notable exception occurred for zoea 1
larvae at station NH03, where the 1971 gut-
fullness is significantly lower than that for the
1970 season.

The onshore-offshore comparison showed that
the greatest gut-fullness for any larval stage oc-

TABLE 8. — A eomparison of Cancer magister larval gut-fullness' between 1970 and 1971 at four New-
port Hydrographic line (NH) stations.^

Stage

Year

NHOl

NH03

NH05

NHIO

Zoea 1

1970

1971

Zoea 2

1970

1971

Zoea 3

1970

1971

Zoea 4

1970

1971

Zoea 5

1970

1971

13.19 ± 2.74(4)

31.23" ± 0,06(126)

19.64

± 0.12(65)

29.53

± 0.07(106)

14.26 ± 0.34(19)

9,86 ± 0,10(78)

24.10"

± 0.03(241)

36.13"

± 0.04(187)

7,56 * 1,97(5)

15.57

± 0.10(72)

12.94

± 0.02(269)

25,00 ± (1)

2329"

■ ± 0,24(38)

30 39"

± 0 11(87)

10,49 ± 10.49(2)

23.76

± 0,09(81)

14.02

± 0,02(212)

1875

± (1)

20 20"

± 0,09(51)

9.40

± 3.36(4)

15.99
19.50

± 1.19(6)
± 1.61(7)

'Gut-fullness is expressed as a reconverted arcsinVpercentage transformed mean followed by its standard error and the number
of observations in parentfieses.

^Tfie station samples in tfiis table represent tfie combined specimens from bothi mesh sizes of the 0.2-m bongo net sampler.
"1% level significant difference bietween yearly means based on a two-sample (-test.

364

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

Table 9.

— A comparison o{ Cancer

• magister larval gut-fullness» between 12 Newport Hydrographic '

line (NH) stations for

1971.2

Station

Zoea 1

Zoea 2

Zoea 3

Zoea 4

Zoea 5

NH01

8.51* ± 1 33(9)

NH03

10.51** ± 0 06(82)

0.79 ± 0.79(2)

0.0 ± (1)

NH05

27.89** ± 0.05(134)

17.57** ± 1.09(14)

NH10

26.51** ± 0.07(137)

29.79" ± 0.07(109)

45.77* ± 0.71(12)

NH15

23.50** ± 0.07(76)

35.1 r* ± 0.33(36)

60.77** ± 2.27(8)

NH20

34.43** ± 0.12(72)

23.69** ± 0.13(69)

20.99 ± 0.12(5)

25 00 ±

(1)

43.75 ± (1)

NH25

6.26** ± 0.25(33)

6.41 ± 0.48(20)

3.01 ± 3.01(3)

NH30

11.51 ±0.07(56)

7.75 ± 0.39(9)

23.41 ± 0.52(4)

25.00 ± (1)

NH35

12.03 ±0.99(3)

17.09 ±0.16(5)

18.75 ±

(1)

NH40

53.14 ± 0.10(2)

0.0 ± (1)

6.70 ± 6.70(2)

6.25 ±

(1)

13.92 ± 5.06(3)

NH50

6.25 ± (1)

0.0 ± 0.0 (2)

19.39 ±

19.39(2)

3.02 ± 1.74(8)

NH60

0.71 ± 0.71(3)

0.13 ± 0.13(10)

0.0 ± 0.0 (4)

0.0 ±

0.0(15)

0.0 ± 0.0 (3)

'Gut-fullness is expressed as a reconverted arcsinVpercentage transformed mean followed by its standard error and the number of observations in
parentheses.
^Tfie station samples in this table are from the 0.7-m bongo net sampler exclusively.
'5% level significant difference between successive station means based on two-sample /-tests.
**1% level significant difference between successive station means based on two-sample f-tests.

curred between stations NH05 and NH20 (Table
9). Any zoeal stage caught within NH03 and
farther ofTshore than NH20 showed a marked de-
crease in gut-fullness.

DISCUSSION

The initial appearance of C magister larvae in
the plankton off the central Oregon coast in late
January and early February occurs at a time
when sea surface temperatures are generally
warming after the yearly mean low in January
(Gonor et al. 1970). High densities of early stage
zoea caught within 3-10 miles of shore are in
agreement with the knowni distribution of the
adults at this time. Relatively few occurrences of
early stage larvae were found beyond 10 miles of
shore during the sampling period as the north-
ward flowing Davidson Current tends to retain
the early developing larvae in the nearshore
area. A very strong onshore component of the
current has been observed within 5 miles of shore
(Keene 1971; Wyatt et al. 1972; Holton and Elliot
1973). During the March and April transition
period when the northward Davidson Current is
replaced by currents flowing to the south and
southwest, the larvae have developed to late
stage zoea and megalopae. The bulk of the C.
magister megalopae settle out of the water and
metamorphose to juveniles by April and May be-
fore the onset of intense coastal upwelling in
June and July, thus reducing the chance of being
carried offshore by the resulting Ekman Current.
During all seasons along the coast, larvae which
occur increasingly closer to shore would be sub-
ject to decreasing current transport either along-
shore or offshore.

1970 Season

It was observed during the 1970 larval season
of C magister that the late zoeal stages "disap-
peared" or were greatly reduced in numbers in
the inshore sampling area, whereupon the
megalopae reappeared after the proper time in-
terval in densities comparable to those of the pre-
viously sampled zoea. Hypotheses to explain their
disappearance and reappearance are as follows:
1) the late zoea were misidentified, 2) some stages
are skipped in development, 3) the sampling in-
terval missed those stages, 4) avoidance of the
samplers increases with zoeal stages of develop-
ment but decreases at megalopal stage, 5) the
late zoeal larvae are carried offshore or
alongshore but upon molting to the megalopal
stages are transported onshore or back to their
original release point, 6) the larvae were very
dispersed at late zoeal development so that the
volume of water filtered was not adequate, or
7) late stage zoea are resting on the bottom or
below the depth sampled.

The late stages of C. magister larvae were not
misidentified as they are morphologically distinct
by this time and are nearly twice the size of any
other local cancrid species. Apparently, the late
larval stages of C. magister were not skipped in
their development since zoea 4 and 5 stages were
collected on the offshore stations in late March
and early April. It is not beheved that the late
zoeal stages have greater swimming ability com-
pared to the early zoea and megalopa which
would permit them to avoid the samplers to a
greater degree. On the contrary, personal obser-
vations of the late stage larvae in laboratory cul-
ture show them to be sluggish swimmers that

365

FISHERY BULLETIN: VOL. 74, NO. 2

spend considerable time resting on the bottom of
the rearing vessel.

A species such as C. magister, which has a lar-
val life of approximately 130 days, could conceiv-
ably be transported northward about 600 miles
along the North Pacific coast as Wyatt et al.
(1972) reported that the winter surface currents
based on drift bottle studies have a mean speed of
0.2 knots, or a drift of 150 miles per month. The
Ekman transport of surface waters due to wind
stress decreases exponentially with depth due to
frictional resistance, so that when the current has
fallen to about one-twenty third that of the sur-
face, this subsurface flow is negligible or reverse
to that of the surface currents (Sverdrup et al.
1942). Recent studies indicate wind driven water
motion extends to a depth of about 10 m (Bourke
et al. 1971). If the larval population resides about
5m below the surface where the wind induced cur-
rent is about one-quarter that of the surface, then
the larvae would only be transported 150 miles in
a linear distance. Larvae located in the water col-
umn below 5m depth, particularly the later zoeal
stages, would experience relatively little trans-
port in any direction. Holton and Elliot (1973)
reported the greatest abundance and density of
zooplankton containing crab larvae occurred at
about 15m depth at nearshore stations off New-
port during the daylight hours. Hypothetically,
larvae released in January-February could be
transported north along the coast in the surface
currents and, after the transition period of cur-
rents in March, travel south a comparable dis-
tance in April and May. Or, taking into considera-
tion the fact that the older stages may reside
deeper into the water column, they could conceiv-
ably travel north in the surface currents as early
zoea and travel south again as late larvae in a
weak underlying countercurrent, but this seems
unlikely. Huyer et al. (1975) reported the north-
ward currents along the central Oregon coast es-
sentially are constant with depth during the win-
ter and southward at all depths in the spring but
stronger at the surface. Larvae occurring within
3-5 miles of the coast probably are caught within a
system of eddies and countercurrents characteris-
tic of this zone, retarding large-scale dispersal in
any direction. The mechanistic concepts of re-
cruitment seem too contrived and unnecessary if
stochastic processes are the general rule for
species producing large numbers of expendable
young. Most investigators would agree that the
great majority of the pelagic larvae of marine in-

vertebrates are lost to the population and that
only a very small percentage of annual recruits
are normally required to maintain a stable popu-
lation for longer-lived adults. Cancer magister
lives 4 or 5 yr so that a population unexploited by
man would only require recruitment every other
year or so. The fact that the adult populations are
not retreating northward supports the view that
at least some of the larvae are retained in the
same general area as their point of origin.

The low densities of late stage larvae collected
in the offshore area indicated that the small vol-
ume of water filtered on the inshore stations
could account for their disappearance or reduced
numbers.

Knowledge of their vertical location within the
water column at different stages of development
is important in understanding their spatial dis-
tribution and local abundance. However, a sepa-
rate study of the larvae within the upper 150 m
was not undertaken. Most crab larvae are photo-
positive to light in their early stages and migrate
to the surface layers, whereas the late stages re-
spond photonegatively and are found in the
deeper layers near the bottom as they prepare to
molt to juveniles (Thorson 1964). The larvae of C
magister appear to follow this same general pat-
tern except that the early megalopal stage shows
anomalous behavior as they have been observed
to "swarm" near the sea surface along the coast
(Cleaver 1949; Gaumer 1971; pers. obs.). Personal
laboratory observations, as well as those by
MacKay (1942) and others, substantiate the fact
that the early zoea and megalopa of C. magister
are generally photopositive in contrast to the late
zoeal stages which are neutral or photonegative.

A scheme is proposed which would explain
their distribution and abundance within 10 miles
of the coast taking into account the differential
behavioral response to light of the various larval
stages. Newly hatched zoeal larvae are strongly
photopositive and swim to the surface where cur-
rent transport during the winter is generally on-
shore. They become progressively heavier and
less photopositive with development until in the
late zoeal stages they are neutral or responding
negatively to light. As a consequence, the late
zoeal stages reside in the deeper layers of water,
possibly within a few meters of the bottom. They
are now maximally dispersed in the nearshore
area. Upon molting to the megalopic stage they
are temporarily strongly photopositive to light
and coupled with their increased powers of

366

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

locomotion, they swarm to the surface again and
are congregated by the prevailing currents usu-
ally in a band w^ithin 5 miles of the coast. If the
late zoeal larvae do in fact reside near the sea
bottom, the onshore drift current within 10-20 m
of the bottom would prevent them from being
transported offshore. Bottom flow in waters less
than 40 m deep is towards the coast in the direc-
tion of wave travel throughout the year (Gross et
al. 1969). The behavior of the larvae within the
water column in relation to the hydrological fea-
tures of the nearshore area under usual cir-
cumstances tends to restrict dispersal of the larvae
to any great degree.

1971 Season

The sparseness of C. magister late zoeal larvae
and megalopae during the 1971 season implies
that a mass mortality occurred in the early zoeal
stages. This apparent mortality was associated
with sea surface temperature and salinity in
analyses of covariance, but larval survival pre-
dicted through response surface methodology and
gut-fullness analysis did not substantially ex-
plain their sparseness. The lack of highly suppor-
tive evidence leads to further speculation as to
the causes of larval mortality in the plankton.

Hypothesis 1: Direct Effects of Temperature
and Salinity

Sea surface temperature, and salinity to a les-
ser degree, were important environmental factors
in explaining the difference in yearly larval popu-
lation means of C. magister by analyses of multi-
ple covariance. However, the statistical impor-
tance of these factors in determining larval
abundance may be misleading. A wide tempera-
ture gradient during a larval season, i.e., a steep
slope, could be statistically significant, but the
range of temperatures may be well within the
tolerance limits of an organism. In contrast, the
salinity gradient during the same larval season is
usually narrow resulting in a statistically non-
significant slope, which may still occur outside
the range tolerated by the larvae. Also, the errat-
ic surface temperature and salinity fluctuations
that occurred during the summer upwelling may
cancel the effect of a signiflcant gradient that oc-
curred earlier in winter and spring.

Cancer magister larvae were reared by Reed
(1969) under various temperature-salinity com-

binations and he concluded that these factors, as
they normally occur off the Oregon coast, would
not significantly affect survival. Response surface
techniques, using Reed's data, predicted about
45% survival under the extreme temperatures
and salinities that occurred during February and
March 1971. The sea surface temperatures and
salinities used in the analysis probably represent
the most extreme long-term conditions that the
larvae could have experienced in the field. Larvae
several meters below the surface may be pro-
tected from the more extreme fluctuations of
temperature and salinity, but some degree of ex-
posure seems certain in view of the fact that ex-
tensive wind mixing occurs in shallow waters
along the coast. The North Pacific is charac-
terized by heavy precipitation during the fall and
winter seasons resulting in considerable land
drainage and river runoff along the nearshore
area. Larvae along the coast, particularly near
the mouths of bays and rivers, may lie in the
low-salinity plume waters before sufficient
mixing occurs. Harder ( 1968) reported that many
planktonic organisms tend to accumulate near
density interfaces that frequently occur in
natural waters. Some species of copepods were
observed under laboratory conditions to react to
extremely small changes in density. Whether C.
magister larvae have the ability to avoid these
low-salinity surface waters that may be detri-
mental to them is not known. The early zoeal lar-
vae would seem most vulnerable to low surface
salinity as their behavioral response directs them
to the surface and their swimming ability is
slight compared to the megalops stage. Early lar-
val ability to avoid low-salinity surface waters
would have to be sufficient to overcome the in-
creased storm-induced mixing during this season.
The mortality rate of C magister larvae reared in
the laboratory under optimum conditions was
constant and minimal throughout development
(Reed 1969). Mortality increased greatly for lar-
vae reared at 20L salinity; early zoeal larvae
were killed within a short period in salinities
less than 20%.. In addition, both the lower range
of salinities and temperatures used in his ex-
periments increased the duration of the larval in-
stars where survival could be monitored for a suf-
ficient time period.

It is difficult to evaluate the extent to which
results from laboratory studies approach reality
in order to understand how environmental vari-
ables may affect survival. Larvae reared at sub-

367

FISHERY BULLETIN: VOL. 74, NO. 2

optimal conditions have been observed to survive
for considerable periods of time, apparently un-
able to molt successfully. These same larvae
eventually die, but laboratory experiments often
are terminated before full mortality can be ob-
served. Low salinity during the winter of 1971
may have been an important factor resulting in
the demise of C. magister larvae that year. Subtle
changes in the flux and composition of the inter-
nal ionic constituents can alter the molting pro-
cess; larvae which appear normal in early de-
velopment may mask deficiencies that express
themselves later in development. Nevertheless,
short-term exposure to extreme conditions may
be just as detrimental as slightly suboptimal con-
ditions over a long period of time (cf Lough and
Gonor 1973a, b). Although the nearshore surf
salinities on a monthly average are in the range
of tolerance by the larvae, daily measurements
occasionally drop below 2Q"L (Gonor et al. 1970).
No larvae survived below 201. salinity in Reed's
(1969) laboratory study.

The effect of low salinity in conjunction with
wider than normal temperatures may play an
important role in larval survival as indicated
from the analyses. Low and high temperatures
greatly accentuated the effects of marginally di-
lute salinities on C. magister larval survival. But
again, the ecological significance of a synergistic
effect has not been fully established in this
study. More detailed, short-term studies of
salinity-temperature variability and larval
monitoring are needed in the nearshore area.
Sastry and McCarthy (1973) observed distinct dif-
ferences in temperature-salinity tolerances and
metabolic responses of the larvae of two species of
Cancer sympatrically distributed along the east
coast of North America. Complete development
for C. irroratus larvae occurred over a wide range
of temperatures, whereas C borealis larvae was
restricted to a narrow range. The metabolic-
temperature pattern of C. irroratus larvae indi-
cated a progessive narrowing in temperature sen-
sitivity. In contrast, the early stages of C borealis
initially were sensitive to warmer temperatures
but in the later stages sensitivity shifted to colder
temperatures. Hatching of the two Cancer species
is separated in time so that the diverse metabolic
responses observed are believed to be adaptations
by larvae of the two species to the different tem-
perature conditions encountered.

The combined effects of salinity and tempera-
ture have been studied under controlled labora-

tory conditions on other species of brachyuran lar-
vae by Costlow et al. (1960, 1962, 1966), Costlow
and Bookhout (1962), and Costlow (1967). Al-
though the adults inhabit euryhaline waters,
specific larval stages have been shown to require
restricted ranges of salinity and temperature to
varying degrees for complete development. In
many cases, both temperature and salinity and
the interaction of various combinations of the two
environmental variables were observed to affect
larval survival and retard development. Salinity
generally has an immediate effect on survival
while temperature appears to play a modifying
role within the extremes of tolerance. Most of
their work indicates that mortality was highest
during the early zoeal stages and that the
megalops stage was the least subject to environ-
mental stress, although exceptions are reported.
Recently, Costlow and Bookhout (1971) investi-
gated the effects of cyclic temperatures compared
to constant temperatures on the larvae of the es-
tuarine mud crab, Rhithropanopeus harrisii.
Duration of larval life and survival were about
the same but survival was enhanced under cyclic
temperatures at the higher end of the experimen-
tal range. Short-term fluctuations in temperature
or other environmental variables throughout the
water column have not been adequately moni-
tored along the North Pacific coast. Their effect
on C. magister larvae is not knowTi and should be
investigated.

Hypothesis 2: Food Quality and Quantity

May (1974) reviewed Hjort's (1914) critical
period concept for fish larvae since Marr's (1956)
evaluation and concluded from recent work that
starvation may be an important cause of mortal-
ity, especially during the period immediately fol-
lowing the yolk sac stage. Although crab larvae
do not have a strictly comparable yolk sac stage
in their planktonic life, adequate food densities
for efficient feeding may be of critical importance
during a brief period following hatching. There is
limited knowledge concerning the types of food
organisms normally available and selected by C.
magister larvae and concerning the densities of
these food organisms sufficient for development.
Most crab larvae are omnivorous, requiring sub-
stantial protein in their diet (Costlow and Sastry
1966; and others). Attempts to distinguish gut
contents of field-caught C. magister larvae were
unsuccessful in the present study. However, the

368

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

specific kind of food organism encountered may
not be as important as its size. The size of food
organisms available for each larval stage must be
within a restricted range in order for a larva to
successfully capture and ingest. The progression
of larval size with development would indicate
that the different larval stages can utilize in-
creasingly greater sizes of food organisms. Reed
(1969) found in laboratory culture that the larvae
of C magister survived well feeding on Artemia
salina (0.475-0.752 mm length) and Balanus
gladula nauplii (0.370-0.420 mm length), but
would only survive for a limited period on smaller
size veliger larvae of Mytilus edulis (0.100-0.300
mm length?). He also reported that unfed C.
magister zoea larvae would only survive for 14
days. This implies that under natural conditions
larvae will not survive if a suitable food organism
is delayed in its appearance by more than 2 wk,
and that certain kinds of food organisms selected
by the larvae are nutritionally inadequate for
their long-term metabolic needs.

Chamberlain (1961, 1962) reared the larvae of
two xanthid crabs, Neopanope texana sayi and
Rhithropanopeus harrisii, on a variety of foods
and found that development was retarded when
larvae were fed on a mixture of nauplii and algae.
Larvae fed algae alone would not molt and only
lived 6-10 days in culture. Algae appeared to be
nutritionally inadequate for successful develop-
ment and restricts the intake of more suitable
food by indiscriminate larval feeding. Costlow
and Sastry (1966) suggested that high mortality
of Callinectes sapidus larvae at the time of the
third zoeal stage in nutritionally inadequate cul-
ture may be due to the initial availability of a
large pool of free amino acids within the eggs
through the first and second zoeal stages. They
also pointed out that the variability in tolerance
to suboptimal conditions may be related to the
size of such a free amino acid pool.

Although the gut-fullness analysis in the pre-
sent study did not provide insight into the differ-
ence in larval abundance between the 2 yr, it did
suggest the existence of an optimum zone for
adequate feeding between 3 and 20 miles offshore
where suitable kinds and densities of food or-
ganisms occur. Zooplankton volumes along the
Washington coast decrease to a minimum level
during the winter and increase to maximum
levels during the spring (Frolander 1962). During
the winter, the volume of zooplankton and abun-
dance of copepods were greater inshore than

offshore as a consequence of the onshore trans-
port of surface waters (Frolander 1962; Anderson
1964; Peterson 1972). Anomalous weather condi-
tions such as occurred during the winter of 1971
may have been ultimately responsible for altera-
tions in the usual types and availability of food
organisms encountered during the first few weeks
of larval feeding.

Hypothesis 3: Predators and Competitors

The importance of the combined or separate ef-
fects of predation and competition on larval popu-
lations is difficult to assess. Predation has gener-
ally been regarded as the major cause of larval
mortality (Thorson 1946, 1950). Lebour (1919a, b,
1920, 1921, 1922, 1923) observed many species of
young fish and medusae to prey upon crab larvae
as well as most other small organisms in the
plankton. Cannibalism is well known in labora-
tory culture. Knudsen (1960) observed in the
laboratory that xanthid first stage zoea were
eaten by older zoea and megalops as well as by
copepods. Other predators known to feed on
marine larvae, such as ctenophores, chaeto-
gnaths, euphausiids, and shrimps, appear sea-
sonally in high densities along the North Pacific
coast. Their effect on larval populations has not
been fully ascertained. Peterson (1972) compared
the ratios of copepod nauplii to total copepods off
the Washington coast and found that more naup-
lii were hatched inshore than offshore throughout
the year, but fewer developed to adults suggest-
ing greater predation in the inshore area. Preda-
tion was reduced during the winter compared to
other seasons within the inshore area. These
findings might similarly apply to relative preda-
tion rates on C. magister larvae along the North
Pacific coast.

Factors in the environment such as abnormally
cold temperatures or lack of food that extend the
pelagic life of the larval phase have been consid-
ered detrimental due to predation. It has been
assumed that the longer the larvae remain in the
plankton the more they will be preyed upon, al-
though predation pressure upon their recruit-
ment to the benthic habitat may be just as great,
or greater (Thorson 1966). Larvae genetically
feeble or weakened by some environmental fac-
tors may be more subject to predation so that
under usual circumstances, the importance of
predation may be secondary in mortality proces-
ses. The effect of predation on larval populations

369

FISHERY BULLETIN: VOL. 74, NO. 2

would not seem to be constant in the heterogene-
ous marine environment, but would more likely
vary in intensity both temporally and spatially.
Predation may only be a dominant factor in un-
usual years and/or on a small-scale basis.

Other members of the plankton community
undoubtedly feed on the same food organisms as
C. magister and competition may become an im-
portant factor when these food organisms become
sparse. One potential competitor was tentatively
identified as C. oregonensis. Its larvae are very
abundant in the inshore area and cooccur wdth
those of C magister. Both species are morphologi-
cally similar and pass through the same number
of larval stages, except that the larvae of C.
magister become increasingly larger with de-
velopment. There are studies showing the an-
tagonistic effects of a mutually shared food re-
source. Brooks and Dodson (1965), in a study of
two species of freshwater Daphnia, concluded
that the larger species was more efficient in col-
lecting both small and large particles and would
competitively exclude the smaller species as long
as size dependent predation was of low intensity.
Conversely, Schoener (1969), in a theoretical
study, concluded that large predators ate an
equal or a greater range of food compared to the
smaller ones as long as food was at some upper
level. But, as food abundance was reduced, the
optimal predator size shifted towards the smaller
predator. Similar situations could conceivably
occur and explain why C. magister larvae were
less numerous in 1971. The interactions of
hierarchies of predators and their prey involving
temporal and spatial changes in densities and
size fi'equencies can be exceedingly complex.

Hypothesis 4: Oceanic currents and
multiple environmental effects

Planktonic organisms have limited means of
locomotion and consequently are subject to the
vagaries of oceanic currents. Changes in the
strength or timing of these currents can be ulti-
mately responsible for the success or failure of
larval populations and their adult stocks (Coe
1956). The transport of entire larval stocks out of
their normal environment can have catastrophic
results for annual recruitment.

During the winter-spring larval period of C.
magister, the major nearshore oceanographic fea-
ture is the northerly intrusion of the Davidson
Current along the Oregon-Washington coast and

its reversal in March- April. The strength and du-
ration of the Davidson Current are critical factors
in the initiation, development, and persistence of
seasonally dominant plankton communities.
Southern neritic zooplankton species appear
abundant off the Oregon and Washington coasts
during fall and winter and are believed to be car-
ried by the northerly surface drift (Cross and
Small 1967; Miller 1972; Frolander et al. 1973).
Frolander (1962) observed widespread anomalous
conditions off the Washington coast during Feb-
ruary 1958, compared to the previous year. Lower
plankton volumes and a change in plankton
species were associated with an increase in the
surface temperatures, a decrease in dissolved in-
organic phosphate, and unusual weather during
the anomalous February. These events were be-
lieved to be the result of southerly offshore waters
moving into the coastal area to a larger extent
that year.

Superimposed upon the nearshore currents
with their characteristic water properties, a dom-
inant modifying process results from precipita-
tion and river runoff. A band of low salinity oc-
curs all along the North Pacific coast. Little
information is available on the effect of the heavy
river runoff on the endemic plankton populations
in the neritic zone, but some studies have been
done concerning the effect of the Columbia River
plume on the physical processes and biota over its
range of influence (Anderson 1972). The Colum-
bia River effluent flows north along the coast of
Washington during the winter in response to the
prevailing southwesterly winds (Barnes et al.
1972). Hobson (1966) and Anderson (1972) ob-
served that chlorophyll and productivity at the
surface of the plume and am.bient waters were
higher than nearby oceanic waters due to the in-
creased stability of the water column providing
an environment where phytoplankton could ac-
cumulate. The major influence of the Columbia
River plume on phytoplankton development is be-
lieved to be in the timing of events. Phytoplank-
ton populations can develop 3-5 wk earlier in the
plume due to the increased stabilization. Hein-
rich (1962, 1968) stated that the seasonal cycle of
phytoplankton communities are less balanced in
the neritic zone and that the phytoplankton popu-
lations in this area can vary depending on the
timing and differential growth of relative copepod
species. Shifts in weather patterns create corres-
ponding changes in nearshore currents resulting
in the intrusion and displacement of endemic

370

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

planktons. Nearshore modifying processes can
change the character of these communities
setting forth new interactions among the
populations.

Anomalous hydrographic and meteorological
conditions were observed along the Oregon coast
during the winter of 1971 in the present study. Its
effect on the plankton populations to date only
have been investigated in relation to C. magister
larvae. To what extent did the relaxation of on-
shore transport of surface waters during January
and February with subsequent increased trans-
port in March 1971, compared to the same period
in 1970, affect the dynamics of the C. magister
larval population? The circumstantial evidence
suggests that heavy mortality of the larvae oc-
curred in 1971. However, the difference in larval
abundance for the 2 yr may not be real if the
larvae were quite localized in their alongshore
distribution and moved out of the study area.
Sampling was not conducted in other areas for
those years to fully answer this point. In addition,
the late zoeal stages were undersampled both
years leaving a gap in our knowledge of their true
numbers, distribution, and condition. Assuming
that a mass mortality of larvae did, in fact, occur
in the study area, what are the most likely en-
vironmental mechanisms? Did the decreased on-
shore surface water transport in early winter of
1971 relative to 1970 allow more larvae to be car-
ried offshore that year where food abundance was
lower, etc.? Any larvae swept off the shelf area
that survived would still probably be beyond suc-
cessful recruitment to the adult nearshore popu-
lation. Did the greater onshore transport of sur-
face waters during late winter of 1971 move the
bulk of the larval population closer to shore into a
suboptimal environment too early in their de-
velopment? What is the effect of the increased
precipitation and river runoff during the winter
of 1971 that reduced nearshore salinities? Was a
phytoplankton bloom initiated earlier in the sea-
son and how did it affect populations of other
planktonic organisms utilized as food for C.
magister larvae? Chamberlain (1961) commented
that, for crab larvae feeding indiscriminately on
both algae and zooplankton, a phytoplankton
bloom initially may retard zoeal development;
however, following the increase of the herbivore
population, more nutritionally adequate food is
available and would accelerate larval develop-
ment. Do the low-salinity Columbia River plume
and other river effluents effectively act as bar-

riers against northerly alongshore transport of
larvae? The lower temperatures and salinities in
1971, particularly in the nearshore area, coupled
with adverse biological pressures, i.e., increased
predation, may have had a synergistic effect on
larval mortality. Many alternatives are open in
marine ecosystems where stochastic processes
prevail producing innumerable permutations.
The indirect effect of physical variables on larval
food organisms and predator-prey relations can
be extremely complex and important. Subtle
changes in these relations may have an ac-
cumulative effect on a larval population already
in a stressed condition and near the point at
which recovery diminishes.

Answers to these questions remain conjectural
and may only be sought through further com-
prehensive and detailed studies. However, in con-
clusion, there is no substantial evidence from this
study that the colder winter of 1971 caused a
delay in the initial appearance and developmen-
tal schedule throughout the larval period of C.
magister. The generally poor appearance of the
early zoeal larvae collected during the 1971 sea-
son suggests that whatever factor(s) responsible
for the apparent mortality appeared to have an
immediate effect on these stages. The first few
zoeal stages may be the critical period in the
early life history of C. magister where the
greatest mortality occurs ultimately determining
future year class strength.

RECOMMENDATIONS FOR
FUTURE RESEARCH

Studies to date have provided a broad overview
of knowledge concerning the initial timing,
abundance, and dispersal of C magister larvae in
relation to major oceanographic events off the
central Oregon coast. First approximation esti-
mates of length of larval life, mortality, and feed-
ing have been achieved, but we are still lacking
detailed insight into the dynamics of the larvae-
plankton-environment matrix. This study points
out our limited knowledge and understanding of
the physical and biological mechanisms affecting
the dispersal and subsequent survival of C.
magister larvae. An understanding of these pro-
cesses is necessary for an understanding of the
stability and long-term productive potential of
the Dungeness crab as a fishery resource in the
Pacific Northwest. By studying processes control-
ling the dispersal and survival of the larvae, we

371

FISHERY BULLETIN: VOL. 74, NO. 2

may be able to gain insight into stock-
recruitment relations and be able to predict the
effects of long-term environmental changes.
Some specific recommendations for further work
are listed below.

1. A minimum of three surveys should be con-
ducted between late January and early June
to monitor initial hatching, production, rate
of development, and dispersal of the larvae. It
is imperative that survey coverage be ex-
tended along the Oregon coast to observe
patchiness and alongshore dispersal. A grid
of stations to within 30 miles of the coast
from at least Cape Blanco, Oreg. to Cape
Flattery, Wash, is recommended. A sufficient
time series of data is required to adequately
assess yearly changes in the larval popula-
tions in order to gain insight into mortality
processes. Also, a long-term series is needed
as a background of knowledge upon which
more specialized short-term studies can be
based. Six or seven years of plankton sam-
pling seems to be a minimum time series for
establishing trends, although 10-15 yr are
required to substantiate significant differ-
ences.

2. Intensive close-order grid sampling on a
short-term basis, following a fairly well-
defined and homogeneous "patch" of larvae,
should be conducted to assess in more detail
mortality and feeding in good and poor areas.

3. This study emphasizes the need for more de-
tailed oceanographic studies in the nearshore
environment and how they affect the popula-
tion dynamics of organisms living in this
zone. In conjunction with larval surveys, cir-
culation studies should be expanded during
the winter and spring along the Oregon coast
to improve the basis for predicting and
evaluating dispersal, primary productivity,
etc. A continuous program of temperature,
salinity, and current measurements are
needed of the nearshore currents during the
larval period from January through June and
particularly the timing and extent of the
March- April transition of the Davidson Cur-
rent.

4. Short-term exposure of the larvae to en-
vironmental variables such as low salinity in
combination with varying temperature, food
density, etc. and subsequent transfer to op-
timum conditions for long-term observations

in the laboratory are needed to properly
evaluate the effects of these factors.

5. Detailed descriptions of the three-
dimensional composition of the associated
plankton communities are needed in terms
of the dominant species, size categories, and
diurnal variability. Investigations into the
contagious distribution of these organisms,
mechanisms of initiation and destruction,
are central to understanding prey-predator
interaction and attempts to model these
phenomena.

6. Fine-mesh (0.165 and 0.053 mm) sampling
with the 0.2-m bongo nets should be used
concurrently with the 0.7-m bongo nets to
examine and answer the questions of food
composition and availability utilized by early
C. magister larvae. In particular, the inver-
tebrate component for both coarse- and
fine-mesh samples should be analyzed ini-
tially between contrasting years or areas of
larval abundance. The use of plankton pumps
may be more amenable in this case as fine-
mesh nets clog rapidly.

7. The vertical distribution and diurnal move-
ments of C magister larvae throughout its
pelagic life is especially important in regard
to sampling variability, dispersal, and feed-
ing, and should be studied. Do most of the
older zoeal larvae, in fact, reside within a few
meters of the bottom in the shallow inshore
area?

8. Laboratory studies should be undertaken to
analyze the phototactic behavior of the lar-
vae at various stages of development to gain
a better understanding of their diurnal
movements as may be modified by tempera-
ture, hunger state, presence of prey and pred-
ators, etc.

9. A new approach is needed in the analysis of
larval gut contents. Biochemical techniques
of gut material may be used to identify food
organisms utilized by the larvae. Energy
budgets should be constructed to determine
minimum food requirements of the various
larval stages. Condition factors indicative of
the physiological well-being of larvae may be
used to evaluate good versus poor areas and
years of feeding.

10. Potential predators that cooccur with C.
magister larvae should be identified and in-
gestion rates determined from field and
laboratory experiments in order to estimate

372

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

their effect on the larval population. Transi-
tional experiments should be carried out in
the field to further assess the reality of
laboratory studies.

ACKNOWLEDGMENT

This research was supported during the years
1969-74 by the National Oceanic and Atmos-
pheric Administration (maintained by the U.S.
Department of Commerce) Institutional Sea
Grant GH-45, GH-97, NOAA-2-35187, NOAA-
04-3-1584.

LITERATURE CITED

ALDERDICE, D. F.

1972. Factor combinations. Responses of marine poikilo-
therms to environmental factors acting in concert. In
O. Kinne (editor), Marine ecology, Vol. 1, Part 3, p.
1659-1722. Wiley-Interscience, Lend.

Anderson, G. C.

1964. The seasonal and geographic distribution of pri-
mary productivity off the Washington and Oregon coasts.
Limnol. Oceanogr. 9:284-302.

1972. Aspects of marine phytoplankton studies near the
Columbia River, with special reference to a subsurface
chlorophyll maximum. In A.T. Pruter and D. L. Alverson
(editors), The Columbia River estuary and adjacent
ocean waters, p. 219-264. Univ. Wash. Press, Seattle.

BAKUN, A.

1973. Coastal upwelling indices, west coast of North
America, 1946-71. U.S. Dep. Commer., NOAA Tech.
Rep. NMFS SSRF-671, 103 p.

Barnes, C. a., a. C. duxbury, and B. A. Morse.

1972. Circulation and selected properties of the Columbia

River effluent at sea. In A. T. Pruter and D. L. Alverson

(editors). The Columbia River estuary and adjacent

ocean waters, p. 41-80. Univ. Wash. Press, Seattle.

BOURKE, R. H., B. GLENNE, and B. W. ADAMS.

1971. The nearshore physical oceanographic environ-
ment of the Pacific Northwest coast. Dep. Oceanogr.
Oreg. State Univ., Ref 71-45, 120 p.

Brooks, J. L., and S. I. Dodson.

1965. Predation, body size, and composition of plankton.
Science (Wash., D.C.) 150:28-35.

Burt, W. v., and B. Wyatt.

1964. Drift bottle observations of the Davidson Current
off Oregon. In K. Yoshida (editor), Studies on oceanog-
raphy, p. 156-165. Univ. Tokyo Press, Tokyo.
Butler, T. H.

1960. Maturity and breeding of the Pacific edible crab,
Cancer magister Dana. J. Fish. Res. Board Can.
17:641-646.

Chamberlain, N. a.

1961. Studies on the larval development of Neopanope
texana sayi (Smith) and other crabs of the family Xan-
thidae (Brachyura). Chesapeake Bay Inst., John Hopkins
Univ., Tech. Rep. 22:1-35.

1962. Ecological studies of the larval development of
Rhithropanopeus harrisii (Xanthidae, Brachyura).

Chesapeake Bay Inst., John Hopkins Univ., Tech. Rep.
■ 28:1-47.

Cleaver, F. C.

1949. Preliminary results of the coastal crab (Cancer
magister) investigation. Wash. Dep. Fish., Biol. Rep.
49A:47-82.

COE, W. R.

1956. Fluctuations in populations of littoral marine in-
vertebrates. J. Mar. Res. 15:212-232.

COLTON, J. B., Jr.

1959. The multiplane kite-otter as a depressor for high-
speed plankton samplers. J. Cons. 25:29-35.

COSTLOW, J. D., JR.

1967. The effect of salinity and temperature on survival
and metamorphosis of megalops of the blue crab
Callinectes sapidus. Helgolander wiss. Meeresunters.
15:84-97.

CosTLow, J. D., Jr., and C. G. BOOKHOUT.

1962. The larval development of Hepatus epheliticus (L.)
under laboratory conditions. J. Elisha Mitchell Sci. See.
78:113-125.

1971. The effect of cyclic temperatures on larval de-
velopment in the mud-crab Rhithropanopeus harrisii.
In D. J. Crisp (editor). Fourth European Marine Biology
Symposium, p. 211-220. Cambridge Univ. Press., Lond.

COSTLOW, J. D., Jr., C. G. Bookhout, and R. Monroe.

1960. The effect of salinity and temperature on larval
development of Sesarma cinereum (Bosc) reared in the
laboratory. Biol. Bull. (Woods Hole) 118:183-202.

1962. Salinity-temperature effects on the larval develop-
ment of the crab Panopeus herbstii Milne-Edwards,
reared in the laboratory. Physiol. Zool. 35:79-93.

1966. Studies on the larval development of the crab,
Rhithropanopeus harrisii (Gould). I. The effect of salinity
and temperature on larval development. Physiol.
Zool. 39:81-100.
COSTLOW, J. D., JR., AND A. N. SASTRY.

1966. Free animo acids in developing stages of two crabs,
Callinectes sapidus Rathbun and Rhithropanopeus
harrisii (Gould). Acta Embryol. Morphol. Exp. 9:44-55.

Cross, F. a., and l. F, Small.

1967. Copepod indicators of surface water movements off
the Oregon coast. Limnol. Oceanogr. 12:60-72.

FROLANDER, H. F.

1962. Quantitative estimations of temporal variations of
zooplankton off the coast of Washington and British
Columbia. J. Fish. Res. Board Can. 19:657-675.
FROLANDER, H. F., C. B. MILLER, M. J. FLYNN, S. C. MEYERS,
AND S. T. ZIMMERMAN.

1973. Seasonal cycles of abundance in zooplankton popu-
lations of Yaquina Bay, Oregon. Mar. Biol. (Berl.)
21:277-288.
GAUMER, T. F.

1971. Controlled rearing of Dungeness crab larvae and
the influence of environmental conditions on their sur-
vival. Commer. Fish. Res. Dev. Act Prog. Rep., Nov. 16,
1965 to June 30, 1971. Fish. Comm. Oreg., Closing
Rep., 43 p.
GONOR, J. J., A. B. THUM, and D, W. ELVIN.

1970. Inshore sea surface temperature sand salinity con-
ditions at Agate Beach, Yaquina Bay and Whale Cove,
Oregon, in 1970. Data Rep. 45. Dep. Oceanogr. Oreg.
State Univ., Ref 70-44, 30 p.

373

Gross, M. G., B.-A. morse, and C. a. Barnes.

1969. Movement of near-bottom waters on the continental
shelf off the northwestern United States. J. Geophys. Res.
74:7044-7047.

Harder, W.

1968. Reactions of plankton organisms to water strati-
fication. Limnol. Oceanogr. 13:156-168.

heinrich, a. K.

1962. The life histories of plankton animals and seasonal

cycles of plankton communities in the oceans. J. Cons.

27:15-24.
1968. On seasonal phenomena in the plankton of the

Northeastern Pacific. Okeanologiya 8:287-296.
HJORT, J.

1914. Fluctuations in the great fisheries of northern Europe

viewed in the light of biological research. Rapp. P.-V.

Reun. Cons. Perm. Int. Explor. Mer. 20:1-228.
1926. Fluctuations in the year classes of important food

fishes. J. Cons. 1:5-38.
HOBSON, L. A.

1966. Some influences of the Columbia River effluent

on marine phytoplankton during January 1961. Limnol.

Oceanogr. 11:223-234.
HOLTON, R. L., AND W. P. ELLIOTT.

1973. The development of methods for studying physical
and biological processes in the nearshore zone on the
Pacific coast of the United States. Dep. Oceanogr.
Oreg. State Univ., R«f 73-3, 18 p.

HUYER, A., R. D. PILLSBURY, AND R. L. SMITH.

1975. Seasonal variation of the alongshore velocity field
over the continental shelf off Oregon. Limnol.
Oceanogr. 20:90-95.
KEENE, D. F.

1971. A physical oceanographic study of the nearshore
zone at Newport, Oregon. M.S. Thesis, Oreg. State Univ.,
Corvallis, 91 p.
KNUDSEN, J. W.

1960. Reproduction, life history and larval ecology of the
California Xanthidae, the pebble crabs. Pac. Sci
14:3-17.
KUKLA, G. J., AND H. J. KUKLA.

1974. Increased surface albedo in the northern hemi-
sphere. Science (Wash., D.C.) 183:709-714.

KULM, L. D., AND J. V. BYRNE.

1966. Sedimentary response to hydrography in an Ore-
gon estuary. Mar. Geol. 4:85-118.

LEBOUR, M. V.

1919a. Feeding habits of some young fish. J. Mar. Biol.

Assoc. U.K. 12:9-21.
1919b. The food of post- larval fish. No. II (1918). J. Mar.

Biol. Assoc. U.K. 12:22-47.

1920. The food of young fish. No. III. (1919). J. Mar.
Biol. Assoc. U.K. 12:261-324.

1921. The food of young clupeoids. J. Mar. Biol. Assoc.
U.K. 12:458-467.

1922. The food of planktonic organisms. J. Mar. Biol.
Assoc. U.K. 12:644-677.

1923. The food of plankton organisms. II. J. Mar. Biol.
Assoc. U.K. 13:70-92.

LONGHURST, A. R., AND D. L. R. SEIBERT.

1967. Skill in the use of Folsom's plankton sample split-
ter. Limnol. Oceanogr. 12:334-335.

LOUGH, R. G.

1975a. A reevaluation of the combined effects of tempera-

374

FISHERY BULLETIN: VOL. 74, NO. 2

ture and salinity on survival and growth of bivalve
larvae using response surface techniques. Fish. Bull.,
U.S. 73:86-94.
1975b. Dynamics of crab larvae (Anomura, Brachyura)
off the central Oregon coast, 1969-1971. Ph.D. Thesis,
Oreg. State Univ., Corvallis, 299 p.
LOUGH, R. G., AND J. J. GONOR.

1973a. A response-surface approach to the combined
effects of temperature and salinity on the larval de-
velopment of Adula californiensis (Pelecypoda: Mytili-
dae). I. Survival and growth of three and fifteen-day old
larvae. Mar. Biol. (Berl.) 22:241-250.
1973b. A response-surface approach to the combined
effects of temperature and salinity on the larval de-
velopment of Adula californiensis (Pelecypoda: Myti-
lidae). II. Long-term larval survival and grovrth in re-
lation to respiration. Mar. Biol. (Berl.) 22:295-305.
MacKAY, D. C. G.

1942. The Pacific edible crab. Cancer magister. Fish.
Res. Board Can., Bull. 62:1-32.
Marr, J. C.

1956. The "critical period" in the early life history of
marine fishes. J. Cons. 21:160-170.
May, R. C.

1974. Larval mortality in marine fishes and the critical
period concept. In J. H. S. Blaxter (editor), The early
life history offish, p. 3-19. Springer- Verlag, N.Y.
MILLER, C. B.

1972. Zooplankton indicators of the seasonal cycle of
currents along the Oregon coast. Trans. Am. Microsc.
Soc. 91:87.
MOOERS, C. N. K.

1970. The interaction of internal tide with the frontal zone
of a coastal upwelling region. Ph.D. Thesis, Oreg. State
Univ., Corvallis, 480 p.
OSTLE, B.

1963. Statistics in research. 2nd ed. Iowa State Univ.
Press, Ames, 585 p.
PATTULO, J. G., AND W. DENNER.

1965. Processes affecting seawater characteristics along
the Oregon coast. Limnol. Oceanogr. 10:443-450.

PEARCY, W. G., AND R. S. MESECAR.

1970. Scattering layers and vertical distribution of oceanic
animals off Oregon. In G. Brooke Farquhar (editor).
Proceedings of an international symposium of biologi-
cal sound scattering in the ocean, p. 381-394. Maury
Cent. Ocean Sci., Dep. Navy, Wash., D.C.

Peterson, w. K.

1972. Distribution of pelagic copepoda off the coasts of
Washington and Oregon during 1961 and 1962. In A. T.
Pruter and D. L. Alverson (editors). The Columbia River
estuary and adjacent ocean waters, p. 313-343. Univ.
Wash. Press, Seattle.
POOLE, R. L.

1966. A description of laboratory-reared zoea of Cancer
magister Dana, and megalopae taken under natural con-
ditions (Decapoda, Brachyura). Crustaceana 11:83-97.

POSGAY, J. A., R. R. MARAK, AND R. C. HENNEMUTH.

1968. Development and tests of new zooplankton sam-
plers. Int. Comm. Northwest Atl. Fish., Res. Doc. 68/85,
ser. 2085, 23 p.

Reed, P. H.

1969. Culture methods and effects of temperature and
salinity on survival and growth of Dungeness crab

LOUGH: LARVAL DYNAMICS OF DUNGENESS CRAB

{Cancer magister) larvae in the laboratory. J. Fish. Res.
Board Can. 26:389-397.
RENFRO, W. C, J. C. MCCAULEY, B. GLENNE, R. H. BOURKE,

D. R. Hancock, and S. W. Hager.

1971. Oceanography of the nearshore coastal waters of
the Pacific Northwest relating to possible pollution.
Vol. I, II. Water Qual. Off, Environ. Prot. Agency,
Grant No. 16070 EOK.
Sastry, a. N., and J. F. McCarthy.

1973. Diversity in metabolic adaptation of pelagic larval
stages of two sympatric species of brachyuran crabs.
Neth. J. Sea Res. 7:434-446.
SCHMITT, W. L.

1921. The marine decapod Crustacea of California.
Univ. Calif Publ. Zool. 23:1-470.
SCHOENER, T. W.

1969. Models of optimum size for solitary predators.
Am. Nat. 103:277-313.
SCHWARTZLOSE, R. A.

1964. Nearshore currents of western United States,
Baja, California, as measured by drift bottles. Calif
Coop. Oceanic Fish. Invest., Prog. Rep. 1 July 1960 -
June 1962, p. 15-22.
SMITH, R. L., J. G. PATTULLO, AND R. K. LANE.

1966. An investigation of the early stage of upwelling
along the Oregon coast. J. Geophys. Res. 71:1135-1140.
SVERDRUP, H. U., M. W. JOHNSON, AND R. H. FLEMING.

1942. The oceans their physics, chemistry, and general
biology. Prentice-Hall, N.Y., 1087 p.

Tate, M. W., and R. C. Clelland.

1957. Nonparametric and shortcut statistics in the social,
biological, and medical sciences. Interstate Printers
and Publishers, Inc., Danville, 111., 171 p.

THORSON, G.

1946. Reproduction and larval development of Danish
marine bottom invertebrates, with special reference to
the planktonic larvae in the sound (Oresund). Medd.
Komm. Dan. Fisk.-og. Havunders. Ser. Plankton.
4(l):l-523.

1950. Reproductive and larval ecology of marine bottom

invertebrates. Biol. Rev. (Camb.) 25:1-45.
1964. Light as an ecological factor in the dispersal and

settlement of larvae of marine bottom invertebrates.

Ophelia 1:167-208.
1966. Some factors influencing the recruitment and

establishment of marine benthic communities. Neth.

J. Sea Res. 3:267-293.

U.S. Environmental Data Service.

1970. Climatological data, Oregon. Environ. Sci. Serv.
Admin., Dep. Commer. 76.

1971. Climatological data, Oregon. Environ. Sci. Serv.
Admin., Dep. Commer. 77.

WALDRON, K. D.

1958. The fishery and biology of the Dungeness crab
(Cancer magister Dana) in Oregon waters. Fish. Comm.
Oreg., Contrib. 24, 43 p.

WiNSOR, C. P., AND G. L. Clarke.

1940. A statistical study of variation in the catch of
plankton nets. J. Mar. Res. 3:1-34.

Wooster, w. S., and J. L. Reid, Jr.

1963. Eastern boundary currents. In M. N. Hill (editor),
The sea. Vol. 2, p. 253-280. John Wiley & Sons, Inc.,
Lond.

Wyatt, B., and W. Gilbert.

1971. Surface temperature and salinity observations at
Pacific Northwest shore stations during 1970. Data Rep.
47. Dep. Oceanogr. Oreg. State Univ., Ref 71-8, 19 p.

1972. Surface temperature and salinity observations at
Pacific Northwest shore stations during 1971. Data
Rep. 51. Dep. Oceanogr. Oreg. State Univ., Ref 72-2,
15 p.

Wyatt, B., D. A. Barstow, W. E. Gilbert, and J. L.
Washburn.

1971. Drift bottle recoveries and releases off the Oregon
coast 1961 through 1970. Data Rep. 50. Dep. Oceanogr.
Oreg. State Univ., Ref 71-36, 57 p.

Wyatt, B., w. v. Burt, and j. g. Pattullo.

1972. Surface currents off Oregon as determined from
drift bottle returns. J. Phys. Oceanogr. 2:286-293.

375

FISHERY BULLETIN; VOL. 74, NO, 2

APPENDIX

Table l. — Cancer magister larval abundance for 1970-71 sea-
sons and associated environmental data used in analyses of
multiple covariance.

0.2 m-bongo

net

Surface

sampler mesh

Larval

temp.

salinity

Date

size (mm)

abundance'

('C)2

C/oo)

1970:

1/29

0571
0233

4,598
2,607

10.3

29.55

2/13

0.571
0.233

6,721
13,276

11.0

30.25

2/25

0.571
0233

65
152

11.4

30.62

3/9

0.571
0.233

536
381

10.8

30.99

4/16

0.571
0.233

88
0

9.7

33.06

4/27

0.571
0.233

7,713
5,267

9.5

32.98

5/1

0.571
0233

1,868
1,847

9.5

32 98

5/6

0.571
0.233

89
39

9.2

32.98

5/22

0.571
0233

1,817
1,697

11.8

32.51

6/4

0.571
0.233

74
32

9.1

33.62

6/23

0.571
0.233

0
26

7.9

33.60

7/2

0.571
0 233

21
38

12.5

32.84

7/16

0 571
0233

56
28

9.6

32.73

7/29

0.571
0.233

0
0

12.7

32.63

1971:

1/18

0.571
0.233

736

1,007

9.9

29.50

2/3

0.571
0233

1,762
1,930

8.6

31.80

2/16

0.571
0.233

2,539
3,408

9.3

31.00

3/20

0 571
0.233

205
21

8.5

32.16

3/30

0.571
0.233

305

316

30.81

4/22

0.571
0233

390
999

9.6

3245

5/03

0571
0.233

0
0

10.4

30.74

5/14

0.571
0.233

0
0

11.5

31,76

5/29

0.571
0.233

26
25

8.8

33.69

6/2

0571
0.233

0
0

10.1

33.52

6/12

0.571

13.4

30.04

6/28

0571

15.1

30.75

7/6

0.571

11.7

31.83

7/21

0.571

9.0

33.53

'Numtjer of larvae per 4,000 m^. Larvae summed over tour instiore stations;
NH01, NH03, NH05, NH10.
^Averaged values over four instiore stations.

376

SUBTIDAL AND INTERTIDAL MARINE FOULING ON

ARTIFICIAL SUBSTRATA IN

NORTHERN PUGET SOUND, WASHINGTON^

Charles H. Hanson^ and Jonathan Bell=*

ABSTRACT

The design and siting of power plant cooling systems requires detailed information concerning the
fouling tendencies of specific organisms on specific construction materials. This study, conducted in the
vicinity of Kiket Island, northern Paget Soimd, Wash., attempts to provide some of this information.
The sessile community characteristics of five materials exposed at three depths and two locations in
the subtidal zone, and of one material in the intertidal zone are described. The degree of biofouling was
least for copper-nickel alloy and progressively greater for Plexiglas, wood, steel, and concrete. Media
decay and biological accumulation was greatest at the near-surface level, decreasing in intensity with
increasing depth. The maximum rate of colonization occurred during the late spring (April-June) and
early fall (mid- August-October). The present study, an analysis of biofouling, indicates that if the
proposed power plant were to be built at Kiket Island, its cooling system intake should be sited in water
deeper than 6 m and should have a safe and adequate fouling control scheme.

The settlement of entrained fouling organisms
seriously affects the proper functioning of indus-
trial cooling systems (Dobson 1946; Beauchamp
1966; Holmes 1970). Thus, the design of a cooling
system requires detailed information concerning
the fouling tendencies of specific organisms on
specific construction materials. The present study
— conducted in the vicinity of Kiket Island, north-
ern Puget Sound, Wash. — attempts to provide
some of this information. At the time, the study
area was the proposed site for a 1,000 MW nuclear
power plant with a once-through cooling system.

The study analyzed the fouling resistances of
several common construction materials both in
the subtidal and in the intertidal zones. Coloniza-
tion in the subtidal zone was examined from April
to November 1972, while colonization in the inter-
tidal zone was examined from December 1971 to
September 1972. Short-term (series I) and long-
term (series II) exposures of test materials pro-
vided information about the rate of fouling ac-
cumulations and progressive community change.
The study also determined the seasonal and verti-
cal distribution of the dominant fouling organisms
endemic to the Kiket Island area. These exposures

'Contribution No. 433 from the College of Fisheries, Univer-
sity of Washington, Seattle, WA 98195.

^Fisheries Research Institute, University of Washington,
Seattle, WA 98195; present address; Department of Animal
Physiology, University of California-Davis, Davis, CA 95616.

^Johns Hopkins University, Baltimore, MD 21218.

also allowed a determination of the periods of
maximum colonization by fouling organisms.

MATERIALS AND METHODS
Subtidal Fouling

Two test sites for the study of subtidal fouling
were established offshore from Kiket and Skagit
islands (Figure 1), in water of a mean depth of 18
m. At each test site five construction materials
were tested for their resistances to fouling. The
materials that were tested included a 90%
copper-10% nickel alloy, steel, Plexiglas'' (an ac-
rylic plastic), white pine wood, and concrete. The
materials were cut into 10 cm x 10 cm squares —
54 squares each of steel, Plexiglas, and wood; 18
squares each of copper-nickel alloy and concrete.
The squares or "plates" had two 12.7-mm holes
drilled into opposite corners of the plate. Rope was
threaded through the holes and the plates were
then separated into 18 "test panels"— each panel
having three plates of steel, Plexiglas, and wood,
and one plate of copper-nickel alloy and concrete.
Within each panel there was a random distribu-
tion of plates.

The test panels were suspended in the water at
mean depths of 1, 6.1, and 15.3 m below the sur-

Manuscripf accepted December 1975.
FISHERY BULLETIN; VOL. 74, NO. 2, 1976.

"Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

377

FISHERY BULLETIN: VOL. 74, NO. 2

123°

.Oecjeption x^j,^_ Skagilj River

Evei ett

FIGURE 1. — Map of western
Washington, with the study areas
shown in inset: 1) Kiket Island sub-
tidal site; 2) Skagit Island subtidal
site: and 3) intertidal site.

face. The 1-m depth test panels were suspended
from a steel surface float. The 6.1- and 15.3-m
depth test panels were suspended between a con-
crete bottom anchor and a steel float moored just
below the extreme low water level. At both test
sites three panels were deployed at each depth
(Figure 2).

Series I test panels were exposed for periods of
41 and 79 days offshore from Skagit Island and
were exposed for 58 and 101 days offshore from
Kiket Island. Series II test panels were exposed
continuously for a period of 8 mo (16 April-29
November 1972) at both locations.

The standard analytical procedure for series I
plates involved identification of the organisms,
estimation of the percent of plate coverage, and, if
possible, a measurement of the size of the or-
ganisms. A central square of each plate, measur-
ing 7 cm X 7 cm, was used for analysis. The fouling
organisms on each 49-cm2 central area were
scraped onto preweighed filter paper, dried at ap-
proximately 100°C for 24 h, and then weighed to
0.01 g. Monthly qualitative observations of series
II plates, anchors, lines, and floats were made using
scuba.

Intertidal Fouling

An examination was made of the settling rate of
intertidal fouling organisms on concrete slabs.
The concrete slabs measured 38 cm wide by 76 cm
long by 15 cm deep. The slabs were uniform in
texture, composition, surface configuration, sta-
bility, and resistance to wave action. They were
anchored to the beach with steel reinforcing bars
imbedded in the concrete. The long dimension was
parallel to the water and the top surface was
placed horizontal to the plane of the water. The
slabs were positioned at the +0.6-, 0-, -0.6-, and
— 1.2-m water levels relative to mean sea level.
Once each month the density of the fouling or-
ganisms was determined from a series of randomly
chosen 49-cm2 areas on each concrete slab.

RESULTS
Physicochemical Environment

Seasonal water quality data for the Kiket Island
area have been described in detail (Stober et al.

378

HANSON and BELL: SUBTIDAL AND INTERTIDAL MARINE FOULING

SURFACE FLOAT

SUFIFACE PANEIS

consisted of soft silt and sediment with a few rock
outcroppings. Acorn barnacles, BaZanas crenatus,
densely covered the few rock outcroppings, but
were otherwise not present. At a depth of 18 m,
light penetration was low and bottom currents
appeared to be generally slow. In contrast, the
bottom at Skagit Island was virtually free of silt
and was predominantly covered with cobble and
rock outcroppings. The cobble and rock were
densely covered with B. crenatus. At 18 m,
light penetration was moderate and the bottom
currents were consistently more rapid than those
at the Kiket Island site.

Fouling Colonization of
the Construction Materials

The fouling resistances of the different test
materials were compared using the dry weights of
organisms collected during periodic sampling. The
dry weight data for the 1-m level are shown in
Table 1. Weight data of the removable material
from the 15.3-m and 6. 1-m levels were negligible
except for the plates of wood and concrete col-
onized by Balanus crenatus (Table 2).

,', BOTTOM PANELS
I

COPPER-NICKEL ALLOY.— There was no
removable material through the first 58 days. The

Figure 2. — A schematic of the array of subtidal test panels used
to measure biofouling with inset showing details of test plate
attachment.

1973). Weekly minimum, mean, and maximum
temperature and salinity readings are presented
in Figure 3. Average weekly temperatures ranged
from 6.2°C to 11.8°C. Average weekly salinities
ranged from 17.5 to 29.7 g/liter; pH ranged from
7.1 to 8.2; and dissolved oxygen concentrations
ranged from 10.5 to 13.3 mg/liter. Lincoln et al.
( 1970) and Bendiner et al. ( 1972) have detailed the
physical oceanography and vertical stratification
of the Kiket Island area. The physicochemical
characteristics of North Skagit Bay led Stober et
al. (1973) to classify the study area as a well mixed
estuary.

Qualitative observations of the study area were
made periodically while scuba diving. The bottom
in the vicinity of the fouling plates at Kiket Island

32 p

30
_ 28
°! 26
>. 24
1 22
o 20
'^ 18

' '"'■i'i'iilil'|i'l"l|| 4" •'■

■■''''

DEC.
1971

JAN FEB MAR, APR

MAY JUNE
1972

JULY AUG, SEPT.

20
19
18
17
16
15
14
13
12
II
10
9
8
7
6
5
4
3
2

E'ffifiijjjipitfo''"

II I I I 1 I I I r [ .

I I I I I I I I ■' I I I I I ■ I

DEC.
1971

FEB. MAR. APR.

MAY JUNE
1972

JULY AUG. SEPT.

30
28
26
24
22
20
IB
16
14
12

68
66
64
62

60
58
56 —

54 t

52 a.

50 ^

48 ^

Figure 3. — Weekly mean, minimum, and maximum salinity
measurements (a) and water temperature (b) recorded in the
Kiket Island area (data from Stober et al. 1973).

379

FISHERY BULLETIN: VOL. 74, NO. 2

Table l. — Dry weight in grams of material collected per square
centimeter of surface area from five artificial media exposed
at the near-surface level for four time periods.

Exposure

Copper-
nickel
alloy

Steel

Plexiglas

Wood

Concrete

May 26

41 -day exposure

0.00

0.15

0.19

0.13

0.12

June 12

58-day exposure
July 3
79-day exposure

0.00
0.01

0.26
0.23

0.09
0.08

0.09
0.07

0.07
0.05

July 25

101 -day exposure

0.01

0.13

0.04

0.05

0.04

Table 2.— Density of the barnacle, Balanus crenatus, per
square centimeter of surface area collected from five artificial
media at three depths.

Copper-

Exposure
and depth

nickel
alloy

Steel

Plexiglas

Wood

Concrete

May 26

41 -day exposure
1 m

0.0

0,0

0.0

6.1 m

0.0

15.3 m

0.0

June 12

58-day exposure

1 m

0.0

0.2

0.1

0.2

2.9

6.1 m

0.0

0.1

0.4

0.3

1.6

15.3 m

0,0

0.0

0.4

July 3

79-day exposure
1 m

0,0

0,3

0.8

0.4

4.9

6.1 m

0,0

0.0

0.5

4.6

11.3

15.3 m

0.0

July 25

101 -day exposure
1 m

0.0

0.1

0.2

0.9

2.1

6.1 m

0.0

0.4

15.3 m

0.0

0.1

0.0

79-day and the 101-day samples had removable
material weighing less than 0.01 g/cm^. Remov-
able material consisted primarily of diatoms, with
small deposits from flaking of the alloy surface. No
mussels, barnacles, or green algae were observed.

STEEL.— After 41 days the dry weight of the
removable material was 0.15 g/cm^, after 58 days
0.26 g/cm2, after 79 days 0.23 g/cm^, and after 101
days 0.13 g/cm^. A high proportion of the remov-
able material was rust; biological accumulations
consisted of diatoms, barnacles, green algae, and
mussels. Balanus crenatus densities at the 1-m
depth ranged from 0.0 on day 41 to 0.3/cm2 on day
19. Balanus density at the 6. 1-m level ranged from
0.0 on day 41 to O.l/cm^ on day 58. No Balanus
were found at the 15.3-m level.

PLEXIGLAS. — After 41 days the dry weight of
the removable material was 0.19 g/cm^, after 58

days 0.09 g/cm^, after 79 days 0.08 g/cm^, and after
101 days 0.04 g/cm^. Removable material con-
sisted of diatoms, green algae, and barnacles.
Mussels were not observed on the Plexiglas media.
The density of Balanus crenatus at the 1-m level
ranged from 0.0 on day 41 to O.S/cm^ on day 79. At
the 6. 1-m \eve\ Balanus densities ranged from 0.0
on day 41 to 0.5/cm2 q^ day 79. No Balanus were
found at the 15.3-m level.

WOOD.— After 41 days the dry weight of the
removable material was 0.13 g/cm^, after 58 days
0.09 g/cm2, after 79 days 0.07 g/cm^, and after 101
days 0.05 g/cm^. Removable material consisted
primarily of diatoms and barnacles with small
amounts of green algae. No mussels were found.
The density of Balanus crenatus at the 1-m level
ranged from 0.0 on day 41 to 0.9/cm2 on day 101.
Balanus density at the 6. 1-m level ranged from 0.0
on day 41 to 4.6/cm2 on day 79. Balanus density at
the 15.3-m level ranged from 0.0 on day 41 to
39/cm2 on day 19. Balanus achieved 100% of plate
coverage on day 79 at the 15.3-m level. No wood
borers were found at any level.

CONCRETE.— After 41 days the weight of the
removable material was 0.12 g/cm^, after 58 days
0.07 g/cm2, after 79 days 0.05 g/cm^, and after 101
days 0.04 g/cm^. Removable material consisted of
diatoms, barnacles, mussels, and green algae. The
density of Balanus crenatus at the 1-m level
ranged from 0.0 on day 41 to 4.9/cm2 on day 79.
Balanus density at the 6. 1-m level ranged from 0.0
on day 41 to 1 1.3/cm2 on day 19. Balanus density at
the 15.3-m level ranged from 0.0 on day 41 to
64/cm2 on day 79. Balanus achieved 100% plate
coverage on day 79 at the 15.3-m level. Bay mus-
sels, Mytilus edulis, achieved a density of 0.4/cm2
at the 1-m level — none were found in the deeper
water samples.

Intertidal Fouling

The colonization of fouling organisms was ob-
served on concrete test slabs positioned at various
levels in the intertidal zone of Kiket Island. The
principal algae species colonizing the slabs were
Fucus distichus and Ulva lactuca. The dominant
animal species included the acorn barnacle,
Balanus glandula, and the bay mussel. A detailed
examination of the natural vertical and seasonal
distribution of the intertidal flora and fauna of
Kiket Island is presented by Houghton (1973).

380

HANSON and BELL: SUBTIDAL AND INTERTIDAL MARINE FOULING

Settlement by barnacles (Figure 4) was the most
rapid during late May. Barnacle density peaked in
June, but subsequently there was a general de-
crease in the density — probably due to intra-
specific competition for the limited growing area.
Barnacle settlement was most successful at the
-0.6-m level. There was limited settlement at the
— 1.2-m level and at the 0.0-m level. No barna-
cles successfully settled at the +0.6-m level.

Settlement by barnacles at the —1.2-m level ap-
peared to be limited by heavy siltation and
diatomaceous growth. The absence of barnacles at
the -1-0. 6-m level was principally caused by the
extensive exposure of the organisms to sunlight.
Successful settlement at the -0.6-m level was the
result of a limited exposure to sunlight and of the
moderate wave action limiting the silt/diatom
buildup.

Settlement by M. edulis was predominant at
the -0.6-m level, where maximum density was
l.S/cm^. Mytilus edulis were present in lower
densities at the 0.0-m and -1.2-m levels. The
same factors affecting settlement by barnacles —
exposure to sunlight and the silt-diatom build-
up— affected settlement by M. edulis. Mussels
were observed to attach primarily in the late
summer and in the fall (July-October); a few in-
dividuals were observed in April and May.

Seasonal Distribution of
Fouling Organisms

The seasonal distribution of the major sessile
fouling organisms found in the Kiket Island area
is presented in Figure 5. Conclusions about the
distribution of these organisms are based on data
collected during a iy2-yr study of intertidal
settlement and an 8-mo study of subtidal foul-
ing. Comparable conclusions were reached by
DePalma (1966) for Admiralty Inlet.

The first diatoms to appear on the study plates
were those of the genus Melosira. These diatoms
remained dominant throughout the study period.
Navicula and Fragilaria, as well as a large
number of unidentified diatoms, also settled on the
plates, but were not nearly as abundant as Melo-
sira. Although the spores of many diatom spe-
cies were present all year, settlement occurred pre-
dominantly from early spring to midsummer.

Four dominant forms of algae settled on the
study plates. Fucus distichus and Ulva lactuca
were dominant in the intertidal zone, while Ulo-

FlGURE 4. — Mean density o{ Balanus glandula attached to
concrete substrata exposed in the intertidal zone of Kiket
Island (tidal level relative to mean sea level).

Diatoms

Algae

B glandula

B. cariosus

B. crenatus

C. dalli

M. edulis

— c

€Z

>-
I>-

— <~zzME:y —

jrr

1972

Figure 5. — Seasonal distribution of predominant subtidal and
intertidal fouling organisms.

thrix sp., Cladophora sp., and Ulva lactuca were
dominant in the subtidal zone. The algae was
abundant seasonally — in the spring and summer
there was an extensive algal cover on the plates,
yet in the fall and winter months the abundance of
algae decreased substantially. Many small crusta-
ceans, including copepods, cladocerans, and am-
phipods, were observed inhabiting the diatoms
and algae covering the test plates.

Although barnacles of the genus Balanus were
present throughout the year, their rate of settle-
ment varied greatly with the different seasons. As
a general rule, maximum settlement occurred
during the late spring (April-June) and in early
fall (mid- August-October). For example, B. glan-
dula settled in the intertidal zone from February

381

through November, but the maximum rate of set-
tlement was observed in May and August. How-
ever, in the subtidal zone,S. crenatus settled from
April to November, but with a peak in late July
and early August. Others, like B. cariosus
and Chthamalus dalli, settles sporadically from
May to November and peaked in August and
September.

Settlement by the bay mussel occurred primar-
ily from August to November, although there was
some settlement during April and May. It appears
that prior settlement by diatoms, algae, and bar-
nacles is necessary for the establishment of a mus-
sel colony. Cleaned test plates were exposed in
both the intertidal and the subtidal zones and
were compared with plates already having an es-
tablished community of diatoms, algae, and bar-
nacles. Only on those plates which were already
fouled was there any settlement by mussels. Coe
(1932) reported the same phenomenon and con-
cluded that the smooth quality of nonfouled sur-
faces was not suitable for attachment by the bys-
sus of young mussels.

Vertical Distribution in
the Subtidal Zone

At both subtidal test sites there was a distinct
vertical pattern to the fouling of the test plates.
The greatest number of species settled at the
near-surface (1-m) level. At that level there were
colonial diatoms of the genera Melosira, Navicula,
and Fragilaria, and three species of the acorn bar-
nacle, Balanus crenatus, B. glandula, and B.
cariosus. Subdominant genera included the green
algae, Ulothrix, Ulva, and Cladophora. Small
numbers of the bay mussel were also found at the
surface level. At the middle depth (6.1 m) the
species composition of the fouling organisms
changed. Green algae became rare and diatoms
were less dense. Mytilus edulis, B. glandula, and
B. cariosus were absent. Balanus crenatus in-
creased in density with increasing depth at Skagit
Island, but not at the Kiket Island site.

The 15.3-m level was very different from the two
upper levels. The plates had no algae or diatoms.
Balanus crenatus was the dominant species. Con-
sistently higher densities of 5. crenatus were ob-
served at the Skagit Island test site. The ratio of
densities between Skagit and Kiket Island for B.
crenatus at the 15.3-m level ranged as high as 50 to
1 for the wood and concrete test plates.

FISHERY BULLETIN: VOL. 74, NO. 2

DISCUSSION

Marine fouling presents one of the most serious
long-term operational problems for power
generating stations using saline waters for cooling
(Powell 1933; Dobson 1946; Holmes 1970). Foul-
ing accumulations reduce the carrying capacity of
cooling system conduits by increasing the fric-
tional resistance and by reducing the pipeline
diameter. In addition, marine fouling reduces the
heat transfer efficiency of steam condenser sys-
tems and promotes severe corrosion of the con-
denser system components. The accumulation of
fouling debris, such as dead shells, adds to the
inefficiency by clogging the condenser tubes.

Data are needed by design engineers in order to
determine the probable construction require-
ments for the control of fouling in a power plant
cooling system. Because marine fouling varies
considerably from one location to another, an on-
site determination of the population dynamics of
fouling organisms is desirable. Each site should be
studied in order to determine: 1) the species com-
position of sessile organisms colonizing specific
construction materials at various subtidal levels,
2) the types of construction materials least likely
to be fouled, 3) the seasonal variations in settle-
ment and abundance, and 4) the times of the year
when antifouling procedures must be considered.
The present study was, in a sense, an attempt to
study all these factors, and although the power
plant for which the study was intended may never
be built, this report should be a useful guide to
future studies of power plant siting.

Data for the present study were collected from
test plates suspended at various depths in the wa-
ter. However, caution must be used in extrapolat-
ing studies carried out with these small static test
plates. Graham and Gay (1945) reported that
plates, 9.8 cm x 9.8 cm, were found to give results
just as reliable as larger ones. Holmes (pers. com-
mun.), however, considers that "edge effects and
top-to-bottom gradients could be very important
in biassing results from such small panels." Al-
though no effort was made in the present study to
determine the reliability of the small plates, a
3-cm border zone surrounding the 49-cm2 exami-
nation area was considered sufficient to eliminate
any edge effect. There was consistently less than
10% variation in the dry weights of the removable
material and in the density of barnacles taken
from different plates of the same media at the
same water level.

382

HANSON and BELL: SUBTIDAL AND INTERTIDAL MARINE FOULING

One must also recognize that data collected from
static test panels can only give a limited indication
of the growth rate of fouling organisms in a
continuous-flow cooling system (Dobson 1946).
Fouling organisms naturally dependent upon
water currents to supply food, may have their
growth rates enhanced by the greater water veloc-
ities of a continuous-flow cooling system (Dob-
son 1946; Benson et al.^). Mawatari^ observed,
however, that test panels exposed in current ve-
locities of 4 to 7 m/s remained totally free of
fouling organisms. Efforts to reduce the influence
of static plates have been made by several authors
(Smith 1946; Doochin and Smith 1951; Wood
1955), but these efforts have produced conflicting
results.

Several additional factors should be mentioned
which influence both the growth rate and the
species composition of sessile organisms coloniz-
ing test plates. The larvae of barnacles and many
other fouling organisms have been found to be
negatively phototrophic when they attach to a sur-
face. Therefore, these organisms prefer to attach
to shaded or dark surfaces (Visscher and Luce
1928; Thorson 1964). Also, surface texture has
been shown to affect the rate of attachment of
settling larvae (Crisp and Ryland 1960; Pomerat
and Weiss'). In general, porous and rough surfaces
have the greatest fouling accumulation.

All of these factors influenced the results ob-
tained by the present study. For example, the test
plates, although they were subjected to natural
flow currents of the marine environment, were not
subjected to the "unnatural" flow currents of a
power plant cooling system. Thus, fouling on the
test plates might be somewhat different from the
fouling of a cooling system. Yet the test plates offer
useful indications as to what will happen in
the actual cooling system and therefore they are
useful for predictive planning of power plant
engineers.

In the present study, vgu-iations in the abun-
dance and species composition of fouling or-
ganisms were observed for the different construc-
tion materials. Accumulation was slow on the

«Benson, P. H., E. L. Littauer, and N. P. Stumbaugh. 1968.
Outlook for rriEirine corrosion and fouling protection. Paper pre-
sented at Symposium on Ocean Technical Problems of the 1970's.
61st Annu. Meet., Los Ang., Calif, Dec. 1968, 42 p.

^awatari, S. 1965. Protection of power plants from biological
fouling. Unpubl. rep. Research Institute for Natural Resources,
Tokyo, Jap.

■'Pomerat, C. M., and C. M. Weiss, 1946. The influence of
texture and composition of surface on the attachment of seden-
tary marine organisms. Unpubl. manuscr.

copper-nickel alloy plates, but was rapid and com-
plete on the concrete and wood plates. Because the
fouling plates were exposed to identical environ-
mental conditions, the differences in fouling resis-
tance must have been dependent upon the differ-
ences between the media. Previous research has
shown the same results — Woods Hole Oceano-
graphic Institution (1952), for example, found
that copper-nickel alloy maintains its fouling
resistance for 10 mo, much longer than concrete
or wood.

Depth was found to have a significant effect
upon the rate of fouling accumulation. For exam-
ple, the dry weight of removable material from all
materials placed below the surface level ( 1 m) was
negligible except for those wood and concrete
plates colonized by Balanus crenatus. Yet at the
surface there was considerable algal and
diatomaceous growth on all media except the
copper-nickel alloy. The only organism which in-
creased in density as the depth increased was B.
crenatus, the only organism colonizing the plates
at the 15.3-m level. Because these results were
similar for all media and because they were cor-
roborated by qualitative examinations made on
the ropes, floats, and anchors, it appears that a
cooling system intake in the Kiket Island area
should be sited in water deeper than about 6 m.
Based on biofouling results, the cooling system
intake should not be sited at the surface because
fouling is greatest at that level.

An analysis of the seasonal distribution of the
fouling organisms showed that there was initially
an accumulation of brown detrital film and bacte-
rial slime on the fouling plates. Soon a filamentous
algae, Enteromorpha, and a diatom, Melosira, be-
came established. As floral density increased,
greater numbers of Crustacea were observed liv-
ing in the growths on the plates. Barnacle and
mussel colonization of the test plates occured
throughout the year, but was greatest from April
through October. For mussels, at least, it ap-
peared that a previous accumulation of fouling
material was required before the mussels would
attach to the test plates. Thus, it would appear
that fouling control should be greatest during the
spring, summer, and early fall. During late fall
and winter fouling control need not be so greatly
emphasized. It must be remembered that the time
for maximum fouling may vary from year to year,
and thus fouling control should be regulated by
routine observations of larval settlement. In ad-
dition, early fouling control may help to deter col-

383

FISHERY BULLETIN: VOL. 74, NO. 2

onization by mussels, which, according to Hoshiai
(1964) and Holmes (1970), are the principal foul-
ing organisms in power plant cooling systems.

The use of intermittent chlorination as a fouling
control agent has been noted by Holmes (1970),
Morris (1971), and Draley (1972). In general, most
investigators feel that the larvae of various
marine fouling organisms are more sensitive to
chronic low-level concentrations of chlorine than
are the adults (Dobson 1946; Turner et al. 1948).
Thus, greatest effectiveness results from repeated
low-level chlorination, which either kills the lar-
vae directly or creates an unfavorable environ-
ment for settlement.

Any fouling control scheme should maintain
adequate precautions against excessive interfer-
ence with organisms inhabiting the receiving
water ecosystem. Chemical toxins such as chlorine
are objected to as antifouling agents primarily
because of the possible detrimental effects on non-
target organisms (Waugh 1964; Hamilton et al.
1970; Stober and Hanson 1974). This effect is par-
ticularly true when the treated effluent is dis-
charged directly into the aquatic environment.

The data presented in this study can only be
called preliminary. Additional tests should be run
which would include at least one complete annual
cycle study of subtidal fouling. Yet the present
study does indicate that if the proposed plant were
to be built at Kiket Island, its cooling system
should be in water deeper than 6 m and should
have a safe and adequate fouling control scheme.
Of the different construction materials tested in
this study, it would appear that copper -nickel
alloy would most effectively deter fouling and that
concrete and wood would be least effective.

It must be emphasized that the present study is
an analysis of biofouling. Prior to the siting and
final design of the cooling water intake structure,
consideration must also be given to the potential
effects of entrainment on zooplankton and larval
and juvenile fish.

ACKNOWLEDGMENTS

The investigation was sponsored in a contract
with Seattle City Light and Snohomish County
P.U.D. as part of a comprehensive biological study
of the Kiket Island nuclear power site. Thanks are
due Q. J. Stober who directed the study and as-
sisted in preparation of the manuscript, Sandi
Hanson for valuable assistance in data collection
and analysis, and E. O. Salo, K. K. Chew, J. P.

384

Houghton, and D. L. Mayer for comments and
suggestions during the study.

LITERATURE CITED

Beauchamp, R.S.

1966. Low-level chlorination for the control of marine foul-
ing. Central Electric Res. Lab., Lab. Memo RD/L/M
147, 16 p.

Bendiner, W. p., T. E. Ewart, and E. H. Linger.

1972. Prediction of excess heat distribution using tracer dye
techniques. Final Rep. APL-UW 7206, Univ. Wash. Appl.
Phys. Lab., Seattle, 83 p.
COE, W. R.

1932 . Season of attachment and rate of growth of sedentary
marine organisms at the pier of the Scripps Institution of
Oceanography, La JoUa, California. Bull. Scripps Inst.
Oceanogr. Tech. Ser. 3:37-87.

Crisp, D. J., and J. S. Ryland.

I960. Influence of filming and of surface texture on the
settlement of marine organisms. Nature (Lond.) 185:119.
DEPALMA, J. R.

1966. A study of the marine fouling and boring organisms
at Admiralty Inlet, Washington. Inf. Ms. Rep. 0-6-66,
Oceanogr. Surv. Dep.. U.S. Nav. Oceanogr. Off, Wash.,
D.C., 32p.
Dobson, J. G.

1946. The control of fouling organisms in fresh- and saltwa-
ter circuits. Trans. Am. Soc. Mech. Eng. 68:247-265.

DoocHiN, H., AND F. G. W. Smith.

1951. Marine boring and fouling in relation to velocity of
water currents. Bull. Mar. Sci. Gulf Caribb. 1:196-208.
DRALEY, J. E.

1972. The treatment of cooling waters with chlo-
rine. ANL/ES-12 Feb. 1972. Argonne National Lab-
oratory, Lemont, 111., 11 P-

Graham, H. W., and H. Gay.

1945. Season of attachment and growth of sedentary
marine organisms at Oakland, California. Ecology
26:375-386.

Hamilton, D. H., Jr., D. a. flemer, C. W. Keefe, and J. A.

MIHURSKY.

1970. Power plants; effects of chlorination on estuarine
primary production. Science (Wash., D.C.) 169:197-198.

Holmes, N.

1970. Marine fouling in power plants. Mar. Pollut. Bull.
1:105-106.
HOSHIAI, T.

1964. Distribution of sessile animals in the intake-duct of
the cooling sea water of the Hachinohe thermal power
station. Asamushi Mar. Biol. Stn. Bull. 12:42-50.
HOUGHTON, J. P.

1973. Intertidal ecology. In Q. J. Stober and E. O. Salo
(editors), Ecological studies of the proposed Kiket Island
nuclear power site, p. 119-257. Final Rep. to Snohomish
County P.U.D. and Seattle City Light. Univ. of Wash.
Coll. Fish., Fish. Res. Inst., Seattle.

LINCOLN, J., E. E. COLLIAS, AND C. S. BARNES.

1970. Skagit Bay study. Prog. Rep. 3. Univ. Wash. Dep.
Oceanogr. Ref M70-111, 88 p.

MORRIS, J. C.

1971. Chlorination and disinfection — state of the art. J.
Am. Water Works Assoc. 63:769:774.

HANSON and BELL: SUBTIDAL AND INTERTIDAL MARINE FOULING

Powell, S. T.

1933. Slime and mussel control in surface condensers and
circulating water tunnels. Combustion (April):7-13.

Smith, F. G. W.

1946. Effect of water currents upon the attachment and
growth of barnacles. Biol. Bull. (Woods Hole) 90:51-70.

Stober, Q. J., AND C. H. Hanson.

1974. Toxicity of chorine and heat to pink (Oncorhynchus
gorhuscha) and chinook salmon (O. tshawytscha). Trans.
Am. Fish. Soc. 103:569-576.

Stober, Q. J., S. J. Walden, and D. T. Griggs.

1973. Seasonal water quality in North Skagit Bay. In Q. J.
Stober and E. O. Salo (editors). Ecological studies of the
proposed Kiket Island nuclear power site, p. 7-34. Final
Rep. to Snohomish County P.U.D. and Seattle City
Light. Univ. of Wash. Coll. Fish., Fish. Res. Inst., Seattle.
THORSON, G.

1964. Light as an ecological factor in the dispersal and
settlement of larvae of marine botttom invertebrates.
Opheha 1:167-208.

Turner, H. J., Jr., d. M. Reynolds, and a. C. Redfield.

1948. Chlorine and sodium pentachlorophenate as fouling
preventives in sea water conduits. Ind. Eng. Chem.
40:450:453.

VISSCHER, J. P., AND R. H. LUCE.

1928. Reactions of the cyprid larvae of barnacles to light
with special reference to spectral colors. Biol. Bull.
(Woods Hole) 54:336-350.

WAUGH, G. D.

1964. Observations on the effects of chlorine on the larvae of
oysters (Ostrea edulis (L. )) and barnacles (Eliminius mod-
estus (Darwin)). Ann. Appl. Biol. 54:423-440.

WOOD, E. J. F.

1955. Effect of temperature and rate of flow on some marine
fouling organisms. Aust. J. Sci. 18:34-37.

WOODS HOLE OCEANOGRAPHIC INSTITUTION.

1952. Marine fouling and its prevention. U.S. Nav. Inst.,
Annapolis, 388 p.

385

OBSERVATIONS ON THE FISH FAUNA ASSOCIATED WITH
OFFSHORE PLATFORMS IN THE NORTHEASTERN GULF OF MEXICO

Robert W. Hastings,^ Larry H. Ogren.^and Michael T. Mabry^

ABSTRACT

The fish fauna associated with two U.S. Navy research platforms. Stage I and Stage 11, in the
northeastern Gulf of Mexico off Panama City, Fla., was studied at irregular intervals from 1970 to
1974. Such platforms function as artificial reef habitats and support diverse and abundant fish
populations not normally characteristic of the open sandy bottoms in the area.

A total of 101 taxa (identified to family or species) was recorded at the two platforms; 61 species
were observed at Stage I in water 32 m deep and 86 taxa at Stage II in water 18 m deep. The greater
number of species recorded at the shallower location may be more a result of the greater number of
observations made there than of differences in the two habitats. The number of species present at the
platforms varies considerably at different times of the day and year. Species numbers are greatest
during the summer and fall, but many species begin to move offshore or southward as the water
temperature drops, and only about 50-60% of those recorded at the platform remain in December. The
number of species diminishes to about 16% in February at Stage II, then increases gradually with the
rising water temperature in the spring.

Major species occupying the platform habitats include fishes usually characteristic of pelagic,
inshore (coastal or estuarine), and rocky reef environments. At the platforms, the pelagic species and
most of the larger predators occupy various levels of the water column, either directly below or
surrounding the structure, while most of the other species are associated either with the pilings and
cross-members of the platform or with the bottom. For some of the species, the platform provides food
and shelter, while for others, it offers only shelter. Some species may be present only to feed on the
numerous fishes and other organisms concentrated there. Diel rhythms of activity are obvious for
many of the fishes, with some species active only during the day, and others only at night.

Offshore oil drilling platforms are known to at-
tract various species of marine fishes and thus
function as artificial reefs (Carlisle et al. 1964;
Treybig 1971). Anglers often recognize such plat-
forms as desirable fishing sites. Carlisle et al.
(1964) documented the development of fish popu-
lations (as well as populations of encrusting or-
ganisms) at two platforms constructed off the
coast of California. The supporting piles and
cross-members of such platforms provide hard
substrates for the settling of pelagic larvae of en-
crusting invertebrates and algae which, with
their associated invertebrate populations, pro-
vide food and shelter for reef fishes attracted to
the structures. In addition, many pelagic fishes
congregate about these platforms, attracted
either by the solid, reeflike nature of the support-
ing structures, or by the numerous smaller forage
organisms in the area.

'Department of Biology, Rutgers University, Camden, NJ
08102.

^Gulf Coastal Fisheries Center Panama City Laboratory, Na-
tional Marine Fisheries Service, NOAA, Panama City, FL
32401.

'Tampa Marine Institute, 1310 Shoreline Drive, Tampa, FL
33605.

Manuscript accepted September 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

Many comparable platforms have been con-
structed in the Gulf of Mexico since the 1940's,
but no studies of their associated fish faunas have
been reported, even though they are known to
attract numerous species of fishes. Current
studies by personnel of the University of South-
western Louisiana have documented the fish
fauna of drilling platforms off the coast of
Louisiana (Sonnier et al. 1976). This paper re-
cords the fish populations observed around two
offshore platforms in the northeastern Gulf of
Mexico off Panama City, Fla.

LOCATION AND TIME OF STUDY

Two research platforms operated by the U.S.
Navy off the coast of Panama City are referred to
as Stage I and Stage II. Stage I is 17.7 km offshore
in water 32 m deep (lat. 30°00.5'N, long.
85°54.2'W). Stage II (Figure 1) is 3.2 km offshore
in water 18 m deep (lat. 30°07.2'N, long.
85°46.4'W). The pilings of Stage I form a square
on the sea bottom with each side measuring 32.6
m, whereas those of Stage II measure 19.1 m. The
two platforms were the sites of biofouling studies

387

FISHERY BULLETIN: VOL. 74, NO. 2

Figure l.— Stage II, the Navy re-
search platform 3.2 km offshore of
Panama City, Fla. (U.S. Navy photo.)

by Pequegnat et al. (1967), Pequegnat and
Pequegnat (1968), and Culpepper and Pequegnat
(1969). Vick (1964) mentioned 13 species of fishes
either collected or reported from the stages and
vicinity.

These platforms were examined occasionally
from 1970 to 1974 by the first and second authors
in connection with studies of reef fishes in the
northern Gulf of Mexico. During the fall of 1970
(Ogren) and summer of 1972 (Hastings and
Mabry), the authors participated in the Scien-
tist-in-the-Sea (SITS I and 11) diving program at
the Naval Coastal Systems Laboratory in Pan-
ama City and were able to make repeated obser-
vations at the platforms.

Between September 1970 and January 1974, 10
dives were made at Stage I (including 1 night
dive) and 21 dives were made at Stage n to de-
termine the composition of the fish populations
under the structures (see Tables 1, 2). During the
SITS II program in 1972, a series of dives made at
various times during four consecutive 24-h
periods (1-4 August) enabled us to determine diel
patterns of concentration of fish schools around
and under the platforms.

METHODS

During each dive an attempt was made to iden-
tify each species of fish present in the area and to
estimate its abundance. At the end of a dive a
debriefing session was held and notes were com-
pared as to species and numbers observed. Divers
often carried hand nets or spears for collecting
unusual or difficult to identify species.

Dives were usually conducted on an irregular
basis, and the length of the observation period
and the area examined varied considerably from
one dive to the next. Consequently, no numerical
values were assigned to these estimates of abun-
dance. Instead, relative terms such as few, sev-
eral, common, and abundant were used, simply to
indicate the impression received by the divers as
to the numbers of each species present. It should
be kept in mind, however, when considering these
estimates, that such relative terms may have dif-
ferent meanings when applied to different species
of fishes. For example, an absolute number such
as 100 individuals might be interpreted as few if
applied to a schooling species such as Harengula
pensacolae , but as abundant if applied to a soli-

388

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

tary reef species such as Chaetodon ocellatus.
These estimates (recorded for all dives in Tables 1
and 2) are admittedly subjective but may be use-
ful in describing the seasonal changes in the fish
populations around the platforms.

Our dives at Stage II during the SITS II pro-
gram were scheduled to occur at approximately
5-h intervals from 1 to 4 August 1972. Although
this schedule was not always followed because of
other diving commitments of the program, we
made 10 dives which included at least 2 dives
during each quarter of the 24-h day. Times of the
dives are presented as Central Standard Time
(CST) in this report. During early August 1972,
the times of sunrise and sunset at Stage II were
approximately 0500 and 1840 CST, respectively.

During the SITS II program, two censusing sta-
tions were set up under Stage II: Station 1 on the
bottom at 18.3 m; and Station 2 directly above,
about 4.6 m below the water surface. Both sta-
tions were the same size and were conveniently
delimited by the cross-members at one corner of
the platform. The stations measured 4.9 x 4.9 x
7.0 m. Counts were restricted to that portion of
the water column estimated to extend 1 m up-
ward from the base of the cross-members (corres-
ponding to the bottom at Station 1). Thus, the
water volume included within each station was
about 12 m^.

During each dive at the censusing stations the
authors attempted to identify and count all
species of fishes present within each station.
Counts were recorded on plastic slates during the
censusing, then transferred to data sheets after
surfacing. Times required to make each census
varied because of the great variation in the num-
bers of fishes present, but were usually about
10-15 min for each station. In most cases both
stations were censused on each dive, and counts
for Station 2 were made about 15-20 min after the
counts for Station 1. On dives 1, 9, and 10, only
one of the two stations was censused. Some
species were so numerous at times that only es-
timates of their abundance could be made. Such
estimates offish abundance by two divers making
counts at the same time were generally in the
same order of magnitude.

During nocturnal diving operations, an under-
water light was suspended near the surface below
a ladder (to facilitate diver return) at the corner
of the platform farthest from the censusing sta-
tions. It was not visible underwater at the census-
ing stations and did not appear to affect our

counts. Although some fishes were attracted to
this light, they were mostly juveniles and larvae
of pelagic species.

Nomenclature and arrangement of the families
in Tables 1, 2, and 3 follow Bailey et al. (1970).

RESULTS AND DISCUSSION

All species recorded at the two stages during
this study are listed in Tables 1 and 2. Table 3 is a
list of the species recorded in the two stations at
Stage n during the SITS II dives. These tables
should be examined in connection with the fol-
lowing synopsis of the results of this study. At
least 101 taxa (identified to family or species)
were recorded during this study at the two plat-
forms; 61 species were recorded at Stage I during
10 dives, and 86 taxa were recorded at Stage II
during 21 dives. The greater number recorded at
Stage II is probably primarily a result of the
greater number of observations there. In gen-
eral, the fish faunas of the two stages are quite
similar, and most of the species recorded at only
one could be expected to occur at both occasion-
ally. Of the 101 taxa recorded, about 75 were
frequently observed during the study and could
be regarded as characteristic members of the
fauna, 41 were recorded as common or abun-
dant. Of the latter group, 27 species were re-
corded at Stage I and 36 at Stage II.

Faunal Composition

The two stages represent an artificial reef
habitat in an area previously characterized by
flat sand bottoms. Thus, the fishes inhabiting the
stage environment are a mixture of faunal types,
including some species usually expected in such
flat, sandy areas, but also including many species
more characteristic of other habitats of the north-
ern Gulf of Mexico, especially fishes which are
attracted to reef environments.

A number of demersal species characteristic of
open sand habitats of the northern Gulf of Mexico
were frequently recorded at the stages, but these
were usually seen over the open sandy areas sur-
rounding the stages. Examples are Dasyatis
sp.. Raja eglanteria, Arius felis, Ogcocephalus
radiatus, Stenotomus caprinus, Hemipteronotus
novacula, Prionotus sp., Paralichthys albigutta,
Lactophrys quadricornis, and Chilomycterus
schoepfi. In addition, many species recorded at
the stages are pelagic fishes characteristic of open

389

FISHERY BULLETIN: VOL. 74, NO. 2

Table l. — Fishes recorded at Stage I off Panama City, Fla., with estimates of usual abundance and

habitat occupied.

Abundance'

Species

Spring
(Aor.)

Summer-fall
(July-Nov.)

Winter

Dec.

Jan.

Habitat'

Carcharhinidae:

Carcharhinus milberti
Dasyatidae:

Dasyatis sp.
Muraenidae:

Gymnothorax nigromarginatus
Clupeidae:

Sardinella anchovia
Ariidae:

Arius felis
Batrachoididae:

Opsanus pardus
Antennarildae:

Antennarius ocellatus
Ogcocephalidae:

Ogcocephalus radiatus
Serranidae:

Centroprlstis ocyurus

Diplectrum formosum

Epinephelus nigritus

Mycteroperca microlepis

Serranus subligarius
Grammistidae:

Rypticus maculatus
Apogonidae:

Apogon pseudomaculatus
Rachycentridae:

Rachycentron canadum
Echeneidae:

Echeneis neucratoides
Carangidae:

Caranx crysos

Caranx hippos

Caranx ruber

Decapterus punclatus

Elagatis bipmnulata

Seriola dumerili

Seriola rivoliana

Trachurus lathami
Lutjanidae:

Lutjanus campechanus

Lutjanus griseus

Rhomboplites aurorubens
Pomadasyidae:

Haemulon aurolineatum

Haemulon plumieri
Spandae:

Archosargus probatocephalus

Lagodon rhomboides
Sciaenidae:

Equetus lanceolatus

Equetus umbrosus

Equetus sp.^
Kyphosidae:

Kyphosus sectatrix
Ephippidae:

Chaetodipterus faber
Chaetodontidae:

Chaetodon ocellatus

Chaetodon sedentarlus

Holacanttius bermudensis
Pomacentrldae:

Abudefduf saxatilis

Chromis enchrysurus

Chromis scotti

Pomacentrus partilus

Pomacentrus variabilis
Labridae:

Halichoeres caudalis

Thalassoma bifasciatum
Sphyraenidae:

Sphyraena barracuda
Blenniidae:

Blennius marmoreus

Hypleurochilus geminatus

sev
few

com
sev

sev

few

sev

few
sev

com
few
sev

com
com

few

com-abun

sev

few

com
sev-com

few-com
sev-com

sev-com

few

few-sev

com
sev-com
few-com
com-abun

sev
few-com

com

few
sev-abun
com-abun

com-abun

few
sev

sev-com

few

sev

sev-com

few

few
sev-com

few-sev

few
sev-com
few-sev
sev-com

sev-com
few-sev

sev-abun

sev

few-com

few
abun

sev

few

com
sev-com

few
sev-com

few

few
abun
sev
com
few

few
few

com-abun
few-sev

sev

few
few
sev

—

com

few

com

—

few

sev

com

B-P

com

B-P

sev

—

0-U

sev

{')

—

o-u

—

few

o-u

com-abun

L-O-U

—

sev

L-U

com

sev

—

L-U

com

—

com

L-U

few

B-P

—

sev

L-U

sev

few

sev

B-P

—

com

B-P

few

sev

few

O-U

B-P

—

390

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS
Table l. — Continued.

Abundance'

Spring
(Apr)

Summer-fall
(July-Nov.)

Winter

Species

Dec.

Jan.

Habitat'

Gobiidae:

Coryphopterus punctipectophorus

few

—

loglossus calliurus

—

sev

—

Acanthuridae:

Acanthurus coeruleus

—

few

Scombridae;

Euthynnus alletteratus

—

sev-com

sev

0-U

Bothldae:

Paralichthys albigutta

—

few

—

Ballstidae:

Balistes capriscus

few

few-sev

few

L-U

Monacanthus hispidus

—

few-sev

—

Ostraclidae:

Lactophrys quadricomis

few

—

Tetraodontidae:

Canthigaster rostrata

—

few

—

Sphoeroides spengleri

—

few

—

Diodontidae;

Chilomycterus schoepfi

few

61 species

21 species

57 species

31 species

32 species

100%

34=o

93=!o

51%

52%

Number of observations

Temperature range

17°-20°C

23-29X

18°-19°C

18°C

'Abbreviations are as follows: sev-several, corn-common, abun-abundant. B-on bottom, L-lower water column, P-on pilings,
0-open water around platform, U-middle to upper water column under platform.

^Echeneis neucratoides on Epinephelus. Sphyaena, Seriola, Balistes. and Caretta.

^Equetus sp. - an undescribed species listed by Bullis and Tfiompson (1965) as "Equetus sp. nov." and by Strufisaker (1969) as
"Blackbar drum Pareques sp. (undescribed)."

waters, which are attracted to solid, reeflike
structures. Smaller baitfishes, such as Harengula
pensacolae, Sardinella anchouia, Etrumeus teres,
Opisthonema oglinum, Decapterus punctatus,
Trachurus lathami, and Scomber japonicus , were
abundant at times and formed dense schools
under the stages. Klima and Wickham (1971)
demonstrated the potential for harvesting com-
mercial quantities of these and other species by
attracting schools to artificial structures. In re-
search conducted near Stage II during 1969, they
found Decapterus and Sardinella more numerous
than Harengula, but did not record the other
species. Harengula pensacolae was usually the
most common species at Stage II during our ob-
servations made in 1972, while Decapterus and
Sardinella were more common in other years.

Larger pelagic species often recorded at the
stages were Rachycentron canadum , Caranx
bartholomaei, C. crysos, C. hippos, C. ruber,
Elagatis bipinnulata, Seriola dumerili, Euthyn-
nus alletteratus, and Sphyraena barracuda. These
species were recorded at the stages often enough
to indicate some attraction to the structures, even
though they are characteristic open-water
species. Part of the attraction for these larger
predators may be the large number of smaller
baitfishes which provide much of their food
(Wickham et al. 1973). Publications on the attrac-

tion of fishes to artificial reefs have noted that
pelagic species such as those listed here are
attracted to artificial structures in greatest num-
bers when the structures extend a considerable
distance above the bottom or even reach the sur-
face, as do these offshore platforms (Unger 1966;
Gooding and Magnuson 1967; Hunter and Mit-
chell 1967, 1968; Klima and Wickham 1971).
Springer and Woodburn (1960) noted thatS. bar-
racuda occurred near shipwrecks off the Tampa
Bay area but not on natural rocky reefs. The oc-
currence of barracuda may be associated with
the higher relief of structures such as shipwrecks
or the stages. In this respect the offshore plat-
forms are ideal for attracting large numbers of
typically open-water fishes.

Sharksuckers (remoras) were often seen as-
sociated with other fish species around the stages
(especially the larger pelagic species such as
Caranx hippos and S. barracuda) but were never
numerous. The species was probably Echeneis
neucratoides , although £J. naucrates could also be
expected in the area. Four Echeneis were also
seen attached to one of two loggerhead turtles,
Caretta caretta caretta, which were observed
asleep on the bottom below Stage I. The remoras
were attached to the turtle's plastron and ventral
margin of the carapace and were inactive except
for movements of their opercula.

391

FISHERY BULLETIN: VOL. 74, NO. 2

Table 2. — Fishes recorded at Stage II off Panama City, Fla., with estimates of usual abundance and

habitat occupied.

Abundance'

Species

Spring
(Apr. -May)

Summer-fall
(June-Nov.)

Winter

Dec.

Feb.

Carcharhinidae
Sphyrnidae:

Sphyrna sp.
Dasyatidae:

Dasyatis sp.
Rajidae:

Raja eglanleria
Muraenidae;

Gymnothorax nigromarginatus
Congridae
Ophichthidae:

Mystnophis Interlinctus
Clupeidae:

Etrumeus teres

Harengula pensacolae

Opisthonema oglinum

Sardinella anchovia
Engraulidae
Ariidae:

Arius felis
Batrachoididae:

Opsanus pardus
Antennariidae;

Antennahus ocellatus
Ogcocephalidae:

Ogcocephalus radiatus
Syngnathidae:

Syngnathus sp.
Serranidae:

Centropristis melana

Centrophstis ocyurus

Centropristis philadelphica

Diplectrum formosum

Epinephelus morio

Epinephelus sp.^

Mycteroperca microlepis

Serranus subligarius
Grammistidae:

Rypticus maculatus
Priacanthidae:

Priacanthus arenatus
Apogonidae:

Apogon pseudomaculatus
Pomatomidae:

Pomatomus saltatrix
Rachycentridae;

Rachycentron canadum
Echeneidae;

Echeneis neucratoides
Carangidae:

Caranx bartholomaei

Caranx crysos

Caranx hippos

Caranx ruber

Decapterus punctatus

Selar crumenophthalmus

Seriola dumerili

Seriola zonata

Trachurus lathami
Lutjanidae:

Lutjanus campechanus

Lutjanus griseus

Lutjanus synagns

Rhomboplites aurorubens
Lobotidae:

Lobotes surinamensis
Pomadasyidae;

Haemulon aurolineatum

Haemulon plumieri

Ortliopnstis chrysoptera
Spandae;

Archosargus probatocephalus

Calamus-Pagrus

Diplodus holbrooki

Lagodon rhomboides

Stenotomus caprinus

few
few

few
few
few
few
few

few

few-sev

few

few
few

few

—

sev-com

—

sev-abun

sev-com

sev

com

—

com-abun

sev-abun

—

com-abun

—

few-abun

—

few-sev

few

few-com

few

few-com

few

—

few

—

few

sev

few-sev

com

com-abun

com

—

few

sev-com

few-corn

sev

few

—

few

—

few

few-sev

sev

sev-com

—

few-corn

few-com

—

few

—

few

few-corn

few-sev

few

—

few-sev

few

—

sev-abun

few

—

com

sev

—

few-com

—

com-abun

abun

com-abun

—

sev-com

—

few

few-sev

sev

few

—

com

few-abun

few-sev

sev

—

sev

few-sev

—

few

—

sev

sev-com

few-sev

—

few

—

com

com-abun

few-com

few-sev

few

com

abun

few-abun

few

sev

few

—

few

—

few-sev

few

com

sev-com

—

com

—

sev
com

sev

few

Habitat'

sev

B
B

L-U

U
L-U

B
B-L

L
B-P

B-P

0-U
0-U

(')

L-U

0-U

L-U

L-O-U

L
L-U

L
L
L

L-U

L-U
B-0

392

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

Table 2. — Continued.

Abundance'

Spring
(Apr.)

Summer-fall
(July-Nov.)

Winter

Species

Dec,

Jan

Habitat'

Sclaenidae:
Equetus lanceolalus
Equetus umbrosus
Leiostomus xanthurus
Sciaenops ocellata

Mullldae

Kyphosldae:
Kyphosus sectatnx

Ephlppidae:
Chaetodipterus faber

Chaetodontldae:
Chaetodon ocellalus
Holacanlhus bermudensis

Pomacentridae:
Pomacentrus variabilis

Labridae:
Halichoeres bivittatus
l-lalichoeres caudalis
Hemipteronotus novacula
Lachnolaimus maximus

Sphyraenldae:
Sphyraena barracuda
Sptiyraena borealis

Polynemldae:
Polydactylus octonemus

Blennlldae:
Blennius marmoreus
Hypleurochilus geminatus

Acanthuridae:
Acanthurus chirurgus

Scombrldae:
Euthynnus alletteratus
Scomber japonicus
Scomberomorus cavalla

Stromateidae:
Peprilus burti

Scorpaenidae:
Scorpaena brasiliensis

Triglldae:
Prionotus sp.

Bothldae;
Paralichthys albigutta
Syacium papillosum

Ballstldae:
Balistes capriscus
Cantherhmes pullus
Monacanthus hispidus

Ostraclidae:
Lactophrys quadricornis

DIodontldae:
Chilomycterus schoepfi

86 taxa
100%

Number of observations

Temperature range

few-sev

few-com

sev

sev-com

few-sev

—

com

sev

few

—

few

—

few-sev

—

sev

few-com

sev

—

few

sev-com

few-com

sev-com

few-sev

few

few-com

few

sev

sev-com

few-sev

—

few

—

few

—

few-sev

—

sev

—

sev

few

few-sev

few

sev-com

—

few

—

sev-com

few-com

com

few

—

sev

—

few-sev

sev

—

few

—

few

—

sev

few-sev

sev

—

few

—

few-sev

few-com

few-sev

—

few

—

few

sev

com

few

A 1 species
48%

17°-20°C

few-sev

81 taxa
94%

20°-30°C

few

57 taxa
66%

15°-19°C

sev

few

13 species
15%

13X

B
B
B
B
O

L-U

B-P

B
B
B

L-O-U
U

B-P

o
u

B
B

L-U

L-P

'Abbreviations are as follows: sev - several, com - common, abun - abundant, B - on bottom, L - lower water column, P - on pilings,
O - open water around platform, U - middle to upper water column under platform.

'Epinephelus sp. - A juvenile apparently either E flavolimbatus or E. niveatus based upon color pattern (brownish with small
white spots on lateral surface and a dark saddle on caudal peduncle. Smith 1971).

^Echeneis neucratoides on Caranx and Sphyraena.

A few species recorded at the stages are typical
inshore fishes which are characteristic of coastal
or estuarine areas. Examples are Orthopristis
chrysoptera, Lagodon rhomboides, and Leios-
tomus xanthurus. These first two species were
important members of the fauna at Stage II,
while L. xanthurus was common at times but
usually remained over the surrounding open sand
bottom.

Most of the species recorded at the stages are
species characteristic of rocky bottom areas
offshore in the Gulf of Mexico. The platforms with
their supporting pilings, as well as litter and
shell hash which has accumulated in the area
immediately surrounding the stages, serve as ar-
tificial reef habitat for such species. Some of the
important reef species are Gymnothorax nig-
romarginatus, Mystriophis intertinctus, Opsanus

393

FISHERY BULLETIN: VOL. 74. NO. 2

TABLE 3.— Counts of fishes at two stations below Stage n off Panama City, Fla., 1-4 August 1972.

(Bold numerals are estimates.)

Time of census (CST)

0129-

0525-

0715-

1042-

1152-

1325-

1519-

1721-

1833-

2308-

Family and species

0216

0604

0755

1133

1228

1405

1553

1800

1913

2342

Station 1 (bottom)

Ophichthidae:

Mystriophis intertinctus

Ciupeidae:

Harengula pensacolae

100

200

Batrachoididae:

Opsanus pardus

Antennariidae:

Antennahus ocellatus

Ogcocephalidae:

Ogcocephalus radiatus

Serranidae:

Centropristis ocyurus

Epinephelus sp.^

Mycteroperca microlepis

Serranus subligarius

Grammistidae:

Rypticus maculatus

Apogonidae:

Apogon pseudomaculatus

Carangidae:

Decapterus punctatus

Seriola dumerili

Lutjanidae:

Rhomboplites aurorubens

Pomadasyidae:

Haemulon aurolineatum

200

300

200

100

Haemulon plumieri

Orthopnstis chrysoptera

100

200

Sparidae:

Lagodon rhomboides

Sciaenidae:

Equetus umbrosus

Kyphosidae:

Kyphosus sectatrix

Ephippidae:

Chaetodipterus faber

Chaetodontidae:

Holacanthus bermudensis

Pomacentridae:

Pomacentrus variabilis

Labridae:

Halichoeres caudalis

Acanthuridae:

Acanthurus chirurgus

Scorpaenidae:

Scorpaena brasiliensis

Balistidae:

Balistes capriscus

Monacanthus hispidus

Diodontidae;

Chilomycterus schoepfi

(')

Station 2 (subsurtace)

Ciupeidae:

Harengula pensacolae

500

300

500

1,000

200

Serranidae:

Serranus subligarius

(')

Grammistidae:

Rypticus maculatus

Carangidae:

Caranx crysos

(')

Caranx ruber

(')

Decapterus punctatus

(')

Sparidae;

Diplodus holbrooki

(')

Lagodon rhomboides

(')

Kyphosidae:

Kyptiosus sectatrix

(')

Blenniidae:

Hypleurochilus geminatus

(')

Acanthuridae:

Acanthurus chirurgus

(')

'No census made.

^Epinephelus sp. -A juvenile apparently either £. flavolimbatus or£. niveatus based upon color pattern (brownish with small white spots
on lateral surface and a dark saddle on caudal peduncle (Smith 1971).

394

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

pardus, Antennarius ocellatus, Centropristis
ocyurus, Diplectrum formosum, Mycteroperca
microlepis , Serranus subligarius, Rypticus mac-
ulatus, Apogon pseudomaculatus , Lutjanus cam-
pechanus, L. griseus, Rhomboplites aurorubens,
Haemulon aurolineatum, H. plumieri, Diplodus
holbrooki, Equetus lanceolatus, E. umbrosus,
Chaetodipterus faber, Chaetodon ocellatus, C.
sedentarius, Holacanthus bermudensis, Chromis
enchrysurus, C. scotti, Pomacentrus variabilis,
Halichoeres bivittatus, H. caudalis, Blennius
marmoreus, Hypleurochilus geminatus, loglos-
sus calliurus, Acanthurus chirurgus, and Balistes
capriscus. A few natural rock outcrops which
support reef faunas occur in the area, especially
offshore from Stage I, but these are characteris-
tically low in relief and are quite distinct in some
ways from the habitats at the stages. They do
support populations of the reef species listed
above (and usually larger numbers than at the
stages), but usually do not attract large masses
of pelagic schooling and predatory species.

A few reef species observed at the stage
habitats (such as Abudefduf saxatilis, Pomacen-
trus partitas, Thalassoma bifasciatum, and Acan-
thurus coeruleus) do not normally occur on the
natural rocky reefs off the northwest Florida
coast, but are tropical coral reef species which
may be carried into the northern Gulf of Mexico
by currents (see Hastings 1972). Such species are
not permanent residents of the northern gulf, but
are apparently usually killed by low winter
temperatures, except for possibly during mild
winters.

Comparison of the Two Stages

Although the fish faunas of the two stages were
quite similar (Tables 1, 2), there were a few nota-
ble differences between the species lists for the
two stages which may be significant. The most
numerous species at Stage II during the summer
and fall were the clupeids, Harengula pensacolae
and Sardinella anchouia, and rather irregularly,
Etrumeus teres and Opisthonema oglinum. These
fishes formed dense schools (Figures 2, 3) below
the platform during daylight hours, usually also
with large numbers of carangids such as Decap-
terus punctatus, Selar crumenophthalmus, and
Trachurus lathami, and the mackerel. Scomber
japonicus. Such schools of baitfishes were consid-
erably less abundant at Stage I except for during
the fall and early winter (especially November

and December) when large numbers o^Sardinella
anchovia and D. punctatus were present. Most of
these had disappeared by January, however.

As might be expected, typical estuarine species,
such as Orthopristis chrysoptera, Lagodon rhom-
boides, and Leiostomus xanthurus, were rare or
absent at Stage I, even though they were quite
numerous at Stage II. In contrast, E/ag^a^is bipin-
nulata, a species typical of open, pelagic waters
(Hiatt and Strasburg 1960), was recorded at
Stage I, but not at Stage II, although Klima and
Wickham (1971) found this species to be the most
common jack congregating about artificial struc-
tures near Stage II in 1969. Other pelagic species
such as Seriola dumerili and Sphyraena bar-
racuda were also more numerous at Stage I. Simi-
larly, some benthic species, which are charac-
teristic of the deeper water, natural reefs in the
northern Gulf of Mexico and may be rare in in-
shore waters as shallow as 18 m, were occasion-
ally recorded at Stage I, but not at Stage II.
Examples are Chaetodon sedentarius, Chromis
enchrysurus, C. scotti, Coryphopterus punctipec-
tophorus, and loglossus calliurus.

The tropical coral reef species, such as Abudef-
duf saxatilis, Pomacentrus partitus, Thalassoma
bifasciatum, Acanthurus coeruleus, and Canth-
igaster rostrata, were recorded only at Stage I.
These tend to be shallow- water species which ap-
parently were able to survive by settling on the
pilings and cross-members near the surface at
Stage I. Such species are occasionally recorded in
inshore artificial reef habitats in the northeast-
ern gulf (Caldwell and Briggs 1957; Caldwell
1959, 1963; Haburay et al. 1969, 1974; Hastings
1972) and should be expected to occur occasion-
ally at both stages.

Winter-Summer Contrast

Seasonal changes in the faunal composition at
the stages were striking in some cases. Water
temperatures recorded during this study ranged
from 17° to 29°C at Stage I and from 13° to 30°C at
Stage II. Lowest temperatures were recorded in
January at Stage I and in February at Stage II.
Highest temperatures were recorded during Au-
gust and September. Changes in the fish fauna
were apparently correlated with temperature,
since the largest percentages of species recorded
(93% at Stage I; 95% at Stage II) were present
during the summer and fall, while the lowest
numbers were recorded during either the winter

395

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 2. — Sardinellaanchovia,Decapteruspunctatus, and Scomber japonicus, in a mixed school, under Stage II ofTPanama City, Fla.

or spring. Estimates of abundance during the
spring, summer-fall, and winter observations
(Tables 1,2) indicate that most species disappear
from the area of the stages during the winter
months, then gradually reappear during the
spring and summer. They apparently either move
offshore to deeper water, or else they migrate
southward along the Florida coast (see Hastings
1972). This decrease in number of species (as well
as number of individuals) occurred at both stages,
but was most profound at Stage II, where temper-
ature extremes were greater. About 509^ of the
number of species recorded at Stage I were pres-
ent in December and January, but at Stage II,
67% were present in December and only 15% in
February. These seasonal changes were most
striking in the schooling clupeids and carangids
(such as H. pensacolae, Sardinella anchovia, and
D. punctatus) which were extremely numerous
during the summer and fall, but usually rare or
absent in January or February (although Decap-
terus was common at Stage II during February).

Habitat Occupation and
Activity Patterns

The usual habitat occupied by each species in
the vicinity of the platforms is indicated in Tables
1 and 2. Station counts for some species at Stage
II, indicating diel changes in activity and occur-
rence at the stage, are shown in Table 3.

The pelagic species which congregate about the
stages normally occupied the upper water col-
umn, either surrounding or below the platform.
The clupeids, H. pensacolae and S. anchovia,
formed dense schools below the platform, usually
near the surface but with Sardinella usually
somewhat deeper. The carangids, D. punctatus
and Trachurus lathami, were also quite numer-
ous, Decapturus normally in mid-water or near
the surface and Trachurus very near the bottom.
At times, these and other schooling species of
comparable size, such as Opisthonema oglinum
and Scomber japonicus, formed mixed schools
under the platform (Figure 2). These species

396

HASTINGS ET AL.; FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

Figure 3. — Large school of Harengula pensacolae surrounding a piling of Stage 11 off Panama City, Fla.

gathered in compact schools below the stage dur-
ing the day apparently as a defense against pre-
dation (Hobson 1965). Station counts at Stage II
for H. pensacolae and D. punctatus indicate that
they left the protection of the platform and moved
into the open areas surrounding the stage at
night. Several species of clupeids and schooling
carangids, including H. pensacolae, Sardinella
anchovia, D. punctatus, and Selar crumen-
ophthalmus, have been described as nocturnal
plankton feeders (Hobson 1965; Starck and
Davis 1966), although some diurnal feeding
activity by Decapterus and Sardinella was ob-
served by us and others (Klima and Wickham
1971). During daylight hours at Stage II, from
about 0500 to about 1800 CST, extensive schools
of if. pensacolae were present around and under
the platform, and, at times, were so dense that
they darkened the area below (Figure 3). Rela-
tively large numbers were present at the census
stations during most daylight dives, but none was
observed during any of the night censuses. Simi-
lar records were obtained for D. punctatus, al-

though the numbers present were considerably
less than for H. pensacolae. In addition, D.
punctatus may have left the vicinity of the plat-
form earlier in the evening (about 1500 CST).

The other pelagic species are, in most cases,
large predators and are continually on the move
in the upper water column surrounding the plat-
forms, occasionally darting into the schools of
smaller fishes to feed. Some, such as Seriola
dumerili, were often seen near the bottom as well.
Most of these pelagic predators probably feed to
some extent at night as well as during the day,
and may follow the bait species, as the bait
species disperse at night. However, studies indi-
cate that many such piscivorous fishes are
primarily crepuscular, with peaks of feeding ac-
tivity at dawn and dusk (Hobson 1965, 1968,
1972, 1974; Starck and Davis 1966).

Only Caranx crysos was consistently present in
the station counts (but only in Station 2 near the
surface). These counts show a pattern similar to
that of//, pensacolae, with fairly large numbers
present during daylight hours and none present

397

FISHERY BULLETIN: VOL. 74, NO. 2

at night. Possibly this jack followed the Haren-
gula as they dispersed, to continue feeding
through the night.

A large number of benthic reef species occupy
the bottom below the platform and also the area
immediately surrounding the stage, where much
litter has accumulated, apparently discarded by
workmen on the platform above. Other benthic
species were observed on the pilings and cross-
members of the platform structure, where en-
crusting invertebrates and algae provided food
and hiding places for smaller species. In addition,
habitat for benthic species may be provided by
accumulations of shell hash at the bases of the
pilings, probably broken loose from the pilings by
storms or by the grazing of fishes or predation by
other organisms. Some of the more important
benthic species at the stages are Gymnothorax
nigromarginatus, Opsanus pardus, Antennarius
ocellatus, Ogcocephalus radiatus, Centropristis
ocyurus, Diplectrum formosum, Serranus sub-
ligarius, Rypticus maculatus, Apogon pseudo-
maculatus, Equetus lanceolatus , E. umhrosus,
Chaetodon ocellatus, Pomacentrus variabilis,
Halichoeres caudalis, Blennius marmoreus,
and Hypleurochilus geminatus. A few of these,
such as S. subligarius, R. maculatus, and R
variabilis, seemed to be equally at home on the
pilings at all levels of the water column, while
others were found only near the bottom (G. nig-
romarginatus, Opsanus pardus, Antennarius
ocellatus, Ogcocephalus radiatus, Centropristis
ocyurus, D. formosum, Apogon pseudomaculatus,
E. lanceolatus, E. umbrosus, and Halichoeres
caudalis) or only on the pilings {Hypleurochilus
geminatus).

An interesting contrast was noticed among
members of the families Pomacentridae and Lab-
ridae at Stage I. Those species which are charac-
teristic and permanent members of the northern
gulf reef fauna {Chromis enchrysurus, C. scotti,P.
variabilis, and Halichoeres caudalis) were most
numerous on the bottom in association with plat-
form supports and other objects. In contrast,
species which are not permanent residents of
reefs in this area, but are apparently tropical
species carried north by currents {Abudefduf
saxatilis, P. partitus, and Thalassoma bifas-
ciatum) were never observed near the bottom, but
were always associated with the pilings and
cross-members within about 10 m of the surface.
These are shallow-water species which appar-

ently do not occur at the greater depths at Stage
I (32 m).

At least two species, G. nigromarginatus and
Mystriophis intertinctus, were usually seen par-
tially buried in the substrate, often with only
their heads protruding.

Several other species occurring on the bottom
were most numerous over the open sandy areas
surrounding the stages. Stenotomus caprinus,
Leiostomus xanthurus, Paralichthys albigutta,
and Lactophrys quadricornis are examples.

This benthic group includes both diurnal and
nocturnal species. Species which are active and
apparently feed at night are R. maculatus, Apo-
gon pseudomaculatus , and E. umbrosus. Benthic
species which are diurnal and inactive at night
are Centropristis ocyurus, D. formosum, Serranus
subligarius, Chaetodon ocellatus, Pomacentrus
variabilis , and H. caudalis. The other species
were not observed enough to determine activity
patterns.

Generally counts of the nocturnal species were
higher during the nocturnal observations. Ryp-
ticus maculatus was more numerous in Station 1
on the bottom under cross-members or other shel-
tering objects, but was counted in Station 2 near
the surface three times, during each of the noc-
turnal counts between about 1730 and 0215 CST
Hiding places on the pilings are rather limited
and can, in most cases, accommodate only small
individuals, so apparently these soapfish were
moving up the pilings at night to feed. Other ref-
erences also report noctural feeding in the
grammistids (Hobson 1965; Starck and Davis
1966).

Apogon pseudomaculatus , when observed at
night, was active, swimming about in open areas
near the bottom, while those observed during the
day were always inactive, hiding among shells or
other debris or under the stage cross-members.
On one occasion a group of about 15 juvenile Apo-
gon was seen associated with a diadematid sea
urchin below a cross-member at Stage I. These
small cardinal fish remained motionless among
the long spines of the urchin. Cardinal fishes in
general are nocturnal predators (Hobson 1965;
Starck and Davis 1966; Livingston 1971).

Species of Equetus (or the related Pareques)
have been reported to remain in small groups in
sheltered areas by day, and then feed individually
in the immediate vicinity at night (Hobson 1965;
Starck and Davis 1966). Similar observations

398

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

were made during this study for E. umbrosus,
which was present during almost every observa-
tion at Station 1.

The smaller demersal sea basses (family Ser-
ranidae) observed at the stages were relatively
inactive fishes which did not exhibit obvious
day-night changes in behavior. However, counts
of Centropristis ocyurus and S. subligarius de-
creased at night, possibly indicating that some
had taken shelter under objects or within shells
or crevices. Literature records indicate that these
and related sea basses are diurnal (Starck and
Davis 1966; Bortone 1971).

Chaetodon ocellatus was usually seen swim-
ming about near the bottom during the day and
frequently in pairs. One individual observed at
night resting on the bottom next to a piling ex-
hibited the typical nocturnal color pattern de-
scribed by Starck and Davis (1966).

Counts of Pomacentrus variabilis and H.
caudalis at Station 1 were considerably higher
during the daylight observations than at night.
Daylight counts for P. variabilis (10-14) were less
variable than those for H. caudalis (6-22).
Pomacentrus variabilis is territorial and probably
remains at the same general location throughout
the day while H. caudalis is less sedentary and
tends to move about more. Starck and Davis
(1966) stated that P. variabilis and other
pomacentrids are diurnal feeders which seek
shelter at night in sponges, rocks, coral, or other
close cover. Most of those at the stage may have
taken shelter in and among the many empty mol-
lusk shells which cover much of the bottom at the
base of the stage. Halichoeres caudalis has not
been studied previously, but several species of
labrids, including H. bivittatus which was also
present at times at Stage II, have been reported to
bury themselves in sand at night (Breder 1951;
Hobson 1965; Starck and Davis 1966), and this
may also be the case with H. caudalis.

The numerous species of free-swimming fishes
occupying the various levels of the water column
under the platform apparently include several
distinct groups based upon activity patterns and
feeding habits. Mycteroperca microlepis is a large
predator which appeared to be continually mov-
ing about under or around the stage, usually near
the bottom, but a few inactive individuals were
observed at night on the bottom resting against
the pilings. Such species are normally described
as being opportunistic feeders with peaks of feed-

ing activity during twilight periods when the
changeover of activity patterns in prey species
makes them more vulnerable (Starck and Davis
1966; Collette and Talbot 1972; Hobson 1972).

Lutjanus griseus (Starck 1971), Haemulon au-
rolineatum, and Orthopristis chrysoptera are ap-
parently nocturnal feeders, which utilize the
stage only as a shelter during daylight hours, and
move out into surrounding areas at night to feed.
Lutjanus griseus was normally seen schooling
during the day in the lower-to-middle water col-
umn under the platform. Haemulon aurolineatum
and O. chrysoptera were two of the most numer-
ous fishes in Station 1 at Stage II (Figure 4), al-
though both were rare or absent during the
nighttime observations. There may be a differ-
ence in the time of major movement for these two
species. Haemulon aurolineatum apparently
began to disperse and move out of the area at or
shortly after sunset, and returned shortly after
sunrise. Orthopristis chrysoptera possibly leaves
the area under the stage earlier in the evening
(just before sunset) and also may return earlier in
the morning. Apparently these grunts feed at
night in the open areas surrounding the stage
and school under the stage as a defense against
diurnal predators (Hobson 1965; Starck and
Davis 1966).

Other species (such as H. plumieri, Diplodus
holbrooki, Lagodon rhomboides, Kyphosus sectat-
rix, Chaetodipterus faber, Holacanthus ber-
mudensis, Acanthurus chirurgus, Balistes capris-
cus, and Monacanthus hispidus) seemed to feed
mostly on benthic organisms attached to the pil-
ings or other objects, and may move up and down
in the water column, grazing upon this material.
However, some of these were more numerous
near the surface (such as D. holbrooki, L. rhom-
boides, and K. sectatrix) while others normally
remained near the bottom (such as Haemulon
plumieri, C. faber, Holacanthus bermudensis , and
B. capriscus). Most of these species are appar-
ently diurnal and become inactive at night. A
few L. rhomboides, H. bermudensis, and B. cap-
riscus were observed near the bottom at night,
either resting on the bottom or in protected places
below cross-members or between pilings and ad-
jacent objects. Kyphosus sectatrix was inactive at
night, but remained in the upper water column.
In contrast, Haemulon plumieri is nocturnal but
seemed to remain in the same general area near
the bottom throughout the day and night. Such

399

FISHERY BULLETIN: VOL. 74, NO. 2

Figure 4. — Haemulon aurolineatum and Orthopristis chrysoptera near the bottom at Stage II off Panama City, Fla.

behavior was also noted by Starck and Davis
(1966).

Starck and Davis (1966) emphasized the impor-
tance of nocturnal foraging migrations and
plankton feeding to the coral reef trophic struc-
ture. Similar feeding patterns may contribute to
the economy of artificial reef structures such as
these offshore platforms, where abundant species
of the families Clupeidae, Carangidae, Lut-
janidae, and Pomadasyidae feed at night in adja-
cent areas, but return to the reef by day, and thus
contribute to the biomass of the community.

In conclusion, the platform pilings and cross-
members, with their encrusting organisms and
associated motile invertebrate fauna, provide
food and shelter for numerous fish species. In ad-
dition, several diurnally schooling species are
abundant beneath the platforms during the day,
where they are afforded some protection from
predation, but disperse into surrounding open
areas at night to feed. Large numbers of piscivor-
ous species also are attracted to the platform

habitat to feed on the numerous smaller fishes
associated with the structure. As the water tem-
perature drops, many species migrate away from
the platforms during the colder months. Repopu-
lation occurs in the spring and summer.

ACKNOWLEDGMENTS

We thank several persons who aided us during
the course of this study. Many individuals as-
sociated with the SITS program were helpful, but
only a few can be mentioned here. Thomas S.
Hopkins, Chief Scientist of the SITS II program,
and Wilbur Eaton, SITS II diving supervisor,
were especially helpftil. Christopher L. Combs,
Sylvia A. Earle, Susan Karl, and Anthony J.
Lewellyn participated in some of the SITS II
dives. Glendle W. Noble, Naval Coastal Systems
Laboratory, kindly made arrangements for us to
accompany him on numerous trips to the stages
during 1971. Thanks are also extended to Eugene
L. Nakamura, of the Gulf Coastal Fisheries Cen-

400

HASTINGS ET AL.: FISH FAUNA ASSOCIATED WITH OFFSHORE PLATFORMS

ter Panama City Laboratory, National Marine
Fisheries Service, NOAA, for allowing us to use
vessels from his laboratory in making some of our
dives at the stages. He and Ralph W. Yerger of
Florida State University read and criticized the
original manuscript. Hastings was partially sup-
ported during this study by a grant from the Sport
Fishing Institute to Florida State University.

LITERATURE CITED

Bailey, R. M., J. E. Fitch, E. S. Herald, E. a. Lachner, C.

C. LINDSEY, C. R. ROBINS, AND W. B. SCOTT.

1970. A list of common and scientific names of fishes from
the United States and Canada. Am. Fish. Soc. Spec.
Publ. 6, 149 p.

BORTONE, S. A.

197 1. Studies on the biology of the sand perch, Diplectrum
formosum (Perciformes: Serranidae). Fla. Dep. Nat. Re-
sour., Div. Mar. Sci., Tech. Ser. 65:1-27.

BREDER, C. M., JR.

1951. Nocturnal and feeding behavior of the labrid fish
Xyrichthys psittacus. Copeia 1951:162-163.
BULLIS, H. R., JR., AND J. R. THOMPSON.

1965. Collections by the exploratory fishing vessels Ore-
gon, Silver Bay, Combat, and Pelican made during 1956
to 1960 in the southwestern North Atlantic. U.S. Fish
Wildl. Serv., Spec. Sci. Rep. Fish. 510, 130 p.
CALDWELL, D. K.

1959. Observations on tropical marine fishes from the
northeastern Gulf of Mexico. Q. J. Fla. Acad. Sci. 22:69-
74.

1963. Tropical marine fishes in the Gulf of Mexico. Q. J.
Fla. Acad. Sci. 26:188-191.

CALDWELL, D. K., AND J. C. BRIGGS.

1957. Range extensions of western North Atlantic fishes

with notes on some soles of the genus Gymnachirus.

Bull. Fla. State Mus., Biol. Ser 2:1-11.
CARLISLE, J., JR., C. TURNER, AND E. EBERT.

1964. Artificial habitat in the marine environment. Calif
Dep. Fish Game, Fish Bull. 124, 93 p.

COLLETTE, B. B., AND F. H. TALBOT.

1972. Activity patterns of coral reef fishes with emphasis
on nocturnal-diurnal changeover. In Results of the Tek-
tite program: Ecology of coral reef fishes, p. 98-124. Nat.
Hist. Mus., Los Ang. Cty., Sci. Bull. 14.

CULPEPPER. T J., AND W. E. PEQUEGNAT

1969. A taxonomic and ecological study of selected ben-
thonic gammarid crustaceans from the northeastern
Gulf of Mexico. Tex. A&M Univ., Dep. Oceanogr. Proj.
286-6, Ref. 69-3T,102 p.
GOODING, R. M., AND J. J. MAGNUSON.

1967. Ecological significance of a drifting object to pelagic
fishes. Pac. Sci. 21:486-497.
HABURAY, K., C. F. CROOKE, AND R. HASTINGS.

1969. Tropical marine fishes from Pensacola, Florida. Q.
J. Fla. Acad. Sci. 31:213-219.
HABURAY, K., R. W. HASTINGS, D. DeVRIES, AND J. MASSEY.
1974. Tropical marine fishes from Pensacola, Florida. Fla.
Sci. 37:105-109.
HASTINGS, R. W.

1972. The origin and seasonality of the fish fauna on a new

jetty in the northeastern Gulf of Mexico. Ph.D. Thesis,
Florida State Univ., Tallahassee, 555 p.

HIATT, R. W., AND D. W. STRASBURG.

1960. Ecological relationships of the fish fauna on coral
reefs of the Marshall Islands. Ecol. Monogr. 30:65-127.

HOBSON, E. S.

1965. Diurnal-nocturnal activity of some inshore fishes in
the Gulf of California. Copeia 1965:291-302.

1968. Predatory behavior of some shore fishes in the Gulf
of CaUfomia. U.S. Fish Wildl. Serv., Res. Rep. 73, 92 p.

1972. Activity of Hawaiian reef fishes during the evening
and morning transitions between daylight and dark-
ness. Fish. BulL, U.S. 70:715-740.

1974. Feeding relationships of teleostean fishes on coral
reefs in Kona, Hawaii. Fish. Bull, U.S. 72:915-1031.

Hunter, J. R., and C. T Mitchell.

1967. Association of fishes with flotsam in the offshore
waters of Central America. U.S. Fish Wildl. Serv., Fish.
Bull. 66:13-29.

1968. Field experiments on the attraction of pelagic fish to
floating objects. J. Cons. 31:427-434.

KLIMA, E. F., and D. a. WICKHAM.

1971. Attraction of coastal pelagic fishes with artificial
structures. Trans. Am. Fish. Soc. 100:86-99.

Livingston, R. J.

1971. Circadian rhythms in the respiration of eight species
of cardinal fishes (Pisces: Apogonidae): Comparative
analysis and adaptive significance. Mar. Biol. (Berl.)
9:253-266.
PEQUEGNAT, W. E., R. S. GAILLE, AND L. H. PEQUEGNAT.

1967. Biofouling studies off Panama City, Florida. IL The
two mile offshore station. Tex. A&M Univ., Dep.
Oceanogr. Proj. 286-6, Ref 67-18T, 47 p.

PEQUEGNAT, W. E., AND L. H. PEQUEGNAT.

1968. Ecological aspects of marine fouling in the north-
eastern Gulf of Mexico. Tex. A&M Univ., Dep. Oceanogr.
Proj. 286-6, Ref. 68-22T, 88 p.

Smith, C. L.

1971. A revision of the American groupers: Epinephelus
and allied genera. Bull. Am. Mus. Nat. Hist. 146:67-241.
Sonnier, F., J. Teerling, and H. D. HOESE.

1976. Observations on the offshore reef and platform fish
fauna of Louisiana. Copeia 1976:105-111.

Springer, v. G., and K. D. Woodburn.

I960. An ecological study of the fishes of the Tampa Bay
area. Fla. State Board Conserv., Mar. Lab., Prof Pap.
Ser. 1, 104 p.
STARCK, W. a. IL

1971. Biology of the gray snapper, Lutjanus griseus (Lin-
naeus), in the Florida Keys. Stud. Trop. Oceanogr., Inst.
Mar. Sci., Univ. Miami 10:11-150.
STARCK, W. A. II, AND W. B. DAVIS.

1966. Night habits of fishes of Alligator Reef, Florida.
Ichthyologica 38(4):3 13-356.

Struhsaker, p.

1969. Demersal fish resources: Composition, distribution,
and commercial potential of the Continental Shelf stocks
off Southeastern United States. U.S. Fish Wildl. Serv.,
Fish. Ind. Res. 4:261-300.

TREYBIG, D. L.

1971. How offshore platforms help fishing. Ocean Ind.
6(4):64-65.

401

FISHERY BULLETIN: VOL. 74, NO. 2

UNGER I. of Mexico. Tex. A&M Univ., Dep. Oceanogr. Meteorol.

1966. Artificial reefs. Am. Littoral See, Spec. Publ. 4, Proj. 286-D, Ref. 64-19T, 77 p.

74 p WICKHAM, D. A., J. W. WATSON, JR., AND L. H. OGREN.

VICK N. G. 1973. The efficacy of midwater artificial structures for

1964. The marine ichthyofauna of St. Andrew Bay, attracting pelagic sport fish. Trans. Am. Fish. Soc.

Florida, and nearshore habitats of the northeastern Gulf 102:563-572.

402

EFFECTS OF INCREASED WATER TEMPERATURE

ON DAPHNIA PULEX

Donovan R. Craddock^

ABSTRACT

Techniques were developed to study the effects of increased water temperature on certain zoo-
plankters; specific studies were conducted on Daphniapulex, an abundant and important zooplankter
of the lower Colimibia River Study methods simulated prolonged exposure to constant high tempera-
tures in thermal discharges and short exposures to increased temperatures in condensers of cooling
systems. Effects were evaluated on the basis of survival and reproduction for periods ranging from 34
to 90 days. The time to death of 50% of the D. pulex, both mature and young, was less than 24 h at
temperatures above 27°C. Temperatures of 27°C and below required an exposure of at least 192 h to
cause 50% mortality. The young females were more tolerant of temperature increases than older
females. The greatest reproduction by older females was at the control temperature (15°C),
whereas reproduction by the young females was low at lower temperatures. No reproduction occurred
above 27°C.

Two groups of D. pulex (one from the Seattle, Wash., area and the other from the Columbia River)
studied at increased temperatures for prolonged periods revealed similar patterns of survival and
reproduction, but the Columbia River group appeared less tolerant of increased temperatures. A
short exposure (15 min) to increased temperatures up to 30°C had little effect on survival and
reproduction.

It was concluded that temperatures should not exceed 26° or 27°C for prolonged periods or 30°C for
more than 15 min to protect D. pulex populations in the river

The lower reaches of the Columbia River (below
Portland, Oreg.) support extensive and valuable
commercial and sport fisheries as well as other
types of recreational activities. This section of the
river is also becoming increasingly industrial-
ized. Associated with the industrialization is
1) the extensive use of river water for cooling pur-
poses and 2) the discharge of heated cooling water
back into the river. This increasing use of the
river for industrial cooling has created concern
that the aquatic biota is endangered by thermal
pollution. North and Adams (1969) have de-
scribed thermal conditions at outfalls and in con-
denser cooling systems of some California plants.
They pointed out that increases of +10°F (5.6°C)
above normal are considered significant biologi-
cally at all seasons of the year. Coutant (1970)
presented a diagram of the hypothetical time-
course of acute thermal shock to any organism
entrained in condenser cooling water systems
that indicates they could be exposed to the
maximum increase (10.8°C) for at least 9 min in
diffuser systems and to substantial increases

from 12 to 20 min in the discharge canal system.
He also noted the average temperature rise re-
ported is about 10.8°C but may be as great as
16°C.

I studied the effect of increased water tempera-
tures on one of the abundant cladocerans of the
area, Daphnia pulex. It has been found to be
important in the diet of valuable stocks of
juvenile chinook salmon, Oncorhynchus tshawyt-
scha, in certain seasons of the year (Craddock et
al.2). Cladocerans may be thermally affected by a
thermal nuclear power plant where, along with
other zooplankton, they may be entrained with
intake cooling water and pass through the con-
denser cooling system encountering sudden and
sizable temperature increases. Increased cooling
water use by industrial and power plants may
increase the temperature of certain areas of the
river (bays and eddies) for extended periods and
also affect zooplankton.

The specific objectives of the study were: 1) to
develop techniques for laboratory study of ther-
mal effects on zooplankton and 2) to assess the

'Northwest Fisheries Center, National Marine Fisheries Ser-
vice, NOAA, 2725 Montlake Boulevard East, Seattle, WA
98112.

^Craddock, D. R., T. A. Blahm, and W. D. Parente, 1974. Occur-
rence and utilization of zooplankton by juvenile chinook salmon
in the lower Columbia River Unpubl. manuscr Northwest Fish.
Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, Wash.

Manuscript accepted October 1975.
FISHERY BULLETIN; VOL. 74, NO. 2, 1976.

403

FISHERY BULLETIN; VOL. 74, NO. 2

effect of both prolonged and short exposure to in-
creased temperatures on survival and reproduc-
tion of D. pulex.

METHODS AND MATERIALS

Two stocks of D. pulex were cultured at two
acclimation temperatures and subjected to three
tjT^es of tests to determine their thermal toler-
ance. One stock was obtained from the Columbia
River and the other from a small pond north of
Seattle, Wash. They were cultured separately and
will be referred to as the Columbia group and the
Seattle group. Stock cultures were maintained in
5-liter battery jars of Lake Washington water
filtered through No. 25 Swiss silk bolting cloth to
remove zooplankton and phytoplankton, but not
bacteria. Taub and Dollar (1968) felt that bac-
teria were important to the nutrition ofDaphnia,
especially in relation to reproduction. Stock cul-
tures were reared and acclimated at either 15° or
20°C in a controlled temperature incubator Con-
tinuous fluorescent lighting (45-50 foot candles,
cool white) provided similar lighting in the in-
cubator and in the laboratory and was consistent
for all animals, test and control. Algae, Chlorella
and Chlamydomonas, were cultured using me-
dium No. 63 developed by Taub and Dollar (1968)
and fed to D. pulex. Water in the test vessels was
changed weekly, and the animals were fed three
times a week.

The test temperatures were maintained by
using primary and secondary water baths and

immersion heaters activated by temperature con-
trollers (Figure 1). The primary bath was a
Plexiglas^ tank 150 x 30 x 23 cm supphed with
flowing water at 10° to 15°C. The secondary baths
consisted of six or seven 5- liter battery jars, 23 x
14 X 17 cm, placed in the primary bath. The
temperature in each of these secondary baths was
raised progressively from the water inlet end to
the outlet end of the primary tank. Temperatures
in the secondary baths could be maintained from
10° to 36°C ± 0.5°. Air continually bubbling into
each secondary bath eliminated stratification.
Experimental subjects were held in 50-ml jars of
filtered lake water suspended in the secondary
baths and equilibrated to the test temperature in
those baths.

Parthenogenetically produced animals of the
same age, either young females (less than 24 h
old) or mature females (approximately 1 wk old),
were selected from the stock cultures and held in
10-ml vials for a day before the start of the exper-
iment to check for handling mortality. At the
start of an experiment, the bulk of the water in
the vials was canted off, and the appropriate
number of test animals was poured directly into
the 50-ml test chamber at the test temperature.
The control groups were treated identically with
the others, except that they were held at acclima-
tion temperatures. A large bore pipette was used
when individual animals were handled.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Figure l. — Experimental equip-
ment used to study temperature ef-
fects on zooplankton showing primary
and secondary water baths, test ves-
sels, and temperature controllers.

404

CRADDOCK: EFFECTS OF WATER TEMPERATURE ON DAPHNIA PULEX

Three experiments were conducted to simulate
thermal conditions that D. pulex might en-
counter. Two experiments studied the effect of in-
creased temperatures that might be encountered
in the area of a heated plant outfall, whereas the
third simulated the thermal conditions small or-
ganisms could encounter in the condenser cooling
system of a thermal power plant.

The first experiment compared the effect of pro-
longed exposure (50-52 days) to constant temper-
atures of 15° (control), 18°, 21°, 24°, 27°, 30°, and
33°C. Test organisms were both mature females
and young females (at the start of the tests) of the
Seattle group acclimated at 15°C. There were 18
mature females per test temperature, 6 per test
jar, and 10 young females were tested per test
temperature and test jar. Ten Daphnia per 50-ml
jar were well below the number that would cause
harmful metabolic waste buildup or oxygen de-
pletion (Pratt 1943); 10 animals has long been
accepted as a standard for bioassays, Doudoroff
(1951), American Public Health Association
(1971), and Sprague (1973).

The second experiment compared the effect of
prolonged exposure (34 days) to temperatures of
20° (control), 23°, 26°, 29°, and 32°C on mature
females of the two groups (Seattle and Columbia)
acclimated at 20°C. There were 10 animals per
test temperature and test jar.

The third experiment subjected mature females
of the Seattle group acclimated to 15°C to a short
exposure (15 min) to temperatures of 15° (con-
trol), 19°, 21°, 24°, 27°, 30°, 33°, and 36°C. Test
organisms were then returned to acclimation
temperature where they were held and observed
for 90 days. Twelve animals were tested at each
temperature.

Test animals were examined frequently to de-
termine the effect of increased temperatures,
usually hourly during the first 8 h of a test. The
next day or two, they were examined two or three
times a day and subsequently once each week
day. During each observation, the mortalities
were noted and removed, and newly born Daph-
nia were counted and removed. The animals were
assumed to be dead when they lay on the bottom
and there was no detectable movement of the an-
tennae, thoracic legs, or the post abdomen.

Temperature effects were evaluated on the
basis of survival and reproduction by animals
tested at the various temperatures. In this study,
my evaluation criterion was the time at a particu-
lar temperature until 50% mortality; therefore, I

use the term TD50 (time to death of 50% of the test
animals at a particular temperature).

RESULTS

Experiments Relating to Discharges
of Heated Water

Seattle Daphnia Acclimated to Water of 15°C

Death occurred rapidly for both mature and
young D. pulex at 33°C. Some animals in both
groups lost equilibrium within the first hour,
TD50 occurred before the third hour, and none
survived the fourth hour of exposure (Table 1).
Mature and young Z). pulex subjected to tempera-
tures above 27°C reached TD50 in less than 24 h.
Temperatures of 27°C and below required an ex-
posure of at least 192 h (8 days) to cause 50%
mortality. The younger females did not succumb
to moderately high temperatures (18°, 21°, and
24°C) as quickly as the older females. Tempera-
tures of 21°, 24°, and 27°C caused TD50 among the
older females after an average of 238 h, whereas
the younger females did not reach TD50 until an
average of 768 h.

Table 1. — Mortality oi Daphnia pulex introduced as mature
and young females and maintained at temperatures of 15° to
33°C (Seattle race, acclimated at 15°C).

Mature

females

Young

females

Test

Hours to

% mortality

Hours to

% mortality

temp

50%

at end of test

50%

at end of test

(X)

mortality'

(50 days)

mortality'

(52 days)

1,008(42)

1,224 (51)

888 (37)

21,248 (51)

259 (9)

1,152 (48)

192 (8)

100

648 (27)

100

264 (11)

100

504 (21)

100

'Days

in parentheses.

250%

mortality not reached.

All animals died before producing young at 30°
and 33°C; rate of reproduction was highest at 24°
and 27°C before all subjects died. Total offspring
produced and rate of reproduction varied for the
two age-groups of females tested at 21°C or below
(Table 2).

First reproduction by the mature females oc-
curred 5 days earlier at test temperatures of 27°,
24°, and 21°C than at the control temperature
(15°C). Only one peak of production occurred at
27 °C before 100% mortality was reached. Repro-
duction at 15°C was stable with peaks occurring

405

FISHERY BULLETIN: VOL. 74, NO. 2

Table 2. — Reproduction of Daphnia piilex introduced as
mature and young females and maintained at temperatures of
15° to 33°C (acclimated at 15°C, Seattle race).

Table 3. — Mortality of Daphnia pulex (Seattle and Columbia
races) acclimated at 20°C and introduced as maturing fe-
males to temperatures of 20° to 32°C.

Mature females

Young females

Test
temp
(°C)

20
23
26
29
32

Seattle race

ColumI

3ia race

Test
temp

rc)

15
18
21
24
27
30
33

Total

young

produced^

1,162
652
244
466
318
0
0

Average no.

young/adult

per day2

2.20
1.05
1.11
2.61
3.02
0.00
0.00

Total

young

produced^

40
286
366
629

Average no.

young/adult

per day^

0.09
0.08
0.74
2.67
3.25
0.00
0.00

Hours to

50%
mortality'

2-816 (34)
648 (27)
120 (5)
120(5)
<24 (1)

% mortality

at end of test

(34 days)

80
100
100
100

Hours to

50%
mortality'

216 (9)

456(19)

48 (2)

120 (5)

<24 (1)

% mortality

at end of test

(34 days)

70
100
100
100
100

'Days
250%

In parentfieses.
mortality not reached.

'Total reproduction from 18 animals during 50 days of experiments.
^Average reproduction based on number of days survivors remained.
^Total reproduction from 10 animals for 52 days of experiments.

regularly at 6-day intervals, whereas at higher
temperatures reproduction was erratic. The
greatest numbers of offspring were produced by
the older females at 15°C and generally decreased
with increasing temperatures. The highest rates
were at 27° and 24°C where the survivors repro-
duced rapidly before they all succumbed on the
13th and 27th day, respectively. High tempera-
tures increased the rate of reproduction for a
short period before total mortality, but the in-
creased rate was short lived and did not match
total production by animals at a more normal
temperature (15°C).

Reproduction by females who were young at
the start of the experiment increased with in-
creasing temperature, contrary to the trend
shown by the older females (Table 2). No repro-
duction occurred at 15°C until the 34th day; at
18° and 21°C, initial reproduction took place on
the 3rd to 6th day but did not resume until the
44th and 22nd day, respectively. At 24° and 27°C,
the first reproduction occurred on the 3rd to 6th
day, stopped for 3 or 4 days, and then continued at
a high rate until the death of all females on the
34th and 27th day, respectively. The low repro-
duction by the younger females at 15° and 18°C is
not explained.

Seattle and Columbia Daphnia Acclimated
to Water of 20°C

Both Seattle and Columbia Daphnia reached
TDso within 24 h at 32°C (Table 3); 90% mortality
occurred in less than 24 h in the Columbia group
and within 48 h in the Seattle group. At 29°C,
both groups reached 50% mortality in 120 h.
There were significant differences in the length of
time to 50% mortality for each of the two groups
at 20°, 23°, 26°, and 29°C (Seattle— x^ = 37.9,

P<0.01, 3 df; Columbia— x^ = 18.8,P<0.01, 3 df).
The Columbia group seemed to succumb more
rapidly than the Seattle group, but the more
rapid demise of the Columbia Daphnia at 20°C
(the acclimation temperature) casts doubt upon
these results. However, a test of homogeneity for
temperatures of 23°, 26°, and 29°C indicated sig-
nificant differences between the two groups in
days to 50% mortahty (x^ = 22.6, P<0.01, 2 df).
Comparatively little reproduction took place at
temperatures of 26°C and above (Table 4). The
greatest reproduction for the Seattle group was at
23°C and for the Columbia group at 20°C. The
Columbia animals remaining after the initial un-
explained mortality at 20°C outproduced the
Seattle animals at the same temperature. The
Seattle animals produced 62% of the total young
produced by the two groups.

Table 4. — Reproduction of Daphnia pulex (Seattle and Colum-
bia races) acclimated at 20°C and introduced as maturing
females to temperatures of 20° to 32°C.

Seattle race

Columbia

a race

Test

Total

Average no.

Total

Average no.

temp

young

young/adult

young

young/adult

(C)

produced'

per day2

produced'

per day2

246

1.12

299

1.82

424

1.68

152

1.02

0.00

0.67

2.14

0.32

32
Total

0
760

000

470

0.00

'Total reproduction by 10 animals for 34 days of experiments.
^Average reproduction based on the number of days survivors remained.

Experiments Relating to Water
Passing Through Cooling Systems

Exposure for 15 min at temperatures of 30°C or
less seemed to have little or no effect upon the
survival of D. pulex (Table 5). The only mor-
talities observed during the exposure period were
at 36°C: within 5 min, over 50% of the animals at
this temperature were dead; all but one died in 15

406

CRADDOCK: EFFECTS OF WATER TEMPERATURE ON DAPHNIA PULEX

Table 5. — Mortality and reproduction of Daphnia pulex
(Seattle race) exposed as maturing females for 15 min to vari-
ous temperatures and returned to acclimation temperature
of 15°C.

Survival

Reproduction

Shock

Hours to

% mortality

Total

Average no.

temp

50%

after

young

young/adult per

rc)

mortality'

90 days

produced^

day of test^

1,178 (49)

2,051

3.64

1.320 (55)

1,340

2.30

1,008 (42)

1,365

3.57

1,512 (63)

2,640

3.50

1,464 (61)

1,716

2.26

1,536 (64)

1,832

2.48

792 (33)

100

363

0.78

0.083

"480

5.33

'Days in parentfieses.
^Total reproduction for 90 days of test.
^Average daily reproduction per surviving adult.

■•Produced by Ifie one survivor of tfie 15-min exposure during the succeed-
ing 90 days.

min. One hour after exposure, one animal had
died at 33°C, but TD50 took 792 h (33 days) at
33°C and 1,008-1,536 h (42-64 days) after expo-
sure to temperatures below 33°C. Time in days to
reach TD50 was not statistically significant (x^ =
6.89, 5 df ) for temperature treatments of 15° to
30°C. A temperature in excess of 30°C for the
15-min exposure was necessary to significantly
increase mortality.

The rate of reproduction was not significantly
changed by an exposure of 15 min to increased
temperatures through 30°C (x^ = 0.79, 5 df ). The
greatest total reproduction was by those D. pulex
tested at 24°C (Table 5) where survival was also
good. Reproduction at 27° and 30°C exceeded the
reproduction at 19°C, so it appears that reproduc-
tion is not materially affected by a short exposure
to temperatures through 30°C that do not seri-
ously affect survival. Reproduction by animals
tested at 33° and 36°C was drastically reduced
because most of the test animals died.

DISCUSSION

In zooplankton sampling of the Prescott-
Kalama section of the Columbia River in 1968-69,
D. pulex was more abundant during periods of
higher water temperature (Craddock et al. see
footnote 2). Numbers of D. pulex were low during
the portion of the year when the temperature re-
mained below 15°C (late fall, winter, and spring),
but as water reached and exceeded this tempera-
ture the population increased rapidly until the
peak abundance was reached at the maximum
water temperature (approximately 21°C). The
mean daily water temperature in August (the

month of highest temperature) ranged from 19.3°
to 22.8°C in 1968 and from 19.7° to 21.1°C in 1969
(Snyder and McConnell 1971). Tauson (1931)
found temperatures of 16°-22°C favorable for
parthenogenetic reproduction by D. pulex, but
above or below this range production was reduced
considerably. The upper limit was 30°C. Ivleva
(1969) reviewed literature on the thermal range
of Daphnia and noted that several researchers
reported the optimum temperature range for de-
velopment of D. pulex as 18°-20°C. Ivleva made
the general observation that the optimal range
varies with age and the young are more resistant
to high temperature than the old, as was indi-
cated by my experiments. Other researchers re-
viewed by Ivleva found that mass mortalities
could occur in the range of 28°-32°C. Some of
these researchers indicated that when Daphnia
species are acclimated to higher or lower temper-
atures over a long period they become more
resistant to further increases or reductions in
temperature.

My experiments to determine the effect of in-
creased temperature on D. pulex that were 1 wk
old and 1 day old (i.e., at the start of the experi-
ment) indicated that the younger animals
adapted better to increased temperatures. Tem-
peratures of 21°C and above seriously reduced the
length of survival of the older females (21°C =
TD50 in 259 h), whereas temperatures of 24° to
27°C or more were required to have the same ef-
fect on the younger females (24°C = TD50 in 648
h; 27°C = TD50 in 504 h). Temperatures above
27°C caused TD50 in a short time (less than 21 h)
for both age-groups.

Although the younger females survived better
at the control and lower test temperatures (15°,
18°, and 21°C), their eventual production of
young was considerably less than that of the ma-
ture animals. This difference was not due solely
to the 1-wk difference in age, and I do not have an
adequate explanation.

My experiment comparing survival and repro-
duction of the Seattle and Columbia races indi-
cated that the Columbia Dap/?7zm may be less re-
sistant to increased temperatures. The results of
the tests of Seattle D. pulex acclimated at 15° and
20°C are not directly comparable and, although
there is some indication that the higher tempera-
ture acclimation increases resistance in the mid-
range (23°-24°C), the effect was not apparent in
the high range (26°-27°C) and no conclusion could
be made.

407

My experiments indicate that an increase of
6°C in the area of an outfall could cause TD50 in
about 168 h (7 days) among important segments
of the reproducing population. To minimize dam-
age to Daphnia populations in the Columbia
River, the temperature should not be raised more
than 6°C above ambient or higher than 26° or
27 °C for any prolonged period.

A short exposure (15 min) to increased temper-
atures as might occur in a condenser cooling sys-
tem did not cause a significant reduction in time
to TD50 or in reproduction unless the temperature
exceeded 30°C. There is a period from mid-July
through September when the lower Columbia
River temperatures may exceed 20°C. In these
instances, the temperature increase in condenser
cooling systems should be less than 10°C if the
Daphnia are to survive. It must be kept in mind
that temperature is only one of several factors
including pressure, abrasion, and toxic chemicals
that could be acting synergistically to damage
zooplankton in a condenser cooling system
(Marcy 1973; Becker and Thatcher'*).

To protect D. pulex populations, water temper-
atures in condenser cooling systems should not
exceed 30°C and passage through the system
should take less than 15 min.

ACKNOWLEDGMENTS

Rufus W. Kiser, CentraUa College, Centralia,
Wash., verified the identification o{ the Daphnia.
Donald D. Worlund and Frank J. Ossiander pro-
vided advice on statistical treatment. Linda Street
McCune assisted in all aspects of culturing and
testing Daphnia

"Becker, C. D., and T. O. Thatcher. 1973. Toxicity of power
plant chemicals to aquatic life. Battelle Mem. Inst., Pac. North-
west Lab., Richland, Wash., rep. for U.S. At. Energy Comm.,
WASH-1249, UC-11, misc. pagination.

FISHERY BULLETIN: VOL. 74, NO. 2

LITERATURE CITED

American public health association.

1971. Standard methods for the examination of water and
wastewater. 13th ed. Am. Publ. Health Assoc, Wash.,
D.C., 874 p.
COUTANT, C. C.

1970. Entrainment in cooHng waters: steps toward predic-
tability. Proc. 50th Annu. Conf. West. Assoc. State Game
Fish Comm., p. 90-105.

DOUDOROFF, P., B. G. ANDERSON, G. E. BURDICK, P. S.
GALTSOFF, W. B. HART, R. PATRICK, E. R. STRONG, E. W.

Surber, and W. M. Van Horn.

1951. Bio-assay methods for the evaluation of acute toxicity
of industrial wastes to fish. Sewage Ind. Wastes
23:1380-1397.
Ivleva, I. V.

1969. Mass cultivation of invertebrates. Biology and
methods. Izd. "Nauka", Moscow. (Translated by Israel
Prog. Sci. Transl., 1973, 148 p.; available U.S. Dep. Com-
mer, Natl. Tech. Inf. Serv., Springfield, VA, as TT 65-
50098.)

Marcy, B. C, Jr.

1973. Vulnerability and survival of young Connecticut
River fish entrained at a nuclear power plant. J. Fish.
Res. Board Can. 30:1195-1203.
NORTH, W. J., AND J. R. ADAMS.

1969. The status of thermal discharges on the Pacific
Coast. Chesapeake Sci. 10:139-144.
PRATT, D. M.

1943. Analysis of population development in Daphnia at
different temperatures. Biol. Bull. (Woods Hole) 85:116-
140.
SNYDER, G. R., AND R. J. MCCONNELL.

1971. Subsurface water temperatures of the Columbia
River at Prescott, Oregon (Hiver mile 72), 1968-69. U.S.
Dep. Commer., NCAA, Natl. Mar. Fish. Serv., Data Rep.
53, 9 p. on 1 microfiche.

SPRAGUE, J. B.

1973. The ABC's ofPollutantBioassay using fish. Am. Soc.
Test. Mater., Spec. Tech. Publ. 528:6-30.

Taub, F. B., and a. M. Dollar

1968. The nutritional inadequacy of Chlorella and
Chlamydomonas as food for Daphnia pulex. Limnol.
Oceanogr. 13:607-617.

Tauson, a.

1931. Die Wirking der ausseren Bedingungen auf die Ver-
anderung des Geschlechts und auf die Entwicklung von
Daphnia pulex De Geer. Wilhelm Roux Arch. Ent-
wicklungsmech. Org. 123:80-131.

408

LIFE HISTORY, ECOLOGY, AND BEHAVIOR OF

LIPARIS INQUILINUS (PISCES: CYCLOPTERIDAE) ASSOCIATED

WITH THE SEA SCALLOP, PLACOPECTEN MAGELLANICUS'

K. W. Able^ and J. A. MusiCK^

ABSTRACT

In the Mid-Atlantic Bight, spawning ofLiparis inquilinus peaked near shore, away from sea scallop
beds, in March and April. In the laboratory, females appeared to initiate spawning activity and each
female probably spawned more than once. The eggs are adhesive and demersal and have been found
attached to hydroids in nature. The larvae were most abundant in plankton collections inshore in May
and averaged 5 mm total length at that time. Larger larvae were found in deeper water, and by 12-13
mm total length they had undergone metamorphosis and descended to the bottom where they became
associated with the sea scallop, Placopecten magellanicus. They maintained this association from
August through December. The population comprises a single year class which leaves the scallops and
migrates inshore to spawn as the fish are entering their second year.

Laboratory and field observations indicated that fish were more abundant in the scallops and more
scallops contained fish during the day. At night, fish left the scallops to feed on small crustaceans.
Liparis inquilinus observed in aquaria used the fin rays of the lower lobe of the pectoral fin to detect
food. These fin rays have taste buds on the surface of each ray.

Liparis inquilinus is probably protected from predation while inside sea scallops since there are few
predators on the scallops of the size usually occupied. Predation while outside the scallop may be
minimized by feeding only at night and then returning as soon as the fish becomes satiated. Sea scallops
seem to suffer no ill effects from the association and they do not compete for food with L. inquilin us since
P. magellanicus is a microplanktonic filter feeder and the former feeds on small crustaceans.

Little is known of the life history of most species
of L/paWs. Most of the meager information avail-
able for North Atlantic Liparis is included in
Bigelow and Schroeder (1953), Andriyashev
(1954), Leim and Scott (1966), and Wheeler
(1969). Unfortunately, taxonomic problems re-
main and some published life history information
may be incorrect because of misidentification. Re-
cently, Detwyler (1963) studied the life history
and reproductive biology of L. atlanticus from
New Hampshire and Maine, and Able (in press)
commented on the life history of a new species of
Liparis from the Gulf of Maine. Elsewhere,
Nizortsev et al. (1963) noted the stomach contents
of L. koefoedi, L. liparis, and L. lapteui in the Ba-
rents Sea; Johnson (1969, 1970) reported on food
habits and age and growth of L. pulchellus off
California; Kosaka (1971) described the food
habits and seasonal distribution of L. tanakae
from Japan; Gibson (1972) mentioned the vertical

'Contribution No. 730 from the Virginia Institute of Marine
Science.

^Biology Department, McGill University, Montreal, Quebec,
Canada.

^Virginia Institute of Marine Science, Gloucester Point, VA
23062.

distribution and feeding of L. montagui; and
Quast (1968) described the food habits of L.
mucosas off California.

The association between Lipam (= L. inquili-
nus, see Able 1973) and the sea scallop, Placopec-
ten magellanicus, has been reported by several
authors (Bean 1884; Goode 1884; Garman 1892;
Goode and Bean 1895; Jordan and Evermann 1898;
Welsh 1915; Burke 1930; Bigelow and Schroeder
1953; Leim and Scott 1966) but information is lack-
ing on most aspects of the assocation. The purpose
of this paper is to report on the life history, ecology,
and behavior of L. inquilinus.

MATERIALS AND METHODS

The life history stages, although often overlap-
ping, are defined as follows: larvae — planktonic
individuals usually 3-13 mm total length (TL),
which have not transformed to adult coloration;
juveniles — sexually immature benthic individu-
als with adult coloration, often associated with
the sea scallop, approximately 14-45 mm TL; and
adults — sexually mature individuals greater
than 33 mm TL. The latter can be distinguished
by the presence of prickles on the body of males

Manuscript accepted December 1975.
FISHERY BULLETIN: VOL. 74, NO, 2. 1976.

409

FISHERY BULLETIN: VOL. 74, NO. 2

and by the enlarged abdomen of females. Scallop
anatomical terminology follows Bourne (1964).

Larval Liparis were examined from monthly
collections of the National Marine Fisheries Ser-
vice (NMFS) laboratory at Sandy Hook in the
Mid-Atlantic Bight during 1966-67 (Clarke et al.
1969) and from routine plankton sampling on the
Woods Hole NMFS RV Albatross IV cruises 69-5
and 72-3 off southern New England, in the Gulf of
Maine, and on Georges Bank. The larvae of L.
inquilinus can be distinguished from those of
other Liparis which occur in the Mid-Atlantic
Bight and the Gulf of Maine by differences in
pigmentation pattern in combination with size at
hatching, disc formation, and notochord flexion
(Able 1974). The eggs were identified on the basis
of their similarity, in size of the egg and me-
lanophore pattern of the embryo, to eggs de-
posited by the laboratory population of L.
inquilinus.

Juvenile L. inquilinus were collected from sea
scallops which were taken in otter trawls during
cruises of the Sea Breeze while on charter to the
Virginia Institute of Marine Science, and Alba-
tross IV cruises 69-11 and 70-6. Other L. in-
quilinus were collected from sea scallops on
Albatross IV cruises 68-14 and 69-8 with a 3-m
scallop dredge with a 5.1-cm ring bag which was
towed for 10 min at each station. On Albatross FV
cruise 68-14, bottom substrate type and amount
were estimated from the scallop dredge catch.
Size and number of scallops and regular hydro-
graphic data were also recorded. On Albatross IV
cruise 69-8, scallop dredge tows were replicated
every 2 h during a 24-h period on 4-5 August 1969.
The same general area was maintained during
sampling by using information from depth record-
ers and loran navigation. The scallop catch at
each station was divided into 5-cm height classes
and a representative number of scallops were ex-
amined for L. inquilinus from each size class.

A large series of adult L. inquilinus collected
off the New Jersey coast in the 1930's was
examined from uncatalogued material of the
Academy of Natural Sciences of Philadelphia.
Other small collections were obtained from a
variety of sources that are too numerous to men-
tion here.

Liparis inquilinus and sea scallops were col-
lected between lat. 39°30' and 40°10'N near Hud-
son Canyon in depths of 36-95 m and maintained
in 10 to 25 gallon aerated aquaria with sand sub-
strates for 15 mo. The aquaria were held in a cold

room at 10°-11°C. Winter temperatures in aquaria
dropped as low as 4°C because of the absence of
heating facilities. Salinity varied from 23 to 42%o.
Illumination was provided by a 60-W bulb in one
corner of the room. This provided approximately
86 to 280-lx illumination for the aquaria, depend-
ing on their location in the room. The light cy-
cle was controlled automatically and approxi-
mated that in nature. Occasional power failures
caused irregular variation in photoperiod and
temperature.

Liparis inquilinus were fed live amphipods,
usually Orchestia platenis and Gammarus mu-
cronatus, and the mysid shrimp Neomysis ameri-
cana and various other small crustaceans. Sea
scallops were fed a mixture of algae, Mono-
chrysis lutheri, Isochrysis galbana, and Phaeodac-
tylum triconutum, that was added to the unfil-
tered aquarium water.

Pectoral fins ofL. inquilinus were sectioned and
stained with Harris' hematoxylin and eosin Y fol-
lowing fixation in 10% Formalin.^

LIFE HISTORY OF
LIPARIS INQUILINUS

In the Mid- Atlantic Bight, spawning of L. in-
quilinus occurs near shore and away from scallop
beds in the winter. In the early 1930's, over 700
adult, sexually mature and maturing L. in-
quilinus were collected from mid-December
through April (Figure 1) off the coast of New Jer-
sey and Delaware. This species was found from
the Brigantine Can Buoy north of Atlantic City,
N.J., to near the mouth of Delaware Bay and in-
side the bay at Old Bare Shoal and in deep holes
off Brandywine (Shoal?) and Lewes, Del. Most of
the collections were in 7-14 m; however, part of
this series was from depths as shallow as 3-4 m
"off New England Creek (near Cape May Co.)."
Unfortunately, we have been unable to locate this
area in New Jersey. Recently (January-March
1971 and January-February 1973) other mature
adults were found off New Jersey, especially off
Little Egg Inlet in depths from 4 to 7 m. Sea scal-
lops were never taken in the vicinity of these col-
lections (D. Thomas pers. commun.).

The average total length of L. inquilinus in-
creases from December through April (Figure 1).
Detwyler (1963) attributed an increase in total

''Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

410

ABLE and MUSICK: LIFE mSTORY AND BEHAVIOR OF LIPARIS INQUILINUS

length of L. atlanticus during the winter to the
replacement of smaller adults by larger adults as
the spawning season progressed. This may occur
for L. inquilinus , but it seems more likely that
sexually immature fish moving inshore in No-
vember, December, and January may continue
to grow as they become sexually mature. In the
laboratory, fish continued to feed during spawn-
ing periods. Although the range in total length
for collections for each month is large, the varia-
tion about and between the means is small (Fig-
ure 1). This probably indicates that a single year
class is present in each sample.

Spawning in L. inquilinus probably peaks in
March and April. A single collection of L. in-
quilinus eggs was made on 9 March 1973, approx-
imately 3.5 nautical miles off Holgate, Long
Beach Island, N.J. Also, the adult fish rep-
resented in Figure 1 were examined for sexual
maturity. The percentage of sexually mature fish
increased from 12% in January, to 44% in Feb-
ruary, and to 67% in March, but decreased to 33%
in April, although this last sample was small.
Hatching times for other Liparis vary from 22-30
days for L. atlanticus (Detwyler 1963) to 6-8 wk
forL. liparis (Breder and Rosen 1966). Therefore,
the occurrence of L. inquilinus larvae averaging 5
mm in May (Figure 1) infers that spawning prob-
ably takes place in March and April, and this is in
agreement with the time of occurrence of sexually
mature adults in inshore waters.

In the laboratory, reproductive activity and
egg laying occurred over many months. During
1969, females distended with eggs and performing
prespawning behavior (see below) were present
from January through August. Eggs with eyed lar-
vae were first found in late April and egg masses
were found through June. Successful hatching oc-
curred only in May. The extensive period of egg
deposition and reproductive behavior observed
in the laboratory does not agree with the limited
reproductive period inferred from field collections.
These differences may be attributable to the occa-
sional power failures which affected photoperiod,
water temperature, and water quality in the
laboratory aquaria or simply to laboratory
confinement.

It is likely that the average size of sexually
mature males and females is similar and that the
sex ratio is 1:1. A single collection of 143 L. in-
quilinus (off New England Creek, 7 m, 22 Feb-
ruary 1933) contained 75 mature and maturing
males (mean 55.3 mm TL, range 37.1-69.6 mm

TL) and 68 females (mean 54.3 mm TL, range
44.3-65.7 mm TL). Neither the ratio of males to
females nor the average total length was sig-
nificantly different.

PLANK TONIC
LARVAE

JUVENILES IN
SEA SCALLOPS

Eggs

407 7

i t

JAN FEB MAR APR MAY JUN JUL AuG SEP OCT NOV DEC

FIGURE 1. — Length-frequency distribution of Liparis inquili-
nus collected from the Mid-Atlantic Bight. For each sample, the
range is represented by the vertical line, mean by the horizontal
line, one standard deviation on each side of the mean by hollow
rectangles and two standard errors on each side of the mean
by solid rectangles. Numbers above figure are sample sizes.
A single collection of L. inquilinus eggs is noted on the hori-
zontal £Lxis.

Female L. inquilinus may spawn more than
once. In the laboratory, the abdomen of individual
fish was observed to decrease in size as more egg
masses were found in aquaria and increase again
later. Also the egg diameters in ovaries of females
from 16 March 1932 and 1933 usually had two
well-defined modes. Fourteen ovaries were ex-
amined and most eggs were either 1.00-1.30 or
0.01-0.50 mm in diameter. The largest eggs were
clear and contained several oil globules and these
were more abundant in the center of the ovary.
When egg diameter modes in the ovary were not
well-defined, egg distribution by size was often
random. Counts for the larger eggs ranged from
105 to 1,135 (mean 447) in seven ovaries from
females raised in the laboratory and from 231 to
563 (mean 342) for females collected off New Jer-
sey. The high count for females raised in the
laboratory may have been due to the failure of the
female to spawn and continued development and
accumulation of the eggs in the ovary because of
disturbances in the laboratory. There seemed to
be no correlation between fish size and egg num-
bers. The average number of eggs is less than the
475 to 700 eggs reported for L. atlanticus (Det-
wyler 1963), a larger species.

411

FISHERY BULLETIN: VOL. 74, NO. 2

Spawning Behavior

Female L. inquilinus may initiate spawning
activity. In laboratory aquaria females with dis-
tended abdomens were the most active. They
often swam in quick dashes around the sides of
the aquarium then up to the surface and down
again. During these dashes, the snout came out of
the water and there was considerable splashing.
Similar behavior has been reported for L. atlan-
ticus females (Detwyler 1963). This activity often
lasted several minutes and on one occasion 7 min
and 20 s. Occasionally during these excited
dashes the females would bump into other fish,
both males and females. In a few instances, this
activity seemed to excite other females and they
also became active. In one instance, a ripe female
repeatedly nudged with her snout a fish of un-
known sex that was attached to the side of the
aquarium. Soon a prominent bulge appeared just
posterior to the genital papilla of the female. This
has been observed just before spawning in L. at-
lanticus (Detwyler 1963) and Cyclopterus lumpus
(Cowan 1929). In this instance, the nudged fish
did not respond and the female swam away. The
bulge receded after about 5 min. Sexually mature
males are covered with numerous prickles while
the females usually lack these or have only a few.
Thus, the female may be able to recognize males
by making contact with them. Breeding tubercles
and contact organs in fishes may function in
maintenance of body contact between the sexes
during spawning and stimulation during breed-
ing (Wiley and Collette 1970). The prickles on L.
inquilinus males may function in these ways also.
Spawning was not observed but is probably simi-
lar to that inL. atlanticus (Detwyler 1963). In the
laboratory, L. inquilinus deposited small clumps
of 20-80 eggs on the bottom of the aquaria and
did not guard them. The eggs collected on 9
March 1973 off New Jersey were attached to
hydroids as has been reported for L. liparis
(Ehrenbaum 1905). The larvae that hatched in
the laboratory did not survive beyond yolk sac
absorption.

Larvae

In the Mid-Atlantic Bight, larvae of L. in-
quilinus are planktonic during the spring. Dur-
ing monthly larval fish surveys in 1966-67 by the
Sandy Hook Laboratory, 98% of the L. inquilinus
larvae were collected in May (Figure 1) from deep
and shallow tows. These averaged 5.1 mm TL

(range 3.2-12.0 mm TL). Larvae were most abun-
dant in samples collected nearest to shore (Figure
2). Other larvae of the same average size have
been collected during May from inshore waters in
the Gulf of Maine and on Georges Bank (Table 1,
Fig. 3). Larvae larger than 13 mm TL were usu-
ally not found in the plankton.

72"

70°

EGEND

NONE

•

- 10

•

- 40

-100

■

..s>

4I<=

70°

40°

39°

\72°
38°

37 =

73°

36°

Figure 2. — Distribution and abundance of larval Liparis
inquilinus from Dolphin cruise D-66-5 during May 1966.

Juveniles

In the Mid-Atlantic Bight, juvenile L. in-
quilinus are associated with sea scallops from
August through December. Stevenson^ reported

^Stevenson, J. A. Fish. Res. Board Can., St. Andrews, New
Brunswick, Manuscr. Rep. 373.

412

ABLE and MUSICK: LIFE HISTORY AND BEHAVIOR OF LIPARIS INQUILINUS

Table l. — Collections of larval Liparis inquilinus from plankton sampling cruises. Mean followed by range in parentheses.

Item

Dolphin 66-3

Dolphin 66-5

Dolphin 66-7

Albatross IV 69-5

Albatross IV 72-3

Locality

Mid-Atlantic Bight

Georges Bank

uthern New England
Gulf of Maine
Georges Bank

Date

8 April 1966

12-20 May 1966

18-27 June 1966

22-26 May 1969

6-16 May 1972

Water depth (m)

—

73.0(65-88)

65.8(36-96)

Number collected

414

Number measured

269

Total length (mm)

3.8

5.0(4.0-11 0)

9.4(3.4-11.5)

5.8(3,5-15.7)

4.6(3.7-10.0)

LEGEND
• ALBATROSS IS. 72-3
O ALBATROSS IE 69-5

Figure 3. — Locations of collections of larval Liparis inquili-
nus from Albatross TV cruises 69-5 and 72-3.

Neoliparis (Liparis) atlanticus from sea scallops
as early as July in the Bay of Fundy off Digby,
Nova Scotia. Specimens we have from sea scal-
lops in that area are allL. inquilinus . JuvenileL.
inquilinus have also been collected from scallops
from Georges Bank in July and may be present in
sea scallops during July in the Mid-Atlantic
Bight as well. The fish found in the scallops dur-
ing August (Figure 1) corresponded in size with
that expected from the earlier collection of
planktonic larvae (Figure 1) and represented the
same year class. The average total length of fish
from scallops increased steadily from August
through November (Figure 1). The small variation
in each collection indicated that there was a single
year class inhabiting sea scallops during a single
year. Liparis inquilinus have been collected from
sea scallops as late as mid-December (17 Dec. 1967,
lat. 38°20'N, long. 73°59'W, 66 m and lat.
38°18'N, long. 74°23'W, 42 m) (Figure 1). The ab-
sence of fish in the scallops collected in January ( 18
Jan. 1968, lat. 38°34.5'N, long. 73°36'W, 62 m; 26
Jan. 1968, lat. 38°05'N, long. 74°13'W, 66 m) cor-
responds with the appearance of L. inquilinus in-
shore off Delaware and New Jersey during the
same periods. These mature and maturing fish

represent the same year class as the juveniles that
were associated with sea scallops. Therefore, L.
inquilinus reproduces when 1 yr old in the Mid-
Atlantic Bight.

Adults may not survive to spawn the following
year. Specimens larger than 50 mm have never
been taken from May through December. The life
history of L. inquilinus in the Mid-Atlantic Bight
is summarized in Figure 4.

DEATH*?

ADULTS

SPAWNING

FEB-APRIL

41 -72 mm TL

MATURING ADULTS
MIGRATING INSHORE
NOV- JAN
> 33 mm TL

LARVAE
PLANKTONIC
APRIL- JUNE
3-13 mm TL

JUVENILES
COMMENSAL IN SEA SCALLOPS
JULY - DEC
14 - 45mm TL

Figure 4. — Schematic presentation of the life history of Liparis
inquilinus in the Mid-Atlantic Bight.

ECOLOGY AND BEHAVIOR OF

LIPARIS INQUILINUS
ASSOCIATED WITH SCALLOP

Resting

In aquaria, L. inquilinus preferred an inverted
resting position with the disc attached to any
smooth substrate such as the side of the
aquarium, the interior of mollusk shells, rocks, or
glass containers. Once attached, the fish flexed its

413

FISHERY BULLETIN: VOL. 74, NO. 2

tail so that the caudal fin was alongside the head.
From 13 November to 20 December 1968, obser-
vations were made on the position offish attached
to four hinged sea scallop shells or "clappers."
These were positioned on the bottom of an
aquarium, with one of each of these pairs placed
with the right valve (flat valve) up and the left
valve down. One value rested on the bottom and
the other was at an angle of approximately 30°-
40°. Of 40 observations, 957c of the fish in shells
were attached upside-down to the top valve of the
clapper with as many as eight attached to the
same valve. The inverted resting position was
also the most commonly observed during the re-
mainder of the time fish were maintained in the
laboratory.

Feeding

Liparis inquilinus has several morphological
and behavioral adaptations which may allow it to
feed at night. In aquaria, fish swimming over the
bottom appeared to depend on reception of tactile
and/or gustatory stimuli received by the head and
pectoral fins. Swimming resulted from the com-
bined action of the tail and the upper lobe of the
pectoral fins. The eight or nine filaments in the
lower lobe of the pectoral fins were extended ver-
tically toward and often touched the bottom.
When amphipods were placed in aquaria, fish did
not appear to respond to visual cues but feeding
usually occurred when the head or the lower lobe
of the pectoral fin touched an amphipod. If food
touched the head, it was immediately ingested. If
food touched the pectoral fin, the fish quickly
backed up or arched its body to the side and
sucked in the prey. The rays in the lower lobe of
the pectoral fin of L. inquilinus contain dark
staining buds along the surface of each ray (Fig-
ure 5A) which are most abundant at the tips
(Figure 5B). They are identified as taste buds
on the basis of their similarity to the figures
presented by Bardach and Case (1965). They
described the sensitivity of the pelvic fins in
Urophycis chuss and the pectoral fins in Pri-
onotus carolinus and P. evolans to gustatory
stimuli. Freihofer (1963) suggested that the par-
ticular pattern of the ramus lateralis accessorius
nerve to the pectoral and pelvic fins in the Li-
paridae allows the development of these fins as
"sensory, locomotor and support appendages."
The well-developed cephalic lateralis system of L.
inquilinus may also function in detecting moving

prey. Occasionally fish sucked in amphipods
which passed within less than 1 inch of the head.
Liparis inquilinus feeds on benthic prey.
Stomachs of fish collected in nature contain al-
most exclusively small crustaceans and small
numbers of sand grains. In the laboratory, sand
from the bottom was frequently sucked in with
food items and then discharged from the gill
opening. A round mouth, as in L. inquilinus, is
well-adapted to sucking in prey (Alexander 1967).

Behavior of Fish Associated
with Sea Scallops

The association between L. inquilinus and sea
scallops is well-developed and both partners show
definite behavioral adaptations. Fish collected
from sea scallops were isolated from them for sev-
eral weeks. Upon reintroduction of fish into
aquaria containing acclimated sea scallops, many
of the fish swam around and over the scallops but
concentrated most of their activity along the scal-
lops' mantles. Most fish alternated between
swimming parallel to the mantle with the lower
lobe of the pectoral fin extended toward it or
swimming with the head oriented directly toward
the mantle. The tentacles on the mantle often
contracted but the valves did not close. On one
occasion a fish "mouthed" a tentacle, an action
similar to the acclimitization behavior of some
pomacentrid fishes associated with anemones
(Mariscal 1966). On two occasions, fish attached
to the mantle, and in one of these instances the
tentacles of the scallop mantle moved over the
body of the fish and depressed the anterior por-
tion of the dorsal fin. There was no reaction by
either partner and eventually the fish attempted
unsuccessfully to enter the scallop.

The tentacles of the sea scallop are tactile and
chemical receptors (Bourne 1964) and may be
able to discriminate between L. inquilinus and
other fishes. In aquaria, sea scallops exposed to
individuals of Gobiosoma bosci and Gobiesox
strumosus reacted negatively, when the mantle of
the scallop was brushed by either species, by clos-
ing the valves. Similar results were observed
when Fundulus heteroclitus and Tautogolabrus
adspersus, were exposed to sea scallops (Musick
1969).

Liparis inquilinus occasionally may enter an
alternate host species. Hoff (1968) reported a
specimen of L. atlanticus from the bay scallop,
Aequipecten irradians, in Buzzards Bay, Mass.

414

ABLE and MUSICK: LIFE HISTORY AND BEHAVIOR OF LIPARIS INQUILINUS

Figure 5. — Section through a fin ray from the lower lobe of the pectoral fin of Liparis inquilinus stained with
hematoxylineosin. Arrows indicate taste buds. A. Taste buds on margin of fleshy portion of fin ray. B. Numer-
ous taste buds at the tips of the fin rays.

We identified a specimen provided by him as L.
inquilinus . Since that initial occurrence he has
collected several other Liparis, which are proba-
bly also L. inquilinus , from bay scallops (pers.
corarann.) . Liparis inquilinus originally collected
from sea scallops were placed in aquaria with bay
scallops to determine if they would attempt to
enter the scallops. These scallops were completely
ignored and the fish made no attempt to enter or

attach to them. When brushed by L. inquilinus,
the bay scallops either showed no response or
closed the valves slightly. Bay scallops are found
in much shallower water than the sea scallops,
and the occurrence of L. inquilinus in depths fre-
quented by bay scallops is unusual. These fish
which occur in shallower water may attempt to
associate with bay scallops in the absence of their
regular host. Confusion in host recognition may

415

FISHERY BULLETIN: VOL. 74, NO. 2

occur where chemical stimulation is important
but other ecological factors usually prevent the
animal from associating with other forms
(Davenport 1955).

Over 30 attempts by L. inquilinus to enter sea
scallops were observed in the laboratory. The
length of time spent swimming along the mantle
of the scallop varied, but some fish were able to
enter in less than 3 s. After swimming along the
mantle most fish turned, placed the head at the
margins of the mantle, and attempted to force
their way inside the scallop with sustained
swimming strokes of the tail. One individual re-
peated this activity 10 times before it gave up.
The point of entry along the mantle appeared to
be selected randomly. Several fish attempted to
enter the incurrent and excurrent opening. The
scallop usually did not react to the fishes' en-
trance and only occasionally responded by closing
the valves slightly. The red hake, Urophycis
chuss, enters and exits the scallop only through
the excurrent opening (Musick 1969).

Perhaps there is individual variation in the ac-
ceptance of fish by scallops. On two occasions,
scallops rejected L. inquilinus after they had en-
tered the scallop by clapping the valves together
and thus forcing the fish out of the mantle cavity.
In each instance, the fish came to rest a few
inches from the edge of the scallop. The fish re-
mained still as the sand stirred up by the scallop's
activity settled over it. Within a few minutes, the
fish returned to the scallop and attempted to
enter again.

Once inside the mantle cavity of the scallop,
the fish attached by their discs in an inverted
position to the mantle tissue of the left valve.
Fish have been observed in this position approx-
imately 20 times, either by viewing through the
excurrent or incurrent opening or picking
the scallop out of the water and looking in as it
clapped. Often several fish were observed in the
same scallop simultaneously. This position in the
scallop is the same as that preferred by fish at-
tached to clapper shells and other smooth sub-
strates. In approximately 100 other instances, L.
inquilinus presence in sea scallops was confirmed
by their absence elsewhere in the aquaria.

Liparis inquilinus and U. chuss apparently
cooccur in sea scallops frequently and in consid-
erable numbers. We have collected these fishes
together in sea scallops from Georges Bank in
September, November, and December. In the
Mid- Atlantic Bight (4 August 1969, lat. 39°40'N,

long. 73°09'W, 40 m) a 141-mm sea scallop con-
tained a red hake (21 mm TL) and 21 L. inquili-
nus which averaged 16.5 mm TL. A 125-mm scal-
lop yielded two U. chuss (43 and 47 mm TL) and
two L. inquilinus (23 and 24 mm TL). Goode
(1884) also reportedL. lineatus (= inquilinus) and
Phycis (= Urophycis) chuss as companions in sea
scallops. These two fishes may not be in direct
competition for this particular habitat since the
L. inquilinus remain attached to the upper sur-
face of the cavity and U. chuss swims in the mid-
dle of or rests on the bottom of the cavity (Musick
1969).

Sea scallops apparently suffer no ill effects
from the association with L. inquilinus. Of sev-
eral thousand host sea scallops opened during
this study, none had noticeable internal damage
which could have been caused by L. inquilinus.
These partners do not compete for food since L.
inquilinus feeds principally on larger crustaceans
and sea scallops are microplanktonic filter feed-
ers (Bourne 1964).

Diel Rhythm in the Fish —
Scallop Association

Juvenile L. inquilinus exhibit a diel rhythm in
their association with sea scallops. In aquaria,
fish were outside of the sea scallops and actively
swimming during periods of darkness. The color
pattern of the fish faded during dark periods but
returned within approximately 5 min after the
lights were turned on. Fish were usually inside of
scallops or attached to some substrate in the
aquarium during light periods. When the lights
went off on their regular cycle, the fish would
often leave the scallops and become active within
5-10 min. These reactions to light and dark were
immediate even when the dark-light cycle was
changed drastically during a single day. Liparis
inquilinus which were collected from sea scallops
during a 24-h period on 4-5 August 1969 near
Hudson Canyon (Figure 6) exhibited the same
pattern (Figure 7). During this period, 3,595 L.
inquilinus, averaging 21.0 mm TL, were collected
from 616 of the 841 scallops examined. In one in-
stance, 32 fish were found inside a 139-mm scal-
lop. Fish were more abundant in scallops and more
scallops contained fish during the day than at
night (Table 2). However, some fish were present
in scallops during every sampling period. The
greatest increase and decrease in the number of
fish per scallop occurred around sunrise and sun-

416

ABLE and MUSICK: LIFE HISTORY AND BEHAVIOR OF LIPARIS INQUILINUS

set respectively. The number offish per scallop was
high during the day (Figure 7) and declined sub-
stantially in the first sample after sunset. After
the initial decrease in the numbers of fish per
scallop after sunset, the number increased regu-
larly up to daytime levels as sunrise approached.
The number offish in scallops was slightly greater
than presented in Figure 7. Fish found outside of
scallops (122 or 3% of the total) in the collecting
buckets or on the deck were not included in the
averages. However, these fish were more abun-
dant at stations where the number offish per scal-
lop was greater so that they did not affect the
comparative data.

7> -^

BANK .•• ••"

•:•>:.••

HUDSON
CANTON

• ALBftTROSS n 68-14
o ALBATROSS IX 69-8

Figure 6. — Locations of sampling sites for fish-scallop asso-
ciation on 5-17 September 1968, Albatross IV cruise 68-14
and on 4-5 August 1969, Albatross IV cruise 69-8.

The majority of L. inquilinus leave scallops to
feed during the night and then return near sun-
rise or as they become satiated. Sixty stomachs
were examined (five from each sampling period)
and were assigned a separate value for relative
fullness (0-4) and state of digestion of contents
(1-3) with the highest numbers given to stomachs
with the most food and the least degree of diges-
tion. When added together, these give a relative
value referred to as the stomach analysis index.
The maximum value possible is 7, the minimum
is 1. The higher values should be from fish which
had recently fed, and digestion had not begun
or had not progressed very far. The stomach
analysis index values increased from 2200 h, with
highest values occurring just before and after
sunrise (Figure 8). The lowest values were found
just before and after sunset (Figure 8). Whole un-
digested amphipods were found in stomachs of
fish taken at night, but after 0800 h stomach con-
tents were in increasingly advanced stages of

1.50-1

102

T 5!
45 7

-1.25-

-J
_1

X <

S ^ 1.00 -

n - ^

5
90

3 56

LaJ
Q-
O

i5 .75-

X ::

-- r

84 T

LOGARIl

MBER OF

r" -

-T

■-

1 1

5 .25-

[■

^ -■

0600 'OSOO' 1000 'l200 ' l«Oo' 16O0' ISC

TIME IN HOURS

O'ZOOO'ZZOO '2400'0200'0400'

SUNSET

SUNRISE

Figure 7. — Number ofLiparis inquilinus per scallop from the
combined total of two 10-min tows taken every 2 h over a
24-h period on 4-5 August 1969 at approximately lat. 39°39'N,
long. 73°08'W. For each sample, the range is represented by the
vertical line, mean by the horizontal line, one standard devia-
tion on each side of the mean by hollow rectangles and two
standard errors on each side of the mean by solid rectangles.
Numbers above each figure represent the number of scallops
sampled.

Table 2. — Comparison of the number ofLiparis inquilinus in
sea scallops during the day and night for a 24-h period.

Number of

Mean number

Percent of

Number of

scallops

of fish

scallops

replicated

Time

examined

per scallop

with fish

stations

Day

489

6.1

86.3

(0503-1 908 h)

Night

352

1.7

57.3

(1909-0502 h)

Total

841

4.2

73.2

0600 0800 tOOO 1200

I600 1800 2000 22 00 2400 0200 0400

Figure 8.— Results of stomach analysis of Liparis inquilinus
taken from scallops over a 24-h period on 4-5 August 1969.
Stomach analysis index value for each stomach was derived
from ranking relative fullness (0-4) which is added to the state
of digestion of the contents (1-3), with the highest numbers
given to stomachs having the most food and the least degree
of digestion.

417

FISHERY BULLETIN: VOL. 74, NO. 2

digestion until only unidentifiable material re-
mained in stomachs collected just before and after
sunset. Fish with full stomachs and undigested
contents were first collected at 2200 and 2400 h.
These were probably returning to scallops as they
became satiated. All fish do not leave the scallops
at sunset (Figure 7). Some may remain if they
still have food in their stomachs. Those fish
examined around 2000 h did not have completely
empty stomachs (Figure 8). None of the fish
examined at 0200 and 0400 h had empty
stomachs.

The number of L. inquilinus occupying sea
scallops probably decreases through the fall and
early winter. During September 1968, 43 collec-
tions near Hudson Canyon (Figure 6), which
overlapped the collecting area in August 1969
(Figure 6), yielded fewer fish per scallop (Table 3)
than in August. These differences could be due to
relative year-class strength or may reflect an ac-
tual change in the number of fish occupying scal-
lops later in the year. Mortality of L. inquilinus
owing to predation or a breakdown in the associa-
tion as the fish grow larger could explain a de-
crease of this magnitude. Small numbers of sea
scallops collected during the fall and early winter
of several years did not yield as many L. in-
quilinus as were collected earlier in the year.

Size of individual sea scallops may be a factor
in their selection by fish. In one instance, a
60-mm scallop contained a 21-mm TL fish, but it
is the larger scallops which contain the largest
number of fish (Figure 9).

Table 3. — Abundance and average total length of Liparis in-
quilinus in sea scallops from August 1969 and September 1968.

Collecting
dates

Number of
scallops
examined

Mean number

of fish

per scallop

Maximum Average

number in TL of fisfi
single scallop (mm)

4-5 Aug. 1969
14-17 Sept. 1968

841
717

4.2
1.7

32
18

21 0
26.1

.. . • : }...

. . :. : I . :. .. . 1...1.
. .■•• J-

. .. . !.:.». .1! t .. .

. .. \:' .:■..!,• , '.=!i*i:-'.l .• . •

>:•■• '•• -X- :•= • + .

. I ij.i ji..' I..:: I. 1.; .

90 100 110 120 130

SC4L1.0P HEIGHT (mm)

Figure 9. — Plot of mean number of fish per scallop versus
scallop height (mm) from daytime collections from Albatross
rV cruise 69-8.

<
o

UJ
Q.

Hi
CD

4 -

1!"

STA 1-149
N. a N.E.
GEORGES BANK

STA 150-206
SOUTHERN
GEORGES

STA 227-314

NEAR
HUDSON CANYON

Geographic Variation in Abundance
of Fish in Scallops

Figure 10. — Plot of mean number of fish per scallop at differ-
ent localities from collections of 5-17 September 1968.

The abundance of fish in scallops varies with
geographic location (Figure 10). On Albatross IV
cruise 68-14, 155 10-min scallop dredge tows were
made as part of a sea scallop survey on Georges
Bank and in the Mid- Atlantic Bight near Hudson
Canyon (Figure 6). From these, 2,274 L. in-
quilinus were collected fi"om 1,228 of the 5,905
sea scallops examined. The mean number of fish
per scallop (Figure 10) and the mean number of

fish per station (Table 4) were highest north of
Hudson Canyon, lowest on the north and north-
west edges of Georges Bank, and intermediate on
southern Georges Bank. Although the greatest
abundance of fish in sea scallops occurred near
Hudson Canyon, where the average depth and
bottom temperature were lowest (Table 4), these
parameters did not seem to be related to abun-
dance in this area (Figures 11, 12). The average

418

ABLE and MUSICK: LIFE HISTORY AND BEHAVIOR OF LIPARIS INQUILINUS

Table 4. — Comparison of the possible parameters affecting 7

Liparis inquilinus abundance in sea scallops over a wide geo-
graphic area. Given as mean followed by range in parentheses.

Item

Northern and

northeastern

Georges Bank

Southern
Georges Bank

Near Hudson
Canyon

No. of stations

Date, 1968
No of fish per

5-10 Sept.

10-12 Sept.

14-17 Sept.

scallop
Scallops with

0.11(0.0-1.1)

0.65(0.0-1.8)

1.74(0.0-6.1)

fish (%)
No. of fish per

10.1(0.0-90.9)

41.0(0.0-78.9)

59.2(0.0-100.0)

station
TL (mm) of fish

4.8(0-31)

20.6(0-64)

27.1(0-82)

in scallops
Depth (m)
Bottom temp ('C)
Clapper shells

29.4(14-47)

77(47-95)

10.0(4.0-14.5)

28.2(16-43)
77(62-90)
9.7(8.2-13.3)

26.1(17-38)
57(37-77)
7.8(6.5-10.1)

(bushels)
No. scallops >60

3.4(1-7)

5.8(2-9)

4.0(2-8)

mm per station

152(3-456)

38(10-79)

68(3-311)

50 55 60 65

DEPTH (METERS)

Figure ll. — Plot of mean number of fish per scallop versus
water depth from daytime collections of 14-17 September
1968 near Hudson Canyon.

number of scallops per station for each area was
not related to increased abundance of fish in scal-
lops (Table 4). Clapper shell abundance, regard-
less of species, was originally hypothesized to be
important in L. inquilinus survival and abun-
dance since L. inquilinus readily occupied shells
in the laboratory, and this habit may offer protec-
tion from predators. A plot of this possible rela-
tionship did not suggest a correlation (Figure 13).
The similarity of abundance estimates for south-
ern Georges Bank and the area near Hudson
Canyon could be attributed to a similarity in bot-
tom types. Both of these areas have smooth bot-
toms and are quite different from the rough topog-
raphy of northern Georges Bank (Uchupi 1968).
Fish living on smooth bottom would have less
chance of concealment and evasion of predators,

'^ p.

<
o

CO
LU

^ 4

UJ
CD

— r-
6

— r-
7

— I—
8

— I—
9

BOTTOM TEMPERATURE (°C)

Figure 12. — Plot of mean number of fish per scallop versus
bottom temperature from daytime collections of 14-17 Septem-
ber 1968 near Hudson Cemyon.

a.
o

I
CO 4

O
cr

UJ
CD

I-

-I 1 1 1 1 1 1 '

I 2 3 4 5 6 7 8

BUSHELS OF SHELLS PER DREDGE CATCH

Figure 13. — Plot of mean number of fish per scallop from
daytime collections versus clapper shell abundance from
collections of 14-17 September 1968 near Hudson Canyon.

which would place a greater selective advantage
on association with scallops. The simplest expla-
nations for observed differences in abundance are

419

FISHERY BULLETIN: VOL. 74, NO. 2

differences in the year-class strength and differ-
ences in actual abundance among different L. in-
quilinus populations.

Possible Advantages of
the Association

Liparis inquilinus probably is protected from
predation by its association with sea scallops. The
only known predators of larger sea scallops which
might also ingest L. inquilinus are Atlantic
wolfish, Anarhichas lupus, and Atlantic cod,
Gadus morhua (Bourne 1964). Wolfish and cod
only feed occasionally on scallops and they are
rare or only winter inhabitants of the Mid-
Atlantic Bight. Also, L. inquilinus is not as-
sociated with scallops during most of the winter.

Individuals of L. inquilinus maximize the
period of protection by associating with sea scal-
lops for most of their demersal life. In the Mid-
Atlantic Bight, L. inquilinus remains associated
with sea scallops from the time they leave the
plankton until they begin to move inshore to
spawn. Also, individuals only leave sea scallops
to feed and then return as soon as they become
satiated. Nocturnal feeding may also decrease the
possibility of detection by predators.

The relative number of scallops may not be a
limiting factor for survival of juvenile L. in-
quilinus. In every sample, at any time of the year
in which L. inquilinus have been taken with sea
scallops, some scallops were always empty. How-
ever, this assumes that all sea scallops will accept
fish. This remains to be proven.

The symbiosis between L. inquilinus and P.
magellanicus should be referred to as a commen-
sal association. Such an association is one in
which the population of the commensal benefits
and the host is unaffected (Odum 1971).

ACKNOWLEDGMENTS

We acknowledge the assistance of the following
for loan of specimens in their care: James E.
Bohlke, Academy of Natural Sciences of
Philadelphia; Roland Wigley and Henry Jensen,
NMFS, Woods Hole; David Thomas and Tom
Tatham, Ichthyological Associates, Middletown,
Del.; James Hoff, Southeastern Massachusetts
University; and for larval material, W. G. Smith,
NMFS, Sandy Hook; Joanne Laroche, Ira C. Dar-
ling Center, Walpole, Maine; Herbert Perkins and
Stanley Chenoweth, NMFS, Boothbay Harbor,

Maine; Thomas Morris, NMFS, Narragansett,
R.I. Many people at NMFS, Woods Hole, provided
aid and facilities for our research, especially
Marvin Grosslein and those who participated in
Albatross IV cruises 68-14 and 69-8. James Hoff
provided information onL. inquilinus in bay scal-
lops and Tom Tatham shared his notes on colora-
tion and development of L. inquilinus eggs col-
lected off New Jersey. We also express apprecia-
tion to the following personnel from the Virginia
Institute of Marine Science: Frank Perkins and
his assistants, especially Patricia Berry, for sec-
tioning and staining the pectoral fins; Juanita
Tutt and her assistants for supplying algal cul-
tures; Michael Castagna for supplying bay scal-
lops; and Charles Barans, Labbish Chao, John
McEachran, Sally Leonard, James Weaver, and
Charles Wenner for collecting Liparis on various
cruises.

LITERATURE CITED

ABLE, K. W.

1973. A new cyclopterid fish, Liparis inquilinus, associated
with the sea scallop, Placopecten magellanicus, in the
western North Atlantic, with notes on the Liparis liparis
complex. Copeia 1973:787-794.

1974. Life history, ecology and behavior of two new Liparis
(Pisces: Cyclop teridae) from the western North Atlan-
tic. Ph.D. Thesis, College of William and Mary, Wil-
liamsburg, 115 p.

In press. A new cyclopterid fish from the western North
Atlantic with notes on life history. Copeia.
ALEXANDER, R. M.

1967. Functional design in fishes. Hutchinson, Lond.,
160 p.
ANDRIYASHEV, A. P.

1954. Fishes of the northern seas of the USSR. Akad.
Nauk SSSR, Zool. Inst., Opredeliteli po Faune SSSR 53,
566 p. (Translated by Israel Program Sci. Transl., 1964,
617 p.; available U.S. Dep. Commer., Natl. Tech. Inf.
Serv., Springfield, Va., as OTS63- 11160.)
BARDACH, J. E., AND J. CASE.

1965. Sensory capabilities of the modified fins of squirrel
hake (Urophycis chuss) and searobins (Prionotus
carolinus and P. evolans). Copeia 1965:194-206.

BEAN, T. H.

1884. List of fishes collected by the U.S. Fish Commission
at Wood's Holl, Massachusetts, during the summer of
1881. Rep. U.S. Comm. Fish Fish. 1882:339-344.
BIGELOW, H. B., AND W. C. SCHROEDER.

1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv.,
Fish. Bull. 53, 577 p.
BOURNE, N.

1964. Scallops and the offshore fishery of the Maritimes.
Fish. Res. Board Can., Bull. 145, 60 p.
BREDER, C. M., JR, AND D. E. ROSEN.

1966. Modes of reproduction in fishes. Nat. Hist. Press,
Garden City, N.Y., 941 p.

420

ABLE and MUSICK: LIFE HISTORY AND BEHAVIOR OF LIPARIS INQUIUNUS

Burke, V.

1930. Revision of the fishes of the family Liparidae. U.S.
Natl. Mus. Bull. 150, 204 p.

Clark, J., W. G. Smith, A. W. Kendall, Jr., and M. P.
Fahay.

1969. Studies of estuarine dependence of Atlantic coastal
fishes. Data Report I: Northern section, Cape Cod to Cape
Lookout. R.V. Dolphin cruises 1965-66: Zooplankton vol-
umes, midwater trawl collections, temperatures and
sahnities. U.S. Bur. Sport Fish. Wildl., Tech. Pap. 28,
132 p.
COWAN, D.

1929. Spawning of the lumpsucker. Rep. Dove Mar. Lab.,
Cullercoats, Northumberland 17:37-38.
DAVENPORT, D.

1955. Specificity and behavior in symbioses. Q. Rev. Biol.
30:29-46.

DETWYLER, R.

1963. Some aspects of the biology of the seasnail, Liparis
atlanticus (Jordan and Evermann). Ph.D. Thesis, Univ.
New Hampshire, Durham, 103 p.

EHRENBAUM, E.

1905. Eier und Larven von Fischen. In Nordisches
Plankton; Zoologischer Teil, Erster Band, p.
1-216. Verlag von Lipsius & Tischer, Kiel und Leipz.

FREIHOFER, W. C.

1963. Patterns of the ramus lateralis accessorius and their
systematic significance in teleostean fishes. Stanford
Ichthyol. Bull. 8:80-189.

GARMAN, S.

1892. The Discoboli. Cyclopteridae, Liparopsidae, and
Liparididae. Mem. Mus. Comp. Zool. 14(2), 96 p.

Gibson, R. N.

1972. The vertical distribution and feeding relationships of
intertidal fish on the Atlantic coast of France. J. Anim.
Ecol. 41:189-207.

GOODE, G. B.

1884. Natural history of useful aquatic animals. Part III,
The food fishes of the United States. In The fisheries and
fishery industries of the United States, Sec. 1:163-682.

GOODE, G. B., AND T H. Bean.

1895. Oceanic ichthyology. U.S. Natl. Mus., Spec. Bull. 2,
553 p.
HOFF, J. G.

1968. Occurrence of Liparis atlanticus in Aequipecten irra-
dians in Buzzards Bay, Massachusetts. Trans. Am. Fish.
Soc. 97:278.

JOHNSON, C. R.

1969. Contributions to the biology of the showy snailfish,
Liparis pulchellus (Liparidae). Copeia 1969:830-835.

1970. Age of the showy snailfish, LJpartspu^cAeZZus. Jap.
J. Ichthyol. 17:121-123.

JORDAN, D. S., AND B. W. EVERMANN.

1898. The fishes of North and Middle America. U.S. Natl
Mus. Bull. 47, Part 2:1241-2183.
KOSAKA, M.

1971. On the ecological niche of the sea-snails, Liparis
tanakai (Gilbert et Burke) in Sendai Bay. J. Coll. Mar.
Sci. Technol., Tokai Univ. 5:27-41.

Leim, a. H., and w. b. Scott.

1966. Fishes of the Atlantic coast of Canada. Fish. Res.
Board Can., Bull. 155, 485 p.
MARISCAL, R. N.

1966. The symbiosis between tropical sea anemones and
fishes: A review. In R. I. Bowman (editor), The
Galapagos, p. 157-171. Univ. Calif Press, Berkeley.
MUSICK, J. A.

1969. The comparative biology of two American Atlantic
hakes, Urophycis chuss and U. tenuis (Pisces: Gad-
idae). Ph.D. Thesis, Harvard Univ., Cambridge, 150 p.

Nizortsev, G. p., V. P. Ponomarenko, and R. a. Shash-

LOVA.

1963. New data on the representatives of Liparis genus in
the Barents Sea. (In Russ.) Zool. Zh. 42:1415-1418.
ODUM, E. P.

1971. Fundamentals of ecology. 3rd ed. W. B. Saunders
Co., Phila., 574 p.
QUAST, J. C.

1968. Observations on the food of the kelp-bed fishes. In
W. J. North and C. L. Hubbs (editors). Utilization of
kelp-bed resources in southern California, p. 109-
142. Calif Dep. Fish Game, Fish Bull. 139.
UCHUPI, E.

1968. Atlantic continental shelf and slope of the United
States — Physiography. U.S. Geol. Surv., Prof. Pap.
529-C, 30 p.

Welsh, w. w.

1915. Note on the habits of the yoimg of the squirrel hake
and sea snail. Copeia 18:2-3.
Wheeler, a.

1969. The fishes of the British Isles and north-west
Europe. Macmillan Co., Lond., 613 p.

WILEY, M. L., AND B. B. COLLETTE.

1970. Breeding tubercules and contact organs in fishes:
Their occurrence, structure, and significance. Bull. Am.
Mus. Nat. Hist. 143:147-216.

421

FURTHER OBSERVATIONS OF THE FEEDING ECOLOGY

OF POSTLARVAL FINFISU, LAGODON RHOMBOIDES ,

AND SPOT, LEIOSTOMUS XANTHURUS^

Martin A. Kjelson and George N. Johnson^

ABSTRACT

The effect of current on feeding, temporal variation in food consumption, and the effect of predator and
prey size on food preferences were evaluated for postlarval stages of pinfish, Lagodon rhomboides
(15-19 mm total length); and spot, Leiostomus xanthurus (16-22 mm). Field and laboratory observations
indicated that pinfish feeding rates decreased as water current velocity increased. Similar behavior
was noted in spot from field observations, but spot feeding rates in the laboratory were highest when a
slight current was present. Mean gut contents of postlarvae collected at midday over a 2-mo period
ranged from 0.4 to 38 copepods/fish. The mean coefficient of variation for the number of copepods per
fish in a single midday sample in = 20 fish ) was 20% . Maximum daily feeding rates were estimated at 17
and 26 copepods/h for spot and pinfish, respectively. Field and laboratory data confirmed that as
postlarvad size increases the size of their prey also increases. Refined estimates of postlarval evacuation
rates and daily rations also are presented. Daily ration estimates as a percent of the fish's wet body
weight were 99c for both species. The ration estimates for both species were greater than metabolic
needs estimated from oxygen consumption measurements.

Information on the feeding ecology of larval fishes
is necessary to understand the role of larvae in
ecosystem energetics and community structure
and the importance of feeding conditions to year
class strength. However, relatively little is
known about the feeding of larval fishes. This
paper reports four major aspects of postlarval
feeding: 1) the effect of current speed on feeding
intensity; 2) temporal variation in postlarval food
consumption; 3) the relation of feeding rate to
food abundance; and 4) the effect of prey and
predator size on postlarval food preferences.
Refined results concerning postlarval evacuation
rates and daily rations also are presented. Our
earlier paper (Kjelson et al. 1975) stressed the
study of food preferences, feeding intensity and
periodicity, evacuation rates, daily rations, and
the effect of handling and capturing the fishes on
their digestive tract contents.

Pinfish, Lagodon rhomboides , and spot, Leios-
tomus xanthurus, constitute a major portion of
the fish biomass of southeastern estuaries of the
Atlantic coast and thus are important to the
structure and function of these ecosystems. Spot
are also an important commercial food species.

'This research was supported under agreement AT (49-7 )-5
between the National Marine Fisheries Service, NOAA, and the
U.S. Energy Research and Development Administration.

^Atlantic Estuarine Fisheries Center, National Marine
Fisheries Service, NOAA, Beaufort, NC 28516.

Both species are primarily winter spawners in
the Atlantic Ocean with larvae migrating inshore
to estuarine waters which serve as nursery
grounds between spring and fall. Larval forms
(here defined as individuals <11 mm) are rarely
found within the estuaries, whereas postlarval
stages (here defined as fish between 11 and 22
mm) occur both in nearshore oceanic and es-
tuarine waters.

METHODS

General

Postlarval pinfish (15-19 mm total length (TD)
and spot (16-22 mm) were collected during
January and February 1974, from the Newport
River estuary, N.C., following their recent im-
migration into the estuary from the offshore
spawning grounds in the Atlantic Ocean. All fish
were collected at Pivers Island, 2.5 km inside the
Beaufort Inlet. Shore samples were collected with
dip nets while those in the adjacent channel were
collected with a channel net (Lewis et al. 1970).
Fish were anesthetized immediately upon cap-
ture in a 0.12 g/liter sea water solution of MS-2223
(tricaine methanesulfonate) and dissected in the

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Manuscript accepted October 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

423

FISHERY BULLETIN: VOL. 74, NO. 2

laboratory. Gut contents included material in the
total digestive tract, from foregut to anus.

Current Speed and Feeding Intensity

Larval pinfish and spot were collected within 2
m of the shore (depth 0-2 m) where refuge from
current was available and in the center of the
adjacent channel (depth 5-7 m) approximately 25
m from shore where a current normally was pres-
ent. Three separate collections were made for
pinfish and two for spot. Twenty fish of each
species from each collection were measured, dis-
sected, and the mean number of copepods per fish
determined. Surface tows for zooplankton were
made at the same time and location using a 30 cm
in diameter, 0.158-mm mesh net with current
meter attached. Observations on copepods
throughout this investigation were restricted to
adult and copepodid stages. Copepod measure-
ments were made of carapace length. Current
velocities were measured with a Gurley current
meter.

Feeding rates of pinfish and spot at varied cur-
rent speeds also were studied in the laboratory.
Fish were captured, placed in four donut-shaped,
11-liter tanks (46 cm in diameter, 10- by 10-cm
cross-sectional area), and allowed to acclimate
overnight in filtered, food-free seawater with no
current flow. Two tanks were used as controls
(zero current flow) and contained 50 and 100 fish,
respectively. The other two tanks, containing 50
fish each, were attached to pumps, providing cur-
rent velocities of 1.7 and 5.1 m/s, respectively.
Current speed was estimated by recording the
amount of time required for a minute innate par-
ticle to complete one revolution of the donut-
shaped tank. At the beginning of each experi-
ment, current flow was started in the two test
chambers and Artemia salina nauplii (1.0/ml)
were provided to each of the four tanks. Fish were
allowed to feed for 1 h with additional food pro-
vided after 30 min to assume a minimum density
of 1 Artemia/ml throughout the experiment.
Twenty fish were sampled from each tank to cal-
culate the mean number of Artemia consumed.

Temporal Variation in
Midday Feeding

Day-to-day variation in the feeding intensity of
larval pinfish and spot was studied at midday
(1100-1300 h) when larval digestive tracts con-

424

tained the greatest amounts of food. Fifl;een col-
lections were made from 21 January to 28 Feb-
ruary at one site within 2 m of the shore. Each
collection consisted of 20 fish of each species.
Total lengths of the fish were measured, the total
number of copepods in each gut counted, and a
geometric mean of the number of copepods per
fish calculated. Geometric means were used as a
measure of central tendency because frequency
distributions of the copepods or Artemia nauplii
per fish showed a positive skewness. In addition,
a geometric mean was used to limit the bias of a
few individuals feeding at a rate not representa-
tive of the population because variation increased
as the mean values increased.

A zooplankton tow was taken at the time and
location of fish capture. The tows were made just
below the surface, against the current, and sam-
pled approximately 5 m^ of water. Estimates of
copepod density were made from three 10-ml sub-
samples of each tow. Twenty copepods per sample
were measured for length frequencies.

Evacuation Rates

To refine our information on larval evacuation
rates of copepods, two laboratory experiments
were performed using pinfish and spot that had
been fed an abundance of natural copepods. Four
to five hundred fish were starved for 8 to 12 h and
then they were allowed to feed for 1 h. Food densi-
ties averaged 2.5 copepods/ml for pinfish and 3.0
copepods/ml for spot. Larvae were acclimated and
experiments run at ambient estuarine tempera-
tures and salinities. Temperature was 12°C for
the pinfish evacuation and 17°C for spot; salinity
was 301.. Following feeding, 30 fish were re-
moved, anesthetized with MS-222 to prevent any
possible regurgitation, dissected, and counts
made of the numbers of copepods per fish. At the
same time, three groups of 100 fish were trans-
ferred to separate food-fi*ee tanks, and the de-
crease in their gut contents observed by sampling
10 fish from each tank at 2-h intervals until more
than one-half of the fish had empty tracts. Instan-
taneous evacuation rates were then calculated
according to the method of Peters and Kjelson
(1975). The amount of food remaining in the
stomach at any time can be predicted from the
following equation:

log C = log A + Bt

where C = content of gastrointestinal tract -i- 1

KJELSON and JOHNSON: FEEDING ECOLOGY OF POSTLARVAL PINFISH AND SPOT

A = amount ingested + 1
B = evacuation rate constant
t = time.

By adding 1 to the amount ingested and to gut
contents we were able to include empty gastroin-
testinal tracts in our calculations. From the
above equation, with log base 10:

2.303 (log A + Bt)

and the instantaneous evacuation rate

dC 2.303 (log A + Bt)

— = 2.303 Be
dt

dC
dt

2.303 fiC.

Feeding Periodicity

Diel periodicity of digestive tract contents indi-
cated the intensity and chronology of feeding by
the fish. Our purpose was to refine the feeding
chronology curve (Kjelson et al. 1975) by taking
samples more frequently than in our previous
study. Ten fish of each species were collected at 2-h
intervals between 0600 and 1800 and at 2100 and
2400 h. Fewer samples were taken at night be-
cause our past observations have shown that lar-
val fish cease feeding during darkness. All fish
were measured, the copepods they contained
counted, and a geometric mean for copepods per
fish calculated for each sample.

Daily Rations

One objective of this research was to re-estimate
the daily ration of larval fish for comparison with
our earlier study. Daily rations were calculated by
the same technique (Kjelson et al. 1975 ) using new
information on diel periodicity of gut contents and
refined measurements of instantaneous evacua-
tion rates. Our method of calculating daily ration
accounts for changes in evacuation rate which ac-
company diel changes in feeding intensity.

To calculate daily ration, we first estimated the
average evacuation rates (in copepods per hour)
for each of the 2-, 3-, or 6-h sampling periods in our
feeding chronology study. This average rate was
the geometric mean of the instantaneous evacua-

tion rates at the beginning and end of each period.
The estimate of food evacuated during any period
is equal to the number of hours in the period mul-
tiplied by the respective average hourly evacua-
tion rate. The total food evacuated per day was
computed by summing the nine respective evacua-
tion estimates, and is an estimate of the daily
ration because the average ingestion rate must
equal the rate at which material in the gut is
assimilated or defecated.

Daily rations were calculated initially as
copepods per fish per day and then transformed to
percent of the larval body weight and calories per
fish per day. Dry weights of ingested copepods
were estimated from the length-weight relation-
ship: W = 6.274L - 0.153 where W is the dry
weight in micrograms andL is the copepod length
in millimeters, based upon Heinle's (1966) data for
all stages oi Acartia tonsa. Copepod dry weights
were converted to wet weights using a factor of 9.1
based upon our measurements of the wet-dry ratio
for zooplankton and were compared with wet
weights of the fish to compute the daily ration as a
percent of live body weight. Daily caloric intake
was computed using our estimation of 0.555 cal/
mg wet weight of an average size copepod, based
on micro-bomb calorimeter measurements of
mixed estuarine zooplankton (Thayer et al. 1974).

RESULTS AND DISCUSSION

Effects of Current Speed on
Feeding Intensity

Pinfish and spot larvae collected along the shore
had more copepods present in their digestive
tracts than those collected in midchannel (Table
1). Previous observations (Kjelson et al. 1975) in-
dicated that neither pinfish nor spot regurgitate or
defecate food under the stress of capture or han-
dling. Thus, differences in collecting techniques

Table l. — Digestive tract contents of larval fishes collected
at midday at midchannel and shore stations in the Newport
River estuary, January to February 1974.

Ivlean number of

Current

copepods/fish ± 1 SE

speed

(m/s) in

Tidal

Date

Species

Shore

Channel

channel'

stage^

29 Jan.

Pinfish

9.1 ± 1.6

0.8 ± 0.2

1.4

30 Jan.

Pinfish

19.6 ± 2.8

4.1± 0.8

0.0

14 Feb.

Pinfish

20.0 ± 2.8

1.8 ± 0.6

3.2

14 Feb.

Spot

14.4 ± 1.7

0.9 ± 0.4

3.2

21 Feb,

Spot

2.7 ± 0.6

0.3 ± 0.2

5.5

'No current was observed along the shore on any sannple date.
2LF = late flood. IVIE = nnid ebb, HS = high slack.

425

(channel net versus dip net) between areas were
not felt to bias the results. We observed no differ-
ences in the length of fish sampled (by species) or
in the density of copepods at the two locations.

These results indicate that larval feeding rates
are limited when the fish are exposed to current.
Current speed ranged from 0 to 5.5 m/s in mid-
channel where feeding was low, to no measur-
able current along the shore where more feeding
occurred (Table 1). In addition, pinfish collected
in the channel on a slack tide contained 4.1 cope-
pods/fish compared with a mean of 1.3 copepods/
fish when there was a current present.

Laboratory experiments indicated that current
speed affected the food consumption rate of both
species (Table 2). Pinfish consumed the most food
when there was no current, but spot ate more at a
current velocity of 1.7 m/s. Pinfish ate the least
food in a 5.1 m/s current while the spot minimum
feeding occurred at varied current speeds. Both
species fed at a higher rate when fish densities
were lower.

The observations from both field and laboratory
studies indicate that postlarval pinfish feeding de-
clines as current speed increases. These results
suggest that current speed influences the ability of
pinfish to capture their prey, although the specific
reasons for such altered behavior are unknown.
The well-known attack behavior of larval fish,
that of visually sighting the prey and of assuming
an S-shape prior to striking (Blaxter and Holliday
1963), may be unattainable by postlarval pinfish
exposed to higher current speeds. Bishai (1959)
found that larval herring drift with a current at
speeds less than the current itself. This may
suggest that the size, shape, and behavior of a
plankter may influence its rate of movement in a
current. Prey organisms may move at a faster rate
than the fish larvae, which in turn may lessen the
ability of the fish to orient to the prey.

Current also may destroy the microstructure of
the prey population. Without a strong current,
food could aggregate in patches thus producing
local areas with high food density and therefore

FISHERY BULLETIN: VOL. 74, NO. 2

increase the rate of ingestion. This latter explana-
tion is probable in the natural environment; how-
ever, it appears unlikely under the laboratory
conditions, because the density of Artemia in the
tanks was very high (1/ml) and prey were re-
plenished to assure that it did not decrease. In
addition, the small cross-sectional area (100 cm^)
and volume of the tanks greatly limited the dis-
tance a larva had to travel to find food even if prey
were in a patch configuration.

Differences in channel versus shoreline feeding
by spot in the natural environment (Table 1) are
similar to those of pinfish; however, feeding by
spot in the laboratory was highly variable and is
difficult to explain. Spot fed at the highest rates
when a slight current was present and even fed at
a high rate when exposed to a maximum current of
5.1 m/s. The spot postlarvae used in the studies
were larger than the pinfish and this may explain
the ability of spot to feed at a high rate when
exposed to current, because increased size usually
improves swimming ability which may improve
the fish's ability to capture their prey. However,
species differences in swimming ability were not
apparent: larvae of both species moved freely
about the tank when current was absent; oriented
into the current or at times drifted with the cur-
rent at the 1.7 m/s speed; and drifted along with
the current in the 5.1 m/s current, although some
individuals oriented into the current briefly. Simi-
lar behavior by larval fishes exposed to varied
current velocities was discussed by Bishai (1959)
and Houde (1969). Ryland (1963) indicated that
the mechanisms by which larval fishes orient to a
current are poorly understood. The lower feeding
rate of spot in no current is unexplainable unless
this species is adapted in some way to be more
effective at capturing prey within a current.
Serebrov (1973) also found differences in the feed-
ing intensity of various species (guppy, Poecilia
reticulata, and European dace, Phoxinus phoxi-
nus) when exposed to different current velocities
and suggested that the differences were due both
to natural adaptation to certain current condi-

TaBLE 2. — Digestive tract contents (mean number oiArtemia nauplii/larva ± 1 SE) of
larval fishes following feeding in the laboratory under several current velocities.

Species

Length
of fish
(mm)

Current

velocity

Date
1974

'5.1 m/s

'1.7 m/s

No
current'

current^

26 Feb.

3 Mar.

14 Mar.

Pinfish

Spot

16-17
20-22
19-20

2.5±1.8

104.7 ±7.5

30.3 ±3.3

22.1 ±3.5

166.0±8.0

90.0 ±5.5

50.9 ±6.3
72.1 ±7.7
39.1 ±4.3

35.0 ±4.0
23.5 ±5.3

23.1 ±3.2

'Fifty fish.

^One hundred fish.

426

KJELSON and JOHNSON: FEEDING ECOLOGY OF POSTLARVAL PINFISH AND SPOT

tions and to the stimulation of food grasping
activity caused by the increased movement of
food in a current.

The highly variable nature of spot feeding in the
laboratory also may be explained by the varied
current conditions within the tank itself, although
conditions were kept as constant as possible dur-
ing the two studies. Current flow may not have
been uniform throughout the tank, although the
importance of this factor upon feeding is unknown.
The larvae in all experiments were distributed
throughout the tank and did not appear to be feed-
ing at specific locations. The low variability in
feeding rate between individual fish in each exper-
iment, as shown by the standard errors (Table 2),
suggests that all individuals were feeding at a
similar rate even though they were dispersed
throughout the tank. The distribution of flow
across the tank vertically was not measured, al-
though such information would be useful (Ryland
1963 ). The two treatment groups of spot postlarvae
were from separate field collections which may
have altered their behavioral characteristics
sufficiently to produce the variable results.

Finally, the apparent necessity for low current
velocity for feeding to take place may restrict
considerably the amount of area suitable for feed-
ing to be successful. This may be particularly true
along the channels linking the oceanic habitat to
that of the estuarine marsh system where our ob-
servations took place. The amount of protected
shoreline and bottom habitat characterized by low
current velocity along these channels is very lim-
ited compared to that present in the broad
reaches of the estuary where cordgrass (Spartina)
marsh shoreline and eelgrass {Zostera) beds are
extensive.

Temporal Variation in
Food Consumption

Considerable day-to-day variation was observed
in the mean number of copepods in the plankton
and in the larval fish collected at midday (Figure
1). Mean pinfish gut contents ranged from 0.4 to 38
copepods/fish while spot contained from 0.5 to 24
copepods/fish. The coefficient of variation for the
number of copepods per fish in single field samples
averaged 20% (range 7-40%) for pinfish, and 17%
(range 8-40%) for spot. The greatest variability
occurred when the average gut contents were low.
Copepod density also fluctuated widely fi'om 477 to
3,262 copepods/m^. These densities are not dis-

z o

z i

< a.

i 8

■^ " 3 000
O I

I 2 JOOO
z o

i S 'OOO

21 26

JANUARY

IS 20

FEBHUABT

Figure l. — Variation in the numbers of copepods per larval
pinfish and spot, and copepods per cubic meter based on mid-
day samples in the Newport River estuary during January and
February 1974.

similar from those observed during the same
months in the open waters of the Newport River
estuary (Thayer et al. 1974). The coefficient of vari-
ation of the copepod counts from five tows at the
site of larval collections was 24%. Such variation
is not high for field sampling and although it rep-
resents the variability for only a single sampling
date, it does suggest that the precision of the esti-
mate of copepod density is acceptable.

One of our goals was to determine if the amount
of food present in larvae was related to copepod
density. In our study, the correlation coefficients
between copepod concentration and gut contents
were very low (r = +0.08 for both pinfish and spot),
indicating that there was no relationship. Other
studies on larval fish populations have shown that
feeding incidence may be correlated with food con-
centration (Berner 1959; Nakai et al. 1966; Bain-
bridge and Forsyth 1971), while Houde (1967)
found no correlation between copepod abundance
and feeding rate by larval walleye.

The number of factors influencing larval feeding

427

FISHERY BULLETIN: VOL. 74, NO. 2

rates in an estuary are undoubtedly numerous;
therefore, it may be difficult through field mea-
surements to establish a relationship between
larval feeding rates and food abundance. For
example, the clumped distribution typical of zoo-
plankton populations may affect larval feeding
rates, with feeding limited primarily to those
periods when the fish are exposed to a dense patch
of copepods.

Comparing naturally occurring mean food
densities with mean gut contents, to establish a re-
lationship between prey abundance and feeding
rate, presents problems if the zooplankton popula-
tions are not randomly distributed or if the fish
collected were not feeding upon the same prey
community sampled by the plankton net (O'Brien
and Vinyard 1974). Furthermore, the aggregation
of zooplankton discussed earlier may be important
in determining the rate of food consumption
(Schumann 1965). Ivlev (1961) indicates that
patchiness in the distribution of the food material
increases the ration by comparison with an even
food distribution when the average concentration
is the same in both cases. High consumption rates
by postlarval pinfish and spot may be possible only
when patches of copepods come within the feeding
range of the larvae. This hypothesis is discussed
by Murphy (1961). The above remarks emphasize
that laboratory investigations may be required in
understanding the relationships between feeding
rates and food abundance.

Size Related Food Preferences

Various investigators have observed selective
feeding by larval fish and, at times, definite pref-
erence for a specific food form is indicated. Much
of the selectivity, however, is due to the size rela-
tionship of the larval fish and the available
zooplankton (Marak 1960). Information gained
from our midday field samples and our laboratory
evacuation experiments enabled us to observe the
relationship between fish size and the size of prey
they consumed. The Wilcoxon test for paired val-
ues (Alder and Roessler 1964) was used to deter-
mine if the mean size of copepods consumed was
significantly different {a = 0.05) from those
collected in the plankton tows or provided in
aquaria.

The spot collected for both field and laboratory
studies were significantly larger than the pinfish.
Both field and laboratory results indicated that
pinfish larvae always ate smaller copepods than

the mean size available to them while the reverse
was true for spot (Table 3). Each species con-
sumed prey that were proportional to their size
with the ratio of the mean copepod length to the
average fish length approximately 1:35 based
upon laboratory measurements to 1:30 based
upon field data.

The above results suggest that, as the larval
fish size increases, the size of the consumed prey
also increases. Many researchers (Blaxter and
Holliday 1963; Blaxter 1965; Ciechomski 1967;
Detwyler and Houde 1970; de Mendiola 1974;
Marak 1974) also have observed this relationship
in a variety of larval fishes. However, the mean
size consumed in each study by either pinfish or
spot varied considerably (Table 3 ). Pinfish of simi-
lar mean sizes (16 and 16.4 mm) fed upon 590-/u.m
copepods in the laboratory, but the 460-/u,m prey
in the field. This difference in prey size may be
explained by the apparent difference in the prey
sizes available to the fish in the two studies;
laboratory prey had a mean size of 663 ixm while
those in the field were only 515 ^tm. Spot size
preferences, on the other hand, are difficult to ex-
plain in the same manner, because spot consumed
larger prey in the laboratory than in the field,
although the prey available in the laboratory
were considerably smaller than those present in
natural waters (Table 3 ).

Table 3. — Mean sizes of copepods eaten by larval pinfish and
spot in the field and laboratory compared to the mean sizes of
copepods present.

Mean length of

Copepods In

Larvae

Copepods eaten

aquaria or net tow

Species

(mm)

(fim ± 1 BE)

(^m ± 1 SE)

Laboratory:

Pinfish

16.0

590 ± 29

663 ± 28

Spot

17.7

669 ±31

491 ±68

Field:

Pinfish

16.4

460 ± 12

515 ± 22

Spot

20.4

581 ± 12

515 ±22

Comparisons of mean size prey from plankton
tows to those from gut contents may be difficult
again due to distributional dissimilarities of both
predator and prey populations during feeding and
prey aggregation patterns (Schumann 1965;
O'Brien and Vinyard 1974). However, these prob-
lems were lessened in laboratory aquaria where
we were able to control the size, density, and
distribution of the predator-prey populations.

Two primary factors appear to explain the in-
crease in prey size as larval fish size increases.

428

KJELSON and JOHNSON: FEEDING ECOLOGY OF POSTLARVAL PINFISH AND SPOT

First, mouth size usually increases as the length
of larvae increases. This relationship has been
documented for larval fish of various species by
Marak (1960), Blaxter (1965), Ciechomski (1967),
Detwyler and Houde (1970), and Shiroto (1970). A
few body measurements of pinfish and spot post-
larvae showed that the gape of the mouth in-
creased as the size of the fish increased. Pinfish of
16 mm TL were estimated to have a mouth gape
of 1.43 mm, while spot of 1.6 mm had a gape of
1.70 mm. The larger gape in spot may explain, in
part, their consumption of larger prey. Sec-
ondly, swimming speed also increases with an in-
crease in the fish's body size (Houde 1969;
Hoagman 1974); hence, the large spot may be
capable of capturing larger copepods.

Although this study emphasized the food size
preferences of postlarval pinfish and spot, a topic
of potential importance in the selective nature of
larval fish feeding deals with the selection of
specific species of copepods. We did not compare
the copepod taxa in the digestive tracts with
those found in the plankton tows, but such effort
should provide valuable information, because
copepod species differences in swimming speed,
vertical position in the water, and aggregation
behavior may be very important in determining
the type of prey available to and finally consumed
by larval fish. However, the dominant genera
present in the estuary during the study period
were Centropages, Temora, Acartia, and Euter-
pina, common forms in the Beaufort area during
winter and early spring (Thayer et al. 1974).
Marak (1960) and Ciechomski (1967) attempted to
assess the selectivity of larval fish for individual
species of copepods, but did not observe any such
preferences.

The size differences in spot and pinfish that we
observed in the Newport River estuary may be
due either to dissimilar spawning times, different
growth rates, or both. Observations made in
another North Carolina estuary (the White Oak
River estuary) during 1969 indicated that es-
tuarine spot and pinfish larval populations dur-
ing January and February differed in size and
that spot were significantly larger than pinfish
(R. M. Lewis, pers. commun., Atlantic Estuarine
Fisheries Center, Beaufort, N.C.); spot average
18.0 mm in length while pinfish were 15.5
mm. Thus, there appears to be consistancy in the
size differences observed in these two species dur-
ing their influx into North Carolina estuarine
waters.

Evacuation Rates

Regression coefficients for the equations de-
scribing the evacuation of copepods by larval
pinfish and spot are shown in Table 4. The
coefficients differ significantly from those calcu-
lated earlier (Kjelson et al. 1975). Copepod evacu-
ation in our previous study was determined using
fish collected in the estuary, placing them in a
food-free environment, and observing evacuation.
Those fish contained limited amounts of food at
the beginning of the experiments apparently due
to a low rate of feeding just prior to capture. Also,
there was a 2°C difference between estuarine and
laboratory water temperatures, and this may
have altered the evacuation rates.

In an effort to measure the evacuation through
a wide range of gut quantities and thus, hope-
fully, achieve a better description of evacuation,
our present study used fish that initially had
their guts full of copepods (21-57 copepods/fish) as
determined from sacrificing 20 fish of each species
at the beginning of the experiment. In addition,
the possible stress of transport and rapid temper-
ature changes in the earlier study were elimi-
nated by using fish that had been acclimated to
laboratory temperatures and that were fed in the
laboratory.

The regression coefficients (slopes) achieved
from our present study (Table 4) were sig-
nificantly different from and approximately twice
those found during the 1972-73 evacuation exper-
iments. We consider the estimates of evacuation
rates in the present study to be more representa-
tive of natural evacuation because the techniques
used in measuring evacuation were more refined
than in the earlier study.

The experimental temperatures, although dif-
ferent for the two species, were within the normal
range for larvae immigrating into North Car-
olina estuaries. The larger negative slope in
the regression model for spot compared with that
for pinfish (Table 4) is probably due in part to the
temperature differences (12°C for pinfish and

Table 4. — Linear regressions describing evacuation of copepods
in pinfish and spot larvae. Y =A + Bt where Y = logio(l + mean
number of copepods per larva) and t = hours since feeding, n =
number of data points.

Species

Size range
(mm)

Temperature

Pinfish
Spot

15-18
16-20

1.30
1.84

-0.18
-0.24

5
5

0.98*
0.98*

•P<0.01

429

FISHERY BULLETIN: VOL. 74, NO. 2

17°C for spot), because Peters et al. (1974) showed
that evacuation rate of juvenile pinfish and spot
is related directly to temperature.

Feeding Periodicity and Feeding Rate

Observations of larval gut contents through 24
h again indicated that these larval fish contained
the greatest amount of food during daylight hours
(Figure 2). Peaks in gut contents for both pinfish
and spot were at 1200 h. Water temperature was
15°C.

The periodicity observed in the gut contents
does not represent the actual feeding periodicity.
However, if our evacuation data and model are
appropriate, the feeding periodicity may be calcu-
lated from the periodicity of gut contents. Gut
contents at the beginning and end of each sam-
pling interval (Figure 2 ) differ by an amount equal
to the amount consumed minus the amount
evacuated during that time period (Peters and
Kjelson 1975). Thus, we can add the amount
evacuated in each interval from the change in gut
content to achieve the amount ingested during
the interval. Maximum hourly feeding rates
(from the 1000-1200-h sampling interval) were 26
copepods/h for pinfish and 17 copepods/h for spot.

Daily Rations

Estimates of daily ration for pinfish and spot
larvae were higher than those obtained from the
1972-73 study. During our earlier study, pinfish
ate 38 copepods/day while the present estimate
indicates 92. Previous estimates for spot were 47
and 99 copepods/day while our present estimate is
115 (Table 5). The increased ration sizes are at-
tributed to the use of higher instantaeous evac-
uation rates, and in the case of pinfish, to the
presence of greater amounts of food during the
feeding periodicity study (Figure 2). Pinfish di-

oeoo 1200

1600 2000

TIME Of DAY

Figure 2. — Diel cycle of digestive tract contents in larval pin-
fish and spot at 15°C based upon the geometric mean of
the number of copepods per fish (n = 10 fish per sampling
time). Vertical bars are equal to two standard errors.

gestive tracts had an average of 47 copepods/fish
at 1200 h during the 1974 sample day whereas
during 1972 they only had 10 copepods/fish. Spot
gut contents at 1200 h averaged 17, 37, and 27,
respectively in the three successive years of the
study.

Based on our daily ration estimates of 1.3 and
2.0 cal/fish per day (Table 5) and the mean
weights of the larvae, this was equal to a con-
sumption rate of approximately 0.05 cal/mg fish
wet weight per day for both species. The similar-
ity is interesting since the average pinfish weight
was only 60% that of spot and suggests that lar-
vae of dissimilar species and sizes have similar
consumption on a unit weight basis. Oxygen con-
sumption measurements by D. E. Hoss (pers.
commun., Atlantic Estuarine Fisheries Center,
Beaufort, N.C.) indicate that similar respiration

Table 5. — Daily rations calculated fi-om feeding studies and O2 consumption mea-
surements at 15°C for larval pinfish and spot in the Newport River estuary, N.C.

Mean
larvae
wet wt

(mg)

Number
copepods/
fish • day

Calories/
fish day

Calories/fish • day from
O2 consumption'

Species

Gilson
respirometer^

Flowing water
respirometer^

Pinfish
Spot

25
42

92
115

1.3
2.0

0.9
1,3

1.0
2.0

'3.38 cal/mg O2.

^Pinfish data from Hoss (1974), spot data from D. E. Hoss (pers. commun , Atlantic Estuarine Fisheries
Center, Beaufort, N.C).

^Pinfish data from W. F. Hettler, Jr. (pers. commun., Atlantic Estuarine Fisheries Center, Beaufort, N.C),
spot data from Hoss et al. (1974).

430

KJELSON and JOHNSON: FEEDING ECOLOGY OF POSTLARVAL PINFISH AND SPOT

values on a per unit weight basis are typical for
larvae of different species. Such similarity, how-
ever, may not exist for all species and size classes.

Measurements of postlarval metabolic expendi-
tures based on oxygen consumptions at 15°C
using a Gilson respirometer (Hoss, pers. com-
mun.) and a flowing water respirometer (W. F.
Hetter, Jr., pers. commun., Atlantic Estuarine
Fisheries Center) are shown in Table 5. In both
cases, fish were deprived of food for 24 h prior to
measurement of their oxygen consumption, and
the oxygen content of the water was near air sat-
uration. Both Hoss and Hettler consider their
measurements to be routine oxygen consumption
as defined by Fry (1971), i.e., the mean rate ob-
served in fish whose metabolic rate is influenced
by random activity under experimental condi-
tions in which movements are presumably some-
what restricted and the fish are protected from
outside stimuli. Postlarval pinfish and spot in the
flowing water respirometer were confined in an
11-liter chamber identical to that used for our
laboratory current- feeding experiments described
earlier and therefore were able to move about
considerably.

A major problem exists in most measurements
offish oxygen consumption due to the uncertainty
as to the animals state of activity (Altman and
Dittmer 1971). Furthermore, measurements of
fish respiration under natural conditions, termed
normal respiration, have been unattainable; and
although many investigators have estimated
normal respiration by doubling routine me-
tabolism, such a process is felt to be too subjec-
tive by Hoss and Peters (in press).

Considering the requirement for information
on fish metabolic needs under natural conditions,
it appears that our method of estimating the daily
rations of postlarval fishes has potential value.
Our estimates of daily rations were higher or
equal to those rations estimated from oxygen con-
sumption measurements. The observed differ-
ences in rations (Table 5) are reasonable if we
assume that oxygen consumption measurements,
particularly those of Hoss, are closer to routine
respiration than to normal. The Hoss data have
the lowest values, followed by the Hettler data.
These differences, although probably not sig-
nificant, are reasonable because the less restric-
tive system provided in the flowing water res-
pirometer allowed the fish to move about in a
manner similar to that in natural water. The lack
of feeding activity by fish during respiration mea-

surements and the respective decrease in oxygen
consumption (Warren and Davis 1967) also
should account for a lesser daily ration.

Based on earlier metabolic measurements,
Thayer et al. (1974) estimated a daily ration of
1.04 cal/fish per day for larval fishes in the New-
port River estuary during January and February.
They indicated that, with larval energy require-
ments of this magnitude and a 90% assimilation
efficiency, the larvae would be required to graze
on an average of 10% of the zooplankton popula-
tion per day. Furthermore, they suggested that
this need may indeed have accounted for de-
creases in zooplankton observed in the estuary
during spring. Our daily rations, based on feeding
periodicity and evacuation (Table 5), are some-
what larger and tend to support the conclusion,
assuming larval densities similar to those pre-
sented by Thayer et al. that larval fishes may
have a significant impact on copepod populations
in this system.

ACKNOWLEDGMENTS

We express our sincere appreciation to Ronald
L. Garner and Jerry D. Watson for their technical
assistance during the entire study.

LITERATURE CITED

ALDER, H. L., AND E. B. ROESSLER.

1964. Introduction to probability and statistics. 3rd ed. W.
H. Freeman and Co., San Franc, 313 p.

ALTMAN, P. L., AND D. S. DiTTMER (EDITORS).

1971. Respiration and circulation. Fed. Am. Soc. Exp.
Biol, Bethesda, 930 p.
BAINBRIDGE, V., AND D. C. T. FORSYTH.

1971. The feeding of herring larvae in the Clyde. Rapp.
P.-V Reun. Cons. Perm. Int. Explor Mer. 160:104-113.

Berner, L., Jr.

1959. The food of the larvae of the northern anchovy
Engraulis mordax. Inter-Am. Trop. Tuna Comm.,
Bull. 4:3-22.
BISHAI, H. M.

1959. The effect of water currents on the survival and dis-
tribution of fish larvae. J. Cons. 25:134-146.
BLAXTER, J. H. S.

1965. The feeding of herring larvae and their ecology in
relation to feeding. Calif Coop. Oceanic Fish. Invest.
Rep. 10:79-88.

BLAXTER, J. H. S., AND F. G. T HOLLIDAY.

1963. The behaviour and physiology of herring and other
clupeids. Adv. Mar. Biol. 1:261-393.
CIECHOMSKL J. DZ., DE.

1967. Investigations of food and feeding habits of larvae
and juveniles of the Argentine anchovy Engraulis an-
choita. Calif. Coop. Oceanic Fish. Invest. Rep. 11:72-81.

431

FISHERY BULLETIN: VOL. 74, NO. 2

DE MENDIOLA, B. R.

1974. Food of the larval anchoveta Engraulis ringens J. In

J. H. S. Blaxter (editor), The early life history of fish, p.

277-285. Springer- Verlag, N. Y.
DETWYLER, R., AND E. D. HOUDE.

1970. Food selection by laboratory-reared larvae of the
scaled sardine Harengula pensacolae (Pisces, Clupeidae)
and the bay anchovy Anchoa mitchilli (Pisces, En-
graulidae). Mar. Biol. (Berl.) 7:214-222.

Fry, f. E. J.

1971. The effect of environmental factors on the physiology
of fish. In W. S. Hoar and D. J. Randall (editors), Fish
physiology, Vol. 6, p. 1-98. Academic Press, N. Y.

Heinle, D. R.

1966. Production of a calanoid copepod, Acartia tonsa in
the Patuxent River estuary. Chesapeake Sci. 7:59-74.

HOAGMAN, W. J.

1974. Vital activity parameters as related to the early life
history of larval and post-larval lake whitefish
(Coregonus clupeaformis). In J. H. S. Blaxter (editor).
The early life history of fish, p. 547-558. Springer-
Verlag, N. Y
HOSS, D. E.

1974. Energy requirements of a population of pinfish
Lagodon rhomboides (Linnaeus). Ecology 55:848-855.

HOSS, D. E., W. F. Hettler, Jr., and L. C. Coston.

1974. Effects of thermal shock on larval estuarine fish -
ecological implications with respect to entrainment in
power plant cooling systems. In J. H. S. Blaxter (editor),
The early life history of fish, p. 357-371. Springer-
Verlag, N. Y.

HOSS, D. E., and D. S. Peters.

In press. Respiratory adaptation: Fish. Proc. Estuarine
Res. Fed. 3rd Int. Bienn. Conf., Oct. 6-9, 1975, Gal-
veston, Tex.
HOLHDE, E. D.

1967. Food of pelagic young of the walleye, Stizostedion
vitreum vitreum, in Onedia Lake, New York. Trans. Am.
Fish. Soc. 96:17-24.

1969. Sustained swimming ability of larvae of walleye
(Stizostedion vitreum vitreum) and yellow perch (Perca
flavescens). J. Fish. Res. Board Can. 26:1647-1659.

IVLEV, V. S.

1961. Experimental ecology of the feeding of fishes
(Translated from Russian by D. Scott) Yale Univ.
Press, New Haven, 302 p.

KjELSON, M. A., D. S. Peters, G. W. Thayer, and G. N.
Johnson

1975. The general feeding ecology of post-larval fishes in
the Newport River estuary. Fish. Bull., U.S. 73:137-144.

Lewis, R. M., W. F. Hettler, Jr., E. p. H. Wilkens. and G. N.
Johnson.

1970. A channel net for catching larval fishes.
Cheasapeake Sci. 11:196-197.

MARAK, R. R.

1960. Food habits of larval cod, haddock, and codfish
in the Gulf of Maine and Georges Bank area. J. Cons.
25:147-157.

1974. Food and feeding of larval redfish in the Gulf of
Main. In J. H. S. Blaxter (editor), The early life history
of fish, p. 267-275. Springer- Verlag, N. Y.

Murphy, G. I.

1961. Oceanography and variations in the Pacific sardine
population. Calif. Coop. Oceanic Fish. Invest. Rep.
8:55-64.

Nakai, Z., K. Honjo, S. Hattori, T. Kidachi, and H. Suzuki.

1966. Further examples of Hjort's hunger theory. 2nd Int.
Oceanogr. Congr., Mosc. Abstr. of pap., p. 263.

O'Brien, W. J., and G. L. vinyard.

1974. Comment on the use of Ivlev's electivity index with
planktivorous fish. J. Fish. Res. Board Can. 31:1427-
1429.

Peters, D. S., and M. a. Kjelson.

1975. Consumption and utilization of food by various post-
larval and juvenile North Carolina estuarine fishes. In L.
E. Cronin (editor), Estuarine research. Vol. 1, p. 448-
472. Academic Press, N.Y

Peters, D. S., M. a. Kjelson, and M. T. Boyd.

1974. The effect of temperature on digestion rate in the
pinfish, Lagodon rhomboides; spot, Leiostomus zan-
thurus; and silverside, Menidia menidia. Proc. 26th
Annu. Conf. Southeast Assoc. Game Fish Comm., p.
637-643.

RYLAND, J. S.

1963. The swimming speeds of plaice larvae. J. Exp. Biol.
40:285-299.

Schumann, G. O.

1965. Some aspects of behavior in clupeid larvae. Calif
Coop. Oceanic Fish. Invest. Rep. 10:71-78.

SEREBROV, L. I.

1973. Effect of a current on the intensity of feeding in cer-
tain fish. Hydrobiol. J. 9(2):68-70.

Shirota, a.

1970. Studies on the mouth size of fish larvae. [In Jap.,
Engl, abstr.] Bull. Jap. Soc. Sci. Fish. 36:353-368.

THAYER, G. W., D. E. HOSS, M. A. KJELSON, W. F. HETTLER, JR.,
AND M. W. LaCROIX.

1947. Biomass of zooplankton in the Newport River estuary
and the influence of postlarval fishes. Chesapeake Sci.
15:9-16.

WARREN, C. E., AND G. E. DAVIS.

1967. Laboratory studies on the feeding, bioenergetics,
and growth of fish. In S. D. Gerking (editor), the bio-
logical basis of fresh water fish production, p.
175-214. Blackwell Sci. Publ., Oxf.

432

THERMAL TOLERANCE AND RESISTANCE OF
THE NORTHERN ANCHOVY, ENGRAULIS MORDAX

Gary D. Brewer^

ABSTRACT

An experimental, flow-through seawater system, constructed to maintain juvenile and adult northern
anchovy, Engraulis mordax, and rear embryos and larvae through yolk-sac absorption under con-
trolled temperature and photoperiod regimes, was used to determine aspects of thermal tolerance,
resistance, rates of acclimation, and some effects of temperature on the development and growth of the
anchovy.

Thermal tolerance was determined for juvenile and adult fish, acclimated to six constant tempera-
tures between 8° and 28°C. Thermal resistance (minutes until death for fish exposed to a lethal
temperature) was independent of photoperiod and fish size; however, females proved more resistant
than males, and resistance decreased at night. Acclimation (as measured by resistance) from 12° to
20°C was nearly complete after 2-day exposure to the higher temperature; acclimation from 20° to 12°C
was nearly complete after 5-day exposure to the lower temperature. Fish subjected to fluctuating water
temperatures between 12° and 20°C proved less resistant to cold than a 12°C (constant) acclimated
group and less resistant to heat than a 20°C (constant) acclimated group.

Thermal tolerance was determined for larvae in the yolk-sac stage, acclimated to four constant
temperatures between 12° and 24°C. Although hatching occurred at temperatures as high as 29.5°C
and as low as 8.5°C, the percentage of normally developed larvae equaled that of controls (incubated at
16°C) only between temperatures of 27.0° and 11.5°C. Embryos in the blastodisc stage proved most
sensitive to acute temperature increases when compared to embryos in the blastopore closure stage and
larvae in the yolk-sac stage. These same three stages proved insensitive to acute temperature de-
creases to 0.5°C for 60-min exposure periods.

Temperature is discussed in relation to anchovy distribution and survival under natural and
artificially created thermal conditions.

Research on the effects of temperature on aquatic
organisms has been given impetus in recent years
as numerous lakes and streams are considered
potential heat reservoirs by electric power
generating plants and other industrial concerns.
As the demands for water as a heat transfer
medium continue to increase dramatically, more
attention will be turned to the marine environ-
ment for large volumes of water and surface areas
necessary for the dissipation of excess heat
(Naylor 1965; de Sylva 1969; Tarzwell 1972). Un-
checked thermal loading of freshwater and near-
shore marine ecosystems will inevitably pose a
serious threat to the homeostasis and well-being of
aquatic communities unless realistic guidelines
are established and enforced. Such guidelines
must be based on knowledge of how aquatic or-
ganisms respond to both acute and chronic tem-
perature changes.

This study details aspects of thermal tolerance
and resistance (as defined by Fry 1971) on the

'Allan Hancock Foundation, University of Southern Cali-
fornia, University Peirk, Los Angeles, CA 90007.

Manuscripted accepted October 1975.
FISHERY BULLETIN: VOL. 74, NO. 2, 1976.

embryo, larval, juvenile, and adult stages of the
northern anchovy, Engraulis mordax Girard. The
study was prompted by the proposed discharge of
thermal effluent into the Los Angeles-Long Beach
Harbor. The biology and fishery of the northern
anchovy in the Los Angeles-Long Beach Harbor
were described by Brewer (1975a).

The general biology of the northern anchovy has
been summarized by Baxter (1967), Messersmith
et al. (1969), the California Department of Fish
and Game (1971), and Brewer (1975a). A dramatic
increase in abundance of E. mordax during the
past 20 yr (Ahlstrom 1967; Smith 1972) has
prompted an intense interest in the biology and
fishery potential of this clupeoid. The California
Department of Fish and Game (1971:48) consid-
ered the anchovy ". . . the most abundant species
with immediate harvest potential in the Califor-
nia Current system."

MATERIALS AND METHODS

Experiments were conducted in a small, tem-

433

FISHERY BULLETIN: VOL. 74, NO. 2

perature and photoperiod controlled, flow-
through sea water system. The system delivered
filtered, ultraviolet-sterilized seawater from the
Los Angeles-Long Beach Harbor to five round,
950-liter, fiber glass aquaria (L5 m in diameter,
0.6 m high) and a single 400-liter rectangular wa-
ter table, all housed in a light-tight aluminum
cargo container (Figure 1). An exchange rate of
2-6 liters/min was maintained in each aquarium,
with overflow drainage provided by standpipes.
Wastewater was not recirculated. Temperatures
were maintained within ±0.5°C.

Above each aquarium were two incandescent
light bulbs controlled by separate dimmer controls
and regulated by a 7-day timer to simulate photo-
periods. The "day" bulb provided 700 Ix and the
"night" bulb provided 16 Ix to the surface of each
aquarium. Oxygen was maintained at or near sat-
uration levels in all acclimation and test tanks by
splashing incoming water at the surface and by
bubbling air stones in the aquaria. Salinity varied
between 31.4 and 33.81. (mean 33. ID during the
study period.

Juveniles and Adults

Juvenile and adult E. mordax were obtained
from a live-bait dealer. The initial transfer from
the bait boat to the 950-liter acclimation tanks
caused 20-30^^ mortality during the first 2-3 days
of confinement. Within 2-4 days, healthy fish be-
gan to feed and were offered a daily ration, equiv-
alent to approximately 4% of the fish's wet weight,

of Trout Chow.^ This ration was supplemented
with chopped anchovy, chopped squid, brine
shrimp, or wild plankton equal to approximately
1% of the fish's wet weight. Adjusted fish ate vora-
ciously and mortality became insignificant in ac-
climation tanks within 1 wk. Acclimation tanks
were stocked with between 3 and 7 kg of anchovy.
The food ration was withheld for a period of 24 h
prior to all thermal tests on juvenile and adult fish.

Ninety-six Hour Tolerance

Standardized techniques for the determination
of lethal temperatures (Fry et al. 1942; Brett 1944;
Fry 1947) call for a series of experiments in which
the animals are acclimated to several different
constant temperatures. Acclimated fish are then
abruptly transferred to test aquaria previously
equilibrated to various high and low temperature
extremes. Mortality is monitored and recorded.

This procedure extends the concept of lethal
temperatures from two extreme end points, to a
family of upper and lower (incipient) lethal levels.
The ultimate upper and lower lethal tempera-
tures, which circumscribe the extreme tolerance
limits, may be determined by graphic
extrapolation — that is, by drawing a line through
those high and low test temperatures that proved
lethal to 50*7^ of the test animals for each acclima-
tion temperature. The extrapolated line will then

^Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

Figure l. — Diagram of the flow-through seawater system used for experiments on Engraulis mordax.

434

BREWER: THERMAL TOLERANCE AND RESISTANCE

intersect a diagonal at the upper and lower ex-
tremes, which represents those points where
the lethal temperatures equal the acclimation
temperatures. Fish cannot be acclimated to tem-
peratures above or below these experimentally
determined ultimate upper and lower lethal
temperatures, respectively.

Juvenile and adult E. mordax between 45 and
139 mm SL (standard length) were held for a
minimum of 3 wk at constant temperatures of 8°,
12°, 16°, 20°, 24°, and 28°C and under a light cycle
of 12 h light and 12 h dark. Unless otherwise
noted, the term "acclimated fish" designates E.
mordax held under such conditions. "Juvenile" re-
fers to metamorphosed fish less than 100 mm SL,
while "adult" refers to fish over 100 mm.

The fish's susceptability to mechanical damage
increased at high and low acclimation tempera-
tures. Therefore, each acclimation temperature-
test series was accompanied by a series of strict
control transfers, and the observed mortalities for
each series were adjusted separately, based on the
respective control mortalities. Fish were consid-
ered dead and were removed when all swimming
movements ceased. Ninety-six hour LDso (mean
lethal dose) temperatures (i.e., incipient lethal
levels) were estimated from regression lines plot-
ted on probit paper (Sokal and Rohlf 1969). Exper-
iments were conducted between February 1973
and November 1974 and included all seasons.
About 20 fish were used for each test.

Resistance as a Function of Size,
Sex, Time, and Photoperiod

To test the potential influence of photoperiod,
sex, size, and diel effects, anchovies were accli-
mated to 20°C, tested by direct transfer to 30°C,
and the time to death (resistance time) deter-
mined. As the fish died, they were removed from
the test aquaria, measured, and adult fish were
sexed. Identical tests were conducted in the morn-
ing (0900 h) and in the evening (2100 h). Tests
were also conducted after fish had been held under
a short-day photoperiod (8 h light) and a long-day
photoperiod ( 16 h light) for periods of 3 wk each.
All thermal resistance tests were run during the
summer and fall.

Rates of Thermal Acclimation

Juvenile and adult anchovy acclimated to 12°C
were subjected to an 8°C temperature change over

a 24-h period to 20°C, and then tested for resist-
ance to 30°C on the same day and after 1-, 2-, and
4-day exposure to the 20°C temperature.
Moreover, fish acclimated to 20°C were subjected
to a temperature decrease to 12°C over a 24-h
period, and then tested for resistance to 6°C on
the same day and after 2-, 5-, and 9-day exposure
to the 12°C temperature. As the fish become accli-
mated to the new higher or lower temperature, one
would expect the mean resistance times for these
fish to approach and eventually equal the mean
resistance times (e.g., reach a steady-state) offish
acclimated to 20° and 12°C and tested at 30° and
6°C, respectively (controls).

Effects of Cycled Temperatures
on Resistance

In view of the observations by Mais (1974) that
E. mordax may undergo diel vertical migrations
and consequently experience fluctuating tempera-
tures, I examined the relative thermal resistance
of anchovies subjected to regular changes in tem-
perature from 12° to 20°C over 48-h intervals. Fish
acclimated to 20°C were gradually subjected to
decreasing temperatures to 12°C over 24 h and
then back to 20°C over the next 24 h. The cycle was
repeated for 25 days, at which time a sample offish
which had just reached 20°C was tested for resist-
ance to 30°C. The following morning, as the re-
maining fish reached 12°C, a sample was tested for
resistance to 6°C.

Embryos and Larvae

Engraulis mordax eggs, caught in plankton
tows in or near the Los Angeles-Long Beach Har-
bor throughout the year, were utilized for experi-
ments on embryos and larvae. Water tempera-
tures, at time of capture, varied between 13° and
18°C. In the laboratory, eggs in the blastodisc
stage were placed into 2-liter glass jars and main-
tained at 12°, 16°, 20°, or 24 °C until transferred to
incubation or test vessels which consisted of
250-ml jars containing 60 ml of seawater. Not
more than five eggs or larvae were tested per jar.

Twenty-four Hour Tolerance

Larvae in the yolk-sac stage were tested within
1 day after hatching at each acclimation tempera-
ture. Larvae were pipetted from each acclimation
temperature directly into test vessels ranging

435

FISHERY BULLETIN: VOL. 74, NO. 2

from 6° to 32°C. Mortality was recorded after 24 h.
Incipient lethal levels for each acclimation level
were estimated as described for the juveniles and
adults.

Hatching and Developmental
Temperature Limits

Eggs in the blastodisc stage were transferred to
a series of incubation vessels, after which the
temperatures were gradually raised or lowered
from the ambient level of 16°C over a period of 60
min in order to avoid possible shock effects to the
developing embryos. Incubation temperatures
were then held constant (±0.5°C) between 6° and
12°C, and 26° and 31°C at 0.5°C intervals. A 16°C
temperature was used as a control. Development
was considered normal only if the larvae were free
of obvious deformities (e.g., spinal curvatures)
until pigmented eyes and functional jaws were
evident, and death had occurred only after yolk
reserves were exhausted.

Resistance to Acute Temperature
Changes

Embryos in the blastodisc stage (ca. 12-14 h
after fertilization at 16°C) and in the blastopore
closure stage (ca. 36-38 h after fertilization), and
larvae in the yolk-sac stage (vdthin 24 h after
hatching) were subjected to temperature shocks
for periods of 1, 3, 5, and 60 min. Embryos and
larvae were pipetted from incubation vessels
maintained at 16°C directly into water at high and
low temperature extremes. After the exposure
period, the embryos and larvae were returned di-
rectly to the incubation vessels at 16°C where they
remained for 48 h after hatching. Mortality and
developmental abnormalities were recorded.

This procedure was an attempt to simulate what
the embryos and larvae might actually experience
if entrained by intake pipes of electrical generat-
ing plants or LNG (Liquified Natural Gas) vapor-
ization plants (or either thermal plums), subjected
to rapid temperature increases and decreases in
the heat exchange systems, and subsequently
flushed back into the natural environmental
temperatures at the outfall.

Development and Growth

Experiments were designed to determine the
temperatures required for optimal growth of an-

chovy larvae. The tests were confined to that
period of larval life between hatching and starva-
tion following exhaustion of all stored yolk re-
serves. No food was offered.

Eggs in the blastodisc stage were reared
through hatching in a series of constant tempera-
ture baths between 10° and 26°C. On the day of
hatching and each subsequent day, approximately
10 larvae were sacrificed from each rearing tem-
perature and measured from the tip of the snout
to the end of the notochord with an ocular microm-
eter to the nearest 0.05 mm. This procedure
was continued until all larvae at each rearing
temperature died of starvation.

RESULTS

Juveniles and Adults

Ninety-six Hour Tolerance

Experiments on juvenile and adult tolerance
encompassed 117 separate 96-h tests and 2,400
fish. Control survival ranged from lows of 81.3 and
87.9% at 8° and 28°C acclimation temperatures,
respectively, to 98.3% at the 16°C acclimation
temperature.

Figure 2 graphically depicts the lethal tempera-
ture relations, with adjusted percent mortality
plotted against test temperatures for acclimation
levels of 8° and 28°C. Adjusted upper and lower
LDso temperatures were plotted against acclima-
tion temperatures in Figure 3 and a thermal toler-
ance polygon constructed (Fry 1947). Ultimate
upper and lower lethal temperatures are esti-
mated by extrapolation (line fitted by eye) to be

TEMPEtATUIE "^

Figure 2.— Effects of acclimation temperatures of 8° and 28°C
on the upper and lower lethal temperatures ofEngraulis mordax
juveniles and adults (original date in Brewer 1975b).

436

BREWER: THERMAL TOLERANCE AND RESISTANCE

32'

30-

28-

_ 2*'
o

Q 24-

S "-I

16-
U-
12-
10-
8-
6'

/
/

^ • *^

^^ •

: .

10 12 u 16 18 20 22 24 26 28 30 32
ACCLIMATION TEMPERATURE °C

Figure 3. — Thermal tolerance polygon for Engraulis mordax
juveniles and adults. Those points where the extrapolated LDso
levels intersect the diagonal represent the extreme (ultimate)
tolerance limits and correspond to 6.5° and 29.5°C.

29.5° and 6.5°C, respectively. These temperatures
represent the maximum tolerance range of E.
mordax juveniles and adults sampled from south-
ern California and maintained under laboratory
conditions as described. Anchovy cannot be accli-
mated to temperatures beyond these extremes.
Attempts were made to slowly acclimate fish to
29.5° and 6.5°C, but they proved futile.

Figure 4 shows the resistance times to median
mortality of juvenile and adult E. mordax, accli-
mated to 8°, 16°, and 28°C, upon exposure to tem-
peratures beyond incipient lethal levels. These
curves were derived by plotting cumulative mor-
tality as percentages against exposure time to es-
timate the time to LDso for each test temperature.

Resistance as a Function of Size,
Sex, Time, and Photoperiod

Results of experiments on thermal resistance to
30°C in relation to size, sex, and potential diel and
photoperiod effects are summarized in Table 1.
Analysis of variance (one-way classification)
showed that resistance times to lethal tempera-
tures of 30°C were not significantly different
(P>0.05) for fishes of different sizes (<79 mm;
80-99 mm; >100 mm) or for fishes maintained
under different photoperiods (8, 12, and 16 h

HOURS TO MEDIAN MORTALITY

Figure 4. — Resistance times to median mortality of juvenile
and adult Engraulis mordax exposed to high (A) and low (B)
lethal temperatures when acclimated to 8°, 16°, and 28°C.

Table l. — Comparison ofresistance times (minutes until death)
for juvenile and adult Engraulis mordax acclimated to 20°C and
immediately transferred to aquaria at 30°C.

Item

Range

(min)

Mean
(min)

Length (mm):

<79 11

80-99 22

>100 12

Sex:

Male 20

Female 36

Time of test:

Morning 45

Evening 38

Photopenod (hours of

8 14

12 34

16 11

49-285
41-302
37-343

6-118

4-401

31-343

8-244

light):

6-401

37-302

31-343

133.9

149.4
141.3

40.9
116,8

143.5
72.6

154.5
141.1
150.8

74.9

62.3

1029

33.1
113.6

76.3
66.7

127.0

67.6

102.3

22.6

13.3
29.7

7.4
18.9

11.4
10.8

34.0
11,6
30.9

light). These results should be verified with larger
sample sizes. Resistance times showed highly sig-
nificant differences (P<0.01) for males compared
with females, and for tests conducted in the morn-
ing as compared with those conducted at night.
Females proved more resistant than males, and
animals tested in the morning showed greater re-
sistance than those tested in the evening.

437

FISHERY BULLETIN: VOL. 74, NO. 2

Rates of Thermal Acclimation

Results suggest that acclimation from 12° to
20°C nears completion within 2 days of exposure to
the higher temperature. Mean resistance times for
day 2 and day 5 samples exceeded the mean resist-
ance time for the control sample. However,
analysis of variance shows that the variation be-
tween day 2, day 5, and control samples is not
significant (P>0.05). The relatively high resist-
ance of some fish in the day 2 and day 5 samples,
which exceeded the resistance of control fish, may
be due to slight temperature variations (± 0.2°C) in
the test aquaria, or possibly to "physiological
overshoots" to the acclimation process (Prosser
1973). Figure 5 shows the progress toward accli-
mation with contined exposure to the higher
temperature. Most noticeable is the change in
shape of the resistance curves with acclimation.
Nonacclimated fish succumb to the lethal 30°C
temperature quickly, probably as a result of
"shock effects" (Scott 1964; Tyler 1966; Allen and
Strawn 1971). Acclimation to the higher tempera-
ture diminishes the shock effects. Apparently, ac-
climated fish die from secondary causes termed
"direct effects" by Fry (1971). The physiological
basis of the shock and direct effects is not clear.
Acclimation from warm to cool water (20° to 12°C)
appears to be nearly complete by day 5 (Figure 6).
As acclimation progresses and resistance to low
temperatures is increased, death rate becomes in-

MINUTCS EXPOSURE

Figure 6. — Cumulative percent mortality ofEngraulis mordax
juveniles and adults as a function of exposure to 6°C. The re-
sponse of a 12°C acclimated control group (C) is compared with
that of a 20°C acclimated group after 0-, 2-, 5-, and 9-day expo-
sures to 12°C (original date in Brewer 1975b).

creasingly regular; the graphs approach a straight
line and the effects of the initial cold shock are
largely eliminated. Because these tests used small
sample sizes, statistical differences cannot be
demonstrated.

Effects of Cycled Temperatures on
Resistance

MINUTES EXPOSURE

Figure 5. — Cumulative percent mortality oiEngraulis mordax
juveniles and adults as a function of exposure to 30°C. The
response of a 20°C acclimated control group (C) is compared with
that of a 12°C acclimated group after 0-, 1-, 2-, and 4-day expo-
sures to 20°C (original data in Brewer 1975b).

Results are summarized in Table 2. Mean re-
sistance times to 6° and 30°C for fish subjected to

Table 2. — Resistance times (minutes until death) of juvenile
and adult Engraulis mordax to 6° and 30°C after being subjected
to temperature fluctuations between 12° and 20°C on a 48-h cycle
for a period of 25 days.

Test

Range

Mean

temp

Group

(min)

6°C

12°C acclimated

643-2,490

1,419.0

589.0

186.3

20° to 12°C

117-1,111

410.6

374.5

118.4

30°C

20°C acclimated

37-302

141.1

67.6

11.6

12° to 20°C

6-68

28.0

22.1

7.0

periodic changes in temperature between 12° and
20°C were well below the mean resistance times of
fish acclimated to a constant 12°C and constant
20°C, respectively. However, the fish have greater
high temperature resistance than those accli-
mated to 12° and greater low temperature resist-
ance than those acclimated to 20°C.

438

BREWER: THERMAL TOLERANCE AND RESISTANCE

Embryos and Larvae

Twenty-four Hour Tolerance

Over 600 larvae were tested in the 24-h toler-
ance experiments. Generally, 10 or more larvae
were tested at each temperature. The percentage
(normal) survival for controls ranged from 72.7 at
the 12°C acclimation level to 86.7 at 16° and 20°C
acclimation temperatures.

Apparently the physiological mechanisms for
thermal acclimation are little developed in E.
mordax larvae in the yolk-sac stage. Figure 7
shows the 24-h lethal temperature relations with
percent adjusted mortality plotted against test
temperatures for acclimation temperatures of 12°
and 24°C. Rearing the yolk-sac larvae in warm
and cold water does little to increase or decrease
their upper or lower lethal temperatures, respec-
tively. Potential effects of parental acclimation
temperatures (Hubbs and Bryan 1974) or the ex-
posure of eggs to acclimation temperatures at the
time of fertilization require investigation.

10 12 14 16 18 20 22 24 26 28 30 32
ACCLIMATION TEMPERATURE t

Figure 8. — Thermal tolerance polygon for Engraulis mordax
larvae in the yolk-sac stage. Those points where the extrapolated
LDso levels intersect the diagonal represent the extreme toler-
ance limits and correspond to 7.0° and 30.2°C.

IfMP(R*IUHf "^

FIGURE 7.— Effects of acclimation temperatures of 12° and 24°C
on the upper and lower lethal temperatures of Engraulis mordax
larvae in the yolk-sac stage (original data in Brewer 1975b).

In Figure 8, adjusted upper and lower LDso
temperatures are plotted against respective ac-
climation temperatures in the construction of a
thermal tolerance polygon. Ultimate upper and
lower lethal temperatures are estimated to be
30.2° and 7.0°C, respectively.

Hatching and Developmental
Temperature Limits

Results of this experiment are given in Table 3.
Although hatching was observed at temperatures

Table 3. — Effects of temperature on hatching and development
of Engraulis mordax through yolk-sac absorption and eye pig-
mentation of larvae. Temperatures were maintained within
±0.5°C of those shown below.

Rearing

With normal development Adjusted

(°C)

hatching

No.

(%)

31.0

0.0

30.0

0.0

29.5

30.0

33.3

29.0

3 .

30.0

33.3

28.5

50.0

55.6

280

76.7

84.9

27.5

70.0

77.8

27.0

90.0

100.0

26.5

80.0

88.9

26.0

90.0

100.0

'16.0

90.0

100.0

7.5

0.0

8.0

0.0

8.5

0.0

9.0

0.0

9.5

0.0

10.0

11.1

10.5

30.0

33.3

11.0

60.0

66.7

11.5

90.0

100.0

12.0

90.0

100.0

'Control.

as high as 29.5°C and as low as 8.5°C, 50% (ad-
justed) normal development occurred between
11.0° and 28.5°C. Only below 27.0°C and above
11.5°C did the percentages of hatching and normal
development approach those for the controls.

439

FISHERY BULLETIN: VOL. 74, NO. 2

Resistance to Acute Temperature
Changes

Resistance to high temperatures is surprisingly
great for embryos and larvae when exposure is of
short duration (Figure 9). Blastodisc stage em-
bryos are least resistant while yolk-sac larvae are
most resistant; LDso values for the 60-min expo-
sure period for the larvae are within 1.3°C of the
extrapolated 24-h tolerance limits determined
from Figure 8. Engraulis mordax embryos and
larvae appear to be insensitive to abrupt tempera-
ture decreases down to 0.5°C for short periods
(Brewer 1975b).

Table 4. — Comparison of the maximum size attained by En-
graulis mordax larvae in the yolk-sac stage before shrinkage due
to starvation.

40-

cH 38-

36'

32-

28-^

3 J 5 a 8 10 20

MINUTES TO MEDIAN MORTALITY

Figure 9. — Minutes to median mortality for the blastodisc
stage (1), blastopore closure stage (2), and yolk-sac stage (3)
subjected to abrupt temperature increases from 16°C (original
data in Brewer 1975b).

Development and Growth

Table 4 summarizes data on the growth of lar-
vae at constant temperatures between 10° and
26°C. The maximum size of larvae attained at any
temperature before shrinkage due to starvation
was 4.16 mm. This is considerably smaller than
the value of 4.8 mm given by Lasker (1964) for jB.
mordax larvae reared under similar conditions.
Variability in egg size may be responsible for this
discrepancy; egg size of the Argentine anchovy, E.
anchoita, is known to vary by season and location
(de Ciechomski 1973).

The highest mean growth response was ob-
tained for larvae reared at 18°C (3.94 mm). Mean
larval lengths less than 3.78 mm were considered

Incubation

temp

Range

Mean

(°C)

(mm)

3.37-3.79

3.63

0.13

0.04

3.31-4.00

3.62

0.21

0.07

3.63-4.10

3.93

0.18

0.06

3.63-4.00

3.81

0.15

0.05

363-4 10

3.94

0.16

0.05

3.52-4.16

3.82

0.18

0.06

3.52-4.00

3.71

0.14

0.04

3.42-3.84

3.56

0.13

0.04

significantly smaller than the maximum response
at 18°C (Least Significant Difference, Sokal and
Rohlf 1969). It seems reasonable to assume that
larvae reared at temperatures of 12°C or lower and
24°C or higher converted yolk into body tissue at
suboptimal levels. Analysis of variance showed
that maximum mean lengths attained by larvae
reared at 14°, 17°, 18°, and 20°C were not sig-
nificantly different (P>0.05).

DISCUSSION

Figure 10 shows a graphic summary of various
field and laboratory-deduced temperature ranges
and limits for the distribution and survival oi E.
mordax. A temperature range of about 4.5°C lies
between the highest temperatures that anchovy
adults have been found in nature (25°C, Baxter
1967) and the experimentally determined upper
lethal temperature for juveniles and adults
(29.5°C). Anchovy had been maintained in the
laboratory at 28°C for weeks with no apparent ill
effects. The fish are extremely active at this tem-
perature and their metabolic requirements are un-
doubtedly considerable. Anchovy maintained at
28°C and fed the standard ration lost weight. The
upper environmental temperature limit and
southern distributional limit of £. mordax may be
dictated by metabolic demands which outweigh
the ration supplied by the environment.
Maximum temperatures off Cabo San Lucas,
which is the southern range limit for E. mordax,
exceeds 25°C (Lynn 1967). Interestingly, 25°C cor-
responds to the highest temperature that juvenile
E. mordax would venture into when tested in
laboratory thermal gradients (Brewer 1974).
Moreover, the plateau in the thermal tolerance
polygon (Figure 3) shows that acclimation tem-
peratures of 24°C and above have little effect on
increasing the incipient upper lethal temperature.
Apparently the anchovy's overall mechanisms for

440

BREWER: THERMAL TOLERANCE AND RESISTANCE

physiological compensation begin to break down
at temperatures above 25°C.

Reid's (1967) observation that E. mordax may
overwinter at temperatures of 7° or 8°C off British
Columbia is of special interest. These fish may be
within less than 1°C of their lower lethal tempera-
ture. Juvenile and adult anchovy acclimated to
8°C in the laboratory and transferred to 7°C made
no effort to consume food offered to them after 5
days at the lower temperature. I have not
confirmed this by stomach examination, but feed-
ing, if it takes place at all, is minimal at this low
temperature.

It is important to consider the possibility that
the thermal tolerance and resistance of £■. mordax
may be different for northern, central, and south-
ern populations. Apparently genetically dis-
tinct, these populations were first identified on the
basis of meristic characters by McHugh ( 195 1 ) and
later on by serum transferrin analysis conducted
by Vrooman and Smith (1971). If the thermal re-
quirements of these populations were distinct, I
would anticipate their reproductive temperature
ranges to vary accordingly. Richardson's (1973)
data on anchovy spawning off Oregon discount
this. In any case, thermal resistance experiments

on samples from each population would be of
interest.

Experiments on the resistance of juvenile and
adult anchovy to a high lethal temperature
showed no significant difference in the mean re-
sistance times for fish of different sizes or for fish
maintained under different photoperiods. How-
ever, females were more resistant than males, and
animals tested in the morning showed greater re-
sistance than those tested in the evening. Inves-
tigators have variously shown significant differ-
ences in one or more of the factors tested here,
depending on the species. Thermal resistance has
been found to vary according to size, with large
Oncorhynchus (Salmonidae) and Carassius (Cyp-
rinidae) more resistant to cold (Brett 1952; Hoar
1955, respectively) and large Clupea (Clupeidae)
less resistant to heat (Brawn 1960). Carassius
maintained under long photoperiods were more
resistant to high temperatures than fish main-
tained under short photoperiods, while resistance
to cold temperatures was greater for the short
photoperiod fish (Hoar 1956). Hoar discovered that
male Carassius were more resistant to low tem-
perature extremes than females. Heath ( 1963) ob-
served slight differences in critical thermal

LOWER LD^ AT
24° ACCLIMATION

UPPER LO^ AT
8° ACCLIMATION

LOWER LD^Q AT
24^^ ACCLIMATION

UPPER LOggAT

t2° ACCLIMATION

1 .

50% HATCH

.. ^

— Maximum Larval Growth ^

°c 6

10 12 14 16 18 20 22 24 26 28 30

-Spawning Limits -

Range of Larvae -

-Range of Adults-

FlGURE 10.— Field and laboratory deduced thermal limits for the distribution and survival ofEngraulis mordax.

441

FISHERY BULLETIN: VOL. 74, NO. 2

maximum temperatures for Saluelinus (Sal-
monidae), depending on the time the test was con-
ducted. He also noticed that maximum tolerance
followed a 24-h cycle and suggested that this was a
physiological adaptation to natural habitats with
24-h variations in temperature.

The present experiments on E. mordax were
conducted in the fall when anchovy presumably
ascend from deep water to warm surface waters in
the evening (California Department of Fish and
Game 1971). If a circadian cycle of thermal resist-
ance existed in anchovy, one might anticipate
maximum resistance to high temperatures to
occur in the evening. The data in Table 1 suggest
that, under laboratory conditions, resistance to
high temperature is reduced in the evening.

The embryonic and larval stages of pelagic
fishes are potentially the most vulnerable ones to
thermal stresses. While juvenile and adult fishes
may detect and avoid unfavorable environmental
conditions (Bull 1928; Doudoroff 1938; Alabaster
and Robertson 1961; Coutant 1969), the eggs and
planktonic larvae of fishes such as E. mordax are
at the mercy of currents which might carry them
into environments unfavorable for growth or sur-
vival. Reviews by de Sylva ( 1969) and Brett ( 1970)
have shown that on the average, marine fish lar-
vae are one-third to one-half as tolerant to thermal
stresses as their conspecific adults. Normal de-
velopment of £. mordax is inhibited below 11.5°C
and above 27.0°C. Larvae held at temperatures
below 11.0°C for short periods become inactive,
making little effort to avoid capture by pipette.

The survival of pelagic larvae is dependent on
the early consumption of prey species and the abil-
ity to avoid predators (Lasker et al. 1970). The
degree to which these two processes can be ac-
complished is largely dependent on the optimal
development of swimming ability, precise biting
reflexes, and visual acuity (Hunter 1972). Since
swimming ability is proportional to larval size, the
development of maximum growth potential
should be of distinct survival value. Maximum
growth of larvae in the yolk-sac stage, in turn, is
dependent on the efficient utilization of the lim-
ited yolk reserve, i.e., its conversion into body
tissues.

Growth of anchovy larvae in the yolk-sac stage
is maximal in experimental temperatures be-
tween 14° and 20°C. Variation within this range
may be highly significant but is difficult to test.
Although growiih rates of anchovy larvae in the
yolk-sac stage increase with increasing tempera-

tures, the maximum size attained by the larvae
decreased at high temperatures.

Thermal tolerance limits have been determined
for anchovy larvae and juveniles and adults by
tests that considered the LD50 as a lethal end
point. LD50 temperatures do not represent "safe"
levels and have been used merely because of con-
vention. Any temperature level that produces a
lethal response significantly greater than the
maximum response at control temperatures
should be considered excessive. This would repre-
sent the most realistic end point to insure en-
vironmental quality. The thermal death of even a
few individuals at any particular temperature
level suggests that the survivors are under severe
stress, leaving them unable to compete success-
fully for limited resources or avoid predation. For
acclimation temperatures of 8°, 12°, 16°, 20°, and
24°C, a range of temperatures encountered by
juveniles and adults in nature, immediate ex-
posure to high temperatures less than 23.0°, 24.0°,
25.5°, 26.5°, and 27.5°C, respectively, would be
tolerated by fish from southern California without
significant mortality from the direct effects of
temperature alone. Likewise, for the same accli-
mation temperatures, juvenile and adult anchovy
could tolerate lows of 7.5°, 10.0°, 12.5°, 13.5°, and
14.5°C, respectively. Larvae in the yolk-sac stage
can tolerate limited exposure (24 h) to any tem-
perature <28.0°C and >12.0°C. Regardless of ac-
climation temperature, larvae in the yolk-sac
stage, juveniles, and adults can endure sudden
temperature increases and decreases between the
limits of 14.5° and 23.0°C wdthout significant le-
thality from direct temperature effects alone.

Although the gross effects of high and low tem-
perature extremes have been quantified, the
physiological and biochemical factors that are
responsible for thermal death and temperature
acclimation are poorly understood. Various
mechanisms to account for these phenomena have
been discussed by Hochachka and Somero (1971)
and Hazel and Prosser (1974). Evidence suggests
that qualitatively different enzymes (isoenzymes)
may be synthesized during thermal acclimation,
and "warm" and "cold" enzyme variants have been
described (Hochachka 1967; Hochachka and
Somero 1968; Hebb et al. 1969). Enzyme inactiva-
tion has been suggested as a cause of thermal
death, but it is ". . . undoubtedly more subtle than
gross protein denaturation" (Hochachka and
Somero 1971:139). The reaction velocities (Kr,,) of
enzymes may drop below certain critical levels at

442

BREWER: THERMAL TOLERANCE AND RESISTANCE

high and low temperatures, resulting in the dis-
ruption of basic physiological functions such as
osmoregulation, respiration, and overall nervous
system integration (Prosser 1973).

It is unlikely that the offshore realm of any
ocean could ever be significantly affected by arti-
ficial thermal input. Projected energy needs for
the decades ahead and their associated require-
ments for immense volumes of water for cooling
(electric power generating) and heating (LNG)
may pose a serious environmental threat in near-
shore areas, especially bays, harbors, and es-
tuaries. As a case-in-point, juvenile northern an-
chovy find the confined waters of the Los
Angeles-Long Beach Harbor a suitable habitat.
Brewer (1975a) found anchovy egg densities as
high as 35/m^ of surface water within 0.5 mile of
the harbor breakwater. These areas will be af-
fected by seawater intake pipes, and thermal
effluent plumes and E. mordax embryos would be
highly susceptible to entrainment. Eggs in the
blastodisc stage are most sensitive to abrupt
changes in temperature. If one considers the high
temperature extremes where mortality begins to
exceed the control mortality as unsafe, anchovy
embryos should not be allowed to remain in tem-
peratures of 35.5°, 30.5°, 30.0°, and 27.5°C for
periods longer than 1, 3, 5, and 60 min, respec-
tively. While embryos proved insensitive to the
effects of temperatures as low as 0.5°C for 60-min
exposures, it is questionable whether these sensi-
tive developmental stages could withstand the
turbulence and mechanical shock associated with
heat exchange systems or thermal effluent out-
falls. In this respect, larvae are most vulnerable,
and Lasker (1964) found this vulnerability in-
creased with decreasing temperatures below 14°C
for Pacific sardine, Sardinops, larvae which are
morphologically similar to anchovy larvae. Their
thin integument and fragile bodies are easily
damaged. Extreme care was taken in the present
study when the larvae were transferred from in-
cubation to test jars, but control survival was only
77.5%. Survival of larvae in experiments that did
not involve transfer to rearing vessels was over
90%. Serious consideration must therefore be
given to the location of intake pipes and effluent
discharge to avoid trapping eggs and larvae. These
stages are probably too small to be excluded by
screening.

Many more experiments are required to under-
stand the dynamics of the thermal requirements of
E. mordax. It may be unreasonable to assume that

there is one optimal temperature for anchovy
well-being. Activity cycles or rhythms (e.g., the
evening spawning cycle) may be present in
natural populations which require diel tempera-
ture changes (e.g., achieved through vertical mi-
gration). Temperature optima for reproduction or
the gi'owth of larvae in the yolk-sac stage may
differ from optima for growth of juveniles and
adults which must respond to fluctuating food
levels. Brett et al. (1969), experimenting with On-
chorhynchus nerka, found that as food rations
were decreased, temperatures required for
maximum growth rates also decreased. When food
rations were not limiting, growth rates increased
as the temperature increased to a certain optimal
level, after which growth rates decreased rapidly.

In conclusion, the potential responses of the
northern anchovy to temperature are many and
varied. They depend upon the degree and rate of
temperature change, length of exposure to a par-
ticular temperature, the previous thermal experi-
ence of the fish, and the effects of interactions
among other environmental variables, both biotic
and abiotic. Furthermore, these responses vary
with ontogeny.

Although expatriated individuals may tem-
porarily tolerate environmental extremes, the dis-
tribution and survival of E. mordax are ultimately
dependent upon those physicochemical charac-
teristics of the environment conducive to spawn-
ing. For the present, such an environment is best
described as that part of the California Current
where surface water temperatures reach 13°-18°C
during at least part of the year (Ahlstrom 1956,
1959, 1966, 1967; Richardson 1973).

ACKNOWLEDGMENTS

The discussions and criticisms of Basil G. Naf-
paktitis, Gerald J. Bakus, John E. Fitch, Bernard
W. Pipkin, and Camm C. Swift are gratefully
acknowledged. Without the interest and coopera-
tion of William Verna, a live-bait dealer at Long
Beach, Calif., this study would not have been pos-
sible. Special thanks go to Dorothy Soule and
Mikihiko Oguri for their support and confidence.

The work was funded, in part, by NOAA-Sea
Grant (No. 04-3-158-36 and 04-3-158-45); the
Army Corps of Engineers; the Resources Agency,
State of California; the Pacific Lighting Service
Company; and a Grant-in-Aid of Research from
the Society of Sigma Xi.

443

LITERATURE CITED

AHLSTROM, E. H.

1956. Eggs and larvae of anchovy, jack mackerel, and
Pacific mackerel. Calif. Coop. Oceanic Fish. Invest.,
Prog. Rep., 1 Apr. 1955 to June 1956:33-42.
1959. Vertical distribution of pelagic fish eggs and larvae
off California and Baja California. U.S. Fish Wildl. Serv.,
Fish Bull. 60:107-146.

1966. Distribution and abundance of sardine and anchovy
larvae in the California Current region off California and
Baja California, 1954-64: A summary. U.S. Fish Wildl.
Serv., Spec. Sci. Rep. Fish. 534, 71 p.

1967. Co-occurrences of sardine and anchovy larvae in the
California Current region off California and Baja
California. Calif. Coop. Oceanic Fish. Invest. Rep.
11:117-135.

ALABASTER, J. S., AND K. G. ROBERTSON.

1961. The effect of diurnal changes in temperature, dis-
solved oxygen, and illumination on the behaviour of the
roach (Rutilus rutilus (L. )), bream {Abramis brama (L.))
and perch iPerca ftuviatilis (L.)). Anim. Behav. 9:187-
196.
ALLEN, K. O., AND E. STRAWN.

1971. Rate of acclimation of juvenile channel catfish, Ic-
talurus punctatus, to high temperatures. Trans. Am.
Fish. Soc. 100:665-671.
BAXTER, J. L.

1967. Summary of biological information on the northern
anchovy Engraulis mordax Girard. Calif Coop. Oceanic
Fish. Invest. Rep. 11:110-116.

Brawn, V. M.

1960. Temperature tolerance of unacclimated herring
{Clupea harengus L.). J. Fish. Res. Board Can. 17:721-
723.

Brett, j. r.

1944. Some lethal temperature relations of Algonquin
Park fishes. Publ. Ont. Fish. Res. Lab. 63:1-49.

1952. Temperature tolerance in young Pacific salmon,
genus Oncorhynchus. J. Fish. Res. Board Can. 9:265-323.

1970. Temperature. In O. Kinne (editor). Marine Ecology,
Vol. 1, Part 1, p. 515-560. Wiley-Interscience, Lond.

Brett, J. R., j. e. Shelbourne, and C. T. Shoop.

1969. Growth rate and body composition of fingerling
sockeye salmon, Oncorhynchus nerka, in relation to
temperature and ration size. J. Fish. Res. Board Can.
26:2363-2394.

Brewer, G. d.

1974. Preliminary observations on the lower minimum
temperature requirements of the northern anchovy. In
D. Soule and M. Oguri (editors). Marine studies of San
Pedro Bay, California, Part III, p. 21-43. Allan Hancock
Found. Publ., USC-SG-1-74.

1975a. The biology and fishery of the northern anchovy in
San Pedro Bay: Potential impact of dredging and landfill.
In D. Soule and M. Oguri (editors), Marine Studies of San
Pedro Bay, California, Part VIII, p. 23-43. Allan Hancock
Found. Publ., USC-SG-1-74.

1975b. The biology of the northern anchovy (Engraulis
mordax), in relation to temperature. Ph.D. Thesis, Univ.
Southern California, 205 p.
BULL, H. O.

1928. Studies on conditioned responses in fishes. Part I. J.
Mar. Biol. Assoc. U.K. 15:485-533.

444

fishery bulletin: vol. 74, no. 2
California Department of Fish and Game.

1971. Pelagic wetfish. Northern anchovy. In H. W. Frey
(editor), California's living marine resources and their
utilization, p. 48-51. Calif. Dep. Fish Game, Resour.
Agency.
COUTANT, C. C.

1969. Temperature, reproduction and behavior.
Chesapeake Sci. 10:261-274.
DE CIECHOMSKI, j. D.

1973. The size of the eggs of the Argentine anchovy En-
graulis anchoita (Hubbs & Marini) in relation to the sea-
son of the year and to the area of spawning. J. Fish Biol.
5:393-398.
DE SYLVA, D. p.

1969. Theoretical considerations of the effects of heated
effluents on marine fishes. In P. A. Krenkel and F. L.
Parker (editors), Biological aspects of thermal pollution,
p. 229-293. Vanderbilt Univ. Press, Portland, Oreg.

Doudoroff, p.

1938. Reactions of marine fishes to temperature gradients.
Biol. Bull. (Woods Hole) 75:494-509.
FRY, F. E. J.

1947. Effects of the environment on animal activity. Univ.

Toronto Stud., Biol. Ser. 55:1-62.
1971. The effect of environmental factors on the physiol-
ogy offish. In W. S. Hoar and D. J. Randall (editors), Fish
physiology. Vol. 6, p. 1-98. Academic Press, N.Y.
FRY, F. E. J., J. R. BRETT, AND G. H. CLAWSON.

1942. Lethal limits of temperature for young goldfish. Rev.
Can. Biol. 1:50-56.
HAZEL, J. R., AND C. L. PROSSER.

1974. Molecular mechanisms of temperature compensa-
tion in poikilotherms. Physiol. Rev. 54:620-677.
HEATH, W. G.

1963. Thermoperiodism in sea-run cutthroat trout (Salmo
clarki clarki). Science (Wash., D.C.) 142:486-488.

Hebb, C, D. Morris, and M. w. Smith.

1969. Choline acetyltransferase activity in the brain of
goldfish acclimated to different temperatures. Comp.
Biochem. Physiol. 28:29-36.
HOAR, W. S.

1955. Seasonal variations in the resistance of goldfish to
temperature. Trans. R. Soc. Can., Ser. Ill 49(5):25-34.

1956. Photoperiodism and thermal resistance of goldfish.
Nature (Lond.) 178:364-365.

HOCHACHKA, P. W.

1967. Organization of metabolism during temperature
compensation. In C. L. Prosser (editor). Molecular
mechanisms of temperature adaptation, p. 177-203. Am.
Assoc. Adv. Sci. Publ. 84.

HOCHACHKA, P. W., AND G. N. SOMERO.

1968. The adaptation of enzymes to temperature. Comp.
Biochem. Physiol. 27:659-668.

1971. Biochemical adaptation to the environment. In W. S.
Hoar and D. J. Randall (editors). Fish physiology. Vol. 6,
p. 99-156. Academic Press, N.Y.

Hubbs, C, and C. Bryan.

1974. Effect of parental temperature experience on ther-
mal tolerance of eggs oiMenidia audens. In J. H. S. Blax-
ter (editor). The early life history of fish, p. 431-435.
Springer- Verlag, N.Y.

Hunter, j. r.

1972. Swimming and feeding behavior of larval anchovy,
Engraulis mordax. Fish. Bull., U.S. 70:821-838.

BREWER: THERMAL TOLERANCE AND RESISTANCE

LASKER, R.

1964. An experimental study of the effect of temperature
on the incubation time, development, and growth of
Pacific sardine embryos and larvae. Copeia 1964:399-
405.

LASKER, R., H. M. FEDER, G. H. THEILACKER, AND R. C. MAY.
1970. Feeding, growth, and survival of Engraulis mordax
larvae reared in the laboratory. Mar. Biol. (Berl.) 5:345-
353.

Lynn, R. J.

1967. Seasonal variation of temperature and salinity at 10
meters in the California Current. Calif. Coop. Oceanic
Fish Invest. Rep. 11:157-187.
MAIS, K. R.

1974. Pelagic fish surveys in the California Current. Calif.
Dep. Fish Game, Fish Bull. 162:1-79.
MCHUGH, J. L.

1951. Meristic variations and populations of northern an-
chovy {Engraulis mordax mordax). Bull. Scripps Inst.
Oceanogr. 6:123-160.
MESSERSMITH, J. D., J. L. BAXTER, AND P. M. ROEDEL.

1969. The anchovy resources of the California Current re-
gion off California and Baja California. Calif. Coop.
Oceanic Fish Invest. Rep. 13:32-38.
NAYLOR, E.

1965. Effects of heated effluents upon marine and estua-
rine organisms. Adv. Mar. Biol. 3:63-103.

PROSSER, C. L. (editor).

1973. Comparative animal physiology, 3rd ed. W. B.
Saunders, Phila., 966 p.

REID, J. L., Jr.

1967. Ocean environments of the genus Engraulis around
the world. Calif Coop. Oceanic Fish. Invest. Rep. 11:29-
33.
RICHARDSON, S. L.

1973. Abundance and distribution of larval fishes in wa-
ters off Oregon, May-October 1969, with special em-
phasis on the northern anchovy, Engraulis mordax. Fish.
Bull., U.S. 71:697-711.
SCOTT, D. P.

1964. Thermal resistance of pike (Esox lucius L.), muskel-
lunge (E. masquinongy Mitchill), and their Fj hybrid. J.
Fish. Res. Board Can. 21:1043-1049.
SMITH, P. E.

1972. The increase in spawning biomass of northern an-
chovy, Engraulis mordax. Fish. Bull., U.S. 70:849-874.
SOKAL, R. R., AND F. J. ROHLF.

1969. Biometry. The principles and practice of statistics in
biological research. W. H. Freeman and Co., San Franc,
776 p.

Tarzwell, cm.

1972. An argument for the open ocean siting of coastal
thermal electric plants. J. Environ. Qual. 1:89-91.

Tyler, A. V.

1966. Some lethal temperature relations of two minnows
of the genus Chrosomus. Can. J. Zool. 44:349-364.
VROOMAN, a. M., AND P. E. SMITH.

1971. Biomass of the subpopulations of the northern an-
chovy Engraulis mordax Girard. Calif. Coop. Oceanic
Fish. Invest. Rep. 15:49-51.

445

NOTES

ISOLATION AND DESCRIPTION OF TWO

VIBRIOS PATHOGENIC TO PACIFIC

SALMON IN PUGET SOUND, WASHINGTON

Vibrio anguillarum (Bergman 1909) is recognized
worldwide as a saltwater pathogen in fish (An-
derson and Conroy 1970). Most epizootics caused
by marine bacteria have been attributed to this
organism (Rucker 1959; Sindermann 1966). This
note describes recent mortalities resulting from
vibriosis of Pacific salmon in the marine waters of
Puget Sound, Wash., and heterogeneity observed
in vibrios isolated from diseased fish.

The National Marine Fisheries Service
(NMFS) is engaged in the experimental culture of
Pacific salmon in salt water at the NMFS Aqua-
culture Experiment Station near Manchester,
Wash. Epizootics caused by marine vibrios have
occurred regularly in cultured salmon during the
spring and summer months; the organisms were
also isolated from diseased fish on a minor scale
in every month during fall and winter (Novotny
1975). Vibrios originally isolated from diseased
fish at Manchester were typical of Vibrio anguil-
larum (Evelyn 1971); strain 775 was represen-
tative.

In November 1973, a commercial salmon farm
in the Manchester area suffered a high mortality
of pen-reared, 0-age, 250-g coho salmon, On-
corhynchus kisutch. Past experience with vib-
riosis in the area indicated that the first serious
outbreaks usually began in April when water
temperatures exceeded 9°C and continued until
water temperatures dropped below 12°C in early
October (Novotny 1975). Water temperatures in
November 1973 were 10° to 11°C; therefore, prob-
lems from vibriosis were not anticipated.

Mortalities also began to occur at about the
same time, although not on an epizootic scale, in
coho salmon held at the NMFS facility at Man-
chester. These fish had been vaccinated in late
spring by injecting a heat-killed bacterin pre-
pared from V anguillarum lib. Oral antibiotics
were administered, but the period required to
bring the disease under control appeared to be
almost twice that usually required for V. an-
guillarum.

Diseased fish sampled from the NMFS pens and

the commercial farm exhibited the common signs
of vibriosis, most notably a hemorrhagic sep-
ticemia. Bacteria characterized as vibrios were
consistently isolated from dead or dying fish, but
the growth rate of the isolated bacteria was mar-
kedly different from that of the typical V. anguil-
larum. Also, this bacterium was not agglutinated
by rabbit anti-V anguillarum lib serum in rapid
slide agglutination tests.

The new isolates were confirmed as pathogens
by injecting pure cultures of them into salmon.
All the injected fish died and the organism was
routinely re-isolated from kidneys. We desig-
nated this bacterium as Vibrio sp. 1669.

In June 1974, NMFS conducted cooperative
vaccination tests with a second commercial salm-
on farm in the Manchester area. Approximately
280,000 coho salmon smolts were injected with a
heat-killed bacterin of V. anguillarum 115 at
least 2 wk prior to their transfer to saltwater
pens. Mortalities were exceptionally low until
late August (less than 6% from all causes and less
than 2% from vibriosis). At that time the rate of
mortality began to increase and Vibrio sp. 1669
was isolated.

Further tests were made in early August 1974,
when 450 0-age sockeye salmon, O. nerka, smolts
were transferred to NMFS saltwater pens. One
pen contained 150 unvaccinated control fish, and
two pens contained 150 fish each that had been
vaccinated in fresh water with a heat-killed V. an-
guillarum lib bacterin. After 50 days in the
saltwater pens, 95% of the unvaccinated fish had
died. During the same period the mortalities in
the vaccinated lots were 9% and 27% (Figure 1).
Vibrios isolated from the vaccinated fish were
only of the 1669 type, based on results of slide
agglutination tests.

Materials and Methods

Samples of kidney, eye, or spleen from freshly
dead or moribund fish were streaked on trypti-
case soy agar (TSA) (Difco)i with 1% salt added,
or on 50% seawater cytophaga agar (Pacha and

'Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

447

100

Figure l. — Comparative rate of
mortality of three lots of 0-age sock-
eye salmon raised in net pens at
Manchester, Wash. Two of the lots
were vaccinated with an intraperi-
toneal injection of a heat-killed
bacterin prepared from Vibrio anguil-
larum 775.

Ordal 1967). Plates were incubated aerobically at
23°C. Presumptive identifications of the bacteria
were based on the following tests: gram stain,
motility and morphology characteristics under
phase contrast microscopy, oxidase test (Kovacs),
fermentation or oxidation of glucose, and sen-
sitivity to the vibriostatic compound 0/129 (2,4-
diamino-6,7-diisopropyl pteridine phosphate).

Further biochemical characterization included
tests in Moeller's media for an alkaline reaction
with arginine and for lysine decarboxylase, the
production of indole, the production of acetyl-
methylcarbinol (Voges-Proskauer test), and the
ability to ferment arabinose, glycerol, mannitol,
sucrose, and galactose. These tests were selected
because they were found to be variable among
marine vibrio groups established by deoxyribo-
nucleic acid homology characteristics (E. J. Or-
dal, University of Washington School of Medicine,
Seattle, pers. commun.). In all of these tests addi-
tional NaCl (1%) was added.

Table l. — Selected properties of Vibrio anguillarum 775 and
Vibrio sp. 1669.

V. anguillarum

Vibno sp

Property

775

1669

Gram reaction

Motility

Oxidase (Kovacs) test

-1-

Fermentative (glucose)

Gas from glucose

Moeller's media;

Arginine-alkaline reaction

Lysine decarboxylase test

Indole production

Voges-Proskauer reaction

Acid from:

Arabinose

Glycerol

Mannitol

Sucrose

Galactose

Antisera for serological comparisons were pre-
pared in both rabbits and coho salmon with
heat-killed bacterins of V. anguillarum lib and
Vibrio sp. 1669 in Freund's complete adjuvant.
Rapid slide agglutination tests with the specific
antisera were used for initial differentiation. The
microtiter system (Cooke Engineering Co.) was
used later to determine agglutinin titers, and
immunodiffusion techniques were used to further
compare antigenic structure and relatedness.
Tests were run with unabsorbed antisera and
with anti-Vibrio sp. 1669 sera absorbed with V.
anguillarum lib.

Results and Discussion

Vibrio sp. 1669 was typical of the marine vibrio
group: it was characterized as a gram negative,
motile, curved, asporogenous rod that was oxi-
dase positive, an anaerogenic fermenter, and sen-
sitive to the vibriostatic compound 0/129. A
slower rate of growth of Vibrio sp. 1669, in com-
parison to V. anguillarum lib, was observed on
TSA, as well as variations in certain culture reac-
tions (Table 1).

Coho salmon anti-V. anguillarum lib serum
with an agglutinin titer of 512 against the ho-
mologous bacterium had a titer of 8 against Vib-
rio sp. 1669. Immunodiffusion also revealed dif-
ferences between the two vibrios. In Figure 2, the
inner precipitin lines demonstrate antigenic
cross-reactivity (reaction of identity). An addi-
tional antigen unique to V. anguillarum lib is
demonstrated by the outer precipitin line which is
not present in reactions with Vibrio sp. 1669.

After all detectable agglutinin activity against

448

Figure 2. — Immunodiffusion comparison of Vibrio anguil-
larum 775 and Vibrio sp. 1669. Wells 1, 3, and 5 contain V. an-
guillarum 775 sonicate and wells 2, 4, and 6 contain Vibrio sp.
1669 sonicate. The center well contains rabbit anti-V. anguil-
larum 775 serum.

V. anguillarum lib in rabbit anti-Viftrto sp. 1669
serum was removed by absorption, a titer of 16
against 1669 remained (Table 2), indicating that
Vibrio sp. 1669 also contains antigenic determin-
ants not present on V. anguillarum 115.

Whether a vaccine containing antigens from
both vibrios would be more protective than vac-
cines containing antigens from only one of the

Table 2. — Agglutinin titers of rabbit anti-Vj6rto sp. 1669
serum unabsorbed and absorbed with V. anguillarum 775
antigen.

Titer

Condition

775

1669

Unabsorbed anti-1669 serum
Anti-1669 serum absorbed with 775

8
0

32
16

vibrios is not known. This possibility is currently
being investigated. Deoxyribonucleic acid homol-
ogy experiments are also in progress to better
clarify the taxonomic relation of the two vibrios.

Literature Cited

ANDERSON. J. I. W., AND D. A. CONROY.

1970. Vibrio disease in marine fishes. In S. F. Snieszko
(editor), A Symposium on Diseases of Fishes and
Shellfishes, p. 266-272. Am. Fish. Soc., Spec. Publ. 5.

EVELYN, T. P. T.

1971. First records of vibriosis in Pacific salmon cultured
in Canada, and taxonomic status of the responsible bac-
terium, Vibrio anguillarum. J. Fish. Res. Board Can.
28:517-525.

NOVOTNY. A. J.

1975. Net-pen culture of Pacific salmon in marine waters.
Mar. Fish. Rev. 37(l):36-47.

Pacha, R. e., and E. j. Ordal.

1967. Histopathology of experimental columnaris disease
in young salmon. J. Comp. Pathol. 77:419-423.
RUCKER, R. R.

1959. Vibrio infections among marine emd fresh-water
fish. Prog. Fish-Cult. 21:22-25.
SiNDERMANN, C. J.

1966. Diseases of marine fishes. Adv. Mar. Biol. 4:1-89.

Lee W. Harrell

anthony j. novotny

Michael H. Schiewe

Harold O. Hodgins

Northwest Fisheries Center
National Marine Fisheries Service, NOAA
2725 Montlake Boulevard East
Seattle, WA 98112

RELATION OF FISH CATCHES IN
GILL NETS TO FRONTAL PERIODS

A study was conducted in 1972 relating gill net
catches of fishes to webbing material, time of day,
and water depth in St. Andrew Bay, Fla. (Pristas
and Trent^). While conducting the study, Pristas
and Trent observed that catches in the nets ap-
peared to be greater when atmospheric fronts
moved through the area in the autumn. We de-
cided to test the hypothesis that catches of fishes
in gill nets increase during fi-ontal periods. Ex-
perimental data were collected in September-
December 1973, and the results of the analysis
are presented in this paper.

'Pristas, P. J., and L. Trent. 1974. Comparisons of catches of
fishes in gill nets in relation to webbing material, time of day,
and water depth in St. Andrew Bay, Florida. Unpubl. manuscr.

449

Study Area and Methods

The study area was located about 300-800 m
northwest of Courtney Point in St. Andrew Bay
(Figure 1). Hydrological, physical, and sedimen-
tological characteristics of the bay system were
described by Ichiye and Jones (1961), Waller
(1961), and Hopkins (1966). The bay system ex-
changes water with the Gulf of Mexico through
East and West passes (Figure 1). Prevailing
winds are from the southwest in the summer,
north and northeast in the autumn, and north
and southeast in the winter and spring. Tides are
usually diurnal with a mean range of about 0.4 m
in St. Andrew Bay (U.S. Department of Com-
merce 1967).

Eleven gill nets of different mesh sizes were
fished for 87 consecutive days from 17 September
to 13 December 1973. Each net was 33.3 m long
and 3.3 m deep. Stretched mesh sizes ranged from
6.4 to 12.7 cm, the mesh sizes increasing by
0.6-cm increments. The nets were made of #208
monofilament nylon webbing hung to the float
and lead lines on the half basis (two lengths of
stretched mesh to one length of float line).

Nets were set parallel to each other about 50 m
apart, perpendicular to shore, and in water
depths (mean low tide) of 2.2 to 2.6 m (Figure 1).
Nets remained in the water continuously except
for 12 brief periods when they were randomly
reset among net locations during the 87-day
period. Damaged webbing never exceeded 5% of
the total surface area of each net.

Fishes were removed from the nets at sunrise
±2 h and occasionally at sunset ±1 h. The total
number of each species caught, including the

ST. ANDREW BAY

\ 11

9 ^f.

'■■■i- i ' .

^ GUIF OF MEXICO

** 8S*'40W

DEPTH CONTOUR

MAGNOLIA BEACH

damaged specimens, was counted. Lengths of the
undamaged specimens were determined on a
measuring board to the nearest 0.5 cm in fork
length (tip of snout to fork of tail) for those fishes
having forked tails and in total length (tip of
snout to extremity of caudal fin) for Atlantic
croaker and sharks.

Total catch and catches of each of the 10 most
abundant species per 24-h period (catches per
day) during and between frontal periods were
compared using a t-test for unpaired observations
(Steel and Torrie 1960). We tested the hypothesis
that the mean catch during frontal periods (n =
23) equaled the mean catch between frontal
periods in = 64). We also used the t-test to test the
hypothesis that the mean lengths of each of the
10 most abundant species caught during and be-
tween frontal periods were equal.

Water temperature was recorded continuously
by a Peabody-Ryan^ thermograph (Model F; accu-
rate within 2% on time and temperature) about 1
m below the water surface at a dockside location
about 100 m from the south end of the study area.
Mean water temperatures per 24-h period were
computed from readings taken every 6 h from the
continuous data. Air temperatures, measured
hourly, were obtained from the weather station at
Tyndall Air Force Base located about 13 km east
of the study area. Air and water temperatures
were averaged over a 24-h period ending at 0600
h. Changes in water temperature per 24-h period
were determined from these means.

Species and Numbers of Fish Caught

A total of 15,398 individuals representing at
least 65 species (not all species o{ Sphyrna and
none o^ Scorpaena were specifically identified) of
marine fishes was caught during the study (Table
1). Catch per day ranged from 10 to 967 individu-
als and from 6 to 25 species; increases and de-
creases in the total number of fish caught per day
were generally accompanied by similar changes
in the number of species of fish caught per day
(Figure 2).

The 10 most abundant species comprised 88%
of the total catch. The 10 were: Gulf menhaden,
Brevoortia patronus; spot, Leiostomus xanthurus;
Atlantic croaker, Micropogon undulatus; pinfish,
Lagodon rhomboides; pigfish, Orthopristis

Figure l . — Study area and net locations in St. Andrew Bay, Fla .

450

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Table l. — Species and numbers of fish caught in gill nets
during September-December 1973 in St. Andrew Bay, Fla.

Species

Number
caught

Gulf menhaden, Brevoortia patronus 3,467

Spot, Leiostomus xanthurus 2,504

Atlantic croaker, Micropogon undulatus 2,335

Pinfish, Lagodon rhomboides 1,661

Pigfish, Orthopristis chrysoptera 905

Sea catfish, Arius felis 853

Bluefish, Pomatomus saltatrix 594

Spanish mackerel, Scomberomorus maculatus 563

Yellowfin menhaden, Brevoortia smithi 473

Gafftopsail catfish, Bagre mannus 239

Crevalle jack, Caranx hippos 212

Blue runner. Caranx crysos 211

Little tunny, Euthynnus alletteratus 170

Inshore lizardfish, Synodus foetens 123

Atlantic sharpnose shark, Rhizoprionodon terraenovae 94

Bonnethead, Sphyrna tiburo 91

Gulf flounder, Paralichthys albigutta 89

Florida pompano, Tracliinotus carolinus 86

Atlantic bumper, Chloroscombrus chrysurus 78

Ladyfish, Elops saurus 74

Cobia, Rachycentron canadum 46

Blacktip shark, Carcharhinus limbatus 40

Blacknose shark, Carcharhinus acronotus 39

Harvestfish, Peprilus alepidotus 34

Yellow jack, Caranx bartholomaei 34

Remora, Remora remora 32

Hybrid menhaden, Brevoortia smithi > patronus 29

Sand seatrout, Cynoscion arenarius 29

Skipjack herring. Alosa chrysochloris 28

Bighead searobin, Pnonotus tribulus 22

Spotted seatrout, Cynoscion nebulosus 22

Striped mullet, Mugil cephalus 22

Leatherjacket, Oligoplites saurus 22

Atlantic thread herring, Opisthonema oglinum 17

Longnose gar, Lepisosteus osseus 16

Florida smoothhound, Mustelus norrisi 15

Black drum, Pogonias cromis 12

Alabama shad, ,4/osa alabamae 11

Gray snapper, Lutianus griseus 10

Atlantic spadefish, Chaetodipterus laber 10

Southern sea bass, Centropristis melana 8

Atlantic threadfin, Polydactylus octonemus 7

Finetooth shark, Aprionodon isodon 7

Sheepshead, Archosargus probatocephalus 6

Gulf toadfish, Opsanus beta 6

Orange filefish, Alulerus schoepfi 5

Gag, Mycteroperca microlepis 5

Sand perch, Diplectrum formosum 5

Atlantic moonfjsh. Vomer setapinnis 5

Hogchoker, Trmectes maculatus 4

White mullet, Mugil curema 4

Hammerhead shark, Sphyrna sp. 3

Southern stargazer, Astroscopus y-graecum 3

Smooth dogfish. Mustelus canis 3

Scorpionfish, Scorpaena sp. 2

Guaguanche, Sphyraena guachancho 2

Striped burrfish, Chilomycterus schoepfi 2

Dusky flounder, Syacium papillosum 2
Tarpon, Megalops atlantica
Bull shark, Carcharhinus leucas
Tripletail, Lobotes sunnamensis
Shrimp eel, Ophichthus gomesi
Sandbar shark, Carcharhinus milberti
Bonefish, Albula vulpes
Halfbeak. Hyporhamphus unifasciatus

Total 15,398

chrysoptera; sea catfish, Arius felis; bluefish,
Pomatomus saltatrix; Spanish mackerel, Scom-
beromorus maculatus; yellowfin menhaden, Bre-
voortia smithi; and gafftopsail catfish, Bagre
marinus (Table 1). Catches per day of each of
these are shown in Figure 3.

Figure 2. — Frontal periods, mean air and water temperatures,
and numbers of species and individuals caught per 24-h period
in St. Andrew Bay, Fla., September-December 1973.

Figure 3. — Frontal periods and number of individuals caught
by species per 24-h period in St. Andrew Bay, Fla., September-
December 1973.

451

Frontal Periods

A frontal period was arbitrarily defined as any
four consecutive days the first of which the water
temperature dropped 2°C or more. Four days
were selected, because fish catches were gener-
ally affected for 2 to 4 days following the initial
temperature drop on the first day of a frontal
period. Six frontal periods occurred in the study
area from 17 September to 13 December (Figure
2). Fronts moved through the study area on 17
October, 28 October, 9 November, 27 November, 5
December, and 10 December (Figure 2). The av-
erage decrease of water and air temperatures per
24-h period for the above dates was 2.5°C and
6.4°C, respectively. In addition to decreases of
temperatures, fronts passing through estuaries of
the northern Gulf of Mexico are also charac-
terized by: 1) rapid changes in barometric pres-
sure, 2) shifts in wind direction and wind speed,
3) changes in tidal heights, and 4) increases in
turbidity and velocity of tidal currents (E. J. Pul-
len, pers. commun., U.S. Corps of Engineers, Gal-
veston, Tex.).

Catch Related to Frontal Periods

Spanish mackerel (Table 2, Figure 3) were the
exceptions. Spanish mackerel was the only
species caught in greatest numbers between fron-
tal periods. Mean catches of the nine species
ranged from 1.7 to 9.5 times greater during fron-
tal periods than between frontal periods.

Mean lengths of fish caught during frontal
periods were not significantly different from
those caught between frontal periods for each of
the 10 most abundant species (Table 2).

These results suggest that many species of
marine fishes become more vulnerable to capture
by gill nets in shallow areas of coastal bays dur-
ing frontal periods in autumn. This increased
vulnerability probably results from increased ac-
tivity, migration, a lessening ability to avoid the
net, and one or more of the factors associated with
fronts, e.g., changes in temperature, tidal height,
turbidity, and current velocity.

Acknowledgments

Our sincere appreciation is extended to J. R.
Lara for fiarnishing climatological data and to D.
B. Jester, M. A. Roessler, and J. Y. Christmas for
their helpful comments.

Each front was characterized by a marked in-
crease in the numbers of individuals caught. Such
a marked increase occurred only once (22-24
November) during a nonfrontal period (Figure 2).
The mean number (all species combined) of fish
caught per day was 354.7 during frontal periods
and 113.1 between frontal periods (Table 2). Mean
catches were significantly higher during frontal
periods for all species combined and for 8 of the 10
most abundant species. Atlantic croaker and

Literature Cited

Hopkins, T. L.

1966. The plankton of the St. Andrew Bay system, Florida.
Publ. Inst. Mar. Sci., Univ. Tex. 11:12-64.

ICHIYE, T., AND M. L. Jones.

1961. On the hydrography of the St. Andrew Bay system,
Florida. Limnol. Oceanogr. 6:302-311.

Steel, R. G. D., and J. H. Torrie.

I960. Principles and procedures of statistics with special
reference to biological sciences. McGraw-Hill, N.Y.,
481 p.

Table 2.— Comparisons of mean catches per day and mean lengths during and between frontal periods, September-December

1973, St. Andrew Bay, Fla.

Mean numbei

• caught per day

Mean

length (cm)

Species group

During frontal

Between frontal

During frontal

Between frontal

or species

periods

f-value

periods

f-value

All fish

354.7

113.1

-6.60"

(')

Gulf menhaden

90.4

21.7

-3.46"

21.0

21.5

1.26

Spot

81.7

9.8

-4.66"

20.2

19.6

-1.85

Atlantic croaker

38.6

22.6

-1.43

26.2

25.6

-1.06

Pinfish

41.6

11.0

-4.46"

17.0

16.5

-0.64

Pigfish

30.4

-5.28"

18.2

18.9

1.36

Sea catfish

16.7

7.3

-5.68"

30.2

309

0.78

Bluefish

10.4

5.5

-2.74"

33.6

35.9

0.89

Spanish mackerel

5.0

7.0

0.70

34.9

36.9

1.15

Yellowfin menhaden

10.5

-2.22-

25.8

26.0

0.65

Gafftopsail catfish

5.0

2.0

-3.98"

42.7

44.2

1 06

'Not determined.

'Significant at 5%

level.

"Significant at 1%

level.

452

U.S. Department of Commerce.

1967. U.S. Coast Pilot 5, Atlantic Coast, 301 p.
Waller, R. A.

1961. Ostracods of the St. Andrew Bay system. M.S
Thesis, Florida State Univ., Tallahassee, 46 p.

nelson may

Lee Trent

Paul J. Pristas

Gulf Coastal Fisheries Center Panama City Laboratory
National Marine Fisheries Service, NOAA
Panama City, FL 32401

PHOSPHOGLUCOMUTASE POLYMORPHISM
IN TWO PENAEID SHRIMPS,
PENAEUS BRASILIENSIS AND
PENAEUS AZTECUS SUBTILIS

In a search for subpopulation differences within
species of penaeid shrimp in the northern Gulf of
Mexico, Proctor et al. (1974) and Marvin and
Caillouet (1976) reported genetically con-
trolled polymorphism in the enzyme phospho-
glucomutase (PGM) in Penaeus aztecus (brown
shrimp) and P. setiferus (white shrimp). The
brown shrimp were collected in the northern Gulf
of Mexico, so they are P. aztecus aztecus Ives, ac-
cording to Perez Farfante (1969). The white
shrimp, collected both from the northern Gulf and

from the North Edisto River, S.C., are P. setiferus
(Linnaeus), according to Perez Farfante (1969).
Our paper describes similar polymorphisms in
PGM in two more penaeids, P. brasiliensis Lat-
reille and P. aztecus subtilis Perez Farfante.

Methods

Specimens were collected off the coasts of
Guyana, Surinam, and French Guiana, South
America, on cruise 49 of the Oregon II, between
lat. 6°13' and 6°29'N and between long. 53°10'
and 53°36'W, at 22-29 fathoms, on 9 and 10 Feb-
ruary 1974. They were stored at -20°C or below
until analyzed. Preparation of abdominal muscle
extracts, electropherograms of general protein
patterns, and PGM zymograms followed pro-
cedures used by Procter et al. (1974). Each speci-
men was identified to species by morphological
characteristics, then their distinctive general
protein patterns (Figure 1) were used to confirm
this identification. To do so, each gel was sliced
horizontally into two halves after electrophoresis
was complete. One half was treated with PGM
specific stain and the other half was stained with
Coomassie Blue.^ Specimens of P. aztecus aztecus

'Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

P. brasiliensis

P. aztecus aztecu

P. setiferus

P. aztecus subtilis

Figure l. — Electropherc^ram showing general protein pattern of Penaeus brasiliensis, P. aztecus subtilis, P. aztecus aztecus, and
P. setiferus. Stain used was Coomassie Blue. Direction (f) of protein migration toward the anode (+) is shown.

453

and P. setiferus collected in the northern Gulf
were included for comparison with P. aztecus sub-
tilis and P. brasiliensis.

Results and Discussion

In P. aztecus subtilis the zymograms of abdomi-
nal muscle extracts exhibited a single region of
PGM activity composed of five anodal bands
which are labelled a, b, c, d, and e. The same was
true for P. brasiliensis with the exception that
band e was not observed. Bands a, b, c, and d are
shown in Figure 2 and bands b, c, and d in Figure
3. Band e, observed in P. aztecus subtilis, is shown
only diagrammatically (Figure 3). Direct com-
parison of PGM bands among P. brasiliensis, P.
aztecus subtilis, P. aztecus aztecus, and P. setiferus
suggested that bands b, c, and d are similar in
these shrimps (Figure 4). This result is supported

by Marvin and Caillouet (1976) who showed
that P. setiferus and P. aztecus aztecus have the
same five PGM bands. These bands are assumed
to be under the control of five codominant allelic
genes designated PGMg through PGMg (Proctor
et al. 1974; Marvin and Caillouet 1976).

Six phenotypes of PGM were observed in P.
brasiliensis and eight in P. aztecus subtilis (Table
1). PGM phenotypes were enumerated from
zymograms to determine numerical distributions
of phenotypes, and allele (PGM band) frequencies
were derived therefi"om (Table 1). Two-banded
phenotypes (Figures 2-4) observed in some indi-
viduals presumably reflect heterozygous indi-
viduals. With PGM phenotypes grouped into
three categories, cc, ex, and xx (where x includes
bands a, b, d, and e), chi-square tests detected no
difference (P>0.05) in phenotype distribution be-
tween the sexes in either species. With the same

—

•

. ^

—

Figure 2. — A. Zymogram showing PGM bands a through d (band e not shown) and phenotypes cd, be, cc, and ac. B. Diagram
showing six PGM phenotypes observed in Penaeus brasiliensis. Direction (f) of protein migration toward the anode ( + ) is shown.

454

Figure 3. — A. Zymogram showing PGM bands b through d (tands a and e not shown) and phenotypes cd, cc and be. B. Diagram
showing eight PGM phenotypes observed inPenaeus aztecus subtilis. Direction (f) of protein migration toward the anode ( + ) is shown.

Table l. — Distribution (number of specimens) of PGM phenotypes and frequency of PGM alleles in samples ofPenaeus brasiliensis,

P. aztecus subtilis, and P. aztecus aztecus.

Total

length'
range
(mm)

Sex

Phenotypes

Alleles

Species

P. brasiliensis

145-185
151-210
145-210

Male

Female

Combined

0
0
0

2
2

1
1

1
0
1

0
0
0

172
161
333

17
14
31

0
0
0

0.0049
0.0081
0.0064

0.0343
0.0215
0.0282

0.9191
0.9301
0.9244

0.0417
0.0403
0.0410

0.0000
0,0000
0,0000

P. aztecus subtilis

102-152
107-175
102-175

Male

Female

Combined

0
1
1

0
0
0

2
0
2

0
1
1

13
13
26

2
0
2

143
119
262

0
2
2

0.0060
0.0035
0.0048

0.0452
0.0556
0.0500

0.9187
0.9062
0.9129

0.0301
0.0278
0.0290

0.0000
0.0069
0.0032

P. aztecus aztecus^

60-100

Combined

211

345

0.0025

0.2175

0.7642

0.0142

0.0017

'Tip of rostrum to tip of telson.

^Data adapted from Proctor et al. (1974).

455

p. brasiliensis P. aztecus subtilis

P. setiferus

P. aztecus aztecus

c
d

«•
%'■•*.

Figure 4. — Zymogram comparing PGM bands b through d (bands a and e not shown) in Penaeus brasiliensis, P. aztecus subtilis, P.
aztecus aztecus, and P. setiferus. Direction (f) of protein migration toward the anode (+) is shown.

phenotype categories, but with data for sexes
combined, the phenotype distribution of P. aztecus
subtilis deviated significantly (x^ = 7.086,
0.025<P<0.05) from that expected from
Hardy-Weinberg equihbrium (Stern 1943). The
reason for this deviation is not known. Johnson et
al. (1974) noted a deviation from Hardy-Weinberg
expectation for PGM phenotype distribution of a
pandahd shrimp Pandalus hypsinotus Brandt, in
Alaska, and they suggested that it might be re-
lated to depth of capture as found in, Pacific
ocean perch, Sebastodes alutus (Johnson et al.
1971).

Our study provided an opportunity to compare
the subspecies P. aztecus subtilis and P. aztecus
aztecus, therefore distribution of PGM pheno-
types and frequency of PGM alleles for the lat-
ter subspecies (data adapted from Proctor et al.
1974) also are shown in Table 1. This comparison
is based on the assumption that bands a and e as
well as bands b, c, and d are similar in the two
species. However, even if this is not the case, the
small frequencies of the rare a and e alleles would
not appreciably affect the comparison. Both sub-
species exhibited eight phenotypes, but not all
were the same. Phenotype ad was detected in P.
aztecus subtilis but not in P. aztecus aztecus.
Phenotype ac was detected in the latter but not in
the former. With phenotypes grouped into
categories cc, ex, and xx, and with sexes com-
bined, a chi-square contingency test detected a
significant (P<0.05) difference in phenotype dis-
tribution between the subspecies, and this result

provides an additional characteristic to existing
evidence of differences between these subspecies
(see Perez Farfante 1969).

This and previous studies by Proctor et al.
(1974) and Marvin and Caillouet (1976) suggest
that zymogram analysis may provide a useful
tool in the study of population genetics of the
Penaeidae. The wide distribution (Mistakidis
1968), commercial importance, and relatively
short generation time of the Penaeidae should
make them particularly attractive subjects of
study by population geneticists.

Acknowledgments

Through initial efforts by Raphael R. Proctor,
Jr., Gulf Coastal Fisheries Center, National
Marine Fisheries Service (NMFS), Galveston,
Tex., this study was made possible. His helpful
suggestions were greatly appreciated. We are
grateful to Albert C. Jones, Alexander Dragovich,
and Donald M. Allen, Southeast Fisheries Center,
NMFS, Miami, Fl., for providing specimens for
this study. Fred M. Utter, Northwest Fisheries
Center, NMFS, Seattle, Wash., reviewed the
manuscript. Frank Patella conducted the statisti-
cal analyses for the study.

Literature Cited

Johnson, a. G., F. M. Utter, and H. O. Hodgins.

1971. Phosphoglucomutase polymorphism in Pacific ocean
perch, Sebastodes alutus. Comp. Biochem. Physiol.
39B:285-290.

456

1974. Electrophoretic comparison of five species of
pandalid shrimp from the northeastern Pacific Ocean.
Fish. Bull., U.S. 72:799-803.

Marvin, K. T., and C. W. Caillouet.

1976. Phosphoglucomutase polymorphism in white
shrimp, Penaeus setiferus. Comp. Biochem. Physiol.
53B:127-131.
MISTAKIDIS, M. N. (editor).

1968. Proceedings of the World Scientific Conference on
the Biology and Culture of Shrimps and Prawns. FAO
Fish. Rep. 57, 4 vol., 1627 p.

PfiREZ FARFANTE. I.

1969. Western Atlantic shrimps of the genus Penaeus.
U.S. Fish Wildl. Serv., Fish. Bull. 67:461-591.

PROCTOR, R. R., K. T. MARVIN, L. M. LANSFORD, AND R. C.

Benton.

1974. Phosphoglucomutase polymorphism in brown shrimp ,
Penaeus aztecus. J. Fish. Res. Board Can. 31:1405-1407.

Stern. C.

1943. The Hardy-Weinberg law. Science (Wash., D.C.)
97:137-138.

Lawrence M. Lansford

Charles W. Caillouet

Kenneth T. Marvin

Gulf Coastal Fisheries Center Galveston Laboratory
National Marine Fisheries Service, NCAA
Galveston, TX 77550

FIRST RECORD OF THE MELON-HEADED
WHALE, PEPONOCEPHALA ELECTRA, IN

THE EASTERN PACIFIC, WITH
A SUMMARY OF WORLD DISTRIBUTION

Peponocephala electra (Gray 1846) is a tropical
pelagic delphinid previously known to occur in the
eastern Atlantic, Indian, and western and central
Pacific oceans. It is also known as the electra dol-
phin, the Hawaiian blackfish, and the many-
toothed blackfish. Since van Bree and Cadenat
(1968; localities 1-4, 6-9, 11, 13, 14, 16, 18, and 19 in

Figure 1) summarized world records, the species
has been reported from Thailand (Pilleri 1973,
locality 17), the Philippine Sea near Cebu (W. H.
Dawbin pers. commun., locality 15), near Towns-
ville, Australia (G. E. Heinsohn pers. commun.,
locality 12), the New Hebrides (Rancurel 1974,
locality 10), and the Tuamotos-Marquesas region
(K. S. Norris pers. commun., locality 5). Records
cited by van Bree and Cadenat (1968) as "in
litteris" or in press, have subsequently been pub-
lished (Dawbin et al. 1970, locality 11; Morzer
Bruyns 1971, localities 6-9). The purpose of this
note is to report a capture that extends the known
range of the species some 3,000 miles into the
eastern tropical Pacific off Central America
(Figure 1; triangle).

The specimen (Figure 2), a male calf 112 cm
long (tip of upper jaw to base of notch in flukes)
and weighing 15 kg, was captured in a tuna purse
seine that had been set on an aggregation of
yellowfin tuna, Thunnus albacares, and dolphins,
Stenella sp., approximately 90 nautical miles
(about 167 km) due west of Champerico, Guate-
mala (lat. 14°20'N, long. 91°52'W), in May 1974.
More precise information on date and locality of
capture is not available. A crew member found the
calf dead in the net, placed it in the ship's freezer,
and on return to port donated it to the National
Marine Fisheries Service, La Jolla. The specimen
was identified using X rays of the dentition. The

high tooth count (

23 +

22 +

23 +

22+

), combined with the

blunt head and dark coloration, is diagnostic of
the species. The specimen was then photo-
graphed, measured, weighed, cast in plastic, and
sent frozen to the U.S. National Museum (USNM),
Washington, D.C, where it was preserved whole

Figure l. — Known distribution of
Peponocephala electra. Triangle is
new record; sources of others in text.
Closed circles are specimen locali-
ties, c^en circles are sightings. Some
circles represent multiple records
from single localities, e.g., Hawaii
and Honshu, Japan.

457

Figure 2. — CaXi oi Peponocephala electm collected in eastern tropical Pacific (USNM 504087).

and placed in the marine mammal collection

(USNM 504087).

Acknowledgments

I thank Edward Kovalchek and Joseph Mad-
ruga for providing the specimen and P. J. H. van
Bree for reading the manuscript.

Literature Cited

dawbin, w. h., b. a. noble, and F. C. fraser.

1970. Observations on the electra dolphin, Peponocephala
electra. Bull. Br Mus. (Nat. Hist.) Zool. 20(6): 175-201.

MORZER BRUYNS, W. F. J.

1971. Field guide of whales and dolphins. C. A. Mees,
Amsterdam, 258 p.

PILLERI, G.

1973. Cetologische Expedition zum Indus und Persischen
Golf und Forschungsreise nach Goa und Thailand im
Jahre 1973. Hirnanatomisches Institut, Waldau-Bern,
Switzerland, 52 p.

Rancurel, p.

1974. Echouage en masse du cetace Peponocephala electra
aux Nouvelles-Hebrides. Biol. Cons. 6:232-234.

Van Bree, P. J. H., and J. Cadenat.

1968. On a skull of Peponocephala electra (Gray, 1846)
(Cetacea, Globicephalinae) from Senegal Beaufortia
14:193-202.

WILLIAM F. PERRIN

Southwest Fisheries Center La Jolla Laboratory
National Marine Fisheries Service, NOAA
La Jolla, CA 92038

FOODS OF JUVENILE SOCKEYE
SALMON, ONCORHYNCHUS NERKA, IN
THE INSHORE COASTAL WATERS OF

BRISTOL BAY, ALASKA, 1966-67

For most living organisms the early portion of life
is most critical in determining survival. Anad-
romous fishes such as Pacific salmon have two
critical periods during early life — development
and growth in fresh water and subsequent adap-
tation to a marine environment. The food of
juvenile salmon during the first few months of
marine life influences growth and condition,
which in turn probably influences parasitism,
predation, and other factors which ultimately
determine marine survival.

Although the sockeye salmon, Oncorhynchus
nerka (Walbaum), is one of the most valuable
commercial fishes in Alaska and has been the ob-
ject of extensive research, little is knovm of its
early life in the sea. Straty (1974) and Straty and
Jaenicke^ have made the only comprehensive
study of early marine life of the sockeye salmon
in Bristol Bay, historically the largest sockeye
fishery in the North Pacific. Documented studies
of sockeye salmon food habits during this period
of life are generally limited to brief accounts of
Soviet research in Kamchatka waters (Synkova
1951), a study in British Columbia (Manzer
1969), examination of a few specimens from Aleu-
tian and Kodiak waters (Chamberlain 1907), and
45 specimens taken off Cape Seniavin in lower

'Straty, R. R., and H. W. Jaenicke. 1971. Studies of the es-
tuarine and early marine life history of sockeye salmon in Bris-
tol Bay, 1965-67. Unpubl. manuscr., 137 p. Northwest Fish.
Cent. Auke Bay Lab., Natl. Mar. Fish. Serv., NOAA, Auke Bay,
AK 99821.

458

Bristol Bay (Dell 1963). Recently, Jaenicke and
Bonnett^ completed an extensive study of the
foods of some 1,200 seaward-migrating sockeye
salmon in Bristol Bay during 1969 and 1970.
Most of their samples were taken over deeper
waters farther offshore than mine — particularly
those off Port Moller.

The purpose of my study was to document the
foods of seaward-migrating sockeye salmon along
the main migration route on the north side of the
Alaska Peninsula in Bristol Bay, Alaska, during
1966 and 1967. Later studies by Straty and
Jaenicke (see footnote 1) and Jaenicke and Bon-
nett (see footnote 2) show that the areas where I
took samples of juvenile sockeye salmon
(Kvichak to Port Moller — Figure 1) were indeed
along the main migration route in the upper and
central parts of the bay (Kvichak to Port Heiden).
In lower Bristol Bay, however, my sampling area
(Port Moller) was inshore from the usual main
migration route. In years when unusually cold
sea water temperatures prevail, the main migra-
tion route in the lower bay shifts to the warmer
inshore waters (Straty 1974). The juvenile sock-
eye salmon I sampled in the Port Moller area
were taken in a year (1967) when normal temp-

^Jaenicke, H. W., and M. B. Bonnett. Food of sockeye salmon
outmigrants in Bristol Bay, Alaska, 1969-70. Unpubl. manuscr,
20 p. Northwest Fish. Cent. Auke Bay Lab., Natl. Mar. Fish.
Serv., NOAA, Auke Bay, AK 99821.

SAMPLING SITES

1966 •

1967 A

-56°

r^- -^f\

Figure l. — Bristol Bay, Alaska, showing locations where
juvenile sockeye salmon were collected in 1966 and 1967 for food
habit analyses. Samiples in the upper bay (Kvichak and Egegik)
were taken in June, and those in the central bay (Ugashik) and
lower bay (Port Heiden and Port Moller) were taken from July to
September.

eratures prevailed and were presumably inshore
from the path followed by most migrants that
year. However, the foods found in 1967 in these
inshore waters may reflect what is usually avail-
able to the main body of outmigrants in colder
years when their path is altered.

Materials and Methods

The samples of juvenile sockeye salmon were
collected in 1966 and 1967 in the following areas
(Figure 1) and times: Kvichak, June of both
years; Egegik, June 1966; Ugashik, July and Sep-
tember 1966 and August 1967; Port Heiden, July
1966 and August 1967; and Port Moller, July and
August 1967. All samples were taken during day-
light, usually between 1000 and 1900 h.

In 1966, the juvenile sockeye salmon were col-
lected with circular tow nets (2.1 m in diameter)
and a small-mesh round haul seine (110 m long
by about 7 m deep); in 1967 they were collected in
a small-mesh lampara seine (183 m long by about
14 m deep). All sampling was done from the 13-m
National Marine Fisheries Service vessel Sock-
eye. Samples were preserved in 10% Formalin^
solution and processed later.

I analyzed the stomach contents of 160 juvenile
sockeye salmon and all but 16 contained food.
These 160 fish represented roughly equal num-
bers of individuals from 1-cm size groups ranging
from 6- to 13-cm fork length and were from all
five areas of Bristol Bay from Kvichak Bay south
to Port Moller — a distance of about 320 km.

The stomach (that portion of the digestive tract
from the anterior end of the esophagus to the
pylorus) of each specimen was removed, and all
food organisms were separated and identified to
the lowest taxonomic level practical. All of the
food items were air dried overnight at room
temperature and weighed to the nearest 0.1 mg
the follovdng day.

The eight major categories of food items:
copepods, fish, larval crustaceans, euphausiids,
amphipods, insects, miscellaneous crustaceans,
and zoofauna, are not mutually exclusive. The
least specific categories merely reflect the di-
gested condition or incidental importance of a
given item, e.g., crustacean remains (recorded as
miscellaneous crustaceans) or arachnids (zoo-
fauna), which occurred only once.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

459

For each sampling area, the weight of each
major food category was calculated as the per-
centage of the total dry weight of all food found.
The percentage of occurrences and weights of
foods were based only on those specimens con-
taining food.

Results

The foods consumed by seaward-migrating
sockeye salmon in Bristol Bay varied in the rela-
tive proportion and occurrence of kinds and
quantities between months during the summer.
The apparent differences between the upper and
lower areas of the bay are largely due to date of
sampling. The 16 empty stomachs found were col-
lected in June from the upper bay — the Kvichak
and Egegik areas.

In early June 1966 in the Kvichak area, 11 of 19
juvenile sockeye salmon contained food. Al-
though fish and insects made up 97% of the bulk
(weight), fish occurred in only 5% of the stomachs
and insects in 53%. By late June in the same
area, 8 of 10 stomachs contained food, most of
which was copepods. They made up 89% of the
bulk and were found in 70% of the stomachs; mis-
cellaneous crustaceans were found in 60%. In
mid- June of the following year (1967), 18 of 21
juvenile sockeye salmon from the Kvichak area
contained food. Fish, insects, and copepods made
up 93% of the bulk; fish occurred in 19% of the
stomachs, insects in 76%, and copepods in 62%.

In mid-June 1966, 20 of 23 stomachs collected
farther seaward at Egegik contained very small
amounts of food. Euphausiids and miscellaneous
crustaceans made up 78% of the bulk, but
euphausiids occurred in only 9% of the stomachs
and miscellaneous crustaceans in 13%. Insects oc-
curred in 48% of the stomachs, but made up only
4% of the bulk.

In mid-July 1966 at Ugashik, all 20 stomachs
collected contained larval crustaceans (79% by

bulk and mostly anomurans). Copepods were in-
significant in terms of bulk but occurred in 70% of
the stomachs. At Port Heiden (farther seaward)
on the same date, fish made up 76% of the bulk of
the contents of the seven stomachs collected. Fish
occurred in 28% of the stomachs, whereas am-
phipods occurred in 71% and insects in 57%.

At Port Moller in lower Bristol Bay throughout
July and on 1 August 1967, copepods made up
71% of the bulk of food in 48 stomachs and oc-
curred in 85%; larval crustaceans occurred in
58%, amphipods in 50%-, and fish in 42%.

By mid-August 1967, when most juvenile sock-
eye salmon have migrated out of Bristol Bay
(Straty 1974), the two juveniles taken at Ugashik
contained only copepods and insects and two
taken at Port Heiden contained mostly fish.

Only eight juvenile sockeye salmon were taken
in September 1966 in the Ugashik area. Copepods
and fish accounted for 86% of the stomach con-
tents, but only copepods occurred frequently
(100%' with copepods vs. 25% with fish).

As the young sockeye salmon migrated sea-
ward over successive months, they ate increasing
amounts of food. In the Kvichak and Egegik areas
during June, 16 of the 73 stomachs examined
were empty and the others had only relatively
small amounts of food (average of 3-6 mg). Later
in the summer and farther at sea (Ugashik and
Port Heiden) the average amount of food per
stomach was much greater (20-24 mg), and still
later in the summer and farther at sea (Port Mol-
ler), the amounts were the highest of all (average
of 82 mg).

In terms of both bulk and frequency of occur-
rence, copepods were the most important food of
juvenile sockeye salmon in inshore Bristol Bay in
1966 and 1967 (Tables 1, 2). Two genera of
calanoid copepods {Eurytemora and Metridia)
made up 98% of the number of copepods in the
stomachs of 50 juveniles taken by Straty and
Jaenicke (see footnote 1) in 1967 at Kvichak and

Table l. — Percentage total dry weight of foods consumed by juvenile sockeye salmon
collected at five areas in Bristol Bay, Alaska, 1966 and 1967.

Kvichak

Egegik

Ugashik

Port Heiden

Port Moller

Food category

W = 50

N =22

A/ = 30

W = 9

N = 48

Copepods

30.3

8.6

25.4

71.2

Fish

45.7

4.1

22.6

80.3

11.8

Larval crustaceans

0.1

0.4

44.6

—

5.7

Euphausiids

—

43.1

0.4

—

5.2

Amphipods

0.6

1.0

1.3

4.8

4.7

Insects

18.6

3.9

0.9

0.8

Miscellaneous crustaceans

2.8

34.9

0 2

0.1

0.5

Zoofauna

2.1

3.3

4.7

6.1

0.2

Other

—

0.8

—

1.8

—

460

Table 2. — Summary of foods eaten by juvenile sockeye salmon
(N = 160) in all regions of Bristol Bay, Alaska, between June
and September 1966 and 1967.

Percentage total

Percentage

Food category

dry weight

occurrence

Copepods

60.4

66.7

Fish

17.4

25.0

Larval crustaceans

9.8

35.4

Euphausiids

4.6

6.3

Amphipods

4.0

29.2

Insects

1.6

41.0

Miscellaneous crustaceans

0.9

22.2

Zoofauna

1.1

18.8

aher

0.1

2.8

Empty stomachs

—

10.0

Port Moller. (The 50 specimens were taken at the
same time and place as my samples.) Fish were
second in importance to copepods in terms of
weight of food, and over half the bulk of these fish
were Pacific sand lance, Ammodytes hexap-
terus. Larval crustaceans were the only other food
of major importance (by bulk) and most of these
were anomuran larvae eaten by juveniles in the
Ugashik area in July 1966. Other items eaten by
juvenile sockeye salmon in significant amounts
during their migration out of Bristol Bay were
euphausiids, amphipods, and insects. Insects and
amphipods occurred frequently in the diet but did
not contribute much bulk.

I looked for differences in food selectivity be-
tween large and small fish among 144 juveniles
(6-13 cm fork length) grouped in 1-cm size
categories, but the results were inconclusive.

Discussion

The results of this study generally agree with
those of other investigators. The importance of
copepods in the diet of juvenile sockeye salmon
near shore in Bristol Bay is paralleled in coastal
waters of British Columbia (Manzer 1969) and
is similar to Kamchatka coasts, where copepods
and cladocerans were the predominant foods of
juvenile sockeye salmon (Synkova 1951). My
findings differ from those of Jaenicke and Ben-
nett (see footnote 2), who sampled mainly over
deeper waters of Bristol Bay farther offshore than
I did, and Dell (1963), who sampled off Port
Moller in Bristol Bay. Jaenicke and Bennett
examined the food of over 1,200 juvenile sockeye
salmon captured during the summers of 1969-70
and found that the main items (in bulk) were
young and larval sand lance and euphausiids.
Similarly, Dell reported that 45 juvenile sockeye

salmon taken in late July 1962 contained mostly
larval sand lance and euphausiids.

Nearly all of the insects I found were from
juvenile sockeye salmon captured in the Kvichak
and Egegik areas in June (Table 1). These areas
are contiguous to many rivers that form part of a
major sockeye salmon reproductive complex of
lakes and steams (Figure 1). According to Hart-
man et al. (1967), most of the migration from
freshwater to Bristol Bay takes place in June.
Most of the insects were probably ingested in
fresh water when the fish were migrating sea-
ward, suggesting that many of the juveniles
taken in these areas were recent immigrants
from fresh water. The occurrence of the only
empty stomachs and small average weight of food
per fish at Kvichak and Egegik suggest that the
juveniles eat very little when they first enter salt
water. Straty (1974) concluded that the young
sockeye salmon did not feed when they entered
Bristol Bay or that food was scarce. Reduction of
feeding could be caused by a number of factors
other than lack of food, including the physiologi-
cal strain of adjusting osmoregulatory ftinction
from a freshwater to a marine environment.

The differences I observed in the types, relative
proportions, and amounts of food eaten over suc-
cessive months by the juvenile sockeye salmon as
they progressed seaward can be largely attrib-
uted to food availability. Near-surface waters in
the Kvichak area contained an average of 27
zooplankters per cubic meter in June, while near
Port Moller in July the density was 1,400-8,100
(see footnote 1). Straty (1974) compared zooplank-
ton abundance in the inner part of Bristol Bay
(above Port Heiden) and the outer part (below Port
Heiden) during 1969-71 and concluded that zoo-
plankton was much more abundant as one pro-
gressed seaward.

Acknowledgements

I thank Walter R. Whitworth, Professor of
Fisheries of the University of Connecticut, for his
advice and help with this manuscript, which in
more inclusive form was accepted as an M.S.
thesis (1968) at the University of Connecticut,
Storrs, and also thank my colleagues at the
Northwest Fisheries Center Auke Bay Labora-
tory, National Marine Fisheries Service, NOAA,
particularly Richard R. Straty and Herbert W.
Jaenicke for their assistance and advice through-
out the study.

461

Literature Cited

Chamberlain, F. M.

1907. Some observations on salmon and trout in Alaska.

Rep. U.S. Comm. Fish. 1906, 112 p. (Bur. Fish. Doc. 627.)
Dell, M. B.

1963. Oceanic feeding habits of the sockeye salmon,

Oncorhynchus nerka (Walbaum), in Aleutian waters.

M.S. Thesis, Univ. Michigan, Ann Arbor, 40 p.

Hartman, W. L., W. R. Heard, and B. Drucker.

1967. Migratory behavior of sockeye salmon fry and
smolts. J. Fish. Res. Board Can. 24:2069-2099
MANZER, J. I.

1969. Stomach contents of juvenile Pacific salmon in
Chatham Sound and adjacent waters. J. Fish. Res.
Board Can. 26:2219-2223.
STRATY, R. R.

1974. Ecology and behavior of juvenile sockeye salmon
(Oncorhynchus nerka) in Bristol Bay and the eastern Ber-
ing Sea. In D. W. Hood and E. J. Kelley (editors), Ocean-
ograhy of the Bering Sea, p. 285-319. Inst. Mar. Sci.,
Univ. Alaska, Fairbanks.
Synkova, a. I.

1951. Food of Pacific salmon in Kamchatka waters. [In
Russ.] Izv. Tikhookean. Nauchno-Issled. Inst. Rybn.
Khoz. Okeanogr. 34:105-121.

H. Richard Carlson

Northwest Fisheries Center Auke Bay Laboratory
National Marine Fisheries Service, NOAA
P.O. Box 155, Auke Bay, AK 99821

OCCURRENCE OF TWO GALATHEID
CRUSTACEANS, MUNIDA FORCEPS AND

MUNIDOPSIS BERMUDEZI, IN THE

CHESAPEAKE BIGHT OF THE WESTERN

NORTH ATLANTIC OCEAN' ^

Living male specimens of Munida forceps A.
Milne-Edwards and Munidopsis bermudezi Chace
(Table 1) were collected on the continental slope
and rise south of Norfolk Canyon off the coast of
Virginia on 18-19 November 1974. An ovigerous
female M. bermudezi was also collected on 14 Sep-
tember 1975 in the Norfolk Canyon. They were
taken with a 15-m shrimp trawl (12-mm stretch
mesh inner liner) towed from the RV James M.
Gillis (University of Miami, Florida).

Munida forceps has been reported from 80 to
338 m within the Gulf of Mexico and in the south-

'Research supported by National Science Foundation Grant
GA-37561, J. A. Musick, Principal Investigator and by U.S.
Department of Commerce, National Marine Fisheries Service
Contract No. 03-4-043-353 for C.E.L. and P.A.H. participation.

^Contribution No. 717, Virginia Institute of Marine Science.

western Atlantic between lat. 22°46.5' and
26°37.0'N (Chace 1940, 1942; Springer and Bullis
1956; Bullis and Thompson 1965). Our find is
consistent with the previously reported depth
range, but it extends the geographic range of
the species northward by 10° latitude.

Munidopsis bermudezi has been reported from
the coast of Cuba (lat. 21°19'N, long. 76°05'W) at a
depth of 2,654 m (Chace 1940, 1942), the Gulf of
Mexico (lat. 25°50.5'N, long. 94°27'W) at 3,294 m
(Pequegnat and Pequegnat 1970), and north of the
Azores (lat. 45°26'N, long. 25°45'W) at 3,171 m
(Sivertsen and Holthuis 1956).

The Munida forceps sample also included the
galatheids M. iris A. Milne-Edwards and M. lon-
gipes A. Milne-Edwards and other decapods in-
cluding Bathynectes superbus (Costa), Cancer
borealis Stimpson, C. irroratus Say, Homarus
americanus H. Milne Edwards, and penaeidean
and caridean shrimps. The association of M. for-
ceps with M. iris and M. longipes in our sample is
previously unreported. Some previous records
have shown associations with M. stimpsoni A.
Milne-Edwards (Chace 1942) and with M. flinti
Benedict and M. irrasa A. Milne-Edwards
(Milne-Edwards 1880 from Pequegnat and
Pequegnat 1970). Others (Benedict 1902; Bullis
and Thompson 1965; Pequegnat and Pequegnat
1970) have not specified association of M. forceps
with other galatheids.

Table l. — Station and morphometric data for Munida forceps
and Munidopsis bermudezi captured near Norfolk Canyon off
the coast of Virginia. Length and width measurements in mil-
limeters.

Item

Munida
forceps

Male

Munidopsis bermudezi

Male

Female

Station
Collection
Location, lat.

long.
Date of collection
Depth (m)

Bottom temperature ("C)
Bottom salinity ( :. )
Total lengtti (rostral tip to

postenor margin of telson)
Carapace width, anterior
posterior
Carapace length (orbit to

posterior margin)
Carapace length (including

rostrum)
Cheliped (right) length
Carpus length
Merus length
Propodus length
Propodus width
Dactylus length
Second left pereopod length

C74-499
36°43.2N
74=38.0W
Nov. 1974
220-310
10.6

34
7.9
10.4

13.5

18.5
45

4.0
15.2
25.6

4.5
15.1
28.8

C74-506
36M1.6'N
73°47.0W
Nov. 1974
2,620-2,650
3.0
34.82

81.4
28.4
31.0

335

44.8
42.4

8.5
14.5
19.3

8.8
10.5
48.7

C74-168
36°57.9'N
73^21. 5'W
Sept. 1975
2,915-2,955
2.3
35.11

83.2
28.8
31.5

335

43.8
40.8

7.5
13.0
14.3

8.0

8.3
46.5

462

In November 1974, Munidopsis bermudezi was
associated with M. curvirostra Whiteaves. Previ-
ous accounts did not indicate association of M.
bermudezi with other galatheids. Other decapods
taken in the November sample were Lithodes
agassizii Smith, Stereomastis sculpta (Smith), and
penaeidean and caridean shrimps, including
Hymenodora gracilis Smith, a species occurring in
the Azores sample (Sivertsen and Holthuis 1956).

In September, M. bermudezi was associated
with M. bairdii (Smith) and M. crassa (Smith), as
well as Lithodes agassizii and caridean shrimp.

The ovigerous M. bermudezi had not shed all
eggs onto the pleopods. The 19 external eggs were
tan and averaged 2.8 mm in diameter. These eggs
were spherical with no visible blastoderm and
were recently extruded. The ovary was tan and
very well developed. It contained 106 ova averag-
ing 2.7 mm in diameter. All eggs were measured
with an ocular micrometer.

We suspect that these species with tropical
affinities are normally present, though rare, in
the Chesapeake Bight; but they could be acciden-
tal migrants. In either case, the probability of de-
tection was raised by the recent increase in sam-
pling intensity in the vicinity of Norfolk Canyon
as compared to other areas of the continental slope
between Florida and North Carolina. The question
of how far north the tropical fauna extends along
the southeastern coast of North America is still
unanswered (Briggs 1974). Cerame- Vivas and
Gray (1966) noted that the inshore fauna of the
North Carolina shelf was warm temperate
(Carolinian) but that the offshore fauna was tropi-
cal. In a study of sea stars of North Carolina, Gray
et al. (1968) found 13 species that occurred in a
northward extension of the Caribbean Province
along the outer shelf and that these species ranged
slightly northward past Cape Hatteras.

The authors are grateful for the assistance of
Fenner A. Chace Jr., in confirming the identifica-
tion of the specimens. All specimens have been
deposited at the United States National Museum,
Washington, D.C.

BULLIS, H. R., AND J. R. THOMPSON.

1965. Collections by the exploratory fishing vessels Oregon,
Silver Bay, Combat and Pelican made during 1956 to 1960
in the southwestern North Atlantic. U.S. Fish Wildl.
Serv., Spec. Sci. Rep. Fish. 510, 130 p.

CERAME-VIVAS, M. J., AND I. E. GRAY.

1966. The distributional pattern of the benthic inverteb-
rates of the continental shelf off North Carolina. Ecology
47:260-270.

Chace, F. a., Jr.

1940. The Atlantis expeditions to the West Indies in 1938
and 1939, under the joint auspices of the University of
Havana and Harvard University. List of Stations. Woods
Hole Oceanogr. Inst., Contrib. 274, 8 p.

1942. Reports on the scientific results of the Atlantis exped-
itions to the West Indies, under the joint auspices of the
University of Havana and Harvard University. The
anomuran Crustacea. I. Galatheidea (Families Chiros-
tylidae, Galatheidae and Porcellanidae). Torreia
(Havana) 11:1-106.
Gray, I. E., M. E. Downey, and M. J. Cerame- Vivas.

1968. Sea-stars of North Carolina. U.S. Fish Wildl. Serv.,
Fish. Bull. 67:127-163.
MiLNE-EDWARDS, A.

1880-81 Reports on the results of dredging under the super-
vision of Alexander Agassiz, in the Gulf of Mexico and in
the Caribbean Sea 1887 '78, '79 by the U.S. Coast Survey
Steamer "Blake," Lieut. -Commander C. D. Sigsbee,
U.S.N. , and Commander J. R. Bartlett, U.S.N. , Command-
ing. VIII. fitudes preliminaries sur les Crustaces. Bull.
Mus. Comp. Zool. Harv. Coll. 8:1-68.
PEQUEGNAT, L, H., AND W E. PEQUEGNAT.

1970. Deep-sea anomurans of Superfamily Galatheoidea
with descriptions of three new species. In W. E. Pequegnat
and F. A. Chace, Jr. (editors). Contributions on the biology
of the Gulf of Mexico, Vol. 1, p. 125-170. Gulf Publ. Co.,
Houston.

Sivertsen, E., and L. B. Holthuis.

1956. Crustacea Decapoda (the Penaeidea and
Stenopodidea excepted). Rep. Sci. Result "Michael Sars"
North Atl. Deep-sea Exped. 1910 5(12):l-54.

Springer, S., and H. r. Bullis, Jr.

1956. Collections by the Oregon in the Gulf of Mexico. U.S.
Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 196, 134 p.

Chae E. Laird

ELIZABETH G. Lewis

Paul a. Haefner, Jr.

Virginia Institute of Marine Science
Gloucester Point, VA 23062

Literature Cited

Benedict J. E.

1902. Descriptions of a new genus and forty-six new species
of crustaceans of the family Galatheidae, writh a list of the
known marine species. Proc. U.S. Natl. Mus. 26:243-334.
BRICXJS, J. C.

1974. Marine zoogeography. McGraw-Hill, N.Y., 475 p.

463

EFFECTS OF MERCURY, CADMIUM,

AND LEAD SALTS ON

REGENERATION AND ECDYSIS IN

THE FIDDLER CRAB, UCA PUGILATOR

Crabs are capable of autotomizing injured limbs at
a preformed breakage plane and subsequently re-
generating them. The regenerating limb bud
grows in a folded position within a layer of cuticle,
and unfolds when the animal molts. The length of
regenerating limb buds is generally expressed in
terms of "R-value" (Bliss 1956) which is length of
limb bud x 100/carapace width. Such a regenera-
tion index is useful for comparisons of crabs of
different sizes. Since regeneration always termi-
nates with a molt, the presence of regenerating
limbs can affect the timing of ecdysis, and factors
which influence ecdysis will also affect regenera-
tion. For example, removal of eyestalks, a source
of molt-inhibiting hormones, is a standard way of
inducing precocious molting. Such animals will
regenerate missing limbs rapidly, but will gener-
ally die at ecdysis. Skinner and Graham (1972)
have shov^Ti that multiple autotomy, producing
many regenerating limb buds, can cause acceler-
ated regeneration, also leading to precocious molt.

Heavy metals as pollutants of the marine envi-
ronment are of great concern. These chemicals are
released as a result of industrial processes and
tend to be toxic and to accumulate in organisms.
Their toxicity to Crustacea has been studied by
Corner and Sparrow (1957), Wisely and Blick
(1967), Eisler (1971), Vernberg and Vernberg
(1972), and O'Hara (1973).

This paper reports on the effects of mercury,
lead, and cadmium on regeneration in the fiddler
crab, Uca pugilator. With its estuarine intertidal
habitat, this crab is likely to be subject to heavy
metal pollution in industrial areas.

Materials and Methods

Fiddler crabs were collected in July and August
from Accabonac Harbor, near East Hampton,
N.Y., and brought into the laboratory. Autotomy
of one chela and six walking legs was induced by
pinching each merus with a hemostat. Im-
mediately after autotomy, crabs were placed in
solutions of Pb(N03)2 (Reagent grade, Fisher Sci-
entific), HgCl2 (Reagent grade, Fisher Scientific),
or anhydrous CdCl2 (Reagent grade, Matheson,
Coleman and Bell) at concentrations of 0.1 or 1.0
mgAiter of the metal ion. Crabs were maintained

in groups of 10 in 1-liter glass aquaria in 200 ml of
filtered seawater (30'L salinity, room temperature
25°C). Twice weekly the aquaria were washed out
and redosed. (In a similar static experimental de-
sign, Jackim et al. (1970) determined that the loss
of metal ion from solution over a 96-h period was
0% for cadmium, 26% for mercury, and 79% for
lead.) Crabs were fed twice weekly with Purina
Fly Chow^ In all experiments, groups were ar-
ranged to have the same mean carapace width and
to have equivalent distribution of males and
females (5/5).

The growth of limb buds was measured twice
weekly under a dissecting microscope with a cali-
brated ocular micrometer. In all cases, the first
walking leg was measured as a representative
limb. Values thus obtained were converted to
R-values, and the means for each group were com-
pared by the use of the ^test. Times of molting
were recorded for all animals. Limb buds reached
R-values of about 20 just prior to ecdysis.

Whole crabs were analyzed for mercury, cad-
mium, and lead following 2 wk of exposure to 0.1
mg/liter. Five crabs were used for each assay,
which was done by New Jersey Department of
Health personnel, using atomic absorption spec-
trophotometry.

Results

In experiment 1, crabs (mean carapace width
15 mm, range 13-16 mm) were exposed to 0.1 vagi
liter of lead, cadmium, and mercury. Ten crabs
were in each group, total biomass about 11 g.
Cadmium had a retarding effect on regeneration
(Table 1 ) although most individuals had molted by
28 days. The majority of controls molted by 21
days, and the rest completed ecdysis by 24 days.
Mercury and lead had no retarding effect.

This experiment was repeated with crabs of a
somewhat smaller size (13 mm carapace width,
range 11-14 mm). Although cadmium again re-
tarded regeneration, the retardation was less and
was not always statistically significant (Table 1).
These crabs reached ecdysis at the same time as
controls (21 days). No effects of lead or mercury
were seen.

In experiment 2, crabs were exposed to lead,
mercury, and cadmium at concentrations of 1.0
mg/liter. Carapace width of crabs was 15 mm

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

464

Table l. — R-values (mean ± standard error) of first walking legs of crabs after
multiple autotomy and treatment with Pb, Hg, and Cd at 0.1 mg/liter.

Days

Chemical

Carapace width 15 mm:

Controls

1.8 ± 0.3

7.3 ± 0.6

12.0 ± 0.7

18.4 ± 0.8

60% molt

2.8 ± 0.5

10.2 ± 0.7

14.7 ± 1.1

20.1 ± 0.6

80% molt

2.3 ± 0.2

8.8 ± 1.2

13.8 ± 1.3

17.7 ± 0.8

70% molt

1.0 ± 3.3

3.3 ± 0.7*

8.6 ± 1.2*

11.0 ± 1.3*

13.5 ± 1.5

Carapace wi

idth 13 mm:

Controls

4.8 ± 0.4

10.6 ± 1.0

17.7 ± 1.0

20.2 ± 0.7
40% molt

70% molt

4.2 ± 0.4

9.2 ± 0.9

17.7 ± 0.8

18.2 ± 0.9
50% molt

gO'-'o molt

3.9 ± 0.7

9.4 ± 0.9

16.2 ± 1.1

17.9 ± 0.5
30% molt

60% molt

3.5 ± 0.6

8.0 ± 1.0

14.2 ± 1.1*

17.0 ± 0.8*
0% molt

70% molt

•P = 0.05

or less.

(range 14-16 mm). At this concentration, cad-
mium retarded regeneration to an even greater
extent. This concentration of mercury was usually
toxic, and the data obtained were from four crabs
which survived the duration of the experiment.
Regeneration did not take place in these crabs
(Table 2). The cadmium, however, was not toxic,
and all crabs survived, the majority (60%) com-
pleting ecdysis by 28 days. There was no mortality
in lead, cadmium, or clean water in any of the
experiments. The majority of controls molted by
24 days. A second group of crabs (carapace width
13 mm, range 12-14 mm) was exposed to cadmium
and mercury at 1.0 mg/liter. Lead was not used in
this experiment. Because of the high mortality in
mercury in the previous experiment, 20 crabs
were exposed to mercury. By the 17th day, the
number surviving in mercury was reduced to
eight, the same percentage as survived the previ-
ous experiment. The amount of growth in these
crabs, though slight, was nevertheless much great-

er than in the previous experiment. Likewise, the
retardation in cadmium was not as striking as in
the earlier experiment (Table 2). The majority of
controls molted by 21 days, whereas the majority
in cadmium molted by 28 days. After 2V2 wk, the
eight crabs remaining in mercury were trans-
ferred to clean water, which was then changed
daily, but they did not show evidence of recovery
within 4 wk after return to clean water, during
which time no significant growth occurred.

Residue analysis revealed that the crabs ex-
posed for 2 wk to 0.1 mg/liter of mercury had ab-
sorbed 0.026 ± 0.001 ppm; those exposed to 0.1
mg/liter of cadmium had absorbed 0.50 ± 0.10
ppm; and those exposed to 0.1 mg/liter of lead had
absorbed 2.04 ± 0.55 ppm.

Discussion

Retardation of regeneration was a specific effect
of cadmium at both 0.1 and 1.0 mg/liter. At 0.1

Table 2. — R-values (mean ± standard error) of first walking legs of crabs after multi
and treatment with Pb, Hg, and Cd at 1.0 mg/liter

pie autotomy

Days

Chemical 7

Carapace width 15 mm:
Controls 4.2 ± 0.4
Pb 2.8 ± 0.6
Hg 0*
Cd 0.3 ± 0.2*

Carapace width 13 mm:
Controls 4,6 ± 0.5
Hg 1.0 ± 0.6*
Cd 3.5 ± 0.2

8.0 ± 0.6

6.2 ± 0.7

2.2 ± 0.8*

10.2 ± 0.7
1.5 ± 0.8*
6.8 ± 0.6*

13.1 ± 1.0

11.4 ± 1.0
0*

4.3 ± 1.2*

15.7 ± 0.9
1,6 ± 0.8*

11.5 ± 1.3*

15.9 ± 0.9
14.5 ± 1.2
0*
5.6 ± 1.5*

18.0 ± 0.6
'2.1 ± 1.0*
13.8 ± 1.5*

18.1 ± 0.3
17.6 ± 0.7
0.01 ± 0.01*
8.3 ± 2.5*

70% molt

2.7 ± 1.1*
16.0 ± 2.0*

70% molt
70% molt
0.01 ± 0.01*
7.6 ± 2.3*
20% molt

90% molt
2.9 ± 1.1*
50% molt

'Returned to clean water.
*P = 0.05 or less.

465

mg/liter, mercury was not toxic and did not have
an effect on the growth of limb buds. At 1.0 mg/
liter, mercury caused almost total inhibition of
limb growth, but also proved lethal to 60% of the
crabs. Therefore, the inhibition of regeneration
may not be a specific effect of the mercury, but just
an indication of the toxicity of the metal to the
crabs. In this light, Uca is seen to be much more
resistant to mercury than the porcelain crab, Pet-
rolisthes armatus, in which the 96 h LC50 (.mean
lethal concentration) was 0.050-0.064 ppm
(Roesijadi et al. 1974). With long-term exposure to
mercury, however, Uca can tolerate only 0.18 ppm
(Vernberg and O'Hara 1972). In the present study,
cadmium might have shown a greater effect than
mercury at 0. 1 mg/liter because it was absorbed to
a much greater extent than the mercury. It is pos-
sible that exposure to mercury at levels between
0.1 and 1.0 mg/liter could inhibit regeneration
without causing mortality. Despite the high
amounts absorbed, lead had no effect on regenera-
tion rate.

At both dose levels of cadmium and the higher
concentration of mercury, the retarding effects
were greater the first time the experiment was
performed (July) than the second (August). Since
these crabs normally molt in August, it is probable
that they have higher titers of ecdysone at that
time, and their progress toward ecdysis cannot be
inhibited to the same extent. A similar seasonal
difference in sensitivity to cadmium was seen in
the shrimp Paratya tasmaniensis, which showed a
threefold higher LC50 value in mid-October than
in early July (Thorp and Lake 1974).

Thurberg et al. (1973) have found that cadmium
reduced the level of oxygen consumption in the
crabs Carcinus maenas and Cancer irroratus. A
reduction of oxygen consumption of the gills of the
mud crab, Eurypanopeus depressus , exposed to
cadmium was found by Collier et al. (1973). Re-
duced metabolism may be responsible for the re-
tardation of regeneration of the crabs in cadmium.
Cadmium has been found to inhibit oxygen con-
sumption and metabolism of fishes (Thurberg and
Dawson 1974; Jackim et al. 1970) and has simi-
larly been found to retard fin regeneration in
fishes (Weis and Weis in press).

In this sort of study it is difficult to extrapolate
laboratory findings to the field. In nature, metals
would tend to be concentrated in the sediments
more than the water, and it would be primarily
from the sediments that these estuarine intertidal
crabs would pick up the metals. Crabs would not

normally be subjected to the loss of many ap-
pendages. Loss of a single limb is not particu-
larly debilitating to a decapod. Should many
limbs be lost, however, the crab's locomotion
would be impaired, and it would be at a disadvan-
tage. It would therefore be advantageous to re-
generate the lost limbs as quickly as possible.
Crabs which could not regenerate as quickly
could be more subject to predation, and the toxic
heavy metal pollutant would then be passed on to
higher trophic levels.

Acknowledgments

Thanks are extended to John C. Baiardi, Direc-
tor of the New York Ocean Science Laboratory, for
making the facilities of the laboratory available
for this study. Appreciation is also extended to
Jennifer and Eric Weis for assistance in collecting
the crabs, and to Linda Mantel and Rita Levin for
their help. This research was supported by a
N.I.H. Biomedical Grant #RR 7059.

Literature Cited

Bliss, D. E.

1956. Neurosecretion and the control of growth in a decapod
crustacean. In K. G. Wingstrand (editor), Bertil Han-
strdm, zoological papers in honour of his sixty-fifth birth-
day, Nov. 20, 1956, p. 56-75. Zool. Inst. Lund, Swed.

COLLIER, R. S., J. E. Miller, M. A. Dawson, and F. P.
Thurberg.

1973. Physiological response of the mud crab,
Eurypanopeus depressus to cadmium. Bull. Environ.
Contam. Toxicol. 10:378-382.
Corner, E. D. S., and B. W. Sparrow.

1957. The modes of action of toxic agents. II. Factors
influencing toxicities of mercury compounds to certain
Crustacea. J. Mar. Biol. Assoc. U.K. 36:459-472.

EISLER, R.

1971. Cadmium poisoning in Fundul us heteroclitus (Pisces:
Cyprinodontidae) and other marine organisms. J. Fish.
Res. Board Can. 28:1225-1234.

Jackim, E., J. M. Hamlin, and S. Sonis.

1970. Effects of metal poisoning on five liver enzymes in the

killifish (Fundulus heteroclitus). J. Fish. Res. Board Can.

27:383-390.
O'Hara, J.

1973. The influence of temperature and salinity on the to-
xicity of cadmium to the fiddler crab, Uca pugilator. Fish.
Bull., U.S. 71:149-153.

ROESIJADI, G., S. R. PETROCELLI, J. W. ANDERSON, B. J.

Presley, and R. Sims.

1974. Survival and chloride ion regulation of the porcelain
crab Petrolisthes armatus exposed to mercury. Mar. Biol.
(Berl.) 27:213-217.

Skinner, D. M., and D. E. Graham.

1972. Loss of limbs as a stimulus to ecdysis in Brachyura
(true crabs). Biol. Bull. (Woods Hole) 143:222-233.

Thorp, V. J., and p. S. Lake.

1974. Toxicity bioassays of cadmium on selected freshwater

466

invertebrates and the interaction of cadmium and zinc on
the freshwater shrimp, Paratya tasmaniensis Riek. Aust.
J. Mar. Freshwater Res. 25:97-104.

Thurberg, F. p., and M. a. Dawson.

1974. Physiological response of the cunner, Tautogolabrus
adspersus, to cadmium. III. Changes in osmoregulation
and oxygen consumption. U.S. Dep. Commer., NOAA
Tech. Rep. NMFS SSRF-681:11-13.
THURBERG, F. P., M. A. DAWSON, AND R. S. COLLIER.

1973. Effects of copper and cadmium on osmoregulation
and oxygen consumption in two species of estuarine crabs.
Mar. Biol. (Berl.) 23:171-175.
VERNBERG, W. B., AND J. O'HARA.

1972. Temperature-salinity stress and mercury uptake in
the fiddler crab, Uca pugilator. J. Fish. Res. Board Can.
29:1491-1494.
VERNBERG, W. B., AND J. VERNBERG.

1972. The synergistic effects of temperature, salinity, and
mercury on survivEd and metabolism of the adult fiddler
crab, Uca pugilator. Fish. Bull., U.S. 70:415-420.
WEIS, P., ANDJ. S. WEIS.

In press. Effects of heavy metals on fin regeneration in the
killifish, Fundulus heteroclitus. Bull. Environ. Contemi.
Toxicol.
WISELY, B., AND R. A. P. BLICK.

1967. Mortality of marine invertebrate larvae in mercury,
copper, and zinc solutions. Aust. J. Mar. Freshwater Res.
18:63-72.

Judith S. Weis

Department of Zoology and Physiology
Rutgers University
Newark, NJ 07102

NOTES ON THE EARLY DEVELOPMENT

OF THE SEA RAVEN,

HEMITRIPTERUS AMERICANUS

Egg and larval characteristics of the sea raven,
Hemitripterus americanus (Gmelin), are distinc-
tive. The species ranges from Labrador to Chesa-
peake Bay but is nowhere abundant (Bigelow and
Welsh 1925). Notes on the fertilized eggs (Bean
1897), newly hatched larvae (Warfel and Merri-
man 1944), and juveniles larger than 45 mm
(Huntsman 1922; Bigelow and Welsh 1925; Bige-
low and Schroeder 1936) have been recorded.
However, there is no available information deal-
ing with specimens between 12 and 45 mm in
length. The present paper attempts, in part, to
bridge this gap in previous observations of these
larvae.

Methods and Materials

lected at the level of the high tide mark at 0930 h
on 9 November 1974. They were placed in an open
system seawater aquarium at the marine station
of Southampton College. In mid-December half of
the eggs were transferred to laboratory facilities
at the Academy of Natural Sciences of Philadel-
phia, where they were held in artificial seawater
(7°C, 32L) with a controlled photoperiod of 10.5
h light and 13.5 h darkness. Crude but effective
temperature control was achieved by placing the
covered rearing container in a water bath. The
water bath and rearing container were then
placed in a refrigerator. The temperature of the
water bath was maintained with a thermostati-
cally controlled aquarium heater. A 7y2-W light
bulb, controlled by an electric timer, was sus-
pended above the rearing container. Moderate
aeration kept the eggs in motion. After hatching,
the larvae were maintained in similar conditions
but without aeration. The strong current result-
ing from aeration appeared to be detrimental to
the fragile larvae. When the yolk was nearly ab-
sorbed, the larvae were presented with food in the
form of Artemia sp. nauplii and small pieces of
Palaemonetes sp. and Littorina sp. flesh. Only
three specimens could be induced to eat the pieces
of flesh by placing the food in their mouths.
Eventually one specimen ate the Artemia sp.
nauplii unassisted.

Measurements were made on live material. Egg
diameters were measured with dial calipers. Total
lengths (TL) of the larvae were measured through
a dissecting microscope using an ocular microm-
eter. Myomere counts were made with the aid of
two Polaroid* HN 38 x 0.3 inch filters placed
above and below the larvae and used in conjunc-
tion with a dissecting microscope and substage
lamp. Final identification of the larvae was based
on a comparison of the largest reared specimen in
this study and the specimens collected in the Gulf
of Maine by Joanne and Wayne Laroche. All 36
preserved specimens were preserved in 5% buf-
fered Formalin and deposited in the Department
of Ichthyology, Academy of Natural Sciences of
Philadelphia (ANSP 131947).

Descriptions

Egg and Embryo

Some of the peripheral eggs in the cluster had

A cluster of nearly 90 eggs was found on the rocky
shore of Montauk Point, N.Y. The eggs were col-

iReference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

467

flattened sides, suggesting the cluster had been
part of a larger mass. A small piece of an encrust-
ing sponge, Halichondria panicea, was found at-
tached to the eggs. Small tubules (0.14 mm in
diameter) were also found on the surface of some
of the eggs and were assumed to belong to some
species of polychaete. Eggs have been described by
Warfel and Merriman (1944). At the time of
collection, embryos were already well developed
in all of the eggs. Pigmentation on the body con-
sisted of melanophores arranged in vertical bars
corresponding to the location of the myomeres.
The retina was black and the iris had a silvery
appearance. The median fin fold and pectoral buds
were formed. The former originated close behind
the hindbrain. By 16 December, large melano-
phores developed on the hindbrain and dorsal half
of the yolk sac. The body pigmentation ended
abruptly on the caudal peduncle about three-
fourths of the total length from the snout. This
characteristic pattern, to be referred to as the
truncated pigmentation pattern, persisted
throughout the development of all specimens. The
mouth was formed and open. The single oil
globule (ca. 0.8 mm in diameter) inside the yolk
sac was located at the anterior confluence of the
abdomen and yolk sac.

Newly Hatched Larvae

The larvae (Figure 1) began hatching on 3
January 1975, 55 days after collecting the already
well developed eggs, and continued through 30
January. The newly hatched larvae averaged 12.8
mm TL (range 11.7-12.7 mm). Warfel and Mer-
riman (1944) noted the larvae emerged head
first. This was not always true in the case of my
material. Nearly one-half of the larvae which

were observed hatching emerged tail first. The
large ovoid yolk extended forward to or beyond the
posterior margin of the eye. The head was not
flexed over the anterior of the yolk sac. Body
pigmentation became more dense and uniform
but was lacking over the forebrain, ventral half of
the yolk sac, and the posterior one-fourth of the
body. Melanophores lined the base of the dorsal fin
fold to the level of the truncated body pigment. A
few melanophores were present along the post-
anal fin fold base, near the posterior margin of the
body pigment. The preanal fin fold was barely
perceptible. No gas bladder developed. The mouth
was very large. The maxillary extended to or
slightly behind the middle of the eye. The lower
jaw contained four sharply pointed, conical teeth
on each side. The fourth tooth was somewhat
smaller and located lower on the dentary. Body
proportions and total myomeres (38 or 39) were
similar to those reported by Warfel and Merriman
(1944) at this stage. The larvae remained mostly
on the bottom of the container, spending much of
the time on their sides possibly as a result of the
enlarged yolk sac. Efforts to swim were very
awkward and only made when the larvae were
disturbed.

Further Development

Near the end of January, the larvae were
observed to be positively phototactic. The yolk of
many of the larvae was absorbed by the end of the
first week in February. The peritoneum appeared
silvery through the skin. The pigmentation be-
came uniform olive grey over the body (Figure 2).
Specimens ranged between 14.0 and 17.0 mm TL
on 6 February. Those longer than 16.1 mm had
rudimentary caudal rays. The larvae were more

Figure l. — Hemitripterus americanus. Prolarva (newly hatched), 8 January 1975: 12.6 mm TL.

468

Figure 2. — HemitHpterus americanus. Early postlarva, 17 February 1975: 15.5 mm TL.

active by this time, but still spent most of the time
on the bottom. By 2 March, the larvae were no
longer attracted to light.

Toward the end of March, the caudal fin had 8 or
9 ray rudiments. Rays began to develop in the
second then first dorsal fins followed by the pec-
toral fins. The caudal peduncle remained unpig-
mented. Spines began to form on the preopercu-
lum. Greyish-tan fleshy tabs developed dorsally
behind the head and around the occiput.

By 20 April, the largest specimen (Figure 3) had
14 and 11 elements in the first and second dorsal
fins, respectively. The anal had 10 rays and the
caudal had 12, at about 20 mm TL. Both the
dentary and premaxilla had 15 teeth on each side.
The preopercular spines became more prominent.
The hypural plate began to form. The ratio of the
head length to total length was 3.6; of the pre-
dorsal length to total length, 3.8; and of the eye
diameter to head length, 2.7. The iris became less
silvery. Dense pigmentation developed on the in-
terradial membrane of the first dorsal fin between
elements 1 through 4 and 8 through 12. Similar
pigmentation developed in the second dorsal fin
(between elements 3 and 7) and the postanal fin

fold (between elements 3 and 6). Few melano-
phores were scattered between these dense areas
of pigment. This larva utilized much of the water
column during active periods and spent little time
on the bottom.

Acknowledgments

I thank Neal R. Foster of the Academy of
Natural Sciences of Philadelphia for providing
laboratory facilities and for his advice on the care
of the larvae and the preparation of the manu-
script, and Joanne and Wayne Laroche of the
School of Oceanography, Oregon State University,
for supplying comparative material to aid in the
identification of the larvae.

Literature Cited

BEAN, T. H.

1897. Notes upon New York fishes received at the New York
Aquarium, 1895—1897. Bull. Am. Mus. Nat. Hist.
9:327-375.
BIGELOW, H. B., AND W. C. SCHROEDER.

1936. Supplemental notes on fishes of the Gulf of
Maine. Bull. U.S. Bur. Fish. 48:319-343.

■'.''. '"■■'■■ *•■• ." .

■i / " /U'r.i'.-T'^^^Sfea!-^ > ■ ■■•■ ■'.•«•'■■.••..■••>:'".■"•■-. .'..'■••■••.•.-...•i: v. ••■'■;■

■iK^l

■■:'*.•

FIGURE 3.— HemitHpterus americanus. Postlarva, 20 April 1975: ca. 20 mm TL (the pectoral fin has been deleted for the sake

of clarity).

469

BIGELOW, H. B., AND W. W. WELSH.

1925. Fishes of the Gulf of Maine. Bull. U.S. Bur. Fish.
40(1), 567 p. (Doc. 965.)
Huntsman, A. G.

1922. The fishes of the Bay of Fundy. Contrib. Can. Biol.
1921:49-72.
WARFEL, H. E., AND D. MERRIMAN.

1944. The spawning habits, eggs, and larvae of the sea
raven, Hemitripterus americanus, in southern New Eng-
land. Copeia 1944:197-205.

LEE A. FUIMAN

Academy of Natural Sciences of Philadelphia

Nineteenth and the Parkway

Philadelphia, PA 19103

Present address: Department of Natural Resources

Cornell University

Ithaca, NY 14853

470

INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN

Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the
following instructions. These are not absolute requirements, of course, but desiderata.

CONTENT OF MANUSCRIPT

The title page should give only the title of the
paper, the author's name, his affiliation, and
mailing address, including Zip code.

The abstract should not exceed one double-
spaced page.

In the text, Fishery Bulletin style, for the most
part, follows that of the Style Manual for Biologi-
cal Journals. Fish names follow the style of the
American Fisheries Society Special Publication
No. 6, A List of Common and Scientific Names of
Fishes from the United States and Canada, Third
Edition, 1970. The Merriam-Webster Third New
International Dictionary is used as the authority
for correct spelling and word division.

Text footnotes should be typed separately
from the text.

Figvires and tables, with their legends and
headings, should be self-explanatory, not requir-
ing reference to the text. Their placement should
be indicated in the right-hand margin of the
manuscript.

Do not send original drawings to the Scientific
Editor; if they, rather than the photographic re-
ductions, are needed by the printer, the Scientific
Publications Staff will request them.

Each table should start on a separate page.
Consistency in headings and format is desirable.
Vertical rules should be avoided, as they make
the tables more expensive to print. Footnotes in
tables should be numbered sequentially in arabic
numerals. To avoid confusion with powers, they
should be placed to the left of numerals.

Acknowledgments, if included, are placed at
the end of the text.

Literature is cited in the text as: Lynn and Reid
(1968) or (Lynn and Reid 1968). All papers re-
ferred to in the text should be listed alphabetically
by the senior author's surname under the heading

Common abbreviations and symbols, such as
mm, m, g, ml, mg, °C (for Celsius), %, % and so
forth, should be used. Abbreviate units of mea-
sure only when used with numerals. Periods are
only rarely used with abbreviations.

We prefer that measurements be given in
metric units; other equivalent units may be given
in parentheses.

FORM OF THE MANUSCRIPT

TITLE PAGE

ABSTRACT

TEXT

LITERATURE CITED

APPENDIX

TEXT FOOTNOTES

TABLES (Each table should be numbered with
an arabic numeral and heading provided)

LIST OF FIGURES (entire figure legends)

FIGURES (Each figure should be numbered
with an arabic numeral; legends are desired)

ADDITIONAL INFORMATION

Send the ribbon copy and two duplicated or
carbon copies of the manuscript to:

Dr Bruce B. Collette, Scientific Editor

Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Fifty separates will be supplied to an author
free of charge and 100 supplied to his organiza-
tion. No covers will be supplied.

%'A.

■.-•■V. '■' ' ^ytJ:

Notes

HARRELL, LEE W., ANTHONY J. NOVOTJJY, MICHAEL H. SCHIEWE, and
HAROLD O. HODGINS. Isolation and description of two vibrios pathogenic to
Pacific salmon in Puget Sound, Washington " 447

MAY, NELSON, LEE TRENT, and PAUL J. PRISTAS. Relation offish catches in gill
nets to frontal periods 449

LANSFORD, LAWRENCE M., CHARLES W. CAILLOUET, and KENNETH T
MARVIN. Phosphoglucomutase polymorphism in two penaeid shrimps, Penaeus
brasiliensis and Penaeus aztecus subtilis 453

PERRIN, WILLIAM F. First record of the melon-headed whale, Peponocephala electra,

in the eastern Pacific, with a summary of world distribution 457 ^

CARLSON, H. RICHARD. Foods of juvenile sockeye salmon, Oncorhynchus nerka, in
the inshore coastal waters of Bristol Bay, Alaska, 1966-67 458

LAIRD, CHAEE., ELIZABETH G. LEWIS, and PAUL A. HAEFNER, JR. Occurrence
of two galatheid crustaceans, Munida forceps and Munidopsis bermudezi, in the
Chesapeake Bight of the western North Atlantic Ocean 462 -^'

WEIS, JUDITH S. Effects of mercury, cadmium, and lead salts on regeneration and

ecdysis in the fiddler crab, Uca pugilator 464 -

FUIMAN, LEE A. Notes on the early development of the sea raven, Hemitripterus
americanus 467

AMERICAS
/^^6.^9-T<& FIRSTINDUSTRY

tV GPO 696-333

Fishery Bulletin

National Oceanic and Atmospheric Administration • National Marine Fisheries Service

5 /VS'I'

MrES o^

Marine uiuloiioal i3Hon.ie.. .

'•V"^

(Continued on back cover)

fl \' COO- ^'

Vol. 74, No. 3 lU—- - July 1976

BREDER, CHARLES M., JR. Fish schools as operational structures 471

PARRACK, M. L. Estimation of fishing effort in the western North Atlantic from |

aerial search data 503 I

ARTHUR, DAVID K. Food and feeding of larvae of three fishes occurring in the f

California Current, Sardinops sagax, Engraidis mordax, and Trachurus sym- I

vietricus 517 I

FLANAGAN, CHRISTINE A., and JOHN R. HENDRICKSON. Observations on the |

\ commercial fishery and reproductive biology of the totoaba, Cynoscion macdonaldi, in |

! the northern Gulf of California 531 t

\ KORN, SID, NINA HIRSCH, and JEANNETTE W. STRUHSAKER. Uptake, |

I distribution, and depuration of ^"^C-benzene in northern anchovy, Engraulis mordax, f

and striped bass, Morone saxatilis 545 :

: SKILLMAN, ROBERT A., and MARIAN Y. Y. YONG. Von Bertalanffy growth i

curves for striped marlin, Tetrapturiis audax, and blue marlin, Makaira nigricans, *

in the central North Pacific Ocean 553

HOBSON, EDMUND S., and JAMES R. CHESS. Trophic interactions among fishes

t and zooplankters near shore at Santa Catalina Island, California 567

* WEINSTEIN, MICHAEL P., and RALPH W. YERGER. Protein taxonomy of the

I Gulf of Mexico and Atlantic Ocean seatrouts, genus Cynoscion 599

j ZWEIFEL, JAMES R., and REUBEN LASKER. Prehatch and posthatch growth of

I fishes-a general model 609

\ MILLER, R. J. North American crab fisheries: Regulations and their rationales . . . 623
j CLARKE, THOMAS A., and PATRICIA J. WAGNER. Vertical distribution and

other aspects of the ecology of certain mesopelagic fishes taken near Hawaii .... 635
MANZER, J. I. Distribution, food, and feeding of the threespine stickleback,
Gasterosteus aculeatus, in Great Central Lake, Vancouver Island, with comments

on competition for food with juvenile sockeye salmon, Oncorhynchus nerka 647

WESTERNHAGEN, HEIN von, and HARALD ROSENTHAL. Predator-prey
relationship between Pacific herring, Clnpea harengus pallasi, larvae and a preda-
tory hyperiid amphipod, Hyperoche medusarum 669

LAURS, R. MICHAEL, WILLIAM H. LENARZ, and ROBERT N. NISHIMOTO.
Estimates of rates of tag shedding by North Pacific albacore, Thunnus alalunga . 675

■ I

Seattle, Washington

U.S. DEPARTMENT OF COMMERCE

Elliot L. Richardson, Secretary

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION

Robert M. White, Administrator

NATIONAL MARINE FISHERIES SERVICE

Robert W. Schoning, Director

Fishery Bulletin

The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and
economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in
1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last
document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a
numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin
instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical,
issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office,
Washington, D.C. 20402. It is also available fi-ee in limited numbers to libraries, research institutions. State and Federal agencies, and
in exchange for other scientific publications.

EDITOR

Dr. Bruce B. Collette

Scientific Editor, Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Editorial Committee

Dr. Elbert H. Ahlstrom

National Marine Fisheries Service

Dr. William H. Bayliff

Inter-American Tropical Tuna Commission

Dr. Roger F. Cressey, Jr.
U.S. National Museum

Mr. John E. Fitch

California Department of Fish and Game

Dr. William W. Fox, Jr.

National Marine Fisheries Service

Dr. Marvin D. Grosslein
National Marine Fisheries Service

Dr. Edward D. Houde
University of Miami

Dr. Merton C. Ingham

National Marine Fisheries Service

Dr. Reuben Lasker

National Marine Fisheries Service

Dr. Sally L. Richardson
Oregon State University

Dr. Paul J. Struhsaker

National Marine Fisheries Service

Dr. Austin Williams

National Marine Fisheries Service

Kiyoshi G. Fukano, Managing Editor

Fishery Bulletin

CONTENTS

Vol.74, No. 3 July 1976

BREDER, CHARLES M., JR. Fish schools as operational structures 471

PARRACK, M. L. Estimation of fishing effort in the western North Atlantic from
aerial search data 503 -^

ARTHUR, DAVID K. Food and feeding of larvae of three fishes occurring in the
California Current, Sardinops sagax, Engraulis mordax, and Trachurus sym-
metricus 517

FLANAGAN, CHRISTINE A., and JOHN R. HENDRICKSON. Observations on the
commercial fishery and reproductive biology of the totoaba, Cynoscion macdonaldi, in
the northern Gulf of California 531

KORN, SID, NINA HIRSCH, and JEANNETTE W. STRUHSAKER. Uptake,
distribution, and depuration of ^^C-benzene in northern anchovy, Engraulis mordax,
and striped bass, Morone saxatilis 545

SKILLMAN, ROBERT A., and MARIAN Y. Y. YONG. Von Bertalanffy growth
curves for striped marlin, Tetrapturus audax, and blue marlin, Makaira nigricans,
in the central North Pacific Ocean 553

HOBSON, EDMUND S., and JAMES R. CHESS. Trophic interactions among fishes
and zooplankters near shore at Santa Catalina Island, California 567

WEINSTEIN, MICHAEL P., and RALPH W. YERGER. Protein taxonomy of the
Gulf of Mexico and Atlantic Ocean seatrouts, genus Cynoscion 599

ZWEIFEL, JAMES R., and REUBEN LASKER. Prehatch and posthatch growth of
fishes-a general model 609 '''

MILLER, R. J. North American crab fisheries: Regulations and their rationales . . . 623 "^

CLARKE, THOMAS A., and PATRICIA J. WAGNER. Vertical distribution and
other aspects of the ecology of certain mesopelagic fishes taken near Hawaii .... 635-^

MANZER, J. I. Distribution, food, and feeding of the threespine stickleback,
Gasterosteus aculeatus, in Great Central Lake, Vancouver Island, with comments
on competition for food with juvenile sockeye salmon, Oncorhynchus nerka 647

WESTERNHAGEN, HEIN von, and HARALD ROSENTHAL. Predator-prey
relationship between Pacific herring, Clupea harengus pallasi, larvae and a preda-
tory hyperiid amphipod, Hyperoche medusarum 669

LAURS, R. MICHAEL, WILLIAM H. LENARZ, and ROBERT N. NISHIMOTO.
Estimates of rates of tag shedding by North Pacific albacore, Thunnus alalunga . 675

(Continued on next page)

Seattle, Washington

For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington,
D.C. 20402-Subscription price: $11.80 per year ($2.95 additional for foreign mailing). Cost per
single issue - $2.95.

Contents— continued

Notes

ZIMMERMAN, STEVEN T., and ROBERT S. McMAHON. Paralytic shellfish poi-
soning in Tenakee, southeastern Alaska: A possible cause 679

COLLINS, JEFF. Oil and grease: A proposed analytical method for fishery waste
effluents 681

GADBOIS, D. F., E. M. RAVESI, and R. C. LUNDSTROM. Occurrence of volatile
N-nitrosamines in Japanese salmon roe 683

WATKINS, WILLIAM A., and WILLIAM E. SCHEVILL. Underwater paint mark-
ing of porpoises 687

EDGAR, ROBERT K., and JAMES G. HOFF. Grazing of freshwater and estuarine,
benthic diatoms by adult Atlantic menhaden, Brevoortia tyrannus 689

JOHNSON, ALLYN G. Electrophoretic evidence of hybrid snow crab, Chionoecetes
bairdi x opilio 693

KORN, SID, JEANNETTE W. STRUHSAKER, and PETE BENVILLE, JR. Effects
of benzene on growth, fat content, and caloric content of striped bass, Morone
saxatilis 694

STAEGER, WILLIAM H., and HOWARD F. HORTON. Fertilization method
quantifying gamete concentrations and maximizing larvae production in Cras-
sostrea gigas 698

Vol. 74, No. 2 was published on 16 July 1976.

The National Marine Fisheries Service (NMFS) does not approve, recommend or
endorse any proprietary product or proprietary material mentioned in this
publication. No reference shall be made to NMFS, or to this publication furnished by
NMFS, in any advertising or sales promotion which would indicate or imply that
NMFS approves, recommends or endorses any proprietary product or proprietary
material mentioned herein, or which has as its purpose an intent to cause directly or
indirectly the advertised product to be used or purchased because of this NMFS
publication.

FISH SCHOOLS AS OPERATIONAL STRUCTURES

Charles M. Breder, Jr.

ABSTRACT

The interaction of a space lattice, vortex trails, and the lubricity of fish surface mucus is shown to be
important to the operation and structure of fish schools and significant in terms of locomotor eflSciency.
This is independent of the various interpretations of possible survival values, protection from
predation, and similar ideas-all of which are extremely difficult to prove, even if valid.

A single type of space lattice is shown to approximate the arrangement of fishes in a school on the
basis of geometrical reasoning. This is supported by empirical data.

The vortex trails left by each fish, when the fishes are deployed according to the "fish school" lattice,
lead each following fish into a series of vortices at a point where the water flow is traveling in the
direction in which they are swimming.

The lubricity of the mucus-water mixture that the fish ahead leaves in its vortices decreases the drag
on the following fish.

The advantages of the regimented life in a school, as against the freedom of action common to the
more or less solitary life, are evidently related to the effectiveness of the drag-reducing mucus in the
vortices. The fishes with the least effective mucus appear to take advantage of the schooling life while
those with the most effective mucus are more likely to be solitary.

The past decade has witnessed^a considerable
increase in output of papers addressed to a better
understanding of the numerous phenomona pre-
sented by fish schools. These documents have
covered a wide variety of the inherent problems.
Nonetheless, there remain some basic questions
that have proved peculiarly elusive, such as the
nature of the evident regularity of the positional
relationships of individuals in well organized
schools and the nature of influences that hold the
school members in their regular patterns. A fish
school is considered here as a group of polarized
individuals that operates as a unit between the
times of its resolution and eventual dissolution.
Initially, the activity of the fishes crowding
together in their polarized pattern creates the
structure of which they form components. Once
established, the school efficiently regulates the
locomotor activities and general comportment of
the organized fishes.

The primary purpose of this paper is to show
that both the geometrical pattern of the space
lattice approximated by schooling fishes and the
surface mucus on their bodies are mutually im-
portant elements in the formation and mainte-
nance of fish schools. The physical bearing of these
two elements is direct and important, each in its

own right, to an understanding of any theory that
attempts to explain the origin of schooling without
recourse to theoretical interpretations.

How much of the schooling phenomenon ob-
served in modern fishes is a result of interactions
between the swimming capabilities of the fishes
and the physical restrictions imposed by their
environment, as compared with other biological
needs, is not readily determined. However, the
experiments described here are in some cases
suggestive. These experiments, primarily under-
taken to establish data relevant to the basic
purposes of this study, in each case, have been
carried only as far as was necessary to make a
point. Many of them could be extended into much
greater refinement with the promise of worth-
while further elucidation.

This work leads to a number of lines of possible
approach to the problems of school organization.
Some of the newer items discussed have had the
benefit of recent studies-remote from schooling
problems and in some instances remote from
biology. This is especially marked in those studies
that are dependent on developments in hydro-
dynamics during the last decade.

FISH SCHOOLS AS SPACE LATTICES

'Mote Marine Laboratory, 9501 Blind Pass Road, Sarasota, FL
33581.

Manuscript accepted Februarj* 1976.
FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

To further the understanding of the physical
organization displayed by schools of fishes, a study

471

FISHERY BULLETIN: VOL. 74, NO. 3

of their geometrical characteristics has been
undertaken. Much of the older literature on the
distribution of individuals of a population, or
smaller group, of animals or plants took for
granted that the deployment is stochastic. Clark
and Evans (1954) stated, "This assumption is no
longer a tenable one and is probably even less
applicable to animal populations." It is, of course,
doubtful if creatures with well organized locomo-
tor abilities and complex sensory systems are ever
distributed in a fully random manner. The systems
encountered in nature seem to be mostly those of
ordered arrays variously distorted by processes of
many kinds, sometimes obvious, but more often
obscure or barely discernible. Attempts to mea-
sure the structure of assemblages of individuals
have been predicated mostly on the idea of show-
ing the extent of their departures from theoretical
randomness. Since fully organized fish schools
have very obvious structure, it is at least equally
appropriate to compare them with mathematically
organized patterns, especially where there are
good theoretical reasons to expect the presence of
some similarity.

Geometrical Models

The establishment of a geometrical model of a
fish school is relatively simple, for whatever else a
fish schooF may be, it is essentially a closely
packed group of very similar individuals united by
their uniformity of orientation. A more explicit
definition has been given by van 01st and Hunter
(1970) who stated, "The principal characteristics of
the organization of fish schools are that the in-
dividuals stay together, tend to head in the same
direction, maintain even spacing, and the activi-
ties of the individuals tend to be synchronized."
Because of the nature of fish locomotion it is
necessary that a certain amount of swimming
room be maintained by each fish (Breder 1965, van
01st and Hunter 1970). Thus each fish and a "shell"
of water about it may be considered as a unit, and
a school as a packing together of these units. Such
structures can be handled by established math-
ematical procedures. The fact the fishes are all
moving forward and, in many instances, often
shifting their relative positions merely makes the

handling of such data a little tedious, but does not
vitiate the basic propositions.

One approach to the analysis of the structure of
a fish school, the empirical, can be made by mea-
suring the distance, angle, or other parameter
between a given fish and the other members of the
school. The mathematical manipulation of such
measurements can establish values that may serve
as an index to the school's organization. One's
imagination alone limits the selection of data.
Papers that have employed this type of approach
include Keenleyside (1955), Breder (1959, 1965),
Cullen et al. (1965), Hunter (1966), van 01st and
Hunter (1970), Symons (1971a, b), Healey and
Prieston (1973), Weihs (1973a), and Pitcher (1973).
Only Cullen et al., Symons, and Pitcher in the
above list attempted complete tridimensional
measurements. Pitcher's paper has important
bearing on the approach developed here on the
basis of abstract reasoning. It will be discussed in
detail later.

A theoretical approach, equally valid, is based on
tridimensional geometrical concepts and con-
structs for purposes of comparison with fish
schools. Since there is an infinite variety of such
constructs possible, only those of some conceivable
application to this study are discussed here. Unlike
the empirical approach, there are evidently no
prior papers that have employed this theoretical
one. The following treatment has been made
especially explicit because of the complex rela-
tionships within both space lattices and space
packings, as some biologists who might consult
these pages may not have instant recall of such
details.

It is necessary to introduce some elementary
data on tridimensional lattices that are essential
to an understanding of their bearing on fish
schools.^ The most readily visualized space lattice
is that in which a cube is the element or cell
(Figure lA). It is not the closest possible packing
of such points: a closer one can be obtained by
figuratively pushing the cubic lattice askew
(Figure IB) so that the special case of cubes with
their 90° angles become rhombohedrons with
other angles. The dotted arrow in Figure IB
indicates the amount of travel of the point in the
upper left front corner of the lattice in attaining

-Definitions of this word as used here are given by Breder
(1959, 1967). For an extended discussion of this and other usages
see Shaw (1970).

'Support of ail geometric statements made in this section may
be found in any formal or informal geometry text covering the
area concerned, such as Hilbert and Cohn-Vossen (1952) and
Lines (1965).

472

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Figure l.-Two space lattices in perspective, each with a single
cell shown as a solid. A. The cubic lattice. B. The rhomboidal
lattice. The arrow indicates the manner of transformation by
which the cubic lattice becomes the rhomboidal lattice.

the transformation from cube to rhombohedron.
All the angles in this rhombic lattice are either 60°
or 120°, SO transformed from the cubic lattice with
only angles of 90°. On the floor of the cubic lattice
in Figure lA, the nearest points to the central one,
in the same plane, are four in number. These are
connected to each other by a dotted line. On the
floor of the rhombohedron in Figure IB, the sides
of which have internal angles of 60° and 120°, the
nearest points to the central point include four at
the corners of the dotted parallelogram plus two
more, indicated by the dark points. These define a
regular hexagon because the parellelograms are
composed of two equilateral triangles.

If models of identical fishes are stationed with
their centers at each lattice point, and if all the
models are in parallel orientation, the group
superficially resembles a fish school. It becomes
immediately apparent however that such a lattice
of fishes has characteristics that are never seen in
a school. If they had ever been seen in such a
formation, their appearance would have been so
striking that the details of the regimentation
would have been recorded long ago. In such a
school, viewed from above, fish would be seen in
horizontal files and these files would be swimming
ahead in rows transverse to their direction of
travel. Viewed from the side, each fish within the
school would have another directly above and
another directly below it, forming columns, except
the two fish marking the upper and lower limits of
the school in each vertical column of fishes. These
two would be without another fish above and
below, respectively. Thus we can temporarily put
this unschoollike lattice aside.

Fish models positioned at the points of the
rhombic lattice do not show the peculiar features
seen in the cubic lattice, but have a more distinct
resemblance to fish schools. It is difficult to deny

that schooling fishes, in most situations, are indeed
approximating this configuration, the details of
which will be discussed later.

Turning now from space lattices to the packing
of space, it is easy to arrive at the above rhombic
lattice by a very different route. As a preliminary
mathematical simplification, fishes and the im-
mediately surrounding water that envelops each
fish individually in a school shall be equated to
spheres, the centers of which are located on the
axis of the fish midway between the end of the
snout and the tip of the tail. Here it is necessary to
describe some of the less obvious geometrical
features of a mass of spheres packed together as
closely as possible. A single layer of identical
spheres on a plane surface packed at maximum
density may be represented on paper by an
equivalent packing of circles (Figure 2). A hexa-
gon may be circumscribed about each circle, one
of which is shown in the lower left corner.

Figure 2.-The closest possible packing of a single layer of
identical spheres or circles, showing the relationships to hexa-
gons and their six equalateral triangles as well as the disposi-
tion of a single diameter in each circle when drawn radiating
from the center of the circle with the circumscribed hexagon.

The individual diameters of each circle as shown
in Figure 2 lie along radiating lines emanating
from the common center of the hexagons.^ Those
lying on the radials passing through the apices of
the larger hexagon are continuous lines (major
axes), while those passing through the equivalent
points on the smaller hexagon are dashed lines
(minor axes). If these diameters are all permitted
to become parallel to one another, a very different

^Although simple, this geometric treatment of transforma-
tions of related diameters of packed circles or spheres is
evidently original here, or at least no approach to this treatment
has been found. No formal proofs are necessarv as the usage here
is simple enough to be self-evident and would be irrelevant to
present purposes.

473

FISHERY BULLETIN: VOL. 74, NO. 3

situation appears. This may be conceptually
treated as though the diameters were under some
common influence, somewhat like iron filings in a
rectilinear magnetic field. Figure 3 shows such an
arrangement, where all diameters are in the first
case at an angle of 30° to a major axis and 15° in
the second case. Obviously the continuous lines of
the major axes of Figure 2 are no longer possible
except when the diameters are at one of the three
angles of the major axes, where in each case such a
drawing would show only a series of continuous
parallel lines. In any of these parallel arrange-
ments the distances of the diameters from end to
end are constant throughout as are the distances
from side to side. These two dimensions change
only if the angle between the diameters and major
axes is changed, as can be seen by comparing
Figures 4 and 5 based on a square with Figures 2,
3A, and 3B based on a hexagon.

These two types of packing may now be con-
sidered in their more complex form in three
dimensional space. The cubic space lattice is very
simple and will be referred to later; the rhombic
spatial array, more likely to be confusing, is

Figure 3. -Parallel diameters drawn on the form of Figure 2. A.
Based on diameters halfway between two major axes, 30° from
either. B. Based on half the angular distance used in A, 15°.

Figure 4.-Cubic packing of a single layer of spheres or circles,
directly comparable with Figure 2.

Figure 5.- Parallel diameters drawn on the frame of Figure 4,
based on diameters halfway between two consecutive axes, 45°
from either. Directly comparable with Figure 3.

discussed in sufficient detail for present needs.
Starting with the single layer of spheres of Figure
2, another layer may be placed upon it so that each
sphere of the second layer rests in the hollow
between three adjacent spheres of the first. The
second layer automatically has a pattern identical
to the first, but the centers of all the spheres of the
upper layer are displaced so as to fall over the
centers of an equilateral triangle connecting the
centers of the supporting first layer spheres. This
is shown in Figure 6 where the centers of the first
layer spheres are indicated by large circles and
those of the second by smaller dark circles. The
dash-line hexagon of Figure 6 indicates the dis-
placement of the second layer centers. It also
shows that just three second layer sphere centers
are within the solid-line hexagon. There are also
shown three similar small open circles forming a
similar pattern within the hexagon, which indicate
the absence of spheres centered by them, and
connected by dotted lines to form a hexagon of
absences. In the upper left corner of this same

474

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Figure 6.-The rhombohedral sphere pack of three layers as
viewed from above. Based on Figures IB and 2. See text for full
description.

figure parts of the adjacent outlines of four
spheres of the first layer are shown by solid lines.
The dotted lines of three overlying spheres of the
second layer are also shown. This indicates clear
vertical passages through the overlying junctures
that permit passage through the two layers of
spheres (small open circles) while two are blocked
by overlying spheres (small dark circles). The
pattern is repeated throughout the system. A third
layer of spheres may be identical with the first, a
fourth layer identical with the second, and so on
indefinitely. This pattern preserves the integrity
of the vertical passages, but this need not be the
case. If the second layer of sphere centers ex-
changes the position of the black and open small
circles, the clear passages occur where the black
circles are now shown and vice versa. As any layer
may be so reversed the passages may be blocked in
many complicated patterns. The shortest possible
passage can be the vertical distance between the
level of the centers of two adjacent layers of
spheres, otherwise the passage may be indefinitely
long.

As these planes, referred to above as layers,
form the faces of the generating rhombohedron
shown in Figure IB, these passages run in three
intersecting directions, as do the three planes of
the lattice. The passages are all interrelated, as
altering the relationships of the sphere centers in
one plane automatically alters those in the two
others.

The above may be simpler to visualize by refer-
ring to the perspective illustration of Figure 7.

Figure 7.-The rhombohedral packing of spheres in perspective,
showing only sphere centers. Two other sets of planes could be
drawn through these centers at angles determined by the sides of
the generating cell, two of which are shown between layers 4 and
5. See text for full explanation.

Here it has been necessary to completely alter the
symbols used in Figure 6 owing to other needs.
Plane 1 of Figure 7 is identical with the first layer
of Figure 6. The hexagon of Figure 6 is shown in
Figure 7 as one of dotted lines. Planes 1, 2, and 3 of
Figure 7 represent the corresponding layers of
Figure 6. The two added planes, 4 and 5, show more
realistically the vertical passage running from A
to A. It has clearance through the first three planes
but is blocked at plane 4 and runs clear through 5.
Note that plane 4 is "reversed" from 2, which is the
reason for the blockage. The passage from B to B is
blocked by planes 1, 3, and 5, but not by 2 and 4.

The indications of the rhombohedral cells by
dotted lines between planes 4 and 5 clearly show
how two additional sets of planes could be passed
through the points.

A perspective view of the simpler cubic packing
of spheres is shown in Figure 8 for comparison
with Figures 1, 4, 5, and 7. Only four planes are
shown, as more are unnecessary. It is evident that
the cubic cell and consequent total right angled
construction precludes any of the rhombic
complications.

These two systems of packing spheres are all

475

FISHERY BULLETIN: VOL. 74, NO. 3

Figure 8.-The cubic packing of spheres, directly comparable
with Figure 7. See text for full explanation.

that will be considered here, as all others are much

looser and are not relevant to this study. The

density of these two and the number of contacts

that interior spheres have with others are given

below.

Percent of volume N umber of contacts
Packing occupied of each sphere

Rhomboidal
Cubic

0.740
0.513

The number of contacts indicated here are iden-
tical with the number of "nearest neighbors"
mentioned in reference to the equivalent space
lattices.

Pitcher's (1973) data on clusters of spheres
presented another way of explaining the com-
plications of close sphere packing. It emphasizes
the measurements from center to center, with
which he was working, rather than the overall
pattern of a larger group, which emphasizes the
layering effect of polarized parallel diameters
discussed here.

Structure and Functioning of
Natural Schools

The series of diagrams in the preceding section
is virtually a key to determining what, if any,

space lattice a given school of fishes could approx-
imate and it clearly indicates what types of space
lattices do not find their embodiment in fish
schools. Reason and observation also indicate that
school-forming fishes establish their schools
rapidly with great unanimity of action. The
schools come to stability only after each individual
has the common orientation, all normally as close
together as the spatial requirements of their
individual propulsive acts permit. The organiza-
tion is strictly one formed in this manner and
without any of the differential behavior that more
complex lattices would require.

Pitcher (1973), by purely empirical means, ar-
rived at the geometrical relationships of a school
of Phoxinus phoxinus (Linnaeus) identical with
the present formal lattice reached by theory. His
fishes fit our theoretical operations even better
than any of the fishes checked for this study. Our
material all showed some attenuation of the lattice
along the axis of travel, which also was the case in
Weihs (1973a). This may simply mean that Phox-
inus keeps a tighter school than any species we
checked, or that there is some small effect here
that relates to speed of fish and their absolute size.
Possibly, however, it may be related to a differ-
ence in behavior between a school swimming
ahead in quiet water and one holding a stationary
position in flowing water, as did Pitcher's fish. In
the latter, optical fixation on fellow fishes and
some background feature is possible, but in the
former, fixation is only possible on other members
of the school as the background apparently drifts
backward. If this effect does modify the spacing of
the fishes, stationary schools in fast flowing rivers
where backgrounds are visible should more closely
approach the theoretical.

Spacing of Fishes

Using the preceding examination of lattices and
the packing of spheres, a preliminary comparison
with fish schools may start by continuing the
equating of fishes in a school to the diameters of
the packed spheres. Schooling fishes should not be
expected to space themselves exactly as spheres
and they do not do so in precise detail, see Pitcher
(1973), but a basic resemblance exists.

If the rigid sphere of geometry be mentally
replaced by a soft rubber ball, the approximation
comes closer to that of a fish embedded in a school
of its fellows. Thus a group of such balls, when
packed together, are subjected to slight flattening

476

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

and to other minor distortions where contacts are
made with other balls, all proportional to the
amount of pressure and its direction. The pattern
of lattice considered here as closest to the spatial
distribution commonly shown by schooling fishes
can be reached by very simple transformations.

The calculations that equated the diameters of
the spheres to the fishes' lengths can be altered.
Here the lengths are changed but the positions of
fishes in space remain the same.

A change that evidently does occur regularly
involves altering the angles in the quadrilateral
mesh composed of two triangles as illustrated in
Figure 9, where A and B represent the quadri-
laterals in Figure 1, and C represents a quadrila-
teral that has been used by Weihs (1973a) in
connection with his studies on vortex streets. It is
called simply a "diamond" by that author. His
model resulted from considerations of energy
saving requirements. The Weihs (1973a) diamond
can be used as a very convenient basic unit or celP
characteristic of the fish school lattices, without
altering any of concepts discussed here. At this
writing, all known changes from the conditions of
regular geometrical figures are on the side of
increased differences between the two pairs of
angles of the diamond. No instances have been
found in real fish schools that would lie between
case A and B of Figures 1 and 9, unless the
widespread separations which have been con-
sidered as degenerating schools are included. All
other variations found are on the far side of B
except for the data of Pitcher (1973), which is
precisely at B. In Figure 9, A shows the square
pattern with 90°, B shows the 60°, 120° rhombus,
and C shows a rhombus with 30°, 150° which
depicts a condition frequently seen in fish schools
and is, as already indicated, the Weihs (1973a)
diamond. Carrying this angular reduction further,
the end is reached as the side to side distance
between fishes is reduced to zero, so that the total
length of the figure becomes a single line equal to
twice the length of a side of the diamond. At the
other end of this series of quadrilaterals, an
increase beyond 90° produces another series. In

^In most schooling fishes two individuals, if isolated from the
others, will swim together side by side or with one diagonally
ahead of the other. If three fish are so isolated, they will normally
form a pattern of three points of a diamond. In this case there is
usually much more shifting around than in the case of two, while
four fish tend to form a diamond. It has been a common practice
for workers in this field to consider these cases of very small
schools. From groups of less than four, it is impossible to make
any reasonable estimate of the shape of the diamond. Some
judgment can normally be obtained from a group of four,
although even that might vary somewhat from a school.

Figure 9. -Three quadrilaterals (lattice elements) as related to
"diameters" or "fish lengths." See text for full explanation.

this case the final result is also a single line, equal

to twice the length of a side but at right angles to
the one reached at the other extreme of the series,
as described above.

Figure 10 shows how these matters relate to the
hexagons and how the quadrilaterals relate to an
entire school. Each small circle in the upper row of
three diagrams represents the midpoint of each
fish. The four fishes, each on a diamond point, are
represented by heavy horizontal lines represent-
ing the fish lengths. The direction of swimming is
understood to be from left to right. All the others,
shown only by the small circles, are moving paral-
lel to and in the same common direction as the four
indicated. Starting at A with a square and passing
to B composed of two equilateral triangles, the
series terminates at C with acute angles of 30°,

Figure lO.-The relations of the three quadrilaterals shown in
Figure 9 to the station points in a school and to the corresponding
hexagons (upper row). The clear turning sectors and those
requiring a too close mutual approach are shown in the lower row.
The latter are marked by their two radii and arc by a heavy solid
line. Their axes, the lines of contact, are marked by dashed radii.
See text for full explanation.

477

FISHERY BULLETIN: VOL. 74, NO. 3

which is the most compressed of the three. Also
shown are the relations between the quadrilater-
als and the corresponding hexagons, as well as the
number of fishes in a given area.

Continuing with Figure 10, it is obvious that the
direction of travel could be in any other horizontal
direction of swimming, than the one shown here.
It should be noted however, that the lattice of each
shows that if the fish turned so as to be parallel
with any edge of their parallelogram, the fishes
would all be brought to the nose to tail position,
something which does not occur.

In the lower row of three corresponding dia-
grams in Figure 10 the dashed radial lines show
the directions of swimming that would place the
fish in contact. The clear spaces indicate where the
passages are unobstructed. The enclosed areas,
which surround the dotted lines of contact, meet
the clear areas at a point halfway between that
line and the centers of the clear areas, except in C
which is not based on a regular quadrilateral or
hexagon. This will be further discussed under
Problems of a School Turning.

In any school, a certain minimum distance from
the nose of a following fish to the tail of a leading
fish is maintained. The evident need for this
separation is natatorial. Requirements differ with
the various types of fishes that form schools.
Although fishes do not leave wakes behind, as does
a motor-propelled ship, there is still the matter of
dying vortices (Rosen 1959; Breder 1965). This
alone could account for the need of a spatial lead.
Conceptually, fishes could swim satisfactorily on
any of the diameters shown in Figures 2 and 3,
except those on the major axial lines. The min-
imum distances between these diameters (fishes) in
a line occur halfway between these axes as in
Figure 2. It is to be noted also that the horizontal
rows of diameters tend to line up so that the
diameters are not all the same distance from each
other as in Figure 3A. This change continues with
angles less than 15° so that when these diameters
become horizontal they are in end-to-end contact,
producing a series of parallel lines. This is merely a
matter of the geometry of the uniform rotation of
the diameters. No schooling fishes would tolerate
this condition, but would adjust their positions to
lie near midway between the positions of those
lateral to them, as shown in the diagrams of
Figure 9. Compare Figure SB with Figure 9C. The
apparent differences between the two are entirely
owing to the fact that the first diagram is based on

rigid circles, or spheres, and the second does not
have that heavy stricture. The three quadrilaterals
in Figure 9 can be considered as making a closed
curvilinear figure, where Figure 9A would be
circumscribed by a circle while Figures 9B and 9C
would both be circumscribed by ellipses, Figure 9C
being much narrower than Figure 9B. This trans-
formation can be brought about by increasing the
head-to-tail distances of the fishes in a single file
and decreasing the distances between adjacent
files.

The greatest width between the tracks of fishes
swimming parallel is also at the halfway angle
between two successive axes, as shown in Figure
3A. As long as all the fishes are swimming in
parallel courses the distance need not vary, as seen
in Figure 3A. The closer this angle approaches an
axis, the smaller becomes the distance between the
parallel tracks, indicated in Figure 3B. The dis-
tance between fishes, head to tail, varies inversely
as an axis is approached.

Still photographs cannot give the sense of a
regular pattern of fishes that is evident on viewing
a school or a motion picture. Because of these
conditions, in those photographs shown here
sufficiently open to see the fishes distinctly, they
appear as rather ragged groups. Thus in Figure 11
of Katsuwonus pelamis (Linnaeus), only frag-
ments of some regularity of pattern can be seen.
Those on the left of center show the pattern of a
loose school while those on the right are breaking
ranks for feeding. This picture, however, indicates
several lines of fish alignment, some running from
top downwards to the right and others to the left,
from which the relationship to the diagram in
Figure 7 can be seen within the limits of a still
picture.

Species attaining very large size, such as Thun-
nus thynnus (Linnaeus), tend to have dispropor-
tionally greater distances between individuals
when large, as compared to their younger and
smaller sizes (see Breder 1965). Contrary to this,
van 01st and Hunter (1970) showed that other
smaller fishes {Scomber, Engraulis, Trachurus,
and Atherinops) tighten their ranks as they grow
from larvae to near adult size, some abruptly and
others gradually.

Hunter (1966) presented some data on the or-
ganization of fish schools for purposes that do not
concern present interests. However these data,
based on motion picture analysis shown in his
figure 2, have a distinct bearing on some features

478

BREDER: PISH SCHOOLS AS OPERATIONAL STRUCTURES

Figure U.-A school of Kafsnwnnus pelamia off the Hawaiian Islands, breaking up for surface feeding. Courtesy of the National

Marine Fisheries Service, Honolulu Laboratory, Honolulu, Hawaii.

of this Study. Figure 12 is based on Hunter's
figure, modified appropriately for this analysis.
Although the small group used, six captive in-

10 CM

dividuals of Trachurus symmetricus (Ayres), is
not large enough to form a well organized school
and even has members that do not always stay
precisely at the same level as the others, it is
exceptionally interesting in that it does display
items pertinent to school structure.

Figure 12 represents the progress of the six fish
covering S% s shown on 100 frames of motion
picture film exposed at a rate of 12 frames/s. The
larger circles indicate the mean values of the eight
positions of the snouts of each of the six fish. These
means are connected serially by straight lines.^
The small circles indicate the patterns of positions
of the six fish's snouts for four of the eight means.
Every other one has been omitted because ad-
jacent patterns overlap enough to be confusing.

Figure 13 indicates the manner in which the
values are related to the trajectory of the group.

Figure 12.-Analysis of the location of six Trachurus sym-
metricus in a school, shown by successive eight steps in their
travel. Based on data of Hunter (1966) and his figure 2. Only the
odd-numbered positions have the individual fish positions
indicated. To show them all would confuse rather than clarify.

''Hunter (1966) recognized three turns in his figure 2. For
present purposes the sequence is given six turns, as indicated in
Figure 12 and Table 1. His three indices, mean separation,
distance to nearest neighbor, and angular deviation represent
other measures of the same activity, all of which relate to the
differences of the mathematical approaches involved.

479

FISHERY BULLETIN: VOL. 74, NO. 3

O o

5 3

6 6°

4 o
5

5 1 „3
— O-

FiGURE 13.-The positions of each of the
six fishes (small circles) and their means
(large circles). The arrowheads indicate
the directions of travel of the group.
These lines pass through the means and
the transverse lines intersect to divide
the field into quadrants.

Also shown is the pattern of each fish's distribu-
tion, together with the means/ the momentary
swimming direction of the school, and a line at
right angles to it intersecting at the mean posi-
tion. This device divides the area in which the
fishes occur into quadrants. The data for this are
given in the first part of Table 1. The precise
positions of the fishes were picked from Hunter's
(1966) figure 2 and have been handled by graphic
methods in the construction of the diagrams
shown in Figures 13 and 14. The numerals attend-

'These means were obtained by separately spreading each of
the eight positions of the six fishes on Cartesian graph paper and
determining their X and Y values and the means.

ing the positions of the fishes, actually the tips of
their snouts, in Figures 12, 13, and 14 are those
used by Hunter (1966) to differentiate the in-
dividuals and they have no other significance here.

It is immediately apparent that fish number 6 is
in the front quadrants continuously. Replotting
this data according to the total number of each fish
separately as in Figure 14A, other features ap-
pear. Figure 14B, which shows the means of
Fig^ure 14A, does indeed approximate the Weihs
(1973a) diamond.

Considering the manner in which the data have
been assembled— captive fishes in a tank, the
curvature of their paths, the difficulties in es-
timating the path of the school, and its generally

Figure 14.-The location of each of the
six fishes (small circles) at each of the
eight positions. A. The larger light
circles with the intersecting lines pass-
ing through them are those shown in
Figure 13. The large dark circles show
the locations of the mean positions of
each of the six fishes. B. The large light
circles represent the means of the six in
A. The small circles show the mean
position of each fish (the dark circles of
A).

«4

•o5
•3

, ° O

6° o

4- J

480

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Table 1. -Location of individual fishes (Trachurus) by quadrants and by halves.
Data based on figure 2 of Hunter (1966). (First letters, L and R = left and right.
Second letters, F and R = front and rear.)

Location of

each fish (Se

e Figure

13)

School's

By quadrants

By h

alves

position

1 3

456

1 23

1 3

456

1 23

235

1 46

235

456

1 2

345 6

1 2

3 56

1356

1 3

1 36

245

1 3

1236

1 36

245

1 3 6

245

Quadrants and halves occupied by indiv

idual fishes (See Figure

14)

Fish

Byqi

uadrants

By h,

alves

no.

35678

1 24

35678

1 24

1 28

3456

1234568

4678

1 23

45678

1 23

1 2

5678

1234

5678

1 23

678

12345

678

1-7

1-8

loose nature- it is remarkable that any such ap-
proximation to a regular figure could be found.
This material indicates that the influence tending
to hold schooling fishes to approximating figures
this close to geometrical regularity is effective
even in assemblages of fishes barely coming with-
in our definition of the word.

Healey and Prieston (1973) brought out a very
interesting feature of schools by the application of
multivariate analysis. This is evidently closely
related to the preceding geometrical study on the
data presented by Hunter (1966). The problem of
the origins or the reasons for the existence of these
individual variations in fish movements within a
school is not yet susceptible to a general solution.
Clearly some are caused by extrinsic stimuli and
some by intrinsic causes, such as the physiological
state of the individual. Healey and Prieston (1973)
wrote that their data suggested, ". . . that there
may be a short-term and a long-term organization
within the school." Possibly this could eventually
be referred to equivalently short- or long-endur-
ing stimuli, not grossly evident to the observer.
The data of McFarland and Moss (1967) and Moss
and McFarland (1970) may represent an intrinsic
short-term event, in this case being a reduction in
oxygen tension. Alekseeva (1963) showed that
various fishes have a greater oxygen consumption
when visually isolated from their fellows. Such
individuals, if able to see the others, do not.
Schuett (1934), Escobar et al. (1936), and Breder
and Nigrelli (1938) indicated that individuals of
Carassius auratus (Linnaeus) swam faster when

alone and when crowded, but slower when with a
few companions. This should be reflected in their
oxygen demand and may account for the results of
Alekseeva (1963).

The very short duration of the Hunter (1966)
data suggests that the details here might be based
on intrinsic sources, as in the case of the fish that
kept the leadership of the school and of the one
that brought up the rear. It is conceivable that
these may be the consequences of the individual
physiological states.

In agreement with Bowen (1931, 1932) and
Radakov (1972), there is no convincing evidence
that the superficial appearance of "leadership," to
be seen occasionally, supports such a view.
Hunter's (1966) data covered only 8V3 s. Breder
(1959) suggested that "white" Carassius auratus
(Linnaeus) seem to take the leadership in schools
otherwise composed only of "yellow" individuals.
This finding of white fishes in leading positions is
apparently related to the greater conspicuousness
of the white fish as compared with the yellow in a
lily pond environment and is not an indication of
leadership by any individual.

Radakov's (1972) data, which was extensive and
important, considered "leadership" in a rather
different sense than the others. He considered
numbers of leaders up to 40% of the number of
fishes comprising a school. The front fishes, with
no other fishes ahead of them, are considered here
as leaders. These fishes are in a different physical
category as they have none of the advantages of
being a following fish.

481

FISHERY BULLETIN; VOL. 74, NO. 3

An exceedingly interesting and simple exper-
iment was undertaken by Radakov (1972) with
21 young Pollachius virens (Linnaeus) of 8 to 9
cm. These were placed in a tank measuring
1.6x7x0.3 m. It was divided into two equal com-
partments by a clear glass partition. All the fish
were placed in one compartment. The experiment
consisted of transferring the fishes, one at a time,
to the other compartment. With 20% or less of the
fishes transferred, the smaller group tried contin-
uously to swim through the glass partition in their
efforts to rejoin the others. Above that percent-
age, the two larger groups, between 30 and 40% of
the fish on both sides tried to form a common
school with the glass partition cutting through it.
Continuing the transferring, a reverse series of
the attitudes described above was obtained.

Movements of Individuals

The study of travel by individual fishes within a

school has difficult and tedious aspects, as is

evident from the preceding. The subject has not

attracted many investigators as witness the

paucity of comments on it in earlier papers. An

examination of Figure 12 shows quickly that such

internal traveling is neither negligible nor slight,

at least in very loosely organized schools, but is

probably much less so in very tight schools.

Because of this, the geometrical properties of

schools have been considered chiefly in a single

layer of fishes, i.e., in terms of plane geometry.

Schools of greater depth present special difficulties

in obtaining adequate field data, as it is necessary

to invoke the complications of the third dimension

while the fishes are often so closely packed that

visual perception within the school is severely

restricted. In addition, there are further problems

incident to the fishes' continual activity. This is

particularly difficult in efforts to recognize the

rhombohedron of Figure IB. The present efforts

have yielded some hints that suggest support to

our thesis.

The vertical structure of schools and vertical
mixing within them is much more diflScult to
handle. This is evidently owing partly to the
greater inherent difficulties in three dimensional
plotting and partly in the nature of fish mor-
phology and methods of propulsion. The influences
of each fish on the others in the same horizontal
plane are greater than in any other direction
because both vision and locomotor mechanics
operate primarily in that plane. That is, optical

axes of schooling fishes lie in that plane and the
propulsive mechanism produces forces operating
in it.**

It is consequently less difficult to compare the
relative amount of shifting about in the horizontal
plane as compared with that in the vertical.
Although we have no clear observations or photo-
graphs of a fish sinking to the layer below it or
rising up from one below, there are many in-
stances of evidently "uncertain" fishes seen
between distinct layers or ones dropping slightly
below, as in Hunter's (1966) figure 2.

Shape and Size of Schools

The closed figure that forms the outline of a
school is a remarkably flexible boundary subject to
continual transformation. These changes are
produced by a large variety of influences both
intrinsic and, by a vastly greater number, extrin-
sic. Obviously, the most important intrinsic factor
in holding a school together is the impulse that
causes fishes of one kind to assemble, respecting
each others necessary swimming room and ac-
cepting a common polarization.

The fishes that are outermost along the sides of
a school do not form a special boundary layer any
more than do those at the front form "leaders."
Those at the side surfaces differ from the rest only
in that they lack fellows on one side. Like those at
the front, they are continually changing as their
aggregating tendency apparently moves them
toward a more central position.

Aside from temporary weakening of the bonds
by such things as vigorous feeding, reproduction,
the coming of a suflRciently dark night, or a
particularly violent disturbance, the basic school
structure is continuous in obligate schoolers. In
facultative schoolers, the school is periodic or of
occasional occurrence. True semipermanent inter-
mediates between these two ordinarily distinct
modes are not easy to find and are uncertain at
best.

The intrinsic influences divide naturally into
two groups, the first being those of nonorganic
elements. Common examples of these are light,
water currents, shoreline, sharply mottled bottom
patterns, and obstructions. Sharp discontinuities
of any of these are especially influential. Organic

*A comparison of fish schools with those of cetaceans should be
illuminating because the propulsive efforts of the latter operate
in the vertical plane.

482

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

factors include other schools, large predatory
fishes, fish-catching birds, and rich plankton
streaks.

Theoretically at least, fish schools could take any
shape. Considered as three dimensional "blobs"
they have been described and photographed in a
wide variety of shapes, including even the nearly
spherical (Breder 1959). The latter mostly occurs in
open water some distance from the influence of the
water's surface and the bottom of the body of
water. These are rare and suggest almost exactly
balanced forces. Under such conditions the school
formation in the ordinary sense breaks down? The
form of organization within such near-spheres has
not been analyzed, nor has their manner of for-
mation or eventual dissolution. Other shapes not

^There is in this case a question as to the propriety of including
this assemblage as a school in any sense. At least the fishes that
form this ball are in a solid school formation as they rush in to
form these structures.

readily described in simple geometrical terms, as
that shown in Figure 15, seem to illustrate the
presence of either spiral arms or "smoke ring"
formations.

Much more frequently encountered are schools
close to the water's surface or the bottom. These
often show a more or less oblate spheroidal form
from which a portion has apparently been planed
off, where near contact with surface or bottom
necessarily caused flattening. Otherwise, the op-
posite side follows the contour of the flattened side
so that the school takes the form of a flattened
sheet of rather uniform thickness. These often
take the form of a sheet one-fish deep, the school
practically reducing to a nearly two-dimensional
figure. These all may occur in open water, either
near the surface or bottom. They are, however,
more usually seen in very shallow water where
both surface and bottom influences impinge on the
school. These schools in which the horizontal
dimensions greatly exceed the small vertical one

Figure 15.- An unusual and not readily explicable maneuver of Jenkensia stolifera, seen at Grand Cayman from scuba gear under very

calm conditions.

483

FISHERY BULLETIN: VOL. 74, NO. 3

are more accessible for study and the data ob-
tained from them is readily handled by much
simpler geometrical methods. Most of the present
knowledge of schools is based on observations and
analyses of these sheetlike schools, treated as a
geometrical surface.

Unless there is mention to the contrary, all
statements in this study refer to small or moder-
ate schools. When schools attain huge dimensions,
some of these statements require modification. A
fish in the central part of such a school, that may
have thousands of others between it and open
water in any direction, is locked in a position that
permits practically no freedom of movement. Such
fish are forced to swerve and swim almost as a
single block. Thus the turns discussed in the
section Problems of a School Turning are not
possible. The section Sizes of Fishes in a School
discusses conditions involving the amount of size
variation of the individuals found in a school. This
reaches its maximum in huge schools where size
variations are often large enough to break up a
lesser school.

Problems of a School Turning

A solitary fish obviously can alter its path from
that of a straight line and swim off in any direc-
tion. The presence of objects, such as neutrally
disposed fishes of the same or other species and
same general size, may make little difference
except for appropriate course altering. Problems
loom as a significant influence only when the
density factor becomes relatively large, as in a
loose unpolarized aggregation. When fishes
become even more crowded by each other, the
ability to swim in any direction is severely re-
stricted by the mere presence of the bodies of
other fishes. In a dense school this manner of
restriction becomes intense. Such closely packed
and regimented fish can swim serenely, parallel to
each other, in a straight line or in large swinging
arcs of a radius down to a value of about as little as
five to ten lengths of the fishes involved as shown
in Table 2A. If, however, a sharp curve of shorter
radius is attempted, complications arise (Table
2B). Such turns are commonly made by small
schools up to sizes that are too large to act as a
completely cohesive unit.^° The data shown in
Table 2 refer only to these small cohesive groups.

Table 2.-Data on two types of turns made by fish schools.

A. Radii ot broad curves in Se/ar crumenophthalmus
Fish lengths in cm

fish lengths Max

Mean

Min fishes

5.5 30.5

25.4

22.6 11

9.3 25.9

25.4

22.8 9

B. Measurements of sharp

curves

Angles of

Species

turn in °

Remarks

Menidia berylina

41.5

Selar crumenophthalmus

135.5
148.0
158.0
160.0
163.0

Shown in Figure 17.

175.±

Turns as in Figure 18.

177.0

Trachurus symmetricus

94.0

Note: The 11 numbers

40.5

not set in boldface

33.0

refer to Figure 16

31.5

25.0

'"See Breder (1967) for a discussion of the vastly greater
complexities inherent in the behavior of enormous schools.

Here some disturbance ahead frequently can set
off an activity among the leading fishes in which
they turn sharply left or right. These are then
followed by the others, making their turns in
substantially the same place. Normally the ma-
neuver is accomplished with a scarcely apparent
and transient slowing of pace. The hydrodynamics
of how sharp turns are made by fishes with a
minimum of deceleration was discussed in detail
by Weihs (1972).

Some of the angles between the initial and
subsequent paths of schools making these sharp
turns are given in Table 2B, picked from motion
picture sequences. Figure 16A indicates that
turning at a certain angle could cause following
fishes to approach the tail tips of those just ahead,
an accident that appears never to happen.

There is nothing inherent in the situation of a
school swimming ahead that concerns angles of
turning. The features of the diagram in Figure
16A are meaningless to the fishes until they begin
to turn. Let the school swim in a straight line and
turn 30° to the right at the center of the diagram.
Each fish will come out in an occluded sector and
find it being brushed by the tail of the fish ahead.
If the Weihs (1973a) diamond is elongate along the
axis of travel, the fishes will fall a little short of
contact but will swim into the wrong side of the
vortices shed by the preceding individual. This is
evidently sufficient to initiate avoidance reactions.

If they turn at 60°, there will be no problem as
they will be well separated by the amount in-
dicated in Figure 3A. The fishes in turning
evidently do so only where there is no danger of

484

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES
0° 0'

////

--7

Figure 16. -Angles of sharp-turning fish schools. A. Angles
compared with the rhomboidal lattice. The four solid radial lines
represent the collision paths of turns if the original path is
represented by the vertical line marked 0°. This direction of a
fish's path is indicated by the arrowhead. The dashed radials
marking the end of each arc separate the clear sectors, without
arcs, from the occluded. The 11 short radial line segments
represent the new path of the fishes after they have made their
sharp turn. The numerical values of the angles are given in Table
2B. The same fish turns compared with the cubic lattice. Here the
fish paths are not limited to the clear sectors. See text for full
explanation.

interfering with each other's swimming. Actual
turns of various species keep well away from the
critical angle. Which particular clear space is
selected is evidently determined, at least in part,
by the strength of the deflection-causing stimulus.
As such a turn is completed, the fish again start to
swim in an essentially straight line while they
regain the positions that were somewhat dis-
turbed in turning and the Weihs (1973a) diamond
appears again. Thus the outlined sectors in Figure
16A become "forbidden" paths. Since the diagram
in this figure is purely a geometrical construction
with the occluded and clear sections having equal
areas, this is not to say that some intrusion into
the outlined sectors is impossible. The axis of the
occluded sectors is the worst position for turning
and that of the clear sector the best, the areas
between grading gradually from one condition to
the other. The dotted radii are halfway between
the center lines of the clear and the occluded areas.
The turns made by real fish schools, measured by
motion picture analysis, and shown in Figure 16A
and Table 2A indicate the absence of intrusion
into the enclosed areas.
This examination of the sharp turnings of fish

schools would not have shown these features if
they had been organized on some pattern other
than that of the hexagonal lattice. If they had been
organized on the square lattice, shown in Figure 4,
there would have been at least some in the "for-
bidden" sectors, as is shown in Figure 16B where
the same data on turning angles have been placed
on a diagram based on the square mesh. Here the
same data show less preferential behavior on the
part of the fishes toward the clear sectors. All the
schools, in the hexagonal case, stayed within the
boundaries of the clear sectors (Figure 16A) while
only 64-1-% did in the square case (Figure 16B).
Also the intrusion into the occluded sections
increased with the increasing angle between the
initial course and the new one. These two items are
additional reasons for considering the lattice to be
basically hexagonal.

A typical turn of the sort discussed is shown in
Figure 17 and in Table 2B. This drawing is based
on a series of seven motion picture frames (0.44 s).
The sequences are of a tight school, the angles
between the straight paths, before and after the
turn, are based on the mean paths of the fishes.
Only a few of the individual fishes are shown in
Figure 17 to indicate the nature of the turn at that
point. Not shown are the many fishes constituting
the bulk of the school.

There is also another type of sharp turn that is
not mentioned in the preceding description. It can
lead to considerable confusion because superficial-
ly it is readily confounded with the foregoing type.
It differs primarily in not being concerned with
angular limitations, which apparently can be
ignored only at the expense of making the turn

Figure 17.-A sharp turn of Selar crnmenophthaimus. Only the
paths before and after the turn are indicated and a few of the
turning fishes. The directions of the two paths are indicated by
arrowheads. See text for full explanation.

485

FISHERY BULLETIN: VOL. 74, NO, 3

with considerably less alacrity. Once the
behavioral differences between the two types of
turn are understood, they can be seen in the field if
one happens to be looking directly at the point of
turning before it begins. This is easier to see in a
relatively large school than in a small one because
the larger the number of individuals involved the
more prolonged the turning maneuver becomes.
Also, it is most noticeable where the sudden
appearance of something large and "threatening"
produces an apparent "panic situation." Instead of
what seems to be the beginning of a tight turn, as
previously discussed, the action is most often seen
as an attempt to retreat over their forward path.
Here there develops a "logjam" and confusion. The
immediate response is for the clump of fishes to
spread out into a more or less circular area, out of
which the school is seen to beat a hasty retreat.
Figure 18 shows such a performance which theo-
retically, at least, could move off in any direction
but, so far as our observations go, has usually been
close to the opposite direction of the abandoned
advance. The conventions of Figure 17 have been
used and the same number of frames cover this
sequence. The seeming difference of speed is
simply that badly frightened fish move faster than
relatively placid ones and therefore make up much
of the time lost in the greater length of their
confusion-imposed travel. This area sometimes
develops a central clear spot devoid of fishes, and a
true "fish mill"^^ is transiently developed.

The angular measurements between the track
of a school before the turn and after it can only be
precise in photographs taken wnth the camera
pointed straight down. This is nearly impossible
with feral fishes because such schools simply move
away from anything directly overhead. The pho-
tographs on which Table 2 are based are those
which approach that position as nearly as possible.
This departure from the vertical naturally tends to
slightly blur the accuracy of the angles and thus
serves to produce a greater spread in the apparent
angles. This effect has less influence on the mean
values of each clustered group. To help counter this
source of error, a transparent dial was prepared
with the sectors shown in Figure 16. Hand held, it
can be tipped at an angle appropriate to the
angular amount of departure from the vertical
with which the eye or camera viewed the scene. A

(

"A term used by Parr (1927) to cover the case of a fish school
swimming in a more or less circular path, where every fish is
following those ahead of it. The clear center that these mills
sometimes develop has been discussed by Breder (1951).

Figure 18.-A slower turn of the school shown in Figure 17, with
confusion at the turning point. The directions of the two paths,
into and away from the melee, are indicated by arrowheads.

variety of items shown in such a film helped
establish the needed correction with sufficient
accuracy for present purposes. Although little use
could be made of it in the direct observations
because of the rapidity of the action, it was
invaluable in studying strips of motion picture
film. Small cues that helped establish the proper
angle of tilt of the viewing dial included prin-
cipally the amount of the sides of the fishes that
could be seen plus other objects incidentally in-
cluded in the photographs.

Absolute turns enforced on schools of Mugil
cephalus Linnaeus and Pollachius virens (Lin-
naeus) by the end of an aquarium were studied by
Radakov (1972). These are in contrast to the
preceding studies of turning in open water where
the actual cause of the turn was often obscure, but
irrelevant to the mechanics of turning. The
aquarium studies show nothing like the "sharp
turns" but are close to, if not identical with, the
present "slower turns" where the school breaks
down and reforms on the retreat path. Here and in
Radakov's (1972) work, there is considerable mix-
ing and the place of individual fishes in the school
after these turns may be grossly altered. In the
confines of an aquarium there is practically no
choice of turning angle and the complex situation
in turning in open water does not exist.

General Traffic Problems

Road traffic of automobiles may seem to be very
remote from a school of fishes. Close examination,
however, reveals that the two have common roots
and that, despite their apparent differences, they
are isomorphic. Both cars on a road and fishes in a
school can be treated as embodiments of math-
ematical expressions concerned with mass
movements of redundant units. The mathematics
of the behavior of automobiles developed along
with their proliferation, following the need for

486

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

increasing specific and detailed control of their
movements. Thus, from a purely empirical begin-
ning these studies have gradually developed into
the present traffic theory, most of which has
developed in the last 10 yr. Introductions to its
considerable literature are given by Ashton (1966)
and Gazis (1967).

Some of the similarities and differences
between cars and fishes in the attainment of an
organization of free-flowing traffic is indicated by
the following comparative listing of the two types
of redundant units.

Fish

1. Fishes, schooling or not,
operate freely in three
dimensions, but most
free-swimming fishes,
especially those that
form schools, operate
mostly parallel (al-
though not necessarily
close) to a usually hor-
izontal surface, either
the surface of the water
or the bottom. These
two mark the vertical
limits within which the
fishes must stay. Hor-
izontal boundaries may
vary from too close for
schools to exist to prac-
tically limitless ex-
panse, in the strictly
physical sense. School-
ing fishes can go in any
direction but only with
their school.

2. Other strictures are
those with which only
schooling fishes are
constrained. Here fishes
all swim in a common
direction, mostly in
parallel paths, and in
single files. Collisions
are rare or absent, their
avoidance evidently be-
ing rooted in their
highly developed sen-
sory mechanisms: vi-
sion, lateral line and
cupulae'- senses, and
hearing. There is no
provision for "night
driving" except in
species carrying their
own illumination.
Others loosen their

A u to7nobiles
Cars, in traffic or not,
are confined to a sur-
face, which is not
necessarily a plane and
is often a warped sur-
face, where the extent
of warping eventually
limits the possibility of
use by cars. Cars must
stay on their roads but
do not necessarily stay
with their fellows. They
may strike out alone
wherever there are con-
nections with other
roads, except where ac-
companied by restric-
tive road signs forbid-
ding a given maneuver
or by the general rules
of behavior.

Other strictures are
those with which only
cars, especially in traffic
are constrained. These
controls are maintained
by laws to run in an in-
dicated direction in sin-
gle files or in parallel
paths, depending on the
width of the road and
its indicated number of
lanes. Collisions occur
with monotonous fre-
quency. The protections
are only the sense or-
gans of vision and
hearing. Night driving
illumination is normally
present.

ranks or break up on
nights sufficiently dark
to eliminate vision.
3. Fishes form well-
defined patterns; for
hydrodynamic reasons
they are quadrilaterals.

Cars form "diamonds,"
or if the road has less
than three one-way
lanes, parts thereof.

In both cases there are valid reasons for not
following closely behind the unit directly ahead
and for not traveling in tandem positions. The
resulting staggered deployment permits passing
and lane shifting with a minimum of confusion. It
is this arrangement of units and their possible
movements that is largely responsible for the
irregularities in any instantaneous structure of
the swimming patterns.

In the case of a traffic jam of cars or the
equivalent conditions of pods^^ of fishes, the
pattern formed by units is nearly obliterated.

The shape of the diamond formed by four cars is
related to the speed of travel and is determined by
the rules of the road covering the increase in
distance to be given the car ahead with an increase
in speed. Also the rules require the passing car to
speed as fast as practicable in passing the slower
car. Thus, the faster the traffic, the farther the
hexagon or diamond departs from the regular,
attenuating along the axis of travel.

That the fish and a car with its human driver are
closely comparable should be clear from the
preceding and the following outline indicating
that the relations between the two dynamic sys-
tems do in fact constitute an isomorphism. Two
central nervous systems, one of a fish whose body
is vehicle, power plant, and pilot, and the other,
that of a human who is the pilot, enveloped in a
capsule comprising the vehicle and power plant,
operationally calls for the same kinematic pattern
and trajectories of behavior. As these are both
systems with feedback in which all essential
variables are evident, the canonical representation
and the ordinary algebraic forms of equations can
be calculated. This will not be done here as it would

In fishes, the same results are obtained by those
ahead leaving both advantageous and disadvan-
tageous water movements in which the followers,
by taking the path of least resistance, fall au-
tomatically into positions that mark out the
diamond. The lengthening of the figures as the
fishes' speed increases is very slight as compared
with that of the distance increase with cars.

^See Cahn (1967) for a survey of the function of these systems.

'^This term has been defined by Breder (1959).

487

FISHERY BULLETIN: VOL. 74, NO. 3

not be relevant to present purposes and would
carry away from the intent of this communication,
although the equation of Breder (1954) should be
useful to such a study.

The main thrusts of the students of traffic flow
have been concerned with such things as problems
of delays, queueing, road junction, traffic signals,
analogies to fluid movements, and follow-the-
leader sequences.

The study of fish schools has not yet reached into
these matters, although they all bear a one-to-one
resemblance to similar items in schools. This
undeveloped area is difl^icult to enter into deeply
partly because there is no facile way to keep track
of each individual. The analysis of the behavior of
individuals in a school based on the data of Hunter
(1966) could be considered as a start in this
direction.

Influence of Body Forms

There is a marked positive relationship between
schooling and the extent of streamlining of the
general contours and of the drag-reducing surface
details of fishes that, in the most advanced ob-
ligates, can be considered exquisite. Parallel to
this is an equally marked negative relationship
between schooling and special surface features of
the eruptive sort. At this end of the series, the
fishes are not schoolers at all, nor even aggrega-
tors, but are usually solitary, neutral, or agonistic
toward their fellows. All of this can be shown to be
related to mechanistic details covering the manner
of life of the individuals involved.

For example, we know of no obligate schoolers
such as clupeids or scombroids that have any
drag-producing extensions, while the vast majori-
ty show beautiful fairing even in the manner that
the maxillary fits into a matching recess when the
mouth is closed and in the slot that the depressed
dorsal fin fits into as shown in Scomberomorus.
Such niceties are not to be found in the facultative
schoolers such as most of the Salmonidae, Cyprin-
idae, and Serranidae. In the essentially non-
schooling fishes, the streamlining often becomes
less effective and outgrowths from the integu-
ment and eruptive structures become more and
more extreme as in Hippocampus, the Scorpaen-
idae, Cyclopteridae, and Diodontidae. With this
comes slower swimming speeds and an increasing
tendency to reduce swimming to a minor roll, as in
some of the Scorpaenidae and all of the
Antennariidae.

The remainder of fishes to be considered here
are those that show a depth^^ equal to or greater
than their lengths. These are often facultative
schoolers. Families in which this is a usual or
frequent condition include the Stromateidae,
Ephippidae (including the extreme platacids),
Chaetodontidae, and Acanthuridae. Many others
show an approach to the condition, as in the
Pomacentridae. In addition to these, there are a
considerable number of families in which one or a
few species have the necessary characteristics, as
the Carangidae and Cichlidae.

The schools that are formed by fishes of great
body depth are superficially very similar to those
formed by fishes with fusiform outlines. A school
of deep-bodied fishes is, however, automatically
tighter because the greater depth of body intrudes
into the swimming areas of the layer of fishes
above as well as the layer below.

There is both mechanical and hydrodynamic
interference and an optical occlusion that is much
more severe because of the greater area of the
sides of these fishes. This leads to greater difficulty
in making sharp turns. These conditions can only
be relieved by loosening the school in the vertical
direction. How much mutual swimming facilita-
tion is lost by this loosening is not known. Figure
19 illustrates these conditions with a head-on
photograph of an extremely loose school of Chae-
todipterusfaber (Broussonet).

The only other fishes known to form schools are
those in which the longitudinal axes do not lie
parallel to their line of travel. They include various
characins, the "head standers" of aquarists, and
some aulostomoids, the best known of which are
Aeoliscus and Macrorhamphosus. These evidently
swim with the head up or down (Atz 1962,
Klausewitz 1963). There is no data on any aspect of
their hydrodynamics nor on their mucus. These
forms, therefore, are not discussed here.

Sizes of Fishes in a School.

The variation in the lengths of individuals in a
school usually reaches no more than 30%. The
difference between the length of the largest fish
minus that of the smallest fish in a given school is
expressed as a percentage in this notation. Data
from Breder (1954), recalculated for present pur-

'^This is not the body depth of taxonomists, but the vertical
depth of the entire profile, including the extent of the dorsal and
anal fin in that dimension.

488

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Figure 19.-A head on view of a loose school of Ckaetodipterus.
faber. From Herald (1961). Photo by Fritz Goro.

poses, yielded the following comparative values:
Harengula humeralis (Cuvier) 12.5, Jenkinsia sp.
24.2, and Atherinomorus stipes (Miiller and Tros-
chel) 25.0. Additional data on Jenkinsia stolifera
(Jordan and Gilbert) taken from Breder and Bird
(1975), based on Grand Cayman fishes, gave 31.7.
All are below the 30% level of variation except the
last. A school of let ahirus nebulosus (LeSueur) still
being herded about by their parents, however, had
42.9. It is known that when several family groups
are present, the young fish often become mixed.
This may well be the cause of this greater varia-
tion, a similar feature being found in extra large
schools of adult clupeids, as discussed by Breder
(1967).

The fate of injured and parasitized schooling
fish has not been given much attention and it has
generally been assumed that such unfortunates do
not long survive. This view has been nurtured by
the fact that a captured school of fish most often
contains no individuals that show either wounds or
evidence of gross parasitism. That there are
striking exceptions to this has been shown by

Guthrie and Kroger (1974). They reported that
individuals of both Brevoortia tyrannus (Latrobe)
and B. patronus Goode, with vitality reduced
because of depletion caused by injury or para-
sitism, are to be found in estuaries schooling with
smaller, younger, but healthy individuals normally
present in these relatively protected areas. Out-
side waters yielded no such composed schools.

The relative sizes of the healthy young fishes
and the handicapped older ones and the ratios
between the largest and smallest individuals are
given below as percentages.

Young

Estuarine
B. tyrannus
B. patronus

Oceanic
Both spp.

52.7
73.6

Old

10.4
44.2

63.4

Only one group has an index of low variation in
lengths, 10.4. The others all have indices of high
variation reaching to the extreme of 73.6. If the
schools of both young and old are each taken as a
whole then all groups would show very high varia-
tion, i.e., 63.6 for B. tyrannus and 80.0 for B.
patronus.

There is only one way these figures can be
interpreted. The schools of both species are a
mixed lot of lesser schools, as would be expected of
fishes that persist in forming enormous schools
that mix broods from different spawning areas
and that are hatched at various times in waters of
different temperatures. This genus would seem to
be the most prodigious gatherer of huge ag-
gregates of a single species on the American
Atlantic coast.

In the usual, more uniform schools, where the
variation is less than about 30%, the geometric
structure is observably more uniform. Theoret-
ically, at least, the smaller the variation in the size
of the fishes the nearer the lattice could approach
geometrical perfection. Schools of fishes where
there is larger variation in size tend to break in
direct proportion to the magnitude of the varia-
tion. In enormous schools with great size varia-
tions breaking up is not always possible but does
lead to considerable churning as individuals of
similar sizes gravitate together.

Effects of Mirrors

The confronting of animals with mirrors has
been practiced for many years, for both trivial and
serious purposes. The vast majority of such pre-
sentations has been made to one subject at a time.

489

FISHERY BULLETIN: VOL. 74, NO. 3

e.g., Svendsen and Armitage (1973). There have
been few cases of mirrors being introduced to
groups, such as fish schools. Pitcher (1973), using
mirrors for certain photographic purposes, noted
some of the reactions of his fish subjects. In both
cases above and almost all others, the studies have
been made on captive animals.

Information on the reactions of individual fishes
vv^ithin a tightly organized school is not readily
obtained. Experiments on captive schools yield
results that are naturally suspect, primarily
because of the usually gross changes in the
behavior of schools confined to small quarters and
the length of time in days or weeks, even in a
relatively enormous container, that it takes to
reach apparent stability. Analysis of motion pic-
tures taken of feral schools cannot be expected to
supply much more than occasionally fortunate
sequences. One difficulty is the interference of
other members of the school or of other species
exterior to it. Mirrors introduce something to
which fishes generally respond and thus the pos-
sibility of reasonably interpreting their responses
exists. The experiments and their results follow.

A submerged mirror, 39 x 57 cm, was hung near
an observation dock or other suitable location
which yielded data on fishes in schools in their
native habitats. The school's presence was in no
way forced, nor were they present because of any
attractiveness mirrors may have, since the sites
selected were normally visited daily by these
schools.

Four species, three of the Clupeoidei and one of
the Mugiloidei, reacted to this mirror, each in a
different manner, as follows.

Anchoa hepsetus (Linnaeus) showed the most
striking reactions. All the schools of this species
were large, at least containing 1,000 fishes and
usually far above that number. The schools ap-
peared at this place only during the daylight hours
and moved off to deeper water for the night. These
movements were independent of the tidal stages.
The horizontal component of the tidal flow clearly
regimented these fishes because at slack tide they
became somewhat disorganized.

If the mirror was submerged while this species
was absent, the fishes schooled on arrival would
regard it simply as any other solid object, such as a
pile, that had to be avoided by changing their
course. In doing this, schooling fishes normally
leave a clear space between them and the object. In
this case it averaged close to 20 fish lengths. If the
fishes were present before the mirror could be

lowered, it was allowed to slide directly into the
school, which produced little disturbance, other
than a few transient "shock waves" as the normal
space was formed around the mirror. It was
noticed early that the distance kept by the fishes
from the back of the mirror, painted black, was a
little greater than that kept from the face of the
mirror.

After the elapse of about 1 h after the introduc-
tion of the mirror, the portion of the school oppo-
site the mirror's face made slight "bulges" toward
it, which were promptly resorbed. Nothing like
this appeared on the part of the school opposite the
black backing of the mirror.

After another hour, the school had moved closer
to the face of the mirror, approximately 10 fish
lengths away. When this was once established,
individual fishes would sally forth from the pe-
rimeter of the school opposite the mirror and swim
to within four fish lengths of the mirror and
momentarily run parallel with their reflection.
This would be followed by a hasty retreat to the
school. The action, repeated frequently by various
individuals, would seem to be explicable as follows.
A peripheral member of the school could see the
school's reflection twice as far as the mirror
surface. To join that "other" school required that
the adventuresome individual had to negotiate
that apparent distance. The fish traveled about
nine fish lengths before it turned back. Here the
fish found that one fish in the reflection is coming
at him and running side-by-side with him, at an
apparent distance of two fish lengths. This kind of
behavior is not the "normal" in the situation of a
few or one fish attempting to join a much larger
group, at least in any of the species under study.
The usual manner in which one or a few fishes join
a large school is to quietly approach the larger
body and pick up its rate of speed and slowly
merge into the main body. There is never any
evident specific act on the part of the affected
fishes of the school. They seem to react to the
"intruders" as they do to the other members of the
school, constantly adjusting their positions by
small amounts.

The above is not true of two schools of more
nearly equal size when in the process of merging.
The smaller will approach the larger at a rate of
speed apparently inversely proportional to the
volume of the smaller school. The larger school will
approach the smaller at a much slower speed also
inversely proportional to its volume. When the two
schools come within a distance equivalent to about

490

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

four fish lengths, both schools show a bulge on the
side closest to the other school and in so doing
automatically loosen their ranks slightly, but
sufficiently to allow the two bulges to merge,
forming a single school where there had been two.
This type of merging can usually be found be-
tween schools that do not differ in size by a factor
as large as four.

The above describes what are evidently the
normal sequences to expect when two schools of
various size relationships have an encounter that
may lead to merging. This leads to the idea that
the "behavior" of a mirror image is sufficiently
unusual to prevent the further development of a
process leading to merging, the fishes evidently
recognizing a difference between another fish and
their own reflection.

Sardinella anchovia Valenciennes and Brevoor-
tia patronus Goode avoided coming close enough
to the mirror for the development of any further
reaction. The first was present frequently in large
schools which tended to stay away from the dock
area in deeper water, but frequently came into the
shallower areas at which time they revealed no
indication of "nervousness." The second was seen
only as young fish in very small schools of not more
than 30 fast-moving individuals, that gave any
solid structure a wide berth, which is characteristic
of this species, at this place at least. Brevoortia in a
10-foot circular concrete tank formed a school of
about 30 individuals that averaged about 10 cm in
length. They had lived there for about 10 mo.
These fish were exposed to the mirror for 1 day in
August and 4 continuous days in November. Prior
to the introduction of the mirror, the school circled
the tank close to its wall. The introduction of the
mirror disrupted this path of the school which then
formed a tight mill as far away from the mirror as
possible. At no time were the fish observed to
approach the mirror. Only dropping food close to
its reflective surface would cause individuals to
move toward the mirror, and then only to snap at
the food and retreat rapidly. The fish fed less
during the presence of the mirror. After the
mirror was removed 6 days passed before the mill
broke up and the former swimming pattern was
resumed. Harengula pensacolae Goode and Bean,
not seen around the dock when the mirror was
used, behaved not unlike the Brevoortia in the
concrete pool.

Mugil curema Valenciennes, in its very young
surface swarming stage of not over 2 cm, forms
very loose schools not at all like those of the adults.

These young, on encountering the mirror, would
try persistently to swim into the mirror, seem-
ingly disregarding their mirror image that just as
persistently "opposed" them. Occasionally when
such a group left the mirror for reasons unknown,
a single fish would remain and continue to try to
swim through the mirror for a long period,
evidently almost to exhaustion.

These observations were carried on from 8 June
to 10 September 1973, weather permitting, and
represent many repetitions of the facts and inter-
pretations. It is impossible to present these notes
in a more formal manner at this time. They clearly
have bearing on the present study and suggest the
desirability of going into this matter further as
another project which would in any case lead away
from present purposes.

The observations indicate that there is a much
wider range of difference in response to the
mirror image than had been expected and there-
fore that the bonds that hold a school together are
not identical for each species, even if the total
result appears as a very similar geometric struc-
ture. It would seem that the response of a fish to a
fellow (here its mirror image) that approaches on a
true and unswerving collision course from which it
will not (cannot) budge is a truly frightening
experience. The difference in response between
Anchoa and Mugil in this case is especially strik-
ing. Anchoa acts in a manner that one might
anticipate, while the action of Mugil in placing
their mouths together has never been seen at any
age or size.

LOCOMOTOR PROBLEMS

With large numbers of fishes of one kind swim-
ming closely together in a common direction, the
locomotor needs of the participants would ob-
viously have influence on the structural nature of
the school, which in turn would also affect some
details of the locomotor efforts. Both classical and
contemporary hydrodynamics have to be invoked
in any attempt to understand this mechanical
aspect of school formation and operation.

Flow Patterns

To answer the question of whether water flow
induced by the propulsive activity of the fishes
themselves can help or hinder other fishes follow-
ing them depends on the direction and strength of
the flow and the angle of entry of a fish encoun-

491

FISHRRY BULLETIN: VOL. 74, NO. 3

toring the flow. The solution of such problems lie in
the realm of classical hydrodynamics. See Lind-
gren (1967) for a brief, but explicit statement of
the hydrodynamics involved. Fishes leave no wake
in the usual sense of the word, but do leave a series
of dying vortices, alternately on either side of the
swimming axis of their producer. The rotational
direction of the flow within the vortices on one side
is always the same and is opposite to the rotation
of those on the other side. The flow within the
vortices is such that, on the side nearest the axis of
the fish producing them, the flow is opposite to the
direction of travel of that fish, while on the side
away from the axis the flow is in the same direc-
tion of travel as the fish. These rotational direc-
tions are opposite to those of vortices formed in a
typical Karman trail produced by a rigid solid. A
following fish thus has the choice of swimming
through the side that would help it on its way or
the other that would retard it. Swimming through
a vortex center would push the head of the fish to
one side before the center was reached and to the
other side after the center had been passed. The
fish that follows is normally found in the water
flow that is in its direction of swimming, see Rosen
(19r)9), Hreder (1%.")), and Weihs (1973a). This
arrangement evidently helps the locomotor efforts
of all but the lead fishes. As the energy in the
vortices dissipates rapidly it is doubtful if more
than the immediately following fishes benefit
significantly. As each fish produces a similar
short-lived set of vortices there is no appreciable
additive effect of successive rows of fish ahead.
Thus all the fishes after the first transverse row
receive approximately the same energy input
from the vortices, so long as they remain in the
specified positions. The value of this has not been
measured as yet or even estimated.

These friction reducing effects evidently in-
fluence small fishes to sometimes closely associate
with much larger, usually solitary, fishes of other
affinities. The small attendant fishes evidently
gain locomotor advantages that are otherwise only
obtained by schooling with their own kind. Many
authors, including Breder (1959, 1965, 1967) and
Aleev (1968), have noted a variety of such fishes.
These fishes station themselves close to and in
definite positions relative to the larger fish, often a
shark. The behavior is habitual, as in Seriola, but
may be occasional, as in Caranx. Shuleikin (1958)
discussed the hydrodynamics of Naucratcs ductor
(Linnaeus) in its persistent association with large
sharks.

Weihs (1973a) indicated additional energy sav-
ing advantages consequent on fish swimming his
diamond pattern; the channeling effect of rows of
similar fishes, the effects of the phase of the
tail-wagging of one fish with respect to the tail
phases of its near neighbors, and the extent of
length variations in the participating fishes. He
calculated this variation as up to 50%. Actually
over 60% variation has been found in unquestion-
able schools (Breder 1954), although it is impossi-
ble from this data to determine the permanency of
such groups or the efficiency loss at this greater
range of variation.

Active fishes, especially schooling types, lack the
protuberances and hollows often present on the
bodies of sluggish fishes. Aleev (1963) enumerated
many instances of the latter. He indicated that
this lack of streamline integrity leads to the
production of minor vortices and that these dis-
turbances, depending on their size and point of
origin, could lower the locomotor efficiency of a
fish. The utility of the larger terminal vortices,
here under discussion, could be reduced or de-
stroyed, thus eliminating one of the advantages of
.school formation.

Turbulent Friction Reduction

Until recently, students of fish locomotion were
not in agreement concerning what function in
relation to swimming, if any, was served by the
presence of the mucus that covers the bodies of
living fishes. Aleev (1963), in a well-documented
review, indicated that he agreed with Richardson
(1936) and Gero (1952) that whatever part it may
play, the effect must be very small. That this could
not be so was mentioned by Rosen (1959) and
Walters and Liu (1967). Recent advances in hy-
drodynamics now indicate clearly that it has a very
considerable role.

Polysaccharides are known to be released by a
variety of aquatic organisms, both plant and
animal. One of the effects of the presence of those
forming long-chain molecules is friction reduction
in turbulent water flow. Some of the history of the
development of this information was recorded by
Newton (1960), Barnaby and Dorey (1965), and
Hoyt (1966, 1968, 1972, 1975). These papers dis-
cussed naturally occurring polysaccharides from
algae as well as synthetic high polymers, some of
the latter being used for very practical purposes as
very efficient reducers of turbulent friction. The
application of extremely small amounts of such
materials can reduce drag by over 60%.

492

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Rosen and Corn ford (1970, 1971) had shown by
means of a special type of rheometer that there are
great diflferences in the friction reducing abilities
of the slime of various species of fishes. See
Jakowska (1968) for a discussion on the extent of
the wide variety of other kinds of utility ascribed
to the mucus of various fishes. It would seem to be
certain that these effects are dependent on the
polysaccharides inherent in fish mucus, although
for present purposes it is not necessary to know
just what components of fish mucus account for
friction reduction.

Successive dilutions of fish slime with the water
of the individual's habitat have been plotted
against reduction of friction in terms of percent
by Rosen and Cornford (1970, 1971). In some cases
the curve rises extremely rapidly, reaching a
reduction of turbulent friction of over 60% with
water dilution to only 5% mucus. Others, with
evidently less potent slime, show a much smaller
rise in friction reduction, reaching 50% or less with
a water dilution to 50% or more of slime. The most
extreme case reaches only 8% reduction in friction
with full strength slime.

It is notable that the two species with the fastest
rise in friction reduction are rapacious and strike
at relatively large prey. These fish can move from
a resting position to their highest speed in a
remarkably short time. The three species at the
other end of the friction reduction series feed on
much smaller organisms in proportion to their own
size, for which violent pursuit is completely un-
necessary. The two species with the most efficient
drag reduction do not form obligate schools and
are often solitary, while the three with the least
effective mucus are schoolers and only one drops to
the facultative status.

The preceding data on the reduction of tur-
bulent friction by means of long-chain polymers,
and the demonstration of the great effectiveness
of the mucus exuded by some fishes, as well as the
geometrical {)atterns in which schooling fishes
arrange themselves, leaves little room for doubt
that the fishes so organized may attain a locomotor
advantage from the mucus trail trapped in the
vortices left by the fishes that preceded them.

The fishes with sharp rise in friction reduction in
Table 8 and P'igure 20 are all nonschoolers or at
most facultative: Paraiichthyi^ californicus
(Ayers), Sphyraena argentea Girard,''' and
Micropterus dolomieui Lacep6de. Those with a
slow rise in friction reduction are all schoolers and
are primarily obligate'" schoolers: Scomber japon-
icus Houttuyn, Sarda chiliensis (Cuvier), with
Saimo frutta Linnaeus and S. gairdneri Richard-
son as facultatives. The nonschoolers are capable
of showing a sudden acceleration from a resting
position and apparently attain their highest pos-
sible speed in a matter of seconds or less. The
hydrodynamic aspects of extreme acceleration
from a position of rest, shown by slender fishes
such as barracuda, are treated by Weihs (1973b).
This can be critical in overtaking relatively large
prey. Schooling fishes that normally swim at a
continued steady pace evidently cannot perform in
such a manner and even the marginal members
seldom try.

Uskova and Chaikovskaya (1975) noted, in a
paper on the chemical nature of the protein com-

'•''It is recognized that the Pacific Sphyraena argentea tends to
form schools more readily than the larjjer Atlantic .S. barracuda
which is usually solitary. The smaller Atlantic conf?cners ap-
proach S. argcnifa in this respect.

"■A term defined by Bredcr (1967).

Tabi.K 3.-DraK reduction by fish mucus, based on data from Rosen and

(Cornford (1970, 1971).

Species

1 Salmo gairdneri Richardson

(Rush Creek)

2 S. gairdneri (Grant Lake)

3 S. gairdneri (Lundy Lake)

4 S. trutta Linneaus

5 Sphyraena argentea Girard

6 Scomber japonicus Houttuyn

7 Sarda chiliensis (Culvier)

8 Micropterus dolomieui Lac6p6de

9 Pomoxis annularis Rafinesque

10 Lepomis machrochirus Rafinesque

11 Paralabrax clathratus (Girard)

12 P. nebuliter (Girard)

13 Parallchthys californicus (Ayres)

Drag

Mucus

reduction

concentration

Length

(%)

(cm)

61.8

62.0

20.5

63.2

65.9

76-79

56.9

38-41

6.4

100

62.0

33-42

61.7

—

60.1

15.3-20.4

58.7

17.4

33±

60.9

493

FISHERY BULLETIN: VOL. 74, NO. 3

.100

-K>6

—

P 80

—

" 60

—

1 n6

5<5l "

o
o

5 A

° \

oe^i^e^"

O 40

—

\ ^ (^

—

\ 4 0 Q

1 r'°^ 1 1

0 20 40 60 80

% DRAG REDUCTION

Figure 20.-Graph of the effectiveness of fish mucus on drag
reduction. Based on the data of Rosen and Cornford (1970, 1971).
The numbers and letters at each point are explained in Table 3,
giving the name and number of each fish in the left hand column.
See text for full explanation.

ponents of fish mucus, that the hydrodynamic
efficiency of the fishes they studied varied directly
with the extent of the basicity of their surface
mucus. The fishes measured were Atlantic bonito,
Sarda sarda (Bloch), sea bass, Serranus scriba
(Linnaeus), and stargazer, Uranoscopus scaber
(Linnaeus), given here in the order of descending
basicity. This is consistent with the present
studies based on the lubricity of certain polymers.
The mucus of a fish in a school does more than
reduce the drag on its producer since it washes
over those that follow. This means that the
"leaders" have only their own mucus to ease their
passages while the "laggards" receive all the
benefits bestowed by those ahead of them. The net
effect is to produce a lubricity gradient from zero
to the maximum which is dependent on the size of
the school. To maintain a steady pace, fishes in the
forepart of the school must use more muscular
power than the others while the last members
require the least effort to hold their positions. As
fatigue sets in, the "front runners" would have a
choice of accelerating their efforts or holding a
steady pace and thus permit those following to
pass ahead of them until they find a place requir-
ing an effort compatible with the magnitude of
their tiring, which could carry them to the trailing

end positions of the school, if necessary. Zuev and
Belyaev (1970) indicated that in a school of Tra-
churus, the individuals in the front part beat their
tails faster than those in the rear. This condition
would naturally follow the lattice-vortex-mucus
thesis as developed here.

Thus, this condition of graded positions in
respect to ease of swimming and the matter of
muscular fatigue may be a large factor in the
maintenance of the integrity of a school and
explain the internal churning so often seen in fish
schools. The very general changes in positions of
individuals within the structure of a school could
thus be impelled to a large extent by the individual
urge to attain a position demanding the least
swimming effort. Also this urge would insure the
usual prompt reassembly of a school after being
violently dispersed and suggests that the closed
figure "mills" of schooling fishes, that would
otheru'ise seem to be trivial and pointless, form a
relatively quiescent rest period in a favored place.
Fish mills have been noted by many students,
beginning with Parr (1927). They can be developed
from many other sources than the one noted above.
Often they are derived directly from extrinsic
events, as discussed by Breder (1965). The
development of an evidently intrinsic mill is
shown there by three photographs that may truly
represent the formation of a true "resting mill" as
suggested above.

There is too little known about the complexities
of fish mucus to permit much further progress into
the details of its relation to school formation and
maintenance or its importance to other matters.
For instance, how constant are its characteristics
and are there rhythmic variations in them related
to season, reproductive periods, or type of food
ingested? Are there changes in the mucus with age
or condition of the fish? Is the mucus of marine
fishes more stable than that of freshwater fishes?
Since ocean water is chemically more uniform
than fresh water it might be expected that these
features were reflected in the mucus.

Experiments with Drag-Reducing Polymers

Fish mucus, in the amounts necessary for these
experiments, is difficult, if not impossible, to
obtain and handle without some decomposition
and reduction of the long-chain molecules. Addi-
tives of some bacteriostatic chemical or refrigera-
tion merely introduces other difficulties that could
make interpretations uncertain.

494

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Furthermore, the drag reduction of a fish slime
diluted with water that produced a 25% reduction
just after its removal from the fish, was inert 3 h
later, according to Hoyt (1975). He also gave a
hydrodynamic explanation on why it is possible for
very small fishes to gain an advantage from their
mucus although the operational mechanics are
diff'erent than those available to larger fishes. This
concerns differences in the boundary-layer transi-
tion from laminar to turbulent flow in relation to
the Reynolds numbers. Fish mucus does not dilute
easily with water by mere contact, but does so
easily with agitation. Rosen (1959) used the term
"reluctance" to designate this condition.
Polymers, especially those manufactured to have
high drag-reducing characteristics as measured on
a rheometer, have drag reduction features that are
comparable to or exceed those of fishes' surface
mucus in the small quantities required to obtain
maximum effects.

The material used was a water soluble resin, a
high polymer of ethylene oxide, from the Union
Carbide Corporation, and generally known by its
trade name Polyox^'^. The significant characteris-
tics, as given by Hoyt (1971) follow

Molecular
weight

6,000,000 (ca.)

Polyox F.R.A. (Lot 1163)

,, J Concentration (%)
Max drag L__

reduction (%)

67.8

Max D.R.
15

V& max D.R.
1

This particular grade of Polyox was used because
of its unusually high molecular weight as the
purpose here was merely to establish whether such
products would induce a change in the swimming
efficiency of the fishes. Hoyt (1975) considered a
minimum molecular weight of 50,000 of the drag-
reducing element to be necessary for friction
reduction to be expected.

Polyox is reported to have very low, if any,
toxicity, (Smyth, et al. 1970, Wade 1970). For the
purposes of this study, toxicity tests were also run
on a variety of fishes. Nothing whatever occurred
that would suggest any physiological disturbance
on any of the test fishes. Both Poecilia reticulata
Peters (fresh water) and Hippocampus erectus
Perry (salt water) produced young when subjected
to concentrations far higher than any required
here. The only item showing obvious adjustments

to the change in lubricity of the water was that
mature examples of Hippocampiis erectus were
unable to use their prehensile tails effectively on
the smaller supports provided in their aquaria.
That is, they simply slipped off plastic rods, of
circular cross section, if the rod diameters were
below a certain magnitude relative to the grasp of
their tails. With larger rods they had no trouble
and were readily able to "grasp" the supports and
hold on in normal fashion. Those that could not
find a suitably sized "perch" coiled their tails so
that about three-quarters of a circle was formed at
right angles to the body axis and then "sat" with
the partial circle laid on the bottom of their
aquarium. Apart from being somewhat restless,
they apparently were just as well off as the others.
The Poecilia moved about in what appeared to be
their normal random manner, but whether they
moved a little faster or not could have only been
determined with great difficulty and would not
have contributed to the problems under study.
None of the fishes tested after the preceding
preliminaries showed any distress from the addi-
tion of Polyox.

The Gulf menhaden, Brevoortia patronus
Goode, was used for tests on drag reduction. This
species is an obligate schooler and, as with many
such schoolers, the ability to spread its caudal fin is
_ severely limited. There is a strong possibility that
none of them exercised this slight ability at all.
Also, these fish accommodate well to aquarium life
if provided adequate swimming room and a few
companion fishes, a feature not common in many
members of this family. The fishes selected for
testing were first established in a circular concrete
tank 4+ m diameter, with a water depth of 1 m.

Specially made aquaria were used for these
experiments. They measured 25 x 25 x 90 cm and
were filled with synthetic seawater^^ to a depth of
20 cm providing a total water volume of 45,000 cm^.
These were established in a perfectly light-tight
room, actually a Navy Sea Van without windows,
remote from vibrations and sounds. Lights were
controlled by a time switch for day and night
effects and a thermostat controlled the tempera-
ture. The test aquarium was placed on the floor
and the others on rocks at a convenient height.
Precautions were taken to protect the fishes from
being startled by motions, vibrations, or other

*'Klndly supplied gratis. Reference to trade names does not
imply endorsement by the National Marine Fisheries Service,
NOAA.

'^'Kindly supplied gratis under the name "Instant Ocean" by
Aquarium Systems, Inc.

495

FISHERY BULLETIN: VOL. 74, NO. 3

disturbances outside their container. Tliere is no
reason to suppose that the results were so
influenced.

Two grams of the dry granular Polyox were
dissolved in a small portion of the synthetic
seawater. This was then returned to the test
aquarium by allowing it to drip back by means of a
siphon tube nearly closed by a screw clamp. The
final concentration of Polyox in the aquarium
became approximately 42 ppm.

A motion picture camera facing down was
erected so that its optical axis was over the
geometric center of the tank. Photo floodlights
were set up as required. The view included most of
the aquarium, omitting only the ends of the tank
where the fishes were forced to turn back, as these
tests must be made with the fishes moving in a
nearly straight line. Also included in the camera's
coverage were tapes marked in centimeters. One
ran along the top edge of the tank and the other
along its bottom, thus providing an index to the
lengths of the fishes and their distances of travel.
The aquarium had its sides blocked with bluish
cardboards, except on the sides toward the lights.
These were higher than the aquarium and off to
one side sufficiently to eliminate reflections into
the camera's lens. The test fish were added and
allowed to adjust to the new situation for about 1
h. The tank in which they had lived for at least 1
wk was identical with the test tank, except that it
had all four sides covered with similar cardboard
guards.

Photographs were taken after the lights had
been turned on gradually to full voltage. It was
found by experience that normal film speed was
fully adequate for our analysis. Sufficient footage
was exposed to insure an adequate number of
straight runs of single fish.

When the above procedures were completed, the

Polyox was allowed to drip into the tank, which
took about 10 min. After 1 h had elapsed, its
mixing was considered completed, for in addition
to the aerating devices, the four very active fishes
provided continuous mixing. After this time in-
terval the photographic procedures were repeated
and the experiment was terminated.

The results of these experiments are given in
Table 4 and their analysis is illustrated by graphs
in Figure 21. Graphs A and C clearly show the
difference between fishes swimming in synthetic
seawater, initially devoid of any long-chain
polymers, and in the same water to which the
polymer has been added. The speed of the fishes is
approximately double in the latter, as are the tail
beats. In this experiment, after the first run (Si)
was made in synthetic seawater, the tank with its
contained fishes was left as it was until 2 days later
when another run (S2) was made. The new speed
readings were a little higher, but the proportional
corrections were not. If more refined measure-
ments show that a small difference is measurable,
it should be due to the additions of organic sub-
stance in the interim, consisting of the body
wastes of the fish as well as their own surface slime
produced in this period. Added to this must be the
dissolved matter from the food given to the fishes.
To minimize all this, all particles not consumed di-
rectly were meticulously removed. The manner of
handling data was that of Bainbridge (1958). The
greater refinements of the methods of Hunter and
Zweifel (1971) were not deemed necessary for the
present simple purposes. Because of the large
differences between the speeds of fishes in the
same water, with and without long-chain
polymers, the slight possible spreading of the
caudal fin in this species could not increase the
area of the tail by more than a negligible amount
in these experiments. Later another set of four

Table 4.-Calculations based on experiments on drag reduction in Brevoortia by polymers. TL = total

length. TB = tail beats.

Date

Water

Fish TL

Run length

Run time

No. of

Speed

cm/s

1973

state

(cm)

(s)

tail beats

(cm/s)

TB/s

13 Sept.

Synthetic

6.30

19.50

1.33

14.58

4.51

2.31

6.60

25.00

1.67

14.90

4.19

2.26

7.10

23.30

1.95

11.99

3.08

1.69

11 Oct.

9.00

23.00

1.44

15.10

2.78

1.68

7.50

15.00

0.89

16.85

4.49

2.25

+ Polyox

7.90

29.00

0.89

32.58

7.87

4.12

7.50

35.00

1.11

31.53

7.21

4.20

8.00

32.00

1.11

28.83

5.41

4.80

7.00

34.00

1.11

30.63

8.11

4.37

6 Nov.

Bay

6.60

10.00

0.78

12.82

3.85

1.94

10.00

36.00

2.00

18.00

2.00

1.80

+ Polyox

6.60

47.00

1.11

42.34

4.50

6.42

10.00

44.00

1.06

41.51

4.55

4.15

496

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

(fi

5 20

o P

/ ■■
/

S2 o

I I '

o p

I I i

—

■:° p

o ■ .

o ■
o

• s

1 II
1 1 . > 1

1 '

B °

02468 02468

BEATS/SEC BEATS/SEC

Figure 21. -Graph of experimental analysis of the relationship
of tail beats to speed in waters of various degrees of lubricity.
S = synthetic sea water. B = bay water. P = Polyox added. S2 in
graph A is shown in graph C by black spots. See text for
explanation.

fishes were similarly tested. These were somewhat
larger than the first. Grossly polluted but sand-
filtered bay water was used. The results were in
good accord with the first set, the readings run-
ning a little higher and the slope of proportionality
being a little steeper.

Measures of the varying amplitudes reached by
swings of the tails were not made as they vary
with the tempo of the cycles, as noted by Bain-
bridge (1958), and contribute no additional infor-
mation germane to this study.

A direct result of these experiments is very
clear. The fishes had a choice of two possible
extreme responses to an increase in the water's
lubricity. They could maintain their former speed
by appropriately reducing the frequency of the tail
cycles, or they could so increase their tail beat rate
and thus their rate of translation. Obviously they

could respond by some intermediate response by
partially using each of the above two responses.
Present data cannot be used to determine these
finer distinctions. However, the amounts of the
speed increase in both cases strongly suggests that
most, if not all, of the gain was by increase of
speed.

It might be thought that the fishes were swim-
ming at their accustomed rate in the situation of
these experiments and so would not change their
rate of swimming even when the changed drag
effects reduced the effort required. The phrase
"accustomed rate" may or may not be the same as
their "optimum speed" as defined by Weihs
(1973a). As they did change their pace it seems
most probable that the fish were swimming close
to their maximum, possibly induced by the in-
creased illumination.

The differences in speed of the fishes between
the nontreated water and that with Polyox added,
expressed in percentages, is impressive. Exper-
iments 1 to 5 and 6 to 9 (synthetic seawater) show a
mean increase of 66+%. Experiments 10 to 11 and
12 to 13 (bay water) show a mean increase of 63+%.
The crude percentage figures show no significant
differences between the two cases. The equivalent
figures, using the correction values for size of the
individual, follow: Synthetic seawater, exper-
iments 1 to 9, mean increase 58+%. Bay water,
experiments 10 to 13, mean increase 35+%. Further
analysis may show this to be a real difference.

Cahn's (1972) studies tend to confirm the impor-
tance of both the hydrodynamic and mucus
elements in the formation and maintenance of
schools of Euthynnus affinis (Cantor). The fish
used by her were about 40 cm in total length and
the project was concerned with lateral line studies.
She found that placing a transparent plastic
partition between two fish that had been swim-
ming in parallel courses with the partition, with
one somewhat ahead of the other as the first point
and one of the side points of Weihs' (1973a)
diamond, resulted in the fishes changing to a side
to side position. Without questioning the value of
the lateral line organs, there is also the value of the
mucus and vortices and the "cues" from them
which may be handled by the lateral line system.
How much these sense organs are directly in-
volved with the maintenance of fish schools is not
yet clear. Williams (1967)1^ did not, "... believe

'^Williams (1964), followed by Hamilton (1971), believes that
schooling is primarily a matter of cover seeking.

497

FISHERY BULLETIN: VOL. 74, NO. 3

that the lateral line is important in schooling
behavior." In the same publication Walters and
Liu (1967) ". . . postulate that the boundary layer
acts as a hydrodynamic amplifier . . ." that is
involved in transferring precise information on
changes in w^ater movement that the fish en-
counters as it swims ahead, reaching the fishes
brain via the lateral line system. In a school, much
of such information concerns the water
movements produced by the swimming activities
of the fishes ahead, probably by the bending of the
cupulae that indicate the direction of flow of the
currents and its strength. Other experiments
carried out by different investigators point the
same way as, for instance, the work of Pitcher
(1973) with mirrors. This is not in discord with the
related work reported here and both can be ac-
counted for by the effects of the lattice pattern and
the hydrodynamic and the mucus cues. Also the
work of Shaw and Tucker (1965) and the interpre-
tation of their results by van 01st and Hunter
(1970), based on an optomotor device, indicated
that the test fish reacted more to the fishes ahead
of it than to the moving target spot.

Another source of possible information has been
pointed out by Smith (1930) in some little-noticed
studies. These have shown that Carassius auratvs
(Linnaeus) can draw samples of the surrounding
water into its lateral line canals and expel them as
new samples are drawn in. This behavior certainly
suggests the possibility of a chemical or other
sensory device that could distinguish the concen-
tration of the mucus of preceding fishes. Present
understanding of the relation of the sensory
possibilities related to schooling organization
clearly suggests that such activity of the lateral
line could be a part, or even an important element,
in a following fish's ability to locate the most
favorable position to be stationed in respect to the
mucus of the preceding individual.

Fish at the front of a school receive locomotor
benefits from only their own production of mucus.
All the rest receive benefits from the mucus of
those ahead; those at the very end of a school thus
receive the most benefit. This is sufficient to
account for the "churning" sometimes seen in
schools, the leaders falling back while others press
ahead, all of which helps maintain the integrity of
the school as previously noted.

The peripheral individuals in a school often keep
trying and usually do eventually attain a more
central position, evidently for reasons similar to

those given above. The rapid reorganization of a
school after violent disruption is apparently
similarly motivated.

The existence of fish mills, as noted in the prior
section, may not be the trivial phenomenon it is
generally thought to be. Instead, in the present
view, it may be a resting device with an important
purpose. If the fishes reach a point of fatigue that
would slow the school down to an extent inimical to
the schools integrity, the mill formation would
supply that necessary respite.

All three of the preceding observable items of
activity, as noted, have a consolidating effect on a
school and none show any tendency toward school
dispersion.

The works of Belyaev and Zuev (1969), Zuev and
Belyaev (1970) and Weihs (1973a) discussed the
hydrodynamic effects of one fish on another in a
school, considering only the water movements
induced by the swimming efforts of each member
of the school. This is all in basic agreement with
the present theoretical treatment of the school
organization. Adding to this, the effects of the
drag reducing abilities of the mucus released by
the fishes involved can only result in much higher
efficiency.

Furthermore, there is no evidence that more
mucus cannot be released by fishes to ease their
muscular efforts when necessary. There are,
however, strong probabilities that such abilities
are indeed present. Species that use their mucus
for other purposes have this faculty developed to
a high degree, as in Rypticus (Maretzki and del
Castillo 1967), that exudes a toxic mucus in great
quantities when attacked or handled or many of
the parrotfishes that envelope themselves in a
"cocoon" of congealed mucus on nightfall (Winn
1955). Quality control is also possible with many
fishes under appropriate stimulation. All calcula-
tions at this time involving mucus production are
somewhat uncertain and must remain so until it is
known whether the mucus is exuded at a rather
steady rate or is subject to wide fluctuations,
somewhat after the manner of perspiration in
various mammals.

It is possible that the closing up of ranks, when a
school is in flight from some danger, may destroy
the assistance of both vortices and mucus. Under
this kind of emergency, involving maximum en-
ergy expenditures, this loss may have to be ac-
cepted. Possibly such a situation could call for an
extra outpouring of mucus.

498

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

DISCUSSION

The two basic purposes of this paper are the
estabUshment of the primary space lattice formed
by schooling fishes and the role that their surface
mucus plays. Both features are supported by
theory and empirical data and both expedite the
swimming efforts of the fishes. This alone gives
sufficient reasons for the formation and the
maintenance of schools.

The question of how much of the schooling
phenomenon is a simple following of the paths of
least resistance, with automatic avoidance of other
fishes, how much is social imitation, and how much
is mediated by communication between in-
dividuals is not answered here. The phrase "social
imitation" is discussed at length by Radakov
(1972) as is the status of the term "communica-
tion" discussed by Tavolga (1974). The latter
indicated that the mechanisms involved can begin
as the optomotor orientations of Shaw (1960, 1961).
He added that possibly the responses of the fishes
". . . even as adults may be primarily taxic." The
rheotactic response to vortices and to fish mucus,
reported here, may be equal to or of greater
influence than the optical response, since they are
fully operable in the dark, but not nearly as precise
as the visual response. This could account for the
fact that schooling fishes do not fully lose contact
with each other in darkness even in species not
given to sound production (Breder 1967).

It is recognized, of course, that there is more to
the activity of any fish than efforts to avoid
possible physical exhaustion. An evaluation of the
importance of other activities or even an enumer-
ation of those that are more evident will not be
attempted here. However, another approach to the
overall problem is noted as follows. The "following
reaction" of Crook (1961), based on bird flocks, has
been discussed in connection with fish schools by
Shaw (1960, 1962), Hemmings (1966), and van 01st
and Hunter (1970). The expression is evidently
very nearly, if not completely, identical with the
"social imitation" of Radakov (1972).

These data suggest a hypothesis that could go as
follows. A group of fossil fishes, not living in
schools, but within swimming distances of each
other, may form the background. One fish crossing
in back of another and encountering its vortex
trail would find that self-propulsion required less
effort. It is not unreasonable to suppose that after
a few such encounters, a tendency to follow would
develop. This might be without any instinct to

follow or imitate but not without prior experience
with the vagaries of water currents, which each
fish encounters on its first feeble swimming at-
tempts as a hatchling; nor is there any reason to
dismiss the alternative, that the order is opposite.
In Recent fishes the latter is most probably the
case. However in the early fishes, which are con-
sidered above, the first move to follow could have
been solely on a hydrodynamic basis. From here
on, with the establishment of a primitive school,
its continued existence and development or ex-
tinction would be regulated by selective processes,
depending basically on whether schooling hin-
dered or enhanced the species' ability to survive. It
is visualized that this process could have taken
place many times in various groups, especially
among fishes with relatively scanty mucus
production. Also, this process would probably be
easily reversible so that fish schools could appear
and disappear according to environmental or
physiological changes that made schooling or a
solitary life favor a species' survival.

Detailed conparisons between schools of various
taxa, or between schools formed by a single species
at various times, or under varied conditions have
not been made. It would seem however, that all
schools are not necessarily isomorphic but are
probably at least homomorphic, in the sense of
Ashby (1956).

In a fully theoretical paper, Hamilton (1971)
supported the view of Williams (1964, 1967), that
most types of animal aggregations owe their
existence basically to each animal (vertebrate or
invertebrate) trying to hide behind another. With
this we have no argument (Breder 1967) and our
presentation here, on the locomotor utility of fish
schools, exists comfortably with or without it. The
question of which came first, hiding or benefiting
from an enhancement of swimming efforts, in-
volves no interference. They could have developed
together or independently, each little advance-
ment of one helping the development of the other.

ACKNOWLEDGMENTS

Assistance came from many sources during
these studies and through the production of the
manuscript. The Director of the Mote Marine
Laboratory, Perry Gilbert, provided excellent
support in facilities and professional assistance
during the course of this investigation. H. David
Baldridge was most helpful in connection with
mathematical and physico-chemical matters. Pa-

499

FISHERY BULLETIN: VOL. 74, NO. 3

tricia M. Bird gave unstinted help with both
experimental and field vi^ork and was responsible
for the necessary underwater field photographs,
one of which is reproduced as Figure 15. Lynn
Erdoesy, amanuensis, prepared the manuscript
with patience and skill and Patricia Morrissey
gave it her special attention through production.
Stewart Springer made possible the reproduction
of Figure 11.

Extended correspondence with Daniel Weihs,
Technion-Israel Institute, Haifa, Israel, was of
great help in connection with hydrodynamic mat-
ters, as were communications with Charles W.
McCutchen of The National Institutes of Health.

Ross F. Nigrelli of the Osborn Marine Labora-
tory, New York Zoological Society, supplied special
data; Carl L. Hubbs of the Scripps Institution of
Oceanography furnished data on some Pacific
fishes; James W. Atz of the American Museum of
Natural History provided bibliographic and other
help; and Phyllis Cahn, Long Island University of
New York, provided advice on several matters in
addition to reading the manuscript. The National
Marine Fisheries Service Honolulu Laboratory
provided the photo for Figure 11. Fritz Goro kindly
provided a fresh print of his remarkable photo-
graph used in Figure 20. To all these people go my
profound thanks.

This project was supported, in part, by National
Science Foundation grant GB-34377 from 1 June
1972 (proposal submitted 1 May 1971).

LITERATURE CITED

Aleev, Y. G.

1963. Function and gross morphology in fish. [In Russ.] Izd.
Akad. Naul< SSSR. Moscow, 245 p. (Translated by Israel
Program Sci. Transl., 1969, 268 p., available U.S. Dep.
Commer., Natl. Tech. Inf. Serv., Springfield, Va., as TT
67-51391.)
Alekseeva, K. D.

1963. Role of vision in group behavior of some Black Sea
fishes. [In Russ.] Vopr. Ikhtiol. 29:726-733.

ASHBY, W. R.

1956. An introduction to cybernetics. Chapman and Hall,
Ltd.,Lond.,295p.
ASHTON, W. D.

1966. The theory of road traffic flow. Methuen's monographs
on applied probability and statistics. John Wiley and
Sons, Inc., N.Y., 178p.
Atz,J. W.

1962. Does the shrimpfish swim head up or head down?
Anim. Kingdom 65:175-179.
Bainbridge, R.

1958. The speed of swimming of fish as related to size and to
the frequency and amplitude of the tail beat. J. Exp. Biol.
35:109-133.

Barnaby, K. C, and A. L. Dorey.

1965. A towing tank storm. Trans. R. Inst. Nav. Archit.
107:265-272.
Belyaev, V. v., and G. V. Zuev.

1969. Hydrodynamic hypothesis of school formation. [In.
Russ.] Vopr. Ikhtiol. 57:716-725.
BOWEN, E. S.

1931. The role of the sense organs in aggregations of
Ameiuru^ melas. Ecol. Monogr. 1:1-35.

1932. Further studies of the aggregating behavior of
Ameiurus melas. Biol. Bull. (Woods Hole) 63:258-270.

Breder, C. M., Jr.

1951. Studies on the structure of the fish school. Bull. Am.

Mus. Nat. Hist. 98:1-27.
1954. Equations descriptive of fish schools and other animal

aggregations. Ecology 35:361-370.
1959. Studies on social groupings in fishes. Bull. Am. Mus.

Nat. Hist. 117:393-482.
1965. Vortices and fish schools. Zoologica (N.Y.)

,50:97-114.
1967. On the survival value of fish schools. Zoologica (N.Y.)
52:25-40.
Breder, C. M., Jr., and P. M. Bird.

1975. Cave entry by schools and associated pigmentary
changes of the marine clupeid, Jenkinsia. Bull. Mar. Sci.
25:377-386.
Breder, C. M., Jr., and R. F. Nigrelli.

1938. The significance of differential locomotor activity as
an index to the mass physiology of fishes. Zoologica (N.Y.)
23:1-29.
Cahn, P. H. (editor).

1967. Lateral line detectors. Indiana Univ. Press,
Bloomington, 496 p.
Cahn, P. H.

1972. Sensory factors in the side-to-side spacing and posi-
tional orientation of the tuna, Euthynnus affinis, during
schooling. Fish. Bull., U.S. 70:197-204.
Clark, P. S., and F. C. Evans.

1954. Distance to nearest neighbor as a measure of spatial
relationships in populations. Ecology 35:445-453.
Crook, J. H.

1961. The basis of flock organization in birds. In W. H.
Thorpe and 0. L. Zangwill (editors). Current problems in
animal behaviour, p. 125-149. Cambridge Univ. Press,
Lond.
CuLLEN, J. M., E. Shaw, and H. A. Baldwin.

1965. Methods for measuring the three-dimensional struc-
ture of fish schools. Anim. Behav. 13:534-543.
Escobar, R. A., R. P. Minahan, and R. J. Shaw.

1936. Motility factors in mass physiology: Locomotor ac-
tivity of fishes under conditions of isolation, homotypic
grouping, and heterotypic grouping. Physiol. Zool. 9:66-78.
Gazis, D. C.

1967. Mathematical theory of automobile traffic. Science
(Wash., D.C.) 157:273-281.
Gero, D. R.

1952. The hydrodynamic aspects of fish propulsion. Am.
Mus. Novit. 1601, 32 p.
Guthrie, J. F., and R. L. Kroger.

1974. Schooling habits of injured and parasitized men-
haden. Ecology 55:208-210.
Hamilton, W.D.

1971. Geometry for the selfish herd. J. Theor. Biol.
31:295-311.

500

BREDER: FISH SCHOOLS AS OPERATIONAL STRUCTURES

Healey, M. C, and R. Prieston.

1973. The interrelationships among individuals in a fish
school. Fish. Res. Board Can. Tech. Rep. 389, 15 p.
Hemmings, C. C.

1966. Olfaction and vision in fish schooling. J. Exp. Biol.
45:449-464.
Herald, E. S.

1961. Living fishes of the world. Doubleday and Co., Inc.,
Garden City, N.Y., 304 p.

HiLBERT, D., AND S. COHN-VOSSEN.

1952. Geometry and the imagination. (Translated from
Ger. by P. Nemenyi.) Chelsea Publ. Co., N.Y., 357 p.
HOYT.J. W. '

1966. The friction-reducing effects of high polymers. Nav.
Res. Rev. 19(10):16-21.

1968. Microorganisms-Their influence on hydrodynamic
testing. Nav. Res. Rev. 21(5):18-21.

1971. Drag-reduction effectiveness of polymer solutions in
the turbulent-flow rheometer: a catalog. Polym. Lett.
9:851-862.

1972. The effect of additives on fluid friction. Trans. Am.
Soc. Mech. Eng., J. Basic Eng. 94D;258-285.

1975. Hydrodynamic drag reduction due to fish slimes, /n T.
V. T. Wu, C. J. Brokaw, and C. Brennen (editors). Swim-
ming and flying in nature. Vol. 2, p. 653-672. Plenum
Press, N.Y.
Hunter, J. R.

1966. Procedure for analysis of schooling behavior. J. Fish.
Res. Board Can. 23:547-562.

Hunter, J. R., and J. R. Zweifel.

1971. Swimming speed, tail beat frequency, tail beat am-
plitude, and size in jack mackerel, Trachurus syinme-
tricus, and other fishes. Fish. Bull., U.S. 69:253-266.
Jakowska, S.

1963. Mucus secretion in fish— a note. Ann. N.Y. Acad. Sci.
106:458-462.
Keenleyside, M. H. a.

1955. Some aspects of the schooling behaviour of fish.
Behaviour 8:183-248.
Klausewitz, W.

1963. Wie schwimmt der Schnepfenfisch? Nat. Mus.
93:69-73.

LiNDGREN, E. R.

1967. The lateral line from the point of view of a fluid
dynamicist. In P. H. Cahn (editor). Lateral line detectors,
p. 448-450. Indiana Univ. Press, Bloomington.

Lines, L.

1965. Solid geometry. Dover Publ., 202 p. Corrected reprint
of the 1935 publ.
McFarland, W. N., and S. A. Moss.

1967. Internal behavior in fish schools. Science (Wash.,
D.C.) 156:260-262.
Maretzki, a.; and J. del Castillo.

1967. A toxin secreted by the soapfish Rypticus
saponaceus. Toxicon 4:245-250.
Moss, S. A., and W. N. McFarland.

1970. The influence of dissolved oxygen and carbon dioxide
in fish schooling behavior. J. Mar. Biol. 5:100-107.
Newton, R. N.

1960. Standard model technique at Admiralty Experiment
Works, Hasler. Trans. R. Inst. Nav. Archit. 102:435-470.
Parr, A. E.

1927. A contribution to the theoretical analysis of the

schooling behavior of fishes. Occas. Pap. Bingham Ocean-
ogr. Collect. 1:1-32.
Pitcher, T. J.

1973. The three-dimensional structure of schools in the
minnow, Phoxinus phoxinus (L.). Anim. Behav. 21:
673-686.
Radakov, D. V.

1972. Schooling in the ecology of fish. Izdatel. "Nauka",
Moscow. [In Russ., Engl, transl., 1973, 173 p. Israel
Program Sci. Transl. Publ. John Wiley and Sons, N.Y.]
Richardson, E. G.

1936. The physical aspects of fish locomotion. J. Exp. Biol.
13:63-74.
Rosen, M. W.

1959. Water flow about a swimming fish. U.S. Naval Ord-
nance Test Stn., China Lake, Calif., Tech. Publ. NOTS TP
2298, 94 p.

Rosen, M. W., and N. E. Cornford.

1970. Fluid friction of the slime of aquatic animals. Nav.
Undersea Res. Dev. Cent., Tech. Publ. NUC TP 193, 45 p.

1971. Fluid friction of fish slimes. Nature (Lond.) 234:49-51.

SCHUETT, F.

1934. Studies in mass physiology: The activity of goldfishes
under different conditions of aggregation. Ecology
15:258-262.
Shaw, E.

1960. The development of schooling behavior in fishes.
Physiol. Zool. 33:79-86.

1961. The development of schooling behavior in fishes.
II. Physiol. Zool. 34:263-272.

1962. The schooling of fishes. Sci. Am. 206:128-134, 137-138.
1970. Schooling in fishes: Critique and review. In L. R.

Aronson, E. Tobach, D. S. Lehrman, and J. S. Rosenblatt
(editors), Development and evolution of behavior, p.
452-480. W. H. Freeman and Co., San Franc.
Shaw, E., and A. Tucker.

1965. The optomotor reaction of schooling carangid fishes.
Anim. Behav. 13:330-336.
Shuleikin, V. V.

1958. How the pilot fish moves with the speed of the
shark. Dokl. Adad. Nauk SSSR 119:140-143.
Smith, G.M.

1930. A mechanism of intake and expulsion of colored fluids
by the lateral line canals as seen experimentally in the
goldfish {Carassius auratus). Biol. Bull. (Woods Hole)
58:313-321.
Smyth, H. F., Jr., C. S. Weil, M. D. Woodside, J. H. Knaak,
L. J. Sullivan, and C. P. Carpenter.
1970. Experimental toxicity of a high molecular weight
poly(ethylene oxide). Toxicol. Appl. Pharmacol. 16:422-
445.

SVENDSEN, G. E., AND K. B. ARMITAGE.

1973. Mirror-image stimulation applied to field behavioral
studies. Ecology 54:623-627.

Symons, p. E. K.

1971a. Spacing and density in schooling threespine stickel-
backs (Gasterosteus aculeatus) and mummichog (Fun-
dulus heteroclitus). J. Fish. Res. Board Can. 28:999-1004.
1971b. Estimating distances between fish schooling in an
aquarium. J. Fish. Res. Board Can. 28:1805-1806.
Tavolga, W. N.

1974. Application of the concept of levels of organization to
the study of animal communication. In L. Krames, T.

501

FISHERY BULLETIN: VOL. 74, NO. 3

Alloway, and P. Pliner (editors), Nonverbal communica-
tion, Vol. 1, p. 51-76. Plenum Press, N.Y.
USKOVA, E. T., AND A. V. Chaikovskaya.

1975. Heterogeneity and chemical nature of protein com-
ponents in sea fish mucus coats. [In Russ.] Gidrobiol. Zh.
ll(3):38-42.
Van Olst, J. C, and J. R. Hunter.

1970. Some aspects of the organization of fish schools. J.
Fish. Res. Board Can. 27:1225-1238.
Wade, R. H.

1970. Pollution potential of drag-reducing polymers. NUC
symposium on environmental preservation. Nav. Under-
sea Cent., Tech. Publ. 215:131-143.
Walters, V., and R. K. Liu.

1967. Hydrodynamics of navigation by fishes in terms of the
mucus-water "interface." In P. H. Cahn (editor), Lateral
line detectors, p. 437-446. Indiana Univ. Press,
Bloomington.

Weihs, D.

1972. A hydrodynamical analysis of fish turning ma-
neouvres. Proc. R. Soc. Lond. B Biol. Sci. 182:59-72.

1973a. Hydromechanics of fish schooling. Nature (Lond.)

241:290-291.
1973b. The mechanism of rapid starting of slender fish.

Biorheology 10:343-350.

Williams, G. C.

1964. Measurement of consociation among fishes and com-
ments on the evolution of schooling. Publ. Mus. Mich.
State Univ., Biol. Ser., 2(7):351-383.

1967. Comments from panelists. In P. H. Cahn (editor),
Lateral line detectors, p. 476-477. Indiana Univ. Press,
Bloomington.

WiNN,H. E.

1955. Formation of a mucous envelope at night by parrot
fishes. Zoologica (N.Y.) 40:145-148.

ZuEV, G. v., and V. V. Belyaev.

1970. An experimental study of group swimming in fish
with reference to horse-mackerel Trachurus mediterran-
eus ponticus Aleev. [In Russ.] Vopr. Ikhtol. 63:733-
739. [Transl. in J. Ichthyol. 10:545-549.]

502

ESTIMATION OF FISHING EFFORT IN THE
WESTERN NORTH ATLANTIC FROM AERIAL SEARCH DATA

M. L. Parrack'
ABSTRACT

Three estimators of days fished were developed from aerial search data obtained by fisheries
surveillance operations over the northwest Atlantic off the northeast coast of the United States. These
algorithms estimate fishing effort by applying functions of past aerial observations and past reported
effort to aerial data from the time period for which effort is to be calculated. An estimator based on the
relation of the average number of fishing vessels that were obsen'ed per flight and days fished as
reported has produced easily calculated estimates of days fished to within -0.50 of the reported value in
90% of all cases, 1971-73. An estimator based on the probability of a day fished if not sighted by fisheries
surveillance operations provided an estimate of fishing effort to within *0.50 in 95% of all cases. An
algorithm based on the probability of a day on fishing grounds, if not actually observed, and on the ratio
of days fished to days on grounds enabled the calculation of days fished with largest error (within -0.50
in approximately 80% of all cases).

Prior to 1961, the waters off the northeast coast of
the United States were fished exclusively by the
domestic fleet. However, in 1961 distant water
fishing fleets of other nations began fishing this
area. Concern for the presence of these fishing
vessels prompted the United States to observe and
record the activities and magnitude of such fleets.
These observations over the 160,000 km^' fishing
grounds were made from land-based aircraft; one
to several flights were made each month. Although
fisheries statistics are reported by fishing nations,
such statistics are only available at least 6 mo after
the close of the reporting period. Overflight ob-
servations are therefore the only available up-to-
date information on that fishery.

The fishery in these waters is regulated by the
International Commission for the Northwest
Atlantic Fisheries (ICNAF), a fisheries man-
agement directed treaty organization. Under the
objective of maintaining a maximum sustained
catch, the Commission sets regulations "to achieve
the optimum utilization of the stocks of those
species of fish which support international fisher-
ies in the convention area."- Intensive fisheries
harvest regulations by that agency^ have required
progressively larger cutbacks in fishing by fleets
other than the United States and Canada in these

'Northeast Fisheries Center, National Marine Fisheries
Service, NOAA, Woods Hole, MA 02543.

2ICNAF. 1974. ICNAF Handbook. Dartmouth, N.S., Can., 78 p.

^ICNAF. 1974. Proceedings of the third special meeting,
October, 1973, N.S., Can., 34 p.

waters. (ICNAF Statistical Subareas 5 and 6,
Figure 1.)

The United States has expressed its concern to
ICNAF as to adherence to these fisheries regula-
tions in 1974.^ This concern originated from
preliminary examination of the fisheries over-
flight data.

As a consequence, stochastic methods to monitor
the fishery through the analysis of overflight data
are of chief importance. In response to such needs
three estimators of fishing effort are presented.
These estimators of days fished are based on the
aerial surveillance data and concomitant reported
fishing effort. (Fishing effort as reported by each
ICNAF member nation is published annually,
usually about 1 yr following the reporting period.
Such statistics used in this study were obtained
from the ICNAF Statistical Bulletin, Vol. 19-23,
Dartmouth, N.S.) In each estimation method,
functions developed from aerial surveillance and
reported data in a previous time interval are used
to calculate fishing effort during a future time
interval for which only aerial surveillance data are
available.

METHOD

Fisheries surveillance flights were approxi-
mately 12 h or less in duration and were carried out

Manuscript accepted March 1976.
FISHERY BULLETIN: VOL. 74, NO. 3.

1976.

^ICNAF. 1975. Proceedings of the fifth special meeting,
November, 1974, N.S., Can., 40 p.

503

FISHERY BULLETIN: VOL. 74, NO. 3

y/.'fi

^■

75°

70°

65° 60"=

Figure l.-ICNAF Subareas 5 and 6.

55°

during daylight hours. The primary objective of
each of the flights vi^as to observe as many vessels
as possible. (U.S. and Canadian vessels were not
considered to be of major concern and therefore
were not sought out.) Flight paths were therefore
not set as required by a probability sampling
scheme; rather, a searching technique was em-
ployed. Flights were first directed to areas of
likely fleet concentration. Such areas were deter-
mined from seasonal fleet locations observed on
overflights in preceding years and the reports of
current fleet locations by U.S. fishers. In the event
that major fleet concentrations were not encoun-
tered at the expected location or on the way to it,
the area was searched as extensively as the range
of the aircraft would permit.

During late winter and early spring these dis-
tant-water fishing fleets were concentrated from
off New Jersey southward, so that fishing surveil-
lance operations based in Virginia covered these

fishing grounds. In late spring, the fleets moved
northward to fishing grounds off New Jersey and
New York, and flight paths were directed to those
areas. During the summer and fall, surveillance
flights originating on Cape Cod, Mass., monitored
fishing areas on Georges Bank, Nantucket Shoals,
and, to some extent, areas in the Gulf of Maine.
The fleets moved southward with winter so that
fisheries surveillance again became concentrated
in areas off New York and New Jersey.

Upon encountering a cluster of fishing vessels,
fisheries surveillance agents recorded the hull
identification number, name, and nationality of
each vessel. Other information including the
fishing gear in use and operational mode (i.e.,
engaged in fishing operations or in other activi-
ties) at the time of sighting was also recorded.
Vessels were judged to be fishing if any evidence
was apparent that fishing had occurred on that
day. Since ICNAF defines a day fished as a day in

504

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

which any fishing occurred, an observation of a
vessel fishing was logically defined as an observed
day fished. Although the majority of vessels
sighted were in some phase of fishing operations,
some vessels were observed to be in other opera-
tional modes such as drifting, steaming, anchored,
loading, unloading, or jogging in heavy seas.

These observations therefore allowed certain
fishing effort variables to be derived by nationality
and gear type. They include: 1) the number of
times any vessel was observed on the grounds (i.e.,
observed days on grounds), 2) the number of times
those same vessels were observed fishing (i.e.,
observed days fished), and 3) the number of
vessels.

These overflight data were subject to limita-
tions which were accounted for in the analysis.
First, some observed vessel days have probably
been incorrectly categorized. Surveillance flights
usually occurred before midday; consequently a
sighted vessel that did not fish until late in the day
was recorded as not fishing. Such an event was
therefore interpreted in the analysis as an ob-
served day on grounds but not as an observed day
fished as would have actually been the case. This
limitation, as will be explained later, has little
effect on the estimation of days fished if such
inaccuracies are constant in magnitude through
time. These data were further limited in that
incorrect vessel identifications sometimes oc-
curred. Adverse weather conditions, dense fleet
concentrations, hull scripts of poor visibility, and
inaccurate interpretations of non-Roman script
resulted in the recording of incorrect individual
vessel identifications. Lists received from certain
countries (Japan, Romania, Spain) made possible
the verification of hull scripts observed during
1974. The comparison of these reported hull iden-
tifications with those recorded on overflights
during 1974 determined that individual vessel
identifications recorded two or more times on
fisheries surveillance operations were almost al-
ways correct; those recorded only once were almost
always incorrect. For example, 40 Spanish stern
trawlers were in ICNAF Subarea 5 and Subarea 6
(Figure 1) during the first 10 mo of 1974^. During
that period, 39 separate Spanish stern trawlers
were recorded on fisheries surveillance activities
more than once. The number of vessels in the area

during a time period of interest was therefore
established by considering only vessel identifica-
tions observed by fisheries surveillance personnel
more than once over the period 1965-74.

Knowledge of each country's fishing effort, both
total and by types of gear used, and of the total
effort expended by all countries, is of prime con-
cern in existing fisheries management regimes.
Separate estimates were therefore made for each
country and vessel-type, as were estimates of each
country's total effort and estimates for each of the
total stern trawl and total side trawl fleets.

In addition it was hypothesized that the relation
between reported and observed effort for the
various gear and nationality components could be
different. Stern trawlers are larger and were
expected to be of greater visibility than smaller
vessels. Also, surveillance searching operations
were likely directed towards certain national
fleets as a result of their greater size, their pres-
ence in an area closed to fishing, or because their
catches were of particular immediate concern. If
such relationships are different, separate es-
timates of functions of sighted and reported effort
for each fleet component would logically increase
estimation accuracy.

Estimator I

The ratio of reported days fished to the average
number of sighted days on grounds per flight were
easily computed for time periods when reported
fishing effort was available:

R=f/g' (1.1)

where

r = (t ^.) -.

^ICNAF. 1974. Comments of the Spanish delegation on the
U.S. memorandum annexed to ICNAF Comm. Doc. 74/41. Special
Meeting ICNAF Comm. Doc. 74/44, Ser. No. 3422.

A = the number of flights made during the time

period,
g ' = the number of sighted days on grounds
during the ? th flight, and
/ = the number of days fished during the time
period as reported to ICNAF.

This ratio may then be applied to aerial observa-
tions in some future time period to estimate days
fished before the value is reported:

f=Rg'. (1.2)

R is computed from previous data and g' is cal-

505

FISHERY BULLETIN: VOL. 74, NO. 3

culated from overflight data from the time period
for which the estimate is to be made.

Estimator II

Days on grounds reported to ICNAF (g) were
correlated with observed days on grounds {g') and
fleet size (V) as established from aerial surveil-
lance data to estimate the probability of a day on
grounds that was not sighted [PiG/N)] for any
desired time period (Af):

P{G/N) = ig-g')^ {{VM)-g'). (2.1)

(See Appendix for the derivation of this
probability and for the resulting estimator for
days fished.) In addition, the relation between
days fished (J^ and days on grounds {g) may also be
established from reported effort:

K=f^g.

(2.2)

Fishing effort may then be estimated for some
future time period from overflight data by as-
suming the A" and Pr{G/N) previously established:

f=K[P{G/N){V-M-g') + g']. (2.3)

Estimator III

The probability of a day fished if not observed
[P{F/N)] may be computed for any time period
{\t) for which reported days fished (/), the number
of vessels present {V, as determined from vessel
identification numbers observed on overflights),
observed days on grounds (g'), and observed days
fished if') are available:

P{F/N) = (f-f)/{{Y-M)-g').

(3.1)

(See Appendix for derivation of this probability
and of the resulting estimator.) If this computed
probability is assumed for some future time period
for which reported effort is not available, days
fished for that time period may be estimated:

f=P{F/N)-{{lt-V)-g')+f'.

(3.2)

In order to develop this algorithm, it was assumed
that a vessel did not fish at all during the day it was
sighted if it was observed in the nonfishing mode.
Since surveillance flights were usually completed
before afternoon, it is possible, as noted earlier,

that evidence of fishing was not observable if the
vessel did not fish until late in the day so that the
above assumption may have been violated in some
cases. If this occurred the P{F/N) is incorrectly
calculated, a situation having no effect on the
estimates of days fished if such inaccuracies are
constant from one time period to another. Since
vessels are usually engaged in fishing operations
whenever sea conditions permit, such inaccuracies
can occur only during days when sea conditions
disallow fishing during morning hours (when
surveillance flights usually occur) and permit
fishing later during the day. If the frequency of
such weather conditions are assumed to be con-
stant the magnitude of these inaccuracies may
also be expected to be unvarying.

RESULTS

Reported effort and aerial observation data
from 1969-73 (Table 1) were used in the various
equations to compute i?(estimator I), P{G/N){es-
timator II), /^(estimator II), and P(F/AO(estima-
tor III). The number of surveillance flights {A,
Equation 1.1) is required to calculate R. The
numbers of flights for 1969-73 were 64, 66, 91, 105,
and 109, respectively. The P{F/N) and R were
computed for each gear-country category, for each
country, and for all stern trawlers and all side
trawlers for each year, 1969-73 (Table 2). Since
days on grounds were not consistently reported
except by the German Democratic Republic (GDR)
and in fact were never reported by some countries,
K and P{G/N) could not be calculated in many
cases.

The variables R, P{F/N), and PiG/N) exhibit no
trends of increase or decline through the years
examined; however, these values varied, at times
substantially, from year to year. Therefore, in
order to decrease estimation error, these variables
were averaged whenever possible over years
preceding the year for which the estimate was
made. The average value was then used to make
the estimate. These variables for 1969-72 were
averaged to make the 1973 estimates; 1969-71 were
averaged to make the 1972 estimates; 1969 and
1970 were averaged to make the 1971 estimates;
and the 1969 values were used to make the 1970

estimates.

As stated above, days on grounds were in-
frequently reported so that such sequential aver-
aging of the P(G/AO(estimator II) was not possible
except in the case of the GDR. The Union of Soviet

506

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

Table L-Reported days fished (/), reported days on grounds (g), observed days fished (/'), observed days on grounds (.9 '

size (v). 1969-73.

and fleet

Country
and gear

1969

1970

1971

USSR

Total

35,922

45,391

2,623

2,901

518

23,856

— 2,167

2,478 .

434

26,673

—

3,049

3,726

511

Side trawl

26,518

32,527

1,987

2,130

401

20,173

— 1,

757

1 ,982 :

331

17,468

—

1,891

2,356

332

Stern trawl

7,342

9,476

528

310

106

2,972

—

392

452

7,880

—

1,095

1,279

172

Purse seine

2,024

3,349

676

—

710

—

Not known

—

615

—

Poland

—

Total

6,880

10,679

448

491

9,346

—

646

768

10,599

—

966

1,221

Side trawl

5,738

9,032

332

366

6,339

—

430

507

5,852

—

498

(

519

Stern trawl

1,142

1,647

116

125

3,007

—

216

261

4,747

—

441

558

Not known

—

GDR'

Total

3,750

4,075

249

289

2,096 2,723

200

237

3,619

4,297

429

511

Side trawl

787

848

778 1

1,022

1,950

2,457

265

;

304

Stern trawl

2,963

3,227

192

228

1,318 1

,701

113

139

1,669

1,840

156

198

Not known

—

FRG2

Stern trawl

1,929

—

127

102

2,093

—

166

205

1,285

—

128

Japan

Stern trawl

1,233

—

1,097

—

1,535

—

Spain

Total

—

909

—

102

116

Stern trawl

—

410

—

Paired trawl

989

1,247

464

—

499

—

Bulgaria

Stern trawl

145

—

217

—

1,261

—

111

147

Romania

Stern trawl

195

—

438

—

TOTAL

Side trawl

33,043

■ —

2,565

2,557

475

27,290

— 2,274

2,587 .

406

25,270

—

2,654

3,279

412

Stern trawl

14,805

—

1,014

1,-

183

217

10,899

— 1,002

1,200 :

205

19,225

—

2,012

2,459

305

Country

1972

1973

and gear

USSR

Total

29,492

39,631

2,264

3,263

498

20,948

29,049

1,147

2,024

392

Side trawl

17,307

22,753

1,214

1,725

278

9,244

13,594

276

601

168

Stern trawl

8,680

10,986

1,016

1,341

197

7,630

10,134

734

1,039

200

Purse seine

2,727

4,629

133

4,074

5,321

133

344

Not known

778

1,263

—

742

—

Poland

Total

10,000

—

754

973

104

6,036

—

358

589

Side trawl

5,058

—

380

451

1,733

—

Stern trawl

4,942

—

371

518

4,303

—

283

484

Not known

—

GDR'

Total

4,954

5,675

543

656

4,220

4,642

285

362

Side trawl

1,825

2,361

242

284

1,427

1,651

121

151

Stern trawl

3,129

3,314

300

360

2,793

2,991

164

209

Not known

—

FRG2

Stern trawl

1,020

—

121

859

—

Japan

Stern trawl

1,787

1,821

2,274

—

112

145

Spain

Total

1,552

1,828

102

114

2,405

2,595

201

222

Stern trawl

1,017

1,-

194

2,024

2,106

107

Paired trawl

535

(

334

381

489

107

115

Bulgaria

Stern trawl

1,325

—

134

189

993

—

Romania

Stern trawl

305

—

333

389

TOTAL

Side trawl

24,190

1,913

2,543

359

12,450

—

516

(

313

238

Stern trawl

22,205

—

2,026

2,657

331

21,209

—

1,504

2,14/

374

'GDR = German Democratic Republic.
2FRG = Federal Republic of Germany.

507

FISHERY BULLETIN: VOL. 74, NO. 3

Table 2.- Estimation parameters for Equations 1.2, 2.3, and 3.2.

Stern trawl

Side trawl

Country

1969

1970

1971

1972

1973

1969

1970

1971

1972

1973

USSR

P{F/N}

0.18

0.08

0.11

0.10

0.17

0.16

0.13

0.16

0.15

P(G/N)

0.23

0.14

0.13

0.21

770.30

434.00

560.70

679.60

800.50

796.80

671.80

674.70

1,053.50

1,676.50

0.77

0.79

0.75

0.82

0.76

0.68

Poland

P{F/N)
P(GIN)

0.12
0.18

0.28

0.29

0.28

0.18

0.29
0.47

0.31

0.27

0.23

0.20

584.70

760.40

774.20

1,001.80

969.10

1,003.40

825.20

860.30

1,177.60

2,031.20

0.69

0.64

GDRi

P(F/N)

0.18

0.12

0.15

0.25

0.22

0.09

0.20

0.19

0.14

P(G/N)

0.20

0.16

0.26

0.23

0.10

0.12

0.25

0.16

831.70

625.80

767.10

912.60

1,456.60

825.70

524.00

583.70

674.70

1,030.10

0,92

0.77

0.91

0.94

0.93

0.76

0.79

0.77

0.86

Spain

P{F/N)

0.18

0.46

0.21

20.14

20.04

20,07

20.05

20.07

P{G/N)

0.54

0.22

20.18

20.06

20.03

1,776.70

2,061.80

2781.40

2850.70

2825.60

2668.80

2851.10

0.95

0.96

20.79

20.84

20.78

Japan

P(F/N)
P{G/N)

0.30

0.18

0.23

0.34
0.35

0.36

1,924.70

731.30

1,790.80

2,468.90

1,709.40

0.98

Bulgaria

P(FIN)

0.19

0.14

0.25

0.24

0.21

1,546.70

842.50

780.60

736.10

1,546.20

FRG3

P(FIN)

0.16

0.18

0.20

0.17

0.12

762.10

673.80

913.60

885.10

1,376.90

Romania

P(F/N)
P(G/N)

0.14
0.14

0.13

0.14

0.11

0.12

1,088.00

1,170.00

797.20

1,455.70

1,451.90

0.93

0.86

Total

P(FIN)

0.18

0.13

0.16

0.17

0.14

0.18

0.17

0.15

0.17

0.14

800.90

598.10

711.50

877.50

1,076.70

827.00

696.20

701.30

998.80

1,486.40

Purse seine

Tot

al of all gears

Country

1969

1970

1971

1972

1973

1969

1970

1971

1972

1973

USSR

P{F/N)

0.50

0.14

0.28

0.31

0.47

0.18

0.14

0.13

0.15

0.14

P{G/N)

0.83

0.54

0.59

0.23

0.20

0.19

1,579.70

1,859.00

1,615.30

1,076.40

2,152.90

792.50

635.40

636.40

949.00

1,128.10

0.60

0.59

0.77

0.79

0.74

0.72

Poland

P{F/N)
P{G/N)

0.24

0.38

0.30

0.28

0.25

0.17

896.80

803.20

789.90

1,079.10

1,117.00

0.64

GDRI

P(F/N)

0.15

0.11

0.17

0.23

0.18

P{G/N)

0.16

0.14

0.20

0.26

0,20

673.40

583.70

644.50

792.90

1,270.70

0.92

0.77

0.84

0.87

0.91

Spain

P(FIN)
P(G/N)

0.14
0.18

0.04

0.06

0.05
0.15

0.02
0.11

781.40

397.70

713.10

1,429.50

1,180.00

0.79

0.85

0.93

Japan

P(F/N)
P\GIN)

0.30

0.18

0.23

0.34
0.35

0.36

1,924.70

731.30

179.80

2,468.90

1,709.40

0.98

Bulgaria

P{F/N)

0.19

0.14

0.25

0.24

0.21

1,546.70

842.50

780.60

736.10

1,546.20

FRG3

P{F/N)

0.16

0.18

0.20

0.17

0.12

762.10

673.80

913.60

885.10

1,376.90

Romania

P(F/N)
P(G/N)

0.14
0.14

0.13

0.14

0.11

0.12
0.14

1,088.00

1,170.00

797.20

1,455.70

1,451.90

0.93

0.86

Total

P(FIN)
R

'GDR = German Democratic Republic.

'Paired trawl.

3FRG = Federal Republic of Germany.

508

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

Socialist Republics (USSR) reported days on
grounds in 1969 and 1972 only, so that estimates
for 1970-72 were based on calculations of P{G/N)
and K from 1969 data. The 1973 estimate was
based on the average of the 1969 and 1972 values.
Spanish paired trawl days on grounds were
reported in these same years so that calculations
via estimator II were achieved in the same way as
for the USSR. Spanish stern trawl and Japanese
days on grounds were first reported in 1972 so that
the 1973 calculation of days fished by estimator II
was based on the 1972 data only. Poland and
Romania reported days on grounds in 1969 only, so
that all calculations by estimator II were based on
P{G/N) and /C values computed from 1969 data.

Estimates of days fished were then made for
each country-gear partition, for each country's
total effort, and for all stern trawlers combined
and all side trawlers combined (Table 3). Estima-
tor II was not used to estimate effort for Bulgaria,
the Federal Republic of Germany (FRG), Japan
for 1970 and 1971, and Spanish stern trawlers in
1972 because of the absence of reported days on
grounds which is required by the estimator.

A coeflficient of estimation error was calculated
to establish a measure of estimator performance:

f = (/-/)//•

(4.0)

This error coefficient, then, is the difference
between the estimated days fished (^0 and the
reported days fished (/) expressed as a proportion

of the reported value. An error coefficient was
computed for each estimate made and these
coefficients (Table 4) were then evaluated to es-
tablish the results of partitioning, to compare the
relative abilities of the three estimators, and to
establish estimator dependability.

Inspection of error coefficients indicated that
they decreased considerably (especially those of
estimators II and III) after 1970, likely as a result
of the averaging of estimation parameters. Since
the error coeflftcients then tended to stabilize, only
values of e for the 1971-73 period were used to
analyze estimator performance.

The frequency distribution of e for estimator II
is slightly negatively skewed, a characteristic also
exhibited by the distribution of t for estimator III
(Figure 2). This indicates a positive bias in both
estimators (approximately 10% in each case). Each
of these two distributions is also noticeably lepto-
kurtic indicating a marked clustering of error
coefficients in the interval *0.10. The distribution
of e for estimator I appears to be approximately
symmetrical and without the pronounced peaked-
ness exhibited by the other two. Statistics were
computed from the calculated error coefficients to
establish the probability that the f came from
normal distributions. (These statistics, a and 6,,
and tables of their probabilities are given by
Pearson and Hartley 1956:61-62, 183.) In the case
of the error coefficients of estimators II and III,
the probability that the error coefficients come
from normal distributions is extremely remote,

Table 3.-Estimated days fished calculated by three different algorithms.

1970

1971

1972

1973

Estimate

III

USSR

29,754

30,069

30,063

29,234

35,896

30,410

21,384

34,848

29,293

13,988

25,635

22,488

Side trawl

23,928

22,156

21,970

19,010

22,462

21,200

11,737

18,696

16,417

4,406

10,184

9,627

Stern trawl

5,275

6,156

6,188

8,462

12,009

9,041

7,375

13,717

9,712

5,825

11,321

9,312

Purse seine

574

2,367

2,397

756

2,367

827

2,134

4,209

2,561

5,684

3,617

2,754

Poland

10,435

7,565

7,624

1 1 ,405

9,162

8,485

7,691

9,620

10,846

4,821

7,671

8,369

Side trawl

7,708

6,047

6,007

6,219

6,303

6,172

3,850

6,986

6,393

825

2,533

2,336

Stern trawl

2,312

1,233

1,452

4,124

2,096

2,230

3,485

2,248

4,155

3,465

2,874

5,595

GDR'

2,418

2,786

2,792

3,530

2,608

2,877

3,960

3,312

3,326

2,237

3,894

3,839

Side trawl

1,226

786

7,567

2,254

941

1,037

1,743

1,243

1,278

903

1,432

1,468

stern trawl

1,752

1,861

1,857

1,585

1,579

1,731

2,542

2,100

1,998

1,504

2,421

2,293

Spain

Paired trawl
Stern trawl

685

1,161

2,182

3,205

2,362

912

1,526

1,513

615

994

661

395
508

1,737

201
956

635
2,619

1,363
4,718

1,027
1,079

Japan

2,887

1,817

1,138

1,617

1,073

1,246

2,300

2,200

1,693

Bulgaria

398

294

1,930

884

1,902

1,098

627

934

FRG2

2,367

1,902

1,010

1,111

902

1,065

504

1,211

Romania

181

203

200

620

204

430

213

359

355

259

361

346

Total side trawl

32,418

28,424

27,444

28,500

17,959

23,600

6,750

15,086
23,002

Total stern trawl

14,563
nocratic Republic

14,255

18,903

18,886

17,802

20,541

14,714

iQDR = German Den

2FRG = Federal Rep

ubiic of Germany.

509

FISHERY BULLETIN: VOL. 74, NO. 3

T.^BLK 4. -Error coefficients of three estimate.s of days fished.

1970

1971

1972

1974

Gear type

Estimator

Country

III

USSR

All

0.247

0.260

0.092

0.346

0.140

-0.275

0.182

-0.007

-0.332

0.223

0.074

Side trawl

.186

.098

.089

.088

.286

.214

-.322

.080

-.051

-.523

.081

.041

Stern trawl

.775

1.071

1.082

.074

.524

.147

-.150

.580

.119

.237

.484

.220

Purse seine

-.151

2,502

2.546

.064

2.502

.164

-.217

.544

-.061

.395

.112

.324

Poland

All

.117

-.191

.184

.076

-.136

-.092

-.231

-.038

.085

-.201

.271

.387

Side trawl

.216

-.046

.052

.063

.077

.026

-.239

.381

.264

-.524

.462

.348

Stern trawl

-.231

-.107

-.517

-.131

-.558

-.259

-.295

-.545

-.195

-.332

.300

GDRi

All

.154

.329

.332

-.025

-.279

-.205

-.201

-.331

-.329

-.470

-.077

-.090

Side trawl

.576

.011

.302

.156

-.052

-.468

-.045

-.319

-.300

-.360

.003

.029

Stern trawl

.314

.412

.342

-.050

-.054

.037

-.187

-.329

-.361

-.462

-.133

-.179

Spain

All

.965

2.289

2.299

-.589

-.252

-.093

.333

-.018

Side trawl

.233

.993

.327

-.263

.119

-.802

.664

2.578

1.697

Stern trawl

-.050

.787

.244

1.331

-.467

Japan

All (stern
trawl only)

1.632

.656

-.258

.053

-.400

-.303

.011

-.032

-.256

Bulgaria

All (stern
trawl only)

.836

.355

.536

.299

.435

-.171

-.368

-.059

FRG2

All (stern
trawl only)

.131

-.091

-.214

.136

-.115

.045

-.413

.410

Romania

All (stern
trawl only)

-.070

.045

.027

.416

.045

-.017

-.306

.179

.166

-.223

.084

.038

All

Side trawl

-.188

.042

-.086

.128

.258

-.024

.458

.211

All

Stern trawl

-.339

.308

.017

-.018

.198

-.075

.306

.085

'GDR = German Democratic Republic.
^FRG = Federal Republic of Germany.

.100

.000

.100

.000

.100

Estpmotor I

I |"l |"i |»-r

Estimotor II

rA.

I ' I ' I

U^^l

^■

' ' ' ' 1.30 ' 1 60 ' Z.bC

I ' I ' I ' I ' I ' I ■ ■ ■..30.,eo'^..o

-1.00 -JBO -.60- -fiO -.20 0 .20 .40 .60 .80 1.00 hAO l.TO 2 60

Figure 2.-0bserved frequency of error coefficients (t) of three
estimators of days fished.

less than 0.01. In the case of error coefficients of
estimator I, that probability is 0.05.

An analysis of variance for a one-way layout
with unequal replication (Steel and Torrie

1960:112-114) was employed to investigate possible
differences in the mean values of estimation error
in regards to the kinds of category estimated.
Here, a separate analysis was carried out for each
of the three estimators. All error coefficients for
estimates of total days fished by country, all gear
types combined, were considered as a single group;
error coefficients of estimates of days fished by
each gear type (i.e., estimates of days fished by all
side trawls and by all stern trawls) as another
group; and error coefficients of estimates of days
fished for gear-country categories as the final
group. F-tests indicated a high probability (>0.25)
that errors were the same among these groups in
the cases of estimators I and III. Estimation of
gear totals (i.e., total stern trawl and total side
trawl) was not possible via estimator II because
days on grounds were not reported for all coun-
tries. The analysis for estimator II therefore
included two groups, i.e., gear-country categories
and country totals. As before, the likelihood that
error rates were the same among these two groups
was high (>0.25).

Although application of the F-test requires that
normality assumptions be made, the test is robust
in regard to violations of these assumptions
(Scheffe 1959:361-364) so that limited deviations
from normality are likely to be of limited con-
sequence. However, the nonparametric Kruskal-

510

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

Wallis one-way analysis of variance (Siegel 1956:
184-194) was also applied and indicated the same
general results. The probability of obtaining the
calculated test statistic under the hypothesis of no
difference among these groups with respect to
means is 0.553, 0.410, and 0.872 for estimators I, II,
and III, respectively. Both parametric and non-
parametric tests, then, indicate that the error
rates of each estimator are of the same magnitude
regardless of the kind of category estimated.

Analyses for possible differences in error
coefficients among estimators were carried out in
the same manner. All error coefficients of estima-
tor I were considered as one group, of estimator II
as another, and of estimator III as the third group.
Both parametric analysis of variance and nonpar-
ametric techniques indicated that the different
estimators probably produced different error
coefficients. The likelihood of obtaining the cal-
culated F statistic under the hypothesis of no
difference among the groups is low (0.006). The
Kruskal-Wallis analysis technique also indicated a
low probability of obtaining the calculated statis-
tic under that hypothesis (0.007).

Cumulative frequency distributions of the c
from 1971-73 estimates were used to compare
estimator performances and to establish estima-
tor dependability. These frequency distributions
were established in the following way. Arbitrary
bounds or intervals (ju) were set up so that the first
bound included error coefficients from -0.049 to
-1-0.049, the second from -0.099 to -1-0.099, and so
on. The number of error coefficients from Table 4
falling in each interval was counted; these counts
were then divided by the total number of
coefficients calculated for that estimator to estab-
lish the percent of occurrences in each interval.
These proportions were then interpreted to be the
likelihood of the error coefficient occurring within
each bound (jli, Table 5). Graphs of these
probabilities (Figure 3) indicate that estimator III
is the most desirable. Its error coefl^cient is most
likely to occur within set error bounds of *0.50 or
less. For error bounds greater than *0.50, estima-
tor I was superior. Estimator II was always infer-
ior to estimator III, but for very narrow error
bounds (-0.20 and less) estimator II was superior
to estimator I.

Although estimator II produced the least desir-
able calculations of days fished, a like algorithm
also based upon P{G/N) estimated days on
grounds acceptably well:

Table 5.-Fre(|uency of error coefficients of estimates of (iay.s
fished, 1971-7:1

Frequency of occurence

1 1 lie 1 vai

Estimator 1

Estimator II

Estima

From

tor III

-A

+fl

Nos.

0.049

0.073

0.103

0.182

.099

.255

.282

.364

.149

.291

.385

.455

.199

.382

.436

.565

.249

.564

.462

.636

.299

.655

.538

.727

.349

.727

.667

.855

.399

.782

.718

.891

.449

.855

.718

.909

.499

.909

.769

.945

.549

.964

.846

.945

.599

.982

.898

.945

.649

.982

.898

.945

.699

1.000

.898

.945

.749

.898

.945

.799

.898

.964

.849

.898

.983

.899

.898

.983

.949

.898

.983

.999

.923

.983

1.049

.949

.983

1.699

.949

1.000

2.549

.974

2.599

1.000

Eslimolor !

,..- ■^ _^

— ',-■■

r /

Est.motof III "^ /

/ /

/" Estpmotor II

///■"'

/7\/

//•■'''

- //J

- u

.10
0

1 ' 1

t .099 ! 299 •

.499 •699 ♦.899

> 1.099

t 1.699 1

INTERVAL OF ERROR COEEFICIENT

Figure 3.- Probabilities of error coefficients occurring within set
bounds for three estimators of days fished.

g = P{G/N)iVM-g')+g'

(5.1)

where g is estimated days on grounds and other
symbols are as before. Comparisons of calculated
days on grounds to reported days on grounds were
made when reported values were available (Table
6). These comparisons indicate that the error was
less than *0.50 for 83% of such estimates.

Approximations of estimation dependability
may be established from the calculations of

511

FISHERY BULLETIN: VOL. 74, NO. 3

T.ABLK 6.-Reported and estimated days on grounds and estimation error rates.

1970

1971

1972

1973

Esti-

Total

Side trawl
Stern trawl
Purse seine

mate

Error

mate Error

mate

Error

mate

Error

USSR

43,998

22,799

17,768

7,015

11.0

0.2

62.5

51.6

32,449

13,400

14,337

6,130

11.7

-1.4

41.5

15.2

GDR

Total
Side trawl

3,028
845

11.2
-17.2

3,387 -21.2
1,238 -49.6

3.941
1,573

30.6
-33.4

4,476
1,859

-3.6
12.6

Stern trawl

2,023

18.9

2,051 11.5

2,308

-30.4

2,576

-13.9

Spain

Total

Paired trawl
Stern trawl

1,779

180.6

3,426
1,181
6,937

32.0

-43.9

1,318.6

if erro:

r (Ou, Table £

i). The m

•obabilit

;.y

DISCUSSION

statement

Pr{-n< e <fi) = $ju.

defines these calculated values. Here, e is the error
coefficient (Equation 4.0), ju is an arbitrary limit of
acceptable error, and ^jx is the likelihood of € being
within the interval -ju to +ju. By substitution (i.e., t
= U'-f)/f) and reduction of terms:

Pr(

1-iU

■/

1-fX

) = /x

Bounds on points estimates can therefore be
calculated:

upper limit = /".

lower limit = f.

1
l + /i

and values of Oju from Table 5 can be used to
approximate the likelihood that the reported days
fished (f) will fall within the interval. For example,
if estimator III calculated days fished to be 100, it
may be stated that there is a 0.945 probability that
the reported value of days fished is within the
interval 67 to 200, i.e..

100

^^^-:o:^>/='rr5:499) = 0.945.

It is important to note, however, that the figures in
Table 5 were computed directly from the
frequency distributions of error rates of estimates
made in the past (i.e., 1971-73). Therefore, estima-
tion bounds may be correctly approximated on
future point estimates only if it is assumed that
future distributions of error coefficients are cor-
rectly represented by these past performance
data.

Of the methods presented, estimator I (based on
the ratio between days fished as reported and
sighted days-on-grounds) and estimator III (based
on the probability of a day fished given that it was
not observed) exhibited the least error. Estimator
III was consistently most accurate (especially in
the last 2 yr) although the difference between the
two is small. This estimator may be expected to
calculate days fished to within *0.50 approximate-
ly 95% of the time. Estimator I has value in that it
does not require sophisticated analysis of over-
flight data (only the numbers of sightings and
numbers of flights are needed) and is less likely
than other estimators to produce error coefficients
greater than 0.50. It may be expected to produce
estimates within *0.50, 90% of the time.

Estimator II, based on the probability of a day
on grounds if it was not sighted, was consistently
the poorest of the three. Its poor performance is
likely the result of insufficient instances of
reported days on grounds. These parameters allow
computations of P{G/N) and K, on which the
estimate is based. In the case where complete data
was available (GDR), eff'ort was estimated very
well by estimator II and, in fact, the error
coefficient did not exceed 0.42. A similar estimator
also based on P{G/N), however, produced accept-
able calculations of days on grounds for all coun-
tries that were within *0.50 of the reported value
in approximately 80% of all cases.

Estimation error can result from sources which
are known to have occurred in the past and are,
therefore, of a magnitude predictable by the
proposed methods for approximating probability
limits on point estimates. The probabilities of
fishing {P{F) =f/{V-At)) from data in Table 1 were
found to be highly correlated with the PiF/N)

512

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

(Table 2) as theoretically should occur. Therefore,
when countries change fishing patterns from one
time period to the next so that PiF) differs,
P(F/N) also changes, thus introducing error in the
estimates made by estimator III, a condition also
true for P{G) and P{G/N) on which estimator II is
based. This results from changes in the mean
number of days fished (or days on grounds in the
case of estimator II) per vessel, a likely occurrence
if a particular fleet experiences difficulties in
finding fish, if weather conditions are unusually
unsuitable for fishing, or if equipment repair or
modifications demand excessive lost time in a
certain time period.

Although these changes theoretically should not
produce changes in the ratio of reported to ob-
served fishing effort {R, estimator I), other factors
can conceivably produce such variation of that
ratio. Changes in visibility due to weather can
likely be an important factor. If fog or other
visibility-restricting weather conditions are more
prevalent in one time period than another, R may
be expected to be larger during that period.
Likewise, varying success of overffights in locat-
ing fleet concentrations is a factor. Unusually
successful searching may be expected to produce
ratios smaller than average while low success will
tend to increase R.

In addition, changes in the accuracy of reported
effort (/andgr) will result in corresponding changes
in the accuracy of calculations of P{F/N), P{G/N),
R, and K for particular time periods. Since
reporting accuracy cannot be measured, such
deviations have been included in the error
coefficients as have the above listed sources of
error.

Although a method of calculating probability
limits on estimates is presented, the methodology
utilizes the observed past performance of each
estimator to establish the probability of error. It
must be assumed, therefore, that the frequency
distribution of estimation error is correctly repre-
sented by these past data. Although this assump-
tion can reasonably be made if fisheries surveil-
lance flight patterns and fishing fleet movements
are generally constant, caution should be exercised
in this regard. If flight patterns or seasonal fleet
movements change drastically, the probabilities of
not sighting fishing effort (estimators II and III)
and the ratio of reported to sighted effort (es-
timator I) will likewise change so that they are not

correctly represented by the range of past values.
Aberrant values will result if the fleets are exten-
sively concentrated in different areas than in the
past. Fleets will not be located by fisheries sur-
veillance flights as well as in the past and, there-
fore, effort will not be observed to the same extent
as in the past. As a result, values of PiF/N), and R
will be much greater than past values. Sizable
underestimates of days fished will occur with
probabilities greater than those represented by
past error frequencies. Conversely, if fleet loca-
tions are anticipated by surveillance flight per-
sonnel much more accurately than in the past,
these estimation constants will be much smaller
than represented by past data, so that probabili-
ties of overestimation will be much greater than
represented by past performance data.

ACKNOWLEDGMENTS

All overflight data utilized in this study was
collected by the National Marine Fisheries Ser-
vice, Law Enforcement and Marine Mammal
Protection Division, Northeast Region. I extend
my thanks to Charles Philbrook and William Beers
of that agency for their advice as to the contents of
those data. I am indebted to Judith Brennan and
Bradford E. Brown of the Northeast Fisheries
Center, National Marine Fisheries Service for
their advice in the development of these meth-
odologies and manuscript. I am especially indebt-
ed to Robert L. Edwards and Richard C. Hen-
nemuth, also of the Northeast Fisheries Center for
the continued support and advice which made this
study possible.

LITERATURE CITED

HOEL, P. G.

1962. Introduction to mathematical statistics, 3rd ed. John
Wiley and Sons, Inc., N.Y., 427 p.
Pearson, E. S., and H. 0. Hartley (editors).

1956. Biometrika tables for statisticians. Vol. I. Cambridge
Univ. Press, Lond., 238 p.
Scheffe, H.

1959. The analysis of variance. John Wiley and Sons, Inc.,

N.Y.,477p.

SlEGEL, S.

1956. Nonparametric statistics for the behavioral sci-
ences. McGraw-Hill, N.Y., 312 p.
Steel, R. G. D., and J. H. Torrie.

1960. Principles and procedures of statistics with special
reference to the biological sciences. McGraw-Hill, N.Y.,
481 p.

513

FISHERY BULLETIN: VOL. 74, NO. 3

APPENDIX

Estimators II and III are based on probabilities
of not sighting daily units of fishing effort (i.e.,
vessel days which were on grounds or on grounds
and fishing). These estimates require the calcula-
tion of these probabilities during some time period
when reported days on grounds and days fished are
available which can be correlated with sighted
days on grounds, sighted days fished, and fleet size
as determined from surveillance overflight data.
These estimators were fashioned by considering
the possible daily events, constructing the
probability space in units of vessel days from
reported and observed effort, and then deriving
needed probabilities from the constructed space.
Let the possible events be symbolized:

G = the daily event of a vessel on the grounds;
F = the daily event of a vessel on the grounds

and fishing;
E = the daily event of a vessel not on the

grounds, i.e., elsewhere;
S = the daily event of a unit of daily effort

observed on overflights; and
N = the daily event of a unit of daily effort not

observed on overflights.

Further, by defining the fleet size during some
time period Af , where t is in days, as the number of
vessels that were present at some time during that
period, the total event space (which is the sum of
all possible events) is easily calculated:

n = Vlt

where n = the total number of all possible
daily events,

It = the time period in days, and
V = the fleet size.

Even though certain cells in the event space (of
little consequence to us) cannot be observed di-
rectly, the possible events defined in units of vessel
days may be broken down:

From effort reported to ICNAF and from that
observed on overflights, the number of daily
events in each cell is easily defined. The number of
events of vessel days fished (/) are reported to
ICNAF and are either observed on overflights (/')
or are not seen on overflights (/"). The number of
events of vessel days on grounds (g) may be
reported to ICNAF and are either observed on
overflights {g') or not observed on overflights {g").
The event of a vessel day on grounds spent not
fishing (o) is either observed (o') or not observed
(o"). It is important to note that if a vessel day on
grounds was fished but was observed on an over-
flight as not in the fishing mode, which may
possibly occur if fishing operations were not
initiated until late in the day after the flight
occurred, that event would be incorrectly categor-
ized as o' rather than as /'. The effect of this
occurrence on estimator III will be discussed; it has
no effect on estimator II. If a possible daily event
were not on the grounds, then it was elsewhere (e)
and was not observed {e"). It was not possible for a
vessel day elsewhere to be observed; overflights
were directed within the fishing grounds so that e'
is zero in all cases. The numbers of daily events,
then, are categorized as on the grounds (g) and
fishing (/) or not fishing (o), or as elsewhere (e).
The numbers of daily events in each category are
symbolized by a single prime ( ') if observed on
overflights, as a double prime (") if not observed,
and without a symbol if the value is a total number
of events.

Estimator II is best explained by considering
the probability of a day on grounds:

''PiG) = PiG,S) + PiG^

= P{G,S) + P{G/N)PiN).

From the above event space, the probabilities of
on grounds and observed [Pr{G,S)], of on grounds
if not observed [Pr(G/AO], and of not observed
[P{N)], are easily defined:

PiG,S) = g'^n,
P{G/N) = g" ^ {n - g'), and
P{N) = {n - g') H- n.

Item Observed (S)

On the grounds (G) g'

Fishing (F) /'

Not fishing <>' = (/' -f"

Elsewhere (E) f" = zero

Total o'

Not .seen (N)

Total

.</" = g -9'

f'=f-f'

o" = 0 - o'

" = /- ff

e = n - 3

Then by substitution
P{G) = {g'^n) + [{g - g') ^ {n - g')\{n - g') ^ n

®For an explanation of probabilities and the theorems used in
the development of these expressions, see Hoel (1962:4-17).

514

PARRACK: FISHING EFFORT FROM AERIAL SEARCH DATA

which reduces to

PiG) =9^n;

so that the equation may be solved for days on
grounds:

9 = nP{G).

Then by substitution

g = n[(g' ^ n) + P(G/N)in - g') ^ n],or
g^g' + P{G/N){n-g').

Estimator II is then derived by inclusion of the
ratio of days fished to days on grounds, K = f -^ g,
so that the above algorithm may be expressed in
terms of days fished:

f=K{g' + PiG/N)in-g')].

An estimate of days fished (./) may be made, then,
from surveillance overflight data if calculations of
R and P{G/N) can be made from past data.

Estimator III is deduced from the event space
according to the same rationale. The likelihood of
an event of a vessel day fished expressed as
observed and not observed is expanded to calculate
days fished. From the event space it is apparent
that:

P{F) = P{F, N) + P{F, S)

where PiF) is the probability of a vessel day fished,
P{F, N) is the probability of a vessel day fished and
not observed on overflights, and P{F, S) is the
probability of a vessel day fished and observed on
overflights. Further, by application of the multi-
plication theorem of probabilities

P{F,K} = P{F/N)-P{N)

where P{F/N) is the probability of a vessel day
fished given that it was not observed on over-
flights, and P{N) is the probability that a possible
vessel day (regardless of location or operational
mode) was not observed on overflights. The first
expression therefore can be written as

P{F) = P(F/N)P{N) + P{F,S).

Although all possible probabilities can be ex-
pressed in terms of the number of events in each
category of the event space, those of interest are:

P(F,S)=f'/n,

P{N) =(« -.g')/«, and
P{F/N)={f-f)/in-g').

By substitution and reduction of terms

PiF) =f/n.

The number of vessel days fished, then, is the
product of the entire event space and the
probability of fishing, i.e.,

/= nPiF).

Then, by substitution, estimator III easily follows
so that days fished are estimable from overflight
data if PiF/N) can be predetermined from past
data:

f = f' + PiF/N)in-g').

From possible algorithms derivable from the
event space, this form makes most use of over-
flight data and is least dependent on functions
calculated from past data.

515

FOOD AND FEEDING OF LARVAE OF THREE FISHES OCCURRING

IN THE CALIFORNIA CURRENT, SARDINOPS SAGAX,

ENGRAULIS MORDAX, AND TRACHURUS SYMMETRICUS^

David K. Arthur-

ABSTRACT

The size, number, and types of food particles eaten by larvae of Pacific sardine, Sardinops sagax;
northern anchovy, Engraulis mordax; and jack mackerel, Trachurus symmetricus, were determined
by an examination of gut contents of larvae captured in plankton samples from the California Current.
Food particles found in larvae of the three fishes were predominantly the eggs, nauplii, and the
copepodid stages of the smaller species of copepods. These increased in width as the larvae grew though
not so uniformly for the anchovy as for sardine and jack mackerel. Particles ingested by anchovies at
first feeding were slightly larger than were those ingested by sardines, while jack mackerel could eat
particles three times wider than sardines of equal length. The smallest individuals of each species were
the most euryphagous, especially anchovies. Feeding incidence of sardine and anchovy declined during
the early larval period while that of jack mackerel increased. Sardine and anchovy larvae fed only
during the day. The data were not analyzed for day-night feeding for jack mackerel.

The relative body depth and relative weight of laboratory-grown anchovy larvae increased
throughout the larval periods examined, whereas, the relative body depth of most ocean-caught
anchovy larvae decreased during the first half of this period, possibly as a result of the poorer ration
obtainable in the ocean. The decline in relative body depth of ocean-caught anchovy larvae may be
related to the decline in feeding incidence and to the apparent lack of increase in size of the food
particles ingested.

Owing to the impending collapse of the Pacific
sardine, Sardinops sagax, fishery, a biological-
oceanographic survey program, which later
became known as the California Cooperative
Oceanic Fisheries Investigation (CalCOFI), was
initiated in March 1949. Instrumental in initiating
a program to study the food of the sardine larva
was the concept developed by Hjort (1914) that the
success of a year's spawning may be determined at
the critical period when the fragile larvae must
secure sufficient food from their environment. For
a recent and thorough review of the literature
concerning this subject, the reader is directed to
May (1974).

To explore the possibilities proposed by Hjort
(1914), 10,408 sardine larvae from 398 samples
were examined. Food of two potential competi-
tors, namely the northern anchovy {Engraulis
mordax, 2,350 specimens, 69 samples) and the jack
mackerel {Trachurus symmetricus, 750

'Based on a portion of a dissertation submitted in partial
satisfaction of the requirements for the Ph.D. degree at the
University of California, Scripps Institution of Oceanography.

^Visiting Scientist, Southwest Fisheries Center, National
Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA
92038.

specimens, 65 samples) was also investigated
(Arthur 1956). Larvae of these three fishes were
supplied to me by Elbert H. Ahlstrom and came
from samples taken during early years of the
CalCOFI program.

Sardines no longer support a viable fishery, but
anchovies have increased in numbers to fill, in
part, the ecologic if not economic void. Increasing
attention, therefore, will be paid in this paper to
this fish and to other species of the genus En-
graulis which occupy coastal environments of
many parts of the world (Reid 1967).

METHODS

Specimens were examined in glycerin because
of its advantages over water. Its clearing qualities
aid in the detection of food particles within the
gut, and the greater viscosity of this medium
dampens the movement of particles during dis-
section. Also, when in glycerin, larvae seem to be
more pliable and the intestinal walls do not tend to
fragment so readily.

At first the entire intestinal tract of each sar-
dine larva was dissected from the body. This

Manuscript accepted Februarv' 1976.
FISHERY BULLETIN: VOL! 74, NO. 3, 1976.

517

FISHERY BULLETIN: VOL. 74, NO. 3

procedure proved to be difficult, unnecessary, and,
at times, actually misleading. It was unnecessary
because no food was ever found in the thin walled
anterior intestine which forms about half of the
total length of the digestive tract. Also, to view the
anterior intestine, the liver which surrounds most
of it must be carefully teased away resulting in the
production of many fragments which may be
confused with possible food particles. Schumann
(1965) observed that food particles pass through
this portion of the gut in about 25 s in labora-
tory-reared sardine larvae. The intestines of jack
mackerel larvae are not as readily observable as in
sardine or anchovy larvae because they are
covered by well-developed pelvic fins and because
of the earlier development of substantial body
walls.

The presence of a single food particle in larval
sardines or anchovies can usually be detected by a
localized swelling of the surrounding gut wall.
When several food particles are present, the
posterior intestine may be highly expanded over
its entire length. Food particles were dissected out
of the gut by means of an instrument consisting of
a pig's eyelash, bevelled cut to form a chisel point,
and mounted in beeswax in one end of a glass tube.
Food particles were identified to taxa as far as
their condition allowed.

Each organism found in the intestine was mea-
sured as to the maximum cross section that the
larva would have to encompass for ingestion.
Herring larvae have been shown to ingest crus-
tacean food particles "head on" by Hardy (1924),
Bowers and Williamson (1951), and Blaxter (1965).
This maximizes the ingestible size of the organism
and positions appendages, spines, and setae to the
rear of the food organism during its transit
through the intestines.

To facilitate a consideration of changes in food
with respect to growth, the size ranges of larvae of
each of the three species of fishes being considered
here have been subdivided into three length
groups. The length intervals used in these sub-
divisions are based on the distribution of sizes in
the collections rather than on any definite changes
in the larvae with respect to age.

FOOD OF SARDINE LARVAE

Table L-Food of sardine larvae.

Size group

End of yolk-
sac stage
to 5.5 mm

No. %

6,0 to
9.5 mm

No. %

10.0 to
25 mm

Food items

No.

Copepod eggs
Copepod nauplii:
Calanoid
Cyclopoid
Harpacticoid
Unidentified

141

179

22.0

149

10.8

3
3
2
5

28.6

Total nauplii

349

54.5

248

76.3

37.1

Copepodid stages:
Calanoid
Cyclopoid
Harpacticoid
Unidentified

1
7
2
1

7
3

Total copepodids

0.3

3.4

31.4

Dinoflagellates
Tintinnids
Foraminifera
Unidentified

crustacean eggs
Unidentifiable

material

120

0.8
3.1
0.5

18.8

9.5

2.9

Total number of
food particles

640

325

are presented in Table 1. Length of sardine larvae
at the end of the yolk-sac stage is variable. One as
small as 3.4 mm contained food, others as long as
5.5 mm still had yolk, but no larva was observed
containing both yolk and ingested food organisms.
Eggs, nauplii, and juvenile stages of copepods
composed almost all of the identifiable food.
Copepodid stages in the diet increased in percent-
age by a factor of 100 through the successive size
groups of the larvae. This is undoubtedly due to
the severe size restrictions placed on the young
larva by the small size of its mouth. As the larva
increased in size, its mouth likewise increased in
gape, and consequently a larger range of the size
spectrum of plankton became available.

Size of Food

Although they are not always of the largest
ingestible size, the food particles increased in size
isometrically with the increased length of sardine
larvae (Figure 1). A larva in doubling its length
from 4 to 8 mm likewise doubled its maximum
ingestible food size from 80 to 160 /xm.

Type of Food

The qualitative results of the examination of
food material from intestines of larval sardines

Feeding Incidence

The percentage of fish containing at least one
food particle is termed the "feeding incidence" for

518

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

250

E 200
§ '50

O 100

UJ
N

c/5 50

' "• ^'^■

. ; i 1 '[X^r^^^"^^^-

— •■•

4 5 6 7 8 9 10 II
LENGTH OF LARVAE (mm)

Figure l.-Food size of Pacific sardine larvae. The line is a least
squares fit to all data points and is expressed by the equation: S
= 13.04L + 5.70, where S is width of food in microns and L is
standard length of larvae in millimeters. The correlation
coefficient r is 0.813 and the coefficient of determination r-
implies that 66% of the variation in food size can be explained by
lan'al size alone.

a particular sample and is considered a measure of
a larva's ability to obtain food in the environmen-
tal circumstances at the time of sampling.

The available data permit an inspection of the
average hour-by-hour series of trophic events for
sardine larvae (Figure 2). The data were divided
into 16 intervals composed of: the first half hour,
the second half hour, the second hour, and the third
hour both before and after sunrise, and both
before and after sunset. There were also midday
and midnight intervals which vary in length
according to the season. Only those intervals in
which at least 50 larvae from at least five samples
were included. Feeding incidence of all three size
groups increased throughout the day. This could
have resulted from accumulation of food in the
gut, or perhaps to the success of larvae in finding
more suitable feeding conditions as the day pro-
gressed. The largest size group demonstrated the
fastest return to a low feeding incidence at night
which probably reflects faster digestive rates for
older larvae, as has been shown for plaice larvae
(Yasunaga 1971). The lower feeding incidence of
older larvae may, therefore, be partly due to an
increased digestive rate.

Figure 2 illustrates the diurnal nature of feed-
ing which is due to the visual feeding sardine larva
requiring light to detect its prey. (Schumann
1965). This results in diurnal changes of the
intestine. The posterior intestine of larvae cap-
tured during the early morning is visibly striated.

UJ
Q
O

Q
UJ
UJ

Sunrise
30

20 -

->Day

-^ Sunset

-^ Night

• - End of yolk-soc to
5 Smnn

O-6 0 to 9 5mm

A-- 10 0mm

> Sunrise
130

-20

Average of

remenfs

123 321 123 321

HOURS BEFORE AND AFTER SUNRISE OR SUNSET

Figure 2.-Diurnality of feeding incidence of Pacific sardine
larvae. Only those intervals in which at least 50 larvae coming
from at least 5 samples are included.

In late afternoon and early evening, the posterior
intestines of many larvae, especially the smaller
ones, are expanded and have no visible striations.
The intestinal wall contains large vacuoles of clear
fluids. During this period of the day, it is common
to capture larvae with greatly expanded intestines
but with no identifiable food organisms. Often
such larvae contain some granular material float-
ing about in the intestinal lumen which is com-
paratively large due to the expansion of the
surrounding wall. During the night, intestinal
expansions disappear and by sunrise almost all of
the larvae have returned to the compact, striated
intestinal condition. This rhythm is most pro-
nounced in the smallest size group where, as
indicated in Figure 2, the amplitude of the diurnal
feeding incidence is at a maximum.

FOOD OF ANCHOVY LARVAE

Type of Food

No larva was found containing both yolk and
ingested food. The diet of anchovy larvae (Table
2) is very similar to that of the sardine. The most
striking difference is that very young anchovies
are more euryphagous. About 40% of their diet (by
numbers) consists of noncrustacean food particles.
A food category entitled "unidentified spheres" is
represented by small (about 20 fim) objects,
probably of plant origin. Copepod nauplii become
increasingly important as anchovy larvae increase
in length and compose the bulk of particulate food
when all sizes of larvae are considered.

Copepod eggs and nauplii were found to be the
most important element in the diet of larvae of

519

FISHERY BULLETIN: VOL. 74, NO. 3

Table 2.— Food of northern anchovy larvae.

Size group

End of yolk-

sac

stage

5.0 to

7.0 to

to 4.5 mm
No. %

6.5

9.0
No.

Food items

No.

Copepod eggs

15.3

14.3

Copepod nauplii:

Calanoid

Cyclopoid

Harpacticoid

Unrecognizable

Total nauplii

42.9

67.9

90.9

Copepod adults:

Calanoid

9.1

Clam larvae

2.0

Foraminifera

2.0

Tintinnids

3.1

Dinoflagellates

7.1

3.6

Ciliates

2.0

7.1

Coccolithophores

4.1

Unidentified spheres

21.4

7.1

TOTAL

Engraulis mordax (Berner 1959), of E. anchoita
(Ciechomski 1967), and E. ringens (Rojas de Men-
diola 1974). Berner and Rojas de Mendiola found
considerably more eggs than nauplii while Cie-
chomski reported about equal numbers.

An unusual example of feeding by both anchovy
and sardine larvae was called to my attention by
Elbert H. Ahlstrom because of the obvious gorged
intestines of some of the larvae. This sample was
taken about 38 km off the coast of central Baja
California approximately 6 h after sunset and IV2
h after setting of a "first quarter" moon. Unusual
aspects of the sample were that most of the larger
larvae of the two species contained food at night
and that they were literally crammed with the
pteropod Limacina bulminoides. Of the larvae
over 10 mm in length, the 26 feeding sardines
averaged 24 pteropods per gut with a maximum of
54 in one 23-mm larva, and the 19 feeding an-
chovies averaged 18 per gut with a maximum of 26
in a 14-mm individual. Compared to the one or two
food particles usually found in a feeding anchovy
or sardine larva, the number of pteropods was
surprising. The only other molluscs found in either
sardine or anchovy larvae in this investigation
were two bivalve larvae, one each in two very
young anchovies. No molluscs were reported in the
extensive investigations of the food of anchovy
larvae by Berner (1959), Ciechomski (1967), or
Rojas de Mendiola (1974). It cannot be determined
whether the larvae found filled with Limacina
reflected beneficial feeding conditions where they

were found, or a hazardous situation in which the
larvae had ingested material they could not digest
and void.

Because the number of pteropods found in this
one sample is larger than the total number of food
particles found in the larger sardine and anchovy
larvae of all other samples examined, they were
not included in the overall tabulations for these
fish. The size of the pteropods was used to establish
the upper ingestible size of food particles for older
anchovies (Figure 3a).

300
^ 250

Q 200

0
0

"!: '50
0

^ inn

T
•••

■r.b

•

N 100 • ^ =. ;

> .. ..•

•• •

5 6 7 8 9 10 II 12
LENGTH OF LARVAE (mm)

1 L200

■

X 1000

^ 800

■

^ 600

•

•'''^Width of mouth

0 400

^ 200

2 n

^^ Largest size ingested

-■""'^ Smollest size ingested
-1 1 — 1 1 1 — J—J — 1 — 1

0 4 8 12 16 20 24 28 32
LENGTH OF LARVAE (mm)

Figure 3.-(a) Food size of northern anchovy larvae. The line is a
least square fit to all data points from 3 to 9 mm inclusive and is
expressed by the equation: S = 16.16L , where S is width of food
in microns and L is standard length of larvae in millimeters. The
correlation coefficient r is 0.473 and the coefficient of the
determination r- implies that 22% of the variation in food size
can be explained by larval size alone, (b) Food size of anchoveta
{Engraulis ringens) larvae. Adapted from Rojas de Mendiola
(1974).

520

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

Size of Food

Food particles of young anchovy larvae are not
selected from those near the largest ingestible size
as are those of young sardine larvae, though there
is a trend to increase particle size as larvae
increase in length. The correlation coefficients
(Figures 1, 3a) suggest that food size of sardines is
more controlled by larval size (0.813- or 66% of
variance explained) than of anchovies (0.4732 or
22% of variance explained). The extensive data,
including many older larvae, reported by Cie-
chomski (1967) and Rojas de Mendolia (1974)
indicate a sharp increase in food size between the
lan-al lengths of about 3 to 4 mm but relatively
little increase for most of the remainder of the
larval period. Rojas de Mendiola's data (Figure
3b), including food sizes of 2,088 feeding larvae 3.1
to 5.0 mm in length, are used to illustrate this
important point. These data indicate that food size
roughly doubles (from approximately 100 to 200
lim) while larvae grew from 4 to 16 mm. Assuming
that both larvae and food particles increased in
size isometrically, then their volumes increased by
the cube of their increase in length or width. Food
particles in doubling in width increased 8 times in
volume, while larvae increasing 4 times in length
increased 64 times in volume. Therefore, the
nutritional equivalent of a 200-/xm food particle to
a 16-mm larva is only one-eight of that of a 100-jum
particle to a 4-mm larva. Although Berner (1959)
measured the length rather than the width of food
particles and his data are not directly comparable,
they do indicate that while anchovy larvae in-
creased in length from 3 to 10 mm (an increase of
37 times in volume) their average food size in-
creased from 68 to 128 jum (an increase of only 6%
times in volume).

Feeding Incidence

Anchovy larvae also are daytime feeders
(Figure 4). The disparity between night and day
values for feeding incidence is greater for an-
chovies than for sardines during their youngest
larval stages. This difference perhaps is due to a
faster digestive rate for the anchovy.

FOOD OF JACK MACKERAL LARVAE
Type of Food

The jack mackerel larva first starts to feed when

5 6 7 8

LENGTH OF LARVAE (mm)

Figure 4.-Diurnality of feeding incidence of northern anchovy

larvae.

it is about 3.25 mm long. By the end of the yolk-sac
stage, the jack mackerel has attained a robustness
which contrasts sharply with the slender early
larval sardine or anchovy. Its body shape, in
general, is more substantial and its mouth is
proportionately larger. No jack mackerel larva
was found with both yolk and ingested food
organisms.

Just as for sardine and anchovy larvae, copepods
contributed the greatest bulk of its food (Table 3).
Eggs and naupliar stages, however, are much less
important. The "egg sacs," appearing under the
title of copepod eggs, were probably ingested
attached to adult copepods and so represented a
coincidental fraction of the food. Copepod nauplii
seemed to be significant in numbers only in larvae
of the smallest size group.

Copepodid stages of copepods make up the bulk
of particulate food, increasingly so as the larva
grows older. By the time the larva is 7.0 mm long,
96.0% (by number) of its food is composed of
various species of copepods. The most significant
feature of the diet is the very high percentage of
occurrence of Microsetella norvegica, one of the
few planktonic species of harpacticoid copepods.
This probably represents a definite selection, as
this species of copepod, though ubiquitous, never
achieved numerical importance in our plankton
hauls. Jack mackerel (Trachurus trachurus) lar-
vae were reported by Sinyukova (1964) to have an
"inborn ability" to select two species of copepods
from the mass of plankton living in the Black Sea.
On the other hand, the respective behavior of the
early jack mackerel larvae and M. norvegica may
cause the two species to be locally aggregated,

521

Table 3.-Food of jack mackerel larvae.

Size group

End of yolk-

sac s

tag©

5.0 to

7.0 to

to 4.5

6.5

10.5

Food items

No.

Copepod eggs:

Single eggs

4.3

4.4

Egg sacs

0.9

2.0

0.6

Copepod nauplii:

Calanoid

Cyclopoid

Harpacticoid

Total nauplii

12.8

3.4

0.6

Copepod adults:

Calanoid:

Calanoid spp.

Metridia sp.

Candacia sp.

Cyclopoid:

Oithona sp.

Corycaeus sp.

Oncaea sp.

Harpacticoid:

Microsetella

norvegica

130

Microsetella

rosea

Clytemnestra

rostrata

Unidentified

Total copepods

55.6

117

86.3

169

96.0

Euphausiid:

Nauplii

0.9

1.1

Calyptopi

1.0

0.6

Cladocera

0.5

Unrecognizable

crustacean remains

8.5

Peteropods

2.6

2.0

1.1

Tintinnids

13.7

0.5

Foraminifera

0.9

Total number of

food particles

117

205

176

perhaps at the surface, thereby allowing the larva
a disproportionate chance of securing individuals
of this copepod.

Jack mackerel larvae may perceive food organ-
isms by their color, since M. norvegica, and species
belonging to the genera Corycaeus and Oncaea,
are among the most brightly colored or least
transparent of copepods. Species of the two latter
genera also enter into the diet of jack mackerel
larvae. Calanoid copepods become more important
in the diet of larger larvae, perhaps because of an
increased visual acuity, or their larger mouths, or a
change in their vertical distribution. Whereas each
feeding sardine or anchovy larva normally con-
tains only one or two food particles, feeding jack
mackerel larvae usually contain more. Some in-
testines contained M. norvegica in numbers as
high as a dozen with no other observable food
items.

FISHERY BULLETIN: VOL. 74, NO. 3

Size of Food

The relatively large mouth of jack mackerel
larvae is reflected in the larger food particles
ingested (Figure 5). The preponderance in
numbers of particles at a size of 120 jum (greatest
cross-sectional dimension) is due to the apparent
selection of M. norvegica. The gape of the larva
apparently increases isometrically with increasing
length of the larva. At 3.5 mm long, it can ingest
particles up to 225 jum in cross section. Doubling its
length to 7.0 mm also doubles its ingesting
capacity to particles of about 435 jum in cross
section.

450

-1

•

— T

—

— •■

• —

•

• •

» — •

—

> — *
r —

_:-.

•*

•

— ■

•

.4 .

" *»

1 :

■f

• • 1

• V

1 ' 1
i [ 1

,.- .

1 1 1 ^j.

•-

1 _.••-. -

•

400

350

£300

8 250

O 200
iij

7r> 150

100
50

3 4 5 6 7 8 9 10 II 12
LENGTH OF LARVAE (mm)

Figure 5.- Food size of jack mackerel larvae.

COMPARISONS

Type of Food

The three species may be characterized as
primarily crustacean feeders as larvae (Table 4)
and the youngest larvae are the most eury-
phagous. Crustacean food is predominant in all
size groups of the larvae of all three species and,
furthermore, becomes increasingly so as the larvae
increase in size. Only in the smallest anchovies is
noncrustacean food an important part of the diet.

522

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

Table 4.-Crustaceans as percentage of total number of iden-
tifiable food particles. Size groups of larvae-small = end of
yolk-sac stage to 4.5 mm, middle = 5.0 to 6.5 mm, and large = 7.0
to 9.0 mm.

Size

group of larvae

Species

Small (%)

Middle (%)

Large (%)

Sardine
Anchovy
Jack mackerel

96
58
83

100
82
98

100
100
99

Size of Food

Figure 6 compares the size ranges of food
particles ingested by the three larvae. Because of
their larger mouth, jack mackerel larvae can
ingest particles about 3 times larger in diameter
than can sardine larvae of the same length. This
represents a difference in bulk of about 27 times
between maximal ingestible sizes for the two
larvae. The small anchovy can ingest particles
about 40 to 50 jum larger than the maximum-sized
particles of the sardine but does not appear to feed
as frequently on organisms near to the maximum
ingestible size as the sardine does.

Feeding Incidence and
Its Relation to Type of Intestine

Feeding incidence increases with length in the
jack mackerel but decreases with length in the
anchovy and sardine (Figure 7). The high percent-
age of jack mackerel larvae containing food may
indicate that either they are more voracious
feeders, or their digestive rate is slower, or per-
haps they are less apt to void their guts while
being caught and preserved. Feeding incidence of
larval fish appears to be associated with the
morphology of the gut. The intestine of the sar-
dine and anchovy remains long and straight with
little observable differentiation until the larva is
about 20 to 25 mm long. On the other hand, when
the jack mackerel has attained a length of about
4.25 mm, a portion of its gut forms a loop. This loop
divides the larval gut into definite functional
parts. Based on a long-range study of feeding
habits of fish in the Black Sea, Duka (1967) clas-
sified the larval gut into three types: long straight,
short straight, and looped. Duka noted also that
larvae with looped guts usually contained much
more food than larvae with straight guts. Cie-
chomski and Weiss (1974) noted that the feeding
incidence of E. anchoita larvae (0-28.0%) was much
lower than of hake larvae (63.3-94.5%) taken in the

A A---A Jock Mackerel

^ • • Sardine

o o Anchovy

34 56 7 8 910 II 12
LENGTH OF LARVAE (mm)

Figure 6.-Size range of food particles ingested by larvae of
Pacific sardine, northern anchovy, and jack mackerel.

100 r

7 9 II

LENGTH OF LARVAE (mm)

Figure 7.-Comparison of feeding incidence of Pacific sardine,
northern anchovy, and jack mackerel larvae. Values for sardine
and anchovy larvae are averages of day and night feeding
incidences. Values for jack mackerel larvae are for all samples
combined.

same plankton samples and that intestines of hake
larvae are not straight but have several folds.

523

FISHERY BULLETIN: VOL. 74, NO. 3

DISCUSSION

Significance of Feeding Incidence

For the past half century, there has been a
discussion in progress concerning the significance
of feeding incidence. Lebour (1921) called atten-
tion to the low feeding incidence of young
clupeoids and attributed this to rapid digestion of
food in the larval intestine. She was soon chal-
lenged by Hardy (1924) who, after observing
herring larvae defecating after capture, assumed
the low value to be an artifact produced by most
larvae voiding their guts. The subject has attract-
ed increasing interest recently. June and Carlson
(1971) and Kjelson et al. (1975) observed older
larvae of the menhaden, Brevoortia tyrannus,
defecating after rough handling and fixation.
Anchovy larvae have been observed defecating
rotifers and Gymnodinium while being handled in
the laboratory (John Hunter pers. commun.).
Gymnodinium is eaten by E. mordax larvae in the
laboratory (Lasker et al. 1970) and probably so in
the ocean (Lasker 1975). Rotifers and the veligers
of various species of molluscs in combination with
Gymnodinium sustain anchovy larvae in the
laboratory up to about 25 days of age (Lasker et al.
1970; Theilacker and McMaster 1971). Blaxter
(1965), however, in attempts to assess the effect of
Formalin -^ on food retention of herring larvae was
able to demonstrate that only 10% of the larvae
empty their guts due to Formalin fixation.
Detwyler and Houde (1970) studying laboratory-
grown larvae of scaled sardine Harengula pen-
sacolae, and bay anchovy, Anchoa, mitchilli, found
almost all of even the first feeding stages con-
tained food after samples of them were taken
from the plankton rich rearing tank and preserved
in 5% Formalin. Feeding incidence of clupeoid
larvae captured in plankton nets has been posi-
tively correlated with the availability of food by
Pavlovskaia (1958), Nakai et al. (1966), Burdick
(1969), Nakai et al. (1969), Bainbridge and Forsyth
(1971), and Schnack (1974). Blaxter (1965) cited the
wide variation and observed feeding incidence in
the literature concerning herring larvae. I believe
that much of the confusion has resulted from
many authors failing to consider the time of day
when larvae were caught (Figure 2) or the age of
the larvae (Figure 4). When these variables are

•'Reference to trade names does not imply- endorsement by the
National Marine Fisheries Service, NOAA.

taken into account, a series of observations of
feeding incidence can reveal valuable insights into
the tropho-dynamics of larvae. Feeding incidence
must be viewed only as an indicator of feeding
success because of the errors likely to be produced
by defecation or to the difficulty in detecting soft
bodied items such as Gymnodinium.

Comparison of the feeding incidence in four
species of Engraulis (Figure 8) shows an increase
in feeding incidence over larval lengths of 3 to 4
mm. Following this relatively high incidence at 4
mm, there is a drastic drop in this value until
lengths of about 7 or 8 mm are reached. The mean
feeding incidences for the four curves in this
length range are 7 times higher for the 4-mm than
for the 8-mm larvae. Feeding incidence remains
low but relatively constant over the length range
from 8 mm to about 14 mm at which point it begins
to increase steadily over the remainder of the
larval period. The value for the 20-mm length of E.
nngens is based on only 12 specimens and, there-
fore, is not as reliable as values for other lengths.

The available data for sardine larvae suggest
the same U-shaped curve. When the values for the
sardine (Figure 7) are compared to Figure 8 it is
seen that feeding incidence in relation to size falls
roughly between E. ringens and E. anchoita,
except that the decrease at intermediate sizes is
not as precipitous. Yamashita (1955) reported the
following feeding incidence values for larval
Sardinops melanosticta: for about 14 mm = 8%, 21
to 30 mm = 56%, and 31 to 40 mm = 81% . The
upward trend of these data is similar to those of
larger anchovy larvae; however, the values are not
comparable because the time of day of sampling
was not reported. It seems significant that the
shape of the curves of the four anchovy species
(Figure 8) are so uniform in their relation to each
other. Engraulis ringens is considerably higher
than all others (except for the value at 20 mm).
This probably is related to the rich plankton
conditions of its habitat.

Clupeoid larvae visually detect prey, approach
it, and then strike from a characteristic S-shaped
posture. Proficiency of capture increases with age
as observed in the laboratory for the larvae of
herring and pilchard (Blaxter and Staines 1971),
sardine (Schumann 1965), and anchovy (Hunter
1972). These investigators also noted that the
volume of water searched increases with larval
age. Feeding incidence should, therefore, increase
markedly with age. Why, then, does the observed
feeding incidence drop so drastically for anchovy

524

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

• - Engraulis mordax, Arthur

o - Engraulis mordax, Berner

A - Engraulis anchoita, CiechomskI

A - Engraulis japonica, Nokai et al

1956
1959
1967
1969

14 18 22 26 30 34

LENGTH OF LARVAE (mm)

Figure 8.-Feeding incidence of larvae of various species of
anchovy. Values are the average of day and night feeding (day
values are divided by two because young anchovy larvae do not
feed at night). Berner's data were recalculated to read "feeding
incidence per length of larva" rather than "percent of feeding
larvae occurring per length."

living in their natural environment? This could be
partly a result of a faster digestive rate of older
larvae as indicated for sardine larvae (Figure 2). It
also could result if either the ambient food density
decreases with time or the larval feeding activity
decreases with age. There are reasons to suspect
that both of these might occur and at the same
time.

Decrease in Food Density

Sardine and anchovy larvae may initiate their
first feeding in higher concentrations of food than
they will experience several days later. Hand and
Berner (1959) found that 74% of the food of adult
sardines, when filter feeding at night, were small
species of copepods, presumably the same species
that produce the small nauplii so important in the
diet of the sardine and anchovy larvae. Further-
more, they found that organisms in stomach
contents had a high correlation with organisms in
plankton samples taken at the same time and
place. The adult anchovy, when feeding at night, is
probably also a filter-feeding zooplanktivore al-
though it does have more omnivorous tendencies
(Loukashkin 1970), and the type of feeding, either
biting or filtering, is controlled by the size of the
food particles available (Leong and O'Connell
1969; O'Connell 1972). Both species also are selec-
tive feeders on larger organisms when visual
conditions permit. As a consequence, filter-feed-

ing adults by actively searching for rich feeding
conditions for themselves also prospect areas
suitable for their larvae. More sardine and an-
chovy larvae were shown to occur in samples where
both species were collected than in hauls where
they occurred alone (Ahlstrom 1967); he concluded
that these samples were collected near centers of
heavier spawning for both species. It would appear
that spawning adults of the two species were
seeking out the same conditions. Sardines (Ahl-
strom 1954), northern anchovies (Bolin 1936), and
Argentine and other anchovies (Ciechomski 1965)
spawn at night. Both spawning and filter feeding
take place at night; therefore, the eggs may be laid
near concentrations of suitably sized copepods
(assuming spawning and feeding occur on the
same night). However, as soon as the eggs have
been spawned, they begin to be dispersed by water
movement from each other and from organisms
they will need for food several days hence. Sardine
eggs are spawned in dense patches according to
Smith (1973), who calculated that the horizontal
mean distance between nearest neighbor eggs is
of the order of 1 to 2 cm at spawning and changes
to 15- to 20-cm mean distance for several-day-old
larvae. These larvae may experience a diminution
of their early feeding conditions as a result of
diffusion as well as of grazing by the various
predators. These ideas are presented to suggest
how a general dilution of the co-occurrence of o^^g
and plankton patches could occur in time. Lasker
(1975) has recorded how rich larval feeding condi-
tions can be destroyed overnight by a single storm.

Condition of Ocean-Caught and Laboratory-
Grown Anchovy Larvae

There are differences in physical condition of the
average ocean-caught and laboratory-grown an-
chovy larvae. These differences are probably a
result of the available food.

Ahlstrom et al.^ have presented a series of
measurements of anchovy larvae and juveniles
taken randomly from samples of the CalCOFI
program. Figure 9 is a scatter diagram of relative
body depths (body depth measured just anterior to
pectoral fin base ^standard length) calculated
from the above date. This diagram demonstrates
that relative body depths of ocean-caught anchovy

^Ahlstrom, E. H., D. Kramer and R. C. Counts. Egg and larval
development of the northern anchovy, Engraulis mordax.
Unpubl. manusc.

525

FISHERY BULLETIN: VOL. 74, NO. 3

OUJ

Jl <

ZOO

•

• ,

180
160

,■.;.- •

•

n..

140

•

120

• * ••

100
080

060
040

■■■■'■ ■■'■■■■'■■■■'■■■■'■■■■'■■■■'■■■■'■■■■'■■■■'■■■■'■■■■' 1 '■■■

0 10 20 30 40 60 60 70 80 90

LENGTH OF LARVAE AND JUVENILES (mm)

100

Figure 9.-Relative body depth of ocean caught northern
anchovy larvae and juveniles calculated from Ahlstrom et al. (see
text footnote 4).

larvae generally decrease until they are 17 to 18
mm long. Figure 10 compares relative body depth,
averaged per millimeter of length, of the above
ocean-caught anchovy larvae to that of larvae
grown in the laboratory. These are larvae grown
by Kramer and Zweifel (1970, experiment 17-11) at
17°C on a diet of wild plankton and with a feeding
incidence described as "high." At the 10-mm length
the two curves are different at the 0.05 sig-

.- .110

<
o

°- .1001-

I-
a.

Q
O
CD

UJ
>

_l
LlJ

.090 -

.080

.070

O- Ocean- cought larvae

o •- Laboratory- grown larvae

• •

•

• •

o • •

• • •

- o o o o

o o o

o o

— 1 1 i 1 1 1 1 1 L 1 \ 1 1 1 1 1 1 1

2 4 6 8 10 12 14 16 IB 20

LENGTH OF LARVAE (mm)

Figure lO.-Comparison of relative body depth of ocean caught
and laboratory grown northern anchovy larvae. Each point
represents an average of at least four larvae.

nificance level (Ntest) and they differ with greater
significance at increasing lengths.

Condition factor (weight -^ length^) for labora-
tory-grown anchovies increases throughout the
larval period as calculated from weight-length
relationships presented by Lasker et al. (1970) and
Hunter (1976).

Condition factor for ocean-caught E. anchoita
larvae as calculated from wet weight data record-
ed by Ciechomski (1965) is at its lowest value
between 15 and 20 mm.

The available data, therefore, indicate that
relative body depths and weights of well-fed
laboratory-grown anchovy larvae increase allo-
metrically throughout the larval period, whereas
these values for average ocean-caught larvae are
at a low value at some midlarval period, followed
by an increase through metamorphosis. This in-
crease is probably related to the start of transfor-
mation to the juvenile stage but may also be
accelerated by improving nutrition.

A relationship between gut thickness and feed-
ing conditions was reported for ocean-caught
larval sardine Sardinops melanosticta (Nakai
1960, 1962). The relationship of body depth to the
nutritional level of fish larvae has been recorded
for E. ja panic Hs (Honjo et al. 1959; Nakai et al.
1969) together with relative body weight for
herring larvae (Blaxter 1965, 1971). Blaxter at-
tributed the low value of body weight for ocean-
caught herring larvae to scarce plankton and to
few feeding hours in the Clyde area at the time of
sampling. The 14- to 15-mm long laboratory-
grown herring larvae when deprived of food died
at relative body weights that were higher than
those of living ocean-caught individuals. This may
be a result of the ocean-caught larvae having
survived on suboptimal rations most of their
existence whereas the laboratory-grown larvae
had ample rations until the time they were sud-
denly deprived of food. The observed decrease in
condition with size might also be an index of the
increasing ability to resist starvation. The rich
feeding conditions of successful laboratory-rear-
ing experiments probably seldom obtain in the
ocean (Lasker 1975; Hunter in press), and this may
be reflected in the condition of the average
ocean-caught larva.

The sardine larva initiates its first feeding
activities in a nutritional deficit (Lasker 1962).
This may also be indicated by the increasing
thinness of the average ocean-caught E. mordax

526

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

larva until about midway through its larval exis-
tence. Further research is required to determine if
the decline in relative physical condition indicates
a state of serious malnutrition and, if so, how far
from the well-fed state can the condition of the
individual vary without resulting in mortality. It
is also possible that laboratory-reared larvae have
abnormally large relative body depths.

Food Size, Feeding Incidence, and
Condition of Anchovy Larvae

The foregoing discussion points to three sig-
nificant trophic features of the average ocean-
caught anchovy larva. These features are:

1. A lack of increase in food particle size propor-

tional to the increase in length for larvae
larger than 4 mm (Figure 3b).

2. A steep decline in feeding incidence beginning

at 4.5 mm followed by an increase in this value
during the second half of the larval period
(Figure 8).

3. A decline in relative morphological condition at

lengths from at least 10 mm to 17 or 18 mm,
followed by an abrupt increase (Figure 9).

Feature 1 must partly reflect the size spectrum
of the available plankton. Arthur (1956) and Beers
and Stewart (1970) have shown that there are far
more food particles of the size taken by the first
feeding larvae (50-100 jum) than there are of
larger particles suitable for older larvae (i.e., 200
/xm). Sardine and jack mackerel larvae, however,
are able to secure increasingly larger food parti-
cles (Figures 1, 5). When features 1 and 2 are
considered together, it would appear that the
average oceanic anchovy larva does not sustain its
original feeding intensity.

Growth of laboratory-grown anchovy larvae
becomes asymptotic at 6 mm long when fed only
Gymnodinium and at 20 mm when fed only a
combination of Gymnodinium and rotifers. This
was noted by Hunter (in press), who concluded
that it is physically impossible for larvae to ingest
enough prey in order to grow when the prey are
below a certain size. Therefore, the decrease in
relative body depth of the ocean-caught anchovy
larva (feature 3) could be directly related to the
insufficient increase in food particle size (feature

1).
Feeding intensity of clupeoid larvae decreases

with malnutrition (Blaxter and Ehrlich 1974;
Hunter in press). If the decline in relative body
depth does denote a condition of malnutrition,
then the decrease in feeding incidence (feature 2)
is correlated with this decline, and might be the
causative factor. This might also result in larvae
spending a longer residence time at these lengths
which would introduce a bias in mortality
estimates.

It is important to keep in mind that we are
considering larvae which have grown in the ocean
and have also been caught by plankton nets. This is
the reason that the expression "ocean-caught"
rather than "ocean-grown" has been used herein.
It might be reasoned that the decline in physical
condition is a sampling artifact produced by the
plankton net catching an increasing percentage of
sick or malnourished specimens of the larger
larvae as a result of the larger healthy larvae being
more capable of dodging the net. The same rea-
soning could be applied to the decline in feeding
incidence. An examination of the physical condi-
tion of over 5,000 sardine larvae (Arthur 1956)
revealed that there is a higher percentage of
larvae in poor shape (e.g., with liver deterioration)
taken in day hauls when healthy larvae can avoid
the plankton net. Such evidence led Isaacs (1964) to
theorize that day-caught sardine and anchovy
larvae represent an approximation of the percent-
age of the population removed by natural mor-
tality. Assuming this sampling bias, however, it
then becomes difficult to explain the increase in
both relative body depth and feeding incidence of
the older larvae taken by the same sampling
methods. Burdick (1969), while examining
Hawaiian anchovy (Stolephorus purpureus) lar-
vae, observed no difference of feeding incidence or
physical condition between samples taken concur-
rently with 1-m net and a plankton purse seine.
Assuming the plankton purse seine captures all
larvae, sick or well, he concluded that there is no
bias produced by only the healthy larvae being
able to avoid the 1-m net.

The average ocean-caught anchovy is signifi-
cantly less robust at its midlarval lengths than its
laboratory counterpart, owing presumably to
differences in their respective rations. The first
feeding (4-day-old) laboratory-reared anchovy
larva spends 85% of the daytime in intermittent
swimming, 7% in feeding, and 4% at rest (Hunter
1972). Perhaps the undernourished average ocean-
caught larva, in response to the usual suboptimal

527

FISHERY BULLETIN: VOL. 74, NO. 3

food densities, conserves its dwindling energy
resources by increased resting and vv'aiting for
prey to appear within its range.

Hjort (1914, 1926) hypothesized that large-scale
mortality will result if the proper food is not
available in sufficient quantity at the "critical
period" when newly hatched fish larvae require
their first feeding, and that the numerical
strength of a year class, therefore, might be
determined at this time. The increasingly thin
shape of young ocean-caught anchovy larvae
suggests that feeding problems may exist for
sometime into the larval period. Saville (1971)
proposed that a "critical stage" might occur at any
stage between hatching and metamorphosis and
that the detection of same would allow one to
specify the earliest stage at which reliable indices
of year-class strength could be determined. The
end of the decline in relative body depth of the
average ocean-caught larva might mark the point
in the larva's development when the danger of
starvation has diminished and perhaps, as sug-
gested by Saville, is the earliest stage at which
estimates of recruitment might be made.

ACKNOWLEDGMENTS

I express my appreciation to Reuben Lasker,
John R. Hunter, and Martin W. Johnson for their
helpful comments and criticisms during the
preparation of the manuscript. Elbert H. Ahlstrom
and Paul E. Smith also furnished valuable sug-
gestions and information. I am indebted to the
Southwest Fisheries Center La Jolla Laboratory,
National Marine Fisheries Service, NOAA, for the
use of facilities and especially to Reuben Lasker
who provided working space and the use of his
personal library. I also convey my gratitude to
Martin W. Johnson and Carl L. Hubbs for their
guidance in the research leading to the completion
of the thesis upon which this paper is largely
based.

LITERATURE CITED

Ahlstrom, E. H.

1954. Distribution and abundance of egg and larval popula-
tions of the Pacific sardine. U.S. Fish Wildl. Serv., Fish.
Bull. 56:83-140.
1967. Co-occurrences of sardine and anchovy larvae in the
California Current region off California and Baja Califor-
nia. Calif. Coop. Oceanic Fish. Invest. Rep. 11:117-135.
Arthur, D. K.

1956. The particulate food and the food resources of the

larvae of three pelagic fishes, especially the Pacific sardine,
Sardinops caerulea (Girard). Ph.D. Thesis, Univ. Califor-
nia, Scripps Inst. Oceanogr., La Jolla, 231 p.

Bainbridge, v., and D. C. T. Forsyth.

1971. The feeding of herring larvae in the Clyde. Rapp.
P.-V. Reun. Cons. Int. Explor. Mer 160:104-113.

Beers, J. R., and G. L. Stewart.

1970. Numerical abundance and estimated biomass of
microzooplankton. In J. D. H. Strickland (editor). The
ecology of the plankton off La Jolla, California, in the
period April through September, 1967, p. 67-87. Bull.
Scripps Inst. Oceanogr. Univ. Calif. 17.

Berner, L., Jr.

1959. The food of the larvae of the northern anchovy
Engraulis mordax. Inter-Am. Trop. Tuna Comm., Bull.
4:3-22.
Blaxter, J. H. S.

1965. The feeding of herring larvae and their ecology in
relation to feeding. Calif. Coop. Oceanic Fish. Invest. Rep.
10:79-88.

1971. Feeding and condition of Clyde herring lar\ae. Rapp.
P-V. Reun. Cons. Int. Explor. Mer 160:128-136.

Blaxter, J. H. S., and K. F. Ehrlich.

1974. Changes in behaviour during star\'ation of herring
and plaice larvae. In J. H. S. Blaxter (editor), The early
life history of fish, p. 575-588. Springer-Verlag, Berl.
Blaxter, J. H. S., and M. E. Staines.

1971. Food searching potential in marine fish larvae. In D.
J. Crisp (editor). Fourth European Marine Biology Sym-
posium, p. 467-485. Cambridge Univ. Press.
BOLIN, R. L.

1936. Embryonic and early larval stages of the California
anchovy, Engraulis mordax Girard. Calif Fish Game
22:314-321.
Bowers, A. B., and D. I. Williamson.

1951. Food of larval and early post-larval stages of au-
tumn-spawned herring in Manx waters. Annu. Rep. Mar.
Biol. Stn. Port Erin 63:17-26.
BURDICK, J. E.

1969. The feeding habits of neju (Hawiian anchovy) lar-
vae. M.S. Thesis, Univ. Hawaii, 54 p.
ClECHOMSKI, J. D. de.

1965. Observaciones sobre la reproduccion, desarrollo em-
brionario y larval de la anchoi'ta argentina {Engraulis
anchoita). Bol. Inst. Biol. Mar. Mar del Plata 9, 29 p.
1967. Investigations of food and feeding habits of larvae
and juveniles of the Argentine anchovy Engraulis an-
choita. Calif. Coop. Oceanic Fish. Invest. Rep. 11:72-81.

ClECHOMSKI, J. D. DE, AND G. WeISS.

1974. Estudios sobre la alimentacion de larvas de la merluza,
Merluccius merluccius hubbsi y de la anchoita, Engraulis
anchoita en la mar. Physis Rev. Asoc. Argent. Cienc. Nat.
33:199-208.

Detwyler, R., and E. D. Houde.

DUKA, L. A.

1967. Struktura I dinamika soobshchestv (Adaptive fea-
tures in the feeding of pelagic larvae of fish living in
different ecological conditions). Akad. Nauk Ukr. SSR, p.
136-143. (Translated by Natl. Lending Libr. Transl. Pro-
gram RTS 5162.)

528

ARTHUR: FOOD AND FEEDING OF LARVAL FISHES

Hand, C. H., and L. Berner, Jr.

1959. Food of the Pacific sardine (Sarrf/Mopscaerw/ea). U.S.
Fish Wild). Serv., Fish. Bui. 60:175-184.
Hardy, A. C.

1924. The herring in relation to its animate environment.
Fish. Invest., Minist. Agric. Fish Food. (G.B.), Ser. II, 7(3),
53 p.
Hjort, J.

1914. Fluctuations in the great fisheries of northern Europe
viewed in the light of biological research. Rapp. P.-V.
Reun. Cons. Perm. Int. Explor. Mer 20:1-228.
1926. Fluctuations in the year classes of important food
fishes. J. Cons. 1:5-38.

HONJO, K., T. KiTACHI, AND H. SUZUKI.

1959. On the food distribution and survival of post larval
iwashi-1 - Distribution of food organisms, the food of the
anchovy and ecologically related species along the south-
western Pacific coast of Honshu, Sept. -Nov. 1958. Reports
on the major coastal fish investigations, and the inves-
tigations for forecasting of oceanographic conditions and
fisheries (preliminary report), February 1959. (Engl,
transl. by S. Hayashi.)
Hunter, J. R.

1972. Swimming and feeding behavior of larval anchovy,

Engraulis mordax. Fish. Bull., U.S. 70:821-838.
1976. Culture and growth of northern anchovy, Engraulis

mordax, larvae. Fish. Bull., U.S. 74:81-88.
In press. Behavior and survival of northern anchovy,
Engraulis mordax, larvae. Symposium on Fishery Sci-
ence. Ciencias Pesquera, Spec. Suppl.
ISSACS, J. D.

1964. Night-caught and day-caught larvae of the California
sardine. Science (Wash., D.C.) 144:1132-1133.
June, F. C, and F. T. Carlson.

1971. Food of young Atlantic menhaden, Breroorfia tyran-
nus, in relation to metamorphosis. Fish. Bull. U.S.
68:493-512.
Kjelson, M. a., D. S. Peters, G. W. Thayer, and G. N. Johnson.
1975. The general feeding ecology of postlarval fishes in the
Newport River estuary. Fish Bull., U.S. 73:137-144.
Kramer, D., and J. R. Zweifel.

1970. Growth of anchovy larvae (Engraulis mordax Girard)
in the laboratory as influenced by temperature. Calif.
Coop. Oceanic Fish. Invest. Rep. 14:84-87.
Lasker, R.

1962. Eflnciency and rate of yolk utilization by developing
embryos and larvae of the Pacific sardine Sardinops
caerulea (Girard). J. Fish. Res. Board Can. 19:867-875.
1975. Field criteria for survival of anchovy larvae: The
relation between inshore chlorophyll ma.ximum layers and
successful first feeding. Fish. Bull"., U.S. 73:4.53-462.

Lasker, R., H. M. Feder, G. H. Theilacker, and R. C. May.

1970. Feeding, growth, and survival of Engraulis mordax
larvae reared in the laboratory. Mar. Biol. (Berl.)
5:345-353.
Lebour, M. V.

1921. The food of young clupeoids. J. Mar. Biol. Assoc. U.K.
12:458-467.
Leong, R. J. H., and C. p. O'Connell.

1969. A laboratory study of particulate and filter feeding of
the northern anchovy (Engraulis mordax). Fish. Res.
Board Can. 26:557-582.

Loukashkin, a. S.

1970. On the diet and feeding behavior of the northern

anchovy, Engraulis mordax (Girard). Proc. Calif. Acad.
Sci., Ser. 4, 37:419-458.

May,R. C.

1974. Larval mortality in marine fishes and the critical
period concept. In J. H. S. Blaxter (editor). The early life
history of fish, p. 3-19. Springer- Verlag, Berl.

Nakai, Z.

1960 Changes in the population and catch of the Far East
sardine area. Proc. World Sci. Meet. Biol. Sardines,
Related Species. 3:807-853.
1962. Studies relevant to mechanisms underlying the fluc-
tuations in catch of the Japanese sardine, Sardinops
melanosticta (Temminck & Schlegel). Jap. J. Icthyol.
9:1-113.

Nakai, Z., H. Koji, H. Shigemasa, K. Takashi, and S. Hideya.

1966. Further examples of Hjorts hunger theory. (Abstr.)
Second Int. Oceanogr. Congr., p. 263.

Nakai, Z., M. Kosaka, M. Ogura, C. Hayashida, and H.
Shimozono.

1969. Feeding habit, and depth of body and diameter of
digestive tract of skirasu, in relation with nutritious
condition. [In Jap., Engl, abstr..] J. Coll. Mar. Sci. Tech.
Tokai Univ. 3:23-34.

O'Connell, C. P.

1972. The interrelation of biting and filtering in the feeding
activity of the northern anchovy (Engraulis mordax). J.
Fish Res. Board Can. 29:285-293.

Pavlovskaia, R. M.

1958. The survival of anchovy larvae in the northwest and
certain other regions of the Black Sea in 1954-55 in relation
to food conditions. Doklady Akad. Nauk SSSR. Biol. Sci.
Sect. (Transl.) 120:346-348. '

Reid, J. L., Jr.

1967. Oceanic environments of the genus Engraulis around
the world. Calif. Coop. Oceanic Fish. Invest. Rep. 11:29-33.

R. DE Mendiola, B.

1974. Food of the larval anchoveta Engraulis ringens J. In
J. H. S. Blaxter (editor). The early life history of fish, p.
277-285. Springer-Verlag, Berl.
Saville, a.

1971. The larval stage. In A. Saville (editor). Symposium on
the biology of early stages and recruitment mechanisms of
herring, p. 52-55. Rapp. P-V. Reun. Cons. Int. Explor. Mer
160.

SCHNACK, D.

1974. On the biology of herring larvae in the Schlei Fjord,
western Baltic. Rapp. P.-V. Reun. Cons. Int. Explor. Mer
166:114-123.
Schumann, G. 0.

1965. Some aspects of behavior in clupeid larvae. Calif.
Coop. Oceanic Fish. Invest. Rep. 10:71-78.

Sinyukova, V. I.

1964. The feeding of Black Sea horsemackerel larvae. Tr.
Sebastopol Biol. Stn. Acad. Sci., USSR. 302-325. (Translat-
ed by Lillian Vlymen.)

Smith, P. E.

1973. The mortality and dispersal of sardine eggs and
larvae. Rapp. P.-V. Reun. Cons. Int. Explor. Mer.
164:282-292.

Theilacker, G. H., and M. F. McMaster.

1971. Mass culture of the rotifer Brachionus plicatilis and
its evaluation as a food for larval anchovies. Mar. Biol.
(Berl.) 10:183-188.

529

FISHERY BULLETIN: VOL, 74. NO. 3

Yamashita, H. Yasunaga, Y.

1955. The feeding habit of sardine, Sardinia melanosticta, 1971. Studies on the feeding habit and growth of the plaice,

in the waters adjacent of Kyushu, with reference to its Paralicktkys olivaceus, in the larval stage. [In Jap., Engl,

growth. [In Jap., Engl. Abstr.] Bull. Jap. Soc. Sci. Fish. Abstr.] Bull. Tokai Reg. Fish. Res. Lab. 68:31-43.
21:471-475.

530

OBSERVATIONS ON THE COMMERCIAL FISHERY AND
REPRODUCTIVE BIOLOGY OF THE TOTOABA, CYNOSCION
MACDONALDI, IN THE NORTHERN GULF OF CALIFORNIA

Christine A. Flanagan and John R. Hendrickson'

ABSTRACT

Information gathered from fishers and records of the failing totoaba, Cynoscio7i macdonaUli,
commercial fishery demonstrate the ability of the three principal ports to fully exploit the dwindling
population during its annual breeding migration to the mouth of the Colorado River. Gonadal
maturation, daily catch, and capture incidence data document the timing and route of the migration,
provide evidence for a tendency toward unisexual schooling in its early phase, and point to the
possibility that totoaba may form large aggregations before spawning is initiated. A trend toward
reduction in the length of the migratory and spawning period, from 5 or 6 mo in 1965 to 1 mo in 1972 is
documented with data from the port of Golfo de Santa Clara. In surveys of the hypothesized nursery
area, 28 juvenile totoaba (6-12 cm standard length) were collected at 4 of 14 sampling sites. The four
collection sites were commonly characterized only by depth (<1 m) and substrate type (soft clay-silt
sediments). Three hypothesized causes of the decline of this commercial fishery are examined by
statistical analyses of Colorado River flow and annual totoaba catch data: overfishing, loss of spawning
grounds, and loss of nursery grounds. Overfishing was found to be the most likely cause of the decline.
Recent trends of catch data among the principal commercial fleets, and evidence that regulatory
measures may have resulted in temporary recovery of totoaba production, provide further support for
the overfishing hypothesis. The journey of the migrant population along a known route and its
concentration into a predictable small area, its hypothesized requirement for dense aggregations prior
to spawning, and the added mortality of juveniles taken by shrimp trawls in the near-delta waters are
important points of vulnerability that render this endemic species particularly susceptible to fishing
pressure. The possibility of the extinction of Cynoscion macdonaldi, without continuation of the newly
decreed prohibition of fishing, is reiterated.

The totoaba,- Cynoscion macdonaldi Gilbert 1891,
is the largest species of the family Sciaenidae,
with maximum reported lengths of almost 2 m
(Berdegue 1956) and weights exceeding 135 kg
(Cannon 1966); the larger females in present-day
commercial catches approximate 1.5 m and 35 kg
(Arvizu and Chavez 1972). The species is endemic
to the Gulf of California, where it used to support a
fishing industry and popular sport fishery based on
its annual spring breeding migration to the shal-
low, formerly brackish waters of the Colorado
River Delta region at the extreme northern end of
the gulf. The major portion of the catch was
exported from Mexico to the United States (prin-
cipally San Diego) and brought a high price per
pound under the influence of apparently unlimited
demand. Presently an indefinite closed season on

'Department of Ecology and Evolutionary Biology, University
of Arizona, Tucson, AZ 85721.

-The common name is often spelled "totuava" by writers from
the United States for no known reason. The spelling used here is
that preferred and used by Mexicans; it should become the
established spelling.

the totoaba, declared by the Government of Mex-
ico on 2 August 1975, prohibits all capture of this
species by both commercial and sport fisheries (H.
Chavez, pers. commun.).

Although the species has been heavily exploited,
its life history, population dynamics, and general
ecology are poorly known. Species accounts are
given in Jordan and Evermann (1898, 1902), Jor-
dan et al. (1930), Gabrielson and Lamonte (1954),
and Lanham (1962). The totoaba was included in
accounts of commercial sciaenid species by Croker
(1932) and Fitch (1949). Aside from these refer-
ences and others cited here, little has been pub-
lished on the totoaba; remaining incidental refer-
ences may be found in Arvizu and Chavez (1972),
the most recent summary of all available infor-
mation on this species. Although notes on the
ecology of the totoaba were first published in 1916
by Jordan, most of the presently accepted life
history information is based on fisher's lore. These
beliefs were first documented by Berdegue in his
1955 study of the fishery in which he also examined
scale annuli series and published the only derived

Manuscript accepted Januar>' 1976.
FISHERY BULLETIN: VOL. 74, NO. 3

531

FISHERY BULLETIN: VOL. 74, NO. 3

growth estimates for this species. His work con-
cluded with a warning that the totoaba is a
declining species, in danger of extinction from a
combination of overfishing and the disappearance
of brackish water spawning grounds due to diver-
sion of Colorado River waters for agricultural and
other purposes. Cause (1969) and Sotomayor (1970)
later echoed this view.

In this paper we present a short history of the
commercial fishery and report new information on
totoaba life history. We summarize what is known
about the ecology of the species and speculate on
consequences of the present small population size
and the intense fishing effort to which the fish have
been exposed. We discuss the three most probable
causes for the decline in the fishery: degradation of
spawning grounds, degradation of nursery
grounds, and overfishing. We examine Colorado
River flow data and annual catch data in the light
of these hypotheses, and discuss our results. In
conclusion, we draw together all these elements in
an attempt to assess the present and future status
of this commercial population.

HISTORY OF THE FISHERY

Until about 1920, commercial exploitation of the
totoaba was confined to export of dried air blad-
ders to the Orient as an ingredient of a gourmet
soup (Chute 1930). Craig (1926) reported the first
export of totoaba flesh to the United States. In
these early, developing years, the totoaba fishery
was directly responsible for the establishment of
three northern gulf fishing villages: Golfo de
Santa Clara and Puerto Penasco in the State of
Sonora, Mexico, and San Felipe in the State of
Baja California Norte (Berdegue 1955). Analysis
of registered catches by all Mexican ports for the
1966-70 period shows that these three ports
produced from 94.9 to 97.7% of the total catch (H.
Chavez, pers. commun.).

From 1929 (when Mexican Government statis-
tics were first collected) onward, the fishery re-
sponded to a growing U.S. market by developing
transportation and refrigeration capabilities and
by improving fishing gear and boat facilities.
Annual yield began to increase rapidly in 1934 and
the catch peaked at 2,261 metric tons^ in 1942

</l J 000 -

■ I ' '

19 10

' ■ I ■

YEAR

Figure 1.- Yield of commercial totoaba fishery, northern Gulf of
California for the 1929-75 period. Figure modified from Arvizu
and Chavez (1972). Data for 1971-75 were obtained from H.
Chdvez (pers. commun.).

(Figure 1). After 1942, despite intensified fishing
effort and increased gear efficiency, the annual
yield exhibited erratic fluctuation to the all-time
minimum of approximately 58 metric tons in 1975
(H. Chavez, pers. commun.^).

Fishing methods evolved from spearing out of
dugout canoes and primitive handlining in the
early years, through dynamiting and primitive gill
netting, to the use of eflficient nylon gill nets. The
usual modern net has a stretched mesh size of
approximately 25 cm and measures 100-200 x 4-5
m. Gill nets were managed from diesel-powered
shrimp trawlers (12-18 m, some temporarily
diverted from shrimping during prime season
totoaba fishing), and from 4.5- to 7.5-m wooden or
fiber glass "pangas" (launches) fitted with out-
board motors. The activities of commercial fishers
have been largely limited to the prime breeding
season (January-March) when the spawning
adults are in the shallow waters of the extreme
northern gulf. Prior to the 1975 total protection of
totoaba, the prime fishing season ended with the
advent of an oflficial closed season, 1 April-15 May
(Arvizu and Chavez 1972), a protective measure
enacted by the Mexican Government in about 1955
(Berdegue 1955).^ At the same time, a sanctuary
was designated at the mouth of the Colorado

^We follow the example of Arvizu and Chavez (1972) in giving
yields as weights of cleaned fish lacking heads and viscera unless
specifically designated otherwise. To convert to whole weights,
multiply by 1.1 (H. Chavez, pers. commun.).

^The 1975 yield reported here is based on catch from principal
ports for the prime season only (through the month of March).
The final figures may be as much as 10% higher.

^According to Berdegue (1955), before 1955 there was a closed
season extending from 20 March to 1 May; the prohibited period
was changed to the later dates because active spawning was
observed after 1 May. In 1969 and 1970 the beginning of the
closed season was delayed 15 days in response to the fishers'
petitions when breeding schools had not appeared by the end of
March (H. Chavez, pers. commun.).

532

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

River. All fishing was prohibited north of an
imaginary line extending from Bahi'a Ometepec
on the Baja California coast to the mouth of the
Rib Santa Clara on the Sonora coast.

In addition to the standard commercial fishery,
the Seri Indians of the Bahi'a Kino and Punta
Chueca areas of Sonora were alleged to capture
totoaba in coastal waters during the fall and
winter, but we have been unable to confirm this by
personal observation. Further pressure was ex-
erted on the stocks by an enthusiastic sport
fishery, based largely on the Baja California side
of the northern gulf, which took unknown
numbers of breeding adults during the prime
season. In recent years when diminution of the
stocks caused the success rate to drop, sport fishing
virtually disappeared. At the peak of the sport
fishery, large numbers of immature fish resident
in the upper gulf waters were also reportedly
taken, usually unrecognized as totoaba. For a time,
a deepwater handline commercial fishery and
accompanying sport fishery continued out of San
Felipe during the summer after the adult fish had
left the spawning grounds, but this activity also
declined in recent years. Craig (1926), Chute
(1928), and Berdegue (1955) provided further
information on the history of the fishery and
contain most of the documented information on
the sport fishery.

METHODS AND MATERIALS

The junior author began field studies on the
species in 1970 with the primary objective of
gathering life history information for conserva-
tion purposes. The results reported here derive
primarily from data collected by the senior author
during three cruises aboard commercial fishing
vessels from Puerto Peiiasco in March and April
1972. Fishing patterns during these cruises in-
cluded most of the Gulf of California north of lat.
31°N; with few exceptions, the locations were
selected by the fishing captain.

The data were gathered by direct observation of
catch and, in a few cases, by reports from fishers
on "companion" vessels (as many as five other
boats in the cooperating group, in one instance).
During 22-24 March 1972 we also collected data
from the panga fleet at Golfo de Santa Clara as the
catch was landed and cleaned at the port. In both
circumstances, our data consisted of information
on location, time and size of catch, number of
operational net hours, time and state of tide.

sexual composition of the catch, and reproductive
state of the individual.

All fish examined by us were breeding adults.
They were classified according to three mutually
exclusive categories of gonadal development: If
not running eggs or milt at the time of capture (or
within 24 h of capture in the case of several
individuals kept alive for a period of hours), they
were classified as "unripe"; if milt or hydrated
eggs ("applesauce" color and texture) could be
expressed with light pressure, they were classified
as "ripe"; females with flaccid ovaries and running
ripe males taken in the same catch with such
females were classified as "spent."

Effort data are reported as the number of
operational net hours rather than total time
(man-hours or boat-hours) spent fishing because
many of the large boats "hunt" for schools suitable
for encircling with their nets during the day and
then set their gill nets in the usual manner to fish
overnight. We believe that recent daytime hunt-
ing for schools to encircle was practiced more in
memory of times past than as a practical matter of
probability. In approximately 50 days aboard such
vessels, we have never seen a school located,
although one heard of such catches each season.
The method persisted because, if successful, it can
yield very high tonnage. The larger, diesel-
powered trawlers with ice-filled holds frequently
stayed at sea for more than a week and commonly
traveled considerable distances back to their home
ports to land the catch. This is in marked contrast
to the methods of the fishers of Golfo de Santa
Clara, who fished primarily from pangas and who
customarily inspected their nets each day by
passing the net over the boat, leaving the weight-
ed ends in place. Such nets "fished" continually,
except for occasions when they were taken up to be
moved to alternate spots. Lacking storage and
refrigeration facilities, the pangas had to return
to port each day with their catch from one or two
gill nets. Catch in kilograms was recorded by a
Mexican government fisheries inspector for each
panga, each day. Although we attempted to cal-
culate catch per unit effort, we were unable to
resolve its heterogeneous nature. Here we present
only our analysis of effort from the panga fishery
of Golfo de Santa Clara.

In late May and early June 1972, a number of
Sonora and Baja California sites around the pe-
rimeter of the extreme northern gulf were sur-
veyed for juvenile totoaba, using both commercial
trawl nets and beach seines. Many of these sites

533

were revisited in June 1973. Observations were
made of water temperature, salinity, turbidity,
and substrate character; associated faunas at each
site were sampled. A few juveniles were trans-
ported alive back to Tucson, Ariz., and maintained
there for about 80 days. Information on distribu-
tion and habitat of the juveniles is presented here;
notes on behavior of the juveniles in captivity will
be reported elsewhere (C. A. Flanagan in prep.).

In our discussion of the hypotheses for the
decline of the totoaba fishery, we present statistics
of Colorado River flow and annual totoaba yield.
The annual yield data are those already presented
(Figure 1). For flow, we have attempted to es-
timate the amount of water delivered to Mexico in
the main river channel at the southerly interna-
tional boundary on the assumption that it will bear
some regular relationship to the volume of fresh
water entering the Gulf of California. This as-
sumption becomes tenuous with the development
of lowland agriculture in Me.xico and with sig-
nificant groundwater pumping in the United
States, both in evidence since about 1960. Suitable
effort data for the totoaba fishery are unavailable
but we have assumed that, following the peak
catch in 1942, effort was constant or increasing.
This assumption is probably warranted given the
demand and high price paid for totoaba flesh. The
limitations imposed by our assumptions are that
no catch datum before 1942 and no flow datum
after 1960 may be considered in these analyses.

FISHERY BULLETIN; VOL. 74, NO. 3

RESULTS AND DISCUSSION

Breeding Migration

The fishers believe that the annual migration of
totoaba is prompted by the urge to reproduce and
is guided by the search for a suitable estuarine
spawning environment. According to their beliefs
the breeding population, seeking areas of reduced
salinity, leaves deep water in the mid-gulf and
follows the Sonora coastline northward; eventually
the schools reach the mouth of the Colorado River,
where they spawn. Following spawning, the to-
toaba supposedly seek out the clearer, deeper
waters to which they are more accustomed and
follow the Baja California coastline on their return
migration southward. These beliefs are based
upon commercial catch experience dating back to
the late 1920's.

Localities and dates of capture observed by the
senior author in 1972 (Figure 2 and Table 1) appear
to document a pattern consonant with the above
hypothesis, as do observations by the junior author
in earlier years. The regular port statistics also
implicitly support the hypothesis, with catches
each year reported chronologically first by Puerto
Penasco, then by Golfo de Santa Clara, and last by
San Felipe fleets. The data in Figure 2 represent
but a small fraction of the total 1972 fishing effort,
however, and in the most conservative interpreta-
tion demonstrate only that experienced fishers

114-

. BREiDING PtESERVE

: OBSERVED CA1CH AREAS

w Golfo de Santa Clara

TT,-

Figure 2.— Locations and dates of
observed commercial catches of to-
toaba during the 1972 prime fishing
season. Catch information in terms of
tonnage by day, boat, and area were
also obtained from Fisheries Inspec-
tors. These latter data are reflected in
the early capture date of 12 February
and the extended capture period of
11-29 March in areas I and III, re-
spectively. Chart shows Gulf of
California north of lat. 31° N (see
locater in upper right-hand corner).

534

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

Table L-Totoaba captures observed in 1972 (see Figure 2 for
areas and timing). The figures shown here are personal observa-
tions of the senior author.

of observed

No.

of successful

Total no. of

Area

net sets

totoaba caught

III

(no data)

'149

(')

'These data are displayed in Figure 5. Although catch data are
available for area III only in terms of tonnage, not head count of
fish taken, use of the 35-kg average per fish would give a con-
servative estimate of at least 2,500 individual fish taken In area
III during 1972.

know where and when to find fish. Ideally, Figure 2
should reflect the results of an even pattern of
standard net sets through the February-June
period.

To our knowledge, no one has investigated the
salinity preferences or tolerances of spawning
adults, but this raises the question of totoaba
spawning sites in estuarine areas of other major
gulf rivers. Spawning totoaba have never been
reported from locations other than the Colorado
River mouth. While further investigation is clear-
ly warranted, at present we accept the fisher's
hypothesis as an adequate predictor of population
migratory patterns.

Spawning Concentration

Because the annual breeding migration results
in a high density of fish within a limited area, it
has become the single most important aspect of
the fishery: total prime season catch is a function
of the number of fish arriving in the spawning
area before 31 March in an average year. The
appearance of migrant schools of totoaba in shal-
low coastal waters, as signaled by catches from
exploratory boats which have ventured out in
anticipation of their arrival, usually occurs in

mid-February, but may take place as early as
December or as late as the end of March.

Three references exist in the literature regard-
ing the spawning period. Nakashima (in Jordan
1916) said that the main spawning period was in
early May, while Berdegue (1955) reported the
reproductive season as extending from the end of
February or early March until early June. Obser-
vations by D. A. Thomson (1969) and the junior
author over the last four seasons indicate peak
spawning as late as April and early May but, more
commonly, in mid- to late March. Historical data
and existing statistics confirm the fisher's claims
that the period of concentrated catch (which
apparently coincides with peak spawning) has
become progressively abbreviated during the past
20 yr. The monthly catch data for Golfo de Santa
Clara for the 1964-72 period show a clear reduction
in length of season from 5 or 6 mo to an ab-
breviated period in March-April at present
(Figure 3). We believe (see below) that the catch of
the Golfo de Santa Clara fleet is a good reflection
of spawning activity and suggest that a pattern of
repeated spawnings formerly extending from
January-February to May and June has collapsed
to a single event which coincides with the old
temporal mode. A small remnant population might
be expected to react more uniformly to environ-
mental cues than would a large one, a factor
leading to progressively shorter migratory and
spawning periods. This is consistent with our
observation of breeding population residence time
of only 18 days on the spawning grounds in 1972
(from 11 March to 29 March, see discussion below).

A limited amount of qualitative data on gonadal
maturation, collected during the 1972 prime
fishing season (Figure 4), indicates that males
ripen before females and retain spawning readi-
ness for longer periods of time-a common occur-
rence among fishes. It also provides evidence for a

400-1
350-

300

g 250 -

- 200 -

J. 150 -

o lOO -

sAl

rn^

J F M i M
1964

1965

1966

W A M J
1967

w a w .

1968

F M A M .
1969

M A M J
1970

F M A W J
1971

W A M J
1972

Figure 3.- Monthly yield in metric
tons of totoaba, port of Golfo de Santa
Clara. Data for 1966-70 from Arvizu
and Chavez (1972); H. Chavez (pers.
commun.) supplied data for 1964-65
and 1971. The 1972 catch data were
obtained from F. Aguilera, Fisheries
Inspector (15 additional metric tons
recorded in 1972 are not shown
because month of capture was
uncertain).

535

FISHERY BULLETIN: VOL. 74, NO. 3

SPENT

■

RIPE

■

UNRIPE

■

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
MARCH 1972

Figure 4.-Degree of gonadal maturation observed on specific
days during the 1972 prime fishing season. See text for explana-
tion of maturation categories. Each symbol represents one or
more individual fish. Open squares represent females; solid
squares represent males.

tendency toward sexually segregated schooling, at
least in the case of male fish. It will be noted that
all the records for the early portion of the period
portrayed in Figure 4 show a single sex per catch,
while the records for the later portion of the period
show both sexes in all but one instance. It must be
clearly understood that some of the catch records
portrayed in Figure 4 represent single fish, mak-
ing those data points meaningless in this context,
but a majority of the data points represent multi-
ple individuals. This apparent sexual separation of
prespawning schools conforms with general ob-
servations by Hendrickson in years before 1972
and with the caption for figure 84 in Chute's (1928)
paper describing the earlier hook-and-line fishery:
"Practically all of the fish in this picture were
males . . . ." (the figure, depicting the butchering
process, shows about 15 large fish caught by three
men in 3 h).

Success of the Golfo de Santa Clara panga
fishery is related to the size of the migrant totoaba
population, the length of the period in residence on
the fishing grounds, and population behavioral
patterns. Because the fishing grounds are identical
with, near to, or in the path to the spawning
grounds, analysis of the panga fishery catch sta-
tistics can yield valuable insight into the breeding
biology of this species. We have used "capture
incidence" as a measure of fishing success, em-
ployed here to give a quantitative indicator of the
presence of breeding adults on or near the sus-
pected spawning grounds (Figure 5). One capture
incident is defined as the catch of at least one
totoaba per panga per day; the daily total reflects
the number of individually successful net sets. We
assume that: 1) fishing effort is constant after a
given date within the prime season and 2) fishers
individually and collectively fish in the same area
each day throughout the period. These assump-
tions are in keeping with the nature of the Golfo
de Santa Clara fishery. This village waited in

MARCH 1972

Figure 5.-Golfo de Santa Clara catch and capture incidence
plotted against days of March for 1972 prime fishing season. See
text for explanation of capture incidence unit. Between 15 and 22
March the number of individually successful nets remained
comparatively constant, despite the peak catch on 22 March.
Official statistics indicate that 15 metric tons in addition to the
total of 71 metric tons shown in Figure 3 were recorded for this
port in 1972 (H. Chavez, pers. commun.). These additional data
cannot be traced to daily catch for inclusion in this figure.

readiness each year for the arrival of the migrant
population and, within a few days of the first
catches by exploratory nets, virtually all available
gill nets were deployed for fishing of totoaba.
Despite daily success or failure, fishing effort
continued at this level until the season closed on 1
April. Most of the panga fishers worked a definable
area of the delta where "canals" (extensions of
Colorado River channels) deep enough to accom-
modate the large totoaba gill nets are separated by
shallow mud bars (see area III in Figure 2).

In 1972 the first catch off Golfo de Santa Clara
occurred on 11 March and was followed by a period
of increasing catch and capture incidence due to
increasing effort until 15 March (Figure 5). During
the March 15-22 period, capture incidence was
relatively constant and high. During this same
period, catch varied somewhat erratically and
peaked on 22 March, after which both catch and
capture incidence fell off drastically despite no
reduction in fishing effort. The 22 March catch
amounted to 27% of Golfo de Santa Clara's yield
for that year and represented 9% of the total
recorded yield from all ports for 1972. The average
Golfo de Santa Clara net must have contained over
twice as many fish on 22 March as on 21 March and
3-5 times as many as on other "good" days in the
prime season.

What factors in totoaba reproductive biology
might explain these results? Catch per net may be
considered an index of migrant arrivals if we

536

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

suppose that, as breeding adults reach the north-
ern end of the gulf, they immediately move up into
the channels at the mouth of the Colorado River.
The arrival of the largest population segment
would then be indicated by the peak catch. Alter-
natively, the peak catch may have signaled a peak
of spawning activity by an already-resident
breeding population, becoming more vulnerable to
the nets by virtue of spatial concentration and/or
behavior. While the data do not allow firm conclu-
sions, we favor the second alternative.

A period of behavioral stimulation in schools to
induce the spawning act is suggested by the fact
that enormous numbers of individuals allegedly
used to gather in this relatively small area to
spawn (Jordan 1916; Berdegue 1955). Although the
population has been drastically reduced, the fish
apparently continue this habit. If the release of
reproductive behavior patterns depends upon
mutual stimulation within large aggregations
(consistent with their sound-producing air blad-
der; see Breder and Rosen 1966), the present small
population might be experiencing some break-
down in the behavioral sequence with consequent
lowered reproductive success.

Although the significant yields of 21 and 22
March may indicate a peak in spawning activity,
this does not preclude the possibility that other fish
were later in arrival and that spawning also
occurred in April (during the closed season). The
Golfo de Santa Clara fleet's near-failure to catch
fish during the 25-31 March period, and our failure
to find adult fish during the April cruise lend doubt
to this possibility, but the existence of more than
one breeding population should not be ruled out.

Juvenile Totoaba Distribution,
Habitat, and Diet

The microhabitat and residence time of juvenile
totoaba on the nursery grounds are largely un-
known. Berdegue (1955) reported that juveniles
remain in the shallow waters near the Colorado
River mouth until they begin a southward migra-
tion to join the parent population. The Colorado
River Delta is heavily exploited by the shrimp
fishery during parts of the year (effort was ob-
served to be especially intense during April, May,
and June), and Berdegue first called attention to
the increased mortality of juvenile totoaba due to
shrimp trawling activity.

In our experience, the juveniles captured in
shrimp trawls are individuals ranging in length

from about 15 cm to about 45 cm. The holotype in
the U.S. National Museum is approximately 25 cm
long and was taken in 20 fathoms of water (Gilbert
1891). To our knowledge, the first collection of
really small juveniles (6-12 cm size range) which
were positively identified as totoaba was made in
1970 near San Felipe, B.C., and described by
Chavez (1973). We surveyed probable northern
gulf sites for the presence of such small juveniles
during May and June 1972 and 1973 (Figure 6).

Substrate and depth appear to be more impor-
tant than either temperature or salinity in char-
acterizing the habitat of the captured juveniles.
For all sites, surface water temperatures ranged
from 25° to 29°C and salinities were recorded
between 35 and 407oo. Sites where we collected
juveniles were shallow as compared to the other
sampling locations, and none were collected from
depths greater than 1 m. Substrates were com-
posed of fine clay-silt sediments, devoid of sand;

1972

1973

Golfo de
Santa Clara

Figure 6.-Sites sampled in northern Gulf of California for
presence of juvenile totoaba in 1972 and 1973. Circled numbers
indicate offshore areas sampled by otter trawl. All other locations
are shore stations sampled by seine. Sites where juvenile totoaba
were found are indicated by stars. Numbers captured are shown
at upper right of each map.

537

FISHERY BULLETIN: VOL. 74, NO. 3

the mud surface layer was very soft. No small
juvenile totoaba were collected over firm mud
sediments or sandy substrates, as is often the case
with the larger individuals taken in shrimp trawls.

Guevara (1974) is presently analyzing the dis-
tribution of juvenile totoaba captures in shrimp
trawls. Most of his specimens are larger than ours,
implying that the fish move into deeper water as
their growth continues. Tidal currents in the area
are extreme and these may also play a significant
role in juvenile distribution.

Juveniles collected in 1972 were examined for
stomach contents. Remains of amphipods and
other small crustaceans common to the habitat
were recognizable, in addition to remains of
juvenile fishes which we identified as Micropogon
sp., Mugil cephalus, and Leuresthes sardina.
Within the limits imposed by size, the diet of
juvenile totoaba as small as about 6 cm standard
length is comparable in these items with the diet
of the large adults.

Decline of the Fishery

We have traced the growth and decline of the
totoaba fishery and discussed its present status
and methodology. We have presented data on
aspects of totoaba life history and raised questions
concerning possible reproductive behaviors which
may have a bearing on reproductive potential.
Although these have significance, if we consider
the resource from a management perspective one
fact becomes clear: the annual breeding migration
to the mouth of the Colorado River emerges as the
primary source of vulnerability for this declining
population. It serves to concentrate adults in a
predictable small area where they may be fished
with efl^ciency during a critical phase of their life
cycle. To recruit, the juveniles must traverse an
area of intense shrimp trawling activity which
artificially increases juvenile mortality and leads
to further reduction of this already-depleted stock.

The precise factors responsible for the decline of
the totoaba stock cannot be identified with cer-
tainty, but we can enumerate the three most
probable causes as: degradation of the spawning
grounds, degradation of the nursery grounds, and
overfishing. The first two are a result of re-
placement of brakish waters by saline waters in
and around the mouth of the Colorado River. Both
alternatives may be explored by examining
Colorado River flow data and annual totoaba yield
over the critical period of declining catch and

significant flow reduction. We might expect a
relationship to exist between flow and annual yield
if the density of the resident breeding population
(as measured by catch) varies with some unknown
but flow-related quality of the spawning ground.
Relation between flow and catch n years later
(with n years corresponding to age at recruitment)
would indicate the importance of some flow-relat-
ed quality of the nursery ground. Although tests of
overfishing using these data are ambiguous, if
catch is statistically related to catch n years later
we might expect a depletion of the breeding
population resulting from lowered recruitment
levels.

The decline in catch with declining, erratic flow
is evident for the 1942-58 period (Figure 7). Fol-
lowing 1958, the catch increased to a secondary
peak and then crashed to the present all-time
minimum, though flow varied little in the same
period. For reasons given earlier, we discuss
separately the pre-1958 and post-1958 periods.

For the years 1942-58 we have plotted catch
against flow (Figure 8). Linear regression of the
data reveals a highly significant correlation of
annual flow and catch for this period (P <0.001).
However, the river flow data are derived from a
diff'erent base after 1951; analysis of these data in
two segments, before and after this change point,
shows no significant relationship between catch
and flow for either the 1942-50 period or the
1951-58 period. These results suggest that the
highly significant correlation of flow and catch for
the total 1942-58 period may be spurious and due
only to the artificial pairing of declining catch and
declining flow functions. Despite these results, we
cannot ignore the fact that the totoaba congregate
only in the Colorado River estuary (so far as
known), and the salient feature distinguishing
this from other estuaries in the northern gulf is
the (former) discharge of large quantities of fresh
water from the Colorado River. Therefore, ac-
cepting the tentative nature of the flow-catch
relationship, we explore its possible biological
basis.

The mechanism could lie in olfactory cues from
the river system (physiological responses to either
fresh water or substrate "odor"). Given the pres-
ent agricultural scene, such cues may no longer be
present. The present Colorado River surface flow
to the Gulf of California is close to zero for all
practical purposes and this situation is likely to
continue in the future. A conspicuous bar now
exists across the channel upstream from the delta

538

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

2,4 00-1

2,0 00

r
o

1,6 00-

1,2 00-

(J 800

4 00

r-1 3

■10

- 8

CATCH

19 30

I T T I I I — [— r

19 35 1940

r I I

1945

1950

1 I r

1955

1960

T-T

■^

I I I I

-6 (0

5 ^-'
4 O

I- 3

1965

1970

19 75

YEAR

Figure 7.-Colorado River flow in thousands of acre-feet and totoaba fishery annual yield in metric tons for the 1930-75 period. Totoaba
annual yield data are those of Figure 1. Flow data were calculated from records published in Water Supply Papers 1313, 1733, and 1926
(U.S. Geological Survey, respectively, 1954, 1964, and 1970). Flow data for 1966-75 are not shown but we do not expect them to deviate
beyond the 1960-65 variation above. We have calculated the flow delivered to Mexico at the southerly international boundary (near San
Luis, Ariz.) as follows (data sources are cited only on first mention): 1930-36: Colorado River at Yuma (1954:710) + Yuma Main Canal
Wasteway (1954:717) + Calif. Drainage Canal (1954:723) - Alamo Canal (1954:724) + Eleven-mile Wasteway (1954:726) + Cooper
Wasteway (1954:726); 1937-50: Colorado River at Rockwood Gate, Calif. (1954:712) - Alamo Canal -i- Eleven-mile Wasteway -(-
Twenty-one Mile Wasteway (1954:727) + Cooper Wasteway; 1951-65: Colorado River at southerly boundary, near San Luis (1964:563;
1970:519-521).

p
«.io
>.io

<.001

6 7 8 9

FLOW (10* acre-feet)

Figure 8. -Plot of annual totoaba
yield and annual Colorado River flow
for the 1942-58 period. Data are those
displayed in Figure 7. The points
below the dashed line represent the
1951-58 flow years. Though the rela-
tionship between catch and flow for
the 1942-58 period are highly sig-
nificant, the disparity between the r^
levels for the component periods
1942-50 and 1951-58 invites caution in
interpretation of these results.

539

FISHERY BULLETIN; VOL. 74, NO. 3

islands, and flow measurements at the southern-
most Mexican hydrographic station known as El
Mari'timo, formerly considered the best single
index of actual surface input to the gulf (Schreiber
1969), were discontinued in 1968 for lack of mean-
ingful data.'' Further, the extensive use of all
available water from the lower Colorado River
drainage system for irrigation has resulted in
hypersalinity of return flows and is a major prob-
lem on both sides of the international boundary.
Water returned to the river channel which may
reach the gulf is now likely to be at least as saline
as the marine water it joins. Thompson (1968:8),
summarizing the history of Colorado River flow
and effects of exploitation on detrital loads, con-
cluded, "Probably little river detritus has reached
the northwestern Gulf of California in the last
55-60 years." Thus, if odor is not carried beyond
upstream dams and fields where the detrital load
stops, it must originate from reworking of the
massive deltaic deposits by the strong tidal cur-
rents of the uppermost gulf. If Thompson's es-
timate is correct, this may have been occurring
during the developing years of the fishery; the rate
of decay of such a process is unknown.

The post-1958 flow and catch data (Figure 7)
contrast with those of the previous period. We feel
that the secondary peak in totoaba production may
be attributable to extraneous factors such as
changes in eff"ort or efficiency (availability of nylon
gill nets?) which produced a temporary increase in
catch. Another possible reason may have been the
enforcement of the 1955 breeding preserve
regulations which offered some temporary relief
from exploitation. If fishing in the sanctuary were
to resume after a period of time, the yield might
recover and fall in the observed manner.

We now consider the second hypothesis, that the
cause of stock depletion is degradation of the
nursery ground. When annual totoaba yield is
compared with river flow in earlier years (e.g., the
1951 totoaba catch compared with the 1942 river
flow, etc.), lag times ranging from 6 to 10 yr all
give significant negative correlations (P<0.05)
using standard linear regression techniques. The
relationship is most distinct (Figure 9) when the

lag time is 9 yr (P<0.01). The 6- to 10-yr periods
correspond with estimated ages of recruitment
employed below. ^ We find this negative relation-
ship of flow and (lagged) totoaba yield highly
interesting, though puzzling. The relation could be
taken to imply that survival of young stages is a
critical factor, since it couples increased river flow
in any one year with reduced recruitment of that
year class to the population. This interpretation
discounts hypotheses of larval and juvenile phys-
iological dependence on waters of lowered salinity
(Berdegue 1955, 1956; Cannon 1966; Cause 1969;
Sotomayor 1970). An alternate analysis using flow
data only for the March-July period over the years
of catch decline would be a better test of the effect
of flow on larvae and small juveniles.

We know that successful reproduction still
continues in the northern gulf as demonstrated by
our ability to find juvenile fish on the nursery
grounds. Despite searching, we have found no
conspicuous subsurface freshwater seeps which
might have provided local areas for limited suc-
cessful spawning. We believe that reproduction
occurs over the entire ancestral spawning
grounds. Thus, we conclude that adverse effects of
salinity changes must operate in a relative and not
an absolute manner. The advantages realized by
potential recruits on the nursery ground may be
those of reduced predation and abundant food

•^Nishikawa-Kinomura, K. A. 1973. Flow of the Colorado
River into the Gulf of California. In S. Alvarez-Borrego et al.,
Preliminary report to the Secretariat of Hydraulic Resources on
the second stage of the chemical study on insecticide contami-
nation at the mouth of the Colorado River, p. 15-19. Unpubl. rep.
Mar. Sci. Unit, Inst. Oceanol. Res., Univ. Baja Calif., Ensenada,
Mex.

^The senior author has reviewed the published estimates of
growth curves and ages of recruitment (see Arvizu and Chavez
1972, for a summary of this literature). Apparent discrepancies
between reported lengths at different ages and serious disa-
greement between Berdegue's (1955) growth estimates and the
distribution of lengths in observed commercial catches in 1963
(Arvizu and Chavez 1972) encouraged closer scrutiny of these
data. The variation in lengths at particular ages and in maximum
lengths reported by different authors and summarized by Arvizu
and Chavez appear to derive from use of both standard length
and total length measurements without discriminating between
the two. The senior author calculated von Bertalanfty growth
curves using a resolved maximum standard length of 1,600 mm
and the intermediate lengths reported by Berdegue (1955). The
new growth curves indicate that the best estimate of recruit-
ment age is 6 or 7 yr; they also produce a length series which
corresponds well with that observed in commercial catches. Male
and female totoaba may vary significantly in growth rates and
therefore may recruit at different ages. This variation allows
extension of the possible recruitment age to 10 yr. J. E. Fitch
(pers. commun.) has examined totoaba otoliths and concluded
that totoaba first spawn at age 8. If totoaba do not accompany
the migrant population until reproductively mature, his results
are consistent with the ages of recruitment used here. However,
his overall ages as read from otoliths indicate that these new
growth curves may contain a wide margin of error in terms of
predicted age at length observed. Fitch has also found that
totoaba scales are of little use for growth studies after about age
8; this may explain the maximum lengths at age 8 or 9 reported
by Nakashima (Jordan 1916), which we now believe to be
erroneous. It also may account for errors in Berdegue's (1955)
estimates, since he relied heavily on age determinations from
scales.

540

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

2 000 -,

^ f LO W YE ARS 1936 196 0

Carchtog r p

I - 5 <_39 6 > 05

.4 4 9 < 0 5

1,600 -

lA
(0

>• 800

< 400

0 '—

1033 - 606 X

7
8
9
10

459 <05

50 3 <.0 5

550 < 0 1

4 5 4 < 0 5

1 2

Figure 9.-Plot of annual totoaba
yield (metric tons) 9 yr following the
annual recorded flow, for the 1936-60
period. This plot displays the sug-
gested relationship between flow and
recruitment level. Linear regression
calculations employing lag times from
1 to 10 yr are significant only for those
years corresponding to estimated
ages of recruitment (6-10 yr).

FLOW (10 acrefeet)

resources, both of which are directly related to
substrate and shallowness, and indirectly related
to flow.

The final cause suggested for the decline in
totoaba stock is overfishing of the breeding
population. We have examined the relationship of
catch with catch n years later. If catch is a good
indicator of population size, then we would expect
a linear, positive relationship between population
size and the size one recruitment age later. Alter-
natively, if catch is partially a function of
sociopolitical constraints (e.g., enforcement of a
preserve area and closed season resulting in a
catch which significantly underrepresents the
population size), we might expect a more com-
plicated plot with a distinct cluster of years
corresponding to periods of fishing regulations.
We have analyzed plots of catch against catch for
recruitment periods ranging from 6 to 10 yr and
have found significant relationships which satisfy
both of the above predictions. Graphs for all
estimated ages of recruitment from 6 to 10 yr
showed essentially the same pattern (Figure 10).
However, we note inconsistencies which advise
against drawing strong conclusions of either
overfishing or the demonstrated worth of enforced
regulatory measures. For example, increases in
catch occur 2-3 yr earlier than is consistent with
our assumption that regulatory measures were not
enforced until 1955, given that the minimum age
of recruitment is 6 yr. The question of a change in

gear efficiency, effecting a realized increase in
effort, serves to confound the analysis; although
catch would increase, this factor alone could not
explain recovery of catch to such high levels. We
can visualize a combination of factors giving rise
to the significant second peak in totoaba produc-
tion (increase in gear efficiency acting on an
increased population size following the period of
regulation) but lacking effort data throughout the
period, our hypothesis must remain speculative.

Support for the overfishing hypothesis may lie
in the recent trend of the relative catches of the
three main totoaba fishing fleets. Historically, the

1,200

800

400 -

Y: 182.9 • .194X
r : .497 p < 05

400

800

1,200
CATCH

1.6 00

2/)00

Figure lO.-Totoaba catch plotted against catch 8 yr previous.
This examines for evidence of reduction in breeding stock by
overfishing. All data are from Figure 1, in metric tons. Points
above dashed line are for the 1952-59 period. The breeding
sanctuary was established in 1955 and the points corresponding
to the years 1955-59 may reflect enforcement of this regulation
(see text).

541

FISHERY BULLETIN; VOL. 74, NO. 3

Table 2.- Prime season catch,' in metric tons, of the three
principal totoaba fishing ports, 1964-75. Data for 1965-70 period
are from Arvizu and Chavez (1972). Data for 1971-75 were
provided by H. Chavez (pers. commun.). The 1974-75 data are
preliminary but are not expected to increase by more than 10*^
from these figures.

Year

Puerto Penasco

Golfo de Santa Clara

San Felipe

1964

72.7

128.4

277.7

1965

97.0

561.2

57.0

1966

177.5

388.6

488.7

1967

188.3

173.4

334.7

1968

37.9

385.7

290.0

1969

60.2

213.8

160.9

1970

27.2

248.7

169.4

1971

69.0

46.0

95.0

1972

52.0

86.0

104.0

1973

88.0

21.0

37.0

1974

51.0

17.0

52.0

1975

49.0

4.0

5.0

'Catch is calculated by adding January-April yields as recorded
in the official statistics.

Puerto Penasco fleet catches fewer totoaba than
either the San Felipe or Golfo de Santa Clara fleets
(Table 2). This is a logical result of the fishing
methods and areas worked by the three fleets.
However during the last 3 yr, Puerto Penasco has
equalled or exceeded the other ports in recorded
totoaba yield despite no apparent increase in
eff'ort. Our interpretation of this new trend is that
the migrant population, encountered first by the
Puerto Penasco fleet along the Sonora shore, is
being decimated before reaching the spawning
grounds.

CONCLUSIONS

Our review of the history of exploitation of the
stock, our data on spawning concentrations,
breeding migration, and juvenile habitat, and our
analyses of proposed hypotheses for the decline of
the fishery have emphasized points of population
vulnerability. Fleets of the three major ports are
highly skilled at finding the migrant schools of
totoaba. They have, in a sense, specialized to
exploit the ascent, resident, and descent phases of
the breeding migration, both by nature of their
vessels and their port facilities, and by con-
sequence of their geographic locations. This level
of exploitation is possible only because the fishers
are able to predict with accuracy the migration
pattern. In the past, the commercial population
level was high and the temporal nature of the port
specializations was not a factor in the ranking of
port yields. Now, when the population level has
reached an all-time low, the Puerto Penasco fleet
seems to have some new advantage.

The totoaba breeding behavior we describe
serves to render the resident spawning population
especially vulnerable to fishing effort. Frenzied
spawning in dense aggregations following a pe-
riod of behavioral stimulation insures that when a
net is encountered, the capture rate will be par-
ticularly high. Our capture incidence, daily catch,
and gonadal maturation data confirm the con-
sequences of these attributes. The bathymetry of
the delta restricts the spawning schools to highly
limited areas. These areas, or channels, are the
prime fishing sites for the Golfo de Santa Clara
fleet. They appear to lie (by our estimate) partially
within the breeding sanctuary established by the
Mexican Government.

We reiterate here the artificial fishing mortality
suffered by juveniles in their forced crossing of the
near-delta waters as they make their way south
from the nursery grounds. We have documented
some known nursery sites and have suggested
characteristics of the juvenile habitat which may
have predictive value in future surveys of the area.

We have examined the three most probable
factors responsible for the decline in totoaba stock.
Subject to the limitations of our catch and flow
data, our results suggest that overfishing has
played the most significant role during the pre-
1958 catch period. We speculate that the low yields
of the 1956-59 period may have been due to en-
forcement of the breeding sanctuary regulation
and that this partial temporary relief from exploita-
tion may, together with increased gear efficiency,
have been responsible for the second peak in
totoaba production. If this is true, then the power
of regulatory measures for recovery of this com-
mercial stock has been demonstrated. The cor-
relation of annual yield with annual Colorado
River flow, though weakened by statistical ir-
regularities, attests to the importance of some
flow-related quality of the spawning grounds.
Degradation of the spawning grounds, possibly in
the ability to provide olfactory cues, also may have
resulted in a decline of the commercial population.
According to our results, degradation of the nur-
sery grounds, through deterioration of some un-
known flow-related quality, has probably not
played a significant role in the fishery's decline.

Although it may be possible to ignore statistical
analyses and the conclusions therefrom, one can-
not deny that the annual yield in 1975 was the
minimum recorded in the history of the fishery, a
mere 2.5% of the highest recorded catch. The area
of the fishery has shrunk to a small fraction of its

542

FLANAGAN and HENDRICKSON: FISHERY AND REPRODUCTIVE BIOLOGY OF TOTOABA

former size, and catch events have become spo-
radic and undependable. The span of the breeding
period has been reduced from several months to a
period of only 18 days within the open season of
1972. Of those 18 days, a majority of the catch
occurred on 21 and 22 March. Although data are
scarce, the average size of adult fish is reduced and
in recent years most commercially caught in-
dividuals have probably been first- or second-year
spawners. These harsh facts are indications of a
fish population struggling unsucessfully for sur-
vival under pressure.

The future of the species is uncertain. Until the
recent action of the Mexican Government in
establishing a total closed season, the outlook was
bleak, indeed. While the commercial fishery was
ready to crash before its legal cancellation (a
number of financial failures were reported to us),
and would presumably never have hunted down
and eliminated the last reproductive pair of these
magnificent animals, the continued rising prices
for totoaba in a seller's market would have guar-
anteed continued maximum pressure. If there are
behavioral elements in the reproductive pattern of
the species which require mutual stimulation in
large schools for reproductive success, a threshold
may have already been crossed which will drive the
totoaba the way of the passenger pigeon. The
trends produced by irreversible change of the
spawning ground may prove more important than
we have speculated. In either of the last two
circumstances, the ability of the stocks to rebound
upon release of fishing pressure may be critically
impaired.

We suggest three meaningful measures at this
stage: 1) Continuation of the total closed season
which has been imposed, until intensive studies
document a strong and vigorously increasing
population. We suggest an enforcement period of
about IV2 times our estimated 6-yr minimum
recruitment age, or 10 yr. 2) Action by the U.S.
Government (the major market area) com-
plementing the Mexican action by declaring the
totoaba an endangered species, to facilitate en-
forcement of the neighbor country laws by
removing much of the stimulus for poaching and
smuggling. 3) Intensive scientific investigation to
provide knowledge of the species' autecology and
behavior with potential application to all facets of
management, ranging from environmental ma-
nipulation to hatchery techniques. Failing these,
we conclude that the probability of extinction of
Cynoscion macdonaldi by the year 2000 is high.

ACKNOWLEDGMENTS

Various research grants and contracts con-
tributed to the support of this work and are here
gratefully acknowledged: International Union for
the Conservation of Nature and Natural Re-
sources/World Wildlife Fund Project No. 623;
National Science Foundation Grants GB29101 and
GB34675; National Aeronautics and Space Ad-
ministration Contract NAS5-21777; and a special
University of Arizona Foundation Grant. We
acknowledge with sincere appreciation the sup-
port and professional participation of Mexican
authorities and scientists in this work. Luis Ka-
suga Osaka, Director of the National Institute of
Fisheries, Mexico, and Pedro Mercado Sanchez of
the Subsecretariat of Fisheries, Ministry of In-
dustry and Commerce, helped us obtain neces-
sary permits (3977, 5344, 8202, and 12183) for work
in the field and in other ways provided important
support and encouragement. Among the many
members of the National Institute of Fisheries
who gave freely of their time and efforts, special
mention must be made of Joaquin Arvizu and
Humberto Chavez, who cooperated in field work
and data analysis. We are particularly indebted to
Biologist Chavez, Head of the Institute's Depart-
ment of Fisheries Resources, for important con-
tributions and professional advice from inception
of field work to manuscript preparation. The
Fishery Inspectors of San Felipe, Baja California
Norte, and Puerto Penasco, Sonora, rendered
valuable assistance in making contacts with
fishers and made date available. Francisco
Aguilera Grijalva, Fishery Inspector at Golfo de
Santa Clara, Sonora, made particular efforts on our
behalf and was of invaluable assistance in collec-
tion of detailed catch data. We thank all the
students of the University of Arizona who ren-
dered assistance on field trips, and all the staff and
students of the Institute of Oceanologic Inves-
tigations, Autonomous University of Baja
California, who have been such productive
partners in most of our work in the northern Gulf
of California. Special recognition is due L. T.
Findley for his field and museum contributions
and enthusiastic interest. We greatly appreciate
the time spent by D. A. Thomson, J. Tash, and G.
Pyke who read the manuscript critically and
consulted on matters of data analysis and inter-
pretation (all responsibility for errors is ours). It is
impossible to overvalue the contribution of Lupe
P. Hendrickson in clerical, translation, and edi-

543

FISHERY BULLETIN: VOL. 74, NO. 3

torial functions throughout the entire course of
the project. Lastly, a particular expression of
gratitude is due to Javier Ramirez of Golfo de
Santa Clara, master fisher and astute student of
nature, who was frequently the key to success in
field projects.

LITERATURE CITED

Arvizu, J., AND H. Chavez.

1972. Sinopsis sobre la biologia de la totoaba, Cynoscion
macdonaldi Gilbert, 1890. FAO (Food Agric. Organ. U.N.)
Fish. Synop. 108, 21 p.

Berdegue, a. J.

1955. La pesqueria de la totoaba (Q/ho.st/o?) macdonaldi
Gilbert) en San Felipe, Baja California. Rev. Soc. Mex.
Hist. Nat. 16:45-78.

1956. Feces de importancia comercial en la costa nor-
occidental de Mexico. Secre. Mar., Dir. Gen. Pesca Ind.
Conexas, 345 p.

Breder, C. M., Jr., and D. E. Rosen.

1966. Modes of reproduction in fishes. Natural History
Press, Garden City, N.Y., 941 p.
Cannon, R.

1966. The Sea of Cortez. Lane Magazine and Book Co.,
Menlo Park, Calif., 283 p.
Chavez, H.

1973. Descripcion de los ejemplares juveniles de totoaba,
Cynoscion macdo)ial(li Gilbert. Rev. Soc. Mex. Hist. Nat.
34:293-300.

Chute, G.R.

1928. The totuava fishery of the California Gulf. Calif. Fish

Game 14:275-281.
1930. Seen Kow, a regal soup-stock. Calif. Fish Game
16:23-35.
Craig, J. A.

1926. A new fishery in Mexico. Calif. Fish Game 12:166-169.
Crocker, R. S.

1932. The white sea-bass and related species that are sold in
California fish markets. Calif. Fish Game 18:318-327.
Fitch, J. E.

1949. Mexican corbina and totuava. In The commercial fish
catch of California for the year 1947 with an historical
review 1916-1947, p. 83-84. Calif. Dep. Fish Game, Fish
Bull. 74.
Gabrielson, I. N., AND F. R. Lamonte.

1954. The fisherman's encyclopedia. Stackpole Co., Harris-
burg, Pa., 730 p.

Cause, C. I.

1969. A fish threatened. Underwater Nat. 6:28-31.
Gilbert, C. H.

1891. Scientific results of the explorations by the U.S. Fish

Commission steamer Albatross. No. XII-A preliminary
report on the fishes collected by the steamer Albatross on
the Pacific coast of North America during the year 1889,
with descriptions of twelve new genera and ninety-two
new species. Proc. U.S. Natl. Mus. 13:49-126.
Guevara, S.

1974. Sobre le eclogi'a de los juveniles de totoabe Cynoscion
macdonaldi Gilbert. (Abstr.) In Resljmenes Quinto
Congreso Nacional de Oceanografia, 22 Oct. 1974, p. 7.
Guaymas, Sonora, Mexico.
Jordan, D. S.

1916. Notes on the totuava (Cynoscion macdonaldi Gilbert).
Copeia 1916:85.
Jordan, D. S., and B. W. Evermann.

1898. The fishes of North and Middle America. Bull. U.S.

Natl. Mus. 47, Part 2:1241-2183.
1902. American food and game fishes. Doubleday, Page &
Co., N.Y., 573 p.
Jordan, D. S., B. W. Evermann, and H. W. Clark.

1930. Check list of the fishes and fishlike vertebrates of
North and Middle America north of the northern boundary
of Venezuela and Colombia. U.S. Bur. Fish., Rep. U.S.
Comm. Fish. 1928. Part II. Append. 10, 670 p. (Doc. 1055.)
Lanham, U.

1962. The fishes. Columbia Univ. Press, N.Y., 116 p.
Schreiber, J. F., Jr.

1969. Changes in Colorado River flow. In D. A. Thomson et
al. (editors), Environmental impact of brine effluents on
Gulf of California, p. 83-87. U.S. Dep. Int., Off. Saline
Water, Res. Dev. Prog. Rep. 387.

Sotomayor, C.

1970. La totoaba, una especia que se extingue lentamen-
te. Tec. Pesq. 3:22-25.

Thompson, R. W.

1968. Tidal flat sedimentation on the Colorado River Delta,
northwestern Gulf of California. Geol. Soc. Am., Mem.
107, 133 p.

Thomson, D. A.

1969. The commercial fisheries industry. In D. A. Thomson
et al. (editors). Environmental impact of brine effluents on
Gulf of California, p. 100-103. U.S. Dep. Int., Off. Saline
Water, Res. Dev. Prog. Rep. 387.

U.S. Geological Survey.

1954. Compilation of records of surface waters of the United
States through September 1950. Part 9. Colorado River
Basin. [U.S.] Geol. Surv. Water-Supply Pap. 1313, 749 p.

1964. Compilation of records of surface waters of the United
States, October 1950 to September 1960. Part 9. Colorado
River Basin. [U.S.] Geol. Surv. Water-Supply Pap. 1733,
586 p.

1970. Surface water supply of the United States 1961-
65. Part 9. Colorado River Basin. Vol. 3. Lower Colorado
River Basin. [U.S.] Geol. Surv. Water-Supply Pap. 1926,
571 p.

544

UPTAKE, DISTRIBUTION, AND DEPURATION OF ^^C-BENZENE

IN NORTHERN ANCHOVY, ENGRAULIS MORDAX, AND

STRIPED BASS, MORONE SAXATILIS

Sid Korn,' Nina Hirsch,^ and Jeannette W. Struhsaker^

ABSTRACT

The uptake, distribution, and depuration of water-soluble, monocyclic hydrocarbon contained in
petroleum and refined products was studied in two species of marine fish. Mature northern anchovy,
Engraulis mordax, and juvenile striped bass, Morone saxatilis, were exposed to sublethal concentra-
tions of '^C-benzene for 48 h. Residues in tissues exhibiting a high lipid content or representing
apparent major metabolic sites were measured during the exposure and afterwards when the fish were
transferred to clean seawater. Fish exhibited a rapid uptake over a wide range of benzene
concentrations in the water column. Accumulation in anchovy was considerably greater than in striped
bass. Results indicate that the pathway of hydrocarbons through the liver, gallbladder, intestines, and
colon is a major depuration route. Residues were depurated rapidly after cessation of exposure; in
striped bass tissues most residues were undetectable by 7 days.

Increased drilling, transportation, and refining of
crude oils near or on coastal waters has led to the
need for research on the effects of oil on estuarine
biota. Considerable public concern has evolved
from such occurrences as tanker spills and the
Santa Barbara well blowout. However, long-term
sublethal effects of low levels of oil in inshore areas
may be of greater importance to marine popula-
tions than short-term lethal effects of high levels
resulting from catastrophic events such as tanker
spills and drilling blowouts. It is important to
study the effects of chronic oil exposure on marine
organisms.

Benzene is a principal aromatic oil component
(up to 6.75 ppm in the water-soluble extract
[Anderson et al. 1974]) that is relatively water
soluble (1,993 jul/liter [Benville and Korn 1974])
and has significant effects on fishes (Brocksen and
Bailey 1973; Korn et al. in press). The preceding
studies demonstrated the effects of benzene on the
nervous system, respiration, and growth of fish.
Brocksen and Bailey showed latent effects of ben-
zene on respiratory response lasting up to 6 days
after fish were placed in clean water.

Concentrations of highly volatile monocyclic
aromatics such as benzene are not thought to be
very high in areas subject to chronic exposure to

^Southwest Fisheries Center Tiburon Laboratory, National
Marine Fisheries Service, NOAA, Tiburon, Calif.; present
address: Northwest Fisheries Center Auke Bay Fisheries
Laboratory, NMFS, NOAA, P.O. Box 155, Auke Bay, AK 99821.

^Southwest Fisheries Center Tiburon Laboratory, NMFS,
NOAA, 3150 Paradise Drive, Tiburon, CA 94920.

oil. However, measurements of monocyclic
aromatics in such situations are scarce. Our
preliminary measurements in San Francisco Bay
indicate a maximum range from 1 to 10 jul/liter
benzene in relatively unpolluted bay areas. Al-
though the chronic levels are low, if fish ac-
cumulate benzene over field concentrations and if
energy is required to metabolize, detoxify, and
depurate accumulated aromatics, detrimental
long-term physiological effects are possible.

Investigators such as Lee, Sauerheber, and
Benson (1972); Lee, Sauerheber, and Dobbs (1972);
Anderson et al. (1974); and Lee (1975) examined
uptake of higher aromatics in invertebrates and
fish, but no work has been done with benzene.

The fish we studied were San Francisco Bay
species but also occur widely in other areas where
chronic oil pollution may pose a problem. Striped
bass, Morone saxatilis, is an important recrea-
tional species on the west and east coasts, while
northern anchovy, Engraulis mordax, is not only a
major forage fish for striped bass but also consti-
tutes the greatest biomass of any fishery in the
California Current.

The objective of this study was to determine the
uptake, distribution, and depuration of benzene in
these two species of fishes.

METHODS

Adult northern anchovies were obtained from a
local bait dealer and acclimated under controlled

Manuscript accepted March 1976.
FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

545

FISHERY BULLETIN; VOL. 74, NO. 3

environmental conditions comparable to those
used in experiments. Juvenile striped bass were
obtained from the water diversion facilities of the
Bureau of Reclamation at Tracy, Calif. Fish were
acclimated in 2,000-liter circular tanks for at least
2 wk before testing and fed ground squid once
daily to satiation.

In all uptake studies, an appropriate number of
fish (Table 1) were transferred into oval 200-liter
test tanks and further acclimated for 1 wk. The
number of fish per tank was limited to the number
(«lg/liter) that could be maintained during a 48-h
static exposure period when oxygen is a limiting
factor. The 48-h static exposure period instead of
an open-system constant exposure was necessi-
tated by the expense of the '^C-benzene required
for a relatively large volume of water. Except for
the 48-h static exposure period, a flow of 1 liter/
min of filtered seawater was maintained through-
out. During flow periods the salinity and temper-
ature of the water were monitored and controlled
by the seawater system components (Korn 1975),
whereas temperature was not controlled during
the static exposure period.

Stock benzene solutions used for dosing the
exposure tanks were prepared as follows: A satur-
ated benzene solution (1 ml benzene in 250 ml
seawater) was prepared in a separatory funnel by
vigorous shaking and then allowed to settle for 1 h.
The resulting solution was analyzed by the gas
chromatography method of Benville and Korn
(1974). Next, "C (99.9% ring-labeled benzene,

specific activity, 85 juCi/mmol) was mixed with
another 200 ml of seawater to make a stock
solution and was kept frozen until used. The
saturated benzene solution was then mixed with
*^C stock solution to the proper specific activity,
and the appropriate volume was poured into each
tank and mixed by gentle stirring. After mixing,
1-ml water samples were added to a scintillator
(10-ml Packard Instagel)^ and the benzene con-
centration was measured. Carbon 14 counting was
done on a Packard Model 2008 Tri-Carb liquid
scintillation spectrometer system. Internal stan-
dardization yielded 85% counting eflficiency, and all
water values were corrected accordingly.

Uptake, distribution, and depuration were de-
termined by sampling fish, rinsing them exter-
nally with methanol to remove adsorbed benzene,
dissecting out tissues, weighing tissue samples
(<200 mg), placing each tissue in a vial with tissue
digester solution (1 ml/100 mg tissue Packard
Soluene-100), and allowing 48-h digestion at room
temperature. Scintillator (10-ml Packard
Dimilume) was added to these samples and '^C
radioactivity measured. Approximate mean
counting efl^ciencies of 60% and 67% were cal-
culated from spiked samples and used to correct
anchovy and striped bass tissue residue values
respectively. Water and tissue samples yielding
below 40 counts per minute were considered below
the detectable limits of our system.

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Table 1. -Summary of experimental conditions for '^C-benzene uptake and depuration tests with northern anchovy and striped bass.

Salinity was 24-267oo

Species and
test number

Initial mean Tanks

Specific benzene per Fish

activity concentration' concen- per

(cpm/nl) (jUl/liter) trafion tank

Time of tissue

sampling^

(days)

Total wet wt (g)

Mean

Type of tissue examined

Northern anchovy:
1

0.11

3.7

0.11

0.0097

0.0048

320

0.00069

Striped bass:
4

0.088

3
4
4
4

8 0.042,0.125,

0,25, 1, 2
8 0.042,0.25,1,2,4

8 0.042,0.25,1,2,4

8 0.25,1,2,3,4,7

0.25, 1,2,3,4,
5,6, 7,8,9

17.03 6.55 Liver, brain, gill, muscle

12.93
11.70
12.74
13.94

5.04
4.74
3.35
4.46

76.87 34.60

Liver, brain, gill, muscle,

gallbladder, intestine
Liver, brain, gill, muscle,

gallbladder, intestine

Liver, brain, gill, muscle,
gallbladder, intestine,
mesenteric fat, colon,
heart, stomach

'Exposure to '■'C-benzene w/as static for 48 h followed by resumption of water flow for the duration. Recent analyses by gas-liquid chro-
matography yielded 0. 00015-0. OOlOjUl/liter background benzene concentration in the seawater at this facility which is not included in these
values.

'One fish per tank at each sampling time.

546

KORN ET AL.: UPTAKE AND DEPURATION OF "C-BENZENE

Water samples were taken first; then tissues
were sampled until open flow was reestablished.
Tissues were sampled as noted for each of the four
experiments included in this report (Table 1). It is
recognized that the residues reported may contain
metabolites and degradation products in addition
to benzene.

The original data on declining seawater con-
centrations of benzene during tests and on
decreasing concentrations of residues in fish tis-
sues during depuration were first analyzed with a
least-squares curve-fitting computer program to
determine if the hypothesized function was the
best fit. Linear regression analyses were then
performed on logarithmically transformed data,
and regression coefficients were tested for sig-
nificance of differences between slopes and a
pooled regression coefficient (Snedecor and Coch-
ran 1968).

RESULTS

There were no deaths during the tests. The
benzene concentration in the seawater declined

Tablk 2,- Benzene concentration during; 4^;-h exposure period

Ex])onential decline (}' = ac

coefficients for each exper-

exponentially (Y = ae

-0.0 183 A'

, where Y is concen-

tration and X is time) during all tests. After 24-h
exposure, 48-65% remained; after 48 h, 30-43%
remained, at which point the water flow was
renewed (Table 2).

In general, accumulation in striped bass was
greatest in the gallbladder, followed by mesen-
teric fat, colon, intestine, liver, brain, gill, heart,

iment from least-squares curve tittinjj.

Benzene

-seawater

Percentage

actual initial

remaining

Test

concentration

after

no.

jal/liter

nl/liter

24 h 48 h

3.7

3,700

3.53

-0.1997

54 30

0.110

110

0.104

-0.02381*

54 31

0.097

9.7

0.094

-0.01262*

65 —

0.0048

4.8

0.00457

-0.01983

48 42

0.00069

0.69 12

0.000692

-0.01847

62 42

0.088

0.0991

-0.01813

65 43

*No significant difference between slopes (at a = 0.05) except
between tests 2a and 2b.

The equation V = ae-ooi83x describes the exponential de-
cline in benzene, using a pooled regression coefficent,

stomach, and muscle (Table 3). Anchovy exhibited
similar results minus the mesenteric fat, colon,
heart, and stomach tissues, which were not sam-
pled. The order of decreasing accumulation varied
slightly according to experiment. The gallbladder
accumulated 53.4-8,450 times the initial water
concentration, while muscle accumulated 1.11-135
times the initial water concentration. Maximum
concentrations were obtained in the tissues from
0.25 to 4 days after starting exposure. Mesenteric
fat, gallbladder, liver, and intestine usually
reached a maximum accumulation later than the
other tissues.

Accumulation in anchovies was considerably
greater than in striped bass in the tissues mea-
sured in both species (Figures 1, 2; Table 3). The
pattern of uptake in the gill and gallbladder was

Table 3.-Mean maximum concentration factors' in various tissues and days elapsed (numbers in parentheses) from beginning of

exposure for northern anchovy and striped bass.

Initial

mean

benzene-

Species
and
test

seawater

Concentration

factor in tissue of

tration

Gall-

number

^Lil/llter)

Gill

Brain

Muscle

Fat Heart

Stomach

Liver

bladder

Intestine

Colon

Northern

anchovy:

3.7

34.3
(2)

30.0
(2)

22.7
(1)

— —

—

45.1
(1)

0.11

41.8

— —

—

54.6

4,360

209

—

(1)

(0.25)

(2)

0.0097

113

29.9

— —

—

309

8,450

505

—

(1)

(0.25)

(2)

0.0048

7.92

7.5

5.42

—

66.7

229

60.4

—

(0.25)

(4)

(2)

(2.3)

0.00069

7.1

9.13

135

— —

—

31.9

116

34.8

—

(0.25)

(4)

(2)

(3)

(2)

Striped

bass:

0.088

5.57

7.16

1.11

1.14 2.95

2.72

9.77

53.4

5.45

14.8

(1)

(0.25)

(0.25) (0.25)

(0.25)

(2)

(0.25)

'Factor X (benzene-seawater concentration in microliter per liter)
liters) / (tissue wet weight in grams).

actual nanoliters per gram mean tissue value or (benzene in nano-

547

FISHERY BULLETIN: VOL. 74, NO. 3

"1 r

24 48

TIME (hours)

24 48

TIME (hours)

Figure l.-Mean '^C-benzene uptake in tissues (nl/g wet weight) in anchovy (solid lines) and striped bass (dashed lines); sample
number: three or four fish. Also shown are mean '^C-benzene concentrations in seawater in anchovy tanks (solid lines) and in striped
bass tanks (dashed lines); sample number in Table 2. The concentrations on the Y-axis are calculated from total '^C radioactivity and
may include metabolites of benzene.

similar between species, while in the brain, liver,
muscle, and intestine, a maximum level was
maintained longer in the anchovy. In both species,
the greatest rate of uptake occurred in the first
6h.

Residues were depurated rapidly after cessation
of exposure (Table 4). Gallbladder, mesenteric fat,
liver, and gill maintained residues the longest.
Depuration appeared to occur more rapidly in
striped bass than in anchovies in some tissues. In
striped bass, depuration is generally described by
the logarithmoc form of a power function (In Y =
In a + 61n X) after cessation of exposure on day 2
until day 4 or 5 (Figure 3). Subsequently, several of
the tissues showed a secondary increase and
decrease in concentration. In muscle tissue, res-
idues were undetectable 24 h after exposure
ended.

DISCUSSION

Accumulation levels are based solely on
radiometric analysis. This analytical technique
does not distinguish between >^C-labeled benzene
and derived ring metabolites. Complementary
analysis by thin-layered chromatography or gas
chromatography could have determined some of
the actual compounds present, but it was not
performed during these experiments. It is
hypothesized that fish are capable of excreting and
metabolizing benzene. Although there is no direct
evidence, the residues reported in selected tissues
may be representative of the unchanged parent
benzene or associated metabolites and degrada-
tion products. Any or all of these may be toxic to
fish.

Benzene and/or metabolites accumulate

548

KORN ET AL.: UPTAKE AND DEPURATION OF '"C-BENZENE

103

■102

Figure 2.-Mean '-"C-benzene depuration from tissues (nl/g wet weight) of striped bass; sample number in Table 4. ND = nondetectable
level (see Methods). The concentrations on the Y-axis are calculated from total '^C radioactivity and may include metabolites of
benzene.

predominantly in tissues that exhibit a high lipid
content or represent apparent major metabolic
sites. Thus, lipid-rich mesenteric fat and brain
tissues had high accumulations; while liver, gall-

bladder, intestine, and colon (which are tissues
associated with the metabolic breakdown and
excretion of benzene) also accumulated benzene to
higher levels (Table 3).

Table 4. -Percent residues' remaining in northern anchovy and striped bass after termination of 48-h exposure to benzene. (Sample

sizes in parentheses.)

Northern
Test 2a

anchovy^
Test 2b

Striped bass^

Test 4

Days from

termination of exposure

Tissue

Gill

61(3)

26(2)

30(3)

25(3)

14(2)

ND3

21(3)

11(1)

Brain

22(4)

42(1)

29(2)

11(1)

95(2)

Muscle

93(3)

Fat

—

28(2)

21(2)

9.0(4)

9.7(4)

11(4)

5.3(4)

3.0(2)

Heart

—

38(1)

Stomach

—

35(1)

106(1)

Liver

70(4)

13(2)

43(3)

20(3)

11(3)

16(3)

26(3)

12(2)

14(1)

Gallbladder

69(4)

63(3)

32(4)

3.4(2)

4.7(3)

6.2(3)

1.4(2)

1.3(1)

Intestine

70(4)

17(3)

42(4)

24(1)

Colon

—

44(3)

26(2)

'Actual nanoliters per gram tissue residues = (benzene in nanoliters) / (tissue wet weight in grams) (mean tissue residue) / 48-h mean
tissue residue) (100% residue).
'The initial mean benzene seawater concentrations (ul/liter) were 0.11 for test 2a, 0.0097 for test 2b, and 0.088 for test 4.
^ND = nondetectable.

549

FISHERY BULLETIN: VOL. 74, NO. 3

BENZENE , INTEGUMENT

GALL BLADDER

(BILE)

-r- STOMACH INTESTINES

METABOLITES

OUT WITH FECES

Figure 3.- Hypothetical pathway and distribution of benzene in

fish.

Figure 3 shows a hypothetical distribution and
pathway of benzene in fish which is substantiated
by our results. Benzene is absorbed across the gills
into the blood where, being lipid soluble, it at-
taches to the erythrocytes and lipoproteins (Ger-
arde 1960). It is then translocated via the blood to
the tissues where it either accumulates or is
metabolized. Parke (1968), Meyers (1970), and Lee,
Sauerheber, and Dobbs (1972) described me-
tabolism of benzene to phenol in the liver of fish
and mammals. The metabolites are excreted from
the liver with the bile and stored in the gallblad-
der. From the gallbladder the bile is excreted into
the intestine and finally eliminated through the
colon with the feces. Our results show high ac-
cumulation in the liver and gallbladder. Lee,
Sauerheber, and Dobbs (1972) also found the
gallbladder of fish to be a storage site for poly-
cyclic aromatic compounds. Our results indicate
that the pathway through the liver, gallbladder,
intestine, and colon is a major depuration route.
These tissues take the longest to accumulate and
depurate. This is probably due to the time needed
for metabolism of benzene. The gill also was one of
the tissues which depurated later. Some un-
changed benzene metabolites are probably ex-
creted across the gills.

The secondary increase in ^^C radioactivity
(days 4-7) observed after initial depuration (days
2-4) in several striped bass tissues (Figure 3) is
difficult to interpret. One explanation may be that
the metabolism of benzene is limited to a certain
rate and that until the initial metabolism is
complete, some benzene accumulates in non-
metabolic tissues and is not totally metabolized
until later. The secondary increase in residues in
fat and brain tissues, however, suggests that
perhaps metabolites such as phenol are ac-
cumulating in the tissues for a period before they
too are depurated. Additional work with uptake

and depuration in herring tissues shows a similar
pattern. Further research must be done to clarify
this point.

The low accumulation tissues such as heart,
muscle, and stomach are also low in lipid content
and apparently do not directly contribute to the
metabolism of benzene. Lee, Sauerheber, and
Dobbs (1972) found similar results with naph-
thalene and benzopyrene in fish. Later work at
Tiburon has demonstrated that little benzene
and/or metabolites accumulate in the kidney
tissue of herring. Because of this and the fact that
fish in salt water excrete little urine, we feel the
kidneys are not a major depuration pathway in
fish from saline waters. Further study of urinary
depuration is needed.

Northern anchovies are schooling fish, and they
swam constantly during the tests-striped bass
were more sedentary. This difference in activity
might explain the higher accumulation in
anchovies.

The short duration of low-level water column
exposures of benzene in these experiments did not
reveal obvious detrimental effects on behavior or
physiology of fish. However, equilibrium ac-
cumulation levels have not been obtained because
of the static exposure with decreasing benzene-
water concentration. During chronic exposures,
higher accumulations of benzene and toxic me-
tabolites (such as phenol) with deleterious effects
are possible. Further, because of the rapid uptake
rate over a wide range of concentrations, it is
conceivable that both species could accumulate
significant benzene levels after brief exposure
during an oil spill. The severity of effects at
chronic and acute levels will depend greatly on the
energy requirements of the fish and the degree of
stress to which they are already subjected. Fish in
spawning condition are particularly susceptible to
additional stress from pollutants (e.g., spawning
Pacific herring [Struhsaker^]). Further study of
uptake in the lipid-rich mature ovaries of fish
should be done.

The rapid depuration of benzene the first day
after exposure ended appears to be due to me-
tabolism and excretion via the liver-intestine
route. Because of this rapid depuration, the pos-
sibility of bio-amplification in fish does not appear

^Struhsaker, J. W. Effects of benzene (a toxic component of
petroleum) on spawning Pacific herring. Manuscr. in prep.
Southwest Fish. Cent. Tiburon Lab., Natl. Mar. Fish. Serv.,
NOAA, Tiburon, CA 94920.

550

KORN ET AL.: UPTAKE AND DEPURATION OF "C-BENZENE

likely, at least after exposure from the water.
Exposure from the ingestion of food organisms
may result in a different metabolic process, how-
ever, and this work should be done before further
conclusions are made. Our results from uptake
studies with a rotifer (Brachionus plicatilus)
(Echeverria^) and those of Lee, Sauerheber, and
Benson (1972) and Lee (1975) with mollusks and
zooplankton indicate that some invertebrates may
be unable to metabolize aromatic hydro-
carbons—accumulating them to very high levels
and depurating them slowly. Fish feeding on such
organisms may be exposed to high and potentially
damaging levels of aromatics.

Additional chronic uptake studies under contin-
uous-flow conditions are needed. Analyses of
metabolites are proceeding and will be reported
later.

ACKNOWLEDGMENTS

We acknowledge the considerable assistance of
other members of the Physiology Investigation,
Tiburon Laboratory, particularly Pete Benville,
Jr., for gas chromatography analyses. We also
thank Stanley Rice, Northwest Fisheries Center
Auke Bay Fisheries Laboratory, NMFS, NOAA;
and Jerry M. Neff, Texas A & M University, for
their critical reviews of the manuscript.

LITERATURE CITED

Anderson, J. W., J. M. Neff, B. A. Cox, H. E. Tatem, and G. M.

HiGHTOWER.

1974. Characteristics of dispersions and water-soluble

•''Echeverria, T. Uptake, storage and depuration of '^C-labeled
benzene in the rotifer, Brachionus plicafilis. Manuscr. in prep.
Southwest Fish. Cent. Tiburon Lab., Natl. Mar. Fish. Serv.,
NOAA, Tiburon, CA 94920.

extracts of crude and refined oils and their toxicity to
estuarine crustaceans and fish. Mar. Biol. (Berl.) 27:75-88.
Benville, P. E., Jr., and S. Korn.

1974. A simple apparatus for metering volatile liquids into
water. J. Fish. Res. Board Can. 31:367-368.

Brocksen, R.W., AND H. T. Bailey.

1973. Respiratory response of juvenile chinook salmon and
striped bass exposed to benzene, a water-soluble compo-
nent of crude oil. In Proceedings of joint conference on
prevention and control of oil spills, p. 783-791. Am. Pet.
Inst., Environ. Prot. Agency, U.S. Coast Guard, Wash.,
D.C.

Gerarde, H. W.

1960. Toxicology and biochemistry of aromatic hydrocar-
bons. Elsevier Publ. Co., N.Y., 329 p.

Korn, S.

1975. Semiclosed seawater system with automatic salinity,
temperature, and turbidity control. U.S. Dep. Commer.,
NOAA Tech. Rep. NMFS SSRF-694, 5 p.

Korn, S., J. W., Struhsaker, and P. Benville, Jr.

1976. Effects of benzene on the growth, fat content, and
caloric content of striped bass, Morone saxatilis. Fish.
Bull., U.S. 74: 694-698.

Lee, R. F.

1975. Fate of petroleum hydrocarbons in marine zooplank-
ton. In American Petroleum Institute, 1975 Conference on
Prevention and Control of Oil Pollution. Proceedings
March 25-27, 1975 San Francisco, Calif., p. 549-553. Am.
Pet. Inst., Environ. Prot. Agency, U.S. Coast Guard.
Lee, R. F., R. Sauerheber, and A. A. Benson.

1972. Petroleum hydrocarbons: Uptake and discharge by the
marine mussel Mytilus edulis. Science (Wash., D.C.)
177:344-346.
Lee, R. F., R. Sauerheber, and G. H. Dobbs.

1972. Uptake, metabolism and discharge of polycyclic
aromatic hydrocarbons by marine fish. Mar. Biol. (Berl.)
17:201-208.
Meyers, F. H.

1970. Review of medical pharmacology. Lange Medical
Publishing Co., Los Altos, Calif., 155 p.
Parke, D. V.

1968. The biochemistry of foreign compounds. Pergamon
Press, Oxf., 269 p.
Snedecor, G. W., and W. G. Cochran.

1968. Statistical methods. 6th ed. Iowa State Univ. Press,
Ames, 593 p.

551

VON BERTALANFFY GROWTH CURVES FOR STRIPED MARLIN,

TETRAPTURUS AUDAX, AND BLUE MARLIN, MAKAIRA NIGRICANS,

IN THE CENTRAL NORTH PACIFIC OCEAN

Robert A. Skillman and Marian Y. Y. Yong^

ABSTRACT

The growth of striped marlin, Tetraptiirus audax, and blue marlin, Makaira nigricans, was described
by fitting von Bertalanffy growth curves (an assumed age model and a length-increment model) to the
progression of age-groups, by quarters, through the Hawaiian longline fishery from 1960 to 1970. For
striped marlin, the sexes grew at about the same rate and had similar predicted asymptotic maximum
fork lengths, 277.4-314.4 cm for males and 288.7-326.2 cm for females. For blue marlin, the sexes grew at
about the same rate until 250 cm. Above this length, the growth rate of males declined and an
asymptotic maximum length of 298.8-368.0 cm was predicted. For females above 250 cm, the growth
continued at a rapid rate; exhibiting little tendency toward an asymptote over the range of ages
available to the study.

While blue marlin, Makaira nigricaus Lacepede,
are important in the U.S. sport fishery in Califor-
nia, Florida, and Hawaii and striped marlin,
Tetraptiirus audax (Philippi), are important in
California and Hawaii, little is known about their
population characteristics or parameters. In par-
ticular, growth of these species has received little
attention. In this paper, the growth of striped and
blue marlins is described by fitting von Ber-
talanffy growth curves to age-groups separated
from length-frequency data collected in Hawaii.

A review of the literature revealed four papers
dealing with the growth of marlins. In them,
growth was examined by plotting the progression
of mean sizes for age-groups separated from
size-frequency data by month or some other time
interval; the fitting of a functional growth model
(e.g., von Bertalanff'y or Gompertz) was not dis-
cussed or attempted. Royce (1957) studied striped
marlin in the Hawaiian longline fishery (1949-52)
and concluded that small (13.6-17.7 kg) and large
(45.3-49.4 kg) size classes grew about 13.6 kg per
year. De Sylva (1957) studied the growth of
sailfish, IstiopJwrus platypterus (Shaw and Nod-
der), in the Atlantic from records obtained
primarily from sport catches. His growth curve,
fitted by eye, showed an extremely rapid rate of
growth: 180 cm total length in the first year of life
and 30 cm in the second. Maximum age was
estimated as IV. De Sylva and Davis (1963) com-

pared data for the white marlin T. albidus Poey,
and the sailfish and concluded that white marlin
live longer than sailfish. Koto and Kodama (1962),
studying the growth of sailfish caught near Japan,
found an annual growth of 35 cm for a 140-175 cm
eye orbit-fork length group, 20 cm for a 176-195 cm
group, and 15 cm for a 196-210 cm group.

The objective of this study was to quantitatively
describe the growth of striped and blue marlins
using a model that adequately followed the ob-
served data and provided estimates of growth
parameters, which could be incorporated into
analytical models of population dynamics. Since
the von Bertalanff'y growth model satisfied these
conditions, it was used in this study. Specifically
for striped marlin, growth parameters were
sought by sex for individual cohorts and then for
data pooled over all years. For blue marlin, growth
parameters were sought by sex only for data
pooled over all years because the data were in-
sufficient to work with individual cohorts. By
pooling across years, the assumption was made
that the populations under study were in or tend-
ing toward a steady state. As such, yearly varia-
tions in mean lengths of age-groups as well as
variations in growth parameters between cohorts
were treated as homogeneous sets of responses to
variations in the environment.

MATERIALS

'Southwest Fisheries Center Honolulu Laboratory, National
Marine Fisheries Service, NOAA, Honolulu, HI 96812.

From April 1960 through April 1970 at the

Manuscript accepted February 1976.
FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

553

FISHERY BULLETIN: VOL. 74, NO. 3

auction blocks of the fresh fish markets in Hono-
lulu, the staff of the Honolulu Laboratory, Na-
tional Marine Fisheries Service, NOAA, collected
weight and sex data from large pelagic species
caught by the Hawaiian longline fleet. Details of
the longline fishery are given by June (1950) and
Otsu (1954). All fish were weighed to the nearest
whole pound. Due to the nearly complete utiliza-
tion of marlin by the dealers, only a small incision
could be made in the abdominal wall in order to
examine the gonads. At best, a small section of
gonad could be cut off and examined; no micro-
scopic determinations were made. Thus, it is possi-
ble that some misidentification of the sex of these
fish occurred, especially preceding sexual maturity
for both species and following spawning for blue
marlin.

METHOD OF ANALYSIS

Briefly, the analyses consisted of 1) transform-
ing the data into usable form by (a) calculating
length-weight relationships using functional re-
gressions (Ricker 1973), (b) converting weights to
lengths, and (c) grouping the lengths by sex,
quarter, year, and length interval; 2) separating
age-groups from the frequency distributions and
estimating their mean lengths; 3) setting up the
progressions of mean lengths and corresponding
age structures; and 4) fitting von BertalanfTy
growth models to the progressions of mean
lengths. Following these steps, tests were per-
formed to determine whether the yearly samples
were homogeneous and could be pooled. These
tests consisted of a series of nonparametric
Friedman two-way analyses of variance (Hol-
lander and Wolfe 1973:139) performed on the
number of age-groups, the mean lengths of age-
groups, and the percent representation of age-
groups separated for each year sampled, as well as
on the growth parameters of the different cohorts.
Also, a sign test (Siegel 1956:68) was used to test
for trends in mean length between sexes; and a
series of one sample runs tests (Siegel 1956:52) was
used to test for trends in mean length among
cohorts. If heterogeneity was not found, the
transformed yearly data were pooled, that is, the
year designation was ignored, and steps 2-4 above
were repeated on the pooled data.

An initial inspection of the blue marlin length-
frequency distributions revealed some unusually
large specimens identified as males weighing up to
328 kg, whereas it had been contended that male

blue marlin in the Atlantic (Erdman 1968) and in
the Pacific (Strasburg 1970) do not exceed about
136 kg. An examination of data collected under
ideal sampling conditions at Hawaiian Interna-
tional Billfish Tournaments revealed only 5 out of
385 individuals in 12 yr that exceeded 136 kg. Four
of these were under 143 kg while one fish weighed
171 kg. On the basis of these data, we accepted
Erdman's and Strasburg's contention as essen-
tially correct, assumed that all males over 143 kg
were misidentified due to the difficult sampling
conditions at the auction sites, and reclassified all
males over 143 kg as females (56 were reclassified
out of 2,710 specimens originally classified as
males).

Transformation of Data

Observed weights were converted from pounds
to kilograms and then to fork lengths in cen-
timeters (tip of bill to middle point on the posterior
margin of the middle caudal rays, FL). Length-
weight relationships used for the latter conver-
sions were calculated as functional regressions
from Skillman and Yong (1974) following the
recommendations of Ricker (1973). Briefly, the
differences between functional and the commonly
used predictive (linear) regressions, which are of
importance to this application, are as follows.
First, the predictive regression applies where it is
hypothesized that one variable is linearly related
to or dependent on a second variable, the in-
dependent variable. Whereas, the functional re-
gression applies where it is hypothesized that two
variables are interdependent, and the effect of one
cannot be disentangled from the effect of the
other. Second, the predictive regression tends to
systematically underestimate the magnitude of
the regression coefl^cient as the sample range
truncates the real range of the variates; the
functional regression does not do so. For striped
marlin, the data were insufficient to calculate
functional length-weight relationships for each
sex; therefore, a single relationship [FL = 73.4429
W 0 28o8) ^^.gg applied to each sex separately. For
blue marlin, separate functional length-weight
relationships were calculated (-^-^female =
65.4502 W 0-3030^ ^nd FLmale = 56.8780 W "-^^is).
As expected, the coefficients of allometry calculat-
ed using functional regressions increased over
those calculated in Skillman and Yong (1974) using
predictive regressions, and the difference between
sexes decreased by 36%.

554

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

In order to efficiently separate age-groups from
the frequency distributions, the data were
grouped and length intervals set up (Simpson et al.
1960). Lengths for striped marlin were grouped by
sex, year, quarter, and 3-cm interval, and these
groupings resulted in a maximum of 96-length
intervals per quarter. Blue marlin lengths were
grouped by sex, quarter, year, and 5-cm interval,
and these groupings resulted in a maximum of
73-size intervals per quarter.

Separation of Age- Groups

The computer program ENORMSEP (Yong and
Skillman 1975) was used to separate the grouped,
length-frequency data into constituent age-
groups and to calculate estimates of the mean
length, variance, percent representation, and
numerical size of the age-groups. Essentially this
computer program automates the Cassie-Harding
probability paper method (Harding 1949; Cassie
1954) and enters intermediate results into
NORMSEP which performs the actual separation
of age-groups.

Progression of Age-Groups

The estimates of mean lengths for age-groups
were plotted by quarter in order to check for
reasonable progression and to assign age. Ages
were assigned by determining the time of peak
spawning, estimating the age at recruitment, and
then merely assigning ages progressively as the
age-groups passed through the fishery.

The time of spawning is not well established for

either striped or blue marlin. For striped marlin,

Nakamura (1949) stated that the time of peak

spawning seemed to be from April to May in the

South China Sea near the Republic of China and

from May to June near the Bonin Islands. Royce

(1957) stated that testes with free flowing milt

were collected in the equatorial central Pacific in

March. Kume and Joseph (1969) estimated, on the

basis of gonad index of females taken in the

eastern tropical North Pacific, that peak spawning

occurs in May and June. From specimens landed in

southern California and northern Mexico, El-

dridge and Wares (1974) indicated that gonad

index was highest in June and July, but they did

not have samples for August or September. Hence,

we took June 1st as the time of peak spawning and

assigned an age of 1.46 yr to the 151-cm male and

152-cm female age-groups recruited in the fourth

quarter.

For blue marlin, Royce (1957) stated that males
with free flowing milt were collected from Feb-
ruary through October in the equatorial Pacific,
and cited Nakamura (1942) as indicating that
spawning occurs east of Luzon (Philippines) from
May to July. Kume and Joseph (1969), on the basis
of gonad index of females taken in the eastern
tropical Pacific, concluded that spawning occurs in
December and January; however, most of their
samples were collected from south of the equator.
We arbitrarily took June 1st as the time of peak
spawning and assigned an age of 0.71 yr to the
55.5-cm female age-group recruited in the first
quarter.

Von Bertalanffy Growth Model

Two computer programs, BGC3 and BGC4,
assembled by Abramson (1971) and written by
Patrick Tomlinson were used in this paper to
obtain estimates of von Bertalanffy growth pa-
rameters. The computer program for model 1,
BGC3, fits the von Bertalanffy model by the least
squares method to lengths from fish of known or,
in this case, assumed age. The basic model is the
familiar equation

A = L^a-e

KU - („)

o))

(1)

where L , = length at age f

Loo = a parameter depicting asymptotic

maximum length
K = a. parameter indicating the rate of pro-
portional growth
f^^ = a parameter depicting the theoretical
age at which the fish has zero length
given the adult growth form.

The computer program for model 2, BGC4, a
version of the size-increment method proposed by
Fabens (1965), fits the von Bertalanffy model by the
least squares method to observed lengths, using
data on growth increment in known time intervals
but making no assumptions about absolute age.
Parameter estimates using this method are in-
cluded in the tables so that any future estimates of
striped or blue marlin growth from tagging data
can be compared directly to our results. For model
2, the von Bertalanffy model is written as

L, =L„(l-6e-'^')

(2a)
555

FISHERY BULLETIN: VOL. 74, NO. 3

where b = 1 --— ^ = a parameter depicting the

°° theoretical proportion of

potential growth in length

that occurs after hatching;

or by substituting t = t + /M and be''"' = 1 - =^

into Equation (2a) ""

A+ A, =1,6-"^' + L^a-e"'^') (2b)

where Af = time increment between points
of measured length.

GROWTH OF STRIPED MARLIN

Results— Analysis of Cohorts

Age-groups were successfully separated by sex
and by quarter, within years, using the computer
program ENORMSEP (Table 1). In general, the
mean length estimates for females were slightly
larger than those for males of the same age-
groups. Quantitatively, the goodness of fit of the
separated age-groups to the observed frequency
distribution can be assessed with the chi-square
values in Table 1. The largeness of the chi-square
values indicated poor fit, but it was found that the
tails of the distribution, having frequencies too
small for the separation of age-groups, contribut-
ed disproportionately to the total chi-square value.
Qualitatively, the goodness of fit was deemed
adequate for the following reasons. In all years
there was close agreement between sexes in the
number of age-groups separated within quarters:
approximately 2, 3, 4, and 3 age-groups for the
first through fourth quarters, respectively. There
was also close agreement among years and
between sexes in the mean lengths and length
composition of age-groups within quarters.

The progressions of mean lengths were set up as
depicted by the connected open circles in Figure 1.
In the third quarter of every year for both sexes,
there was an age-group with mean length of about
167 cm that did not fit into the progression of
age-groups. By assigning the same assumed age to
this age-group as to a similar size group in the first
quarter, and allowing the same time between
spawning and the attainment of this age, this
age-group could have resulted from a spawning in
January. On the basis of gonad indices, Kume and
Joseph (1969) believed that striped marlin from
the eastern South Pacific spawn from November to

December, and Royce (1957) indicated that striped
marlin with ripe gonads have been collected in
February in the equatorial region of the central
Pacific. Hence, we concluded that this age-group
belonged to a different spawning stock and should
not be used in the calculation of the von Ber-
talanffy growth curves for the central North
Pacific stock. Also, for females there were two
age-groups having mean lengths of 267.5 and 200.0
cm in the third quarter 1964 and 1966, and for
males there were four age-groups having mean
lengths of 204.6 and 272.0 cm in the third quarter
of 1968, 271.8 cm in the fourth quarter 1968, and
266.0 cm in the second quarter 1969. These could
not be assigned with certainty to any cohort.

Several qualitative aspects of the observed
growth of the cohorts were noteworthy (Figure 1).
First, there seemed to be a cyclical pattern in the
mean size at recruitment but no upward or down-
ward trend. Second, the progression of age-groups
during the first year in the fishery was fairly
smooth and consistent between cohorts. Third,
after about a year and a half in the fishery, there
seemed to be a regression or slowing down in the
apparent growth that persisted for two or three
quarters. Fourth, the mean length of the last
age-group in each cohort varied considerably.

There were a sufficient number of age-groups
for the 1959-65 cohorts to fit a von Bertalanffy
growth curve but not for the 1957-58 and 1966-68
cohorts. The calculated growth curves were shown
as smooth curves in Figure 1. As expected from the
variation shown in the progression of the observed
mean sizes in Figure 1, the standard errors of
estimates were moderate, and there was variation
in parameter estimates between cohorts (Table 2).

To investigate the variation in estimates of
mean length for age-groups shown in Table 1 and
in growth parameters shown in Table 2, a series of
nonparametric Friedman two-way analyses of
variance was performed. No difference in the
number of age-groups by sex, quarter, or sample
year could be demonstrated (all 5 test statistics
were insignificant with probability P>0.05, Hol-
lander and Wolfe 1973:139). In testing the effect of
sex on the mean length of age-groups within
quarters, three significant effects (S = 5.44,
P50.05; S = 8.99, PsQ.oi; and S = 5.44, Ps0.05)
were found out of 16 comparisons. However, using
a sign test for trends in mean length between
sexes showed that females tended to be larger
than males (test statistic P = 0.00005, or
probability P^O.Ol, for data from all years). In

556

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

Table 1. -Estimates of striped marlin age-groups (numbers do not indicate age) by sex and by year and (juarter of sam])ling. Estimates of
mean forl< length.FL; percent of total sample comprising a particular age-group, %; and chi-stiuare goodness of fit value. ,\--', were obtained
from the computer program ENORMSEP.

Age-group 5

1960

1962

1963

1965

1966

1967

1968

Year Quarter Sex FL % FL % FL % FL % FL % X^

Age-group 1

Age-g

roup 2

Age-group 3

Age-

group 4

Sex

150.9

3.4

205.7

86.6

220.8

10.0

—

154.0

3.4

208.2

85.5

228.7

11.1

—

160.2

50.4

208.7

49.6

—

162.2

65.2

213.5

34.8

—

176.5

27.5

215.7

72.5

—

172.5

26.2

223,5

56.0

231.1

17.8

—

161.1

18.7

184.7

26.9

203.3

14.8

223.6

39.6

159.3

39.4

197.4

17.6

224.2

31.6

247.4

11.4

156.7

11.2

207.8

72.8

220.6

16.0

—

158.0

13.4

211.1

71,8

225.8

14.8

—

175.8

49.4

211.9

50.6

—

170.6

45.9

213.7

54.1

—

181.7

15.9

215,2

81,0

238.4

3.1

—

175.9

17.6

217,9

40.4

232.3

42.0

—

169.0

28.7

191.4

37.1

213.4

30.7

225.0

3.5

170.3

39.9

193.5

41.5

212.6

18.6

—

149.7

45.4

209.1

49.4

218.8

5.2

—

150.2

47.1

209,6

48,4

241.1

4.5

—

162.9

76.6

210.3

23.4

—

163.1

80.2

210.7

19,8

—

173.9

35.4

217.8

62.8

230.7

1.8

—

168.4

32.9

225.4

33.4

228.8

33.7

—

171,0

3.4

187.2

19.9

217.4

76.7

—

173.0

9.0

194.8

33.0

219.9

21.0

239.5

37.0

151.0

14.7

202,0

59.0

216.5

26.3

—

153.4

24.3

206.2

53.5

219.0

22.2

—

173.5

76.7

206.3

23.3

—

173.7

69.3

198.7

30.7

—

186.3

38,8

209.7

52.8

227.3

8.4

—

179.5

35,5

216.9

56.0

242.1

8.5

—

164.3

26.0

191.3

21.3

205.3

49.1

228.7

3,6

168.8

48.7

192,3

10.3

215.2

39.4

267,5

1.6

160.7

1.9

203,5

94.2

217.2

3.9

—

158.6

0.8

206,0

91.5

222.2

7.7

—

159.8

3.2

203.2

96.8

—

164.1

4.7

206,8

95,3

—

172.7

7.5

212,5

90,9

232.4

1.6

—

171.4

11.5

220,3

84,7

235.0

3.8

—

164.6

27.0

193.8

18.3

211.7

40.2

230.6

14.5

164.3

29,8

206.1

40.4

224.0

28.0

248.0

1,8

144.5

3.0

205.9

88.5

225.5

8.6

—

148.4

4.4

207.9

83.9

222.8

11.7

—

158.6

14.3

207.7

85.7

—

165.1

18.2

212,1

79.0

250.5

2.8

—

182.3

23.6

215.5

76.0

249.6

0.4

—

177.7

25.7

220,8

64,8

244.0

9.5

—

171.3

17.0

198.6

69.2

218.6

10.8

234.6

3.0

193.8

74.2

200.0

16.5

209.8

2.8

240.8

6.5

147.9

3.4

206,5

94.0

236.3

2.6

—

131.0

3.6

208.8

84.0

214.6

12.4

—

159.4

6.2

204.3

93.8

—

165.7

4.3

207.8

95.7

—

171.9

1.5

208.2

97.8

233.0

0.7

—

174.1

5.5

221.4

87.3

249.1

7.2

—

168.3

27.7

192.8

53.7

213.4

3.5

232,5

15.1

166.9

30.1

185.1

25.6

204.5

44.3

—

168.7

0.7

202.7

93.4

218.9

5.9

—

162.8

0.4

202,8

97.4

225.8

2.2

—

161.7

1.6

200.7

98.4

—

165.6

3.1

202.3

94.8

242.6

2.1

—

174.6

2.2

208.1

97.5

242.0

0.3

—

182.0

2.3

215.8

97.3

222.7

2.4

—

169.0

9.4

194.7

63.7

217.1

16.0

204,6

6.7

164.5

9.5

193.8

63.6

213.4

19.5

235.1

7.4

157.7

5.3

205.5

93.9

271.8

0.7

—

153.9

8.0

209,4

89.3

242.5

2.7

—

^~"

— — 64.2

— — 44.7

1961 1 M 160.2 50.4 208.7 49.6 — — _ _ _ _ 16.4

— — 32.1

— — 70.6

— — 39.6

— — 20.8

— — 35.1

— — 8.6

— — 28.0

— — 18.7

— — 32.5

— — 47.7

— — 28.5

— — 24.8

— — 24.2

— — 79.5

— — 67.3

— — 61.9

— — 47.1

— — 66.2

— — 61.8

— — 10,8

— — 12.4

— — 32.0

— — 58.6

1964 1 M 173.5 76.7 206.3 23.3 — — _ _ _ _ 77.4

— — 97.4

— — 63.2

— — 35.6

— — 15.3

— — 43.6

— — 99.8

— — 137.1

— — 49.3

— — 44.6

— — 34.3

— — 22.0

— — 6.7

— — 26.4

— — 47.3

— — 117.9

— — 37.0
_ — 30.0

— — 32.1

— — 26.4

— — 15.5

— — 21.1

— — 142.0

— — 52.4

— — 44.8

— — 40.1
_ — 24.8

— — 60.2

— — 10.5
_ — 9.8

— — 82.9

— — 98.4

_ — 37.6

_ — 48.4

— — 58.1

— — 45.6
272.0 4.2 30.9

_ — 18.4

_ — 79.4

— — 66.4

557

FISHERY BULLETIN: VOL. 74, NO. 3

Table 1. — Continued.

Age-

group 1

Age-

group 2

Age-

group 3

Age -

group 4

Age-group 5

Year

Quarter

Sex

FL %

1969

157.8

8.9

203.5

91.1

31.1

156.6

12.6

206.6

87.4

—

— —

32.2

179.0

9.1

211.8

90.6

266.0

0.3

—

— —

34.1

182.0

0.4

212.0

97.0

221.9

2.6

—

— —

39.4

161.0

0.5

195.4

62.2

220.0

25.2

235.9

12.1

— —

15.5

174.7

8.3

196.8

71.2

221.0

13.8

241.4

6.7

— —

40.2

154.5

5.8

206.6

92.5

233.0

1.7

—

— —

33.9

155.3

4.8

206.0

93.4

236.2

1.8

—

— —

70.9

1970

164.1

45.3

210.4

54.7

—

— —

28.2

164.2

47.4

208.2

52.6

—

— —

35.9

testing significance of differences between the
growth parameters, the analysis of variance failed
to demonstrate an effect of either cohort or sex on
Loo, K, or to (all 5 values insignificant with P>0.05).
We concluded that there were no significant

effects of either cohort or se.x on the number of
age-groups or growth parameters, and that
females tended to be statistically larger at suc-
cessive ages even though the magnitude of the
differences was not significant.

Figure 1. -Striped marlin von Bertalanffy
growth curves (heavy lines) by sex for the
1959-65 cohorts. Observed mean lengths for
age-groups, represented by circles con-
nected with light lines, were used in fitting
the growth curves.

q:
o

E
u

I I I I I I I

558

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

Table 2.-Striped inarlin von Bertalanffy growth parameters hy
cohort and sex. The parameter estimates, L^ (asymptotic
maximum fork length), K (rate of proportional growth), and („
(theoretical time at which the tish would have zero length), are
given for model 1 (upper row) and mode! 2 (lower row).

Year-

Standard error

class

Sex

i-x (cm)

t, (yr)

of estimate

1959

Male

241.0

0.812

0.297

5.1

251.1

0.651

6.1

Female

245.0

0.809

0.294

6.6

255.1

0.626

7.3

1960

Male

233.7

0.929

0.260

3.6

233.5

0.986

4.7

Female

261.7

0.584

-0.096

6.0

290.0

0.437

8.0

1961

Male

247.0

0.640

0.007

4.6

250.9

0.630

5.3

Female

283.3

0.428

-0.304

6.0

265.7

0.574

8.6

1962

Male

274.4

0.448

-0.435

6.8

248.2

0.873

9.0

Female

265.7

0.554

-0.125

3.7

261.9

0.664

6.0

1963

Male

256.0

0.622

-0.004

6.2

244.2

0.655

7.9

Female

245.5

0.729

0.091

9.1

258.7

0.607

10.9

1964

Male

243.7

0.818

0.351

7.1

275.5

0.570

6.7

Female

243.3

0.803

0.282

7.0

234.9

1.067

9.8

1965

Male

253.3

0.630

0.086

4.3

258.1

0.585

5.3

Female

239.8

0.914

0.542

6.5

231.5

1.267

8.9

Results— Analysis of Pooled Data

In the preceding section, it was shown that there
were no demonstrable differences in the growth
parameters between different cohorts; hence, the
estimates for L^, K, and /q could be averaged to
provide a pooled estimate. In addition, it was not
possible to consistently show significant differ-
ences between quarters or years in the number of
age-groups separated (S= 11.78, P>0.05), the mean
lengths of age-groups (5= 15.51, PsO.05, for males
in third quarter, all seven remaining 5 values had
P>0.05), or the percent representation of age-
groups (S= 57.18, PsO.Ol, for males in the second
quarter, S = 49.57, P^O.Ol, for females in the third
quarter, all six remaining S values with P>0.05).
Neither was it possible to show a trend in mean
lengths among cohorts using a series of sample
runs tests (one test out of 18 deviated from
random at the 0.05 level). We interpreted these
results to mean that the yearly samples were
homogeneous and that at least approximately a
steady state existed. Therefore, the yearly
frequency data were pooled; and ENORMSEP was
used to separate age-groups.

The mean lengths and percent representation of
age-groups by quarter (Table 3) were quite similar
to the values found using the yearly data (Table 1).
Quantitatively evaluating the goodness of fit, the
chi-square values were found again to be rather
large. Qualitatively, however, the shape of the
frequency distributions was consistent between
sexes within quarters and generally so between
quarters within each sex (Figure 2). The plots for
the third quarter, the off-season for striped marlin
in Hawaii, did not show much of a pattern at all.
Also, the shapes of the plots for the pooled data
analysis were similar to those for the analysis of
individual cohorts (not shown).

Mean lengths for females and males exhibited a
fairly smooth progression (Figure 3). As in the
analysis of cohorts, there was an age-group with
mean length of about 167 cm in the third quarter
that did not fit into the progression of age-groups.
Again, it was assumed that this age-group
belonged to a different spawning stock and should
not be used in the calculation of the growth curve.
For females, there were 11 age-groups in the
progression whereas there were 10 or 12 but never
11 for the analysis of cohorts. For males, there
were 12 age-groups in the progression whereas
there were 9 to 11 in the analysis of cohorts. The
smallest fish were recruited into the fishery in the
fourth quarter and progressed through the fishery
until the largest fish passed from it in the third

Table 3.-Statistics for striped marlin age-groups by quarter an^d
sex, pooled over all years. Estimates of mean fork length, FL\
percent representation of the age-group, %; the numerical
sample for size of the age-group, n; the total sample size, A^; and
the chi-square goodness of fit value, x^, were obtained from the
computer program ENORMSEP.

Age-

Male

Female

Quarter

group

166.6

36.2

1,480

167.1

42.6

1,568

204.7

63.8

2,613

207.2

57.4

2,115

4,093

3,683

114.1

183.8

177.3

17.5

1,020

174.4

21.0

852

212.5

81.5

4,746

219.7

69.7

2,827

228.0

1.0

238.4

9.3

377

5,823

4,056

97.2

54.0

167.1

17.0

178

166.7

22.7

190

195.3

52.6

551

194.0

50.1

419

215.2

19.9

208

217.5

19.4

162

231.6

9.1

236.2

7.8

260.0

1.4

—

1,047

836

71.6

45.1

151.3

7.1

421

152.2

8.7

431

204.7

87.0

5.162

207.0

83.0

4,113

222.6

5.9

352

222.7

8.3

413

5,935

4,957

130.1

192.3

559

MALE

FISHERY BULLETIN: VOL. 74, NO. 3

FEMALE

400

300

>-
o
z
^ 200

100

700

T 1 1 1 1 I I I r

FIRST QUARTER

1 I r

-| 1 1 1 1 r

>
o

100

1 1 1 1 1 1
- THIRD QUARTER

1111

1 1 I 1 1

1 1 oottP*'! ^.X^

>°^o^o„

t^AOes fiQD an 1 1

n 1 1 r

I 1 I

>-
O

700

600

500

400-

300-

200

100

125 150 175 200 225 250 275 300 125 150 175 200 225 250 275 300

LENGTH INTERVAL (cm) LENGTH INTERVAL (cm)

Figure 2.-Striped marlin length composition by sex and quarter for the analysis of pooled data. The smooth curves represent
age-groups separated by the computer program ENORMSEP, and circles represent observed values.

560

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

270
260
250
240

230
220
210
200
190
180
170
160
150
140
130

-; \ 1 1 1 1 \ 1 1 1 1 r

MALE(«) X

FEMALE (o)

MALE(>)

_i I i_

-I I I L.

1.0 15 20 2.5 30 3.5 40 4.5 5.0 55

AGE IN YEARS

Figure 3.— Striped marlin von Bertalanffy growth curves by sex,
for the analysis of pooled data. Male and female growth curves
having the same length were obtained using 11 age-groups while
the longer male curve was obtained using in addition an
age-group 1 yr older. Observed mean lengths for male and female
age-groups are given. As explained in the text, the outliers at age
2.46 yr were not used in fitting the growth curves.

quarter 3 yr later for females and 4 yr later for
males.

Discussion

We felt that the parameter estimates from the
analysis of pooled data provided better estimates
of population parameters than those from the
analysis of individual cohorts because pooling
smooths out variation in individual curves. In
addition, the less well-represented age-groups,
having small and large mean lengths and being
recruited to or escaping from the fishery, were
estimated more accurately given the larger sam-
ple size after pooling.

Since estimates of mean lengths at age were
used in the fitting of the growth curves, estimates
of growth parameters are in terms of these aver-
age values. Likewise, estimates of length at age
derived from these models (Table 4) will actually
be average values and should only be calculated for
the range of ages used in fitting the models. In
fact, the greatest utility of these models will be to
predict length at ages within the range of the
observed data. The accuracy of the estimation of
L^ is dependent on the range of values used, and
it should be remembered that the data used here
included fish only up to the onset of sexual
maturity.

Table 4.-Striped marlin von Bertalanffy growth parameters by
sex for analysis of pooled data. The parameter estimates, L„
(asymptotic maximum fork length), K (rate of proportional
growth), and ^o (theoretical time at which the marlin would have
zero length) are given for model 1 (upper row) and model 2(lower
row).

Standard

error of

Sex

Case

Loo (cm)

fc (yo

estimate

Male

All age-groups

277.4

0.417

-0.521

5.5

314.4

0.315

5.5

Male

Less oldest

239.7

0.810

0.235

2.8

age-group

240,0

0.809

3.7

Female

All age-groups

251.0

0.696

0.136

4.2

251.8

0.709

6.2

The estimation of population L^^ was complicat-
ed by the fact that males and females were not
represented in the fishery for the same length of
time. Obviously, the length of time (number of
age-groups) that the sexes were in the fishery had
an eff"ect on the estimation of L^o, as well as K and
^0 (Table 4). We believed that the parameter
estimates for males using all 12 age-groups
provided the most accurate estimates because a
greater part of the growth curve was measured.
We must admit that the estimates may not be very
precise because the estimated sample size of the
last age-group was 14 individuals, and the stan-
dard error of the estimate for the growth curve
was larger when all age-groups were included. The
estimates of L^ for males using models 1 and 2
were 277.4 and 314.4 cm, respectively. In order to
obtain estimates for females that can be compared
to those for males, the differences found between
females and males using 11 age-groups, 11.3 and
11.8 cm for models 1 and 2, respectively, were
added to the estimates of L^ for males, giving
288.7 and 326.2 cm, respectively. These estimates
seemed reasonable when compared to the largest
striped marlin measured by personnel from the
Honolulu Laboratory: 296 cm for males, 305 cm for
females, and 310 cm for sex undetermined.

Both the analysis of cohorts and the pooled data
analysis (though less so) were plagued by apparent
negative growth or at least by very slow growth
during some quarters about a year and a half after
recruitment. Accepting the general growth pro-
gression as valid, this problem probably biased the
estimates of L^ downward and K upward and
contributed to the size of the standard errors of
estimates. We do not believe that this period of
apparent negative growth resulted from some
physiological change in the form of growth for
which the von Bertalanffy model could not account

561

FISHERY BULLETIN: VOL. 74, NO. 3

and, therefore, is invalid, but rather that the
apparent change in growth was caused by seasonal
changes in availability of the stock due to some
seasonal size-age migration phenomenon or possi-
bly by changes in the selectivity of the fishery.
Preliminary examination of Japanese longline
statistics suggested that the striped marlin stock
available to the Hawaiian fishery shifts its center
of density northward in the months with the
warmest water temperature and becomes less
available to the local fishery (Heeny S. H. Yuen,
Southwest Fisheries Center, pers. commun.). The
decreased growth most commonly seen in the third
quarter might then be due to smaller fish being
associated with the periphery of the stock.

There are of course other possible e.xplanations.
For example, one reviewer suggested that this
period of slow or negative growth represented an
asymptote followed by the initiation of a new
growth phase. Fitting a two-cycle Gompertz
growth curve to the pooled data, this reviewer
found both sexes tending toward an asymptote at
age 2.46 yr followed by another growth phase
where females tended toward an asymptote at
320 cm, but no solution was found for males. Such
changes in growth phase are common at sexual
maturity and at other times when body form
changes. Change in growth form is commonly
accompanied by a corresponding change in the
length-weight relationship, and Skillman and
Yong (1974) found no indication of a change in the
length-weight relationship over the range
142.2-310.1 cm. Also, the age of fish in the
Hawaiian fishery, having a calculated mean length
corresponding to length at first maturity found by
Eldridge and Wares (1974) and Kume and Joseph
(1969) for the eastern tropical Pacific, was 4.2 yr.
This age is nearly double the hypothesized age of
first asymptotic growth. Thus, while it is possible
to fit a segmented growth curve to the data,
biological evidence given above does not support
such a procedure.

Another possible explanation involves the
separation of age-groups. The aberrant growth
occurred most frequently in the third quarter, and
since the sample size was smallest in this quarter,
the precision of the estimates is probably less than
for the other quarters. However, the aberrant
growth did not always occur in this quarter, and its
repeated occurrence among cohorts suggested
that it was real and not an artifact of the estima-
tion procedure per se. With any probabilistic

means of separating age-groups from a mixed
distribution, there is always the danger that
age-groups from diff"erent cohorts of the same
spawning stock will be so similar in size that they
cannot be separated, especially with increasing
age and varying growth rates of the cohorts. We
acknowledge that this may be a problem, but if it
is, it would seem from Figure 1 to be more impor-
tant for the growth period following the period of
aberrant growth. This problem would be increased
if there were more than one spawning stock
involved, and this seems to be the case for some
quarters. In spite of the small sample sizes in the
third quarter, there seemed to be little doubt that
the 167-cm age-group was real, since its mean
length is quite removed from that of the next
age-group at about 194 cm and since the age-group
was found for the pooled data analysis and for 8
out of 9 yr for females and for all years for males in
the analysis of yearly data. Because the two
spawning stocks would continue to have quite
different lengths for the next couple of quarters
and no comparable age-groups were separated in
these quarters, it was reasonable to assume that
this other secondary stock was not represented in
the catches in the subsequent fourth, first, and
probably second quarters. But what about the
following third quarters? If similar growth curves
are assumed for this other stock, then the 200.0-cm
female age-group in the third quarter of 1966 and
the 204.6-cm male age-group in the third quarter
of 1968 could also belong to the secondary spawn-
ing stock. If this secondary stock was present in
other years but in numbers too small to be separ-
ated out, it would tend to bias downward the
estimates of the similar-sized age-group of the
primary spawning stock. With the accuracy of the
present set of data, it is impossible to comment on
the likelihood or importance of this possibility.

The occurrence of these age-groups at approx-
imately 167 cm in the third quarter presents an
additional problem. Where do they come from? If
the male and female growth curves are used to
back calculate the probable time of spawning for
the age-group at approximately 167 cm in the
pooled data analysis, January is estimated as the
time of peak spawning. We hypothesize that these
fish could come from a stock spawning in the
equatorial region, probably north of the equator,
during months corresponding to the southern
summer. It is hard to visualize a hypothesis that
would account for a stock spawned 6 mo out of

562

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

phase migrating poleward, but at lower latitudes,
at the same time as the primary stock. Possibly
these fish associate with blue marlin of about the
same size that migrate into Hawaiian waters in
the third quarter.

Estimates of von Bertalanffy growth pa-
rameters for both sexes were first obtained using
for females and using 12 and 11 (deleting oldest)
age-groups for males (Table 4, smooth curves in
Figure 3). The standard errors of estimates were
slightly smaller than those for the individual
cohorts but still not what could be considered
small. When the oldest age-group for males was
deleted from the calculations, L^ for females was
11 to 12 cm greater than for males. Using all of the
age-groups for males, the estimate of L^ increased
substantially.

Although this paper deals with growth, the
length composition and age of striped marlin as
found in this study have some relevancy to the
problem of migration. First, Matsumoto and
Kazama (1974) hypothesized that striped marlin
migrate out of Hawaiian waters to spawn, most
likely to the western North Pacific. The calculated
mean length of the last female age-group found in
the Hawaiian fishery (age 4.2 yr) corresponded to
the length at first maturity found by Eldridge and
Wares (1974) and Kume and Joseph (1969) for the
eastern tropical North Pacific. Thus, our data
established that as striped marlin reached the
length corresponding to sexual maturity, they
became unavailable to the local fishery. Second,
Kume and Joseph (1969) indicated that there was a
tendency for average length to increase in the
southern areas of the Japanese longline fishery in
the eastern tropical North Pacific, and it seemed to
us from their charts that there was also a western
component to the increasing average lengths.
Eldridge and Wares (1974) believed that maturing
striped marlin moved out of the range of the sport
fisheries based in southern California and Mexico;
and Squire (1974) suggested that the movement of
striped marlin away from the Baja California area
might be to the area of the Revilla Gigedo Islands
where fish with ripe gonads have been collected
and where behavior suggestive of spawning ac-
tivity has been observed by the Japanese. While
the range of our length data was similar to that
found in the eastern tropical Pacific, the last
age-group recognizable in our data comprised less
than 10% of the total frequency whereas similar-
sized fish seemed to be well represented in the

southern and western areas of the eastern tropical
North Pacific longline fishery. Thus, it seems
apparent that the fish leaving the fishery off the
American coast do not migrate through the
Hawaiian fishery in any appreciable numbers.
However, the capture of a striped marlin, tagged
off Baja California, 322 km southwest of the
Hawaiian Islands indicates that some eastern
Pacific fish move into the vicinity of the Hawaiian
Islands. Finally, our analyses do not provide any
information on the direction of emigration from
the Hawaiian fishery.

GROWTH OF BLUE MARLIN

Results— Analysis of Pooled Data

The number of age-groups, as separated by the
computer program ENORMSEP, varied from
three in the third quarter for males to as many as
eight in the first quarter for females (Table 5). The

Table 5.-Statistics for blue marlin age-groups by quarter and
sex for analysis of pooled data. Estimates of mean fork length,
FL\ percent representation of the age-group, %; the numerical
sample size for the group, ti; the total sample size, A^; and the
chi-square goodness of fit value, x". were obtained from the
computer program ENORMSEP.

Age-group

Male

Female

Quarter

123.0

3.3

55.5

0.6

172.7

8.7

145.9

1.8

225.0

57.2

190.5

3.0

240.5

20.0

232.8

7.4

281.8

10.8

286.8

31.5

—

333.5

32.9

—

366.1

22.2

—

415.5

0.6

166

16.8

33.2

180.5

0.3

205.7

5.9

220.0

65.4

172

298.3

76.2

365

250.1

31.7

345.4

13.0

278.0

2.6

377.2

4.9

264

479

22.0

34.1

163.6

1.4

158.8

0.4

227.7

77.6

486

213.5

7.3

255.8

21.0

131

292.6

53.1

352

327.1

24.2

160

362.9

15.0

100

626

663

33.2

38.5

175.5

8.4

101.7

0.9

228.7

80.7

304

180.8

9.2

264.0

10.4

225.1

11.3

285.5

0.5

280.0

33.5

151

307.6

14.2

—

342.1

27.4

123

390.8

3.5

377

449

28.2

69.3

563

FISHERY BULLETIN: VOL. 74, NO. 3

shape of the frequency distribution differed sub-
stanially between sexes (Figure 4). The 240.5-cm
male age-group separated in the first quarter that
had zero variance (Table 5) was regarded as false
and was not used in subsequent calculations.
Quantitatively, the chi-square values do not in-
dicate very good fit; however, as was the case with
striped marlin, the tails of the frequency distribu-
tions, having frequencies too small for the separ-
ation of age-groups, contributed disproportiona-
tely to the total chi-square value. Qualitatively, as
can be seen from Figure 4, the shape of the
frequency distribution was similar from quarter
to quarter, especially for males.

Progressions of mean lengths were set up for
males and females as depicted in Figure 5. The
smallest blue marlin recruited into the fishery in

the first quarter were females, with males being
recruited 1 yr later. Males were present in the
fishery for 3% yr and females for 7 yr. Several
age-groups represented by one or two individuals
were separated for both sexes. The existence of
these age-groups was tentatively accepted, but the
accuracy of their estimated mean lengths was
viewed with skepticism in calculating growth
parameters. The mean length estimates of male
and female age-groups were in close agreement
until about 250 cm {S = 0.50, P>Om). From 250 to
300 cm, the mean lengths for female age-groups
were larger than estimates for males. Above 300
cm, only female age-groups were found, and these
formed an irregularly increasing progression.

Estimates of von Bertalanffy growth pa-
rameters for both sexes were first obtained using

MALE

Figure 4-Blue marlin length
composition by sex and
quarter for the analysis of
pooled data. The smooth
curves represent age-groups
separated by the computer
program ENORMSEP, and
circles represent observed
values.

>-
o

FIRST QUARTER

FEMALE

- 1 ° I

1 1 ' \ 1 : 1 r

SECOND QUARTER "o

100 150 200 250 300 350

LENGTH INTERVAL (cm)

100 150 200 250 300 350 400 450

LENGTH INTERVAL (cm)

564

SKILLMAN and YONG: GROWTH CURVES FOR TWO MARLINS

450

400

350

-~ 300

250

g 200
O

ISO

100

T I I r

FEMALE (o)

MALE (O

qI 1 1 1 L

-J 1 1 I I I I 1 l_

AGE IN YEARS

Figure 5.-Blue marlin von Bertalanffy growth curves by se.x for
the analysis of pooled data. Observed mean lengths for female
and male age-groups are given.

all age-groups, and then using only those with
estimated numerical representations greater than
two individuals. Further analyses were done using
age-groups for females, over the same age span as
for males, wnth numerical representation greater
than two individuals (Table 6). For males, pa-
rameter estimates were similar using all age-
groups and those age-groups represented by more
than two individuals. The standard error of es-
timate was smaller for the latter than it was for
the case using all age-groups. For females, again
the parameter estimates were similar for the two
data sets, and the standard errors of estimates did
not change appreciably. The estimates for L^ were
nearly doubled those for males. In addition, the
estimate of L^ for females, using the same age-
groups as for males with age-groups represented
by more than two individuals, was nearly 3V2 times
that for males.

Discussion

The estimates of von Bertalanffy growth pa-
rameters for male blue marlin differed little
whether all of the age-groups were used or
whether the less well-represented age-groups
were deleted (Table 6). Because the standard
errors of estimates were generally smaller for the
reduced data sets, we felt that these fits provided

Table 6.- Blue marlin von Bertalanffy growth parameters by se.x
for the analysis of pooled data. The parameter estimates, L^
(asymptotic maximum fork length), K (rate of proportional
growth), and ^,( theoretical time at which the fish would have zero
length) are given for model 1 (upper row) and model 2 (lower
row).

Sex

Case

Lx (cm)

fo (yr)

Standard
error of
estimate

Male

All age-groups

371.1
282.3

0.285
0.815

0.106

12.7
18.6

Male

Age-groups with
more than two
Individuals

368.0
298.8

0.315
0.560

0.390

9.9

15.0

Female

All age-groups

659.1
807.8

0.116
0.091

-0.161

10.2
13.8

Female

Age-groups with
more than two
individuals

626.6
540.2

0.123
0.175

-0.202

9.1
14.0

Female

Same age-groups
as for males with
more than two
individuals

1,248.1
875.2

0.048
0.086

-0.674

4.0
5.2

better estimates of parameters. Although the
standard errors of estimates were larger than
desirable, they varied from less than 1% to only 7%
of the estimated Lr^. Thus, the von Bertalanffy
growth model described the data satisfactorily.
The mean length estimates for the poorly repre-
sented age-groups, which were the youngest and
oldest in our samples, should be viewed as
approximate.

For males, estimates of L^, 368.0 and 298.8 cm
for models 1 and 2, respectively, bracketed the
commonly accepted asymptotic length of about
300 cm. If our assumption of a knife-edge limit of
143 kg (approximately 300 cm) for males was
incorrect, the progression of age-groups would
have been expected to increase in length up to this
point without approaching an asymptote. Since an
asymptote was found, we felt our assumption was
valid.

For females, the von Bertalanffy growth curves
seemed adequate for describing the data, but the
estimates of growth parameters were not
biologically reasonable. Using the same range of
age-groups as used for males, the estimates of Lr^
were around 1,000 cm, confirming the visual im-
pression that there was little tendency towards an
asymptote over this range of ages. Using all of the
data, estimates of L^ were 626.6 and 540.2 cm for
models 1 and 2, respectively, or approximately
1,729 and 1,060 kg, respectively. While these
results suggested that there was some tendency
towards an asymptote, which is not visually ap-
parent in the data, we do not believe that enough
older age-groups were included in the regressions

565

FISHERY BULLETIN: VOL. 74, NO. :?

to obtain valid estimates of L^ or K. Among
fishery biologists, there seems to be less of a
consensus on the maximum size of females than of
males, but generally it is thought that female blue
marlin have a maximum weight of less than 900
kg. Hence, our estimates of L^ seemed too large.

ACKNOWLEDGMENTS

The efforts of many people collecting data over a
10-yr period from the fresh fish auctions in Ho-
nolulu have made this paper possible. Thanks are
extended to James B. Reynolds, Luis R. Rivas,
Michael F. Tillman, and the anonymous reviewers
for their helpful comments on an earlier version of
this manuscript.

LITERATURE CITED

Abramson, N. J. (compiler).

1971. Computer programs for fish stock assessment. FAO
Fish. Tech. Pap. 101, [149 p.]
Cassie, R. M.

1954. Some uses of probability paper in the analysis of size
frequency distributions. Aust. J. Mar. Freshwater Res.
5:513-522.
DE Sylva, D. p.

1957. Studies on the age and growth of the Atlantic sailfish,
iMiophorus americanus (Cuvier), using length-frequency
curves. Bull. Mar. Sci. Gulf Caribb. 7:1-20.
DE Sylva, D. P., and W. P. Davis.

1963. White marlin, Tetrapturus alhidiis, in the Middle
Atlantic Bight, with observations on the hydrography of
the fishing grounds. Copeia 1963:81-99.
Eldridge, M. B., and p. G. Wares.

1974. Some biological observations of billfishes taken in the
eastern Pacific Ocean, 1967-1970. In R. S. Shomura and F.
Williams (editors). Proceedings of the International
Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August
1972. Part 2. Review and contributed papers, p.
89-101. U.S. Dep. Commer., NOAA Tech. Rep. NMFS
SSRF-675.
Erdman, D. S.

1968. Spawning cycle, sex ratio, and weights of blue marlin
off Puerto Rico and the Virgin Islands. Trans. Am. Fish.
Soc. 97:131-137.
Fabens, A. J.

1965. Properties and fitting of the von Bertalanffy growth
cur\-e. Growth 29:265-289.
Harding, J. P.

1949. The use of probability paper for the graphical analysis
of polymodal frequency distributions. J. Mar. Biol. Assoc.
U.K. 28:141-153.

Hollander, M., and D. A. Wolfe.

1973. Nonparametric statistical methods. John Wiley and
Sons, Inc., N.Y., 503p.
June, F. C.

1950. Preliminarv fisheries survey of the Hawaiian-Line

Islands area. Part I - The Hawaiian long-line fishery.
Commer. Fish. Rev. 12(l):l-23.
Koto, T., and K. Kodama.

1962. Some considerations on the growth of Marlins, using
size-frequencies in commercial catches. I. Attempts to
estimate the growth of Sailfish. [In Jap., Engl, abstr.] Rep.
Nankai Reg. Fish. Res. Lab. 15:97-108.
Kume, S., and J. Joseph.

1969. Size composition and sexual maturity of billfish caught
by the Japanese longline fishery in the Pacific Ocean east
of 130°W. Bull. Far Seas Fish. Res. Lab. (Shimizu)
2:11.5-162.
Matsumoto, W. M., and T. K. Kazama.

1974. Occurrence of young billfishes in the central Pacific
Ocean. In R. S. Shomura and F. Williams (editors).
Proceedings of the International Billfish Symposium,
Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review
and contributed papers, p. 238-251. U.S. Dep. Commer.,
NOAA Tech. Rep. NMFS SSRF-675.
Nakamura, H.

1942. On the ecology of the istiophorid fishes of Taiwan

waters. [In Jap.] Suisan Gakkwai Ho 9:45-51.
1949. The tunas and their fisheries. Takeuchi Shobo, Tokyo,
118 p. (Translated from Jap. by W. G. Van Campen. 1952.
U. S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 82, 115 p.)
Otsu, T.

1954. Analysis of the Hawaiian long-line fishery,
1948-52. Commer. Fish. Rev. 16(9):1-17.
Ricker.W. E.

1973. Linear regressions in fishery research. J. Fish. Res.
Board Can. 30:409-434.

Royce, W. F.

1957. Observations on the spearfishes of the central
Pacific. U.S. Fish Wildl. Serv., Fish. Bull. .57:497-.554.
Siegel, S.

1956. Nonparametric statistics for the behavorial sci-
ences. McGraw-Hill, N.Y., 312 p.
Simpson, G. G., A. Roe, and R. C. Lewontin.

1960. Quantitative zoology. Revised ed. Harcourt, Brace
and Co., N.Y., 440 p.
Skillman, R. a., and M. Y. Y. Yong.

1974. Length-weight relationships for six species of
billfishes in the central Pacific Ocean. In R. S. Shomura
and F. Williams (editors). Proceedings of the Internation-
al Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August
1972. Part 2. Review and contributed papers, p. 126-137.
U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-675.

Squire, J. L., Jr.

1974. Migration patterns of Istiophoridae in the Pacific
Ocean as determined by cooperative tagging pro-
grams, hi R. S. Shomura and F. Williams (editors).
Proceedings of the International Billfish Symposium,
Kailua-Kona, Hawaii, 9-12 August 1972. Part 2. Review
and contributed papers, p. 226-237. U.S. Dep. Commer.,
NOAA Tech. Rep. NMFS SSRF-67.5.

Strasburg, D. W.

1970. A report on the billfishes of the central Pacific Ocean.
Bull. Mar. Sci. 20:575-604.
Yong, M. Y. Y., and R, A. Skillman.

1975. A computer program for analysis of polymodal
frequency distributions (ENORMSEP), FORTRAN IV.
Fish. Bull., U.S. 73:681.

566

TROPHIC INTERACTIONS AMONG FISHES AND ZOOPLANKTERS
NEAR SHORE AT SANTA CATALINA ISLAND, CALIFORNIA^

Edmund S. Hobson and James R. Chess^

ABSTRACT

Predation pressures from fishes have influenced major evolutionary trends among shallow-water
zooplankters, as concluded from study at Santa Catalina Island, Calif. The predominant zooplank-
tivorous fishes near shore are actinopterygians, an evolutionary line that has centered around
generalized visually feeding, large-mouthed predators. Historically, zooplankters threatened by these
fishes have faced selective pressures favoring reduced size, transparency, and/or nocturnal planktonic
habits. At present, most zooplankters in the nearshore water column by day are very small (<2 mm,
approximately); included are cladocerans, copepods, and various larval forms. Their small size precludes
capture by most large-mouthed fishes, thus providing protection in daylight, when the visual sense of
generalized predatory fishes is most effective. Larger zooplankters in the water column by day, for
example chaetognaths, tend to be transparent. The advantage of transparency to organisms
threatened by visually feeding predators is obvious, and is only briefly mentioned here. Zooplankters
having sizes (most >2 mm) and other features making them vulnerable to large-mouthed fishes tend to
enter the water column only at night, when darkness off'ers some security from visually feeding
predators. Included are polychaetes, mysids, cumaceans, gammaridean and caprellid amphipods,
tanaids, isopods, and carideans.

Because successful defensive features of prey create pressures that modify the offensive features of
predators, the tendencies toward reduced size and nocturnal habits among zooplankters have generated
appropriate adaptations among planktivorous fishes. Fishes that prey as adults on zooplankters during
the day (e.g., blacksmith, Chromis pxinctipinnis) have specialized features, including a small highly
modified mouth, that permit even relatively large individuals to take the tiny organisms which
constitute the daytime zooplankton. Some other fishes are diurnal planktivores only as small juveniles
and assume different feeding habits as they grow larger (e.g., kelp perch, Brachyistius frenatus;
senorita, Oxyjulis californica; smaller juvenile olive rockfish, Sebastes serranoides). Fishes that prey on
zooplankters at night (e.g., larger juvenile olive rockfish; kelp rockfish, Sebastes atrovirens; queenfish,
Seriphus poHtus; walleye surfperch, Hyperprosopon argeyiteum; and salema, Xenistius californiensis)
take the larger organisms that join the zooplankton after dark. In their feeding morphologies and body
form, these large-mouthed fishes have diverged less than their diurnal counterparts from the
generalized predators that give rise to them all. They have, however, acquired specialized features,
including large eyes, suited to detect and capture prey in the dark.

Interactions among predators and their prey
are best recognized by viewing assemblages of
animals that occur together in nature. Further-
more, many trophic interactions become apparent
only upon considering the changes that occur from
day to night, and from one season to another.
These convictions shaped studies of feeding rela-
tions among tropical reef fishes undertaken
between 1962 and 1970 (Hobson 1965, 1968, 1972,
1974), and similarly influenced work done in warm
temperate waters from 1972 to 1975. This more
recent work centered on the inshore habitats at
Santa Catalina Island, Calif, (lat. 33°28'N, long.

'Contribution no. 21 from the Catalina Marine Science Center,
University of Southern California.

^Southwest Fisheries Center Tiburon Laboratory, National
Marine Fisheries Service, NOAA, Tiburon, CA 94920.

118°29'W), where most of the attention was di-
rected at fishes that forage on the benthos (Hobson
and Chess in prep.). The present report, however,
deals with that segment of the work involving
certain fishes and trophically related zooplankters
that interact in the water column near shore.

Only a few studies have considered feeding
habits in natural assemblages of marine fishes.
Limbaugh (1955) and Quast (1968) made the major
contributions in southern California, but these
important studies represent only a beginning.

The present study goes beyond earlier inves-
tigations by considering the organisms taken by
the fishes as prey against a broader consideration
of the array of similar forms present that would
seem to have been equally accessible. The selection
of specific prey, however, is only partially
developed in discussing these data. Selectivity will

Manuscript accepted March 1976.
FISHERY BULLETIN; VOL. 74, No. 3, 1976.

567

FISHERY BULLETIN: VOL. 74, NO. 3

be treated in depth later (Hobson and Chess in
prep.), when these data can be reconsidered along
with the data on organisms simultaneously acces-
sible on the sea floor and other substrata, as well as
in the water column above habitats different from
those described here.

TERMINOLOGY

In this report, the term zooplankton encom-
passes all the varied small organisms we collected
with a plankton net during day and night, and
(most important to this study) which proved to
include the major foods of a well-defined assem-
blage of fishes. All the organisms that we consider
within this definition belong to groups included in
most general accounts of the zooplankton (e.g.,
Newell and Newell 1963; Wickstead 1965).

Nevertheless, some planktologists would ex-
clude from zooplankton forms like large caridean
shrimps that irregularly enter the water column at
night. But among crustaceans such distinctions
fail to establish where, along the continua of size,
mobility, and time spent in the water column,
forms like large carideans are apart from those
minute calanoid copepods that are zooplankters by
any definition. A number of terms defining certain
ecological categories among zooplankton have
been proposed (e.g., holoplankton, meroplankton,
tychoplankton, etc.), but while such terms are
useful in certain contexts, we have seen none that
define categories meaningful to the concepts
developed in this paper (see Discussion).

STUDY AREA

The study area is 25 to 75 m off the western shore

of Big Fisherman's Cove (Figure 1). Most of the
area is open water about 5 to 15 m deep over a
sandy sea floor largely overgrown by the brown
alga Dicfyopteris zonariodes (most of which is
anchored to tubes of the polychaete Chaetopterus
variopedatus (Figure 2)). From the seaward edge
of the study area, the bottom falls sharply to the
greater depths (more than 30 m) that lie at the
center of the cove. Shoreward, and at the mouth of
the cove, lies a forest of giant kelp, Macrocystis
pyrifera. This large brown alga grows to the
water's surface from a rocky bottom that slopes up
to the shoreline from depths of about 8 m (Figure
3). Water temperatures during the study ranged
from lows around 12°C in spring, to highs around
20°C in late summer.

FISHES STUDIED

The fishes studied are those that, during either
day or night, swim in the water column and feed
principally on zooplankters. They are:

Family Scorpaenidae: scorpionfishes
Olive rockfish, Sebastes serranoides (Eigen-

mann and Eigenmann)
Kelp rockfish, S. atrovirens (Jordan and
Gilbert)
Family Pomadasyidae: grunts
Salema, Xenistius californiensis (Stein-
dachner)
Family Sciaenidae: drums

Queenfish, Seriphus politus Ayres
Family Embiotocidae: surfperches
Walleye surfperch, Hyperprosopon argen-

teum Gibbons
Kelp perch, Brachyistius frenatus Gill

.^-T.

Figure l.-Big Fisherman's Cove,
Santa Catalina Island. The study site
lies near the opposite shore, between
the buoy and the headland.

568

HOBSON and CHESS: TROPHIC INTERACTIONS

Figure 2.-The Dictyopteris field,
bordered by the Macrocystis forest.

^^^-

•^'^■?5^:

Figure 3.-The study area at Santa Catalina Island.

Family Labridae: wrasses

Senorita, Oxyjulius californica Giinther)
Family Pomacentridae: damselfishes

Blacksmith, Chromis punctipinnis (Cooper)

Only two other species in the study area have
similar zooplanktivorous habits: the topsmelt,
Antherinops affinis (Ayres), and the shiner perch,
Cymatogaster aggregata Gibbons. These two,
however, are more characteristic of other habitats,
where the species composition of available prey is
different. Although for this reason they will be
described in separate reports later, their activities
are entirely consistent with what is reported and
discussed below.

METHODS

Direct Observations

We used scuba and snorkeling (167 h under-
water) to observe activity of the fishes and tro-
phically related organisms during all periods of
day and night. Except when collecting specimens,
we tried to avoid influencing the organisms or
their environment.

Collecting Zooplankters

During the same period that we collected fishes
for the food-habit study, organisms in the water
column that might be their prey were sampled
with a 1-m plankton net (0.333-mm mesh) that we
pushed through the water for 5-min periods
(Figure 4). In this way, a series of paired collec-
tions sampled the waters above the Dictyopteris
field during September 1973, February 1974, and
May 1974. Of each pair, the first sampled the water
column midway between the water's surface and
the sea floor (in 10 to 15 m of water); and the
second, which followed immediately, sampled the
base of the water column to within about 10 cm
above the bottom. During each sampling month,
we made a set of eight collections-four at full
moon, and four at new moon. Each set included a
pair at night (between 2 and 4 h after last evening
light), and a pair the following day (between 1200
and 1400 h). In addition, we made one set of
collections in the kelp forest bordering the study

569

FISHERY BULLETIN: VOL. 74, NO. 3

Figure 4.-Collecting plankton at
middepths.

area: a pair at night, under a full moon (when there
was enough light to maneuver among the kelp
columns), and a pair the following midday.

Because diving lights probably influence organ-
isms in the water column, we turned them off when
collecting with the plankton net at night. At these
times the moon provided ample light to navigate
when it was present, but on dark nights we
depended on the luminous dials of our compasses
and depth gauges.

Collecting Fishes

To determine the food habits of the fishes, we
speared 521 specimens of the eight species and
then examined their gut contents. All specimens
were taken in the study area between September
1973 and May 1974— the same period over which we
sampled the zooplankton. Most of these specimens
were collected either at night, within the 2 h
before sunrise, or during the afternoon— times
that best show differential day or night feeding.
All measurements of fish size noted in this report
are of standard length.

Sample Analysis

Zooplankton Samples

Generally the samples were analyzed within 2
wk after collection. Sample volumes, which ranged
from 0.2 to 36.0 ml (x = 8.3), were determined

after they had settled for 5 min in a graduated
cylinder. The entire sample was analyzed when its
volume was less than 5 ml. When the sample
was larger, 5-ml aliquots were analyzed, and
numbers for the entire sample then extrapolated.
Whenever less than the entire sample was an-
alyzed, the balance was searched for forms miss-
ing from the aliquot; when found-always in small
numbers— these were counted and added to the
list.

Fish Gut-Content Samples

The digestive tract of each fish specimen was
removed immediately after collection, and pre-
served in a 10% Formalin-^ solution. For analysis,
the contents were examined under a binocular
dissecting microscope, and, when necessary, a
binocular compound microscope. A note was made
of the position in the digestive tract of the various
items. A list was then composed of the items in the
gut, with the species identified when feasible. The
following data were then noted for the items in
each listed category: 1) their number; 2) their size
range; 3) the extent they had been digested (sub-
jectively assessed on a scale of five, from fresh to
well-digested); and 4) an estimate of their repre-
sentation in the gut as percent by volume of the
contents.

^Reference to trade names does not imply endorsement by
the National Marine Fisheries Service, NOAA.

570

HOBSON and CHESS: TROPHIC INTERACTIONS

RESULTS

Volumes of Zooplankters, Day and Night

Our collections with the plankton net were too
few, and spaced over too much time, to attach
much significance to the differences in volume
between various samples. Nevertheless, certain
characteristics probably are meaningful.

The volumes of samples taken day and night at
full moon compared with day and night at new
moon are as follows: FULL MOON-daytime,
middepths {n = 3): 1 to 18 ml, x = 8.3; daytime,
near bottom {n = 3): 0.2 to 5 ml, x = 4; nighttime,
middepths (n = 3): 2 to 13 ml, x = 4; nighttime,
near bottom {n = 3): 5 to 15 ml, x = 10.3. NEW
MOON-daytime, middepths (n = 3): 2 to 5 ml, x =
4; daytime near bottom (n = 3): 1.2 to 5 ml, x = 3.7;
nighttime, middepths {n = 3): 3 to 13 ml, x = 8;
nighttime, near bottom (n = 3): 9 to 36 ml, x =
19.6.

Thus, during the day the volumes of collections
made at the middepths generally were greater
than those made near the bottom, whereas the
situation was reversed at night. Furthermore,
volumes tended to be greater at night than during
the day, with the greatest volumes of all taken
near the bottom on dark nights.

Activity Patterns of
the Zooplankters, Day and Night

The zooplankters are here grouped into a series
of categories (Tables 1, 2),^ most of which represent
phylogenetic classes or subclasses.

Radiolahans

Based on the collections made with the plankton
net (Tables 1, 2), radiolarians are consistently
present in the water column during both day and
night, sometimes in large numbers.

Polychaetes, Swimming

We saw polychaetes in the water column only at
night. Highly motile epitokous nereids were

*The data in Tables 1 and 2 are from collections with the
plankton net made above the open field of low benthic algae
adjacent to the kelp forest. A set of day-night collections was also
made within the forest (see Methods), where the fishes discussed
below spend part (in some cases most) of their time. Because the
data from these collections are essentially like those shown in the
tables, they are not presented.

especially prominent when they swam at mid-
depths during reproductive periods. Polychaetes
are underrepresented in the plankton collections
(Tables 1, 2), however, because their mobility
permitted many to evade our net.

Mollusk Larvae

Based on specimens taken in the plankton net
(Tables 1, 2), mollusk veligers occur in the water
column in similar numbers during both day and
night.

Cladocerans

Cladocerans (Figure 5C) were consistently
present in the collections during both day and
night (Tables 1, 2), although they were more
numerous in the daytime collections.

Ostracods

We saw ostracods in the water column at night,
but never during the day. Our daytime plankton
collections took only a few individuals, these close
to the bottom (Table 3). At night, however, several
species were consistently numerous in both mid-
depth and near-bottom collections (Table 3). The
most numerous ostracod, Parasterope sp. A
(Figure 5H), was numerous in the surface layers of
the sand during the day (Hobson and Chess in
prep.), and during the middle of the night we
observed and collected it concentrated at the
water's surface.

Calanoid and Cyclopoid Copepods

Calanoid and cyclopoid copepods were numerous
in the water column during both day and night,
based on our observations in the water as well as
on our collections (Tables 1, 2). Indeed, calanoids
were the most numerous of all organisms larger
than about 1 mm taken in the net. Calanoids and
cyclopoids were collected in greater numbers at
night (Table 2), but because the plankton is gen-
erally richer after dark, they represented a smaller
percentage of the sample volumes at night than
during the day (Table 1).

The vast majority of calanoids and cyclopoids in
the collections were subadults, and some species
could be recognized only as adults. Of those iden-
tified, the major calanoids were Acartia tonsa
(Figure 5F) and Calanus pacificus, with others

571

FISHERY BULLETIN: VOL. 74, NO. 3

Table 1. -Organisms collected in the plankton net, day and night, showing size and mean percent of total
volume represented by organisms in major taxonomic categories.

Day

Night

Middepth

Near-bottom

Middepth

Near-bottom

Size

collections

Organism category

(mm)

(" = 6)

{n = 6)

(n = 6)

Radiolarians

0.1- 1.0

5.9

7.2

2.1

0.6

Polychaetes

1.0-55.0

0.3

2.1

Mollusk larvae

0.4- 0.8

0.2

0.1

Cladocerans

0.3- 1.0

11.2

1.0

0.9

0.5

Ostracods

0.5- 2.0

0.1

<0.1

0.8

1.9

Calanoids and cycloipods

0.6- 4.0

62.5

66.0

29.1

15.3

Harpacticoids

0.6- 1.0

0.2

1.3

0.6

0.8

Other copepods

1.0- 3.0

0.2

<0.1

0.1

Cirripedian larvae

0.6- 1.0

0.3

1.6

0.2

0.1

Nebaliaceans

3.0- 8.0

0.0

0.3

0.2

Mysids

1.0-12.0

0.0

0.7

39.1

47.3

Cumaceans

2.0- 5.0

0.0

2.0

3.7

Tanalds

1.0- 3.0

0.0

0.3

Isopods

1.0-10.0

0.0

0.3

1.7

2.4

Gammarideans

1.0- 5.0

0.1

0.2

10.0

14.2

Caprellids

3.0-18.0

0.0

0.4

0.8

Euphausid larvae

1.0- 3.0

0.6

1.0

0.5

0.3

Euphausid adults and juveniles

12.0-14.0

0.0

0.2

0.1

Caridean larvae

1.0- 5.0

2.0

2.3

4.8

3.9

Carldean adults and juveniles

4.0-10.0

0.0

1.0

0.7

Reptantian zoea

0.5- 4.0

0.4

0.6

0.9

0.7

Brachyuran megalops

2.0- 3.0

0.0

0.1

0.4

0.5

Bryozoan larvae

0.5- 1.0

0.7

0.6

0.7

0.5

Chaetognaths

4.0-10.0

1.3

3.4

0.6

0.3

Larvaceans'

1.0- 4.0

0.5

0.4

0.1

Fish eggs

0.6- 2.0

12.1

12.2

1.1

1.3

Fishes

3.0-11.0

1.6

1.9

1.5

1.2

'Underrepresented in collections, see text.

Table 2.-0rganisms collected in the plankton net day and night, showing occurrence and mean number of
individuals of organisms in major taxonomic categories.

Day

Night

Middepth

Near-

■bottom

Middepth

Near-

bottom

collections

(n '

= 6)

= 6)
Mean

= 6)

Mean

no.

Organism category

freq.

indlv.

freq.

indlv.

freq.

Indlv.

freq.

indiv.

Radiolarians

100

-500

100

-550

-3,500

-1 ,800

Polychaetes

Mollusk larvae

Cladocerans

100

1,227

100

146

Ostracods

100

Calanoids and cyclopoids

100

2,414

100

1,978

100

3,730

100

2,203

Harpacticoids

100

Other copepods

Cirripedian larvae

Nebaliaceans

Mysids

100

1,100

100

1,721

Cumaceans

100

105

Tanaids

Isopods

100

Gammarideans

100

436

100

2,121

Caprellids

Euphausid larvae

100

Euphausid adults and juveniles

Caridean larvae

100

200

100

220

Caridean adults and Juveniles

Reptantian zoea

100

Brachyuran megalops

•=1

Bryozoan larvae

100

117

Chaetognaths

100

Larvaceans'

Fish eggs

100

137

100

Fishes

'Underrepresented In collections, see text.

572

HOBSON and CHESS: TROPHIC INTERACTIONS

Table 3.-Ostracods collected in the water column, day and night.

Day

Night

Midd

epth

Near-

•bottom

ddepth

Near-

-bottom

collections

col

lections

collections

{n =

= 6)

(

n = 6)

= 6)

Mean

Size

no.

Species

(mm)

freq.

indiv.

freq.

indiv.

freq

indiv.

freq.

indiv.

Parasterope sp. A

1-2

100

10.2

33.8

Cycloleberis lobiancoi

1-2

3.5

18.2

Vargula americana

1-4

4.8

100

12.2

Philomedes sp. A

0.5

3.5

Unidentified

species A

0.8

0.2

Euphilomedes carcharodonta

1-2

0.5

Cythereis sp.

0.2

0.5

Philomedes sp. B

0.5

Conchoecia sp.

0.3

Unidentified species B

0.2

present including Candacia spp., Clausocalanus
sp., Ctenocalanus sp., Euchaeta sp., Labidocera
spp., Lucicutia sp., Metridia pacificus, Pa-
racalaniis sp., and Rhincalanus na^utu^. The
major cyclopoid was Coryceus sp. (Figure 5E), but
others, including Oithona sp., were present.

Harpacticoid Copepods

Our daytime collections took relatively few
harpacticoids, all near the bottom. They were more
numerous in the night collections, however, when
they appeared in both middepth and near-bottom
samples. One form predominated, a species of
Porcellidium, probably undescribed, designated
Porcellidium species A (Figure 5G). Our night
middepth collections {n = 6) took x = 21.6
specimens of this species, whereas the near-bot-
tom collections {n = 6) took x = 37.3. During the
day Porcellidium species A was absent in all
middepth collections {n = 6), but the near-bottom
collections {n = 6) took x = 16. Only one other
harpacticoid was collected in daylight, a form here
designated as harpacticoid species A. Our daytime
middepth collections (n = 6) took x = 1.8
specimens of this species, but it was absent in all
daytime near-bottom collections, and all collec-
tions made at night. Three other forms— a second
species of Porcellidium, and two species of Eupel-
ta (all probably undescribed)-were taken only at
night: a combined mean of 0.7 in the middepth
collections, and a combined mean of 3.2 in near-
bottom collections.

Other Copepods
No other copepods were seen in the water

column, and very few were taken in the plankton
net. An occasional caligoid or monstrilloid ap-
peared in the collections, but were too few to
suggest a pattern.

Cirripedian Larvae

Most of the tiny cypris larvae of the barnacles
(Figure 5D) are smaller than 1 mm. Their occur-
rence in the collections (Tables 1, 2) was irregular,
and without consistent differences between day
and night, or between middepth and near-bottom
samples.

Nebaliaceans

At night we occasionally observed and collected
one species, probably Nebalia pugettensis (Figure
5J; see Smith and Carlton 1975). However, they
were neither seen nor taken during the day.

Mysids

Siriella pacifica (Figure 5M) was the most
widespread mysid over the study area. It remained
sheltered on the sea floor and close to kelp during
the day, but during late twilight moved into open
water, where it spent the night (Table 4). On five
evenings we noted when S. pacifica had first risen
as much as 1 m above the bottom, and found this
level attained 29 to 42 (x = 37.6) min after sunset.
On six mornings, the last individual 1 m above the
bottom was seen 32 to 50 {x = 38.7) min before
sunrise. The stomach contents of 30 S. pacifica
collected during day and night were examined:
DAYTIME-of 10 (8.5-10.5 mm, x = 9.6) collected
amid giant kelp during midafternoon (5 from the

573

FISHERY BULLETIN: VOL. 74, NO. 3

Figure 5.-Relative sizes of some of the zooplankters involved in this study. A to F are full-time inhabitants of the water column; G to W
are species that rise into the water column after dark (R is occasionally there during daylight). A. larvacean Oikopleura sp.; B. bryozoan
larva, cyphonautes; C. cladoceran Evadne sp.; D. barnacle larva, cypris; E. cyclopoid copepod Coryceus sp; F. calanoid copepod Acartia
tonsa; G. harpacticoid copepod Porcellidium sp.; H. ostracod Parsterope sp. A; I. ostracod Cycloleberis lobiancoi; J. nebaliacean Nebalia
pugettensis; K. mysid Acanthomysis sculpta; L. mysid erythropinid sp.; M. mysid Siriella pacifica; N. cumacean Cumella sp. A; 0.
cumacean Cyclaspis nubila; P. isopod Paracercies sp. ($); Q. isopod Cirolana diminuta; R. isopod gnathiid ( $ ); S. gammaridean
amphipod Batea transversa; T. gammaridean amphipod Gitaiiopsis vilordes; U. gammaridean amphipod Erictkonias braziliensis; V.
caprellid amphipod Caprella pilidigita; W. caridean decapod Hippolyte clarki.

574

HOBSON and CHESS: TROPHIC INTERACTIONS

canopy, 5 from the lower portions of the plants), 7
were empty and 3, whose stomachs averaged 13%
full, contained crustacean fragments (55% of total
volume) and unidentified material. NIGHT-
TIME-20 individuals (7-12 mm, x = 10.5) taken at

middepth 2 h before first morning light had
stomachs averaging 82% full, and containing
crustacean fragments (100% of total volume),
including copepods and cladocerans. Clearly, S.
pacifica is a nocturnal predator.

575

Species

FISHERY BULLETIN: VOL. 74, NO. 3

Table 4.-Mysids collected in the water column, day and night.

Day

Night

Middepth

collections

{n = 6)

Near-bottom

collections

{n = 6)

Middepth

collections

{n = 6)

Size
(mm)

%
freq.

Mean

no.

indiv.

%
freq.

Mean

no.

indiv.

%
freq.

Mean

no.

indiv.

Near-bottom

collections

{n = 6)

%
freq.

Mean

no.
indiv.

Siriella pacilica

2-12

100

693.0

100

1,242.2

Erythropinid sp.

1- 6

5.3

100

400.3

100

468.3

Acanthomysis sculpta

4- 8

6.2

2.0

Unidentified sp.

3- 8

1.0

An unidentified erythropinid species (Figure
5L) behaved much like iS. pacifica, but was seen
less often. Although one daytime near-bottom
collection took 32 individuals (probably the net
sampled a diurnal aggregation close to the sea
floor), generally the species was taken in the
plankton net only at night. During the day we
found it numerous amid the flocculent material
that often accumulates in shallow depressions on
sandy bottom (Hobson and Chess in prep.)

The predominant mysid observed and collected
in the canopy of the kelp forest was Acanthomysis
sculpta (Figure 5K), which aggregated in small
openings among the kelp fronds during the day.
{Siriella pacifica also was numerous in the kelp
canopy, but not the erythropinid.) At night some
A. sculpta moved out over the adjacent open
regions sampled by our net (Table 4), but most
stayed close to the kelp. The stomach contents of
20 A. sculpta collected during day and night were
examined: DAYTIME-All 10 (8.5-11 mm, x =
10.0) collected amid the canopy of giant kelp
during midafternoon contained food, with their
stomachs averaging 85% full. All 10 contained
plant material, apparently Mocrocystis (69% of
diet volume), while 7 contained crustacean frag-
ments, mostly copepods, (30% of diet volume).
NIGHTTIME-All 10 (8-11 mm, .f = 9.6) collected
in the kelp canopy 30 min before first morning
light contained food, with their stomachs averag-
ing 82% full. All 10 contained plant material.

apparently Mocrocystis (56% of diet volume), and 9
contained crustacean fragments, mostly copepods
(44% of diet volume). Thus, A. sculpta, which does
not join the other two mysid species in their mass
movement into open water after dark, seems to
feed on plants and animals during both day and
night.

Cumaceans

Cumaceans were numerous in the water column
at night, but absent there during the day. On four
evenings we noted the first one to rise as much as 1
m above the bottom, and found this level attained
26 to 41 (x = 32.3) min after sunset. On four
mornings we noted the last individual 1 m above
the bottom, and recorded this event 37 to 50 {x =
41.3) min before sunrise. Usually we were unable
to determine the species of cumaceans seen swim-
ming in the water, but our plankton collections
(Table 5) took only two species in substantial
numbers: Cumella sp. A (Figure 5N) and Cyclaspis
nubila (Figure 50). Both of these species were
numerous in samples of sand taken from the
surface of the sea floor during the day (Hobson and
Chess in prep.).

Tanaids

Tanaids were absent in the daytime collections,
but one species, Leptochelia dubia, was collected at

Table 5.-Cumaceans collected in the water column, day and night.

Size
(mm)

Day

Night

Middepth

collections

(n = 6)

Near-

C0ll€

■bottom
!ctions
= 6)

Midde

collect

{n =

■pth

ions

Near-bottom

collections

{n = 6)

Species

%
freq.

Mean

no.

indiv.

%
freq.

Mean

no.

indiv.

%
freq.

Mean

no.

indiv.

%
freq.

Mean

no.

indiv.

Cumella sp.A
Cyclaspis nubila
Unidentified sp.

1-2
2-5
2

0
0
0

100
67
50

11.7

17.8

0.7

100

73.5
29.2

1.7

576

HOBSON and CHESS: TROPHIC INTERACTIONS

night. The nighttime middepth collections (n = 6)
took X = 5.8 L. dubia, and the nighttime near-
bottom collections (w = 6) took x = 9.5. We found
this species in tubes of cemented sand grains in
daytime dredge samples from sandy bottom
(Hobson and Chess in prep.).

Isopods

Isopods generally were absent from the water
column during the day, although the plankton
collections show that at least some juvenile and
female gnathiids (Figure 5R) are present. After
dark, however, a number of isopods occurred in the
mid-waters (Table 6). Paracercies spp., in par-
ticular, were numerous. Most of the specimens of
Paracercies were juveniles or females (Figure 5P),
and their identity remains uncertain. Based on the
occurrence of males, P. cordata is by far the most

numerous species of this genus in the study area,
but at least one other is present.

Gammaridean Amphipods

Gammaridean amphipods were generally ab-
sent from the water column during the day,
although Gitanopsis vilordes, which lived prin-
cipally amid the dense surface canopy of the kelp
forest bordering the study area (Hobson and Chess
in prep.), was collected in small numbers (Table 7).
At night, however, we saw gammarideans
throughout the water column, and, with Batea
transversa (Figure 5S) predominating, they were
a major component of our catch in the nighttime
collections (Table 7). Batea transversa was
numerous during the day amid the low benthic
algae that floors most of the study area (Hobson
and Chess in prep.).

Table 6. -Isopods collected in the water column, day and night.

Day

Night

Middepth

Near-

■bottom

Middepth

Near-

•bottom

collections

collecfions

collections

(n = e)

= 6)

(n =

= 6)

Mean

Size

% no.

no.

Species

(mm)

freq. indiv.

freq.

indiv.

freq.

indiv.

freq.

indiv.

Gnathiid juveniles

1-3

0 0

100

9.8

100

32.5

Paracerces sp.

1-3

0 0

11.3

11.2

Cirolana harfordi

0 0

1.5

2.0

Cirolana diminuta

3-5

0 0

0.2

0.8

Eurydice caudate

2-3

0 0

0.8

Excorailana kathae

0 0

0.3

0.2

Cirolanid sp.

0 0

0.2

Exospheroma rhomburum

0 0

0.2

Limnoria sp.

0 0

0.2

Table 7.-

-Gammaridean amphipods

. collected in the water column, day and night.

Day

Night

Middepth

Near

-bottom

Middi

2pth

Near-bottom

collections

llections

= 6)

(n =

(

n = 6)

Mean

Size

no.

Species
Batea transversa

(mm)
1-4

freq.

indiv.

freq.

indi

freq.

indiv.

freq.

indiv.

100

400.7

100

1,978.2

Gitanopsis

vilordes

1-3

0.8

2.3

11.3

100

41.7

Ericthonias braziliensis

2-4

0.7

2.0

Synchelidit

jm sp.

0.3

1.3

Orchomene sp.

2-4

0.2

1.3

Aoroides columbaie

3-4

1.3

0.5

0.3

Pleustes pi

atypa

0.7

Ampithoe spp.
Podocerus cristatus

8
2

0
0

17
0

0.2
0

17
33
17

Oedocerotid so.

Phoxocepli

alid sD.

0.2

Unidentified'

1-4

79.7

100

95.5

'Many of the unidentified specimens are juveniles, probably at
above.

east many being of the species listed

577

FISHERY BULLETIN: VOL. 74, NO. 3

On four evenings we noted the first gammar-
idean seen as much as 1 m above the bottom, and
found this level attained 27 to 39 (.r = 34) min after
sunset. On each occasion, individuals had been
visible close among the bottom algae for about 5
min before any of them rose to the 1-m level. The
final return to the sea floor at daybreak was
monitored on four mornings, when the last in-
dividual was seen 1 m above the bottom 26 to 41 (.f
= 35) min before sunrise. Similar to the evening
situation, individuals continued to be visible close
above the bottom algae for an additional 5 min, or
so.

To roughly determine the proportion of gam-
marideans that rise from the sea floor at night, we
compared the amphipods in a sample of benthic
algae at night, with a similar sample taken in the
same place the following day (both samples,
loosely packed in a 2.3-liter plastic bag, were taken
immediately after plankton collections). Both
samples contained 2.5 ml of animals (including
other forms besides amphipods). Nevertheless, the
limited data (Table 8) indicate that the numbers of
some gammarideans on the algae, notably B.
transverm, dropped sharply after dark, those of
others, including Ericthonias braziliensis (Figure
5U), experienced a lesser decline, and those of still
others remained essentially unchanged. Data
from the collections (Table 7) and direct observa-
tions indicate that there were fewer amphipods on
benthic algae after dark because many have risen
into the water column. But the tendency to leave
the sea floor clearly varies between species and in
perhaps no species is it absolute. Probably at least

Table 8.-Gammaridean amphipods collected in samples of
benthic algae, day and night.

No. of

Individuals

Species

Day

Night

Species known from plankton collections

Batea transversa

Ericthonias braziliensis

Ampithoe spp.

Aoroides columbiae

Pleustes platypa

Podocerus cristatus

Total

Species unknown from plankton collections

Hyale nigra

Photis brevipes

Elasmopus antennatus

Heteropf)lias seclusus

Total

Unidentified forms'

'At least some of the unidenitfied forms probably are juveniles
of tfie species listed above.

many individuals make only short excursions into
the water column.

Caprellid Amphipods

We never saw caprellids above the bottom
during the day, but saw them, though infrequent-
ly, in the water column at night. Consistent with
these observations, caprellids were collected in the
plankton net at night, but never during the day.
The nighttime middepth collections (n = 6) took x
= 3.2 Caprella pilidigita and x = 0.2 C. califor-
nica, whereas the near-bottom collections {n = 6)
took X = 9.7 C. pilidigita and x = 5.2 C. califor-
nica. In addition, a single unidentified juvenile
was taken in one nighttime middepth collection.
Size ranges of specimens: C. pilidigita 4 to 18 mm,
C. californica 6 to 10 mm, and the unidentified
juvenile 3 mm. Both C. californica and C. pilidigi-
ta (Figure 5V) were at all times numerous amid
the low benthic algae that floors most of the study
area.

Euphausid Larvae

The calyptopis larvae of euphausids occurred
regularly in both day and night collections.

Euphausid Adults and Juveniles

Euphasid adults and juveniles were neither seen
nor collected in the water column during the day,
but occasionally swarmed around our lights at
night. The few individuals collected in the plank-
ton net (Tables 1, 2) are of one species: Thijsan-
oessa spinifera. The numbers collected, however,
underrepresent the numbers we saw in the water
(all of which appeared to be T. spinifera), probably
because this relatively large, motile animal effec-
tively evaded our net. Rather than rising from the
sea floor at nightfall, as do so many other noctur-
nal components of the plankton discussed above,
this euphausid seems to move in from deeper
water. Unlike the other forms, euphausids were
not taken in our extensive diurnal sampling of the
benthos (Hobson and Chess in prep.).

Caridean Larvae

Based on the collections (Tables 1, 2), caridean
larvae are numerous in the plankton during both
day and night, but more so at night. Furthermore,
there are more larger individuals in the water

578

HOBSON and CHESS: TROPHIC INTERACTIONS

column after dark. We made no attempt to iden-
tify our specimens to species, but probably many
are larvae of the two species discussed as adults
and juveniles, below.

Caridean Adults and Juveniles

Caridean adults and juveniles were observed in
the water column only at night. On the one even-
ing that the event was noted, the first individual
seen rising as much as 1 m above the bottom
attained this level 39 min after sunset. Adult and
juvenile carideans are absent in the daytime
collections, but Hippohjte clarki (Figure 5W), was
sometimes numerous in collections made after
dark. The nighttime middepth collections (n = 6)
took .r = 25.2 H. clarki, and the near-bottom
collections {n = 6) took x = 10. Only one other
adult caridean was collected, this a single Eualus
herdmani in a nighttime near-bottom sample. The
specimens of H. clarki were 4 to 10 mm long, the
single E. herdmani 12 mm. Hippohjte clarki is
numerous during the day in the kelp forest bor-
dering the study area, where it concentrated in the
dense surface canopy and upper regions of these
massive plants. At the same time E. herdmani
predominated in the lower regions of the same
plants (Hobson and Chess in prep.).

The stomach contents of 20 H. clarki collected
during day and night were examined. DAY-
TIME-Of 10 (8-16 mm, .f = 10.8) collected from
giant kelp plants during midafternoon, 4 were
empty, and the other 6, whose stomachs averaged
17% full, contained only extensively macerated
fragments. NIGHTTIME-Of 10 (8-17 mm,. r = 12)
collected close to giant kelp 1 h before first morn-
ing light, 1 was empty, and the other 9, whose
stomachs averaged 34% full, contained a variety of
prey, some of it fresh: mollusk veligers in 4 (28% of
total volume); foraminiferans in 3 (9% of total
volume); shrimp larvae in 1 (11% of total volume);
and extensively macerated material in 7 (52% of
total volume). These limited data indicate this
animal is primarily a nocturnal predator, but only
a relatively few seem to swim far from algal cover.

Reptantian Zoea

Based on the collections (Tables 1, 2), zoea were
consistently present in moderate numbers at all
levels of the water column during both day and
night, but were most numerous there after dark.
Usually we failed to notice zoea in the water, but

one night observed them in dense swarms close to
the bottom.

Brachyuran Megalops

Our plankton collections (Tables 1, 2) indicate
that brachyuran megalops were frequently pres-
ent, if not numerous, in the water column at night,
but only infrequently present during the day.

Bryozoan Larvae

The cyphonautes larvae of bryozoans (Figure
5B) were consistently taken in substantial
numbers by middepth and near-bottom collections
both day and night.

Chaetognaths

Our collections regularly took chaetognaths both
day and night (Tables 1, 2), but even though these
animals are relatively large, we failed to see them
in the water, presumably because they are largely
transparent. Chaetognaths probably were more
numerous in the study area than our collection
data indicate, owing to a mobility that would
permit many to evade our net.

Larvaceans

We collected larvaceans in our plankton net both
day and night, but only in small numbers (Tables 1,
2). It became clear that these numbers far under-
represented the numbers present, however, when
we examined the gut contents of the blacksmith
(recounted below). Most larvaceans in the area are
less than 0.5 mm in diameter, and apparently their
pliable bodies readily squeeze through the
0.333-mm mesh of our net. So we made a midday
tow in the study area using a 0.25-m net with a
0.253-mm mesh. Significantly, larvaceans, most
being of the genus Oikopleura (Figure 5A), made
up 20% of the sample. There was one larvacean to
about every six copepods (calanoids and cy-
clopoids), and they ranged from 1 to 3 mm long,
with a diameter of about 0.2 to 0.5 mm. Sig-
nificantly, active individuals throughout much of
this size range were observed passing through a
piece of 0.333-mm mesh net placed among them in
a petri dish. Because these animals are transpar-
ent, and so small, we failed to see them in the
water.

579

FISHERY BULLETIN: VOL. 74, NO. 3

Fish Eggs

Fish eggs were a regular component of both
middepth and near-bottom plankton collections
during both day and night (Tables 1, 2). Owing to
their small size and transparency, however, they
went unseen by us in the water.

Fish Larvae

Fish larvae were consistently seen and collected
at middepths and near the bottom both day and
night (Tables 1, 2).

Activity Patterns of
Planktivorous Fishes, Day and Night

Having described the zooplankters that occur in
the water column during both day and night, we
now consider the feeding activities of the fishes
that find prey there.

Sebastes serrayioides—oWve rockfish

Small juveniles of this species first appeared
inshore during midsummer when about 30 mm
long. They remained here throughout the ensuing
year, growing to about 100 to 110 mm long. Al-
though their numbers declined sharply during the
following summer, when the next crop of small
juveniles arrived, many remained in the area well
into a second winter, and some stayed longer.
Nevertheless, few olive rockfish exceeding about
120 mm occurred in the study area. Larger in-
dividuals (to well over 200 mm) were numerous in
deeper water, but were not considered in this
study. Limbaugh (1955) noted: "The young appear
in large schools, from May through September.
The schools form behind protective reefs, in bay
entrances, and in the lee of islands." Other data on
this species presented by Limbaugh, and also by
Quast (1968), pertain generally to individuals
larger than those discussed here. The species is
reported to reach 610 mm (Miller and Lea 1972).

The activity pattern of this fish changes mark-
edly during its first year inshore. Most of the
smaller juveniles are active by day and relatively
inactive at night. Beginning among those about 55
mm long, however, there is a general shift toward
feeding after dark. Nocturnal habits are charac-
teristic among individuals larger than about 65
mm (to at least 120 mm— the largest considered

here). This report, therefore, recognizes three size
categories, and treats each separately:

1) small juveniles, which are predominantly diur-
nal, are those shorter than 55 mm; 2) intermediate
juveniles, which represent a transition to the
nocturnal mode, are those between 55 and 64 mm;
and 3) large juveniles, most of which are nocturnal,
are those 65 mm and longer.

SMALL JUVENILES.-During daylight, the
small juveniles generally hovered in small ag-
gregations at middepths in less than 5 m of water.
In the study area they were most numerous along
the shoreward margin of the kelp forest, close to
rising stands of Macrocystis and other large algae.

The small juveniles appeared in the water
column each morning, beginning about 40 min
before sunrise, after a night spent sheltered under
cover of algae or rocks. They occurred first as
solitary individuals, but soon assembled in ag-
gregations that were well-formed by 30 min
before sunrise. Only after sunrise, however, did
they feed appreciably. Then, sporatically at first,
but with steadily increasing frequency, they
began to snap at objects in the water indistin-
guishable to a human observer a few meters away.

The onset of feeding in the morning is illus-
trated by the decreasing incidence of empty guts
in specimens collected during this period from the
mid-water aggregations. Empty guts occurred in
84% of those sampled during the 40 min before
sunrise (52 of 62 specimens; 42-54 mm, x = 49), in
58% of those collected during the 15 min following
sunrise (7 of 12 specimens; 41-53 mm, .f = 48), in
25% of those taken 15 to 30 min after sunrise (2 of 8
specimens; 45-53 mm, .f = 50), and in none of those
collected 30 to 60 min after sunrise (10 specimens;
41-54 mm, .f = 50).

Intermittent observations throughout the day
showed consistent feeding activity. The guts were
full in all 11 specimens (40-51 mm, x = 45) sampled
from aggregations during midafternoon. Items
they had taken, combined with items taken by the
31 specimens containing food that were collected
during early morning (a total sample of 42 fish),
document the food habits of these small juveniles.

Prey of the 44 small juveniles that had iden-
tifiable material in their guts are listed below in
order of their rank as prey. (The same format is
used in presenting the gut contents of the other
fish species, below.) In this list, the major num-

580

HOBSON and CHESS: TROPHIC INTERACTIONS

bered categories are the same as those in which the
zooplankters are organized in Tables 1 and 2 and in
the text above. The few additional major categor-
ies include various nonplanktonic organisms that
some of these fishes had taken in small numbers.
Listed under each major category, according to
rank within that category, are the species and
species groups that are the actual prey of the fish.
Following most entries throughout the listing are
sets of three values in parentheses; these values
relate certain characteristics of the entry to the
food habits of the fish. (The values were derived
from calculations based only on fish that contained
identifiable material. Fish with empty guts or
containing only unidentifiable material were not
considered.) The first value in parentheses is the
percent of fish that contained the item(s); the
second value is the mean number of individuals of
the item(s) that were taken, and the third value is
the mean percent of the diet volume represented
by the item(s). Rank as prey was determined by a
ranking index, which is not shown, but which is the
product of the first and third values in
parentheses.

Following the above format, the prey organisms
are:

1. CALANOID AND CYCLOPOID COPEPODS (83: 44.4: 59.5)

calanoids, including Acartia tonsa and Lahidocera sp. (81:
40.3: 54.9); cyclopoids, including Corycaeus sp. (38: 4.1: 4.6).

2. GAMMARIDEAN AMPHIPODS (29: 0.4: 11.4)

Batea transversa (18: 0.3: 7.6); unidentified fragments (11:
0.1:3.8).

3. CARIDEAN LARVAE (20: 0.4: 3.2)

unidentified species.

4. MYSIDS(11:0.3:4.5)

Acanthomysis sculpta (3: 0.1: 2.0);erythropinid sp. (3: 0.1:
0.5); unidentified fragments (7: 0.1: 2.0).

5. CLADOCERANS (20: 1.1: 2.4)

Evadnc sp.

6. OTHER COPEPODS (18: 0.3: 2.2)

unidentified monstrilloids.

7. EUPHAUSID ADULTS AND JUVENILES (11: 0.1: 1.3)

unidentified fragments.

8. BRACHYURAN MEGALOPS (5: 0.1: 4.8)

unidentified.

9. HARPACTICOID COPEPODS (18: 0.4: 1.2)

harpacticoid sp. A (7: 0.1: 0.7); PorcelUdium sp. B (5: 0.1: 0.2);
Porcellidium sp. A (2; 0.1: 0.1); unidentified fragments (5:
0.1:0.2).

10. TANAIDS(10:0.2:2.4)

Leptochelia dubia (7: 0.1: 1.2); unidentified fragments (3: 0.1:
1.2).

11. REPTANTIAN ZOEA (11: 0.1: 1.3)

unidentified.

12. FISHES (5: 0.2: 2.0)

unidentified larvae.

13. ISOPODS (2: <0.1: 1.1)

Paracercies sp.

14. GASTROPODS (2: 0.4: 0.5)

Tricolia sp.
1.5. EUPHAUSID LARVAE (2: <0.1: 0.1)
calyptopsis.

16. CIRRIPEDIAN LARVAE (5:<0.1: <0.1)

cypris.

17. BRYOZOAN LARVAE (2: <0.1:<0.1)

cyphonautes.

Small juveniles took calanoid copepods as their
major prey from the time they began feeding at
sunrise until they ceased feeding at the end of the
day. In 10 specimens collected during May and
June (the only times for which calanoids in this
material were identified to species), about 22% of
the calanoids were Acartia tonsa, and although
the rest remained unidentified (except for a single
specimen of Lahidocera sp.), many probably were
immature individuals of this same species.

A number of the prey listed above occurred only
in specimens collected during early morning.
These are: the gamaridean amphipods, the tan-
aids, the euphausids, the lone isopod, the megalops,
and all mysids except those in one individual (see
below). Most of these items were extensively
digested, in sharp contrast to the freshness of the
calanoids and other food materials in the early-
morning specimens. Clearly, they had been in the
guts for some time, probably since the previous
night. Nevertheless, judging from the empty guts
in most individuals of this size at daybreak it
would seem that nocturnal feeding is insignificant.

Only later than about 30 min after sunrise did
the olive rockfish begin taking Evadne sp., but this
cladoceran then became a consistant component of
the diet for the rest of the day. Evadne is slightly
smaller and more transparent than the other prey
organisms, and to capture it the rockfish may need
more light. The only mysid taken during the day
was Acanthomysis sculpta, of which two in-
dividuals that appeared recently ingested were
found in one olive rockfish during midafternoon.

INTERMEDIATE JUVENILES. -Individuals
between about 55 and 65 mm long were highly
inconsistent in so far as whether they fed by day or
by night (many did both). The nocturnal situation
among intermediate individuals is represented by
18 specimens (55-63 mm, x = 58) collected before
sunrise from open water during the hour before
first morning light, and also from developing

581

FISHERY BULLETIN: VOL. 74, NO. 3

aggregations of juveniles more than 30 min before
sunrise. Of these, 13 (72%) were full of prey in
varying stages of digestion, demonstrating noc-
turnal feeding, whereas 5 (28%) were empty,
indicating they had been inactive that night. All
the empty fish were from the developing ag-
gregations, but many of those containing food
were also taken from those aggregations. Items
taken at night by the 13 intermediate juveniles
containing food were as follows, with the format
being that used for the small juveniles, above.

1. GAMMARIDEAN AMPHIPODS (69: 2.9: 28.5)

including Batea transversa and Ericfhonias braziliensis.

2. CUMACEANS (54: 2.7: 26.3)

Cycla.'^pis nxhila (46: 2.6: 26.1); Cumella sp. A (8: 0.1: 0.2).

3. MYSIDS(38:?: 16.2)

Siriella pacifica (23: 0.3: 13.1); unidentified fragments (15: ?:
3.1).

4. FISHES (15: 7.1: 5.4)

unidentified larvae.

5. CAPRELLID AMPHIPODS (8: 0.5: 6.2)

Caprella pilidigita.

6. POLYCHAETES, SWIMMING (8: ?: .5.0)

unidentified fragments.

7. OSTRACODS (8: 0.1: 0.8)

Paraaterope sp. A

8. BRACHYURAN MEGALOPS (8: 0.5: 0.4)

unidentified.

The diurnal feeding situation, as well as the
changeover from day to night, is represented by 12
individuals (55-62 mm, x = 58), all with full guts,
collected from among feeding aggregations of
small juveniles within 1 h after sunrise. Almost all
the food items in this sample were either fresh or
well-digested-there was little in between. Pre-
sumably, the fresh items were those that had been
taken after feeding began within the previous
hour, whereas the extensively damaged items had
been taken during the night before. (One would
expect specimens taken as early in the morning as
these to contain evidence of any nocturnal feeding
they might have done, and this proved true here.)
Seven of the 12 individuals sampled contained
both fresh and well-digested material in large
numbers, always with the fresh items forward in
the gut (often in the esophagus), and the well-
digested items well back in the posterior region.
Clearly, these individuals had fed substantially
during both day and night (a conclusion strength-
ened by the kinds of prey among the fresh and
well-digested segments of the diet, see below).
Three of the other five specimens contained only
fresh items, indicating diurnal feeding exclusive-
ly, whereas two contained just well-digested ma-

terial, indicating only nocturnal feeding. Food
items in this material are identified below, but
with fresh items listed separately from well-
digested items.

FRESH ITEMS

1. CALANOID AND CYCLOPOID COPEPODS (83: 65.9: 47.8)

calanoids (83: 65.7: 47.5); cyclopoids (17: 0.2: 0.3).

2. CLADOCERANS (33: 0.8: 0.8)

Evadne sp.

3. OSTRACODS (8: 0.1: 1.7)

Cycloleberis lohiancoi.

4. OTHER COPEPODS (8: 0.1: 0.4)

monstrilloids.

5. ISOPODS(8:0.1:0.4)

gnathiid juveniles.

6. HARPACTICOID COPEPODS (8: 0.2: 0.2)

Porcellidium sp. A.

7. CARIDEAN LARVAE (8: 0.1: 0.3)

unidentified.

WELL-DIGESTED ITEMS

1. GAMMARIDEAN AMPHIPODS (.50: 1.3: 1.1)

including Batea transversa.

2. CARIDEAN LARVAE (33: 2.2: 12.9)

unidentified.

3. EUPHAUSID ADULTS AND JUVENILES (17: 0.7: 10.7)

unidentified fragments.

4. FISHES (17: 1.0: 9.2)

unidentified larvae.

5. REPTANTIANZOEA(17:0.3:2.0)

unidentified.

6. BRACHYURAN MEGALOPS (8: 0.3: 1.3)

unidentified.

7. INSECTS (8: 0.1: 0.8)

unidentified.

8. CAPRELLID AMPHIPODS (8: 0.1: 0.4)

unidentified.

The fresh items apparently represent diurnal
feeding, the well-digested items nocturnal feed-
ing. Thus, among individuals within the inter-
mediate size range there obviously are many that
forage during both day and night.

LARGE JUVENILES.-During the day, olive
rockfish more than about 65 mm long generally
hovered in small aggregations low in the water
column beneath the kelp canopy within the
seaward part of the forest (Figure 6). Aggrega-
tions composed of relatively large individuals
(exceeding a length of about 100 mm) sometimes
hovered above others of the same size seated on
the rocks below.

In contrast to the small individuals described
above, large juveniles generally showed no sign of
feeding during the day, an observation supported

582

HOBSON and CHESS: TROPHIC INTERACTIONS

Figure 6.-A daytime aggregation of large juvenile olive rockfish, Sebastcs serranoides. Many nocturnal fishes spend the day in

quiet schools.

by examination of gut contents. Of 42 specimens
(65-120 mm, x = 91) collected from aggregations
during midafternoon, 28 (67%) had empty guts,
and 8 (19%) contained only well-digested frag-
ments. Six (14%), however, contained relatively
fresh prey probably captured earlier that day: the
mysid Acanthomysis sculpta (50: 15.7: 32.5); the
caridean shrimps Hippolyte clarki (50: 0.5: 14.2)
and Euahis herdmani (17: 0.2: 5.2); the cladoceran
Evadne sp. (17: 4.3: 10.0); calanoid copepods (17: 1.5:
5.2); euphausid larvae, calyptopis stage (17: 0.5:
1.6); and harpacticoid copepod Porcellidium sp. A
(17: 0.2: 0.1). Also present were extensively
digested fragments of cumaceans, tanaids, eu-
phausids, and mysids (33: ?: 20.5) that probably had
been taken the night before (a judgment in-
fluenced by knowledge of nocturnal food habits,
defined below). All of the cladocerans, calanoids,
and euphausid larvae among this material consti-
tuted the entire gut contents of one 82-mm in-
dividual, and the contents suggest a mode of
feeding like that of the small juveniles above.

Beginning about 20 min after sunset, large
juveniles began leaving the sites of their daytime
aggregations. They moved away from the kelp
forest, and dispersed over the adjacent field of
Dictyopteris. Many of them rose into the upper
part of the water column, but most remained
within 5 m of the sea floor. They remained in these
positions throughout the night, often assuming a
tail-down attitude, now and then darting a few
centimeters forward and snapping at objects in
the dark water. The few that remained in the kelp
forest usually hovered high in the water column
beneath sizeable breaks in the kelp canopy. They
began returning to the forest at first morning
light, and by 30 min before sunrise were back in
their daytime aggregations.

Clearly, olive rockfish of this size feed chiefly at
night. This conclusion is supported by study of gut
contents from 72 specimens (65-157 mm, x = 85)
collected in this area at night-later than 4 h after
sunset, and before first morning light. Only two of
these (less than 3%) had an empty gut, a contrast

583

to the high incidence of empty guts (67%) among
specimens collected during the afternoon. More
significant, the gut of all 70 other specimens
contained many fresh items, all organisms present
in the water column after dark.

Major categories of prey with included species
and species groups, are listed below in order of
their rank as prey.

I. GAMMARIDEAN AMPHIPODS (90: 16.9: 43.9)

Batia transversa 76: 8.5: 21.9); Ericthonias braziliensis (19:
1.1: 2.6); Ampithoe spp. (20: 1.3: 2.4); Photis hrevipe^ (14: 1.2:
O.S); Ampelisca sp. (3:<0.1: 1.6); Synchelidium sp. (9: 0.1: 0.6);
Aoroides columbiae (9: 0.1: 0.4); Hi/ale nigra (3: 0.1: 0.3);
Monoculoides sp. (3: 0.3: 0.2); Podocerus cristatus (4: <0.1:
0.1); phoxocephalid sp. (3: <0.1: 0.1); lysianassid spp. (1: <0.1:
0.1);Pa/-n/)/io.r((.ssp. (1:<0.1: 0.1); Pleiistes phti/pa (1: <0.1:
0.1); unidentified gammarideans, including unknown forms
and those unrecognized due to damage (73: 3.9: 12.6).

2. MYSIDS (69: 2.7: 12.5)

Siriella pacijica (47: 1.7: 9.2); erythropinid sp. (40: 0.9: 3.1);
Acanthomysis sculpta (3: <0.1: 0.2).

3. CUMACEANS (57: 4.9: 8.4)

Cyclaspis nubila (37: 4.1: 7.2); Cnmella sp. (40: 0.8: 1.1);
unidentified (3: <0.1: 0.1).

4. POLYCHAETES. SWIMMING (36: 0.5: 8.6)

at least most of them nereids.

5. CAPRELLID AMPHIPODS (36: 1.4: 7.0)

Caprella pilidigita (24: 0.8: 4.2); C. californicn (19: 0.5: 2.6);
C. brerirostis (1: <0.1: 0.1); unidentified species (1: 0.1: 0.1).

6. OSTRACODS (43: 1.6: 3.8)

Parasterope sp. A (37: 1.0: 2.9); Vargida amerkana (9: 0.3:
0.5); Phihnnedes sp. (4: 0.1: 0.2); Cycloleheris lobiancoi (3:
<0.1: 0.1); unidentified (1: <0.1: 0.1).

7. ISOPODS (39: 1.7: 3.3)

Paracercies sp. (27: 0.8: 2.1); gnathiid juveniles and females
(21: 0.8: 0.8); Idotea spp. (4: 0.1: 0.1); Cirolana diminuta (3:
<0.1: <0.1); Limnoria lignorum (1: <0.1: 0.1); Excorallana
kathae ( 1: <:0.1: <0.1).

8. CARIDEAN ADULTS AND JUVENILES (24: 0.4: 4.1)

Hippolyte clarki (20: 0.2: 2.2); Eualus herdmani (6: 0.2: 1.9).

9. TANAIDS(26:0.5:1.2)

Leptochelia diibia (25: 0.4: 1.0); unidentified (4: 0.1: 0.2).
10. EUPHAUSID ADULTS AND JUVENILES (7: 0.1: 2.0)
Thys:anoessa sp. (1: <0.1: 0.7); unidentified (6: 0.1:1.3).

II. FISHES (9: <0.1: 1.6)

unidentified larvae.

12. BRACHYURAN MEGALOPS (10: 0.2: 0.7)

unidentified.

13. CARIDEAN LARVAE (9: 4.0: 0.6)

unidentified.

14. HARPACTICOID COPEPODS (13: 0.2: 0.2)

Porcellidinm sp. A.

15. REPTANTIAN ZOEA (6: 0.6: 0.4)

unidentified.

16. CALANOID AND CYCLOPOID COPEPODS (4: <0.1: 0.2)

unidentified cyclopoids.

17. OTHER COPEPODS (1: <0.1: 0.1)

unidentified caligoids.

18. NEBALIACEANS(1:<0.1:<0.1)

Nebalia pugettensis.

FISHERY BULLETIN: VOL, 74, NO. 3
Sebastes atrovirens— kelp rockfish

The kelp rockfish, which may attain a length of
425 mm (Miller and Lea 1972), was the most
numerous adult scorpaenid in the study area.
During the day, a few individuals hovered above
the sea floor in shadow under the kelp canopy, but
most spent the daytime seated on rocky substrata
within the forest-quiet but alert. At night this
fish generally hovered in mid-water close to the
rising kelp stipes (Figure 7), and often amid the
kelp canopy, near the water's surface. Occasionally
at night it hovered in open water close along the
seaward margin of the forest. Diff"erences in
activities between day and night have gone un-
noted in previous reports of this species. Lim-
baugh (1955) reported that it lives in the lower
levels of the kelp and among the rocks, and feeds
on "crustaceans and small fish." Quast (1968), on
the other hand, reported that it ranges all the way
from the bottom to the kelp canopy and apparently
utilizes "all available foods in these regions."

Of 29 specimens (89-240 mm, x = 175) collected
for study of food habits, all 6 (100%) taken during
midafternoon were empty, whereas only 3 of 23
(13%) taken at night (more than 4 h after sunset)
were empty. Clearly, this fish is predominantly a
nocturnal feeder. Quast (1968) noted that many of
the kelp rockfish he examined had an empty
stomach but did not suggest nocturnal feeding. He
noted that his specimens "appeared quite thin"
and though recognizing this may be a natural
condition, thought perhaps "the high frequencies
of empty stomachs and the broad variety of food
items found may indicate malnutrition." The kelp
rockfish of our study area, we have noted, often
have deeply concaved bellies during the day, which
we assume is due to the emptiness of their guts at
this time.

Almost all food materials taken by this fish were
from the water column. The major food categories,
which included species and species groups, are
listed below in order of their rank as prey.

1. MYSIDS (90: 22.3: 39.5)

Acanthomysis sculpta (60: 18.3: 30.1); Siriella pacijica (65:
3.6: 9.2); erythropinid sp. (15: 0.4: 0.2).

2. CARIDEAN ADULTS AND JUVENILES (85: 7.0: 16.2)

Hippolyte clarki (65: 4.4: 10.0); Eualus herdmani (40: 2.6:
6.2).

3. GAMMARIDEAN AMPHIPODS (95: 13.8: 13.7)

Batea transversa (95: 9.3: 9.5); lysianassid spp. (50: 1.2: 1.1)
Ampelisca sp. (10: 1.4: 1.3); Pleustes platypa (25: 0.3: 0.3)
Podocerus cristatus (5: 0.1: 0.2); Ampithoe tea (5: 0.1: 0.2)

584

HOBSON and CHESS: TROPHIC INTERACTIONS

Figure 7.-A solitary kelp rockfish,
Sebastes atrovirens, close to rising
stipes of a giant kelp plant at night.

Aoroides columbiae (15: 0.2: 0.1); Hyale nigra (5: 0.1: 0.1);
Ericthonias braziliensis (5: 0.2: 0.1); unidentified (40: 0.9:
0.8).

4. ISOPODS (75: 3.7: 14.3)

Paracerciefi sp. (75: 3.3: 11.8); Pentidotea resecata (5: 0.1:
1.8); gnathiid juveniles (10: 0.1: 0.3); Cirolana harfordi (10:
<0.1: 0.2); Idotea rectolinenta (5: <0.1: 0.2).

5. POLYCHAETES, SWIMMING (20: 0.3: 7.1)

unidentified, but only certain epitokous nereids were
significant, these being prominent in the guts on nights
when they swam in mid-water.

6. BRACHYURAN ADULTS (10: 0.1: 4.2)

all Pugettia prodiicta.

7. OSTRACODS (30: 0.5: 0.9)

Cycloleberis lobiancoi (20: 0.2: 0.5); Vargula americana (10:
0.2: 0.3); Parasite rope sp. A (5: 0.1: 0.1).

8. FISHES (15: 0.1: 1.2)

larvae (10: 0.1: 0.8); scales (5: ?: 0.4).

9. NEBALIACEANS(5:0.1:1.1)

Nebalia pugettensis.

10. CUMACEANS (5: 0.3: 0.7)

all Cgclaxpia nubila.

11. GASTROPODS (5: 0.1: 0.1)

Lacuna unifasciafa.

12. EUPHAUSID ADULTS AND JUVENILES (5: 0.1: 0.1)

unidentified.

Xenistius californiensis—siilemz

We never saw salema in the study area during
the day, but at night frequently encountered
solitary individuals (Figure 8), or loosely spaced
groups of four to six. Usually they swam high in
the mid-waters above the open fields of Dictyop-
teris within 10 m of the forest. Their first appear-
ance in the evening consistently occurred about 40
min after sunset, apparently after they had come
from some distance away. The relatively few
times we saw this species in daylight (always more
than 400 m from the study area), it swam in schools
of more than 50 individuals, closely spaced and
seemingly inactive, at middepths within the for-
est. Reportedly this fish reaches 255 mm (Miller
and Lea 1972).'

Fresh material filled the stomachs of all five
specimens (163-180 mm, x = 170) collected for
study of food habits. They were taken at night,
more than 3 h after sunset, and before daybreak,
and so nocturnal feeding is apparent. All three

Figure 8. -A solitary salema, Xenistius
californiensis, swims above the sea floor at
night.

585

FISHERY BULLETIN: VOL. 74, NO, 3

taken before midnight had their intestines empty;
this, considering also the inactive appearance of
those in diurnal schools, suggests lack of daytime
feeding. Quast (1968) reported a high incidence of
empty stomachs in specimens that he collected
during the day, but did not relate this to nocturnal
feeding.

All food material in the guts of specimens
collected during this study are organisms that
occured in the water column. Major categories of
prey, which included species and species groups,
are listed below in order of their rank as prey.

1. GAMMARIDEAN AMPHIPODS (100: 44.8: 38.2)

Bafea transversa (100: 26.0: 30.0); Ampithoe phimulosa (20:
5.2: 3.0); Encthonias braziliensis (20: 2.0: 1.0); lysianassid spp.
(20: 0.2: 0.4); Gitannpsis vilordes (20: 0.2: 0.2); Ampithoe spp.
(20: 4.0: 1.0); unidentified species (60: 7.2: 2.6).

2. MYSIDS (100: 22.0: 28.0)

Siriella pacifiica (100: 20.2: 26.8); erythropinid sp. (60: 1.8:
1.2).

3. POLYCHAETES. SWIMMING (40: ?: 20.0)

unidentified species, mostly epitokus nereids.

4. CUMACEANS (60: 2.0: 2.6) "

Cijclaspis nuhila (60: 1.6: 2.2); unidentified juveniles (20: 0.4:
0.4).

5. CAPRELLID AMPHIPODS (40: 5.6: 3.0)

Caprella pUkUgita (40: 4.4: 1.8); C. californica (40: 1.2: 1.2).

6. OSTRACODS (80: 2.2: 1.0)

Parasterope sp. A (60: 1.2: 0.6); Cycloleberis lobiancoi (20: 0.6:
0.2); Vargula americana (20: 0.4: 0.2).

7. NEBALIACEANS(20:1.0:3.0)

Nebalia piigettensis.

8. ISOPODS (40: 2.4: 1.2)

Cirolana harfordi (20: 0.8: 0.4); Paracercies sp. (20: 1.0: 0.2);
Excorallana kathae (20: 0.4: 0.2); gnathiid juveniles (20: 0.2:
0.4).

9. FISHES (20: ?: 1.8)
scales.

10. CARIDEAN LARVAE (20: 1.2: 0.6)

unidentified.

11. CARIDEAN ADULTS AND JUVENILES (20: 1.2: 0.6)

unidentified.

12. REPTANTI AN ZOEA (20: 3.6: 0.4)

unidentified.

13. CALANOID AND CYCLOPOID COPEPODS (20: 0.2: 0.2)

calanoid, Labidocera sp.

This list indicates a diet much like that of
salema collected from a kelp bed near La Jolla by
Quast (1968), although Quast questioned the
validity of his data because of the collecting
methods used.

Seriphus pol/tus—queenfish

The queenfish, which can grow to 304 mm (Miller
and Lea 1972), consistently appeared in the study
area about 40 min after sunset and remained
active there throughout the night. Generally,
solitary individuals, or loosely spaced groups of
two to six swam several meters above the sea floor,
usually close to the seaward edge of the kelp
forest, but frequently above the open fields of
Dictyopteris. Then, shortly after first morning
light, 40 to 50 min before sunrise, they abruptly
left the area.

During the day queenfish hover in dense, rela-
tively inactive schools close to shore (Figure 9), but
we have not seen them within 1.5 km of the study
site in daylight. Limbaugh (1955), presumably
assessing the daytime situation, stated:

Figure 9.-A daytime aggregation of
queenfish, Seriphus politus.

586

HOBSON and CHESS: TROPHIC INTERACTIONS

"Queenfish school in tightly packed aggregations
over sandy bottom."

Four of five individuals (124-171 mm, x = 148)
collected shortly after they had arrived in the
study area at nightfall had an empty gut, and the
fifth contained just a single freshly ingested
shrimp (unidentified). We conclude that these
individuals had passed the previous day without
feeding. The evidence further suggests they do
not feed while en route from daytime schooling
sites to their feeding ground in the study area.

All 31 specimens (114-193 mm, x = 151) sampled
in the study area at night, later than 3 h after
sunset and before first morning light, had material
in their guts— much of it fresh. All prey belonged
to groups known to occur in the water column.
Limbaugh (1955) reported that this species feeds
on "small free-swimming crustaceans and fish."
Below are ranked the species and species groups
taken as prey by this fish.

1. MYSIDS (84: 22.5: 44.7)

Siriella pacifica (84: 21.0: 39.6); Acanthomysis sculpta (52:
1.4: 5.0); erythropinid sp. (6: 0.1:<0.1).

2. GAMMARIDEAN AMPHIPODS (89: 16.0: 21.6)

Batea transversa (84: 15.6: 20.2); Ampelisca sp. (9: 0.2: 0.3);
lynsianassid spp. (6: 0.1: 0.2); Ampithoc sp. (3: 0.1: 0.1);
unidentified (3: <0.1: 0.8).

3. POLYCHAETES, SWIMMING (31: 0.8: 21.8)

Epitokous nereids (22: 0.7: 18.4); unidentified (9: <0.1: 3.4).

4. CARIDEAN ADULTS AND JUVENILES (44: 0.7: 5.9)

Eualus herdmani (28: 0.4: 2.0); Hippolyte clarki (19: 0.3: 2.0);
unidentified (3: <0.1: 1.9).

5. ISOPODS (34: 0.7: 3.6)

Paracercies sp. (22: 0.5: 1.7); gnathiid juveniles (13: 0.1: 0.2);
Limnoria sp. (3: 0.1: 1.1); Excorallana kathae (3: <0.1: 0.5);
Cirolana harfordi (3: <0.1: 0.1).

6. FISHES (6: 0.6: 1.4)

scales.

7. NEBALIACEANS(6:0.1:0.2)

Nebalia pugettensis.

8. OSTRACODS(13:0.3:0.2)

Vargula americana (6: 0.1: <0.1); Cycloeberis lobiancoi (3:
0.1: <0,1); Parasterope sp. A (3: 0.1: <0.1).

9. CARIDEAN LARVAE (3: <0.1: 0.3)

unidentified.

10. BRACHYURAN MEGALOPS (3: <0.1: 0.2)

unidentified.

11. EUPHAUSID ADULTS AND JUVENILES (3: <0.1: <0.1)

unidentified.

12. CUMACEANS (3: <0.1: <0.1)

Cyclaspis nuhila.

A single small juvenile queenfish, 38 mm long,
was collected on 2 November shortly before first
morning light as it swam alone close over the sand.
Its full gut contained mysid Siriella pacifica,
gammaridean amphipod Batea transversa, and
isopod Limnoria sp. All of these forms are also

prey of larger queenfish, but those taken by this
small individual were less than half the size of prey
routinely taken by the larger fish.

Material that we collected at La Jolla in 1971
included some information on smaller juveniles.
Ten individuals (10-27 mm, x = 19) were collected
on the same day during the hour before first
morning light-all from the stomachs of larger
individuals of their own species. Of these, only the
two largest, 23 and 27 mm, contained prey of the
types taken by larger conspecifics: mysids and
gammaridean amphipods constituted 99% of the
diet of these two, with calanoid copepods repre-
senting the remainder. In contrast, calanoid
copepods were the major prey of the seven smaller
individuals (in six, 80% of the total diet). Fish
larvae (in one, 11% of the total diet), and cladocer-
ans (in one, 9% of the total diet), constituted the
rest. These limited data indicate that the
queenfish, like the olive rockfish above, changes as
it grows from a diet of copepods to one of mysids
and other plankters that appear after dark. The
queenfish, however, seems to make the change at a
smaller size, perhaps because it has a larger mouth.
Moreover, the data fail to show that the queenfish,
like the olive rockfish, feeds by day when subsist-
ing on copepods.

Hyperprosopoti argenteutn— walleye surfperch

The walleye surfperch, which can grow to 304
mm (Miller and Lea 1972), consistently schooled
during the day in about 2 to 5 m of water over sand
at the edge of the forest at the head of Fisher-
men's Cove. Usually these schools included 20 to
more than 100 closely spaced individuals. Members
of these schools appeared inactive, an impression
supported by the eight empty guts found in nine
individuals (115-173 mm, .f = 140) taken during
midafternoon (and the ninth contained only
well-digested fragments). Presumably describing
the daytime situation throughout southern
California, Limbaugh (1955) stated: "They school
in an aggregate cloud . . . over sand patches among
rocks."

The schools dispersed at nightfall, and many
individuals spread along the seaward edge of the
forest at the perimeters of the cove. They swam
individually (Figure 10) or in small groups 1 to 3 m
above the bottom, usually over sand within a few
meters of, but sometimes within, the forest. Of the
35 (60-151 mm, x = 111) collected in the study area
between 4 h after sunset and daybreak, only one

587

FISHERY BULLETIN: VOL. 74, NO. 3

Figure lO.-A solitary walleye surfperch, Hyperproaopon argen-
teiim, swims in the water column at night.

was empty; the rest were full of prey, much of it
fresh.

Clearly, this is a nocturnal fish. Those seen in the
study area at night, however, tended to be smaller
on the average than those seen in the diurnal
schools, suggesting that the larger fish might
range farther away. All prey in the 34 individuals
containing identifiable material were organisms
that occur in the water column, as listed below.

3. ISOPODS (72: 21.1: 10.2)

Paracercies sp. (65: 19.5: 7.6); gnathiid juveniles (21: 1.0: 0.6);
Pentidotea resecata (15: 0.2: 0.6); Excorallana kathae (3: 0.1
1.0); Cirolana diminuta (15: 0.2: 0.2); Rocinella belliceps (6:
<0.1: 0.2); ExoRpheroma sp. (6: <0.1: <0.1); idoteid sp. (3: <0.1
0.1).

4. CAPRELLID AMPHIPODS (41: 2.1: 6.0)

Caprella pilidigita (24: 1.4: 4.7); C. caUforn tea (21: 0.7: 1.2); C.
penantiK (3: <0.1: <0.1); Tritella laevis (3: <0.1: <0.1).

5. POLYCHAETES, SWIMMING (35: <0.7: 6.4)

epitokous nereids (9: 0.7: 5.5); unidentified fragments (26: ?:
0.9).

6. OSTRACODS (62: 1.7: 1.6)

Paraatvrope sp. A (38: 0.7: 0.6); Ciichleberia lohiancoi (23: 0.5:
0.4); Philomedef! sp. (9: 0.4: 0.4); species 0 (3: <0.1: <0.1);
species N (3: <0.1: <0.1).

7. MYSIDS(21:0.6:1.3)

Siriella pacifica (15: 0.5: 1.2); Acanthomysis sculpfa (3: <0.1:
<0.1); unidentified fragments (3: 0.1: <0.1).

8. CARIDEAN ADULTS AND JUVENILES (21: 0.9: 0.9)

Hippolyte clarki (3: 0.2: 0.1); unidentified (24: 0.7: 0.8).

9. BRACHYURAN MEGALOPS (26: 0.7: 0.5)

unidentified.
TANAIDS(15:0.5:0.5)

Leptochelia duhia (6: 0.2: 0.3); unidentified (9: 0.3: 0.2).
NEBALIACEANS (6: <0.1: <0.1)

Nebalia ptigettensis.
CARIDEAN LARVAE (3: <0.1: <0.1)

unidentified.

1. GAMMARIDEAN AMPHIPODS (100: 63.6: 47.0)

Bafea transversa (85: 39.8: 24.2); Am pit hoe spp. (41: 3.3: 3.8);
Hyale nigra (9: 2.8: 2.9); Ericthonias braziliensis (15: 0.5:
1.1); Ampelisca sp. (15: 0.6: 1.0); Synchelidium sp. (24: 0.4
0.3)rlysianassid spp. (15: 0.6: 0.2); Heterophilias seclusus (6
0.1: 0.1); Photis sp. (3: <0.1: 0.1); Paraphoxus sp. (3: <0.1
<0.1); Aoroides colu mbiae (3: <0.1: 0.1); unidentified (91: 15.4
13.2).

2. CUMACEANS (85: 52.9: 25.2)

Cyclaspis nubila (76: 51.2: 24.8); Cumella sp. A (18: 1.7: 0.4).

B rachyistius frenatus— kel p perch

The kelp perch, which Miller and Lea (1972)
claimed can attain a length of 214 mm, was
numerous close among the rising stands of giant
kelp. It often aggregated immediately under the
canopy (Figure 11), but occurred along the entire
length of the plants from water's surface to the

Figure 11. -Kelp perch, Brackyistius
frenatus, aggregated close to kelp,
pluck zooplankters from the water
column during the day.

588

HOBSON and CHESS: TROPHIC INTERACTIONS

sea floor, with larger individuals mostly in the
lower regions. Far fewer numbers also occurred
close above low fields of benthic algae some dis-
tance from the kelp forest. It assumed similar
attitudes in the same places during both day and
night, but after dark there seemed to be more of
them in the mid-waters along the outer edge of the
kelp.

Most kelp perch feed by plucking material from
the surface of algae, but plankton-feeding is
widespread, especially among those aggregated in
the mid-waters at the edges of the forests. Lim-
baugh (1955) reported that the kelp perch feeds on
small crustaceans, particularly those that occur on
giant kelp. Quast (1968) also reported a predomi-
nantly crustacean diet, with a preponderance of
amphipods, but also including mollusks and
bryozoans.

Preliminary assessment of our food-habit data,
along with direct observations, showed that in this
species it is primarily the smaller individuals that
feed on plankton. Consequently, we consider for
this paper only those less than 100 mm long,
leaving the larger individuals for a later paper.
This point is drawn somewhat arbitrarily, al-
though plankters generally become noticeably less
prevalent in the diet at about this size. With kelp
perch more so than with the other species treated
in this paper, however, many of the individuals
considered had mixed a diet of plankters with
organisms plucked from a substrate. Bray and
Ebeling (1975) reported that kelp perch feed
mainly on tiny plankters, mostly copepods, based
on a sample of predominantly small individuals
(43-142 mm, ,r = 103).

All 35 specimens (40-99 mm .f = 81) collected
during the afternoon as they swam over various
locations in the study area, usually close to kelp,
contained food, much of it fresh. On the other
hand, of 34 specimens (38-99 mm, .r = 76) collected
during the 2 h of night before first morning light
25 (74%) were empty. The other nine, however,
contained food, including relatively fresh items.
Thus, although the kelp perch within this size
range clearly fed mostly by day, some apparently
fed at night. Individuals evidencing nocturnal
feeding ranged from 81 to 99 {x = 95) mm long,
and so were among the larger ones in the sample.
Recognizing that the contrasting conditions
between day and night undoubtedly influenced the
composition of the diet, food data from individuals
collected during the afternoon (when presumably

most fresh items in the gut had been taken by day)
were considered separately from food data from
individuals collected during the last hours of the
night (when presumably most fresh items in the
gut had been taken after dark).

In addition to the high incidence of empty guts
in kelp perch collected at night, the guts of those
that had taken prey after dark averaged 50% full,
compared with an average of 72% full for the day
feeders. Furthermore, the night feeders contained
an average of 38 prey items, compared with an
average of 252 for the day feeders (at least in part,
however, this difference reflects the larger size of
nocturnal prey). These data strengthen our con-
clusion that over the size range studied, nocturnal
feeding is relatively unimportant to this species.
Bray and Ebeling (1975) also noted that kelp perch
feed mainly by day.

Foods taken by individuals that had been feed-
ing during the day are ranked below:

1. CALANOID AND CYCLOPOID COPEPODS (94: 157.7: 49.1)

calanoids, including Calannn pacijicax, and Rhincalanus
nasutus (71: 137.2: 44.2); cyclopoids, including Corycaeus sp.
and Oticea sp. (74: 20.5: 4.9).

2. GAMMARIDEAN AMPHIPODS (63: 57.9: 37.0)

Microjassa litodes (46: 23.1: 15.1); Ericthonias braziliensis
(14: 3.9: 2.6); Gitanopsifi vilordes (11: 3.0: O.l); Ampithoe spp.
(3: 0.1: 0.2); Hyale nigra (3: 0.3: 0.2); Batea transversa (3: 0.1:
0.1); unidentified (63: 27.4: 18.7).

3. CLADOCERANS(37:26.1:6.9)

Evadne sp.

4. CIRRIPEDIAN LARVAE (31: 1.8: 0.9)

cvpris stage.

5. POLYCHAETES, NONSWIMMING (11: 0.4: 1.9)

Spirorbis sp. (9: 0.4: 1.8); unidentified (3: 0.1: 0.1).

6. HARPACTICOID COPEPODS (14: 1.6: 0.7)

Porcellkiium sp. A (11: 1.5: 0.6); Porcellidium sp. B (3: 0.1:
0.1).

7. OSTRACODS (26: 0.6: 0.6)

Cythereis sp. (17: 0.3: 0.2); Philomedes sp. B (11: 0.2: 0.1);
unidentified sp. C (3: 0.1: 0.3).

8. CAPRELLID AMPHIPODS (9: 0.3: 0.7)

Caprella pUidigita (6: 0.2: 0.4); C. califurnica (3: <0.1: 0.3).

9. FISH EGGS (14: 0.4: 0.3)

unidentified.

10. PELECYPODS(11:0.4:0.3)

Hiatella arctica (9: 0.3: 0.3); Halodakra brunnea (3: <0.1:
<0.1).

11. ISOPODS(14:0.7:0.2)

Paracercies sp. (6: 0.5: 0.1); gnathiid juveniles (6: 0.1: <0.1);
unidentified fragments (3: 0.1: 0.1).

12. BRYOZOAN LARVAE (9: 0.1: 0.2)

cyphonautes.

13. CARIDEAN LARVAE (9: 0.2: <0.1)

unidentified.

14. FISHES (6: 0.2: <0.1)

unidentified larvae.

589

FISHERY BULLETIN: VOL. 74, NO. 3

15. MYSIDS(3:<0.1:<0.1)

Siriclla pacifica.

16. CUMACEANS(3:<0.1:<0.1)

CyclaspiK niihiki.

17. CARIDEAN ADULTS AND JUVENILES (3: <0.1: <0.1)

Hippolyte clarki.

Although these fish preyed heavily on zoo-
plankters, clearly many of the organisms in the
above list were plucked from a substrate. The
major gammaridean, Microjas.sa litodes, was
never seen or taken by us in the water column, but
was a predominent form on the surface of giant
kelp (Hobson and Chess in prep.) Similarly, the
many forms known to occur in the water column
only at night, e.g., Siriella pacifica, Cyclaspi>;
nubila, Pa race tries sp., Batea transversa, and
Hippolyte clarki were probably plucked by these
day feeders from the algae or sand where they
occur in the daytime.

Foods taken by individuals that had been feed-
ing at night are ranked below.

1. GAMMARIDEAN AMPHIPODS (100: 3.3.8: 71.1)

Batea tram^rcrsa (66: 8.1: 22.1); Ericthuiiiax hrazilieusis
(44: 0.9: 2.6); Micnijasya lito(h'!< {22: 1.6: 1.7); Ampithoe spp.
(22: 0.3: 0.9); Hjiah nigra (11: 0.4: 0.9); Aoroides columhiat
(11: 0.1: 0.3); unidentified, at least some probably juveniles
of the above (100: 22.4: 42.6).

2. CAPRELLID AMPHIPODS (66: 4.4: 15.0)

Caprella califoniica (55: 2.0: 9.5); C. pilidigifa (11: 2.0: 4.1);
unidentified (22: 0.4: 1.4).

3. ISOPODS (44: 0.5: 6.6)

Paracercies sp.

4. CARIDEAN ADULTS AND JUVENILES (55: 0.7: 5.1)

unidentified.

5. MYSIDS(22:0.2:1.1)

Siriella pacifica (12: 0.1: 1.0); erythropinid sp. (11: 0.1: 0.1).

6. POLYCHAETES, NONSWIMMING (22: 0.2: 0.3)

Spirnrbis sp.

7. OSTRACODS (22: 0.2: 0.2)

Parasterope sp. A (12: 0.1: 0.1); Cytlicre.^ia sp. (11: 0.1: 0.1).

8. FORAMINIFERANS(11:0.1:0.2)

unidentified.

9. HARPACTICOID COPEPODS (11: 1.0: 0.1)

Porcellidium sp. A.

Two specimens also contained fragments of
algae {Macrocystis in one, Sargassiun in the other)
that probably had been taken incidentally along
with prey. Clearly this fish took at least some of its
nocturnal prey from a substrate- Spirorbis sp., for
example. Nevertheless, because the diet is com-
prised mostly of organisms that swim in the water
column at night, we believe this was probably
where most of them were taken. Most of these
prey organisms also occurred on rocks and algae
after dark, but if substrate-feeding had

predominated, we would have expected a greater
proportion of strictly substrate-dwelling forms.

Oxyjulh californica—sehont2i

The senorita, which can attain a length of 250
mm (Miller and Lea 1972), is perhaps the most
widespread fish in nearshore habitats at Santa
Catalina Island. It is strictly a diurnal species that,
like other labrids, rests under cover on the sea floor
at night (Hobson 1971). Often during the day it
swims in large assemblages 1 to 2 m above the sea
floor (Figure 12).

Most senoritas feed by plucking material from
the surface of algae-often from algae drifting as
fragments in the mid-waters-but plankton-feed-
ing is widespread, and predominates in smaller
juveniles. Limbaugh (1955) concluded that the
senorita is an omnivorous carnivore that feeds "on
almost any animal protein." Hobson (1971) found
that specimens between 110 and 195 mm long had
fed primarily on bryozoans that encrust algae, and
on caprellid amphipods. Quast (1968) reported the
principal foods to be small gastropods and crus-
taceans commonly associated with algae, but
noted that specimens 50 to 60 mm long had fed
heavily on copepods, ostracods, and bryozoan
larvae.

Direct observations, complemented by our food
habit data (see below), agree that smaller in-
dividuals mostly pluck their prey from the water
column, whereas larger individuals mostly pluck
their prey from some substrate. In this respect,
then, the senorita is similar to the kelp perch,
described above. So, as with the kelp perch, this
paper considers only those individuals less than
100 mm long, leaving the larger individuals for a
later paper. We have better reason for drawing
the dichotomy at this point with the sefiorita than
with the kelp perch: the smallest senorita we found
containing prey obviously plucked from a sub-
strate was 101 mm long, and although planktivo-
rous habits predominated in certain individuals up
to 175 mm (which were among the largest taken),
most over 100 mm seemed to feed primarily on a
substrate. So unlike the diverse feeding habits of
smaller kelp perch, the smaller senoritas seemed
strictly planktivorous. Bray and Ebeling (1975)
stated: "Unlike kelp perch, senoritas did not
exploit the plankton as a major source of food."
Although this view would seem to disagree with
our findings, their samples of the two species were
not comparable on this point. Most of their kelp

590

HOBSON and CHESS: TROPHIC INTERACTIONS

Figure 12.- An aggregation of senoritas, Oxyjulis californica, passes along the edge of a kelp forest during the day.

perch were small, as noted above, whereas their
senoritas were relatively large (110-223 mm, i =
169).

All 24 specimens (19-99 mm, x = 51) collected
during the afternoon as they swam in groups
above the sea floor were full of relatively fresh
prey, as ranked below:

1. CALANOID AND CYCLOPOID COPEPODS (100: 68.7: 74.1)

calanoids (75: 44.3: 43.8); cyclopoids, including Corycaeus sp.
and Oncaea sp. (67: 24.4: 30.3).

2. BRYOZOAN LARVAE (58: 7.3: 4.3)

cyphonautes.

3. HARPACTICOID COPEPODS (42: 7.1: 4.8)

Microsetella sp. (25: 3.1: 2.1); Porcellidium sp. A (8: 0.1: 0.2);
unidentified spp. (21: 3.9: 2.5).

4. CIRRIPEDIAN LARVAE (46: 2.8: 2.8)

cypris stage.

5. GAMMARIDEAN AMPHIPODS (25: 1.8: 4.7)

unidentified fragments.

6. CLADOCERANS (21: 4.8: 5.4)

Evadne sp.

7. MOLLUSK LARVAE (46: 1.1: 1.0)

veligers.

8. FISH EGGS (4: 0.1: 0.3)

unidentified.

9. RADIOLARIANS(4:0,1:
unidentified.

.0.1)

With the likely exception of the gammarideans,
which were unidentifiable, all of the items in the
above list are organisms present in the water
column at the time these fish were feeding.

Chromis pi/nctip/nm's—hhcksmith

The blacksmith, which can attain a length of 300
mm (Miller and Lea 1972), is probably the most
numerous fish in the nearshore waters at Santa
Catalina Island. During the day it concentrated
along the seaward edge of the kelp forests, but
occurred in varying numbers in most nearshore
habitats, usually aggregated in the mid-waters
(Figure 13). At nightfall it sheltered among the
rocks, often considerable distances inshore from
where it spent the day.

Other species of the genus Chromis are wide-
spread in tropical seas, where they are known to be
planktivores, e.g.: West Indies (Randall 1967); Gulf

591

FISHERY BULLETIN: VOL. 74, NO. 3

Figure 13.-Blacksmiths, Chromis punctipinnis, aggregated at the edge of a kelp forest, pluck zooplankters from the water column

during the day.

of California (Hobson 1968); Hawaii (Swerdloff
1970; Hobson 1974). It is also well-known that C.
pxucfipinnis is a planktivore. Limbaugh (1955)
noted that it feeds on "particulate plankton such
as small fish, squid, and crustaceans^ and "may
materially affect the amount of plankton entering
kelp beds because they eat it as it enters." Quast
(1968) listed the principal food of the blacksmith as
"minute swimming Crustacea and crustacean eggs
and larvae gleaned from open water species of
kelp beds and over rocky areas." In taking its tiny
prey from the water column in what seems a
visually directed action, the blacksmith suddenly
thrusts both of its highly protrusible jaws forward,
then immediately retracts them, presumably
sucking plankters into its rapidly expanding oral
cavity. This way of feeding has also been noted
among its tropical congeners (Swerdloff 1970;
Hobson 1974).

Aggregations of blacksmiths feeding on plank-
ton occurred throughout the water column, with
each member of an aggregation acting indepen-

dently. They aggregated according to size: the
discrete aggregations of small juveniles (which
first appeared inshore during late summer, when
about 15 to 25 mm long) generally were closer to
the sea floor than were aggregations of the adults.
Blacksmiths fed throughout the day, but the
rate at which they ingested prey varied. In the
tropical Atlantic, Eupomacentrus partitus, an-
other planktivorous pomacentrid, feeds more
rapidly with increased light and with increased
current (Stevenson 1972). Blacksmiths, too, feed
more rapidly in a current than at slack water,
presumably (as Stevenson suggested of E. parti-
tus) because they receive more plankters. To
measure this effect, we counted the characteristic
mouth movements of feeding adult blacksmiths,
first in a moderate current, and then at slack
water. The observations were made during
midaf ternoon under a clear sky at a depth of 5 m in
10 m of water. The fish were part of an aggrega-
tion with members ranging from about 109 to 130
mm long (these being the sizes of the two in-

592

HOBSON and CHESS: TROPHIC INTERACTIONS

dividuals collected later judged to be the largest
and smallest in the group). In the moderate cur-
rent, with the giant kelp lying over at about 25°
(attempts to measure the current in this habitat
proved unsatisfactory owing to complex eddy
systems), 10 individuals (selected haphazardly)
plucked at plankters 50 to 73 {x = 58) times during
1-min periods. One hour later, when there was no
discernible current, 10 individuals in a similar, if
not the same aggregation, each plucked at
plankters 30 to 51 (.f = 39) times during 1-min
periods. Probably there is an optimum current
speed beyond which the fish find the increasing
diflSculty of maintaining station outweighs the
advantage of added food. We lack data on feeding
rates, but have noted that in strong currents
blacksmiths abandon the open places within the
forest, where they had been dispersed and feed-
ing, and concentrate in dense numbers close in the
lee of individual kelp columns.

Of 41 adults (92-145 mm, .r = 118) collected from
aggregations in the water column throughout the
study area during the afternoons, 36 were full of
food, much of it fresh. The other 5, all collected
during midafternoon along the margin of the
forest bordering the inshore edge of the Dictijop-
teris field, were empty. All prey taken by these
blacksmiths, ranked below, are forms we have
collected in the water column during the day.

1. LARVACEANS (100: 448.1: 57.5)

most of them Oikopleura spp.

2. CALANOID AND CYCLOPOID COPEPODS (100: 256.3:

33.7)

calanoids, including Calanus pacificus, Acartia tonsa,
Labidocera sp., and Rhincalanus nasutits (100: 2.53.6: 32.4);
cyclopoids, including Conjcaeus sp., Oncaea sp., and Oith-
ona sp. (72: 2.7: 1.3).

3. FISH EGGS (69: 17.9: 4.1)

unidentified.

4. CLADOCERANS (75: 24.6: 2.5)

Evadne spp.
.5. CARIDEAN LARVAE (33: 1.4: 0.9)
unidentified.

6. EUPHAUSID LARVAE (33: .5.1: 0.6)

calyptopis stage.

7. CIRRIPEDIAN LARVAE (33: 3.5: 0.6)

cypris stage.

8. BRYOZOAN LARVAE (17: 1.2: 0.2)

cyphonautes.

9. CHAETOGNATHS(3:0.1:0.5)

unidentified.

10. REPTANTIAN ZOEA (3: 0.4: 0.1)

unidentified.

11. HARPACTICOID COPEPODS (3: <0.1: <0.1)

Microsetella sp. A.

12. FISHES (3: <0.1:<0.1)

unidentified larvae.

13. ISOPODS(3:<:0.1:<0.1)
gnathiid juvenile.

In feeding so heavily on larvaceans, their major
prey, these adult blacksmiths differ from other
species treated in this report. Significantly, how-
ever, larvaceans are also major prey of other
species of Chromis elsewhere, e.g., in Hawaii
(Hobson 1974) and in the West Indies (Randall
1967). Probably larvaceans are important food of
the blacksmith throughout its range, even though
they have gone unreported in previous food-habit
studies of this species. Larvaceans are difficult to
recognize, especially if digestion is advanced or
preservation faulty, and this may account for
them going unreported.

Because juvenile blacksmiths were in feeding
aggregations distinct from those of the adults, we
analyzed their gut contents separately. Of 14
juveniles (16-47 mm, .f = 34) collected from ag-
gregations during the afternoon, all were full of
food, much of it fresh. All prey, ranked below, are
forms that we have collected from the water
column during the day.

1. CALANOID AND CYCLOPOID COPEPODS (100: 394.4:

.54.1)

calanoids, including Acartia tonxa (100: .366.2: 50.5); cy-
clopoids, including Corijcaeua sp. and Oncaea sp. (93: 28.2:
3.6).

2. LARVACEANS (93: 48.9: 26.4)

most of them Oikopleura spp.

3. CLADOCERANS (100: 108.8: 12.5)

Evadne spp.

4. CIRRIPEDIAN LARVAE (79: 63.6: 4.6)

cypris stage.

5. BRYOZOAN LARVAE (79: 63.6: 4.6)

cyphonautes.

6. HARPACTICOID COPEPODS (.57: 6.4: 0.8)

MicroKetella sp. A.

7. FISH EGGS (43: 2.1: 0.7)

unidentified.

Diff"erences in the diet between juvenile and
adult blacksmiths can be related to the sizes of the
various organisms in the water column. Most prey
of the juveniles were less than 0.5 mm long.
Compared to the prey of adults these included
more cladocerans, copepods, and larvae of barna-
cles and bryozoans, but fewer larvaceans and fish
eggs (there were no larvaceans in the smallest
blacksmith, 16 mm long, and no fish eggs in all six
<34 mm).

During the day, the heaviest concentrations of
adult blacksmiths in the vicinity of the study area
were at the mouth of the cove seaward of the kelp

593

forest. The sea floor in this region is sand, and lies
under more than 30 m of water. Because black-
smiths habitually settled among rocks at night,
the offshore feeders migrated to resting areas
inshore at day's end. At the migration's peak,
groups of 100 or more blacksmiths spaced perhaps
50 m apart streamed along established routes.

As the migrators swam between feeding
grounds and shelter areas, they passed among
many other blacksmiths, most of which were
actively feeding and which gave the migrators no
overt notice. Most of the blacksmiths in the vicin-
ity of the study area were nonmigrators that
found nocturnal shelter among rocks lying below
their mid-water feeding grounds.

Most of the blacksmiths within the forest bor-
dering the study area began descending toward
the sea floor by sunset, and by 35 min after sunset
the vast majority had taken shelter among the
rocks. They rested here throughout the night, and
their lack of feeding during this period is indicated
by the empty guts we found in all 11 individuals
(111-143 mm, .r = 122) collected among rocks
during the 2 h immediately before first morning
light.

In the morning, blacksmiths among the rocks

FISHERY BULLETIN: VOL. 74, NO. 3

became noticeably active about 40 min before
sunrise. They began to rise among the kelp
columns about 25 to 30 min before sunrise, and to
feed about 5 to 10 min later. At about the same
time that blacksmiths within the forest were
rising into the water column, the migrating in-
dividuals streamed along their courses to the
ofi'shore feeding grounds, reversing the courses
they had followed inshore the night before.

DISCUSSION

Trophic relationships among the fishes and
zooplankters near shore at Santa Catalina Island
diff"er strikingly between day and night (Table 9),
broadly paralleling the situation described earlier
in the water column above tropical reefs (Hobson
1965, 1968, 1972, 1973, 1974). This section discusses
these difi'erences and their evolutionary
implications.

The Mid-Waters in Daylight

Zooplankters populating the nearshore water
column at Santa Catalina during the day-
including radiolarians, cladocerans, copepods, and

Table 9.- Percent of each fish species that took prey in each major food category.

Day feeders

Night feeders

olive rockfish

< 55 mm

, n = 42

olive rockfish

>65 mm, n =

kelp perch <

100 mm, 1

n = 35

kelp roc

;kfish,

n = 20

senorita

« 100 mm, n =

= 24

salema.

n = 5

blacksmith >

90 mm, n

= 36

queenfish, n ^

= 32

blacksmith <

50 mm, n

= 14

walleye

surfperch, n =

= 34

kelp perch ■=

100 mm.

n = 9

Major food category

Radiolarians

Polychaetes, swimming

Mollusk larvae

Cladocerans

100

Ostracods

Calanoid and cyclopoid copepods

100

Harpacticoid copepods

Other copepods

Cirripedian larvae

Nebeliaceans

Mysids

100

Cumaceans

Tanaids

Isopods

Gammaridean amphipods

100

Caprellid amphipods

Euphausid larvae

Euphausid adults and juveniles

Caridean larvae

Caridean adults and juveniles

Reptantian zoea

Brachyuran megalops

Bryozoan larvae

Chaetognaths

Larvaceans

100

Fish eggs

Fishes

Other

594

HOBSON and CHESS: TROPHIC INTERACTIONS

various larval forms (see Tables 1, 2)-tend to be
less than 2 mm in their greatest dimension. Forms
appreciably larger than this— including chaetog-
naths and some larvaceans-tend to be transpar-
ent. These organisms are equally numerous in the
water column at night, and none are residents of
the study area. The species are widespread in the
water columns of the various inshore habitats, and
also offshore.

This assemblage resists a common label. Most of
the species have been considered holoplankton
(planktonic throughout the whole of their life
histories: Newell and Newell 1963), but this term
excludes the larval forms so prominent here. The
larval forms generally are considered meroplank-
ton (planktonic during some stage in their life
histories, but benthonic during some other: Newell
and Newell 1963), but this term has been used in
general reference to organisms that are plank-
tonic at night, but benthonic by day (e.g., Williams
and Bynum 1972). As noted above, we do not use
these terms because they fail to define categories
meaningful to the concepts developed in this
paper.

Fishes that forage in the water column by day
have certain characteristics relating to the prob-
lems they face as diurnal planktivores. Sig-
nificantly, of the four diurnal planktivores studied
at Santa Catalina, three-the senorita, the kelp
perch, and the small juvenile olive
rockfish— outgrow this habit. Apparently as they
grow larger they find the tiny organisms in the
mid-waters increasingly inappropriate as prey.
We believe that each is limited in taking very
small prey by the size and structure of its mouth, a
problem solved by changing either feeding place,
or feeding time. Thus, the senorita and kelp perch
(noted by Hubbs and Hubbs 1954, to have similar
dentition and feeding habits) increasingly aban-
don the water column as a hunting ground as they
grow and shift to prey on organisms that live on
algae. The small juvenile olive rockfish, on the
other hand, continues to feed in the water column,
but assumes nocturnal habits that bring it into
contact with the larger organisms that rise above
the sea floor at night (see below). The senorita and
kelp perch, both relatively small-mouthed species,
generally shift their food habits when about 100
mm long; the olive rockfish, with a much larger
mouth (compare Figures 6, 11, 12), generally shifts
when under 65 mm long.

The fourth diurnal planktivore studied at Santa
Catalina, the blacksmith, retains its planktivorous

diet through adulthood. It does so despite growing
to a relatively large size because it has, among
other adaptive features, a small mouth specialized
for this habit. Judging from its numbers, the
blacksmith is highly successful in the warm tem-
perate waters of southern California. But it does
not range far into the colder waters northward,
and all its many congeners live in the tropics. The
blacksmith embodies morphological features un-
characteristic of temperate-zone fishes, but which
are widespread among tropical species. In writing
of reef fishes in the tropical Atlantic Ocean, Davis
and Birdsong (1973) described morphological
specializations adaptive for foraging on small
organisms in the mid-waters, and although they
do not make the point, all their examples are
species that feed by day. Especially striking are
the modifications of head and jaws, including
dentition, that permit even relatively large in-
dividuals to effectively capture tiny prey in open
water.

The Mid-Waters at Night

The nocturnal zooplankton include, in addition
to the organisms also present during the day, a
large array of forms that rise at the onset of
darkness from daytime shelters on, in, or close to
the sea floor or other cover. These nocturnal
additions to the zooplankton include various poly-
chaetes, mysids, cumaceans, gammaridean and
caprellid amphipods, isopods, tanaids, carideans,
and others (see Tables 1, 2). Most exceed 2 mm in
their greatest dimension, and many are 7 to 10
mm, and longer. Unlike the full-time zooplankters,
which have no particular relation to the study area,
these part-time zooplankters are local residents
that rise at night from substrata they are closely
associated with during the day.

The nocturnal components of the zooplankton
seem to have reasons for rising into the water
column at night that are as diverse as their
morphologies. Because they have diverse habits
that are little known, we feel that terms defining
ecological categories among them are premature.
Bousfield (1973), and others, have referred to many
such forms as tychoplankton, but this term implies
that presence in the water column is by chance, or
accident-a description that fits very few, if any, of
the forms considered here. Many are nocturnal
feeders; e.g., when the mysid Siriella pacifica
moves into the mid-waters after dark, it feeds on
copepods and other smaller zooplankters. Similar-

595

FISHERY BULLETIN: VOL. 74, NO. 3

ly, nocturnal foraging may be the rule among
species like the ostracod Perasterope sp. A, the
cumacean Cumella sp. A, and the amphipod Batea
tmm^ret'i^a, where it seems the majority spend
most of the night in the water column. Working in
the tropical Atlantic Ocean, Emery (1968) noted
that polychaetes, cumaceans, and zoea rise into the
water column at night after spending the day
under reef shelter and speculated that they make
this ascent to forage. But most of the polychaetes
entering the water column after dark at Santa
Catalina are epitokus nereids, whose mid-water
activities probably relate to reproduction.

Williams and Bynum (1972) doubted the nightly
ascents of amphipods in North Carolina estuaries
relate to feeding because they subsist on detritus.
But detritus can be available to zooplankters in the
water column, as reported by Gerber and Marshall
(1974) from a coral atoll in the central Pacific.
Significantly, however, many of the amphipods
that enter the mid-waters at night appear mor-
phologically maladapted for swimming. The
oedicerotids (including Monoculodes and Sijn-
chelidium), for example, are modified for burrow-
ing in unstable sand (Bousfield 1973), and the
caprellids (Figure 5V) seem especially unsuited for
existence in mid-water. It is unlikely that such
forms are in the water column to feed, especially
as relatively small proportions of their populations
are up there. Probably these and similar forms
make only brief, or infrequent excursions into the
water column for reasons yet undetermined. Wil-
liams and Bynum (1972) suggested that relative
numbers of caprellids entering the water column
may correlate with seasonal deterioration of their
benthic habitats. They also felt that among gam-
maridean amphipods the tubicolous forms (e.g.,
Ampelisca and Ericthonias) may facilitate re-
production by entering the water column, where
mating pairs would have free access to each other.
We reject Williams and Bynum's additional
suggestion that the nightly ascent may be an
attempt to escape from predatory bottom-feeding
fishes. Most bottom-feeding fishes that prey
heavily on amphipods (and other similar organ-
isms) are diurnal. (Some of the relatively few
fishes that prey on amphipods at night, and the
circumstances surrounding this predation, will be
discussed later; Hobson and Chess in prep.) Most
predaceous fishes feed visually, and tiny, cryp-
tically hued forms on dark substrata seem to go
unseen in the low levels of illumination that
prevail at night (Hobson and Chess in prep.; see

also Hobson 1968, 1974, for accounts of the same
situation on tropical reefs). For whatever the
reason, most amphipods that fall prey to predatory
fishes after dark are species that rise into the
water column.

Compared to their diurnal counterparts, the
nocturnal planktivorous fishes are of relatively
large sizes and have large mouths— both charac-
teristics reflecting the relatively large size and
accessibility of organisms in the mid-waters after
dark.

Evolutionary Implications

Since early in the Mesozoic period, the evolution
of actinopterigian fishes has centered on a main-
stream of generalized predators (Schaeff'er and
Rosen 1961). Because such predators are adapted
for straightforward attacks at prey in exposed
positions, the water column, with its absence of
cover, has been a risky place for smaller organisms
throughout the evolution of modern nearshore
marine communities. In discussing the earliest
actinopterigian fishes, Schaeffer and Rosen stated
that food was probably first obtained by biting and
was swallowed whole. Although advances in mouth
structure have refined their means of seizing food
(Schaeffer and Rosen 1961; Gosline 1971), gener-
alized teleosts have continued to take their prey
intact. Consequently, these fishes have found
appropriate prey to be organisms large enough for
them to entrap in their relatively large mouths,
yet small enough for them to swallow whole. As
demonstrated at Santa Catalina, prey of appro-
priate size include animals that rise into the
nearshore water column after dark-mysids, am-
phipods, isopods, and others.

The present situation at Santa Catalina Island
suggests that since early times predation pres-
sures from large-mouthed, generalized predators
have influenced major evolutionary trends among
shallow-water zooplankters. Most apparent are
trends toward nocturnal planktonic activity in
those zooplankters that would spend only part of
their time in the water column, and toward
reduced size among those zooplankters that would
spend all of their time in the water column. At the
same time it would appear that each of these
trends has elicited an evolutionary response
among planktivorous fishes, as discussed below.
We do not suggested that pressures exerted in
predator-prey interactions have been the only
force shaping these trends, but we believe their

596

HOBSON and CHESS: TROPHIC INTERACTIONS

impact has been significant. (A trend toward
transparency, most developed in the larger zoo-
plankters present in the water column during the
day, is obviously adaptive for organisms threat-
ened by visually feeding predators. Although this
trend is only briefly mentioned here, its impor-
tance in defense against predators is emphasized
by Hamner et al. 1975.)

Because most generalized predaceous fishes
probably have been visual feeders since early
times, their prey would have long since found
water-column activities most safely performed
under cover of darkness. Not surprisingly, the
zooplankters that are vulnerable to large-mouthed
fishes are mostly nocturnal forms that spend the
daytime amid benthic cover. But only organisms
that have the capacity to leave the water column
can enjoy the advantage of cover during vulnera-
ble periods.

The organisms that spend all their time in the
water column have had to adapt to being fully
exposed during daylight, when the visual sense of
their predators is most eff'ective. Under this cir-
cumstance one would expect long-established
selective pressures favoring sizes smaller than
those that can be entrapped by the relatively large
mouths of generalized predators. That such selec-
tive pressures operate today among zooplankters
in daylight is well documented. Brooks and Dotson
(1965), for example, described the larger zoo-
plankters in a lake being eliminated by the plank-
tivorous clupeoid fish Alosa pseudoharengus,
reported by Emery (1973) to feed by day.

Because successful defenses in prey create
pressures that modify the offenses of predators,
early tendencies in prey toward nocturnal habits
or reduced size would have generated appropriate
evolutionary responses among predators. Cer-
tainly a long-standing selection for nocturnal
capabilities is evidenced by the many large-
mouthed predaceous fishes that forage in the
mid-waters at Santa Catalina after dark, includ-
ing the walleye, the salema, and the queenfish.
Large eyes, an obvious advantage in predators
that hunt at night, have been widely acquired by
these fishes (see Figures 6-10). Similarly, the small
mouth and other specialized features of diurnal
planktivorous fishes, like the blacksmith, clearly
are adaptive for feeding on the very small organ-
isms that constitute the diurnal zooplankton (see
Figures 11-13). In their feeding morphologies and
body forms, the nocturnal planktivores have
diverged less than have their diurnal counterparts

from the generalized predators that gave rise to
them all. This does not necessarily mean that the
diurnal planktivores are more recently evolved.
Each is the product of an equally long evolution,
and while the diurnal planktivores have been
molded by selective pressures favoring the
capacity to take tiny organisms, the nocturnal
species have been influenced during the same
period by selective pressures favoring the capacity
to detect and capture prey in the dark.

ACKNOWLEDGMENTS

We thank Russell Zimmer, and his staff at the
Catalina Marine Science Center, University of
Southern California, for making facilities availa-
ble. We also acknowledge with gratitude the
following individuals for identifying specimens of
various species cited in this study: Robert Given
and Kristian Fauchald, University of Southern
California; Abraham Fleminger and David This-
tle, Scripps Institution of Oceanography; Diana
Laubitz, National Museum of Canada; Bruce Ben-
edict and Bradley Meyers, Marine Biological
Consultants; and Jack Word, Southern California
Coastal Water Research Project. For constructive
criticism of the manuscript we thank Richard
Rosenblatt, Scripps Institution of Oceanography;
Richard Vance, University of California, Los
Angeles; John E. G. Raymont, University of South
Hampton, South Hampton, England; and E. Bous-
field, National Museum of Natural Sciences, Ot-
tawa, Canada. Johanna Alban, Southwest Fisher-
ies Center Tiburon Laboratory, National Marine
Fisheries Service (NMFS), NOAA, typed the
manuscript, and Kenneth Raymond, Southwest
Fisheries Center La Jolla Laboratory, NMFS,
NOAA, drew Figure 2.

LITERATURE CITED

BOUSFIELD, E. L.

1973. Shallow-water gammaridean Amphipoda of New
England. Cornell Univ. Press, Ithaca, 312 p.
Bray, R. N., and A. W. Ebeling.

197.5. Food, activity, and habitat of three "picker-type"
micro-carnivorous fishes in the kelp forests off Santa
Barbara, California. Fish. Bull., U.S. 73:815-829.
Brooks, J. L., and S. I. Dodson.

1965. Predation, body size, and composition of plankton.
Science (Wash., D.C.) 150:28-35.
Davis, W. P., and R. S. Birdsong.

1973. Coral reef fishes which forage in the water column.
Helgolander wiss. Meeresunters. 24:292-306.

597

FISHERY BULLETIN: VOL. 74, NO. 3

Emery, A. R.

1968. Preliminary observations on coral reef plankton.
Limnol. Oceanogr. 13:293-303.

1973. Preliminary comparisons of day and night habits of
freshwater fish in Ontario lakes. J. Fish. Res. Board Can.
30:761-774.

Gerber, R. p., and N. Marshall.

1974. Ingestion of detritus by the lagoon pelagic community
at Eniwetok Atoll. Limnol. Oceanogr. 19:815-824.

GOSLINE, W. A.

1971. Functional morphology and classification of teleos-
tean fishes. Univ. Press Hawaii, Honolulu, 208 p.
Hamner, W. M., L. p., Madin, A. L. Alldredge, R. W. Gilmer,
AND P. P. Hamner.

1975. Underwater observations of gelatinous zooplankton:
Sampling problems, feeding biology, and behavior. Lim-
nol. Oceanogr. 20:907-917.

HOBSON, E. S.

1965. Diurnal-nocturnal activity of some inshore fishes in

the Gulf of California. Copeia 1965:291-302.
1968. Predatory behavior of some shore fishes in the Gulf of
California. U.S. Fish Wildl. Serv., Res. Rep.73, 92 p.

1971. Cleaning symbiosis among California inshore fishes.
Fish. Bull., U.S. 69:491-523.

1972. Activity of Hawaiian reef fishes during the evening
and morning transitions between daylight and darkness.
Fish. Bull., U.S. 70:715-740.

1973. Diel feeding migrations in tropical reef fishes. Hel-
golander wiss. Meeresunters. 24:361-370.

1974. Feeding relationships of teleostean fishes on coral
reefs in Kona, Hawaii. Fish. Bull., U.S. 72:915-1031.

HUBBS, C. L., AND L. C. HUBBS.

1954. Data on the life history, variation, ecology, and
relationships of the kelp perch, Brachyistius frenatus, an
embiotocid fish of the Californias. Calif. Fish Game
40:183-198.

Limbaugh, C.

1955. Fish life in the kelp beds and the effects of kelp

harvesting. Univ. Calif. Inst. Mar. Resour., IMR Ref. 55-9,
158 p.
Miller, D. J., and R. N. Lea.

1972. Guide to the coastal marine fishes of California. Calif.
Dep. Fish Game, Fish Bull. 157, 235 p.
Newell, G. E., and R. C. Newell.

1963. Marine plankton. Hutchinson, Lond., 221 p.
Quasi, J. C.

1968. Observations on the food of the kelp-bed fishes. In W.
J. North and C. L. Hubbs (editors), Utilization of kelp-bed
resources in southern California, p. 109-142. Calif. Dep.
Fish Game, Fish Bull. 139.
Randall, J. E.

1967. Food habits of reef fishes of the West Indies. Stud.
Trop. Oceanogr. (Miami) 5:665-847.

SCHAEFFER, B., AND D. E. ROSEN.

1961. Major adaptive levels in the evolution of the actinop-
terygian feeding mechanism. Am. Zool. 1:187-204.
Smith, R. I., and J. T. Carlton (editors).

1975. Light's manual: Intertidal invertebrates of the central
California coast. 3rd ed. Univ. Calif. Press, Berkeley,
716 p.
Stevenson, R. A., Jr.

1972. Regulation of feeding behavior of the bicolor dam-
selfish {Eupomacentrxs: partitus Poey) by environmental
factors. In H. E. Winn and B. L. 011a (editors). Behavior of
marine animals. Vol. 2: Vertebrates, p. 278-302. Plenum
Press, N.Y.
Swerdloff, S. N.

1970. The comparative biology of two Hawaiian species of
the damselfish genus Chromis (Pomacentridae). Ph.D.
Thesis, Univ. Hawaii, 202 p.

WiCKSTEAD, J. H.

1965. An introduction to the study of tropical plankton.

Hutchinson and Co., Lond., 160 p.
Williams, A. B., and K. H. Bynum.

1972. A ten-year study of meroplankton in North Carolina

estuaries: Amphipods. Chesapeake Sci. 13:175-192.

598

PROTEIN TAXONOMY OF THE GULF OF MEXICO AND
ATLANTIC OCEAN SEATROUTS, GENUS CYNOSCION

Michael P. Weinstein' and Ralph W. Yerger-

ABSTRACT

Taxonomic relationships among the western North Atlantic seatrouts, genus Cynoscion, were
investigated utilizing acrylamide gel electrophoresis. Several tissues (blood serum, eye lens, and
muscle) were incorporated in this study to gain a better taxonomic overview than would be attainable
with a single protein system.

Blood serum exhibited considerable variation in banding patterns. Because direct interspecific
comparisons were not possible, a phenetic analysis was employed. Eye lens and muscle patterns,
however, were directly comparable. Based on the overall results, three taxonomic conclusions may be
drawn. First, with the exception of a single taxonomic distance (di^) value calculated in the phenetic
analysis, the relationships established by electrophoresis reflect the phyletic relationships proposed by
Ginsburg. This "aberrant" value is believed to result from the small sample size and the possibility of
ecological convergence. Second, the data indicate that Cynoscion nehulosus is the most divergent of the
four forms, supporting previous morphological and ecological conclusions. Third, as suggested by
previous studies, the taxonomic status of C. arenarius as a distinct species is again questioned.
Electrophoretic patterns indicate that it should be regarded as a subspecies of C. regaiis.

Investigation of general protein systems has often
proven useful in elucidating taxonomic relation-
ships. Species-specific banding patterns have been
reported for numerous taxa including fishes
(Tsuyuki and Roberts 1965; Perrier et al. 1973).
Nyman and Westin (1969) studied serum patterns
of cottid fishes from the Baltic Sea and concluded
that the patterns reflected the commonly accepted
scheme. Species and group (genus, family, class)
specificities have also been described for eye lens
proteins of several fishes (Rabaey 1964, Bon et al.
1964, Cobb et al. 1968). Recently Eckroat (1974)
compared members of the pike family (Esocidae)
using this tissue. Myogens have proven par-
ticularly useful in reviews of several groups in the
families Catostomidae (Tsuyuki, Roberts, and
Vanstone 1965; Tsuyuki et al. 1967; Huntsman
1970), Salmonidae (Tsuyuki, Roberts, Vanstone,
and Markert 1965; Tsuyuki et al. 1966) and Scor-
paenidae (Tsuyuki et al. 1968; Johnson et al. 1972).
In this study we have investigated taxonomic
affinities among the western North Atantic sea-
trouts, genus Cynoscion. Four species are current-
ly recognized: spotted seatrout, C. nebulosus
(Cuvier); weakfish, C. regaiis (Bloch and

'Lawler, Matusky & Skelly Engineers, 415 Route 303, Tappan,
NY 10983.

^Department of Biological Science, Florida State University,
Tallahassee, FL 32306.

Manuscript accepted March 1976.

FISHERY BULLETIN: VOL. 74, NO. 3, 1976,

Schneider); silver seatrout, C. nothus (Holbrook);
and sand seatrout, C. arenarius Ginsburg. Cynos-
cion arenarius is restricted to the Gulf of Mexico;
specimens have been captured from Campeche,
Mexico, eastward to the southwest coast of
Florida. Cynoscion regaiis has been generally
considered to be limited to the Atlantic coast.
Guest and Gunter (1958) described its southern-
most occurrence as the St. Lucie estuary, Fla. We
now have evidence which conclusively proves its
presence in the Gulf of Mexico.

Cynoscion nehulosus occurs from New York to
Mexico (Bay of Campeche); its center of abun-
dance is in Florida and the Gulf States (Pearson
1929). Cynoscion nothus is found from Chesapeake
Bay, Md., to the Bay of Campeche but is uncom-
mon at the southern extremity of its range. It is
relatively abundant on the gulf coast and from the
east coast of Florida to North Carolina.

Several tissues (blood serum, eye lens, and
muscle) were utilized in order to achieve a better
taxonomic overview. Since it is difficult (if not
impossible) to construct a phylogeny solely on the
basis of biochemical differences, our results have
been compared with the existing phylogenetic
schemes of Ginsburg (1929) who recognized C.
arenarius and C. regaiis as cognate species, and
Mohsin (1973) who placed C. arenarius and C.
nehulosus in one phyletic line, and C. regaiis and C
nothus in another.

599

FISHERY BULLETIN: VOL. 74, NO. 3

MATERIALS AND METHODS

Spotted seatrout were obtained by hook and line
at seven localities from Corpus Christi, Tex., to
Indian River, Fla. Weakfish were caught by hook
and line in Peconic Bay, N.Y., and together with
silver seatrout in otter trawls in Wassaw Sound,
Ga. Sand seatrout were collected by hook and line
at Pensacola, Fla., and by shrimp trawl in the
vicinity of Carrabelle, Fla.

Preparation of serum and eye lens samples and
electrophoretic methods are identical to those
recently described by Weinstein and Yerger (in
press). Samples were electrophoresed in 7%
acrylamide gel, using a modified Davis (1964)
technique. Diluting tissue preparations with 10%
glycerol avoided the tedius requirement of
producing three-layered gels, yet allowed highly
satisfactory resolution.

Soluble muscle proteins were prepared by
homogenizing 1-g tissue samples with 2 volumes of
ice cold 0.05 M phosphate buffer (pH 7.4).
Homogenates were centrifuged in a SorvalF RC-2
refrigerated centrifuge at 20,000 rpm for 20 min.
Fifty microliters of supernate were combined with
an equal volume of 10% glycerol, and 50 jul of the
mixture layered on each gel. During electrophore-
sis the dye band was allowed to migrate to within
0.5 cm of the end of each gel.

RESULTS

4«^

■ <Mi*SS«

B G

Figure 1. -Serum protein electropherograms derived from whole
sera of four seatrouts. (A) Cynoscion nothus, (B) C. arenarius,
(C) C. regaliif, (D) C. nebulosus.

tion study on C. nebulosus (see Weinstein 1975).

Because of the widespread variation observed in
the blood serum patterns, direct comparison
between the species investigated was difficult. In
order to "sum" the intraspecific variation observed
and subsequently to use the composites for direct
comparison, the taxonomic distance (dji,) measure
of Sokal (1961) was utilized. In this formula

di2 =

E i^n-^i2f

i=i

Serum Proteins

Although serum protein patterns varied intra-
specifically both in the frequency of occurrence of
particular bands, and occasionally in their compo-
sition (intensity), species specificity was evident.
Differences among patterns were not so pro-
nounced as to prevent assigning a given pattern to
the proper taxon. Typical results obtained from
the four seatrouts are shown in Figure 1, and are
diagrammed in Figure 2. All bands observed in the
total number of electrophoresed samples are in-
dicated. Their position on the diagrams is also an
accurate representation of the relative distances
(on the gels) that each band migrated.

We follow the standard method of defining
protein zones (a, (i, y, albumin, prealbumin). The
various designations were derived from a popula-

♦ y

:/?

alb

p alb

alb
p alb

*alb
*

palb

ijjliljjji

--i

alb

p alb

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Figure 2.-Diagrammatic representation of protein bands
occurring in serum electropherograms of four seatrouts. Protein
zones are as follows: y = immunoglobin zone; /? = /3-globulin
zone; « = a-globulin zone; alb = albumins; p-alb = prealbumins.
* indicates band present in 100% of samples. (A) Cynoscion
iiotkus, (B) C. arenarius, (C) C. regalis, (D) C. nebulosus.

600

WEINSTEIN and YERGER: PROTEIN TAXONOMY OF CYNOSCION

all of the bands observed on the gels were taken as
"characters" and their percent occurrence as
"character states." The data utilized in computing
taxonomic distances are summarized in Table 1,
and the results of such an analysis in Table 2. All
data were "standardized" to have a mean of 0 and
a variance of 1 (indicated by £)• Values of d are
interpreted as follows, "The larger the distance,
the smaller the degrees of association or correla-

Table 1. -Percent occurrence of banding patterns derived from
whole serum samples of seatrouts (Cynosciou). A dash indicates
the absence of that band.

Serum C. nothus C. arenarius C. regalis C. nebulosus

band n = 34 n = 19 n = 19 n = 500

81.1

83.3

100

74.5

70.3

75.0

7.7

—

81.1

91.7

76.9

96.5

—

100

30.8

—

100

50.0

84.6

100

8.1

—

54.1

—

100

—

67.6

75.0

84.6

30.0

10.8

—

92.4

—

23.9

86.5

50.0

92.3

18.4

—

88.9

78.4

100

53.8

94.3

—

8.3

76.9

11.2

—

67.3

78.4

50.0

15.4

70.6

59.5

50.0

—

76.0

45.9

—

66.4

—

63.9

43.2

—

92.3

46.1

24.3

100

—

24.3

—

89.1

100

—

15.4

—

84.6

—

91.7

69.2

100

—

38.5

89.8

48.6

16.7

100

20.0

—

91.7

15.4

90.4

94.6

66.7

100

—

8.4

91.7

—

100

75.7

91.7

100

—

24.3

16.7

—

100

91.9

50.0

92.3

13.8

—

77.6

78.4

66.3

23.0

24.8

—

2.1

73.0

91.7

—

80.5

21.6

—

100

—

Table 2.-Taxonomic distances (d,^) calculated for interspecific
comparisons among four seatrouts (Cynoscion). The larger the
value, the smaller the degree of association or correlation
between taxa (Sokal 1961).

Species compared dj

C. nothus versus C. arenarius
C. nothus versus C. regalis
C. nothus versus C. nebulosus
C. arenarius versus C. regalis
C. arenarius versus C. nebulosus
C. regalis versus C nebulosus

63.8
44.0
78.5
54.7
67.3
72.9

tion between taxa!' (Sokal 1961). Based on the
calculated distances (Table 2), C. nebulosus has
diverged to a larger extent than any other
member of the genus and differs from the others
by about the same order of magnitude. Cynoscion
nothus and C. regalis apparently share a closer
relationship in blood protein patterns than do C.
regalis and C. arenarius, a result comparable to
that based on osteological similarity (Mohsin
1973). It also seems apparent that the close cor-
relation between C. yiothus and C. regalis (44.0)
and between C. arenarius and C. regalis (54.7)
should imply a similar distance between C. nothus
and C. arenarius. Such is not the case; and the
value 63.8 apparently indicates that their differ-
ences are even greater than their similarities.

Eye Lens Proteins

Eye lens preparations exhibited considerable
uniformity of pattern (Table 3, Figure 3). Four
bands designated by arable numerals were shared
in common by the seatrouts; however, the amount
of protein in each band differed significantly. For
example, C. regalis had a greater protein concen-
tration in band 1 than did any of the others. The
quantity of protein in this band was not observed
to differ significantly in any of the samples
processed. Bands 1 and A in C. nebulosus {n = 275)

Table 3.-Percent occurrence of banding patterns derived from
eye lens nuclei of seatrouts {Cynagcion). A dash indicates the
absence of that band.

Band

C. nothus C. arenarius C. regalis C. nebulosus
n = 35 n = 12 n = 16 n ^ 275

100

—

12.7

100

—

14.1

33.0

—

100

17.0

42.0

—

59.7

—

19.0

1
2

4 4

■■

3 3

^_g

■■i

■"

+ A B C D

Figure 3.-Diagrammatic representation of protein bands
occurring in eye lenses of four seatrouts. Arabic numerals
indicate bands shared by all taxa (similar electrophoretic
mobility). Letters indicate bands that are either unique or not
shared by all members of the genus. (A) Cynoscion nothus, (B) C.
arenarius, (C) C. regalis, (D) C. nebulosus.

601

FISHERY BULLETIN: VOL. 74, NO. 3

together contained approximately the same
quantity of protein as found in band 1 of C.
arenarius. We believe that band 1 (100% occur-
rence) in C. nebulosus contains at least one protein
which exhibits polymorphism. Since other proteins
(frequency 100%) in this band mask the identity of
the protein in question, it is not possible at the
level of sensitivity of this system to distinguish
the mode of inheritance for this polymorphism.
The same situation seems to be true of bands 3 and
B in this species.

Band 1 is consistently found in lower concen-
tration in C. nothus {n = 35), but the reverse is
true of band 2, which exhibits continuously greater
concentration than the comparable band in any
other seatrout. Band 3 is found in the highest
concentration in C. arenarius (n = 12); a slightly
lower concentration occurs in the composite of
bands 3 and B in C. nebulosus, and a still lower
concentration is found in C. regalis and C. nothus.
Band 4 is present in approximately the same
concentration in all four species. It should be
emphasized that these are average values; small
intraspecific differences were noted from sample
to sample.

Qualitative pattern differences were also noted.
Bands A and B are unique to C. neb2ilosus. A third
band, designated C, was found in 2 of 12 samples of
C. arenarius, but not in any other species. A
fourth variant, designated D, was found in C.
nebulosus, C. arenarius, and C. nothus, but not in
C. regalis. Lastly, a band migrating farthest
anodally in C. nebulosus was designated E. These
qualitative as well as quantitative differences in
eye lens patterns are summarized in Table 3.

Myogens

Electropherograms derived from soluble muscle
proteins provided the most clearly discernible
measure of biochemical relationship. Compared
with serum patterns, only minor intraspecific
variations were evident. A typical grouping from
the four seatrouts is shown in Figure 4, and the
patterns are diagrammed in Figure 5. The broken
lines indicate two minor bands that occurred in a
variable manner and in relatively low frequencies;
hence, they were not considered further. All other
bands occurred in 100% of the samples and are
designated as comprising the typical species-
specific patterns. A remarkable degree of similar-
ity in the patterns is obtained for C. regalis and C.
arenarius; they share not only 12 and 13 bands in

^^BB^*

A BCD

Figure 4. -Electropherograms derived from protein extracts of
epaxial musculature. (A) Cynoscion nothus, (B) C. arenarius, (C)
C. regalis, (D) C. nebulosus.

0 ^^^

5 ^-^

B
C

■■

■^

Figure 5.-Diagrammatic representation of the protein bands
occurring in myogen extracts of four seatrouts. (A) Cynoscion
tiothus, (B) C. arenarius, (C) C. regalis, (D) C. nebulosus.

common (as indicated by electrophoretic mobility
and sieving characteristics), but also compare
favorably in the quantities of protein comprising
each band (Figure 6). Although some variation
occurred in relative peak heights from sample to
sample (within a species), the densitometer trac-
ings shown in Figure 6 are representative of each
species. A close relationship clearly exists between
C. regalis and C. arenarius both in the distance of
migration and in the quantity of protein making
up the individual bands (Figures 5, 6). Although
the other two species shared the same general
generic pattern, they varied in the composition of
several major bands. Cynoscion nebulosus always
has a high concentration in band 2 in each of the 12
samples processed, and has a second band im-
mediately adjacent of the same thickness (denoted
J on Figure 5). These bands (2 and J) were not
resolved as separate peaks in a series of densi-

602

WEINSTEIN and YERGER: PROTEIN TAXONOMY OF CYNOSCION

C- nolhus

Figure 6.- Densitometer tracings of representative myogen
patterns of four seatrouts (Cynoscion). Intensity of particular
bands are indicated by relative peak heights.

tometer tracings; however, observation of gels and
photographs clearly indicated their double nature.
Band 1 was not present in any of the samples of C.
nebulosus; however, a band designated as H oc-
curred in a more cathodal direction (above position

I).

Bands D and E in C. nothus are slightly dis-
placed; i.e., they have a slightly different electro-
phoretic mobility from their "counterparts" (F
and G) in the other three species. This difference
could be an artifact, but duplicate experiments
indicate otherwise. Cynoscion nothus also lacks
bands B and C found in the other species. Band A
is absent in C. nebulosus, but present in the other
seatrouts.

DISCUSSION

Morphological Taxonomy

In his review of the seatrouts of the Atlantic and
Gulf coasts of the United States, Ginsburg (1929)
recognized C. nebulosus as the most distinctive
morphologically on the basis of its color pattern
and its scaleless dorsal and anal fins. Cynoscion
nebulosus also differs ecologically from the other
Cynoscion; it is primarily an estuarine form while
the others have a closer affinity to the marine
environment.

The remaining species are less easily distin-
guished. Of the many criteria used, size and color
are most important. Cynoscion regalis is readily
recognized in the adult stage by the longitudinal
rows of small spots on its back, which produce a
mottled appearance. The paler, C. arenarius of the
gulf lacks conspicuous pigmentation. Cynoscion
nothus is similar in color to C. arenarius, but
differs in several other respects including verte-
bral and anal-fin ray counts. Cynoscion nothus
may not attain as large a maximum size as C.
arenarius, although this observation may be a
sampling artifact. Gunter (1945) noted that C.
nothus occurs at slightly greater depths than the
other seatrouts. Therefore, the main populations
of C. nothus may not have been adequately
sampled.

Taxonomically, the status of C. arenarius has
never been satisfactorily resolved. Guest and
Gunter (1958) accorded full species rank for C.
arenarius, as does the current list of the American
Fisheries Society (Bailey et al. 1970), and the
recent investigation by Mohsin (1973). However,
the original description leaves room for consider-
able doubt. Ginsburg (1929) stated in a footnote
that, "An unbiased study of the data here pre-
sented shows, I believe, that there is room for
difference of opinion as to the degree of difference
between this form [C. arenarius] and ragalis
[regalis] from the Atlantic coast-whether they
should be regarded as species or subspecies."
Furthermore, by Ginsburg's (1938) own criteria of
the "arithmetical" definition of a species, the 18%
intergradation of the most "divergent" character
(the number of articulated dorsal rays) would give
the two forms only subspecific status.

Protein Taxonomy

Our primary purpose in this study has been to

603

FISHERY BULLETIN: VOL. 74, NO. 3

provide biochemical evidence for the taxonomic
relationship among four members of the genus
Cynoscion (including the degree of divergence),
and to compare this information with existing
phylogenetic schemes. Although no attempt has
been made to construct a phylogeny based on
biochemical data, qualitative differences (and
similarities) allow some taxonomic conclusions to
be drawn.

Serum Proteins

Environmentally induced changes in blood ser-
um components have been well substantiated
(Thurston 1967). This evidence, nonetheless, would
not preclude blood serum patterns from being a
useful taxonomic tool if one additional step is
taken. It is obvious that the classical mor-
phologists in comparing populations of animals (or
plants) are including the influence of the environ-
ment in the range of variation they are describing.
For example, it is commonly observed that counts
of meristic characters (fin rays, scales, etc.) in-
crease in the northerly direction of the animal's
range (in the Northern Hemisphere). This, how-
ever, will not affect the conclusions drawn as long
as sufficient samples are taken to cover the full (or
nearly so) range of variation in the population.
Once adequate samples are obtained, accurate
modes may be calculated for each character and
the relationship between two forms established.
Within this framework utilization of highly var-
iable patterns such as that found for serum pro-
teins are justified.

In this study we have been able to sample only a
relatively small number of each species, with the
exception of C. nebulosus (Table 2). Hence, any
conclusions regarding the biochemical relation-
ship among the four taxa must be provisional.

Although the blood patterns of the species of
Cynoscion are somewhat more variable than has
been reported for many fishes and other verte-
brates, we can present evidence for relationships
among the Gulf of Mexico and Atlantic Ocean
seatrouts. The Taxonomic distances calculated for
members of this genus are listed in Table 2. The
value (54.7) for the alleged cognates, C. arenarius
and C. regalis, is surpassed only by the value (43.9)
for C. nothus and C. regalis. Only 10 bands of the
41 present were unique to one of the four species; 7
were found in C. nebulosus, 2 in C. nothus, and 1 in
C. regalis. Cynoscion arenarius did not display

any species-specific bands. Therefore, a consider-
able portion of the differences among the four
seatrouts, as expressed by 4^, are generated by
different percentage compositions of the serum
proteins.

The similar values obtained for C. regalis and C.
nothus may be interpreted in three ways: 1) these
species may actually be more closely related than
are C. regalis and C. arenarius; 2) similar envi-
ronmental selection pressures have produced an
example of ecological convergence; 3) sample size
may be insuflRcient to yield accurate results. Three
of the 19 samples of C. regalis were taken from the
same estuary (Wassaw Sound) as were all samples
of C nothus; the remaining sera from C. regalis
were collected in an estuary (Peconic Bay) sharing
several physical and chemical parameters with
Wassaw Sound (Odum et al. 1974). Thus, a measure
of ecological convergence may be involved. Similar
reasoning might explain the d,f, calculated for C.
arenarius versus C. nothus; the value (63.8) might
be reduced if several other gulf populations of C.
nothus were added to the total sample.

It could be argued that the much larger sample
of C. nebulosus (n = 500) was responsible for most
of the difference in the taxonomic distance value
since rare bands are being included. This could
only be the case for band 38 which occurred in only
2.1% of the specimens sampled. The values (per-
cent occurrence) of the remaining six unique
bands (8%, 23%, 66%, 67%, 89%, 89%) argue against
this possibility. The average value of 72.8 is
therefore taken to mean that C. nebulosus is the
most divergent of the four species investigated.
Possible reasons for this observation have been
elaborated previously.

A significant observation in our study is that
relatively few species-specific (i.e., unique) pro-
teins have been detected, a phenomenon not
without precedence, however (Lewontin 1974). In
a study of 10 species of Drosophila, the number of
unique proteins ranged from 2.6 to 28.2%, with an
average of 14.3% (Hubby and Throckmorton 1968).
Our own figures compare favorably with these: C
nebulosus, 23%; C nothus, 7%; C. regalis, 5%; and C.
arenarius, 0%.

Eye Lens Proteins

In a review of intraspecific variation in lens
proteins. Day and Clayton (1973) detected no
polymorphisms and concluded that observed dif f er-

604

WEINSTEIN and YERGER: PROTEIN TAXONOMY OF CYNOSCION

ences were almost wholly quantitative rather
than qualitative. Data from other studies indicate
two further conclusions. First, lens proteins on the
whole express a high degree of conservatism.
Secondly, in cases where evidence of polymor-
phisms have been obtained, fishes have been most
often implicated. Smith and Goldstein (1967),
Smith (1969, 1971), and Smith and Clemens (1973)
reported intraspecific variations in the lens pat-
terns of numerous species. Barrett and Williams
(1967) detected a polymorphism in the lens pro-
teins of the bonito Sarda cluliensis. Eckroat and
Wright (1969) and Eckroat (1973) provided direct
evidence of polymorphisms in the eye lens of the
brook trout, Salvelinus fontinalis, and demon-
strated simple Mendelian inheritance for several
characters.

Previous observations for eye lens proteins and
the conclusions stated above are reflected in our
work on the patterns derived from the genus
Cijnoscion. Lens protein patterns displayed con-
siderable convervatism among the four seatrouts.
Four bands from a total of eight occur in all taxa
and are probably high molecular weight a- and
/8-crystallins. Only a single band (E in C. nebulo-
sus, Figure 3) is unique and is found in either very
low frequency or not at all in four of the seven
estuaries sampled. Its relatively high frequencies
in Corpus Christi, Galveston, and Florida Bay (36,
39, and 50%, respectively) indicate a possible
relationship to high turbidity and low light inten-
sities characteristic of these three areas.

Although intensity patterns did not vary sig-
nificantly within a species (with the exception of
two bands involved in a suspected polymorphism
in C. nebulosus), the quantities of protein in bands
with the same mobility were quite different and
species-specific (Figure 3). The selective forces
which control the quantity of protein present in a
given band are not easily recognized. The geogra-
phic ranges of these four species overlap con-
siderably althouth their centers of abundance are
quite different. Cijnoscion nothus is found farther
offshore than its congeners; C. nebulosus is
primarily restricted to the estuarine habitat. All
seatrouts probably experience a similar range of
water color and turbidities in their respective
habitats. None is considered to be more diurnal or
nocturnal than the others. Their temperature
ranges overlap considerably. Therefore, it is
somewhat puzzling as to the cause of the common
observation that variations in patterns both with-

in a species and between them is restricted mainly
to intensity differences. Presently the advantages
of difTerent proportions of crystallins and other
eye lens protein in a particular species are poorly
known.

Myogen Proteins

The general application of myogen proteins to
systematic studies has been reviewed by Tsuyuki
(1974). Perhaps no other tissue investigated has
displayed such an overall lack of intraspecific
variations. Only a few species of fishes have
exhibited detectable polymorphisms (e.g., Nyman
1967; Tsuyuki et al. 1968; Gray and McKenzie
1970), and it is noteworthy that most of these are
"tetraploid" species. The majority of investiga-
tions on other forms reveal virtually no intra-
specific variation, an observation in direct contrast
with other protein systems which generally dis-
play polymorphisms. Various estimates of propor-
tions of polymorphic alleles in vertebrate species
are placed at from 10 to 20% (Selander and Kauf-
man 1973). The constancy maintained in myogen
proteins in the presence of selective forces is
indeed remarkable.

The general conservatism displayed in myogen
patterns was observed in our own work, but with
several important differences. As previously de-
scribed, C. nothus and C. nebulosus differed in the
presence or absence of one or more major (by
staining intensity) bands. Band J (Figure 5),
unique in C. nebulosus, is found in all samples, and
produces a large characteristic peak on densi-
tometer tracings. The absence of several bands,
notably B and C (Figure 5), characterizes C.
nothus.

On the basis of myogen patterns, we suggest
that C. arenarius and C. regalis are more closely
related to each other than are any other combina-
tion of species under consideration and should be
treated as conspecific. Thus, we reject the
phylogeny based on slight osteological differences
proposed by Mohsin (1973). The gulf form (C. are-
narius) should be recognized as a subspecies of C.
regalis, a conclusion strengthened by recent
confirmation of specimens of C. regalis from the
Gulf of Mexico.

Earlier reports of C. regalis in the gulf generally
lacked documentation, or were misidentifications
of C. arenarius. The report of Jordan and Eigen-
mann (1889) from Mobile Bay, Ala., was based on
specimens of C. arenarius, a form not recognized

605

until 45 yr later. Rivas (1954) mentioned the
weakfish in the gulf but provided no specific data.
Hutton et al. (1956) reported C. regalis from Boca
Ciega Bay at St. Petersburg, Fla., but Springer
and Woodburn (1960) listed only C. arenarius
from Tampa Bay. No specimens of C. regalis from
the gulf are in the reference fish collection of the
Department of Natural Resources at the St. Pe-
tersburg Marine Laboratory (Moe et al. 1966).

Two adult C. regalis (266 and 298 mm standard
length) were captured by personnel from the
Marco Ecology Laboratory in the vicinity of Marco
Island, on the southwest coast of Florida on 21 July
1972 (Florida State University Fish Collection,
catalog number 24023). The documentation of the
weakfish in the Gulf of Mexico together with the
extremely close morphological and biochemical
characteristics shared by C. regalis and C. are-
narius suggest that gene exchange between the
Atlantic Ocean and gulf coast populations is
feasible although we have no proof of their inter-
breeding. Nevertheless, the evidence points to the
same series of events which characterize the
evolutionary history of other marine geminate
species in Florida. When the peninsula split the
ancestral population into two, the Gulf population
differentiated from that in the Atlantic (see
Ginsburg 1952; Walters and Robins 1961).
Whether or not isolation was complete or only
partial, the present distribution indicates that at
least one form (C. regalis) has been successful
in moving around the tip of the peninsula into
southeastern gulf waters and in establishing
secondary contact with the other (C. arenarius).
The status of C. arenarius should be investigated
in depth. Perhaps an extensive enzyme study
would be appropriate, the results of which could be
compared by statistical analyses (Avise 1974) to
determine the level of differentiation between two
forms.

ACKNOWLEDGMENTS

The authors extend their thanks to the many
individuals who aided in collection of seatrout,
particularly Robert Stickney of the Skidaway
Institute, Skidaway Island, Ga., and Rufus Messer
of Carrabelle, Fla. Our work was supported in part
by the Florida Department of Natural Resources;
we thank Edwin Joyce for his support and
encouragement.

FISHERY BULLETIN; VOL. 74, NO. 3

LITERATURE CITED

Avise, J. C.

1974. Systematic value of electrophoretic data. Syst. Zool.
23:465-481.
Bailey, R. M., J. E. Fitch, E. S. Herald, E. A. Lachner, C. C.
LiNDSEY, C. R. Robins, and W. B. Scott.
1970. A list of common and scientific names of fishes from
the United States and Canada. Am. Fish. Soc, Spec. Publ.
6, 150 p.
Barrett, I., and A. A. Williams.

1967. Soluble lens proteins of some scombroid fishes. Copeia
1967: 468-471.

Bon, W. F., P. L. Swanborn, G. J. C. M. Ruttenberg, and

A. DOHRN.

1964. Comparative investigation of the soluble proteins of
the eyelens of fish. Publ. Stn. Zool. Napoli 34:59-65.
Cobb, B. F., Ill, L. Carter, and V. L. Koenig.

1968. The distribution of the soluble protein components in
the crystalline lenses of fishes. Comp. Biochem. Physiol.
24:817-826.

Davis, B. J.

1964. Disc electrophoresis-II. Method and application to
human serum proteins. Ann. N.Y. Acad. Sci. 121:404-427.
Day, T. H., and R. M. Clayton.

1973. Intraspecific variation in the lens proteins. Biochem.
Genet. 8:187-203.
Eckroat, L. R.

1973. Allele frequency analysis of five soluble protein loci in
brook trout, Salvclinus fontinalis (Mitchill). Trans. Am.
Fish. Soc. 102:335-340.

1974. Interspecific comparisons of lens proteins of
Esocidae. Copeia 1974: 977-978.

Eckroat, L. R., and J. E. Wright.

1969. Genetic analysis of soluble lens protein polymorphism
in brook trout, Salvelinus fontinalis. Copeia 1969:466-473.

Ginsburg, I.

1929. Review of the weakfishes (Cynoscion) of the Atlantic
and Gulf coasts of the United States, w^ith a description of
a new species. Bull. U.S. Bur. Fish. 45:71-85.

1938. Arithmetical definition of the species, subspecies and
race concept, with a proposal for a modified nomenclature.
Zoologica (N.Y.) 23:253-286.

1952. Eight new fishes from the Gulf coast of the United
States, with two new genera and notes on geographic
distribution. J. Wash. Acad. Sci. 42:84-101.
Gray, R.W., and J. A. McKenzie.

1970. Muscle protein electrophoresis in the genus Salmo of
eastern Canada. J. Fish. Res. Board Can. 27:2109-2112.

Guest, W. C, and G. Gunter.

1958. The sea trout or weakfishes (genus Cynoscion) of the
Gulf of Me.xico. Gulf States Mar. Fish. Comm., Tech.
Summ. 1, 40 p.
Gunter, G.

1945. Studies on marine fishes of Texas. Publ. Inst. Mar.
Sci. Univ. Tex. 1:1-190
Hubby, J. L., and L. H. Throckmorton.

1968. Protein differences in Drosophila. IV. A study of
sibling species. Am. Nat. 102:193-205.
Huntsman, G. R.

1970. Disc gel electrophoresis of blood sera and muscle
extracts from some catostomid fishes. Copeia 1970:
457-467.

606

WEINSTEIN and YERGER: PROTEIN TAXONOMY OF CYNOSCION

HuTTON, R. F., B. Eldred, K. D. Woodburn, and R. M. Ingle.
1956. The ecology of Boca Ciega Bay with special reference

to dredging and filling operations. Part I. Fla. State Board

Conserv., Tech. Ser. 17, 87 p.
Johnson, A. G., F. M. Utter, and H. 0. Hodgins.

1972. Electrophoretic investigation of the family Scorpaen-
idae. Fish. Bull., U.S. 70:403-413.

Jordan, D. S., and C. H. Eigenmann.

1889. A review of the Sciaenidae of America and Europe.
Rep. Comm., U.S. Comm. Fish Fish. 1886: 343-451.
Lewontin, R. C.

1974. The genetic basis of evolutionary change. Columbia
Univ. Press, N.Y., 346 p.
MoE, M. A., Jr., P. C. Heemstra, J. E. Tyler, and H. Wahlquist.

1966. An annotated listing of the fish reference collection at
the Florida Board of Conservation Marine Laboratory.
Fla. State Board Conserv. Mar. Lab., Spec. Sci. Rep. 10, 121

P-
MoHSiN, A. K. M.

1973. Comparative osteology of the weakfishes (Cynoscion)
of the Atlantic and Gulf coasts of the United States
(Pisces-Sciaenidae). Ph.D. Thesis, Te.xas A&M Univ., 147

P-
Nyman, C

1967. Protein variations in Salmonidae. Rep. Inst. Fresh-
water Res. Drottningholm 47:5-38.

Nyman, L., and L. Westin.

1969. Blood protein systematics of Cottidae in the Baltic
drainage area. Rep. Inst. Freshwater Res. Drottning-
holm 49:164-174.

Odum, H. T., B. J. CoPELAND, AND E. A. McMahan (editors).

1974. Coastal ecological systems of the United States. Vol.
I. Conservation Foundation, Wash.,D. C, 533 p.

Pearson, J. C.

1929. Natural history and conservation of redfish and other
commercial sciaenids on the Texas coast. Bull. U.S. Bur.
Fish. 44:129-214.
Perrier, H., J. P. Delcroix, C. Perrier, and J. Gras.

1973. An attempt to classify the plasma proteins of the
rainbow trout (Salmo gairdnerii Richardson) using disc
electrophoresis, gel filtration and salt solubility fractiona-
tion. Comp. Biochem. Physiol. 46B: 475-482.
Rabaey, M.

1964. Comparative study of tissue proteins (lens and mus-
cle) in fish. Protides Biol. Fluids, Proc. Colloq. 12:273-277.
RiVAS, L. R.

1954. The origin, relationships, and geographical distribu-
tion of the marine fishes of the Gulf of Mexico. In P. S.
Galtsoff (coordinator). Gulf of Mexico, its origin, waters,
and marine life, p. 503-505. U.S. Fish Wildl. Ser\-., Fish.
Bull. 55.
Selander, R. K., and D. W. Kaufman.

1973. Genie variability and strategies of adaptation in
animals. Proc. Natl. Acad. Sci. 70:1875-1877.
Smith, A. C.

1969. Protein variation in the eye lens nucleus of the
mackeral scad {Decapteriis piniiidatus). Comp. Biochem.
Physiol. 28:1161-1168.
1971. Genetic and evolutionary analysis of protein variation
in eye lens nuclei of rainbow trout (Salmo gairdnerii). Int.
J. Biochem. 2:384-388.
Smith, A. C, and H. B. Clemens.

1973. A population study by proteins from the nucleus of

bluefin tuna eye lens. Trans. Am. Fish. Soc. 102:578-583.
Smith, A. C, and R. A. Goldstein.

1967. Variation in protein composition of the eye lens
nucleus in ocean whitefish, Caulolatilus princeps. Comp.
Biochem. Physiol. 23:533-539.
SOKAL, R. R.

1961. Distance as a measure of taxonomic similarity. Syst.
Zool. 10:70-79.
Springer, V. G., and K. D. Woodburn.

1960. An ecological study of the fishes of the Tampa Bay
area. Fla. State Board Conserv., Mar. Lab., Prof. Pap. Ser.
104 p.

Thurston, R. V.

1967. Electrophoretic patterns of blood serum proteins from

rainbow trout (Salmo gairdneri). J. Fish. Res. Board Can.

24:2169-2188.
TSUYUKI, H.

1974. Muscle proteins of fishes. In M. Florkin and B. T.
Scheer (editors). Chemical zoology. Vol. 8, p. 287-305.
Academic Press, N.Y.

TsuYUKi, H., and E. Roberts.

1965. Zone electrophoretic comparison of muscle myogens

and blood proteins of artificial hybrids of Salmonidae with

their parental species. J. Fish. Res. Board Can. 22:767-773.

TsuYUKi, H., E. Roberts, R. H. Kerr, J. F. Uthe, and L. W.

Clarke.

1967. Comparative electropherograms of the family Cato-
stomidae. J. Fish. Res. Board Can. 24:299-304.

TsuYUKi, H., E. Roberts, R. H. Lowes, W. Hadaway, and S. J.
Westrheim.

1968. Contribution of protein electrophoresis to rockfish
(Scorpaenidae) systematics. J. Fish. Res. Board Can.
25:2477-2501.

TsuYUKi, H., E. Roberts, and W. E. Vanstone.

1965. Comparative zone electropherograms of muscle
myogens and blood hemoglobins of marine and freshwater
vertebrates and their application to biochemical systema-
tics. J. Fish. Res. Board Can. 22:203-213.

TsuYUKi, H., E. Roberts, W. E. Vanstone, and J. R. Markert.

1965. The species specificity and constancy of muscle
myogen and hemoglobin electropherograms of Oncorhyn-
chus. J. Fish. Res. Board Can. 22:215-217.

TsuYUKi, H., J. F. Uthe, E. Roberts, and L. W. Clarke.

1966. Comparative electropherograms of Coregonus
chipeaformis, Salvelinus namaycush, S. alpinus, S. mal-
ma, and S.fontinalis from the family Salmonidae. J. Fish.
Res. Board Can. 23:1599-1606.

Walters, V., and C. R. Robins.

1961. A new toadfish (Batrachoididae) considered to be a
glacial relict in the West Indies. Am. Mus. Novit. 2047, 24

P-
Weinstein, M. p.

1975. Electrophoretic investigation of the Gulf of Mexico
and Atlantic Ocean seatrouts of the genus Cynoscion, with
special reference to the population structure of the spotted
seatrout, Cynoscion nebulosus. Ph.D. Thesis, Florida State
Univ., 115 p.

Weinstein, M. P., and R. W. Yerger.

In press. Electrophoretic investigation of subpopulations of
the spotted seatrout, Cynoscion nebulosus (Cuvier), in the
Gulf of Mexico and Atlantic Coast of Florida. Comp.
Biochem. Physiol.

607

PREHATCH AND POSTHATCH GROWTH OF FISHES-

A GENERAL MODEL

James R. Zweifel and Reuben Lasker^

ABSTRACT

The developmental stages of fish eggs and the growth of larval fishes of several species can be
represented by a Gompertz-type curve based on the observation that in widely different living systems,
exponential growth tends to undergo exponential decay with time. Further, experimental studies and
field observations have shown that the effect of temperature on the growth process follows the same
pattern, i.e., the rate of growth declines exponentially with increasing temperature. Evidence suggests
that prehatch growth rates determine ideal or optimum trajectories which are maintained after hatch
in the middle temperature range but not at either extreme. Also, posthatch growth exhibits a
temperature optimum which is not apparent in the incubation period. These studies have also .shown
that for the same spawn both the prehatch and yolk-sac growth curves reach asymptotic limits
independent of temperature. Other biological events (e.g., jaw development) occur at the same size for
all temperatures.

The growth of post-yolk-sac larvae follows a curve of the same type and hence the posthatch growth
trajectory may be represented by a two-stage curve. For starving larvae, the second stage shows a
decline in size but maintains the same form, i.e., the rate of exponential decline decreases exponentially
with time.

Recent success in spawning and rearing marine
fish larvae at the Southwest Fisheries Center
(SWFC) (Lasker et al. 1970; May 1971; Leong 1971)
has made possible a much more fundamental
examination of the growth process than has here-
tofore been possible. Controlled laboratory exper-
iments can now be utilized to investigate both the
inherent nature of the growth process as well as
the effect of some environmental factors.

Considerable care is required, however, in con-
structing a model- which is meaningful both
mathematically and biologically. For example,
almost all growth models currently in use can be
derived as variations of the differential equation:

= r]W -kW

dL n

dt '

K'L'

(1)

(la)

(von Bertalanffy 1938; Beverton and Holt 1957;
Richards 1959; Chapman 1961; Taylor 1962) where

'Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

^A model is here conceived to be a mathematical representa-
tion of change in length or weight with time under measureable
environmental conditions.

W is weight, L is length, and tj, k, m, n, m', and n'are
arbitrary constants. These are the equations used
most often to describe growth as a function of
anabolic and catabolic processes of metabolism.
The rate of anabolism, tj, is considered to be
proportional to W"' and the rate of catabolism, k,
proportional to W". Equation (la) requires, in
addition, the allometric relationship W = qL" ,
where again q and p are arbitrary constants. In
practice a dilemma arises from the fact that while
such models yield a good empirical fit to the data,
the estimates of parameters r\ and k are often
negative, thereby negating the assumptions on
which the model is based. For n = \ and m = 0, 1,
2, respectively. Equation (1) gives rise to the von
Bertalanffy growth in length, Gompertz, and
logistic growth functions. Although we have not as
yet made any extensive comparisons, the fact that
for m>l and n = I, t] and k must be negative,
suggests that in many instances the Gompertz and
logistic rather than the von Bertalanffy functions
may provide more appropriate models of fish
growth. In particular, the simple von Bertalanffy
growth model has no inflection point and hence
curves such as the generalized von Bertalanffy,
Gompertz, or logistic must be used when an in-
flection in the growth trajectory is evident.

Laird et al. (1965) have presented a Gompertz-
type mathematical model of growth based on the

Manuscript accepted March 1976.

FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

609

FISHERY BULLETIN: VOL. 74, NO. 3

observation that the specific growth rate dW/ Wdt
of animals and their parts tends to decay expo-
nentially with increasing age. They have shown
that this relation offers a practical means of
analyzing the growth of parts of embryonic and
postnatal animals (Laird 1965a), the growth of
tumors (Laird 1964, 1965b), whole embryos of a
number of avian and mammalian species (Laird
1966a), and early stages of postnatal growth of a
variety of mammalian and avian organisms (Laird
1966b). Further, Laird (1966b, 1967) has shown
that postnatal growth of a variety of mammalian
and avian organisms can be fitted by compounding
this model with a linear growth process beginning
at birth and extending on beyond the asymptotic
limit of the underlying Gompertz growth process.
Overall growth is assumed to be genetically de-
termined by programming only the initial specific
growth rate and the rate of exponential decay,
these governing growth processes then act on a
genetically determined original mass to produce
the observed course of growth to a final limiting
size characteristic of the species and individual.

Mathematically, these assumptions are de-
scribed by the two equations:

dWjt)
dt

y{t)Wit)

and^

dm = -ay{t)

which have the solution

W{t) = Woe

-at.

(2)

where W^ is weight at f = 0, Aq is the specific
growth rate at ^ = 0, a is the rate of exponential
decay and the specific growth rate at time t, A, =
Aoe-'".

Laird et al. (1965) indicated that an additional
growth component not included in the Gompertz
equation may be due to the accumulation of
products that are not self-reproducing or to
renewal systems that are not in exact phys-
iological equilibrium and suggested the com-
pound growth curve:

W= W^ + (i

t w,

. M

(3)

where Wq is the mass due to the Gompertz growth
process, /3 is the rate of linear growth, and M is the
asymptotic limit of the growth process. She also
suggests that this linear process starts in the early
embryonic period, if not at conception. For the age
interval covered in this paper, however, the linear
growth component {W - W^) was not found to be
important.

Several characteristics of the curve are worthy
of mention:

The asymptotic limit M is Wq Exp (^ ,)/«)•
The point of inflection {t,, W,) =

[^-

iAo/a),WoExp{^
a

»]

3. The zero point on the time scale may
be shifted to any point t^ without changing the
form of the equation with new parameters W^ =
W(l), A^ = Aoe'"^ where a remains un-
altered.

The fundamental concept of the Laird-
Gompertz model is one of change in weight or
mass with time, being due primarily to the self-
multiplication of cells and genetically determined
limitations on the growth parameters. The use of
length as the measured variable is thus a matter of
convenience due to the fact that weight measure-
ments are much more time consuming, especially
in early larval growth, but also in juvenile and
adult fishes. As indicated in Equation (la), if a true
allometric relationship existed, the choice would be
unimportant. However, all experimental evidence
indicates that both length and weight can be
described by a Gompertz-type curve. Hence, it can
be shown that 1) the growth rate for both changes
continually with time and 2) the form of the
length-weight relationship will change continually
except for two special instances. Laird et al. (1968)
has shown that this occurs only when the rates of
exponential decay are the same and either the two
measured variables begin growth at difi'erent
times at the same initial rate or at difi'erent rates
at the same time. In all other cases the allometric
plot will be nonlinear. For

■'In the usual Gompertz representation the rate of exponential
growth is assumed to decline logarithmically as W approaches

the asymptote M = WqC » , i.e.,

dW ^

amn(M/W).

Kl(^

and

L = Loe

W = WoB^^^^

rPt)

^-at\

610

ZWEIFEL and LASKER: PREHATCH AND POSTHATCH GROWTH OF FISHES

the length-weight relationship is

then

Wit) = Me

-at

i„pf„ + ^i-k:!^)""j

(4) or ln[-ln(PF(0/M)] = In/C-a^,

Only when a = ft does the relationship reduce to
the linear form

\nW= \nWo + ^ln(L/Lo).

As shown in Figure 1, departure from linearity
will not always be great, but for extrapolation the
effect of overestimation at larger sizes may
become serious.

Throughout this paper, growth will, by necessi-
ty, be measured in terms of length rather than
weight even though the model equation is
developed from the opposite point of view. It
should be remembered, however, that no allomet-
ric relationship is assumed, i.e., no relationships
among the two sets of parameters are assumed
except as they are jointly a function of age.

100
80

60
40

E
S

I
I-
o

004 006 01

0 2 04 06 08 I

WEIGHT (mg)

Figure 1. -Length-weight relationship in larval anchovies: Solid
line fitted from log W = a + 6 log L; dashed line fitted from
Equation (4); estimates are coincident up to 10 mm.

INITIAL ESTIMATES

Equation (2) may be rewritten as follows:
Let K = Ao/a

and

M = Woe^,

and hence the logarithm of the logarithm of the
ratio of size to the asymptotic limit M with the
sign changed will be linearly related to time f with
parameters In K and -a. Wq may be obtained
from the relationship In M = In Wq + K. Note: For
decreasing curves, use the reciprocal of the ob-
served values.

VARIABILITY, ESTIMATION, AND
TRANSFORMATION BIAS

It is an unfortunate circumstance that the
determination of the "best" estimation procedure
can rarely be separated from the determination of
the "best" mathematical model, i.e., there is no
recognized best estimation procedure except in
some specialized instances. This is brought about
by the fact that almost all parametric estimation
procedures assume some information concerning
the form and stability of the "error" distribution.
This requires, at the very least, the knowledge that
the variance is constant and, at the most, the exact
form of the error distribution. Since the term
"error" in the biological sciences takes a meaning
quite different from that in the physical and
mathematical sciences in that it represents, in the
main, natural variability rather than measure-
ment or experimental error and since natural
variability is large (especially so in cold-blooded
organisms), few a priori assumptions can be made.

Since most estimation procedures assume a
normal distribution of errors at each point along
the curve with equal variance (homoscedasticity),
the obvious approach, when no more plausible
alternative is available, is to fit the situation to
this mold.

Some general recommendations are helpful.
"Although no clear rule may be safely offered for
the taking of logarithms to reduce data to man-
ageable configurations, nevertheless, this trans-
formation (logs) is probably the most common of
all. Almost all data that arise from growth phe-
nomenon, where the change in a datum is likely to
be proportional to its size and hence errors are
similarly afl^hcted, are improved by transforms to
their logarithms" (Acton 1959: 223). Specifically, it
can be shown that the logarithmic transformation
will induce homoscedasticity in those instances

611

FISHERY BULLETIN: VOL. 74, NO. 3

where the standard deviation is proportional to
the population mean, i.e., a = fi^i or log a = log ft +
log ju. Hence, a plot of log o on log ju will have a slope
of unity and the antilog of the intercept will define
the proportionality constant. Plots of log o on log n
were made for several experiments where data
were available for extended periods of time. None
of the regression coefficients was significantly
different from unity. These experiments cover a
variety of life stages and environmental situa-
tions from controlled laboratory experiments on
larval anchovies (Lasker et al. 1970) to large tank
feeding of anchovies captured from the wild at 75
mm (Paloma, SWFC, unpubl. data) to samples of
adult sardines obtained from bait boats (Lasker
1970). Growth for the 75-mm anchovies was slow
and much more uniform than for the other exper-
iments as indicated by the mean square errors in
Table 1. The analysis of covariance (Table 1) shows
no difference in slope for either length or weight
from larval, juvenile, and adult fishes. The average
slopes are 0.9981 for larvae and adults and 1.1061
for juveniles. With a slope of unity, the propor-
tionality constant can be estimated by Exp (Ina -
Inju). The results from the several experiments are
shown below:

Lasker et al. (1970):

Experiment 1

Experiment 2
Paloma^
Lasker (1970)

Not unexpectedly, variation in weight exceeds
that of length and both decrease with increasing
age.

The question of normality and its relationship to
homoscedasticity is more tenuous, but again some
help is available. In practical work, it is generally
assumed that both ,v and log x can be regarded as
normally distributed as long as the coefficient of
variation C = o/\i< Vs or a,,,,, j. <0.14 (Hald 1952:
164). This allows transformation for one desidera-
tum without noticeably affecting another.

Paloma (see footnote 4) collected one or two
samples per month of laboratory-reared anchovies
for a period of nearly 2 yr. Approximately 25 fish
were taken for each sample. We examined nor-
mality in terms of skewness (Gj) and kurtosis
(mean absolute deviation A). Although sample

Table l.-The relationship of mean and standard deviation for
both length and weight measurements in fishes.

loga=a +j8 log/x

Analysis of covariance

deviations from

regression

s.s.

m.s.

Larvae and adults:

Lengtfi' exp. 1

-1.5568

1.6979

0.3308

0.0551

exp. 2

-0.8003

0.8281

0.7167

0.0896

Weighti exp. 1

-0.4192

1.0373

0.1572

0.0262

exp. 2

-0.4852

1.0077

0.4241

0.0530

Length^

-1.6093

1.0848

2.2933

0.0382

Weight^

-0.4748

0.7906

2.5913

0.0432

Within

148

6.5134

0.0440

Betv^een

0.1425

0.0285

Common

153

6.6559

F = 0.0285/0.0440

= 0.65

Juveniles:

Length'

-1.3975

1.1644

0.3658

0.0118

Weight^

-0.8000

1.1029

0.1511

0.0048

Within

0.5169

0.0083

Between

0.0002

Common

0.5171

F = 0.0002/0.0083

= 0.02

'Lasker et al. (1970), larval anchovies.
2Lasl<er (1970), adult sardines.

^Paloma: unpublished data available at SWFC, juvenile an-
chovies.

sizes are small, in terms of positive (>mean) and
negative (< mean) coefficients, the transformation
was effective in normalizing both fish weight and

0/]U

lengin as snowi

ength Weight

Gi(mgi = 0)

0.12 0.39

0.12 0.33

Aiiij^ = 0.7979)

0.06 0.20

0.04 0.13

T7<_„ il _„

logL

■» Paloma, P. Unpublished data available at SWFC.

For these same samples, length and weight were
assumed bivariate-log normal and confidence
regions were calculated for each sample. On the
average, 96% of the observations fell within the
95% confidence ellipse.

In summary, there is strong evidence that the
logarithmic transformation will be required to
stabilize the variability in all phases of fish growth
and that such a transformation will support the
assumption of a normal distribution at least in the
intermediate size range (75-100 mm) and most
likely at other sizes as well.

Seemingly then, the conditions have been met
for implementation of either the maximum
likelihood or least squares estimation process.
However, two problems remain, neither of which
has an entirely satisfactory solution. The first, the
absence of an explicit solution of the normal
equations, arises because the parameters enter the
model in a nonlinear manner and, as is usual in

612

ZWEIFEL and LASKER: PREHATCH AND POSTHATCH GROWTH OF FISHES

situations of this kind, an iterative procedure is
required. Tlie one employed for this paper is
Marquardt's algorithm (Conway et al. 1970).
Procedures such as this are usually justified on the
basis that for large samples and independent
observations the estimates obtained are "very
close" to those which would be obtained by plot-
ting the likelihood function itself (Box and Jen-
kins 1970: 213). In truth, the small sample bias and
variability of such estimates remains unknown. In
growth data the second problem is that sequential
obsen^ations are not likely to arise from entirely
independent processes. This fact is usually man-
ifested as a series of runs above and below a fitted
curve rather than random variation. One simple
explanation is that growth is in reality a series of
asymptotic curves and that oscillations around a
fitted curve indicate more than one growth cycle.
In this case, the basic assumption of the estima-
tion procedure and the likelihood function itself
will not be met. No satisfactory solution to this
problem has been proposed and none is proffered
here. However, since the same larvae were not
measured at different ages and since correlated
observations usually have little effect on the
estimates of mean values, such estimates will
likely not be seriously biased. Using these es-
timates, "goodness of fit" is examined through the
magnitude of the residual mean square and the
pattern of residuals along the growth curve, rather
than using significance tests or confidence
intervals.

One further point often considered but left
unsaid is the effect of transformations on the
estimated means. Such changes of scale can lead to
serious biases and errors in interpretation,
especially when the coefficient of variation is
large. When the exact form of the error distribu-
tion is known the bias can usually be determined
mathematically. For the log normal, for example,
it is necessary to add one-half of the error mean
square before calculating the antilog mean. Un-
fortunately, in practical work, it is generally
impossible without very large samples, to deter-
mine the distributional form. As stated above, for
many situations, x and log x can both be considered
to be normally distributed. In these intermediate
cases, however, the bias correction for log x will be
small so, that as a general rule, one can state that
whenever a transformation is made, the correction
for transformation bias should be used.

RESULTS
Growth Cycles

Previous work on the growth of larval anchovies
(Kramer and Zweifel 1970) suggested that the
Laird form of the Gompertz equation might
provide a useful model of larval growth. Figure 2
reveals several phenomena found to be almost
universal in larval growth: 1) there is a moderate
increase in length during the interval following
hatch that is followed by 2) a period of minimal
growth accompanied by nearly uniform variabili-
ty, and 3) at the onset of feeding, the mean size
increases rapidly with variability proportional to
the square of the mean size.

Farris (1959) noted the rapid leveling off in
growth following hatch for the Pacific sardine and
three other species and approximated the growth
rate by two discontinuous curves and indicated
that "a more detailed study would probably reveal
a nonlogarithmic continuous growth function."

J meon ± 2 S 0

az'C

T ^^

,ir

1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 J. J

. I7°C

I 2''

S 20

16
12

qI I I ' I I I ■ ' I I I 1 1— I ' 1 1— I

0 4 8 12 16 20 24 28 32 36

DAYS AFTER HATCHING

Figure 2.-Change in length of yolk-sac and feeding larval
anchovies at two temperatures, 17° and 22°C from Kramer and
Zweifel (1970); curves are two-cycle Laird-Gompertz.

613

FISHERY BULLETIN: VOL. 74, NO. 3

Although the single stage model used by
Kramer and Zweifel (1970) provides an adequate
grow^th curve, tM^o growth cycles are evident: one
extending from hatching to the depletion of the
yolk sac and the other a more rapid growth at the
onset of feeding. Thus, a two-stage model was
used to obtain the curves in Figure 2. The fitting
procedure is outlined in the Appendix.

It is evident that early larval growth of this
species can be represented by a two-stage Laird
growth curve. The charcteristics of the growth
curves of feeding larvae, i.e., the second cycle, may
be related to several environmental factors of
which the two most important are probably food
ration and temperature. However, an examination
of data available on nonfeeding larvae (Figure 3)
indicated that even in food-limited situations,
change in size may be represented by the two-
stage Laird curve.

Growth From Hatch to Depletion
of Yolk Sac

The characteristics of the early posthatch
growth of larval fishes is more completely de-
scribed by Lasker (1964). In this series of exper-
iments, growth in length of the Pacific sardine,
Sardinops sagax, was measured for up to 10 days
following hatching at 12 temperatures in the
range 11°-21.3°C. The parameters of a single
stage Laird curve (Equation 2) were estimated for
each of these experiments. Data only up to the day
preceding the first decrease in size were used in
the calculations.

Even though for such short time series, the
parameters are highly correlated due to near-
redundancy of one of the parameters, two obser-
vations were striking; there was a nearly constant
estimated hatching length of about 3.75 mm and
a nearly constant estimated maximum length of
about 6.1 mm. Accordingly, those experiments
with hatching lengths near 3.75 mm and a mea-
sured increase in size of at least 3 days were fitted
to the reparameterized model:

L{t)T=Loe^^'-''"'"^

6.51-

where

K = Ao / a

and the T subscript indicates temperature in °C. A
plot of a-p on temperature revealed another
Laird-Gompertz curve approaching an asymptote
at higher temperatures.
A five parameter model:

o JACK MACKEREL

A SENORITA

• SQUARETAIL

□ SARDINE

2 4 6 8 10 12

DAYS AFTER HATCHING

Figure 3.-Change in length of yolk-sac and starving larvae;
curves are two-cycle Laird-Gompertz.

where

(5)
(5a)

was used to fit the growth data from all exper-
iments and provided an excellent fit except at the
highest temperature where growth was always
overestimated. This suggested a temperature
optimum with growth rates decreasing as the
absolute difference | T - T^pi \ increases. Following
Stinner et al. (1974), who used a different temper-
ature function, we assumed symmetry around the
optimum.

Using Equation (5a), the origin of the tempera-
ture scale may easily be shifted to the optimum
Topt by the relationships:

«opt = «oe

m(l-e'^^opt)

and

rw-opt = me'/^'^opt

614

ZWEIFEL and LASKER: PREHATCH AND POSTHATCH GROWTH OF FISHES

and letting 1=\t-T,J

we have the symmetric relationship

Oy = a

opt

e"'opt'

(5b)

Substituting Equation (5b) for Equation (5a) and
treating T^,p^ as an unknown parameter, a six
parameter model was fitted to the growth data
with the results shown in Table 2.

Table 2.-Growth in length of yolk-sac lan'ae of the Pacific
sardine at several temperatures.

Length

Age
(days)

Temper-
ature

Observed

1 Estimated^

(°C)

3.76

3.72

0.15

0.00

11.00

4.30

4.27

0.27

1.00

4.78

4.71

0.50

2.00

4.97

5.06

0.25

3.00

3.77

3.72

0.20

0.00

12.00

4.50

4.40

0.24

1.00

4.71

4.91

0.29

2.00

5.04

5.28

0.44

3.00

5.50

5.54

0.36

4.00

3.73

3.72

0.16

0.00

13.00

4.50

4.55

0.23

1.00

4.97

5.12

0.41

2.00

5.46

5.49

0.45

3.00

4.80

4.72

0.20

1.00

14.00

5.39

5.33

0.27

2.00

5.65

5.67

0.36

3.00

3.93

4.08

0.13

0.30

14.20

4.08

4.09

0.13

0.30

14.30

5.14

4.89

0.44

1.00

15.00

5.59

5.51

0.35

2.00

5.96

5.81

0.32

3.00

3.71

3.72

0.25

0.00

16.00

5.01

5.07

0.25

1.00

5.68

5.67

0.26

2.00

5.99

5.91

0.15

3.00

6.23

6.00

0.11

4.00

3.74

3.72

0.22

0.00

16.80

5.20

5.21

0.16

1.00

5.77

5.78

0.20

2.00

6.14

5.97

0.20

3.00

3.69

3.97

0.10

17.80

5.27

5.38

0.19

1.00

5.86

5.88

0.23

2.00

6.06

6.01

0.22

3.00

3.71

3.72

0.21

0.00

18.80

5.46

5.53

0.18

1.00

5.98

5.95

0.21

2.00

6.09

6.04

0.15

3.00

3.73

3.72

0.10

0.00

19.60

5.36

5.58

0.19

1.00

5.73

5.97

0.17

2.00

5.93

6.04

0.25

3.00

5.10

4.83

0.12

0.50

20.50

5.46

5.45

0.16

1.00

5.43

5.32

0.03

1.00

21.30

5.90

6.00

0.13

3.00

'From Lasker (1964).

^Calculated from Equations (5) and (5b) with parameters L^ =
3.716, K = 0.4872, a , = 1.8523, m = 3.3878, ft = 0.0490, and

opt

19.38.

Growth From Fertilization to Hatch

Coincident to the investigation of early larval

growth, a study of the incubation times for the
sardine showed that they also could be character-
ized by a Laird-Gompertz curve. The fitting of
Equation (5a) with uj- being incubation time
showed no bias at any point along the curve
(Figure 4). Unlike the posthatch growth curves,
however, no evidence of a temperature optimum
was found, i.e., incubation time did not increase at
high temperatures. One possible explanation is
that larvae which expire cannot be included and
hence mortality introduces a negative bias in the
estimate of average or median incubation time.

The question arises whether changes in growth
rates occur at hatching, i.e., is there a single curve
from fertilization to onset of feeding? It can be
shown that under the Laird-Gompertz model
w^here growth is approaching a common asymp-
tote from a common origin, i.e. fertilization, the
incubation time It is simple multiple of the decay
rate a j. From Equation (5) we may solve for the
time to hatch Ij at size L^ to obtain:

K-\n{L„/L,)

,/a.

Since incubation times were not available for all
temperatures used in the growth experiment, the
sardine curve from Figure 4 was used to convert
all data taken at temperatures less than optimum
to time from fertilization and fitted to Equation
(5).

The results for sardines indicated an increasing
size at hatch with increasing temperature which
was not evidenced by the observed data and an
overestimate of size at temperatures less than
14°C. It was thus concluded that a change in
growth rate occurs at hatch, the more noticeably at
extreme temperatures and that the prehatch curve
must be estimated separately.

The parameters of the prehatch growth curves
were obtained by fitting the equation

L„ =l^e^(^-e-"TiT)

(6)

to only data obtained less than 12 h following
hatch. The average estimated hatching size was
3.73 mm and the asymptotic limit was 6.13 mm.
The plot for several selected temperatures is
shown in Figure 5. Laird (1965a) has shown that
the length scale may be standardized and logically
simplified by expressing size relative to the
asymptotic limit. Biological events such as

615

FISHERY BULLETIN: VOL. 74, NO. 3

20.

BAIRDIELLA
°\ (6537,-7.1015,0.0650)*

J I

15 25 35 45

O
o

28r

24-

20-

16-

LlI

liJ 8

^ 22r

LlI

14-

24-

SHEEPHEAD
_ ^ (830.3,-78980,0.0250)*

ANCHOVY
(1861,-5.4572,0.0626)

I I I I I

JACK MACKEREL -
*° (6910,-6.1172,0.0599)

I I I I

SANDDAB
(1159,-5.1563,0.0629)*

J I I I I I I I L

HAKE
\(780.3,-4.l383,0.058l)

SARDINE
.(69IO,-6.2026,0£)905)

J L

TURBOT
(1012,-5.2103,0.0444)^

J I I I \ i I I I 1 1 1

PACIFIC MACKEREL |-
(3646,-6.4705,0.05351

J L

SENORITA
(8293 ,-7.6534 ,0.0624 )*

50 100 150

100

150

50 100

150 200

HOURS

Figure 4.-0bserved (0) and estimated (curve) (parameters in parentheses are /„, ni, and fi for the

-m(l - f"^ ^)

equation /t = /of and * indicates time from stage III eggs) incubation times for anchovy,

Engraulis mordax [combined data from Lasker (1964) and Kramer (unpubl. data) available at
SWFC]; hake, Merluccius productus; sheephead, Pimelometopon pulchrum; bairdiella, Bairdiella
icistia; jack mackerel, Trachurus symmetricus; sardine, Sardinops sagax; Pacific mackerel,
Scomber japonic us (Watanabe 1970); sanddab, Ciiharichthys stigmaeus; turbot, Pleuronichthys
decurrens; senorita, Oxyjulis californica.

developmental egg stages, hatching, and develop-
ment of the functional jav^^ occur at fixed points
along the curves. Ahlstrom (1943) reported time to
several developmental egg stages at different

temperatures from field observations. In addition,
Lasker (1964) showed incubation times and time to
the development of the functional jaw for a wider
range of temperatures. Each of these events can

616

ZWEIFEL and LASKER: PREHATCH AND POSTHATCH GROWTH OF FISHES

be identified as a fixed percentage point in Figure
5 or the estimated value

t,^ = In

(

K - In {Lt/Lo)

)

/a,

(7)

as shown in Table 3.

Lasker (1964) found that the functional jaw did
not develop at the lowest two temperatures in
agreement with the result that the critical size
would not be reached until well after yolk
absorption.

^° 'I°=.6.I3
5.97

E
E

2 4 6 8 10 12 14

DAYS FROM FERTILIZATION

Figure 5.-Prehatch growth curves estimated from Equations (5)
and (5a) for the Pacific sardine.

Incubation Times

Incubation times were available for several
other species. The fitting of Equation (5a) for each
species showed no clear bias at any point along the
curve (Figure 4). As for the sardine, no evidence of
a temperature optimum appeared for any of the
species in the temperature ranges used in the
experiments. However, it was observed that the
decay parameter was relatively constant varying
from 0.03 to 0.09 with a mean value of 0.05. When
Equation (5a) was fitted with the temperature
decay parameter, /?, the same for all species,
incubation times were closely approximated by
Equation (5a) with parameters as shown in Table
4.

The incubation curves used here differ sig-
nificantly from those calculated from the classical
Arrhenius equation: log (incubation time) = a +
6/absolute temperature. Using this method, near-
ly all species showed a characteristic under-
estimate at the temperature extremes and over-
estimates in the middle range as shown for the
northern anchovy, Engraulis mordax (Figure 6).

Prehatch Growth Curves for
Other Species

In addition to incubation times for the northern
anchovy, Kramer'' recorded time to several
developmental egg stages. Also, Lasker (1964)
provided time to hatch from stage IV^ (Table 5).
Further, Hunter (pers. commun.) indicates that

•■^Unpublished data available at SWFC.

''Stages of embryological development are those described by
Ahlstrom (1943).

Table 3.-0bserved (Obs.) and estimated (Est.)' time in hours to developmental egg stages^, hatch, and appearance of the functional jaw

of the Pacific sardine.

Ahlsf

rom (1943)

Lasker (1964)

Temp.

Stage
Obs.

III
Est.

Stage VI

Stages VIII-IX
Obs. Est.

Stage

Temp.
{°C)

Incubation time

Functior
Obs.

al Jaw

(°C)

Obs.

Est.

Obs.

Est.

Obs.

Est.

13.5

20.4

20.1

41.8

42.9

62.5

63.2

82.6

85.4

140

135

—

14.0

18.9

18.6

39.1

39.7

58.3

58.5

77.2

79.0

115

114

—

14.5

17.4

17.3

36.6

36.8

59.4

54.3

72.2

73.3

213

216

15.0

16.2

16.1

34.3

34.2

50.7

50.4

67.5

68.1

78.5

82.4

179

185

15.5

14.9

15.0

32.1

31.9

47.2

46.9

63.1

63.4

68.1

71.0

156

160

16.0

13.8

14.0

30.0

29.7

44.0

43.7

59.0

59.1

60.2

61.6

136

138

16.5

28.1

27.7

41.1

40.8

55.1

55.2

53.7

53.8

119

121

17.0

26.3

26.0

51.5

51.6

18
19
20
21

48.4
43.2
39.3
34.0

47.3
41.8
39.2
33.2

105
94
85
77

106
94
84
75

'/.^ = 0.0341, K = 5.20, a„ = 0.0317, m = 6.19, and /?
^Egg stages are defined in Ahlstrom (1943).

0.0489.

617

Table 4.- Parameters for estimating incubation time / at
centigrade temperature T from the relationship Ir =
/oe"^' "' ° ' for several fishes where P is the same for all species.

'Time from stage III eggs.

2.2r

2.0

P 1.8

<
CD

1.6

1.4

J-

OBSERVED

LAIRD -GOMPERTZ
— ARRHENIUS EQUATION
® COINCIDENT POINTS

0.00330

0.00340

0.00350

0.00360

^TEMPERATURE (°K)

Figure 6.-A comparison of two methods of fitting the tempera-
ture-incubation time relationship in the northern anchovy.

larval anchovy, on the average, hatch at about 2.9
mm.

Prehatch growth curves were obtained by
fitting Equation (6) to hatch sizes of 2.85 at all
observed temperatures as shown in Figure 7.

FISHERY BULLETIN: VOL. 74, NO. .3

21° 19° 17° 15° 13° 11°

Species

Sefiorifa

£
£

Oxyjulis calitornicus

'6,103

-7.9531

0.0527

Bairdiella

Bairdiella icistia

'3,170

-6.8216

0.0527

Pacific mackerel

Scomber japonicus

3,580

-6.4896

0.0527

•|-

Jack mackerel

Trachurus symmetricus

1,854

-6.2486

0.0527

Pacific sardine

Sardinops sagax

2,121

-6.2322

0.0527

Northern anchovy

Engraulis mordax

1,389

-5.5218

0.0527

Speckled sanddab

Citharichthys siigmaeus

'984.6

-5.4258

0.0527

Ijj

California sheephead

Pimelometopon pulchrum

'1,316

-5.4194

0.0527

Turbot

Pleuronlchthys decurrens

'1,065

-4.7059

0.0527

Pacific hake

Merluccius productus

699.2

-4.1772

0.0527

0.8

E
E

I
h-
o
z
llJ

2 4 6 8 10 12

DAYS FROM FERTILIZATION

Figure 7. -Prehatch growth curves estimated from Equations (5)
and (5a) for the northern anchovy.

Comparison with the sardine curves indicate that
similar events (i.e., stages of development) occur
relatively later for the anchovy. Observed and
estimated event times are shown in Table 5.

Except for size at hatch, development data for
the prehatch stage was not available for any other
species. The curves may, if desired, be easily
constructed from the parameters as shown in
Table 6.

DISCUSSION

Nothing seems more true than the statement of
Thompson (1942:158), "Every growth-problem
becomes at last a specific one, running its own
course for its own reasons. Our curves of growth
are all alike-but no two are ever the same. Growth
keeps calling our attention to its own complexity.
. . . not least in those composite populations whose
own parts aid or hamper one another, in any form
or aspect of the struggle for existence."

The truth of this statement has been realized in
the disappointing search for growth models de-
rived from physiochemical processes. While it is
true that the mathematical form of some equa-
tions arrived at from metabolic considerations are
the same as those derived in other ways, more

618

ZWEIFEL and LASKER: PREHATCH AND POSTHATCH GROWTH OF FISHES

Table 5.-0bserved (Obs.) and estimated (Est.y time in hours to developmental egg stages^, hatch, and appearance of the functional jaw

of the northern anchovy.

Kramer

(unpubl

data)

Lasker (1964)

Temp.
("C)

Stage

III

Stage VI

Stage
Obs.

VIII
Est.

Stag

B XI

Incubat
Obs.

on time
Est.

Temp
(°C)

Stage IV t(
Obs.

D hatch

Ops.

Est.

Obs.

Est.

Obs.

Est.

11.1

—

113

118.8

12.5

—

95.1

13.8

15.2

41.8

59.4

77.1

78.2

15.2

12.6

34.6

49.1

63.7

64.7

16.8

16.6

10.6

29.0

41.2

53.4

54.2

17.8

18.0

9.0

24.6

35.0

45.4

46.0

18.8

19.4

7.7

21.1

30.0

39.0

39.5

19.6

20.8

6.7

18.4

26.1

33.8

34.3

20.5

'Estimates obtained from Equation (7) with parameters as shown in Table 6.
'Egg stages are defined by Ahlstrom (1943).

Table 6.-Mathematical parameters for prehatch growth curves of six fishes. See text for

notation.

Average

size at

Species

L„

a ,

hatching

Trachurus symmetricus

0.0005

9.0986

0.0226

5.8338

0.0588

1.95

Sardinops sagax

0.0341

5.1918

0.0317

6.1876

0.0490

3.74

Engraulis mordax

0.0250

5.1493

0.0412

5.5338

0.0546

2.86

CitharicMhys stigmaeus

0.1814

5.0600

0.0270

6.2898

0.0319

1.97

Oxyjulis calilornicus

0.0425

4.7164

0.0572

7.2126

0.0260

1.89

Pleuronichthys decurrens

0.1843

3.2915

0.0480

4.5184

0.0528

3.00

often than not no meaningful biological interpre-
tation of the metabolic parameters can be made.
The essence of the growth equation used here is
genetically programmed processes of exponential
growth and of exponential decay of the specific
growth rate. The most probable source of expo-
nential growth is, of course, self-multiplication of
cells, the causes of decay are many but not well
understood. Laird (1964, 1965a, b, i966a, b, 1967)
has shown that this kind of relationship offers a
practical means of analyzing growth of all tumors,
as well as embryonic and postnatal growth of a
number of avian and mammalian species. We have
shown that at least the early stages of the growth
of fishes follows a similar pattern.

As with other organisms, several growth cycles
exist in fishes. The number of such cycles which
will be recognized is determined by the time scale
of measurements. We have used three cycles: 1)
from fertilization to hatching, 2) from hatch to
onset of feeding, and 3) feeding larvae.

In addition, we have observed that the temper-
ature specific growth follows a similar pattern, i.e.,
exponential increase with an exponential decay of
the temperature specific growth rate. In some
instances a temperature optimum exists beyond
which the specific growth rate begins to decline,
although this may be related to food requirements
at onset. of feeding. Further, we have observed

that for the same spawn 1) the asymptotic limit of
each growth cycle is independent of temperature
and 2) the biological events such as developmental
egg stages, hatching, functional jaw development,
etc., occur at the same size at all temperatures.

Figure 8 shows posthatch growth curves of the
sardine as 1) extrapolated from the prehatch
curves and 2) obtained from posthatch data. Al-
though the curves are quite similar at higher
temperatures, differences in the lower tempera-
ture range are large. Nevertheless, the time to
development of the functional jaw is much more
accurately determined from the extrapolated
curve, indicating an intrinsic process independent

5.97

I
I-
o

3.73

2 4 6 8 10 12

DAYS FROM FERTILIZATION

Figure 8.-Posthatch growth curves of the northern anchovy.
Solid lines are extrapolated from prehatch curve. Broken lines
are fitted to actual growth data.

619

of actual realized size. Comparison of the ex-
trapolated curves for the sardine and anchovy,
Figures 5 and 7, shows that for the same temper-
ature and relative to the asymptotic size, hatching
occurs later for the anchovy, but jaw development
and first feeding occur at about the same time.

In summary, each growth cycle may be repre-
sented by an equation of the form

L = L„e^''^-'""^"

with

o„e '"(i-'-/^^'

«r = «opt^ "1'

\T- T.

,(\-.-Ph

OJlt

when a temperature optimum exists. The time
required to attain a given size S is

t^ =

^='^ [kTi^J/"^

which has the same form as the original equation.

Most of the data available were from studies of
two species, Sardinops sagax and Engranlh mor-
dax, so that generalizations must be made with
caution. Nevertheless, incubation times for sever-
al other species fit the model well.

Finally, it seems worthwhile to repeat that
every growth problem becomes at last a specific
one depending on many factors known or un-
known, measureable or not. For example, time of
fertilization will often not be known and age
determinations will be inexact. Further, Hunter
and Lenarz' have shown that egg size is a mea-
surable and probably important factor in growth
and survival of anchovy larvae. For feeding larvae,
the quantity and quality of food is critical. Egg
size appears to afl'ect growth by a simple scale
factor, all events being shifted up or down in
proportion to the egg size. Variation in food may
result in many "artificial" cycles when nutritional
and caloric requirements are not met. Neverthe-
less, it seems clear that at least the early growth of
many fishes may be described in terms of genet-
ically determined but dynamically changing
growth rates as defined by the Laird-Gompertz
growth function.

'Hunter, J., and W. Lenarz. 1974. A discussion on the adaptive
values of variation of fish egg sizes. Unpubl. manuscr., 7 p.
Southwest Fisheries Center, Tiburon Laboratory, National
Marine Fisheries Service, NOAA, Tiburon, CA 94920.

P^ISHERY BULl.KTI.N: VOL. 74, NO. 3

ACKNOWLEDGMENTS

We express our appreciation to David Kramer
and Pete Paloma of the National Marine Fisheries
Service for making unpublished data available to
us, to Michel Coirat for her diligent eff"orts in the
laboratory, to Lorraine Downing for her typing
skill and patience with mathematical formulae,
and to John R. Hunter for his advice on this work.
Special thanks are due to Fisherii Bulletin
reviewers for checking our mathematics and for
constructive criticism of the manuscript.

LITERATURE CITED

Acton, F. S.

1959. Analysis of straight-line data. John Wiley and Sons,
Inc.,N.Y!.267p.

AHL.STKOM, E. H.

1943. Studies on the Pacific pilchard or sardine {Sardiiwps
caerulea). 4. -Influence of temperature on the rate of
development of pilchard eggs in nature. U.S. Fish Wildl.
Serv., Spec. Sci. Rep. 23, 26 p.
Beverton, R. J. H., AND S. J. Holt.

19.57. On the dynamics of e.xploited fish populations. Fish.
Invest. Minist. Agric, Fish. Food (G.B.), Ser. II, 19, .533 p.
Bo.x, G. E. P., AND G. M. Jenkins.
1970. Time series analysis forecasting and control. Holden-
Day, San Franc, 553 p.
Chapman, D. G.

1961. Statistical problems in the dynamics of e.xploited
fisheries populations. Proc. Fourth Berkeley Symposium
on Mathematical Statistics and Probability 4:153-168.
Conway, G. R., N. R. Glass, and J. C. Wilcox.

1970. Fitting nonlinear models to biological data by Mar-
quardt's algorithm. Ecology 51:503-507.
Farris, D. a.

1959. A change in the early growth rates of four larval
marine fishes. Limnol. Oceanogr. 4:29-36.
Hald, a.

1952. Statistical theory with engineering applications. John
Wiley and Sons, Inc., N.Y., 783 p.
Kramer, D., and J. Zweifel.

1970. Growth of anchovy larvae (Engraulis mordax Girard)
in the laboratory as influenced by temperature. Calif.
Coop. Oceanic Fish. Invest. Rep. 14:84-87.
Laird, A. K.

1964. Dynamics of tumor growth. Br. J. Cancer 18:490-502.
1965a. Dynamics of relative growth. Growth 29:249-263.
1965b. Dynamics of tumor growth: Comparison of growth

rates and extrapolation of growth curve to one cell. Br. J.

Cancer 19:278-291.
1966a. Dynamics of embryonic growth. Growth 30:263-275.
1966b. Postnatal growth of birds and mammals. Growth

30:349-363.

1967. Evolution of the human growth curve. Growth
31:345-355.

Laird, A. K., A. D. Barton, and S. A. Tyler.

1968. Growth and time: An interpretation of allometry.
Growth 32:347-354.

Laird, A. K., S. A. Tyler, and A. D. Barton.

1965. Dynamics of normal growth. Growth 29:233-248.

620

ZWEIFEL and LASKER; PREHATCH ANU FOSTHATCH GROWTH OF FISHES

Lasker, R.

1964. An experimental study of the effect of temperature on
the incubation time, development, and growth of Pacific
sardine embryos and larvae. Copeia 1964:399-405.
1970. Utilization of zooplankton energy by a Pacific sardine
population in the California Current, hi J. H. Steele
(editor), Marine food chain, p. 265-284. Oliver and Boyd,
Edinb.
Lasker, R., H. M. Feder, G. H. Theil.\cker, and R. C. May.

1970. Feeding, growth, and survival of Engrauli!< morda.r
larvae reared in the laboratory. Mar. Biol. (Berl.)
5:345-353.

Leong, R.

1971. Induced spawning of the northern anchovy, Engraulis
mordax Girard. Fish. Bull., U.S. 69:357-360.

May, R. C.

1971. Effects of delayed initial feeding on larvae of the
grunion, Leiiref<fhcs tenuis (Ayres). Fish. Bull, U.S.
69:411-425.
Richards, F.J.

1959. A flexible growth function for empirical use. J. Exp.
Bot. 10:290-800.
Stinner, R. E., a. p. Gutierrez, and G. D. Butler, Jr.

1974. An algorithm fortemperature-dependent growth rate
simulation. Can. Entomol. 106:519-524.
Taylor, C. C.

1962. Growth equations with metabolic parameters. J.
Cons. 27:270-286.
Thompson, D'Arcy Wentworth.

1942. On growth and form. 2nd ed. Cambridge Univ. Press,
464 p.
vonBertalanffy, L.

1938. A quantitative theory of organic growth (inquiries on
growth laws. II). Human Biol. 10:181-213.
Watanabe, T.

1970. Morphology and ecology of early stages of life in
Japanese common mackerel, Scunibcrjapdniciis Houttuyn,
with special reference to fluctuation of population. [In
Engl, and Jap.] Bull. Tokai Reg. Fish. Res. Lab. 62, 283 p.

APPENDIX

The estimation procedure of Conway et al.
(1970) is a least squares procedure which requires
only the definition of the functional relationship
and the first derivative with respect to each
parameter. Although not stated explicitly, con-
stant variance is assumed and, hence, the loga-
rithmic form will be used throughout. For a sin-
gle-cycle Laird-Gompertz curve the equations are
as follows:

InF = InFo + A[l - Ex-p{-at)]/a

■^0

6lnF
A

6\nF

= [1 - Exp(-aO]/«

= A[{at + 1) Exp(-aO - l]/a2

For a two-cycle curve with the second cycle begin-
ning at t = t* the equations are:

\r\F = InF,, + ^[l-Exp(«Ai)]/a
+ B[l - Expi-0^2)]//3

^^ ^ VF

6Fo

*1^ = [1 - Exp{-aA,)]/a

dlnF
ba

6lnF
h\nF

b\nF
bt*

where

= A[(al, + 1) Exp(-aAi) - l]/a2

= [l-Exp(-/8A,)]//3

-yS[(y5A2 + 1) Exp(-/?A,,) - l]/yS2

= [A Exp(-aAi) - B Exp(-^2)]

Al = MlNitJ*)

A, = MAX(^-r ,0).

FORTRAN programs are available for fitting
single-cycle, temperature-dependent and multi-
cycle, temperature-dependent curves at SWFC.

621

NORTH AMERICAN CRAB FISHERIES: REGULATIONS
AND THEIR RATIONALES

R. J. Milleri

ABSTRACT

Because of similarities in species' life histories, fishing and processing methods, economics of fishing
and processing, and political systems among jurisdictions, managers of North American crab fisheries
share many common problems. This review is presented to suggest options to those delegated the
responsibility for managing crab fisheries.

The review is organized by fishery and management problems. Six fisheries in 12 government
jurisdictions are included. Regulations are grouped into management problems of 1) conservation, 2)
allocation of landings among commercial fishermen, 3) stability of landings, 4) conflict over grounds or
resource, 5) processing economics, and 6) administration. A final section discusses procedures in eight
jurisdictions by which public or government representatives may effect changes in regulations.

If the rationale for each regulation (or at least each new one) and the name of the group requesting it
are appended to copies distributed to users, more informed discussion of management problems and
more reasoned support for regulations may result.

Problems of managing crab fisheries change as
established fisheries develop and new fisheries
emerge. Because of similarities in species' life
histories, fishing and processing methods, eco-
nomics of fishing and processing, and political
systems within which both government employees
and fishing industries must operate, managers of
North American crab fisheries share many com-
mon problems. This review of North American
crab fishery regulations and their rationales is
presented to suggest options to those delegated
the responsibility for managing crab fisheries.
While these regulations may not be optimum
according to either biological or economic criteria,
they have met the very demanding test of political
feasibility.

This review is organized by fishery and man-
agement problems. The classification of man-
agement problems is necessarily arbitrary. Jus-
tifications for a given regulation may make it
applicable to more than one problem in the same
governmental jurisdiction or applicable to differ-
ent problems in different jurisdictions. The clas-
sification is an attempt to make the presentation
more user-oriented, as a search for regulations is
commonly prompted by a particular management
problem. A final section contains procedures for
eight jurisdictions whereby either public or

'Department of the Environment, Fisheries, and Marine
Service Biological Station, St. John's, Newfoundland, Canada,
AIC lAl.

government representatives may recommend
changes in regulations.

METHODS

Information on regulations and their rationales
was provided by government biologists and re-
source managers in interviews on the west coast
and in correspondence on the east coast. Table 1
lists these contacts and their agencies.

The regulations are not a complete set for any
jurisdiction but do represent a large sample of the
types of controls in force. Some regulations are
omitted because I judged them not to be of general
application or my contacts did not know their
rationales; the latter is understandable consider-
ing the time period over which most sets of
regulations evolved. Sampling was least complete
for the blue crab fishery. There are 16 States with
regulations governing this fishery and several
were not included because of the similarity among
their regulations.

I have not commented on the success of en-
forcement of regulations as this would have
required firsthand knowledge of each fishery or
extensive field interviews with enforcement
oflScers and fishery participants.

The management problems into which regula-
tions have been grouped are listed and defined
below.

Conservation: to prevent resource waste,

Manu.script accepted March 1976.

FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

623

FISHERY BULLETIN: VOL. 74, NO. 3

Table 1. -Sources of information.

Fishery

Contact

Agency

Alaska king crab

Guy C. Powell
Duane E. Phinney

Alaska snow crab

Duane E. Phinney
Guy C. Powell

West coast Dungeness

crab

Duane E. Phinney
T. H. Butler
Herb C. Tegelberg
C. Dale Snow
Walter Dahlstrom

Eastern Canada snow

crab

Author (R. J. Miller)

East coast blue crab

Richard W. Cole

William A. Outten
Dennis L. Spitsbergen
Edwin A. Joyce
Edgar A. Hughes
Terrance R. Leary

Florida stone crab

Edwin A. Joyce

Alaska Dap. Fish and Game

Alaska Dep. Fish and Game

Canadian Fisheries & Marine Serv.

Washington Dep. Fisheries

Oregon Fish and Wildlife Commission

California Dep. Fish and Game

Canadian Fisheries & Marine Serv.

Delaware Dep. Natural

Resources & Environmental Control

Maryland Dep. Natural Resources

North Carolina Dep. Natural and Economic Resources

Florida Dep. Natural Resources

Alabama Dep. Conservation & Natural Resources

Texas Parks and Wildlife Dep.

Florida Dep. Natural Resources

principally by fishermen, and to maximize
physical yield.

Allocation of landings among commercial
fishermen: to partition the annual catch of
a single species among participants, usually
by area or gear restrictions.

Stability of landings: to even out annual
landings over good and bad years of re-
source recruitment.

Conflict over grounds or resource: to resolve
competition among classes of users for
fishing grounds or fishery resources.
Sport-commercial conflicts over the same
species are included in this category.

Processing economics: to limit landings to
crabs that can be processed at a profit
acceptable to processors.

Administration: licensing and registration of
boats, men, and gear, and collection of
statistics.

RESULTS

Alaska King Crabs {Paralithodes

camtschatica, P. platypus, P. brevipes,

and Lithodes aequispina)

Exploitation of American stocks of king crabs
was sporadic prior to 1953, but annual landings
increased from 5 to 159 million pounds from 1953
to 1966, sharply decreased, then recovered to 90
million pounds in 1974 and 1975 (Rothschild et al.
1970; D. E. Phinney, pers. commun.). The Amer-
ican fishery is pursued along most of the Alaskan
coast from the Bering Sea pack ice to the northern
end of Vancouver Island over a depth range of 20
to 100 fathoms (Idyll 1971). The catch is

predominately P. camtschatica. Although Japan-
ese and Russian fleets formerly took large catches
from the Bering Sea, their efforts are now con-
centrated in the western North Pacific. The fol-
lowing regulations are those of Alaska.

Conservation

Seasons are set to prevent the taking of soft-
shelled crabs during and immediately after the
molting season. Soft-shelled crabs provide a low
meat yield, the quality of meat is poor, and the
handling mortality is high.

Harvest levels (i.e., variable catch quotas) and
minimum sizes ensure that enough mature males
are left on the fishing grounds for breeding. Males
are sexually mature for an average of 2 yr before
reaching the minimum size. The minimum size
also helps maximize yield as determined from
growth and mortality rates.

Most females are protected by the minimum size
but a separate regulation prevents their being
retained to leave them for reproduction.

Gear type is limited to traps, ring nets (a type of
baited trap), and diving to prevent use of destruc-
tive trawls and tangle nets. The latter two types of
gear result in unacceptable levels of mortality of
noncommercial crabs returned to the water and of
commercial crabs before they reach the processing
plant.

One nursery area is always closed to fishing to
prevent repeated handling of undersized and
female crabs.

To help enforce seasons and harvest levels,
tunnel eyes, i.e., entrances of traps, must be at
least 5 inches high. This is so that king and snow

624

MILLER: NORTH AMERICAN CRAB FISHERIES REGULATIONS

crab traps can be distinguished and to prevent
king crab fishing in seasons for snow crab fishing
only.

A second season in a year in the Kodiak area has
a larger minimum crab size than the primary
season. This encourages boats to fish areas where
catch per trap is lower than in more productive
areas but where large, old crabs have accumulated
because the areas have been underfished.

A subsistence or sport limit of six crabs per day
limits waste that might result from higher
catches.

Allocation of Landings Among Commercial Fishermen

There are seven exclusive and two nonexclusive
registration areas. A boat may register before the
season opens to fish in only one exclusive area but
in either or both nonexclusive areas in addition. To
enforce this regulation and to prevent fishing
before the season opens, a boat must have its hold
inspected prior to fishing to verify that no king
crabs are on board. If a boat wishes to land its
catch outside its registration area, it must report
by radio to a designated authority the size of its
catch, and it may be required to submit to a hold
inspection before leaving its registration area.
The boat may at the time of landing have no more
or less king crabs on board than were present at
the time of reporting or inspection. To revalidate
its registration, a boat must be reinspected in its
registration area prior to resuming fishing. As the
exclusive areas are more accessible to harbors and
population centers, they are easier to fish than
nonexclusive areas. By limiting boats to one
exclusive area the larger, more mobile boats must
take part of their catch from more remote areas
less accessible to small boats. The small-boat
operations are economically viable because of their
versatility to participate in other fisheries, e.g.,
salmon, halibut, and shrimp. A boat operator is
limited to operating only one boat in one exclusive
area although he may operate the same or addi-
tional boats in nonexclusive areas. This excludes
one-operator fleets from exclusive areas.

Trap limits per boat in some areas favor small
boats because large boats cannot operate as eco-
nomically if their fishing power is restricted.

Local boats are favored as an ancillary effect of
the second season mentioned above. The catch per
trap is lower and the weather less favorable than
in the primary season, and nonlocal boats are often
unwilling to fish for the lower returns.

Stability of Landings

Harvest levels are set to ensure that at least two
year-classes are well represented in any year's
landings. This helps dampen the effect on landings
of uneven annual recruitment to commercial size.

Conflict Over Grounds or Resource

Trap sanctuaries off limits to towed gear have
been negotiated with foreign groundfish trawlers.
Foreign trawlers have also agreed to area closures
and to use rollers on trawls to reduce the catch of
king crabs. Domestic shrimp trawlers and scallop
draggers are excluded from some prime king crab
grounds.

Processing Economics

Crabs are hard shelled much longer than the
time required for the fishery to take the annual
harvest levels. There is, however, a slight im-
provement in meat yield as the hard-shelled period
progresses with the best yield occurring in most
areas at times when the weather is unfavorable
for fishing. The season opening within the hard-
shelled period is a compromise between the re-
spective interests of fishermen and processors.

Administration

Boats are licensed and registered each year and
boats and crab-trap buoys must clearly display
registration numbers. Plants are obligated to
report area of catch, number of trap lifts, and
landings by boat. These regulations are necessary
to enforce fishing-area and harvest-level restric-
tions as well as to provide economic and biological
data on the fishery.

Alaska Snow Crabs
{Chionoecetes bairdi, C. opilio)

Although Chionoecetes bairdi, C. opilio, and C.
tanneri are all referred to as snow crabs, the
current domestic fishery consists of about 98% C.
bairdi and 2% C. opilio. Alaskan landings have
increased rapidly from 3 million pounds in 1968 to
61 million pounds in 1973. This fishery operates
from the Bering Sea to southeastern Alaska over a
depth range of 20 to 140 fathoms (Brown'-'). As in

^Brown, R. B. The development of the Alaskan fishery for
tanner crab, Chionoecetes species, with particular reference to
the Kodiak area. Unpubl. manuscr., 15 p. Alaska Dep. Fish Game,
Kodiak.

625

FISHERY BULLETIN: VOL. 74, NO. 3

the king crab fisheries, Japan and the USSR
formerly took arge quantities from the Bering
Sea, but the USSR has not fished since 1971 and
the Japanese catch is limited by bilateral treaty to
about 22 million pounds per year (D. E. Phinney,
pers. commun.). Alaska is the only North Amer-
ican jurisdiction with regulations for this
fishery.

Conservation

The following regulations serve the same pur-
pose as in the king crab fishery. Seasons prevent
fishing when crabs are soft shelled; fishing gear is
limited to traps, ring nets, and diving; harvest
levels help ensure enough males are left on the
grounds for breeding; females may not be taken;
and subsistence fishing is limited to 30 crabs per
day. Trap tunnel eyes must be less than 5 inches
high when the king crab season is closed to distin-
guish between snow and king crab traps and to
reduce the incidental catch of king crabs.

Cone-shaped traps with a single top entrance
may be used for snow crabs in addition to the
rectangular king crab trap modified with a smaller
tunnel entrance.

Location of Landings Among
Commercial Fishermen

As in the king crab fishery, there are exclusive
(two) and nonexclusive (three) fishing areas. A
boat may register for either one exclusive area or
any number of nonexclusive areas. A boat must
have its hold inspected to validate its registration
and must report prior to landing its catch in an
area other than where it is fishing. There are also
trap limits for some areas. The rationale for these
is the same as in the king crab fishery.

Stability of Landings

Annual harvest levels by area dampen the effect
on landings of variable recruitment to commercial
size.

Conflict Over Grounds or Resource

As with the king crab fishery, foreign trawlers
have agreed to area closures and to use rollers on
trawls to restrict the incidental catch of snow
crabs. The trap sanctuaries for king crabs also
protect the snow crab fishery in many cases.

Processing Economics

The season within the hard-shelled period is set
for the convenience of fishermen and processors.

Although there is no minimum size restriction,
most immature males are returned to the water on
the fishing grounds because they are too small to
be processed economically.

Administration

The regulations are similar to those for the king
crab fishery.

West Coast Dungeness Crab
( Cancer magister )

This is an old fishery with commercial exploita-
tion since at least 1917 (Cleaver 1949). Landings
are quite variable ranging from 14 to 60 million
pounds in the 1970's alone.'' The fishery operates
from southwest Alaska to central California over a
depth range of 1 to 20 fathoms.^ Only United
States and Canada fish this species.

Conservation

Closed seasons for the commercial fishery pro-
tect soft-shelled crabs in at least some areas of all
jurisdictions. Seasons also apply to the sport
fishery in California and ocean beaches in Oregon.
In addition to a season, Washington specifically
prohibits the landings of soft-shelled crabs: "A
soft-shelled crab is defined as a crab whose shell,
including covering of the legs, is not fully hard-
ened and said shell is flexible and depresses to
digital pressure!' This regulation has been upheld
in Washington courts.

Females may be retained by commercial fisher-
men only in British Columbia and by sportsmen
only in California and British Columbia provided
they exceed the minimum legal size. They are
protected for breeding purposes (Alaska,
Washington, Oregon, California) and because of
processing considerations (Washington, Oregon).

Traps left unattended for over 2 wk must have
bait removed and doors secured open as protection
against ghost fishing (Alaska).

^Anon. 1974. Crab review. Fisheries and Fish Prod. Div.,
Fisheries and Food Prod. Br., Dep. Industry, Trade, and Com-
merce, Ottawa, 83 p.

'Anon. 1972. Pacific edible crab. Fishery Fact Sheet, 2 p. Dep.
Environ., Ottawa.

626

MILLER: NORTH AMERICAN CRAB FISHERIES REGULATIONS

Types of gear are regulated by stating either
what may or what may not be used. The effect is to
Hmitthe commercial fishery to traps and ring nets,
and the sport fishery to traps, ring nets, dip nets,
handlines, and diving. Sharp instruments, tangle
nets, and usually trawls are excluded to avoid
unacceptable levels of crab mortality. To allow
escapement of subcommercial-sized crabs, one or
two rings of at least 4-inch diameter must be set in
the trap mesh in all jurisdictions. This is usually
required to be in the upper half of the trap to
reduce the chance of openings being covered by
drifting sand.

The minimum size is regulated in all jurisdic-
tions. It allows males to mate at least once before
reaching legal size although opinions among jur-
isdictions differ as to whether their respective
minimum sizes are biologically optimum. To help
enforce size regulations, crabs must be landed
whole.

Allocation of Landings Among
Commercial Fishermen

Alaska has trap limits which vary considerably
among areas. The low limits discourage participa-
tion of large boats and reserve the resource for
small and local boats. British Columbia limits
commercial gear in one area to ring nets or dip
nets and traps are excluded in five bays in Oregon
to eliminate large commercial operators.

Alaska has both exclusive and nonexclusive
fishing areas as in the king and snow crab fisher-
ies, for the same reasons and with the same
supporting regulations. As in the snow crab
fishery, a boat may not be registered in both
exclusive and nonexclusive areas whereas in the
king crab fishery a boat may register in one
exclusive plus nonexclusive areas.

Conflict Over Grounds or Resource

All jurisdictions have a small catch quota for
sport fishermen, ranging from 20 crabs per day in
Alaska to 6 per day in British Columbia and
Washington. Sport fishermen are limited to three
traps or three ring nets in Oregon and two traps or
two ring nets in Washington. These regulations
serve to differentiate between sport and commer-
cial fishermen and, in some areas, to divide the
available catch among many sport fishermen.

There are a number of concessions to sport
fishermen in British Columbia, Washington,

Oregon, and California for this very accessible
species. The fishery is open to only sport fishermen
in a marine park in British Columbia, in Hood
Canal in western Puget Sound in Washington, and
in bays, harbors, and near jetties in California. A
20-trap commercial limit imposed in Dungeness
Bay, Wash., controls competition with sport
fishermen. A slightly smaller minimum crab size is
applied to sport than to commercial catches in
Washington and Oregon to increase the sport
share of the catch. This size difference is sig-
nificant because in areas available to the commer-
cial fishery over 80% of the legal-sized crabs are
generally caught in the first few months of the
season.

Salmon troller operators and crab fishermen in
Oregon have an informal agreement to divide the
grounds at the 15-fathom contour to resolve in-
compatible use of the fishing grounds. California
trawlers are permitted to land up to 500 pounds of
legal-sized male crabs per trip during the crab
season. This discourages trawling directed at
Dungeness crabs but allows them to retain in-
cidental catches.

Processing Economics

In addition to protection of their reproductive
role, females may not be retained in Washington
and Oregon because the meat yield and quality are
lower than for males.

Some areas near the City of Vancouver, B.C., are
closed to both commercial and sport crab fishing
because of polluted water.

Administration

A commercial fishing license specifically for
crabs is required in Alaska and Washington, while
only a general commercial license is required in
British Columbia, Oregon, and California. Al-
though this provides a small amount of revenue, it
is primarily for records on participants.

Eastern Canada Snow Crab
{Chionoecetes opilio)

This fishery has two centers of operation, the
western Gulf of St. Lawrence and eastern New-
foundland. The first significant commercial land-
ings were taken in the Gulf of St. Lawrence in 1967
and in Newfoundland in 1969. Landings from both
areas totaled 23 million pounds in 1974 with

627

FISHERY BULLETIN: VOL. 74, NO. 3

further expansion expected only in Newfound-
land. The fishery operates in depths of 40 to 70
fathoms in the Gulf of St. Lawrence and 90 to 200
fathoms in Newfoundland. Only Canada is en-
gaged in this fishery and there is no sport fishery.
Regulations for this fishery have only recently
been implemented, so, although they have cleared
public service and political hurdles, they have yet
to be tested by performance.

Conservation

Any specified area may be closed to the fishery
at any time for conservation reasons. Justifica-
tions could be an abundance of soft-shelled or
sublegal-sized crabs in the catches. Periodicity of
soft-shelled abundance is not predictable enough
to set annual seasons.

Fishing is permitted only by traps to exclude the
wasteful bottom trawl and tangle net gears. A
minimum mesh-size regulation allows escapement
and eliminates handling of a large portion of the
sublegal-sized crabs. A minimum crab size is set
(Newfoundland only) in hope of maximizing the
yield per recruit, to ensure the presence of enough
mature males for mating success, and to satisfy
processing requirements. The minimum size ex-
cludes all females.

A regulation requires that soft-shelled crabs be
returned to the water on the fishing ground. They
are unacceptable for processing because of low
meat yield, poor quality meat, and poor survival
while being held for processing. If landed, they are
discarded by processors.

Allocation of Landings Among
Commercial Fishermen

Trap limits in the Gulf of St. Lawrence limit the
fishing effort per boat.

Any new boats entering the fishery after 1974
must be recommended by a crab management
committee composed of representatives from
fishermen, processors. Provincial governments,
and the Federal government, and approved by a
Regional Director of Fisheries. New entrants are
considered for underexploited areas only.

Stability of Landings

A single quota for all of Newfoundland is in-
tended to dampen the effects on landings of
variable recruitment to commercial size.

Processing Economics

With present technology and product prices,
crabs smaller than the legal minimum cannot be
processed economically.

Administration

Crab boats must be licensed specifically for crab
fishing to control entry and to provide economic
data on the fishery. In Newfoundland, boats must
report their fishing area to provide a history of
yields by area.

East Coast Blue Crab
{Callinectes sapidus)

This species has supported a commercial fishery
since at least 1890 (Newcombe 1945). Landings
have been about 140 million pounds annually in the
1970's (footnote 3). The fishery operates along
most of the U.S. Atlantic coast and all of the Gulf
of Mexico coast, but the bulk of the landings and
the most extensive fishery regulations are from
the mid-Atlantic States. The depth range of the
fishery is between less than 1 to 10 fathoms and
the fishery is prosecuted entirely from the United
States.

Conservation

Generally, egg-bearing females must be re-
turned to the water to allow them to release their
progeny. This is requested by the fishing industry
(Delaware, Maryland, Florida, Texas) although
there may be no biological evidence establishing a
relationship between the size of the parent stock
and strength of the resulting year-classes
(Delaware, Florida, Texas).

To allow escapement of small crabs, Maryland
requires that the wire mesh covering traps be a
minimum of 1 by 1 inch, Florida requires that an
escape hole near the bottom of traps be a minimum
of 2 by 2 inches, and Texas requires that crab
trawls have a minimum mesh size of 5 inches
(stretch measure).

Seines must be hauled up in the water rather
than on shore in Maryland to help ensure that
unwanted animals such as small crabs and fish are
returned to the water rather than left on the
beach.

Hard-shelled crabs must be a minimum of 5-inch

628

MILLER: NORTH AMERICAN CRAB FISHERIES REGULATIONS

width in Delaware, Maryland, and one county in
Florida. This is slightly over the average size at
maturity and allows crabs to spawn at least once
before being subject to depletion by the fishery.
One bushel of undersized crabs is permitted in a
daily catch in Delaware.

A 150-trap limit is enforced in small bays just
inland from Maryland's ocean beaches as the
peeler-crab (crab about to molt) fishery has
recently become intensive on the small popula-
tions in these bays. Baiting traps with live males
ensures a high proportion of females in the catch
which molt, then copulate. Effort control through
trap limits is an attempt to prevent rates of
female removal that would affect the ability of the
population to replace itself.

Allocation of Landings Among
Commercial Fishermen

A 150-trap limit for some areas in Delaware is
near the maximum most boats can fish per day and
controls the fishing power of the few fishermen
who would choose to fish more traps.

North Carolina prevents the use of dredges
operated by power winches in one area. This
controls the fishing power of boats using dredges
for any species but was aimed primarily at the
oyster fishery.

Conflict Over Grounds or Resource

Sport fishermen have gear and catch limits to
control their competition with commercial fisher-
men and to distribute the available landings
among many sport fishermen. Limits are two
traps, four handlines, and one-bushel catch per day
in Delaware; one-bushel catch per day in Mary-
land; one trap which may not be fished from a boat
in North Carolina; and five traps in Florida.

Most Maryland streams emptying into Chesa-
peake Bay are off" limits to traps to reserve the
areas for crab fishermen using trot lines. In North
Carolina, crab traps are excluded from some areas
from 1 April through 30 November to reserve the
areas for haul seines and shrimp trawls. Other
areas in North Carolina are designated for fixed
gears only, however, to protect them from towed
gears.

Traps may not be set in marked navigation
channels (Delaware, Maryland, North Carolina),
may not be set in water shallower than 4 feet at
mean low tide (Maryland), or may not be larger

than 24 inches on a side (Maryland) because of the
hazard to navigation. Traps may not be set near
bathing beaches in Maryland because the presence
of fishing boats, discarded bait, and discarded dead
crabs interfere with the recreational use of the
beaches.

Crab dredges (also rakes and scrapes) may not
be used on bottom leased for oyster propagation
except by the lease holder (Delaware) or on public
oyster grounds where oysters or shells have been
planted by the State (North Carolina). Dredging is
not permitted from 16 March to 15 December
(Delaware) or 1 April to 30 November (North
Carolina) since crabs are not buried in the bottom
during this time, and dredging is destructive to
both commercial molluscs and noncommercial
benthos. Maximum dredge weight is 100 pounds in
North Carolina and 40 pounds in Maryland to limit
destruction of bottom organisms by the gear.

North Carolina and Texas restrict crab trawling
because of possible damage to shrimp stocks. Some
shrimp nursery areas are closed to crab trawling
since the resulting turbidity may be lethal to
shrimp (North Carolina). Mesh size on trawls may
not be smaller than 3-inch stretch measure when
used for hard-shelled crabs nor smaller than 2-inch
stretch measure when used for soft-shelled crabs
to allow shrimp to escape (North Carolina). The
5-inch minimum mentioned earlier serves the
same purpose in Texas. Trawls used for soft-
shelled and peeler crabs are limited to 25 feet in
width (float line length) to control damage to sea
grass beds where most of this fishing occurs (North
Carolina).

Crab trawling is prohibited from 2000 h on
Saturday to 2000 h on Sunday to eliminate the
time conflict with fishers of men (North Carolina).

Processing Economics

Processors have requested that egg-bearing
females not be landed because of their low meat
yield (Texas). North Carolina prohibits trawling in
ocean inlets to interior bays from 1 April to 31
August because females incubating eggs are
concentrated in these areas: females are uneco-
nomical to process.

Minimum shell width for hard-shelled crabs is 4
inches (Alabama) or 5 inches (Maryland, North
Carolina-10% undersized permitted). This pro-
tects the processor from pressure to accept small
crabs which are unprofitable to process.

629

FISHERY BULLETIN: VOL. 74, NO. 3

Soft-shelled crabs and peelers have legal min-
ima of 3V2 inches and 3 inches, respectively, in
Delaware and Maryland compared to a 5-inch
minimum for hard-shelled crabs. Because of the
greater molting frequency of smaller crabs, the
smaller size limit permits a greater volume of this
relatively high-priced product. Soft-shelled crabs
are sold and eaten whole so the economics of meat
extraction is not a consideration. Crabs with a
shell just starting to harden (paper shell) may not
be landed as they are not suitable for the soft-shell
market and the meat yield is too low for processing
as hard-shelled crabs (Maryland).

In Maryland, dredges are permitted only from
15 April to 31 October (compare with summer
closure in Delaware and North Carolina in the
previous section). The crabs are not buried in the
sediment in this period and have had time to clean
themselves of attached mud making them a more
desirable product.

Administration

A commercial license is required specifically for
blue crab in Delaware and Maryland, while a
general commercial license will sufl'ice in North
Carolina and Texas. No license is required in
Alabama and only a permit number is required in
Florida. Traps are generally required to be buoyed
and must have the boat permit or license number
displayed on buoys in Florida, Maryland, and
Delaware. This is to reduce the navigation hazard
of traps and to help enforce seasons and registra-
tion requirements.

Florida Stone Crab
{Menippe mercenaria)

This species has supported a commercial fishery
in Florida for approximately 25 yr. Landings have
recently increased from 1 million pounds in 1965 to
2.1 million pounds in 1973. The fishery operates
around most of Florida's coast over less than 1 to 8
fathoms depth, but 80% of the landings are taken
from the Keys and the southwest coast. The
fishery does not exist in other areas of the United
States but does extend into the Caribbean.

Conservation

This fishery has a unique regulation requiring
that only the claw may be retained. The remainder
of the crab must be returned live to the water. The
crab market accepts only the claw. A small per-

centage of declawed animals survive to spawn and
a small percentage regenerate the claw to com-
mercial size. The minimum size for propodus
length of the claw is 2% inches. Data on growth
and natural mortality indicate that this is near the
optimum size for maximum yield per recruit.

Crab fishing is closed for 5 mo over the spawning
season. Fishermen reason that this closure yields
better recruitment to the fishery although this is
not supported by present biological data. The
eff'ort restriction does produce higher catches per
unit effort during the open season, however.

It is unlawful to fish with spears, hooks, or other
gear that might kill the crabs.

Administration

Each trap must have a buoy, and traps and boats
must be clearly marked with a permit number and
color code unique to each boat. These regulations
help in enforcement of seasons and boat registra-
tion requirements. Traps marked with buoys also
reduce their hazards to navigation. Boats must be
registered specifically for the stone crab fishery.

PROCEDURES FOR CHANGING
LAWS AND REGULATIONS

To this point no distinction has been made
between laws and regulations. Laws are passed by
an elected legislative body whereas regulations
are approved by a department's secretary or
minister, or an appointed commission. Recom-
mendations for changes in laws or regulations
usually follow the same route whether they
originate within the public service or the fishing
industry.

Alaska

Regulations are made by a seven-member Board
of Fisheries composed of fishermen and business-
men and appointed by the State Governor.
Proposed changes for regulations are submitted to
the board by the Department of Fish and Game
staff and the public at least 60 days before their
annual shellfish meeting. Thirty days before the
meeting a printed list of all proposals is sent to
fishermen, processors, government representa-
tives, and any other interested parties. During its
meeting, which is public, the board solicits com-
ments from the public and the staff of the depart-
ment on each recommendation. Following the

630

MILLER: NORTH AMERICAN CRAB FISHERIES REGULATIONS

discussion, each recommendation is voted upon by
the board in the meeting before proceeding to the
next item.

District management officers have authority to
adjust seasons and harvest levels and to open and
close fishing areas by field announcement.

British Columbia and Eastern Canada

Regulations in these areas, excluding Quebec,
are under Federal control. Proposals from any
party are submitted to the regional resource
management group who drafts regulations. These
are forwarded to a resource management group in
Ottawa who checks for consistency with existing
regulations and considers the justification offered
in light of their experience. The Justice Depart-
ment then checks for contraventions of existing
laws, especially the human rights code. It then
passes through senior management levels of the
Fisheries and Marine Service to the Minister of
State for Fisheries. If approval is granted, the
Minister finally seeks approval from the Federal
Cabinet. Regional Directors of Fisheries have
authority to adjust seasons and quotas.

Washington

The Director of the Department of Fisheries has
authority to establish many types of fishery
regulations, e.g., seasons, gear restrictions, and
size limits, after holding public hearings. The
State legislature has exclusive authority in setting
license fees and can legislate in areas normally the
responsibility of the Director.

Oregon

Staff biologists submit proposals to the Marine
Fisheries Regional Supervisor who in turn for-
wards them to the State Fisheries Director. If
approved at both these levels, proposals are sub-
mitted to a seven-member commission at a public
hearing. The commission hears staff and public
testimony and accepts, rejects, or modifies the
proposal. If accepted, it is registered with the
Secretary of State and goes in force. Any citizen of
the State may request a public hearing of the
commission to consider his views on fisheries
regulations. The commissioners are appointed by
the Governor and may be any private citizens of
the State except an officer in a sportsmen's or-
ganization or an affiliate of the commercial fishing
industry.

California

A staff biologist submits his proposed law
change to his regional manager of the Department
of Fish and Game, who in turn submits it to the
Department Director. The Director enlists the
cooperation of a State senator or representative to
sponsor a bill in the legislature where it must be
passed by both houses and signed by the Governor.
An industry representative may begin at any level
in this sequence.

Delaware

The Division of Fish and Game drafts new laws
at their own initiative or in response to requests
from the public. These drafts of new laws are
submitted to a Natural Resources Committee
composed of State legislators who in turn brings
the recommendations to the legislature for a vote.
The laws that have been passed by the legislature
are finally signed by the State Governor.

The Division of Fish and Game may also initiate
resolutions. These are not enforceable but are
desirable policy in the view of the division. Hear-
ings are held by the division to solicit public
opinion. Final approval is required from only the
Secretary of State.

Maryland

Recommendations for changes in regulations
are submitted to a Fisheries Administration
staffed by government employees. After the
legality of the submission is ensured, public hear-
ings are held by the Fisheries Administration in
areas which would be affected by the change. A
legislative board of review composed of State
legislators must finally approve changes.

Fishery laws are dealt with in the State legisla-
ture and are submitted for their consideration by
either government or private sources. A legisla-
tive committee holds public hearings on proposed
changes before they are brought to a vote in the
legislature.

North Carolina

A nine-member Fisheries Advisory Board ap-
pointed by the Governor is staffed by three repre-
sentatives each from recreational fisheries, com-
mercial fisheries, and the scientific community.
This is a source group which advises a seven-
member Fisheries Commission. The latter group.

631

FISHERY BULLETIN: VOL. 74, NO. 3

also appointed by the Governor, includes repre-
sentatives from the tourist industry, seafood
processors, sport fishermen, commercial fisher-
men, and scientists. Recommendations for
changes in regulations may be brought to either
body although only the Fisheries Commission has
authority to make changes. They are obligated to
hold public hearings on changes and advertise the
hearings in news media a minimum of 10 days
before a hearing. Following the hearings, and
solicitation of advice from the Fisheries Advisory
Board and the Division of Commercial and Sports
Fisheries, changes are passed by majority vote.

Texas

The Parks and Wildlife Department accepts
recommendations from its staff or the public. Once
each year a public hearing is held by the Parks and
Wildlife Department in each county affected by
suggested changes. A six-man Parks and Wildlife
Commission appointed by the Governor then
considers staff recommendations and the public
reaction at its monthly public meeting when
setting regulations. Laws are enacted in the State
legislature in response to requests from the
government or the public. Fisheries in 15 of 19
coastal counties are controlled by regulations
while those in the remaining 4 are controlled by
laws.

SUMMARY AND CONCLUSIONS

Management problems and applicable regula-
tions are summarized below.

1. Conservation:

a. soft-shelled crabs: season protecting soft
shells, taking soft shells prohibited;

b. protection of breeding crabs: no females, no
egg-bearing females, fishing closed during
spawning season, trap limits to control
catches of mature females, minimum size
which excludes some mature males, catch
quotas to leave a significant portion of
commercial-sized males, all stone crabs re-
turned live to water;

c. ghost fishing: traps must be attended at
least every 2 wk;

d. handling subcommercial sized crabs: escape
holes in trap mesh, minimum mesh size in
traps and trawls, fishing excluded in nursery

areas, seines must be hauled up in water
rather than on shore;

e. wasteful gear: sharp instruments, tangle
nets, and trawls excluded; and

f. optimize yield per recruit: minimumn crab
size, second season with larger minimum
size.

2. Allocation of landings among commercial

fishermen: trap limits, trap type (ring nets
only), registration area, limited entry of
boats.

3. Stability of landings: catch quotas (harvest

levels) by area.

4. Conflict over grounds or resource: areas re-

served for sport fishery; smaller size limit for
sport; limits on catch, gear type, and gear
quantity for sport; areas for traps only, set
gear only (traps or lines), or mobile gear only
(trawls or seines); traps may not be set in
navigation channels, in less than 4 feet
depth, or near beaches; limit on crab catch by
groundfish trawlers; weight limit on
dredges; seasons and area limits for crab
dredging; shrimp nursery areas closed to
crab trawling; shrimp trawling, groundfish
trawling, and scallop dragging excluded
from good crab fishing areas.

5. Processing economics: minimum crab size,

females or egg-bearing females excluded,
areas of female concentration closed, soft-
shelled crabs excluded, areas of polluted
water closed, dredging prohibited when
crabs buried in sediment.

6. Administration: registration of boats, men, and

gear; marking boats and gear with regis-
tration number; reporting fishing area,
number of trap lifts, and quantity of
landings.

Many resource managers agree that some
regulations are unsupportable on either conserva-
tion or economic grounds. This is understandable
since there is an inevitable time lag between the
collection of information and the updating of
regulations, and since groups or individuals are
sometimes able to influence regulations by weight
of authority without supporting rationale.

I recommend that copies of regulations (or at
least each new regulation) provided to en-
forcement officers, fishermen, processors, etc.,
have the rationale for each regulation as well as
the group requesting it appended. This procedure
has. the following possible benefits:

632

MILLER: NORTH AMERICAN CRAB FISHERIES REGULATIONS

1. Fishery participants would be informed as to

the benefits of the regulations, i.e., why they
are expected to observe them.

2. They could be at least partially educated to the

tools and rationale of fisheries management.

3. Providing participants with a background for

informed discussion should help to involve
them in managing their fishery.

4. Making the concerns of different vested inter-

ests public would hopefully provoke the fishing
industry, regulatory authorities, and legisla-
tors to provide reasoned support for
regulations.

ACKNOWLEDGMENTS

I thank the individuals cited in the Methods
section of the text for their generous cooperation
in supplying the bulk of the information included
in this review. William R. Beckman and I. B. Byrd
of the U.S. National Marine Fisheries Service,
NOAA, were helpful in supplying contacts and
regulations for States on the east and south coasts.

M. C. Mercer, Duane E. Phinney, and R. G. Bug-
geln constructively criticized the manuscript.

Considering the amount of detail in the sets of
regulations included and the unusual (for someone
not trained in the field) legal terminology em-
ployed, some errors are inevitable. I accept re-
sponsibility for these.

LITERATURE CITED

Cleaver, F. C.

1949. Preliminary results of the coastal crab {Cancer
magi!<fer) investigation. Wash. Dep. Fish., Biol. Rep.
49A:47-82.
Idyll, C. P.

1971. The crab that shakes hands. Natl. Geogr. 139:254-271.
Newcombe, L. L.

1945. The biology and conservation of the blue crab, Calli-
nectex sapidiis Rathbun. Va. Inst. Mar. Sci., Educ. Ser. 4,
39 p.
Rothschild, B. J., G. Powell, J. Joseph, N. J. Abramson, J. A.

Buss, AND P. ElDRIDGE.

1970. A survey of the population dynamics of king crab in
Alaska with particular reference to the Kodiak
area. Alaska Dep. Fish Game, Inf. Leafl. 147, 149 p.

633

VERTICAL DISTRIBUTION AND OTHER ASPECTS OF THE ECOLOGY OF
CERTAIN MESOPELAGIC FISHES TAKEN NEAR HAWAII

Thomas A. Clarke^ and Patricia J. Wagner-

ABSTRACT

Data on abundance, size, depth and time of capture, and sexual development are presented for 37
species of 15 families of rare to moderately abundant mesopelagic fishes taken in the central Pacific.
These exhibit a wide variety of patterns of vertical distribution and diel migration. Several undertake
migrations similar in extent to those of most myctophids and migrating stomiatoids, while others
remain at depth both day and night. In between are species where occurrence or extent of migration are
related to size. A trend for juveniles to occur shallower than adults, already noted in myctophids and
stomiatoids, is present in most species covered here regardless of migration pattern. Sexual differences
in adult size and uneven sex ratios are indicated for several species. The interplay between sexual
dimorphism, size difference, and sex ratio and the consequences to reproductive strategy are briefly
discussed.

Most ecological studies of mesoplagic fishes have
dealt primarily or exclusively with the two groups
which dominate the fauna in most parts of the
ocean— the family Myctophidae and the stomiatoid
fishes. Because other forms are generally collected
in small numbers, our knowledge of their ecology
is limited to minor parts of general reports (e.g.,
Badcock 1970) or short notes on a few new
specimens. Systematic or zoogeographic studies
have assembled data from earlier collections, but
in most cases the ecological value of such data is
limited because sampling programs were not
designed with ecological objectives in mind. Also
the gear used was in many cases undoubtedly
ineff"ective at sampling many forms and was
fished without really good knowledge of depth of
tow.

Recent, ecologically designed collections in the
central Pacific Ocean near Hawaii by our program
and that of R. E. Young have yielded a large
amount of material involving some 225-250 species
of mesopelagic fishes. Data on the myctophids and
certain stomiatoids have already been reported
(Clarke 1973, 1974), and material including many
of the remaining species, which has been passed to
other investigators, will eventually be covered in
broader reports, e.g., family revisions, etc.

'University of Hawaii, Hawaii Institute of Marine Biologv,
P.O. Box 1346, Kaneohe, HI 96744.

^University of Hawaii, Hawaii Institute of Marine Biology;
present address: University of Alaska, Institute of Marine
Science, Fairbanks, AK 99701.

Manuscript accepted March 1976.

FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

In this paper we report on a rather heterogen-
eous group of rare to moderately abundant fishes
taken in these collections. Included are represen-
tatives of several families which are present and
often moderately abundant in most parts of the
world ocean, but of which knowledge of even the
depth distribution is rather poor. Even though we
are able to consider only a few other ecological
parameters in detail for most of these, we feel that
presentation of this data contributes to a broader
understanding of the patterns of life history
exhibited by the diverse mesopelagic fauna.

MATERIALS AND METHODS

Most of the specimens considered herein were
taken near Oahu, Hawaii in a series of collections
described in detail in Clarke (1973). These included
six approximately quarterly series of extended
horizontal tows with 2-m (one series) and 3-m
Isaacs-Kidd midwater trawls (IK) and a series of
samples in the upper 250 m with the larger Cobb
Pelagic Trawl (CT). Because the program was
designed primarily for study of vertically migrat-
ing species, the upper 250 m at night and the
400-1,200 m zone by day were covered most thor-
oughly; effort in the deeper zone at night was
roughly one-fourth that by day. Thus for some of
the nonmigrating species considered here, we have
examined deep night collections made by R. E.
Young with an opening-closing Tucker Trawl
(TT). We have also examined collections from

635

FISHERY BULLETIN: VOL. 74, NO. 3

seven TT collections taken between 1,200 and 2,300
m. Finally, to present as much data as possible on
rare species, we have included more recent collec-
tions with 3- and 5-m IK, deep-day and shallow-
night TT collections, and a variety of CT and IK
tows made by the National Marine Fisheries
Service (NMFS).

The lower depth limits given are subject to
question because most of the specimens con-
sidered were taken by the IK or CT, neither of
which were equipped with opening-closing
devices. As discussed in Clarke (1973, 1974), this
problem is not great in dealing with abundant
species, but for rare species-as were many con-
sidered here-there is no basis upon which to
discriminate captures made at the principal tow-
ing depth from those made in transit to and from
towing depth. Consequently, some of our conclu-
sions about depth ranges and vertical migration
must be regarded as tentative.

Lengths of all specimens were measured to the
nearest millimeter. For each species discussed
individually, the total number examined and the
length range in millimeters are given in paren-
theses after the species name. (Unless noted as
TL-total length, SL or standard length is used.)
Gonads were examined under a dissecting
microscope to estimate size at maturity (defined
here as the smallest female with well-developed
ova), seasonal trends in gonad development, and
sex ratios. For rarely taken species, we examined
all specimens and have reported results for un-
damaged specimens large enough to be reliably
sexed by our routine technique. In species where
we examined only a fraction of the total available
material, we selected tows taken from throughout
the day and night depth range and from all
seasons and examined all individuals of the size
range of interest from these selected tows.
Hopefully this procedure minimized any potential
bias due to sexual differences in depth distribu-
tion, etc. In all cases where we discuss population
sex ratio, at least 50% (usually 70-80%) of the total
individuals of the appropriate size range were
sexed.

Numbers of specimens captured per tow were,
for all species considered here, too low to treat in a
rigorous quantitative fashion. Consequently, we
have pooled data from all seasons to estimate
depth ranges, pooled fish from all depths to con-
sider seasonal changes in abundance, etc. Since
both depth coverage and eff^ort were roughly equal
for each of the seasonal series of collections, it is

unlikely that any serious bias resulted from our
procedures.

Specimens of all species considered here will be
deposited in the National Museum of Natural
History, Washington, D.C.

Table 1. -Lengths and capture data for 11 species taken near
Hawaii (lat. 22°20-30'N; long. 158°20-.30'W) and two specimens
(*) taken in the central equatorial Pacific (lat. 3°30'N; long.
145°W). Total length is given for Isistius and Snijderidia;
standard length for the others. For horizontal tows, the most
frequently fished depth is given; oblique tows are noted by
0-ma.ximum depth. Catches by opening-closing trawl are noted
by (OC).

Length (mm)

; depth (m)

Family, species

Night (2000-0500)

Day (0700-1800)

Squalidae:

Isistius

brasiiiensis

355; 70
398; 0-500
492; 170

Argentinidae:

Microstoma

microstoma

89; 490

Nansenia sp.

85; 525 (OC)

46; 560
73; 0-1,100
86; 725
107; 620

*Xenophthalmich-

thys danae

ca. 75; 300

Opisthoproctidae:

Rtiynchotiyalus

natalensis

61; 600

80.0-1,100

156; 0-400

85; 0-800
107; 530 (OC)

*Winteria telescopa

41 ; 450

Rondeletiidae:

Rondeletia

loricata

15; 190

30; 800

21; 200

33; 1,150

28; 175

47,69 (2); 775

38; 450

86; 0-1,900

85; 100

52; 925

Barbourisiidae:

Barbourisia rufa

125; 800
276; 750
314; 0-1,200

67; 750

Zoarcidae:

Synderidla canina

176; 0-350

79; 0-1,160

188; 150

188,247 (2); 800

227; 500

ca. 240 (3); 0-500

Gempylidae:

Lepidocybium

flavobrunneum

15; 200
29; 0-350

Nesiarchus nasutus

22; 370
67; 170

Scombrolabrax

heterolepis

10; 190

18,26 (2); 800

26; 25

27; 1,000

26 (2); 250

27,29 (2); 750

Trichiuridae:

Aplianopus carbo

19; 170

197; 1,100

40; 150

252; 660

59; 190

636

CLARKE and WAGNER: VERTICAL DISTRIBUTION OF MESOPELAGIC FISHES

RESULTS

Table 1 gives length and capture data for 11
rare and sporadically taken species from our
collections near Hawaii. We also have included in
Table 1 capture data for two rather infrequently
collected species of argentinoid fishes which were
taken by a series of IK trawls in the central
equatorial Pacific (cruise 47 of the NMFS RV
Townsend Cromwell). Other species are considered
under individual headings below.

Opisthoproctidae

Opisthoproctus soleatus Vaillant (150; 28-84 mm)

Almost all 0. soleatus were taken at 450-600 m
during the day; highest catch rates were at 500-550
m. Large (>50mm) fish were caught throughout the
day depth range, but smaller fish were taken
mostly above 550 m. Only six specimens were
taken at night-also in the same depth range. The
daymight ratio of total trawling time in this depth
range was about 4:1; thus the difference in catch is
only partially explained by differences in effort.
Since the night catches did not indicate that 0.
soleatus is spread more thinly over a broader depth
range, the difference in catch per effort indicates
that this species avoids the net better at night.

Female 0. soleatus mature at about 60 mm. Data
for each season were few, but there was no in-
dication of seasonality in gonad ripeness, size
composition, or abundance.

Opisthoproctus sp. (3; 11-17 mm)

A 17-mm specimen, tentatively identified as 0.
grimaldii Zugmayer, was taken in a day tow at
500 m in September. Two smaller specimens (11
and 15 mm) taken in June are apparently 0.
soleatus. One was taken at night in an oblique tow
from 0 to 350 m; the other was caught, possibly in
transit, by a day tow which fished at 725 m.

Alepocephalidae

Photostylus pycnopterus Beebe (12; 62-1 13 mm)

Photostylus pycnopterus was taken within the
same depth range day and night. Five day catches
were at 750-975 m, and the two night catches at 750
and 875 m. Five other day catches were from
oblique tows which fished to 800-1,000 m.

Photostylus pyncopterus appears to mature at
about 100 mm and to spawn relatively few but
large eggs. Goodyear (1969) recorded a 93-mm
female with eggs 1.4 mm in diameter and two
specimens (84 and 96 mm) with much smaller eggs.
Our three largest females (101-113 mm) carried
eggs about 1.75 mm in diameter. One undamaged
specimen had only 80 eggs in the ovaries. Another
apparently had spawned some already; there were
26 eggs— mostly in the anterior sections of the
ovaries. The gonads of two large males (106 and
110 mm) filled most of the body cavity. The
remaining specimens (62-89 mm) were clearly
immature.

The eggs of P. pycnopterus, both absolutely and
relative to body size, were larger than those of any
other species examined from our collections. Mead
et al. (1964) have pointed out that other species of
Alepocephalidae also have large eggs.

Giganturidae

Bathyleptus lisae Walters (89; 49-195 mm)

Although a few B. lisae were caught as shallow
as 500 m, the majority were taken at 750-1,000 m
both day and night. Of the 70 specimens taken in
horizontal tows, only 7 were taken above this
range and 3 deeper. There was no apparent trend
in size with depth.

Female B. lisae appear to reach much greater
size than males. Of 26 fish sexed, there were 14
females of all sizes (67-195 mm) and 12 males-all
between 63-81 mm. All nine specimens over 81 mm
were females. Of these, only one (171 mm) ap-
peared mature.

Eurypharyngidae

Eurypharynx pelecanoides Vaillant
(34; 89-575 mm TL)

Except for two day captures of small individuals
(126 and 155 mm at 425 and 550 m, respectively) E.
pelecanoides was taken between 650 and 1,300 m.
Twenty-five specimens were taken during the day
within this range. Of the 14 less than 300 mm, only
2 were taken below 1,000 m, and all over 300 mm
were taken below 1,000 m. Thus, the small fish
appear to occur shallower than the large ones.
There were only seven night catches in horizontal
tows, but these agreed with the size-depth pattern
apparent in the day data.

637

FISHERY Bl'LLKTIN: VOL. 74. NO. :i

Among 19 specimens that could be sexed, there
were 8 males and 11 females. The seven largest
fish included two males (440 and 507 mm), three
apparently mature females (ca. 490-575 mm), and
two nearly mature females (464 and 494 mm).

Bregmacerotidae

Two species of the genus Bregmaceros were
taken in our collections. One fits published data
(D'Ancona and Cavinato 1965) on B. japonicus
Tanaka reasonably well, while the other is closest
to but not identical with B. macdellandi Thomp-
son. The latter is apparently distinct from another
form similar to B. maccleUandi which has been
taken in the southern Pacific (E. H. Ahlstrom and
J. E. Fitch, pers. commun.). The exact identity of
all of these must await a thorough review of this
badly confused genus.

The two Hawaiian forms were, however, quite
distinct from each other. Dorsal and pectoral ray
counts were 56-62 and 19-21 for B. mocclelknidi vs.
50-54 and 17-19 (rarely 20 or 21), respectively, for
B. japouicus. The latter was the more slender
species with SL/greatest body depth of 7.3-10.0 vs.
6.5-7.3 in B. maccleUandi. Bregmaceros japonicus
adults were distinctly darker dorsally, while B.
maccleUandi were not countershaded. The isth-
mus and pelvic fins of all larger B. maccleUandi
were grey, while in juveniles (<25-30 mm), the
isthmus was covered with small melanophores. In
most B. japonicus the isthmus and pelvics were
totally unpigmented. A few small (ca. 20-25 mm)
specimens whose counts fit B. japonicus had a few
large melanophores on the isthmus.

Bregmaceros japonicus T^iniki (284; 18-52 mm)

The great majority of B. japonicus were taken
at 25-125 m at night; however, 40 specimens,
possibly contaminants, were taken at 125-200 m.
Those under 30 mm were taken mostly above 100
m, while larger individuals were taken with
roughly equal frequency throughout the 25- to
125-m range. Only 32 specimens were taken dur-
ing the day; most (25) were large individuals (>35
mm) and taken at 600-800 m. This suggests that
during the day the juveniles may occur shallower
than the upper limit of our day samples (ca. 300 m).

Female B. japonicus appear to mature at about
40 mm, and almost all specimens over this size
carried well-developed ova at all seasons. Small

fish (<30 mm) were most abundant in March; they
made up about 50^?^ of the catch then as opposed to
less than 10% at other seasons.

Bregmaceros cf. maccleUandi Thompson
(274; 14-94 mm)

Bregmaceros maccleUandi occurred between 100
and 250 m at night. Most individuals less than 30
mm were caught about 150 m, and most 30-50 mm
above 175 m, but larger fish were taken with
roughly equal frequency throughout the night
depth range. Day catches were mostly between
600 and 1,000 m with those less than 30 mm
occurring above 800 m. Seven specimens (65-80
mm) were taken in tows that fished between 1,200
and 1,400 m; three of these were from an open-
ing-closing trawl.

Bregmaceros maccleUandi over about 35 mm
appear to avoid the IK. Of the total specimens, 152
were taken by the CT in March 1971. Of these only
about 12% were less than 35 mm, whereas, about
half of the IK specimens were less than 35 mm for
either the March data alone (12/23) or the total IK
collection (56/122).

Females mature at about 60 mm. There were so
few mature females in most series that no trends
in gonad ripeness could be ascertained. The size
composition of the catch showed no obvious sea-
sonal changes.

Melamphaidae

Scopelogadus mizolepis mizolepis (Giinther)
(201; 7-74 mm)

Ebeling and Weed (1963, 1973) concluded from
their data that S. m izolepis does not undertake diel
vertical migrations and gave the upper depth limit
of "adults" (66-94 mm) as 500 m. Our data, in
contrast, clearly indicate that 5. mizolepis of all
sizes undertake a definite vertical migration.
During the day, S. mizolepis occurred between 600
and 1,000 m and possibly deeper (the few tows
below 1,000 m do not allow us to guess whether IK
catches there were made in transit). Most of the
fish less than 25 mm were taken between 600 and
800 m, and most larger ones at 700-1,000 m. At
night the smallest fish occurred at 100-180 m, those
25-50 mm mostly at 150-250 m, and the larger ones
at 200-400 m. There were no night catches between
400 and 600 m, but several specimens of all sizes

638

CLARKE and WAGNER: VERTICAL DLSTRIBUTION OF MESOPELAGIC FLSHES

were taken at night within the daytime depth
range, suggesting that a small fraction of the
population does not migrate.

We examined the gonads of 127 specimens. Of
39 females (19-74 mm), those less than 50 mm were
clearly immature, 2 56-mm fish were nearly ma-
ture, and 8 of the 10 largest (57-74 mm) carried
well-developed ova. The 88 males were 18-60 mm.

There were too few mature females to consider
any seasonal trends in ripeness. Juveniles (7-12
mm) were taken in March, June, and September,
and made up the largest fraction of the catch (59*%)
in March. There were other peaks in size-
frequency distributions at all seasons, but none
could be clearly traced from season to season.

Poromitra crassiceps (Gunther) (57; 16-130 mm)

All sizes of P. crassiceps occurred shallower at
night than during the day. Day catches were
between 750 and 1,000-1,200 m. No specimens over
60 mm were caught above 900 m. At night, two
small fish were caught near the day depth, the
remaining small fish (19-51 mm) at 150-400 m, and
the larger fish (84-128 mm) between 340 and 825 m.

The seven smallest fish (16-25 mm) were taken
in March, June, or July, and 16 intermediate-sized
individuals (27-35 mm) were all taken in Sep-
tember. The others (39-130 mm) were scattered
seasonally. Twenty-four specimens were 80 mm or
larger. Nineteen of these (80-101 mm) were males;
several of those over 90 mm appeared, subjective-
ly, to be mature or nearly so. The five females were
97-130 mm, and none were mature.

Poromitra mega/ops (Liitken) (56; 13-41 mm)

All but one P. megalops were either 13-21 mm or
28-41 mm. Four of the small fish were caught at
625-1,000 m during the day. At night, five were
taken at 250-380 m. and five at 690-775 m. Of
the large fish, 27 were taken at 725-1,000 m during
the day and 13 at 640-850 m at night. Thus some of
the small fish undertake a fairly substantial up-
ward migration at night, but the large fish appear
to shift upwards only slightly, if at all. There were
no obvious seasonal trends in size composition of
the catches; specimens of both size groups were
present at all seasons.

Of the 34 specimens sexed, there were 18
females (26-41 mm) and 16 males (28-39 mm). The 5
smallest females (26-35 mm) were immature,
while the 13 large ones (37-41 mm) appeared

mature. Ebeling and Weed (1973) reported the size
range of mature P. megalops as 45-62 mm. Possi-
bly, P. megalops matures at a smaller size in
certain parts of its range. (Ebeling and Weed did
not give specific geographic data for their mature
specimens.)

Poromitra oscitans Ebeling (19; 44-71 mm)

Poromitra oscitans is a deep-living, nonmigrat-
ing species (Ebeling 1975). It occurred only at the
lower edge of the depth range sampled in detail.
One specimen each was taken at 750 and 850 m; the
others were caught in nine tows all of which fished
below 1,000 m. Four of these were taken in open-
ing-closing TT tows which fished only below 1,350
m. Three were males (44-53 mm), and the others,
immature females (45-71 mm).

Scopeloberyx opisthopterus (Parr) (93; 14-38 mm)

Scopelohergx opisthopterus was taken between
540 and 1,200 m during the day. Night catches by
the IK were at 650-1,175 m, and one specimen was
taken by the TT fished open between 1,300 and
1,450 m. There was thus no evidence of any diel
change in depth range. Most small specimens (<25
mm) were taken above 800 m and most large ones
below 750 m.

Out of 55 specimens (25-38 mm) sexed, there
were 10 immature females (26-30 mm), 24 mature
females (31-38 mm), 7 males (27-33 mm), and 4
(25-29 mm) that were too small to sex with cer-
tainty but which were probably males. Mature
females were taken at all seasons except
December (when only four 5. opisthopterus were
taken). There were two rough size groups in the
catch; all but seven specimens were either 14-20
mm or over 26 mm. Representatives of the smaller
group were absent from samples taken in July and
nearly absent in June, suggesting possible sea-
sonality in recruitment.

Scopeloberyx robustus {Gunther) (120; 12-31 mm)

Scopeloberyx robustus was taken at 550-1,200 m
during the day. With the exception of three small
(14-20 mm) specimens taken at 340-425 m, the
night depth range was similar-600-1,175 m. Thus
there is no indication that any but the small S.
robustus vertically migrate. There was a distinct
increase in size with depth. With few exceptions,
fish less than 20 mm were caught above 800 m,

639

FISHERY BULLETIN: VOL. 74, NO. 3

those 20-25 mm at 750-1,000 m, and those over 25
mm below 900 m.

Of 41 specimens (21-31 mm) sexed, 24 were
females of which 6 (29-31 mm) were mature. The
17 males were 22-30 mm. Fish less than 16 mm
were taken only in July and September. There
were no seasonal trends in abundance of the larger
fish.

Melatuphaes danae Ebeling (627; 1 1-22 mm)

During the day, M. danac occurred principally at
750-1,200 m; a few were taken as shallow as 650 m.
Fish less than 15 mm were almost all taken above
1,000 m, while larger ones occurred throughout the
day depth range. Most night captures were
between 75 and 200 m; the small fish were mostly
taken at 75-100 m, while the larger ones occurred
throughout the depth range. There were no night
captures between 400 and 650 m, but 27 specimens
of all sizes were taken at night in tows that fished
within the day depth range. Although this catch
was numerically small, and possibly due to in
transit captures, it was large enough relative to
effort to suggest that a small, but not insignificant
fraction of the population did not regularly
migrate.

Female M. danae matured at about 17-18 mm.
Mature females were present in comparable
numbers and percentages at ail seasons, but the
size composition of the catches indicated that
juveniles were recruited primarily in the spring
and early summer. For the series where the proper
depth ranges were adequately and roughly
equivalently sampled, the small (11-14 mm) fish
made up 27% of the total catch in March, 15% in
June, and 42% in July as opposed to 1% in Sep-
tember and 2.5% in December.

Melamphaes sitnus Ebeling (4; 14-24 mm)

Data on M. f^iinus indicate little more than that
it is present in low abundance in the area. The two
night captures were at 300 and 800 m, while the
two day captures were at ca. 700-800 m (the latter
depths are estimates from wire out; depth records
for both day tows were invalid).

Melamphaes indicus Ebeling (20; 16-5 5 mm)

Eleven M. indicus were taken at night at
125-150 m-nine of these in one tow. Nine
specimens were taken during the day at 640-900 m.

Two large females (51 and 55 mm) were mature,
and four males (47-53 mm) appeared mature or
nearly so.

Melamphaes sp. (janael Ebeling) (10; 17-54 mm)

Seven M. "janae" were taken at night at 190-250
m and three during the day between 650 and 900
m. All were taken in September or November. The
two largest specimens (43 and 54 mm) were both
males and larger than the ma.ximum size of this
species given by Ebeling (1962). Ebeling, however,
did note geographic differences in size at maturity.
Our specimens fit the description of M. janae in
other respects and could be reliably distinguished
from similar-sized individuals of M. iyidicus.
Study of more specimens will be necessary to
determine whether M. janae is more variable in
size than Ebeling noted or more than one species is
involved.

Melamphaes sp. {lotigh'elis? Parr) (2; 18, 20 mm)

Two small specimens of the "typhlops" group
are tentatively identified as M. longivelis. The
smaller was taken at 625 m at night, the larger at
640 by day.

Melamphaes po/ylep/s Eheling (10; 17-57 mm)

One M. pohjiepis was taken at night at 930 m;
the remainder were taken during the day at
640-1,150 m. They included two mature females (56
and 57 mm), six males (46-56 mm), and two
juveniles (17 and 19 mm).

Anoplogasteridae

Auoplogaster cornuta (Valenciennes)
(93; 3-126 mm)

Juvenile A. cornuta undertake a substantial
upward migration at night. At least some of the
large fish also move upwards, but occur deeper
than the juveniles both day and night. Seventy-
two specimens were small (3-24 mm) and were
taken in February-March. Fifty-eight were taken
at night between 135 and 185 m; the remaining 14
were taken during the day- 12 at 650 m and 2 at ca.
800 m. Larger specimens (all >70 mm) were taken
throughout the year. At night six (77-87 mm) were
taken between 275 and 475 m, one each at ca. 600 m
(108 mm), 900 m (109 mm), and in an oblique tow to

640

CLARKE and WAGNER: VERTICAL DISTRIBUTION OF MESOPELAGIC FISHES

980 m (94 mm). The 12 large (77-126 mm) in-
dividuals taken during the day were from
750-1,150 m.

Among the 18 fish sexed, there were 12 males
(80-109 mm) and 6 females (78-126 mm). None
appeared mature. The collection of so many small
individuals in one of the seasonal series indicates
that A. cornuta has a rather short spawning
season.

Stylephoridae

Stylephorus chordatus Shaw ( 19; ca. 60-31 5 mm)

Seven S. chordatus (ca. 60-315 mm) were taken
at night between 300 and 600 m. Eleven (63-282
mm) were taken between 625 and 800 m during the
day, and one at dusk at 500 m. Thus S. chordatus
appear to migrate about 200-300 m upward at
night.

Two females (282 and 315 mm) appeared ma-
ture; the next largest female was 147 mm. The four
largest males were 235-243 mm.

Gempylidae

Gempylus serpens Cuvier (29; 7-148 mm)

All but two G. serpens were taken at night in the
upper 250 m; 19 were from 30-100 m. During the
day, a 60-mm specimen was taken at 450 m and a
30-mm one at 800 m. It seems likely that the latter
or both of the day catches were made in transit
and that G. serpens migrates downward only a
short distance, if at all, during the day. None were
near maturity.

Nealotus tripes Johnson (95; 7-173 mm)

Most A'', tripes were small (9-41 mm) taken at
50-200 m at night. Seventy-three were taken in
December-58 in three tows at 170-200 m and 12 in
a tow at 250 m. The CT collected four large
specimens at night, three (75, 168, 173 mm) at 100
m and one (68 mm) at 250 m, while only one (49 mm
at 150 m) was taken by the IK. No small fish and
only three large ones were taken during the day.
The CT captured a 49-mm individual at ca. 350 m,
and the IK took two (63, 105 mm) in separate tows
at ca. 750 m. The small N. tripes apparently stay in
the upper layers both day and night. Since the
larger fish were obviously inadequately sampled
by the IK and there were no deep day tows made

with the CT, it is not clear whether adults descend
or not. The two deep day catches by the IK may
well have been coincidentally taken in transit by
tows which fished the same depth.

Diplospinus multistriatus Maul (224; 8-239 mm)

Most of the D. multistriatus were small in-
dividuals caught at night at two depth zones and at
two separate seasons. Of the 100 specimens taken
in December, 78 (8-30 mm) were taken in three
tows at 170-200 m. In July, 62 specimens were
taken; 31 (7-18 mm) were from four tows at 100-110
m. Other small fish taken at night were mostly
from the upper 200 m with a few, probably cap-
tured in transit, taken in deeper tows. Of the 18
larger (35-239 mm) specimens taken at night, 13
were taken in the upper 130 m, 4 at 200-300 m, and
1 probably captured in transit, at 500 m.

Only 37 were taken in day tows, all but 2
between 500 and 1,000 m. Only one of these (11
mm) was in the size range which dominated the
night catches. Three specimens were slightly
larger (36-42 mm) and the remainder 68-221 mm.
Most less than 140 mm were taken above 800 m,
and most over 140 mm were taken below 700 m.

The near absence of small D. multistriatus in
the day samples suggests that they either remain
in the upper layers during the day (and were not
sampled by our tows) or occur deeper and avoid the
net during the day. The latter seems improbable
for such small fish. Assuming the former is true
and considering the data on larger fish, it appears
that D. multistriatus occurs in the upper 100-200
m at night and that the larger sizes migrate to
progressively greater depths by day.

Of the 46 specimens sexed, 12 were males (93-207
mm) and 34 females (75-239 mm). Eight females
(163-239 mm) were mature.

DISCUSSION

Vertical Distribution and Migration

The diverse group of fishes considered here, as
might be expected, exhibit a greater array of ver-
tical distribution patterns than the myctophids
and stomiatoids which occur in the study area.
Most species of the latter two groups undertake
substantial diel vertical migrations. The remain-
ing species do not vertically migrate at all. Among
the species considered here, migrators and non-
migrators are about equally represented, and

641

FISHERY BULLETIN: VOL. 74, NO. 3

almost every conceivable intermediate pattern is
represented as well.

Four of the common species, Bregmaceros
japoHicKs, B. macclelland} , Scopelogadus
mizolepis, and Melaniphaes danae, are typical
migrators. Both juveniles and adults move from
well below 500 m during the day into the upper 250
m at night. The data indicate that four rarer
species, Rondeletia loricata, Melaniphaes indicus,
M. "janae" and Scombrolahrax heterolepis proba-
bly perform similar migrations.

The first four species were the most abundant of
all considered here and ranked with all but the 8-10
most abundant myctophids and migrating
stomiatoids (see Clarke 1973, 1974). The night
size-depth patterns of the four were similar to the
general types observed in the latter groups. Breg-
maceros japouicKs cooccurred with similar-sized
individuals of several abundant myctophid species
and Vinciguerria nimharia, whWe B. macclellandi
and the melamphaids had patterns similar to those
of deeper-living species, e.g., Lanipauycfus niger
and Gonostoma spp. In the case of the Bregmaceros
spp. and M. danae, the adults occurred throughout
the depth range instead of primarily at the lower
end as was usually the case with the other fishes.

During the day, the four migrating species
exhibited a trend for increased size with depth.
The day depth range of B.japon reus was similar to
those of many other migrating species, but the
other three were the only migrating species be-
sides the myctophid Lampangctus nobilis whose
day depth range extended well below 1,000 m.
La})ipa}n/ct}is nobilis, B. macclellandi, and
Scopelogadus mizolepis are relatively large
species, but M. danae is one of the smallest species
of fishes encountered in our study area.

The species for which there was no indication of
diel change in vertical distribution are a rather
heterogeneous group. Opisthoproctiis soleatus
inhabited a relatively shallow depth range and
cooccurred with several stomiatoid species with
similar, and probably convergent, morphological
features (see Clarke 1974). Other nonmigrating
species {Scopeloberyx spp.; Poromitra oscitans,
Photostylus pycnopterus, Bathyleptus lisae, Eury-
pharynx pelecanoides, and probably Barbourisia
rnfa) occurred mostly below 600 m. Many of these
species are commonly referred to with the too
casually used adjective "bathypelagic," which has
the connotation (if not always the denotation) of
extremely great depths well removed from direct

influences of surface phenomena. Our data in-
dicate that most of these should more properly be
considered members of the mesopelagic commun-
ity. Even taking into account the relatively few
hours of sampling below 1,000-1,200 m, the only
species which appear to occur in any abundance
below this depth are Poromitra oscitans and E.
pelecanoides (of course, other fishes not covered
here do occur deeper and some, e.g., certain cera-
tioidsand the eel Cyema appear to occur only below
1,000-1,200 m). In "fact, B. lisae, the Scopeloberyx
spp., and probably all the others except P. oscitans
have their primary centers of abundance above
1,000-1,200 m. During the day, they cooccur and
presumably interact with vertically migrating
species. Thus at least some aspects of their ecology
must be aff'ected by diel light changes.

Four species showed limited diel changes in
depth distribution. Sfylephorvs chordatus moved
somewhat shallower at night, but did not occur in
the upper 250 m. Juveniles of P. crassiceps and
Anoplogaster cornnta undertook fairly substantial
upward migrations at night, but the adults shifted
only slightly shallower. Juvenile P. megalops
occurred somewhat shallower at night, but there
was no conclusive evidence that the adults moved
at all. Juvenile Scopeloberyx robustus, considered
a "nonmigrator" above, may also move up at
night. Since only P. crassiceps and P. megalops
were taken in even moderate numbers, the pat-
terns for the other species must be regarded as
tentative. Size-related diff"erences in migration
have been noted for some myctophids and stomia-
toids (Clarke 1973, 1974). As examples, the adults
of Bolinichthys distofax (identified as B. superla-
teralis in Clarke 1973) appear not to migrate while
the juveniles do, and the larger individuals of
Gonostoma elongatum appear to occasionally
remain at depth during the night.

Interpretation of data on the gempylid-tri-
chiurid species is limited because, with the excep-
tion of Diplospinus, only the small juveniles were
collected, and even these either avoided the net
during the day or occurred so shallow that they
were not sampled by our program during the day.
The data indicate that all sizes of Gempylus
serpens (to 148 mm) and Nealotus tripes (to 173
mm) collected probably remain in the upper layers
during the day. Although the deep day catches of
small Scombrolabrax heterolepis may have been
made in transit, the absence of this species from
day tows above 750 m suggests that it may mi-

642

CLARKE and WAGNER: VERTICAL DISTRIBUTION OF MESOPELAGIC FISHES

grate. Diplospinus mnlfistriatiis exhibited a
pattern opposite to that of P. crassiceps; the small
fish either remain in the upper layers or descend
only slightly during the day while the larger
juveniles (>ca. 60 mm) and adults undertake a
substantial migration.

Avoidance

With the exception of the gempylid-trichiurid
species, there were few obvious indications of
sampling error due to avoidance, but in most cases
data were too few to even discuss the subject. The
failure to capture mature specimens of the two
large Poroniitra spp. indicates avoidance by these
and probably a fraction of the populations of
other large melamphaids. Bregmaceros japonicus
was apparently undersampled during the day, and
the large B. macclellandi were sampled better by
the CT than by the IK. It is not unexpected that
avoidance was indicated for the larger, more
"solidly built" species rather than for small species
such as M. danae and the Scopeloheryx spp. or

species such as Bafhi/h'pti(s lime and Euryphar-
ipix pek'canoiiles which do not appear "designed"
for swimming ability. The most puzzling indica-
tion of avoidance was that suggested for OpintJio-
proftus mleatus. This species not only has few
characteristics indicating swimming prowess, but
was undersampled at night rather than during the
day as one might expect if vision were involved.

Sexual Dimorphism and Sex Ratio

In several species, the males appeared to be
smaller than females (Table 2). The extreme case
was Bafhyleptus lisae where the largest female
was about 2.5 times longer than the largest male.
In Scopelogadus mizolepis, the females mature at
about the size of the largest males observed and
reach somewhat larger maximum size. A similar
trend is suggested by the data for Anoplogasfer
cornuta and two other large melamphaids, Po-
romitra crassiceps and P. oscitans, but the
numbers involved are too small to confirm it here.
In two smaller species of the same family, P.

Table 2.-Summary of data on sex ratio and sexual differences in size for 10 species of fishes.
Under Population sex ratio and left hand column gives the number and size ranges of all males
in the population with 95"? confidence limits (read to the nearest 0.01 from Chart 3 in Tate and
Clelland 19.57). Sex ratio was considered significantly different from 1:1 if these limits did not
cross 0.50. Under Size Difference similar figures are given for only those specimens larger than
the smallest mature female (since all Melamphaex danae were as large or larger than the
smallest mature female, the data are the same for both pairs of columns). For Bathylepiut^ linae,
where the smallest mature female was much larger than the largest male, and for three
melamphaids, where no mature females were taken, we have given only the number and size
range of females larger than the largest male.

Population sex ratio

Size difference

No. examined

Proportion

No. examined

Proportion

(Size range,

of males

(Size range,

of males

Species

mm)

(95% limits)
0.46

mm)
99 (90-195)

(95% limits)

Bathyleptus lisae

12(? (63-81)

—

149 (67-195)

(0.26-0.67)

Scopelogadus mizolepis

88 <? (18-60)

0.69

2 S (60)

0.17

399 (19-74)

(0.60-0.77)

109 (57-74)

(0.02-0.47)

Poromitra crassiceps

19<? (80-101)
59 (97-130)

0.79
(0.57-0.96)

49 (121-130)

Poromitra megalops

16<J (26-39)

0.47

3(? (37-39)

0.19

189 (28-41)

(0.29-0.66)

13 9 (37-41)

(0.04-0.47)

Poromitra oscitar)s

3(? (44-53)
169 (45-71)

0.16
(0.03-0.41)

109 (53-71)

Scopeloheryx opisthopterus

21 (? (25-33)

0.38

8<? (31-33)

0.25

349 (26-38)

(0.25-0.53)

249 (31-38)

(0.11-0.45)

Scopeloberyx robustus

17<J (22-30)

0.41

4<? (29-30)

0.40

249 (21-31)

(0.24-0.58)

69 (29-31)

(0.12-0.70)

Melamphaes danae

282^ (17-22)
1449 (17-22)

0.65
(0.60-0.70)

Anoplogasfer cornuta

12^ (80-109)
69 (78-126)

0.67
(0.41-0.87)

39 (110-126)

Diplospinus multistriatus

12(? (93-207)

0.26

3<? (184-207)

0.23

349 (75-239)

(0.13-0.42)

109 (163-239)

(0.05-0.55)

643

FISHERY BULLETIN: VOL. 74, NO. 3

megalops and Scopeloberi/x opisthopterus, the
maximum size of females was only slightly greater
than that of males, but there were relatively few
males larger than the smallest mature female.
Thus there appears to be a slight but real differ-
ence in size of the two sexes. The two smallest
melamphaids, Mehmphaet^ danae and S. rohustus
showed no sexual differences in size. In all cases,
except P. crassiceps and P. osciiam^, there were
sufficient small females to indicate that the size
differences were not due to protandrous
hermaphroditism.

Sexual differences in size have been reported for
many species of dioecious mesopelagic fishes.
Large differences comparable to that observed in
Bafhyleptus occur in the ceratioid anglerfishes
(Bertelsen 1951) and the stomiatoid Idiacauthus
(Gibbs 1964). Differences of the order observed for
some of the melamphaids are known for several
stomiatoids: Stomias (Gibbs 1969), Echiosfoma
(Krueger and Gibbs 1966), and Cyclothone
(Kobayashi 1973). The usual explanation of the
adaptive significance of smaller males (Marshall
1971) is that in a food limited environment-such
as the deep-sea probably is-the energy required
by the population is lowered without diminished
fecundity.

Sexual dimorphism (as opposed to differences
only in size) is quite common among meso- and
bathypelagic fishes. Males of several groups ex-
hibit better developed swimming muscles or sen-
sory apparatus than the females (Marshall 1971).
In many myctophids and stomiatoids, there are
sexual differences in light organs. In most cases,
sexual dimorphism seems related to increasing
reproductive success by increasing the probability
of heterosexual encounter.

No obvious external sexual dimorphism was
observed in any of the species considered here
(with the exception of Isisfius), but at least two
species appear to have uneven sex ratios (Table 2),
an adaptation which, like the dimorphisms noted
above, serves to increase the probability of a
female meeting a conspecific male. Actually, the
sex ratios of five species were significantly differ-
ent from 1:1, and Anoplogaster cornnia showed a
nearly significant trend. However, it seems wise to
view the estimates for Poromitra crassiceps, P.
oscita7is, and Diplospinus nudtistriatits with
suspicion since numbers were rather low and
biases due to inadequate sampling and avoidance
may be involved.

For both Melamphaes danae and Scopelagadus
mizolepis, the numbers involved are relatively
high and there is no indication that the popula-
tions were not adequately sampled. The estimated
sex ratios for these two species indicate that the
probability of an individual female encountering a
male is about twice that expected for a population
with the same density of females and 1:1 sex ratio.
(The probability of an individual male encounter-
ing a female is lowered, but this has no con-
sequences to population reproductive success.) In
the case of M. da nae, where the sexes are the same
size, population fecundity would be less than that
of a population of equal total biomass and 1:1 sex
ratio because about two-thirds of the biomass are
males. The males of S. mizolepis are, however,
smaller than the females. Consequently, the effect
of uneven sex ratio on population fecundity is to
some extent balanced by the more nearly even
division of population biomass between males and
females. Better data on stages of matu-
rity-particularly for males-and size distri-
bution of mature fish of each sex would be
needed to quantitatively describe the "trade off"
between uneven sex ratio and sexual size
difference.

The difference between M. danae and S.
mizolepis may simply be due to the fact that M.
danae is already a "dwarf" species-the smallest at
maturity of all mesopelagic fishes in our collec-
tions. There may be other factors which select
against the males being smaller than the already
tiny females. On the other hand, M. danae may in
some sense be less "food-limited" than S. m izolepis
and thus able as a population to afford having
two-thirds of the biomass as males. Further study
of the interplay between sexual dimorphism,
differences in size, and departure of sex ratio from
1:1 might prove to be a fruitful approach toward
understanding the diverse life history features
shown by mesopelagic fishes.

ACKNOWLEDGMENTS

We are indebted to the captain and crew of the
RV Teritu and to the many people who assisted in
collecting and rough-sorting the material. We also
thank A. W. Ebeling for confirming our iden-
tifications of representatives of each of the
melamphaid species considered here. This re-
search was supported by NSF GB23931, NSF
GA38423, and funds from the University of
Hawaii, Hawaii Institute of Marine Biology.

644

CLARKE and WAGNER: VERTICAL DISTRIBUTION OF MESOPELAGIC FISHES

LITERATURE CITED

Badcock, J.

1970. The vertical distribution of mesopelagic fishes collect-
ed on the Sond cruise. J. Mar. Biol. Assoc. U.K.
50:1001-1044.
Bertelsen, E.

1951. The ceratioid fishes: Ontogeny, taxonomy, distribu-
tion, and biology. Dana Rep. Carlsberg Found. 39,
276 p.
Clarke, T. A.

1973. Some aspects of the ecology of lanternfishes (Mycto-
phidae) in the Pacific Ocean near Hawaii. Fish. Bull., U.S.
71:401-434.

1974. Some aspects of the ecology of stomiatoid fishes in the
Pacific Ocean near Hawaii. Fish. Bull., U.S. 72:337-351.

D'Ancona, U., and G. Cavinato.

1965. The fishes of the family Bregmacerotidae. Dana Rep.
Carlsberg Found. 64, 92 p.
Ebeling, a. W.

1962. Melamphaidae. I. Systematics and zoogeography of
the species in the bathypelagic fish genus Melamphaes
Gijnther. Dana Rep. Carlsberg Found. 58, 164 p.

1975. A new Indo-Pacific bathypelagic-fish species of Po-
romitra and a key to the genus. Copeia 1975:306-315.

Ebeling, A. W., and W. H. Weed III.

1963. Melamphaidae. III. Systematics and distribution of
the species in the bathypelagic fish genus Scopelogadus
Vaillant. Dana Rep. Carlsberg Found. 60, 58 p.

1973. Order Xenoberyces (Stephanoberyciformes). In

Fishes of the western North Atlantic. Part Six, p.
397-478. Mem. Sears Found. Mar. Res., Yale Univ. 1.
GiBBS, R. H., Jr.

1964. Family Idiacanthidae. /» Fishes of the western North
Atlantic. Part Four, p. 512-522. Mem. Sears Found. Mar.
Res., Yale Univ. 1.
1969. Taxonomy, sexual dimorphism, vertical distribution,
and evolutionary zoogeography of the bathypelagic fish
genus Stomias (Stomiatidae). Smithson. Contrib. Zool. 31,
25 p.
Goodyear, R. H.

1969. Records of the alepocephalid fish Photoatylus pycnop-
teruft in the Indian and Pacific oceans. Copeia
1969:398-400.

KOBAYASHI, B. N.

1973. Systematics, zoogeography, and aspects of the biology
of the bathypelagic fish genus Cyclothone in the Pacific
Ocean. Ph.D. Thesis, Univ. California, San Diego, 487 p.
Krueger, W. H., AND R. H. Gibbs, Jr.

1966. Growth changes and sexual dimorphism in the
stomiatoid fish Echiosfoma harbafum. Copeia 1966:42-49.
Marshall, N. B.

1971. Explorations in the life of fishes. Harvard Univ.
Press.,Camb.,204p.
Mead, G. W., E. Bertelsen, and D. M. Cohen.

1964. Reproduction among deep-sea fishes. Deep-Sea Res.
11:569-596.
Tate, M. W., and R. C. Clelland.

1957. Nonparametric and shortcut statistics in the social,
biological, and medical sciences. Interstate Printers and
Publishers, Danville, 171 p.

645

DISTRIBUTION, FOOD, AND FEEDING OF THE THREESPINE

STICKLEBACK, GASTEROSTEUS ACULEATUS, IN GREAT

CENTRAL LAKE, VANCOUVER ISLAND, WITH COMMENTS ON

COMPETITION FOR FOOD WITH JUVENILE SOCKEYE SALMON,

ONCORHYNCHUS NERKA

J. I. Manzer'

ABSTRACT

The distribution, relative abundance, and food of the threespine stickleback, Gasterosteus aculeatus,
was studied in Great Central Lake on Vancouver Island, B.C., in 1970 and 1971 as part of a
multidisciplinary study on the production of sockeye salmon, Oncorhynchus nerka, following controlled
additions of inorganic nutrients (1970-73 inclusive) to an oligotrophic sockeye nursery lake. Stickleback
appeared along shore in relatively low numbers prior to mid-April and most were between 30 and 59 mm
long. Following spawning in June and July, initially stickleback were smaller, but as fish of the year
became more available, both the number and average size of stickleback increased. They were absent in
the littoral zone by November, but their presence in the pelagic zone in winter could not be established.
Over the diel cycle the larger individuals apparently move offshore during the day. The populations in
the 2 yr did not differ greatly in size.

In each of the 2 yr stickleback had a wide but similar diet. They predominantly fed on two cladocerans
(Holopediuni gibberum, Bosmina coregoni), two copepods (Epischura nevadensis, Diaptomus
oregonensis), and a cyclopoid copepod {Cyclops bicuspidatuf:). Larvae and pupae of the family
Chironomidae were also of some importance. Other food items, but of minor importance, included
harpacticoid copepods, insects, pelecypods, ostracods, acarids, Araneida, planaria, Odonata, and fish.
Variations in diet in relation to season, size and sexual maturity of stickleback, and time of day were
observed. The daily ration for stickleback was estimated to be 6.55% of their body weight in July and
7.80% in October.

Stickleback and juvenile sockeye salmon in the littoral zone exhibited considerable dietary overlap,
especially during the late spring and summer. However, since sockeye salmon in this zone are relatively
few in number, and stickleback do not inhabit the limnetic zone, serious interspecific competition for
food in the lake is probably lacking, especially in years of abundant food supply.

Along the Pacific coast of North America, three-
spine stickleback, Gasterosteus aculeatus, here-
after referred to as stickleback, occur in many
coastal lakes, rivers, and streams ranging from
western Alaska to lower California (McPhail and
Lindsey 1970). In British Columbia and Alaska,
large populations have been reported in some
nursery lakes of young sockeye salmon, Onco-
rhynchus nerka (Greenbank and Nelson 1959;
Ruggles 1965). Separate studies on the food of
young sockeye salmon (Ricker 1937; Narver 1970;
Barraclough and Robinson 1972) and stickleback
(Greenbank and Nelson 1959) in British Columbia
and Alaska lakes have generally shown that both
species feed mainly on planktonic crustaceans and
insects. Rogers (1968) compared the food of both

'Pacific Biological Station, Department of the Environment,
Nanaimo, B.C.. Canada V9R 5K6.

Manuscript accepted January 1976.
FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

species residing in the same lake in Alaska and,
after finding a great similarity in diet, concluded
that potential interspecific competition for food
exists. Krogius and Krokhin (1956) and Ruggles
(1965) studied production of young sockeye salmon
in different lakes where the two species were
present and observed that sockeye salmon
production was inversely related to stickleback
abundance.

In 1969, the Pacific Biological Station of
the Department of the Environment, Canada,
started a multidisciplinary investigation to deter-
mine if the production of juvenile sockeye salmon
in Great Central Lake on Vancouver Island, B.C.,
(Figure 1) would be increased by controlled addi-
tions of inorganic nutrients. Approximately 100
tons of inorganic nutrients were added from late
spring through summer for 4 yr beginning in 1970,
usually in 5-ton weekly lots with the ultimate

647

FISHERY BULLETIN: VOL. 74, NO. 3

Figure l.-Map of Great Central Lake, British Columbia,
showing the location of beach seining (numbers) and mid-water
trawling (lines) stations.

purpose of increasing the food resource for young
sockeye salmon without significantly altering
water quality. Preliminary results for 1970 when
compared with 1969 (untreated year), indicate
that primary production was increased without
substantially changing the nature of the food
chain (Parsons et al. 1970; Parsons et al. 1972).
Zooplankton standing stock from May through
October was approximately 10 times higher (Le-
Brasseur and Kennedy 1972). Young sockeye
salmon generally consumed the important zoo-
plankters in the lake and underyearling sockeye
salmon were 30% heavier in weight (Barraclough
and Robinson 1972). Considering the results of
earlier studies by other investigators on the food
of young sockeye salmon and stickleback, and the
uncertainty of the response of the stickleback
population to lake enrichment, a study on the
biology of stickleback with special emphasis on
diet and feeding habits was carried out in 1970 and
1971 as part of the overall fertilization experiment
in Great Central Lake. This paper reports on the
results of studies on distribution, relative abun-
dance, and food and feeding of stickleback, and in
addition contains comments on stickleback as a
competitor of juvenile sockeye salmon for the food
resource in the lake.

DESCRIPTION OF GREAT
CENTRAL LAKE

Great Central Lake is an ultra-oligotrophic lake
situated in the central part of Vancouver Island,
B.C. The lake is approximately 34 km (21 miles)
long, varies between 1 and 2.5 km (0.6 and 1.5
miles) in width, and has a surface area of 5,100
hectares. The maximum depth is approximately
250 m (800 feet). The shoreline varies from gentle
sloping beaches to rocky, precipitous ledges. The
littoral area in comparison to lake perimeter is

relatively small and depths of 25 m or more only a
few meters from shore are common. Beach cover
ranges from small pebbles to rocks and boulders.
Water inflow is by two major streams at the west
end and several minor streams around the lake, as
well as by snow melt in the spring months. The
lake is drained at its east end by the Stamp River,
which flows approximately 30 km before emptying
into the sea at the head of Alberni Inlet. Surface
water temperatures in the lake ranged from 4° to
21°C in 1970 and from 4° to 24°C in 1971. Minimal
temperatures occur in February; maximal tem-
peratures in late July. In general warm-up is
slower in the western end, but once maximum
temperatures are reached in July, surface water
cools off at approximately the same rate. In some
winters, the lake is ice-covered for varying periods
of time, more often at the western end.

The fish community consists of at least eight
species. Young sockeye salmon are by far the most
abundant, followed by stickleback. Other species
caught in considerably fewer numbers are juvenile
coho salmon, 0. kisutch; cutthroat trout, Salmo
clarki; rainbow trout, 5. gairdneri: Dolly Varden,
Salvelinus malma; prickly sculpin, Cottus asper;
pumpkinseed, Lepomis gibbosus; and river lam-
prey, Lampetra ayresi.

TAXONOMY

Two morphologically different forms of G.
aculeatus occur along the Pacific coast of North
America: a heavily plated form, trachurus, that is
usually marine, and a partially plated freshwater
form, leiurus. McPhail and Lindsey (1970) provid-
ed nomenclatural and taxonomic details regarding
the G. aculeatus complex. Recent studies on
isolated freshwater populations indicate consider-
able geographic variation with the result that
their taxonomic status is of considerable uncer-
tainty and interest (Hagen 1967; Narver 1969;
Miller and Hubbs 1969; Hagen and McPhail 1970;
Hagen and Gilbertson 1972). Hagen and Gilbert-
son (1972) consider that at least three plate morphs
are present in permanent freshwater populations
of Gasterosteus, namely low plated (3-7), partially
plated (8-29), and fully plated (30-35).

The stickleback morph in Great Central Lake
was identified from samples collected prior to the
spawning season at four stations (3, 5, 13, and 14,
see Figure 1) located along the length of the lake.
The individual samples contained from 14 to 20

648

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

stickleback. The length of the stickleback in the
combined samples ranged from 45 to 79 mm.
Lateral plates along the left side and caudal keel
were counted, using a probe and binocular micro-
scope. Since all individuals in the samples exceeded
30 mm in length, plate formation was considered
complete (Hagen and Gilbertson 1972). Analysis of
variance revealed no significant difference in plate
counts between stations (F = 3.15; df = 3, 66;
P>0.05). The mean plate count for the combined
samples was 25.17. Considering plate counts, it can
be concluded that the stickleback population in
Great Central Lake is a freshwater population
more representative of the trachiiriis than the
leiuru>i form.

Descriptid)!

Gentle slope, gravel bottom
Gentle slope, rocks and boulders
Rock slope, sharp dropoff

Station

1, 4, 5, 9, 10, 14
6, 12, 15, 16, 17. 18
2,3, 7. 8, II, 13, 19

Information on the winter distribution of stick-
leback was obtained from purse seining operations
carried out on 18 February and from mid-water
trawling on 23 and 24 March in the pelagic zone,
using a mid-water trawl routinely employed to
catch age-0 sockeye salmon in the lake (Barra-
clough and Robinson 1972). Ice cover restricted
fishing to the eastern one-half of the lake.

Results

DISTRIBUTION AND
RELATIVE ABUNDANCE

Methods

Distribution and estimates of relative abun-
dance of stickleback were determined from
catches made with a purse seine used as such in
mid-lake waters or as a beach seine along the
shoreline, in 1970 and 1971. A description of the
gear and its operation as a beach seine was
provided by Manzer (1971). The net sampled an
area between 450 and 550 m-, or approximately 10
m of shoreline.

The field program in 1970 was carried out over
eight surveys between 22 April and 27 November.
Some purse seining and sighting were carried out
in the early season but most eff'ort was devoted to
beach seining along shore. Here 18 different loca-
tions representing typical but different shoreline
habitats were fished between 0830 and 1730 h
(Pacific daylight time). Eleven of these stations
were established as key stations. Coverage was
more complete between late June and late August
when surveys were conducted at 2-wk intervals.
The fishing program in 1971 was essentially the
same as in 1970. Five secondary stations sampled
in 1970 were deleted and one new station was
added to provide better coverage of the lake.
Seven surveys were carried out between 18 Feb-
ruary and 30 November, approximately at month-
ly intervals beginning in May. Fishing was con-
ducted between 0630 and 1830 h. No fishing was
done in September in either 1970 or 1971. The
beach seining stations are shown in Figure 1 and
grouped by character below, the stations in bold-
face being key stations.

Sighting surveys, purse seining, and beach
seining were conducted in the eastern part of the
lake in April and June 1970. The purpose of these
operations was to determine the distribution of
stickleback in proximity to the shoreline. It was
considered that the results of these operations
would be applicable to the lake as a whole. Stick-
leback were readily observed in varying numbers
close to shore apparently moving at random and
feeding in waters from less than 1 foot (0.3 m) to
several feet (ca. 2 m) deep. They were rarely seen in
offshore waters. This general pattern of distribu-
tion was confirmed by purse seine and beach seine
catches. Eight "blind" (i.e., uncorroborated by
sightings) purse seine sets in the limnetic zone
yielded three stickleback. The net was considered
effective to a depth of 3-4 m. In contrast, 16 beach
seine sets at shore areas ranging from shallow
beaches to precipitous slopes yielded stickleback
on all but three occasions. As many as 350 stick-
leback were caught in a single set. Their virtual
absence in offshore waters was indicated by the
results of townetting for young sockeye salmon in
the lake. A total of 480 tows made during 1969-73
in the limnetic zone of the lake at various depths
(0-60 m) with trawl nets with mouth openings of
approximately 18 m- and 4 m- yielded 21 stickle-
back (D.G. Robinson pers. commun.). From the.se
operations it is concluded that stickleback were
primarily concentrated close to shore.

Catches of stickleback by beach seining opera-
tions are given in Table 1 by survey and location.
Catches in each year ranged from zero or a few fish
to estimates of 2,500. In 1970, 105 sets were made
and 10,727 stickleback were caught. Twenty-one
sets failed to catch stickleback. In 1971, 89 sets
were made and 10,806 stickleback were caught. Of

649

FISHERY BULLETIN: VOL. 74, NO. 3

w
e«

o
_o

35 .2

-a g

rt o

4J ■ —

c ■*-'

ca
I

^ O

o g"

~ a)

C\J (J

- D

m <

<o O)

- 3
Ul <

. o.
eg <

(OCMOOOCMin ^^-o |io»- leg
T-" T-" n

» I I

CO (o o>

in o CO

CO
CO

C\J

2!.

1° I I

(^ h~ in
eg r~ cj

I I I IS I I

CO O -^ O O CJ)

eg T- eg o o eg
CO in

I eg •» in CO CO CO
I 1- eg •<r eg 1-

<n <» CO CO o CD

CO CO CSJ o '-

O >- CO I CD 00

o r- CM I

1 CO o
1 in T-

o
eg

1 CD 1-
1 CO 1-

CJ)
CO

CJJ

h- S- CO
O r-

CO
CO

O »- CD
O CO C3)

in eg

CO ■«r eg
CO in

o 1^ I

Tt in CD

■<r

CO O T-

in o o

ei o in

in cvj

IS I

in C3> h-

t- CO 1-

cocD^-^cocMOOegTf
CO eg '-

inino>egoegcoo>mo5mN.h*egcocgr**o
h*cDh^coejinin»— ininoOr-cDincoegT—
T-in^^. egeg_ oco^

^" '- ci eg"

r-- in
eg o

eg eg CM

o CO o eg o

CO o o o o

CD o o o o

CD CO eg o

^ in CO o r^

eg eg in o CO

eg "- o eg

eg eg
o o o o

o o

o
o
in

eg eg eg eg eg
eginO'<tt^inegin
>- o eg eg '-eg

CO r-

t- h- in O) OT
CO o> h- ■*

CM CO

<DOO)0)T-o)egO'-
m egincor~T-ineg

ocD^ooinegco^o

mincOr-T- r-T-O

o f^

o CO

CM '-

CO o
C3> eg

eg o oo >-

s I i-s

CM^
O

I I I

r^ o

CM 1-

in (O

CJ> CO

I II II I II I I I I 11^'

csjcO"*mcDr>-coo)0'»-CMCO'^m<os.ooo5

i2 E
a>

o c
2 <

<o CO

cm'

m o
en T-

1- o>

in y-

T—

T-

co'

CO
eg

c
CD

650

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

these, 12 sets failed to catch stickleback. Most of
the sets which failed to catch stickleback (21 of 33)
were made in February and November. Arithme-
tic and geometric means of the numbers caught in
each survey are also provided. The latter are
included because of the skewness of the catch data
and were obtained from log (» + 1) transforma-
tion of the data where )i is the stickleback count in
each set. This transformation permitted utiliza-
tion of zero catches in the computations: in all
likelihood during the spring to fall months at least
one stickleback would have been caught had
fishing been repeated.

The distribution and relative abundance of
stickleback and size composition of the catches
according to small (<30 mm), medium (30-59 mm),
and large (60+ mm) stickleback are illustrated in
Figure 2. (The size-groups were arbitrarily chosen
but in general approximate age-groups: <30 mm =
0 age; 30-59 = lyr old; 60-1- mm = 2 yr and older.)
Gear efficiency was assumed to be reasonably
constant, although a few sets were made under
conditions of relatively strong wind and current.
It was further assumed that after spawning (July
and later) stickleback were catchable regardless of
size. Abundance levels just prior to spawning may
have been higher than catches indicate because of
the decreased vulnerability of mature individuals,
especially males which repair to nesting areas.

Some differences in survey dates, especially in
the early part of the year, and some changes in the
sampling sites in the 2 yr prevent a strict time and
place comparison of the data. Nevertheless some
general conclusions on distribution and relative
abundance can be made from Table 1 and Figure 2.
Seasonally, stickleback appeared along shore prior
to mid-April. Their abundance at this time was low
and appeared to vary between locations. Most
stickleback in almost all localities ranged in length
between 30 and 59 mm. A few larger individuals
were caught but none smaller. In both years it was
obvious that in all areas stickleback progressively
increased in numbers, from July through October,
although apparently they were less abundant off
rock slopes than on gentle sloping beaches covered
by either gravel or boulders. This increase is due to
the recruitment of fish of the year as evidenced by
the large proportion of fish less than 30 mm in July
and August. The average seasonal catch was
largest in October and consisted of stickleback
measuring between 80 and 59 mm long. Fish
belonging to the small and large size groups also
were present in considerable numbers, and in

some areas small fish predominated (for example,
the central part of the north shore). The small or
zero catches made in November suggest that
stickleback prior to winter had abandoned the
shore areas.

Observations on diel size variation in stick-
leback along the shore were made in conjunction
with diel feeding habits, which are described in a
later section. Paired samples taken 100 m and 15
min apart were collected at station 1 at 3-h inter-
vals between 0700 and 1900 h in October 1970 and
through the 24-h cycle in July 1971. Diel size
changes observed during each series are illustrat-
ed in Figure 3 using the graphic method developed
by Dice and Leraas (Simpson and Roe 1939). At
each site and date the size of stickleback decreased
from dawn to midday and then increased again by
dusk, suggesting that the large fish are less
available in the littoral area during the day. This
trend is most clearly shown by fish in July at site B.
Here, stickleback at midday are significantly
smaller than at either dawn or dusk.

Stickleback virtually abandon the shore areas by
November, but their presence in numbers in the
pelagic zone of the lake during the winter could
not be established. Limited purse seining (four
sets) in February in the pelagic zone of the eastern
part of the lake failed to yield any stickleback.
Mid-water trawling in March, along transverse
and longitudinal axes of the lake over a lineal
distance of 22 km and at depths ranging from 10 to
100 m in the eastern half of the lake, resulted in
the capture of one stickleback; ice cover prevented
trawling in the western half of the lake. This
stickleback measured 37 mm long and could have
been caught at some depth down to 50 m. From the
results of these fishing operations stickleback
apparently either leave the lake or retreat to areas
where they cannot be caught for the winter
months, becoming available again between Feb-
ruary and April.

Reliable estimates of the size of the stickleback
population could not be made from the available
catch data. Within any survey, catches varied
widely between locations. In addition, local var-
iance in the catches is not precisely known, al-
though judging from a few instances when two
sets were made in the same location the numbers
caught can vary greatly. The catch data are
considered more informative for the period
beginning in July when coverage was more com-
plete and stickleback availability increased. As-
suming that factors contributing to variability in

651

49*25'

FISHERY BULLETIN: VOL. 74, NO. 3

125*25' 20' 15' 10' 05' 125*00' 125*25' 20' 15' 10' 05' 125*00'

49* 25'

49* 20

49*25'

49* 20'

49*25'

49* 20

49*25

49* 20

49*25'

49* 20'

49*25

49* 20'

49* 25'

49*20

49*25

49*20 -

125*25' 20' 15' 10' 05' 125*00'

49* 20

49*25

49* 20

49*25

49*20

49*25

49*20

49*25

49*20

49*25'

49*20

49*25

49* 20

125*25' 20' 15' 10' 05' 125*00'

NUMBER LENGTH (mm)

500

100-500
25-99

O 1-24

o 0

^ <30

^ 30-59

A 60-t-

Q NO DATA

Figure 2.- Distribution and size composition of catches of threespine stickleback in Great Central Lake, 1970 and 197L

652

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

40-

60-

40-

OCTOBER 1970

(12)

-MAXIMUM LENGTH

- +2 S.E OF MEAN
-MEAN LENGTH

-2 S.E OF MEAN

- MINIMUM LENGTH
(251- SAMPLE SIZE

(25)

(22)

I 1 1 1 1 1 1 r-

JULY 1971

n r'(25)

n(l6) ,

(22)

(19)

0700

1300 1900

HOUR

0100

Figure 3.-Diel changes in the length of threespine stickleback
as indicated by paired catches at station 1, October 1970 and July
1971. Open bars = length range; solid bars = ±2 SE of mean;
dash = mean length. Site A = left bar, site B = right bar.

the catches in the 2 yr averaged out, the mean
catch for surveys in 1971 was consistently higher
than that for the same period in 1970. The dif-
ference between yearly mean catches was only
20%, suggesting that the stickleback populations
in the 2 yr were approximately similar in size.

Discussion

Seasonal changes in abundance and distribution
have been described for several lake populations of
threespine stickleback. Greenbank and Nelson
(1959), on the basis of beach seine catches, reported
that in Bare and Karluk lakes, Alaska, from late
May into September stickleback in varying
numbers essentially inhabited shallow waters. A
few were sighted on the surface of Karluk Lake at
a considerable distance from shore, and some were
caught by fyke nets at depths of 30 and 80 feet
(approximately 9 and 25 m) but not in sets at 126 or
200 feet (approximately 39 and 61 m). Ruggles
(1965), while studying juvenile sockeye salmon in
Lake Owikeno, B.C., observed that during April to

October, stickleback were most abundant in areas
suitable for spawning and were taken in two
netting operations in midlake surface waters in
considerable numbers. Stickleback fry were
caught throughout the spring to fall seasons but
largest catches were made in the spring. In some
years, a secondary increase in abundance occurred
in the fall. In Lake Aleknagik, Alaska, Rogers et
al. (1963), and Rogers (1968) using beach seines,
trawls, and tow nets, observed stickleback in the
spring and early summer to inhabit mainly the
littoral area. By midsummer, fish of age I and II
became pelagic while age 0 and III tended to
remain inshore. Observations on stickleback dis-
tribution, movement, or numbers during the late
fall and winter are lacking for these lakes, pre-
sumably because of ice cover. Markovtsev (1972),
however, in Lake Dalnee from January through
August observed that stickleback are present over
winter in the pelagic zone and the population
started moving from the pelagic to the littoral
zone about May and resumed pelagic residence in
the summer.

The seasonal occurrence of threespine stick-
leback in Great Central Lake is generally similar
to those described for other lake populations along
the Pacific coast, but their distribution during
summer appears to be somewhat diff"erent. In
other lakes, beginning in midsummer, some stick-
leback leave the littoral area to inhabit pelagic
waters; those in Great Central Lake remain rela-
tively close to or along the shore throughout lake
residence. The reason for this apparent difference
in distribution patterns is not known although it
seems unlikely that it is the result of diff'erent
fishing gears and methods employed by various
investigators. The distribution patterns in the
different lakes may be related to lake bathymetry.
By camparison with other lakes studied Great
Central Lake has relatively little littoral area.
Expanses of water exceeding 25 m or more in
depth only a few meters from shore are common.
This bathymetric feature may provide stickleback
with a food supply close to shore thus making it
unnecessary for them to move into offshore feed-
ing areas.

The virtual absence of stickleback in the pelagic
zone in Great Central Lake does not conflict with
the documented onshore-offshore movements of
larger individuals during midsummer and fall.
Offshore movement during the day and corre-
sponding onshore movements at night were
reported for marine threespine stickleback in the

653

FISHERY BULLETIN: VOL. 74, NO. 3

Baltic (Meek 1916). The stimulus for this size-
related behavioral difference remains unknown. In
Great Central Lake some survival or feeding
advantage may accrue to smaller individuals
remaining close to shore but the affinity for shore
shown by large individuals in July is probably
associated with reproduction because virtually all
these fish were physically mature or gravid.

FOOD AND FEEDING

Methods

Feeding Relationships

percent of body weight was used as an index of
feeding intensity. Gravid females were excluded
from the analyses because they appeared to feed
less intensively, judging from the occluded
stomachs of many individuals. Supplementary
information on feeding periodicity was also ob-
tained by subjectively classifying stomachs as
either full, three-fourths full, one-half full, one-
fourth full, trace of food, or empty, and noting
whether the contents were fresh, partially digest-
ed, or digested and therefore unidentifiable. The
basic data are reported by Manzer (1971, 1972).

Three methods were used to determine the
importance of organisms as food:

Seasonal and spatial differences in stickleback
diet were determined from catches or samples of
catches, if large, made during each fishing survey
in 1970 and 1971. By coincidence, stomachs from
544 stickleback, or approximately 5% of the total
number caught in each year, were examined for
content. Stickleback examined in 1970 ranged in
length from 15 to 78 mm; in 1971, from 14 to 86
mm. The numbers of fish examined from each
station and by survey in the 2 yr are given in Table
2.

Fork length (millimeters), body weight (milli-
grams, minus the weight of the body cavity para-
site, Schistocephalus, if present), and stomach
content weight (to nearest 0.2 mg) were obtained.
Stomach content weight was determined by first
weighing the stomach with food and then without.
The stomach contents were identified to species
when possible, and counted using a binocular
microscope. The content weight expressed as a

a. Occurrence-the percent of stickleback feeding
on a particular organism.

b. Numerical-mean number of a particular or-
ganism per stomach.

c. Points— relative importance of organisms con-
sidering size and numbers.

The relative merits of these methods have been
discussed by Hynes (1950) and Windell (1968). For
the points method, the equivalent units assigned
different organisms are given in Table 3. The units
for common planktonic Crustacea are in the ratio
of their wet weight, as determined from zoo-
plankton studies in Great Central Lake (LeBras-
seur and Kennedy 1972). Equivalent units for
other organisms, including insects, were deter-
mined by inspection and assigned the same unit
value as other organisms or groups of organisms of
similar volume, assuming a common specific
gravity. Since individual size of a given organism

Table 2.

-Numbers of

threespine

stickleback stomachs examined,
1971.

survey

and location,

1970 and

Location

Survey

Date 1

2 3 4

5 6 7 9

10 11 12

19 Total

1970:

7
Total

1971:

6
Total

22, 30 Apr.
24, 25 June
8, 9 July
22, 23 July
5, 6 Aug.
19, 20 Aug.
2 Oct.

12,20 May
10, 17 June

9 July

10 Aug.
14 Oct.
30 Nov.

15 — — — — — — — — — — — — — — — — _ 15

— — 23 8 — — — — — — — — — — — 15 10— 56

— 13 — — — — — — — — — — — — — — — — 13

— 30 20 10 10 — — — 10 — — — — — — — — — 80

19 — 15 32 36 — — 10 25 15 13 12 10 10 — — — — 197

— 13 20 13 22 10 — — 10 10 10 — — 10 — — — — 118

— 15 10 10 10 10 — — — 10 — — — — — — — — 65

34 71 88 73 78 20 — 10 45 35 23 12 10 20 — 15 10 — 544

— 20 20 3 8

10 10 — 10 10 10

— 10 —

10 20 15 6

10 10 12 11

10 10 12 10

8 8 10 —

10 —

— 11

— 10 10
10 10 10

— 10 10
10 10

10 10

6 8

40 68 68 62 45 20 31 40 43

26 28 — 53 —

— 74
10 110
10 130

— 105

— 99

— 26
20 544

654

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

Table 3. -Equivalent units of important dietaries.

Organism

Bulk units

Organism

Bulk units

Alona

Nauplius

0.2

Holopedium

Copepodids

Bosmina

Harpactlcoid

Daphnia

Chironomid larva

Epischura

Chironomid pupa

Diaptomus

Egg, zooplankton

Cyclops

Egg, stickleback

was reasonably uniform with time, seasonal ad-
justment of equivalent units appeared unneces-
sary. Items which averaged less than one per
stomach or less than 1.0% of the bulk were record-
ed as trace (T) quantities.

Stomachs of large stickleback frequently con-
tained several hundred organisms. In such cases,
contents were identified and enumerated from a
weighed portion of the total bolus and the result-
ing counts were then prorated to the total weight
to estimate the numbers of organisms consumed.
The remaining portion of the bolus was examined
for food organisms not represented in the sample.
Correlation analysis indicated a very significant
positive relationship between actual and estimat-
ed counts of major food items (r = -i- 0.964, P<0.01,
n = 15).

Major features of the stickleback diet were
adequately described from examinations of 10
stomachs per sample. In a few cases, smaller
numbers were examined to eliminate gaps in time
or place. On the basis of two separate tests of
association between stomach contents of 10 and 25
fish samples from the same catch, ranked by
numbers, Spearman's rank correlation test (Siegel
1956) gave r, values of -1-0.943 and 1.000. The
extent of lake coverage in the 2 yr, especially 1970,
differed between surveys. The dietary agreement
among stickleback taken at different locations
within surveys was examined using Kendall's
coefficient of concordance test (Siegel 1956). For
each survey, the most common food items at each
location were ranked according to mean number in
the sample, excluding material rendered uniden-
tifiable through digestion. Corrections were made
for items tied in rank and W, the index of diver-
gence of observed from perfect agreement, and
related chi-square values were calculated. For
eight of the nine surveys tested (four in 1970 and
five in 1971) the agreement observed in rankings
of dietaries among locations was higher than it
would be by chance {P = 0.05) (Table 4). Therefore,
it seemed reasonable to combine the data for all
locations by survey to facilitate detection of

possible seasonal changes in diet. From plankton
studies conducted in Great Central Lake in 1970,
LeBrasseur and Kennedy (1972) stated that "the
epilimnion is well mixed, thus assuring a nearly
uniform dispersal of planktonic organisms along
the lake."

Diet in relation to sexual maturity was deter-
mined from combined samples of stickleback
caught during the first three surveys (mid-May to
early June) in 1971. Mature and immature females
were separated on the basis of size, 60 mm being
used as the dividing length. Of 54 females 60 mm
or larger examined, 4 were immature and 50 were
mature. Of the latter group, 28 were ripe. Blue
coloration of the iris and red coloration of the
pelvic region were used to separate mature from
immature males (Craig-Bennett 1931; Greenbank
and Nelson 1959). Because female sticklebacks are
larger than males of equivalent age (Greenbank
and Nelson 1959; van Mullem and van der Vlugt
1964) males larger than 60 mm were considered to
be sexually mature. From testes inspection, ripe
males were few in number and accordingly no
attempt was made to treat functional and non-
functional males separately. The relative scarcity
of ripe males is believed due to their behavior of
attending spawning females or nests.

Diel Feeding Rhythm

Diel feeding periodicity and food composition
studies were based on paired catches made at
station 1 on 1-2 October 1970, and 21-22 July 1971
at two sites (A and B), approximately 100 m apart.
In the October series, fishing started at 1300 h 1
October and during the next 24-h period was
conducted at 1600, 1900, 2200, 0630, and 1000 h.

Table 4.— Summary of results of Kendall coefficient of concor-
dance ( WO tests (Siegel 1956) for agreement in diet of threespine
stickleback at different sampling locations.

Number

of food

Survey

locations

categories

Chi-

date

(N)

square

P level

1970:

22-23 July

0.566

31.13

0.01

5-6 Aug.

0.383

26.12

0.001

19-20 Aug.

0.498

49.30

0.001

2 Oct.

0.581

41.83

0.001

1971:

12, 20 May

0.220

8.80

0.70

10, 17 June

0.317

25.36

0.01

9 July

0.450

43.20

0.001

10 Aug.

0.230

11.50

0.05

14 Oct.

0.629

45.29

0.001

655

FISHERY BULLETIN: VOL. 74, NO. 3

Fishing at 2200 h at each site failed to yield any
stickleback, presumably because of inefficient
operations under conditions of total darkness. As a
consequence further sampling was suspended
until 0630 h 2 October. Fishing during the July
series began at 0700 h and was repeated at 1000,
1300, 1600, 1900, 2200, 0100, and 0400 h. Gear
problems precluded fishing at site B at 1900 h.
During each series, the time interval between
fishing at the two sites at any time of day was
approximately 15 min and for practical purposes
can be considered concurrent.

The target sample size for each site and time of
day was 25 fish. Except for sampling times already
indicated, this number was achieved or closely
approximated. The smallest sample contained 12
fish (site B, 1900 h). All fish in the sample were
processed in accordance with methods described
earlier and 10 fish, selected at random, were
examined for stomach contents. A total of 226
stickleback were examined for the October series,
334 for the July series. The sizes of stickleback by
sample are illustrated in Figure 3.

Mean feeding intensity indices (food weight/

body weight X 100) for paired samples were
similar, and data for each series were pooled by
time of day.

Daily Ration and Maximum Meal Size

In this study, daily ration is defined as the
weight of food consumed over a 24-h period ex-
pressed as a percent of body weight. Daily rations
were estimated from the diel feeding rhythm
curve, using a modification of the method
developed by Keast and Welsh (1968). Essentially,
diff'erences between maximal and minimal feed-
ing indices during successive periods over a 24-h
cycle were determined and these values and the
residual content were summed. The method is
most applicable to species which completely empty
their stomachs between meals.

Maximum individual meal size was determined
from regression analysis of stickleback taken
during the maximal feeding period in July and
which were judged to have "full" stomachs ac-
cording to the subjective "fullness" scale described
earlier.

Table 5. -Seasonal chang-e in the diet of threespine stickleback in Great Central Lake. 1970.

Date

22, 30 Ap

24-25 June

8-9 J L

ily

22-23 Ji

uly

5-6 Aug

19-20 Aug.

2 Oct.

No. fish examined

197

118

% empty

3.6

15.4

13.7

3.6

3.4

Size range (mm)

33-63

36-72

16-70

18-75

15-74

19-76

27-76

Mean size (mm)

Mean content wt (mg)

34.6

27.4

14.6

25.4

15.3

14.3

17.4

Organism

Rotifera

—

Ciadocera:

Holopedium

—

115

141

129

Bosmina

100

2,419

235

116

Daphnia

—

Alona

—

Copepoda:

Epischura

Diaptomus

—

Cyclops

100

—

Copepodids

—

Nauplii

—

Harpacticoid

—

Insecta:

Chironomid larvae

—

Chironomid pupae

—

Other

—

Eggs - zooplankton

Other:

Pelecypoda

—

Ostracoda

—

Acari

—

Planaria

—

Odonata

—

Fish

—

Unidentifiable %

'% stomachs with item.

^Mean no. items per stomach examined.

Mtem = % of total bulk units. T = Trace

< 1 organism or <1%.

656

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

RESULTS

Feeding Relationships
Seasonal Variations in Diet

Data on size and stomach contents of stick-
leback examined in 1970 and 1971 are summarized
by survey in Tables 5 and 6. The predominant
features regarding seasonal change in diet are
depicted in Figure 4. Observations for 1970, except
for August when almost all stations were sampled,
are based mainly on samples taken from the
eastern part of the lake. Observations for 1971 are
based on samples from most of the key sampling
stations except in November when fishing was
confined to the eastern end of the lake.

Although the numbers of stickleback examined
differed by survey, a similar seasonal trend in the
proportion of fish with empty stomachs was ob-
served for the 2 yr: low in the spring and early
summer, highest in midsummer, and again low in
the fall. The mean weight of stomach contents

fluctuated in each year but generally was higher in
the spring and early summer. The higher mean
values in the early part of the year are probably
related to fish size. On the average, stickleback
were larger in the spring and early summer than
in the late summer and fall. The relatively high
proportion of fish with empty stomachs in mid-
season can be explained by feeding behavioral
diff"erences associated with sexual maturity.

In each of the 2 yr stickleback had a wide but
rather similar diet. They predominately fed on five
species of organisms: two cladocerans
{Holopedium gibberum, Bosmina coregoni), two
copepods {Epischura nevadensis, Diaptomus
oregonensis), and a cyclopoid copepod {Cifclops
biciispidatus). Larvae and pupae of the family
Chironomidae were also of some importance. The
distinction between zooplankton eggs and fish
eggs in 1971 represents a qualitative refinement in
analysis of the data, rather than any difference in
diet. Other kinds of organisms consumed at var-
ious times but of minor importance were harpac-
ticoid copepods, insects, pelecypods, ostracods.

Table 6.-Seasonal change in the diet of threespine stickleback in Great Central Lake, 1971.

Date 12, 20 May 10, 17 June 9 July 10 Aug. 14 Oct. 30 Nov.

No. fish examined 74 110 130 105 99 26

% empty 2.7 4.5 18.5 19.1 8.1 3.8

Size range (mm) 29-86 33-82 15-86 14-80 23-77 24-78

Mean size (mm) 54 54 58 33 38 34

Mean content wt (mg) 26.1 45.3 28.0 16.5 18.6 13.9

Organism 1122 3^1 2 3 123 123 1 23 123

Rotifer _ _ _ 10 1 T 40 15 T 44 27 T 74 35 T 31 12 T

Cladocera:

Holopedium 1 T T 34 14 1 59 47 16 57 22 28 89 163 67 65 20 35

Bosmina 47 10 2 37 2 T 19 T T 40 10 4 89 85 12 69 43 25

Daphnia — — — — — — — — — — — — 19 1 T — — —

Alona 5 T T 4 1 T 2 3 T 54 17 14 20 1 T 23 1 1

Copepoda:

Epischura 1 7 19 56 286 95 50 55 68 6 6 28 19 1 1 — — —

Diaptomus — — — 10 2 1— — — 5 2 2 71 28 8 54 10 12

Cyclops 40 75 19 49 68 2 35 4 T 4 T T 65 20 3 42 6 3

Copepodids 40 13 3 23 8 T 40 12 1 6 T T 48 52 7 50 38 22

Nauplii 32 22T— — — 13 6T— — — 28 2 1— — —

Harpactlcoid 10 1T9 13T3TT8TT1 TT4TT

Insecta:

Chironomid larvae 19 IT 6 TT9TT11TT 1 TT4T —

Chironomid pupae 23 3 38 11 TT 7 1 6 2TT — — — — — —

Other 12 1 13 9 TT 16 16 51 21 3 TT4TT

Eggs - zooplankton 5 1 T 19 3 T 39 22 2 1 T T 54 15 2 31 3 2

Fish ITT 3 TT 4TT ITT — — — — — —

Other:

Amphipoda 4TT 2 TT 2TT— — — — — — 4TT

Pelecypoda 8 1 2 3 TT TTT — — — — — — — — —

Ostracoda — — — 2 TT4TT17 1T — — — 8TT

Acari 8TT4 TT— TT22 12 — — — — — —

Aranelda 1 T T — — — — — — — — — — — — — — —

Fish ____________ 1 TT— — —

Coleoptera 2 T T — — — — — — — — — — — — — — —

Ceratopogonidae 11 T T 3 T T — — — — — — — — — — — —

Isopoda 1 T T — — — — — — — — — — — — — — —

Unidentifiable % 51 47 38 24 36 34

'% stomachs with item.

^Mean no. items per stomach examined.

3|tem = % of total bulk units. T = Trace =<1 organism or <1%.

657

FISHERY BULLETIN: VOL. 74, NO. 3

19 70

NUMERICAL POINTS

22,30 APRIL

1971

NUMERICAL POINTS
12,20 MAY

24,25 JUNE

10,17 JUNE

22,23 JULY

5,6 AUGUST

10, AUGUST

19,20 AUGUST

20 OCTOBER

14 OCTOBER

30 NOVEMBER

65^

LEGEND

gi^K BOSMINA

HOLOPEDIUM

EPISCHURA

DIAPTOMUS

CYCLOPS

EGGS

COPEPODIDS
NAUPLII

CHIRONOMID
PUPAE

ALONA

OTHER

ORGANISMS

Figure 4.-Seasonai change in the predominant food items of threespine stickleback in Great Central
Lake, 1970 and 1971. Figures in the periphery of each pie diagram represent the percent of
stickleback stomachs containing the particular item.

658

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

acarids, Araneida, planaria, Odonata, and fish
(cottids).

The different food organisms differed seasonally
in their dietary importance. Considering items of
major importance, in 1970 in late April, virtually
all stickleback stomachs examined contained Bos-
mina, Cyclops, and zooplankton eggs, but Bos-
mina was most important, averaging 2,419 in-
dividuals per stomach and making up 95% of the
bulk. By late June, Bosmina was still the dominant
food item but had declined somewhat in impor-
tance as indicated by an increasing proportion of
stickleback feeding on Epischura (55%),
Hohpedium (41%), and chironomids, especially
pupae. Of these Epischura was most important,
forming almost 50% of the bulk. Through July and
August, Bosmina was consumed by a high
proportion of stickleback (no less than 60%) but
Holopedium progressively became the dominant
food organism (approximately 60% by bulk). Dur-
ing these two months, the number of stickleback
feeding on Alona, copepod copepodids and nauplii,
and Diaptonms increased but none of these items
was important quantitatively. In October,
Holopedium continued to be the dominant food
item in terms of bulk, but more stickleback fed on
Bosmina (98%). Diaptomus and Cyclops were
present in about 50% of the stomachs examined
and were of minor importance. Rotifers and eggs
were present virtually throughout the study pe-
riod, the former item occurred rather frequently
(26-76%), but were unimportant in terms of bulk.
Judging from size, the eggs were from both
zooplankton and stickleback. Since stickleback
spawn between late June and early August, eggs
encountered at other times of the year presumably
were zooplankton eggs.

In May 1971, about one-half of the stickleback
had Bosmina, Cyclops, and copepodids in their
stomachs. Cyclops was most important in terms of
numbers per stomach (75) but chironomid pupae,
because of relative size of individuals, was impor-
tant in terms of bulk (38%). By mid-June, more
stickleback were feeding on Epischura (56%) and
Holopedium (34%), but Epischura was the domi-
nant food organism (95% of total stomach con-
tents). About the same number (49%) of stick-
leback fed on Cyclops as in May, and although the
item ranked second in incidence, it accounted for
only 2% of the total stomach content. In July,
Epischura declined in importance but still main-
tained dominant position among the other food
organisms. Holopedium continued to increase in

importance. This inverse trend in the importance
of these two food items was observed into October.
In October, Holopedium was the dominant food
item and Bosmina ranked second in bulk and were
consumed by as many stickleback as were
Holopedium. In terms of occurrence, Diaptomus
(71%), Cyclops (65%), copepod copepodids (48%),
and zooplankton eggs (54%) were of secondary
importance. At the end of November, Holopedium,
Bosmina, and copepod copepodids formed the
major part of the diet of stickleback and in-
dividually were of about equal importance.

The stickleback diet in 2 yr showed some marked
seasonal similarities and differences. Bosm ina was
not as important in the early part of 1971 as in
1970. Another difference is the greater importance
of Epischura later into 1971 than 1970, and the
greater importance of Holopedium in July and
August in 1970. A feature common to both years is
the late season resurgence of Bosmina as an
important food organism. It is not known for
certain whether these differences and similarities
represent annual differences in abundance levels
of the various kinds of organisms or in sampling
dates.

Diet in Relation to Stickleback Size

A total of 205 stickleback taken from the eastern
end of the lake on 22 July and 5 August 1970, and
ranging in length from 15 to 78 mm were ex-
amined for diet differences in relation to size. The
stickleback were arbitrarily divided into four size
groups: <30 mm, 30-49 mm, 50-69 mm, 70-1- mm.
Data on diet for the same size group for the 2 days
were pooled since samples were obtained in the
same general area within a short time interval
(Table 7).

A high proportion of the stickleback (75, 65, and
68% respectively) in the <30 mm group consumed
Bosmina, Rotifera, and Holopedium. Alona, Epis-
chura, and chironomid larvae occurred in about
one-half of the stomachs. Of the remaining items
consumed only copepod nauplii, chironomid pupae,
and zooplankton eggs were of any importance,
occurring in 18, 16, and 13% of the stomachs,
respectively. Larger stickleback, excluding the
70-1- mm group of which only 11 were examined,
tended to feed more on Holopedium, Epischura,
chironomid pupae, and zooplankton eggs, and less
on Rotifera (except those in the 30-49 mm group),
Bosmina and Alona. Copepod nauplii apparently
were not consumed by larger stickleback, but fish

659

FISHERY BULLETIN: VOL. 74, NO. 3

Table 7.-0ccurrence (percent) of different organisms in the diet
of threespine sticklel)ack of four size groups- <30 mm group
contained 100 fish; 30-49 mm, 34 fish; 50-69 mm, 60 fish; and 70-1-
mm, 11 fish. Based on samples taken on 22 July and 5 August
1970.

Table 8.- Relative importance (percent) of food organisms of
different bulk units in the diet of threespine stickleback of four
size groups- <30 mm group contained 100 fish; 30-49 mm, 34 fish;
50-69 mm, 60 fish; and 70-1- mm, 11 fish. Based on samples taken
on 22 July and 5 August 1970.

Size group

(mm)

Bulk
units

ize group (mm)

Organism

<30
65

30-49
85

50-69
49

70-1-
9

<30

30-49

50-69

70 +

Item

Rotifera

< 1

Rotifera, nauplius, Bosmina,

Cladocera:
Holopedium
Bosmina

68
75

81
36

71
28

Cyclops, copepodids,
harpacticoids, zooplankton
eggs

Alona

—

Alona, Daphnia, Diaptomus,

Copepoda:

stickleback eggs

Epischura

Holopedium

Diaptomus

—

Chironomid larvae

Cyclops

—

Epischura

Copepodids

—

>50

Chironomid pupae, fish

Nauplii
Harpacticoid

_..

—

Insecta:
Chironomid larvae

Table 9

.-Stomach i

contents of nongravid and gravid females

Chironomid pupae

—

and sexually mature

male

threespi

ne stickleback, Great Central

Other
Eggs - zooplankton

7
13

5
29

7
23

Lake, 12

May-

■9 July

1971.

Other:

Pelecypoda

Female Male

Ostracoda

6
2

—

Non-gravid

Gravid

Acari

Araneida

Fish

10
3

No. examined
Percent with food

Organism

28 81
61 90

Isopoda

2^ 33

12 3 12 3

larvae were, the largest, a cottid, measuring 14
mm. It is reasonably clear that a positive rela-
tionship exists between food size and stickleback
size. This relationship is also apparent when for
each stickleback size-group the different food
organisms, especially the common items (namely,
Bosmina, Holopedium, and Epischura), are ex-
pressed as a percent of the total stomach content
for that group (Table 8).

Diet in Relation to Sexual Maturity

Mature males showed a higher incidence of
feeding (90%) than did gravid females (61%)
(Table 9), the difference being statistically sig-
nificant, (x" = 13.811, n = 2,P= <0.01).

Nongravid females, gravid females, and mature
males fed on a variety of similar kinds of organ-
isms (Table 9) and, except for Epischufa, none of
the items were of great importance as food. Since
Epischura is the largest planktonic form, its
predominance in the diet of large individuals is not
unexpected. Epischura formed more than 90% of
the bulk units and the mean number ingested was
very much higher than for any other single item.
In contrast to 54% of gravid females which had
eaten this item, its occurrence in nongravid
females and in males was considerably less, 18 and
16%, respectively. Planktonic crustaceans, insects.

Rotifera 92T ___ 10 4T
Cladocera:

Holopedium 23 6 1 32 75 7 17 13 1

Bosmina — — — 4TT 61T

Alona 51T ___ 22T
Copepoda:

Epischura 18 141 94 54 268 92 16 230 94

Diaptomus — — — — — — ITT

Cyclops 14 3 T 11 T T 22 11 T

Harpacticoid 5TT — — — 4TT

Copepodids 52T — — — 91T
Insecta:

Chironomidae L 27 4 1 4 T T 17 1 T

Chironomidae P 18 T 1 11 T T 15 2 3

Coleoptera 5TT — — — ITT

Ceratopogonidae 14 1 2 — — — 11 T T

Other 18 T T 14 T T 15 T T

Araneida — — — — — — ITT

Acari ___ ___ 9TT

Ostracoda 5TT ___ 5TT

Pelecypoda — — — — — — 15 TT

Isopoda — — — — — — ITT

Amphipoda 5TT 4TT 7TT
Eggs:

Zooplankton 9 5 T 14 13 T 19 8 T

Stickleback ___ 4TT 92T

Detritus — — — 1— — 13 — —

'Percentage of stomachs vi^ith item.

2f^ean number of items per stomach examined.

'Item = percent of total bulk units. T = Trace = < 1 organism

or <1%.

eggs of zooplankton and stickleback, and other
miscellaneous taxonomic groups, some of which
are littoral in habitat, made up most of the
remainder of the stomach contents. Males ate
more benthic and epibenthic forms, as well as
detritus (mainly sand and twigs), than did
females. Detritus in individual male stomachs
made up from 10 to 100% of the contents and was

660

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

present in 13 stomachs, compared with 1 for
females. The ingestion of detritus by males is
probably related to its role in nest building and not
to feeding behavior per se.

Diel Feeding Rhythm and Variation in Diet

Despite some size differences in stickleback at
sites A and B (station 1), feeding intensity indices
(food weight/body weight x 100) for stickleback
caught at a specific sampling time were similar
during October and July. Active feeding took place
mainly during postdawn and predusk hours, lead-
ing to two daily alternating feeding and "non-
feeding" periods (Figure 5). Differences between
the mean indices for different times of day in the
October and July series were subjected to the
Kruskal-Wallis test (Siegel 1956) and found to be
significant (October, H = 25.71, 4 df , P = <0.0001;
July, H = 28.97, 7 df , P = <0.001). This periodicity
in feeding was corroborated by the mean number
of organisms present in stomachs at different
times of day (Table 10).

The kinds of organisms consumed and their
importance at different times of the diel cycle are
presented in Table 10 for both the October and
July series. Information for October is based on
stickleback ranging in mean length from 37 to 44
mm. Stickleback examined in the July series were
less uniform in size and ranged in mean length
from 49 to 63 mm.

Considering the important food items, the com-
position of the diet changed through the daily
cycle in October and July (Table 10). In October,
Bosmina and Holopedium occurred in a very high
percentage of the stomachs examined, regardless
of sampling time. In terms of numbers consumed
and bulk units, Holopedium was the dominant
item, especially between 0700 and 1000 h. Between
1300 and 1900 h the relative importance of
Holopedium was reduced somewhat by the in-
creased consumption of Bosmina, Alona, Epis-
chura, and eggs of zooplankton.

In July, Holopedium was the dominant food
organism throughout the daily cycle. Although not
as important as Holopedium in terms of numbers
or bulk, eggs of zooplankton were present in a
large proportion of the stomachs examined, rang-
ing from 40% (0100 h) to 100% (1000 h), with
consumption being greatest in the morning.
Epischura was present in stomachs at most times
of the day, but their contribution to the diet was
highest during peak feeding times.

1000 1300 1600 1900 2200

HOUR OF DAILY CYCLE (PS T )

0100 0400

Figure 5.- Diel fluctuations in feeding intensity of threespine
stickleback in October 1970 (closed circles) and July 1971 (open
circles). The number associated with each datum point repre-
sents sample size. The horizontal bars indicate periods of
daylight and darkness.

Rotifera were present in a large proportion of
the stomachs throughout the diel cycle in October
and July and were numerous compared to most
other items. Their individual small size would tend
to depress their importance as a food item.

Daily Ration and Maximal Meal Size

The described diel fluctuations in feeding in-
tensity indicate that in July at least, consumption
and evacuation occurred alternately over periods
of approximately 6-h duration. On the average, a
particle of food required about 6 h to pass through
the stomach. Stomachs were least full at 0400 h
when the contents amounted to 0.65% of the mean
body weight but they were, on the average, never
devoid of food, suggesting that feeding was con-
tinuous in the population. Freshly ingested or-
ganisms were present in some stomachs even
during dark hours.

Recognizing two periods of consumption and
stomach evacuation each of approximately 6-h
duration, and the presence of "residual" content,
the daily ratio (DR) in July can be calculated by the
formula:

DR = R + Pi + P2

where R = residual content x food particle evac-
uation time.
Pi = Major feeding index - residual content,
P2 = Minor feeding index - residual content.

Substituting actual values indicated in Figure 5,
the food consumed by stickleback in July amount-
ed to (0.65 X 24/6) + (2.80 - 0.65) + (2.45 - 0.65) =

661

FISHERY BULLETIN: VOL. 74, NO. 3

-a

•o

■a

t-.

-tJ

*«

«(-l

.!->

!4H

o
I

Ed
03

(D to (- «3

»"

H CM CO 1-

C*J

HKH^

1- 1- H 1-

in <N h- CM
in T- Tt

T- CO

00 ■* O) H

C\l CO

•-•-^-Si

1- 1- H H "3
o

O) ^ o uo

CVJ

in h* CO CO

CO CO CM CJ)

,- (O CO T-

O CO

C\J CO

<35

CO CM I— <o
CO <J)

ai eg h- CO
CM <o

I- t- h- (- CD I- HI

T- 1- Tt H to K I- I
■* T- eg CD '

CMCOCOH CDCO I— COCD 0|

T- ^— CO r^

H K

m h-

•^ ^. <D

r~.

CJ)

-r K

o o>
CM in

<3>

o
■<>■

•fl;

o •<t

CM m

(D
CD

■*

O CO
CM ^

CO CO L_ CO

CM O

■- H '- CO

in ,, _ ,

1^ CO CM "^ 1^

o in in in

CM r- 1- r^

h~ CD I CD

in •>- CO o 1

- 1

1 ^

m o m o 1

CM >- CM <D 1

in 1

1 S

eg "^ CO

o CO o
eg CO (O

o
CO

CM CO

CJ) »- I CO
T- CO I •*

o o I o

CO C» I CD

CO I- o
CO CO

CO 1- o
CO -^

I o m o
I 00 CO

T- CO I- in
CO

Tf s. H °>
1- ■^

in in in o

^ C7> CJ>

in I-

■* I- I- CM t-
CM

I- H 1- K

in in in in
CO

I- I- I eg

.- (D

I (J)

o o

1- CO

o
1^

I I I

|HH
I in in

H h-

m I in I

I- I-

•fl-l-l-egi- |-l-|eg

H I- H 5

in in o o

1- CO 1- •T

^^ I 5

CM
CO

in in in o o
in in eg

H '- 00 t- I-

in in o o in
1- m in eg CO

o o I m
CM -r- I •<3-

I- CO

in o

y- CO

O 05
CM CO

eg I- K |_ I- H <J)

<D f- o i_ h- H eg

o in o o in in o
CO CM r^ T- CO

H K 1-

2 o K

5^ CM in

CD
00

CO T- H m

in eg

in o o in

•V T- y- Oi

I I Sog S

in
CO

O H
in

CO 1-

o o
in 1-

■<»■ I-

in o m I o o
eg 1- I eg 1-

^^K I ^ I

CO CD I- I CJ>
CO I

o o in I in
in CO >- I CO

I- I- I t^ I-

<0 CO h- I O I-

T- I eg

o in m I o m
eg T- in I ^^

H K eg

in in o

CO Tt

f- K

o o in

y- t

<o
eg

I -

h- 1- H

m in o

T- CM

: O) g CO
w .2

•^ <1> i/} ^ o o

,c:

<» CD '£ S:

CO "O

C CL
O

« 2

■B 2 ^ -^

o S E 3 t^

Q) —
CO (U C
Q. CO O
« 3 -o —

J5 d-E-^

ra o

■o -o

E E
o o

c
o o

CO
O OJ

Z20

c '^
CO o

So «?: OQ Q 3:

O CO

E
E

_ m ra ■"

CO 0) .5 T3 C

■O CO != o 5

o .c i5 a <D r-

o " — - "^ -c

^N tU ri *" _ ^*

3
<D C E O-

■5

is S ?'t;

oicc o

aS o S--2 - o - - " ' l: i: o>-e » o " <= E S o S wo.2

SOiLSQ.CoJi^-C;C(U«a,(0^30,=0«i>Oogg--5
Q-OOQUjX SoOO ^j:00|-Q. E ,j. c^S^OQQI

O - UJO CO . S SjOC O

o 1* »-

*" _ o ' C "^

6^ &° ^^ss;^

3 Q. O

Q. w e-.s

<o o

o
O

o m -
z20

>•
a

5» to o Ss-2 5. CO
g-Z O O Q Uj X

E J>: to

o c .Q

c CO »

■££«§ =

O O g>N W
HI

c
■o "
o 2
o E
to >_
l; i: -c o

£ O u. I

O CO

c
o

Q.
CO

c
to

C3)

(1)

E
c

CJ
V

662

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

6.55% of their body weight. Some digestion would
have occurred during consumption so this is a
minimal value.

For October, failure to obtain feeding indices
between 1900 and 0700 h over the diel cycle pre-
cluded similar estimation of the daily ration.
However, if the residual content is assumed to be
0.65% of the body weight during periods lacking
observations, the daily food consumption can be
estimated to be 2.60 + 2.90 + 2.30 = 7.80% of the
mean body weight.

Estimates of maximum meal size were obtained
by plotting feeding indices for only those fish
which were judged to have "full" stomachs during
the postdawn feeding period (i.e., the most inten-
sive feeding time of day) against length (Figure
6A). Data for stickleback in July were used

o
o

o
o
m

I
o
u

a
o
O

I
u

o
o
o

»' = 8.l8-0.077Jir
A =-0.788

. Hatching size

30 40 50

FISH LENGTH (mm)

Figure 6.-The relation between maximum size of single meal
(A) and weight of stomach contents (B) with length of threespine
stickleback.

because of their wide range in length. Despite
considerable individual variation between fish of
the same length obviously feeding intensity was
inversely related to length (r = -0.788, df = 26, P
<0.01). From the regression line fitted by the
method of least squares, it can be predicted by
extrapolation that, on the average, larval stick-
leback, which measure approximately 8 mm upon
hatching, consume 7.5% of their body weight in a
single meal, and that consumption in relation to
body weight decreases 0.8% per 10 mm increase in
length. As would be expected, large fish in a single
meal eat more than do small fish and the relation-
ship is of the positive exponential form (Figure
6B).

For stickleback in October and July (assuming
mean lengths of 40 mm and 60 mm, respectively),
the average meal size was approximately 5 and
3.5% of their body weight, respectively. Assuming
two feeding periods per day, the daily ration
becomes 10 and 7% of body weight. These values
are in reasonable agreement with daily ration
estimates based on diel fluctuations in stomach
contents.

Discussion

During 1970 and 1971, the first 2 yr of a fer-
tilization program attempting to increase sockeye
salmon production in Great Central Lake, stick-
leback were observed to feed on a variety of
organisms with planktonic crustaceans (cladocer-
ans and copepods) and insects (chironomid pupae
and larvae), to a lesser degree, being the main food
organisms. These findings are consistent with
observations on food of stickleback in a variety of
freshwater habitats made by other investigators
(Hartley 1948; Hynes 1950; Greenbank and Nelson
1959; Rogers 1968). From a trophic standpoint, the
species is a secondary consumer.

The literature on feeding of fishes in both
laboratory and in nature is replete with evidence
that consumption is influenced by a multitude of
factors. In the present study eff"ort was focussed on
examining seasonal and diel changes in feeding
habits, possible influencing factors being limited
to size and sexual maturity.

The most pronounced feature observed in the
feeding of stickleback was the seasonal change in
the importance of different kinds of organisms
consumed. Although the food resource was not
sampled in conjunction with the food studies, some
general comments on food availability and selec-

663

FISHERY BULLETIN: VOL. 74, NO. 3

tivity by stickleback in 1970 can be made using
results of zooplankton studies by LeBrasseur and
Kennedy (1972) (Figure 7). A more precise method
of measuring the use of major planktonic forms in
relation to availability would have been the em-
ployment of Ivlev's( 1961) "electivity index," taking
into account the comments of O'Brien and Vin-
yard (1974) regarding distribution of predator and
prey. Bosmina, Holopedium, and Diaptomus were
consumed approximately in relation to their
abundance, although in the early part of the year
relative utilization was highest for Boxmina.
Cijclops and Bosmuia were approximately equally
abundant and exhibited somewhat similar sea-
sonal fluctuations but utilization of Bosmina was
sharply restricted during July and early August
whereas Cyclops was relatively unutilized
throughout the summer. Consumption of Epis-
chura, a less abundant form which occurred
mainly between May and September, was highest
in June during the early part of the "bloom."

The reasons for the apparent differences in the
relative utilization of the major food items would
appear to differ. The shift from Epischura, despite
rather uniform abundance, to smaller organisms,
mainly Holopedium and Bosmina, through the
season may be due to the decrease in average size
of stickleback that occurred in midsummer. Epis-
chura, which equals 11 bulk units compared with 3
and 1 for Holopedium and Bosmina, respectively,
may have been too large an item to be consumed
by the majority of stickleback present after July.
Greenbank and Nelson (1959) and Rogers (1968)
observed that feeding habits of G. aculeatus in
Alaskan lakes changed through the summer and
differed between individuals of different size. The
disparity in relative utilization of Bosmina and
Cyclops, which were of comparable abundance and
individual size, cannot be thus explained. Rather,
it would appear that the difference in their dietary
importance may be explained by differences in
spatial distribution affecting availability: Cyclops

1000

DIAPTOMUS

HOLOPEDIUM

OUG SEPT OCT NOV DEC

(r
LU
m
5

o
<

MAY JUNE JULY AUG SEPT OCT NOV DEC MAY JUNE JULY AUG SEPT OCT NOV DEC

Figure 7.— Seasonal change in the biomass (unbroken line) of important prey species for threespine stickleback and the average number
present per stomach. Graphs representing biomass were taken from LeBrasseur and Kennedy (1972) and are shown in logarithmic scale.

664

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

were hypolimnetic whereas Bos^mina were mainly
epilemnetic (LeBrasseur and Kennedy 1972,
Figure 2). If temperature influences their dis-
tribution as is suggested by their distribution in
relation to the thermocline, one might reasonably
infer that Cyclops was less available than Bosm ina
to stickleback inhabiting the littoral and near-
shore areas where water temperatures generally
are highest. Consumption of Holopedium in-
creased rapidly as the summer progressed. Con-
sumption of Diaptomus increased during August
and September. The appearance of these approx-
imately similar sized species in the diet of stick-
leback paralleled their occurrence in population
succession and maximum abundance in the rela-
tively warm surface waters.

The diel feeding rhythm observed during July
and October has not been described for G. actdea-
tus in fresh water but the pattern is exemplary of
feeding periodicities described for a variety of
freshwater and marine fishes. The association of
peak feeding with postdawn and predusk periods
in summer and late fall when the number of
daylight hours differs suggests that feeding is
light-dependent.

The literature on meal size and daily ration for
G. aculeatus is rather scant considering the
number of studies on the feeding biology of the
species. Krokhin (1957) using the O2 consumption
method estimated that stickleback averaging 4.5 g
in summer (August) consumed 5.1% of their body
weight daily. Beukema (1968) feeding stickleback
(2.5 g mean weight) Tubifex worms concluded that
the contents of a well-filled stomach equalled 5.5%
of the body weight, and that daily intake amount-
ed to 12% of the body weight. Beukema recognized
that the daily ration was rather high for adult fish
and suggested that rapid digestibility of the food
offered may have been responsible for the rather
high food intake value obtained. The mean daily
ration estimated in the present study from diel
feeding rhythm curves for stickleback in July
(mean length = 55 mm, mean weight = 2.4 g) and
in October (mean length = 39 mm, mean weight =
0.7 g) was 6.5% and 7.8%, respectively, of their body
weight. These estimates are only slightly less than
those derived by doubling the maximum meal size
of individuals of corresponding length (see Figure
6A), namely 7.8% and 10%. Considering that food
intake is influenced by several factors such as size,
physiology and behavior of individual, food de-
privation, previous meal size, temperature, and
prey digestibility (Darnell and Meierotto 1962;

Davis and Warren 1968; Keast and Welsh 1968;
Swenson and Smith 1973), one may conclude that
the mean daily rations determined in this study
are in close agreement with those obtained from
experimental studies.

FEEDING RELATIONSHIP
BETWEEN STICKLEBACK AND
JUVENILE SOCKEYE SALMON

Information on competition between stick-
leback and juvenile sockeye salmon for food must
be based on samples of each species from the same
catch. Further, it must be assumed that in-
dividuals of each species taken together fed in the
same area. In 1970, 7 of 105, or 6.6%. of the sets
yielded both species. Sockeye salmon equalled 5%
of the two species combined. In 1971, the two
species were caught together in 18 of 89, or 20.2%
of the sets, and sockeye salmon equalled 2.2% of
the combined catch.

Sockeye salmon and stickleback caught in the
littoral zone in October 1970 and May-July 1971
were used in this comparative study (Tables 11 and
12). Only catches containing 5 or more individuals
of each species were considered and a maximum
number of 10 individuals of each species was
examined from any one catch. For convenience,
the catches were grouped according to the follow-
ing time periods: October 1970, May-June 1971,
and July 1971.

Stickleback through this period increased in
average size as a result of seasonal growth. By
contrast, sockeye salmon, although larger,
decreased in average size. This decrease in size
reflects the emigration from the lake of the larger
individuals as smolts in the following spring. The
relatively high percentage (20%) of stickleback
with empty stomachs in July can be explained by
the presence of the gravid females.

In general, stickleback and young sockeye sal-
mon taken together exhibited considerable die-
tary overlap (Tables 11 and 12). Stomach contents
of sockeye salmon were treated and analysed in
accordance with methods used for stickleback. The
degree of similarity in diet during each period was
determined from occurrence data using Spear-
man's rank correlation coeflficient, r, (Siegel 1956).
The r, value indicates agreement in rank of food
items and can range from -i- 1.0 for complete
agreement to -1.0 for total disagreement. The
tests were restricted to items which were not
rendered unidentifiable through digestion and

665

FISHERY BULLETIN: VOL. 74, NO. 3

Table 11. -Stomach contents of threespine stickleback in Great Central Lake, October 1970-July 1971.

Date

October 1970

May-June 1971

July 1971

No. examined

Percent empty

Size range (mm)

27-76

40-86

42-86

Mean length (mm)

39.5

54.0

59.8

Percent of

Average

% Of

Percent of

Average

% of

Percent of

Average

%of

stomachs

no. 2 per

total

stomachs

no. per

total

stomachs

no. per

total

Food item

with item

stomach

bulk

with item

stomach

bulk

with item

stomach

bulk

Rotlfera

Cladocera:

Holopedium

100

Bosmina

100

Daphnia'

—

Alona

Copepoda:

Epischura

109

Diaptomus

—

Cyclops

Copepodids

Nauplii

—

Harpacticoid

Insecta:

Chironomjd L

Chironomid P

Other

Mites

Eggs:

Zooplankton

—

. 26

Fish

—

Other:

Amphipoda

—

Pelecypoda

—

Ostracoda

—

Unidentifiable

0.0

52.0

72.0

Total

100.0

'Mainly D. pulex.

'Based on stomachs in whicfi condition of contents permitted counts of various dietaries.

3T = Trace = < 1% of bulk.

which were present in at least 10% of the stomachs
of one or the other foraging species. Infrequent
ties in rank were broken in favor of the larger food
item.

The r, values for May-June and July samples
were significant at P = 0.05 but that for October
was not (Table 13). In October Bosmina, Cyclops,
and copepodids were common items in the diet of
stickleback compared to the larger Epischura and
Holopedium in the sockeye salmon diet. A possible
explanation for the diff'erence between stickleback
and sockeye diets in October may be that larger
predators feed on larger prey: in October, sockeye
salmon on the average measured 74.6 mm, stick-
leback 39.5 mm.

The observed dietary overlap indicates the
existence of potential competition between stick-
leback and sockeye salmon for food in May-June
and July. Accurate assessment of actual competi-
tion is contingent not only on information on food
and feeding habits of the two foraging species but
on other factors, such as their temporal and spatial
associations during different life history stages
and their abundance and growth in relation to
food supply. For this study, data essential for

quantitative assessment of competition during
different seasons are inadequate or unavailable,
although competition in winter is precluded by the
apparent absence of stickleback. It is known
however that when the two species occur together
it is near shore or in the littoral zone, and that
relative to stickleback sockeye salmon are few in
number: sockeye salmon are almost the exclusive
inhabitants of the limnetic zone (D. Robinson,
pers. commun.). From the distribution patterns of
the two species, it can be inferred that stickleback
in Great Central Lake are not serious competitors
of sockeye salmon for food despite their similarity
in diet. Additionally, during this study the zoo-
plankton abundance had increased substantially
as a result of nutrient additions (LeBrasseur and
Kennedy 1972) and the growth rate in sockeye
salmon was faster than that observed under
untreated lake conditions (Barraclough and
Robinson 1972). However, in lakes where both
species are abundant and overlap extensively in
spatial distribution, utilization of a common food
resource may affect production of one or both of
the foraging species, especially during periods of
reduced or limited food supply.

666

MANZER: DISTRIBUTION AND FOOD OF STICKLEBACK

Table 12.-Stomach contents of young sockeye salmon in Great Central Lake, October 1970-July 1971.

Date

October 1970

May-June 1971

July 1971

No, examined

Percent empty

Size range (mm)

58-95

28-82

37-75

Mean length (mm)

74.6

63.0

60.0

Percent of

Average

% of

Percent of

Average

% of

Percent of

Average

%of

stomachs

no.' per

total

stomachs

no. per

total

stomachs

no. per

total

Food item

with item

stomach

bulk

with item

stomach

bulk

with item

stomach

bulk

Rotifera

—

Cladocera:

Holopedium

360

Bosmina

Daphnia'

—

Alona

—

Copepoda:

Epischura

100

234

Diaptomus

—

Cyclops

Copepodids

Nauplii

—

Harpacticoid

—

Insecta:

Chironomid L

—

Chironomid P

—

Diptera (pupae & adult)

Araneida

—

Remains

—

Other

—

Eggs - Zooplankton

Unidentifiable

21.0

57.0

62.0

Total

100.0

'Mainly D. pulex.

'Based on stomachs in which condition of contents permitted counts of various dietaries.

3T = Trace = < 1% of bulk.

Table 13.-Similarity in diet of threespine stickleback and young
sockeye salmon in the littoral zone, Great Central Lake, October
1970-July 1971. Similarity was measured by Spearman's rank
correlation coefficient (r,). Rotifers are excluded from the
calculations.

Time period

No. food items
considered

October 1970
May-June 1971
July 1971

-0.068
0.629*
0.738*

*Significant at P = 0.05.

ACKNOWLEDGMENTS

J. C. Mason and R. J. LeBrasseur read the
manuscript and offered valuable suggestions for
its improvement.

LITERATURE CITED

Barraclough, W. E., and D. G. Robinson.

1972. The fertilization of Great Central Lake. III. Effect on
juvenile sockeye salmon. Fish. Bull., U.S. 70:37-48.
BUKEMA, J. J.

1968. Predation by the three-spined stickleback {Gasteros-
teus aculeatus L.): The influence of hunger and experi-
ence. Behaviour 31:1-126.
Craig-Bennett, A.

1931. The reproductive cycle of the three-spined stickleback,

Gasterosteus aculeatus, Linn. Philos. Trans. R. Soc. Lond.,
Ser. B, Biol. Sci. 219:197-279

Darnell, R. M., and R. R. Meierotto.

1962. Determination of feeding chronology in fishes. Trans.
Am. Fish. Soc. 91:313-320.

Davis, G. E., and C. E. Warren.

1968. Estimation of food consumption rates, hi W. E.
Ricker (editor). Methods for assessment of fish production
in fresh waters, p. 204-225. IBP (Int. Biol. Programme)
Handbook No. 3. Blackwell Sci. Publ., Oxf., Engl.

Greenbank, J., and P. R. Nelson.

1959. Life history of the threespine stickleback Gasterosteus
aculeatus Linnaeus in Karluk Lake and Bare Lake,
Kodiak Island, Alaska. U.S. Fish Wildl. Serv., Fish. Bull.
59:537-559.

Hagen.D. W.

1967. Isolating mechanisms in threespine sticklebacks
(Gasterosteus). J. Fish. Res. Board Can. 24:1637-1692.
Hagen, D. W., and L. G. Gilbertson.

1972. Geographic variation and environmental selection in
Gasterosteus aculeatus L. in the Pacific Northwest,
America. Evolution 26:32-51.
Hagen, D. W., and J. D. McPhail.

1970. The species problem within Gasterosteus aculeatus on
the Pacific coast of North America. J. Fish. Res. Board
Can. 27:147-155.
Hartley, P. H.T.

1948. Food and feeding relationships in a community of
fresh-water fishes. J. Anim. Ecol. 17:1-14.
Hynes, H.B.N.

1950. The food of fresh-water sticklebacks (Gasterosteus
aculeatus and Pygosteus pungitius), with a review of

667

FISHERY BULLETIN: VOL. 74, NO. 3

methods used in studies of the food of fishes. J. Anim.

Ecol. 19:36-58.
IVLEV, V. S.

1961. Experimental ecology of the feeding of fishes.

(Translated from the Russian by Douglas Scott.) Yale

Univ. Press, New Haven, 302 p.
Keast, a. H., and L. Welsh.

1968. Daily feeding periodicities, food uptake rates, and
dietary changes with hour of day in some lake fishes. J.
Fish. Res. Board Can. 25:1133-1144.

Krogius, F. v., and E. M. Krokhin.

1957. Causes of the fluctuations in abundance of sockeye
salmon in Kamchatka. (Tr. Probl. Temat. Soveshch. Zool.
Inst. Akad. Nauk SSSR, 1956, 6:144-149.) Fish. Res. Board
Can. Transl. Ser. 92, 5 p.
Krokhin, E. M.

1957. [Determination of the daily food ration of young
sockeye and three-spined stickleback by the respiration
method.] [In Russ.] Izv. Tikhookean. Nauchno-Issled. Inst.
Rybn. Khoz. Okeanogr. 44:97-110.
LeBrasseur, R. J., AND 0. D. Kennedy.

1972. The fertilization of Great Central Lake. II. Zoo-
plankton standing stock. Fish. Bull., U.S. 70:25-36.
Manzer, J. I.

1971. Data on catch, size and food of the three-spined
stickleback in Great Central Lake, British Columbia in

1970. Fish. Res. Board Can. Manuscr. Rep. 1145, 43 p.

1972. Data on catches, size, and food of the three-spined
stickleback in Great Central Lake, British Columbia, in

1971. Fish. Res. Board Can. Manuscr. Rep. 1166, 52 p.
Markovtsev, V. G.

1972. Feeding and food relationships of young sockeye and
three-spine stickleback in Lake Dalnee. Izv. Tinro
82:227-233.

McPhAIL, J. D., AND C. C. LiNDSEY.

1970. Freshwater fishes of northwestern Canada and Alas-
ka. Fish. Res. Board Can., Bull. 173, 381 p.
Meek, A.

1916. The migrations of fish. Edward Arnold, Lond., 427 p.
Miller, R. R., and C. L. Hubbs.

1969. Systematics of Gasterosteus aculeatiis, with par-
ticular reference to intergradation and introgression
along the Pacific Coast of North America: A commentary
on a recent contribution. Copeia 1969:52-69.

MULLEM, p. J. VAN, and J. C. VAN DER VlUGT.

1964. On the age, growth and migration of the anadromous
stickleback Gasterosteus aculeatus L. investigated in
mixed populations. Arch. Neerl. Zool. 16:111-139.
Narver, D. W.

1969. Phenotypic variation in threespine sticklebacks {Gas-

terosteus aculeatus) of the Chignik River system, Alas-
ka. J. Fish. Res. Board Can. 26:405-412.
1970. Diel vertical movements and feeding of underyearling
sockeye salmon and the limnetic zooplankton in Babine
Lake, British Columbia. J. Fish. Res. Board Can.
27:281-316.
O'Brien, W. J., and G. L. Vinyard.

1974. Comment on the use of Ivlev's electivity index with
planktivorous fish. J. Fish. Res. Board Can. 31:1427-1429.
Parsons, T. R., C. D. McAllister, R. J. LeBrasseur, and W. E.
Barraclough.
1972. The use of nutrients in the enrichment of sockeye
salmon nursery lakes (a preliminary report). In M. Ruivo
(editor). Marine pollution and sea life, p. 519-525. Fishing
News (Books) Ltd., Surrey and Lond., Engl.
Parsons, T. R., K. Stephens, and M. Takahashi.

1972. The fertilization of Great Central Lake. I. Effect on
primary production. Fish. Bull, U.S. 70:13-23.

Ricker, W. E.

1937. The food and the food supply of sockeye salmon
{Oncorhynchus nerka Walbaum) in Cultus Lake, British
Columbia. J. Biol. Board Can. 3:450-468.
Rogers, D. E.

1968. A comparison of the food of sockeye salmon fry and
threespine sticklebacks in the Wood River lakes. In R. L.
Burgner (editor). Further studies of Alaska sockeye
salmon, p. 1-43. Univ. Wash. Publ. Fish., New Ser. 3.
Rogers, D. E., M. 0. Nelson, J. J. Pella, and R. L. Burgner.
1963. Relative abundance and distribution of fish species in
Lake Aleknagik. Res. Fish. Fish. Res. Inst. Univ. Wash.
1962:14-15. (Univ. Wash. Coll. Fish. Contrib. 147.)
Ruggles, C. P.

1965. Juvenile sockeye studies in Owikeno Lake, British
Columbia. Can. Fish Cult. 36:3-21.
Siegel, S.

1956. Nonparametric statistics for the behavioral sci-
ences. McGraw-Hill Book Co., Toronto, 312 p.
Simpson, G. G., and A. Roe.

1939. Quantitative zoology. McGraw-Hill Book Co., Inc.,
N.Y.,414p.
Swenson, W. a., and L. L. Smith, Jr.

1973. Gastric digestion, food consumption, feeding pe-
riodicity, and food conversion efl[iciency in walleye
{Stizostedion vitreum vitreum). J. Fish. Res. Board Can.
30:1327-1336.

Windell,J.T.

1968. Food analysis and rate of digestion. In W. E. Ricker
(editor), Methods for assessment of fish production in
fresh waters, p. 197-203. IBP (Int. Biol. Programme)
Handbook No. 3. Blackwell Sci. Publ., Oxf., Engl.

668

PREDATOR-PREY RELATIONSHIP BETWEEN PACIFIC HERRING,

CLUPEA HARENGUS PALLASI, LARVAE AND

A PREDATORY HYPERIID AMPHIPOD, HYPEROCHE MEDUSARUM^

Hein von Westernhagen and Harald Rosenthal-

ABSTRACT

Predatory efficiency of Hyperoche medusarum (Hyperiida, Amphipoda) on yolk-sac larvae of Pacific
herring, Clupea harengus pallasi, was studied in the laboratory under continuous light conditions: 1, 5,
10, and 50 herring larvae were exposed to 1, 2, 4, 8, and 16 hyperiids in 500-ml beakers. It was found that
the number of attacked larvae per unit time increased with rising predatory and/or prey density.
Individual mean predation rate was found to decline with increasing predator as well as prey densities,
prolonged exposure times, and the presence of alternative prey.

Aside from starvation (Sette 1943; Schnack 1972),
one major cause of mortality in marine fish larvae
is assumed to be predation (Stevenson 1962), the
predators frequently being crustaceans, as de-
scribed by Garstang (1900), Lebour (1925), Davis
(1959), Lillelund (1967), Rosenthal (1967), Kabata
(1970), Lillelund and Lasker (1971), Theilacker and
Lasker (1974), and others. The pelagic hyperiid
amphipod Hyperoche medusarum occurs com-
monly off the Oregon coast (Lorz and Pearcy 1975),
in Californian waters (Hurley 1956), in the North
Atlantic (Shoemaker 1930; Bowman et al. 1963;
Dunbar 1963), in the North Sea (Sars 1895; Evans
and Sheader 1972), and in New Zealand waters
(Hurley 1955). In British Columbia waters it occurs
commonly in the upper layers (<30 m) of the water
column (Bowman 1953), and in Departure Bay
(Vancouver Island) its juveniles are frequently
found clinging to the exumbrellae of hydromedu-
sae (Westernhagen 1976).

The cooccurrence of large numbers of juvenile
H. medusarum with newly hatched larvae of the
Pacific herring, Clupea harengus pallasi, was
incidentally discovered in 1974 at the pier of the
Pacific Biological Station, Departure Bay. Field
observations indicated that Hyperoche juveniles
preyed on herring larvae and occasionally on other
fish larvae. Since this was the first record on a
possible predator-prey relationship between H.

^This study was sponsored by the international bureau of the
Gesellschaft ftir Kernenergieverwertung in Schiffbau and
Schiffahrt in connection with the German-Canadian agreement
on scientific and technical cooperation.

-Biologische Anstalt Helgoland (Zentrale, 2 Hamburg 50,
Palmaille 9, Germany (Federal Republic of Germany)).

medusarum and marine fish larvae, this study was
initiated to shed some light on the predatory
efficiency of this amphipod.

MATERIAL AND METHODS

For prey, yolk-sac larvae (8.0-9.5 mm TL (total
length)) of the Pacific herring incubated in the
laboratory were used. Immature H. medusarum
(1.48-1.80 mm TL) which had aggregated beneath
a light at night were caught with a pail and
separated from other plankton organisms with a
large bore pipette.

Experiments were performed in filtered
seawater in 500-ml beakers (salinity 28''/oo; tem-
perature 9°C; constant light). The water surface of
the beakers was covered with 300-/xm mesh size
nylon gauze in order to keep the amphipods from
breaking through the surface. Because Hyperoche
specimens in their natural habitat were occasion-
ally found resting on the exumbrellae of medusae,
a strip of nylon gauze (50x20 mm) hanging from
the surface cover provided attachment for the
amphipods when needed.

Different numbers of herring larvae 1, 5, 10, and
50 were exposed to 1, 2, 4, 8, and 16 hyperiids for
three exposure periods (2, 4, and 8 h). The number
of replicates for all predator/prey ratios were 4, 6,
and 5 for the 2-, 4-, and 8-h exposure periods. Some
additional experiments with 6- and 10-h exposure
periods were used for the computation of a mean
attack rate on the basis of HI h of observation.
Eleven trials using 25 herring and 25 flatfish
larvae with 16 amphipods were also conducted.
One additional control vessel (50 herring larvae, no

Manuscript accepted February 1976.
FISHERY BULLETIN: VOL 74, NO. 3, 1976.

669

FISHERY BULLETIN: VOL. 74, NO. 3

amphipods) was used for each exposure period.
Mortality of larvae was measured every 2, 4, and 8
h by means of direct counts. All remaining larvae
were removed, and healthy, wounded, and dead
larvae were counted. The original number of
herring larvae then was restored before a new
experiment was started. Between experiments,
the hyperiids were provided with food in order to
reduce cannibalism.

RESULTS

Swimming and Feeding Behavior of
Hyperoche medusarum

Two modes of swimming were observed: 1)
quick darting movements with the body kept in a
horizontal position; and 2) slow hovering, in which
the body was held in a vertical position, and the
pleopods beat continuously. The latter mode of
swimming was maintained for periods longer than
20 min, but the speed of swimming was slow
(about 10 cm/min). It was only during swimming
that Hyperoche would, by chance encounter, cap-
ture a herring larva. The amphipod usually
grasped the tail but attacks at the head and the
midportion of the larva also occurred. An attacked
larva did not survive long. The larva attempted to
shake the amphipod off for a few minutes, then
sank to the bottom where it was eaten by the
Hyperoche. Larvae were not always consumed.
Frequently, amphipods clung to a larva for only a
few seconds but the wound inflicted during this
process inevitably lead to the death of the larva.
Wounded larvae which were removed after ter-
mination of the experiment never survived for
more than 4-5 h when kept in separate beakers.

Between swimming activities, the amphipods
either remained on the bottom (probably an ar-
tifact due to the small size of the beakers-in large
enough containers Hyperoche juveniles swam
continuously (Westernhagen 1976)), or attached
themselves with the posterior pereiopods to the
nylon gauze provided in the beakers for this
purpose and assumed a resting posture. This
posture has been described for Hyperia galba by
Bowman et al. (1963) and for Hyperoche ynedusar-
iim by Evans and Sheader (1972). The latter
authors defined the posture as an "inactive curled
position head and urus directed away from the
substrate it (the animal) sits on." Larvae that
bumped into resting amphipods were not pursued
or captured.

Predatory Efficiency of
Hyperoche medusarum

The results of all experiments were summarized
and presented as the number of larvae attacked
per hour at different predator and prey densities
(Figure 1). The number of wounded and killed
larvae was dependent on two factors, the density
of the herring larvae and the density of hyperiids.
With increasing numbers (predator or prey) larval
mortality per hour increased, reaching a value of
more than two larvae killed or wounded per hour at
the 16 Hyperoche and 50 herring larvae
combination.

The number of larvae attacked per unit time (1
h) depended to a great extend on the duration of
the experiment (Figure 2). Experiments with
short exposure times (2 h) yielded for all larvae
and hyperiid combinations higher attack rates per
hour than experiments lasting 4 or 8 h. The mean
predatory efficiency of the hyperiids was affected
also by their density in each beaker. The number
of larvae attacked per unit time decreased as the
density of the predators increased (Figure 3). It is
for this reason that there are different values for
the number of herring larvae attacked per hour by
one hyperiid (Figure 4), (A) for the observation of
one single hyperiid, and (B) for the calculated
mean predation rate of a hyperiid from exper-
iments with 1, 2, 4, 8, and 16 Hyperoche. Yet both
curves show that an increase of a potential prey in
a constant environment beyond a certain density

01 5 10

Number of h#rr.ng larvae / 500 ml

Figure 1. -Predatory efficiency of Hyperoche medusarum on
yolk-sac larvae of Clupea harengus pallasi at different predator
and prey densities. Water temperature: 9°C; total observation
time: 111 h; observation periods: 20.

670

WESTERNHAGEN and ROSENTHAL: PREDATOR-PREY RELATIONSHIP

5.0
1 larva /500ml

</ZZl7i

0 12 4 8

Number of Hyperoche/500ml

0 12 U

Num ber of Hyperoche/ 500ml

Figure 2.-Mean number of yolk-sac larvae of Clupea harengus pallasi attacked by Hyperoche medusarum after different exposure

times. Water temperature 9°C.

does not necessarily lead to a corresponding in-
crease in predation. At herring larvae densities of
5/500 ml and 10/500 ml, one individual hyperiid
attacked 0.1 larvae/h and 0.16 larvae/h, respec-
tively. At 50 larvae/500 ml the attack rate was 0.45
larvae/h. Assuming a linear increase in attack
rate, we would have expected rates of 1.0 and 0.8
larvae/h.

Alterations in predation rates of Hyperoche
were obtained when heterogenous prey was
offered (25 herring larvae -i- 25 flatfish larvae), and
Figure 3 shows that predation on larvae was
remarkably reduced. Of the 0.07 larvae attacked
per hour by one hyperiid, 0.055 (78%) were herring
larvae and 0.015 (22%) flatfish larvae, thereby
showing a pronounced preference for herring.

DISCUSSION

Figure 1 shows a clear, direct relationship
between number of attacked larvae and both

larval and hyperiid density. Increase in larval as
well as predator density lead to increasing attack
rates per hour. Because searching and contacting
are random, this response was expected and has
been described by Murdoch (1971) for predator-
prey interaction. That relatively more larvae are
attacked per hour during short exposure periods
than during long ones (Figure 2) can be partially
explained by a rapid thinning out eff'ect on prey in
confined containers, a problem discussed by Mur-
doch (1969) for the predation of Thais and
Acanthina on Mytilus and Balanus. These data
suggest that short observation periods are prefer-
able in experiments of this type, a point
frequently neglected in experiments with expo-
sure times of 20 and more hours (Lillelund 1967;
Lillelund and Lasker 1971; Theilacker and Lasker
1974; Ambler and Frost 1974), leading to an under-
estimate of the actual possible predation rate.
An additional factor may be the degree of satia-
tion, which could be shown for invertebrates to

671

FISHERY BULLETIN: VOL. 74, NO. 3

0.5 -1

O.A

a;
O
>

-o
o

"o
"o

D
Z

0.3

0.2

0.1 - o

• _• 1 larva /500ml

o — o 5 larvae/ 500ml

»-* 10 larvae/ 500ml

-i-a 50 larvae/ 500ml

♦ 50 larvae (mixed)

1 — I \ 1 \

0 1 2 ^ 8 16

Number of Hyperoche/500ml

Figure 3. - Mean number of yolk-sac larvae of Clupea harengus
pallaxi attacked by Hyperoche medusa rum during 1-h exposure
time in an experimental volume of 500 ml at different lanal
concentrations. Water temperature 9°C. the "mixed" trial was
provided with 25 herring and 25 flatfish lanae (11 replicates, 64 h
total obser\'ation time).

reduce the rate of predation (Holling 1966; Brandl
and Fernando 1974).

It became evident through the experiments that
predation rate was also influenced by the number
of predators present in an experimental beaker
(Figure 3). Calculated mean individual predation
rates in experiments using 50, 10, and 5 larvae
decreased as the number of hyperiids in one
container increased. Lillelund (1967) observed the
same phenomenon in his experiment using cy-
clopids preying on larvae of Osvierus eperlanus,
and Salt (1967) noted the same trend in exper-
iments using the predatory protozoan Woodruffia
metabolica preying on Paramecium. We con-
sider this phenomenon an artifact caused by more
than one predator feeding on the same prey, an
event frequently observed at higher predator
densities. This is unlikely to occur in the natural
habitat, because a herring larva once killed by its
predator which is still attached to it would sink
down to the bottom out of the reach of the other
Hyperoche.

0,5 -,

0.4 -

SO.3

5 0.2

0.1 -

Number of herring larvae / 500 ml

Figure 4. -Mean number of yolk-sac larvae of Ctupca hannquK
pallasi attacked per hour by one Hyjwrochi' at different larval
densities:

A. data of actual experiments with single hyperiids;

B. data obtained from mean values for experiments with 1, 2, 4, 8,
and 16 hyperiids/50() ml.

The number of herring larvae attacked did not
increase proportionally with an increase of her-
ring larvae available for the predators (Figure 4).
This phenomenon has been termed "functional
response" (type 2 response) by Holling (1966), and
is believed to occur commonly in preying inverte-
brates. Similar responses are displayed by the
house cricket, Acheta dojuei^ticus (Pimentel and
Cranston 1960); Podiscus maculive7itri.s (Morris
1963); Acanthina sp. (Murdock 1969); Tortanus
discaudat^is (Ambler and Frost 1974); and En-
phausia pacifica (Theilacker and Lasker 1974). In
a typical functional response curve, the number of
prey eaten or attacked per predator increases to
reach or approach a maximum at an asymptote
(Murdoch 1971). Although the curves in Figure 4 do
not yet approach an asymptote due to insuflficient
prey density, the trend towards a maximum at-
tacking rate at a given prey density is noticeable.

Hyperoche medumrum exposed to two species
of fish larvae clearly discriminated disproportion-
ately between these two. In Figure 3, the total
number of larvae attacked in trials providing
alternate prey at equal densities is given as 0.7
individuals/h. Of these, 0.055 were herring larvae
and 0.015 flatfish larvae. Discrimination between
two prey species, which is likely to occur only in
predators with searching and food selection

672

VVESTERNHAGEN and ROSENTHAL: PREDATOR-PREY RELATIONSHIP

behavior (Murdoch and Marks 1973), might be
either caused by different distribution of prey
species (Oaten and Murdoch 1975), differences in
palatibility (Rolling 1965), avoidance behavior of
the prey, or conditioning and/or training of the
predator (Murdoch 1969; Oaten and Murdoch 1975)
in cases of weak preferences.

Although generally H. medusaritm was con-
sidered to lead a parasitic life on medusae (Sars
1895) such as Cyanea capillata (Bowman et al.
1963) or Plenrohrachia pileus (Evans and Sheader
1972), the results of our experiments show that
even in the presence of alternate prey this am-
phipod displays considerable predation on herring
larvae.

Unlike another carnivorous hyperiid, Pam-
themisto gaudichaudi, which hunts moving plank-
ton visually (Sheader and Evans 1975), H. medu-
sarum depends on random encounters with its
prey. Many carnivorous copepods display the same
behavior (Dziuban 1937; Fryer 1957; Lillelund
1967; Rosenthal 1972; Brandl and Fernando 1974;
Ambler and Frost 1974). This mode of hunting
requires a relatively high density of prey in-
dividuals which at times is provided by the enor-
mous numbers of newly hatched herring larvae.
During this investigation, herring larvae density
during the day at the water surface was
frequently above 2 larvae/ 100 cm'-' (direct obser-
vations). Simultaneous mass occurrences of H.
medusarum suggest that the amphipods could
possibly contribute considerably to herring larvae
mortality, especially since conditioning to abun-
dant prey organisms is comprehensible as could be
shown by Sheader and Evans (1975) for P. gaudi-
chaudi and its feeding on fish larvae. In fact
stomach-content analyses of H. medusarum cap-
tured during this study period revealed that the
amphipods had eaten considerable amounts of fish
larvae (Westernhagen 1976).

ACKNOWLEDGMENTS

We are indebted to D. F. Alderdice for providing
laboratory space, to J. Klinckmann and G. Fiir-
stenberg for expert technical assistance and to M.
Blake for advice on the preparation of the
manuscript.

LITERATURE CITED

Ambler, J. W., and B. W. Frost.

1974. The feeding behavior of a predatory planktonic

copepod, Tarfanus dincaiidatus. Limnol. Oceanogr.
19:446-451.

Bowman, T. E.

19.53. The systematics and distribution of pelagic am-
phipods of the families Vihiliidae, Paraphronimidae,
Hyperiidae, Dairellidae, and Phrosinnidae from the
northeastern Pacific. Ph.D. Thesis, Univ. California, Los
Ang., 430 p.

Bowman, T. E., C. D. Meyers, and S. D. Hicks.

1963. Notes on associations between hyperiid amphipods
and medusae in Chesapeake and Narragansett Bays and
the Niantic River. Chesapeake Sci. 4:141-146.

Brandl, Z., and C. H. Fernando.

1974. Feeding of the copepod AcantliocijclopK rernaliy on
the cladoceren Ciriodaphiiia reticulata under laboratory
conditions. Can. J. Zool. 52:99-105.

Davis, C. C.

1959. Damage to fish fry by cyclopoid copepods. Ohio J. Sci.
.59:101-102.
Dunbar, M.J.

1963. Amphipoda. Sub-order: Hyperiidea Family: Hyper-
iidae. Cons. Perm. Int. Explor. Mer, Fiches Identification
Zooplankton 103:2-4.
Dziuban, N. A.

1937. On the nutrition of some Cj/clopidae (Crustacea). C.R.
(Dokl.) Akad. Sci. URSS, Nour. Ser. 17:319-322.

Evans, F., and M. Sheader.

1972. Host species of the hyperiid amphipod Hyperoche
medusarum (Kr0yer) in the North Sea. Crustaceana
(Suppl.) 3:275-276.
Fryer, G.

1957. The feeding mechanism of some freshwater cyclopoid
copepods. Proc. Zool. Soc. Lond. 129:1-25.
Garstang, W.

1900. Preliminary experiments on the rearing of sea-fish
lanae. J. Mar. Biol. Assoc. U.K. 6:70-93.

HOLLING, C. S.

1965. The functional response of predators to prey density
and its role in mimicry and population regulation. Mem.
Entomol. Soc. Can. 45:1-60.

1966. The functional response of invertebrate predators to
prey density. Mem. Entomol. Soc. Can. 48:1-86.

Hurley, D. E.

1955. Pelagic amphipods of the sub-order Hi/periidae in
New Zealand waters. I. -Systematics. Trans. R. Soc. N.Z.
83:119-194.

1956. Bathypelagic and other Hyperiidae from Californian
waters. Allan Hancock Found., Occas. Pap. 18, 25 p.

Kabata, Z,

1970. Crustacea as enemies of fishes. In S. F. Snieszko and
H. A. Axelrod (editors). Diseases of fishes. Vol. 1, p.
128-133. T.F.H. Publ. Inc., Jersey City, N.J.

Lebour,M. V.

1925. Young anglers in captivity and some of their en-
emies. A study in a plunger jar. J. Mar. Biol. Assoc. U.K.
13:721-734.
Lillelund, K.

1967. Experimentelle Untersuchungen uber den Einfluss
carnivorer Cyclopiden auf die Sterblichkeit der Fisch-
brut. Z. Fisch. Hilfswiss. 15:29-43.

Lillelund, K., and R. Lasker.

1971. Laboratory studies of predation by marine copepods
on fish larvae. Vish. Bull., U.S. 69:655-667.

673

FISHERY BULLETIN; VOL. 74, NO. 3

LoRZ, H. v., AND W. G. Pearcy.

1975. Distribution of hyperiid amphipods off the Oregon
coast. J. Fish. Res. Board Can. 32:1442-1447.
Morris, R. F.

1963. The effect of predator age and prey defense on the
functional response of Podisus maculiventris Say to the
density of Hyphantria cornea Drury. Can. Entomol.
95:1009-1020.
Murdoch, W. W.

1969. Switching in general predators: experiments on
predator specificity and stability of prey populations.
Ecol. Monogr. 39:335-354.

1971. The developmental response of predators to changes
in prey density. Ecology 52:132-137.

Murdoch, W. W., and J. R. Marks.

1973. Predation by coccinellid beetles: Experiments on
switching. Ecology 54:160-167.
Oaten, A., and W. W. Murdoch.

1975. Switching, functional response, and stability in
predator-prey systems. Am. Nat. 109:299-318.

PiMENTEL, D., AND F. CRANSTON.

1960. The house cricket, Acheta domes^ticuii, and the house
fly, Musca domestica, as a model predator-prey system. J.
Econ. Entomol. 53:171-172.
Rosenthal, H.

1967. Parasites in larvae of the herring (Clupea harengus
L.) fed with wild plankton. Mar. Biol. (Berl.) 1:10-15.

1972. ijber die Geschwindigkeit der Sprungbewegungen bei
Cyclops sfrenuus (Copepoda). Int. Rev. Gesamten Hy-
drobiol. 57:157-167.

Salt.G. W.

1967. Predation in an experimental protozoan population
(Woodruffia-Paramecium). Ecol. Monogr. 37:113-144.

Sars, G. 0.

1895. An account of the crustacea of Norway. Vol. 1.
Amphipoda. Alb. Cammermeyers Forlag, Copenh.,
711 p.

SCHNACK, D.

1972. Nahrungsokologische Untersuchungen an Herings-
larven. Ber. Dtsch. wiss. Komm. Meeresforsch. 22:273-
343.
Sette, 0. E.

1943. Biology of the Atlantic mackerel {Scomber scombrus)
of North America. Part I: Early life history, including the
growth, drift, and mortality of the egg and larval popula-
tions. U.S. Fish Wildl. Ser\-., Fish. Bull. 50:149-237.
Sheader, M., and F. Evans.

1975. Feeding and gut structure of Parathemisto gaudi-
chaudi (Guerin) (Amphipoda, Hyperiidea). J. Mar. Biol.
Assoc. U.K. 55:641-656.

Shoemaker, C. R.

1930. The amphipoda of the Cheticamp expedition of

1971. Contrib. Can. Biol. Fish. 5:221-359.

Stevenson, J. C.

1962. Distribution and survival of herring larvae {Clvpea

pallasi Valenciennes) in British Columbia waters. J. Fish.

Res. Board Can. 19:735-810.

Theilacker, G. H., and R. Lasker.

1974. Laboratory studies of predation by euphausiid

shrimps on fish larvae. In J. H. S. Blaxter (editor). The

early life history of fish, p. 287-299. Springer-Verlag,

Berl.
Westernhagen, H. von.

1976. Some aspects of the biology of the hyperiid amphipod
Hyperoche medasarum. Helgolander wiss. Meeresunters.
28:43-50.

674

ESTIMATES OF RATES OF TAG SHEDDING BY
NORTH PACIFIC ALBACORE, THUNNUS ALALUNGA

R. Michael Laurs, William H. Lenarz, and Robert N. Nishimoto '

ABSTRACT

Type-I (immediate) and Type-II (instantaneous) rates of tag shedding by North Pacific albacore,
Thunnus alalunga, are estimated using data from a double-tagging experiment. Type-I shedding is
estimated to be about 0.12 and Type-II to be between 0.086 and 0.098 on an annual basis. The paper also
contains a discussion on the accuracy of the estimates, and a method is developed to estimate possible
bias due to fishermen reporting double tag recoveries as single tag recoveries. The possible bias is
estimated to be low.

A tagging program was initiated in 1971, and is
continuing, on Nortli Pacific albacore, Thunnus
alalunga (Bonnaterre), to examine their migra-
tion patterns, to obtain information for use in
population studies, and to estimate rates of mor-
tality. Because loss of tags through shedding can
cause estimates of mortality to be biased upwards
unless corrected for, part of the tagging program
in 1972 consisted of an experiment in which 788
albacore were double-tagged to evaluate tag
shedding by this species.

Chapman et al. (1965) developed a formulation
of the return of single- and double-tagged fish
which includes instantaneous loss rates due to
fishing mortality, other mortality, and tag shed-
ding. They then solved for the instantaneous rate
of tag shedding given data from double-tagging
experiments. Bayliff and Mobrand (1972) e.xtended
the work of Chapman et al. to provide estimates of
the portion of tags which are retained after
immediate shedding occurs. Results of the use of
the Bayliff and Mobrand procedure to estimate
rates of tag shedding from the double-tagging
experiment on North Pacific albacore are pre-
sented in this paper.

METHODS

The tagging program is being conducted jointly
by the National Marine Fisheries Service-

' Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038. Authorship
is alphabetical. Laurs was the investigation leader and responsi-
ble for the overall tagging program and was aided by Nishimoto.
Lenarz was responsible for the analytical aspects of the study.

■ Southwest Fisheries Center, La Jolla, CA 92038.

(NMFS), NOAA, and the albacore fishing industry
through the American Fishermen's Research
Foundation' (AFRF).

Albacore were caught by commercial jig boats
and a bait boat on charter to the AFRF. Fishing
operations on jig boats were conducted with
standard commercial albacore feathered jig-
fishing equipment and commercial trolling meth-
ods. Most of the fish that were tagged and released
from the bait boat were caught by the "winging"
method of live-bait, pole-and-line fishing, whereby
a fish is caught on an anchovy-baited barbless hook
on the end of a short line attached to a stout pole.
Immediately after hooking, the fish is lifted out of
the water, swung toward the fisher, and caught
under the arm of the fisher, who then removes the
hook. A small number of the fish tagged from the
bait boat were taken by trolling feathered jigs and
on rod-and-reel using live anchovy as bait.

Special care was exercised to tag and release
only fish judged to be in very good condition. Fish
which showed signs of severe bleeding, which were
hooked through the roof of the mouth or which
showed signs of extreme exhaustion, were not
tagged. For each tagged and released fish records
were kept of the number of the tag, the date and
time of tagging, the length of fish to the nearest
lower centimeter, condition of fish, and sea surface
temperature. A fish caught by pole and line was
measured with a large caliper and tagged with two
tags inserted almost simultaneously by a tech-
nician while the fisher held the fish under his arm.
A fish caught on trolling gear and rod-and-reel was

Manuscript accepted January 1976.
FISHERY BULLETIN: VOL. 74, NO. 3, 1976.

^AFRF administers revenues derived from an assessment paid
on U.S. - landed albacore.

675

FISHERY BULLETIN: VOL. 74, NO. 3

measured on a Naugahyde-covered foam measur-
ing pad and tagged by a technician while it was on
the pad. In order to tag an albacore on each side,
using this method, the fish had to be turned from
side-to-side.

Spaghetti-dart type Floy^ tags are being used in
the tagging program. The tags are made of yellow
Resinite tubing, 12 to 13 cm long and similar in
structure to those described by Yamashita and
Waldron (1958) and identical to those used by Fink
(1965). The tags were inserted on both sides of the
fish below the second dorsal fin with the aid of a
beveled stainless steel piece of tubing, 14 to 16 cm
long and 0.135- or 0.156-inch inside diameter. The
tags were inserted so that the barb of the tag was
lodged around the pterygiophores of the second
dorsal fin.

We estimated rates of tag shedding using the
notation and methodology of Bayliff and Mol)rand
(1972) for yellowfin tuna as did Lenarz et al. (1973)
in a similar study on bluefin tuna. Bayliff and
Mobrand's equations for returns of tags are:

(1)

ndrffc=FTiVD7rp''e-(^+A' + 2LK.

and

nd,k = 2FtN^7tp (1 - pe-^'k)e-^^ + ^ + ^".

(2)

where

No
77

time at the middle of the kth recovery

period of length t days {k = 1, 2);

number of returns of double-tagged

fish retaining both tags during the

period centered at t^ ;

number of returns of double-tagged

fish retaining only one tag during the

period centered at 4;

number fish released with double tags;

portion of tagged fish which remain
alive after the immediate mortality,
including Type Itagging mortality, has
taken place;

= portion of the tags which are retained
after Type-I (immediate) shedding
has taken place;

= instantaneous rate of fishing mor-
tality;

••Floy Manufacturing Company, Seattle, Wash. Reference to
trade names does not imply endorsement by the National Marine
Fisheries Service, NOAA.

X = instantaneous rate of other mortality
(other included natural mortality,
Type-II (long-term) tagging mortali-
ty, and apparent mortality due to
migrations from the fishery); and

L = instantaneous rate of tag shedding
(Type-II shedding).

Bayliff and Mobrand (1972), using Equations (1)
and (2), showed that

27id,

Odsfc '^^^idk

= -L4 -h In p = ^^

(3)

where //,, is an estimate of the natural logarithm of
the proportion of tags retained up to time ^;^. Note
that the first factor of the right-hand side of
Equation (2) is the integer 2. Both Bayliff and
Mobrand (1972) and Lenarz et al. (1973) mis-
takenly left this multiplier out of the equation in
their papers. However, the error was typo-
graphical and did not affect their derivations or
results. Given n,,,,^, n,,,,,^, and f,. , L and p are
estimated using simple linear regression; or as in
the case of this study when only two recovery
periods are used, the solution of two simultaneous
equations. Equations (1) and (2) assume that L and
the total of F and X are constant over t,^ . Since the
albacore fishery is seasonal, the assumption is
likely to be violated. The effect of the violation has
not been examined.

RESULTS

Release and return data through 1973 are shown
in Table 1. The number of returns in 1974 was
insufficient for analysis. A chi-square test indicat-
ed that gear type did not have a significant effect
on the proportions of single- and double-tag returns
in 1972 (x- = 1.117, df = 1). Data from both gears
were combined for the remainder of the analysis.

Estimates of p and L are shown in Table 2. Only
returns that could be specified to the nearest week
are included in Table 1. Precise dates of recovery

Table l.-Tag releases and returns with information on date of
recovery for North Pacific albacore and double-tag study.

1972
double-

1972 returns

1973 returns

Gear
type

Average Average

tag Double Single days out Double Single days out

releases In^^,) (n^J

(',)

("..2) ("ds)

('2)

Jig 330

Bait 448

Total 778

10
22
32

— 12

— 2
54.71 14

451.55

676

LAURS ET AL.: TAG SHEDDING BY ALBACORE

Table 2.-Estimates of rates of tag shedding, L (on an annual
basis), retention, p, from 1972 North Pacific albacore double-tag
study-

Table 3.-Tag releases and returns from North Pacific albacore
single-tag studies.

Item

Undated returns excluded
Undated returns included

0.098
0.086

0.88
0.88

could not be assigned to seven double-tag and two
single-tag returns in 1972 and one double-tag
return in 1973. We assumed that 4 was the same
for the returns shown in Table 1 and the returns
with unspecified recovery dates and included the
10 additional returns in a recalculation of p and L
The results of the recalculations are similar to the
original (Table 2). We estimated p to be about 0.88
and L on an annual basis to be between 0.086 and
0.098. This means that if no mortality occurs, 8.2 to
9.3% of all unrecovered tags are expected to be lost
through shedding annually.

Our estimate of p is similar to the results
obtained for yellowfin tuna (p = 0.913) by Bayliff
and Mobrand (1972) and bluefin tuna (p = 0.973) by
Lenarz et al. (1973). However, our estimate of L is
considerably lower than that obtained for yel-
lowfin tuna (L = 0.278) and bluefin tuna (L =
0.310).

Methodology for estimation of the variance of L
and p when only two periods of recovery are
available has not been published. However, we
believe that the number of tag returns available
for this study is too low for accurate estimates of p
and L. We made the following calculations to
illustrate the relative level of accuracy. If we
arbitrarily assume that the returns of double- and
single-tagged fish in 1973 were from a binomial
distribution with the probability of a returned fish
having only one tag being 0.5, the probability of
having 8 or fewer fish returned with only one tag
out of a sample of 22 fish from such a population is
about 0.14. If 11 fish were returned with single
tags (the expected value from the assumed dis-
tribution) instead of the 8 observed, our estimates
of p would be 0.895 and our estimate of L would be
0.172. Thus it appears that there is a reasonable
chance that our estimate of L (about 0.09) could be
considerably lower than the true value.

We are not aware of any other data available
from double-tag studies on albacore. However,
there is a considerable amount of data available
from single-tag studies conducted in recent years
on albacore in the eastern North Pacific (Table 3).
Return rates in the year after release were 0.018

Year of

Number
released

Number returned

release

1971

1972

1973

197.

1971
1972
1973
1974

887
1,304
1,806
2,490

16
27

11
47
13

6
14
59
35

for the 1971 releases, 0.036 for the 1972 releases,
and 0.033 for the 1973 releases of single-tagged
fish, for an average of 0.029. If the return rates are
divided by 0.88 to account for Type-I tag shedding,
the average becomes 0.033. The return rate in the
year after release for the double-tag study was
0.027. If the rate is divided by 0.99 (1 - (I - p)'^) to
account for Type-I shedding of both tags, the
return rate is 0.027. Thus the return rates from the
single-tag studies give further evidence that
Type-II shedding is insignificant, because if it
were not, return rates adjusted for Type-I shed-
ding from the single-tag releases should be lower
than return rates from the double-tag releases,
provided mortality rates were similar for these
years.

The above estimates are based on the assump-
tion that all double-tag recoveries are reported as
double-tag recoveries. A possible source of error is
that some fishers may return only one tag from a
double-tag recovery. These fishers might return
only one tag because of their interest in albacore
migrations, but retain the second tag as a souve-
nir. This would result in our underestimating the
value of p. To illustrate the extreme case assume
that p is actually 1.0, but we estimate it to be 0.88
because of incomplete reporting. Then assuming p
= 1, Equations (1) and (2) become

(4)

UMk = FTA^ovr5e-(^+^^ + 2LK.

and

na^K - 2F T iVo TT (1 - e-^'k)e-(^ + ^ + ^'^ + (1 - 5)

FrNo-rre-^^^^^^^^''- (5)

where B = minimum proportion of double-tag
recoveries that are reported as double-
tag recoveries.

Manipulation of Equations (4) and (5) results in
(«dd2 + ^^ds2) (nddi) 2e^'2 - 1

{riddi + Wdsl) i'Hdd2)

2e^'i-l

(6)

and

677

FISHERY BULLETIN: VOL. 74, NO. 3

B =

(2e^'^ - 1) (^rf J

(7)

An estimate of L is obtained from an iterative
solution of E(iuation (6). An estimate of the
minimum value of B is obtained from substitution
of the estimate of L into Equation (7). Our es-
timate of L and the minimum value of B, where
only returns with specified dates are included in
the calculations, are 0.087 and 0.78, respectively.
When all of the return dates are included we
estimate L to be 0.077 and B to be 0.78. Thus, it
appears that the rate of reporting double-tag
recoveries as single-tag recoveries is less than 0.22
(\-B).

However, we have no evidence to indicate that
fishers have returned only one tag from fish
recovered with two tags. We believe that fishers
have turned in both tags of fish recovered with two
tags based on interviews with those who have
recovered tagged fish, the very good cooperation
that we have received from them during the
tagging program, and the fact that tags from
recovered fish may be returned to the fisher if he
wishes to have them.

ACKNOWLEDGMENTS

We thank the skippers, crewmen, and NMFS
technicians who participated in the albacore tag-
ging and release program; members of AFRF for

supporting the tagging charters; albacore fishers
and fish processors for their excellent cooperation
in reporting tag recoveries. We also thank R.
Francis of the Inter-American Tropical Tuna
Commission, J. Wetherall of the Southwest Fish-
eries Center Honolulu Laboratory, and J. Zweifel
of the Southwest Fisheries Center La Jolla
Laboratory, for critically reviewing the paper. J.
Wetherall and M. Yong have developed maximum
likelihood estimators of the variance of p, L, and B
and intend to publish their work.

LITERATURE CITED

B.WLIFF, W. H.. .AND L. M. MOBR.\ND.

1972. E.stimates of the rates of shedding of dart tags from
yellowtin tuna. [In Engl, and Span.] Inter-Am. Trop.
Tuna Comm., Bull. 15:441-462.

Chapman, D. G., B. D. Fink, and E. B. Bennett.

1965. A method for estimating the rate of shedding of tags
from yellowfin tuna. [In Engl, and Span.] Inter-Am.
Trop. Tuna Comm., Bull. 10:333-352.
Fink. B. D.

1965. A technique, and the equipment used, for tagging
tunas caught by the pole and line method. J. Cons.
29:335-339.
Lenarz, W. H., F. J. Mather III, J. S. Beckett, A. C. Jones, and
J. M. Mason, Jr.

1973. Estimation of rates of tag shedding by northwest
Atlantic bluefin tuna. Fish. Bull., U.S. 71:1103-1105.

Yamashita, D. T., and K. D. Waldron.

1958. An all-plastic dart-type fish tag. Calif. Fish Game
44:311-317.

678

NOTES

PARALYTIC SHELLFISH POISONING

IN TENAKEE, SOUTHEASTERN ALASKA:

A POSSIBLE CAUSE

PSP (paralytic shellfish poisoning) has been
reported from much of the west coast of North
America. Recent reviews (Halstead 1965: 157-240;
Quayle 1969) summarizing many aspects of the
problem have emphasized its causative organism,
Gonyaulax catenella (and possibly G. acatenella).

Chemical studies (Schantz and Magnusson 1964)
indicate that the poison is chemically similar
throughout the range of G. catenella— Califorma
through Alaska. Because of this similarity, and the
reported occurrence of G. catenella in Alaska
(Meyers and Hilliard 1955; Sparks 1966; Neal 1967),
it has often been assumed that this species is the
cause of PSP in Alaska. This assumption has not
been well verified, however. A 2-yr study in
southeastern Alaska by the University of Alaska
failed to find a significant correlation between the
occurrence of PSP and G. catenella (Chang 1971).
Sparks (1966) and Neal (1967) reported a correla-
tion in their occurrence near Ketchikan, but the
number of G. catenella was so low that very long
toxification periods would have been required to
cause lethal clams.

The difficulty in verifying the relationship re-
sults, in part, from the very low densities of G.
catenella in Alaska plankton (Schantz 1966; Chang
1971). Sparks (1966) stated that it has even been
difficult to demonstrate that G. catenella occurs in
Alaska waters. Since toxic shellfish occur quite
frequently in southeastern Alaska, some observers
(Schantz and Magnusson 1964; Neal 1967; Chang
1971) have concluded that organisms other than G.
catenella might also cause PSP.

We believe the events reported in this paper
provide the first demonstration of a localized G.
catenella bloom followed by a PSP outbreak in
Alaska waters.

Methods and Results

On 20 September 1973, 5 days before an out-
break of shellfish poisoning in humans occurred,
very high bioluminescence was seen in Tenakee
Harbor (lat. 57°48'N; long. 135°14'W). During
darkness, glowing outlines of large individual fish

and schools of fish were clearly seen moving in the
water. Long-time residents remarked that it was
the greatest amount of "phosphorous" (biolumin-
escence) they had ever seen there.

The RV Maybeso, Alaska Department of Envi-
ronmental Conservation, was in the area at the
time, and curiosity about the bioluminescence
prompted the crew to collect a small (lOO-cm"*)
water sample, which was preserved with
Formalin.' Water temperature at the time of
collection was 11.5°C, and salinity was 22.18"/(xi.
The water could not be microscopically examined
until 1 October, when the Mayba^o returned to
Juneau. At that time the sample was given to the
senior author, who was coordinating a PSP re-
search program for the Department of Environ-
mental Conservation. Large numbers
(235,000/liter) of G. catenella were found in the
sample. Other dinoflagellate species were present
but only in trace amounts. No organism other than
G. catenella was found in high enough numbers to
cause intense bioluminescence.

We learned that on 25 September 1973, several
families had dug the butter clam, Saxidomus
giganteus, near the boat harbor in Tenakee. After
eating the clams, two people reported severe
symptoms of PSP to the Alaska Department of
Health and Social Services. When interviewed, the
victims, as well as other Tenakee residents, stated
that they had eaten clams from the same area
earlier in the year without any toxic reactions.

Using conventional methods (Quayle 1969; Pra-
kash et al. 1971), the Alaska Division of Public
Health Southeast Regional Laboratory deter-
mined that the level of toxin in the uneaten
portion of some of the cooked clams from Tenakee
was 4,550 jug/lOO g. The toxin was distributed
throughout the body and was not concentrated in
the siphons. Indeed, one of the illnesses was caused
by ingesting clams from which the siphons had
been removed before cooking.

We flew to Tenakee on 5 October, about 2 wk
after the outbreak, but found no G. catenella in the
water. We did not test any clams for toxin levels at
that time, but the mussel, Mytilus edulis, growing
on harbor pilings had high levels of toxin (2,300
/xg/lOOg).

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

679

Discussion

The fact that toxin was distributed throughout
the bodies of the clams, rather than being concen-
trated in the siphons, indicates that the contact
between the clams and the toxin-producing or-
ganisms had been recent. The lack of a concentra-
tion of toxin in the siphon may even indicate that
toxification was in progress (Quayle 1969). The high
toxin levels in mussels also reinforces the
probability that toxification had occurred recently;
mussels lose their toxin rapidly (Prakash et al. 1971)
and the high levels indicate that the toxicity was
acquired shortly before our sampling.

There is presently no information on the pump-
ing rate, particle retention, or assimilation
efficiency of Saxidomus giganteus (K. Chew pers.
commun.). Pumping rates of the American oyster,
Crassostrea virgi)iica,ca.n be as high as 20 liters/h
and probably average about 10 liters/h (Loosanoff
and Engle 1947; Galtsoff 1964). By using the rate
of 10 liters/h, which is conservative for the larger
S. giganteus, and assuming a particle retention of
25%, which is also conservative when particles the
size of G. catenella, 25-55 jum, are ingested (Loo-
sanoff and Engle 1947), a toxification period may
be calculated.

Approximately 3,000 G. catenella will produce
one mouse unit (approximately equal to 0.2 i-tg) of
toxin (see discussion in Neal 1967). Filtering 10
liters/h of water containing 235,000 G. catenella
/liter and retaining 25% of the G. catenella will
result in an increase of 40 jug of toxin/h in each
clam. The Saxidomus sampled at Tenakee con-
tained 4,500 jug/100 g or approximately 2,250
jLig/clam (an average clam probably weighs less
than 50 g). Thus, using these conservative figures,
it would have taken slightly more than 2 days (57
h) of filtering to reach the levels found in Tenakee
clams.

From the known background of this event, it is
apparent that the shellfish must have become toxic
shortly before the illnesses were reported. The
occurrence of the G. catenella bloom approximate-
ly 1 wk before the PSP outbreak indicates that
even though this species is normally found in very
low densities in Alaska, it can occur in high enough
numbers to rapidly toxify clams.
Acknowledgments

Donald Gerber and William Goodman, Alaska
Department of Health and Social Services, made
their investigation of the Tenakee outbreak

available to us. Louisa Norris and Richard Norris,
University of Washington, confirmed the iden-
tification of Gonijaulax catenella. Kenneth Chew,
University of Washington, provided a very ben-
eficial description of the present status of research
on Saxidomus giganteus and PSP.

Literature Cited

Chang.J.C. C.

1971. An ecological study of butter clam {Sai-idomus gigan-
teus) toxicity in Southeast Alaska. M.S. Thesis, Univ.
Alaska, College, 94 p.
Galtsoff, P. S.

1964. The American oyster Craaaostrea rirginica Gmelin.
U.S. Fish Wildl. Serv', Fish. Bull. 64:1-480.

Halstead, B. W. (editor).

1965. Poisonous and venomous marine animals of the world.
Vol. 1. U. S. Gov. Print. Off., Wash., D.C., 994 p.

Loosanoff, V. L., and J. B. Engle.

1947. Effect of different concentrations of micro-organisms
on the feeding of oysters (0. rirginica). U. S. Fish Wildl.
Serv., Fish. Bull. ,51:30-57.
Meyers, H. F., and D. R. Hilliard.

1955. Shellfish poisoning episode in False Pass, Alas-
ka. Public Health Rep. 70:419-420.
Neal, R. A.

1967. Fluctuations in the levels of paralytic shellfish to.xin in
four species of lamellibranch molluscs near Ketchikan,
Alaska, 1963-1965. Ph.D. Thesis, Univ. Washington, Seat-
tle, 164 p.
Prakash, A., J. C. Medcof, and A. D. Tennant.

1971. Paralytic shellfish poisoning in eastern Canada. Fish.
Res. Board Can., Bull. 177, 87 p.
Quayle, D. B.

1969. Paralytic shellfish poisoning in British Colum-
bia. Fish. Res. Board Can., Bull. 168, 68 p.

SCHANTZ, E.J.

1966. Chemical studies on shellfish poisons. In W. Felsing
(editor). Proceedings of joint sanitation seminar on North
Pacific clams, p. 18-21. Alaska Dep. Health Welfare and U.
S. Dep. Health, Educ, Welfare, Public Health Serv.

SCHANTZ, E. J., AND H. W. MaGNUSSON.

1964. Observations on the origin of the paralytic shellfish
poison in Alaska butter clams. J. Protozool. 11:239-242.
Sparks, A. D.

1966. Physiological ecology of the causative organisms
including mechanisms of toxin accumulation in shell-
fish. In W. Felsing (editor). Proceedings of joint sanitation
seminar on North Pacific clams, p. 10-11. Alaska Dep.
Health Welfare and U.S. Dep. Health, Educ, Welfare,
Public Health Serv.

Steven T. Zimmerman

Northivest Fisheries Center Auke Bay Fisheries Laboratory
National Marine Fisheries Service, NOAA
P.O. Box 1.5.5, Auke Bay. AK 99821

Robert S. McMahon

Alaska Department of Environmental Conservation
Pouch 0, Juneau, AK 99801

680

OIL AND GREASE: A PROPOSED
ANALYTICAL METHOD FOR
FISHERY WASTE EFFLUENTS

The published procedures (American Public
Health Association 1971:407-413; Environmental
Protection Agency 1974) for determining oil and
grease in industrial wastes are generally unsuit-
able for fish-processing waste effluents, especially
for such high-load effluents as occur during the
processing of salmon for canning. These wastes
cannot be filtered satisfactorily by the method
described. In addition, a Soxhlet extraction of the
fish proteinlike material after drying for 30 min
gives low values because of the inefficient extrac-
tion of protein-bound lipids.

These inadequacies of the published methods for
the analysis of oil in fish-processing waste streams
indicate a need for an alternate method that is
simple and accurate. Accordingly, a method was
worked out using portions of the published oil and
grease methods and using techniques developed
by Kelley and Harmon (1972) for the analysis of
carotenoids. The method involves a precipitation
of protein and particulate matter to allow easy
filtration and subsequent extraction of oil from
the residue under anhydrous conditions, using
2-propanol (IPA) and petroleum ether (PE). The
method is proposed as an alternate method for
determining oil and grease in fishery waste
effluents.

Materials and Methods

Reagents and Equipment

Celite^ 503, Johns-Mansville (filter aid): For best
results, Celite should be washed with water and
solvents because a slight oil residue may carry over
into the oil fraction. Blend about 100 parts of
Celite by weight with 500 parts water, filter,
reblend with 500 parts (vol) IPA, filter, reblend
with 500 parts (vol) PE, filter and apply suction
until reasonably dry. Air dry and store in a jar.
Filter paper dispersion: Blend 20 7-cm filter paper
disks (Whatman 1 or 40) with distilled water in a
blender for 5 - 10 min. Bring volume to 2,000 ml.
Sodium hexametaphosphate in water: 250 mg/ml,
use 1 ml per analysis, i.e., 250 ppm. Other materials
required are: filter flasks (250 ml and 2,000 ml).

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

graduated cylinder (1,000 ml), filter pump (water
aspirator), filter funnel (fritted disc, 350 ml coarse,
150 ml medium), blender and jars (Virtis Model 23
and 200-ml blender jars), rotating evaporator with
250-ml flask, film to seal cylinder (parafilm "M"
American Can Company, Marathon Products),
50% acetic acid, anhydrous magnesium sulfate
(powdered), reagent grade IPA, and reagent
grade PE (bp 40°-60°C).

Preparation of Filter Funnel

Assemble filter flask and a 350-ml "c" sintered
glass filter funnel. Add about 3 g filter aid and 100
ml filter paper dispersion directly to the funnel.
Fill funnel with water, stir and allow to partly
drain without vacuum. Apply vacuum, rinse brief-
ly, and press down along edge of mat to ensure a
good seal.

Preparation of Sample and Filtering Step

Pour well-mixed sample of effluent to the
1,000-ml mark in the graduated cylinder. Add 3 to
6 g filter aid to aid precipitation. In its absence,
flotation and precipitation both occurred. Add 1 ml
hexametaphosphate solution, seal cylinder with
film, and mix by inverting cylinder about 12 times.
Add 2 ml acetic acid. The amount of acid will vary
with the type of effluent and is not critical provid-
ed enough is added; the pH must be lower than 4.2,
but precipitation works equally well at several
levels between pH 2.1 and 4.2. Invert three or four
times. Excessive mixing inhibits rate of precipi-
tation. Wait about 2 min and add more acid if top
inch or so is not clear. Solids in salmon waste
effluents are slow to settle and are best handled by
allowing the mixture to settle overnight in the
refrigerator. Salmon waste, after 2-h settling, can
be filtered but with difficulty. If filtration is
started too soon, the sample often must be dis-
carded because it will not filter. Shrimp and crab
waste usually can be filtered in 15 to 30 min. Filter
clear supernatant fluid under vacuum through the
prepared filter funnel (very rapid), and transfer
more slowly the precipitate (50-75 ml vol) and
rinsings to the funnel. Use about 200 ml water to
remove excess acid and to rinse graduate and
filter. Continue vacuum 5 to 10 min to remove as
much water as possible because the next step, the
extraction, must be anhydrous.

Extraction of Oil

Carefully transfer solid material, including

681

Celite and filter paper, to the 200-ml blender jar
plus about 15 g anhydrous MgSO, and 75 ml I PA.
The desiccating step with MgS04 is not effective if
volumes of IPA are excessive. In addition, all
volumes should be maintained as specified to allow
rinsing without exceeding the capacity of the
250-ml evaporating flask. The IPA should be
measured in the liter graduate and shaken or
rotated to wash cylinder. Blend at high speed for 5
min, then pour contents of blender jar into 150-ml
dry filter funnel (M-porosity), apply vacuum until
dripping ceases, rinse briefly with PE (wash
bottle), then repeat extraction with 75 ml PE. The
second extraction with PE removes about 2.5% of
the total oil.

Quantitatively transfer filtrate to a pre-dried
and weighed 250-ml 24/40 standard taper round-
bottom flask, and flash evaporate using a rotating
vacuum evaporator and warm water bath. This
method takes from 5 to 10 min, but other tech-
niques of evaporating would be suitable. When
solvents are removed, add about 10 ml PE to
determine if water or solid materials are present.
If clean, evaporate to dryness, wipe outside of
flask, and place in drying oven for exactly 30 min
to remove traces of solvent or water. Cool in air for
1 h and weigh. Subtract tare weight and record
weight of oil directly as milligrams per liter. The
common practice of storing the dry flasks in a
desiccator was not necessary because there was
little change in weight with subsequent exposure
to air. The oil apparently reached nearly constant
weight (oxidation) during the 0.5-h drying step.
Exposure of the dry oil and the flasks to air for 15
and 40 min resulted in 2.2 and 2.6 mg gain in
weight for 1,684 mg oil and only 3.2 mg gain with
overnight exposure. Consequently, because the
250-ml round-bottom flasks were difficult to weigh
in a rapid manner, weights were obtained after
oven drying for 0.5 h and air cooling for 1 h.

If the above PE solution is not free of water or
solid particles, add 10-15 g anhydrous sodium
sulfate and sufficient PE to mix well. Let sit a few
minutes, and filter through sodium sulfate on a
60-ml medium- or fine-porosity fritted-glass fun-
nel, rinse with PE, and transfer back to evaporat-
ing flask. The pre-weighed 250-ml flask should be
washed out with water and solvents before reuse.
This step is time-consuming and is never neces-
sary if the previous extraction and desiccating
steps are done properly.

Accuracy and Precision

The results of replicate analyses on eight
effluent samples indicate that the proposed meth-
od gives acceptable precision (Table 1).

The mean standard deviation of these data on
three difl:'erent species is 5.3, and the mean is 552
mg/liter. The published mean standard deviation
for the three methods given in the Environmental
Protection Agency (EPA) manual is 1.1, with a
mean of 15.0 mg/liter. To compare standard
deviations with different means, the coeflficient of
variation (CF) is used, and for the data in this
paper the CV is 1 as compared with 7 for the data
given in the EPA manual. This means that a
sample of waste effluent having 100 mg oil and
grease/liter will have a comparative standard
deviation of 1 or 7 mg/liter, depending on the
method used.

The accuracy of the proposed method was
evaluated by comparing the EPA Soxhlet method
with the method given in this paper, using seven
grab samples of king crab, snow crab, and shrimp
waste efl^uents. The data in Table 2 show that the
oflicial EPA Freon 113 Soxhlet method gave oil
and grease values that were consistently low,
varying from 6 to 48% and averaging about 30%.

The filtrates from the EPA method of filtration
from samples 3, 4, 5, 6, 7 were precipitated and

Table l.-Oil and grease values expressed as milligrams per liter
for eight effluent samples.

Replicate oil and grease values

Sample

III

Snow crab effluent

158

154

153

Snow crab effluent

251

250

248

Shrimp effluent

397

399

Shrimp effluent

432

404

Salmon effluent

844

847

Salmon effluent

231

221

Salmon effluent

923

925

Salmon effluent

1,200

1,190

Table 2.-Comparison of oil and grease values expressed as
milligrams per liter determined by the EPA Soxhlet method and
the proposed method.

EPA

method

Proposed

method

Sample analyzed

King crab effluent

(')

225

King crab effluent

(')

164

221

225

Shrimp effluent

179

182

215

209

Snow crab effluent

161

164

174

Snow crab effluent

'Samples 3A, 3B, and 4A could not be filtered except by
changing filters.

682

extracted by the method of this paper to give
recoveries of 49 mg (237c), 56 mg (25%), 18 mg
(107f), and 6 mg (507c), respectively. Thus, the
official method of filtration resulted in an average
loss of oil and grease of 25% of the values deter-
mined by the proposed method.

Two effluents (3 and 4) were precipitated by the
method in this paper but extracted by the Soxhlet
method and gave 16 and 5% low values, respec-
tively. In addition, contamination of the oil frac-
tion with Celite and fiber is apparent in the EPA
Soxhlet method and oil and grease values are
estimated to be 5-10 mg lower than reported.

Discussion

Different precipitation techniques were used in
developing this method and gave valid results for
specific waste effluents. For freshwater-processed
shrimp, Celite, alum (200 ppm), and Magnafloc
835A (2 ppm) resulted in complete precipitation in
about 15 min. The alum technique also worked on
waste effluents from saltwater-processed shrimp
and on snow crab, but precipitation was slower and
filtration was more difficult. In general, the hexa-
metaphosphate precipitation is the preferred
technique because it resulted in a more firm, dense
floe that filtered more rapidly than the alum
system. In addition, the soluble proteins along
with their oil content are recovered in the hexa-
metaphosphate precipitate and included in the
analysis. The soluble proteins generally are not
recovered with the alum system or by the EPA
method. Presumably, any reagent can be used for
precipitation provided there is no carry-over into
the oil fraction. Sulfuric acid was used to develop
this method, but it occasionally resulted in a dark
oil after drying. Consequently, the use of sulfuric
acid was discontinued in favor of acetic acid. The
proposed method should be tested further in
comparison with the standard EPA methods for oil
and grease to determine its applicability to other
fishery waste effluents.

Literature Cited

American Public Health Association.

1971. Standards for the examination of water and
wastewater. 13th ed. Am. Publ. Health Assoc, Wash., D.
C, 874 p.
Environmental Protection Agency.

1974. Oil and grease. //; Methods for chemical analysis of
water and wastes, p. 226-235. Environ. Prot. Agency, Off.
Technol. Transfer, Wash. D.C.

Kelley, C. E., and A. W. Harmon.

1972. Method of determining carotenoid contents of Alaska
pink shrimp and representative values for several shrimp
products. Fish. Bull, U.S. 70:111-113.

Jeff Collins

Pacific Utilization Research Center Kodiak
Utilization Research Laboratory
National Marine Fisheries Service, NOAA
P.O. Box 1638, Kodiak, AK 99615

OCCURRENCE OF

VOLATILE N-NITROSAMINES IN

JAPANESE SALMON ROE

Consumer interest and concern about food addi-
tives is as strong in Japan as in the United States.
The possibility that secondary or tertiary amines
and nitrites in fish roe products (sujiko) might
combine to produce A^'-nitrosamines, known car-
cinogens, has received much attention and pub-
licity. If the use of nitrites is curtailed in Japan,
American salmon canners would be hurt because
of loss of sales or decreased prices for roe sold to
Japanese processors operating in the Pacific
Northwest. The value of this business to the U.S.
salmon industry is from $10 to $15 million each
year.

Investigations by Howard et al. (1970) and
Fazio, Howard, White, and Watts (1971) showing
trace quantities of A'^-nitrosodimethylamine
(NDMA) from samples of chub, sable, salmon, and
shad prompted the National Marine Fisheries
Service (NMFS) to be concerned about A^-nitro-
samines in smoked nitrite-treated fishery
products. This concern was shared by the National
Canners Association (NCA) in connection with
nitrite-treated salmon roe products. Various sam-
ples of salmon roe commercially produced in
canneries in the northwestern United States and
Alaska were obtained by the NCA for analysis of
volatile A^-nitrosamines.

In addition to the analysis for nitrosamines
which was carried out by NMFS, samples were also
analyzed by NCA for residual nitrite and chloride
concentrations. The results of these findings are
presented in this report.

Experimental

Background

For a number of years, Japanese companies

683

have maintained salmon roe processing operations
at canneries in the northwestern United States
and Alaska. The processing of salmon roe is an art
rather than a formulated production procedure,
and numerous minor differences are found in the
various recipes employed. The following is, of
necessity, a generalized description of the produc-
tion operation.

Roe from the butchered salmon is received in the
egg house, cleaned of extraneous fish material, and
rinsed to remove blood. From 27 to 38 kg of roe are
placed in a vat containing 200 liters of saturated
brine into which has been added either 0.02-0.05%
nitrate or 0.05-0.07% nitrite (equivalent to 500-700
ppm.).

The mix is agitated mechanically for approx-
imately 20 min. The actual length of time is
determined by technicians who consider a range of
variables, such as the size of roe, the freshness of
fish from which roe was obtained, and temperature
of brine solution. Larger roe, as from king or chum
salmon, are held in the brine longer. Brine batches
may be used for several changes of roe; normally,
they are changed four or five times in an 8-h day.

After removal from the vat, the roe are drained
and graded by size and color. Nitrite level of the
roe at this time is about 50 ppm. The roe are then
packed in 10-kg wooden boxes which are lined with
sheets of plastic. After each layer is packed, it is
lightly salted with a fine grind sodium chloride.
The boxes are slightly overfilled, and the lids
placed on without nailing. They are then stacked
with weights on top to form a press. The boxes are
cured in this fashion for as long as 7 to 10 days,
depending on ambient temperature conditions.
During the curing period, the desirable red color of
sujiko develops, and nitrite residuals drop to less
than 5 ppm. It is possible that the color enhancing
action of the nitrite may be due to its inhibiting
effect on color destroying oxidative enzymes in the
roe.

Following pressing and curing, the product is
inspected. If satisfactory, the lids are nailed down,
and the boxes are stored at -5°F (-20.6°C) at the
cannery and placed aboard transport vessels to
Japan. In Japan, the same storage conditions
apply until the product is sold to the retail
markets.

Production Survey

Duplicate 10-kg samples of commercially
produced red and pink salmon roe products were

obtained from four of the five major sujiko
processors. The processing plants were located on
Kodiak Island in the Gulf of Alaska, southwest of
Anchorage; Hawk Inlet in the Admiralty Islands,
west of Juneau; Cook Inlet, large inlet which
Anchorage is at the head of; and Ketchikan,
southeast Alaska on the south side of Revil-
lagigedo Island. Duplicate 10-kg samples of roe
from three species of salmon— red, chum, and
king— were obtained from the fifth major producer
located at Puget Sound, Wash. All of these samples
were obtained after their delivery to Japan. It was
decided to sample the roe in Japan so that storage
conditions would be more nearly identical to those
received by the product going to consumers. Upon
return of the samples to this country, NCA
delivered them to NMFS. The samples were com-
posited in a Hobart silent cutter, packaged in
Mylar' bags, and sealed. A portion of the compos-
ite sample was returned to NCA for determina-
tions of residual nitrites and NaCl content.

Experimental Pack

Using roe from the same batch of fish, one test
pack and one control pack of salmon roe were
prepared by NCA. The test pack was prepared in a
saturated brine containing 700-ppm. nitrite, while
only a saturated brine was used to prepare the
control pack. The packs were cured at a tempera-
ture of 60°F (15.6°C) for 7 days and then stored for
6moat-5°F(-20.6°C).

Materials

The solvents-methylene chloride, pentane, and
ethyl ether-were purified by distillation. Sol-
vents, silica gel, and Celite 545 were tested prior to
use to assure the absence of interfering peaks.

Analytical

The multidetection method for the analysis of
volatile A^-nitrosamines in foods developed by
Fazio, Howard, and White (1971) was used in this
investigation. Because of the high phospholipid
content of the salmon egg samples, William T.
Roubal of the Northwest Fisheries Center, NMFS,
NOAA, found it necessary to make some
preliminary modifications in the procedure (Fazio,

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

684

Howard, and White 1971). Very briefly, these
modifications were as follows:

1. Initial digestion of product-30 g of KOH were
employed in the digestion, and the methanolic
KOH solution was re fluxed for 2-5 h.

2. Distillation step— 8 g of Ba(OH)o was utilized.
The distillation was carried out with the aid of a
magnetic stirrer.

Briefly, the procedure involved digestion of the
sample in methanolic KOH, liquid-liquid extrac-
tion of an aliquot of the digest with methylene
chloride, distillation of the nitrosamines from
alkaline solution with further cleanup by solvent
partitioning and column chromatography on silica
gel and Celite 545 columns followed by GLC
(gas-liquid chromatography) analysis.

A victoreen Model 4000 GLC Chromatograph
equipped with a Coulson Electrolytic Conductivity
Detector and an Autolab System IV Computing
Integrator was employed in the analysis of salmon
roe extracts. A 9 foot {2.74 m) x 4 mm inside
diameter glass column coated with 10% Carbowax
1540 + 3% KOH on 80/100 mesh gas chrom Q was
used. The following parameters were maintained
throughout all analyses.

Temperature of injector block - 190°C
Carrier gas (helium) flow rate - 70 ml/min
GC (gas chromatograph) oven' temperature -
ambient for 540 s; GC oven door was closed
and brought to 80°C (held at 80°C for 180 s);
80°-180°C at a program rate of 5°C/min.

Conditions of Coulson Electrolytic Conductivity
Detector operated in reductive mode were:

Hydrogen flow rate - 83 ml/min
Venting helium flow - 70 ml/min
Furnace temperature - 820°C
Venting block temperature - 190°C
Conductivity bridge - 30 V
Attenuation - 1.

Moisture, nitrite, and chloride determinations
were made according to the official methods of
analysis of the Association of OflScial Analytical
Chemists.

ducted. A mixture of six A/'-nitrosamines was used.
The A^-nitroso compounds were NDMA, dieth-
ylamine (NDEA), dipropylamine (NDPA), dibu-
tylamine (NDBA), piperidine (NPi), and pyr-
rolidine (NPy). Prior to recovery runs, however,
the salmon roe samples were examined for A'-ni-
trosamines. Several of the cleaner samples were
fortified at the 10-ppb (parts per billion) level. In
instances where a nitrosamine was found under
study, appropriate adjustments were made in the
recovery values. Recovery of the A''-nitrosamines
at the 10-ppb level ranged from 67 to 88%.

Representative chromatograms obtained from a
fortified pink salmon roe extract together with
those obtained from the corresponding unfortified
samples are shown in Figure 1. This figure shows
the recovery of six nitrosamines after the silica gel
cleanup step. Usually, the interferences occurring
at a retention time of NPy were removed by
further cleanup on the acid-Celite column. During
the course of this investigation, blank runs (with-
out a salmon roe sample) were made, and the
minute GLC peak (3-15 mm) with the same reten-
tion time of NDMA observed with all roe samples
was not apparent in the blank. As shown in Table
1; if the peaks are calculated as NDMA, the levels
range from 0 to 3 ppb. Residual nitrite and chloride
concentration are also shown.

A total of 24 salmon roe samples were analyzed
in duplicate. All samples contained less than 5 ppb
of NDMA. The demonstrated sensitivity of the
method was shown to be 10 ppb. A peak with a
retention time of NDEA was found (< 1 ppb). No
attempt was made to confirm the identity of
NDMA or NDEA in any of the samples since all
were too low for mass spectrometric confirmation.
Some samples were carefully concentrated down

Results and Discussion
During the survey, recovery studies were con-

FiGURE l.-Gas chromatograms of spiked and unspiked extracts
of pink salmon.

685

Table l.-DMNA, nitrite, and NaCl content of salmon roe
samples prepared from different species of salmon at various
processing plants in the Pacific Northwest.

Location of

Residual

Apparent DMNA

processing

NaCI

nitrite

found

Species

plant

(%)

(ppm.)

(ppb)

roduction survey

Red

Kodiak Island

8.42

0.3

Red

Kodiak Island

8.01

Trace

Red

Cook Inlet

8.75

Trace

Red

Cook Inlet

7.07

Trace

Red

Ketchikan

8.73

Trace

1.5

Red

Ketchikan

8.50

0.85

1.5

Red

Hawk Inlet

7.89

Trace

Red

Hawk Inlet

8.56

Trace

Pink

Kodiak Island

9.38

0.3

1.8

Pink

Kodiak Island

7.94

Trace

1.5

Pink

Cook Inlet

9.34

0.2

2.5

Pink

Cook Inlet

9.16

0.3

2.5

Pink

Ketchikan

9.01

0.60

Pink

Ketchikan

9.58

0.3

Pink

Hawk Inlet

9.58

0.3

Pink

Hawk Inlet

8.82

Trace

King

Puget Sound

5.19

Trace

King

Puget Sound

5.30

Trace

Chum

Puget Sound

9.53

Trace

Chum

Puget Sound

10.97

Trace

Red

Puget Sound

8.46

0.3

Red

Puget Sound

9.16

0.3

uperimental pack

Red

Control, NCA

9.5

Trace

Red

Test, NCA

8.8

Trace

to 100 jul. The chromatograms showed few, if any,
indications of other volatile A^-nitrosamines
studied.

The multidetection method of Fazio, Howard,
and White (1971) was used to prepare four runs of
the same sample. The eluants from four silica gel
columns were combined into a 1-liter Kuderna-
Danish apparatus and concentrated to 1 ml and an
aliquot injected for GLC analysis. The concentrate
represented 100 g/ml of roe instead of the usual 25
g/ml. The increase in the area and height of the
NDMA peak was very pronounced. The extract
was submitted to an acid-Celite column cleanup.
Interferring peaks were removed, but the sus-
pected NDMA was still present (Figure 2).

In view of the above findings, it can be concluded
that less than 5 ppb of apparent NDMA was found
in salmon roe products of the different species of
salmon having been processed at five major loca-
tions, and that no other nitrosamines were
evident.

Acknowledgments

We express our appreciation to Donald M. Cros-
grove of NCA, Northwest Laboratories, for fur-
nishing us with the salmon roe samples and the
respective residual nitrite and chloride data.

lOOg roe/ml concentrate

TIME

Figure 2.-Chromatograms of an extract of nitrite-treated
salmon roe before and after being cleaned up on a column of acid
Celite.

Literature Cited

Fazio, T., J. N. Damico, J. W. Howard, R. H. White, and J. 0.
Watts.
1971. Gas chromatographic determination and mass spec-
trometric confirmation of A^-nitrosodimethylamine in
smoke-processed marine fish. J. Agric. Food Chem.
19:250-2.53.
Fazio, T., J. W. Howard, and R. H. White.

1971. Multidetection method for analysis of volatile A^-
nitrosamines in foods. Proc. Heidelb. Meet. Nitrosamines,
Int. Agency Res. Chem., 13-15 Oct.
Howard, J. W., T. Fazio, and J. 0. Watts.

1970. E.xtraction and gas chromatographic determination of
A^-nitrosodimethylamine in smoked fish: Application to
smoked nitrite-treated chub. J. Assoc. Off. Anal. Chem.
53:269-274.

D. F. Gadbois
E. M. Ravesi

R. C. LUNDSTROM

Nortliea.'<t Utilization Reaearch Center
National Marine Fisheries Service, NOAA
P.O. Box 61
Gloucester. MA 019.i0

686

UNDERWATER PAINT MARKING
OF PORPOISES^

Identification of individual animals has always
been a problem in cetacean behavioral research.
Only a small part of the animal is ordinarily
visible, and individuals within a pod of whales or
porpoises may all look very much alike, and, for
that matter, very much like all the individuals in
all neighboring pods. How does one mark (or label)
an animal at sea?

Our radio tagging experiments and flashing
light systems (Schevill and Watkins-) were design-
ed to provide a partial solution to this problem,
and, more recently, radio transmitters have been
attached to animals by means of harnesses or
other fastenings (Evans 1974; Norris and Gentry
1974; Norris et al. 1974). Conspicuous visual marks
have often been suggested, and a few have been
successfully contrived for particular experiments,
including freeze-branding, brightly colored
buoyant lines, buoys, and plastic numbered but-
tons toggled through dorsal fins (Norris and Pryor
1970; Evans et al. 1972)

We have been loath to use acoustic tags on
animals that react to the noise of ships, and even to
low-level pingers (Watkins and Schevill 1975).
Frequencies that are above their hearing would be
useful only at short ranges because of attenuation
of high frequencies in seawater.

Ideally, we wanted a mark that was highly
visible, that could be varied, that had no effect on
the behavior of the animal, that would last for long
periods of time, and that was easy to apply at sea.
Even a temporary mark permitting positive iden-
tification for only a few hours would be a boon.
Paint seemed an answer (Schevill 1966).

Materials and Methods

Several standard paint formulations were tried;
some could be applied to a wet surface, and some
would set relatively quickly underwater. Applica-
tion of these paints was easiest by pressurized
spray. We experimented with spray volumes,
velocities, propellants, and methods of controlling
the paint. A propellant that mixed well with the
paint carried it in a discrete stream, preventing

immediate mixing with the water, and higher
volumes of the paint mixture provided more
effective displacement of the water on the surface
to be painted. In our most satisfactory marking
system, we used 186-g (6-ounce) pressurized cans
of paint with a fire-extinguisher type of valve to
deliver short bursts of paint at about 125 g/s. A
nozzle 3 cm long with a 3.5-mm orifice was
fabricated to actuate the valve and direct the paint
in a coherent stream (in air, 2 or 3 m horizontally).
An internal modification to the standard container
removed the dip tube so that the can could be used
in an inverted position. For ease in handling and to
allow the stream of paint to be brought close to a
passing animal (as from the bow of a ship), a
holder for the paint can was mounted at the end of
a pole.

Paint bounced off most hard-surface materials
before it could set underwater, unlike human or
porpoise skin which appeared to have approxi-
mately equivalent temporary reactions to paint.
But paper masking tape (3M-Scotch 183),^ which
has a softer surface, reacted somewhat like skin to
both the paint and the water, and was used as an
underwater test surface.

Two paints were selected: a red lacquer based on
a nitrocellulose/alkyd vehicle and a red-orange
fluorescent based on an acrylic ester resin vehicle."*
These paints solidify by removal of the solvents
rather than by oxidation, as in the usual paint
preparations. The paint containers were capped at
about 4.2 kg/cm- (60 Ib/in^) at room temperature.
A 5% change of pressure can be expected with
each 5°C change in ambient temperature; can
temperature is critical for adequate pressure.

Tests were conducted in a 3-m-^ tank of flowing
seawater, and water temperatures were controlled
from 20.9°C in steps of a degree or less to 3.45°C,
and a comparison was made for each temperature
at several depths. Both paints penetrated the
water in a coherent stream, adhered to the test
surface, and set (hardened) underwater. The red
lacquer set within a second or two, but was con-
siderably dulled when applied through the water.
The fluorescent red-orange was largely unaffected
by underwater application, except that its setting
time was extended by 10-15 min. Patches of both

'Contribution No. 3586 from the Woods Hole Oceanographic
Institution.

-Schevill, W. E., and W. A. Watkins. 1966. Radio-tagging of
whales. Unpubl. manuscr., 15 p. Woods Hole Oceanogr. Inst. Ref.
No. 66-17.

^Reference to trade names and manufacturers does not imply
endorsement by the National Marine Fisheries Service, NOAA.

-•These two paints are similar to formulation AL-98 and V-129
by Lenmar, Inc., 150 South Calverton Road, Baltimore, Md.
These and other formulations and colors recommended by
Lenmar have been tested and appear to have equivalent under-
water characteristics.

687

paints applied underwater and kept immersed for
13 days and 7 h at 3.5°C showed only slight
differences from short-term tests. There was little
difference in the painted surfaces down to a
temperature of 4.0°C. Below this, less paint ad-
hered, and the color of the painted surface was
duller. The paint maintained a coherent stream to
greater depths in warmer water, perhaps because
of associated higher air temperature and there-
fore greater pressure in the can. With the appar-
atus above water, penetration and marking (at
15°C) occurred to a depth of about 40 cm, and with
increasing depth progressively less paint adhered.
Comparisons of application of these paints in both
seawater and fresh water showed little difference,
at least on a short-term basis.

Since only a portion of the paint actually ad-
hered underwater, the residue of these paints
floated as an inert scum in temperatures of 5°C or
warmer, generally not sticking to anything. This
was in sharp contrast to many other paints that
often floated as soft globules on the water, and for
hours thereafter would coat any objects they
contacted.

A Lagenorhynchus acutus was successfully
marked in the open ocean on 8 May 1975, 8 to 10 km
northeast of Race Point, Cape Cod, Mass. Though
L. acutuf^ usually is shy of ships and diflicult to
approach (Schevill 1956), we found about 30 of
these animals and were able to get close to a
subgroup of six porpoises. They would not surface
within reach of our vessel (13-m RV Asterias), so
the paint was applied through 15 to 20 cm of water.
The paint mark was a 10-cm circular red spot at
the after part of the buff-colored stripe on the right
side. We were able to follow this porpoise for only
30 min, but during this time, the mark provided a
highly visible tag which permitted rapid iden-
tification of the marked individual as well as the
subgroup of animals. This subgroup appeared to
stay together even when mingling with others of
the larger porpoise aggregation. Again, the paint
mark appeared to be ignored by all of the animals.
The next day, two schools of L. actus (probably
including the same animals) were studied, but no
mark could be found.

Discussion

Results

On 16 December 1974, we tested both paints on a
captive Tursiops (one of two in a tank) at the
Naval Undersea Center, San Diego, Calif. The
porpoise swam slowly past with all but its dorsal
fin underwater. The holder for the paint can was
hand-held about 20 cm above the water, and the
paint stream was directed downward at the an-
imal, about 20° from the vertical. The stream
penetrated the water by as much as 15-20 cm,
marking a streak 6-8 cm wide (at each pass) on the
animal's back, as well as on the right side of the
dorsal fin.

The paint contrasted sharply with the dark gray
color of the animal and provided a conspicuous
mark that was brightly visible 8 h after applica-
tion, although patches of it had disappeared.
Twenty-four hours after painting, only a small
strip of paint (at the leading edge of the dorsal fin)
remained, and much of this residue was still there
56 h after application, though quite dulled.

Of the two Tursiops in the tank only one was
painted, yet no obvious behavioral changes could
be noted; they both seemed to ignore the whole
process and behaved as before. There was no
. obvious reaction to either the painted animals or to
the excess paint floating on the water.

We suppose that the paint on the leading edge of
the dorsal fin of the captive porpoise persisted
longer than elsewhere because of the roughness
and scarring of the skin there. The disappearance
of the paint from the smooth surfaces on both the
captive and wild animals was apparently because
of the normally rapid sloughing of surface layers
of skin. Palmer and Weddell (1964: 555) noted that
cells in Tursiops skin undergo mitosis 250 to 290
times as rapidly as human skin, and Harrison and
Thurley (1972) also reported that cells in the
surface layer are desquamated in large numbers.
Presumably, the paint came loose because the
surface cells sloughed off. The relative stiffness
and greater mass of the cells coated with paint
would have accelerated their removal, but after
the paint had worn off, no difference in the skin
surface could be noted. We could find no indica-
tions of any adverse effects. Since the paint lasted
so much longer on the rough part of the fin, we
anticipate that similar nonsloughing surfaces on
the other cetacean species also would hold a paint
mark well (e.g., the highly barnacled portions of a
gray whale, or perhaps right whale bonnets). In
addition, we anticipate that such paints could
usefully mark other aquatic animals (turtles, seals,
manatees, etc.).

Little is known about color vision in porpoises.

688

though it has been assumed that they could see
color because of the relative numbers and ar-
rangement of rods and cones in the retina of
Tursiops (Perez et al. 1972). But since very little in
the animals' open ocean experience involves much
color, the painted marks may hold small sig-
nificance for them.

Since our purpose was to test the feasibility of
paint marking of porpoises, no attempt was made
to create an ideal paint, though a paint formulated
specifically for marking doubtless would have been
better than those we used. Our experiments began
with available paints, and those that were found to
coat wet surfaces were modified for use in pres-
surized containers with high volume valves. Paint
manufacturers generally are prepared to process
only large volume orders, but we found that
smaller specialty companies were able to prepare
formulations to order and modify small quantities
of pressurized paint containers.

Conclusions

Paint marking of porpoises provides a satisfac-
tory short-term tag that can be applied at sea. The
paint has not modified the animals' behavior and it
seems not to be detrimental in any way. The high
visibility of the colors we tried often made it
possible to locate the marked animal when other
porpoises of the school were obscured. The under-
water paint marking technique would appear to be
potentially useful in the study of other aquatic
animals.

Acknowledgments

We appreciate the help and advice extended by
G.V. Cass of Krylon Department, Borden, Inc.,
and Helene R. Johnson of Lenmar, Inc. We are
grateful also to the Naval Undersea Center, San
Diego, for their hospitality and good nature in
allowing our paint experiment, especially J. C.
Sweeney, Sam H. Ridgway, and William E. Evans.
Teresa Bray participated in laboratory test and
manuscript preparation. Support for this work
was from the Oceanic Biology Program of the
Oflfice of Naval Research, contract N00014-
74-C-0262.

Literature Cited

Evans, W.E.

1974. Radio-telemetric studies of two species of small

odontocete cetaceans. In W. E. Schevill (editor), The whale
problem, p. .385-394. Harvard Univ. Press, Camh., Mass.
Evans, W. E., J. D. Hall, A. B. Irvine, and J. S. Leatherwood.
1972. Methods for tagging small cetaceans. Fish. Bull U S
70:61-6.5.

Harrison, R. J., and K. W. Thurley.

1972. Fine structural features of delphinid epidermis.
(Abstr.) J. Anat. 111:498-500.
NoRRis, K. S., W. E. Evans, and G. C. Ray.

1974. New tagging and tracking methods for the study of
marine mammal biology and migration, hi W. E. Schevill
(editor). The whale problem, p. 395-408. Harvard Univ.
Press, Camb., Mass.
NoRRis, K. S., and R. L. Gentry.

1974. Capture and harnessing of young California gray
whales, Eschrhichtins rohustttx. Mar. Fish. Rev. 36(4):
58-64.

NoRRis, K. S., and K. W. Pryor.

1970. A tagging method for small cataceans. J. Mammal.
51:609-610.
Palmer, E., and G. Weddell.

1964. The relationship between structure, innervation and
function of the skin of the bottle nose dolphin (Tursiops
truncatiis). Proc. Zool. Soc. Lond. 143:553-567.
Perez, .J. M., W. W. Dawson, and D. Landau.

1972. Retinal anatomy of the bottlenosed dolphin (Tursiops
fruncatus). Cetology 11:1-11.
Schevill, W. E.

1956. Lagenorhynchus acutus off Cape Cod. J. Mammal.

37:128-129.
1966. Comments. In K. S. Norris (editor). Whales, dolphins,
and porpoises, p. 487. Univ. Calif. Press, Berkeley and Los
Ang.

W.ATKINS, W. A., AND W. E. SCHEVILL.

1975. Sperm whales (Physeter catodon) react to pingers.
Deep-Sea Res. 22:123-129.

William A. Watkins
William E. Schevill

Woods Hole Oceanographic Institution
Woods Hole, MA 02543

GRAZING OF FRESHWATER AND ESTUARINE,

BENTHIC DIATOMS BY ADULT ATLANTIC

MENHADEN, BREVOORTIA TYRANNUS

The diet of the Atlantic menhaden, Brevoortia
tymnnus (Latrobe), varies with stages in meta-
morphosis and the availability of food resources,
but it has been characterized consistently in the
literature as derived from the particulate organic
components of planktonic ecosystems (Reintjes
1969; June and Carlson 1971; Jeffries 1975; Peters
and Kjelson 1975; Durbin and Durbin 1975). Men-
haden larvae feed primarily by selective predation
on the larger estuarine zooplankters. Their meta-
morphosis into prejuveniles brings about the

689

development of a functional branchial filtering
apparatus which promotes a grazing of phyto-
plankton and suspended detritus. Late juveniles
and adults are primarily herbivores also but retain
the ability to eat zooplankton.

The stimulus for this investigation was a shore-
line observation of adult menhaden grazing di-
rectly on the benthic microbial communities cover-
ing the rocks in the headwaters of a Massachusetts
estuary. The fish were observed to bite or rip off
chunks of the benthic community film and swallow
them. This film was composed primarily of dia-
toms and detritus. Subsequent gut analyses of the
fish and the epilithic diatom assemblage confirmed
the field observations. Additionally, ingestion of
these benthic primary producers and their as-
sociated detritus by juvenile menhaden is pos-
tulated from a reinterpretation of previous
reports on their diet.

Methods

In the early afternoon of 19 September 1974,
nine adult menhaden (25-34 cm fork length) were
collected in the oligohaline region of the Slocum
River estuary, Mass. (Hoff et al. 1969). The fish
were sampled with a 10-m, 64-mm mesh haul seine
from a school of about 150, which was observed
feeding on the bottom within a 500-m'- area about
1 m deep for the 15 min prior to collection. The
pyloric stomachs were excised, opened, and their
fullness visually estimated. The stomach contents
of each fish were maintained and examined
separately; they were preserved in 3*^ formal-
dehyde solution. A preliminary microscopic ex-
amination of the contents was made to determine
the presence of diatoms and other components of
the diet. Diatoms were prepared for detailed
examination by a nitric acid-dichromate oxidation
of an aliquot of the sample followed by washing of
the cleaned frustules and mounting in Hyrax^
(Hohn and Hellerman 1963). Diatom populations
in each sample were identified and enumerated
from a random sample of about 200 frustules,
which were observed using oil-immersion phase-
contrast optics at a magnification of 1000 x .

On 21 September 1974, a 20-cm diameter rock
was removed from the same region of the estuary
in which the menhaden had been observed feed-
ing. The diatom assemblage on the rock was
air-dried, then scraped off and subjected to the

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

same procedures of preparation and examination
as those derived from the stomachs.

All samples and slides have been deposited in
the Hellerman Diatom Herbarium at Southeast-
ern Massachusetts University according to the
following collection numbers: HH918-HH926
(stomach samples) and HH927 (epilithic sample).

The diatom populations were classified as
freshwater, brackish, or marine based on the
habitat in which they grow optimally. This clas-
sification was derived primarily from the works of
Hustedt (1937-1938, 1939), Patrick and Reimer
(1966), Foged (1947, 1954), and Cleve-Euler
(1951-1955). Only those populations identified
without reservation to the species level were
classified ecologically. Additionally, in an
ecological classification of diatoms, identification
of populations to the level of variety is desirable
among multivarietal species, because frequently
different varieties of the same species have differ-
ent optimal habitats.

The terms "common" and "rare" as employed in
this paper, differentiate diatom populations hav-
ing greater than 1% or less than 1% mean relative
abundance, respectively, in the stomach samples.

Results

All fish stomachs were completely full or nearly
so. Amorphous detritus and diatoms composed the
bulk of the material with the detritus accounting
for the greater portion, but as estimated micro-
scopically, from 5 to 25^ of the volume was dia-
tomaceous. Most larger diatom cells were broken
and without contents, but many smaller diatoms
retained their chromatophores in structurally
intact frustules. Other microorganisms, par-
ticularly filamentous blue-green algae and nema-
todes, were evident, and the remnants of some
microcrustaceans were noted in a few stomachs.

The examination of about 1,800 diatom in-
dividuals from the stomachs revealed 163 popula-
tions of which 134 were identified to species or
variety. Twenty-three populations were common
and only three of them were not assignable to a
particular species (Table 1). The rare populations
which were unidentified constituted less than 2%
of all individuals. Practically all the populations
are benthic. Eight of them, particularly Skeleton-
ema costatum and Thalassiosira spp., are con-
sidered planktonic, but they contained less than
7% of all individuals and were found also on the
rock. Freshwater populations composed 50*^ of the
common and nearly 70% of all populations (Table

690

Table l.-The relative abundance (%) and the optimal habitats of the common diatoms occurring both in the
stomachs of Atlantic menhaden and on a rock from the Slocum River estuary, Mass. Only those populations
having greater than 1% mean relative abundance in the stomachs are listed.

Diatoms

Fish stomachs

Rock

Optimal
habitat'

Mean

Range

8.0

4.1-

13.3

2.3

8.0

1.5-

13.5

0.5

6.8

3.6-

9.6

5.0

6.7

1.8-

18.9

1.4

4.5

1.9-

6.3

0.5

B/M

4.2

1.5-

8.7

4.6

3.0

1.4-

4.5

2.7

2.9

0.5-

5.9

1.4

2.8

0.9-

5.1

1.8

2.8

0.5-

5.8

5.5

2.8

0.9-

5.0

0.5

2.3

0.0-

4.1

0.5

2.2

0.5-

4.0

1.4

2.1

0.9-

5.9

2.7

2.0

0.0-

5.0

2.7

2.0

0.0-

4.0

1.4

1.9

1.4-

2.7

1.8

F/B

1.7

0.9-

4.4

1.4

1.7

0.5-

1.9

0.5

1.5

0.9-

2.8

0.9

1.5

0.0-

1.5

25.6

1.3

0.5-

3.0

4.1

1.1

0.0-

2.7

1.8

Nitzschia Irustulum var. perminuta Grun.

N. subtilis var. paleacea Grun.

Navicula cincta (Ehr.) Ralfs

Melosira nummuloides (Dillw.) Ag,

Skeletonema costatum (Grev.) CI.

Cyclotella sp. cf. glomerata Bachm.

Eunotia pectinalis (Dillw.) Rabh. var. pectinalis

Achnanthes minutissima Kutz.

Bacillaria paradoxa Gmel.

Cyclotella sp. cf. atomus Hust.

Melosira varians Ag.

Navicula diserta Hust.

Navicula capitata var. hungarica (Grun.) Ross

Eunotia pectinalis var. minor (KiJtz.) Rabh.

Fragilaria construens var. venter (Ehr.) Grun.

Navicula gregaria Donk.

Rhoicosphenia curvata (KiJtz.) Grun.

Nitzschia Sigma W. Sm.

Thalassiosira sp. cf. nana Hust.

Achnanthes wellsiae Reim.

Nitzschia parvula Lewfis

Cyclotella striata (KiJtz.) Grun.

Fragilaria construens var. intercedens (Grun.) Hust.

Total

73.8

71.0

'F = freshwater, B = bracl<ish, M = marine.

2). They accounted also for more than 50% of all
individuals. Brackish and marine populations were
present in about equal numbers, but more common
populations were brackish. Nearly 35% of all
individuals belonged to brackish populations. All
common populations in the stomachs were at least
present on the rock, and 17 of the 23 also accounted
for greater than 1% relative abundance in the
epilithic assemblage (Table 1). Additionally, 24
rare populations were found in both the stomach
and the epilithic samples. The greater number of
rare populations found in the stomachs as com-
pared to the rock is attributable to the greater
sample size associated with the stomachs. These
rare populations were primarily species of Ach-
nanthes, Amphora, Cocconeis, CymbeUa, Eunotia,

Table 2.-The distribution of numbers of common and rare
diatom populations from the stomach and epilithic assemblages
among their optimal habitats (F = freshwater, B = brackish, M
= marine). Populations interpreted as growing equally well in
two habitats are divided equally between them.

Diatom

Stomachs'

Rock^

populations

B M

Total

Common

11.5

6.0 2.5

20.0

11.5

6.0

2.5

20.0

Rare

79.5

15.0 19.5

114.0

12.0

2.5

8.5

23.0

All

91.0

21.0 22.0

134.0

23.5

8.5

11.0

43.0

'Total sample size ~ 1,800 individuals.
^Total sample size = 200 individuals.

Fragilaria, Gomphonema, Navicula, Nitzschia,
and Synedra.

Based on the examination of about 200 in-
dividuals from the epilithic assemblage, 43
populations were identified to species or variety.
Twenty identified and three unidentified popula-
tions were common in the fish stomachs (Table 1).
Only four other populations were unidentified, and
they represented less than 3% of all individuals in
the sample. All populations are benthic. Given the
means of collection of the epilithic assemblage,
those populations usually considered planktonic
were clearly benthic. They accounted for about 7%
of the total number of individuals in the assemb-
lage, as they did in the stomach samples. Nearly
90% of all populations found on the rock were
recorded in the stomachs. Freshwater populations
accounted for about 50% of both the common and
all populations (Table 2). Brackish and marine
populations occurred equally among all popula-
tions, but among the common ones, brackish
populations were more frequent. A population of
the freshwater diatom, Nitzschia parvula, consti-
tuted 25% of the whole assemblage.

Discussion

The benthic microbial communities of estuaries
and the adjacent freshwater reaches of rivers, as

691

well as probably those of shallow marine coastal
waters, are utilized directly as a food resource by
adult and juvenile menhaden. Our field observa-
tions of their grazing habits, the preponderance of
benthic diatoms in their stomachs, and the taxo-
nomic and ecological similarity of the diatom
assemblages in their stomachs with that of the
benthos support this conclusion. The composition
of the stomach and epilithic samples is commen-
surate with the expectations of random sampling
of the benthos in this region of the estuary. The
quantitative characteristics of estuarine benthic
diatom assemblages can be extremely variable
within a small space, even on similar substrates
(Mclntire and Overton 1971; Round 1971; Main and
Mclntire 1974), and so the expectation of quaniti-
tative identity among random samples is low. But,
much greater quantitative similarity is expected
of samples from similar substrates in the same
area.

The data of other investigators, but not their
conclusions, support our findings. In a study of the
diet of juvenile menhaden collected between April
and June 1961, in Delaware, June and Carlson
(1971) found most frequently eight genera of
diatoms present in their guts: "Pleurosigma,
Navicula, Nitzschia, Cyclotella, Melo^ira, Am-
phora, Gyrosigma, and SurireUa.'" All these gen-
era are characteristically benthic in marine and
estuarine ecosystems. Compared to the list of
diatom genera they reported from the phyto-
plankton, which they sampled between November
1960 and May 1961, in the same area, the eight
genera accounted on the average for less than 10%
of the total number of diatom phytoplankters.
Furthermore, they reported that Skeletonema,
Coscinodiscus, Rhizosolenia, Thalassiosira, and
Thalassiothrix composed on the average 75% of
the diatom phytoplankton, but all were unrecorded
from their gut analyses of the fish. We conclude
from their data that the juvenile menhaden they
collected were not grazing primarily on the
plankton but rather on the benthos. Likewise,
Mulkana (1966) reported six diatom genera from
the stomachs of juvenile menhaden collected in
Rhode Island estuaries, and four of the six are
usually benthic: Gyrosigma, Grammatophora,
Achnanthes, and Navicula. Although the diatoms,
whether planktonic or benthic, appear to consti-
tute a less significant portion of the diet of
juveniles and adults in estuaries than does detritus
(Jeffries 1975; Peters and Kjelson 1975), they
accurately reflect the immediate source of the

detritus, because they are good habitat labels
(Round 1964, 1971).

Both juvenile and adult menhaden tolerate
salinities of less than T/oo (Reintjes 1969), but we
know of no records other than our own of their
feeding on primarily freshwater or oligohaline
resources.

The Atlantic menhaden is among the commer-
cially most important species in the United States
fishery, and consequently, the factors which
regulate its population size are of considerable
interest. Assuming that human and other preda-
tors are prudent, trophic energy availability is
likely to be limiting. McHugh (1967) has postulated
that "the rate of plankton production will limit the
numbers of menhaden . . . that a particular body of
water can support." If we interpret the concept of
the plankton liberally, including the living organ-
isms plus the suspended detritus, the idea is
certainly tenable; however, it is conditional upon
the menhaden's grazing being restricted to the
plankton. Also, adult menhaden's minimum-size
threshold for filtration of particles appears to be
around 15 p.m with the consequence that a sub-
stantial portion of the phytoplankton will be
unavailable to them (Durbin and Durbin 1975).
But, considering the productivity of benthic
primary producers and the quantities of
sedimented detritus in shallow estuaries (Darnell
1967; Odum 1971; Smayda 1973), the menhaden's
exploitation of the benthos, potentially, at least,
doubles the energy available to it. Unfortunately,
the quantitative significance of their benthic
grazing habits and their ability to assimilate the
ingested materials during the estuarine portions
of their life cycle are unassessed.

Jeffries (1975) has characterized the menhaden
as an adaptable species capable of switching from
one food resource to another, and thus compen-
sating for the variability in the availability of
estuarine food resources. In general, this apparent
switching in juveniles and adults is more the
product of a fine-grain feeding in resource-dif-
ferent habitats than of coarse-grain feeding on
the plankton. Our observations extend this mode
of feeding in menhaden to include benthic
habitats.

Literature Cited

Cleve-Euler, a.

1951-55. Die Diatomeen von Schweden und Finnland. K.
Sven. Vetenskaps akad. Handl., Ser. 4, 1,2(l):l-163, (1951);

692

II,4(1):1-158, (1953); III,4(5):l-255, (1953); IV,5(4):l-232,
(1955); V,3(3):l-153, (1952).
Darnell, R. M.

1967. Organic detritus in relation to the estuarine ecosys-
tem. In G. H. Lauff (editor), Estuaries, p. 376-382. Am.
Assoc. Adv. Sci. Publ. 83.

DURBIN, A. G., AND E. G. DURBIN.

1975. Grazing rates of the Atlantic menhaden Brevoortia
tyrannus as a function of particle size and concentration.
Mar. Biol. (Berl.) 33: 265-277.
FOGED, N.

1947. Diatoms in water-courses in Funen. II. Lindved AA
(The Lindved Brook). III. Odense AA (The Odense Brook).
Dan. Bot. Ark 12(6):l-69. 1954. On the diatom flora of some
Funen lakes. Folia Limnol. Scand. 6, 75 p.
HoFF, J. G., P. Barrow, and D. A. McGill.

1969. Some aspects of the hydrography of a relatively
unpolluted estuary in southeastern Massachusetts. Proc.
24th Ind. Waste Conf., Purdue Univ. Eng. E.xten. Ser.
135:87-98.
HoHN, M. H., AND J. Hellerman.

1963. The taxonomy and structure of diatom populations
from three eastern North American rivers using three
sampling methods. Trans. Am. Microsc. Soc. 82:250-329.
HUSTEDT, F.

1937-1938. Systematische und okologische Untersuchungen
uber die Diatomeen-Flora von Java, Bali and Sumatra.
Arkiv Hydrobiol. Suppl.-Bd. XV: "Tropische Binnenge-
wasser" 506 p.
1939. Die Diatomeenflora des Kustengebeites der Nordsee
vom Dollart bis zur Elbe-mundung. Abh. Naturwiss. Ver.
Bremen 31:572-677.
Jeffries, H. P.

1975. Diets of juvenile Atlantic menhaden {Brevoortia
tyrannus) in three estuarine habitats as determined from
fatty acid composition of gut contents. J. Fish. Res. Board
Can. 32:587-592.
June, F. C., and F. T. Carlson.

1971. Food of young Atlantic menhaden, Brevoortia tyran-
nus, in relation to metamorphosis. Fish. Bull., U.S.
68:493-512.
Main, S. P., and C. D. McIntire.

1974. The distribution of epiphytic diatoms in Yaquina
Estuary, Oregon (U.S.A.). Bot. Mar. 17:88-99.
McHuGH, J. L.

1967. Estuarine nekton. In G. H. Lauff (editor), Estuaries,
p. 581-620. Am. Assoc. Adv. Sci. Publ. 83.
McIntire, C. D., and W. S. Overton.

1971. Distributional patterns in assemblages of attached
diatoms from Yaquina Estuary, Oregon. Ecology 52:
758-777.
Mulkana, M. S.

1966. The growth and feeding habits of juvenile fishes in
two Rhode Island estuaries. Gulf Res. Rep. 2:97-167.
Odum, E. p.

1971. Fundamentals of ecology. 3rd ed. W. B. Saunders,
Phila., 574 p.
Patrick, R., and C. W. Reimer.

1966. The diatoms of the United States exclusive of Alaska
and Hawaii. Vol. 1: Fragilariaceae, Eunotiaceae, Ach-
nanthaceae, Naviculaceae. Monogr. Acad. Nat. Sci. Phila.
13, 688 p.
Peters, D. S., and M. A. Kjelson.

1975. Composition and utilization of food by postlarval and

juvenile fishes of North Carolina estuaries. In L. E. Cronin
(editor), Estuarine Research. Vol. 1, p. 448-472. Academic
Press, N.Y.
Reintjes, J. W.

1969. Synopsis of biological data on the Atlantic menhaden,
Brevoortia tyrannus. U.S. Fish Wildl. Serv., Circ. 320, 30 p.
Round, F. E.

1964. The ecology of benthic algae. In D. F. Jackson (edi-
tor), Algae and man, p. 138-184. Plenum Press, N.Y.
1971. Benthic marine diatoms. Oceanogr. Mar. Biol., Annu.
Rev. 9:83-139.
Smayda, T. J.

1973. Phytoplankton. In Coastal and offshore environmen-
tal inventory: Cape Hatteras to Nantucket Shoals, Sec. 3,
"~~--JQO^ Mar. Publ. Ser. 2, Univ. Rhode Island, Kingston.

Robert K. Edgar
James G. Hoff

Hellerman Diatom Herbarium
Southeastern Massachusetts University
North Dartmouth. MA 0271,7

ELECTROPHORETIC EVIDENCE OF

HYBRID SNOW CRAB,

CHIONOECETES BAIRDI X OPILIO

Karinen and Hoopes (1971) and Hoopes et al.
(1972) reported finding snow (Tanner) crabs in the
southeastern Bering Sea which possessed mor-
phological characteristics that were atypical for
either Chionoecetes bairdi or C. opilio and, instead,
were intermediate. The females of this form
appeared to have reduced reproductivity, as many
were nongravid at maturity, and those that were
gravid possessed abnormally small egg clutches
containing large numbers of dead eggs. These
conditions were presented as evidence of hy-
bridization. Hybrid-type males constituted 1.0% of
all male snow crabs captured, while hybrid-type
females made up 0.4% of the females captured.

Karinen (1974) confirmed the above reports and
found that hybrids made up 4.6% of the snow crabs
collected in the Bering Sea and were most abun-
dant west of lat. 166°W. The carapace width
frequency of the hybrids was intermediate
between C. bairdi and C. opilio-prowiding addi-
tional evidence of hybridization.

The purpose of the present study was to deter-
mine if electrophoretic differences between the
parent species and the hybrid could be detected.

The samples used were collected from the
southeastern Bering Sea in July 1974, identified,
and frozen by National Marine Fisheries Service
(NMFS) personnel. The general proteins of leg

693

A
AB

Hybrids C.opilio C.bairdi Bands

Origin -

Figure l.-Electropherogram of starch gel showing general muscle protein patterns of CJi inoei-ftex ha inJi. C. opilio, and hybrids.

muscle tissue from 10 C. hairdi, 5 hybrids, and 10
C. opilio were examined electrophoretically using
the methods of Johnson et al. (1972) and the buffer
system of Ridgway et al. (1970).

The electrophoretic patterns of general muscle
proteins are shown in Figure 1. All C. opilio
patterns possessed a single band (A), while all C.
hairdi showed a slower anodally migrating band
(B). The five hybrids possessed three bands: A, B,
and an intermediate band AB which indicates
hybridization between C. bairdi and C. opilio.

The intermediate band (AB) was less intense
than either of the other bands (A or B). A 1:2:1
ratio is expected in random combination of
dimeric protein. I thus assume that there is non-
random association between the protein units.

Further investigation is needed to determine if
the electrophoretic patterns reported here are
evident in all possible crosses between the two
parent species and that the parental patterns are
invariant throughout their ranges.

Acknowledgments

I thank Robert J. Wolotira, Jr. (Northwest
Fisheries Center, NMFS, NOAA, Seattle, Wash.)
for providing identified crab samples for this
report.

Literature Cited

HooPES, D. T., J. F. Karinen, and M. J. Pelto.

1972. King and Tanner crab research. Int. North Pac. Fish.
Comm., Annu. Rep. 1970:110-120.
Johnson, A. G., F. M. Utter, and H. 0. Hodgins.

1972. Electrophoretic investigation of the family Scorpaen-
idae. Fish. Bull., U.S. 70:403-413.

Karinen,.!. F.

1974. King and Tanner crab research, 1971. Int. North Pac.
Fish. Comm., Annu. Rep. 1972:102-111.
Karinen, J. F., and D. T. Hoopes.

1971. Occurrence of Tanner crabs {Cliionoecetea sp.) in the
eastern Bering Sea with characteristics intermediate
between C. hairdi and C. opilio. (Abstr.) Proc. Natl.
Shellfish Assoc. 61:8-9.
Ridgway, G. J., S. W. Sherburne, and R. D. Lewis.

1970. Polymorphism in the esterases of Atlantic herring.
Trans. Am. Fish. Soc. 99:147-151.

Allyn G. .Johnson

Northurst Fiftheriex Center

National Marine Fisheries Service, NOAA

Seattle, Wash.

Present address: Gulf Coastal Fisheries Center Port

Aransas Laboratory, NMFS, NQAA

West Port Street, Port Aransas, TX 78373

EFFECTS OF BENZENE ON GROWTH,

FAT CONTENT, AND CALORIC CONTENT

OF STRIPED BASS, MORONE SAXATILIS

The San Francisco Bay area is a major terminus
and refinery area for crude oil, and oil-related
activities in the area are expected to increase
because of the Alaska pipeline and expanded
drilling on the outer continental shelves of
California and Alaska. The San Francisco Bay-
delta region supports a number of fisheries, in-
cluding the most important recreational striped
bass, Morone saxatilis, fishery on the west coast.
Information on the toxicity of aromatics in crude
oil to striped bass and other fisheries is needed.

694

The aromatic hydrocarbon, benzene, is one of
the major water-soluble components of crude oil.
Anderson et al. (1974) reported 6.75 and 3.36"/»o in
the water-soluble fractions of south Louisiana and
Kuwait crude oil standards respectively. In addi-
tion to being relatively soluble in water (1,780 "/ou -
McAuliffe 1966), benzene is one of the most
toxic components of petroleum.

The acute 96-h, TL-50 lethal level (10-11 nl/Yiter)
of constant benzene exposure for juvenile striped
bass was determined previously at our laboratory
by Meyerhoff (1975). The objective of experiments
described here was to see if sublethal levels of
benzene, although not inducing death, would
inhibit efficient energy utilization by the fish as
measured by growth (wet weight, dry weight), fat
content, and caloric content. Because the exper-
imental period of 4 wk was relatively short, the
juvenile striped bass were exposed to mean high-
sublethal concentrations (3.5 /il/liter, SD 1.4; 6.0
,ul/liter, SD 1.6) to determine the efi'ects of benzene
on growth.

Methods

Juvenile striped bass (mean standard length
18.1 cm, SD 2.3; mean total wet weight 3.39 g, SD
1.1) were obtained from the Tracy pumping plant
operated by the Bureau of Reclamation, Tracy,
Calif. After being transported by truck to our
facility (Korn 1975) the fish were changed to saline
water (267»ii) during a 3-day period. Juvenile fish
occur naturally at this salinity as well as in fresh
water. The fish were acclimated for 2 wk to test
conditions (salinity 26"/n(i, temperature 15°-16°C,
pH 7.8). Thirty-five fish were then placed into each
of nine 80-liter fiber glass aquariums and ac-
climated for one more week. Halver's diet (1957) in
pelleted form (5.350 kcal/g) was fed at the rate of
3% of fish body weight per day.

Benzene concentrations were maintained in
three aquariums at 3.5 jul/liter benzene and in
three at 6 jul/liter benzene; three others served as
controls (0 jul/liter). Relatively constant benzene
concentrations were maintained using the method
of Benville and Korn (1974). The input of ben-
zene-saturated air was balanced by a continuous 2
liters/min water flow through the aquariums.

Benzene concentrations were monitored daily
using the gas chromatograph procedure of Ben-
ville and Korn (1974). Water quality conditions
during the test were as follows: temperature,
15.2°-16.4°C; oxygen, 7.5-7.9 mg/liter; salinity,
25-267to; pH, 7.7-7.8; ammonia, <0.5 mg/liter.

Seven fish were sampled from each aquarium at
0, 7, 14, 21, and 28 days. The animals were anesthe-
tized with MS-222,' killed by severing the spinal
cord, blotted dry, weighed individually, dried in a
70°C oven for 4 days, cooled in a desiccator, and
reweighed. Three of the fish were then processed
for caloric analyses and four for fat analyses.

Calorimetric content was analyzed by in-
dividually processing three fish in a Parr adiabatic
calorimeter, model 1241.

For fat analyses, the four dried fish were blend-
ed with 150-ml MF Freon (monoflourotrichloro-
methane) in a high-speed blender. The mixture
was poured and rinsed into a Buchner vacuum
filter through No. 1 filter paper. The filtrate was
put into preweighed beakers and evaporated in a
hood to dryness. After reweighing the beakers, fat
content was calculated.

Data were analyzed with an analysis of variance
for factorial design program (BMD 02V— Dixon
1973). The independent factors of tank, week,
concentration, and their interactions were tested
for significance of effect on the dependent varia-
bles of wet weight, dry weight, fat content, and
caloric value. Duncan's new multiple-range test
(Duncan 1955; Pachares 1959) was used to deter-
mine the significant diflferences between means of
levels for treatments found significant in the
analysis of variance.

Results

Benzene concentrations varied because of fluc-
tuations in water flow caused by particulate ma-
terial clogging the valves. The high-level treat-
ment varied from 3.6 to 8.1 jul/liter during the 4-wk
test; the low-level treatment varied from 1.5 to 5.4
jul/liter. Analysis of variance of the benzene water
concentration showed a significant (P<0.01) in-
crease at both levels over the test period. However,
the means of low (3.5 jul/liter, SD 1.4) and high (6.0
jul/liter, SD 1.6) concentrations were significantly
different (P<0.01).

The start of benzene exposure caused pro-
nounced hyperactivity at the high level and a
moderate effect at the low level. The fish reacted
by attempting to jump out of the water. Fish
exposed to the high level attempted to feed but
were unable to locate and consume their ration.
Random jerking movements were observed when

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

695

food was introduced. Fish exposed to the low level
had some success locating the food, and approx-
imately 50% of the pellets were consumed. Control
fish successfully consumed all of their ration
within 5 min.

After 1 wk, feeding success on low- and
high-dose fish started improving gradually. At the
end of the study, the control and low-level fish fed
normally, while the high-level fish consumed 50%
of their ration.

Analyses of variance of wet weight, dry weight,
kilocalories per gram ash-free dry weight, and
percent fat between concentrations, weeks, and
tanks yielded the following results (Tables 1, 2).
There was a significant decrease in wet weight
(P<0.05), dry weight (P<0.01), and percent fat
(P<0.01) with increasing concentration (Table 2).
Concentration levels varied significantly (P<0.05)
(Table 1): Wet weight was less at 6.0jLii/liter than
controls and did not vary significantly between 3.5
jul/liter and controls or between 3.5 and 6.0 n\/\\ter.
Dry weight was less at 6.0 jul/liter than at 3.5
/xl/liter and controls but did not vary significantly
between controls and 3.5 jul/liter. Percent fat was
less at 6.0 and 3.5 /xl/liter than in controls. There
was no significant difference in percent fat
between 6.0 and 3.5 fxl/liter.

There was a significant increase in dry weight
(P<0.05) during the last week at all exposures
(Table 2, Figure 1). There was no significant
difference between treatments in kilocalories per
gram ash-free dry weight (Table 2). The sig-
nificant interaction between concentration and
tank (P<0.05-Table 2) is a result of experimental
design in which certain tanks were always at a

Table l.-Mean wet wet and dry weights and fat caloric content
of one control and two test groups of striped bass, Morone
sajcatilis, exposed to benzene for 4 wk.

Treatment

mean

concentration

Oi I/liter)

Variable^

Wet
weight

(g)

Dry
weight

(g)

Fat
(%)

Ash-free

dry weight

(kcal/g)

Control

Low level (3.5)

High level (6.0)

Total number
of fish

12.7135
12.6062
2.3951

315

10.8721
10.8137
0.7242

315

39.2

6.8123

34.1

6.8435

32.2

6.7451

'The three treatments used three replicate tanks per treatment
sampled at 0, 1, 2, 3, and 4 wk. Tests for wet and dry weights had
seven fish/tank per week; tests for percent fat had four fish/tank
per week; and tests for kilocalories/gram ash-free dry weight had
three fish/tank per week.

^Duncan's new multiple-range test of differences between
means of treatment levels was performed. Means grouped above
with same bar are not significantly different at the 5% level.
Means not grouped with same bar are significantly different at
the 5% level (Duncan 1955; Pachares 1959).

high or low concentration. No significant variation
occurred between tanks.

Discussion

Although acclimated, fish in all treatments were
stressed from crowding and insufficient water
movement. This was unavoidable because space
and equipment were limited. Consequently, the
control fish did not grow at the same rate as
similar fish held in larger tanks at this facility. In
spite of these limitations, significant relative
changes in growth rate and fat content did occur
between exposure treatments. Wet weight, dry
weight, and fat content decreased with increasing
concentration as expected. This was probably due

Table 2.-Analysis of variance of treatment effects of benzene
concentration (>il/liter), week, and tank number on wet weight
(g), dr>- weight (g), kilocalories per gram ash-free dry weight,
and percent fat of juvenile striped bass, Morone saxatilis.

Dependent variable and
source of variation

Sum of
squares

Mean
square

F
ratio

Proba-
bility

Wet weight:
Concentration
Week
Tank

Concentration-week
Concentration-tank
Week-tank
Concentration-week-
tank
Within (error)

Total

Dry weight:
Concentration
Week
Tank

Concentration-week
Concentration-tank
Week-tank
Concentration-week-
tank
Within (error)

Total

Kilocalories per gram
ash-free dry weight:

Concentration

Week

Tank

Concentration-week

Concentration-tank

Week-tank

Residual (error)

Total

Percent fat:
Concentration
Week
Tank

Concentration-week
Concentration-tank
Week-tank
Residual (error)
Total

2
4
2

8
4
8

16
270

5.511

7.750

3.050

11.673

11.255

11.673

15.516
235.675

2.756
1.938
1.525
0.909
2.814
1.459

0.970
0.873

3.16
2.20
1.75
1.04
3.22
1.67

P< 0.05

1.11 NS

314 302.103 — — —

4
2
8
4
8

16
270

314

2
4
2

8
4
8

1.165
1.232
0.420
0.933
1.214
1.367

2.166
29.137

37.634

0.076
0.404
0.284
1.177
0.147
0.667
1.222

3.977

0.583
0.308
0.210
0.117
0.304
0.171

0.135
0.108

0.038
0.101
0.142
0.147
0.037
0.083
0.076

5.40
2.85
1.94
1.08
2.81
1.58

P< 0.01

P< 0.05

1.25 NS

0.50
1.33
1.87
1.93
0.49
1.09

NS
NS
NS
NS
NS
NS

2 383.902 191.951 13.42 P< 0.01

4
2
8
4
8
16

99.539

52,878

138.843

55.000

190.733

228.929

24.885
26.439
17.355
13.750
23.842
14.308

1.74
1.85
1.21
0.96
1.67

NS
NS
NS
NS
NS

44 1,149.824 —

NS = not significant.

696

X
(3

>-
01

00-

96-

• CONTROL
■ LOW LEVEL
» HIGH LEVEL

WEEK

Figure 1. -Average weight of each of three groups of striped
bass, Morone saxatilis, exposed to three concentrations of
benzene (0, 3.5, and 6.0 n\/\\ier) for 4 wk. The dry weight of
high-level exposure fish was significantly less (P^O.Ol) than the
other two groups at the end of the first week and thereafter. The
dry weight of the three groups combined was significantly higher
(P<0.05) than in previous weeks.

mostly to impaired food localization at higher
concentrations. A similar effect on the nervous
system is documented by Brocksen and Bailey
(1973). The energy required to metabolize benzene
could also decrease efficient utilization of energy
for growth and fat deposition.

There was an apparent acclimation of the fish to
benzene at the low level (3.5 jul/liter) by the end of
the 4-wk exposure, as reflected by the dry weight
of the fish (Figure 1). After 4 wk at high level (6.0
jLil/liter), fish also appeared to begin to recover
from effects. This was substantiated by observa-
tions of improved feeding response in exposed fish
as the experiment progressed. Nevertheless, de-
finite effects of benzene on growth parameters
were noted at 6.0- and 3.5-jMl/liter levels of ben-
zene. Although the fish may be able to adapt by
metabolic detoxification and depuration of ben-
zene and metabolites, after more prolonged peri-
ods the competitive effects on energy utilization
may not only decrease growth but also increase
mortality or reduce ability to withstand environ-
mental stress.

The parameters measured in this study show-
effects at the low /xl/liter levels. In most situations,
it is unlikely that fish would be exposed to benzene
above the nl/liter level except shortly after catas-
trophic spills. Anderson et al. (1974) obtained a
concentration of several jal/liter benzene in
water-soluble extracts of crude oils. In the marine
environment, dilution and volatilization of ben-
zene would probably lower the concentration of
benzene rapidly. Research on effects at the nl/liter
level is needed along with monitoring information

on actual concentrations of benzene in chronically
polluted environments. Such situations may in-
duce a reduction in growth rate and fat deposition
which would have implications in the reproductive
potential of exposed species. Studies of chronic
effects of low concentrations of benzene on re-
production, including fecundity, egg size, em-
bryonic development, and larval survival, are
indicated. Some of these studies have been com-
pleted at the Tiburon Laboratory and will be
reported on later.

Acknowledgments

We thank Tina Echeverria and Richard Paris of
the NMFS Southwest Fisheries Center Tiburon
Laboratory, for calorimetric analyses and assist-
ance in data processing respectively. We also
thank John Hunter, Southwest Fisheries Center
La Jolla Laboratory, and Stanley Rice, Northwest
Fisheries Center Auke Bay Fisheries Laboratory,
for reviewing the manuscript.

Literature Cited

Anderson, J. W., J. M. Neff, B. A. Cox, H. E. Tatem, and G. M.

HiGHTOWER.

1974. Characteristics of dispersions and water-soluble
extracts of crude and refined oils and their toxicity to
estuarine crustaceans and fish. Mar. Biol. (Berl.) 27:75-88.
Be.wille, p. E., Jr. and S. Korn.

1974. A simple apparatus for metering volatile liquids into
water. J. Fish. Res. Board Can. 31:367-368.

Brocksen, R. W., and H. T. Bailey.

1973. Respiratory response of juvenile chinook salmon and
striped bass exposed to benzene, a water-soluble compo-
nent of crude oil /» Proceedings of Joint Conference on
Prevention and Control of Oil Spills, p. 783-791. Am. Pet.
Inst., Environ. Prot. Agency, U.S. Coast Guard, Wash.,
D.C.
Dixon, W. J. (editor).

1973. BMD biomedical computer programs. Univ. Calif.
Press, Berkeley, 773 p.
Duncan, D. B.

1955. Multiple range and multiple F tests. Biometrics
11:1-42.
Halver, J. E.

1957. Nutrition of salmonid fishes. III. Water-soluble
vitamin requirements of chinook salmon. J. Nutr.
62:225-243.
KoRN, S.

1975. Semiclosed seawater system with automatic salinity,
temperature, and turbidity control. U.S. Dep. Commer.,
NOAA Tech. Rep. NMFS SSRF-694, 5 p.

McAULIFFE, C.

1966. Solubility in water of paraflin, c\'cloparaflin, olefin,
acetylene, cycloolefin, and aromatic hydrocarbons. J.
Phys. Chem. 70:1267-1275.
Meyerhoff, R. D.

1975. Acute toxicity of benzene, a component of crude oil, to

697

juvenile striped bass {Morone saxatilis). J. Fish. Res.
Board Can. 32:1864-1866.
Pachares, J.

1959. Table of the upper 10% points of the studentized
range. Biometrika 46:461-466.
Parr Instrument Co.

1969. Instructions for D41 and D42 adiabatic calorimeters.
Man. 142. Parr Instrum. Co., 211 53d St., Moline, 111., 23 p.

Sid Korn

Southwest Fisheries Center Tiburon Laboratory

National Marine Fisheries Service, NOAA

Tiburon, Calif.

Present address:

North west Fisheries Center Auke Bay Fisheries

Laboratory, NMFS, NOAA

P.O. Box 155, Auke Bay, AK 99821

Jeannette W. Struhsaker
Pete Benville, Jr.

Southwest Fisheries Center Tiburon Laboratory
National Marine Fisheries Service, NOAA
3150 Paridise Drive
Tiburon, CA 91920

Marine Science Center (MSC) at Newport, exposed
to ultraviolet light (3.785 liters/min), diluted
(when necessary) to 25"/uo with distilled water, and
stored in Nalgene carboys. This salinity is within
the range recommended for C. virginica by Davis
and Calabrese (1964), and was used for mainte-
nance of oysters and for experiments on fertiliza-
tion and early larval development. In laboratory
procedures, all glassware was initially acid-
washed; used glassware was carefully cleaned and
rinsed several times first in tap water and then in
distilled water; all polyethylene tubing was
Tygon^ R3606 (nontoxic by bioassay, Breese, MSC,
unpubl. data); gametes and larvae were confined in
glass containers only (except for momentary
exposure to stainless steel syringe needles and
nylon screen); all seawater used in fertilization
experiments was Millipore-filtered (0.47 jum) and
stored in glass screw-cap bottles with Parafilm-
lined caps (nontoxic by bioassay, Breese unpubl.
data).

FERTILIZATION METHOD QUANTIFYING

GAMETE CONCENTRATIONS AND

MAXIMIZING LARVAE PRODUCTION

IN CRASSOSTREA GIG AS

Most workers obtain oyster larvae by using ex-
perimental methods similar to those reported by
Galtsoff" (1964). Although useful in most hatchery
or laboratory investigations, these methods do not
quantify gamete concentrations. To obtain
specific larval concentrations, most researchers
dilute dense postfertilization concentrations.

This paper reports on a method of estimating
sperm concentrations of Pacific oyster, Crassos-
trea gigas, using colorimetric techniques, and on a
method of fertilization using small volumes of
seawater and known gamete concentrations. We
also present an index which may be useful in
evaluating the efficiency of fertilization. These
methods were developed during 1973 and should
prove useful in the study and production of cul-
tured oysters.

Materials and Methods

Pacific oysters were obtained from Fowler Oys-
ter Co. on Yaquina Bay, Newport, Oreg. Sand-
filtered seawater of 25-32"/(ki salinity and pH 7.0-8.1
was collected at the Oregon State University

Procurement of Gametes

To enhance gonad development, we conditioned
mature oysters in seawater at 16.0° ± 1.0° C for 3-6
wk (Loosanoff and Davis 1963). To identify test
oysters, we drilled a 0.8-mm (1/32-inch) hole in the
umbo and attached a 6.4- x 15.9-mm numbered
plastic tag (Howitt Plastics Co., Mollala, Oreg.)
with monofilament. After conditioning, access to
the gonads was made by drilling a 1.2-mm
(3/64-inch) hole in the posterodorsal region of the
right valve. We extracted gametes with a 2.5-cm^
glass syringe fitted with a 20-gauge 38-mm needle
containing about 0.5 ml of seawater (Lannan
1971). Oysters containing either intensively motile
sperm or eggs greater than or equal to 36 jum were
kept for fertilization experiments. To prevent
spawning after extractions, we isolated individual
oysters for 12-24 h in 3-liter beakers containing
seawater at 12°C.

Prior to gamete extraction we raised the tem-
perature of all donor oysters to 27.0° ± 0.5°C, a
temperature within the range recommended by
Davis and Calabrese (1964) for fertilization and
larval development. Oysters were transferred
from the conditioning tray to an 18.9-liter
(5-gallon) tank containing 11.4 liters (3 gallons) of
seawater at 16.0° ± 1.0°C; a 100-W aquarium heater

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

698

connected to a thermostat slowly increased the
temperature to 27.0° ± 0.5°C. After 30-60 min at
this temperature, we extracted gametes by syr-
inge (as above) for use in fertilization exper-
iments. To protect gametes from large pressure
changes during gonad extractions, we maintained
a gentle and constant negative pressure in the
syringe; gametes drawn abruptly into the syringe
were discarded. Only extractions which entered
the syringe as dense white cords were used; more
diffuse or cloudy extractions were also discarded.
Samples of gonad extractions and all seawater and
glassware used were stored at 27.0° ± 0.5°C.

Gamete Concentrations and Fertilization

Extractions from the gonads of 3-5 males were
transferred to and lightly agitated in a Klett-
Summerson sample tube containing 5-8 ml of
seawater. Using a Klett-Summerson colorimeter
with a green (#54) filter, we measured light
diffusion through diluted extractions. We then
diluted subsamples, placed them in a hemacy-
tometer, and counted the number of sperm. Com-
parisons were then made between the Klett read-
ing (K) and the actual sperm counts.

Gonad extractions from 3-5 females were
similarly pooled, transferred to a Nytex screen
(36-;um mesh), and rinsed with seawater to remove
small debris and reduce the possibility that some
component of the eggs (released from any ova
broken during extraction) would cause sperm to
agglutinate (Galtsoff 1964). We then rinsed the
cleaned eggs into a 250-ml beaker containing 20-50
ml of seawater, counted samples of eggs using a
dissection microscope, diluted samples with
seawater until reaching the desired concentration,
and maintained the egg-seawater suspension at
27.0° ± 0.5°C. We discarded eggs remaining in
seawater for more than 1 h to reduce the pos-
sibility of sperm agglutination resulting from
secretions.

Using a Pasteur pipette (45 drops of seawa-
ter/ml), we transferred various sperm concentra-
tions (Table 1) to numbered Syracuse watch
glasses, then with an automatic pipette added 0.2
ml of egg-seawater suspension containing 100 ± 4
eggs. The pH of 7.8 ± 0.1 was in the range recom-
mended by Humphrey (1950). After introduction
of the egg-seawater suspension, we added 7 ml of
seawater at different time intervals (flooding
time. Table 2) to dilute the sperm concentrations
and to reduce the possibility of polyspermy. The

Table 1. -Range and mean of percent fertilization (%Z) and
percent of larvae developing to the D-shape stage (%D) resulting
from different sperm concentrations combined with fresh eggs
(100 ± 4/0.2 ml) of Crassostrea gigas.^

Sperm

Lx^

Concn.

Vol (ml)

Range

Mean

Range

Mean

1.1 X 10''

0.02

0- 58

27.6

0-56

25.0

2.60

0.74

3.3 X 10"

0.07

10- 55

34.4

10-50

29.2

5.20

0.64

5.5 X 10"

0.11

41- 86

57.4

36-73

48.0

9.40

0.99

1.1 X 105

0.22

55-100

77.2

50-93

58.6

18.60

0.91

2.4 X 105

0.02

60-100

72.6

47-88

58.2

14.40

1.05

7.3 X 105

0.07

75-100

87.0

51-94

68.8

18.20

1.16

1.2 X 10(>

0.11

78-100

89.0

35-70

45.2

43.80

0.33

2.4 X 10'

0.22

80-100

89.2

13-72

29.0

60.20

0.12

5.0 X 106

0.50

80-100

92.4

8-43

21.9

70.50

0.06

1.1 X 10'

1.00

87-100

96.0

0-23

15.0

81.00

0.03

'Seawater of salinity 25%o and pH 7.8 it 0.1; temperature 27°
± 0.5"C; gametes diluted with 7 ml of seawater at 10 min post-
fertilization; 5 repetitions per sperm concentration.

Mean percent larvae losses = (%Z- minus %Dj).

^Cl = Compatibility index

/l -(%Z -)

X ^o^.

Table 2.-The range and mean of percent fertilization (%Z) and
percent of larvae developing to the D-shaped stage (%D)
obtained by different flooding times i after combining the
gametes of Crassost rea gigan.'^

Flooding
(min)

Range

Mean

Range

Mean

Lv^

41- 88

67.2

41-72

62,0

5.2

73-100

85.4

65-94

73.0

12.4

75-100

87.8

51-94

68.2

19.6

78-100

89.8

0-68

47.0

42.8

82-100

96.4

0-20

8.8

87.6

'Flooding time = time (min) between the combination of ga-
metes and the addition of 7 ml of seawater to the gamete mixture.

^Using seawater of salinity 25%oand pH 7.8 ± 0.1, 100 ± 4
eggs in 0.2 ml of seawater were added to 7.3 X 105 sperm in
0.07 ml of seawater; 5 repetitions/flooding time were used-

'L - = Mean percent larvae losses

(%Z - minus %0j^).

watch glasses were stacked to reduce evaporation
and incubated at 27.0° ± 0.5°C. Because the number
of swimming lavae did not increase after 6 h
postfertilization time, the number of fertilized
eggs was obtained by counting unfertilized eggs
remaining on the bottom at 6 h and subtracting
this figure from 100 (the number of eggs originally
present). After 24 h we transferred the watch
glasses to a 4°C refrigerator; within 30 min the
D-shaped (straight-hinged) larvae settled to the
bottom and were easily counted.

Although none of the 460 oysters examined
appeared hermaphroditic, sperm-free controls
were used in all experiments. We did not observe
fertilization in any of the controls.

Results and Discussion
The relationship between K and the number of

699

sperm counted is linear (r = 0.996) from about K
= 10 to about K = 80 (Figure 1). Because this
method of estimation is sufficiently precise and
accurate and because attempts to minimize go-
nadal debris (and thus minimize a variable in
colorimetric evaluation) by gravity filtration or
centrifugation usually resulted in broken tails and
agglutination, respectively, we consider our
methods of sperm procurement and estimation
useful. Measuring light diffusion through a sample
of C. gigas eggs did not accurately estimate egg
numbers because they settled rapidly.

We estimate that about one-half the number of
sperm counted had little or no observable motility;
we may have withdrawn immature sperm or
damaged mature sperm during extraction. Inac-
tive sperm were not agglutinated, an indication
that the acrosome reaction was not the major
cause of immotility. Although not directly equat-
ing fertilization capacity with high motility, our
assumption is that relatively immotile sperm are
incapable of fertilizing viable eggs. Similarly,
extractions from females often included small and
presumably immature eggs (Galtsoff 1964). Sperm
concentrations reported in Figure 1 and Tables 1
and 2 are observed values and do not reflect
estimates of immotile cells; only "mature-sized"
eggs were used because eggs less than 36 /xm were
rinsed through the cleaning screen.

Within the limits of this investigation, mean
percent fertilization (%Zj:) increased as the.
number of sperm/ 100 eggs increased (Table 1).
The mean percent of larvae developing to the
D-shape stage {%I)j) increased until 7.3 x 10^
sperm were used; %Dj: decreased with further
increases of sperm concentration (Table 1).
Because Glatsoff (1964) reported that high sperm
concentrations may result in polyspermy and
because in our experiments resulting in large

I- 40

» = 0.7 » ia.3(«)
r = 0.996

0 15 2.0 2.5 3 0

NUMBER OF SPERM /ml « lO'

Figure 1. -Correlation between mean (A^ = 5) number of
Crassostrea gigas sperm and Klett units (light diffusion read-
ings) on a Klett-Summe.rson colorimeter.

losses of larvae (Lf [where Lj = ?cZ^ minus %Dj])
aberrant forms were observed (e.g., swimming
chains of cells, and trochophores persisting beyond
48 h), we assume polyspermy was responsible for
the increasing L^.

Using 7.3 X 10-^ sperm/100 eggs, we observed
that %Zj increased as flodding time increased
(Table 2). Lf also increased as flooding time in-
creased, and maximum %Dj was obtained using a
flooding time of 5 min.

Although most workers need only to maximize
%'Djr without regard to L?, some investigators may
need to minimize L? due to limited spawning stock
or other problems. Thus, to achieve maximum
efficiency it is necessary to maximize %!)? and
minimize L/. Under different conditions (e.g.,
water quality and gamete viability may differ at
different locations or at different times), the
optimal sperm concentration and flooding time
will vary in response to the environment. In-
creases in ^Djr (by increasing sperm concentra-
tion or flooding time) also produce undesirable
increases in L^, thus a subjective decision usually
is made to evaluate the efficiency of fertilization
and larvae production. To reduce the subjectivity
of this evaluation, we suggest the following for-
mula reflects a compatability between maximum
%Dj and minimum L^:

(%T) )'
Compatability index (CI) ===== x 10-^.

/L; (%Zj)
In our lab, values greater than or equal to 1 were
desirable, and 1.16 was the maximim value ob-
tained (Table 1). CI values can be high for rela-
tively low %Dj: if L? is unusually low (e.g., where
%Zj = 30, and %Dj = 28, CI = 1.01). Low L, values
will normally be associated with low XD? ; however,
if a low Lj: occurs concommittantly with a "rea-
sonable" %l)j, we assume that the evaluation
would be based more on the desired %Dj rather
than on CI. Further, due to the often dramatic
differences in conditions at different labs and
hatcheries, or at different times, attempts to
establish a desirable CI value or range under
specified conditions may prove useful.

During a 4- to 6-wk period we made 8-12 ex-
tractions from individual oysters, but did not
observe a deterioration of gametes. Data from
experiments using gametes from initial extrac-
tions were consistent with those of later extrac-
tions. The pooling of extractions may have reduced
observable changes. After about 8 wk, eggs were
easily broken and we noticed free yolk in extrac-

700

tions. Although deterioration of male gonads was
less evident, we noted that the sperm concentra-
tion decreased after about 8 wk, presumably as a
result of resorption. The mortality rate for oysters
repeatedly used for gamete extractions and
maintained without food or biological filters in
113.6-liter (30-gallon) tanks containing recirculat-
ing seawater (25"/o()) at 16.0° ± 1.0°C was about
10% during the 8-wk period.

Because high concentrations and large numbers
of gametes can repeatedly be extracted from the
gonads of individual oysters without apparent
detriment and because gamete extraction ob-
viates artificial spawning and its inherent prob-
lems, we suggest our method of gamete pro-
curement can be useful in many investigations
and hatchery situations. Our method also permits
repeated use of the gametes of selected oysters,
and this together with the possible use of
cryopreserved sperm (Staeger 1974) reduces var-
iability and increases control and management of
hatchery production or biological investigations.

Acknowledgments

Thanks are due James Lannan, Raymond Mil-
lemann, and James Rybock for their assistance.
This work is a result of research sponsored by the
Oregon State University Sea Grant College Pro-

gram, supported by NOAA Oflfice of Sea Grant,
U.S. Department of Commerce, under Grant
number 04-3-158-4

Literature Cited

Davis, H. C, and A. Calabrese.

1964. Combined effects of temperature and salinity on
development of eggs and growth of larvae of M. mercen-
aria and C. rirginica. U.S. Fish Wildl. Serv., Fish. Bull.
63:643-655.
Galtsoff, p. S.

1964. The American oyster Cratisostrea virgin ka Gmelin.
U.S. Fish Wildl. Serv.. Fish. Bull. 64:1-480.
Humphrey, G. F.

1950. The metabolism of oyster spermatozoa. Aust. J. Exp.
Biol. Med. Sci. 28:1-13.
Lannan, J. E.

1971. Experimental self-fertilization of the Pacific oyster,
Crasf!Of:trea gifjaf:, utilizing cryopreserved sperm. Gene-
tics 68:599-601.

LOOSANOFF, V. L., AND H. C. DaVIS.

1963. Rearing of bivalve mollusks. Adv. Mar. Biol. 1:1-136.
Staeger, W. H.

1974. Cryobiological investigations of the gametes of the

Pacific oyster, Crattsofttrea gigas. M.S. Thesis, Oregon

State Univ., Corvallis, 45 p.

William H. Staeger
Howard F. Horton

Department of Fisheries and Wildlife
Oregon State University
Corvallis, OR 973S1

701

3J.,,i^i':^6

INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN

Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the
following instructions. These are not absolute requirements, of course, but desiderata.

CONTENT OF MANUSCRIPT

The title page should give only the title of the
paper, the author's name, his affiliation, and
mailing address, including Zip code.

The abstract should not exceed one double-
spaced page.

Text footnotes should be typed separately
from the text.

Figures and tables, with their legends and
headings, should be self-explanatory, not requir-
ing reference to the text. Their placement should
be indicated in the right-hand margin of the
manuscript.

Do not send original drawings to the Scientific
Editor; if they, rather than the photographic re-
ductions, are needed by the printer, the Scientific
Publications Staff will request them.

Acknowledgments, if included, are placed at
the end of the text.

Literature is cited in the text as: Lynn and Reid
(1968) or (Lynn and Reid 1968). All papers re-
ferred to in the text should be listed alphabetically
by the senior author's surname under the heading

Common abbreviations and symbols, such as
mm, m, g, ml, mg, °C (for Celsius), %, "L and so
forth, should be used. Abbreviate units of mea-
sure only when used with numerals. Periods are
only rarely used with abbreviations.

We prefer that measurements be given in
metric units; other equivalent units may be given
in parentheses.

FORM OF THE MANUSCRIPT

TITLE PAGE

ABSTRACT

TEXT

LITERATURE CITED

APPENDIX

TEXT FOOTNOTES

TABLES (Each table should be numbered with
an arabic numeral and heading provided)

LIST OF FIGURES (entire figure legends)

FIGURES (Each figure should be numbered
with an arabic numeral; legends are desired)

ADDITIONAL INFORMATION

Send the ribbon copy and two duplicated or
carbon copies of the manuscript to:

Dr. Bruce B. Collette, Scientific Editor

Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Fifty separates will be supplied to an author
free of charge and 100 supplied to his organiza-
tion. No covers will be supplied.

Contents— continued

Notes

ZIMMERMAN, STEVEN T., and ROBERT S. McMAHON. Paralytic shellfish poi-
soning in Tenakee, southeastern Alaska: A possible cause 679

COLLINS, JEFF. Oil and grease: A proposed analytical method for fishery waste
effluents 681

GADBOIS, D. F., E. M. RAVESI, and R. C. LUNDSTROM. Occurrence of volatile
N-nitrosamines in Japanese salmon roe 683

WATKINS, WILLIAM A., and WILLIAM E. SCHEVILL. Underwater paint mark-
ing of porpoises 687

EDGAR, ROBERT K., and JAMES G. HOFF. Grazing of freshwater and estuarine,
benthic diatoms by adult Atlantic menhaden, Brevoortia tyrannus 689

JOHNSON, ALLYN G. Electrophoretic evidence of hybrid snow crab, Chionoecetes
bairdi X opilio 693

KORN, SID, JEANNETTE W. STRUHSAKER, and PETE BENVILLE, JR. Effects
of benzene on growth, fat content, and caloric content of striped bass, Morone
saxatilis 694

STAEGER, WILLIAM H., and HOWARD F. HORTON. Fertilization method
quantifying gamete concentrations and maximizing larvae production in Cras-
sostrea gigas 698

AMERICAS
7> ^c, FIRSTINDUSTRY

■j!!r GPO 696-333

,^^^'°'Co,

Fishery Bulletin

National Oceanic and Atmospheric Administration • National Marine Fisheries Service

^^ATES O^ ^

• » I

i ' -" i

I i

Vol.74, No. 4 I V/oods Ho! J, iviass. j October 1976

EBERLING, ALFRED W., and RICHARD N. BRAY. Day versus night activity of reef

fishes in a kelp forest off. Santa Barbara, California 703

KROUSE, JAY S. Incidence of cull lobsters, Homarus americanus, in commercial and

research catches off the Maine coast 719

COLLINS, JEFF, and RICHARD D. TENNEY. Fishery waste effluents: A method to

determine relationships between chemical oxygen demand and residue 725

BRINTON, EDWARD. Population biology of Euphausia pacifica off southern

California 733

BISSON, PETER A., and GERALD E. DAVIS. Production of juvenile chinook salmon,

Oncorhynchus tshawytscha, in a heated model stream 763

MANOOCH, CHARLES S., III. Reproductive cycle, fecundity, and sex ratios of the red

porgy, Pagrus pagrus. (Pisces: Sparidae) in North Carolina 775

HALL, ALICE S., FUAD M. TEENY, LAURA G. LEWIS, WILLIAM H. HARDMAN,

and ERICH J. GAUGLITZ, JR. Mercury in fish and shellfish of the northeast

Pacific. I. Pacific halibut, Hippoglossus stenolepis 783"^'

HALL, ALICE S., FUAD M. TEENY, and ERICH J. GAUGLITZ, JR. Mercury in fish

and shellfish of the northeast Pacific. II. Sablefish, Anoplopoma fimbria 791

WALTERS, JOHN F. Ecology of Hawaiian sergestid shrimps (Penaeidea: Serges-

tidae) 799

LORD, GARY E. Decision theory applied to the simulated data acquisition and

management of a salmon fishery 837

HUNTER, JOHN R., and CAROL SANCHEZ. Diel changes in swim bladder inflation

of the larvae of the northern anchovy, Engraulis mordax 847

WIDERSTEN, BERNT. Ceriantharia, Zoanthidea, Corallimorpharia, and Actiniaria

from the continental shelf and slope off the eastern coast of the United States. . . . 857
ALVAREZ, JOSE, CHRIS 0. ANDREW, and FRED J. PROCHASKA. Dual structural

equilibrium in the Florida shrimp processing industry 879

BLACKBURN, MAURICE, and WALTER NELLEN. Distribution and ecology of

pelagic fishes studied from eggs and larvae in an upwelling area off Spanish Sahara 885
CRONE, RICHARD A., and CARL E. BOND. Life history of coho salmon, Oncorhyn-
chus kisutch, in Sashin Creek, southeastern Alaska 897

YOUNGBLUTH, MARSH J. Vertical distribution and diel migration of euphausiids in

the central region of the California Current 925

(Continued on back cover)

Seattle, Washington

U.S. DEPARTMENT OF COMMERCE

Elliot L. Richardson, Secretary

NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION

Robert M. White, Administrator

NATIONAL MARINE FISHERIES SERVICE

Robert W. Schoning, Director

Fishery Bulletin

The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and
economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in
1904 and the Fishery Bulletin of the Fish and Wildbfe Service in 1941. Separates were issued as documents through volume 46; the last
document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a
numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin
instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical,
issued quarterly. In this form, it is available by subscription from the Superintendent of Docvunents, U.S. Government Printing Office,
Washington, D.C. 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and
in exchange for other scientific publications.

EDITOR

Dr. Bruce B. Collette

Scientific Editor, Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Editorial Committee

Dr. Elbert H. Ahlstrom

National Marine Fisheries Service

Dr. WilUam H. Bayliff

Inter-American Tropical Tuna Commission

Dr. Roger F. Cressey, Jr.
U.S. National Museum

Mr. John E. Fitch

California Department of Fish and Game

Dr. William W. Fox, Jr.

National Marine Fisheries Service

Dr. Marvin D. Grosslein
National Marine Fisheries Service

Dr. Edward D. Houde
University of Miami

Dr. Merton C. Ingham

National Marine Fisheries Service

Dr. Reuben Lasker

National Marine Fisheries Service

Dr. Sally L. Richardson
Oregon State University

Dr. Paul J. Struhsaker

National Marine Fisheries Service

Dr. Austin Williams

National Marine Fisheries Service

Kiyoshi G. Fukano, Managing Editor

The Fishery Bulletin is published quarterly by Scientific Publications Staff, National Marine Fisheries Service, NOAA,
Room 450, 1107 NE 45th Street, Seattle, WA 98105. Controlled circulation postage paid at Seattle, Wash.

Fishery Bulletin

CONTENTS

Vol. 74, No. 4 October 1 976

EBERLING, ALFRED W., and RICHARD N. BRAY. Day versus night activity of reef

fishes in a kelp forest off Santa Barbara, California 703

KROUSE, JAY S. Incidence of cull lobsters, Homanis americanns, in commercial and

research catches off the Maine coast 719

COLLINS, JEFF, and RICHARD D. TENNEY. Fishery waste eflluents: A method to

determine relationships between chemical oxygen demand and residue 725

BRINTON, EDWARD. Population biology of Euphausia pacifica off southern

California 733

BISSON, PETER A., and GERALD E. DAVIS. Production of juvenile chinook salmon,

Oncorhynchus tshawytscha, in a heated model stream 763

MANOOCH, CHARLES S., III. Reproductive cycle, fecundity, and sex ratios of the red

porgy, Pagrus pagrus (Pisces: Sparidae) in North Carolina 775

HALL, ALICE S., FUAD M. TEENY, LAURA G. LEWIS, WILLIAM H. HARDMAN,

and ERICH J. GAUGLITZ, JR. Mercury in fish and shellfish of the northeast

Pacific. I. Pacific halibut, Hippoglossus stenolepis 783

HALL, ALICE S., FUAD M. TEENY, and ERICH J. GAUGLITZ, JR. Mercury in fish

and shellfish of the northeast Pacific. II. Sablefish, Anoplopoma fimbria 791

WALTERS, JOHN F. Ecology of Hawaiian sergestid shrimps (Penaeidea: Serges-

tidae) 799

LORD, GARY E. Decision theory applied to the simulated data acquisition and

management of a salmon fishery 837

HUNTER, JOHN R., and CAROL SANCHEZ. Diel changes in swim bladder inflation

of the larvae of the northern anchovy, Engraulis mordax 847

WIDERSTEN, BERNT. Ceriantharia, Zoanthidea, Corallimorpharia, and Actiniaria

from the continental shelf and slope off the eastern coast of the United States 857

ALVAREZ, JOSE, CHRIS 0. ANDREW, and FRED J. P«(OCHASKA. Dual structural

equilibrium in the Florida shrimp processing industry 879

BLACKBURN, MAURICE, and WALTER NELLEN. Distribution and ecology of

YOUNGBLUTH, MARSH J. Vertical distribution and diel migration of euphausiids in

the central region of the California Current 925

(Continued on next page)

Seattle, Washington
1977

For sale by the Supenntendent of Documents, U.S. Government Printmg Office, Washing-
ton, D.C. 20402 — Subscription price: $1 1.80 per year ($2.95 additional for foreign mail-
ing). Cost per single issue ■ $2.95.

Contents— confiynted

SCHERBA, STEPHEN, JR., and VINCENT F. GALLUCCI. The application of
systematic sampling to a study of infauna variation in a soft substrate environ-
ment 937

KROUSE, JAY S. Size composition and growth of young rock crab, Cancer irroratus,
on a rocky beach in Maine 949

DOTSON, RONALD C. Minimum swimming speed of albacore, Thunnus alalunga . . 955

BAILEY, JACK E., JEROME J. PELLA, and SIDNEY G. TAYLOR. Production of fry
and adults of the 1972 brood of pink salmon, Oncorhynchus gorbuscha, from gravel
incubators and natural spawaning at Auke Creek, Alaska 961

KEENE, DONALD F., and WILLIAM G. PEARCY. Comparison of the most
successful and least successful west coast albacore troll fishermen 973

Notes

GOLDBERG, STEPHEN R. Seasonal spawning cycles of the sciaenid fishes Genyone-
mus lineatus and Seriphus politus 983

KRAVITZ, MICHAEL J., WILLIAM G. PEARCY, and M. P. GUIN. Food of five
species of cooccurring flatfishes on Oregon's continental shelf 984

RALSTON, STEPHEN. Age determination of a tropical reef butterflyfish utilizing
daily growth rings of otoliths 990

ROTHLISBERG, PETER C, and WILLIAM G. PEARCY. An epibenthic sampler used
to study the ontogeny of vertical migration of Pandalus jordani (Decapoda,
Caridea) 994

CARR, WILLIAM E. S., and THOMAS B. CHANEY. Harness for attachment of an

ultrasonic transmitter to the red drum, Sciaenops ocellata 998

INDEX, VOLUME 74 1001

Vol. 74, No. 3 was published on 16 September 1976.

The National Marine Fisheries Service (NMFS) does not approve, recommend or
endorse any proprietary product or proprietary material mentioned in this
publication. No reference shall be made to NMFS, or to this publication furnished by
NMFS, in any advertising or sales promption which would indicate or imply that
NMFS approves, recommends or endorses any proprietary product or proprietary
material mentioned herein, or which has as its purpose an intent to cause directly or
indirectly the advertised product to be used or purchased because of this NMFS
publication.

DAY VERSUS NIGHT ACTIVITY OF REEF FISHES IN
A KELP FOREST OFF SANTA BARBARA, CALIFORNIA

Alfred W. Ebeling and Richard N. Bray'

ABSTRACT

Vertical distributions and feeding activities of residential kelp-bed fishes were compared between day
and night in an area of reef and kelp off Santa Barbara, Calif. Abundances and positions of fishes within
four vertically oriented zones were obsen-ed during 42 paired day and night scuba transects made
throughout the year along a line secured to a high-relief rocky reef located about 1.6 km offshore.
Feeding activity was determined for surfperches (Embiotocidae) from the proportion of fish collected
during the day or night having empty "foreguts" and inferred for other fishes from general
observations of individuals. Although almost all of the 25 common fish species recorded were seen both
day and night, the number seen and the degree of activity of most of these species decreased
considerably at night. Many fishes that fed and moved about in mid-water well above the reef during
the day were found in holes and crevices in the reef at night; others that foraged on or just above the
bottom during the day showed little change in their position; and still others tended to disperse to
adjacent areas of the reef. Daytime aggregations of fishes centered around the crest of the reef and
other productive prominences and invariably dispersed at night. Unlike tropical communities of reef
fishes, the kelp-bed community included neither a broad replacement for diurnal planktivores in the
night shift nor a contingent that moves out over nearby sand flats to forage at night. Kelp-bed fishes
showed considerable intraspecific variability in behavior. Thus, the kelp-bed community appears to be
more loosely "programmed" even though it follows the same basic pattern of diel activity as the
tropical-reef community. The kelp-bed species that belong to primarily tropical families tended to be
quite specialized in their nocturnal sheltering behavior. Yet the primarily temperate surfperches, for
example, simply became somewhat lethargic and remained exposed at night.

Day-night variations in the activities of reef
fishes have received considerable attention
recently, especially as these variations may relate
to foraging methods (Hobson 1975), and to sharing
of limited space (Smith and Tyler 1972). Direct
observations of coral reef fishes have shown that,
although most species are active mainly during the
day, a substantial number are active only during
the night (Hobson 1965, 1968, 1974; Starck and
Schroeder 1965; Smith and Tyler 1972). In both
instances, fish forage mainly during their active
periods and school and/or seek shelter during their
inactive periods (Hobson 1972). Dawn and dusk are
important transitional periods when fishes that
had been active seek shelter, when fishes that had
been resting begin foraging, and when piscivores
become most effective (Hobson 1972; Collette and
Talbot 1972; Domm and Domm 1973).

The assemblage of fishes at a particular place on
a tropical reef at night may differ markedly from
the assemblage gathering at the same place dur-
ing the day because foraging and sheltering

'Marine Science Institute and Department of Biological
Sciences, University of California, Santa Barbara, CA 93106.

activities do not always occur in the same area. For
example, some surgeonfishes (Acanthuridae) and
damselfishes (Pomacentridae), which shelter at
night in the shallower portions of coral reefs,
migrate offshore at dawn to their feeding grounds
in deeper water (Hobson 1972). Nocturnal preda-
tors may undergo even more extensive migra-
tions. Some snappers (Lutjanidae) and grunts
(Pomadasyidae) are among a considerable number
of species that move away from the reef at night
to forage over surrounding sand flats (Hobson
1968, 1972). For many planktivores, however, the
change in activity simply involves vertical
movements from foraging areas in the water
column to underlying sheltering sites (Hobson
1973). Thus the important events during the
transition period between day and night include
vertical as well as horizontal movements of fish.

Less is known about the nocturnal activities of
temperate kelp-bed fishes. Some information has
been available on a few species: the garibaldi,
Hypsypops rubicundus (By Clarke 1970); the
California sheephead, Pimelometopon pulchrum
(by Wiley 1974); the kelp perch, Brachyistius
frenatus; white seaperch, Phanerodon furcatus,

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

703

FISHERY BULLETIN: VOL. 74, NO. 4

and senorita, Oxyjulis californica (by Bray and
Ebeling 1975); and the horn shark, Heterodontus
francisci, and swell shark, Cephaloscy Ilium ven-
triosum (by Nelson and Johnson 1970). More
recently, Hobson and Chess (1976) presented more
comprehensive comparisons of the day and night
feeding activities of fish off Santa Catalina Island:
in particular, of the blacksmith, Chromis punc-
tipinnis; the walleye surfperch, Hyperprosopon
argenteum; the kelp rockfish, Sebastes atroviretis;
the olive rockfish, S. serranoides; the queenfish,
Seriphus politus; and the salema, Xenistius
californiensis, as well as some of the others
mentioned above. Therefore, we initiated a com-
parative day and night survey of the fishes in-
habiting an area of reef and kelp off Santa Bar-
bara, Calif., to see if the fish community under-
goes a substantial diel change in its composition,
vertical distribution, and activity.

METHODS

Naples Reef is a large rocky outcrop located 24
km west of Santa Barbara (lat. 34°25'N, long.
119°57'W). The reef measures 275 m by 80 m (2.2
hectares) and lies 1.6 km offshore. The substratum
consists of a series of sandstone rills and ridges
that run parallel to the coast. Depths across the
reef average 8 to 10 m, although some prominences
project to within 5 m of the surface. The bottom
surrounding the reef is 16 to 20 m deep and is
comprised of sand with rocky outcrops inshore, or
sand and cobbles offshore. The assemblage of plant
and animal life on and about the reef is among the
richest along the Santa Barbara coast. Giant kelp
(Macrocystis) is always present on the reef, al-
though kelp densities fluctuated throughout the
study period. Temperatures along the top of the
reef ranged from 11°C in the spring to 19°C in the
fall. Underwater visibility averaged 5 m at the
transect line.

A transect line consisting of two 40-m segments
was staked along either side, shoreward and
seaward, of a high-relief ridge with a crest at 6 m.
Day and night counts of fishes along the line were
made by scuba divers. For each day-night pair of
samples, we counted fish within 2 m on either side
of the line. To minimize the effect of nondiel
fluctuations on our observations, we always made
the night transect member of a pair within 12 h of
the day transect. A special effort was made to
insure that the night counts of fish were made
throughout approximately the same reef area and

overlying volume of water as were the day counts.
Powerful 10-cell underwater hand lights, fitted
with reflectors to illuminate the data sheets, were
used intermittently during the day to inspect
holes, and used continuously throughout the night
dives.

We evaluated the diel activities of fish species by
observing the fishes' vertical distribution and
feeding habits. During the transects, fish sight-
ings were tallied in separate columns on our plastic
data sheets according to the zone in which each
individual was observed (Table 1).

The use of dive lights at night may have at-
tracted or repelled fish depending on the species
and/or altered their state of activity. Yet fishes
normally inactive at night did not seem to be
affected by brief exposures to dive lights. Species
normally active at night responded in various
ways, from showing hyperactivity to apparent
immobilization. Other nighttime observations of
reef fishes off California (Nelson and Johnson
1970) and in tropical waters (Hobson 1965; Starck
and Davis 1966; Smith and Tyler 1972) also in-
dicate that night-active fishes often respond un-
predictably to artificial illumination.

Day and night differences in the feeding habits
of many species were inferred either from direct
observations of foragers or from changes in the
fishes' vertical distribution and activity level (i.e.,
whether the fish were exposed and responsive to
our presence or sheltered and unresponsive). We
feel that such observations of fish activity by
themselves were sufficient to distinguish feeding
from nonfeeding periods for many of the more
prominent species. However, such observations
proved to be inadequate indicators of foraging
activity for surfperches (Embiotocidae), which
comprise the most abundant and diverse foraging
guild of the fishes on Naples Reef. To test for diel
differences in feeding activity of surfperches,
therefore, we speared during all hours of day and
night approximately 400 adults of the five common
demersal species: the black perch, Emhiotoca
jacksoni (median standard length 195 mm, range

Table 1.— Zones of vertical orientation in which fish were
observed along a transect line

Zone

Extent of zone

IV Mid-water

III Suprabenthic
II Bottom

I Shelter

Greater than 1.0 m above the bottom, in open

water and/or near kelp stipes
Within 1.0 m of the bottom
In physical contact with the bottom yet

exposed
In holes, crevices, or under ledges

704

EBELING and BRAY: ACTIVITY OF REEF FISHES

86-244 mm); striped seaperch, E. lateralis (200,
110-280); rubberlip seaperch, Rhacochilus toxotes
(279, 165-400); pile perch, Damalichfhys vacca (210,
97-260); and rainbow seaperch, Hypsurus caryi
(159, 114-253). Immediately after each dive, in-
dividuals were either iced and later frozen, or slit
ventrally and fixed in 10% Formalin. ^ The
procedure for gut analysis followed the method of
Bray and Ebeling (1975), except that the
surfperch's gut, which is simple and tubular and
lacks a well defined "stomach," was divided into
quarters. Fullness of the "foregut," defined as the
first quarter of the length of the entire gut, was
scored subjectively from 1 (empty) to 5 (full), and
plotted against time of collection. Since fish were
sampled throughout the year, their times of col-
lection were seasonally adjusted relative to actual
times of sunrise and sunset as determined from
solar tables.

RESULTS

We identified 25 species of fishes from 21 paired
day-night transects made between April 1972 and
September 1973. Most of the fishes seen along the
transect line were adults. The only abundant
juveniles were of the blue rockfish, Sebastes mys-
timis. Hence for blue rockfish only, juveniles and
adults were counted separately. We excluded from
the analysis all species that could not be consis-
tently observed, such as some of the more cryptic
and secretive fishes that blend with their sur-
roundings and hide in kelp and rocks, and species
that occur only near the water surface outside our
field of vision.

It appeared that our visual counts adequately
sampled all of the more conspicuous kelp-bed
fishes. The rank order of abundance of fishes
recorded in the 21 daytime transects was highly
correlated with that of fishes observed in a photo-
graphic survey consisting of 125, 2.5-min motion
pictures (Ebeling, Larson, and Alevizon unpubl.
data) filmed over the same area (Kendall's tau
coefl^cient of rank correlation = 0.65; P <0.001).

The species composition of seasonally pooled
samples and the relative abundances of the dif-
ferent species varied surprisingly little during the
17-mo study period. Almost all species were seen
throughout the year, and rank orders of species
abundances, pooled over day and night samples,

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

were significantly concordant among seasonal
periods that correspond roughly to annual oceano-
graphic periods defined by Brown (1974) (Table 2).

During the day, almost 4,000 fishes representing
11 families of teleosts and 1 family of sharks were
counted along the transect line. The two dominant
groups-surfperches (Embiotocidae) and rock-
fishes (Sebastes) -were represented by six species
each. The most abundant species was 5. mystinus
whose juveniles accounted for 44% of the in-
dividuals sighted during the day (Table 3).

Most individuals of all species of fishes (66%)
were observed in the mid-water zone higher than 1
m off the bottom (Table 4). The two most abundant
species in the mid-water zone, S. mystinus and
Chromis punctipinnis often formed large, mixed
aggregations above rocky prominences and
around columns of giant kelp. Besides S. mystinus
and C. punctipinnis, more than 80% of the in-
dividuals in several other species were observed in
the mid-water zone: the kelp bass, Paralabrax
dathratus; Oxyjidis californica; opaleye, Girella
nigricans; and S. serranoides (Table 4). But 10 of
the total of 19 species recorded from the mid-water
zone were more abundant in other zones.

Some 25% of the total individuals of all species
were observed in the suprabenthic zone, within 1
m of the rocky bottom (Table 4). This zone included
the most species (21) and was dominated by
surfperches: 71% of the individuals observed in the
suprabenthic zone were surfperches, as compared
with but 12% in the mid-water zone. Nearly half
the individuals were Embiotoca jacksoni or E.
lateralis.

Less than 10% of the total individuals recorded
during the day were observed either in the bottom
zone, contacting the reef in an exposed position, or
in the shelter zone, occupying a crevice or hole
(Table 4). Most of these were demersal,
"ambusher-type" predators, e.g., rockfishes and
sculpins (Cottidae), although a few of the mid-
water species, e.g., S. mystinus and C. punctipin-
nis, were also observed in these zones in small
numbers.

We recorded substantially fewer individuals at
night than during the day (Table 3). Day to night
decreases in total numbers were consistently
significant among the 21 pairs of day-night sam-
ples (Wilcoxon signed-ranks test for paired
observations, P<0.005). Also, lists of species,
ranked by abundance, differed at night. All 21 rank
correlations for the day-night sample pairs (tau =
-0.32 to -1-0.22), as well as the single rank correla-

705

FISHERY BULLETIN: VOL. 74, NO. 4

Table 2.-Seasonal variation in relative abundance of fishes observed along a transect line.
Samples are pooled over day and night transects by trimonthly intervals generally
coinciding with periods of oceanographic change off Santa Barbara. Kendall's H^ coefficient
of rank concordance among seasons = 0.77 (P«0.005).

Species

Sebastes mystinus (Juvenile)
Chromis punctipinnis
Embiotoca jacksoni
Oxylebius pictus
Embiotoca lateralis
S. carnatus

Hyperprosopon argenteum
S. mystinus (adult)
Pimelometopon pulchrum
Damalichthys vacca
Rhacochilus toxotes
Hypsypops rubicundus
Coryphopterus nicholsii
Paralabrax clathratus
S. chrysomelas
Medialuna calilorniensis
Scorpaenichthys marmoratus
Hypsurus caryi
Oxyjulis californica
Phanerodon furcatus
Ophiodon elongatus
Girella nigricans
Sebastes serriceps
S. serranoides
Cephaloscyllium ventriosum
S. atrovirens

Total no. of individuals
Total no. of transects

Percent of total individuals by season

Dec. -Feb.

Mar. -May

June-Aug.

Sept. -Nov.

29.5

29.4

45.3

25.2

21,3

11.1

14.5

14.9

14.5

10.9

8.75

14.5

5.61

3.97

2.05

3.71

4.97

2.34

1.49

1.86

3.71

4.69

2.33

3.12

3.29

2.71

1.40

8.27

2.09

6.40

1.77

9.20

1.95

0.36

0.84

0.42

1.86

3.07

2.98

2.36

1.44

1.89

0.47

1.27

1.30

1.17

1.86

1.94

1.25

0.18

0.84

1.21

0.36

1.30

1.86

1.21

0.81

0.65

1.10

0.97

0.09

0.28

0.17

0.84

0.27

0.09

0.08

0.56

13.6

3.17

1.86

0.56

2.34

1.12

2.62

0.51

2.89

4.28

0.17

0.46

0.18

0.09

0.37

0.27

2.70

3.04

0.19

0.36

0.37

0.14

0.09

1.77

0.76

0.19

0.09

0.34

0.05

0.45

0.37

0.34

2,150

1,109

1,074

1,185

Table 3.— Day-night variation in abundance of fishes observed along a transect line. Samples
are pooled over seasonal intervals (see Table 2). Symbols in the "Difference" column indicate for
each species whether the numbers of individuals observed during the day, ordered among all
transects, were significantly greater than (>), less than (<), not significantly different from ( = ),
or too few to compare with (NC) the numbers observed at night (Wilcoxon signed-rank test for
paired observations, P^O.05).

No. of in(

jividuals

Percent of total
individuals

Difference

day vs.

night

Species

Day

Night

Day

Night

Sebastes mystinus (juvenile)

1,730

43.80

0.51

Embiotoca jacksoni

492

207

12.40

13.20

Chromis punctipinnis

253

662

6.40

42.30

Oxylebius pictus

215

5.44

1.02

S. mystinus (adult)

210

5.31

2.17

Hypsurus caryi

188

4.75

1.98

Embiotoca lateralis

134

3.39

2.37

Damalichthys vacca

112

2.83

1.41

Phanerodon lurcatus

2.30

Oxyjulis californica

2.05

Girella nigricans

1.77

0.38

Paralabrax clathratus

1.64

0.13

S. carnatus

138

1.42

8.82

Hypsypops rubicundus

1.39

1.85

Pimelometopon pulchrum

1.04

1.21

Coryphopterus nicholsii

0.96

0.06

Rhacochilus toxotes

0.78

2.62

S. serranoides

0.53

0.70

Medialuna calilorniensis

0.48

0.51

S. chrysomelas

0.43

2.43

Scorpaenichthys marmoratus

0.30

0.70

Ophiodon elongatus

0.25

0.19

Sebastes atrovirens

0.18

0.45

S. serriceps

0.10

0.51

Cephaloscyllium ventriosum

0.05

0.70

Hyperprosopon argenteum

214

13.70

Total no. of individuals

3,954

1,564

100.0

Total no. of transects

706

EBELING and BRAY: ACTIVITY OF REEF FISHES

Table 4.— Vertical-zone variation in numbers of fishes observed along a transect line compared between day and night. Vertical zones
are defined in Table 1; the M measure of a species' change in vertical position between day and night is defined in the text.

Day

Night

Supra-

Family and species

Mid-water

benthic

Bottom

Shelter

Mid-water

benthic

Bottom

Shelter

A/7

Scyliorhinidae;

Cephaloscyllium ventriosum

0.27

Serranidae:

Paralabrax clathratus

1.33

Kyphosid-like fishes:

Girella nigricans

1.60

Medialuna californiensis

1.07

Embiotocidae:

Damalichthys vacca

-0.13

Embiotoca jacksoni

101

386

0.18

E. lateralis

108

0.30

Hyperprosopon argenteum

213

—

Hypsurus caiyi

0.93

Phanerodon furcatus

—

Rhacochilus toxotes

-0.12

Pomacentridae:

Chromis punctipinnis

210

610

2.62

Hypsypops rubicundus

1.66

Labridae:

Oxyjulis calilornica

'2.91

Pimelometopon pulchrum

2.61

Gobiidae:

Coryphopterus nicholsii

-0.16

Scorpaenidae:

Sebastes atrovirens

0.14

S. carnatus

0.19

S. chrysomelas

0.09

S. mystinus (adult)

178

2.08

S. mystinus (juvenile)

1,606

119

2.55

S. serranoides

0.40

S. serriceps

Hexagrammidae:

Ophiodon elongatus

1.80

Oxylebius pictus

179

0.74

Coftidae:

Scorpaenichthys marmoratus

0.09

Total no. of individuals

2,586

997

295

343

166

269

786

Percent of day or night

total

65.4

25.2

7.46

1.92

21.9

10.6

17.2

50.3

Total no. of transects

'Individuals are assumed to bury themselves at night.

tion for the day-night contrast with samples
pooled (tau = 0.13), were nonsignificant (P>0.05).
A Wilcoxon signed-ranks test for paired (day-
night) observations indicated that numbers of
eight species did not differ significantly between
day and night, while numbers of four species
actually increased (Table 3).

Two species commonly observed during the day
were either seldom or not seen at night: Phaner-
odon furcatus and Oxyjulis californica. Although
we often saw individuals of P. furcatus browsing
on bryozoan-encrusted algae (mainly Gelidium
sp.) along a crest of the reef during the day, we
rarely observed them at night along the crest and
never observed them during regular transects. We
commonly saw small groups of 0. californica in the
mid-water zone above the transect lines during the
day. At dusk, however, Oxyjulis individuals bury
themselves in rubble and sand on the reef and

remain covered until dawn (Herald 1961; Feder et
al. 1974; Bray and Ebeling 1975).

Only one species was seen at night but never
during the day. Hyperprosopon argenteum was the
second-most abundant species recorded at night,
although it was never seen around the transect
line during daylight hours. In over 6 yr of obser-
vations, we have seen this species in kelp beds on
only a few occasions during the day. Schools of H.
argenteum commonly occur in shallow waters
along sandy beaches and shallow reefs during the
day, so it appears that at least some of the larger
individuals migrate offshore to kelp beds at dusk.
To reach Naples Reef, fish near the surf would
have to swim approximately 1.6 km offshore.

Resemblance between the day and night sam-
ples of S = 25 species within each of the four
vertical zones was measured by coefficients of
similarity or "overlap" (cf. Colwell and Futuyma

707

FISHERY BULLETIN: VOL. 74, NO. 4

1971). Similarity (Q is scaled from 0 (no resem-
blance at all) to 1.0:

C= 1.0 - V2 12\Pu - Pj,
I = 1

where P,j = the proportionate abundance of
species i in day sample j, and P,,^ = that in night
sample k.

Though the mid-water zone abounded with
fishes during the day, it appeared sparsely
populated at night (Table 4). Day-night similarity
within the mid-water zone was the least (C = 0.12)
for the four zones. Six of the 10 species recorded
from the mid-water zone at night were surf-
perches, while three of the remaining four were
rockfishes. Hyperprosopon argenteum accounted
for 62% of the total fish recorded in this zone.
Damalichthys vacca, along with Sebastes serra-
tioides and adult S. mystinus, were often seen
scattered in the water column at night.

Although the suprabenthic zone underwent a
substantial reduction in fish abundance at night,
its day-night species similarity was the highest
(C = 0.67) for the four zones. During both day and
night, the suprabenthic zone was dominated by
surfperches. At night, surfperches comprised the
four most abundant species, accounting for almost
80% of the total fishes observed in the suprabenthic
zone (Table 4). Although Pacific electric rays
{Torpedo californica) were never recorded over the
transect lines, they were often encountered
nearby, swimming slowly and hovering above the
bottom (Bray, Hixon, and Ebeling unpubl. data).
Swell sharks {Cephaloscylliu m ventriosu m), whose
nocturnal activities were investigated by Nelson
and Johnson (1970), were occasionally seen swim-
ming just above the reef at night.

Fish observed in the bottom zone increased from
7.4% of the total individuals recorded from all
zones during the day to 17.1% of the total at night
(Table 4). The zone's relatively low day-night
species similarity (C = 0.28) was due to variations
in numbers of the demersal ambusher-type
predators and increases in numbers of "resting"
surfperches. Among the ambusher-type species,
e.g., numbers of painted greenling, Oxylehius
pictus, decreased from 179 counted during the day
to only 6 at night, and numbers of two common
rockfishes increased: the black-and-yellow, S.
chrysomelas, almost doubled and the gopher, S.
carnatus, almost tripled (Table 4).

At night, most fishes were observed in the
shelter zone. Although only 2% of the day total of
fishes were seen in holes and crevices, 50% of the
night total were observed there (Table 4). Day-
night species similarity was fairly low (C = 0.36),
largely because of the increase in numbers of
individuals of Chromis punctipinnis observed in
holes: from only 13 counted during the day to 610
counted at night (Table 4). Individuals of
Pimelomefopon pulchrum and S. mystinus were
also commonly seen in the shelter zone at
night.

These counts of fishes inhabiting holes,
especially at night, may be conservative because
we could not completely census the numerous deep
holes and crevices along the transect line. This
problem certainly influenced our counts of in-
dividuals of 0. pictus and juvenile S. mystinus.
Nocturnal counts of both species were much lower
than those made during the day, and the in-
dividuals that were observed at night were invar-
iably hiding deep in holes. Subsequent nighttime
applications of small amounts of the anesthetic
quinaldine to holes that first appeared vacant
often yielded several 0. pictus and 5 to 20 juvenile
S. mystinus. Similar applications of this anes-
thetic during the daytime occasionally revealed
these fishes, but in far smaller numbers.

The vertical positions of the 25 species of fishes
during the day and night are summarized in Table
4. Data on some species are fragmentary because
individuals of these species were rarely encoun-
tered along the transect line. However, general
observations made during hundreds of hours of
diving during both day and night tend to sub-
stantiate conclusions based on these data. For
example, we saw but two kelp bass along the
transect line at night, one in mid-water, the other
on the bottom. In surrounding areas, we saw many
individuals resting on the bottom, several in
mid-water, but very few in holes. Eighteen of 24
species recorded during the day were most com-
mon in the suprabenthic and mid-water zones
above the reef. Only the treefish, .S. serriceps, was
most common in the holes of the shelter zone. Of
the 23 species recorded at night 16 were most
common in contact with the reef, either in the open
positions of the bottom zone or in the holes of the
shelter zone. Only two species, Hyperprosopon
argenteum and S. serranoides, were most common
in the mid-water zone.

The day-night differences in the activities of
many species involved considerable shifts among

708

EBELING and BRAY: ACTIVITY OF REEF FISHES

the four zones. These shifts are measured in Table
4 by values of A/;:

= 2 liPi

1 = 1*-

'iday)(l) - iPi

night) (^)J ,

where p,day is the proportion of individuals of a
species observed during the day in zone / (i = 1,2,
3, or 4 for the shelter through mid-water zones,
respectively) and p, nj^ht is the proportion observed
at night. The Ih's range from +3.0, when all
observed individuals of a species undergo a max-
imum shift downward from the mid-water zone
during the day to the shelter zone at night, to -3.0,
when all individuals undergo the reverse max-
imum shift upward. A Ih = 0.0 indicates little or
no shift, in that the species' proportional distribu-
tion among zones does not change from day to
night.

Fish species varied considerably in the degree to
which they changed zones between day and night,
although the patterns of shifting upward or
downward were similar within families (Table 4).
Some species changed their vertical position little
if at all: several species of rockfishes; the cabezon,
Scorpaenichthys marmoratus; Rhacochilus tox-
otes; Damalichthys vacca; and blackeye goby,
Coryphopterus nicholsii. Other species changed
their vertical position markedly between day and
night. Individuals of Chromis punctipinnis and
Pimelometopon pulchrum, which had near-max-
imum positive values of Ih, move about in the
water column during the day and shelter in holes
at night. No individuals of Oxyjulis californica
were seen at night (recall that they descend from
mid-water to bury themselves in sand or gravel
patches). Assuming that burying individuals are
in the "shelter zone!' Ih for Oxyjulis = 2.91. No
species had a large negative value of Ih, i.e., no
species mostly contained individuals that rose
from the bottom to mid-water at night. Hobson
and Chess (1976) noted that during the day most
Sebastes atrovirens were "seated on rocky strata"
whereas at night they "hovered in mid-water." In
the present study, the A/i of S. atrovirens was
small but positive (Table 4); however, this species
was relatively rare in our transects.

Several lines of evidence indicate that many of
the kelp-bed fishes observed become less active
and do not regularly feed at night. The levels of
activity often could be inferred from direct obser-
vations. Many species that swam about and fed on
or above the reef during the day were found deep
in holes and crevices at night and would flee from

their shelter only when vigorously disturbed.
These species included Hypsypops rubicundus, C.
punctipinnis, P. pulchrum, and juvenile S. mys-
tinus. Some individuals of P. pulchrum reportedly
secrete a mucous envelope about themselves
(Wiley 1974), and we often found this fish wedged
deep in crevices in an apparent state of torpor at
night. Individuals of Girella nigricans were also
found in holes or on the bottom but were more
responsive to our presence. Previous diel analyses
of gut contents substantiate our present impres-
sions that the following species are strictly day-
time feeders: H. rubicundus (by Clarke 1970), 0.
californica (by Bray and Ebeling 1975), juvenile S.
mystinus (by Thomas Bailey unpubl. data), and C.
punctipinnis (by Hobson and Chess 1976; Bray
unpubl. data).

Our analyses of fish-gut emptiness revealed that
even many of the kelp-bed fishes not undergoing
such obvious diel changes in vertical position may
stop feeding at dusk (Figure 1, Table 5). Although
all five demersal surf perches {Embiotoca jacksoni,
E. lateralis, Hypsurus caryi, Damalichthys vacca,
and Rhacochilus toxotes) generally remain in the
suprabenthic and bottom zones both day and
night, their diel patterns of gut emptiness indicate
that all but R. toxotes do not feed at night. Median
scores of gut fullness for E. jacksoni reached
maximum values in the afternoon and declined
after sunset; at dawn, all guts examined were
empty (Figure la). Fully 88% of the fishes speared
during daylight hours contained food in their
foreguts (Table 5). Although 39% of the fishes
collected at night contained food, 89% of these
were collected before midnight. Thus it is likely
that the food contained in the foreguts of these
individuals was eaten before nightfall and had not
yet passed into the second quarter of their guts.
Foreguts of E. lateralis, H. caryi, and D. vacca
show the same pattern (Figure Ib-d). In fact, all
four species had significantly less food in their
guts during the night than during the day (Table
5). Gnose (1968) also observed that individuals of
E. lateralis collected from off" Oregon had empty
guts at dawn. Additionally, two other kelp-bed
surfperches that commonly occur in mid-water,
Phanerodon furcatus and Brachyistius frenatus,
which is rare at Naples Reef, feed mainly during
the day (Bray and Ebeling 1975; Hobson and Chess
1976).

In contrast, median scores for fullness of R.
toxotes reached maximum values at night, and
many foreguts were empty during the day (Figure

709

FISHERY BULLETIN: VOL. 74, NO. 4

3
2

I h

!'-•'-•'

0600

1200

1800

2400

0600

\'-i'

3
2

.'„■-

, if-ii'^^'

0600

1200

1800

2400 0600

./T\

►«

>ll

ivi'->'

0600
5 r

4
3 ■

1200

.Rn

1800

2400 060O

♦^-•^^-•^

0600

1200

1800

t'.

'•v4'g-»

0600 1200 1800 2400 0600

Time, h

Figure l.-Scored fullness (1, empty-5, full) of foreguts of five
demersal surfperches: a, Embiotoca jacksoni; b, E. lateralis; c,
Hypsurus caryi; d, Damalichthys vacca; and e, Rhacochilus
toxotes. Each point represents the median score and each vertical
line the range of scores for (n) individuals collected over a 2-h
interval. Time is measured relative to sunrise (0600 h) and sunset
(1800 h).

le). Some 67% of the fish collected at night con-
tained food in their foreguts, and 49% of those
collected during the day also contained food,

Table 5.- Day-night variation in "foregut" emptiness for the
five species of demersal surfperches. Values of chi-square with 1
df were calculated from day-night, empty-not empty values in
contingency tables for each species.

Day

ght

Day

No.

vs. night

exam-

exam

- %

Species

ined

empty

ined

empty

Embiotoca jacksoni

26.4

<0.005

E. lateralis

5.9

23.0

<0.005

Hypsurus caryi

18.6

<0.005

Damalichthys vacca

8.3

27.9

< 0.005

Rhacochilus toxotes

3.2

«0.07

although this difference was not significant (Table
5).

It is likely that many of the large-mouthed
demersal species feed any time that suitable prey
are available. Included among these species are
various rockfishes {Sebastes carnatus, S. chry-
somelas, and S. serriceps) and Scorpaenichthys
marmoratus, all of which are cryptically patterned
and probably ambush much of their prey. Analyses
of gut fullness and states of digestion relative to
time of day may be of little value in determining
the feeding chronology of these fishes, especially
larger individuals. Kariya (1969) showed that food
items may take days rather than hours to pass
through the stomach of Sebastes: inermis, a species
from Japan; and Larson (pers. commun.) found
that small majid crabs (10 mm carapace width)
were still intact in the stomachs of adults of S.
carnatus up to 10 h after ingestion. However,
other lines of evidence indicate that they feed at
night. We saw more individuals of 5. carnatus and
S. chrysomelas at night, probably because they
were more active then, and the types of food items
included in their diets suggest they feed at night
as well as during the day. Their diets include
medium-sized crustaceans (crabs, shrimps, etc.)
and cephalopods (Quast 1968b; Larson 1972); both
prey were far more active and exposed along the
transect line at night. We have observed in-
dividuals of S. carnatus and 5. chrysomelas at
night with live, struggling crabs and octopi pro-
truding from their mouths. Also, individuals of
these species often consumed small fishes that
escaped from our collecting spears during night
dives. Finally, all of these fishes can be caught by
hook and line at night as well as during the day
(Milton Love, pers. commun.).

Along the transect line during the day, fishes
congregate in mid-water to pick plankton and
browse on kelp surfaces. At night, on the other
hand, almost all of the foraging by fishes occurs on

710

EBELING and BRAY: ACTIVITY OF REEF FISHES

or near the bottom. Hyperprosopon argenteum
was an exception in tliat individuals of this species
occurred alone or in small, loose groups in mid-
water at night. Hobson and Chess (1976) and Bray
(unpubl. data) found that guts of specimens
speared at night were full of recently ingested
prey, whereas almost all guts from individuals
speared during the day were empty. However, the
fact that this fish constitutes a large portion of the
catch made by shore fishermen (Frey 1971) in-
dicates that at least some individuals feed during
the day.

We know little about the feeding periods of the
remaining six species seen along the transect line.
The half moon, Medialuna calif orniensis, often
appeared to be more sensitive to our presence than
were individuals of other species near the bot-
tom, and we cannot deny the possibility that
Medialuna feeds at night. It reportedly eats
mainly algae supporting a variety of attached
epiphytic animals and much smaller quantities of
free animals (Quast 1968b; FoUett et al. 1960). The
fact that two small demersal species, Coryphop-
terus nicholsii and Oxylehius pictus, were seen
much less often at night suggests that they retreat
deep into holes and crevices then. Larger in-
dividuals of lingcod, Ophiodon elongatus; Parala-
brax clathratus; and S. mystinus eat cephalopods
as well as fishes and other prey (Love 1974; Miller
and Geibel 1973; Quast 1968c), so it is reasonable to
suspect that they too feed, at least occasionally, at
night.

DISCUSSION

During the day, large numbers of fishes pervade
the study area of reef and kelp off Santa Barbara.
Most fishes inhabit the mid-water zone well off the
bottom, while smaller numbers of ambusher-type
predators remain in contact with the reef bottom.
In contrast, the same kelp forest appears almost
abandoned at night. Most notably, large numbers
of fishes disappear from mid-water, while the
numbers of fishes increase markedly in the shelter
zone of holes and crevices.

Although day-night changes in fish abundance
may be partly attributable to sampling error
caused by our use of lights at night, etc., these
changes most certainly reflect differences in the
fishes' requirements and distributional patterns
between their periods of activity and inactivity.
During the day, the area in the vicinity of our
transect line seems to constitute a focal point of

fish activity. Daytime fish diversity and abun-
dance appeared to be greater along the transect
line than in adjacent areas, 5 to 10 m away. Loose
aggregations of juvenile 5. mystinus and, less
frequently, Chromis punctipinnis, P. clathratus,
and Girella nigricans, formed in the water column
above the transect line. Likewise, other fishes
gathered closer to the bottom. Perhaps this local
richness relates to the position of the transect
about the reef crest. The transect line was at-
tached to one of the highest rocky prominences on
the reef, and it ran along the inner margin of a
dense stand of giant kelp. Quast (1968a) noted that
the combination of high-relief rocks and kelp
augments the surface area available for inverte-
brates, the principal food of the fishes, and serve as
orientation points for fishes throughout the water
column. Also, inshore and offshore margins of kelp
beds often demonstrate the "edge-effect," in that
the fauna is richer there than in areas on either
side (Feder et al. 1974). At Naples Reef, surf-
perches, especially individuals of Embiotoca later-
alis and Phanerodon furcatus, tend to congregate
about the reef crest and the south dropoff 20 m
away where thick stands of Gelidium and other
red algae flourish. Individuals of E. lateralis gorge
themselves on the caprellid amphipods that occur
in great numbers amongst the algae (Robert
Cowen and David Laur, pers. commun.). Also we
noticed that fishes tend to aggregate in sunlit
areas like the reef crest and avoid the shaded areas
on either side. In tropical reefs, diurnally schooling
fishes that migrate to adjacent sand flats at night
return in daylight to the same prominent topo-
graphic features on the reef (Hobson 1973). Other
factors that influence local fish abundance and
diversity are: availability of food (e.g., Hobson
1968, 1972, 1974), proximity to shelter (e.g., Low
1971; Sale 1972), and the presence of "cleaner" fish
that rid larger fish of their ectoparasites (Slobod-
kin and Fishelson 1974). At least some of these
factors may also have contributed to the high
numbers of fish along our transect line.

After dark, fishes that seek shelter and/or
become inactive are no longer attracted by richer
feeding grounds and orientation points charac-
teristic of the reef crest. As darkness falls, mid-
water aggregations of C. punctipinnis, Parala-
brax clathratus, Girella nigricans, and young S.
mystinus dissolve as the fish disperse singly or in
small groups over the bottom to shelter in the
many holes and crevices in surrounding areas.
During the day, for example, individuals of C.

711

FISHERY BULLETIN: VOL. 74, NO. 4

pnnctipinnis occur patchily in small mid-water
aggregations over the transect line and in much
larger aggregations along the outer margins of
the kelp bed. At night, however, they shelter in
holes throughout the entire study area. In fresh-
water lakes of Ontario at night, day-active fishes
move into shallow water where there is sufficient
cover for sheltering (Emery 1973), but in the
tropics, most day-active reef fishes shelter in holes
deep in the coral and so their exposed numbers
decrease at night (see Hobson 1974).

The decrease in nocturnal abundance of fishes in
the transect area might have been caused by their
migrations to nearby areas of sand. Over coral
reefs, many of the more prominent fishes seen in
large stationary schools during the day are ac-
tually nocturnal species that leave the reef at dusk
(Hobson 1968). Among these are croakers (Sciaen-
idae), snappers (Lutjanidae), and grunts
(Pomadasyidae), which move to surrounding sand
flats to feed on their invertebrate prey during the
night (see papers by Hobson). However, we found
no evidence of a pronounced nocturnal migration
of fishes from reef and kelp to the surrounding
sand. Essentially all of the fishes observed during
the day were accounted for at one part or another
of the reef at night. On several occasions at night,
while swimming considerable distances over the
surrounding sand flats, we saw only species that
occur commonly at kelp-bed margins and do not
actively forage at night (e.g., Phanerodonfurcatus
and Damalichthys vacca), or that typically inhabit
sandy bottoms (e.g., the spotted cusk-eel Chilara
taylori, and various skates and rays). We have
occasionally seen relatively inactive schools of
black croaker, Oieilotrema saturnum, on the reef
during the day, and although we have not seen the
fish at night, it is possible that they migrate to
adjacent sandy areas to feed. Limbaugh (1961)
reported that they are most active at night.

In a study of the night habits of coral reef fishes,
Starck and Davis (1966) noted that the feeding
times of reef fishes are closely related to the type
and activities of their prey. Microcarnivorous and
omnivorous fishes that browse and pick at sessile
organisms are generally active only during the
day. Mesocarnivorous fishes (i.e., those that feed
on larger motile invertebrate prey) are largely
nocturnal, because their prey (e.g., crustaceans)
are active and exposed at night. Planktivorous
fishes feed during both day and night, the noctur-
nal species having larger eyes than their diurnal
counterparts. Piscivorous fishes feed opportunis-

712

tically during the day and night, but are most
active at dawn and dusk when their prey fish are
exposed while moving to and from foraging and
sheltering areas. This feeding pattern has also
been observed in other tropical areas (Hiatt and
Strasburg 1960; Collette and Talbot 1972; Hobson
1974) and in freshwater lakes (Emery 1973).

Kelp-bed fishes also tend to show this general
feeding pattern, though perhaps not so distinctly.
Small-mouthed microcarnivores that pick or graze
sessile invertebrates and hidden prey from off" the
bottom and other substrates are generally active
only during daylight hours. Such foragers, includ-
ing most of the surf perches, as well as Oxyjvlis
californica, Pimelometopon pulchrum, and Hyp-
sypops rnbicnndns, readily converge on urchins
broken open during the day, but completely ignore
such chum at night.

Also as in the tropics, though less extensively so,
different planktivores feed in the mid-water zone
of kelp beds during the day and night. The most
visible daytime planktivores, Chromis punctipin-
nis and juvenile S. m.ystinus, often form mixed
aggregations of individuals that pick small zoo-
plankton from the incoming currents. At night,
neither was seen exposed outside its shelter, and
individuals collected by spear and later examined
had empty stomachs. Instead, the mid-water zone
is dominated at night by the large-eyed species,
Hyperprosopon argenteum, which darts about,
actively feeding throughout the water column.
Though we have little data on kelp-bed mesocar-
nivores, some, such as various rockfishes, Scor-
paenichthys marmoratus, and Ophiodon elonga-
tus, may feed at night.

We emphasize the fact that many kelp-bed
fishes show considerable intraspecific variability
in vertical distribution and feeding activity. Al-
though a large majority of the population of C.
punctipinnis usually feeds in mid-water during
the day, e.g., a few individuals can usually be
found in holes. Likewise, a small proportion of the
day-sampled individuals of Embiotoca jacksoni
had empty foreguts even though the species is
strictly a diurnal forager. Even more variable is
the feeding schedule of Rhacochilus toxotes. Most
individuals probably have empty guts at any
daylight hour, although others are satiated. We
have observed that at any given time during the
day, most of these surfperch assemble as schools of
varying sizes just above the bottom or even in
mid-water (see also Alevizon 1975). However, a
lone individual may suddenly leave the school to

EBELING and BRAY: ACTIVITY OF REEF FISHES

feed rapidly over the bottom for several minutes
before rejoining the same or another school of
lazily swimming, nonfeeding fish. But we do not
know yet if any particular individuals tend to feed
in this sporadic manner during the day to a
greater extent than do most others which may
feed more consistently during the night. Hobson
(1971, 1976) stressed the probably widespread
occurrence of individual variation in the tendency
of fishes to "clean" ectoparasites from larger host
species. Specifically, Hobson (1976) observed that
even though cleaning is not considered to be
characteristic of the rock wrasse, Halichoeres
semicinctus, this feeding mode was repeatedly a
major activity in what was probably the same
individual. Thus, as far as fish activities are
concerned, the behavior of an individual is not
always predictable from the general characteris-
tics of its species.

Temperate-Tropical Differences

Some phenomena that characterize the day-
night change in activities of fishes inhabiting
tropical coral reefs appear less well developed or
absent in the activity cycles of the Santa Barbara
kelp-bed fishes. For one thing, no kelp-bed species
that we observed forms inactive schools over the
reef during the day and disperses elsewhere to
feed at night, as do snappers and grunts in tropical
systems. Another noticeable lack in the kelp
forests is the widespread replacement of daytime
mid-water planktivores by nighttime counter-
parts. In tropical areas, this replacement involves
more species and, to some extent, occurs vertically:
at night, the diurnal planktivores (a few
pomacentrids, the unusual labrid Clepticus, etc.)
take refuge in reefs that had provided shelter for
the nocturnal planktivores (some holocentrids,
apogonids, priacanthids, etc.) during the day
(Hobson 1968, 1972; Collette and Talbot 1972). In
our kelp-bed system at Naples Reef, however, the
only noticeable replacement of the abundant
daytime mid-water planktivores is Hyperprosopon
argenteum, which, moreover, is probably a "hor-
izontal replacement" from inshore areas. In this
system, the only common fish that shelters during
the day and may emerge at night is Cephaloscyl-
lium ventriosum, a rather slow-moving piscivore
that probably eats sheltering or inactive prey
(Nelson and Johnson 1970). The cryptic demersal
mesocarnivores (i.e., carnivores that feed on
medium-sized prey, e.g., rockfishes, Scorpaen-

ichthys marmoratus, etc.) may shelter either day
or night between feeding bouts.

However, feeding on plankton at night may be
more widespread in areas farther south. Hobson
and Chess (1976) concluded that several species eat
plankton at night off Santa Catalina Island.
Though they are relatively rare at Naples Reef, for
example, individuals of Sebastes atrovirens and
larger juveniles of S. serranoides are important
mid-water planktivores at night in kelp beds off
Santa Catalina. Also, Xenistivs californiensis
picks plankton in the relatively clear waters
around this island, but this species does not com-
monly occur as far north as Santa Barbara. Naples
Reef is located just south of a faunal boundary at
Point Conception (cf. Hubbs 1960; Quast 1968;
Briggs 1974). Also, our mainland assemblage differs
noticeably from nearby insular communities
(Ebeling et al. unpubl. data). Nonetheless, it is
reassuring to find that many of our results parallel
those of Hobson and Chess.

Figure 2 summarizes the day-night distribu-
tions of kelp-bed fishes from an evolutionary point
of view. Fish are depicted as being distributed
vertically, based on their proportionate abun-
dances in each of the four zones, from the mid-
water zone to the shelter zone, and as comprising
four ecological groups, based on their habits and
phylogenetic origins. Belonging to taxa with
temperate origins, all species in group A are
demersal species of the bottom-habitat group,
which generally move but little from their perches
on the bottom during the day or night (see Table 4,
\h = 0.42). Groups B, C, and D are composed of
more active species that commonly occur in the
suprabenthic zone and in mid-water during the
day, but there the similarity ends. Also with
temperate origins, species in group B are large-
mouthed generalized predators, which can switch
from plankton to larger prey including small fishes
as the occasion arises (Love 1974), and simply
descend to rest on the bottom at night {Ih = 1.92).
Group C and D species are small-mouthed
microcarnivores of mixed origins, which either
forage over the substrate or pick plankton from
mid-water. Group C fishes are all surfperches with
a common temperate origin, whose day-night
change in vertical position is relatively slight
{ih ^ 0.22), and whose nocturnal behavior is rela-
tively unspecialized. in that the fish simply slow
down over the bottom and do not generally seek
shelter in holes and crevices. But in contrast with
all the others, group D fishes appear to be rela-

713

FISHERY BULLETIN: VOL. 74, NO. 4

Figure 2.-Day and night positions of four ecological groups of
fishes inhabiting Santa Barbara kelp beds: A) demersal species
{Coryphopterus nicholsii, Ophiodon elongatus, Oxylebius pictiis,
Sebastes carnatus, S. chrysomelas, S. serriceps, and Scorpaen-
ichthys marmoratus); B) large-mouthed generalized predators
{Paralabrajc clathratus, Sebastes serranoides, and adult S.
mystinus); C) surfperches {Embiotoca jacksoni, E. lateralis,
Hypsurus caryi, Pkanerodon furcatus, Rha^ochihis toxotes, and
Damalickthys vacca); D) small-mouthed grazing and picking
tropical derivatives (Chromis punctipinnis, Hypsypops
nibicundus, Medialuna califomiensis, Oxyjulis californica, and
Pimelometapon pulchrum). Vertical zones (I-IV) are defined in
Table .' Each fish symbol represents 10% of the total individuals
in the group expressed proportionally to the relative abundances
of the different species in the group.

tively recent derivatives of primarily tropical
families (Pomacentridae, Labridae, etc.), and
essentially all show extreme changes in their
vertical distribution {\h = 2.07) as they actively
seek the shelter zone refuge. Some at least, like
Chromis punctipinnis, are specialized to the
extent that they tend to "home" to the same hole
on successive nights (Bray unpubl. data).

Thus, in the kelp beds, there is no broad re-
placement for the "day shift" of fishes at night,
even though the fishes' invertebrate prey appear
to be more active and exposed then. And, in
general, after the dusk period of intensified ac-
tivity, the notably lackluster night life gives the
kelp forest an aura of desolation, as compared with

the pictures of renewed (albeit lessened) activity
painted of the community of coral reef and out-
lying sand-flat fishes at night (Starck and Davis
1966; Collette and Talbot 1972; etc.). Perhaps the
relatively clear and well-lighted tropical waters
are more conducive to nocturnal activity for the
many visually oriented fish. Denied much of the
moonlight by the dense kelp canopy and frequent
low clouds, the relatively turbid, temperate waters
are often a dark and gloomy place at night. In fact,
even during the day when the water is particularly
turbid, the usually active planktivores, grazers,
and browsers tend to stop foraging and often seek
shelter, as do their tropical counterparts under
similar conditions (Collette and Talbot 1972).

It is paradoxical that the "tropical derivatives"
(Figure 2D) persist in their complex nocturnal
shelter-seeking while many primarily temperate
fishes remain exposed. One explanation assumes
that selection pressures brought about by noctur-
nal (or crepuscular) predation are either different
or more relaxed in our temperate system of kelp
forest and reef than in the tropical reef system.
Observing a similar set of circumstances, Hobson
(1972) noted that Hawaiian reef fishes, which
enjoy a relative dearth of crepuscular predators,
show the same specialized sheltering behavior
during twilight as do their close relatives in the
Gulf of California, which have many such preda-
tors. He suggested that these complex behavior
patterns may evidence historic selection pressures
from predators. These patterns may persist on
Hawaiian reefs today even though they are cur-
rently perhaps less critical to the survival of the
refuge-seeking species than in reef systems else-
where. An alternative explanation holds that
crepuscular and nocturnal predation by, e.g., the
Pacific electric ray, is important in kelp beds, but
that the tropical derivatives compete more suc-
cessfully against the primarily temperate species
for shelter.

CONCLUSIONS

As indicated by paired day-night observations
along a transect line, kelp-bed fishes occur in about
the same relative abundances throughout the year
in an area of reef and kelp along the mainland side
of the Santa Barbara Channel. During the day,
most fishes occupy the "mid-water zone" higher
than 1 m off the bottom. Far fewer are "exposed on
the bottom" or in the "shelter zone" of holes and
crevices in the reef itself. During the night, when

714

EBELING and BRAY: ACTIVITY OF REEF FISHES

the number of individuals appears reduced by
more than half, most fishes occupy the bottom and
shelter zones.

Thus, like that of tropical reefs, the vertical
distribution of fishes changes markedly between
day and night. Planktivores that pack the mid-
water zone during the day virtually abandon the
area at night to rest on the bottom or seek shelter
in reef holes. The vacated mid-water space is only
partly reoccupied by a relatively sparse population
of nocturnal planktivores and a few remaining
generalized carnivores. The largest relative in-
crease of individuals occurs in the shelter zone,
where superabundant daytime planktivores, such
as the blacksmith, hide at night. With so many
fishes commuting extensively between the mid-
water and shelter zones, it is understandable that
the intervening suprabenthic zone shows the
greatest species similarity between day and night.
Many ambusher-type foragers are always orient-
ed to the bottom and change their positions rela-
tively little for the night shift.

It seems likely, therefore, that at night feeding
on plankton decreases and most of the foraging by
fishes takes place over the bottom. The large-
mouthed demersal ambushers— various rockfishes,
the cabezon, and others-probably feed almost any
time that suitable prey are available. The rubber-
lip seaperch may actually feed more actively at
night. Nonetheless, many of the fishes that
wander over the bottom at night may stop feeding
at dusk. Most demersal surfperches remain ex-
posed at night, although their foreguts soon
empty, and the fish appear more lethargic than
they do during the day when they are actively
foraging.

Focal points of daytime fish activity, such as the
productive crest of the reef and other prominent
landmarks, appear to lose their attractiveness at
night. Most aggregations disappear at dusk as
fishes generally disperse out over the reef bottom.
But unlike many tropical-reef fishes, kelp-bed
species do not normally move off the reef to forage
over the adjacent sand flats.

Kelp-bed fishes often show considerable
intraspecific variation in vertical distribution and
feeding activity. During the day, e.g., noticeable
numbers of typically mid-water species invariably
seek shelter, while at night some individuals
remain in the water column. And fishes differ in
the intensity at which they feed during any given
period during the day. All this suggests that
certain individuals may assume and even main-

tain distinctive habits that differ from the species
"norm," i.e., the behavior of a particular fish is not
always predictable from the general characteris-
tics of its species.

Thus, in comparing the diel behavior of kelp-bed
fishes as a group with that of their tropical coun-
terparts, it becomes apparent that even though
both groups follow the same basic patterns, the
kelp-bed community is the more loosely
"programmed." In the kelp-bed system, for exam-
ple, there is less large-scale replacement of fishes
between discrete areas or vertical zones at dusk.
Here, the night shift offers no real substitute for
the dense aggregations of daytime planktivores or
demersal microcarnivores, even though these
fishes' invertebrate prey are active and exposed at
night. Perhaps the better lighted tropical waters
allow more specialized activities because here the
visually oriented fishes can better see what they
are doing, even by moonlight. In the kelp forest,
the level of fish activity decreases even during the
day when the water becomes very turbid, as often
happens with the onset of dense blooms of phyto-
plankton during the spring and summer.

The kelp-bed species that belong primarily to
tropical families tend to show the same specialized
pattern of nocturnal shelter seeking as do their
close tropical relatives, even though the general
program of diel activity in the kelp forest appears
to be comparatively unstructured. Perhaps the
specialized refuge-seeking procedures of kelp-bed
pomacentrids and labrids are simply
"evolutionary holdovers" that contribute relative-
ly little to the present fitness of these fishes. But
alternatively, the "tropical derivatives" may ac-
tually compete more successfully against primari-
ly temperate species such as surfperches for
shelter on the reef. Even though the intensity of
predation at twilight and perhaps at dark may be
somewhat less in our temperate system than in the
tropics, a few ingenious and effective predators,
such as the Pacific electric ray, patrol the Santa
Barbara kelp forests throughout the night.

ACKNOWLEDGMENTS

We thank Edmund Hobson, Ralph Larson, and
Robert Warner for critically reading the manu-
script and offering helpful suggestions. James
Cook and several students, especially Larry Asa-
kawa, Craig Fusaro, David Laur, Gary Morris,
Paul Reilly, Michael Rode, and Dale Sarver, helped
with the diving operations. Steve Edwards and M.

715

FISHERY BULLETIN: VOL. 74, NO. 4

Rode assisted with the fish-gut analyses. Norm
Lammer provided invaluable technical assistance
with equipment and boating operations, and
Cindy Nissley drafted the illustrations. This work
is a result of research sponsored by NOAA, Ofl^ce
of Sea Grant, Department of Commerce, under
grant no. 2-35208-6 and 04-3-158-22, R-FA-14; and
by NSF Grant GA 38588 and Sea Grants GH 43
and GH 95. Supplementary funding was provided
by a U.C.S.B. Faculty Research Committee grant
(No. 369) for Computer Center user services, and
by the Marine Science Institute through the
courtesy of Henry Offen, Acting Director, for
interim project support.

LITERATURE CITED

Alevizon, W. S.

1975. Spatial overlap and competition in congeneric surf-
perches (Embiotocidae) off Santa Barbara, Califor-
nia. Copeia 197.5:352-356.
Bray, R. N., and A. W. Ebeling.

1975. Food, activity, and habitat of three "picker-type"
microcarnivorous fishes in the kelp forests off Santa
Barbara, California. Fish. Bull., U.S. 73:815-829.
Briggs, J. C.

1974. Marine zoogeography. McGraw-Hill, N.Y., 475 p.
Brown, D. W.

1974. Hydrography and midwater fishes of three contiguous
oceanic areas off Santa Barbara, California. Los Ang. Cty.
Mus. Contrib. Sci. 261:1-30.
Clarke, T. A.

1970. Territorial behavior and population dynamics of a
pomacentrid fish, the garibaldi, Hypsypops ruhicunda.
Ecol. Monogr. 40:189-212.

COLLETTE, B. B., AND F. H. TaLBOT.

1972. Activity patterns of coral reef fishes with emphasis on
nocturnal-diurnal changeover. In B. B. Collette and S. A.
Earle (editors), Results of the Tektite program: Ecology of
coral reef fishes, p. 98-124. Bull. Los Ang. Cty. Nat. Hist.
Mus. Sci. 14.

COLWELL, R. K., AND D. J. FUTUYMA.

1971. On the measurement of niche breadth and over-
lap. Ecology 52:567-576.

DOMM, S. B., AND A. J. DOMM.

1973. The sequence of appearance at dawn and disappear-
ance at dusk of some coral reef fishes. Pac. Sci. 27:128-135.

Emery, A. R.

1973. Preliminary comparisons of day and night habits of
freshwater fish in Ontario lakes. J. Fish. Res. Board Can.
30:761-774.

Feder, H. M., C. H. Turner, and C. Limbaugh.

1974. Observations on fishes associated with kelp beds in
Southern California. Calif. Dep. Fish Game, Fish Bull.
160, 144 p.

FOLLETT, W. I., D. GOTSHALL, AND J. G. SMITH.

1960. Northerly occurrences of the scorpid fish Medialuna
californiensis (Steindachner), with meristic data, life-
history notes, and discussion of the fisheries. Calif. Fish
Game 46:165-175.

Frey, H. W. (editor).

1971. California's living marine resources and their utiliza-
tion. Calif. Dep. Fish Game Spec. Publ., 148 p.
Gnose, E. C.

1968. Ecology of the striped seaperch, Embiotoca lateralis,
in Yaquina Bay, Oregon. M.S. Thesis, Oregon State
Univ., 51 p.
Herald, E. S.

1961. Living fishes of the world. Doubleday and Co., Inc.,
Garden City, N.Y., 304 p.
Hiatt, R. W., and D. W. Strasburg.

1960. Ecological relationships of the fish fauna on coral reefs
of the Marshall Islands. Ecol. Monogr. 30:65-127.
Hobson,E. S.

1965. Diurnal-nocturnal activity of some inshore fishes in
the Gulf of California. Copeia 1965:291-302.

1968. Predatory behavior of some shore fishes in the Gulf of
California. U.S. Fish Wildl. Serv., Res. Rep. 73, 92 p.

1971. Cleaning symbiosis among California inshore fishes.
Fish. Bull., U.S. 69:491-523.

1972. Activity of Hawaiian reef fishes during the evening
and morning transitions between daylight and darkness.
Fish. Bull., U.S. 70:715-740.

1973. Diel feeding migrations in tropical reef fishes. Hel-
golander wiss. Meeresunters. 24:361-370.

1974. Feeding relationships of teleostean fishes on coral
reefs in Kona, Hawaii. Fish. Bull., U.S. 72:951-1031.

1975. Feeding patterns among tropical reef fishes. Am. Sci.
63:382-392.

1976. The rock wrasse, Halichoeres semicinctus, as a cleaner
fish. Calif. Fish Game 62:73-78.

HoBSON, E. S., AND J. R. Chess.

1976. Trophic interactions among fishes and zooplankters

near shore at Santa Catalina Island, California. Fish.

Bull., U 5. 74:567-598.
HUBBS, C. L.

1960. The marine vertebrates of the outer coast. In Sympo-
sium: The biogeography of Baja California and adjacent
seas. Part 2, p. 134-147. Syst. Zool. 9.

Kariya, T.

1969. The relationship of food intake to the amount of
stomach contents in mebaru, Sebastes inermis. Bull. Jap.
Soc. Sci. Fish. 35:533-536.

Larson, R.J.

1972. The food habits of four kelp-bed rockfishes (Scor-
paenidae, Sebastes) off Santa Barbara, California. M.A.
Thesis, Univ. California Santa Barbara, 56 p.

Limbaugh, C.

1961. Life-history and ecologic notes on the black croaker.
Calif. Fish Game 47:163-174.

Low, R. M.

1971. Interspecific territoriality in a pomacentrid reef fish.
Pomace ntrus flavicauda Whitley. Ecology 52:648-654.
Love, M. S.

1974. Food habits of three midwater kelp-bed predators.
M.A. Thesis, Univ. California Santa Barbara, 50 p.
Miller, D. J., and J. J. Geibel.

1973. Summary of blue rockfish and lingcod life histories; a
reef ecology study; and giant kelp, Macrocystis pyrifera,
experiments in Monterey Bay, California. Calif. Dep.
Fish Game, Fish Bull. 158, 137 p.

Nelson, D. R., and R. H. Johnson.

1970. Diel activity rhythms in the nocturnal, bottom-

716

EBELING and BRAY: ACTIVITY OF REEF FISHES

dwelling sharks, Heterodontus francisci and Cephaloscyl-
lium ventriosum. Copeia 1970:732-739.

QUAST, J. C.

1968a. Fish fauna of the rocky inshore zone. In W. J. North
and C. L. Hubbs (editors), Utilization of kelp-bed re-
sources in Southern California, p. 35-79. Calif. Dep. Fish
Game, Fish Bull. 139.
1968b. Observations on the food of the kelp-bed fishes. In W.
J. North and C. L. Hubbs (editors). Utilization of kelp-bed
resources in Southern California, p. 109-142. Calif. Dep.
Fish Game, Fish Bull. 139.
1968c. Observations on the food and biology of the kelp bass,
Paralabrax clathratus, with notes on its sportfishery at
San Diego, California. In W. J. North and C. L. Hubbs
(editors). Utilization of kelp-bed resources in Southern
California, p. 81-108. Calif. Dep. Fish Game, Fish Bull.
139.

Sale, P. F.

1972. Influence of corals in the dispersion of the jwrnacen-
trid fish, Dascyllus aruanus. Ecology 53:741-744.

Slobodkin, L. B., and I. Fishelson.

1974. The effect of the cleaner-fish Labroides dimidiatus on
the point diversity of fishes on the reef front at Eilat. Am.
Nat. 108:369-376.

Smith, D. L., and J. C. Tyler.

1972. Space resource sharing in a coral reef fish communi-
ty. In B. B. Collette and S. A. Earle (editors). Results of
the Tektite program: Ecology of coral reef fishes, p.
125-170. Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 14.

Starck, W. a., II, and R. E. Schroeder.

1965. A coral reef at night. Sea Front. 11:66-79.

Starck, W. A., II, and W. P. Davis.

1966. Night habits of fishes of Alligator Reef, Florida.
Ichthyol. Aquarium J. 38:313-356.

Wiley, J. W.

1974. Observations on the use of mucus envelopes by the
California sheephead, Pimelometopon pulchrum, on
southern California rock reefs. Copeia 1974:789-790.

717

INCIDENCE OF CULL LOBSTERS, HOMARUS AMERICANUS, IN
COMMERCIAL AND RESEARCH CATCHES OFF THE MAINE COAST^

Jay S. Krouse^

ABSTRACT

Data obtained by port sampling the Maine commercial lobster catch (1968-74) and the natural lobster
population near Boothbay Harbor, Maine, with research gear (1969-74) indicate that 6.5% of the
commercially harvested lobsters have lost at least one claw while 21.0% of the lobsters (all sizes) in the
natural population have missing and/or regenerating claws. An assessment of variations in cull
frequencies associated with different seasons, fishing localities, and lobster size distributions suggests a
direct relationship between fishing intensity and the incidence of culls. This information further
supports Krouse and Thomas' recommendation that all lobster traps be equipped with an escape vent
thus minimizing fishermen's needless handling of excessive numbers of sublegal-sized lobsters.

Over the years the occurrence of American lobster,
Homarus americantis, with a missing and /or
regenerating cheHped in the commercial landings
has undoubtedly resulted in a significant financial
loss to the fishing industry due to the culls' reduced
weight and marketability (retail price of culls is
less per pound). Scarratt (1973) reported that
commercially caught lobsters from ports off Nova
Scotia and Prince Edward Island had incidences of
missing claws ranging from 5 to 19%. Although
claw loss could not be attributed to a single factor,
causes related to fishing such as rough handling by
fishermen and movement of traps over the seabed
were cited. Recognizing the importance of this
situation, I have analyzed cull data provided by the
Maine Department of Marine Resources Lobster
Research Program's research catches (Krouse
1973) and sampling of the commercial catch
(Thomas 1973). In this paper I attempt to assess
the magnitude of the cull problem along the Maine
coast, some of its causes, and a possible solution to
diminish the number of culls.

METHODS

From June 1969 through December 1974, the
occurrence of lobsters with missing and/or regen-
erating claw(s) in daily catches of research gear

'This study was conducted in cooperation with the U.S.
Department of Commerce, National Marine Fisheries Service,
under Public Law 88-309, as amended. Commercial Fisheries
Research and Development Act, Project 3-153-R.

^Maine Department of Marine Resources, West Boothbay
Harbor, ME 04575.

was noted. Carapace length in millimeters, weight
in grams, and sex were recorded for each lobster.
Wire lobster traps (2.54 x 2.54 cm mesh) were
fished throughout the 6-yr period, whereas
modified wooden traps with plastic escape vents of
3.81, 4.13, and 4.45 cm were not used until July
1972. Most experimental fishing was conducted in
the vicinity of Capitol, Squirrel, and Damariscove
islands in the Boothbav region of Maine (Figure

1).

Information pertaining to the frequency of culls
in the Maine commercial catch from 1968 through
1974 was obtained from the probability sampling
program described by Thomas (1973).

A length-weight relationship was calculated for
297 lobster's with a regenerating claw and for 225
lobsters with a missing claw collected near Booth-
bay Harbor, 1972-73. All lobster culls used in this
determination had one normal sized claw. The
regression of weight on carapace length for these
two cull categories was fitted by the method of
least squares using the logarithmic transforma-
tion logio W = logio a + b logio L.

RESULTS AND DISCUSSION

Seasonal and Size Variation in •
Cull Frequency

From the research catches I have calculated the
percentage of culls by month and 5-mm size groups
for 1969 through 1974 (Tables 1, 2). Fluctuations in
the monthly percentages of culls seem to follow a
seasonal pattern, i.e., the number of culls peaked

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

719

FISHERY BULLETIN: VOL. 74, NO. 4

Table L-Monthly incidence of lobster culls in research catches
near Boothbay Harbor, Maine, 1969-74.

69''40'

69035

Figure l.-Map showing the areas fished near Boothbay Harbor,

Maine.

in the catch during winter-spring (26.9-31.3%),
then subsided in July-August (18.1-19.4%), and
increased in the fall (17.7-25.6%— variations in fall
percentages may be due to sporadic shedding
during this season). These seasonal changes might
be related to the time of molting (July-September)
associated with temporal fluctuations in fishing
pressure. If we assume that a high percentage of
the total number of culls are caused by fishing
operations as suggested by earlier observations of
fishermen's needless handling of excessive
numbers of sublegal size lobsters (Krouse and
Thomas 1975), then the decline in cull frequency
during July and August (peak of shedding) may be
explained, in part, by: 1) some culls losing this
status after shedding and regenerating normal
size claws, and 2) small lobsters (usually nonculls
by virtue of their nonvulnerability to fishing gear)
being recruited into the fishery at this time as a
result of shedding. Even though the catch data of
this study reveals that a high percentage of
lobsters <45 mm carapace length were culls, most
of these smaller culls probably acquired this con-
dition while confined in our lobster traps.

Total no.

Culls

Total no.

Culls

Month

examined

(%)

Month

examined

(%)

Jan.

31.3

Aug.

2,164

19.4

Feb.

26.9

Sept.

1,266

23.1

Mar.

29.4

Oct.

504

19.0

Apr.

26.5

Nov.

129

25.6

May

296

24.0

Dec.

17.7

June

805

24.0

July

1,032

18.1

Total

6,417

21.0

Table 2.— Percentage of lobster culls by 5-mm size groups in
research catches near Boothbay Harbor, Maine, 1969-74.

Carapace
length
(mm)

Total
number
caught

Culls
(%)

Carapace
length
(mm)

Total
number
caught

Culls

{%)

36- 40

26.3

76- 80

1,403

22.7

41- 45

33.8

81- 85

406

22.7

46- 50

160

21.9

86- 90

240

17.9

51- 55

333

19.5

91- 95

119

8.4

56- 60

542

23.2

96-100

10.0

61- 65

802

22.8

^101

11.1

66- 70

1,046

19.3

71- 75

1,232

19.2

Total

6,417

21.0

The frequency of culls by 5-mm increments
(Table 2) indicated that culls are most prevalent at
carapace lengths ^45 mm and progressively less
numerous at lengths ^86 mm. The high incidence
of culls for small lobsters can be attributed, at
least in part, to these lobsters being particularly
defenseless to claw loss inflicted by larger lobsters
wnthin the trap. On several occasions we have
either caught small lobsters with recent claw
losses in traps containing larger lobsters or ac-
tually witnessed larger lobsters destroying the
claw of their diminutive opponent. To further
substantiate this explanation of the two cull
categories, i.e., regenerating and missing claws,
the missing claw group predominated for lobsters
^50 mm; however, for sizes ^51 mm, lobsters with
regenerating claws usually outnumbered those
without claws (Table 3). This disparity was more
pronounced for wood traps within the 81- to 85-
and 86- to 90-mm groupings. Considering that
legal-sized lobsters are handled only once and
therefore are probably less prone to claw loss, then
one would expect these larger lobsters to have a
higher incidence of regenerating claws. Con-
versely, those sublegal-sized lobsters between 76
and 80 mm that are repeatedly discarded from the
fishermen's catch have a preponderance of missing
claws for catches with wire and wood traps.

The decline in the incidence of culls at the legal
sizes (Maine minimum legal size is 81 mm carapace
length) is manifested not only by the research

720

KROUSE: INCIDENCE OF CULL LOBSTERS IN CATCHES

Table 3. -Percentage of lobster culls with missing claws by 5-mm size
groups for research catches of wire and wooden traps.

Wire traps

Wood traps

Number of lobsters .,. .

Carapace Missing

length Regenerat- Missing claw

(mm) ing claw claw (%)

Number of lobsters

Regenerat-
ing claw

Missing
claw

Missing
claw
{%)

36-40

60.0

—

41-45

88.2

—

46-50

68.2

—

51-55

50.0

—

56-60

40.5

—

61-65

42.1

— .

—

66-70

44.0

80.0

71-75

42.6

48.4

76-80

56.5

52.8

81-85

54.8

36.2

86-90

33.3

i^91

50.0

Total

353

327

48.1

132

121

47.8

catches (Table 2) but also by the commercial
catches for 1968-74 (Table 4). For both catches,
more legal-sized culls occurred in the 81- to 85-mm
size group while the percentage of culls gradually
decreased for carapace lengths >85 mm. If, once
again, it is assumed that fishing operations often
cause culled lobsters and knowing that legal
lobsters are handled only once and not repeatedly
as may be the case for sublegal-sized lobsters, one
would expect fewer culls amongst legal lobsters
along with a gradual reduction in culls for sizes >85
mm. Since this study's data demonstrate this very
pattern, my contention concerning the possible
injurious effects of fishing activities on lobsters
<81 mm is strengthened. Certainly there is a
greater likelihood of a lobster becoming injured
when as a result of fishing operations this lobster
is: 1) crowded with other cannibalistic lobsters in a
trap; 2) held captive in a trap which may undergo
rigorous movement during a storm; 3) hauled
boatside with appendages dangling between the
trap's laths; and 4) removed from the trap while
clinging to the trap, fishermen, or another lobster
and eventually released for a descent to the ocean
floor during which predation may occur.

Table 4.-Incidence of lobsters with missing claws by 5-mm size
groups occurring in the commercial catch along the Maine coast
(1968-74).

Carapace

Total

Carapace

Total

length

number

Culls

length

number

Culls

(mm)

caught

(%)

(mm)

caught

(%)

81- 85

5,322

8.1

106-110

219

3.7

86- 90

7,373

6.2

111-115

109

1.8

91- 95

5,580

5.2

116-120

4.9

96-100

1,208

3.7

= 121

101-105

368

4.3

Total

20,226

6.5

Effect of Fishing Intensity on
Cull Frequency

The relationship of fishing intensity and its
influence on cull incidence was investigated by
calculating the frequencies of culls caught with
wire and wood lobster traps (Table 5) at three
different fishing sites near Boothbay Harbor
(Figure 1). In addition, length-frequency histo-
grams were constructed by 1-mm increments of
the catches for each of the sampling sites (Figure
2). This analysis revealed that catches off Capitol
Island, the most intensively fished area, contained
more culls (23.3% for wire and 22.0% for wood
traps) than either the catch of Damariscove
(21.2%) or Squirrel islands (12.7% for wire and
17.9% for wood traps) which had the fewest culls.
Although Damariscove and Squirrel islands ap-
peared to have similar trap concentrations based
on visual sightings of pot buoys, appreciably more
culls were trapped at Damariscove. Possible rea-
sons for this difference may be related to: 1)
lobsters being maimed by excessive movement of
traps over the substrate during storms off the
more exposed seaward shoreline of Damariscove
(waters fished at Squirrel were more sheltered); 2)
Damariscove's greater abundance of small lob-
sters which are more vulnerable to injury [average
size of lobsters in Damariscove catch was smaller
than those of the other two areas (Figure 2), and
the percentage of lobsters with missing claws was
highest at Damariscove (Table 5)]; and 3) perhaps,
an error in our rather subjective determination of
nearly equal fishing intensities for both islands.
Nevertheless, there does appear to be a positive

721

FISHERY BULLETIN: VOL. 74, NO. 4

CAPITOL ISLAND

104— E

WOOD TRAPS 1972-1974
X = 80 5

70 80 90 100

CARAPACE LENGTH , mn

120 130

DAMARISCOVE ISLAND

0 WIRE TRAPS 1969-1974

35 40

MINIMUM LEGAL SIZE

7b 80 90 13o

CARAPACE LENGTH, mm

55"^ — So

SQUIRREL ISLAND

112-^

[~j WOOD TRAPS 1972-1974

Figure 2.-Length-frequency histo-
grams for the lobster catches with wire
and wood traps for each of the three
sampling sites near Boothbay Harbor,
Maine.

80 90 l6o

CARAPACE LENGTH , mm

130

722

KROUSE: INCIDENCE OF CULL LOBSTERS IN CATCHES

Table 5-Comparison of the incidence of lobster culls in catches of wire and wood traps for various areas near Boothbay

Harbor, Maine, 1969-74.

Total

Regenerat

ng Both claws

Missing

Both claws

Regenerating and

no.

Culls

claw

regenerating

claw

missing

missing claws

Gear and area

lobsters

(%)

Wire traps:

Capitol

1,627

23.3

10.6

1.8

9.1

1.1

0.7

Damariscove

920

21.2

8.4

1.0

10.0

1.0

0.5

Squirrel

787

12.7

6.9

0.5

5.1

0.1

Wood traps:

Capitol

1,125

22.t)

10.0

1.4

8.4

1.3

0.8

Squirrel

162

17.9

8.6

1.9

5.6

1.9

correlation between fishing intensity and in-
cidence of culls; however, this does not preclude
other factors such as predation, intraspecific
competition, molting difficulties, and storm relat-
ed damages.

Loss of Value of Catch Due to Culls

At the beginning of this paper I mentioned that
culls have perennially detracted from the landed
value of the lobster catch. To assess this situation,
the regressions of weight for lobsters with missing
and regenerating claws on carapace length for
sublegal- and legal-sized lobsters were calculated.
These curves were then compared to the length-
weight relationship for noncull lobsters (Krouse
1973) (Figure 3). These comparisons reveal that
noncull lobsters are about 14 to 20% heavier than
those lobsters with regenerating and missing
claws. Knowing these weight differentials and
that about 6.5% of the lobsters in the commercial
catch are missing at least one claw (Table 4) and
that at least an equal percentage (6.5) of lobsters
must have regenerating claws, the cull loss to the
fishery can now be quantified. From the 1974
Maine Landings which reported a lobster catch of
16,457,666 pounds valued at $23,212,808, 1 estimat-
ed that the annual catch without any culls could
have been increased by about 363,700 pounds
(2.2%), adding $512,800 to the landed value. Un-
fortunately, there probably is no way to eliminate
culls completely; however, proper size escape vents
in all traps would be beneficial in effecting a
marked reduction in the incidence of culls (Krouse
and Thomas 1975). This reduction in culls would be
the result of decidedly fewer numbers of sub-
legal-sized lobsters being handled by fishermen as
indicated by the conspicuous disparity between
the size composition of research catches with wire
and vented wooden traps (Figure 2). Even if the
cull loss could be lessened by only 25%, the industry

60 70

CARAPACE LENGTH, MM

Figure 3.-Comparison of the calculated length-weight relation-
ships for lobsters with regenerating, missing, and normal claws
(noncull). The regression equations are: 1) regenerating claws:
logio ^= -2.99-h2.91 logio L; 2) missing claws: log,o
W = -3.03 + 2.92 logio L; and 3) noncull: logio W' = -2.91-1-2.90
logic L-

would still realize an annual increase of about
$128,000.

ACKNOWLEDGMENTS

I am grateful to the summer aides who assisted
in the collection of field data. Thanks are due to
David A. Libby for his help with data complila-
tions and figure drafting. I also extend my ap-

723

preciation to Robert L. Dow and James C. Thomas
for reviewing this paper.

LITERATURE CITED

Krouse, J.S.

1973. Maturity, sex ratio, and size composition of the
natural population of American lobster, Homarus amer-
icanus, along the Maine coast. Fish. Bull., U.S. 71:165-173.
Krouse, J. S., and J. C. Thomas.

1975. Effects of trap selectivity and some population pa-

FISHERY BULLETIN: VOL. 74, NO. 4

rameters on size composition of the American lobster,
Homarus americanus, catch along the Maine coast. Fish.
Bull, U. 8.73:862-871.

SCARRATT, D. J.

1973. Claw loss and other wounds in commercially caught
lobsters (Homarus americanus). J. Fish. Res. Board Can.
30:1370-1373.
Thomas, J. C.

1973. An analysis of the commercial lobster (Homarus
americanus) fishery along the coast of Maine, August 1966
through December 1970. U.S. Dep. Commer., NOAA Tech.
Rep. NMFSSSRF-667,57p.

724

FISHERY WASTE EFFLUENTS: A METHOD TO DETERMINE
RELATIONSHIPS BETWEEN CHEMICAL OXYGEN DEMAND AND RESIDUE

Jeff Collins and Richard D. Tenney'
ABSTRACT

Researchers and the fishing industry have experienced difficulty in applying the Environmental
Protection Agency's standard tests to industrial fishing waste effluents, especially for total suspended
and settleable solids, and oil and grease.

The relationship between chemical oxygen demand and residue was determined on a limited number
of samples from four types of screened waste effluents from November 1973 to September 1974: shrimp
using fresh or salt water processing, snow crab, and canned salmon. In addition to chemical oxygen
demand and residue, tests for settleable solids, total suspended and settleable solids, oil and grease,
protein, and salt were also performed. Based on these relationships, a method is suggested to develop a
system for the analysis of pollutants that will be more economic and give more meaningful data than
currently obtainable under Environmental Protection Agency's methods. The method requires that
base data on a plant be obtained to relate chemical oxygen demand with residue values using regression
lines and equations. A subsequent routine monitoring program need only test for total residue and
chemical oxygen demand of the filterable residue. Substitution into the equations gives the other
residue fractions and their chemical oxygen demand values, i.e., total chemical oxygen demand,
chemical oxygen demand of the particulate matter, filterable residue, and nonfilterable residue.

This laboratory has modified and studied in detail
a number of analytical techniques to measure
pollutants (Tenney)^. We have considered the
methods of testing specified by the Environmental
Protection Agency (EPA) to monitor fishery pol-
lutants and are of the opinion that the monitoring
program and analytical methods specified under
the National Pollutant Discharge Elimination
System (NPDES) program could be improved for
application to seafood-processing effluents
(Pojasek 1975). The purpose of this paper is to
suggest different tests for monitoring efl^uents
with certain prerequisites that would satisfy the
intent of the law, yet recognize both the technical
and economic problems associated with the fishing
industry's efforts to comply with the monitoring
regulations.

Since laboratory space, equipment, and labor
necessary to conduct a waste-monitoring program
are quite expensive to the fishing industry, eco-
nomics suggest the use of a minimum number of
tests to do the job, and where possible, the use of

Pacific Utilization Research Center Kodiak Utilization Re-
search Laboratory, National Marine Fisheries Service, NOAA,
P.O. Box 1638, Kodiak, AK 99615.

^Tenney, R. D. 1972. COD for Industrial Waste Water, Tech.
Rep. 97, 5 p.; 1972. Chemical Oxygen Demand, Tech. Rep. 101, 12
p.; 1973. Shrimp Waste Streams and COD, Tech. Rep. 104, 3 p.
Unpublished, intralaboratory reports, Kodiak Utilization Re-
search Laboratory.

inexpensive equipment. In some analyses, the
time required to complete any analysis is impor-
tant, as in the 5-day test for biological oxygen
demand (BOD). In this instance, the chemical test
(chemical oxygen demand-COD) provides quick
results and has better application. The limited
level of laboratory experience and equipment
generally found in seafood-processing plants and
their diverse and often remote locations also
suggest that the regulations and permit system
should reflect these limitations and require only
fairly simple tests to measure pollutants. At the
same time, however, analytical techniques used to
measure pollutants must be accurate, have good
precision, and be a meaningful measure of pollu-
tants.

In this study we have evaluated the relationship
between COD and residue of the screened eflfluents
of four plants. Based on these correlations, a
monitoring system is suggested that enables the
results of two analyses to provide data on six
pollutant parameters.

EXPERIMENTAL

Identification and Definition of Terms

BOD {Biochemical oxygen demand): oxidation
by bacteria.

Manuscript accepted April 1976.

FISHERY bulletin! VOL. 74, NO. 4, 1976.

725

FISHERY BULLETIN: VOL, 74, NO. 4

COD {Chemical oxygen demand): oxidation by
potassium dichromate.

Residue: This term does not necessarily mean
solids, rather it is the results of or the substance
remaining from a separation process such as
filtering or drying. For example, if a solvent is
evaporated from oil, the resulting residue is a
liquid, not a solid.

TR {Total residue): is the weight of material
remaining from a sample of the original screened
effluent after overnight drying at 103°C.

FR {Filterable residue): is the residue of the
filtrate (GF/A glass filter) dried at 103°C. Drying
seafood effluents at 180°C (Environmental Pro-
tection Agency 1974) produced results that could
not be related to the TR and nonfilterable residue.

NFR {Nonfilterable residue): is the residue
remaining on the glass filter after drying at 103°C.
Since the three residue terms are related and
provided drying conditions are the same, NFR can
be determined indirectly, i.e., TR - FR.

5S {Settleable and floatable solids): This term has
caused considerable trouble to the industry and
researchers. By custom, the volume of the settled
portion in the Imhoff cone is measured and con-
sidered SS. However, this measurement does not
actually measure SS, because floatables are not
included in the reading. The term only has correct
meaning when SS is determined in milli-
grams/liter by difference: the NFR minus the
NFR of a sample taken from near the center of the
Imhoff cone after 1 h of settling.

Sus. Sol. {Suspended solids): are the particulate
matter suspended in the center of the Imhoff cone,
i.e., the NFR of that area.

TSS {Total suspended nonfilterable solids): This
term has also caused confusion. It means the dry
weight of all particulate matter (settleable, sus-
pended, floatable), i.e., the NFR. For both tech-
nical and grammatical reasons, NFR is the
preferred term.

O&G {Oil and grease): content was determined
by a method in which the precipitated, filtered-
solids material plus Celite^ (used as a precipitation
aid) is extracted directly under anhydrous condi-
tions, using 2-propanol and petroleum ether
(Collins 1976). This technique extracts all lipidlike
material, including carotenoids.

Protein: The nitrogen content was determined
by the macro-Kjeldahl method on 100- to 200-g

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

726

samples and expressed as protein by multiplying
N by 6.25 (Horwitz 1965:273).

Salt: Chloride was determined by the standard
AgNOs method and expressed as NaCl (Horwitz
1965:273).

Subscripts: In this paper, we use subscripts to
identify the particular portion of the sample
tested. For example, CODtr is the COD of the
screened waste effluent, and COD^r is the COD of
the FR, i.e., the filtrate, not the actual dried FR. If
no subscript is used, we are referring to the test in
general or to the test on the original screened
sample, i.e., COD is the same as COD-tr.

Industrially screened shrimp and crab effluents
were obtained from November 1973 through Feb-
ruary 1974 and from salmon effluents July
through September 1974. Since our purpose was to
compare data rather than characterize the level of
pollution in a plant, we took grab samples at
specific times during the production to get a useful
range of values. The following analyses were
made: COD^r, CODpR, TR, FR, NFR (i.e., TSS), SS,
protein, O&G, salt, and the COD of a sample from
the center of the Imhoff cone after 1 h of settling.

In conducting these analyses we used the meth-
ods of the Environmental Protection Agency
(1974), unless otherwise indicated. The particulate
matter in our samples of fishery waste was so high
that the filter clogged frequently before the entire
sample had been filtered. For this reason, sample
sizes were reduced, where necessary, to 25 ml.

The degree of pollutant in an effluent is affected
by the processes employed, species processed, and
the use of fresh or salt water in varying degrees
during processing. Mechanical shrimp peelers use
about 7 gallons of water per pound of shrimp. Salt
water from wells close to the shore or from the
ocean is sometimes used on the mechanical peelers.
The two main types of peelers vary in their
relative waste load. The Model A peeler peels raw
shrimp and generally has a higher waste load than
the Model PCA peeler that peels a steam-blanched
shrimp.

RESULTS

Study I— Shrimp: Analyses of effluents from a
shrimp plant processing with fresh water and
mechanical peelers (Model A).

Over a 10-day period in December 1973, eight
samples of waste effluents were taken from the
underflow of the Bauer Hydrasieve (tangential
screen, 0.04-inch) and analyzed (Table 1). Aver-
ages for COD by analysis are as follows:

COLLINS and TENNEY: FISHERY WASTE EFFLUENTS

a. COD, screened effluent

b. COD, center of Imhoff cone

c. COD, filtrate of NFR test

3,257 mg/liter
3,043 mg/liter
1,616 mg/liter

By calculation, the COD of the particulate
matter and its percentage contribution to the total
COD are:

COD of NFR (a - c) = 1,641 mg/liter 50.4%
COD of SS by weight

(a - b) = 214 mg/liter 6.6%
COD of Sus. Sol.

[(a - c) - (a - b)] = 1,426 mg/liter 43.8%

By analysis and calculation, data were also
obtained for the means of other residue tests:

By analysis:

d. Total residue (TR)

e. Filterable residue (FR)

f. Settleable solids (SS)

g. Nonfilterable residue (NFR)

2,381 mg/liter

1,577 mg/liter

5.6 ml/liter

769 mg/liter

By weight, the FR was 66.3% of the TR, but the
COD of the FR was only 49.6% of the total COD.
The NFR, however, contributed only 33.7% to the
TR by weight but contributed 50.4% as COD.

The standard deviations (SD) in Table 1 show
relatively large values in agreement with practical
experience. The higher average concentration and
lower SD for the NFR determined by difference
suggests that this is a better method for deter-
mining the concentration of NFR than is the
direct analysis.

Study 2— Shrimp: Analyses of effluents from a
shrimp plant processing with salt water and
mechanical peelers (Models A and PCA).

Ten samples of waste effluent were taken from
the underflow of the 0.7-mm Dorr-Oliver screen.
The individual results are given in Table 2, and the
average analytical data are as follows:

By calculation: NFR, i.e., (d
mg/liter.

e) or TR - FR = 804

a. COD, screened effluent

b. COD, center of Imhoff cone

c. COD, filtrate from NFR test

d. Settleable solids (SS)

e. Nonfilterable residue (NFR)

2,643 mg/liter

2,338 mg/liter

1,519 mg/liter

7.8 ml/liter

684 mg/liter

Table 1. -Analyses of shrimp waste effluents from a plant processing with fresh water and
mechanical peelers (Model A). All values are in milligrams per liter except SS in milliliters per
liter.

Screened

Filtrate

Center

Direct

Date of
sample

efflu

ent

(glass filter)
CODfR FR

difference

Imhoff

SS
Vol

analysis

Dec. 1973

COD,,

COD^,,

NFR

COD

NFR

3,070

2,370

1,492

1,640

1,578

730

2,856

8.0

880

3,364

2,660

2,040

1,990

1,324

670

3.212

7.0

840

3,068

2,290

1,580

1,540

1,488

750

2,912

3.0

656

2,516

1,970

1,240

1,350

1,276

620

2.312

7.0

796

3,353

2,280

1,405

1,360

1,948

920

2,956

11.0

892

2,660

1.790

1,080

1,010

1,580

780

2,418

2.0

120

2,962

2,040

1,588

1,420

1,374

620

2,841

2.5

660

5,065

3,650

2,500

2,310

2,565

1,340

4,836

4.0

1,308

Mean

3,257

2,381

1,616

1,577

1,642

804

3,043

5.6

769

789

578

455

407

428

238

781

3.2

332

Table 2.-Analyses of shrimp waste effluents from a plant processing with salt water and
mechanical peelers (Models A and PCA). All values are in milligrams per liter except SS in
milliliters per liter.

Screened

Filtrate

Center

Direct

effluent

(glass

filter)

difference

Imhoff

analysis

Date of

COD,,

Vol

sample

COD,,

COD^,,

NFR

COD

NFR

18 Nov. 1973

3,264

33,500

—

2,915

8.0

993

27 Nov.

4,050

—

2,690

—

1,360

—

3,883

3.0

—

7 Dec.

2,090

25,550

1,212

25,360

878

190

1,882

4.0

580

10 Dec.

3,161

34,090

1,729

33,780

1,432

310

2,935

9.0

1,212

2 Jan. 1974

3,143

27,730

1,733

—

1,410

—

2,849

8.0

1,008

9 Jan.

2,364

23,314

1,353

—

1,011

—

2,021

9.0

180

1 Feb.

2,890

23,300

1,363

23,100

1,527

200

2,487

10.0

616

4 Feb.

1,948

26,610

1,100

—

848

—

1,640

9.5

476

7 Feb.

2,442

23,940

1,659

—

783

—

1,806

9.5

896

15 Feb.

1,080

25,240

828

25,200

252

960

8.0

192

Mean

2,643

1,519

1,056

2,338

7.8

684

836

534

415

839

2.4

367

121

FISHERY BULLETIN: VOL. 74, NO. 4

By calculation, the COD of the particulate
matter and its percentage contribution to the total
COD are:

underflow of Dorr-Oliver 0.4-mm screen and an-
alyzed (Table 3). Average values by analysis are as
follows:

COD of NFR (a - c) = 1,124 mg/liter 42.5%

COD of SS (a - b) = 305 mg/liter 11.5%

COD of Sus. Sol.

[(a - c) - (a - b)] = 819 mg/liter 31.0%

Some figures were also collected on the concen-
tration of residues by direct analysis and are
included in the table to illustrate the problems
associated with monitoring plants that process
with salt water. Residue values were not deter-
mined by calculation because of the high and
variable salt content. It is questionable that
meaningful data for NFR can be obtained because
of errors that can occur when the salt values of
about 25,000 mg/liter are subtracted from the
mean TR values of about 27,000 mg/liter.

Study 3— Snow Crab: Analyses of effluents from a
plant processing both meats and sections in fresh
water.

Over a 2-wk period in February 1974, six sam-
ples of waste effluent from a plant processing snow
crab using fresh water were taken from the

a. COD, screened effluent

b. COD, center of Imhoff cone

c. COD, filtrate from NFR test

d. Total residue (TR)

e. Filterable residue (FR)

f. Settleable solids (SS)

g. Nonfilterable residue (NFR)

1,426 mg/liter
1,332 mg/liter

824 mg/liter

1,393 mg/liter

1,086 mg/liter

4.2 ml/liter

277 mg/liter

By calculation, the mean values for COD of the
particulate matter and its percentage contribution
to the total were:

COD of NFR (a - c) = 602 mg/liter 42.2%

COD of SS (a - b) = 94 mg/liter 6.6%
COD of Sus. Sol.

[(a - c) - (a - b)] = 503 mg/liter 42.2%

By calculation, the mean value for NFR is:
(d - e) = 307 mg/liter.

Study 4— Salmon: Analyses of effluents from a plant
processing canned salmon.

During the summer of 1974, ten samples of

Table 3.— Analyses of snow crab waste effluents from a plant processing both meats and
sections in fresh water. All values are in milligrams per liter e.xcept SS in milliliters per liter.

Screened

Filtrate

Center

Direct

Date of

effh

jent

{glass filter)

difference

Imhoff

analysis

sample

Feb. 1974

COD,,

COD^,,

NFR

COD

Vol

NFR

680

880

506

770

174

110

599

1.3

126

888

960

650

850

238

110

868

0.5

1,056

1,230

746

1,030

310

200

974

5.0

143

1,560

1,590

870

1,280

690

310

1,408

4.0

462

1,988

1,900

1,077

1,500

911

400

1,889

7.5

540

2,383

1,800

1,093

—

1,290

—

2,254

7.0

348

Mean

1,426

1,393

824

1,086

602

226

1,332

4.2

277

668

433

235

303

442

127

640

2.9

202

Table 4.- Analyses of salmon waste effluents from a plant processing canned salmon. All values are in milligrams

per liter.

Date of

Screened effk

lent

Flit

rate — glass filte

By difference

sample

Salmon

1974

COD,,

Protein

O&G

Salt

CODpp

Protein

Salt

COD^,,

NFR

30 June

Red

5,716

3,695

2,197

1,190

574

1,365

1,513

1,044

545

4,351

2,182

7 July

Red

2,908

2,076

1,500

330

373

1,212

1,135

656

273

1,696

941

8 July

Red

4,069

2,368

1,453

918

253

1,131

1,078

744

247

2,938

1,290

11 July

Chum

2,070

1,125

1,179

308

—

797

350

531

453

1,273

775

14 July

Chum

6,294

4,450

2,980

—

728

2,687

2,560

1,775

436

3,607

1,890

17 July

Pink

9,513

7,102

—

596

4,020

3,655

3,346

465

5,493

3,447

30 July

Chum

9,101

6,315

3,346

1,407

397

3,420

2,813

—

5,681

3,502

13 Aug.

Pink

5,236

3,595

2,378

845

493

1,462

1,465

1,009

459

3,774

2,130

14 Aug.

Pink

2,647

2,148

1,518

226

344

1,822

1,570

1,168

292

825

578

22 Aug.

Mixed

6,219

4,874

3,263

924

642

2,615

2,722

1,938

556

3,604

2,152

Mean

5,377

3,775

2,201

769

489

2,053

1,886

1,357

414

3,324

1,889

2,557

1,937

840

437

157

1,078

1,007

885

115

1.664

1,026

728

COLLINS and TENNEY: FISHERY WASTE EFFLUENTS

salmon cannery waste effluent were taken from
the underflow of a Bauer screen (0.03-inch) (Table
4). The average values by analysis and calculation
are as follows:

COD, screened effluent

5,377 mg/liter

COD, filtrate of NFR test

2,053 mg/liter

COD,ofNFR(a-b)

3,324 mg/liter

Total residue

3,775 mg/liter

Filterable residue

1,886 mg/liter

Nonfilterable residue (d - e)

1,889 mg/liter

O&G, screened effluent

769 mg/liter

Protein, screened effluent

2,201 mg/liter

Salt, screened effluent

489 mg/liter

The NFR is 50% of the TR, but the COD of the
NFR is 62% of the total COD.

DISCUSSION

The following discussion is concerned with
monitoring parameters previously suggested or
currently in effect under EPA effluent limitations
for seafood processing and with the suggestion of
a more precise and simpler monitoring system.
The present EPA requirements, however, for use
of alternative analytical methods must be con-
sidered. Under EPA rules (Title 40 "Code of
Federal Regulations" Parts 136.4 and 136.5), any
person wishing to use alternative analytical
methods for the parameters listed must follow
variance procedures specified under the NPDES
permit system.

Current permits require monitoring for SS,
COD (i.e., CODtr), TSS (i.e., NFR), O&G, flow, and
pH. SS is imprecise and contributes so little to the
pollution load in seafood processing that it has
relatively little value as a measure of pollution,
although it has merit as a check on the efficiency of
screen operation. As discussed later, total COD can
be determined more accurately in an indirect
manner. The O&G analysis is difficult to do, and
this value, too, can be obtained more accurately
through calculation. Data in this paper suggest
that the indirect analysis for NFR (i.e., TSS) was
more accurate than the direct method. The FR is
an important parameter because this fraction
contributed about 50% to the total COD or TR and
will need to be considered in the design of future
treatment systems.

To develop an improved monitoring system, we
plotted the COD and residue data of Table 1 to
illustrate the correlation between the COD of the

residue and the concentration of the residue
(Figure 1). The regression lines and equations
were determined by the method of least squares.
The TR and FR regression lines were obtained
through direct analyses, and the NFR line was
obtained by difference. The maximum deviation of
any COD value from the regression line was 260
mg/liter. This is slightly less than the possible
error of the analytical method ( ± 8%) (Moore et al.
1949). On the average, the individual values were
within 107 mg COD of the regression line.

The correlations shown in Figure 1 can be used
to calculate COD and residue values. In the fol-
lowing, the first three equations are the regression
lines of Figure 1 and the next three are derived
equations to solve for residue rather than for COD.
Of course, these equations are valid only for this
group of data and for this particular plant. If the
TR and CODpK are determined by analysis, the
other values can be derived from the equations or
the regression line and from the expression
TR = FR -h NFR.

CODtr
CODpR

CODkpk

1.32 TR + 113
1.08 FR - 96
1.78 NFR + 210

(1)
(2)
(3)

5.000

4,000 -

1000

2,000

1,000

1000

2000 3,000

RESIDUE (mg/l)

4,000

Figure l.-Relationship between the COD of the residue and the
concentration of the residue from shrimp processed on Model A
peelers using fresh water.

729

FISHERY BULLETIN: VOL. 74, NO. 4

TR = 0.76 CODtr - 86

FR = 0.93 CODpR - 89

NFR = 0.56 CODj,FR - 118

(4)
(5)
(6)

When salt water was used in processing, such as
in the second plant study (Table 2), the residue
values included salt. Since salt values were not
determined, COD and residue data were not cor-
related for this plant.

In the third plant study of snow crab effluent,
the data (Table 3) were plotted similarly to the
shrimp data (Figure 2). The basic equations for
snow crab can also be used to calculate from two
analyses the other COD or residue values. The
equations are listed in Figure 2.

Data for the fourth plant study of salmon-waste
effluents (Table 4) were also plotted, and the
regression lines and equations were similarly
determined (Figure 3). The regression lines for
salmon are less precise because of the variable salt
content of the effluent and the high levels of COD
and residue. Salt varied because of the erratic
operation of the salmon egg-processing room.
These regression lines (salmon) should not be used
to calculate or interpolate COD or residue values
unless a check is first made on salt content. If salt
content of the effluent is about normal (500
mg/liter), the calculation is valid since these
equations are derived from data with a high
standard deviation for salt. A check is made to
ensure that the level is not 1 or 2% as it could be if a
brine tank were dumped. A routine composite

3,000

\ 2,000
01
E

o
O

2.65NFR - 1 26

1,000 2,000

RESIDUE (mg/l)

Figure 2.-Relationship between the COD of the residue and the
concentration of residue from the processing of snow crab meats
and sections.

730

9,000 '

epoo-

zooo

<lOOO

-i 5,000

<J 4.000

JflOO-

2.000

1,000 •

coo,. = 1.30TR + 4 56

COD^„= 1.59NFR+ 330

1000 2,000 3,000 4.000 5,000 40OO 7000
RESIDUE (mg/ l)

Figure 3.-Relationship between the COD of the residue and the
concentration of the residue for canned salmon processing.

sampling program for the plant, of course, would
reduce salt variation.

A SIMPLIFIED MONITORING SYSTEM

The data of the first plant study (Table 1) and
the six equations listed earlier may be used to
illustrate how a simplified monitoring system can
be set up for a particular plant.

Since COD is difficult to determine on the
original effluent (particulate matter causes dilu-
tion problems) and impractical to determine on a
solid sample, COD should be determined on the
filterable residue sample before drying. Equation
(5) is then used to calculate FR in milligrams per
liter. It is not necessary to actually finish the FR
test. The next analysis most logically should be the
total residue test. It is an easy test to do and is
accurate. Equation (1) is used to calculate the COD
of the TR, and the previously calculated FR is
subtracted from TR to give the NFR in milligrams
per liter. Equation (3) is then used to calculate the
COD of the NFR. Thus, two analyses plus several
calculations give three COD and three residue
values.

The two analyses recommended (CODpp and
TR) are logically the most accurate of the six

COLLINS and TENNEY: FISHERY WASTE EFFLUENTS

possible, thus the other calculated values that are
based on an ideal regression line should be more
valid than those obtainable by direct analysis.
Although this system may suggest doing the FR
rather than the CODpR, we believe that one direct
analysis for COD is desirable, since the effect of
oxygen demand on the receiving water is an
important parameter of a monitoring program.
Although O&G were not specifically considered
except for salmon, for which we had limited data,
the COD and residue data imply that O&G are
related and that a regression line could be cal-
culated.

In conclusion, it appears that in-plant monitor-
ing for CODpK and TR and the application of
proper correlation factors and equations
previously determined for the plant effluent will
give reportable data on CODxr, COD^fr. CODpR,
TR, FR, and NFR. The suggested analyses can be
done at reasonable cost with simple equipment,
are capable of good precision and accuracy, and
can be conducted by quality assurance personnel
in the fishing industry. We suggest, recognizing
the limitations of our data and obvious and known
differences between processing plants and
processing methods, that if regression lines or

correlations similar to those given in this paper
were determined, the resulting monitoring system
would be simpler and more accurate than that
currently in use.

In a subsequent paper, we will report regression
data for protein and O&G similar to that suggest-
ed in this paper and a method using a simultan-
eous equation to calculate protein and O&G from
TR and COD data.

LITERATURE CITED

Colons, J.

1976. Oil and grease: A proposed analytical method for
fishery waste effluents. Fish. Bull., U.S. 74:
Environmental Protection Agency.

1974. Methods for chemical analysis of water and wastes.
Environ. Prot. Agency, OflF. Tech. Transfer, Meth. Dev.
Qual. Assurance Res. Lab., Natl. Environ. Res. Cent. 298 p.

HoRWiTZ, W. (chairman and editor).

1965. Official methods of analysis of the Association of
Official Agricultural Chemists. 10th ed. Association of
Official Agricultural Chemists, Wash., D.C., xx+ 957 p.
Moore, W. A., R. C. Kroner, and C. C. Ruchhoft.

1949. Dichromate reflux method for determination of ox-
ygen consumed. Anal. Chem. 21:953-957.
Pojasek, R. B.

1975. NPDES permits and water analyses. Environ. Sci.
Technol. 9:320-324.

731

POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A
OFF SOUTHERN CALIFORNIA

Edward Brinton^

ABSTRACT

Euphausia pacifica was observed with respect to reproduction, growth and development of cohorts, and
successions in population structure and biomass during 4 yr, 1953-56. The southern California eddy and
its upwelling regime serve as a reproduction refuge for a warm-temperate population of this
euphausiid. Three size classes spawn there during a year-the largest in April-June, an intermediate in
June-February, and small, newly mature females usually in August-January. There were year-to-year
differences.

The largest densities of larvae were observed about a month after egg peaks (one survey later) or
appeared coincident with them. In 1953 there was strong spring recruitment, abruptly subsiding with
an early decline in upwelling-the index of environmental enrichment used. During 1954 only one
substantial cohort was recorded, in June at the height of a poor upwelling season. In 1955 repeated
spawning occurred during the long upwelling season, but recruitment after July was poor. The year of
most intense upwelling, 1956, yielded three strong cohorts— the last, July-October, being exceptionally
strong. Smallest larvae were usually in 12°-16°C waters. Ripe females were concentrated at high
densities at these same temperatures during August-March but were distributed over a broader range
at 10.5°-19°C during April-July.

Growth was estimated to be about 3 mm body length per month, slowing during September-January
or after about 17 mm. Females appeared to grow slower in breeding seasons. Maturity can be at 11 mm,
but reproduction is not general until 15-16 mm. Here, maximum size was 21 mm after about 7 mo for
early-year recruits and a year for summer recruits. Survival rates appeared higher in the latter. Growth
rates were similar to those reported for E. pacifica off Oregon and higher than in the subarctic Pacific.
Survivorship was lowest for furcilia larvae, increased in juvenile and young adult phases, then
decreased after reproduction became regular. Slowed growth and increased survivorship at life
interphases appeared to cause regular frequency and biomass maxima at lengths of 7, 10-12, and 15
mm. Sex ratio favored females. Males apparently accomplished multiple fertilizations.

Euphausia pacifica Hansen is a temperate North
Pacific euphausiid crustacean, composing a sub-
stantial part of the zooplankton of the North
Pacific Drift, lat. 40°-50°N, and ranging south-
ward along the coast of North America as far as
lat. 25°N (Brinton 1962a). In the cooler part of the
California Current, it occurs in association with
the euphausiids Nematoscelis difficilis and Thy-
sanoessa gregaria. Depth ranges of the three
species overlap daily as E. pacifica and A'', difiicilis
engage in distinctive vertical migrations while T.
gregaria does not migrate (Brinton 1967a). Hor-
izontal ranges are sufficiently similar so that
these species, together with E. gibboides, were
considered the euphausiids of a California Cur-
rent-Transition Zone plankton assemblage
(Brinton 1962a).

Euphausia pacifica performs extensive vertical
migrations. Off California it lives at daytime
depths of 200-400 m, entering the surface layer at

'Scripps Institution of Oceanography, La Jolla, CA 98093.

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

?-5? ' ^i Z-

night. It is an omnivore (Lasker 1966) and pos-
sesses thoracic food-gathering limbs which are
nearly uniform in length and in setation of the
filtering screens.

Euphausia pacifica is usually the most abun-
dant euphausiid. Its maximum densities are often
centered relatively near to the coast of California.
The low-latitude part of the population of E.
pacifica is the object of this study. Aspects of its
life history have been observed in the more
typically temperate regime to the north of lat.
40°N (Nemoto 1957; Ponomareva 1963; Smiles and
Pearcy 1971) where environmental characteristics
show stronger seasonality than to the south. The
extent to which the downstream portion of this
distributional range is maintained by local
processes has not been previously investigated.

The study was organized in relation to existing
knowledge of the physical-chemical characteris-
tics of the area and of the species distributions.

Surveys of the region of the California Current
since 1949 have provided a reservoir of hydro-
graphic data and plankton samples that lend

733

FISHERY BULLETIN: VOL. 74, NO. 4

themselves to time-series studies of biological and
environmental developments. The CalCOFI
(California Cooperative Oceanic Fisheries Inves-
tigations) Atlas series (Numbers 1-24) presented
varied material, including euphausiid distribu-
tions derived from the program. Charts of dis-
tributions of E. pacifica based on the data that are
the subject of the present analysis are included in
Brinton and Wyllie (in press). Smith (1971) de-
scribed the distribution of zooplankton biomass.

Description of the Study Area

The southern California eddy is the southern-
most area in which E. pacifica is still both abun-
dant (commonly 10-1,000 individuals beneath 1
m^ of sea = 10-1,000 mg wet weight) and domi-
nant among the larger zooplankters (Brinton 1967a,
b). The eddy may be considered bounded on the
north by Point Conception, lat. 34°N, and on the
south by about lat. 30°N. Its east-west extent is
about 250 km; beyond its western limits, flow is
consistently from the north and apparently con-
tributes relatively little water and biota to the
eddy.

The sluggish circulation off southern California
evidently permits substantial autonomy for the
resident populations. The currents are commonly
5-10 cm/s and rarely as much as 25 cm/s, both at
the surface and at 200 m depth (Wyllie 1966).
Direction of flow sometimes reverses between
these two levels. These are, respectively, the night
and day depth levels occupied by vertically mi-
grating E. pacifica (juvenile and adult) in the area;
larvae remain near the surface day and night
(Brinton 1967a).

Circulation of the eddy is cyclonic. Within it,
therefore, there is upward transport of enriched
water. The center of the eddy (no surface flow) is,
on the average, near San Nicholas Island (lat.
33°15'N, long. 119°30'W), 100 km off the midpoint
of the southern California coast. The study area
was centered here. Farther east, mean flow is
northwesterly along the coast. To the west, flow is
southeasterly, angling toward the coast near lat.
30°N.

About 150 km south of Point Conception, mean
geostrophic flow approaches 135°, averaging 10
cm/s. A parcel of water entering the eddy from the
northwest would, at that speed, take 100 days
to move around the eddy back to Point Conception,
flow permitting. Average velocities within the
eddy are much less. Places where substantial

734

advection takes place across margins of the area
are determinable from the flow diagrams in a
relative sense. Northerly surface flow into and out
of the area is characteristic of winter months
when the Davidson Countercurrent is developed.
Southerly flow into or through the western part of
the area is usually strongest in April-July. The
eddy persisted in almost all of the months studied.

Upwelling enhances the temperate character of
the area during spring and summer, usually in-
tensifying during April-June (Bakun 1973) when
annual temperature minima are usually found. It
is responsible for much of the local nutrient
enrichment (Reid et al. 1958). Seasonal periodicity
is evident when water temperature is averaged for
the area of the eddy as a whole: August-October is
generally warmest and January-April coolest
(Anonymous 1963). The area contains a scatter of
islands which provide substantial shoal grounds,
regarded off Oregon to be areas best suited for E.
pacifica (Smiles and Pearcy 1971). Such islands
also provide topography for the formation of
downstream eddies which are enrichment centers
(Uda and Ishino 1958). They also serve as centers
of upwelling. Here, upwelling is less dependent on
the direction of the wind than on its intensity.
However, the coast from Point Conception
eastward remains the main focus of upwelling
during the period of prevailing northwest
winds, February-June. According to the indices
derived by Bakun from extrapolated atmospheric
pressure gradients at the sea surface, upwelling
off southern California is the most intense to be
found in the California Current.

For this initial life-history study, the period
chosen (1953-56) was one of generally stable
oceanic climate and hydrographic conditions,
compared with the years immediately following,
which included times of more extreme fluctuations
in temperature and flow characteristics. During 2
of the 4 yr, 1955 and 1956, upwelling was inferred
by Bakun (1973) to be more intense than the
1946-71 mean; however, during 1954 it was less,
and during 1953 upwelling commenced early but
barely achieved the June peak of mean intensity
and was greatly diminished in the summer
months.

Thus it was anticipated that the study period
would yield observations of low annual variability
in the population of E. pacifica, thereby providing
a baseline against which eventually to measure
events in years of known extremes in ocean
climate, e.g., 1957-59 (Brinton 1960).

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

Previous Investigations

In addition to the observations on the life
history of E. pacifica (Nemoto 1957; Ponomareva
1963; Smiles and Pearcy 1971), aspects of the
energy budget and physiology of this species have
been studied. Lasker (1964, 1966) measured
moulting frequency, feeding rates, respiration
and carbon utilization by specimens maintained in
the laboratory, and observed growth rate in
juveniles and adults. Fowler et al. (1971) con-
sidered effects of temperature and size on moult-
ing. Small et al. (1966) measured respiration at
different temperatures and discussed energy flow,
while Small (1967) further examined energy flow.
Paranjape (1967) made observations on moulting
and respiration. Aspects of depth-habitat and
pressure in relation to respiration were considered
by Small and Hebard (1967), Pearcy and Small
(1968), and Childress (1971). Gilfillan (1972) studied
oxygen uptake in relation to laboratory controlled
temperatures and salinities.

Total oocytes in a large female were counted by
Ponomareva (1963). Clutch size estimates and the
vertical distribution of different age groups were
given in Brinton (1962b and 1967a, respectively).

limitations of the Study

Understanding the population biology of an
oceanic species depends in large part upon the
extent to which a representative part of the
population can be representatively sampled. In the
planktonic environment, currents not only tend to
transport the organisms across an observer's
horizon, but also cause relative horizontal dis-
placement of life stages because, in many species,
the various stages of development live at different
depths and experience different horizontal trans-
port. This is true of euphausiids. Species under-
going both ontogenetic and daily vertical migra-
tions, such as E. pacifica, are further subject to
differential horizontal transport. Thus, water
movement is a variable which complicates any
plan for temporal continuity in sampling a
population. The area covered and the time spent in
carrying out an assessment of a population does
not need to be great if the waters are restricted
geographically and if growth and development of
the population is measurable between successive
assessments. Clearly, a gyre of circulation, such as
the eddy lying off southern California, may be
expected to harbor elements of a population that

persists locally. This study area has proven prac-
tical in size according to the logistics of CalCOFI.

MATERIALS AND METHODS

Samples were obtained by oblique tows, 0-140 m
depth (except where the water was shallower),
using the CalCOFI standard net, 1-m mouth
diameter and 0.55-mm mesh width (Ahlstrom
1948). The mesh width of the cod end and of a
40-cm section in front of it was 0.25 mm. The
volume of water strained through a net was
determined with a TSK (Tsurumi-Seiki Kosaku-
sho) flowmeter.- Most volumes were in the range
of 300-400 m^ The net was towed at about 75
cm/s. The 1953-56 cruises provided month-to-
month data, including more frequent surveys off
southern California in late 1955 (four in Sep-
tember, three in November). Station positions and
collecting data together with displacement
volumes of the plankton samples are from annual
listings of CalCOFI plankton sampling 1953-56
(South Pacific Fishery Investigations 1954, 1955,
1956; Thrailkill 1957).

Specimens smaller than 3 mm in length are able
to pass through the meshes of the net and there-
fore were not representatively sampled. Smaller
specimens (2 mm) are nevertheless retained by the
fine meshes of the cod end of the net and counts of
these are included as indicative of the presence of
the small calyptopis larvae. Free floating eggs of
E. pacifica are not retained by this net. Estimates
of egg production are derived from examination of
the ripe females sampled, as described in the
discussion of fecundity below.

A total of 819 samples from 48 cruises, 5301
(January 1953) through 5612 (December 1956),
were examined (Figure 4d). Only nighttime sam-
ples were used since juveniles and adults are not
representatively sampled in the daytime, owing to
vertical migration and avoidance of the net
(Brinton 1967a). Between 7 and 43 nighttime
samples were collected in the study area during
each cruise. "Night" was considered to be the
period from 1 h after sunset to 1 h before sunrise.
A few sunrise and sunset samples were analysed if
they were collected under overcast skies. A sample
marginal to, but outside of, the area was studied
when such a sample was from a locality nearer to
the closest boundary of the area than any of the

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

735

FISHERY BULLETIN: VOL. 74, NO. 4

sampled localities within the area. Thus, some
samples from station lines 77 (northern) and 97
(southern), or designated 80.80 (western), were
occasionally used.

The samples were examined in the following
manner. An aliquot containing 100-200 E. pacifica
was counted; the specimens were measured to the
nearest millimeter of body length (tip of frontal
plate to tip of telson)]_adults were sexed; and the
degree to which the reproductive products were
developed was recorded. If, for adults (specimens
>10.5 mm in length), the initial aliquot contained
fewer than three specimens of any particular
length, a second aliquot of equal size was ex-
amined for specimens of that size or larger. In this
way, increasingly large fractions of the sample
were examined for specimens of those length
intervals which were progressively determined to
be fewest in the sample. This procedure made it
possible to count the rarer, large specimens with a
degree of accuracy comparable with that to which
the consistently more abundant small specimens
were counted. Usually, the entire sample was
examined for specimens of more than 14-mm body
length. This procedure was facilitated by the use
of the Folsom plankton splitter which, through
successive splitting operations, provides aliquots
of V2, Vi, Vs, . . . \/n. All counts were standardized
for 1,000 m-^ of water strained by the net.

After standardization, the counts for a sample
(station) were weighted according to the propor-
tion of the survey area represented by that sta-
tion. When the nearest area surrounding the
station was equal to a 65 x 65 km square (a usual
spacing for CalCOFI stations), the weighting
factor was 1.0. When areas represented by sta-
tions were greater or less than 65x65 km,
weighting factors were proportionally greater or
less than 1.0. The study area was equal to 19 65 x 65
km squares. Therefore the sum of the weighted
abundances (for each size of E. pacifica) was
divided by 19, providing a mean standardized
abundance for the area for the given survey. (The
night stations were not at the same localities on
each cruise, though tracks followed by the vessels
were generally repeated. Furthermore, as is to be
expected, clusters of day stations tend to alternate
with clusters of night stations. Unsampled parts
of the area are expected to be better represented
by samples from stations nearest to that unsam-
pled part than by samples from more distant
localities.)

Females were classified as 1) with ripe eggs

736

(Mauchline's [1968] egg phase IV) and with at-
tached spermatophore, 2) with ripe eggs and no
spermatophore, 3) with ripening eggs (approx.
Mauchline's phase II), or 4) ovary weakly
developed. Adult males were categorized as 1) with
ripe spermatophores, either protruding or inter-
nal, or 2) without ripe spermatophores.

Biomass was calculated using abundance at each
body length (1-mm increment). Values are in
terms of wet displacement volume (wet weight) of
E. pacifica, given per body length increment in
Miller (1966). The following conversion factors
from Lasker (1966) may be applied:

Dry weight = 17.2% of wet weight
Carbon = 42 ± 1.7% of dry weight

Carbon = 7.2% of wet weight

RESULTS

Southern California Eddy in Relation
to the Rest of the California Current

October 1955 data (cruise 5510) illustrated char-
acteristics of flow and temperature in the cur-
rent, and occurrences of E. pacifica larvae (Figure
la-c). These were general to fall-winter 1953-56
and placed the southern California area in broader
geographical perspective. At that time the land-
ward portion of the current, slow and cool, sup-
ported five centers of recruitment of E. pacifica
(Figure Ic): 1) off San Francisco, probably related
to the September peak off Oregon obsrved by
Smiles and Pearcy (1971), 2) north of Point Con-
ception, 3) southern California, 4) Point Colnett
(lat. 31°N), and 5) Point Canoas (lat. 29°N). The
three centers off California were then associated
with current reversals while the two centers off
Baja California were places w^here upwelling was
conspicuous. A Punta Eugenia center, farther
south (lat. 27°-28°N), usually supports E. pacifica
earlier, during the local peak of spring coastal
upwelling, May-June.

Direction and intensity of coastal flow tends to
vary on a seasonal basis. During cruise 5510 and
through ensuing fall and winter months, coastal
currents off California provided means of north-
erly transport for portions of southern popula-
tions. During spring and summer, intensified
southerly currents off northern California are
expected to bring elements of the northern
population into the southern California area via
the offshore route west of Point Conception,
diverting shoreward near lat. 32°N.

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

1 1 r

10 METER TEMPERATURE

CONTOUR INTERVAL I.O°C

J I I L

120° 115°

Euphausia pacifica

AVERAGE FOR ALL STATIONS
LINES 60-123. CRUISE 5510

Cc 1000

500-

E AVERAGE FOR SOUTHERN AREA ONLY.
LINES 80-93. CRUISE 5510

October 1955 ( Cruise 5510)

2 4 6 8 10 12 14 16 18 20 22
BODY LENGTH (mm)

Figure l.-October 1955 data (cruise 5510). a, Surface streamlines showing areas of current reversals off California and b, 10-m
temperatures indicating upwelling centers along Baja California, both associated with c, aggregations of Euphausia pacifica larvae.
Length-frequency distributions of E. pacifix:a are averages for d, all nighttime stations and e, stations within southern California area.

The length-frequency (L-F) diagram for E.
pacifica in the California Current as a whole
(cruise 5510) shows four modes present in the
overall population: 3-4 mm (larvae), 7-8 mm, 10-12
mm, and 15-16 mm (Figure Id). Time progressions
in such modes are used below to estimate popula-

tion development, including growth and mortality.
The southern California part of the population is
characterized by small (10-12 mm) and large adults
(15-16 mm). It will be shown below that each of
these two October 1955 modes is distinguishable
within a month-to-month L-F sequence of cohort

737

FISHERY BULLETIN: VOL. 74. NO. 4

development; the 10-12 mm group, most charac-
teristic of the southern California area in Figure le,
was of a cohort which remained locally dominant
from its inception in July 1955 until January 1956.

The 7-8, 9-12, and 15-16 mm modes are described
below as being common to E. pacifica because they
are at body lengths at which life-phase changes
and growth slows; therefore frequencies of those
sizes increase, particularly during fall-winter
periods of reduced food supply.

L-F curves for individual stations show the clear
7-8 mm mode along an "offshore" north-south
track (Figure 2a, c) in the axis of the fastest part
of the current (Figure la). It dominates the 9-10
mm mode as the transect, following the steam-
lines, angles shoreward along the southern edge of
the southern California area, until lat. 35.5°N
(station 97.50) where the 7-8 mm mode becomes
inconspicuous and the 9-10 mm mode assumes
dominance. Thus offshore, where southerly
population transport would be expected on the
basis of the observed current, dissipation of the
L-F characteristic of the northern population
takes place along the western limit of the study
area. This is considered evidence that such trans-
port then contributed little to the area's popula-
tion, relative to more nearshore, local
contributions.

Individual stations along a "nearshore" north-
south transect (Figure 2b) showed a dense heter-
ogeneous population of E. pacifica off San
Francisco (station 63.55, lat. 37°N). Off central
California (stations 70.55, 77.55), 7-8 mm juveniles
became conspicuous (cf. Figure 3). Farther south,
particularly in the southern California area
(stations 83.51-90.28), 7-10 mm individuals were
much reduced in numbers, while the frequency of
the 11-12 mm size increased, appearing as a clear
L-F mode. In October, larvae were few off north-
ernmost Baja California where oceanic water
typically moves eastware compressing shoreward
the faunistic connection of the southern California
area to more southern upwelling centers. To the
south along the Mexican coast, the 11-12 mm mode
characteristic of the study area reappeared, coin-
cident with areas of production of larvae. Farthest
south (off Punta Eugenia; stations 120.45, 123.40),
modes were at 9-10 mm and at 3-mm larvae. These
9-10 mm specimens may be poorly nourished
individuals, corresponding in age to 11-12 mm
individuals occupying the area immediately to the
north— an area which appears relatively fertile
with respect to production of larvae. The same

738

relationship was observed locally off northernmost
Baja California; there the population having a 9-10
mm mode included few larvae (Figure Ic) and
occupied an easterly incursion of oceanic water
(Figure la), being bounded on the north and south
by cooler and presumably more fertile areas in
which both 11-12 mm and larval modes were again
conspicuous.

At this time (October 1955) the range of E.
pacifica terminated near Punta Eugenia, but it
can extend to lat. 23°S (Brinton 1967b). These
far downstream parts of the population appear
reproductive, but to the south of southern Califor-
nia they are impermanent (Brinton 1967b, 1973).
Mature or maturing individuals are expected to be
intermittently injected from the north, par-
ticularly during the March-June period when
southerly flow is intensified. These individuals
may find local places of refuge in cool, slowly
moving, productive coastal waters from Point
Conception southward in association with up-
welling centers. The southern California eddy is
the largest such refuge, serving also as a major
population center which has both coastal and
oceanic dimensions.

Spawning and Recruitment

Spawning intensity was estimated indirectly
since free-floating eggs were not sampled.
Females bearing ripe eggs provided a means of
estimating incipient spawning. All females hav-
ing an attached spermatophore also carried ripe
eggs in the ovary. From the several thousands of
these counted, 373 of different body lengths were
examined with respect to number of ripe eggs
carried. The relationship between body length and
mean number of ripe eggs was linear between 11
and 20 mm length (means were encompassed by
95^ confidence limits of regression line), with the
mean number of eggs extending from 20 to 212
across this range (Figure 3). Disproportionately
small numbers of eggs were observed in the
largest (>20 mm) females. Mean values for each
body length were applied to the numbers of each
length of ripe female counted in the plankton
samples to estimate the spawning potential for
each sampling period. These are underestimates
since, for 60% of the surveys, the predicted values
are not high enough to have produced the density
of larvae found at the time of the next
surv^ey— even presuming only 50% mortality
between surveys (Figure 4c). Evidently some eggs

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

Q OFFSHORE STATIONS

3000—,

STA 60 80

E
o
o
o

UJ
Q.

500 — )

100 —

10 —

8090

I I I 1 I

9090

I 1 I I I ' 1 ' M I

500 -n

100 —

/\ 97 50

10 —

/ \

[3-4 1 7-8 1 II -I2I 15-16 I19-20I
-2 5-6 9-0 13-W 17-B 21-22

BODY LENGTH (mm)

1^ NEARSHORE STATIONS

5000—1

Figure 2. -Length-frequency distribution of
Euphaiisia pacifica along north-south axes of
California Current (cruise 5510). a, Approx-
imately 250 km offshore; b, <100 km offshore;
c, positions of stations.

100 —

100 -

100-

STA 63 55

7055

I I I 1 1

77 55

,83 51

87 45

ql I M

90 28

100-

10-

1 ' I ' 1 I I ' I

93 40

I'll

10 —

I I I I

113.35

I I I

11735

I I I I

I 1 I

12045

I I I I

123.40

BODY LENGTH (mm)

739

FISHERY BULLETIN: VOL. 74, NO. 4

en
•a

■0
«

d
U

300

200

lOO —

12 14 16 18 20

Body Length (mm) Of Ovigerous Females

Figure 3.-Number of mature eggs in ripe spermatophore-bear-
ing Euphausia pacifica in relation to body length. Numbers of
individuals examined are indicated.

found to be immature at time of counting, either
in ripe or other females, mature in time to con-
tribute to the monthly spawn. The egg estimates
are therefore regarded as only relative, month to
month.

The production of eggs and larvae in each year
(Figure 4c) was considered in relation to four
parameters: 1) annual upwelling cycle in the
southern California area inferred from atmo-
spheric parameters (Figure 4a) and from min-
imum water temperatures (Figure 4b), 2) size
structure of the spawning stock (Figure 4d), 3)
zooplankton biomass (Figure 5a), and 4) E. pacifica
biomass (Figure 5b).

1953

Upwelling began early (February, cruise 5302)
with above-average intensity, accompanied by
spawning in February and April. The February

740

spawn, mainly by females of medium length
(12.6-16.5 mm), led to discernable recruitment of
larvae in March. The April spawn, mainly by large
females (16.6-21.5 mm) led to the year's maximum
recruitment in May-June. Upwelling peaked in
June, and diminished to an unseasonably low
intensity thereafter (Figure 4a), accompanied by
local variability in water temperature through
October (Figure 4b).

Substantial egg production during June-
August, by medium-sized and small (10.6-12.5 mm)
spawners, led to less recruitment than in April
when spawning was of similar intensity. April was
the start of the general spring zooplankton bloom
(Figure 5a), presumably a response to the greater
availability of phytoplankton food in the spring.
Spawning diminished after August although lar-
vae were evident in September and November.
Small females became predominant after Sep-
tember when they became important contributors
to the production of eggs.

These estimates of relative spawning are sup-
ported by a consistent relationship of egg peaks to
larva peaks. Three of the four egg peaks in 1953
were followed by larva peaks a month later. Under
conditions of laboratory hatching and rearing,
euphausiids live as larvae for about 29 days
(Gopalakrishnan 1973).

1954

Upwelling commenced in March (Figure 4a), a
month later than in 1953. Local temperature
minima, however, showed that this process was
not obvious until April (Figure 4b). By both
criteria, spring upwelling in 1954 was the least
intense to be observed during 1953-56. (According
to Bakun (1973), it was the least observed during
1953-71, though substantially greater than during
1947-52.) Production of eggs was initiated in March,
evidently by a stock of large females derived from
the September 1953 recruitment (see sections on
growth and survival below, and Figure 9).
Recruitment became intense only during June-
July, associated with the one peak in spawning
observed during 1954.

1955

As in 1954, upwelling started in March
(following Bakun 1973, Figure 4a) or in April
(using temperature minima. Figure 4b). There
was a gradual increase in egg production begin-

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

NO OF
SAMPLES

CRUISE NUMBER (YEAR & MONTH )

I I I I I I I I I I I I I I I I I I I I I I I I I I III I III! I II I II I III I

12 17 13 14 16 21 17 17 7 12 10 14 24 24 18 21 19 15 16 16 17 18 12 14 14 13 20 18 18 774 14 7616 14 II 15 1220 15 21

I I I
19 18 20

Figure 4.-a, Inferred monthly index of upwelling intensity per 100 m of southern California coastline, 1953-56 (from Bakun 1973). b.
Temperature range and mean, by cruise, in study area, c. Estimated densities of ripe eggs and <;4.5 mm larvae of Euphausia pacifica in
area, d, Densities of ripe females, three body-length groups. Number of samples examined are indicated by cruise.

ning in December 1954, and the first significant
recruitment was in February (cruise 5502). This
increase in spawning continued through March,
but recruitment did not increase markedly until
May, following an April egg maximum. Thereafter

egg production peaked in alternate months, June
(the annual upwelling maximum), September, and
November-but recruitment was generally low
(<2,500 larvae/ 1,000 m^ in the area) except during
May and July. July yielded the year's peak in

741

1000

FISHERY BULLETIN: VOL. 74. NO. 4

1000

5301 03 05 07 09 11 5401 03 05 07 10 5501 03 05 07 09 1011125601 03 04 0607 !0 II 12

Cruise Number ( Year & Month )

Figure 5.-Mean biomass, by cruise, of a, zoopiankton in southern California area, based on same samples examined foreuphausiids; and
b, biomass of Euphausia pacifica and its proportion of zoopiankton biomass, with dominant body lengths indicated.

742

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

lar\'ae. Large and medium-sized spawners were
substantial contributors to this recruitment. The
latter were predominant and continued to be
throughout 1955. This differed from 1953, 1954, and
1956 when small or large spawners were
predominant during at least part of the year.

In September a brief increase in larvae closely
followed the year's peak in potential egg produc-
tion, as observed during four September cruises
closely spaced in time. This was at the time of
maximum water temperatures (Figure 4b). The
November peak in eggs, to which small spawning
females contributed importantly for the only time
in 1955, led to a slight increase in recruitment in
December. This November activity was associated
with residual upwelling that was significantly
more intense than the 20-yr November mean.

1956

Upwelling began early in February as in 1953.
February spawning was also high, as in 1953 and
differing from 1954-55. Spawners were small and
medium-sized females (Figure 4d). Larvae peaked
during the same month. Following a March decline
in eggs and larvae, April spawning returned to the
February level associated with the usual spring
appearance of large spawners. This egg maximum
was followed in May by a small peak in larvae. In
June, egg production reached a peak for the 4-yr
period (13,000 eggs/1,000 m'^) at the same time as a
4-yr peak in the upwelling index which, however,
was not confirmed by the observed temperature
minima (Figure 4a, b). In July 1956, larvae showed
strong survival from the June spawn with a
density of 17,000/1,000 m\ While the upwelling
index continued to be well above average through
August, CalCOFI sampling did not resume until
October. Therefore, August-September recruit-
ment was not recorded. High numbers of larv^ae
observed from 28 September to 5 October
(5,000/1,000 m^) together with record numbers of
8-11 mm juveniles appearing in November-
December (Figure 10) indicated that August-Sep-
tember spawning was heavy and greater than the
substantial August-September spawn of 1953.

An increase in egg production in November 1956
resulted in little recruitment in December, after
upwelling had stopped. An explanation may be
inferred from the fact that, though zooplankton
biomass had peaked earlier (May-July, 5505-07),
the euphausiid part of the biomass became ex-
tremely high (24-41 g/1,000 m^) only in

November-December, consisting largely of 8-12
mm juveniles and young adults (Figure 5b). Eu-
phausia pacijica then made up a larger proportion
than ever before of the total biomass (15-20%),
indicating a diminished amount of organisms of
other taxa, such as salps and copepods. These, like
larval euphausiids^ depend heavily upon primary
production for food. Their reduced numbers sug-
gest diminished phytoplankton food (unless their
mortality was not food related), hence the poor
December survival of E. pacijica larvae emerging
from the November spawn. Additional evidence of
diminished food in November-December will be
seen in the negligible rate of growth during
November-December of the massive population of
8-12 mm E. pacijica. Alternatively, this population
may have consumed the November larvae as well
as their food, but this presumption is not support-
ed by its low growth rate.

Recruitment Efficiency and
Spatial Aggregation of Eggs

The relationship of spawning potential (density
of ripe eggs) to larvae subsequently recruited is
irregular, although a trend (Figure 6) indicated
that eflSciency of recruitment from available eggs
was better during spring and summer (March-
September) than during fall and winter (October-
February). In 1953 the spring-summer peaks in

RELATION OF NO OF LARVAE RECRUITED IN GIVEN MONTH
TO NO OF RIPE EGGS OBSERVED IN PREVIOUS MONTH
• SPWNG-SOmUCR (MARCH EGGS VS APRIL LARVAE, Tt«OUGH AUGUST EGGS VS SEPT LARVAE)
o AUTUMN- WINTER (SEPT EGGS VS OCT LARVRE.Tt«OOGH FEB EGGS VS MARCH LARVAE)

13.000

5000 10,000

NO OF LARVM (PER lOOOm")

Figure 6.-Density of <4.5 mm Euphausia pacijica larvae in
given month in relation to ripe unspawned eggs observed
previous month, 1953-56 data. Regressions (Bartlett's test) for
spring-summer (March-September) and autumn-winter
(October-February) data are not significantly different.

743

FISHERY BULLETIN: VOL. 74, NO. 4

egg production were followed by proportionately
high peaks in larvae, relative to 1955 and early
1956 (January-April) (Figure 4c). Both the one
peak in eggs in 1954 (June) and the highest peak in
1956 (June) led to particularly heavy recruitment.
Incipient spawners and larvae were both un-
evenly distributed in the study area, the larvae
usually more patchy than the spawners (Brinton
and Wyllie in press). A possible effect of relative
aggregation of spawners on recruitment was
considered. A monthly index of survival of newly
hatched larvae was determined as the ratio of the
mean density of larvae observed on a given cruise
to the density of ripe eggs calculated for the
previous month— usually one cruise earlier. (As
noted above, this ratio is >1.0 in about one-third of
the instances, indicating that spawning is under-
estimated. The indices are, therefore, regarded
only as relative to each other.) Cruise-to-cruise
differences in patchiness of spawners were es-
timated by comparing, among cruises, variances
of number of ripe eggs carried by incipient
spawners. Each variance was derived by use of
numbers from all stations of a cruise. The regres-
sion of patchiness in relation to survival of calyp-
topis larvae showed a slope not significantly
different from zero (Figure 7). Evidently,
differences in the degree of aggregation of
spawners on the scale observed (32-64 km between
stations) did not affect sun-ival of newly hatched
larvae.

S £

w
to

<
a.

RELATIVE RECRUITMENT

^0. LARVAE, l-4mm, PER CRUISE, PER lOOOm^
NO EGGS PER PREVIOUS CRUISE. PER lOOOm'

Figure 7.— Index of patchiness of ripe unspawned eggs of
Euphausia pacifica in relation to index of recruitment during
succeeding month. Slope of regression not significantly different
from 0 (P>0.05, <-test). Standard deviation is used as a measure of
dispersion and in no way assumes normality of the data.

Temperature Relationships of
Spawners and Larvae

Abundances of spawners and recently hatched
larvae (calyptopes of <2.5 mm) were plotted in
relation to ambient temperature at 10 m depth
(Figure 8). A relationship of spring-summer up-
welling to maxima in reproduction, however indi-
rect, was evident in foregoing observations. There-
fore, data for the months of strong upwelling
(April-July) are separated from those of the other
months.

Both spawners and larvae occurred across a
range 10°-21.6°C, virtually the available range.
When lumped by 0.5°C increments, close to 40% of
the stations yielded some calyptopis larvae and
40% yielded incipient spawners. During August-
March (Figure 10a) larvae were most concentrated
within the range of 12°-16.5°C, the same as the
spawners. There, mean densities of larvae were
50-200/1,000 m\ During April-July (Figure 8b),
maximum densities of larvae, 200-7,000/1,000 m'^.

TEMPERATURE RELATIONSHIPS (1953-56 DATA) OF
Jl^^ SMALL (<2 5mr,i) E poatKa LARVAE
fi'*\ RIPE (SPERMATAPHORE-BEARING) FEMALES
e'o MAXIMUM NUMBER

WEAK-UPWELLING MONTHS (AUGUST -MARCH)

S: 7000^

loo-
se-

12" 14" 16"

STRONG- UPWELLING MONTHS (APRIL -JULY)

5" §•

i I

12" 14" 16" 18" 20"

TEVf=ERATURE ("C AT Om DEPTH)

Figure 8.— Densities of Euphausia pacifica larvae <2.5 mm
length and ripe females in relation to water temperature at 10 m
depth, a, August-March; b, April-July.

744

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

were within a somewhat narrow range of tem-
perature, 12°-15.5°C, as compared with the weak-
upwelling period, and most spawners were within
a broader range, 11.5°-17.5°C.

By years, during 1953, 1955, and 1956 the mean
maxima of larvae were at 14°-15°C. In 1954 there
were maxima at both 13°-14°C and 14.5°-16°C.
Occurrences of larvae and spawners at tempera-
tures >18.5°C during January-March and August-
December were only during 1955, the year in which
spawning extended on into September and
November. Occurrences of April-July larvae at
temperatures >18°C were all during 1954, the year
of weakest upwelling, except for a single record in
1953.

Overall frequency of spawners did not differ
between the periods of strong and weak up-
welling, in contrast to large differences in the
frequency of recruits. This implies that factors
other than temperature are important to
recruitment— probably the production of food
associated with the upwelling period. Patchy local
increases in surface nutrients associated with the
upwelling season of 1969 are described in the
Discussion.

It is also noteworthy that during periods of both
strong and weak upwelling, mean maxima of
spawners occurred at or just outside the limits of
the optimal temperature range for recruits: dur-
ing weak upwelling months, at 12°-12.5°C and
15.5°-16°C; and during strong upwelling months,
at 11.5°-12°C and 17°-17.5°C. This implies that
stations showing maximum densities of larvae
and those showing maxima of spawners were
mutually exclusive— an impression gained earlier
during counting. Removal of adults from the
region where they might fortuitously feed upon
their young could be brought about by the vertical
migration of the adults and their consequent
differential transport at greater daytime depths,
in accordance with the hypothesis of Hardy (1956).

Growth

Monthly L-F polygons for E. pacifica consis-
tently peak at larvae 3-4 mm in length (Figures 9,
10). In the stream of continuous recruitment, a
month-class is first distinguishable as high
numbers of larvae relative to those produced in the
months before and after. Subsequent growth can
be traced through successive months as an L-F
mode, either in the form of a crest, irregularity in
slope, or change in height relative to the month

before. Observations of growth and survival ap-
pear most reliable when cohorts are traced that
begin as densities in excess of 2,000 larvae per 1
mm length increment per 1,000 m^.

A cohort is designated by the year-month (e.g.,
5303) in which its larva maximum is observed.
Presumed relationships of egg maxima (Figure 4c)
to subsequent recruitment are indicated in Figure
12.

When presented in terms of biomass (Figures 9,
10), population composition differs from that
indicated by length frequency. For example,
biomass modes may increase in height with time
owing to growth, while corresponding L-F modes
decrease in height because of mortality. As a
consequence, cohorts are often more conspicuous
when plotted as biomass. Biomass is plotted on a
linear scale while abundances (length frequency)
were plotted on a logarithmic scale to accom-
modate fluctuations in the many larvae and the
few large adults. The biomass of larvae was
generally low but periods of heavy recruitment are
conspicuous.

1953 Cohorts

A small February 1953 cohort (Figures 9, 12) was
tentatively traced through April as 10-11 mm
adolescents. More substantial recruitment oc-
curred in March from the February egg maximum,
followed by little recruitment in April; growth
appears to have been to 7-8 mm in April, 10 mm in
May, 13 mm in June, 15-17 mm in July, 18 mm in
August, and 18-19 mm in September.

Production of larvae first became intense during
May-July 1953 (cuises 5305-5307), resulting in a
broad mode recognizable as 3-7 mm in June
(enclosed by a pair of dashed lines in Figure 11, one
originating at 3 mm in May and the other at 3 mm
in June). July larvae appeared to show poor sur-
vival, as shown in the reduced 5-7 mm component
of the population in August. This is interpreted as
leading to graphic separation of the May-June
cohort as a conspicuous L-F mode, first observed in
August (5308) as 8-13 mm juveniles and young
adults (Figures 9, 11a), persisting into September
at 12-15 mm, and perhaps surviving without
growth into October, though decimated. Develop-
ment of this cohort is even more conspicuous
through the sequence of biomass modes.

An increase in recruitment in September (5309)
over August, followed by low production in Oc-
tober, yielded a particularly conspicuous cohort

745

FISHERY BULLETIN: VOL. 74, NO. 4

JAN '53

FEB '5}

MARCH 53

SOOO-

MAY 53

'.->'-

JUNE '53

r-\V

2 5 , \ 10 ', 15', 20 23

JllMlNiL

OCT '53

NOV '53

/ I,

wu,

JAN '51

FEB b'>

MARCH '54

r-^

r."j

AUG '54

Hk^

SEPT 54

NQ1.».

DEC 54

TOTAL LENCTH (mm)

Figure 9.-Length-frequency (histograms) and biomass (line graphs) distributions, of Euphausia
pacifica, 1953-54 cruises. Dotted bwxes appended to histograms for body lengths 16-21 mm are
corrections for net avoidance using Isaacs' (1965) factors derived for anchovy larvae of those sizes.
Corrections are not applied to biomass. Arrows trace development of cohorts. Solid arrows trace
sequences considered clear, dashed arrows trace those less clear.

746

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

SEPT 4-6 55

MARCH 8-16 '56

SEPT ll-24'55

OCT 21- 30 '55

Ih^

NOV 8-13 '55

NOV l6-e'55

MAY 30-JUrC 13

JULY 7-l4'56

Ite

DEC 1-7 55

JAN 5-9'56

SEPT 28 1

FEB 5-9 56

OCT 29
-NOV 4

DEC 5-M 56

2,5 10

2 5 10

TOTAL LENGTH (mm)

Figure lO.-Length-frequency (histograms) and biomass (line graphs) distributions, of
Euphausia pacifica, 1955-56 cruises. Dotted boxes appended to histograms for body lengths 16-21
mm are corrections for net avoidance using Isaacs' (1965) factors derived for anchovy larvae of
those sizes. Corrections are not applied to biomass. Arrows trace development of cohorts. Solid
arrows trace sequences considered clear, dashed arrows trace those less clear.

747

FISHERY BULLETIN: VOL. 74, NO. 4

50001

1000 =

100 =

5305

10 =

3UUU-

1000 =

»— '

100 =

10 E

5000

1000 =

5401
5402

5000

1000 =

100:

\0:

5000:

1000

100:

IOh

5000

1000 =

100:

10=

I ■

5506
r5507

10
-JULY 7-13,1955

SEPT 4-6
SEPT 11-24

0CT2I-30
NOV 8-13
N0VI6-I9
DEC 1-7
JAN 5-9, 1956
FEB 5-9

-SEPT 28 -OCT 5 5509-10
—OCT 29 -NOV 4
r-DEC5-ll

— I —
10

— I —
15

TOTAL LENGTH (mm)

Figure 11. -Cohort development of Euphausia pacifica, shown as progressions of length-frequency modes. Curves are
three-point running averages of portions of histograms in Figures 9 and 10. Cohort is identified by date (cruise) at
appearance of conspicuous mode of 2-3 mm larvae, a, 5305; b, 5309; c, 5406; d, 5502; e, 5505; f, 5509-10.

748

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

traceable for 10 mo through July 1954 (5407) when
it had achieved large-adult size, 17-20 mm (Figures
9, lib, 12). Separate L-F curves for males and
females (Figure 13), commencing at the onset of
maturity ca. 11 mm, show that the modes for the
5309 cohort illustrated in Figures 9 and lib ac-
tually are made up of paired overlapping peaks,
for females regularly at a larger body-length
increment by about 1 mm and for males where the
difference in absolute frequency between males
and females is greatest.

It is not likely that females, upon maturity, have
undergone sudden, relatively rapid growth so as to
exceed males in size. The curves (Figure 13) show
larger females to be at a relatively greater
frequency than males and the converse would be
expected with increased female growth-rate.
(Average male/female ratio is probably 1:1 at on-
set of adulthood, discussed under Sex Ratio.)
Rather, the most mature females— those at the
leading edge of the mode-cohort at the onset of
February-March breeding— are growing slower
than before, thereby appearing more numerous.
At the same time, decreasing relative male sur-
vivorship could contribute to the increasing in-
equality in sex ratio. At body lengths >16 mm,
females tend to dominate by 2:1 or more, indicat-
ing that they then spend twice as long as males at
given sizes, at least while breeding, or that their
survivorship is then greater, or that males remain
below sampling depths at night. These alterna-
tives are considered in the discussion of Sex Ratio,
below.

GROWTH CURVES OF THE COHORTS CONSIDERED TRACEABLE

J FMAMJ J ASONOJ FMAMJJ A
1953 1954

D|JFM>MJJ SONDJFUAMJJ OND

I9S5 1956

Figure 12.-Growth curves of Euphausia pacijica inferred from
length-frequency modes. Clear (solid lines) and unclear (dashed
lines) sequences as in Figures 9 and 10. Times of egg production
are extrapolated, see Figure 4c. Fall-winter period of slowed
growth is crosshatched.

1954 Cohorts

The single intense spawn of 1954 (June) led to
strong June-July recruitment, establishing a

cohort (5406) that was followed through a 10-mo
period to 17-19 mm in April 1955 and, with less
certainty, to 20 mm in June (Figures 9, 10, lie, 13).

1955 Cohorts

Conspicuous 1955 cohorts arose in February
(5502) and July (5507). The former appeared to
attain 18 mm after 7 mo (September) and the
latter reached 17-18 mm after 8 mo, following
slowed growth during October-January (Figures
lld-e, 12). This cohort appeared at too-low density
in October (5510) relative to a month later. This
may be due either to sampling variability or to
"piling up" at the 11-12 mm increment in
November owing to growth being faster into the
newly adult phase than out of it, energy then
being diverted to gonad development. Neverthe-
less, it is noteworthy that the 5502 and 5507
cohorts appeared to be distinguishable in October
(5510) as modes of 10-12 mm and 15-16 mm, Figure
le, discussed earlier when the southern California
area was compared with the California Current as
a whole.

The December 1955 cohort was the only distinct
year-end cohort observed during 1953-56 (Figures
10, 12, 13). It grew rapidly at 4 to 5 mm/mo during
December-February and 3 to 4 mm /mo during
February-April, apparently attaining 18 to 20 mm
length by June 1956.

1956 Cohorts

These were scarcely traceable except for that
appearing as 8-11 mm individuals in early
November and as 8-12 mm in December. This
mode doubtless derives from extremely dense
larvae sampled during 5507 and 5509-10, its crest
appearing to relate mainly to the latter. The small
biomass peak at 10-11 mm during 29 September-5
October is clearly derived from the very heavy July
recruitment. It subsequently becomes indistin-
guishable during November and December from
the biomass of 8-12 mm juvenile-adults considered
to have grown from 5509-10 larvae. The 29 October
-4 November peak appears most likely to have
derived from the 5509-10 larvae.

Survivorship

The average L-F distribution for all samples
(Figure 14) shows that decline in density with
body length is roughly exponential. The decline is

749

5406 COHORT

-"^ TOTAL FEMALES

NON-GRAVID FEMALES
TOTAL MALES
-5309 COHORT

', ' 5506 sg

10 12 14 16 IS 20 22

FISHERY BULLETIN: VOL. 74. NO. 4

ADULT MALE . FEMALE
Euphausia pacifica

PATHWAYS OF GROWTH

AND ABUNDANCE OF

REPRESENTATIVE COHORTS

200r

5507 50

10 12 14 16 18 20 22

TOTAL LENGTH (mm)

5512 COHORT

10 12 14 16 18 20 22

Figure 13. -Length-frequencies of adult males and females of Euphausia pacifica. Dashed lines trace development in males and
females. Frequencies of females without ripe eggs are indicated (pertinent to discussion of Sex Ratio).

rapid during the larval phase and slower there-
after until large adulthood, 18-19 mm. Positive
perturbations appear at 6-7 mm, 9-10 mm, and
14-15 mm. Average survivorship is 16% during the
1 mo furcilia lan^a phase, as seen in the decline in
mean population density from 1,850 to 300/1,000
m-' (Figure 14) between 3 mm and about 6 mm in
body length which Boden (1950) has shown to be

750

larval phase. For juveniles, 6 mm through 9 mm,
survivorship is near 67%/mo over about 2 mo.

For adolescents and young adults of 9-14 mm,
average survivorship remains nearly the same,
64%/ mo, then decreasing to 60% /mo through 18
mm. After that, population decline appears rapid,
possibly because sampling of such large in-
dividuals is not representative. Apparent

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

E pacifica SURVIVORSHIP

RAPID INOtEASE
IN % FERTUTY

E
O
O
O—

S§

10 C M 16

BODY LENGTH (mm)

Figure 14. -Catch cur\-es for all Euphausia pacifica sampled,
densities shown on logarithmic and linear (adults only) scales.
Periods of changing slope (changing sun'ivorship, net avoidance
and/or growth rates) indicated as related to life phases. Scale
used for density of sexed adults (right) is doubled for lumped
immatures (left).

differences in survivorship between males and
females (Figure 14) are discussed below under Sex
Ratio.

Survival rates for individual cohorts were ap-
proximated from relative amplitudes of month-
to-month modes in the sequences used to trace
growth (e.g.. Figure 11). Percent survivorship
plotted against estimated age shows cohort cun-es
to be similar (Figure 15a). A positive change in
slope consistently occurs within the range of 8-12
mm body length encompassing adolescence. How-
ever, regressions of logio density on age take two
forms:

1) Mean life-span survival rate calculated as a
single linear regression for individual cohorts is
highest among those recruited during June-
December (06-12). For example, it is 51%/mo for
the 5512 cohort, 58^c for 5610, and o9^c for 5309. In
such late-year cohorts most of the juvenile-adult
phase is during August-March, the period of
reduced food and slowed growth. For example, the
cohort 5507 attained adolescence (9-10 mm) in
September and large adulthood (17 mm) in March
(Figure 15b), having an estimated life span of 10
mo. (Egg stage to 3 or 4 mm length is considered
the first month.) The cohort 5406 (Figure 15c)
attained adolescence at 9 mm in August, appeared
to show strong survival through 15-16 mm in
February-, and was distinguishable at 20 mm size
in June-a life span of 13 mo. Thus those cohorts
which attained 15-16 mm with densities >50/l,(X)0

100

80-

X
CO

geo

<« 40

20- ||-l2mm ^^

*("' \ \V9-I0rrm^
- 8- IOn»n^\J<''J<''e
August

0-l2n¥n, Januory

2 34 5 67 89 10 II 12

2 3 4 5 6 7 8 9 10 II 12 13
ESTtMATED AGE (MONTHS)

FiGiniE 15.-Survivorship of cohorts of Euphausia pacifica, from
amplitudes of length-frequency modes, a, Percent survivorship
showing rapid decline until adolescence, ca. 9-11 mm. b, c,
Age-frequency distributions of 06-12 cohorts smoothed for
apparent piling up at times of slowed growth, d, e, Age-
frequency distributions of 02-05 cohorts, f, Curves seen in b-e,
clustered, g. Average slopes (from straight line regressions) for
02-05 cohorts seen as steeper than for 06-12 cohorts.

751

FISHERY BULLETIN: VOL. 74. NO. 4

m^ by February-March continued to be evident on
into the spring bloom.

Two exceptionally large cohorts, 5406 and 5507,
were initiated during late June-July. At first,
these survived poorly, 8-10'^/mo for 5406 through
August-September and 40%/mo for 5507 through
October (Figures lie, f; 15b, c). Growth apparently
then stopped after 9-11 mm body length, and the
density had declined to 100/1,000 m^ This took
place when the onset of maturity was in Sep-
tember-October. This is presumably the start of
the fall-winter period during which food supply is
inadequate to permit both gonad development and
size increase. During October-December, the 10-12
mm sizes increased in frequency, indicating con-
tinuing growth into that range by younger
elements of the overall population and much
reduced growth out of it. Therefore survivorship
of the 5406 and 5507 cohorts during September-
December could not be determined, but it appears
to have been high. By January, body-length
growth of these cohorts, now numerically en-
hanced, resumed. Survivorship of "5406" prevailed
at about 47'^/mo through June 1955 (21 mm), and
for "5507" at 40% /mo as before September.

The large 5607 and 5610 cohorts appear to have
undegone similar development (Figure 10), ap-
pearing to coalesce at 9-12 mm during November-
December, with much increased frequencies at
those body lengths.

2) Survival rate is poorer, 26-45%/mo, for
recruits produced earlier in the year, February-
June. Mean life-span survival was 43%/mo for the
5303 cohort, 26% for 5305, 37% for 5306, 30% for
5404, 44% for 5502, and 45% for 5605. Nonlinear
details of survivorship in these cohorts are depict-
ed in Figure 15d, e, while differences between
early-year and late-year cohorts in mean slope of
survivorship regressions are seen in Figure 15g.
Coincidence of the juvenile-adult phases of early-
year cohorts with the productive period May-Sep-
tember evidently accounts for the observed rapid
growth during this period, hence the poor survival
rate. These cohorts were traced to body lengths of
16-18 mm after 7.5-8 mo (5502, 5303, 5605) or to
13-15 mm after 4-5 mo (5404, 5305). Having de-
clined to densities <10/ 1,000 m-^ during summer-
fall, they were no longer recognizable in winter
sampling.

Annual Biomass

Annual biomass by body length shows year-to-
752

Euphausia pacifica BIOMASS

I4 4gm per lOCXJm'

4 6 8 10 12 14 16 18 20 22

BODY LENGTH ( mm )

Figure 16. -Biomass, annual mean values for Euphausia
pacifica and distributions per 1 mm body length, a, b, Uniform
distributions for 1953, 1954, with modes at 3-4 mm, 7 mm (onset
of juvenile phase), 10-12 mm (onset of adulthood) and 15-16 mm
(start of ma.ximum egg production, cf. Figure 21b). c, d. Dis-
tributions, strongly peaked at adolescence, biased by large 5507
and 5609-10 cohorts respectively.

year similarities (Figure 16). Peaks are at 1) 3-4
mm, owing to consistent abundances of larvae in
early furcilia phase; 2) 7 mm (except 1956), the
onset of juvenile phase; 3) 10-12 mm (9-11 mm in
1956) the onset of adult phase; and 4) 15 mm, early
in the peak reproductive phase. It was noted
(Figure 5b) that monthly biomass peaks were
usually dominated by one or another of these four
body lengths. The larva peak occurs in spite of
rapid early growth. The other three peaks are at
ages when slowed body-length growth would be

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

expected: onset of juvenile phase, onset of gonad
development, and time of maximum gamete
production.

Biomass on body-length distribution was most
even during 1953 and 1954 (Figure 16a, b).
Recruitment in May and September 1953 led to the
7-mm peak of that year, and the September cohort
was the main contributor to the 10-12 mm peak.
The 1954 crest at 10-12 mm stemmed mainly from
October and December sampling of the June 1954
cohort.

In 1955 and 1956, 3-4 mm larvae were reduced in
average biomass compared with 1953 and 1954
while biomass of 9-12 mm adolescents was 'two
times greater. The November 1955 stock of 11-12
mm stages (5507 cohort) was mainly responsible
for the 1955 biomass peaks. The November-
December 1956 stock of 9-11 mm stages (5609-10
cohort) provided much of the 1956 peak.

Large 18-20 mm adults showed their greatest
biomass in 1956 following the strong upwelling
year 1955, and lowest in 1955 following least
productive year 1954.

Monthly changes in biomass are traced for each
of three conspicuous sizes (Figure 17). Small (7 mm)
juvenile bulk is greatest within May-July follow-
ing spring recruitment. Other high values for the
7-mm size are not consistent seasonally, occurring
during October-March.

Adolescents (10 mm), considered representative
of the 9-12 mm juvenile-adult phase change, tend
to be at greatest volume during August-January
(when the smallest spawners, 10.6-12.5 mm, were
also observed to peak, Figure 4d). Increased sur-
vivorship and slowed growth during fall-winter
maturation of spring cohorts, discussed above, are
considered responsible.

Subsequent February-March peaking of
biomass at 15-mm size occurs as egg development
accelerates. (This is preliminary to the appearance
of the large >16.5 mm spawners during April-
June, Figure 4d.)

A close relationship is evident (Figure 17)
between biomass of each of the three sizes and
their percent of the total E. pacifica biomass. This
indicates that a given month's increase in biomass
of the 7-mm size (or of the 10-mm or 15-mm size) is
not accompanied by proportionate increase in the
composite biomass of all other sizes. Therefore, the
periodic peaks in biomass shown in Figure 5b
should be largely due to peaks at these or very
similar sizes, which was indeed the case.

7 mm SIZE

15mm SIZE

01 02 03 04 05 06 07
MONTH

10 II

<
O

Q
O

E
O

o
o

Figure 17. -Annual length-biomass modes of Eupkausia
pacifica analyzed by months. The 7-mm size peaks heavily in
June-July (May-June in 1956), with other peaks in March and
November-January; the 10-mm size peaks September-January;
and 15-mm size peaks February or March.

There are variations from this relationship: 1)
moderate increase in biomass of 10-mm size dur-
ing August-December 1954 caused a dispropor-
tionately large percent-increase in it-an effect of
the single large 1954 cohort (5406-07) developing

753

unaccompanied by other substantial cohorts
(Figure 16b); 2) the converse, when November-
December 1955 biomass of 10-mm size (together
with 11-12 mm, Figure 16c) increased extremely
but the percent increase did not keep pace because
of strong survival from extended July-September
recruitment, seen as piling up in December across
8-12 mm range.

Rate of growth (body length) was seen, above, to
be generally steady (Figure 12). Slowed growth
was commonest when adolescence or late adult-
hood took place during fall-winter. Exceptionally
high biomass of 10-12 mm sizes in 1955 and 1956
was attributed to greatly slowed growth of
adolescents of large cohorts during November-
December of both years.

Regular, less extreme peaking of biomass at the
four body lengths just descibed as prominent may
be interpreted in terms of differing survival rates
among life phases:

If body-length growth is steady during a given
life phase, such as the larval period, biomass
growth would proceed as the cube of body length,
while population size would be expected to decline
exponentially. This inequality leads to a biomass
peak at a particular body length which depends on
survival rate (Figure 18a). A survival rate of about
24%/mo for the larval phase is found to yield such a
peak at 4 mm length in the biomass on body-length
distribution, a size at which biomass regularly
peaks during E. pacijica development.

Other survival rates were extrapolated from a
cluster of age-density curves so as to yield biomass
peaks which coincide with real average peaks
shown in Figure 16: 43%/mo was found to peak at 7
mm, 54%/mo at 10-11 mm, and 66%/mo at 15 mm.
A derived age-biomass distribution, linear scale
(Figure 18b), is composed of segments based on
the above sequence of survival rates. Segments
end at 5.8 mm (end of larval phase), 9.3 mm (end of
juvenile phase), and 13.2 mm (start of intensive
reproduction, after Figure 21b).

The derived distribution is similar in shape to
the observed average annual biomass distribu-
tions for 1953 and 1954 (Figure 16a, b). (Growth
rates of 1953 cohorts were relatively steady,
Figure 12. Those of 1954 appeared less steady but
were still without the massive November-
December pile-ups of adolescents noted in 1955
and 1956.) However, except for the larval period
for which the derived and observed mean survival
rates (from Figure 14) were both about 23%/mo;
other derived rates had to be different from the

754

lo-

CO
LiJ
CC
C3

g 10

z)
(/)

tr
O

t 10

E
O

o
o

FISHERY BULLETIN: VOL. 74, NO. 4

cr
o

01 o

001

0001

Body Length
3-5Bmm .K

End

Larvol Phose

Survi»ol22%/mo ^5B-93mni ^^

*3^ . 93-l32mm

Begin Gonod O«v«lopm«nt

O 100

^ Begin Intensive Reproduction

13.2 mm -

- 2

E
o
o
o

4 6 8

AGE (MONTHS)

I I I \ \ L

II 15

BODY LENGTH (mm)

Figure 18. -Hypothetical age-frequency and age-biomass dis-
tributions of Euphausia pacijica assuming uniform body-length
growth, a, Constant survivorship at each of four rates, selected to
yield biomass peaking at 4, 7, 10, and 15 mm, respectively, b, An
approximation of annual length-biomass distributions shown in
Figure 17, obtained by changing survivorship at life-phase
change.

mean observed rates so as to yield the observed
peaks at 7, 10-11, and 15 mm length. These were
lower by 24% and 10% for the juvenile and young
adult phases respectively, and higher by 6% for the
14-18 mm sizes. This means that after the larval
phase observed, mean survivorship decreased
phase-to-phase by about 4%/mo, whereas in the
derived distribution it increased by 11-12%/mo at
phase change. This is attributed to deviations
from evenness in real growth rates. However,
there is a tendency toward progressively positive
inflexion with age in certain of the survivorship
curves of individual cohorts (Figure 15b-g).

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

Sex Ratio

Fifty percent of the estimates of prespawned
eggs were two to four times greater than the
estimates of larvae in the plankton a month later.
The other 50% of egg/larva ratios were even lower
than two (Figures 4d, 6). Further evidence that
spawning was underestimated is seen in an ex-
amination of ratio of the sexes and state of their
reproductive products.

The ripe male E. pacifica stores two spermato-
phores in a pair of ducts. The fertilized female
possesses a single attached spermatophore
(Brinton in press). This discrepancy might be
attributed to a sex ratio in which females
predominate, or to a need for more than one
fertilization when spawning is protracted or in-
termittent across days, intermolt periods or
longer. If such multiple fertilizations take place,
males transfer one spermatophore to each of two
females, probably quickly because single ripe
spermatophores were not observed in males. The
paired spermatophores in males were observed
always to be of equal size, color, and readiness for
extrusion. (Ready spermatophores may be easily
expelled with gentle external pressure in the
laboratory.) A continuing preponderance of ripe
males, as shown in Figure 19, would tend to insure
fertilization of females whenever they ripen.
Mauchline and Fisher (1969) have explained, with
reference to Meganyctiphanes norvegica, that
fully formed spermatophores may be stored in the
ejaculatory ducts for some time.

Here, ripe and unripe females outnumber males
by about 1.5 times at 15 mm, and 3 times at 20 mm
(See Figure 21a). Ponomareva's (1963) data on E.
pacifica from the Sea of Japan showed females to
be 56% of the adult population, and from the
Okhotsk Sea 63% in April, 62% in June-July, and
44% in October. Four factors may contribute to the
apparently greater number of females:

1) In the present data, apparent dominance by
females (all body lengths lumped, Figure 20) is
partly due to periods in which the population
included late-maturing individuals of 10.5-11.5
mm length, some males of which were as yet
without petasmas and were therefore categorized
as females. (Secondary sexual characters of E.
pacifica are usually evident at this size.) For
example, this apparently happened during count-
ing of material from cruises 5401 and 5402 (Figure
13), and cruises 5610-12 (Figure 20) when

MAXIMUM LARVAE

FEa-APfi MAX

MEDIUM ' SIZED

SPAWNERS

MAXIMUM EG6S

APR.,JUN

OOAL MAXIMA

FOR

LARGE SMWNERS

SUMMER MAX.

MEDIUM -SIZED

SPAWNERS

FALL MAX.
SMALL SMWNERS

Figure 19.-a, Densities of ripe female Euphausia pacifica by
months, three body-length groups, 1953-56 data combined from
Figure 6d. b. Densities of males with ready spermatophores,
same body-length groups.

"females" dominated the dense population of 8-12
mm individuals.

2) Increasing mortality in males relative to that
in females may take place after 12 mm body
length. Since the ratio of males to females
decreases with body length, multiple fertilizations
by males would be increasingly important with
increasing size. (Mates are probably of similar
size, in view of large spermatophores being at-
tached to large females and small spermatophores
to small females.)

3) Large males and unripe females may be more
underestimated than egg-bearing females if the
latter are less able to avoid net capture. For
anchovy larvae, Isaacs (1965) hypothesized that
avoidance of the 1-m net becomes significant after
15 mm body length. Similar differential avoidance
might contribute to the female/male bias here.
For 3 of the 4 yr, the average percentage of
females that are ripe crested at 15-16 mm (Figure
21b). It remained high, 40-60%, through the larger
size groups. The 1954 data differed in that the
proportion of ripe/unripe females remained low
through 16 mm body length. This is also seen in
Figure 4d in which the 12.6-16.5 mm group showed

755

FISHERY BULLETIN: VOL. 74, NO. 4

500

60r

a 02 03 04 05 06 07

CRUISE

10 II 12

Figure 20.— Mean densities of presumed mature (>10.5 mm) male
and female Euphausia pacifica, by month.

low reproductive activity. Furthermore, in 1954
the sex ratios for 13, 14, and 16 mm body length
were 1:1, as compared with other years (Figure
21a). However, no relationship was seen (Figure
13) between numbers of gravid females of a given
size and the difference between numbers of males
and total females of the same size. Therefore, the
observed increase with body length (at least to
15-16 mm) in the ratio of gravid to nongravid
females appears natural, attributable either to

C3 50
.<o

;§

I 40

§ 30

I
l3 20

y
'b 10

- a

1954

^^^d"" '" "' '

-^\

-r

I956\ /

1953

^^>-^

4- YEAR

AVERAGE

1955^., '

1 1 1

1 1

1 1 1

II 12 13 14 15 16 17 18 19 20 21

70 1-

Ch-

60-

< 50-

I-

1. 401-

o
0+

^20
s9 10

.).9.55..
1956

/
/

/..■■■
/ —

/l954

11-12 13-14 15-16 17-18 19-21

SIZE OF ADULT Euphausia pacifica ( mm )

Figure 21.— a, Sex ratios for Euphausia pacifica by 1 mm body
length, all data for each year averaged and 4-yr average, b,
Annual gravidity ratios, by body length.

higher frequency or longer duration of egg
production with increasing body length.

4) An increase with body length in female/male
ratio may be due to their differing growth rates.
Both sexes tend to mature at the same size, ca. 11
mm. Thereafter, females grow slower, appearing
increasingly numerous relative to males at suc-
cessive body-length increments (Figures 14, 21a).
Nemoto's (1957) data suggested that the adult
male of E. pacifica tends to be smaller than the
female of the same age, and Mauchline (1960)
stated this to be the case in Meganyctiphanes
norvegica. Slower growth rate in females indicates
shorter life span for males, probably by 1 or 2 mo,
since females grow to 21 mm length off southern
California (rarely more) compared with 20 mm for
males.

In summary, reasons were sought for a) un-
derestimation of spawning, b) paired spermato-
phores in males, and c) apparent imbalance in sex
ratio. These explanations were considered: eggs
can ripen and females can spawn more often than
the frequency of the surveys, applicable to a) and
b); the bias in sex ratio favoring females is real
and develops either with higher male mortality at

756

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFIC A

all sizes after maturity, applicable to b) and c), or
because females grow slower and live longer, also
applicable to b) and c); and the bias is an artifact of
reduced net avoidance by ripe females and of
observations during seasons when some males
mature relatively late, resembly females for a
time, applicable to c). Evidence supports each of
the above. With regard to the increasing
female/male ratio with body size, there are par-
ticularly strong indications of relatively slow
growth in females, apparently leading to better
survival than in males at given sizes and ages.

DISCUSSION

The predominance of E. pacifica among zoo-
plankters off southern California appears related
to the spring-early summer upwelling regime,
which coincides with heaviest spawning. Recruit-
ment consistently crested during May-July fol-
lowing annual surface temperature minima in
April or May. Although this species may range
southward along California and Mexico because of
currents and the cool (10°-18°C) water, sole
dependence on temperature effect in the southern
California area for reproductivity is not likely
because the area as a whole is coolest during
December- April and the most substantial recruit-
ment is later.

That the dependence is partly effected by food,
as indicated by the seasonal pattern in availability
of nutrients (plants), was shown in charts of the
California Current region for 1969 for nutrients
(Thomas and Seibert 1974) and for chlorophyll a
(Owen 1974). The assumption is made that timing
of the seasons in 1969 agrees enough with 1953-56
so that the April-June buildup in upwelling applies
to both periods. Off southern California nutrient
concentrations intensified in April and peaked in
May in a patchy distribution corresponding to the
areas of low surface temperature. For example,
PO^-P (integrated through 0-50 m depth) was in
the range of 10-40 mg-at./m- during January-
February and August-December but increased to
40-60 mg-at./m- during April-June. Silicate-Si
peaked at 400-1,000 mg-at./m- during April-June;
during the other months concentrations >400
mg-at./m- were rare.

Correspondingly, during the main upwelling
period, April-September, chlorophyll a in the
surface waters inhabited by newly hatched eu-
phausiid larvae showed the patchy pattern of
extreme concentrations shown also by the nu-

trients. Values peaked at 3.0 mg/m' during April-
September compared with 2.0 mg/m^ for Jan-
uary-March and 0.5 mg/m-^ for October-December.
The possible importance of shallow (12-19.5 m),
intense (to 50 iug/liter) chlorophyll maxima-
particularly those containing the dinoflagellate
Gymnodinium splendens-to first feeding of an-
chovy larvae was put forward by Lasker (1975).
These maxima were found during March-April
1974 within 15 km of the southern California coast.
Such layering of food particles could have broad
significance to feeding and survival of zooplankton
larvae.

Most larvae of E. pacifica are found in nearshore
areas described above as recruitment refuges
where upwelling prevails and currents are slug-
gish. Similarly, off Oregon (Smiles and Pearcy
1971), more larvae were in nearshore upwellings
than in offshore water characterized by a summer
productivity minimum typical of the region. Also
working off Oregon, Peterson and Miller (1975)
found no relationship between year-to-year
(1969-71) intensity of summer upwelling and
abundances of euphausiid eggs and larvae (not
identified to species).

Evidence that larvae occupying southern
California waters are produced locally is seen in
the time of the upwelling season along the coast.
Upwelling peaks off southern Baja California in
February-March. Progressing northward, its
maximum off Oregon is during August-Sep-
tember. Hence maximum spawning and recruit-
ment, if upwelling induced, should develop along
the same northerly track, counter to the direction
of main flow in the California Current during this
period of relatively consistent northeast winds.
This is the case: recruitment off mid-Baja Califor-
nia, lat. 27°-29°N, is mainly February-April
(Brinton 1967b, 1973), in Monterey Bay it is both
spring and summer (Barham 1957), off southern
California it is mainly May-July, and off Oregon,
August-December.

Although ripening of ovaries, spawning, and
recruitment reach maxima as consequence of
upwelling-associated events, the southern
California population includes ripe females and
newly hatched larvae year-round (Figures 4c, d;
19). Off Oregon, E. pacifica also includes some
larvae at all times (Smiles and Pearcy 1971); while
in the Sea of Japan (lat. 40°-50°N) and south of
Kamchatka, Alaska (lat. 50°-55°N), in areas en-
riched by winter mixing of the water column, E.
pacifica possesses ripe gonads in May-June, the

757

FISHERY BULLETIN: VOL. 74, NO. 4

presumed breeding period (Ponomareva 1963),
though eggs were abundant in August nearby in
the Sea of Okhotsk. To the north and south of the
eastern Aleutian Islands, Nemoto (1957) found
females of E. pacifica with attached spermato-
phores during July.

Dominance of the southern California popula-
tion by the particular cohorts followed in the
analyses of growth tends to obscure the regular
contribution of small classes, including those of
fall-winter in which densities of larvae are usually
1,000-2,000/1,000 m'.

Such continuous recruitment of variable inten-
sity is seen as an adaptation to midlatitude ir-
regularity in oceanographic conditions, both sea-
sonal and year-to-year, as compared with cycles at
high latitudes. Continuous recruitment permits
the stock to always include a wide spectrum of
sizes and maturity stages, providing a potential
for one or another to adapt to periods of poor
climate or food availability, of differing duration
or amplitude. For example, in 1954, a year of weak
upwelling, recruitment was all but limited to
June-July; nevertheless, spawning resumed at
high intensity during four different periods in
1955.

Periodicity was observed in maxima of spawn-
ing and recruitment, and recruitment is appro-
priately out of phase with the inferred spawning
(Figure 4c), implying substantial synchrony
among breeders. Spawning apparently pulses at a
2-mo frequency during the period of maximum
gamete generation, which also must be the period
of maximum food use by breeders and larvae. This
is to be compared with the annual (or at most,
semiannual) frequency of breeding noted in the
subarctic North Pacific. Thus it appears possible
that, under optimal feeding conditions off south-
ern California, a female might spawn every 2 mo:
first at about 11.5 mm length (20-50 eggs), second
at 16 mm (50-200 eggs), and third at 20 mm
(100-400 eggs), during which time an individual
might be expected to produce a maximum of 650
eggs. This is compatible with an observation of
1,400 oocytes (all stages of development) in ovaries
of an E. pacifica in the springtime in the north-
eastern Pacific (Ponomareva 1963) where spawn-
ing is concentrated into one season, and with
Lasker's observation, reported in Mauchline and
Fisher (1969), that an E. pacifica from southern
California shed 230 eggs after capture.

The long duration of maturity— probably half of
this species' life expectancy-further contributes

to population stability and continuity. In conjunc-
tion with substantial horizontal transport, the
capacity to breed several times enhances genetic
integration across the distributional range.

The first observations on growth in E. pacifica
were from specimens maintained in the laboratory
by Lasker (1966) at 10°C with excess food. In small
juveniles, growth was steady at 2.5 or 2.9 mm
during 2 mo, from about 5 to 8 mm length. In the
southern California field populations, growth of
juveniles of this size was consistently in the range
of 3-3.5 mm /mo. However, the 5309 cohort, having
reached 5 mm by the start of the fall period of
reduced growth, then grew only 3 mm in 1.5 mo.

Larger E. pacifica were observed by Lasker to
grow somewhat slower. A 6.5-mm specimen grew
1.5 mm in 70 days, but added only 1.5 mm in 230
more days before dying, not having reached fully
adult size. A 7.9-mm specimen grew 1.5 mm in 75
days, an 8.0-mm specimen grew 1.5 mm in 130
days, and an 8.4-mm specimen grew 1.0 mm in 160
days. These rates are smaller than those for the
local field populations. They are closer to those
supposed for E. pacifica in the northeastern Pacific
where environmental enrichment is not by inter-
mittent upwelling but by winter mixing followed
by spring stability in the water column, hence not
a continuing process.

In the analysis of growth, cohorts are considered
as normal L-F distributions representing broods
continuously hatched during a few days to a month
or more. Observation on duration of reproduction
is limited by the character of the sampling, here in
approximately 1-wk period with a 2-3 wk interval
between surveys. Only in a few of the months can
a pulse in recruitment by recognized as distinct to
that month. In most months, the larvae derive
from the beginning, continuation, or end of a
period of cohort formation which extends beyond
one survey period and into another. Recruitment
found less than in past or succeeding months is
neither recognizable initially as a cohort nor
traceable thereafter.

The area's population, therefore, is constantly
polymodal in character, being compounded of
individuals belonging to different age-groups and
sexes. The possible difference in size between the
sexes after about 15 mm length was not taken into
consideration in the growth study.

The simplest method of analyzing growth and
survival is that of following obvious modes, survey
to survey. This is probably the most significant
means biologically. Nevertheless, certain im-

758

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

precise trends in development of presumed cohorts
provide growtli rates which corroborate the more
obvious trends. Some pathways of development
indicated in Figures 9 and 10 may appear imagi-
nary unless the shapes and amplitudes of the
related L-F distributions, adjacent in time, are
closely compared. When such indistinct modes are
followed, precision and accuracy in recognizing
rates of development are reduced. Graphical
procedures for mathematically defining cohorts
composing irregular L-F polygons (e.g., Harding
1949) required some subjectivity in recognizing
modes and were employed only in an exploratory
way.

There can be important inaccuracies in field
estimates of growth rate when reliance is upon
time-sequences in L-F modes. Even with steady,
uniform recruitment, peaks or troughs would
appear in the L-F distribution owing to differing
growth rates and survivorship among life phases
or between sexes. With unsteady recruitment,
such peaks may sometimes lie in phase with the
cohort being traced, but the cohort nevertheless
becomes compounded by younger individuals
when its growth is differentially slowed or by older
individuals when accelerated.

It is possible that the individuals composing a
mode could be totally replaced in the course of its
time progression, although the modal assemblage
persists as a size group, presumably feeding and
mating as a unit. I have noted above that spring-
summer cohorts tend to "pile up" in fall-winter
when growth of adolescents appears to be food
limited.

In tracing growth, reliance is therefore upon the
more substantial cohorts. Although these can be
masked, their frequent appearance as modes at
sizes not associated with life-phase changes gives
credence to the method.

Growth rates of E. pacifica off southern
California appear similar to those off Oregon
(Smiles and Pearcy 1971). Figure 22 shows gen-
eralized growth cun^es for this species from four
areas in the North Pacific. The Oregon population
showed steady growth after September recruit-
ment. The juvenile and adolescent phases were
during the winter and 13 mm was reached by
February. About 22 mm was attained after 1 yr.
This parallels development of a winter cohort
(5512) off southern California which grew to 12 mm
in 3 mo and was traced to about 21 mm after 8 mo.

Spring (5406) and summer (5309) cohorts off
southern California grew at rates similar to the

NORTHEAST
PACIFIC

OREGON- >.'^'' ^(Nemoto)

(Smiles a Pearcy)/^

OREGON
(Smiles a Pearcy)

SO CALIF
5512

>■ I I I I I 1,1 I I I I I I I I I I I I I Ill

6 8 10 12 2 4 6 8 10 12 2 4 6 8 10
I MONTHS I

Figure 22.-Representative growth curves from southern
California area compared with curves previously derived for
Euphausia pacifica and illustrated by Smiles and Pearcy (1971).

winter cohort, except for slowing during October-
December-5406 during adolescence and 5309 dur-
ing the juvenile phase.

Here, life expectancy appears to be about 8 mo
for winter and early-spring cohorts, to sizes of
18-20 mm by August-October. December-January
populations never included individuals larger than
19 mm. Life expectancy is up to 12 mo for late-
spring and summer cohorts, which grew to 21-22
mm by the following April-July. This agrees with
estimates of 12 mo for September cohorts off
Oregon.

Growth in other euphausiid species, mostly
summarized in Smiles and Pearcy (1971), is
similar. Several reach about 22 mm after 1 yr: E.
superba (Ruud 1932; Bargmann 1945; Marr 1962),
E. triacantha (Baker 1959), Thysanoessa raschii
(Mauchline 1966), Meganyctiphanes norvegica
(Ruud 1936; Einarsson 1945; Mauchline 1960;
Matthews 1973), and Thysanopoda acutifrons
(Einarsson 1945). Most of these species have a life
expectancy of 2 yr, reproducing in each and grow-
ing slowly or not at all in winter.

During winter in the westernmost North Pacific
(Sea of Okhotsk), Ponomareva (1963) found E.
pacifica to be 8 mm (considered to have hatched
the previous summer) and 14-15 mm (considered 2
yr old). In the spring it was 12-13 mm (1 yr old) and
19 mm (2 yr old). Both groups bred in June. Off
nearby Kamchatka in the summer, Nemoto (1957)
found a size range of 12-22 mm, much like that
found by Ponomareva, but with most at 14-20 mm.
There were no larvae, but females with spermato-
phores were present in September, as off Oregon.

759

FISHERY BULLETIN: VOL. 74. NO. 4

South of the Aleutian Islands in September, he
found a 6-12 mm group interpreted as having
hatched in the spring or early summer. Maximum
numbers of adult females were 16-19 mm in May,
17-21 mm in June, and 18-22 mm in September.

Thus growth of E. pacifica is inferred to be
slower and of longer duration in the Subarctic seas
than off Oregon and California (Figure 22).
Nemoto's (1957) estimate from south of the Aleu-
tians was intermediate between Ponomareva's
(1963) from the western Pacific and those from the
American coast. Ponomareva's finding that sexual
maturity is attained by 15-17 mm, with some
mature at only 11-12 mm, agreed with the obser-
vations off southern California.

During E. pacifica's main reproductive season
there is similarity in surface water temperatures
(Sverdrup et al. 1942; Anonymous 1963) among the
five North Pacific areas from which information on
life history comes; there is less agreement in
winter temperatures:

Sea of Okhotsk
Off Kamchatka
South of Aleutians
Off Oregon

10-13°C (Aug.),

9-ll°C (Aug.),

10-12°C (Aug.),

10-14°C (Sept.),

Off southern Califor- 10-18°C (June),
nia

0°C (Feb.)

0°- l°C(Feb.)

2°- 4°C (Feb.)

9°-ll°C(Feb.)

12°-15°C (Feb.)

The intense densities of E. pacifica at 8-12 mm,
also appearing as conspicuous biomass peaks, are
the rule rather than the exception. Therefore, such
regular concentrating at the adolescence inter-
phase, particularly in fall-winter, may be other
than an incidental consequence of reduced food. It
appears as a means of increasing size uniformity
in the population, hence improved breeding
efficiency, by the time of the spring bloom-a
condition fulfilled by stricter seasonality in the
high-latitude populations of E. pacifica.

ACKNOWLEDGMENTS

I thank E. W. Fager, J. D. Isaacs, and R. Lasker
for providing insight into aspects of this problem;
E. L. Venrick and M. M. Mullin for reading the
manuscript and offering suggestions; A. Town-
send, S. Drais, and T. Stewart for assistance in
many phases of the work; and F. Crowe and B.
Thomas for drafting most of the figures. Support
was provided by the Marine Life Research Pro-
gram, the Scripps Institution of Oceanography's
component of the California Cooperative Oceanic
Fisheries Investigations, a project under sponsor-
ship of the Marine Research Committee of
California, and by the National Science
Foundation.

Winter temperatures in the three subarctic
areas are near 0°C whereas off Oregon and
California they differ little from spring-summer
temperatures influenced by upwelling. An overall
temperature regime for E. pacifica. is thereby
described in which low temperature does not limit
occupancy but in which 9°-16°C is suitable for
reproduction, food permitting. In the subarctic
region reproduction takes place at 9-13°C, the
highest annual temperatures there. To the south
of the California Current off mainland Mexico,
food seems to be abundant, but other factors
(temperatures >20°C, oxygen concentrations <0.1
ml/liter, different current systems) appear there
to curtail the species' range.

The serial biomass representations included
here clearly show rise and decline of cohorts, but
are less exact than length frequency in determin-
ing growth and do not serve in estimating sur-
vivorship. It is evident that biomass of the species
fluctuates month-to-month, with recruitment and
growth not balancing mortality in any regular
way. However in 34 of the 48 mo, the biomass was
within the range of 8-22 g/ 1,000 m^.

LITERATURE CITED

Ahlstro.m, E. H.

1948. A record of pilchard eggs and larvae collected during
surveys made in 1939 to 1941. U.S. Fish. Wild!. Serv., Spec.
Sci. Rep. Fish. 54. 75 p.
Anonymous

1963. CalCOFI atlas of 10-meter temperatures and salini-
ties 1949 through 1959. Calif. Coop. Oceanic Fish. Invest.,
Atlas 1, 297 p.
Baker, A. de C.

1959. The distribution and life history of Euphausia
tricantha Holt and Tattersall. Discovery Rep. 29:309-340.

Bakun, a.

1973. Coastal upwelling indices, west coast of North Amer-
ica. 1946-71. U.S. Dep. Commer.. NOAA Tech. Rep. NMFS
SSRF-671. 103 p.
Bargmann, H. E.

1945. The development and life-history of adolescent and
adult krill, Euphausia superha. Discovery Rep. 23:103-176.
Barham, E. G.

1957. The ecology of sonic scattering layers in the Monterey
Bay area. California. Ph.D. Thesis, Stanford Univ., 192 p.
BODEN. B. P.

1950. The post-naupliar stages of the crustacean Euphausia
pacifica. Trans. Am. Microsc. Soc. 69:373-386.
Brinton, E.

1960. Changes in the distribution of euphausiid crustaceans

760

BRINTON: POPULATION BIOLOGY OF EUPHAUSIA PACIFICA

in the region of the California Current. Calif. Coop.

Oceanic Fish. Invest. Rep. 7:137-146.
1962a. The distribution of Pacific euphausiids. Bull. Scripps

Inst. Oceanogr., Univ. Calif. 8: .51-269.
1962b. Two new species of Euphausiacea, Eupliausia nana

and Stylocheiron robustum from the Pacific. Crustaceana

4:167-179.
1967a. Vertical migration and avoidance capability of

euphausiids in the California Current. Limnol. Oceanogr.

12:451-483.
1967b. Distributional atlas of Euphausiacea (Crustacea) in

the California Current Region, Part I. Calif Coop. Oceanic

Fish. Invest. Atlas .5, 275 p.
1973. Distributional atlas of Euphausiacea (Crustacea) in

the California Current region, Part II. Calif. Coop.

Oceanic Fish. Invest. Atlas 18, 336 p.
In press. Observations of spermatophores attached to

pleopods of male euphausiids. Crustaceana.

Brinton, E., and J. G. Wyllie.

In press. Distributional atlas of Euphausiacea, Part III.
Developmental stages of euphausiids in the region of the
Soi^thern California eddy, 1953-1956. Calif. Coop. Oceanic
Fish. Invest., Atlas 24.
Childress, J. J.

1971. Respiratory rate and depth of occurrence of midwater
animals. Limnol. Oceanogr. 16:104-106.

EiNARSSON. H.

1945. Euphausiacea. 1. North Atlantic species. Dana Rep.
Carlsberg Found. 27, 185 p.
Fowler, S. W., L. F. Small, and S. Keckes.

1971. Effects of temperature and size on molting of eu-
phausiid crustaceans. Mar. Biol. (Berl.) 11:45-51.

Gilfillan, E.

1972. Reactions of Eupliausia pacijica Hansen (Crustacea)
from oceanic, mixed oceanic-coastal and coastal waters of
British Columbia to experimental changes in temperature
and salinity. J. Exp. Mar. Biol. Ecol. 10:29-40.

Gopalakrishnan, K.

1973. Developmental and growth studies of the euphausiid
Nematoscelis difficiliti (Crustacea) based on rearing. Bull.
Scripps Inst. Oceanogr., Univ. Calif. 20:1-87.

Harding, J. P.

1949. The use of probability paper for the graphical analysis
of polymodal frequency distributions. J. Mar. Biol. Assoc.
U.K. 28:141-153.
Hardy, A. C.

1956. The open sea. Its natural history: the world of plank-
ton. Houghton Mifllin Co., Boston, 335 p.
Isaacs, J. D.

1965. Larval sardine and anchovy interrelationships. Calif.
Coop. Oceanic Fish. Invest. Rep. 10:102-140.

Lasker, R.

1964. Molting frequency of a deep-sea crustacean, Euphau-
sia pacifica. Nature (Lond.) 203:96.

1966. Feeding, growth, respiration, and carbon utilization
of euphausiid crustacean. J. Fish. Res. Board Can.
23:1291-1317.

1975. Field criteria for survival of anchovy larvae: The
relation between inshore chlorophyll maximum layers and
successful first feeding. Fish. Bull, U.S. 73:453-462.
Marr, J.

1962. The natural history and geography of the Antarctic
krill (Euphausia superba Dana). Discovery Rep. 32:33-464.

Matthews, J. B. L.

1973. Ecological studies on the deep-water pelagic com-
munity of Korsfjorden, western Norway. Population
dynamics of Meganyctiphanes norvegica (Crustacea,
Euphausiacea) in 1968 and 1969. Sarsia 54:75-90.

Mauchline, J.

1960. The biology of the euphausiid crustacean, Meganycti-
phanes norvegica (M. Sars). Proc. R. Soc. Edinb., Sect. B.,
Biol. 67:141-179.

1966. The biology of Thysanoessa raschii (M. Sars), with a
comparison of its diet with that of Meganyctiphanes
norvegica (M. Sars). In H. Barnes (editor). Some contem-
porary studies in marine science, p. 493-510. George Allen
and Unwin Ltd., Lond.

1968. The development of the eggs in the ovaries of eu-
phausiids and estimation of fecundity. Crustaceana
14:155-163.

Mauchline, J., and L. R. Fisher.

1969. The biology of euphausiids. Adv. Mar. Biol. 7, 454 p.

Miller, J. K.

1966. Biomass determinations of selected zooplankters
found in the California Cooperative Oceanic Fisheries
Investigations. Scripps Inst. Oceanogr. SIO Ref. 66-15, 18

P-
Nemoto, T.

1957. Foods of baleen whales in the northern Pacific. Sci.
Rep. Whales Res. Inst. 12:33-90.

Owen, R. W.,jR.

1974. Distribution of primary production, plant pigments
and Secchi depth in the California Current region, 1969.
Calif. Coop. Oceanic Fish. Invest. Atlas 20:xi-xiv, charts
98-117.

Paranjape, M. a.

1967. Molting and respiration of euphausiids. J. Fish. Res.
Board Can. 24:1229-1240.

Pearcy, W. G., and L. F. Small.

1968. Effects of pressure on the respiration of vertically
migrating crustaceans. J. Fish. Res. Board Can.
25:1311-1316.

Peterson, W. T., .\nd C. B. Miller.

1975. Year-to-year variations in the planktonology of the
Oregon upwelling zone. Fish. Bull, U. S. 73:642-653.

PONOMAREVA, L. A.

1963. Euphausiids of the North Pacific, their distribution
and ecology. Akad. Nauk SSSR Inst. Okeanol. (Translated
by Israel Program for Scientific Translations, Jerusalem
1966, IPST Catalog No. 1368, 154 p.)

Reid, J. L., Jr., G. I. Roden, and J. G. Wyllie.

1958. Studies of the California Current system. Calif. Coop.
Oceanic Fish. Invest., Prog. Rep., 1 July 1956 to 1 Jan. 1958,
p. 27-57.

RUUD.J. T.

1932. On the biology of southern Euphausiidae. Hvalradets

Skr. 2, 105 p.
1936. Euphausiacea. Rep. Dan. Oceanogr. Exped. 1908-1910
Mediter. Adjacent Seas 2D6 (Biol.), 86 p.
Small, L. F.

1967. Energy flow in Euphausia pacifica. Nature (Lond.)
215:515-516.
Small, L. F., and J. F. Hebard.

1967. Respiration of a vertically migrating marine crus-
tacean Euphausia pacifica Hansen. Limnol. Oceanogr.
12:272-280.

761

FISHERY BULLETIN: VOL. 74, NO. 4

Small, L. F., J. F. Hebard, and C. D. McIntire.

1966. Respiration in euphausiids. Nature
211:1210-1211.

(Lond.)

Smiles, M. C, Jr., and W. G. Pearcy.

1971. Size structure and growth rate of Euphausia pacifica
off the Oregon coast. Fish. Bull., U.S. 69:79-86.

Smith, P. E.

1971. Distributional atlas of zooplankton volume in the
California Current region, 1951 through 1966. Calif. Coop.
Oceanic Fish. Invest., Atlas 13: v-xvi, 144 p.

South Pacific Fishery Investigations.

1954. Zooplankton volumes off the Pacific Coast, 1953. U.S.
Fish Wildl. Serv., Spec. Sci. Rep. Fish. 132, 38 p.

1955. Zooplankton volumes off the Pacific coast, 1954. U.S.
Fish Wildl. Serv., Spec. Sci. Rep. Fish. 161, 35 p.

1956. Zooplankton volumes off the Pacific coast, 1955. U.S.
Fish Wildl. Serv., Spec. Sci. Rep. Fish. 177, 32 p.

Sverdrup, H. U., M. W. Johnson, and R. H. Fleming.

1942. The oceans, their physics, chemistry and general
biology. Prentice Hall, N.Y., 1087 p.
Thomas, W. H., and D. L. R. Seibert.

1974. Distribution of nitrate, nitrite, phosphate and silicate
in the California Current region, 1969. Calif. Coop.
Oceanic Fish. Invest., Atlas 20:vii-,\, charts 2-97.
Thrailkill, J. R.

1957. Zooplankton volumes off the Pacific coast, 1956. U.S.
Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 232, 50 p.

Uda, M., and M. Ishino.

1958. Enrichment pattern resulting from eddy system in
relation to fishing grounds. J. Tokyo Univ. Fish.
44:105-119.

Wyllie,J.G.

1966. Geostrophic flow of the California Current at the
surface and at 200 meters. Calif. Coop. Oceanic Fish.
Invest. Atlas 4, 288 p.

762

PRODUCTION OF JUVENILE CHINOOK SALMON,
ONCORHYNCHUS TSHA WYTSCHA, IN A HEATED MODEL STREAM^

Peter A. Bisson-and Gerald E. Davis'

ABSTRACT

Temperature was elevated approximately 4°C in a model stream, compared with an unheated but
otherwise similar control stream. The streams were located outdoors and received identical amounts of
exchange water from a nearby creek. Diel and seasonal temperature fluctuations were similar to those
of area streams. Juvenile spring chinook salmon, Oncorhynchus tshawytscha, were introduced into each
stream either as eyed eggs or fry and allowed to remain for approximately 1 yr. Two consecutive year
classes of juvenile salmon were studied. Their production was measured triweekly and related to
changes in temperature, food availability, and other environmental factors. Ancillary experiments
utilizing water from the model streams permitted measurement of diflferences in growth rate of salmon
fed various rations.

Salmon production in the control stream exceeded that in the heated stream. In 1972, total production
in the control stream was twofold greater and, in 1973, it was approximately 30% greater than in the
heated stream. Elevated temperature resulted in reduced growth rates of the fish especially as food
became less abundant and at times also resulted in lower biomasses of food organisms, either because
the temperature increase directly affected survival and growth of benthic invertebrates or because
increased sedimentation associated with heavier growth of filamentous algae made riffle substrate less
suitable for certain species. Beneficial effects of increased temperature appeared to include protection
from infestation by a trematode parasite (Nanophyetus salmincola) and, possibly, increased tendencies
of some invertebrates to enter the drift.

Studies of the effects of elevated temperature on
stream dwelling organisms have been largely
confined to short-term laboratory experiments or
to field surveys associated vi^ith man-caused ther-
mal increases. We have employed two large model
streams, one heated and one unheated, to examine
the effects of constantly elevated temperature on
production of juvenile chinook salmon, Oncorhyn-
chus tshawytscha (Walbaum), under conditions
similar to natural streams, but where temperature
could be controlled. Identifying the factors
governing productivity of the streams that were
influenced by increased temperature and measur-
ing the impact of the addition of a known amount
of heat on chinook salmon production were the two
main objectives of the research.

Temperature change can affect salmonid fishes
in two general ways. First are the direct effects,
e.g., accelerated developmental rates, altered food
conversion efficiencies, and, under certain condi-

'Technical Paper No. 4078, Oregon Agricultural Experiment
Station, Oregon State University, Corvallis, OR 97331.

-Department of Fisheries and Wildlife, Oregon State Univer-
sity, Corvallis, OR 97331; present address: Weyerhaeuser Com-
pany, Longview, WA 98632.

•"Department of Fisheries and Wildlife, Oregon State Univer-
sity, Corvallis, OR 97331.

tions, lethality. These kinds of effects have
received considerable attention in laboratory
experiments. Less well understood are the indirect
effects, one of the most important being resultant
changes in the abundance of food organisms. In a
previous study involving the same streams, Iver-
son (1972) found that the production of juvenile
coho salmon, 0. kisutch, was significantly reduced
in the heated stream compared with the unheated
control, and he attributed this reduction mainly to
lower biomasses of immature stages of insects in
the heated stream. Evaluating the importance of
indirect consequences of temperature elevation on
juvenile chinook thus became one of our major
concerns, for water quality guidelines relating to
the temperature requirements of salmon and trout
are based primarily upon knowledge of direct
effects and to a much lesser extent upon possible
indirect or secondary effects.

MATERIALS AND METHODS

Physical Characteristics of the Streams

The model streams were located at the Oak
Creek Laboratory of Biology near Corvallis in

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

763

FISHERY BULLETIN: VOL. 74, NO. 4

western Oregon. They consisted of two large
wooden channels interconnected at the ends by
pipes (Figure 1). Within each stream were four
riffle-pool sections of equal size; the total surface
area available to fish and other organisms was 22
m^'. Minor differences in substrate composition,
water velocity and depth, and shading from ter-
restrial vegetation existed among the riffle-pool
sections. These variations were sufficient to
prevent the sections from being treated as rep-
licates; therefore, samples from each of the four
sections were composited.

The slope of each stream was approximately
1.9% so that water pumped into the upstream ends
flowed downstream at velocities typical of natural
streams (approximately 60 cm/s in the riffles to
near 0 cm/s at the bottom of the pools). A 2-
horsepower centrifugal pump forced water from
the downstream end of one channel to the up-
stream end of the other. A gate valve controlled
the flow rate, which was maintained at approx-
imately 1.35 m'^/min.

Complete freedom of movement for the fish was
allowed between the two channels. Individuals
could pass downstream or upstream through the
pipe from one side to the other; they were, how-
ever, prevented from entering the pump by a
screen at the downstream end of the lowermost
pool. Movement of the fish from the streams took
place through a 6-cm diameter outlet pipe that
originated at the screen and terminated in a
partitioned trap. Fish that entered the trap were
returned to the uppermost riffle both to avoid
fortuitous losses and to provide the fish with an
adequate opportunity to establish residence.

Substrate consisted of a layer of rocks approx-

imately 7 cm deep. Following Cummins' (1962)
terminology, cobbles and pebbles composed more
than 95% of the substrate, both in the riffles and
pools, while larger sand was almost absent. No
large boulders were present, although a few cob-
bles projected above the water. A difference in the
amount of very fine sediments existed between
the two streams; this difference will be discussed in
connection with their invertebrate faunas.

Temperature Regulation

Water temperature in the unheated control
followed natural diel and seasonal cycles (Figure
2). Two 6-kw stainless steel heaters regulated by a
variable input timer facilitated temperature
elevation in the heated stream. Continuous
recordings of the temperature were made by
Partlow RFT thermographs.^ Differences between
monthly means ranged from 3.3°C (August 1972)
to 4.9°C (December 1972): the average tempera-
ture difference between the streams was 3.9°C.

Both streams received 10-20 liters/min of
unfiltered water from a small spring-fed creek
that contained aquatic invertebrates and algae,
but no fishes. During periods of low stream flow,
the water supply was supplemented by a mixture
of well water and unfiltered water pumped from a
large nearby creek. The model streams have been
operating continuously at approximately the same
temperature differential since completion of con-
struction in 1969 (Iverson 1972). However, in
December 1972 unusually cold weather caused

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Figure L-Top. Plan view of model
streams. Arrows indicate direction of
water flow and black squares in the
heated stream denote location of
heating units. Bottom. Cross section
of one of the channels in a model
stream.

Riffle

Downstream
Screen

_ «^-'
'^^:

■ P

HEATED
STREAM

.:-r\MM ■

Eichance

yi^-si/i^^iSi!^

Outflow Exchange

pcaag;-T^::iE'.C»c>v -

V^^gigSI, , , , ;

mmM^^

■vti«f''iic-,-;iaB.,

764

BISSON and DAVIS: PRODUCTION OF JUVENILE CHINOOK SALMON

?0-

y *■

. 0-

EI6-

Heated Stream

1 -- " T -- 1

1 1 -- T ' -- --

2Q. Control Stream

t +

i + +

\\\\\u

DJFMAMJJASONDJFMAMJJASOND
1972 1973

Figure 2. -Temperature conditions in the model streams.
Horizontal and vertical lines represent monthly means and
ranges, respectively.

glass observation windows in the control stream to
break; as a result, the entire riffle substrate was
exposed for several days while new windows were
being installed and some losses of periphyton and
invertebrates occurred. One window was also
replaced in the heated stream at this time, neces-
sitating exposure of one riffle.

Associated Flora

Vegetation surrounding the streams included
red alder, Alnus nibra, and apple, Malus sp. These
trees contributed leaves, catkins, and flowers as
well as a variety of terrestrial invertebrates.

Periphytic algae composed the bulk of living
plant material within the streams. The same plant
species were found in both heated and control
streams, although differences in biomass and
temporal succession occurred. The dominant
species from late spring to fall was Cladophora
glomerata, a filamentous green alga that attached
to large particles in the riffles and often trailed
into pool areas. Various diatoms also made up a
significant proportion of the flora. Two species
exhibited especially heavy seasonal blooms. In
early spring, filaments of a colonial diatom, Melo-
sira varians, covered both riffles and pools; this
species was noticeably more common in the control
than in the heated stream. In summer and fall
Synedra ulna became the dominant diatom, oc-
curring both in the water mass and among living
and dead algae on the bottom. Blue-green algae
were generally found in late spring and summer.
Calothrix and Nostoc were more abundant and
appeared earlier in the heated stream than in the
control. An unidentified dense moss colonized
some of the large cobbles in the riffles. Diatoms

and desmids, in addition to plant materials from
terrestrial sources, were common in the drift. The
desmid Clostermm lunula was abundant in spring
and early summer and was found to be an impor-
tant food resource for filter-feeding invertebrates.

Benthos and Drift

Benthic plants and animals were sampled

triweekly. Wire mesh baskets 20 x 20 x 6 cm

painted with nontoxic paint and having wood

bottoms were filled with substrate and placed

against supporting blocks in the riffles. The mesh

size (2 cm) was small enough to retain most of the

particles and large enough to allow movement of

invertebrates into and out of the baskets. Each

riffle in the streams contained four baskets placed

about 1 m apart from upstream to downstream

end. One basket was selected from a different

location in each riffle, the contents emptied into a

bucket, and all large particles cleaned with a

plastic scrub brush. The combined samples from

four baskets (0.16 m- total) were then collected in a

200-jam mesh bag. One sample was taken from a

pool in each channel, and collected material was

combined and preserved in 10% Formalin.

Drifting organisms were collected triweekly by
means of 333-iLim mesh drift nets (Anderson 1967)
that were suspended at the downstream end of the
riffles. Two nets were fished in each stream (one
per channel) for a 24-h period. Samples were
removed and preserved at approximately sunrise
and sunset so that diel differences in drift rates
could be measured. Current velocity was measured
at each sampling position and the amount of water
passing through the nets during an interval was
determined by multiplying this velocity by the
cross sectional area of the water (330 cm^) at the
mouth of the net. During periods when considera-
ble masses of leaves or algae were present in the
drift, usually late summer and fall, some clogging
took place and the volume of water entering the
nets was overestimated.

All samples were allowed to remain in Formalin
for 1 or 2 days, after which they were washed
briefly with water. Drift samples were transferred
directly to 70% ethanol prior to enumeration, while
bottom samples were first sorted to remove in-
vertebrates larger than 4 mm, and then subsam-
pled (10% by volume) and preserved in ethanol. All
organisms were measured to the nearest mil-
limeter by means of a metric grid placed on the
stage of the microscope. We assumed that no

765

FISHERY BULLETIN: VOL. 74, NO. 4

length changes occurred during preservation. The
number of individuals for each species in each size
interval was recorded for every sample.

The remaining 90% of a bottom sample-that
not sorted under magnification-was dried at 70°C
for 4 days and then ashed at 600° C. Ten percent of
its organic weight was arbitrarily assumed to have
been lost during preservation. Subtracting the
estimated biomass of small (<4 mm) invertebrates
within this subsample from the total loss of igni-
tion yielded the ash-free dry weight of filamentous
algae, some diatoms, detritus, and organisms too
small to be seen during the sorting process. Con-
version to energy units (kcal) was accomplished by
multiplying the plant-detritus biomass by 4.05, the
mean of five samples combusted in an oxygen
bomb calorimeter.

Computations of invertebrate biomasses were
based on live specimens collected from a nearby
stream, grouped according to size and species, and
weighed after drying 4 days at 70° C. Their aver-
age dry weights were converted to calories by
values obtained from Cummins and Wuycheck
(1971) or determined directly by calorimetry.
When no representatives of a certain size were
available, a value for that interval was estimated
by interpolation. Very similar forms were as-
sumed to have identical values. For bottom sam-
ples, the biomass (kcal/m-) of each size class of
each taxon was taken as the product of the number
of individuals in that class, the estimated caloric
value for individuals of that size, and the appro-
priate area conversion factor. The product of the
number of individuals and the caloric value was
divided by the total amount of water passing
through the nets to give biomass estimates per
unit volume (cal/m^) for the drift samples. Sum-
ming the values of all size intervals gave the total
caloric content for each taxon.

Fish

Fertilized chinook salmon eggs were obtained
from the Marion Forks Salmon Hatchery of the
Oregon Department of Fish and Wildlife. Eggs for
the 1972 experiment, taken 3 October 1971, were
from a single pair mating. Eggs used in the 1973
experiment, taken 1 October 1972, were obtained
by crossing three females with four males. This
was done in order to increase genetic heteroge-
neity among fish in the 1973 experiment. Follow-
ing fertilization, the eggs were transported im-
mediately to holding facilities where they were
incubated at 12°C.

In 1971, eggs were introduced into the streams
when they reached the eyed stage. They were
hatched in floating baskets and the fry were
released shortly before yolk absorption was com-
pleted. Owing to accelerated development in
warmer water, fish in the heated stream were
released sooner than those in the control, although
the initial number of individuals placed in the two
streams was identical (425). A 10-wk recoloniza-
tion period following repairs delayed introduction
of salmon until mid-March 1973, when 200 fry were
released simultaneously into each stream.

When the fish had reached approximately 0.4 g
wet weight, they were all removed from the
streams for measurement of individual length and
weight every 3 wk until an experiment was ter-
minated. From 5 to 20 fish were randomly drawn
from the populations for stomach analyses. A
blunted 22-gauge needle on a 5-ml syringe was
inserted through the esophagus of an anesthetized
fish into the anterior limb of the stomach. Several
milliliters of water were gently injected into the
stomach, forcing the contents out through the
mouth into a collecting beaker. The combined
whole food organisms and identified fragments of
each taxon were weighed to the nearest 0.1 mg,
and each taxon was assigned a percentage of the
diet based on its fraction of the total wet weight of
the sample.

Direct effects of the model stream temperature
regimes on chinook salmon growth rates at
different levels of food availability were studied in
concurrent experiments. Fish of the same paren-
tage and size as those in the model streams were
placed in insulated streamside troughs, where
they were fed live Tubifex at rations ranging from
near maintenance to near repletion. The troughs
received water directly from^ the model streams,
and temperature differences between the troughs
and streams were never greater than 0.3°C. Ten-
day growth experiments were carried out once
each season during 1973. Each experiment was
preceded by a 10-day period of acclimation to
temperature and ration size. Numbers of in-
dividuals tested at each ration level ranged from
10 to 20 depending upon fish size.

Average relative growth rates (Warren 1971) of
the salmon were calculated as:

ARG =

W, - n\

0.5 (W, + \V.,)'t
where ARG represented growth, Wi and W2

766

BISSON and DAVIS: PRODUCTION OF JUVENILE CHINOOK SALMON

represented the mean weights of the fish at the
beginning and end of the sampling interval, and t
was the sampling interval in days. Growth was
assumed to be linear over the relatively short 3-wk
period. Relative growth rates, which were essen-
tially the same as instantaneous growth rates,
were considered more appropriate for comparison
with relative food consumption rates.
Average biomass (B) was calculated as:

B =

B, + B-.

where B^ and Bo represented the total weights of
the fish at the beginning and end of the sampling
interval.

Production during each sampling interval was
calculated as the product of average relative
growth rate (ARG) and average biomass {B).

The conversion of wet weights to calories was
accomplished by relating caloric content of tissue
to condition factors of the fish, where condition

factor was taken as 100 times a fish's weight (g)
divided by the cube of its fork length (cm). Figure
9 of Warren et al. (1964:630), describing this
relationship for cutthroat trout, Salmo clarki, was
used for graphical estimates of calories per gram
of wet weight for juvenile chinook salmon.

RESULTS

Temporal Changes in Production

Total production of chinook salmon in the heat-
ed stream was less than half that of the control in
1972 (Table 1). During the following year, produc-
tion in the control stream was approximately 30%
higher than in the heated stream. Mortality was
greatest immediately after release into the
streams, with populations attaining fairly stable
levels by late summer. Population biomasses rose
during winter and spring, were highest during
late spring, and gradually declined through sum-
mer and fall. The mean annual biomass in the

Table l.-Mean production statistics of experimental chinook salmon populations. H = heated stream,

C = control stream.

Individual

Popu

lation

Mortality

size

rate

Biomass

Growth rate

Production

Tir

ne
rval

(kc

al)

(no.)

(%/day)
H C

(kca

l/m2)
C

(calkc
H

al/day)
C

(kca
H

il/m2)

inte

1972:

20 Dec.

-24 Jan.

0.31

302

1.61

4.12

1.85

0.27

25 Jan.

-14 Feb.

0.49

169

0.46

3.09

21.48

1.39

15 Feb.

- 7 Mar.

0.84

0.48

140

358

1.22

0.62

4.20

7.24

26.50

3.08

2.23

1.14

8 Mar.

-27 Mar.

1.18

0.92

112

271

0.74

0.81

5.14

8.62

16.00

33.03

1.73

5.42

28 Mar.

-15 Apr.

1.41

1.32

233

1.23

0.73

5.38

11.91

9.31

19.21

0.95

4.26

16 Apr.

- 6 May

2.18

2.01

184

1.63

1.39

5.32

14.04

23.80

47.42

2.27

5.89

7 May

-27 May

2.77

2.79

145

0.92

0.50

5.29

15.82

11.35

15.48

1.26

5.14

28 May

-16 June

3.61

3.36

125

0.48

0.84

5.80

17.61

13.17

9.27

1.53

3.27

17 June

- 7 July

4.36

3.75

105

0.88

0.75

6.25

16.97

8.96

5.22

1.18

1.86

8 July

-28 July

4.53

4.32

0.01

0.35

6.26

16.97

1.82

6.73

0.24

2.40

29 July

-22 Aug.

4.82

4.35

1.24

0.34

5.26

16.65

2.48

0.28

0.34

0.12

23 Aug.

- 8 Sept.

5.68

4.74

0.84

2.03

4.65

13.84

9.64

5.05

0.76

1.19

9 Sept.

- 3 Oct.

5.79

5.01

0.22

0.15

4.56

11.52

0.77

2.22

0.09

0.64

4 Oct.

-19 Oct.

5.61

5.76

0.74

0.61

4.15

11.87

-1.44

9.29

-0.13

1.65

20 Oct.

- 7 Nov.

6.15

5.73

0.35

0.01

3.88

12.01

4.83

-0.27

0.36

-0.06

8 Nov.

-30 Nov.

6.45

5.50

0.00

0.01

4.01

11.49

2.07

-1.78

0.19

-0.47

Total

14.66

32.45

Mean

4.84

13.33

1973:

16 Mar.

- 7 Apr.

0.98

1.02

187

185

0.57

0.66

7.14

7.48

14.49

12.70

2.38

2.19

8 Apr.

-26 Apr.

1.57

1.48

154

168

1.23

0.15

8.93

9.55

24.36

19.37

4.13

3.51

27 Apr.

-18 May

3.33

2.75

109

150

1.74

0.89

12.14

14.42

32.65

27.30

8.72

8.66

19 May

- 7 June

4.69

3.63

106

2.42

2.26

11.85

15.37

17.85

14.52

4.02

4.24

8 June

-27 June

6.92

4.98

1.09

1.75

10.82

12.33

19.31

15.68

4.16

3.87

28 June

-19 July

7.31

5.99

0.42

0.90

11.00

11.22

2.49

8.37

0.60

2.07

20 July

- 9 Aug.

6.87

6.96

0.30

0.01

10.31

12.07

-2.96

7.13

-0.64

1.81

10 Aug.

-30 Aug.

7.00

7.10

0.15

0.12

9.46

12.78

0.89

0.95

0.18

0.25

31 Aug.

-19 Sept.

6.76

7.89

0.17

0.75

9.07

12.61

-1.66

5.02

-0.32

1.33

20 Sept.

-10 Oct.

6.48

8.23

0.00

0.28

8.73

12.09

-2.01

2.01

-0.37

0.51

11 Oct.

-31 Oct.

6.35

9.00

0.86

0.78

7.58

11.36

-0.97

4.26

-0.15

1.02

1 Nov.

-21 Nov.

6.65

8.71

0.59

0.53

6.50

10.06

2.18

-1.56

0.30

-0.33

22 Nov.

-12 Dec.

6.07

8.66

0.91

0.20

5.49

9.08

-4.34

-0.27

-0.50

-0.05

Total

22.51

29.08

Mean

9.16

11.57

767

FISHERY BULLETIN: VOL. 74, NO. 4

heated stream was about twice as high in 1973 as
in 1972, while average biomasses were slightly
reduced following repairs in the control stream.
Peak production in both streams occurred from
April to June (Table 1, Figure 3), this being related
to the high growth rates that took place during
spring. Differences in production between the
streams, however, were related primarily to
higher population biomasses maintained in the
control stream than in the heated stream, rather
than to differences in growth rate.

Production of salmon in the heated stream
during the spring, 1973, was higher than in the
spring 1972 (Figure 3). The fish were stocked as fry
in 1973, whereas in 1972 they were introduced as
eyed eggs. The low average growth rate and
survival (Table 1) of fish reared in the heated
stream from the egg stage suggest that produc-
tion was influenced by conditions during early
development. Some individuals grew very rapidly
during their first few weeks of residence; others
apparently did not make the transition to feeding
in the heated stream and died from the effects of
starvation. Negative production occurred during
fall months, when many fish had stopped growing
and some were losing weight.

3Sr

30-

.2 20

S 15

10-

A - Meoied
a - Coo'fol

J F M AM JJASONDJ FMAMJ JASOND
1972 1973

Figure 3.-Cumulative production of juvenile chinook salmon
during 1972 and 1973.

Direct Temperature Effects on Growth

Relationships between average relative growth
rate and food consumption rate of juvenile chinook
salmon held in water from the model stream
(Figure 4) showed that differences between fish
held in heated and unheated water were greatest
at low rations and least at high rations. At low
rations, control individuals were most efficient; at
high levels, there was no appreciable difference

768

except during spring when the elevated tempera-
ture facilitated increased food consumption and
growth efficiency. The highest rations were close
to the maximum amount of food that the young
salmon would eat at one feeding in a day, and the
graphs for summer and fall indicate that max-
imum consumption declined as individuals' size
increased.

The relationships observed in the experiments
between temperature, ration level, and fish size
were consistent with the results of laboratory
studies of sockeye salmon, 0. nerka (Brett et al.
1969; Brett and Shelbourn 1975); coho salmon
(Averett 1969); and steelhead trout, Salmo gaird-
neri, (Wurtsbaugh 1973). At low levels of food
availability, increased metabolic requirements
associated with elevated temperature resulted in
reduced growth rates; at high levels of food
availability, growth rates were not appreciably
altered by thermal increases. If responses of
juvenile chinook to the range of ration levels in the
aquarium growth experiments approximated
growth of fish in the model streams at differing
consumption rates (Carline and Hall 1973), the
growth rate data of Table 1 suggest that during
most of the year the fish were feeding well below
their maximum possible consumption. Only during
certain periods in late winter and spring did
growth rates approximate the maximum rates
shown in Figure 4. From this we concluded that,

.. ,15

J — i — 6 — t~is — rt — <r

25 3

30 2-

15 1-
lO

»r

Food Consumption Role ( X / day I

Figure 4. -Seasonal changes in the growth rates of juvenile
chinook salmon. E.xperiments continued for 10 days and were
preceded by 10 days of acclimation to temperature and ration
size. Plotted values of growth rate at each feeding level were
based upon the following numbers of fish: winter - 20; spring - 20;
summer - 12; and fall - 10. Mean caloric contents (kcal) of the fish
at the beginning of each experiment were; winter - 0.59; spring -
1.26; and summer - 7.05; and fall - 8.37.

BISSON and DAVIS: PRODUCTION OF JUVENILE CHINOOK SALMON

during most of the year, the experimentally
elevated temperature contributed directly to the
reduced growth and production of the fish.

Disease

An unexpected indirect effect of elevated tem-
perature was apparent protection from infesta-
tion by an intermediate stage of the trematode
Nanophyetus salmincola, which was present in the
streams from late spring through fall. Infective
cercaria emerged from the snail Oxytrema silicula
to encyst in the skin and tissues of juvenile
Chinook as metacercaria. The distinction between
heavy vs. light infestation was made visually and
was somewhat arbitrary (Figure 5): conspicuous
bumps at the base of the caudal peduncle, darken-
ing of fins, and papules on the body surface were
considered symptoms of heavy infestation. While
the parasite was obviously present in 1972, it was
not until after its appearance in 1973 that at-
tempts were made to quantify its effects.

Infestation rates in the heated stream remained
low through summer and early fall and increased
until termination of the experiment. Heavy in-
festations were present in most of the control fish

HEATED STREAM

CONTROL STREAM

..°-^.

a— a

J J A S O N D

-D— L.9M

a. ^

v^^

J J A S O N D

19 73

Figure 5.-Infestation rates and weight differences of juvenile
Chinook salmon infested by metacercaria of Nanophyetus
salmincola.

soon after cercaria had begun emerging from the
snails. In addition, a greater difference existed
between the mean weights of heavily and lightly
infested individuals in the control stream than in
the heated stream. The impact of this parasite
thus appeared to be more severe in the control
than in the heated stream.

Food Availability

An understanding of changes in food availabili-
ty required: 1) that preferred food items be
identified, 2) that it be determined when they were
available for consumption, and 3) that their rela-
tive abundance was estimated under comparable
circumstances. In this study, the second require-
ment was met through observation; food organ-
isms became available only when they entered the
drift and then mainly during daylight. Unlike
many other salmonids, juvenile spring chinook
salmon placed in the model streams were never
seen feeding on invertebrates in the benthos. The
extent of feeding during darkness was not deter-
mined, but was believed to be small. Identical
sampling procedures were assumed to fulfill the
third requirement, although differential con-
sumption of food before it entered the drift nets
could have caused some error.

Oligochaetes were almost completely excluded
from the diet of large fish even though they
composed an important fraction of the drift (Table
2). Mollusca (exclusively Gyraulus sp.) and Tri-
choptera were comparatively large food items and
were consumed more readily by large fish than by
small fish. Ostracod Herpetocypris chevreuxi was
taken throughout the year in proportion to its
relative abundance, while Ephemeroptera and
Chironomidae— generally small organisms that
were usually numerous in the drift— were
preferred by smaller fish although these groups
were always major components of the diet. In
general, differences in food habits between
populations in the streams were related to
differences in the relative abundance of various
food groups. One exception was the greater con-
sumption of terrestrial forms (primarily aphids
and spiders) by fish in the heated stream, despite
approximately equal input of these invertebrates
into both streams.

Measurements of food organisms drifting dur-
ing daylight hours (Figure 6) were not well cor-
related with measurements of the biomass of
those organisms in the riffle benthos (Figure 7).

769

FISHERY BULLETIN: VOL. 74, NO. 4

Table 2.-Average percentages of different taxa (by weight) in the food of juvenile chinook salmon
compared with percentages of those organisms in the day drift (in parentheses). H = heated stream,
C = control stream.

Oligo

chaeta

Mollusca

Ostracoda

Collembola

Ephem
H

eroptera

Season

1972:

Winter

0( 45)

<1( 24)

K 4)

0( 5)

5( 4)

0( 0)

1(<1)

0( 4)

2( 4)

30 ( 2)

Spring

0( 43)

0( 37)

0( 13)

0( 4)

18( 7)

4( 2)

<1(<1)

< 1(<1)

11( 3)

45( 24)

Summer

1( 34)

<1( 56)

0( 5)

0( 2)

32( 28)

3( 3)

<1( 1)

0(<1)

10(<1)

10( 8)

Fall

<1( 7)

0( 13)

2( 18)

0( 4)

7( 4)

4( 5)

7( 1)

4( 1)

12( 15)

4( 6)

1973:

Spring

13( 14)

15( 43)

0( 6)

<1(<1)

4( 7)

1(<1)

<1(<1)

0(<1)

22( 11)

25( 6)

Summer

6( 3)

K 8)

11( 7)

1( 2)

12( 8)

4( 11)

<1(<1)

10( 8)

30( 18)

Fall

<1( 6)

^1( 1)

54( 33)

4( 5)

4( 4)

<1( 1)

6( 2)

4( 2)

4( 3)

18( 19)

Plecoptera

Trichoptera

Chironomidae

Terrestrials

Miscellaneous

Season

1972:

Winter

25( 9)

52( 42)

0( 0)

51 ( 26)

16( 3)

14( 10)

2( 14)

<1(<1)

< 1(<1)

Spring

12( 7)

6( 3)

3(<1)

5( 21)

27( 24)

36 ( 6)

27( 2)

7( 2)

2(<1)

1(<1)

Summei

K 1)

4(<1)

<1(<1)

9( 1)

49( 65)

67( 36)

2( 2)

5{ 8)

2( 5)

2( 1)

Fall

2( 1)

33( 13)

21( 1)

25 ( 4)

33( 39)

27( 33)

14( 10)

1( 18)

<1( 4)

K 1)

1973:

Spring

<1(<1)

1(<1)

9(<1)

3(-^1)

44( 30)

■49( 32)

5( 23)

2( 16)

2( 8)

3( 1)

Summer<l( 3)

11( 2)

11{ 1)

12( 6)

38{ 65)

26( 36)

8( 2)

11( 16)

2( 3)

4( 2)

Fall

<1( 1)

9( 5)

1( 4)

33( 10)

11( 35)

25( 55)

20( 6)

7( 2)

1( 6)

K 1)

J FMAMJJASONDJ FMAMJ JASOND
1972 1973

Figure 6.-Seasonal changes in the biomass of food organisms
present in the day drift. Each point is the mean of two triweekly
samples. ^

Moreover, seasonal patterns in drift differed
greatly between 1972 and 1973, with both streams
exhibiting higher drift biomasses during the
second year than during the first. Although
benthic biomasses were significantly greater in
the control than in the heated stream, (P<0.001,
paired ^-test), these differences were often not
translated into drift; in fact, during the latter part
of 1972 and spring 1973, more food was available in
the heated stream. No explanation was found for
increased drift in 1973 relative to 1972, but it
appeared that increased food availability in 1973

-D

Q£

A- Heated
D - Control

J FMAMJ J A SON D J FMAMJ JASOND
1972 1973

Figure 7.— Seasonal changes in the biomass of food organisms
present in the riffle benthos. Each point is the mean of two
triweekly samples.

resulted in more growth, higher biomasses, and
increased production of fish in the heated stream.
Why production in the control stream population
did not reflect the greater abundance of food is not

770

BISSON and DAVIS: PRODUCTION OF JUVENILE CHINOOK SALMON

Table 3.- Annual average biomasses (cal/m^), drift rates (cal/m^), and drift ratios of selected aquatic taxa, excluding
winged adults. Drift ratios were calculated according to the formula (day drift/riffle biomass) x 10-''. Asterisks denote
values for the heated stream that were significantly different (P<0.05, single classification analysis of variance) from
the control.

1972

1973

Heated

Contro

Heated

Control

Riffle

Day

Drift

Riffle

Day

Drift

Riffle

Day

Drift

Riffle

Day

Drift

Taxon

biomass drift

ratio

biomass drift

ratio

biomass drift

ratio

biomass drift

ratio

Oligochaeta

8,532

0.117

1.74

10,698

0.106

1.43

4,112

0.036

1.26

4,034

0.043

1.56

Mollusca'

778

0.031

2.71

127

0.008

6.98

906

0.104

13.18

132

0.029

29.08

Ostracoda

703

0.055

23.63

177

0.010

8.87

273

0.036

29.23

218

0.030

28.40

Ephemeroptera

308

0.024

12,14*

7,612

0.011

0.23

430

0.031

8.46

4,150

0.102

3.12

Plecoptera

378

0.009

7.95*

5,966

0.033

0.59

173

0.002

20.37*

3,717

0.035

1.32

Trichoptera

516

0.001

0.22

1,643

0.018

0.36

570

0,010

5.86

1,569

0.039

2,48

Chironomidae

1,202

0.021

3.26*

1.661

0.011

0.95

2,549

0.052

2.20

1,979

0.045

3.67

'Gyraulus sp.

known, although severity of infestation by Nano-
phyetufi was not compared over the 2 yr and may
have been more serious in 1973.

In 1972, drift ratios (the ratio of drift to
biomass) of several invertebrate taxa were higher
in the heated stream than in the control (Table 3).
The next year some of the drift ratios increased,
and although many were higher in the control
stream, the differences were not statistically
significant. Of taxa showing increased drift ratios
in the heated stream, Ephemeroptera and Plecop-
tera were most consistently influenced by elevated
temperature. In 1972, Chironomidae also exhibited
a significantly greater tendency to drift in the
heated stream than in the control. These three
groups were important components of both the
day drift and the diet of juvenile salmon and often
contributed to the greater availability of food in
the heated stream than in the control during
certain periods.

Fewer macroinvertebrate taxa were present in
the heated stream than in the control. Paired
f-tests indicated that number of taxa were
significantly different in both riffles (treatment
mean = 21, control mean = 34; P<0.001) and pools
(treatment mean = 16, control mean = 19,
P<0.01). Most of those taxa that were unique to one
stream or the other were very rare and contributed
little to fish production. Major biomass differences
arose because many taxa had greater population
densities in the control while only a few fared
better in the heated stream. The several taxa that
did exhibit higher biomass in the heated stream
were very abundant and tended to dominate the
bottom fauna to a greater extent than did common
taxa in the control. The two most abundant in-
vertebrates in the heated stream were Oxytrema
silicula in the riffles and Limnodrilus sp. in the
pools. Neither of these two species was consumed

in significant quantities by the young salmon;
thus, increased dominance in the heated stream
did not give rise to greater food availability.

Periphyton Biomass and Sedimentation

Plants and detritus were significantly more
abundant in the heated stream (P<0.001, paired
f-test) than in the control (Figure 8). The greater
amounts of plants and detritus in the heated

Meo'ed

Control

RIFFLES

O
O

- 1

^ 20

c
0

r 15

POOLS

I ■ I ' 1 I I — 1_

JFMAMJJASONDJFMAMJJ A S O N D
1972 1973

Figure 8.-Biomasses of plants and detritus in riffles and pools of
the model streams. Each point is the mean of two triweekly
samples.

771

FISHERY BULLETIN: VOL. 74, NO. 4

stream than in the control were due to the high
densities of filamentous algae in the riffles and the
considerable accumulation of organic detritus in
the pools. Increased primary production associated
with elevated temperature in laboratory streams
has been measured by Kevern and Ball (1965) and
Phinney and Mclntire (1965). The dominant algal
species in our model streams, Cladophora glome-
rata, grows rapidly at high temperatures
(Whitton 1971; Adams and Stone 1973).

Heavy growths of algae on the riffles apparently
accelerated sedimentation rates in the heated
stream (Table 4) by acting as filters to trap and
consolidate fine particles introduced with ex-
change water. In the pools, where filamentous
algae did not grow, fine sediment levels in both
streams were similar. By indirectly enhancing
sediment accumulation, elevated temperature
probably had an important effect on the numbers
of food organisms available to salmon in the
heated stream. Hynes (1960) described how silta-
tion alters the habitat of many invertebrates, with
the result usually being a reduction in benthic
biomass (Cordone and Kelly 1961). Greatly
reduced mean annual biomasses of Ephemerop-
tera, Plecoptera, and Trichoptera in the heated
stream (Table 3) compared with the control sug-
gest that these groups were influenced by the
amount of fine sediments in the substrate, and
these insects were often preferred food items of
the fish (Table 2).

Table 4. -Levels of fine sediments, expressed as grams dry
weight per square meter, in the model streams during May 1974.
The figures in parentheses refer to the amount of time that had
elapsed since a major disturbance to the riffles.

Parlic

le size (mm)

Item

0.175-1

0.088-0.175

0.088

Riffles:

Control (17

mo)

169

Heated (17

mo)

147

943

Heated (31

mo)

167

1,443

Pools:

Control

1,219

1,746

Heated

1,064

1,728

DISCUSSION

Our study was designed to examine the effects
of elevated temperature on the production of
juvenile chinook salmon. The constantly elevated
temperature was not meant to simulate a par-
ticular type of thermal increase, but was within
the range of temperature elevations caused by
heated discharges into running waters (Wilber

1969, Parker and Krenkel 1970), irrigation runoff
(Eldridge 1963), and removal of streamside vege-
tation (Brown and Krygier 1970). It was also
within the limits of temperature increase legally
allowed by some regulations (Burd 1969).

Both direct and indirect temperature effects
influenced chinook salmon production, but the
magnitude of these effects varied seasonally.
Production was high in spring because tempera-
ture was in a range that was favorable to growth,
parasitism had not yet become an important
factor, and the small fish were able to efficiently
exploit available food. Summer was generally a
period of declining production because high tem-
peratures resulted in an increase in maintenance
requirements and, for the control stream, because
parasites had attacked the majority of the
population. Low production during late summer
and fall was associated with high levels of infes-
tation and the ineffectiveness of large fish in
exploiting small organisms that were abundant in
the drift.

The lack of correlation that existed between
growth rates (Table 1) and food availability
(Figure 6) may have been related to the species
composition of drifting invertebrates. A high
percentage of summer and fall drift was composed
of very small forms such as oligochaetes (Nais
communis) and chironomids (Table 2). During
those seasons, tiny organisms were not preferred
food items of the young salmon, which were larger
and less numerous than during the spring. High
growth rates exhibited by fish during winter and
spring when drift rates were comparatively low
suggest that smaller, more abundant fish were
able to utilize the entire range of sizes of inverte-
brate species that left the substrate. It was im-
possible to determine whether food size prefer-
ence affected fish in the two streams identically,
but based on overall invertebrate composition
(Table 3), taxa containing species of large size
(Ephemeroptera, Plecoptera, Trichoptera) were
more abundant in the control than in the heated
stream. This was reflected in higher growth rates
of salmon in the control than in the heated stream
during summer and fall. Clearly, more intensive
examination of the relationship between prey size
and prey selection by salmonids is needed.

Low benthic invertebrate biomasses in the
heated stream were associated with increased
sedimentation rates and reduced numbers of taxa.
Iverson (1972) suggested that the poor success of
certain invertebrates in the heated stream was

772

BISSON and DAVIS: PRODUCTION OF JUVENILE CHINOOK SALMON

due to their being cold-adapted species. No large
scale mortality of larvae or pupae was detected in
the heated stream, even during summer months.
However, very early developmental stages and life
history patterns may have been altered (Macan
1961a, b; Hynes 1970).

The tendency of certain invertebrates in the
heated stream to enter the drift in greater
proportion to their benthic biomasses (Table 3)
was probably related both to elevated temperature
and to fine sediment levels. Increased drift as-
sociated with increasing temperature was de-
scribed for certain invertebrates by Miiller (1963),
Waters (1968), and Pearson and Franklin (1968). In
other studies, significant positive correlations
between drift and temperature have not been
detected (Bishop and Hynes 1969; Wojtalik and
Waters 1970; Muller 1970; Reisen and Prins 1972).
Experimental additions of sediments to a stream
were found by Rosenberg and Weins (1975) to
significantly increase the drift of some inverte-
brate taxa and to have inconsistent effects on

others.

Although the influence of elevated temperature
on the production of juvenile chinook salmon was
complex, we were able to identify both beneficial
and harmful effects. The fish benefited in several
ways. First, the temperature increase may have
stimulated higher consumption rates when suit-
able food was very abundant, although this con-
dition was rarely achieved. Second, higher temper-
atures afforded protection from infestation by a
trematode parasite, which heavily infested the
majority of individuals in the control stream.
Third, certain invertebrates may have been
stimulated to enter the drift and thus became
more available as food. Fish were harmed in at
least two ways. First, growth efficiencies were
lowered at all but the highest consumption levels.
Second, despite high drift ratios of some taxa, food
availability was generally reduced because
preferred food organisms were much less abun-
dant in the substrate of the heated stream than in
the control. The net result was that salmon
production in the heated stream was about 50%
less in 1972 and 25% less in 1973 compared with the
unheated stream.

ACKNOWLEDGMENTS

Support for the study was provided by the Office
of Water Research and Technology, U.S. Depart-
ment of Interior, under provisions of Public Law

88-379. C. E. Warren, C. B. Schreck, N. H. Ander-
son, and C. D. Mclntire offered suggestions for the
manuscript, and C. D. Mclntire and H. K. Phinney
aided in identifying stream flora. David Neiss,
Tim Joyce, Howard Worley, John Toman, Steve
Ross, Eric Johansen, Jean McRae, and Mary
Buckman assisted with field and analytical work.

LITERATURE CITED

Adams, M. S., and W. Stone.

1973. Field studies on photosynthesis of Cladophora glome-
rata (Chlorophyta) in Green Bay, Lake Michigan. Ecology
54:853-862.
Anderson, N. H.

1967. Biology and downstream drift of some Oregon
Trichoptera. Can. Entomol. 99:507-521.
AVERETT, R. C.

1969. Influence of temperature on energy and material
utilization by juvenile coho salmon. Ph.D. Thesis, Oregon
State Univ., Corvallis, 75 p.
Bishop, J. E., and H. B. N. Hynes.

1969. Downstream drift of the invertebrate fauna in a
stream ecosystem. Arch Hydrobiol. 66:56-90.
Brett, J. R., and J. E. Shelbourn.

1975. Growth rate of young sockeye salmon, Oncorhynchus
nerka, in relation to fish size and ration level. J. Fish. Res.
Board Can. 32:2103-2110.
Brett, J. R., J. E. Shelbourn, and C. T. Shoop.

1969. Growth rate and body composition of fingerling
sockeye salmon, Oncorhynchus nerka, in relation to
temperature and ration size. J. Fish. Res. Board Can.
26:2363-2394.

Brown, G. W., and J. T. Krygier.

1970. Effects of clear-cutting on stream temperature.
Water Resour. Res. 6:1133-1139.

Bird, R. S.

1969. Water quality standards for temperature. In F. L.
Parker and P. A. Krenkel (editor), Engineering aspects of
thermal pollution, p. 72-77. Vanderbilt Univ. Press,
Nashville.

Carline, R. F., and J. D. Hall.

1973. Evaluation of a method for estimating food consump-
tion rates of fish. J. Fish. Res. Board Can. 30:623-629.
Cordone, a. J., AND D. W. Kelley.

1961. The influences of inorganic sediment on the aquatic
life of streams. Calif. Fish Game 47:189-228.

Cummins, K. W.

1962. An evaluation of some techniques for the collection
and analysis of benthic samples with special emphasis on
lotic waters. Am. Midi. Nat. 67:477-504.

Cummins, K. W., and J. C. Wuycheck.

1971. Caloric equivalents for investigations in ecological
energetics. Mitt. Int. Ver. Theor. Angew. Limnol.
18:1-158.

Eldridge, E. F.

1963. Irrigation as a source of water pollution. J. Water
Pollut. Control. Fed. 35:614-625.

Hynes, H. B. N.

1960. The biology of polluted waters. Liverpool Univ. Press,
Liverpool, 202 p.

1970. The ecology of running waters. Univ. Toronto Press,
Toronto, 555 p.

773

FISHERY BULLETIN: VOL. 74, NO. 4

IVERSON, R. A.

1972. Effects of elevated temperature on juvenile coho
salmon and benthic invertebrates in model stream
communities. Ph.D. Thesis, Oregon State Univ., Corvallis,
98 p.
Kevern, N. R., and R. C. Ball.

1965. Primary productivity and energy relationships in
artificial streams. Limnol. Oceanogr. 10:74-87.
Macan, T. T.

1961a. Factors that limit the range of freshwater animals.

Biol. Rev. (Camb.) 36:151-198.
1961b. A review of running water studies. Int. Ver. Theor.
Angew. Limnol. Verh. 14:587-602.
MULLER, K.

1963. Diurnal rhythm in organic drift of Gammarus

puler. Nature (Lond.) 198:806-807.
1970. Die Drift von Insektenlarven in Nord - und
Mitteleuropa. Osterr. Fisch. 23:111-117.
Parker, F. L., and P. A. Krenkel.

1970. Physical and engineering aspects of thermal
pollution. CRC Press, Cleveland, 100 p.
Pearson, W. D., and D. R. Franklin.

1968. Some factors affecting drift rates of Baefis and
Simuliidae in a large river. Ecology 49:75-81.
Phinney, H. K., and C. D. McIntire.

1965. Effect of temperature on metabolism of periphyton
communities developed in laboratory streams. Limnol.
Oceanogr. 10:341-344.
Reisen, W. K.. and R. Prins.

1972. Some ecological relationships of the invertebrate drift

in Praters Creek, Pickens County, South Carolina. Ecology
53:876-884.
Rosenberg, D. M., and A. P. Weins.

1975. Experimental sediment addition studies on the Harris
River, N.W.T., Canada: the effect on macro-invertebrate
drift. Int. Ver. Theor. Angew. Limnol. Verh. 19:1568-
1574.
Warren, C.E.

1971. Biology and water pollution control. W. B. Saunders
Co., Phila., 434 p.
Warren, C. E., J. H. Wales, G. E. Davis, and P. Doudoroff.
1964. Trout production in an experimental stream enriched
with sucrose. J. Wildl. Manage. 28:617-660.
Waters, T. F.

1968. Diurnal periodicity in the drift of a day-active stream
invertebrate. Ecology 49:152-153.

Whitton, B. a.

1971. Terrestrial and freshwater algae of Aldabra. Philos.
Trans. R. Soc. Lond., Ser. B., Biol. Sci. 260:249-255.

WlLBER,C.G.

1969. The biological aspects of water pollution. Charles C.
Thomas, Springfield, 111., 296 p.

WOJTALIK, T. A., AND T. F. WATERS.

1970. Some effects of heated water on the drift of two
species of stream invertebrates. Trans. Am. Fish. Soc.
99:782-788.

WURTSBAUGH, W. A.

1973. Effects of temperature, ration, and size on the growth
of juvenile steelhead trout, Salmo gairdneri. M.S. Thesis,
Oregon State Univ., Corvallis, 69 p.

774

REPRODUCTIVE CYCLE, FECUNDITY, AND SEX RATIOS OF
THE RED PORGY, PAGRUS PAGRUS (PISCES: SPARIDAE)

IN NORTH CAROLINA

Charles S. Manooch, IIP

ABSTRACT

Macroscopic examination of gonads and gonad indices demonstrated that March and April were the
peak spawning months in Raleigh and Onslow bays, N.C. Ripe fish were collected over irregular bottom
from January to April in water ranging from 21 to 100 m in depth. Bottom temperatures during the
spawning period ranged from 16.4° to 21.5°C. Three predictors of fecundity, total length, weight, and
age were evaluated and regression equations derived. Fish weight proved to be the most precise
predictor of fecundity: In fecundity = 1.7369 + 1.5178 (In weight of the fish) where fecundity is the
total number of eggs in both ovaries. Fecundity estimates ranged from 48,660 for a 304-mm (390-g) red
porgy to 488,600 for a 516-mm (1,783-g) fish. Although some individuals reached sexual maturity at age
II, most spawn for the first time at age III. Chi-square tests revealed a significant departure from the
expected 1:1 sex ratio when data were stratified by month, year, and size. Females were encountered
more frequently each month for all 3 yr, and in the smaller size intervals.

The red porgy, Pagrus pagrns Linnaeus, is one of
the most important demersal marine fishes taken
by recreational anglers fishing from headboats-
between Cape Hatteras, N.C, and Charleston, S.C.
In 1972 and 1973, 513,700 red porgy weighing 1.3
million pounds were taken by this sport fishery
(Sekavec and Huntsman 1972; Huntsman 1976). In
spite of the importance of the species, published
information on the red porgy in the western
Atlantic is scarce. Dias et al. (1972) described the
length-weight relationship for Pagrus collected off"
South Carolina; Ciechomski and Weiss (1973)
reported on egg, embryo, and larval development
of red porgy from the Argentine Sea; and Man-
ooch et al. (in press); Manooch (in press), discussed
the taxonomic status and the food habits of P.
pagrus, respectively.

This study investigated reproduction of red
porgy in North Carolina to determine: 1) spawning
season, 2) size and age of females at sexual matu-
rity, 3) prediction equations for estimating fecun-
dity, 4) sex ratios by month and size, 5) spawning
ecology, and 6) a description of the eggs and
young. This research is part of a National Marine
Fisheries Service project which is studying the
bottom fishes of the outer continental shelf of the
Carolinas.

'Atlantic Estuarine Fisheries Center, National Marine Fish-
eries Service, NOAA, Beaufort, NC 28516.

-Headboats are those that charge for a day's fishing on a per
person basis.

MATERIALS AND METHODS

Length, weight, sex, stage of gonad develop-
ment, and gonad length and weight were recorded
for fish sampled from North Carolina headboats
and by experimental fishing aboard the RV 0ns-
lotv Bay from 1972 to 1974. Gonads were preserved
in 10% Formalin-^ and macroscopically examined to
determine maturity using modified criteria from
Orange (1961): Stage 1-S: infantile, gonads small
and ribbonlike (sex determination by gross ex-
amination not possible); Stage 1: immature, go-
nads elongated, slender, but sex discernible by gross
examination; Stage 2: early maturing, gonads
slightly enlarged, individual ova not visible to
naked eye; Stage 3: late maturing, gonads en-
larged, individual ova visible to naked eye; Stage 4:
ripe, ovary greatly enlarged, many ova trans-
lucent and easily dislodged from follicles or loose
in lumen of ovary; and Stage 5: spawned, includes
recently spawned fish with mature ova occurring
as remnants in various stages of reabsorption.

Time of spawning based on 243 females was
determined by using: 1) the gonad index (G.I.) of
Schaefer and Orange (1956), and 2) the index:
lOOG.W./F.W., where G.W. is the fresh gonad
weight to the nearest 0.01 g and F.W. is the body
weight of the fish to the nearest 1.0 g. Mean values

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

775

FISHERY BULLETIN: VOL. 74, NO. 4

of these indices were plotted monthly, to thus
indicate duration and peak of spawning, and age
and size at sexual maturity. The linear regressions
fecundity on length, weight, and age were cal-
culated based on mature (Stage 4) ovaries from 50
females (ages II-IX, 304 to 520 mm TL) collected
from January through March for the years 1973
and 1974. One ovary randomly selected from each
pair was blotted dry and weighed to the nearest
0.01 g. The selected ovary was crumbled and all
ovarian tissue removed. The eggs were then
filtered, blotted dry, and weighed. One sample
from each ovary of 0.2-0.4 g was weighed to the
nearest 0.001 g and placed in a 6 x 6 counting grid
and all ova were counted. The formula:

(W,){w)

was used to estimate the number of eggs in the
ovaries, where Y = total number of eggs in both
ovaries, W = weight of both ovaries, W^ = weight
of selected ovary, W^' = weight of ovary after
removal of ovarian tissue, iv = weight of sample,
and // = number of eggs in the sample (Lassiter
1962).

RESULTS AND DISCUSSION

Sexual Maturity

Ovary condition progressed from ripe. Stage 4,
dominant from January through March, to
spawned. Stage 5, dominant from May through
June, indicating that peak spawning occurred in
March and April (Figure 1). Ovaries collected in
April and May were flaccid and showed resorption
of eggs. By June all of the fish were early matur-
ing. The ovaries gradually became more firm after

Gonod Ind.. (°°,"°t(T' * '°°) "=423
Gonod lnd«« I r* X 10 I N =242

... I I I I I I

SEPI OCT NOV DEC JAN FEB MAD API! MAT JUN JUl AUG

Figure l.-Mean monthly gonad indices for female Pagrus
pagrus collected from Onslow Bay, N.C., 1972 to 1974.

resorption in early summer and little change in
gonad condition was noted in late summer and
early fall. Maturation of ovaries occurred between
October and January. Stratification of the sexual
maturity data by month supported the hypothesis
of late winter to early spring spawning. Approx-
imately 23% of the fish examined in January were
late maturing and 77% were ripe. By February,
12.5% were classified as late maturing, and 87.5%
were ripe. The first spawned (Stage 5) fish were
collected in March and their frequency of occur-
rence increased to 60.5% in April (Figure 2).
Walker (1950) reported ripe P. pagrus in January
and February off North Carolina, and Ranzi (1969)
found that they were sexually mature from April
to June in the Mediterranean Sea off Algeria.

Early maturing and ripe stages of males were
easily discernible by gross examination of the
testes, but the late maturing and ripe classes were
difficult to separate. Milt could be pressed from the

100
80-
60-
40-
20-
0

^^. ,

Late Maturing, Stage 3
N = 31

Zi
O

>-
u

Ripe, Stage 4
N=77

Spawned, Stage 5
N = 33

Maturing , Stage 2
N=91

MONTHS

Figure 2.-Percentage of female red porgy at various stages of
sexual maturity, collected in Onslow and Raleigh bays, N.C., by
month.

776

MANOOCH: REPRODUCTIVE CYCLE OF RED PORGY

central canal of testes from January through
April.

Female red porgy were separated into two
maturity classes: Immature fish, and mature
(maturing, ripe, and spawned). No individuals less
than 260 mm and all fish greater than 360 mm were
sexually mature (Table 1). The linear regression of
percent maturity (iO on total length (X):

Y = -211.2946 + 0.8576X, r = 0.94,

was significant at cc = 0.01. Half the females
were mature at 304 mm. By inserting age data
(Manooch 1975) to the graph, age at sexual matu-
rity was determined. Regression of age with
length suggests that none of the age I fish, 37% of
the age II, 81% of the age III, and 100% of the age
IV fish were mature. Some age II and III females
apparently showed the characteristic, seasonal
maturation of ovaries but did not spawn the first
year, because several specimens had ovaries con-
taining absorbed ova during the peak spawning
period.

Table L-Number and percentage of female red porgy, grouped
into 20-mm size categories, staged as immature and mature
(maturing, ripe, and spawned) off North Carolina 1972-74.

Total length

Immature

(mm)

(no.)

<220

220-239

240-259

260-279

280-299

300-319

320-339

340-359

360-379

380-399

400-419

420-600

Mature
(no.)

Mature
(%)

Total

214

502

0.0
0.0

0.0

33.3

15.4

58.3

87.9

87.2

100.0

soo-

z
<

a
z

300-

200-

100-

300

450 500 550

TOTAL LENGTH IN MILLIMETERS

500-

< 400-

2 300-

Z 200-

In F*cundity = l 7369 + t 5178 (Ln Wl)
,' = 70

I I

I I ' — I— ' — r

90C
WEIGHT IN GRAMS

- 1 . I I I

300 500 700 900 IIOO 1300 1500 1700 1900

Figure 3.-Relationship between fecundity and two predictors:
(top) length and (bottom) weight of 50 red porgy collected in
Onslow Bav, N.C.

In Fecundity =
In Fecundity =

1.7369 + 1.5178(ln Wt),
r2 = 0.70 and
-14.1325 + 4.3598(ln TL),
r~ = 0.66.

Fecundity

Regression analyses indicated total length,
weight, or age could be used to predict fecundity of
red porgy, but weight proved to be the best
predictor of fecundity (r- = 0.70) and had the
lowest error mean square. Combinations of two
independent variables, weight and length, im-
proved predictability only slightly, therefore,
separate equations were derived by using weight
on fecundity, and length on fecundity. The equa-
tions describing the relationships (Figure 3) and
coefficients of determination (r^) are:

The 95% confidence limits have also been calculat-
ed. Predicted fecundity ranges from 48,660 eggs
for fish 304 mm TL and 390 g in weight to 488,600
ova for fish 516 mm TL and 1,783 g. Theoretically, a
600-mm red porgy which is not uncommon in the
sport catch, could produce approximately 943,000
eggs if maximum ova production is not obtained at
a smaller size.

Sex Ratios and Hermaphroditism

Sex of 736 red porgy collected in 1972, 1973, and
1974 was grouped by year and month, and data

777

FISHERY BULLETIN: VOL. 74, NO. 4

revealed females to be more abundant in the catch
than males (Table 2). The sex ratio was not 1:1
males to females as hypothesized but actually 1:2.1
when the years were combined. Data for each year
analyzed separately also provided significant
deviations from expected. The sex ratio for each
year was 1:2.1, 1:1.9, and 1:3.3 for 1972, 1973, and
1974, respectively (Table 2). The overall, higher
deviation from 1:1 for 1974 is because most of the
fish were collected in late winter and spring of
that year, months which reflected the greatest
deviation from 1:1. Of all months examined, only
August, September, and November were
nonsignificant, revealing equal number of males
and females. The ratio for October could not be
tested because of insufl^cient data. During the
spawning season, chi-square values were very high
and perhaps reflect monosexual schooling.

Sex ratios for males and females grouped into
50-mm length intervals had significant departures
from the expected 1:1 ratio for most size categories
(Table 3). In general, females predominated in the
smaller size classes, whereas males predominated
in the larger size classes. The nonsignificant value
for the smallest size interval is probably unrealis-
tic since the sample is very small and the sequen-
tial intervals are highly significant in favor of
females.

Both protandrous and protogynous hermaph-
roditism are relatively common among the sparids
(D'Ancona 1950, 1956). Pagrtis pagrus collected
from the west coast of Florida appear to display
protogynous hermaphroditism although data

Table 2.-Number of male and female red porgy collected by
month during 1972, 1973, and 1974 with chi-square values
obtained from testing a 1:1 sex ratio in each month (a), and each
year (b).

Table 3.— Number of male and female red porgy grouped into
50-mm size categories with chi-square values assuming a 1:1 sex
ratio.

Year

Total" d

Month

1972

1973

1974

f X^

May

9:12

17:41

0:5

26:58 2 12.08**

June

16:22

26:45

—

42:67

5.74*

July

9:27

27:44

—

36:71

11.44**

Aug.

11:15

30:33

—

41:48

.55

Sept.

3:9

21:26

—

24:35

1 2.04

Oct.

1:9

—

1:9

Nov.

2:3

12:15

—

14:18

.50

Dec.

5:15

3:11

—

8:26

9.52**

Jan.

—

0:2

9:22

9:24

6.82**

Feb.

—

5:12

8:32

13:44

16.86**

Mar.

—

5:26

4:32

9:58

17.92**

Apr.

—

6:29

10:10

16:39

9.62**

Totalb

56:112

152:284

31:101

239:497

18.6**

39.96**

37.12**

90.44**

* __

P<0.05.

• • =

P<0.01.

Length

Male

Female

Total

<300

3.57

300-350

47.28**

351-400

157

167

129.40**

401-450

161

209

61.10**

451-500

124

207

8.12**

501-550

7.38**

551-600

3.76

>600

—

Totals

240

497

737

* = P<0.05.
** = P<0.01.

available are insuflficient for quantitative descrip-
tion (D. S. Beaumariage pers. commun.).

The predominance of females at smaller size
intervals in this study and discovery of individuals
with both ovarian and testicular tissue supports
the theory of protogyny. Although hermaphroditic
red porgy were found by macroscopic examina-
tion, only 16 specimens of the 752 examined (2%)
contained both male and female gonadal tissues.
Hermaphroditic red porgy ranged in size from 325
mm to 424 mm TL (x = 400 mm); possibly the
length range over which sexual transition takes
place. In each fish the ovaries were dominant with
only redundant testicular tissue present. From
preliminary studies with red porgy in the Gulf, M.
A. Moe (pers. commun.) reported that the male
portion of the gonad develops in the muscular
tunica of the gonad wall and eventually completely
takes over the gonad.

Spawning

Ripe red porgy were collected over irregular
bottom from January through April in water
depth ranging from 21 to 100 m and bottom
temperatures of 16.4° to 21.5°C (Figure 4). Pagrus
pagrus spawns from December through January
in the Argentine Sea when water temperature is
approximately 20° to 21°C (Ciechomski and Weiss
1973).

The relationship of the gonad index to photo-
period and bottom temperature were plotted
monthly (Figure 4). By inspecting this figure one
could conclude that photoperiod is more directly
correlated to gonad maturation and spawning.
Similarly, gonad maturation of red grouper,
Epinephelus mono, another demersal reef species,
was unrelated to bottom temperature (Moe 1969).
Harrington (1956) demonstrated the importance
of photoperiod to gonad maturation and spawning

778

MANOOCH; REPRODUCTIVE CYCLE OF RED PORGY

2 Of

O e£

30n

20-

10-

Bottom Temperature (°C)

I I 1 1 r

Z 3

'H o

> ^

< z

o -

14.0-

12.0-

/ \

Photoperiod

10.0-

1 1 1

1 1 1 1 1 1

1 1 1

<
z
o
o

6.0-

4.0

2.0-

Gonad Index

-T-
M

A M

I I

J J

-I r

A S

—\ 1 1

O N D

MONTHS

Figure 4.-Mean gonad indices for female Pagrus pagriis for
each month compared with photoperiod and bottom
temperatures.

for the banded sunfish, Enneacanthus obesus.
Pagrus pagrus spawns between January and
April, when photoperiod increases rapidly, but
when bottom temperatures fluctuate irregularly.
Gonads were in spent and resting stages during
maximum photoperiod, May to August, and began
developing as photoperiod decreased. The graphs
suggest that seasonal increase in photoperiod in
late winter and early spring initiates final matu-
ration of ovaries and ultimately, the spawning of
P. pagrus.

Eggs and Young

Red porgy eggs are pelagic, spherical, without
appendages and contain a single oil droplet. Pre-
served eggs were generally yellow to orange in
color, they measured 0.31 to 0.94 mm in diameter
and the oil droplet was 0.20 to 0.32 mm in diameter.

This size description is similar to the unfertilized
eggs of another sparid, Stenotomus chrysops,
which were 0.66 to 0.95 mm and had an oil droplet
0.17 to 0.40 mm in diameter (Finkelstein 1969). I
induced three females (355-560 mm TL) to release
ova in aquaria in March 1975. These eggs appeared
transparent and were noticeably larger than those
described above. Since I considered these eggs to
be most representative of mature, unfertilized
eggs, I recorded size for 10 eggs from each fish.
Their mean size was 0.88 mm in diameter and
ranged from 0.64 to 0.92 mm; the oil droplet aver-
aged 0.25 mm in diameter. Very little difference
was found in egg size for each fish.

Prejuvenile red porgy were collected in April off
South Carolina. An 18-mm specimen had minute
spines along the dorsal and ventral outlines of the
body, and five to six vertical pigment bands
(Figure 5). These bars appeared red on stressed
adults. Ranzi (1969) described young P. pagrus
from the Bay of Naples and referred to the vertical
bands in specimens 13 mm and larger.

Forty-four juvenile P. pagrus ranging in length
from 42 to 59 mm {X = 51 mm) were collected by
trawl off Charleston in relatively shallow water
(9-20 m); bottom temperatures ranged from 17.5°
to 18.5°C. The fish were also collected in April,
indicating spawning may occur slightly earlier in
that area compared with Onslow Bay and Raleigh
Bay, N.C.

SUMMARY AND CONCLUSIONS

Red porgy spawn in North Carolina waters from
January through April with a peak in spawning
activity between March and April. Maturation of
gonads and spawning appear to be correlated with
increased photoperiod. Spawning fish were col-
lected over irregular bottom ranging from 21 to
100 m in depth. Bottom temperatures at these
depths ranged from 16.4° to 21.5°C. Collection of
relatively large juveniles off Charleston in April
indicates that spawning may occur earlier there.

Some female P. pagrus attain sexual maturity
as 2-yr-old fish; however, the majority mature at 3
yr. AH of the fish examined had reached sexual
maturity by the fourth year. Approximately 50%
of the females were mature at 304 mm TL, and 75%
were mature at 334 mm. All fish 364 mm or more in
length were sexually mature. Evidently, some of
the age II and III fish experience regular, seasonal
maturation of gonads but do not spawn that first
year. This conclusion is based upon several fish I

779

FISHERY BULLETIN: VOL. 74, NO. 4

Figure 5.- Young red porgy, 18 mm total length, collected by trawl off Charleston, S.C, in April 1974 (drawing by Herbert Gordy,

National Marine Fisheries Service, NOAA).

observed which had ovaries containing absorbed
ova during the peak spawning period.

Fecundity estimation for red porgy ranged
from 48,656 eggs for a 304-mm female to 488,600
ova for a 516-mm fish. Larger fish (>600 mm TL),
which occasionally appear in the sport fishery, may
produce over 900,000 eggs. Eggs removed from
ripe females ranged in size from 0.31 to 0.94 mm in
diameter. The developed P. pagrus eggs averaged
0.88 mm in diameter and contained a single oil
droplet averaging 0.25 mm in diameter. While
fecundity was correlated to three predictors;
length, weight, and age, weight was the most
accurate predictor of fecundity. Although age was
not as satisfactory a predictor of fecundity as
weight and length, it should not be overlooked,
because the age-fecundity relation can have useful
application in population modeling. High vari-
ability in fecundity estimates for age-groups is
expected due to range in size and variation in
gonad size among fish of the same size (Bagenal
1967).

Sex ratios for red porgy were usually un-
balanced in favor of females. Analyzing data by
month, year, and size, I observed a domination by
females. The overall sex ratio observed was 1:2.
The occurrence of females was higher during the
spawning season. This predominance may be

attributed to difference in feeding behavior of ripe
fish, or to true population differences in the areas
sampled. I do not believe gear selectivity
influenced sex ratios. The dominance of females
for the smaller size classes and actual documenta-
tion of hermaphroditic red porgy in the study
lends some support to the theory of protogynous
hermaphroditism reported for the species in the
Gulf of Mexico (Beaumariage pers. commun.).
Both protandrous and protogynous hermaphrodi-
tism are relatively common among the sparids
(D'Ancona 1950, 1956). Although only 2% of the fish
examined were obviously hermaphroditic, a com-
plete histological study of gonadal development is
needed to determine if the species displays sex
reversal. Protogynous hermaphroditism may have
selective advantages as Atz (1964:224) mentioned
providing an endocrinologically better balanced
fish, assuring presence of both sexes in isolated,
insular areas, or a mechanism of population con-
trol. For the latter purpose, certain population
pressures presumably stimulate sexual transition.
Probably more applicable to red porgy hermaph-
roditism is the "size advantage model" proposed
by Ghiselin (1969). The theory explains sequential
hermaphroditism as occurring when an organism
reproduces more efficiently as one sex when small
and the opposite sex when larger. A male's poten-

780

MANOOCH: REPRODUCTIVE CYCLE OF RED PORGY

tial, theoretically, is higher than a female's at
larger sizes, and conversely, a female's reproduc-
tive potential is higher than a male's at smaller
sizes. The female reproductive capabilities could
continue to increase with age. Perhaps males
function more efficiently at larger sizes because
they can mate with numerous females. Evolution-
ary factors which favor protogyny are those which
tend to depress male reproductive potential at
early ages, such as inexperience, territoriality, or
female mate selection (Warner 1975). Without
additional information on the spawning behavior
of Pagrus, it would be difficult to eliminate any of
these factors.

ACKNOWLEDGMENTS

I thank W. W. Hassler, Department of Zoology,
North Carolina State University, G. R. Huntsman,
Task Leader of the Offshore Bottom Fisheries
Task, Atlantic Estuarine Fisheries Center, NOAA,
and David Colby, Atlantic Estuarine Fisheries
Center for their technical assistance and critical
review of the manuscript. I am grateful to Charles
A. Barans and F. H. Berry, South Carolina Wild-
life and Marine Resources Department for col-
lecting juvenile red porgy, D. S. Beaumariage,
Florida Department of Natural Resources and M.
A. Moe, Aqualife Research, St. Petersburg, Fla.,
for their comments on Pagrus hermaphroditism.

LITERATURE CITED

Atz.J. W.

1964. Intersexuality in fishes. In C. N. Armstrong and A. J.
Marshall (editors), Interse.xuality in vertebrates including
man, p. 145-232. Academic Press, N.Y.
Bagenal,T. B.

1967. A short review of fish fecundity. In S. D. Gerking
(editor). The biological basis of freshwater fish production,
p. 89-111. Wiley, N.Y.

ClECHOMSKI, J. D. DE, AND G. WeISS.

1973. Desove y desarollo embrionario y larval de besugo,
Pagrus pagrus (Linne) en el Mar Argentina (Pisces,
Sparidae). Physis Rev. Asoc. Argent. Cienc. Nat., Secc. A
32:481-487.
D'Ancona, U.

1950. Determination e differenciation du sexe chez les
poissons. Arch. Anat. Microsc. Morphol. Exp. 39:274-294.

1956. Inversion spontanees et experimentales dans les
gonades des Teleosteens. Annee Biol., Ser. 3, 32:89-99.
DiAS, R. K., J. K. DiAS, AND W. D. Anderson, Jr.

1972. Relationships of lengths (standard, fork and total) and

lengths to weight in the red porgy, Pagrus sedecim
(Perciformes, Sparidae), caught off South Carolina.
Trans. Am. Fish. Soc. 101:503-506.
Finkelstein, S. L.

1969. Age at maturity of scup from New York waters. N.Y.
Fish Game J. 16:224-237.
Ghiselin, M. T.

1969. The evolution of hermaphroditism among animals. Q.
Rev. Biol. 44:189-208.
Harrington, R. W., Jr.

1956. An experiment on the effects of contrasting daily
photoperiods on gametogenesis and reproduction in the
centrarchid fish, Enneacanthus ohesus (Girard). J. Exp.
Zool. 131:203-223.
Huntsman, G. R.

1976. Offshore headboat fishing in North Carolina and South
Carolina. Mar. Fish. Rev. 38(3):13-23.
Lassiter, R. R.

1962. Life history aspects of the bluefish, Pomatomus
salatrix (Linnaeus), from the coast of North Carolina.
M.S. Thesis, North Carolina State Univ.. Raleigh, 68 p.
Manooch, C. S.

1975. A study of the taxonomy, exploitation, life history,
ecology and tagging of the red porgy, Pagrus pagrus
Linnaeus off the Carolinas. Ph.D. Thesis, North Carolina
State Univ., Raleigh, 275 p.
In press. Food habits of the red porgy, Pagrus pagrus
Linnaeus (Pisces: Sparidae) off North Carolina and South
Carolina, U.S.A. Bull. Mar. Sci.
Manooch, C. S., G. R. Huntsman, B. Sullivan, and J. Elliott.
In press. Conspecific status of the sparid fishes Pagrus
sedecim Ginsberg and Pagrus pagrus Linnaeus. Copeia.
Moe, M. A., Jr.

1969. Biology of the red grouper, Epinephelus morio
(Valenciennes), from the eastern Gulf of Mexico. Fla.
Dep. Nat. Resour., Mar. Res. Lab.. Prof. Pap. Ser. 10, 95 p.
Orange, C. J.

1961. Spawning of yellowfin tuna and skipjack in the
eastern tropical Pacific, as inferred from studies on gonad
development. [In Engl, and Span.] Inter-Am. Trop. Tuna
Comm., Bull. 5:457-526.
Ranzi, S.

1969. Sparidae. In S. Lo Bianco, Eggs, larvae, and juvenile
stages of Teleostei, Parts I and II, p. 330-37.5. Fauna and
flora of the Bay of Naples, Monograph No. 38. (Translated
from Ital., TT68-50346.)
Schaefer, M. B., and C. J. Orange.

1956. Studies of the sexual development and spawning of
yellowfin tuna {Neothunnus macroptervs) and skipjack
(Katsuivonus pelamis) in three areas of the eastern Pacific
Ocean, by examination of gonads. [In Engl, and Span.]
Inter-Am. Trop. Tuna Comm. Bull 1:281-349.
Sekavec, G. B., and G. R. Huntsman.

1972. Reef fishing on the Carolina Continental Shelf. Proc.
1.5th Annu. Int. Game Fish Res. Conf., p. 76-86.
Walker, E. T.

1950. Spawning records of fishes seldom reported from
North Carolina waters. Copeia 1950:319.
Warner, R. R.

1975. The adaptive significance of sequential hermaphrodi-
tism in animals. Am. Nat. 109:61-82.

781

MERCURY IN FISH AND SHELLFISH OF THE NORTHEAST PACIFIC.
I. PACIFIC HALIBUT, HIPPOGLOSSUS STENOLEPIS

Alice S. Hall/ Fuad M. Teeny,i Laura G. Lewis/ William H. Hardman,^ and Erich J. Gauglitz, Jr.'

ABSTRACT

A total of 1,227 Pacific halibut, Hippoglossus stenolepis, were analyzed for mercury content in the edible
muscle tissue. These fish were obtained from five geographical areas within the species range: the
Bering Sea, Gulf of Alaska, southeast Alaska, British Columbia, and Washington-Oregon. Mercury was
found to be uniformly distributed from nape to tail in the edible muscle tissue. Within each
geographical area the mercury concentration increased as the size of the fish increased. The mercury
concentration also increased in fish of the same size from the northern to the southern part of the
species range.

In the past few years, numerous investigators
have examined the distribution and levels of
mercury in food, including aquatic food animals,
because of the potential health hazards involved.
The U.S. Food and Drug Administration estab-
lished an administrative guideline of 0.50 ppm
mercury in fish and shellfish in 1969. Since that
time, the guideline has been the subject of several
reviews and recently has been proposed as a
formal action level (Schmidt 1974).

Since 1970, the Pacific Utilization Research
Center (PURC) and the Southeast Utilization
Research Center (SEURC) at College Park, Md.,
have been conducting extensive studies of fish and
shellfish taken from marine and inland waters of
the United States to determine the extent to
which mercury exceeds the guideline in our aquat-
ic resources. This paper reports our findings on
mercury in the edible tissue of the Pacific halibut,
Hippoglossus stenolepis Schmidt.

EXPERIMENTAL PROCEDURE
AND METHODS

Halibut were obtained from commercial fishing
vessels, fish processing companies, and research
vessels of the International Pacific Halibut Com-
mission (IPHC). Data were obtained on area and
date of catch, and weight or length of each fish

'Pacific Utilization Research Center, National Marine Fish-
eries Service, NOAA, 2725 Montlake Blvd. East, Seattle, WA
98112.

^International Pacific Halibut Commission, P.O. Box 5009,
University Station, Seattle, WA 98105.

analyzed. Data were also obtained on age and sex
when possible.

The five areas of catch were: Washington-
Oregon, British Columbia, southeast Alaska, Gulf
of Alaska, and the Bering Sea (Figure 1). Com-
mercial halibut are eviscerated at sea, landed as a
heads-on eviscerated product, and then beheaded
for marketing as fresh or frozen fish. Weights
reported here are in pounds for heads-off eviscer-
ated fish because this is the standard practice of
the halibut industry. For convenience of some
readers who do not normally use our measurement
system, approximate metric equivalents in kilo-
grams are given in the tables and figures. When
actual weights were impractical to obtain, the
lengths of the heads-on fish were used, and heads-
off eviscerated weights were estimated using
length-weight conversion tables of the IPHC. Age
was determined, as described by Hardman and
Southward (1965), from otoliths collected at the
landing site when circumstances permitted and on
all halibut taken by IPHC research vessels.

Before setting up sampling procedures, exper-
iments were carried out to determine the unifor-
mity of distribution of mercury in the muscle of
individual fish. No significant differences in con-
centration of mercury (deviation did not exceed
± 0.03 ppm) were noted in muscle tissue taken from
nape, midbody, or tail sections.

Analytical samples consisted of skinned and
deboned edible muscle tissue that was normally
taken from the nape section just behind the head.
Some samples, however, were in the form of steaks
and a few consisted of the entire fillets of small
fish. Portions, usually about 400 g, taken from the

Manuscript accepted March 1976.
FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

783

FISHERY BULLETIN: VOL. 74, NO. 4

Figure l.-Mean mercury levels in
Pacific halibut by area of catch.

'V° • "^'-st^

.=a-=*

BERING SEA

0.15

WASHINGTON -OREGON
0.45

nape section were ground in a Hobart grinder^
equipped with a Vg-inch (3.2-mm) hole stainless
steel plate. Larger steaks and fillets were ground
in a Hobart Silent Food Cutter (Model 84181). The
comminuted flesh was mixed thoroughly before
subsampling for analysis. Because samples were
often collected more rapidly than they could be
analyzed, they were stored at -29°C until analysis.
No change in mercury content was observed in
halibut that were analyzed immediately or that
had been held in frozen storage in either glass
vials or aluminum containers if dehydration was
prevented. A halibut sample stored in the above
manner and used as an analytical control showed a
mean mercury content of 0.88 ±0.02 ppm over a
2-yr period. This control was analyzed routinely to
verify both accuracy and precision of the method.
Total mercury was determined at the PURC by
either the method of Munns and Holland (1971) or
Malaiyandi and Barrette (1970) as modified by
Munns (1972). The former method uses sulfuric,
nitric, and perchloric acids for digestion with
sodium molybdate as a catalyst, while the Munns'
modification utilizes nitric and sulfuric acids for
digestion and vanadium pentoxide as a catalyst.
Some samples were analyzed at the SEURC by the
method of Hatch and Ott (1968) as modified by
Uthe et al. (1970). This method uses sulfuric acid
for digestion and potassium permanganate as an
oxidizing agent.

■'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Final quantitation was by flameless spectroscopy
using a Perkin-Elmer Model 403 Atomic Absorp-
tion Spectrophotometer at the PURC and by a
Varian Techtron Model AA5 at the SEURC. In a
collaborative study, the mean deviation between
laboratories and methods did not exceed ± 0.02 ppm
Hg. All samples were analyzed in duplicate or
triplicate, depending upon the method of analysis
used. We consider ± 0.05 ppm a significant devia-
tion; therefore, when differences between rep-
licates exceeded this level the samples were
reanalyzed. Results are stated in parts per million
wet weight.

RESULTS AND DISCUSSION

A total of 1,227 halibut were analyzed for mer-
cury content. Results indicated a relationship
between mercury levels and area of catch, age, and
size of fish. The results are broken down by the
previously described catch areas (Figure 1). The
fish taken from each area were separated by
weight classes that approximate those used in the
halibut industry; the low, high, and mean mercury
values for each weight class are given with a
frequency distribution of the fish by increasing
mercury concentration (Tables 1 through 5).
Because we thought that large fish would be more
likely to exhibit higher concentrations of mercury,
we attempted to obtain as many large fish as was
practicable. For this reason our sampling contains
a greater percentage of large fish than do the
commercial catches from most of the areas dis-

784

HALL ET AL.: MERCURY IN PACIFIC HALIBUT

cussed here. Therefore, these data cannot be
interpreted to indicate the approximate percent-
age of the commercial catch that is likely to
contain mercury in concentrations over the
guideline.

In 152 halibut taken from the Bering Sea, the
mercury level in the muscle of 7 fish (5% of the
sample) was over the guideline (Table 1). The
incidence (percentage over the guideline of the
total number of fish within a weight range) was
highest among fish weighing more than 80 pounds.

Most of our samples, 761 fish, were taken from
the Gulf of Alaska. We found that mercury in the
muscle of 38 fish (5% of the sample) exceeded the
guideline (Table 2). The highest incidence occurred
in fish weighing more than 80 pounds. The weight
ranges contributing most to the incidence were
those of 126 to 150 pounds and those of more than
150 pounds. These two weight ranges contribute
21% and 32%, respectively, in contrast with only 3%
in each of the weight ranges 81 to 100 pounds and
101 to 125 pounds.

The analytical data on 70 fish taken from south-

east Alaska area showed that mercury in the
muscle of 9 fish (13% of the sample) was 0.50 ppm
or higher (Table 3). The small number of fish in the
larger weight ranges makes it impossible to be
definitive, but it is reasonably clear that in this
group, too, the incidence of mercury levels over the
guideline was greatest among the largest fish.

Analyses on 163 fish from the British Columbia
area showed that 44 of these (27% of the sample)
were over the guideline (Table 4). In addition to
this relatively high incidence, we saw for the first
time the presence of significant numbers of high-
mercury-level fish in all weight groups, i.e., 10% of
the fish were over the guideline in the 5- to
60-pound range, 75% in the 61- to 80-pound range,
73% in the 81- to 100-pound range, 100% in the 101-
to 125-pound range, and 67% in the 126- to 150-
pound range. We also saw that the concentration
of mercury tended to increase with an increase in
the incidence of fish that were over the guideline.

The analytical results on 81 fish taken from the
Washington-Oregon area, the most southerly area
of the range of the Pacific halibut, showed 29 fish

Table 1. -Mercury concentration in heads-off eviscerated Pacific halibut from the Bering Sea.

Weight range

1 No.

Mercury (pprr

I) in 1

edible muscle

tissue

Pounds

0.25

- 0.40-

- 0.50-

- 0.60-

0.70-

- 0.80-

0.90-

-1.00-

(kg)

fish

Low

High

Mean

<0,25

0.39

0.49

0.59

0.69

0.79

0.89

0.99

1.49

• - -Nurp^^'' '^' fieh - _ .

5-60

0.02

0.78

0.11

1 ttji ii

(2-27)

61-80

0.06

0.42

0.15

(28-36)

81-100

0.09

0.55

0.19

(37-45)

101-125

0.08

1.00

0.32

(46-57)

126-150

0.22

0.35

0.27

(57-68)

Total

152

0.02

1.00

0.15

134

Table 2.-Mercury concentration in

heads-off eviscerated Pacific halibut from the Gulf of Alaska

Weight range
Pounds

No.
of

Mercury (ppi

m) in edible muscle

tissue

0.25-

- 0.40-

0.50-

- 0.60-

0.70-

- 0.80-

0.90-

-1.00-

(kg)

fish

Low

High

Mean

<0.25

0.39

0.49

0.59

0.69

0.79

0.89

0.99

1.49

- hliifYy^^f '^' //oh

5-60

378

0.01

0.50

0.11

371

(2-27)

61-80

0.05

0.47

0.18

(28-36)

81-100

0.05

1.10

0.25

(37-45)

101-125

0.03

0.74

0.29

(46-57)

126-150

0.12

1.28

0.38

(57-68)

Over 151

0.14

1.05

0.45

(68)

Total

761

0.01

1.28

0.20

561

107

785

FISHERY BULLETIN: VOL. 74, NO. 4
Table 3.-Mercury concentration in heads-off eviscerated Pacific halibut from southeast Alaska.

Weight range

No.
of

Mercury (ppm) m i

;dible r

nuscle

tissue

Pounds

0.25-

- 0.40-

- 0.50-

- 0.60-

0.70-

0.80-

0.90-

-1.00-

(kg)

fish

Low

High

Mean

<0.25

0.39

0.49

0.59

0,69

0.79

0.89

0,99

1.49

5-60

0.04

0.34

0.12

(2-27)
61-80

0.09

1.30

0.33

(28-36)
81-100

0.09

0.59

0.28

(37-45)
101-125

0.22

0.95

0.46

(46-57)
126-150

0.26

0.36

0.31

(57-68)
Over 151

0.50

1.10

0.80

(68)
Total

0.04

1.30

0.26

Table 4

.-Mercury concentration i

n heads-off eviscerated Pacific halibut from British Columbia.

Weight range

No.
of

Mercury (ppm) in (

edible 1

muscle

tissue

Pounds

0.25-

- 0.40

- 0.50-

- 0.60-

0.70-

■ 0.80-

0.90-

-1.00-

(kg)

fish

Low

High

Mean

<0.25

0.39

0.49

0.59

0.69

0.79

0.89

0.99

1.49

5-60

122

0.04

1.04

0.19

(2-27)
61-80

0.12

1.23

0.69

(28-36)
81-100

0.10

1.22

0.66

(37-45)
101-125

0.50

1.46

0.96

(46-57)
126-150

0,25

0.77

0.52

(57-68)
Total

163

0.04

1.46

0.32

102

Table 5.-

Mercury

concentration in

heads-off

eviscerated Pacific halibi

Jt from Washi

ington-Oregon.

Weight range

No.
of

Mercury (ppi

m) in (

sdible 1

muscle

tissue

Pounds.

0.25-

- 0.40

- 0.50-

- 0.60-

0.70-

■ 0.80-

0.90-

-1.00-

(kg)

fish

Low

High

Mean

<0.25

0.39

0.49

0.59

0.69

0.79

0.89

0.99

1.49

5-60

0.10

1.43

0.42

(2-27)
61-80

0.70

1.13

0.88

(28-36)
Total

0.10

1.43

0.45

(36% of the sample) were over the guidehne (Table
5). None of these fish weighed more than 80
pounds, and only six weighed more than 60 pounds;
31% of the 5- to 60-pound fish and all of the 61- to
80-pound fish were over the guideline. In fish from
this area, as in those from British Columbia, the
concentrations of mercury increased with the
incidence of fish over the guideline.

It is apparent that the mean level of mercury in
the edible tissue and the incidence of fish over the
guideline increases from the northern to the
southern part of the range of the Pacific halibut
(Figure 1, Table 6). There is also a relationship

786

between the size of fish and the level of mercury in
the muscle. Because of the sex-size relationship of
halibut, i.e., males rarely exceed 80 pounds
regardless of age, the correlation of mercury to
age should be closer than that of mercury to size.
However, age data were collected on only 76% of
the total sampling, whereas weight was obtained
on all samples. For this reason, and as a guide to
industry, we have worked mostly with the mer-
cury-size relationship. Evaluation of the data by
regression analyses showed that the data are well
described by the exponential function {y = ax*).
Comparisons of the weights of halibut against

HALL ET AL.: MERCURY IN PACIFIC HALIBUT

Table 6.-Summary of mercury concentration in Pacific halibut.

1.50

1.00-

Number
of
Area of catch fish

Mean
lb

weight
kg

Me
Low

rcury (p
High

pm)
Mean

0.15
0.20
0.26
0.32
0.45

Percent of
samples

exceeding
0.50 ppm

4.6

5.0
12.8
27.0
35.8

Bering Sea
Gulf of Alaska
Southeast Alaska
British Columbia
Washington-Oregon

152
761

70
163

54.6
71.8
67.6
39.3
30.3

24.8

32.6
30.7
17.8
13.8

0.02
0.01
0.04
0.04
0.10

1.00
1.28
1.30
1.46
1.43

WEIGHT(kg)
50 75

100

WEIGHT (kg)
50

0.50-

50 100 150 200 250

WEIGHT (pounds)

Figure 2. -Relationship between heads-off eviscerated weight
and mercury concentration in the edible muscle tissue of Pacific
halibut from the Bering Sea.

Q.
Ql

1.50

1.00-

0.50

100

100 150

WEIGHT(pounds)

250

Figure 4. -Relationship between heads-off eviscerated weight
and mercury concentration in the edible muscle tissue of Pacific
halibut from southeast Alaska.

WEIGHT (kg)

1.50

100

>
IT

1.00

+
+

f +

>>vt

+ +*

0.50

^ *.(... ..*■+

+ +

+■

V^t,^^>■*^r

*4t

+ *+

, J t jHvj, ff%,^JU^i^

= 0.730

^^X*^*^^

V* +

0 50 too 150 200 250

WEIGHT (pounds)

Figure 3.-Relationship between heads-off eviscerated weight
and mercurj' concentration in the edible muscle tissue of Pacific
halibut from the Gulf of Alaska.

mercury concentrations in the edible tissue for
each area are shown in Figures 2 through 6.
Correlation coefficients (r values) are shown on

each plot and are significant at the 0.1% level.
Correlation coefficients between length and mer-
cury were also significant at the 0.1% level within
each area and were essentially identical to the
correlation coefficients between weight and mer-
cury. This would be expected from the weight-
length relationship. Correlation between age and
mercury was higher than between weight or
length and mercury for fish from the Bering Sea,
the Gulf of Alaska, and southeast Alaska; the same
for fish from British Columbia; and lower for fish
from Washington-Oregon. These correlation
coefficients between age and mercury were also
significant at the 0.1% level in all areas.

In evaluating the data, areas were used that are
either the same as the fishery management areas
defined by the International Pacific Halibut Com-
mission (1974) or subdivisions of a management
area. This was both logical and practical for the
purpose of providing useful information to the
halibut industry. The plots of mercury concentra-
tion in the edible muscle against weight of fish
taken from both the Bering Sea and the Gulf of

787

FISHERY BULLETIN: VOL. 74, NO. 4

WEIGHT (kg;
25 50

100

1.50

E
a.

> 1.00
cr

0.50

' [

■ I 1

_ +

+
-t

>^= 0.766

+ +-H^

•:v

^ +

+? + -^ +>
+. '^>''^
* /+*

+
+

J^^^ 4-

+
1

0 50 100 150 200 250

WEIGHT (pounds)

Figure 5.-Relationship between heads-off eviscerated weight
and mercury concentration in the edible muscle tissue of Pacific
halibut from British Columbia.

WEIGHT(kg)
50 75

0 50 100 150 200

WEIGHT (pounds)

250

Figure 6.-Relationship between heads-off eviscerated weight
and mercury concentration in the edible muscle tissue of Pacific
halibut from Washington-Oregon.

Alaska (Figures 2, 3) are so similar as to suggest
that the environmental and biological factors that
determine the rate and extent of deposition of
mercury in the muscle are the same in both areas.
In any case, the mean level of mercury and the
incidence of fish exceeding the guideline increases,
while the size of the fish decreases, from north to
south.

Increasing concentrations of mercury have been
noted in other marine animals as one moves south
from the Bering Sea. Anas (1974) pointed out that
the harbor seal, Phoca vitulina richardi, which is a
nonmigratory, inshore carnivore that feeds prin-
cipally on fish, provides geographical information
on local concentrations of contaminants. The livers
of harbor seals taken from the Bering Sea con-
tained lower levels of mercury than did those from
Washington and Oregon, and those from southern
California contained the highest levels. Sablefish,
A noplopoma fimbria (Pallas), also shows a similar
pattern and will be the subject of another paper in
this series.

These observations suggest that the total mer-
cury contamination in the ocean environment
(natural plus man-made) increases in a north-to-
south direction. Unfortunately, conclusive data to
substantiate this hypothesis are not available.
Eggerman and Mar (1972), in a review of the
research that has been conducted on the various
aspects of mercury transport, state that there is a
paucity of available data, especially on the
biological transport of mercury in marine waters.

ACKNOWLEDGMENTS

We thank Lyle Morimoto and Michael Bienn,
formerly of the PURC and the SEURC, for assis-
tance in mercury analyses; Virginia Stout of the
PURC and Murray Amos and Ernest Decorvet of
the Northwest Fisheries Center for their help with
data processing; and Bernard Skud, Director,
IPHC, for his cooperation in this investigation.

LITERATURE CITED

Anas, R. E.

1974. Heavy metals in the northern fur seal, Callorhinus
ursinus, and the harbor seal, Phoca vitulina richardi.
Fish. Bull., U.S. 72:133-137.
Eggerman, T., and B. Mar.

1972. Mercury in the North Pacific-NORFISH NFOl. Cent.
Quant. Sci., Univ. Wash., Seattle, 15 p.
Hardman, W. H., and G. M. Southward.

1965. Sampling the commercial catch and use of calculated
lengths in stock composition studies of Pacific halibut.
Int. Pac. Halibut Comm. Rep. 37, 32 p.
Hatch, W. R., and W. L. Ott.

1968. Determination of sub-microgram quantities of mer-
cury by atomic absorption spectrophotometry. Anal.
Chem. 40:2085-2087.
International Pacific Halibut Commission.

1974. Pacific Halibut Fishery Regulations 1974. Int. Pac.
Halibut Comm., Seattle, Wash., 5 p.

788

HALL ET AL: MERCURY IN PACIFIC HALIBUT

Malaiyandi, M., and J. P. Barrette.

1970. Determination of submicro quantities of mercury in
biological materials. Anal. Lett. 3:579-584.

MuNNS, R. K.

1972. Mercury in fish by cold vapor AA using sulfuric-nitric

acid/VoO.i digestion. FDA (Food Drug Admin.) Lab. Inf.

Bull. 1500, 8 p.
MuNNS, R. K. AND D. C. Holland.

1971. Determination of mercury in fish by flameless atomic

absorption: A collaborative study. J. Assoc. Off. Anal.
Chem. 54:202-205.
Schmidt, A. M.

1974. Action level for mercury in fish and shellfish. Fed.
Regist., 39 (236) Part II: 42738-42740.
Uthe, J. F., F. A. J. Arnstrong, and M. P. Stainton.

1970. Mercury determination in fish samples by wet diges-
tion and flameless atomic absorption spectrophotometry. J.
Fish. Res. Board Can. 27:805-811.

789

MERCURY IN FISH AND SHELLFISH OF THE NORTHEAST PACIFIC.
II. SABLEFISH, ANOPLOPOMA FIMBRIA

Alice S. Hall, Fuad M. Teeny, and Erich J. Gauglitz, Jr.*

ABSTRACT

Sablefish, Anoplopoma fimbria, collected from several locations in Alaska, Washington, Oregon, and
California were analyzed for their mercury content. Mean mercury level in this species varied with the
geographical location of catch, showing a gradual increase in magnitude from north to south; the
average size of the specimens decreased in the same pattern, north to south. Of the 692 specimens
analyzed in this study, approximately 30% exceeded the U.S. Food and Drug Administration action
level of 0.50 ppm mercury. Significant relationships between the size of the fish and mercury content
were observed.

Following the Canadian disclosure in March 1970
of high mercury levels in fish caught in Lake St.
Clair (Hearnden 1970), the National Marine
Fisheries Service (NMFS) initiated studies to
determine the distribution and level of mercury in
our marine resources. Since that time, the Pacific
Utilization Research Center, NMFS, has been
conducting extensive screening studies of fish and
shellfish of the northeast Pacific in order to evluate
the mercury problem as it relates to those species
taken by both commercial and sport fisheries. The
main objectives were to determine which species
contained mercury in excess of the Food and Drug
Administration (FDA) action level of 0.50 ppm
(Schmidt 1974) and the severity of the problem.

During our preliminary screening of Pacific
species, we found that the edible muscle tissue of a
number of sablefish contained mercury in excess
of the FDA action level. This species ranges from
southern California to the Bering Sea (Clemens
and Wilby 1961:240). Domestic landings in 1971
were about 6 million pounds (2.7 x 10'' kg)
(Thompson 1971) but its high value as a smoked
product and the availability to the fishermen of
additional supplies of this species suggests that
landings will increase.

This paper is the second in a series and reports
our findings on mercury in the edible muscle tissue
of sablefish, Anoplopoma fimbria (Pallas). The first
paper in the series is on the Pacific halibut, Hip-
poglossus stenolepis Schmidt (Hall et al. 1976).

'Pacific Utilization Research Center, National Marine Fish-
eries Ser\'ice, NOAA, 2725 Montlake Boulevard East, Seattle,
WA 98112.

EXPERIMENTAL PROCEDURE
AND METHODS

Most of the sablefish used in this study were
obtained by NMFS personnel aboard National
Oceanic and Atmospheric Administration
(NOAA) research vessels. Some samples were
obtained from commercial lots through the coop-
eration of fish processors in order to cover the range
of this species. Samples were obtained from the
waters off Alaska, Washington, Oregon, and
California. Date and location of catch were
recorded for all specimens.

Weights and lengths are reported for heads-off
eviscerated fish because this is the standard prac-
tice for landing sablefish. Round weights and
lengths were converted to the heads-off eviscerat-
ed values using conversion tables. Where possible,
sex was determined by physical examination when
the specimens were eviscerated. Age was deter-
mined from the otoliths which were removed at
the same time.

Analytical samples consisted of the entire fillets
of each fish. The edible muscle tissue was ground in
a Hobart grinder'^ equipped with a stainless steel
plate perforated with holes Vs inch (3.2 mm) in
diameter. The comminuted flesh was mixed thor-
oughly; subsamples were removed, packaged, and
stored at -29°C until analysis.

Total mercury was determined by either the
FDA method of Munns and Holland (1971) or
Malaiyandi and Barrette (1970) as modified by

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

791

FISHERY BULLETIN: VOL. 74, NO. 4

Munns (1972). Final quantitation was by flameless
spectroscopy using a Perkin-Elmer Model 403
Atomic Absorption Spectrophotometer.

Results are stated in parts per million wet
weight. All samples were analyzed in duplicate
and where the deviation for replicates exceeded
±0.05 ppm, the analyses were repeated. Control
samples of known value were analyzed routinely to
verify both accuracy and precision of the method.

RESULTS AND DISCUSSION

A total of 692 sablefish taken from the Bering
Sea and coastal waters of the Pacific Ocean from
Kodiak Island, Alaska, to San Diego, Calif., were
analyzed for individual mercury content. The
specific locations of catch and the mean mercury
levels by area are shown in Figure 1. The mean
mercury levels show a general increase from north
to south, as does the percentage of fish that exceed
the FDA action level of 0.50 ppm (Table 1).

Table 1.— Summary of mercury concentration in sablefish.

Mean

weight

No.

Mercury (DOm)

samples

fish

Pounds
(kg)

over

Area of catch

Low

High

Mean

0.50 ppm

Bering Sea-

2.02

0.02

0.11

0.04

Kodiak Island

(0.92)

Southeast

120

5.22

0.06

0.77

0.28

Alaska

(2.37)

Washington

121

5.27
(2.39)

0.06

1.28

0.37

Oregon

174

4.33
(1.96)

0.06

1.23

0.40

Northern

3.14

0.03

0.95

0.26

California

(1.42)

Central

5.68

0.08

0.79

0.47

California

(2.58)

Southern

119

2.93

0.04

2.11

0.60

California

(1.33)

Effect of the Geographical Location

The fish caught in the Bering Sea and in the
vicinity of Kodiak Island were all small (less than 3
pounds [1.4 kg]) and contained very low levels of

Chichogof island
Baranof Island

Betim Conal

Neoh Bay

Long Beoct>
Tillamook Heod
Cope Lookout

N. CALIFORNIA 026 } Eureko

Fori Bragg

C. CALIFORNIA 0.47

S. CALIFORNIA 0.60

Sonta Cruz
Po<nt Sal

Figure 1.— Mean mercury levels (parts per million) in sablefish by area and the specific locations of the catches.

792

HALL ET AL.: MERCURY IN SABLEFISH

mercury (0.02-0.11, x 0.04 ppm). The data for the
specimens from these two areas were combined
since the samples were relatively few in number,
and there was no evidence of any significant
differences based on area (Table 1).

A much better weight distribution is seen in the
120 fish from southeast Alaska (Table 2). The fish
taken from several locations around Baranof and
Chichagof islands (45 specimens) contained a
significantly lower mean level of mercury (0.19
ppm) than did the 75 fish taken from the Behm
Canal area (0.34 ppm). The only fish (5% of the total
sample) from southeast Alaska that exceeded 0.50
ppm mercury were caught off Betton Island, which
is in the north arm of Behm Canal. This would
indicate a higher level of mercury contamination
in the inland waters than in the offshore waters
around the outer islands.

Analyses of 121 fish from Washington showed
that 23% (28 fish) of the sample exceeded the action
level (Table 2). The fish taken from the northern
coast off Neah Bay and those taken from the
southern coast off Long Beach showed little
difference in mercury content.

Of the 174 fish from Oregon, 51 or 29% exceeded
the action level (Table 2), which is an increase over
that observed in previously discussed areas. A
significant part of the total sample (39%) consisted
of fish weighing less than 3 pounds (1.4 kg) and of
these small fish we observed an increase in the
percentage that exceeded 0.50 ppm mercury.

The sampling from northern California (Table
3) consisted of 98 fish of which 62% (61 fish)
weighed less than 3 pounds (1.4 kg) and contained
low levels of mercury. Only one of these small fish
exceeded 0.50 ppm mercury. Of the remaining 37
larger fish, the mercury level of 20 fish exceeded
0.50 ppm. The mean mercury level of the total lot
of 98 fish was 0.26 ppm, and 21 fish or 21% exceeded
the action level. Considering that this lot repre-
sented an atypical weight distribution, it seems
likely that both the mean and the percentage of
fish exceeding the action level would be higher in a
sampling where the number of fish are more
uniformly distributed over the weight range.

The 30 fish collected in central California were
well distributed over the weight range (Table 3)
and 43% of these fish exceeded the action level.

Analytical data on 119 fish from the southern
California area showed that 72% (86 fish) exceeded
the action level (Table 3). Of this group, 47%
weighed less than 3 pounds (1.4 kg). Here, as in
Oregon, we saw that smaller fish contained high
levels of mercury in comparison to other areas. The
weight range of the fish from southern California
was small (from 0.5 to 5.5 pounds [0.2-2.5 kg]), but
the mercury levels were higher than were ob-
served in any other area.

Effect of Size of Fish

The observations on mercury levels and size of

Table 2.-Mercury concentration in heads-oflF eviscerated sablefish from southeast Alaska, Washington, and Oregon.

South

east Al

aska

Washington

Oregon

Weight range

ish

rcury (ppm)

■ish

rcury (ppm)

ish

Mercury (p

ipm)

Pounds

% over

(kg)

No.

0.5 ppm

Low

High

Mean

No.

0.5 ppm

Low

High

Mean

No.

0.5 ppm

Low

High

Mean

0,5-2.99

0.06

0.23

0.12

0.07

0.52

0.27

0.06

0.69

0.25

(0.23-1.36)

3.0-3.99

0.07

0.48

0.29

0.12

0.82

0.35

0.20

0.63

0.43

(1.36-1.81)

4.0-4.99

0.12

0.40

0.27

0.06

0.52

0.29

0.22

0.71

0.48

(1.82-2.26)

5.0-5.99

0.19

046

0.33

0.26

0.61

0.41

0.18

0.72

0.45

(2.27-2.72)

6.0-6.99

0.23

0.61

0.39

0.12

0.62

0.33

0.06

1.23

0.51

(2.72-3.17)

7.0-7.99

0.18

0.66

0.36

0.08

0.72

0.41

0.33

0.84

0.55

(3.18-3.63)

8.0-8.99

0.20

0.47

0.35

0.48

0.65

0.58

0.33

0.72

0.52

(3.63-4.09)

9.0-9.99

0.24

0.48

0.37

0.15

1.28

0.61

0.44

1.02

0.68

(4.09-4.54)

10.0-10.99

0.46

0.77

0.61

100

0.52

0.90

0.72

—

(4.54-4.99)

11.0-11.99

100

0.56

100

0.50

0.67

0.58

100

1.15

(4.99-5.44)

12.0-12.99

0.38

0.48

0.43

—

(5.45-5.90)

793

FISHERY BULLETIN: VOL. 74, NO. 4

Table 3. -Mercury concentration in heads-off eviscerated sablefish from California.

Northern California Central California Southern California

Fish Mercury (ppm) Fish Mercury (ppm) Fish Mercury (ppm)

% over % over % over

No. 0.5 ppm Low High Mean No. 0 5 ppm Low High Mean No. 0.5 ppm Low High Mean

61 2 0.03 0.65 0.12 4 0 0.08 0.18 0.13 68 59 0.04 1.74 0.50

8 50 0.20 0.85 0.45 5 40 0.36 0.54 0.46 35 86 0.38 2.11 0.73
14 50 0.22 0.73 0.47 7 29 0.42 0.79 0.53 12 100 0.53 0.88 0.72

9 44 0.22 0.75 0.49 4 25 0.13 0.57 0.40 4 100 0.54 0.83 0.71
2 100 0.70 0.75 0.73 2 0 0.31 0.44 0.37 — — _ _ _
2 50 0.32 0.89 0.61 1 100 0.58 0.58 0.58 — — _ _ _
1 100 0.95 0.95 0.95 3 100 0.62 0.69 0.66 — — _ _ _
1 100 0.53 0.53 0.53 2 100 0.61 0.68 0.64 — — _ _ _

— — _ _ _ 1 100 0.71 0.71 0.71 — — _ _ _

— — _ _ _ 1 100 0.61 0.61 0.61 — — _ _ _

Weight range

Pounds
(kg)

0.5-
(0.23-

3.0-
(1.36-

4.0-
(1.82-

5.0-
(2.27-

6.0-
(2.72-

7.0-
(3.18-

8.0-
(3.63-

9.0-

(4.09-

12.0-

(5.45-

13.0-

(5.90-

2.99

1.36)

3.99

1.81)

4.99

2.26)

5.99

■2.72)

6.99

3.17)

7.99

3.63)

8.99

4.09)

9.99

4.54)

12.99

5.90)

13.99

6.35)

sablefish are analogous to what was found in
Pacific halibut (Hall et al. 1976); i.e., mercury levels
increased from north to south until at the southern
part of the range even small fish exhibited high
mercury levels. Anas (1974) observed a similar
pattern in the harbor seal, Phoca vitulina
richardi.

There appears to be a direct relationship
between the size of the sablefish and the mercury
level found in the muscle. Comparisons between

WEIGHT (kg)

0.16^

) 0.5

1.0 1.5

■■ ■ 1

- 0.12
a.

•^

■t-

Q^ 0.08

— +

■¥

^ +

0.04

H--ff .

-"' ' ' r= 0.1 02

^^^^^ '" " f+

J- + +

•«■ -H

2 3

WEIGHT (pounds)

Figure 2.- Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from the
Bering Sea-Kodiak Island, Alaska.

WEIGHT (kg)
4 6

5 10 15

WEIGHT(pounds)

Figure 3.— Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
southeast Alaska.

weight and mercury level for fish from each area
are given in Figures 2-8. The exponential function,
y = aj*, was used for the statistical evaluation of
the data. Length-to-mercury relationships are
very similar to those for weight-to-mercury and
are not shown for this reason. Correlation
coeflRcients (r values) are shown on each plot, and
the relationship between weight and mercury was
highly significant (0.1% level) in all areas except in
the Bering Sea-Kodiak Island area where the
relationship was not significant.

794

HALL ET AL.: MERCURY IN SABLEFISH

WEIGHT (kg)

1.60'

|l.20

4 6

>
q:

^ 080

+
+ +

040

+ + +"■ +
'~ l_j

.-^- "^r =0.376

+
+

1 1

5 10 15

WElGHT(pounds)

Figure 4.-Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
Washington.

1.60

1 20-

WEIGHT (kg)
4 6

o 0 80
cr

040

♦

1 1

. ., ^M>x

'r = 0 606

t 1 L_

5 10 15

WEIGHT (pounds)

Figure 5.-Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
Oregon.

Effect of Age

Since the female sablefish grows faster and
attains a larger size than the male (Clemens and
Wilby 1961), it would seem logical to assume that
the correlation between age and mercury level
might be better than that of weight and mercury
level. However, higher correlation coefficients
exist between weight and mercury than between
age and mercury in all areas except Oregon.

WEIGHT (kg)
1 2 3

1.00

0.80

^ 0.60

UJ 040

0.20

r=0811

2 4 6 8

WEIGHT (pounds)

Figure 6.- Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
northern California.

WEIGHT (kg)
2 4 6

10 15

WEIGHT (pounds)

Figure 7.-Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
central California.

Relationships between age and mercury are
significant in all areas except Washington (Table
4). Age was not obtained on fish from the Bering
Sea-Kodiak Island area.

795

FISHERY BULLETIN: VOL. 74, NO. 4

Table 4.-Correlation coefficients for relationship of mercury level in the edible flesh to weight, age, and

sex of sablefish.'

Number

Weight

Number

Age

Number

Weight

Number

Weight

Area of catch

fish

mercury

fish

mercury

females

mercury

males

mercury

Bering Sea-Kodiak Island

20.102

—

Southeast Alaska

120

0.772

103

0.684

0.868

0.762

Washington

121

0.376

'0.179

0.731

^0.349

Oregon

174

0.606

0.693

116

0.657

0.480

Northern California

0.811

0.558

—

Central California

0.741

30.430

■>0.630

<0.758

Southern California

119

0.748

0.439

0.661

20.203

'Correlation coefficients significant at the 0.1% level unless otherwise indicated.
2Not significant.
^Significant at 5% level.
■•Significant at 1% level.

WEIGHT (kg)
1 2

2 00-

— I—

r=0.748

4 6 8

WEIGHT(pounds)

Figure 8.-Relationship between heads-off eviscerated weight
and mercury concentration in muscle tissue of sablefish from
southern California.

Effect of Sex

The females show better correlation between
weight and mercury than do the males, and cor-
relation coefficients are significant for females
from all areas (Table 4). Correlation coefficients
for weight to mercury are also significant for
males in all areas except Washington and southern
California. Sex was not obtained on fish from the
Bering Sea-Kodiak Island or northern California.

Effect on Utilization of Sablefish

It is apparent that sablefish can accumulate
mercury in amounts that exceed the maximum
level permitted in fish by the FDA. Spinelli et al.

(1973) noted that fish withheld from food use due
to high mercury levels constitute a significant loss
to the industry and showed that such losses could
be reduced by using a cysteine treatment to lower
the mercury content of the fish during processing.
Teeny et al. (1974) conducted a similar study on the
reduction of mercury in sablefish, and found that
up to 80% of the mercury present in the edible
tissue could be removed. Processing techniques of
this type could result in all sablefish being ac-
ceptable for human consumption.

ACKNOWLEDGMENTS

We thank Laura G. Lewis of the Pacific Utiliza-
tion Research Center; Lyle Morimoto and Michael
Bienn, formerly of the Pacific Utilization Research
Center for assistance in mercury analyses; and
Richard L. Major of the Northwest Fisheries
Center for determining the age of the specimens.

LITERATURE CITED

Anas, R. E.

1974. Heavy metals in the northern fur seal, Callorhinus
iirsinus and harbor seal, Phoca vitvlina richardi. Fish.
Bull., U.S., 72:133-137.
Clemens, W. A., and G. V. Wilby.

1961. Fishes of the Pacific Coast of Canada. 2d ed. Fish.
Res. Board Can., Bull. 68, 443 p.
Hall, A. S., F. M. Teeny, L. G. Lewis, W. H. Hardman, and E. J.
Gauglitz, Jr.
1976. Mercury in fish and shellfish of the northeast Pacific. I.
Pacific halibut, Hippoglossus stenolepis. Fish. Bull., U.S.
74:783-789.
Hearnden, E. H.

1970. Mercury pollution Fisheries Department acts quickly
to safeguard public health. Fish. Can. 22(10):3-6.
Malaiyandi, M., and J. P. Barrette.

1970. Determination of submicro quantities of mercury in
biological materials. Anal. Lett. 3:579-584.

796

HALL ET AL.: MERCURY IN SABLEFISH

MUNNS, R. K.

1972. Mercury in fish by cold vapor AA using sulfuric-nitric
acid/VjOj digestion. FDA (Food Drug Admin.) Lab. Inf.
Bull. 1500, 8 p.

MuNNS, R. K., AND D. C. Holland.

1971. Determination of mercury in fish by flameless atomic
absorption: A collaborative study. J. Assoc. OflF. Anal.
Chem. 54:202-205.

Schmidt, A. M.

1974. Action level for mercury in fish and shellfish. Fed.
Regist. 39(236) Part 11:42738-42740.

Spinelli, J., M. A. Steinberg, R. Miller, A. Hall, and L.
Lehman.

1973. Reduction of mercury with cystein in comminuted
halibut and hake fish protein concentrate. J. Agric. Food
Chem. 21:264-268.

Teeny, F. M., A. S. Hall, and E. J. Gauglitz, Jr.

1974. Reduction of mercury in sablefish (Anophpoma
fimbria) and the use of the treated flesh in smoked
products. Mar. Fish. Rev. 36(5):15-19.

Thompson, B. G.

1971. Fishery Statistics of the United States. U.S. Dep.
Commer., Natl. Mar. Fish. Serv., Stat. Dig. 65, 424 p.

797

ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS
(PENAEIDEA: SERGESTIDAE)

John F. Walters'

ABSTRACT

This paper describes the vertical distribution and migration, population size, seasonal size-frequency
distribution, and diet of 20 species of sergestid shrimps collected between 1970 and 1973 in the vicinity
of Oahu, Hawaii.

During the daytime, half-red sergestids live between 450 and 725 m, while all-red sergestids range
from 650 to at least 1,200 m. At night all but two species migrate into the 0- to 300-m region, half-red and
all-red groups mixing together. One nighttime group lives above 100 m, another lives between 125 and
300 m. Moonlight depresses the shallow group below 150 m; it has little effect on the deep group. In
addition, some species stop migrating around full moon, remaining at their daytime depths.

All -species examined eat zooplanktonic Crustacea in the 1- to 3-mm size range. Some species can also
utilize smaller zooplankton around 0.4-0.6 mm. This ability is unrelated to the enlarged maxillipeds
found in some species.

Most species appear to spawn mostly during the spring, although ovigerous females can be found at
any time of the year. Life span appears to be 1 yr for all species except Sergia bisulcata, which lives 2 yr.
One species does not reproduce in Hawaiian waters.

Hawaiian sergestids are specialized by size, morphology, and vertical distribution. The most closely
related species pairs are always separated by size. The Hawaiian sergestid assemblage is very similar to
assemblages reported from two areas of the tropical Atlantic.

Shrimps of the family Sergestidae (Decapoda,
Penaeidea) are one of the most characteristic
groups of micronekton over much of the open
ocean. They dominate the crustacean micronekton
over large areas of the North Pacific, where they
form sound-scattering layers (Barham 1957) and
feed baleen whales (Omori et al. 1972). Two
speciose sergestid assemblages have been de-
scribed from the subtropical Atlantic by Foxton
(1970) and Donaldson (1973, 1975). This paper
examines the sergestid assemblage from the
central Pacific near the Hawaiian Islands, report-
ing vertical distribution and migration, abun-
dance, growth and reproduction, and diet.

MATERIALS AND METHODS

Sampling Area

All the sergestids examined in this study were
collected off the leeward (west) coast of Oahu,
Hawaii at about lat. 21°30'N, long. 158°20'W. Most
trawling was done 10-25 km offshore in water
1,500-4,(X)0 m deep. Physical and chemical data for

'Department of Oceanography, University of Hawaii, 2525
Correa Road, Honolulu, HI 96822.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

this area, as well as the nearby Gollum Station (lat.
22°10'N, long. 158°00'W), have been reported by
Gundersen et al. (1972) and Gordon (1970). The
mixed layer is 50-80 m thick with a temperature of
23°-26°C. The annual variation in temperature of
the mixed layer is only about 3°C (Gordon 1970). A
broad thermocline extends to approximately 500
m, where the temperature is 5°-7°C. Salinity
varies from 34.0 "/oo at 400-500 m to 35.2 '7m at 100 m;
oxygen varies from 7 mg/liter at 100 m to 1
mg/liter at 700-900 m. The water is very clear. In
situ measurements of irradiance to 500 m at lat.
28°29'N, long. 155°14'W in August 1972 gave an
extinction coeflRcient of 0.029 m" ' at a wavelength
of 471 nm for depths below 200 m; surface ir-
radiance at 471 nm was 7 x 10^ jnW/cm'^ per nm,
decreasing to 1 x l(^^ /nW/cm^ per nm at 500 m
(E. M. Kampa, pers. commun.). Annual net
primary productivity has been estimated at 50 g
C/m2 (S. A. Cattell in T. A. Clarke 1973:431).
Nakamura (1967) found an annual mean standing
crop of zooplankton of 2.6 g/m^ in the upper 200 m.
The sampling area was chosen as the deep water
nearest to Honolulu. It has the further advantage
of being in the lee of Oahu under normal
tradewind conditions, an important practical
consideration when working from RV Teritu. In

799

FISHERY BULLETIN: VOL. 74, NO. 4

spite of its proximity to land, the area appears to
be representative of the open waters of the central
North Pacific. Meroplankton is sometimes abun-
dant, particularly larval stomatopods, but never
dominates the zooplankton. The light regime at
night may be affected by light from the urbanized
areas of Oahu, although direct light from Honolulu
is shielded by mountains. Doty and Oguri (1956)
found enhanced values of primary productivity
near the Hawaiian Islands (the "island mass
effect"), but Gilmartin and Revelante (1974) found
this effect only within about 1 km of land. The
advantages of nearness to port and convenience of
study greatly outweigh the potential disadvan-
tages of being affected by nearshore processes.

Vertical Distribution:
Teuthis Sampling Program

Most of the material studied was collected
during the "Teuthis" program, a series of 23
cruises during 1971-73 by the University of
Hawaii's RV Teritii. The primary objective of the
program was to determine the vertical distribu-
tions of the various species of micronekton during
the daytime and at night. For this purpose an
extensive series of horizontal tows was made
using a modified Tucker trawl (MT) with a mouth 3
m wide. The trawl can be opened and closed at the
desired sampling depth, avoiding contamination
of the sample by organisms from shallower depths
during setting and retrieval. It is lined with
knotless nylon mesh, with apertures about 7 mm in
diameter. The cod end is a 1-m plankton net of
303-ium Nitex.- Mounted on the trawl is a time-
depth recorder (Benthos 1170) which provides a
record of the depths sampled by the trawl.

This basic configuration was extensively
modified during the course of the sampling pro-
gram to obtain more reliable operation and better
data. The original acoustic-controlled opening-
closing system (Inter-Ocean) was replaced by a
more reliable messenger-operated double-trip
mechanism (modified General Oceanics No. 4020).
A digital flowmeter (General Oceanics No. 2030)
was added at the beginning of 1972, giving a more
accurate estimate of the volume of water sampled
by the trawl. An acoustic telemeter (AMF No.
1024) allowed real-time monitoring of trawl depth
beginning in November 1972; earlier tows wan-

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

dered vertically over 10-20% of their maximum
depth.

The limitations of time and unreliability of
sampling gear forced abandonment of plans for a
uniform series of standard tows. Each cruise
attempted instead to sample depths not yet sam-
pled or to answer questions raised by previous
sampling. Informal as this protocol was, the actual
depths sampled often differed greatly from the
plan. Before a telemeter was available, the sam-
pling depth was set by the amount of wire paid
out; two tows with the same amount of wire out
often showed a twofold variation in modal depth.
Over the course of the program the upper 1,200 m
was sampled rather thoroughly, with a few deeper
tows down to 2,300 m.

A typical cruise lasted 4 days. On each day two
tows were made during the daytime and two at
night, avoiding the twilight periods when many
mid-water animals are migrating. Tows sampled
for 3 h at a towing speed of about 4 knots. The
catch was immediately placed in chilled seawater,
and live specimens were removed to an aquarium
for observation. The rest of the catch was sorted
and preserved in buffered 5% Formalin seawater.
The inside of the net was picked clean of animals
after each tow to prevent contamination of sub-
sequent tows. Physical conditions recorded in-
cluded ship's position at the beginning and end of
sampling, weather conditions and sea state, time
of sunrise and sunset, and lunar phase. Bathy-
thermograph casts were made during the early
cruises, later replaced with expendable bathy-
thermograph casts; at least one was taken per
cruise (Maynard et al. 1975). The 1973 cruises also
recorded biological sound scattering at 25 kHz and
surface light irradiance (Walters in prep.). In the
laboratory, the sergestids were sorted to species,
sexed, and counted, and the carapace length (CL)
from the base of the rostrum to the posterior
margin of the carapace at the dorsal midline was
measured to the nearest 0.1 mm with an eyepiece
micrometer in a dissecting microscope.

Between February 1971 and June 1973, 16
cruises produced 160 horizontal tows (Table 1).
Daytime (DAY) tows were lumped together, but
nighttime tows were divided into tows during the
dark of the moon or with the moon obscured by
clouds (NIGHT) and tows made under substantial
amounts of moonlight (MOON). Total trawling
time for each 25-m interval of the water column to
1,500 m for the entire series was calculated from
time-depth records (Table 2).

800

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Cruise

Table l.-Summary of sampling, 1970-73.

Dates

Samples

Remarks

70-12

8-10, 13-17 Dec. 1970

Teuthis IV

19-21 Feb. 1971

Teuthis V

15-19 Mar. 1971

Teuthis VI

22-26 Apr. 1971

Teuthis VII

24-26 May 1971

Teuthis VIII

21-25 June 1971

Teuthis IX

30 July-1 Aug. 1971

Teuthis X

22-24 Sept. 1971

Teuthis XI

28 Jan.-I Feb. 1972

Teuthis XII

25-29 Feb. 1972

Teuthis XIII

25-29 Mar. 1972

Teuthis XV

23-27 May 1972

Teuthis XVI

29 June-2 July 1972

Teuthis XVII

1-5 Aug. 1972

Teuthis XIX

3-7 Nov. 1972

Teuthis XXI

4-7 May 1973

Teuthis XXIII

13-17 June 1973

Teuthis XVIII

30 Sept.-4 Oct. 1972

Teuthis XXII

23-27 May 1972

DSB III

2-3 Feb. 1973

Echo IV

5-11 Dec. 1973

3-m IKMT' horizontal open tows

3-m M'P horizontal opening-closing tows

3-m IKMT shallow and deep oblique open tows

3-m MT horizontal and oblique open tows
3-m IKMT stratified oblique open tows

||KMT = Isaacs-Kidd midwater trawl.
2MT = modified Tucker trawl.

Oblique Series:
Teuthis XVIII and XXII

Teuthis XVIII, 30 September to 4 October 1972,
represented a departure from our normal sam-
pling program. It consisted of a series of oblique
tows vi^ith a 3-m Isaacs-Kidd midwater trawl
(IKMT) designed to assess the relative importance
in numbers and biomass of the various groups of
micronekton, and also to determine the proportion
of the mid-water community undergoing diurnal
vertical migration (Table 1). Two series of oblique
tows were taken: "deep" tows from the surface to
1,200 m, and "shallow" tows from the surface to
400 m. The catches were preserved unsorted in 5%
Formalin seawater and returned to the laboratory,
where they were sorted into the major taxa,
blotted dry, counted, and weighed. Further details
of sampling methods and results can be found in
Maynard et al. (1975). The sergestids were divided
into half-red and all-red types, counted, and
weighed. They were later separated by species,
counted, and sexed, and the carapace length
measured.

Teuthis XXII, 23-27 May 1973, followed the
same sampling protocol as Teuthis XVIII, with
series of shallow and deep oblique tows. Sergestids
from this cruise were separated by species, count-
ed, and sexed, and the carapace length measured.

Eflfects of Moon: 70-12 and Echo IV
The Teuthis cruises were unevenly spaced in

time, making it difficult to use the data for study-
ing growth rates and other aspects of population
dynamics. In particular there were no cruises at all
between early November and late January. To fill
this gap in the seasonal coverage, I examined the
sergestids from the December 1970 cruise of T. A.
Clarke (70-12). This cruise used a 3-m IKMT for an
extensive series of 2- and 3-h horizontal open tows
in the upper 1,250 m of the water column (Table 1).
Further details of sampling can be found in T. A.
Cla'rke (1973). While the material from this cruise
helped balance the seasonal data, it raised new
questions about the vertical distribution of ser-
gestids. Many of the species in the 70-12 samples
showed abnormal vertical distributions. Since the
cruise took place near full moon, it appeared that
the abnormalities, in particular the absence of
normal vertical migration patterns in some
species until the final two nights of the cruise,
were related to lunar phase. Unfortunately, shal-
low and deep tows were not taken on the same
night, so it was unclear whether entire populations
were affected and on which night normal behavior
resumed.

The sampling program of Echo IV attempted to
clarify these problems. We planned to make shal-
low and deep oblique tows with a 3-m IKMT from
first quarter to full moon in an attempt to find
when vertical migration ceased. Mechanical
difficulties postponed the cruise until three nights
before full moon; migration had already ceased by
this time. The sampling protocol called for a
shallow tow, either 0-200 m or 200-400 m; an

801

FISHERY BULLETIN: VOL. 74, NO. 4

Table 2.-Number of tow.s and total towing time for each depth interval, 0-1,500 m, Teuthis sampling program.

DAY

NIGHT

MOON

Depth
(m)

No. of

tows

Total

tows

>10 min

min

124

106

121

105

116

122

115

150

302

160

291

609

406

508

260

370

492

559

225

274

371

388

459

341

449

379

206

190

165

185

179

197

148

166

209

126

104

No. of

tows

Total

tows

>10 min

min

433

425

313

385

381

336

486

306

131

337

149

118

123

266

135

241

140

216

147

172

249

119

—

No. of

tows

Total

tows

>10 min

min

514

350

604

124

530

324

204

335

180

208

199

162

—

211

236

293

205

210

115

143

114

—

^_.

—

100 -

200 -

300

400 -

500 -

600

700

800

900

1,000 -

1,100 -

1,200 -

1,300 -

1,400 -

1,500 -

802

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

intermediate tow of 400-600 m; and a deep tow of
400-1,200 m, eacli night of the cruise. Daytime
trawling investigated possible moon-related
changes in the daytime distribution of sergestids
and included a 400- to 800-m tow and a 600- to
1,000-m tow each day. The actual depths sampled
by the trawl deviated somewhat from the protocol,
as we used no telemetry on the trawl. The last
daytime tow was an all-day affair sampling from
1,100 to 1,900 m. The sergestids from this cruise
were identified to species and counted, but not
sexed or measured.

Feeding Study: DSB III

An important problem in any study of feeding
in mid-water animals is the effect of the sampling
gear on feeding behavior. A mid-water trawl
concentrates animals in the cod end to unnaturally
high densities. Often the trawl lumps together
animals from different depth zones. A predator
feeding on the contents of the cod end is likely to
eat prey it would not normally take in the natural
state, either because predator and prey do not
occur at the same depth or because the prey can
normally escape the predator. Examination of
sergestid stomach contents from the Teuthis
series suggested that many shrimp had been
feeding in the trawl. A modification of the trawl
became necessary to get reliable feeding data.

The DSB III cruise of 2-3 February 1973 was
designed to investigate the feeding behavior of
mid-water animals. The MT was modified by
tieing off the cod end ahead of the plankton net,
allowing zooplankton to escape through the
meshes. The trawl mouth was tied open. Daytime
and nighttime oblique and horizontal tows were
taken, the main objective being to obtain as large
and varied a collection of mid-water animals as
possible without much concern for their depth of
capture (Table 1). The samples were preserved in
5% Formalin seawater and returned to the labo-
ratory, where the sergestids were sorted out and
their stomach contents identified.

Using the MT in this fashion produced one
unexpected bonus. In addition to flushing out
prey-sized zooplankton, the water current forced
the catch and the inner lining of the net through
the coarse outer net in pockets. Within each pocket
the animals were firmly held by the force of the
water, preventing movement and feeding. Future
feeding studies might profit from deliberately
designing this effect into the sampling gear.

Analysis of Vertical Distribution Data:
The Contamination Problem

Most previous studies of vertical distribution
(e.g., Foxton 1970, T. A. Clarke 1973, Donaldson
1975) have assumed that all the animals captured
in a horizontal tow were taken at a single depth.
While such an assumption simplifies the presen-
tation and interpretation of the data, it can
produce a misleading picture of the vertical struc-
ture of the mid-water community if the tows
actually fish over a substantial depth range. Open
trawls like the IKMT are the most susceptible to
contamination of the catch by animals from other
depths, since they fish during setting and re-
trieval. In this case, contamination usually takes
the form of shallow-living animals appearing to
have been captured below their normal depth.
Rapid setting and retrieval can minimize but not
eliminate the problem (T.A. Clarke 1973). Foxton
(1970) and Donaldson (1975) have shown that
animals from other depths can contaminate IKMT
samples even when the trawl is fitted with open-
ing-closing cod end buckets. Some animals become
temporarily entangled in the net early in the tow.
When they break free later on, the trawl may be
fishing at a different depth, resulting in a sample
that mixes shallow and deep animals in an un-
known proportion.

Even an opening-closing trawl like the MT can
give misleading results if it is allowed to wander
vertically while open. In such a case, assigning the
entire catch to the modal depth broadens out the
apparent vertical range in both directions. Our
experience has shown that towing the MT deeper
than 200 m results in substantial vertical wander-
ing unless its depth is constantly monitored and
adjusted. Since a working telemeter was available
only during the latter part of our program, most of
our "horizontal" tows actually have a vertical
range of 50-100 m. The problem increases with
depth; tows below 800 m commonly wander 200 m
or more. Assigning the catch to a modal depth
would produce a misleading vertical distribution
pattern.

The vertical distribution diagrams presented in
this paper allow for vertical wandering of the
trawl and for unequal sampling time with depth.
Only horizontal tows are considered. The water
column is divided into 25-m zones, and the amount
of time each tow spent in each zone is determined
from the various depth zones in proportion to the
time towed in each zone. Let q be the number of

803

FISHERY BULLETIN: VOL. 74, NO. 4

shrimp captured by the iih tow and ^,. ^ be the
amount of time tow / spent in the Jth depth zone.
Then the proportional catch c,. ^ from the iih tow in
thejth depth zone is

Ci.j =

(1)

For each depth zone ./, summing proportional
catches from all tows and dividing by total
trawling time in the zone gives the catch rate /• :

rj =

(2)

Ideally, the catch rate is proportional to the
population density, so that dividing the catch rate
by trawl filtering rate gives an estimated popula-
tion density; i.e.,

D, =

M, •/• V

(3)

where D, is the estimated population density in the
/th zone, M,. is the effective mouth area of the trawl
(because of the design of the trawl, this quantity
decreases with increasing towing speed),,/" is the
filtering efficiency of the trawl, and r is the towing
speed.

Proportional allotment of the catch by this
method assumes that a particular shrimp is equal-
ly likely to have been captured at any instant
during the tow. This assumption is clearly false for
tows that spend only part of their time in the
shrimp's actual depth range. However, spurious
catch rates outside the actual depth range are
minimized by additional tows in these zones that
do not enter the actual depth range and do not
catch shrimp; these tows increase the denominator
of Equation (2) without increasing the numerator.
It follows that this method of estimating vertical
distributions works best when each depth zone is
sampled many times.

Table 2 shows that during the daytime all depth
zones between 400 and 1,075 m were sampled at
least five times and that at least five tows spent
more than 10 min in all zones between 425 and 950
m. Nighttime sampling was less thorough because
tows were split into two groups on the basis of
moonlight. In both groups all zones in the upper
200 m were sampled at least five times, as was the
600- to 700-m range (NIGHT) and 650- to 725-m

range (MOON). NIGHT tows in the 200- to 225-m
zone sampled only 2 min; estimated population
densities for this zone, while generally plausible-
looking, should be regarded cautiously. The 0- to
25-m zone for MOON tows were sampled many
times for brief periods by open tows that spent
nearly all their time at depths of 50-150 m, but was
never sampled extensively by any tow. Many
species show spuriously high estimated population
densities in this zone. There were no NIGHT tows
between 1,150 and 1,300 m, and no MOON tows
between 375 and 400 m or below 1,175 m. Night-
time sampling was generally sparse below 800 m,
and the estimated population densities for this
region are very crude.

A second major assumption of this method of
presenting vertical distribution data is that the
vertical distribution remains constant throughout
the sampling period, allowing data from many
different cruises to be summed together. The
resulting estimated population densities repre-
sent an average over the entire sampling period.
The actual vertical structure on any given cruise
may vary considerably from this average. The
separation of nighttime tows into NIGHT and
MOON tows is the only systematic attempt to
show variations in vertical distribution; other
variations are discussed in the species accounts.

Presentation of Results

A brief explanation will aid in interpreting the
vertical distribution figures that follow (e.g.,
Figure 1). Catch rates were converted to estimated
population densities in numbers per 10"' m' by
assuming an average trawling speed of 2 m/s,
effective trawl mouth area of 5.1 m- (at 2 m/s), and
filtering efficiency of 90%. DAY, NIGHT, and
MOON (see above) distributions are shown for the
entire population as histograms on the right side
of the figure. The number to the right of each
histogram is the sample size. In addition, the
catches were divided into size classes, and popula-
tion densities were estimated by the same method
for each size class. Species with a maximum
carapace length less than 17.0 mm were divided
into 0.5-mm classes, while larger species were
divided into 1.0-mm classes. The result was an
array of estimated population densities as a
function of size and depth. Interpolation produced
a series of contours of equal population density.
The lowest contour level represents 0.2 shrimp per
10"' m"* per mm CL; each successive contour level

804

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

represents a tenfold increase over the previous
one.

The oblique tows of September 1972 (Teuthis
XVIII) and May 1973 (Teuthis XXII) provided
data that yield two estimates of the population
densities of the various species, using the method
of Maynard et al. (1975). Summing over the entire
water column the depth-specific population densi-
ties obtained from the horizontal tows provides a
third estimate of population densities. The results
of these estimates are reported as numbers per 100
m- of ocean surface in Table 3. Sample sizes and
standard deviations are given for the mean values
of the oblique series. Because of the nature of the
calculations for the horizontal tows, no standard
deviations can be figured, but the variation is
probably of the same order as those of the oblique
series, since horizontal tows sampled each depth
interval about the same number of times and for
roughly the same total amount of time as the
oblique tows.

The Teuthis data are poorly suited for inves-
tigating growth and reproduction of sergestids.
The sampling program was designed primarily to
investigate the vertical distribution of mid-water
animals. Depth coverage varied widely from cruise
to cruise, and the cruises were spaced irregularly
throughout the year. In order to smooth the
irregularities as much as possible, the data are
lumped into 3-mo periods. The cruises invoved are:

Jan. -Mar. T4, T5, Til, T12, T13

Apr. - June T6, T7, T8, T15, T16, T21, T23

July - Sept. T9, TIO, T17, T18

Oct. - Dec. T19, 70-12.

Histograms show the size-frequency distribution
of males and females for each species. For Ser-
gestespectinatus only, data from the oblique series
of May 1973 (Teuthis XXII) are added into the
second quarter histogram.

Because of the problem of feeding in the trawl
(discussed above), only the stomach content data
from DSB III (February 1973) are presented.
Table 4 shows the average condition of the
stomach contents for each tow. The two indices
reported represent the quantity of food present
and its state of digestion. Both are based on an
arbitrary scale of 1 to 5:

2. More than half

full.

3. 25-50% full.

4. Less than 25%

full.

5. Empty.

Body still mostly intact, ap-
pendages separated, some
digestion of soft parts.

All soft parts digested, cuti-
cle remaining, usually dis-
articulated.

Cuticle broken into small
fragments.

Empty.

Contents

1. Packed full,
distended.

Digestion

Whole animal, with little
evidence of digestion.

Stomach contents with a digestion state of 1 were
seldom found in the DSB III samples but were
rather common in the Teuthis material, probably
because of feeding in the trawl.

Table 5 shows the kind and number of food items
found in the stomachs of each species. Often the
stomach contents were too well digested for
identification. Food items were not identified
beyond the general categories presented except
for the calanoid copepod genus Pleuromamma,
which has a prominent shiny knob on the side of
the metasome that is highly resistant to digestion.

RESULTS

Sergestid species occurring in Hawaiian waters
are listed in Table 6, along with the total number
caught. Serge^fes and Sergia until recently were
considered to be subgenera of genus Sergesfes s.l.;
however, Omori (1974) has rightly elevated the
subgenera to full genera. This paper follows his
usage but gives feminine endings to species of
Sergia. A paper presenting systematic descrip-
tions of Hawaiian species is in preparation.

Sergestes atlanticus Milne Edwards 1830

Vertical Distribution (Figure 1)

The normal daytime range of S. atlanticus was
550 to 725 m. Small individuals had a more re-
stricted range than the larger ones; shrimp less
than 5.5 mm CL stayed between 550 and 650 m.
Sergestes atlanticus was occasionally taken at 800
m or below. The small concentrations between 800
and 1,050 m in Figure 1 all resulted from the June
1973 cruise. In addition, the December 1970 cruise
took seven shrimp in an 800-m tow. At night S.
atlanticus occurred over a wide range from the
surface to about 300 m. The large concentration in
the upper 25 m resulted from a single large capture
in May 1973. This depth interval was extensively
sampled by only three tows, so it is unclear

805

FISHERY BULLETIN: VOL. 74, NO. 4

11 6 8 10 12 14 15 6 16

200
400
600
800
1000
1200

ORT

' ^-L-^S^k^^

NIGHT

"-m —

MOON

200
400
600
800
1000
1200

200
400
600
800

1000
1200

200
400
600
800
1000
1200

6 8 10 12 14
CPRPPRCE LENGTH (MHI

16 8 16
NO. PER 10= m3

Figure 1. -Vertical distribution of Sergestes atlanticus.

whether the bulk of the population normally occurs
so shallow. A number of night tows in the 75- to
150-m range took small numbers of 5. atlanticus,
and one tow at about 250 m captured five in-
dividuals. There appeared to be no significant
variation of depth with size. Tows on moonlit
nights took small numbers of S. atlanticus from
the surface to about 350 m, with a possible con-
centration around 150-175 m. Several captures
were made between 600 and 925 m, suggesting
that at least part of the population did not always
migrate. The December 1970 cruise took substan-
tial numbers (up to seven) near full moon between
250 and 300 m, and also between 550 and 800 m,
indicating that only part of the population was
migrating. Later in the cruise when the moon was
waning, tows between 30 and 100 m took S. atlan-
ticus in moderate numbers (four to six).

Population Size, Growth, and Reproduction (Figure 2)

Sergestes atlanticus was moderately abundant,
turning up regularly in tows at the appropriate

Table 3.-Estimated population sizes of Hawaiian sergestids from all horizontal tows and two series of oblique

tows (no. per 100 m^).

Teulhis XVIII

TeuthisXXII

Species

(')

Horizontal

(')

Sergestes atlanticus

Sergestes erectus

Sergestes armatus

Sergestes vigilax

Sergestes orientalis

Sergestes consobrinus

Sergestes sargassi

Sergestes pectinatus

DD
DN
SN

4.36
2.66
4.28

3.66
1.31
3.09

28
19
47

3.79

2.75

7.22

3.06

5.30

3.83

5.46

2.61

5.85

3.03

143

12.37

14.47

12.16

7.69

7.61

3.61

10.22

8.28

250

3.33

2.52

2.30

1.55

4.63

1.42

3.58

1.95

5.54

4.12

5.11

2.00

12.45

5.03

130

8.43

5.29

202

6.72

3.70

3.30

1.28

5.82

4.13

5.26

3.46

129

10.37

7.85

3.53

1.40

4.91

2.16

5.84

4.73

145

23.24

17.36

141

24.14

14.13

178

30.44

20.86

330

26.67

17.30

649

DD
DN
SN

1.17
1.10
0.28

1.67
1.47
0.56

0.88

1.33

1.79

1.25

1.91

1.00

2.28

0.80

1.97

1.01

1.18

1.85

1.46

1.70

0.87

0.80

1.15

1.45

0.76

0.96

0.72

0.81

0.99

0.83

0.61

0.75

0.65

0.92

0.18

0.31

0.99

0.83

0.64

0.80

2.48

1.42

2.10

1.96

1.45

1.03

2.08

1.40

1.55

1.41

1.68

0.95

2.06

1.60

1.74

1.29

6.95

4.04

2.58

1.34

2.89

2.04

4.70

3.59

D
N

2.10
0.84
0,46

180
68
19

1.31

267

5.55

542

2.02

156

2.83

151

3.81

849

2.48

251

2.63

141

1.83

2.35

433

0.52

0.28

0.15

0.35

100

1.61

160

1.39

130

0.75

42.

1.32

332

0.72

1.99

231

0.57

1.05

325

0.64

1.07

0.40

0.70

174

2.01

245

1.54

136

1.40

1.71

467

806

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Table 3.-Continued.

'DD— Deep Day tows (0-1,200 m): T18 lour tows, T22 six tows.
DN — Deep Night tows (0-1,200 m): T18 five tows, T22 three tows.
SN — Shallow Night tows (0-400 m): T18 seven tows, T22 five tows.
D — DAY Horizontal tows.
N — NIGHT Horizontal tows.
M— MOON Horizontal tows.

Teuthis

XVIII

Teuth

is XXII

Horizontal

Species

(')

Sergia tulgens

DD
DN
SN

0.35
0.29
0.63

0.69
0.40
0.60

DD
DN
SN

24.15
8.63
9.15

19.01
4.17
5.08

186
46

D
N
M

1.55
2.21
3.57

185
237
134

0.45

0.55

15.95

14.91

296

2.26

556

Sergia scintillans

DD
DN
SN

9.86

6.80

12.14

9.90

8.55
5.17
6.95

6.83

130

221

DD
DN
SN

5.12
0.72
3.80
1.90

4.41
0.81
2.17

2.25

D
N
M

3.69
3.59
2.35

355
329
151

3.31

835

Sergia gardlnerl

DD
DN
SN

8.80

11.74

9.44

2.13
9.23
3.88

100

DD
DN
SN

5.12
0.72
0.56

4.41
0.81
0.45

D
N
M

11.22
9.95
2.67

964

793

10.00

5.58

243

1.90

2.25

8.65

1,808

Sergia bigemmea

DD
DN
SN

2.50
3.19
1.67

2.04
2.57
0.95

16
23
18

DD
DN
SN

0.29
0.76
0.00

0.45
0.84
0.00

4
0

D
N
M

0.37
1.48
0.19

116

2.35

1.85

0.31

0.53

0.64

145

Sergia inequalis

DD
DN

2.01
0.61
0.57

1.76
1.37
0.60

12
5
6

DD
DN
SN

.0.38
0.57
0.00

0.63
0.56
0.00

3
3
0

D
N
M

0.63
0.76
0.19

16
50
10

0.94

1.29

0.31

0.52

0.55

Sergia bisulcata

DD
DN
SN

1.15
1.31

1.12

1.09
1.76
0.48

DD
DN
SN

1.03
0.37
0.82

1.19
0.64
0.90

8
2
6

D
N
M

1.04
1.32
1.96

91
68
38

1.19

1.08

0.82

0.96

1.35

197

Sergia tenuiremis

DD
DN

0.43
0.86
0.67

0.52
1.17

0.91

3
6
9

0.77
0.55

0.70

0.97
0.96
0.91

6
3
9

D
N
M

0.93
0.99
0.73

39
25
14

0.89

Petalidium suspiriosum

DD
DN

1.09
1.30
1.21

0.73
1.09
0.90

7
9

DD
DN

1.50
0.93

1.31

2.29
0.58

1.85

10
5

D
N
M

1.94
0.82
2.84

1.84

53
13
22

Table 4.-Feeding chronology of sergestids from DSB III.

DAY

NIGHT

(Tow no. 1-3,

12)

(Tow no

. 5-10)

Number

Empty

Number

Empty

Species

examined

(%) Content'

Digestion'

examined

(%)

Content'

Digestion'

Sergestes atlanticus

1.5

3.5

Sergestes erectus

3.7

3.9

3.3

3.8

Sergestes armatus

4.8

4.7

3.7

4.3

Sergestes vigilax

—

100

5.0

Sergestes orientalis

100

5.0

3.0

4.0

Sergestes sargassi

—

3.5

4.0

Sergestes pectinatus

—

2.7

3.6

Sergia tulgens

4.0

4.5

4.1

4.2

Sergia scintillans

2.8

3.8

Sergia gardineri

3.7

4.1

3.2

4.2

Sergia bigemmea

3.0

4.0

3.0

3.6

Sergia inequalis

—

4.0

3.5

Sergia bisulcata

3.2

4.0

3.0

3.6

Sergia tenuiremis

100

5.0

—

Total sample

4.1

4.2

205

3.1

3.7

'See text.

depths, but seldom in numbers greater than five or
six for a 3-h tow. The average population density
estimated from all horizontal tows was 1.31 per 100

m'-. Daytime tows caught larger numbers than
nighttime tows, the population density from
daytime horizontal tows being 2.10 per 100 m-. The

807

FISHERY BULLETIN: VOL. 74, NO, 4

Table 5.-Diet of sergestids from DSB III.

«9

c
5

<0
3

5
o>

to
to
to
o>

to
c

2
o

■p

o
to

Number

of shrimp

0>l

containing

<1>
to

(1>
to

Calanoid copepods

Pleuromamma'

Cyclopoid copepods

Amphipods

Ostracods

Euphausilds

Decapod larvae

Bivalve larvae

Foraminifera

Chaefognath spines

Unidentified Crustacea

Fibrous matter

Others

Empty

Total number examined

'Included witfi calanoid copepods.

^Including gastropod larvae, radiolarians, pteropods, fish eggs, and fish scales.

Table 6. -Numbers of Hawaiian sergestids captured, 1970-73.

Half-red species

Sergestes atlanticus Milne-Edwards
Sergestes cornutus Kr^yer
Sergestes erectus Burkenroad
Sergestes armatus Kr^yer
Sergestes vigilax Stimpson
Sergestes orientalis Hansen
Sergestes tantillus Burkenroad
Sergestes consobrinus Milne
Sergestes sargassi Ortmann
Sergestes pectinatus Sund
Sergia fulger)s (Hansen)
Sergia scir^tillans (Burkenroad)

All-red species
Sergia gardineri (Kemp)
Sergia bigemmea (Burkenroad)
Sergia inegualis (Burkenroad)
Sergia bisulcata (Wood-Mason)
Sergia maxima (Burkenroad)
Sergia tenuiremis (Kr0yer)
Petal idium suspiriosum Burkenroad

546

1,371

1,113

271

1,030

647

497

1,541

1,118

1,610

3,096
398
149
350
2
147
170

SERGESTES ATLANTICUS

'O"! JAN -MAR

September 1972 oblique series gave a figure of 3.79
per 100 m'-, and the May 1973 oblique series yielded
0.88 per 100 m'-'; these figures are probably close to
the maximum and minimum population density.

Recruitment was highest during the third
quarter (July-September), the only time of year
when immature shrimp less than 4 mm CL were
taken. The largest shrimp were most abundant
during the second quarter (April-June).

Diet (Table 5)

Only three individuals examined had recogniz-
able stomach contents: a calanoid copepod {Pleu-

808

25-
20-
15-

OCT- DEC

nJ"

10-
5-

-1

CARAPACE LENGTH Imm)

Figure 2.-Quarterly size-frequency distribution of Sergestes

atlanticus.

romamma), an amphipod, and fragments of larval
bivalve shells.

Sergestes cornutus Kr«Syer 1855
Vertical Distribution

Only four individuals were captured in horizon-

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

tal closing tows; these were all daytime tows
between 450 and 550 m. Several were captured in 0-
to 400-m oblique night tows, indicating that S.
cornutus is a vertical migrator. Donaldson (1975)
found S. conmtus mostly in the upper 50 m at
night.

Population structure and diet were not studied
because of the small sample size.

tures in the 350- to 475-m range; immatures were
found between 175 and 250 m. The peak in the
upper 25 m is a sampling artifact. There was no
positive evidence of full moon nonmigration in the
horizontal samples, but the December 1970 cruise
took a dozen shrimp in the 575- to 700-m range,
suggesting that about 20-30% of the population
was not migrating.

Sergestes erectus Burkenrozd 1940
Vertical Distribution (Figure 3)

Sergestes erectus was abundant in our collection,
but nearly half of the shrimp came from daytime
tows of the November 1972 cruise. The daytime
vertical range was about 550 to 800 m, with
maximum catches between 625 and 750 m. Imma-
ture shrimp did not occur below 750 m. The night-
time range varied with size. Small immature
shrimp less than 12 mm CL occurred between the
surface and 200 m, mostly below 125 m. Inter-
mediate-sized shrimp between 12 and 16 mm CL,
including immature and newly mature shrimp,
ranged between 150 and 250 m. Adult shrimp
larger than 16 mm CL were found between 250 and
325 m. Moonlight depressed the vertical range of
adults very little, although there were some cap-

H 8 12

16 . 20 24 28

200

DAT

UOO

54«

600

2>3'

^---

■

800
1000

1200

•

"^ " o

200

s^ — * ' —

fc=n:D

1>»

cs cl*Nj|

MOO

NIGHT

DEPTH (M

1000

■

1200

■

200

MOON

151

400

600

■

800

1000

1200

200
1400
600
800
1000
1200

200
UOO
600
800
1000
1200

200
100
600
800
1000
1200

8 12 16 20 2M 28
CflflRPRCE LENGTH IMMl

20 HO
NO. PER 10^ m3

Figure 3.— Vertical distribution of Sergestes erectus.

Population Size, Growth, and
Reproduction (Figure 4)

Sergestes erectus was the second most abundant
species in the horizontal series, the average
population density estimated from all horizontal
tows amounting to 3.81 per 100 m-. Like S. atlan-
ticus, it was taken in larger numbers during the
daytime than at night, the population density
estimated from daytime horizontal tows amount-
ing to 5.55 per 100 m-. These numbers reflect its
extreme abundance during the November 1972
cruise, when as many as 157 were taken in a single
3-h tow. The oblique series of September 1972 and

SERGESTES ERECTUS

.oA

p ''< I — ^

6 10 14 18

APR -JUN

JAN - MAR

^V^^T>.

o-" [—' — I 1 r

6 10 14 18 22 26

r-^ :i^->y%

"T 1 r

6 10 14 18

6 10 14 18 22 26

>" 0 ■ 'i^— I 1 P" — 0 '"'i 1 1 1 T 1

6 10 14 18

6 10 14 18 22 26

CARAPACE LENGTH (mm)

Figure 4.-Quarterly size-frequency distribution of Sergestes

erectus.

809

May 1973 yielded figures of 5.85 and 1.97 per 100
m^, respectively.

Recruitment was not noticeably high during
any particular quarter. However, medium-sized
shrimp in the 10- to 14-mm CL range were
significantly more abundant during the fourth
quarter (October-December) than at other times
of the year (Kolmogorov-Smirnov test, P<0.05).

Diet (Table 5)

Calanoid copepods made up the bulk of the
stomach contents of the S. erectus from DSB III. A
few amphipods and a single euphausiid were also
found. One individual had some material very
tentatively identified as a small fish, the only one
found in the DSB III collection. No food items in
the 0.4- to 0.6-mm size range were found.

Sergestes armatus KrtSyer 1855

Vertical Distribution (Figure 5)

The daytime vertical distribution of 5. armatus
varied somewhat with size. Immature shrimp
ranged between 450 and 600 m; adults were gener-
ally between 550 and 650 m, but sometimes as
shallow as 450 m. One tow in November 1972 took
13 shrimp at about 675 m. The December 1970

1>4

200

100

600

800
1000
1200 •

200
400
600
800
1000
1200

DfiT

NIGHT

200
MOO
600
800
1000
1200

200
400
600
800
1000
1200

6 8 10 12
CflfifiPfiCE LENGTH IMM)

10 20
NO. PER 10^ m3

Figure 5.- Vertical distribution of Sergestes armatus.

FISHERY BULLETIN: VOL. 74, NO. 4

cruise took nine shrimp in open tows below 800 m;
most of these were probably contaminants. The
nighttime range also varied with size; shrimp
smaller than 8 mm CL usually occurred between
100 and 200 m, while adults were found mostly
between 150 and 300 m, with occasional captures
as deep as 450 m. Moonlight did not depress the
vertical distribution of 5. armatus. The peak in the
upper 25 m is a sampling artifact. The open tows of
the December 1970 and December 1973 cruises
took small to moderate numbers of S. armatus at
the daytime depth. If these shrimp were not
contaminants, they suggest that about 5-20% of
the December 1973 population was not migrating.

Population Size, Growth, and Reproduction (Figure 6)

Sergetes armatus was abundant in the horizon-
tal series, the average population density of 2.35
per 100 m'-' estimated from all horizontal tows
making it the fourth most abundant sergestid.
The catch was even greater during the September

1972 oblique series, which yielded a figure of 10.22
per 100 m'-\ second only to S. pectinatus. The May

1973 oblique series took much smaller numbers,
amounting to only 1.51 per 100 m^'.

Recruitment was much higher during the second
quarter (April-June) than during the rest of the
year. Large individuals were most abundant dur-
ing the fourth quarter (October-December).

SERGESTES ARMATUS

- jAN-MAR

..^A

-t JAN -MAR

"T T "'I 1 r

,.4^-n

Pu-'Jk

q[\-nArin/x^

30-

JUL - SEP

20 H

JUl

10-

-1

-n rfl n, ,-Y^

, r— V-

T 1 ■ — t

14 2

CARAPACE LENGTH (mm)

ja»

Figure 6.-Quarterly size-frequency distribution of Sergestes

armatus.

810

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Diet (Table 5)

DSB III took over 40 S. armatus, but two-thirds
had empty stomachs. Food items included calanoid
copepods, amphipods, an euphausiid, and an
unidentified decapod larva. Prey in the 0.4- to
0.6-mm size range, such as foraminifera or bivalve
larvae, were not found.

Sergestes vigilax Stimpson 1 860

Vertical Distribution (Figure 7)

The daytime vertical range of 5. vigilax was
about 550 to 725 m, with a concentration at about
675 m. Nighttime captures were all in the 0- to
200-m depth range, peaking at about 50-75 m.
Moonlight depressed the peak to 150-200 m, but
some individuals remained shallower. There was
no evidence of full moon nonmigration.

Population Size, Growth, and Reproduction (Figure 8)

Sergestes vigilax was not abundant in Hawaiian
waters. The average population density estimated
from all horizontal tows was only 0.35 per 100 m-.
Daytime catches were larger than night catches,
the population estimate from the day tows being

4 6 8 10

12 14 16

2 4

200

ORT

: ,

600

F "^^

800

1000

■

1200

■

^yr? C3

NIGHT '

200

^ r :

400

■

600

■

800

■

1000

•

1200

200

400

■ ■ 'm^ . . . .

MOON

^ ■

600

800

1000

1200

200
400
500
800
1000
1200

200
400
600
800
1000
1200

6 8 10 12 14
CBRflPHCE LENGTH (MM)

16 2 4
NO. PER 10^ m3

Figure 7.- Vertical distribution of Sergestes vigilax.

0.52 per 100 m-. The oblique series of September
1972 and May 1973 took larger numbers, yielding
estimates of 3.57 and 0.61 per 100 m-, respectively.
These larger population sizes may indicate that
the finer mesh of the I KMT sampled S. vigilax, a
relatively small species, more efficiently than did
the MT used for the horizontal tows.

The seasonal size-frequency histograms are not
significantly different from one another. Shrimp
less than 5 mm CL were most abundant in the
third quarter (July-September).

Diet (Table 5)

Only a single individual was examined; it had an
empty stomach.

Sergestes orientalis Hansen 1919

Vertical Distribution (Figure 9)

The daytime vertical distribution of S. orien-
talis varied with size; small shrimp less than 6.5
mm CL were taken from 450 to 575 m, while larger
ones were found between 500 and 625 m, mostly
between 550 and 600 m. The nighttime range was
from the surface to 125 m, with largest numbers in
the 25- to 50-m and 75- to 100-m zones. Small
shrimp less than 6 mm CL stayed above 75 m.
Moonlight depressed most of the nighttime

SERGESTES VIGILAX

JAN - MAR

, JAN - MAR

4-
0

■-■ n "^ n

4-
0

o e-

'6' OCT- DEC

12-

C4RAPACE LENGTH (mml

Figure 8.-Quarterly size-frequency distribution of Sergestes

vigilax.

811

population into the 100- to 200-m range, although
some remained shallower. A few nighttime cap-
tures were made in the daytime depth range. The
December 1970 cruise took large numbers near full
moon at 550-600 m (up to 40), and also at 150-200 m
(up to 21). Apparently at least 50^;^ of the popula-
tion was not migrating. Later in the cruise, when
the moon was waning, large numbers of S. orien-
folis were taken in tows between 30 and 120 m (up
to 70). There was no evidence of full moon nonmi-
gration during the December 1973 cruise.

Population Size, Growth, and
Reproduction (Figure 10)

Sergci^tes orientalis was moderately abundant
in Hawaiian waters. The average population
density estimated by all the horizontal tows was
1.32 per 100 m'-, daytime and night tows giving
similar figures. The oblique series of September
1972 yielded a higher figure of 8.43 per 100 m'-', S.
oriental h being the second most abundant species
in the shallow night tows. On. the other hand, it
was much scarcer during the oblique series of May
1973, which gave a population density of only 0.64
per 100 m'-. Sergestes orientalis was particularly
abundant during the December 1970 cruise, when
as many as 70 were taken in a single 3-h IKMT
tow.

The seasonal size-frequency histograms are all
very similar to one another. Shrimp smaller than 6
mm CL were proportionally most abundant during
the first quarter (January-March), but the dif-
ference was not statistically significant
(Kolmogorov-Smirnov test, P> 0.05).

Diet (Table 5)

Only two individuals from DSB III were ex-
amined. One had an empty stomach; the other had
eaten an ostracod.

Sergestes tantillus Burkenroad 1940

Vertical Distribution

Because of the rarity of 5. tantillus, little can be
inferred about its vertical distribution. Single
shrimp were taken in daytime tows between 410
and 915 m. The largest night catch was at 50 m
(four shrimp), with individual captures to about
200 m. A tow between 635 and 715 m on a moonlit
night took six shrimp.

812

FISHERY BULLETIN: VOL. 74; NO. 4

U 6 8 10 12 14 16 8 16

200
400
600

800
1000
1200

200
MOO
600
800
1000
1200

200
400
600
800
1000
1200

DRT

o mt^

NIGHT

MOON

■r
■

200

uoo

500

800

1000

1200

200
400
600
800
1000
1200

• 200

400

600

800

1000

1200

8 10 12 14 15

15
nS m3

CflRRPHCE LENGTH IMMl NO. PER 10= M^

Figure 9.- Vertical distribution of Sergestes orientalis.

SERGESTES ORIENTALIS

60-

OCT ■

D£C

40-

•—I

20-

iJ ' *^^— 1 ' 1-

CAR4PACE LENGTH (mml

Figure lO.-Quarterly size-frequency distribution of Sergestes

orientalis.

Growth, reproduction, and diet were not studied
because of the small sample size.

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Sergestes consobrinus Milne 1 968

Vertical Distribution (Figure 1 1)

Nearly two-thirds of the captures during the
Teuthis series were from shallow night tows
during the May 1973 cruise (Teuthis XXI); it was
also fairly abundant in the oblique series of
September 1972 and May 1973 (Teuthis XXII).

10 20

200

MOO ■

500

800
1000
1200

200

400

600

800
1000
1200

200
400
600
800
1000
1200

DfiT

NIGHT

€S>

-fer^

MOON

<s>

200
400
600
800
1000
1200

200
• 400
600
800
1000
1200

200
400
600
800
1000
1200

6 8 10 12
CRRHPRCE LENGTH (MMl

10 20
NO. PER 10^ m3

Figure IL-Vertical distribution of Sergestes consobrinus.

Sergestes consobrinus was broadly distributed
during the daytime, from 450 to 725 m. The
population maximum appeared to be around 600
m, but most daytime catches were small. A few
captures were made between 800 and 950 m; these
may have been contaminants. The nighttime
distribution showed a broad peak from the surface
to 75 m, with lesser numbers to 125 m. These
numbers were strongly influenced by the May 1973
captures. Moonlight depressed most of the
population to 100-150 m, with a substantial
number remaining at the daytime depth. The
December 1970 cruise took S. consobrinus near full
moon in tows between 140 and 180 m, and also in a
700- to 800-m tow. Later catches when the moon
was waning were in the upper 120 m, with a large
catch at 30 m.

Population Size, Growth, and
Reproduction (Figure 12)

Like S. vigilax, S. consobrinus appears to have
been undersampled by the MT. The average
population size estimated by all horizontal tows
was 1.05 per 100 m'-. The figure for only the night
tows was 1.99 per 100 m-, reflecting the large night
catches of the May 1973 cruise (Teuthis XXI)-up
to 76 in a single 3-h tow. The oblique IKMT series
of September 1972 and May 1973 (Teuthis XXII)
yielded higher figures of 5.42 and 2.08 per 100 m-,
respectively, presumably because the finer mesh of
the IKMT retained more of the small shrimp.

The seasonal size-frequency histograms show a
maximum proportion of small individuals in the
third quarter (July-September). The largest
shrimp were taken in the first and second quarters,
although first quarter catches were small.

Diet was not examined, since none were taken
during DSB III.

Sergestes sargassi Ortmznn 1893

Vertical Distribution (Figure 13)

With the possible exception of S. cornutus,

SERGESTES CONSOBRINUS

15-
10-

JAN -MAR

30-

20-

10
CO

_) 0

30
20

10-

5soH

JUL-SEP

- JUL-SEP
30-

20
10-

30-
20 -

OCT-OEC

|— 1 '-

10 -

0 -J

=L_^

2 *

30-

OCT-OCC

20-

10-
0-

— ' I

CARAPACE LENGTH (mm)

Figure 12.-Quarterly size-frequency distribution of Sergestes
consobrinus.

813

FISHERY BULLETIN: VOL. 74, NO. 4

14 6 8 10 12 lU 16

200

OflT ■

400
600

: "^^<X^

800

1000

r 2

•

1200

■

\ ^'^^p^c

NIGHT ;

^ \

200

MOO

600

BOO

1000

■

1200

200
HOO

m m>

MOON

600

800

1000

1200

■

200
MOO
600
800
1000
1200

200
400
600
800
1000
1200

SERGESTES SAR0AS3

M 6 8 10 12 14 16 6 12
CRRHPfiCE LENGTH (MMl NO. PER 10= M^

Figure 13.- Vertical distribution of Sergestex xargaasi.

which is very rare in Hawaiian waters, S. sargastii
had the shallowest daytime range of the local
species: 450-575 m, with a maximum around 475 m.
No significant variation of depth with size was
noted, perhaps because of the small number
caught. One immature shrimp was captured
between 340 and 425 m, and oblique tows from the
surface to about 350 m took a few immature
specimens in September 1972. There was a pro-
nounced variation of size with depth at night.
Immature individuals less than 6 mm CL occurred
between 100 and 200 m, mostly in the 125- to 150-m
range. Larger shrimp were found from 125 to 300
m, mostly from 225 to 275 m. Most of the adults
captured at night were males; the few females
were mostly taken between 125 and 175 m. This
apparent segregation by sex was probably a
sampling artifact, since the December 1970 cruise
took both males and females in tows from 250 to
300 m. Moonlight had little effect on the depth
range of adults; immature individuals were
depressed to about 150-200 m. The peak in the
upper 25 m is a sampling artifact. There was no
evidence of full moon nonmigration.

Population Size, Growth, and Reproduction (Figure 14)
Sergestes sargassi was not very abundant in the

^^^

^^[\

(J " T I [

2 20.

20-

JUL

SEP

10 -

nHJ

""-1-^

,_r^

■ ■■!

t 1

CARAPACE LENGTH (mm)

Figure 14. -Quarterly size-frequency distribution of Sergestes

sargassi.

horizontal collections, the estimated average
population density being only 0.70 per 100 m-. The
night tows gave a higher total than the daytime
tows, 1.07 and 0.64 per 100 m-, respectively. The
two IKMT oblique series of September 1972 and
May 1973 produced higher estimates, 5.84 and 1.74
per 100 m-', respectively; the daytime figure of
10.37 per 100 m- in September was the third
highest such total for that cruise. These higher
figures for the IKMT tows were not the result of
more eflRcient filtering, as appears to be the case
for the smaller species, since S. sargassi is a
moderately large species, about the same size as
Sergia scintillans and Sergia gardineri, neither of
which showed any signs of significant undersam-
pling by the MT.

Recruitment was highest during the second
quarter (April-June), when nearly 70% of the
population was immature. Growth during the
summer was about 1.0-1.3 mm CL per month,
slowing during the third and fourth quarters to
0.2-0.5 mm CL per month. Maximum sizes were
attained in December and the first quarter
(January-March).

Diet (Table 5)

DSB III took 10 S. sargassi. Three had empty
stomachs. The rest had eaten zooplanktonic

814

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Crustacea, including calanoid copepods and a
euphausiid.

Sergestes pectinatus Snnd 1920

Vertical Distribution (Figure 15)

The daytime range of 5. pectinatus was broad,
extending from 425 to 725 m. The peak at 425-450
m came from a single tow in the June 1971 cruise,
and the peak at 650-675 m was also from a single
tow in the November 1972 cruise. Most large
catches centered around 575 to 625 m. There was a
poorly defined size-depth trend. Small shrimp less
than 4 mm CL seldom occurred below 600 m, while
the very large females seldom occurred above 550
m. At night the size-depth trend was pronounced.
Males less than 3.5 mm CL were found in the upper
100 m, mostly between 25 and 75 m. From 4 to 5
mm CL, maximum catches were in the 75- to 250-m
range, peaking around 150 m. The largest males
were taken in the 200- to 275-m range. Females
showed a similar trend; maximum catches of
shrimp less than 4.5 mm CL occurred around 50 m,
increasing to 150 m for shrimp between 4.5 and 6
mm CL, and 200 m for shrimp larger than 6 mm
CL. A few shrimp were taken below 300 m; these
may have been contaminants. The moon depressed

most of the population to about 150-250 m. The
peak in the upper 25 m is a sampling artifact.
None of the Teuthis samples showed any indica-
tions of full moon nonmigration. The December
1970 cruise took 14 specimens in three open tows
between 400 and 600 m, probably representing less
than 10% of the population.

Population Size, Growth and Reproduction (Figure 16)

Sergestes pectinatus appeared to be signif-
icantly undersampled by the MT. The average
population density estimated by all horizontal
tows was 1.71 per 100 m-. The IKMT with its finer
mesh captured many more shrimp than the MT.
Sergestes pectinatus was the most abundant ser-
gestid in the September 1972 oblique series, which
yielded a population density estimate of 26.67 per
100 m^. The shrimp from this cruise composed
nearly 40% of the entire catch of 5. pectinatus. The
May 1973 series gave a figure of 4.70 per 100 m'^,
second only to S.fulgens. In both cases, the aver-
age size of an individual was considerably smaller
than in a typical MT tow. Interpretation of the
size-frequency histograms is complicated by the
undersampling problem. For 5. pectinatus only,
data from the May 1973 oblique series were added
to the second quarter horizontal data. This means

n 6 8 10 12 in 15

200

DRT

■

1400

24S

600

. 5 •

800

■

1000

•

1200

•

NIGHT ;

■ r^ \

200
UOO

r-5-

3 13»

600

■

800

■

1000

■

1200

•

^ a^

MOON ;

)
1

■

200

>100

■

600

•

800

•

1000

•

1200

•

200

noo

600

800

1000

1200

200

noo

600

800

1000

1200

200
400
600
800
1000
1200

M 6 8 10 12 14 16 6 12
CRRRPRCE LENGTH (MM) NO. PER 10= M^

Figure 15.- Vertical distribution of Sergestes pectinatus.

SERGESTES PECTINATUS

JAN->UR

^:x

rO.

6 8 2

CARA«C£ LENGTH (mml

Including TnMilt 1001

Figure 16.-Quarterly size-frequency distribution of Sergestes
pectinatus. April-June quarter includes data from Teuthis XXII.

815

FISHERY BULLETIN: VOL. 74, NO. 4

that I KMT data were included in three of the four
quarters, only the first quarter (January-March)
lacking IKMT data. Small shrimp were propor-
tionately most abundant during the third quarter
(July-September), and large shrimp were most
abundant during the first quarter, although lack of
IKMT data probably affected the shape of the first
quarter histogram.

Diet (Table 5)

Seven of the twelve shrimp from DSB III had
eaten calanoid cope pods, mostly Pleuromamma
spp.

Sergia fulgens (Hansen 1919)

Vertical Distribution (Figure 17)

Because of the peculiar fluctuations in abun-
dance during the course of the sampling program,
the vertical distribution patterns of 5. fidgeits
derived from the data should be regarded strictly
as estimates. All the daytime captures lay between
550 and 625 m; there was no variation in depth
with increasing size. The open tows of the
December 1970 cruise took nine specimens
between 525 and 630 m. Most nighttime captures
lay between 75 and 125 m for immature shrimp
less than 8 mm CL, with some as shallow as 25-50
m. Nearly all the adults came from a single tow at
150-200 m; a few captures came as shallow as 75 m.
Almost all of the captures near full moon came
during the June 1973 cruise, which took immature
shrimp between 250 and 475 m; there were three
captures of adults between 150 and 325 m. The
peak in the upper 25 m is a sampling artifact. The
December 1970 cruise took nine adults in open
tows between 160 and 300 m and one adult at 400
m. There was no evidence of full moon non-
migration.

Population Size, Growth, and
Reproduction (Figure 18)

Sergia fulgens fluctuated drastically in abun-
dance during the sampling program. The first 13
cruises of the Teuthis series (Teuthis IV-XVII,
February 1971-August 1972) caught a total of 13
specimens. After the September 1972 cruise it
turned up in many tows, often in very large
numbers. Nearly all the specimens were immature
shrimp less than 10 mm CL. However, one hor-

M 6

10 12 14 16

20 UO

200

DRT

400

•

185

600

' — I—

800

1000

1200

P , 1

200

;j7

400

t NICMT

600

■

800

•

1000

•

1200

■

200

^^^^^ £> MOON

1)2

' . ' ■

400

-■-^

600

•

800

1000

1200

4 6 8 10 12 14 16
CHRfiPfiCE LENGTH IMMl

NO.

20 40
PER 10= m3

200
400
600
800
1000
1200

200
400
600
800

1000
1200

200
400
600
800
1000
1200

Figure 17. -Vertical distribution of Serbia fulgens

SERGIA FULGENS

JAN-UAR

1 1 ^ 1 r

JAN

WAR

'-'

1 " 1

3 0-

^ ^,

iia

.=^50-

-D£C

■ r"^n ■?=-

CARAPACE LENGTH (mm)

Figure 18.-Quarterly size-frequency distribution of Sergia

fulgens.

izontal night tow in May 1973 took 89 adults. In
addition, the December 1970 cruise caught a total
of 21 S. fulgens; 19 of these were adults. Combining
the very low numbers from the first 13 horizontal
series with the very high numbers from the last 3
horizontal series gives an average population

816

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

density of 2.26 per 100 m-, an estimate of doubtful
meaning. The two oblique series produced very
different estimates. The September 1972 cruise
gave a figure of only 0.43 per 100 m'-, making S.
fulgens the least abundant of the 16 regularly
occurring species. On the other hand, the May 1973
cruise gave a figure of 15.95 per 100 m-, more than
3 times greater than any other species, and 37
times the September figure.

The seasonal size-frequency histograms reflect
the fact that nearly all S. fulgens were caught in
the second and third quarters. The second quarter
(April-June) histogram is trimodal. The peaks at
5.5 and 8 mm CL represent the same cohort as
sampled in May and June (the Teuthis XXII
oblique series of late May took 6.5-mm shrimp),
giving a growth rate of 2.1-2.2 mm CL per month
for immature shrimp in this size range. The
assumption that the peak at 13-15 mm in May is
the same cohort as the peak at 7.5 mm from the
preceding November yields a growth rate of
1.0-1.2 mm CL per month, reflecting a slowing of
the growth rate as the shrimp approach maturity.
The presence of large numbers of immature
shrimp in the second and fourth quarters implies
that S. fulgens either has a very broad spawning
period or has two widely separated spawning
peaks.

Diet (Table 5)

DSB III took nine S. fulgens. Seven of these had
food in their stomachs, including a calanoid
copepod, an amphipod, and an ostracod, plus
smaller prey including larval bivalve and
foraminifera.

Sergia scintillans (Burkenroad 1940)

Vertical Distribution (Figure 19)

The vertical distribution of 5. scintillans
showed a slight tendency for smaller shrimp to live
deeper than larger ones, both day and night.
Daytime ranges were about 575 to 700 m for
individuals less than 7 mm CL and 525 to 650 m for
those larger than 7 mm CL, with maximum catches
between 575 and 625 m. The small peak at 325-350
m resulted from two shrimp taken in a tow that
dipped as deep as 480 m; they were probably
captured at the deep end of the tow. At night the
adults were mostly between 25 and 125 m, but
immature shrimp less than 6 mm CL ranged

200

voo

600

800

1000

1200

200
MOO

600

800

1000

1200

200
UOG
600
800

1000
1200

8 10 12 m 16

' » T I I T T T — <

DRY

NIGHT

MOON

200
MOO
600
800
1000
1200

200
UOO
600
800
1000
1200

200
WO
600
800
1000
1200

H 6 8 10 12 lU 16 MO 80
CflRfiPflCE LENGTH (MMl NO. PER 10^ m3

Figure 19.- Vertical distribution of Sergia scintillans.

between 50 and 225 m. Although the population
was centered at 100-125 m for both sexes, few
males occurred shallower than 50 m. Three shallow
tows from the May 1973 cruise that caught 40
females and 12 males are primarily responsible for
this difl!"erence. Moonlight depressed the depth of
most of the population to 100-275 m, peaking
around 200 m. The peak in the upper 25 m is a
sampling artifact. There was no evidence of full
moon nonmigration.

Population Size, Growth, and
Reproduction (Figure 20)

Sergia scintillans was one of the most abundant
sergestids in Hawaiian waters. The average
population density estimated by all horizontal
tows was 3.31 per 100 m'-', the daytime and night-
time figures being similar. It was particularly
abundant in the shallow night tows of the May
1973 cruise (Teuthis XXI), one 3-h tow taking 179
shrimp. The oblique series of September 1972 and
May 1973 (Teuthis XXII) produced figures of 9.90
and 3.70 per 100 m-, respectively.

Small shrimp were proportionally most abun-
dant in the third quarter (July-September). First
and second quarter populations were similar in
size-frequency, although the larger females oc-
curred in the second quarter (April-June).

817

FISHERY BULLETIN: VOL. 74, NO. 4

SERGIA SCINTILLANS

40 - JUL - SEP

1 0
0

60 -

OCT-O^C

40 -

-1 ^^

20 -

"U ,

CARAPACE LENGTH (mm)

Figure 20.-Quarterly size-frequency distribution of Seryia

ncintillans.

Diet (Table 5)

The DSB III material showed that 5. scintiUans
ate the usual variety of zooplanktonic Crustacea,
including calanoid copepods, amphipods, and an
ostracod. The 0.4- to 0.6-mm size fraction was also
taken; bivalve larvae, foraminifera, and cyclopoid
copepods were found in many individuals. Other
food items included the large cyclopoid copepod
Sapphirina, a larval decapod, and masses of an
unidentified greenish, fibrous material.

Sergi'a gardineri (Kemp 1913)

Vertical Distribution (Figure 21)

Sergia gardineri was usually found between 650
and 775 m during the daytime, although shrimp
smaller than 5 mm CL seldom occurred below 700
m. The extremely high values in this range were
largely due to the catches of the November 1972
cruise. On certain occasions the population seemed
to extend downward to at least 1,200 m. The June
1973 cruise took 59 specimens in three tows
between 850 and 1,050 m, and only 8 specimens in
four tows between 650 and 850 m. The December
1970 cruise caught only nine specimens in an open
tow at 650-680 m, but tows below 800 m caught

large numbers, including 77 in a tow from 1,150 to
1,250 m. On the other hand, all four daytime tows
on the May 1972 cruise between 650 and 950 m took
only one shrimp.

The nighttime distribution was strongly
influenced by large catches from the May 1973
cruise. It showed a concentration in the upper 150
m, with shrimp less than 6 mm CL restricted to
25-100 m. All large shrimp in the upper 25 m were
females, the result of a single tow in May 1973 that
fished between 15 and 45 m, taking 36 adult
females and 1 very small male. A tow at 20 m on
the same cruise took no S. gardineri, indicating
that this species probably does not reach the
surface. There were a few captures below the
normal range on moonless nights, notably a 250-m
tow in September 1971 that took four, and a 480- to
550-m tow in November 1972 that took five.

Most captures of 5. gardineri on nights with
much moonlight were at the daytime depth, except
for the March 1971 cruise, which took 16 shrimp at
320-340 m and 20 shrimp at 100-150 m, although a
tow at 170-200 m did not take any. Three open tows
near full moon on the December 1970 cruise took
207 S. gardineri between 700 and 1,000 m, while a
550- to 600-m tow took 9. Later in the cruise when
the moon was waning, they were captured at 80
and 30 m (but not at 100-110 or 50 m!). The

6 8 10 13

CflRflPflCE LENGTH (MM)

16 60 \i

NO. PER 10=

Figure 21.- Vertical distribution of Sergia gardineri.

818

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

December 1973 cruise took only two immature
specimens at night in 10 oblique tows less than
650-m maximum depth. Six oblique tows from 400
to 1,200 m took a total of 62 shrimp, although high
seas resulted in some catch spillage in two cases.
Sergia gardineri clearly does not migrate near full
moon.

Population Size, Growth, and
Reproduction (Figure 22)

Sergia gardineri was by far the most numerous
sergestid in the Teuthis collections, the average
population density of 8.65 per 100 m- estimated
from all horizontal tows being more than twice as
high as the next most abundant species. Horizon-
tal tows taking more than 100 shrimp occurred in
February 1971 (night), June 1971 (day and night),
November 1972 (day), and May 1973 (night). In
addition, the open tows of the December 1970
cruise took large numbers, including 129 in a night
tow. Sergia gardineri appeared to have been much
less abundant during the first half of 1972, al-
though most of these cruises occurred near full
moon, when the normal vertical distribution pat-
terns seem to be disrupted. The estimate from
night tows affected by moonlight, 2.66 per 100 m-,
was much lower than the daytime or moonless
night estimates. The oblique series of September

SERGIA GARDINERI

40-

JAN - MAII

■^

40-

JAN- HAR

_^r--|

4 6

•

160-

APH - JUN

160-

APR - JUN

120-

•0-

""'"

•0-

40-

_H-LJ^-^

40-

4 •

1 1 —

• •

JUL- it*

juL-sep

40-

-L^_

0-1 1 1 —

4 •

•

6 B

200-

OCT -OtC

200-

OCT-OCC

i_n

l«0-

IM-

120-

■0-

to-

40-

H "-i

_^^_^

l-rJ

-• k.

0-1

4 «

•

CARAPACC

LCI

«GTH (mm)

Figure 22.-QuarterIy size-frequency distribution of Sergia

gardineri.

1972 and May 1973 gave figures of 10.00 and 1.90
per 100 m'-', respectively.

Recruitment was highest during the third
quarter (July-September), although small shrimp
began to enter the population in June. The median
carapace length increased from 4.9 mm to 6.4 mm
between the September 1972 and November 1972
cruises, giving a growth rate of about 1.2 mm CL
per month. From November to May the growth
rate was much lower, about 0.25 mm CL per month.
The average size of females was largest in May,
although a few very large females were still
present in June. Sergia gardineri has a total life
span of about 1 yr.

Diet (Table 5)

Thirteen of the nineteen specimens of S. gar-
dineri taken during DSB III had empty stomachs.
The others contained calanoid copepods, an
ostracod, a larval decapod, bivalve larvae, fora-
minifera, and greenish fibrous matter.

Sergia bigemmea (Burkenroad 1940)

Vertical Distribution (Figure 23)

Most of the few daytime captures of S. bigem-
mea during the Teuthis series were of immature
shrimp less than 8 mm CL. A tow between 610 and
690 m took 15 in July 1971; 5 were caught in
November 1972 in a tow probably around 750 m.
The peak around 1,100 m resulted from two tows
that fished as shallow as 820 m. Two of the three
daytime captures of adults during the Teuthis
series were between 1,000 and 1,100 m; the other
was around 750-850 m. The December 1970 cruise
took 20 adults in open tows between 800 and 1,200
m. The nighttime distribution varied with size;
shrimp smaller than 10 mm CL generally occurred
between 50 and 225 m, while the adults ranged
between 125 and 250 m. The February 1973 cruise
(DSB III) took several large hauls of S. bigemmea,
including 49 specimens in a 1-h tow at 150-175 m.
Only a few were caught under moonlit conditions;
most of these were between 250 and 350 m. The
December 1970 cruise took 5 S. bigemmea at 250 m
and 11 at 750 m, indicating that much of the
population was not migrating.

The vertical distribution patterns of S. bigem-
mea appeared to be affected by avoidance. While
the females of most sergestid species grow con-
siderably larger than the males, in S. bigemmea

819

FISHERY BULLETIN: VOL. 74, NO. 4

6 8 10

12 14 16

5 10

200

DOT

■

400

■

600

<^^

^ .

800

•

1000
1200

■

*^-^^ii

^^^

-^ 1

200

J '~

118

400

NIGHT

600

800

1000

1200

200

MOON

400

600

800

1000

1200

200
400
600
800
1000
1200

6 8 10 12 14
CRRfiPfiCE LENGTH (MM)

5 10
NO. PER 10^ m3

Figure 23. -Vertical distribution of Sergia bigemmea.

the maximum size was the same in both sexes,
suggesting that the largest shrimp were escaping
capture. Daytime catches were much smaller than
nighttime catches, indicating that avoidance was
more effective during the day. However, the max-
imum size captured was the same during the
daytime as at night. It is curious that neither
Sergestes erectus nor Sergia fulgens showed any
signs of avoidance, though those caught are larger
than S. bigemmea; perhaps S. bigemmea is par-
ticularly fast for its size or better at sensing the
approach of the trawl.

Population Size, Growth, and
Reproduction (Figure 24)

Sergia bigemmea was one of the less common
sergestids in our collection; the average population
density estimated from all horizontal tows was
only 0.64 per 100 m-. Most catches occurred at
night, the figure for nighttime tows being 1.48 per
100 m'-. Very few S. bigemmea were captured
during the first half of 1972, when most sampling
was done near full moon. The oblique series of
September 1972 took moderate numbers, produc-
ing a population density figure of 2.35 per 100 m-,
higher than any other all-red sergestid except S.
gardineri. It was also moderately abundant dur-

SERGIA BIGEMMEA
d

JAN - MAR

jj-q-

-x-A

m!^

- JUL - SEP

r^' — I r

£ 10

JUL- SEP _

n nlh^n

10 14

£lU

Ar^

CARAPACE LENGTH (mm)

Figure 24.-Quarterly size- frequency distribution of Sergia

bigemmea.

ing the December 1970 cruise, which had the only
large daytime catch: 23 in an 800- to 900-m open
IKMT tow. The largest catches of S. bigemmea
occurred during the February 1973 cruise (DSB
III) when it was the most abundant species taken,
with 49 in a 1-h open tow. The May 1973 oblique
series took only a handful, giving a population
density estimate of 0.31 per 100 m-.

None of the seasonal size-frequency histograms
are significantly different from the others
(Kolmogorov-Smirnov test: P>0.05). Females
larger than 12 mm CL were proportionately most
abundant in the third quarter (July-September).

Diet (Table 5)

The surprisingly large catch of S. bigemmea
during DSB III produced a more detailed picture
of its diet than for the other species. Only 11 of the
88 shrimp had empty stomachs. Sergia bigemmea
ate crustacean zooplankton, including calanoid
copepods, amphipods, and ostracods; ostracods
■appeared to be a more important prey item than in
the other species. Smaller prey were also

820

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

eaten-larval bivalves, small cyclopoid copepods,
and foraminifera commonly occurring in the diet.
Sergia bigemmea was the only species in which
chaetognath spines were found. Other food items
included the large cyclopoid copepod Sapph iriua, a
single larval decapod, and unidentified fibrous
matter.

Sergia inequalis (Bur ken road 1940)

Vertical Distribution (Figure 25)

As with S. bigemmea, S. inequalis may have
avoided the trawl. The few daytime captures were
nearly all below 750 m; the peak near 550 m
resulted from a tow in June 1971 that dipped to 760
m. Maximum daytime depth appeared to be 1,100-
1,200 m. The December 1970 cruise took S. in-
equalis in open tows between 800 and 1,250 m-
seven in a 950- to 1,000-m tow and four in a 1,150-
to 1,250-m tow. The nighttime distribution varied
with size. Small shrimp less than 12 mm CL were
found in the upper 100 m; larger shrimp occurred
between 100 and 250 m. Moonlight did not
significantly affect the adults; there were no
captures of small shrimp under these conditions.
The Teuthis series showed no evidence of full
moon nonmigration, but the December 1970 cruise

200

DfiT

1400

600

800
1000

•
i ©•

•

k ,.

1200

,fc=^-^°-

■L 1. « 1 l,..t_

<. . ^

^mj=^

200

.— '

■" ■ »

400

NIGHT

■

600

800

■

1000

■

1200

■

200

MOON ;

:^ '■ ■

400

600

■

800

■

1000

■

1200

4 8 12 16 20 24
CflRflPflCE LENGTH (MM)

2 4
N6. PER IQS m3

200
400
600
800
1000
1200

Figure 25.- Vertical distribution of Sergia inequalis.

took five specimens at night between 550 and
800 m.

Population Size, Growth, and
Reproduction (Figure 26)

Sergia inequalis was not abundant in Hawaiian
waters; the average population density estimated
from all horizontal tows was only 0.55 per 100 m^,
less than any other regularly occurring all-red
sergestid. The largest catch of adults was only
seven, from an open tow in December 1970. The
oblique series of September 1972 and May 1973
gave estimates of 0.94 and 0.31 per 100 m-,
respectively.

In spite of its relative rarity, S. inequalis
showed a clear seasonal cycle of growth, although
because of the small sample size, the differences
among histograms are only marginally significant
statistically (Kolmogorov-Smirnov test, II dif-
ferent from III, 0.10 >P>0.05). Recruitment was
greatest in the second quarter (April-June), and

SERGIA INEQUALIS

c5 o

rHiH

4 8 12 16 20
APR - JUN

I ■ 1 r

4 e 12 l« 20

■II

-| r— — 1 1 r—

4 8 12 16 20 20

8 - JUL- SEP

-1 1 1 1 r

4 8 12 16 20

■•— ^ 1 I I

-1 1 r

4 8 12 lb 20 24

OCT

DEC

4 -

_r^

-AA

-1 — I I

4 8 12 '6 20 24

CARAPACE LENGTH (mm)

Figure 26.-Quarterly size-frequency distribution of Sergia

inequalis.

821

FISHERY BULLETIN: VOL. 74, NO. 4

the population increased in average size of in-
dividuals in succeeding quarters, the largest
females being proportionately most abundant in
the first quarter (January-March).

Diet (Table 5)

The single S. inequalis taken by DSB III had a
calanoid copepod in its stomach.

Sergja hisulcata (Wood-Mason 1891)

Vertical Distribution (Figure 27)

As with S. bigemmea, equality in size of the
sexes and small daytime catches indicate that 5.
hisulcata was avoiding the trawl. Immature
shrimp were mostly taken between 675 and 750 m
during the daytime, adults mostly from 700 to 900
m, with a few catches as deep as 1,100 m. The
December 1970 cruise took 19 individuals, includ-
ing both immatures and adults, in an open tow
from 650 to 680 m, with much smaller catches down
to 1,200 m. At night, immature shrimp occurred
between 175 and 300 m, adults mostly from 225 to
350 m, with occasional captures as deep as 450 m.
Moonlight depressed the population below 300 m;
two tows at 450 m during the June 1973 cruise took

M 8

12 16 20

211

10 20

200

OflT

400

•

600

_r-' 91

800

=\^:,-^

1000

1200

■

200

^Z^^

NIGHT ■

1 TS'i':';',H3

. >

1400

r^ :

600

800

1000

1200

200

MOON

»»--'*^''''"-^'taaaa>

400

ci"*1Do

S^ .,

600

^^j- — ^^-^

800

■

1000

1200

200
400
600
800
1000
1200

4 8 12 16 20 24 28
CFWflPflCE LENGTH IMM)

10 20
NO. PEH 105 m3

10 and 8 individuals, respectively. There was no
evidence of full moon nonmigration.

Population Size, Growth, and
Reproduction (Figure 28)

Sergia bisiilcata was the second most abundant
all-red sergestid in the Teuthis collection, though
far below S. gardineri in numbers. The average
population density figure from all horizontal tows
was 1.35 per 100 m'-. The figure for tows on moonlit
nights was higher, 1.96 per 100 m-', probably a
sampling artifact. The two oblique series produced
similar numbers; September 1972 gave 1.19 and
May 1973 gave 0.82 per 100 m-.

While quarterly variations in the size-frequency
distributions of most Hawaiian sergestids suggest
that they live about 1 yr, only in S. hisulcata is
there evidence for a longer life span. Small imma-
ture shrimp around 7-9 mm CL were recruited in
the second quarter (April-June) and grew to
sexual maturity at about 14-18 mm CL in 1 yr.
They continued to grow at a rate of approximately

SERGIA BISULCATA

-\ JAN - MAR

OH 1 1—"— 1 — ' *! " ^ T

6 10 14 18 22 26

gj APR-JUN

r In,

10 14 18 22 26

JUL -SEP

P . np ,r

6 10 14 18 22

10 14 18 22 26

_ OCT -DEC

j£y

1 I r^ T r^

18 22 6 10

CARAPACE LENGTH (mm)

-r

Figure 27.-Vertical distribution of Sergia hisulcata.

Figure 28.— Quarterly size-frequency distribution of Sergia

hisulcata.

822

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

0.6 mm CL per month for up to one additional year.
Few males appeared to live beyond 18 mo, but a
few large females greater than 22 mm CL were
probably a full 2 yr old. The pattern of growth
appears clear cut, but the small sample size means
that the data should be treated cautiously. The
Kolmogorov-Smirnov test showed that only the
third and fourth quarter size-frequency curves for
the females were significantly different from each
other (0.05>P>0.01).

Diet (Table 5)

Seven of the eight 5. bisulcafa taken by DSB III
had food in their stomachs. Food items included
ostracods. an amphipod, foraminifera, bivalve
larva, and crustacean remains probably including
a euphausiid and a larval decapod. No copepods
were found, probably because of the small sample
size.

Sergia maxima (Bur ken road 1940)

Only two individuals of this species were cap-
tured, one on the March 1972 cruise in a 480- to
615-m daytime tow and the other on the December
1973 cruise in an open tow between 400 and 550 m.
Both individuals were immature males.

Sergia tenuiremis (KreSyer 1855)

Vertical Distribution (Figure 29)

During the daytime most of the population was
below 800 m, although shrimp were sometimes
taken as shallow as 700-750 m. A single immature
individual was taken in June 1971 between 610 and
690 m. The deepest capture was in a tow between
1,220 and 1,500 m in August 1972. Tows below 1,500
m did not capture S. tenuiremis, but total trawling
time in this region was rather small. Immature
shrimp less than 15 mm CL were vertical migra-
tors, moving up to 300-500 m at night. The adult
population did not migrate as a whole, but part
spread upward at night as shallow at 550-600 m.
Moonlight had no effect on the nighttime vertical
distribution of S. tenuiremis.

Population Size, Growth, and
Reproduction (Figure 30)

Sergia tenuiremis is not abundant in Hawaiian
waters. The average population density estimated

8 12 16 20 24 28

8 12 16 20 2U 28
CflRflPfiCE LENGTH IMM)

2 M
NO. PER 105 n3

Figure 29.- Vertical distribution of Sergia tenuiremis. NIGHT
and MOON data combined.

SERGIA TENUIREMIS

-I \ 1 r

8 16

APR - JUN

« 4 -

M n -

1^, ,

8 16

■■T

"T 1 1 ^ • — r— — 1 r

8 16 24 32

O 4-

JUL -SEP

1 1

1 I ' I '

a 16

a„

-I — I I I I

8 16 24 32

,r~i I

24 8 16

CARAPACE LENGTH (mm)

Figure 30.-QuarterIy size-frequency distribution of Sergia

tenuiremis.

by all horizontal tows was 0.89 per 100 m^, day and
night values being similar. The two oblique series
produced slightly smaller values, 0.67 per 100 m- in
September 1972 and 0.70 in May 1978. Since these

823

series sampled only to 1,200 m, they may have
missed the deeper portion of the population.

The seasonal size-frequency histograms are not
significantly different from one another.

Diet (Table 5)

DSB III took only three specimens, all with
empty stomachs.

Sergia laminata (Burkenroad 1940)

Vertical Distribution

Only four individuals were captured in closing
tows, all in daytime tows during the November
1972 cruise. A tow at 650-725 m took three shrimp,
and a tow at about 750-800 m took one shrimp. At
night oblique tows in the upper 400 m took single
shrimp during the September 1972 and May 1973
cruises, suggesting that S. laminata may be a
vertical migrator. On the other hand, the
December 1970 cruise captured one shrimp in an
open horizontal tow at 550-600 m at night, sug-
gesting that S. laminata may not migrate near
full moon.

The small sample size did not allow studies of
growth, reproduction, or diet.

Petalidium suspiriosum Burkenroad 1937

Vertical Distribution (Figure 31)

A deep-living nonmigrator, P. suspiriosum
generally stayed below 800 m day and night. The
shallowest captures came during the June 1972
cruise, which took six in a 750- to 800-m day tow
and five in two night tow^s between 630 and 720 m.
Maximum depth appeared to be at least 1,500 m; as
with 5. tenuiremis, limited trawling below 1,500 m
did not catch any P. suspiriosum.

Population Size, Growth, and
Reproduction (Figure 32)

Petalidium suspiriostim is more abundant than
its small numbers in our collection would seem to
indicate, since the depths below 800 m where it
lives were not as thoroughly sampled as the shal-
lower waters. The average population density
estimated from all horizontal tows was 1.84 per 100
m-, making it the second most abundant all-red
sergestid. Like S. tenuiremis, the oblique series

FISHERY BULLETIN: VOL. 74, NO. 4

10 12 14 16

U 6 8 10 12 14 16
CmfiPfiCE LENGTH (MM)

NO. PER 10=' M

5 m3

Figure 31. -Vertical distribution of Petalidium suspiriosum.
NIGHT and MOON data combined.

PETALIDIUM SUSPIRIOSUM

^^il

I — I

(/>

< 4 -

2 -

V) 0

n, nn

JZL-Q.

__L oJ , n . I h n I . In

^^i^

10 12

CARAPACE LENGTH (mm)

Figure 32.— Quarterly size-frequency distribution of Petalidium
suspiriosum.

gave lower numbers, 1.21 per 100 m- in September
1972 and 1.31 in May 1973, probably because some
of the population was below the 1,200 m maximum
of the oblique tows.

Because of its susceptibility to damage, it was
possible to make accurate measurements of car-
apace length on only about two-thirds of the
specimens in the collection. There was no

824

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

significant seasonal trend in the size-frequency
distributions of P. suspiriosum.

DSB III did not take P. suspiriosum, so its diet
was not examined.

DISCUSSION

Color Pattern and

Daytime Vertical Distribution:

Role of Countershading

Sergestids display two basic color patterns. One
group, including Sergestes and species of Sergia in
Yaldwyn's (1957) "S. challengeri" species group, is
"half-red!' that is, its members are semitrans-
parent except for the eyes and viscera, with red,
stellate, subcuticular chromatophores scattered
over the body and appendages, most concentrated
on the cephalothorax. All half-red sergestids have
well-developed photophores; Sergestes species
have internal photophores, the organs of Pesta,
and the half-red Sergia species have external
cuticular lensed photophores. The other group,
including the remaining species of Sergia and
Petal id i inn, is "all-redT that is, its members are
covered with a relatively uniform red cuticular
pigment. All-red sergestids have simple lensless
cuticular photophores or else lack photophores
altogether.

Foxton (1970) showed that most mid-water
decapods in the Fuerteventura area (Canary Is-
lands) are either half-red or all-red. He found that
half-red shrimps generally live shallower than 700
m during the daytime, while all-red shrimps
generally live below 700 m. He concluded that the
half-red color pattern and complex photophores
are adaptations for concealment by countershad-
ing to match the light intensity of the surrounding
waters when viewed from any angle, the photo-
phores producing a ventrally directed beam of
light to fill in the shadow of the animal. He
suggested that the half-red pattern gives way to
the all-red pattern at the depth where biolumi-
nescent light becomes more important than pene-
trating surface light. Although many all-red
decapods have simple photophores, he concluded
that their function does not involve daytime
countershading. Donaldson (1975) did not discuss
this phenomenon, but an examination of his ver-
tical distribution data for the Bermuda area shows
the same daytime pattern of shallower half-red
sergestids and deeper all-red sergestids, the
dividing line again being approximately 700 m.

Other mid-water animals show similar depth-
related changes in color patterns during the
daytime. Badcock (1970) noted that mesopelagic
fishes in the Fuerteventura area tend to be silvery
above 650-700 m and dark below that depth.
Amesbury (1975) found the same pattern in
Hawaiian mesopelagic fishes, several independent
analyses of community structure locating a major
faunal boundary at 675-700 m between mostly
silvery shallow mesopelagic fishes and mostly dark
deep mesopelagic fishes.

Figure 33 shows how half-red and all-red ser-
gestids difi'er in depth during the daytime in
Hawaiian waters. The half-red species range from
425 to 725 m, with maximum abundance in the 600-
to 625-m interval. The all-red species range from
625 to 1,500 m, with maximum abundance at
700-725 m.

Rather surprisingly, the depths of maximum
abundance for the two types are only 100 m apart,
and there is a large amount of overlap in their
ranges, particularly in the zone between 650 and
725 m. Nearly half of the half-red sergestids below
650 m are Sergestes erectus, a species often taken
in large numbers in tows that also take large

NUMBER /lO^M^
HALF-RED ALL-RED

150 100 50 0 50 100 150

-1 1

300-

1500-'

Figure 33.-Daytime vertical distribution of half-red and all-red
Hawaiian sergestids. Half-red species are on the left (half-red
Sergia spp. crosshatched), all-red species on the right. Scale of
light intensity is from unpublished data of E. M. Kampa, at lat.

28°N.

825

FISHERY BULLETIN: VOL. 74. NO. 4

numbers of Sergia gardineri, the most abundant
all-red sergestid. To see if this overlap is real and
not an artifact produced by vertical excursions of
the trawl or seasonal variations in the position of a
sharper transition depth, Teuthis XIX extensively
sampled the 600- to 800-m zone in November 1972,
using depth telemtery to try and maintain the
trawl within a 25- to 50-m depth range. One tow
between 630 and 680 m took 139 Sergesfes erectus
and 31 Sergia gardiyieri, another from 650 to 730 m
took 157 Sergesfes erectus and 289 Sergia gardin-
eri, and a third from 700 to 740 m took 19 Sergestes
erectus and 312 Sergia gardineri. On this occasion,
at least, substantial numbers of both color pat-
terns were living between 650 and 725 m.

Other investigators have found similar transi-
tion zones. In Hawaii, Riggs (pers. commun.) has
found that the all-red species Gennadas propin-
quiis (Penaeidae, Benthesicymae) lives as shallow-
as 600 m, with maximum numbers at 650-675 m.
Ziemann (1975) obtained similar results for an-
other all-red shrimp, Systellaspis debilis (Caridea,
Oplophoridae), 75% of the adult population being
found above 650 m on one occasion. In the Atlantic,
Foxton's (1970) data show the half-red Sergestes
coruiculiirn (closely related to S. erectus) extend-
ing to at least 800 m, overlapping the ranges of the
all-red species Sergia rohusta and Systellaspis
debilis (although most of the catch of the latter
species were lighly pigmented juveniles).
Donaldson's (1975) data show a transition zone
from 650 to 800 m occupied by the half-red Ser-
gestes atlanticus and S. corniculum and the all-red
Sergia grandis. In view of this extensive overlap
in the distribution of half-red and all-red
decapods, it is necessary to review the conditions
under which countershading is an effective con-
cealment strategy and, in particular, Foxton's
conclusion that only half-red decapods counter-
shade.

The angular distribution of light in the meso-
pelagic environment is independent of solar
elevation and depth (Denton and Nicol 1965). At
any given point, the background light intensity is
highest directly overhead, falling off rapidly to the
sides, with a very low light intensity of back-
scattered light from below. The intensity of the
background light 90° from the vertical is only 3-4%
of the zenith value, decreasing to 0.3-0.5% at 180°
from the zenith (Tyler and Preisendorfer 1962).
Changes in surface irradiance or depth change the
intensity but not its angular distribution. Coun-
tershading mechanisms match the animal to this

background pattern; thus mid-water fishes use a
dark dorsal surface, silvery sides, and ventral
photophores for countershading (W. D. Clarke
1963; Nicol 1967; Badcock 1970). Foxton (1970)
concluded that the half-red coloration of shallow
mesopelagic decapods is a countershading mech-
anism using transparency rather than reflectors
for lateral countershading. I propose that some
deep mesopelagic all-red decapods also counter-
shade ventrally and that ventral countershading
can be effective below the transition zone from
half-red to all-red decapods.

As depth increases and the intensity of the
penetrating light dwindles, bioluminescence
becomes relatively more and more important as a
source of light in the mesopelagic environment.
Bioluminescent light has a much different tempo-
ral and spatial distribution from the penetrating
surface light. The bioluminescent light field is the
sum of glows and flashes from many point sources
whose angular distribution is more or less random.
Countershading is an ineffective concealment
strategy against bioluminescence; the silvery
sides which camouflage a mid-water fish against
the penetrating sunlight may in deeper water
reflect a bioluminescent flash and reveal the fish
against a black background. The best strategy of
concealment in an environment lit only by random
flashes is to be as nonreflective as possible. The
dark brown or black fishes and all-red Crustacea of
the deep mesopelagic zone reflect blue light poorly
(Nicol 1958), presumably indicating their use of
this strategy.

Another effect of increasing depth is that the
penetrating light eventually becomes too dim
to be seen. The absolute visual threshold for
deepsea fishes has been estimated as about
3 X 10-20 ^w / cm2 by Clarke and Denton (1962), a
figure that undoubtedly varies in other groups of
animals correlated with the degree of develop-
ment of the eye. A slightly higher intensity is
required before countershading becomes neces-
sary. The maximum depth of effective counter-
shading depends on the angular distribution of the
penetrating light; thus in Hawaiian waters the
threshold of lateral countershading is reached
110-120 m higher in the water column than the
threshold for ventral countershading. Between
these two depths lateral countershading is not
needed but ventral countershading can still be
effective.

The all-red sergestids with photophores appear
to combine an antibioluminescent color pattern

826

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

with a ventral array of simple photophores for
low-intensity ventral countershading. This inter-
pretation implies that the transition from half-red
to all-red sergestids at 650-725 m marks the upper
limit of bioluminescence as an important source of
ambient light. The lower limit of all-red sergestids
with photophores should then mark the threshold
of ventral countershading. Unfortunately, this
study produced good daytime vertical distribution
data for only one such species, Sergia gardineri. If
its lower limit under normal conditions is typical
of the other species, then the threshold of ventral
countershading should lie at approximately 775 m.
This depth is also the approximate upper limit of
the two sergestids, Sergia tenuiremis and Pef-
alidiuni siispiriosum, that lack photophores
(Figures 29, 31). If 775 m is the threshold of ventral
countershading, then the threshold for lateral
countershading should be 110-120 m higher or
about 660 m, approximately the depth of the
transition from half-red to all-red sergestids.

It thus appears that the transition from half-red
to all-red sergestids does not mark the absolute
lower limit of countershading, but is related to the
depth at which lateral countershading becomes
ineffective and bioluminescent light forces a
change in concealment strategy. Although the
all-red color pattern hides the shrimp from
bioluminescent flashes, enough of the penetrating
light remains directly overhead that ventral coun-
tershading continues to be eff"ective more than 100
m below the transition zone. The simple lensless
photophores of the all-red sergestids presumably
produce low levels of light in this dimly lit region.
Other mid-water animals that lack lateral coun-
tershading mechanisms but have ventral arrays of
photophores, such as many of the black stomiatoid
fishes, may have evolved the same kind of
camouflage.

Nighttime Vertical Distribution
and Migration

The structure of the sergestid assemblage
changes drastically as day gives way to night. All
Hawaiian sergestid species except Sergia tenui-
remis and Petalidiuni suspiriosum migrate into
the upper 300 m of the water column. As figure 34
shows, the division into a shallow half-red and a
deep all-red mesopelagic sergestid assemblage
disappears on moonless nights. The species fall
with little overlap into a shallow and a deep
migratory group, adults of the shallow group

NUMBER /IO^m'
HAIF-RED ALL-RED

J50 100 50 0 50 100 150

NIGHT

1200-"

Figure 34.-Nighttime vertical distribution of half-red and
all-red Hawaiian sergestids (moonless conditions). Hachure, etc.,
as in Figure 33. Light intensity estimated from unpublished
daytime data of E. M. Kampa at lat. 28°N, using G. L. Clarke's
(1968) values for relative intensity of day vs. night.

living in the upper 100 m, adults of the deep group
living from 125 to 300 m. The shallow group
includes Sergestes vigilax, S. consobrimis, Sergia
scintillans, and 5. gardineri. The deep group
includes Sergestes erectus, S. armatus, S. sargassi,
Sergia bigemmea, S. inequalis, and S. bisulcata.
The single large nighttime capture of adult Sergia
fulgens is at about 175 m, probably placing this
species also in the deep group. In addition, Ser-
gestes pectinatus is broadly distributed from 25 to
250 m, and S. atlanticus may likewise be broadly
distributed if the single large catch in the upper 25
m is not representative of its normal distribution.
T. A. Clarke (1973) found a similar pattern in the
nighttime distribution of Hawaiian myctophid
fishes, with a shallow group down to 125 m, a deep
group at 150-250 m, and a few species broadly
distributed in the upper 250 m. Closely related
pairs of myctophid species separate into a shallow
species and a deep species. Most sergestid species
pairs are found at the same nighttime depths,
except for Sergestes armatus and S. vigilax, and
probably Sergia scintillans and S. fulgens.

The division of the nighttime sergestid as-
semblage at 100-125 m may possibly be related to

827

FISHERY BULLETIN: VOL. 74, NO. 4

the penetrating surface light. The scale at the left
of Figure 34 is an estimate, derived by assuming
the value of light intensity at the surface on a
moonless night to be 10' times fainter than during
the day, a figure used by G. L. Clarke (1968), and
applying this correction to the daytime light curve
of E. M. Kampa (unpubl. data) used in Figure 33.
The lower limit of daytime ventral countershad-
ing, estimated above as approximately 775 m, is
equivalent to a nighttime depth of approximately
125-150 m, suggesting that the shallow group, but
not the deep group may countershade at night.
The lower limit of daytime lateral countershading,
estimated as about 660 m, is equivalent to about 50
m at night, the approximate upper limit of S.
gardineri, the shallowest all-red sergestid at
night. Although these figures admittedly pile
estimate on estimate, they suggest that light may
influence the vertical distribution of sergestids at
night as well as during the daytime.

W. D. Clarke (m Barham 1970:118) and Foxton
(1970) have suggested that countershading may
occur primarily at night in some mid-water an-
imals. While nighttime ventral countershading
appears feasible for some species of Hawaiian
sergestids, these species all may need to counter-
shade during the daytime also (Table 7). A number
of species maintain approximately constant illu-
mination day and night. Some species live in much
brighter waters during the daytime than at night.
No species, however, lives in brighter waters at
night than during the daytime, as would be ex-
pected if countershading were occurring only at
night.

Only two species of Hawaiian sergestids
definitely do not migrate. Sergia tenuiremis ap-
pears to migrate to 300-400 m when less than 15

Table 7.-Estimated light intensities for daytime and dark night
habitats of Hawaiian sergestids. Numbers are negative loga-
rithms of light intensity (smaller numbers mean brighter light).

Daytime

Night

Species

habitat

Sergestes atlanticus

5.0- 6.7

5.5- 6.0

D o*' N

Sergestes erectus

5.5- 6.7

9.0-10.0

D >>N

Sergestes armatus

4.7- 6.0

8.2- 9.7

D >>N

Sergestes vigilax

5.0- 6.7

5.5- 7.7

D /~ N

Sergestes orientalis

4.0- 5.2

5.5- 7.2

D ^ N

Sergestes consobrinus

4.7- 5.2

5.5- 6.2

D > N

Sergestes sargassi

3.5- 5.0

7.7- 9.5

D >>N

Sergestes pectinatus

4.2- 6.2

7.5- 9.0

D > N

Sergia fulgens

4.7- 5.5 ?

7.7- 8.2 ?

D > N ?

Sergia scirttillarts

4.2- 6.5

5.7- 7.2

D '-' N

Sergia gardineri

4.7- 7.2

5.7- 7.2

D /-* N

Sergia bigemmea

(7.0-11.0)

7.7- 9.0

D z-* N

Sergia inequalis

(7.0-11.0)

7.7- 9.0

D z-' N

Sergia bisulcata

6.5- 8.7

8.7-10.2

D > N

mm CL. The adult population spreads upward
from an upper limit of 750-800 m during the day to
about 600 m at night, although many shrimp
remain in the daytime depth range. Petal id ium
sHspiriosKm remains below 750-800 m both day
and night. Few Hawaiian sergestids occupy the
depths between 300 and 600 m at night, in contrast
to the Atlantic, where Donaldson (1973) found
Sergestes corniculum and Sergia grandis and
Foxton (1970) found Sergestes corniculum and
Sergia robusta in this depth range. The reasons for
this difference are unknown.

Although considerable evidence links diurnal
vertical migration to the diurnal light cycle (e.g.,
Marshall 1954), the exact relation of light to
vertical migration is complex and poorly under-
stood. The simplest scheme, merely maintaining a
constant light intensity around the clock, is not
used by all Hawaiian sergestids, as Table 7 shows.
The daytime sergestid assemblage cannot shift en
masse to equivalent light levels at night, because
the light intensity at the surface on a moonless
night is approximately equivalent to that at 600 m
during the daytime. Instead, we find species with
similar daytime ranges but different nighttime
ranges, such as Sergestes armatus and S. conso-
brinus; species with similar nighttime ranges but
different daytime ranges, such as Sergia scintil-
lans and S. gardineri; and species that exchange
relative positions, such as Sergestes sargassi and
Sergia gardineri. Vertical migration is a more
complicated behavior than merely maintaining a
constant light level.

A further complication of the vertical migration
mechanism involves the response of sergestids to
moonlight. When the moon increases the night-
time surface irradiance, the two groups of migra-
tors react in different ways. The deep group
remains relatively unaffected by moonlight, the
young often moving downward to the depth of the
adults. Moonlight drastically affects the shallow
group, depressing most of the species below 150 m.
The two assemblages, which separate by depth on
dark nights, mix together on moonlit nights.

In addition to the normal response of sergestids
to moonlight, there appears to be a period of about
a week around full moon when some species stop
migrating entirely, remaining at their daytime
depths. This behavior is poorly shown by the
results of the Teuthis cruises, showing up better in
the supplementary data from 70-12 and Echo IV.
Not all species react the same way to the full moon
period. Sergia gardineri and probably Sergestes

828

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

aflanticus appear to stop migrating altogether.
During cruise 70-12 part of tlie populations of
Sergestes orientalis and S. consohrinus migrated,
while the rest of the populations remained at their
daytime ranges; during Echo IV both species
appeared to migrate normally. Other species, such
as Sergestes vigilax and Sergia scintillans, have
shown no indications of nonmigratory behavior.
Species showing the best evidence of nonmigra-
tion are all members of the shallow migratory
group, but sampling was inadequate to determine
definitely whether any species in the deep group
are also nonmigrators. The data from cruise 70-12
and the June 1973 Teuthis cruise further suggest
that when a species is not migrating its daytime
depth can also be abnormal. Sergia gardineri,
normally found between 650 and 775 m during the
daytime, was taken as deep as 1,200 m on these
cruises, and Sergestes atlanticus, normally found
between 600 and 725 m, was taken down to 800 m.

The nonmigration of some sergestids around
full moon is a separate behavior from the moon-
light depression below 150 m. Nonmigration is not
a direct effect of increased light levels. During
Echo IV the moon was often heavily obscured by
cloud, yet the nonmigratory species remained
deep. During Cruise 70-12, nonmigratory species
remained deep until the next-to-last night, when
normal migration resumed, although light inten-
sity in the surface waters could not have been
radically different than on the previous night.
Nonmigratory behavior has been observed in
December and June, suggesting that it occurs
during most seasons of the year.

Studies of seasonal variation in vertical migra-
tion can be complicated by moon effects. For
example, Donaldson (1973) found abnormally deep
distributions day and night for Sergia splendens
during a February 1972 cruise. He also cited data
from the same cruise showing that sergestid
numbers were strongly influenced by moonlight in
the upper 100 m, both at the quarter and at full.
Knowing only that the moon was in various phases
during the February 1972 cruise, it is impossible to
separate seasonal effects from moonlight effects
for 5. splendens. Other mid-water groups show
nonmigratory behavior not tied to lunar phase.
Riggs (pers. commun.) found that species of the
penaeid shrimp Gennadas did not migrate during
our November 1972 cruise, which sampled near
new moon when sergestids appeared to be mi-
grating normally, and concluded that a seasonal
factor was involved. In summary, the depth struc-

ture of the mesopelagic community changes in a
bewilderingly complex manner under the
influence of ambient light, lunar phase, season,
and probably other undiscovered effects.

Feeding Chronology and Diet

In studies of the diets of mesopelagic animals,
the time of day when feeding takes place is as
interesting a datum as the kinds of prey eaten.
The most widely accepted theory of the function of
vertical migration holds that mesopelagic animals
move into the food-rich shallow water at night to
feed in the dark and retreat into deeper water at
sunrise to escape the efficient visual predators of
the epipelagic zone (Marshall 1954). If this theory
is correct, an examination of the feeding chron-
ology of vertical migrators should reveal that the
majority, at least, of feeding occurs at night.

Table 4 compares the stomach contents of day-
caught with night-caught sergestids from the
DSB III cruise of February 1973. The night sam-
ples as a whole had a lower percentage of empty
stomachs, a greater amount of food in the
stomachs, and a lesser degree of digestion than the
day samples, indicating that most feeding oc-
curred at night. Unfortunately, only two species
were abundant both day and night. Sergestes
armatus fed more at night than during the day-
time, although most specimens had empty
stomachs regardless of time of day. Sergestes
erectus actually had a lower percentage of empty
stomachs during the daytime than at night, but
the night specimens on the average were fuller
than the day specimens. Other studies of feeding
chronology in sergestids, notably those of Omori
(1969) on Sergia lucens, Judkins and Fleminger

(1972) on Sergestes similis, and Foxton and Roe
(1974) on a variety of Atlantic species, also in-
dicated that most feeding occurs at night. How-
ever, the DSB III day samples contained a number
of individuals with appreciable amounts of food in
their stomachs, showing that a certain amount of
feeding occurs during the daytime. Donaldson

(1973) found that Sergestes sargassi, S. pectinatus,
and Sergia japonica appeared to feed around the
clock. The first two species also live in Hawaiian
waters; unfortunately, they only occurred in the
night samples of DSB III, so this study could not
test his observations.

If Hawaiian sergestid species have specialized
by dietary preference, they might be expected to
exhibit specialized structures for catching prey.

829

FISHERY BULLETIN: VOL. 74, NO. 4

One important systematic character, the third
maxilliped, appears directly related to feeding.
Many species of Sergestes have greatly enlarged
third maxillipeds, armed with stout spines and
varying in length and development among the
different species. Sergestes pectinatus in par-
ticular has highly modified third maxillipeds, with
a series of short, comblike setae between the
longer spines. The division of Hawaiian sergestids
into a long-maxilliped group and a short-maxil-
liped group would seem logically to indicate a
difference in diet between the two groups.

The results of the DSB III study (Table 5) are
rather unexpected. All the species captured fed
largely on zooplankton-sized Crustacea in the 1- to
3-mm size range, chiefly calanoid copepods,
myodocopid ostracods, and hyperiid amphipods.
Some species also ate smaller zooplankton in the
0.4- to 0.6-mm size range, chiefly larval bivalves,
foraminifera, and cyclopoid copepods. Ability to
utilize prey in the small size range appeared to
depend not on the length of the third maxillipeds
but on the degree of setation of the first three
pairs of pereiopods and (when not enlarged) the
third maxillipeds. Species feeding on small zoo-
plankton all have long setae spaced about 0.3-0.4
mm apart. All well-sampled species in the short-
maxilliped group except Sergestes erectus fed on
the small zooplankton. Within the long-maxilliped
group, there is a gradation in degree of setation of
the pereiopods from S. annatus, which has very
short, sparse setae, through the S. orientalis
group, which have somewhat longer, more numer-
ous, but still rather sparse setae, to S. sargassi and
S. 'pectinatus, which have rather long setae spaced
about 0.5-0.6 mm apart. Of this group only S.
armatus was captured in quantity during DSB
III; its diet definitely lacked small zooplankton. A
few specimens of 5. sargassi and S. pectinatus
were captured; none contained small zooplankton,
but with the small sample size their status remains
in doubt.

The dietary specializations of Hawaiian serges-
tids thus appear more related to size than to type
of prey. The variety of copepods, amphipods, and
ostracods that compose the large zooplankton
fraction all seem to be equally acceptable as prey.
The various modifications of the third maxilliped
may reflect specialized methods of capturing prey
rather than a specialized diet. In particular, the
diet of S. pectinatus lacks any distinctive charac-
teristics which can be associated with its unusual
maxillipeds. While large zooplankters are prob-

ably seized individually, the small zooplankton
appear to be sieved from the water onto the long
setae, spaced so as to retain zooplankton and pass
water, a process akin to filter feeding.

The small zooplankton probably represent a
supplementary rather than a primary resource for
sergestids. Larval bivalves, as meroplankton, are
unlikely to be abundant all year around (they were
abundant in the zooplankton during the December
1973 cruise; I have not examined other zooplankton
samples) and are unlikely to be abundant far from
land. Many of the individuals containing small
zooplankton also contain masses of an unidentified
greenish, fibrous material. Judkins and Fleminger
(1972) reported similar material in Sergestes
similis, and Foxton and Roe (1974) reported
similar material in a number of Atlantic species. If
this material is detritus and not the digested
remains of some unidentified organism, it would
represent another resource available to sieving
sergestids, potentially very important when small
zooplankton is sparse. The inefficient-looking
sieving mechanism of the Hawaiian sergestids are
a reminder that none of these species feed solely,
or even primarily, on the small zooplankton. Any
modifications for increased sieving ability must
not hamper the animal's ability to seize large
zooplankton.

The results from DSB III are quite different
from those reported by Donaldson (1975). He
found a much larger proportion of large prey, such
as euphausiids and fishes, and also many more
chaetognaths. Part of the difference is due to his
much larger sample, where infrequently eaten
prey are more likely to turn up. Large sergestids
captured in very small numbers during the DSB
III cruise, particularly Sergia bisulcata, S. in-
equalis, and perhaps 5. tenuiremis, are likely to
eat larger prey than is reported here. However,
some of the difference between Donaldson's re-
sults and the DSB III results may be due to a
higher degree of feeding in the trawl during
Donaldson's study. The abnormal conditions in the
cod end of a mid-water trawl are apt to lead to
abnormal feeding. Judkins and Fleminger found a
much lower proportion of euphausiids in the
stomachs of sergestids caught by albacore than in
trawl-caught shrimp. They also found fish scales
only in trawl-caught shrimp, an unlikely food item
under natural conditions. These results emphasize
the need for future feeding studies to take what-
ever steps are necessary to minimize or eliminate
feeding in the trawl.

830

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

Reproduction and Growth

It is diffcult to determine when sergestids
spawn. Copulation occurs long before spawning;
female Sergia gardineri whose ovaries have not
started to mature often bear spermatophores.
Eggs are spawned directly into the water rather
than being carried on the appendages as in the
carideans. Some spawning probably occurs year
around, as sexually mature females can be cap-
tured at any time of the year. In Sergesfeff, the
anterior lobes of the ovary vary greatly in size,
filling much of the carapace at maximum develop-
ment. However, there is no correlation of ovary
development with carapace length in adult
females. One possible explanation is that a female
may spawn several batches of eggs over a period of
several months, the ovaries regressing in size
between batches.

Recruitment to a catchable size can be deter-
mined from the quarterly size-frequency histo-
grams. Omori (1969) found that Sergia lucens
required about 2 mo from spawning to recruit-
ment; assuming the time is similar for Hawaiian
species, a maximum in recruitment implies a
maximum in spawning about one quarter earlier.
Most Hawaiian sergestids showed peak recruit-
ment in either the second (April-June) or third
(July-September) quarter. Species with maximum
recruitment in the second quarter included Ser-
gestes annatu^, S. sargassi, Sergia inequalis^, and
S. bisulcafa. Species with maximum recruitment
in the third quarter included Sergefites atlanticus,
S. consobrinus, S. pectinatui^, Sergia scintillans,
and S. gardineri. Some species showed no signif-
icant difference from one quarter to the next;
these included Sergestes orientalis, Sergia bigeni-
mea, and probably S. tenuiremifi and Petalidium
snspiriosiim. Sergestes vigilax had somewhat
higher recruitment in quarters two and three than
during the rest of the year. Sergestes erectus
showed no particular recruitment maximum, but
intermediate-sized shrimp were most abundant in
the fourth quarter. Sergia fulgens is a peculiar
case, to be discussed later.

Size-frequency data indicate that most
Hawaiian sergestids appear to live about 1 yr, in
agreement with most other studies (Pearcy and
Forss 1969; Omori 1969; Donaldson 1973). The
size-frequency histograms of Sergia bisculcata
indicate that this species has a 2-yr life span,
though the conclusion is based on a small sample.
Donaldson concluded that S. robusta may also live

2 yr. Genthe (1969) arrived at a 2-yr life span for
Sergestes similis off California, though Pearcy and
Forss found a 1-yr life span for the same species off
Oregon. Genthe asserted that juveniles less than 5
mm CL are 9 to 11 mo old, which seems too old. His
data support a 1-yr life span if a 2- to 3-mo larval
development time is assumed. Probably only a few
large all-red species live more than a single year.

Sergia fulgens differed from all other Hawaiian
species by showing an extremely modal size-
frequency distribution and varying drastically in
abundance from one month to the next (Figure
18). This behavior can best be explained by as-
suming that S. fulgens is an expatriate species
occasionally moving into Hawaiian waters from
elsewhere. Adult females from the December 1970
cruise and the May 1973 cruise (Teuthis XXI) had
small ovaries with eggs about 150 jum in diameter.
Mature females of the closely related but smaller
species S. scintillans had proportionately larger
ovaries with eggs about 260 jum in diameter. Omari
(1969) reported an average diameter of 255 /xm for
another closely related species, S. lucens. It thus
appears that the large female S. fulgens are not
ripe. While it is possible that female S. fulgens
continue to grow to 18 or 20 mm CL before
spawning and that ripe females have never been
captured, it seems more likely that S. fulgens do
not reproduce in Hawaiian waters and that the
local population is carried in by currents from its
normal breeding range. Unfortunately, the geo-
graphic range of S. fulgens is almost totally
unknown; in addition, it is very similar or identical
to S. talismani in the Atlantic. Influxes of S.
fulgens did not coincide with captures of Sergestes
tantillus, an equatorial species occasionally found
in Hawaiian waters, but little more can be in-
ferred about the source of the local population of S.
fulgens.

Interspecific Relationships

The 20 species of Hawaiian sergestids exhibit a
variety of specializations in morphology and habit
that appear to minimize interspecific competition
and allow them to coexist as a stable assemblage.
Most obvious is the division into half-red and
all-red species, related to shallow and deep day-
time depth ranges and the different concealment
strategies required. The all-red sergestids are
subdivided by size and nighttime vertical dis-
tribution, as are the half-red sergestids, which are
also further subdivided by photophore type and

831

FISHERY BULLETIN: VOL. 74, NO. 4

length of third maxillipeds. Finally, nearly all
species cooccur with at least one other species that
is much more closely related than any of the other
Hawaiian sergestids. Interspecific competition
should be strongest between members of a species
pair; the ways in which two closely related serges-
tids divide up the mid-water environment should
suggest the kinds of competition that occur in the
mid-water environment and how competition is
minimized.

Table 8 shows some observed parameters of
Hawaiian sergestids. The dendrogramlike pattern
at the left is a subjective representation of the
affinities among the species, based on morpho-
logical features. Some differences will be noted in
the vertical distribution patterns of the species
pairs; for example, Sergestes vigilax is more
broadly distributed than S. armatus and tends to
live shallower at night. However, most species
pairs are commonly found together over much of
their vertical ranges. The most striking difference
among closely related species is adult size. In every
case the most closely related species show little or
no overlap in the adult size range. For example,
Sergia scintillans appears nearly identical to S.
fulgens, differing chiefly in the number of photo-
phores on the antennal scale and exopod of the
uropod. However, adult S. scintillans vary from
5.5 to 10.5 mm CL, while adult S. fulgens vary from
11 to 16.5 mm CL. The only exception to this rule,
the species triplet Sergestes orientalis-S. tantil-
lus-S. consobrinus, is a revealing case. Sergestes

orientalis is well separated in size from S. conso-
brinus, the largest females of S. consobrinus
overlapping only slightly with the smallest males
of S. orientalis. However, S. tantillus, while
somewhat smaller in average size than S. orien-
talis, still overlaps considerably in size with the
larger species. In this case it turns out that S.
tantillus is primarily an equatorial species
(Judkins 1972), occurring only rarely in Hawaiian
waters. Mac Arthur (1972) has shown on theoretical
grounds that when three similar species differ in
only one parameter, such as body size, the compe-
tition pressures are strongest on the middle
species. One of the factors determining the north-
ern limit of S. tantillus may be this competition
from both a larger and a smaller species.

Specialization solely by adult size could still
result in competition if adults of the small species
cooccur with similar-sized juveniles of the large
species. In this case, other specializations appear
to become important. When the species have
similar vertical ranges, the juveniles may live
shallower than the adults. For example, adult
Sergia bigemmea and adult 5. inequalis both occur
at about 150 to 225 m at night. Juvenile S. in-
equalis in the 10- to 13-mm CL range, the size of
adult 5. bigemmea, are mostly found between 50
and 150 m, so that similar-sized individuals of the
two species seldom occur together.

Competition could occur if the large species lives
somewhat deeper than the small species, so that
juveniles of the large species live at about the

Table 8.— Characteristics of Hawaiian sergestid species. Dendrogram shows estimated phylogenetic affinities among

species.

Species

Adult size Day depth Night depth Population size

(CL, mm) (m) (m) (no./100 m')

I — Sergestes orientalis
_n — Sergestes tantillus
' Sergestes consobrinus

Sergestes armatus
Sergestes vigilax

Sergestes atlanticus

Sergestes cornutus

Sergestes erectus

Sergestes sargassi

Sergestes pectinatus

J Sergia fulgens

' Sergia scintillans

Sergia gardineri
Sergia bigemmea
Sergia inequalis

I Sergia bisulcata

' Sergia maxima

Sergia tenuiremis

Sergia laminata

Petalidium suspiriosum

5.5-10

500-

625

125

1.32

5.5- 8

450-

650?

100?

<0.10

3.8- 6

450-

725

1.05

9 -14.5

550-

650

150-

300

2.35

6 - 8.5

550-

725

200

0.35

5 - 9

550-

725

300

1.31

3.5- 5

450-

550?

(0-

50)?

<0.10

13 -24.5

550-

800

250-

325

3.81

7 -10.5

450-

575

125-

300

0.70

3.2- 7.5

450-

725

75-

275

1.71

11 -16.5

550-

625?

75-

200?

2.26

5.5-10.5

525-

700

25-

125

3.31

4.5- 9

650-

775

25-

150

8.65

9.5-14.5

750-1

,100?

125-

250

0.64

13.5-22

750-1

,100?

100-

250

0.55

16.5-23

700-

900

225-

350

1.35

(41.5)

<0.10

18.5-29

750-1,300 +

550-1,200 +

0.89

7 -10

700-

800?

<0.10

8.5-12

800-1,300 +

800-1,200 +

1.84

832

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

same depth as adults of the small species. The
clearest example of this type in Hawaiian waters
is Sergestes vigilax and S. armatus, where most
juvenile S. armafus in the 6- to 8.5-mm CL range
live around 100 to 150 m at night, overlapping
somewhat with adult 5. vigilax in the same size
range (most 5. vigilax live above 100 m, but adults
often occur somewhat deeper). However, it ap-
pears that adult 5. vigilax are most abundant from
October to March, while juvenile S. armatu.-^ in the
same size range are most abundant from April to
June (Figures 6, 8). Thus the actual overlap at any
one time is probably small.

The Hawaiian sergestid assemblage can thus be
described by size, morphology, and vertical dis-
tribution. Consider the half-red species first. Those
with short maxillipeds are divided into a pair of
species with lensed cuticular photophores and
three species with organs of Pesta. The pair with
cuticular photophores includes a large species,
Sergiafulgens, and a small species, S. scintillans;
these species may also live at different depths at
night. Of the three species with organs of Pesta,
Sergestes erectus is very large, distantly related to
the other two, and lives deeper at night; 5. atlan-
ticus is larger than S. cornutus and may live
deeper during the daytime. Long third maxillipeds
appear to have evolved at least twice, possibly
three times, in Sergestes (Burkenroad 1937; Foxton
1972). The 5. sargassi-S. pectinatus pair is distinct

from the others; again, 5. sargassi is large and 5.
pectinatus is small, with specialized maxillipeds
and broader vertical distributions day and night.
The other two groups are more closely related, but
the S. armatus-S. vigilax pair has longer maxil-
lipeds than the S. orientalis-S. tantillus-S. conso-
hrinus triad. Sergestes armatus is larger than 5.
vigilax and lives deeper at night; the other group
has been discussed above.

Among the all-red Hawaiian sergestids, a
similar organization prevails. The Sergia in-
equalis-S. bigemmea-S. gardineri group are re-
spectively large, medium-sized, and small; in
addition, S. gardineri lives shallower than the
other two at night, and perhaps during the day-
time. The 5. hisidcata-S. maxima pair is related to
the above triad, but S. hisulcata is somewhat
larger than 5. inequalis and lives deeper at night,
while the rare S. maxima is extremely large. Two
species without photophores are nonmigrators; S.
tenuiremis is much larger than Petalidium suspi-
riosum. The rare 5. laminafa, while related to S.
tenuiremis, is smaller, has photophores (Walters
1975), and appears to migrate.

Studies of sergestid assemblages in the sub-
tropical Atlantic by Foxton (1970) near Fuerte-
ventura (Canary Islands) and Donaldson (1975)
near Bermuda showed interesting parallels to the
present study in the subtropical Pacific (Table 9).
The two Atlantic areas were verv similar to one

Table 9.-Atlantic and Hawaiian sergestid assemblages.

Hawaii

Sergestes consobrinus
Sergestes orientalis \
(Sergestes tantillus) f

Sergestes armatus
Sergestes vigilax

Sergestes atlanticus
(Sergestes cornutus)

Sergestes erectus

Sergestes sargassi
Sergestes pectinatus

Sergia fulgens
Sergia scintillans
Sergia gardineri \
Sergia bigemmea J
Serga inequalis
Sergia bisulcata \
Sergia maxima J

(Sergia laminata)
Sergia tenuiremis

n.e.

Petalidium suspiriosum

Bermuda'

(Sergestes edwardsii)

Sergestes armatus
Sergestes vigilax

Sergestes atlanticus
Sergestes cornutus
"Sergestes corniculum'

Sergestes sargassi
Sergestes pectinatus

(Sergia talismani)
n.e.

Sergia splendens
Sergia robusta

Sergia grandis

(Sergia lillcta)
Sergia tenuiremis

Sergia japonica

'From Donaldson (1975).
2From Foxton (1970).
3n.e. — no equivalent.
<See text.

Fuerteventura'

n.e. 3

n.e.

Sergestes armatus
Sergestes vigilax

(Sergestes atlanticus)

n.e.
"Sergesfes corniculum'

Sergestes sargassi
Sergestes pectinatus

n.e.
n.e.

Sergia splendens
Sergia robusta

n.e.

n.e.
Sergia tenuiremis

Sergia japonica

833

FISHERY BULLETIN: VOL. 74, NO. 4

another except for the abundance of Scrge^tes
aflaiiticKs and Sergio grand is (Sund 1920) in
Bermuda relative to Fuerteventura (the Fuerte-
ventura material all came from a single cruise and
may have lacked some of the less-abundant
species). More surprising, the Atlantic sergestids
were very similar to the Hawaiian species, par-
ticularly the half-red types. "Sergestex cornicidum
Kr^iyer 1855"' replaced its close relative, S. erectits
in the Atlantic, and two rare Bermuda species, S.
edirardsii Kr6yer 1855 and Sergia talismaui
(Barnard 1947), had close relatives, Sergesfes
c())isohri H IIS and Sergia fidgens, in Hawaiian
waters; otherwise, all the half-red species in the
two Atlantic studies also occurred in the present
study. There were some differences in abundance
and vertical distribution, partly real and partly
due to differences in sampling. Sergestes vigilax
was more abundant than 5. armotiis in the Atlan-
tic studies, and S. sargassi was more abundant
than 5. pecfiuotiis; the opposite was true in
Hawaiian waters. S. atlaiificus was more abun-
dant near Bermuda and less abundant near Fuerte-
ventura than near Hawaii. Sergestes coniiciiliou
was more broadly distributed at night than its
Hawaiian counterpart, S. ereetus. The biggest
differences were the rarity or absence in the
Atlantic collections of the S. oriental is types and
the half-red Sergia species, both of which were
abundant in Hawaiian waters. Still, the similar-
ities between the subtropical Atlantic and Pacific
were considerable: one or more large species with
short third maxillipeds and with fairly deep
nighttime distributions, one or two smaller species
with short maxillipeds and living shallower at
night (in Bermuda), and a variety of species with
long maxillipeds occurring in closely related
groups of large and small species.

The all-red sergestids also showed similarities
between the subtropical Atlantic and Pacific,
although the parallelism was not as striking as in
the half-red types. Sergia ten u irem is was found in
all three areas. The role of S. gardineri was filled
in the Atlantic by the closely related S. splendens
(Sund 1920). It was somewhat larger than 5.
gardineri, exceeding 11 mm CL, but had no po-

■'Crosnier and Forest (1973) have reviewed the systematics of
Atlantic species of Yaldwin's "Sergestex coniiciiliini" species
group. They replaced S. coniiciiliini Kreiyer with S. heu^eni
(Ortmann 1893) and three new species-S. parasi'miiiiidux, S.
pediformiff, and S. curratux. Donaldson's figure of 5. coniiciiltnu
corresponded to S. curratuf:. Foxton gave no drawings of S.
corniculinii, but a later study in the same area, Foxton and Roe
(1974) found S. Iienseni and S. ciirvatiig.

tential competition in the 10- to 15-mm CL size
range like the Hawaiian S. higemmea. The nearest
Atlantic equivalents of S. ineqiialis and S. bisul-
eata, respectively S. robusta (Smith 1882) and S.
grandis, lived much deeper at night, in the 400- to
600-m zone, which was nearly devoid of sergestids
around Hawaii. Sergia filicta (Burkenroad 1940)
may be the Atlantic counterpart of S. laminata,
but very little is known about either species.
Sergia japoniea (Bate 1881) had no Hawaiian
equivalent, and S. maxima had no Atlantic
equivalent. Neither Atlantic study mentioned
Petalidium, so it is unclear whether there is an
Atlantic counterpart to the Hawaiian Petalidium
siispiriosum {P.faliaceiim Bate 1881 occurs in the
South Atlantic (Kensley 1971) ). Both oceans thus
contain an all-red assemblage consisting of one or
more nonmigrators, a small, abundant species
with a shallow nighttime range, and several larger
species living deeper at night. In general, the
Hawaiian area appears to have more half-red and
fewer all-red sergestid species than the sub-
tropical Atlantic.

While the parameters of Table 8 indicate that
Hawaiian sergestids have partitioned the mid-
water environment, this study has left unclear the
ecological significance of most of the parameters.
Differences in size, length of third maxilliped, and
nighttime vertical range are presumably related
to diet, but the data on feeding show little dietary
specialization other than the ability of some
species to eat submillimeter-sized zooplankton. A
more elaborate study may reveal more subtle
variations in diet, perhaps related to vertical
distribution of prey or differences in hunting
strategies. Daytime vertical distribution and color
pattern seem most likely related to predation.
Virtually nothing is known about predation on
Hawaiian sergestids. The division of half-red
sergestids into species with organs of Pesta and
species with lensed cuticular photophores has an
unknown ecological significance. Cuticular photo-
phores are fixed in position, but I have observed
sergestids with organs of Pesta rotating them
through nearly 180°, maintaining a vertical orien-
tation of the photophores regardless of the atti-
tude of the animal (see also Omori 1974). Studies of
live sergestids may reveal differences in behavior
between the two groups related to the need for
ventral countershading. Hawaiian sergestids ap-
pear to occupy distinct niches, but the niches
cannot be defined yet in an ecologically meaning-
ful way.

834

WALTERS: ECOLOGY OF HAWAIIAN SERGESTID SHRIMPS

ACKNOWLEDGMENTS

This paper is based on a Ph.D. dissertation
submitted to the Department of Oceanography,
University of Hawaii. I thank my committee,
Richard Young (Chairman), John Caperon,
Thomas Clarke, Jed Hirota, and John Stimpson for
their advice and criticism. Steven Amesbury,
Sherwood Maynard, Fletcher Riggs, and Richard
Spencer also helped in collecting and analyzing
data. David Judkins and Henry Donaldson sup-
plied data from other places; Judkins identified
some of the more obscure sergestid species.
Elizabeth Kampa furnished unpublished light
intensity measurements from north of Oahu.
Dennis Kam supplied contouring subroutines for
the vertical distribution figures and helped debug
my computer programs. This work was supported
in part by NSF grants GB-20993 and GA-33659.

LITERATURE CITED

Amesbury, S. S.

1975. The vertical structure of the micronektonic fish
community off leeward Oahu. Ph.D. Thesis, Univ. Hawaii,
Honolulu, 176 p.
Badcock, J.

1970. The vertical distribution of mesopelagic fishes collect-
ed on the Send Cruise. J. Mar. Biol. Assoc. U.K.
50:1001-1044.
Barham, E. G.

1957. The ecology of sonic scattering layers in the Monterey
Bay area. Stanford Univ., Hopkins Mar. Stn. Tech. Rep.
1, 182 p.

1970. Deep-sea fishes: Lethargy and vertical orientation. In
G. B. Farquar (editor). Proceedings of an International
Symposium on Biological Sound Scattering in the Ocean,
p. 100-118. Maury Cent. Ocean Sci. Rep. 005.

BURKENROAD, M. D.

1937. The Templeton Crocker Expedition. XII. Sergestidae
(Crustacea Decapoda) from the Lower Californian region,
with descriptions of two new species and some remarks on
the organs of Pesta in Sergestes. Zoologica (N.Y.)
22:315-329.
Clarke, G. L.

1968. Light measurements. BioScience 18:965-966.
Clarke, G. L., and E. J. Denton.

1962. Light and animal life. In M. N. Hill (editor), The sea,
Vol. 1, p. 456-468. Interscience, N.Y.

Clarke, T. A.

1973. Some aspects of the ecology of lanternfishes

(Myctophidae) in the Pacific Ocean near Hawaii. Fish.

Bull, U.S. 71:401-434.
Clarke, W. D.

1963. Function of bioluminescence in mesopelagic
organisms. Nature (Lond.) 198:1244-1246.

Crosnier, A., AND J. Forest.

1973. Les crevettes profondes de I'Atlantique oriental
tropical. Faune Tropicale, Tome XIX. ORSTOM (Off.
Rech. Sci. Tech. Outre-Mer), Paris, 409 p.

Denton, E. J., and J. A. C. Nicol.

1965. Refle.xion of light by external surfaces of the herring,
Clupea harengii.'^. J. Mar. Biol. Assoc. U.K. 45:711-738.
Donaldson, H. A.

1973. Ecology of the Genus Sergestes (Decapoda, Natantia)
in an area near Bermuda. Ph.D. Thesis, Univ. Rhode
Island.

1975. Vertical distribution and feeding of sergestid shrimps

(Decapoda: Natantia) collected near Bermuda. Mar. Biol.

(Berl.) 31:37-50.
Doty, M. S., and M. Ogurl

1956. The island mass effect. J. Cons. 22:33-37.
FOXTON, P.

1970. The vertical distribution of pelagic decapods

(Crustacea: Natantia) collected on the Sond Cruise 1965.

II. The Penaeidea and general discussion. J. Mar. Biol.

Assoc. U.K. 50:961-1000.

1972. Further evidence of the taxonomic importance of the
organs of Pesta in the genus Sergestes (Natantia,
Penaeidea). Crustaceana 22:181-189.
Fo.xTON, P., and H. S. J. Roe.

1974. Observations on the nocturnal feeding of some meso-
pelagic decapod crustacea. Mar. Biol. (Berl.) 28:37-49.

Genthe, H. C, Jr.

1969. The reproductive biology of Sergestes similis
(Decapoda, Natantia). Mar. Biol. (Berl.) 2:203-217.

Gilmartin, M., and N. Revelante.

1974. The 'island mass' effect on the phytoplankton and
primary production of the Hawaiian Islands. J. Exp. Mar.
Biol. Ecol. 16:181-204.

Gordon, D. C, Jr.

1970. Chemical and biological observations at station Gol-
lum, an Oceanic Station near Hawaii, January 1969 to
June 1970. Hawaii Inst. Geophys., Tech. Rep. HIG-70-22,
44 p.

Glnderson, K., C. W. Mountain, D. Taylor, R. Ohye, and J.
■Shen.

1972. Some chemical and microbiological observations in the
Pacific Ocean off the Hawaiian Islands. Limnol. Ocean-
ogr. 17:524-531.
Judkins, D. C.

1972. A revision of the decapod crustacean genus Sergestes
(Natantia, Penaeidea) sensu lato, with emphasis on the
systematics and geographical distribution of Neoser-
gestes, new genus. Ph.D. Thesis, Univ. California, San
Diego, 289 p.
Judkins, D. C, and A. Fleminger.

1972. Comparison of foregut contents of Sergestes similis
obtained from net collections and albacore stomachs. Fish.
Bull., U.S. 70:217-223.
Kensley, B. F.

1971. The family Sergestidae in the waters around Southern
Africa (Crustacea, Decapoda, Natantia). Ann. S.-Afr.
Mus. 57:215-264.

MacArthur, R. H.

1972. Geographical ecology: Patterns in the distribution of
species. Harper & Row, Publishers, Inc., N.Y., 269 p.

Marshall, N. B.

1954. Aspects of deep sea biology. Philosophical Library,
N.Y., 380 p.
Maynard, S. D., F. V. Riggs, and J. F. Walters.

1975. Mesopelagic micronekton in Hawaiian waters: Faunal
composition, standing stock, and die! vertical migration.
Fish Bull., U.S. 73:726-736.

835

FISHERY BULLETIN; VOL. 74, NO. 4

Nakamura, E. L.

1967. Abundance and distribution of zooplankton in
Hawaiian waters 1955-56. U.S. Fish Wild!. Serv., Spec.
Sci. Rep. Fish. 544, 37 p.
NicOL, J. A. C.

1958. Observations on luminescence in pelagic animals. J.
Mar. Biol. Assoc. U.K. 37:705-752.

1967. The luminescence of fishes. Symp. Zool. Soc. Lond.
19:27-55.
Omori, M.

1969. The biology of a sergestid shrimp, Sergestes lucens
Hansen. Univ. Tokyo, Bull. Ocean Res. Inst. 4, 83 p.

1974. Thebiology of pelagic shrimps in the ocean. Adv. Mar.
Biol. 12:233-324.
Omori, M., A. Kawamura, and Y. Aizawa.

1972. Sergestex nimilin Hansen, its distribution and impor-
tance as food of fin and sei whales in the North Pacific
Ocean. In A. Takenouti (editor), Biological oceanography
of the northern North Pacific Ocean, p. 373-391. Idemitsu
Shoten, Tokvo.

Pearcy, W. G., and C. a. Forss.

1969. The oceanic shrimp Sergesfes siniilis off the Oregon
coast. Limnol. Oceanogr. 14:755-765.
Tyler, J. E., and R. W. Preisendorfer.

1962. Transmission of energy within the sea. In M. N. Hill
(editor). The sea, Vol. 1, p. 397-451. Interscience, N.Y.

Walters, J. F.

1975. Ecology of Hawaiian sergestid shrimps (Penaeidea:
Sergestidae). Ph.D. Thesis, Univ. Hawaii, 227 p.

Yaldwyn.J.C.

1957. Deep-water Crustacea of the genus Sergestes
(Decapoda, Natantia) from Cook Strait, New Zealand.
Zool. Publ, Victoria Univ., Wellington (N.Z.) 22:1-27.

Ziemann, D. a.

1975. Patterns of vertical distribution, vertical migration,
and reproduction in the Hawaiian mesopelagic shrimp of
the family Oplophoridae. Ph.D. Thesis, Univ. Hawaii,
112 p.

836

DECISION THEORY APPLIED TO THE SIMULATED DATA
ACQUISITION AND MANAGEMENT OF A SALMON FISHERY^

Gary E. Lord^

ABSTRACT

A salmon fishery management model utilizing statistical decision theory has been constructed. The
model provides for the successive acquisition of data that can be used to formulate and maintain an
optimum management strategy. The Bayes risk is defined as the e.xpected economic loss resulting from
a set of fishery management decisions and the criterion of optimality is taken to be the strategy that
minimizes the Bayes risk. Specific functional forms are assumed where necessary in order to obtain a
closed form expression for the Bayes risk. The Bayes risk, in units of numbers of fish, can then be
computed for any particular sequence of fishery management decisions.

This paper represents a continuation of an earlier
effort (Lord 1973) in which statistical decision
theory was applied to the data acquisition and
management of a salmon fishery. The crucial
feature was not that the species considered was
salmon but that the assumed fishery was both
dynamic and subject to errors in the population
estimation. The population is assumed to be sub-
ject to continuing assessment, however, so that as
the season progresses it is possible to make re-
peatedly more refined estimates of the true state
of nature. The management strategy may thus be
modified successively to reflect the additional data
as they become available.

The development was quite abstract and pre-
sented only the basic theory in a relatively general
way. The present paper represents an inter-
mediate situation in which the theory is applied to
a specific model constructed to represent such a
fishery. The principal features of this model are: 1)
a Ricker spawner-return relationship, 2) simulated
sampling for population estimation purposes, and
3) an economic loss function based on maximum
substained yield (MSY).

A limitation of the present model is that it is
constructed in such a manner that a closed analytic
form is obtained without recourse to Monte Carlo
or other approximate methods of analysis. In other
words, the Bayes risk may be computed exactly
upon the specification of well defined sets of

'Contribution No. 456, College of Fisheries, University of
Washington, Seattle, WA 98195.

^Fisheries Research Institute, College of Fisheries, University
of Washington; Present address: Applied Physics Laboratory,
University of Washington, Seattle, WA 98195.

parameters. The imposition of such analytical
requirements constrains the choice of functions to
those that are mathematically tractable. An-
ticipating the final results. Equations (18) and (20),
I feel that about the maximum degree of gen-
erality has been retained consistent with analyt-
ical tractability. It is likely that models possess-
ing a greater degree of fidelity to the actual fishery
situations will require the use of Monte Carlo
methods as Mathews (1966) used in his simulation
of the cannery portion of the Bristol Bay fishery.

ANALYSIS

The notation used in Lord (1973), with only
minor changes, will be retained here. In this
section I will discuss the Bayes risk for a particular
fisheries model based on the Ricker spawner-
return relation. The criterion of optimality will be
taken as MSY. Economic losses will accrue as the
actual management strategies depart from the
optimum. Generally these losses will be reflected
in either a decreased present catch or in dimin-
ished future returns due to prior overfishing.

A loss function proportional to the difference
between the optimum catch and the actual catch,
on an MSY basis, will be assumed. This is a simple
and intuitively reasonable concept but, nonethe-
less, a unique formulation of the loss function from
this criterion is no simple task. The difficulty arises
from the use of a spawner-return relation which
reflects the biological fact that the present state of
the system is necessarily the result of past actions
and, similarly, that future conditions will depend
on present actions. In the case of sockeye salmon,

Manuscript accepted March 1976.
FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

837

FISHERY BULLETIN: VOL. 74, NO. 4

an added complication is provided by the fact that
the run in any year represents the progeny of
several spawning groups.

The Ricker spawner-return relation, for a single
spawning group, is given by

R^ = aE^_,,e

-bE.

(1)

where R,, is the return in year n resulting from an
escapement E„_,, k years prior, e is the base of
natural logarithms, and a and b are parameters
assumed unique for any river system or spawning
group. We may generalize Equation (1) to the case
of multiple spawning groups to give

Rn = ^a.E^.^e-''^'-^ (2)

k - 1

where the relevant spawning occurs over the years
(w - 1) through (w - K). In Equation (2) the
coefficients (o^.) now reflect not only the magni-
tude of the run, as in Equation (1), but the
proportion of the run arising from each spawning
group. Specifically we can write a^. = a/\ where a
is the parameter in Equation (1) and ip), is propor-
tion of the run in year n arising from spawners in

year (n - k). We have the condition X ft = 1

fc = 1

from which it follows that 7^ (h = a.

k = 1

The return as given by Equation (1) or Equation
(2) is a deterministic function of the parameters.
In actual practice, however, the return, from the
biologist's point of view, is a random variable in
which case some additive or multiplicative error
term must be appended to Equation (1). Thus, at
an appropriate point in the analysis, the return
will be assumed to be a random variable whose
expected value is given by Equation (1).

Let Xn be the catch in year n. Then

X„ = a ^ PkE„.

i-e

- bE

E„

k = 1

Let Xtot be the total catch over some fixed but
otherwise arbitrary number, say n*, of fishing
seasons. Then

-^tot —

n' r- K

2 I 2

n = 1 Lk = I

OkK-k^'

(3)

If we attempt to maximize Xy^^ with respect to the
yearly escapements E^, E2,...E„*, it turns out
that as n* ^ oo: a) a steady state solution exists

and b) the optimum steady state escapement, £"0,
is that which maximizes the function iaEe~ ''^ - E).
Let L„ denote the economic loss in any year n and
define L^ as the difference between the optimum
catch, Xopx, and the actual catch J^act. i-e-,

L„ — X,

opt

Xact- From Equation (3) we obtain

L„ = {aEoe- "^u -E^

fc = 1

PkE„-ke

- bE

" £•)

(4)

Eq is fixed and all of the escapements
{E„ _/;}(A; = 1 . . . K) have already occurred thus
leaving only E„ at our disposal. L„ is clearly
minimized by setting E„ = 0 but since this would
eliminate a portion of the run in future years the
subsequent loss would be high indeed. Consider
now the combined loss for two successive years n
and (n + 1). Proceeding along the same lines that
led to Equation (4) we obtain

L„ + L,, ^,= 2{aEoe-'>^.-Eo)

(a 2 p, £•„_,.- "--£'„)

k = 1
K

-(a 2 thK+x-ke-^-'-'-^-E,, ^,).

A: = 1

If this is treated as a function of the single
variable E^,, an optimum value can be obtained.
However, this loss also depends on .£^ + 1 which has
not yet occurred. Let us extend this process
through year (n + K), which is a convenient stop-
ping point since it represents the completion of a
cycle starting at year n. The total loss over this
period is given by

/>«.« + A = iK+ l)(a£'o6'-''^„-£'o)

- 2 a2 PkE^-ke-'^'^-El (5)

) = ,1 *- k = 1 -■

The loss given by Equation (5) depends not only on
past and present escapements but on the future
values E^ +1, E„ + 2, - - • -^i + a as well. Thus, when
formulating a policy for any particular year one
must take into account future policies also. From a
mathematical point of view what we have emerg-
ing here is another dynamic program, i.e., the
optimum year-to-year allocation as well as the
within-year allocation is in the form of a dynamic
program. This is too great an anlytical burden to

838

LORD: DECISION THEORY APPLIED TO SALMON FISHERY

bear. However, we can invoke the "Principle of
Optimality" (Bellman 1957:83) to specify that
En +j = £"() for all j al, i.e., all future escapements
are assumed to be the optimum MSY escapement.
This is a reasonable assumption since the principle
of optimality states that an optimal policy is one
which, given the present state of the system,
establishes and maintains an optimal policy for all
future time periods. Since £"0 represents such an
optimum steady state escapement it follows that
E„ +1 = Eq for future optimality. In this case
Equation (5) takes the form

^n.n + K

= L{E„) = (K+ DiaEoe-'^o-Eo)
- {aE„e~ ^^"- E„) + (terms depending

on Eq and past escapements only). (6)

From Equation (6) it appears that the optimiza-
tion will be over a total of {K + 1) seasons. This is
not actually the case since, as noted above, the
constraint E„ +^ = £"0 has been imposed and the
analytical procedures used in year n will be ap-
plicable in year {n + 1), etc. Note also the intu-
itively reasonable result that the loss given by
Equation (6) is minimized by setting E„ = Eq, the
optimum MSY escapement.

The analysis thus far has assumed that all
quantities are deterministic. Random variables
will now be introduced to simulate the situation
actually existing in salmon fishery assessment and
management. Let N„ denote the run size resulting
from the (known) escapements {£"„ _j){j = 1, . . .
K) and let A',, be a random variable which, for
definiteness, will be assumed to have the two-
parameter gamma density

/3o.Vn

(7)

where F denotes a gamma function. The pa-
rameters (oq, /Sq) are subscripted to denote that
they are applicable prior to the start of the run and
/is subscripted by one to denote that it is applica-
ble to the first fishing period. The quantity ijo is a
symbolic conditioning variable denoting the pre-
season information that is available for the
specification of (oq. I^o)- Anticipating the dynamic
nature of the fishery and its management the
probability density of A^„ will be conditioned
successively to reflect the data obtained after the
start of the run.

We assume now that the expected value, shown
as E[N„] is that given by the Ricker relation, i.e.,

Rn =E[N„] = a 2 PkK-ke-

k - 1

(8)

The variance of A^„ may be estimated from his-
torical data, e.g., smolt outmigrations, high seas
catches, etc. Knowledge of the mean and variance
is sufficient to determine the parameters (a,,, /?o).

At this point it might be well to justify, or at
least explain, the assumption of a gamma density
for A^„ . Clearly, one cannot obtain Equation (7) on
the basis of biological arguments. On the other
hand, a gamma density does not do particular
violence to one's intuition concerning the dis-
tribution of population sizes. In particular. Equa-
tion (7) confines A^„ to positive values with scale
and location specified by («(,, Po)- In salmon
population estimation, it is rare that parameters
beyond mean and variance are available from
whatever source. It is in this spirit that Equation
(7) is introduced. Further, the gamma distribution,
not coincidentally, has the added virtue that it is
an analytically convenient function. Similar ar-
guments will be used to justify some of the func-
tions to be introduced subsequently.

For the remainder of the analysis, only events in
year n will be considered so that the subscript may
be omitted from A^„ . The fishing season is assumed
to consist of m nonoverlapping time periods dur-
ing each of which a management decision, 8, must
be made. Let {8,} (i = 1, . . . m) be an arbitrary
sequence of decisions where each of the 8^ is a
member of some finite set of possible management
decisions.'^ Assume now that during the t'th period
a fraction p, of the total run enters the fishery. The
set {p,){i = 1, . . . m), which is assumed to be
known, may be obtained from such sources as the
almanac prepared by Royce (1965). The (ft) must
obviously satisfy the condition

2 ^' = 1-

i = 1

Corresponding to any actual realization of the
run, N, there exists some unique set of optimum
catch-escapement allocations <t], ) (i = 1, . . . m).
Rothschild and Balsiger (1971) used linear pro-

'A typical set of management decisions consists of such actions
as opening or closing the fishery, the imposition of gear limita-
tions, waiting periods, etc.

839

FISHERY BULLETIN: VOL. 74, NO. 4

gramming to determine an optimum set of such
allocations. Such fine-scale is not practical here so
that the individual tj, are irrelevant here except
that they must satisfy the condition

I = 1

Let{TJj} {i= 1, ... m) be the actual allocations
where each tjj will be assumed to be a random
function of the management decision 5j taken
during the /th period. It will be assumed that the
{tj,} {i = 1, . . .m) have independent beta distribu-
tions where the beta parameters ( t; , ju^ ) are
uniquely determined by the management decision
S, that is taken during period /. Thus we have

giVi^Vz ...n™ 181,82, ...5„) = llgdn,\8,) (9a)

i = 1

where

9ii^i |8i) =

r(»',)r(iu,)

Ml-^)

tf.-l

(9b)

This is a reasonable probability density to assume
since it confines tj; to the interval (0,1) and the
parameter choice permits, within appropriate
limits, the specification of the mean and variance^
of r}i .

Return now to the central feature of the analy-
sis which is to take into account the dynamics of
the fisher}'. Equation (7) is the probability density
of N appropriate for the first period of the fishery
during which only preseason conditioning infor-
mation, denoted symbolically by yo, is available.
Assume now that, during the first and subsequent
time periods, additional population data, z/i, 2/2, .• .
become successively available. This data may then
be used to condition the probability density of N,
hopefully in such a manner that our knowledge of
the true value of A'^, as measured by its variance,
improves as more data are gathered. At each stage
of the fishing season we compute the Bayes risk
with respect to the then current probability den-
sity of N and adopt a strategy that takes into
account all available data and all previous man-
agement decision. This will be formalized analyt-

■•The conditioning of tj by 5 only is probably an
oversimplification. There is evidence to indicate that ^ also

depends on the number of fish that enter the fishery during any
fishing period.

ically upon the specification of an appropriate
sampling distribution for the {?/; ) {i = I, 2, . . .
m - 1). t/^ is irrelevant since it is obtained after the
final decision 8„, will have been made.

Assume that during each stage of the run some
fixed fraction e of the total number of fish entering
the fishery is vulnerable to sampling. For example,
if the sampling is done by gill nets e may be
determined from knowledge of the length, the
time of soak, and the efficiency of the net. With
such a sampling scheme, it is reasonable to assume
that the samples ?/,, 2/2, ... Vk -\ will have in-
dependent Poisson densities with parameters Aj,
^2. • • • K -1 where Aj = tpjN, i.e., ep^N is the
expected sample size for the ith period and

PiYi =yi\N)^e

_ c- f 0 A'

(10)

where y^ = 0, 1, .... Equation (10) and Bayes
theorem may be utilized to modify or update
Equation (7) to reflect the additional information
that is assumed to have become available. Assume
that the system is now at the start of the second
stage and that the sample y^ is now available.
Bayes theorem gives

..xn ^ PiYr = y.\N)fmyo) ....

XP(Y,=y,\N')AiN'\yo)dN'

Substituting Equations (7) and (10) in Equation
(11) gives, after dividing common factors.

fiiNlyo.yi) =

(12)

The integral in the denominator of Equation (12)
is a standard form expressible in terms of gamma
functions which gives

r(ai)

(13)

where oj = ao + ?/i and (i^ = /3q + epj. The updat-
ed probability density for A^ given by Equation
(13) is, like the prior density given by Equation (7),
a gamma density but with modified parameters a^
and /?!. The process by which Equation (13) was
obtained may be repeated indefinitely to give

fkiN\yo.yi

2/fc-i)

_ (^fc-lT'-xrV,-! -^.-.'V

n^k-i)

(14)

840

LORD: DECISION THEORY APPLIED TO SALMON FISHERY

as the posterior density for A^ at the start of the
kth fishing period. The parameters are given by

Ok - 1 = ao + iji + y2 + ■ ■ • +yk - 1 and fik -i =
fio + e(p2 + P2 + . . . + Pit-i)- At this point it is
appropriate to observe that, as time progresses
and additional population data are obtained, the
distribution of A^, as specified by the parameters
afc_i and fik-i> will more and more reflect the
in-season sampling data with a corresponding
decrease in the relevance of the preseason infor-
mation implied by a^ and Pq.

The probability densities given by Equations
(7), (10), and (14) enjoy a peculiar relationship in
which the posterior density of A'^, given by Equa-
tion (14), is from the same family as the prior
density, Equation (7), for the particular sampling
distribution given by Equation (10). Such pairs of
densities are called conjugate pairs (DeGroot
1970:159-166). It is obvious that one cannot, in
general, be so fortunate as to have parameter and
sampling distributions that form a conjugate pair
as in the model assumed' here. However, DeGroot
does outline some somewhat ad hoc procedures for
constructing reasonable posterior probability
densities.

All of the quantities and distributions necessary
to compute the average or expected loss, i.e., the
Bayes risk, are now available. The expression for
the Bayes risk, to be evaluated at the start of the
A:th fishing period, may be written formally as

Rki^u ^2, •••5m 1^0,2/1. ■■■yk -l)

= J^fkiN\yo,yu ■ • •yk-i)dH/^dri, . .

1 m

Jd-nmUEJ n ^.(^. 1^)

(15)

i = 1

where Qi , L(£^), and^. are given by Equations (9),
(6), and (14) respectively. Notice that the Bayes
risk as given by Equation (15) is a function not
only of the decisions already made, 6i, 82, ...
6;, _ 1 and the decision about to be made, 6;^ , but of
all future decisions 6^+1, ... 8^ as well. This
dependence on all decisions, past, present, and
future, reflects the assumption that the loss is a
function primarily of the final state of the system,
i.e., to a first approximation one cannot ascribe
values to individual units of escapement during
the season but only to the final total escapement.
This presents no particular analytical difficulties
since any particular sequence of optimum future
decisions S^ +1, . . . 5„ is certainly subject to revi-

sion as time passes and additional information
becomes available.

Substituting Equations (6), (9), and (14) in
Equation (15) gives

Rk{8i,82, ...8„ 12/0,2/1, •• •2/fc-i)

L- (Eo.E,. E..„)-iali,e-i"<-E.)

^ = ' r(.jr(/i,)

■n;

«", - 1

(l-7j)«ri

(16)

where L'{Eq, E„ _i, E,^_2, ■ ■ ■ K-k) denotes that
portion of the loss function that does not depend
on E„. Thus L' is a fixed quantity and may be
removed from the integral signs. This leaves only
probability densities, which must integrate out to
unity, so that

Rki8i,82,..-8m\ 2/0,2/1,... 2/k-l)

(/Sfc_i)^-.

1 («fc - 1)

where the escapement £„ has been expressed as

En = N^ pfl, .
i = 1

The integrations in Equation (17) cannot be
performed as expressed. If the order of the inte-
grations is reversed, the integration with respect
to A^ may be performed but the remaining in-
tegrations over fji, 7J2, ... t)„ will be virtually
impossible. However, if the exponential term

exp (-6A^2 P,^,) is expanded in its Maclaurin

series and if the resulting multinomials of the

/ III \n

form -^I-bN^ p,Vij {n = 0, 1, . . .) are ex-
panded according to the multinomial theorem, the
integrand in Equation (17) will be in a completely
factored form. As a result of this factorization, the
integrals take the form of various moments about

841

FISHERY BULLETIN: VOL. 74, NO. 4

the origin. These integrals are all standard forms
(c.f., Bierens de Haan 1939). The reader will be
spared the details of this reduction and the ensu-
ing integrations. The final expression for the
Bayes risk is

Rk (5i,52 5„ \yQ,y^ Vk -\)

= L'iE^,E,_„...E„-K) +

tti

■k -1

Pi";

fik - 1 j^i J'i + HAj
. n

b \ r(afc _ 1 + n - 1)

fik - 1 n=o \ ^k-ij n\ r(afc _ i)

2fc, = n
m

m
\k1IC2 ■ ■ . fCfn/i = 1

'i + th + ki

fr r{v, +11,) n., -hA^) ^k,

r=\ r(.,)r(., +11, + k,) '

(18)

where ( kik2 ... k„,) denotes a multinomial
coefl[icient.

A slightly different form for the risk may be
obtained under an alternate set of assumptions.
Considerable emphasis has heretofore been placed
on the conjugacy of the gamma-Poisson families
of distributions. The gamma-Poisson assumption
is a reasonable one and the resulting conjugacy
lends a certain elegance. However, this line of
analysis results in posterior gamma parameters
(«fc, Pk) that, among other things, depend on the
run fractions (ft) {i = 1, . . . k). This parameter
dependence on the run fractions virtually pre-
cludes treating the set (ft) (i = 1, . . . k) as any-
thing but fixed quantities; i.e., once a variable
becomes the argument of a gamma function one
has usually arrived at an analytical dead end. In
actual practice, however, the quantities (p; ) (i = 1,
. . . m.) are random variables since there may be
considerable year-to-year variation in the time
profile of the run. Such temporal variation may be
of considerable importance in Bristol Bay because
of the large magnitude of the run and its short
duration.

It has been suggested (0. A. Mathisen, pers.
commun. and others) that the probability density
of A^ is most appropriately conditional upon the
catch-per-unit-effort (CPUE) observed during the
course of the run. In so doing one can remove the
explicit dependence of (a^ , P^ ) on (ft ) {i = 1, . . . k).
An implicit dependence remains, however, since
the CPUE will be a function of the run fractions.
One can formally bypass this dependence, how-
ever, by relating the density of A^ directly to the

842

CPUE. In so doing one can then introduce tempo-
ral variability in the set (ft) (i = 1, . . . m) and in
evaluating the Bayes risk an additional expecta-
tion with respect to the density of these random
variables must be taken.

An almost ideal probability density to describe
the run fractions is the Dirichlet density defined

,, . r(Yi + Y2 + ... + Ym)

/l(Pl,P2, . ..pm) =

r(Yi)r(Y2) . . . r(Y„;

ip ^.

y - 1

(19)

where ft^O for all i. As written this density is
singular since the variates must satisfy the side

condition V Pj = 1. The choice of the pa-

1 = 1
rameters (yi, Y2. .•• Ym ) then permits the
specification of any m of the means, variances, and
covariances of the (ft) (i = 1, . . . m). If Equation
(19) is substituted in Equation (16) the integra-
tions with respect to A'' and (tji, tjo, . . . tj,„ ) may be
done as before. The remaining integrals over (pi,
P2) ••• Pm) are all Dirichlet integrals (Wilks
1962:177, et seq.) for which the values are readily
determined. The resulting Bayes risk for this case
may then be shown to be given by

/?,(6,.5o. ...5„, ICPUE)

= L'iEo-E,, _ 1. . . . E„ _a) + -^ V — '-^—

a ^ / b Y r(a, _, + n + l
k-o\kik2.. ./c,J

fj r(f, +^•)^(., +iu,)r(Y, + k)

i = 1

r(.,)r(., -Hrt +^)r(Y,)

y (Y, + k,){p, + k,)

f, + Ph + k,

) = 1

(20)

where G = X

i = 1

Equations (18) and (20) are somewhat in-
timidating, particularly if one were to attempt to
infer the qualitative behavior of the system as the
parameters descriptive of the fishery and its
management are varied. Indeed, Equations (18)
and (20) are virtually useless for this purpose with
the exception of the determination of certain

LORD: DECISION THEORY APPLIED TO SALMON FISHERY

limiting behavior as the appropriate parameters
assume their extreme values. However, Equations
(18) and (20) do have the virtue that, in closed
form, the most crucial features of the fishery
dynamics and statistics are accommodated in a
quantitative and, hopefully, reasonably accurate
fashion.

A NUMERICAL EXAMPLE

The foregoing mathematical model was applied
to the simulated management of the Wood River
system of Bristol Bay. It should be emphasized at
the outset, however, that the assumptions, meth-
ods, and results presented here should in no way be
construed as representing a management scheme
preferable to those currently in use. The Wood
River was chosen simply because, based on Math-
ews' (1966) data, it seemed to follow the Ricker
spawner-return curve reasonably well.

In the example considered here, the model was
limited to a fishing season of five time periods
during each of which a choice of two management
decisions was possible. This limitation was neces-
sary to avoid inordinately lengthy calculations.
Ricker parameter values of a = 4.077 and b =
0.8 X 10~^, which were used by Mathews, were
used here. The return was assumed to consist of
only the progeny of a single spawning group K
years prior where K is arbitrary, i.e..

Pi =

1 i = K
0 i jt K

All prior escapements were assumed to be the
optimum escapement £"0 so that the loss function
given by Equation (5) becomes

A = {aEoe- *^o -Eo)- (aE,, c ''^„ - £" ) .

For the above values of the Ricker parameters,
the MSY escapement is given by £"0 = 709,000.
The expected value and standard deviation of a
r(«o. Po) variate are given by ao//So and aQ^Z/^Q,
respectively. In terms of the Ricker parameters,
the expected run size is given by aEo exp(- bEo)
which determines the ratio oq/Pq = 1.64 x 10*". An
initial (i.e., preseason) standard deviation of
one-half the expected run size was assumed. In
terms of the gamma parameters this gives
ao'V^o= ao/2fio or oq = 4.0 and /?o = 2.44 X 10^.
The two management strategies assumed were
complete closure (option 2) and one level of open-

ing (option 1). In terms of the beta parameters,
closure is simulated merely by setting /X2 = 0 with
an arbitrary positive value for ^2. During fishery
opening it was assumed that an average of 80% of
the available fish are caught with a standard
deviationof 0.25. This gives (;' J, fij) = (0.312,1.248)
as the appropriate beta parameters. The set of run
fractions {ft) (1 = 1, . . . 5) was determined from
the time profile proposed by Royce (1965). Values
of 0.156, 0.282, 0.348, 0.160, and 0.054, using five
equal length time intervals, were obtained. No
attempt was made to treat the run fractions as
random variables. All of the parameter values
were chosen to reflect reasonably well the known
behavior of the system.

The fishery dynamics were treated by two
distinct methods. The first method utilized the
gamma prior density for A'^ with a Poisson sam-
pling density thus, through conjugacy, giving a
gamma posterior density. A gamma posterior
distribution was also assumed in the second
method but the posterior gamma parameters were
back-calculated after introducing prescribed
stage-to-stage trends in the population mean and
standard deviation.

The Bayes risk at each stage was computed for
each of the 2^ = 32 total possible sequences of
decisions, past, present, and future; i.e., no at-
tempt was made to formulate and solve the func-
tional equation associated with dynamic pro-
gramming.^ While relatively unsophisticated, this
approach does permit one to use hindsight to
determine, ex post facto, what an optimum
previous strategy would have been, given the
information currently available. In real life, of
course, "what might have been" is irrelevant in the
management of a dynamic system-one must
optimize the system as it exists in real time in
accordance with the principal of optimality, the
relevant homily for which might well be "what's
past is prologue."

The numerical results are summarized in Tables
1 to 3. Tables 1 and 2 give the optimum strategies
and corresponding minimum Bayes risks for a
gamma prior run size distribution with simulated

^Subsequent to the submission of this paper, C. J. Walters
(1975) published a paper in which the ideas of dynamic pro-
gramming were applied to the optimum year to year man-
agement of a salmon fishery. His work is of considerable interest,
particularly since he managed to impose the principle of op-
timality and carry out the backward recursive scheme proposed
by Bellman (1957). It remains to be seen if this method can be
applied to the decision theoretic model presented here, but I am
no longer as pessimistic as I formerly was.

843

FISHERY BULLETIN: VOL. 74, NO. 4

Table 1. -Optimum strategies and minimum Bayes risks for a five-period, two-decision fisiiery with

a sampling fraction c = 1 x 10"^.

Time period (j)

Run fraction (p, )

0.156

0.282

0.348

0.160

0.054

Poisson parameter (X,)

256

462

570

262

—

yi=\

Simulated samples (y,)

256

462

570

262

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

3.49 X 10*

3.33 X

lO"

3.29 X 10"

3.28 X 10*

y. =2\

Simulated samples {y,)

512

924

1,140

564

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

1.89 X 105

1.70 X

105

1.90 X 105

y, ='^A,

Simulated samples (y,)

128

231

285

131

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

6.90 X 10'

3.35 X

103

2.48 X 105

2.29 X 10'

Table 2.-0ptimum strategies and minimum Bayes risks for a five-period, two-decision fishery with

a sampling fraction « = 1 x 1(H.

Time period (0

Run fraction (p,)

0.156

0.282

0.348

0.160

0.054

Poisson parameter (A,)

—

y,=\

Simulated samples (y,)

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

5.86 X 10*

4.55 X

10*

4.19 X

10*

4.11 X 10*

y. =2\

Simulated samples (i/i)

114

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

1.86 X 105

1.87 X

105

1.89 X

105

1.88 X 105

y. ='^\

Simulated samples (y,)

—

Optimum strategy

open

Minimum Bayes risk

1.80 X

105

4.14 X 10*

1.85 X

10*

1.12X

10*

9.65 X 10J

Table 3.-0ptimum strategies and minimum Bayes risks for a five-period, two-decision fishery with
linear stage-to-stage trends in the expected run size and the run size standard deviation with
preseason parameters oq = 4.0 and Pq = 2.44 x 10"^.

Time period (0

Run fraction (p, )

0.156

0.282

0.348

0.160

0.054

Constant expected run size:

«, - 1 / A - 1

1.64 X 104
8.20 X 105

1.64 X 104
6.89 X 105

1.64 X 104
5.57 X 105

1.64 X 104
4.26 X 105

1.64 X 10*

V«, - . / ft - .

2.95 X 105

Optimum strategy

open

Minimum Bayes risk

1.80 X 105

1.41 X 105

1.07 X 105

7.87 X 10*

5.74 X 10*

Increasing expected run size;

«. - 1 / A - 1

1.64 X 104

1.97 X 10^

2.30 X 104

2.63 X 104

2.95 X 104

V«. - , / ft - 1

8.20 X 105

6.89 X 105

5.57 X 105

4.26 X 105

2.95 X 105

Optimum strategy

open

Minimum Bayes risk

1.80 X 105

1.40 X 105

1.26 X 105

1.43 X 105

1.90 X 105

Decreasing expected run size

«. -1/ A -.

1.64 X 104

1.48 X 104

1.31 X 104

1.15 X 104

9.84 X 105

v'«, - 1 / A - 1

8.20 X 105

6.89 X 105

5.57x105

4.26 X 105

2.95 X 105

Optimum strategy

open

Minimum Bayes risk

1.80 X 105

1.52 X 105

1.25 X 1X)5

9.57 X 10*

6.68 X 10*

Poisson sampling. The sampling was intended to
simulate actual run sizes equal to, greater than, or
less than the preseason estimate of the run size,
ao/fio. The Poisson sampling was done by brute
force in which sample values exactly equal to the
desired expected values were chosen. For example,
to simulate an actual run size twice that based on
the preseason parameters we choose ijj = 2Xi

844

where X, = ep^ uq/Pq is the Poisson parameter for
the ?th period obtained from the preseason pa-
rameters. The deterministic samples (which is
really a contradiction in terms) permit one to elicit
the response of the system to specified input
stimuli.

The Bayes risks are all in units of numbers of
fish. The optimum strategy is that strategy which

LORD: DECISION THEORY APPLIED TO SALMON FISHERY

minimizes the Bayes risk given that all prior
decisions were optimum for the time periods in
which they were made. In other words, the
"hindsight" feature was not utilized to "improve"
a past decision— once made any decision is retained
through all subsequent stages.

The mathematical machinery developed gen-
erally gives intuitively reasonable results.
Specifically, the tendency toward larger or smaller
run sizes results in optimum strategies that tend
successively toward more or fewer open periods
respectively. The Bayes risk generally, but not
always, decreases as the season progresses, largely
reflecting the decreasing variances in the es-
timates of the run size. Increases in the Bayes risk
can usually be attributed to past decisions that, in
the light of subsequent sampling, are no longer
optimum thus requiring corrective action.

CONCLUSIONS

The mathematical models assumed and
developed here for the objective management of a
typical salmon fishery, as previously noted, are
based on quite specific functional forms and thus
represent somewhat of an idealized situation.
However, these functions were chosen to reflect
the behavior of the system insofar as the knowl-
edge of such behavior is available. Indeed, the
acquisition of such detailed knowledge is an im-
portant area of current research and subsequent
refinements of the statistics will be possible as
more data are gathered.

Of more concern than the accuracy of the fine-
scale mathematical behavior of the system is the
appropriateness of the basic mathematical theory
upon which the models are built. I feel that statis-
tical decision theory is a most natural framework
on which to base an objective management model.
The nomenclature lends support to this view. For
example, the equivalence of a management deci-
sion and a statistical decision is obvious.^ The term
risk, in the economic if not the strict Bayesian
sense, is frequently used in discussions of fishery
management. Finally, Bayes theorem provides a
convenient and theoretically appropriate method
for accommodating the combined data acquisition
and dynamics of the fishery.

®This equivalence is not always evident even within decision
theory itself. For example, it requires a slight mental contortion
to treat statistical estimation as an application of decision theory
as the statisticians have done.

Advantage has been taken of some powerful
analytical tools to characterize salmon fishery
management. However, any enthusiasm for these
quite contemporary methods should be tempered
somewhat by consideration of some of the specific
practical difl^culties likely to be encountered. One
of these, mentioned in Lord (1973), is the difl^culty
associated with multistage dynamic processes.
While the fishery management problem under
discussion falls very naturally into a class of
stochastic dynamic programs it is not yet obvious
whether the functional equation arising from the
imposition of the principal of optimality can be
formulated or solved in a useful fashion. The
calculations done here were more of the brute
force variety in which all strategy combinations,
optimal or not, were considered. In other words,
the backward recurrence scheme central to dy-
namic programming was not used to reduce the
total number of possible strategies to be con-
sidered. In so doing, the "Curse of Dimensionali-
ty" about which Bellman (1957:6) so aptly warned,
proved to be a limiting condition. To evaluate
completely the five-stage, two-decision fishery
considered here required from 10 to 15 min of
Control Data Corporation^ 6400 central processor
time for each set of input parameters. This is not a
trivial numerical effort and should give one pause
when considering more elaborate models.

In conclusion I feel that advantage should be
taken of the appropriate analytical tools as they
are made available by the mathematicians or, at
the very least, such tools should be investigated.
However, the availability of such methods in no
way indicates their eventual practicality for any
specific problem. For this careful additional in-
vestigation is necessary.

LITERATURE CITED

Bellman, R.

1957. Dynamic programming. Princeton Univ. Press, 340 p.

BlERENS DE HaAN, D.

1939. Nouvelles tables d'integrules definies. G. E. Stechert
& Co., 716 p.
DeGroot, M. H.

1970. Optimal statistical decisions. McGraw-Hill Book Co.,
Inc., 489 p.
Lord, G. E.

1973. Characterization of the optimum data acquisition and
management of a salmon fishery as a stochastic dynamic
program. Fish. Bull., U.S. 71:1029-1037.

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

845

FISHERY BULLETIN: VOL. 74, NO. 4

Mathews, S. B. Royce, W. F.

1966. The economic consequences of forecasting sockeye 1965. Almanac of Bristol Bay sockeye salmon. Univ. Wash,

salmon {Oncorhynchus nerka, Walbaum) runs to Bristol Fish. Res. Inst. Circ. 235, 48 p.

Bay, Alaska: A computer simulation study of the potential Walters, C. J.

benefits to a salmon canning industry from accurate 1975. Optimal harvest strategies for salmon in relation to

forecasts of the runs. Ph.D. Thesis, Univ. Washington, environmental variability and uncertain production

Seattle, 238 p. parameters. J. Fish. Res. Board Can. 32:1777-1784.

Rothschild, B. J., and J. W. Balsiger. Wilks, S. S.

1971. A linear-programming solution to salmon 1962. Mathematical statistics. John Wiley & Sons, Inc.,

management. Fish. Bull., U.S. 69:117-140. N.Y., 644 p.

846

DIEL CHANGES IN SWIM BLADDER INFLATION OF THE
LARVAE OF THE NORTHERN ANCHOVY, ENGRAULIS MORDAX

John R. Hunter and Carol Sanchez'

ABSTRACT

Laboratory and field studies demonstrated that larval anchovy 10 mm standard length and larger
inflate their swim bladders each night and deflate them in the day. Maximum night levels of inflation
were attained 2 h after the onset of dark and typical day levels occurred about 2 h after the onset of
light. Laboratory experiments indicated that larvae fill their bladders at night by swallowing air at the
water surface and the vertical distribution of sea-caught larvae suggested that they migrate to the
surface each night to fill their swim bladders. Gas is released by passing bubbles through the pneumatic
duct into the alimentary canal. The diel rhythm of inflation was viewed as an energy sparing
mechanism. Measurements of sinking speed of larvae with and without inflated bladders suggested
that the energy saved at night by inflation of the swim bladder would exceed the cost of vertical
migration to the surface and that the migratory range over which energy savings are possible would be
greater as larvae increased in length.

Northern anchovy, Engraulis mordax Girard, are
more vulnerable to starvation in the larval stage
than at any other time of life, consequently,
energy sparing mechanisms may be critical to
their survival. In a recent paper Uotani (1973)
showed that the larvae of several clupeoid fishes,
Engraulis japonicus (Houttuyn), Sardinops
melanosticta (Temminck and Schlegel), and
Etrumeus teres (DeKay) have expanded swim
bladders when captured at night in the sea and
deflated ones when captured during the day.
Energy conservation is certainly one of the possi-
ble adaptive advantages of such behavior, but the
energy saved must be evaluated in terms of the
energy cost of daily filling the bladder. This
requires that the mechanism of filling be known.
The object of the present study was to determine
if the larvae of the northern anchovy display a
similar rhythm and to evaluate this behavior as a
possible energy sparing mechanism.

The swim bladder in adult northern anchovy is a
tubular vesicle that extends the length of the body
cavity. It is connected to the alimentary canal by a
pneumatic duct which originates from the dorsal
wall of the cardiac stomach; no anal duct exists as
it does in some clupeoids (O'Connell 1955). Two
tubules on each side of the body extend from the
anterior end of the bladder into the cranium where
they expand into two pairs of capsules, termed

'Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

prootic and pterotic bullae (O'Connell 1955). The
swim bladder of the larva is basically similar to
that of the adult. At the time of initial filling of the
swim bladder, the pneumatic duct is functional
and the bullae become filled with gas. No his-
tological evidence exists for gas secretion in adult
E. mordax nor for the larvae (O'Connell 1955, and
pers. commun.).

The swim bladder is deflated by passing gas
bubbles through the pneumatic duct into the
alimentary canal and out the anus. On a number of
occasions we have observed this process while
examining a live anchovy larva under a dissection
microscope. We have also captured larvae with
deflated swim bladders that had gas bubbles in the
alimentary canal.

METHODS

Fertilized anchovy eggs were obtained from a
captive population of adults maintained in
spawning condition in the laboratory (Leong 1971)
and the larvae were reared using the techniques,
foods, and tanks described by Hunter (1976). The
larvae were reared at temperatures of 16.5° ±
0.2°C and 16.9° ± 0.9°C. A 12-h photoperiod was
used without a dawn or dusk transition in light
intensity. Incident light at the surface was about
2,000 Ix in the day and at night no light was
provided in the closed room which contained the
rearing tanks.

Larvae reared in the laboratory were sampled at

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

847

FISHERY BULLETIN: VOL. 74, NO. 4

various times of day commencing at age 1 day to
determine if a daily rliytiim of inflation existed
and to determine the larval length at which the
swim bladder was inflated. Samples of preserved
specimens from California Cooperative Oceanic
Fisheries Investigations (CalCOFI) ichthyoplank-
ton collections were also examined to determine if
differences existed in swim bladder inflation in
sea-caught specimens.

The standard length was measured to the near-
est 0.1 mm and the maximum width and length of
the swim bladder to the nearest 0.02 mm. The
volume of the swim bladder was calculated by
using the equation for a prolate spheroid,
V = i/STrab'-, where a is half the maximum blad-
der length and b is half the maximum width. For
larvae 16 mm and larger, the calculated swim
bladder volume may be converted to actual gas
volume by multiplying it by the coefficient 0.82
(Figure 1). This conversion is based on data ob-
tained while measuring the composition of swim
bladder gas. The larvae used in that experiment
were larger (mean length 15.6 to 29.6 mm) than
most of the larvae in the rest of the experiments.
For this reason we have used the calculated swim
bladder volume in all computations.

We also sampled larvae reared in the laboratory
to determine the eff"ect of swim bladder develop-

3 -

=J
_l

o
>

<
O

I-

<
q:

/'^ •

Y = 0 82 X

/•

•/

1 1 1 1

1 1

1 1 1

2 -

CALCULATED VOLUME (mm^)

Figure 1.- Relation between the volume of the swim bladder
calculated from the equation V = 4/3 :r ab- and the actual volume
of gas extracted from the swim bladder for northern anchovy
larvae of mean length 15.6 to 29.6 mm. Each point is the mean
volume of the swim bladder calculated for a sample of two to
eight larvae taken at night and the average volume of gas
extracted from that sample. Sample means were weighted by
their variances to calculate the regression line; intercept for line
did not differ from 0; and the standard error of line was 0.0428.

ment and swim bladder inflation on sinking rate.
The method of Blaxter and Ehrlich (1974) was used
to measure sinking rates of larvae. Larvae an-
esthetized in MS 222^ were measured and added to
a 1-liter graduated cylinder without contact with
the air. The larvae were allowed to sink a few
centimeters, then the rate of descent was timed
with a stopwatch for a distance of 7 to 35 cm. Only
one measurement was made per larva and larvae
were reexamined after the test to determine if
they were still alive (dead larvae sank faster than
live ones) and if any gas had been lost from their
bladders. Fresh seawater was used in the
graduated cylinder for each day's run and the
specific gravity and temperature of the seawater
were measured before each larva was tested. The
specific gravity averaged 1.0262 and ranged from
1.0259 to 1.0266. The graduated cylinder was
immersed in a temperature-controlled water bath
which was maintained within 1°C of the rearing
temperature. One rearing group was tested at
15.9° ± 0.2°C and another at 18.0° ± 0.1°C. In the
Results section we have combined the data from
these two rearing groups because covariance
analysis indicated that the diflferences in sinking
speed when adjusted for swim bladder volume and
larval length were not significant.

To determine if anchovy larvae filled the swim
bladder by gulping air at the water surface, the
following experiment was performed. Commenc-
ing 4 h after the onset of dark, larvae in a 400-liter
rearing tank were sampled and the lengths and
dimensions of the swim bladder of each larva in
the sample measured. Just before the onset of dark
on the following day, the surface of the tank was
sealed with a 0.5-cm layer of mineral oil. A second
sample was taken commencing at 2400 h, 4 h after
the onset of dark and ending just before the
beginning of light at 0800. A third sampling was
taken of lar\-ae in the sealed tank during the day
beginning at 1000 h, 2 h after the onset of light,
and ending at 1400.

The gas content of the swim bladders of labora-
tory-reared larvae captured in the dark was an-
alyzed using the micro gasometric method and
apparatus described by Scholander et al. (1955).
Swim bladders were dissected from the larvae in
acid citrate solution, removed with a Pasteur
pipette, and injected into an acid citrate filled
capillary tube sealed at one end. After two to eight

-Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

848

HUNTER and SANCHEZ: CHANGES IN SWIM BLADDER INFLATION

swim bladders had been collected, they were
macerated in the tube and the gas withdrawn with
a Pasteur pipette and inserted into the syringe gas
analyzer (Scholander et al. 1955). In the analyzer,
the carbon dioxide was absorbed with alkaline
citrate, oxygen by pyrogallol, and the volume of
gas determined before and after each treatment.
The remaining gas was considered to be nitrogen.
The volume of gas was read under a dissecting
microscope using an optical micrometer. Reading
error was about ± 0.09 ii\ or from 1 to 2% depend-
ing on the volume of the sample.

RESULTS
Diel Rhythm in Swim Bladder Inflation

The volume of the swim bladder of larvae cap-
tured at night in the sea was greater than that of
larvae collected in the day (Table 1). Similarly, the
volume of the swim bladder of larvae reared in the
laboratory was greater at night than in the day. To
illustrate these daily changes for laboratory-
reared larvae, the mean volume of the swim
bladder for 2-h intervals was calculated for each of
three length classes (10.0 to 11.9 mm, A^ = 121; 12.0
to 13.9 mm, N = 202; 14.0 to 15.9 mm,N = 129). No
evidence existed for anticipation of the onset of
dark at 2200 h nor for the onset of light at 1000 h
(Figure 2). In all three length classes the mean
volume did not return to the daytime level until
about 2 h after the onset of light nor did they reach
the maximum at night until about 2 h after the
onset of dark.

The swim bladder of larvae at night was
frequently so inflated that it constricted the gut
(see fig. 8 in Uotani 1973). Larvae in the dark with
filled swim bladders were motionless or slowly
sinking. The body was oriented head down at an
oblique angle to the water surface. After sinking a
short distance, the larvae swam back to the water

Table 1.— Swim bladder volume (mm^) of preserved northern
anchovy larvae from standard CalCOFI oblique plankton tows
taken at night and in the day in southern California inshore
waters.

0.06 1-

DAY

NIGHT

DAY

I0.0-Il.9nnm

ight samples

Day samples

Length class
(mm)

Swim bladder vol
(mean ± 2 SE)

11.0-11.9
12.0-12.9
13.0-13.9
14.0-14.9

23
20
24
14

0.044 ± 0.007
0.073 ±0.015
0.124 ±0.011
0.128 ±0.011

0.018 ± 0.008
0.015 ± 0.004
0.030 ±0.016
0.029 ± 0.003

II II I 111 I I I I yr'\"\ I I I I I I I I I

19 21 23 I 3 5 7 9 II 13 15 17
TIME OF DAY (hours)

Figure 2.-Mean swim bladder volume ± 2 SE of mean for 2-h
class intervals plotted at midpoint of each 2-h class. Data shown
for three length classes of laboratory-reared northern anchovy
lar\-ae; the onset of dark was at 2200 h and onset of light at 1000 h.
No transitional level of illumination existed between night and
day.

surface, a behavior closely resembling that of
yolk-sac larvae (Hunter 1972).

Specimens with obviously inflated swim blad-
ders occurred occasionally in day samples from the
sea and laboratory but these were only a few
percent of the larvae examined if the first 2 h after
the onset of light are excluded. On the other hand,
the occurrence of larvae with deflated bladders at
night was more common. About 10% of the wild
larvae and 20% of the laboratory-reared (12.0 to
12.9 mm) larvae had swim bladder volumes at
night comparable to those in the day (Figure 3).
The proportion of larvae with deflated bladders at
night decreased with larval length.

849

FISHERY BULLETIN: VOL. 74, NO. 4

llJ
<
>

u.
o

o
ir

UJ
Q.

I—

WILD LARVAE

_ r--

1 C] DAY

0 NIGHT

20
10

'-M

- — -I

^..f^

\V///////A

y//////A.^

^^^/,V^y(//?/////^y

Y////{/////^

70
60
50
40
30
20
iOl-

" 1

LABORATORY- REARED
LARVAE

0.005 0.025 0.045 0.065 0.085 0.105
SWIM BLADDER VOLUME (mm^)

Figure 3. -Percent of northern anchovy larvae. 11.9 to 12.0 mm,
at night (solid bars) and in day (dashed bars) having swim
bladder volumes in various 0.01-mm' classes; numbers on
abscissa are midpoints of swim bladder volume classes. Upper
panel, larvae from CalCOFI ichthyoplankton collections
(preserved specimen length), A^ = 20 for night, and .V = 30 for
day. Lower panel, larvae reared in laboratory (live specimen
length), .V = 49 for night, and .V = 29 for day. Data from 2-h
after onset of dark and 2-h after onset of light were excluded in
laboratory-reared larvae.

The mean swim bladder volume at night was
greater for wild than for laboratory-reared larvae
of the same length. The effect of preservation on
larval length for larvae of this size is not known
but a shrinkage of about lOSt in length in the
Formalin-preserved ichthyoplankton specimens
would account for this difference. The effect of
preservation on swim bladder volume is also
unknown. In some of the preserved specimens, we
noticed the bladder was filled with fluid but we did
not routinely make an examination of the bladder
contents.

Swim Bladder Inflation and Larval Length

The swim bladder was fully formed when larvae
reached 8 to 9 mm but it usually was not inflated.
To determine the larval size at which nightly
inflation commenced, night and day samples from
the laboratory were grouped into 1-mm length
classes (9.0 to 9.9 mm, 10.0 to 10.9 mm, etc.), and the
mean volume for day and night samples for each

850

class calculated, and compared using the t test. The
first 2 h after the onset of dark and the onset of
light were excluded from the classes.

Some of the 9.0 to 9.9 mm larvae appeared to
have inflated swim bladders at night but the
night-day difference in swim bladder volume was
not significant (0.2>P>0.1). Mean volumes for day
and night samples were different in larvae 10.0 to
10.9 mm as were those for larvae in all succeeding
length classes (P<0.001). Thus, the threshold larval
length for nightly inflation of the swim bladder
occurred at about 10 mm, the point at which the
means for day and night volumes diverge (Figure
4).-^ From this point, mean volume of night sam-
ples increased exponentially with length whereas
that for day samples increased linearly.

Relation Between Sinking Speed,
Swim Bladder Volume, and Larval Length

We observed that larvae with inflated bladders
sank more slowly than those with uninflated

•'Swim bladder inflation is reported to occur at 7 mm in E.
japoiiiciif! (Uotani 1973). Comparison of his illustrations to those
of Uchida et al. (1958) suggests Uotani's reported lengths are in
error and that E. Japan iciis also inflates the bladder at about 10
mm.

020r

o
>

Q
Q
<

0 18
016
0 14
0 12
010
008
006
004
002

000

O DAY
• NIGHT

^ * - °

I? °

i , I I I

~\ — I — I — \ — I — r

0 8 9 10 II

"~T \ 1 1 1 1 1 1 1 — I \ 1

12 13 14 15 16 17

LENGTH (mm)

Figure 4.- Mean swim bladder volume ± 2 SE for laboratory-
reared northern anchovy larvae for 1-mm classes of length
plotted at the midpoint of each class. Solid circles are night (first
2-h after onset of dark omitted) and open circles day (first 2-h
after onset of light omitted).

HUNTER and SANCHEZ: CHANGES IN SWIM BLADDER INFLATION

bladders. At night, larvae were occasionally neu-
trally buoyant but most were slightly negatively
buoyant.

To develop an equation for expressing sinking
speed in terms of larval length and swim bladder
volume, the data on sinking speeds were grouped
into four classes of larval length: 10.0 to 11.9 mm,
N = 30; 12.0 to 13.9 mm, A^ = 41; 14.0 to 15.9 mm,
A'' = 54; and 16.0 to 17.9 mm, N = 14. A regression
of sinking speed on swim bladder volume for each
length class yielded the following slopes and
standard errors for the regression lines: -3.040,
SE = 2.339; -4.001, SE = 1.297; -4.8796,
SE = 0.616; and -5.070, SE = 1.680, respectively.
Covariance analysis of these data indicated that
the slopes were not different whereas the inter-
cepts for the regression lines were statistically
different (P = 0.01). Since no difference existed in
the slopes among the four groups, the common
slope from the covariance analysis, -4.769,
SE = 0.487, was used to express the relation
between sinking rate and swim bladder volume for
each length class (Figure 5, lower panel). When
adjusted for the common slope, the sinking rate
intercepts of the four regression lines showed a
precise linear relationship when plotted against
the midpoints of their respective length classes
(Figure 5, upper panel). The equation for the
intercept-length relationship was y = O.ISL - 1.51
where L is larval length (the midpoints of the
larval length classes) and y is the intercept for the
regression of sinking rate on swim bladder volume
(the sinking rate at F = 0 in Figure 5). This
equation was combined with the common slope to
provide the equation given below:

S = 0.18L- 1.51- 4.77 F

where S = sinking speed in centimeters per sec-
ond
L = larval length in millimeters
V = swim bladder volume (outside
dimensions) in cubic millimeters.

We examined the changes in sinking speed of
larvae from the time of hatching through the
development of the swim bladder. These changes
are of interest because they illustrate the timing
of swim bladder development, its effect on
buoyancy, and the advantage of a nightly inflation
cycle. Data for sinking rates for larvae 4.0 to 9.9
mm were grouped into 1-mm classes and the
means plotted at the midpoints of the class inter-

002 004 006 008 010 012 014
SWIM BLADDER VOLUME (V) mm^

016

Figure 5.- The relation in larval northern anchovy between
sinking speed (S), swim bladder volume ( V), and larval length (L).
Lower panel, regression lines show relation between sinking
speed and swim bladder volume for the four classes of larva!
length indicated in the figure, when a common slope of -4.769 is
used (see text). Upper panel, the regression of the y intercepts (S
at F = 0) of the four regression lines on larval length (midpoints
of the four length classes); equation for intercept line was
y = 0.18L - L5L Final equation is S = 0.18L - L51 - 4.777.

vals except for the yolk-sac larvae (3.7 mm) which
were all about the same length. For larvae 10.0 mm
or larger, we calculated sinking speeds from the
mean swim bladder volume given in Figure 4
using the equation given in the preceding
paragraph.

Sinking speed increased exponentially with
length, when larvae sampled at night are excluded
(Figure 6). The increase is roughly proportional to
the cube of the length (curved line in Figure 6).
This might be expected since sinking speed is
dependent upon buoyancy which varies with the
volume (L^) and the difference in specific gravity
between the fish and its medium. For estimating
mean sinking speed for larvae with swim bladders
in the day, or for those without swim bladders the
equation S = 0.094 -i- 0.000264^^ where L is
length in millimeters and S is sinking speed in
centimeters per second, gives a good fit to the data.

The length threshold for filling the swim bladder
(about 10 mm) coincides with a rapid acceleration

851

FISHERY BULLETIN: VOL. 74, NO. 4

SINKING SPEED
• BEFORE INFLATION

/°

(rt

o AFTER INFLATION (DAY)

A AFTER INFLATION (NIGHT)

O 08

« 06

i 04

' ' /

(/I

\y»

J — — ; 1 1 1 1 1

1 —

5,000

- 4,000

Table 2. -Mean swim bladder volume and mean length of
northern anchovy larvae in sealed and unsealed containers.

1,000

0 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17
LENGTH (mm)

FiGfRE 6. -Change in sinking speed from hatching through
threshold of the initial inflation of swim bladder (arrow) of
northern anchovy larvae. For larvae <10 mm, mean sinking speed
± 2 SE were plotted against the midpoint of 1-mm length
classes, except for the first point which was for yolk-sac larvae
and is plotted at average length for the class. For larvae >10 mm,
sinking rates are estimated from the mean swim bladder volume
given in Figure 3 using the general equation for sinking speed
given in Figure 5. Open circles are estimates for the day and solid
triangles for the night. Line is the cube of the length (U) plotted
against length.

in the volume of the larva and its sinking speed.
Thus, the timing of the swim bladder inflation
may be related to these events.

The effect of night inflation of the swim bladder
is also illustrated in Figure 6. For larvae 12 mm
and larger, the average sinking speed at night
appears to be relatively constant at about 0.6 £ 0.1
cm/s (22 m/h) while the sinking speed in the day
increased exponentially with length. A larva 16.5
mm long had a sinking speed at night nearly half
that of the day. Swimming speeds of larval an-
chovy while searching for food in the day, range
from about 0.6 to 1.0 body length/s (Hunter in
press). If a larva did not inflate the swim bladder
at night, the swimming required just to maintain
a position in the water would be equivalent to that
used in the search for food in the day. Since larvae
do not feed at night, filling the swim bladder would
clearly be advantageous as an energy conserving
mechanism.

Mechanism of Swim Bladder Inflation

It was not possible to determine if larvae in the
tank sealed with mineral oil swam just below the
layer of oil or into it because our view was from
above rather than from the side. However, the
mean swim bladder volume for larvae sampled at
night in the sealed tank was less than that for
larvae sampled on the previous night when the

Mean length

of larvae

ean swim bladder vol

Treatment

(mm ±2 SE)

(mm3 ± 2 SE)

Unsealeij tank:

Night

13.9 ±0.37

0.094 ± 0.037

Sealed tank:

Night

14.1 ± 0.41

0.035 ± 0.008

Day

13.8 ±0.86

0.026 ±0.019

tank was not sealed {\ test, P = 0.001, Table 2).
The mean volumes of the swim bladders for larvae
in the day and at night in the sealed tank were not
different. This experiment suggests that anchovy
larvae in the laboratory fill their swim bladders by
swallowing air at the surface.

An analysis of the oxygen content of the swim
bladder could suggest whether or not the gas in the
swim bladder was secreted or taken from the air.
Newly secreted gas would be expected to be
oxygen (Wittenberg 1958), but if larvae were
swallowing air the concentration should be about
lY7c oxygen. Our analysis did not agree with either
pattern even though some measurements were
made 30 min after the onset of dark. Samples
averaged about 11% oxygen, consistently less than
the atmospheric concentration (Table 3). Carbon
dioxide levels (0.9 to Ifl'Jc) were higher than
atmospheric levels but little can be concluded
because our experimental reading error was 1 to
2% owing to the small volumes used. It is probable
that oxygen was lower than atmospheric concen-
tration because it was absorbed from the bladder
by the larva. Except for the first two observations
in Table 3, oxygen concentration tended to
decrease with time from the onset of dark. It
should be noted that preferential removal of
oxygen from swim bladder gases is not unique to
anchovy larvae but is found in most fishes which
have been studied (Wittenberg 1958).

The rate at which the swim bladder was filled
also suggests that the filling is accomplished by
gulping air. Larvae with filled swim bladders were
captured 20 to 30 min after the onset of dark and
the means were at a maximum by 2 h after dark.
Uotani (1973) reported ioxE. japonic us that filling
was completed by 1 h in the sea. Fishes that fill the
swim bladder by secretion require much more time
to fill the bladder, for example, Stenotomus ver-
sicolor (Mitchill) requires 10 to 12 h; Anguilla
rostrata (LeSueur), 12 to 24 h; Opsanus tan
(Linnaeus), Prionotus carolinus (Linnaeus), and
P. evolans (Linnaeus), 24 h; and Tautoga onitis

852

HUNTER and SANCHEZ: CHANGES IN SWIM BLADDER INFLATION

Table 3- Percent composition of swim bladder gas of laboratory-reared northern
anchovy larvae sampled at night, listed in order of time of sampling.

Elapsed time
after onset
of darkness

Number
of

Mean
larval
length

Composition of
swim bladder gas (%)

Sample
volume

Oxygen

tank

(h) (min)

larvae

(mm)

CO,

(mD

(ml/liter)

0 30

27.1

1.2

9.6

89.2

5.54

5.4

0 30

29.6

1.6

8.2

90.2

8.12

0 50

22.1

1.3

14.2

84.5

4.99

5.7

3 15

21.0

1.8

13.1

85.0

6.08

5.0

4 20

25.3

2.2

11.1

86.7

7.18

5.1

6 25

19.3

0.9

12.7

86.4

4.85

4.6

6 25

18.9

0.6

12.6

86.9

4.05

5.0

7 35

22.1

1.4

9.1

89.5

6.06

4.9

8 10

20.1

0.6

9.5

89.9

3.70

4.8

Linnaeus, about 24 h (Wittenberg 1958). Con-
sidering the evidence presented here, and the
apparent lack of gas secretion in clupeoid fishes in
general (Brawn 1962), the most tenable hypothesis
is that swim bladder inflation is accomplished in
larval anchovy by taking in air at the water
surface.

Vertical Migration

If anchovy larvae fill their swim bladders each
night by swallowing air, they must either remain
near the surface throughout the day and night or
migrate to the surface at dusk.

We reexamined the original data collected by
Ahlstrom (1959) to determine if any evidence
existed for vertical movements in northern an-
chovy larvae. Ahlstrom (1959) made extensive
horizontal tows for fish larvae with opening and
closing nets and presented the average number of
larvae of all lengths at various depths. We sepa-
rated his original length data into two length
classes: larvae <11.75 mm (preserved standard
length) and larvae 211.75 mm for night and day
collections; we omitted those collections occurring
near dawn and dusk. Unfortunately, only 14 larvae
sll.75 mm were taken in the day while 279 were
taken at night but the depth pattern in the day
collections was relatively consistent. Larvae
< 11.75 mm were more abundant: N = 6,456, night;
and A^ = 331, day.

At night, over 50% of the larvae £11.75 mm were
taken in the upper 10 m whereas in the day the
upper 10 m contained less than 10% of the larger
larvae (Figure 7). About 50% of the larvae <11.75
mm occurred in the upper 10 m, but no obvious
difference between day and night samples existed.
These results are in general agreement with those
of Ida (1972) who studied the vertical distribution
of the Japanese anchovy, E. japonicus, a closely

30 40 50
DEPTH (m)

60 70

Figure 7.-The vertical distribution in the sea of northern larvae
at night and in the day for two length classes; length <11.5 mm,
upper panel, and length 211.5 mm, lower panel (lengths for
preserved specimens). Numbers of larvae taken at each depth
were cumulated starting at the shallowest tow (5 m) and
expressed as the cumulated percent of the total larvae taken.
Data are from Ahlstrom (1959).

related species that also has a diel rhythm in swim
bladder inflation (Uotani 1973). Ida (1972) found a
striking diel change in the vertical distribution of
E. japonicus with the maximum numbers occur-
ring at the surface at night and at 20 to 30 m
during the day with the movement to the surface
occurring at twilight. Examination of the size
frequency histograms from some of the collections

853

FISHERY BULLETIN: VOL. 74, NO. 4

led Ida to conclude that the diurnal change was
caused by the vertical migration of the larger
larvae (10 to 15 mm).

The diel vertical movements that appear in
larval anchovy at the time of swim bladder
inflation probably persist into adult life. The
adults, however, are quite variable in their
behavior which changes with size of school and
season (Mais 1974). Vertical migration is most
noticeable in large schools which are deep during
the day (119 to 220 m) and rise to the surface and
disappear as sonar targets at dusk. These schools
reform and descend at first light in the morning
(Mais 1974).

Possible Adaptive Advantages

Inflation of the swim bladder reduces the energy
required for maintaining a position in the water
column. This reduction in sinking speed could
represent an important energy savings for larval
anchovy because they do not feed at night and
swimming can not be used in the search for food.
The major energy cost of a diel rhythm of swim
bladder inflation is the required vertical migration
to the surface. Laboratory work suggests that
anchovy larvae, by modification of swimming
speed and direction of turning, are able to find and
remain in area of high food density (Hunter and
Thomas 1974). Thus, it is possible that a larva could
follow an upward and downward movement of
food at dusk and dawn. In this case the added cost
for vertical movements would be slight since the
energy spent in swimming could be used in
searching for food. It is unlikely, however, that
this condition could always be met. Thus, the
energy saved at night by inflation of the swim
bladder should exceed that used in vertical migra-
tion. Assuming the energy used per centimeter
swum is the same for vertical migration as for
maintaining a position in the water at night, the
energy used in a round trip vertical migration of
100 m would be equivalent to that used to maintain
a position for 10 h at night when the sinking speed
was 0.28 cm/s. Thus, the difference between day
and night sinking speeds would have to exceed 0.28
cm/s before a 100-m round trip could be considered
an energy sparing mechanism. The difference in
sinking rates exceeds 0.28 cm/s for larvae 13.5 mm
and larger (Figure 6). This difference increases
with larval length suggesting that the vertical
range of migration over which energy savings are

854

possible increases with length. In addition, the
difference between day and night sinking speeds
may be underestimated because sinking speeds
were measured at the surface. If larvae descend
during the day the gases in the swim bladder
would be compressed, increasing body density and
thereby increasing the sinking speed for larvae in
the day.

These calculations are, of course, a great
oversimplification, but they do illustrate that the
energy saved by inflation of the swim bladder at
night could exceed the cost of a vertical migration
and that the possible range of migration could be
greater for larger larvae.

The energy costs of maintaining a position in
the water column for fish with and without swim
bladders have been calculated by Alexander
(1972). His calculations are not appropriate for
anchovy larvae at night because he considered fish
without a bladder to be continuously swimming
and gaining lift from the pectoral fins. The
behavior of an anchovy at night that failed to
inflate the swim bladder would probably resemble
one with an inflated bladder. It would sink motion-
less at an oblique angle to the water surface and
interrupt sinking by bursts of near vertical swim-
ming. To maintain a position, these bursts of
swimming would have to be of longer duration or
of greater frequency than if the swim bladder
were filled.

In addition to an energy sparing mechanism, a
nightly pattern of swim bladder inflation could
possibly reduce predation. Some predators of
larval fishes, for example chaetognaths and
medusae, use the movement or turbulence
produced by prey for detection and attack
(Horridge 1966; Newbury 1972). Thus, the reduc-
tion of activity produced by slower sinking speeds
could reduce predation. The vertical migration of
the larvae could also result in exposure to different
and possibly less hazardous predators at night. It
would also serve to aggregate larvae, thus facili-
tating social contacts necessary for the develop-
ment of schooling which begins at about 15 mm.

ACKNOWLEDGMENTS

Harold Dorr and Sharon Hendrix assisted in the
laboratory work. James Zweifel provided statis-
tical advice and Reuben Lasker and Paul Smith
reviewed the manuscript. E. H. Ahlstrom allowed
us to present original data on vertical distribution
of anchovy larvae.

HUNTER and SANCHEZ: CHANGES IN SWIM BLADDER INFLATION

LITERATURE CITED

Ahlstrom, E. H.

1959. Vertical distribution of pelagic fish eggs and larvae off
California and Baja California. U.S. Fish Wildl. Serv.,
Fish. Bull. 60:107-146.
Alexander, R. McN.

1972. The energetics of vertical migration by fishes. Symp.
Soc. E.\T). Biol. 26:273-294.
BlAXTER, J. H. S., AND K. F. Ehrlich.

1974. Changes in behaviour during starvation of herring
and plaice larvae. In J. H. S. Blaxter (editor), The early
life history of fish, p. 575-588. Springer-Verlag, Berl.
Brawn, V. M.

1962. Physical properties and hydrostatic function of the
swim-bladder of herring {Clupea harengus L.) J. Fish.
Res. Board Can. 19:635-656.
HORRIDGE, G. A.

1966. Some recently discovered underwater vibration
receptors in invertebrates. In H. Barnes (editor). Some
contemporary studies in marine science, p. 345-405.
George Allen & Unwin Ltd., Lond.
Hunter, J. R.

1972. Swimming and feeding behavior of larval anchovy,

Engraulis mordax. Fish. Bull., U.S. 70:821-838.
1976. Culture and growth of northern anchovy, Engraulis

mordax, larvae. Fish. Bull., U.S. 74:81-88.
In press. Behavior and survival of northern anchovy En-
graulis mordax larvae. Symposium on Fishery Science
Cienca Marinas, Spec. Suppl.
Hunter, J. R., and G. L. Thomas.

1974. Effect of prey distribution and density on the search-
ing and feeding behaviour of larval anchovy Engraulis

mordax Girard. In J. H. S. Blaxter (editor), The early life
history of fish, d. 559-574. Springer-Verlag, N.Y.

Ida, H.

1972. Some ecological aspects of larval fishes in waters off
central Japan. Bull. Jap. Soc. Sci. Fish. 38:981-994.

Leong, R.

1971. Induced spawning of the northern anchovy, Engraulis
mordax Girard. Fish. Bull., U.S. 69:357-360.

Mais, K. F.

1974. Pelagic fish surveys in the California Current, Calif.
Dep. Fish Game, Fish Bull. 162, 79 p.
Newbury, T. K.

1972. Vibration perception by chaetognaths. Nature
(Lond.) 236:459-460.

O'Connell, C. p.

1955. The gas bladder and its relation to the inner ear in
Sardinops caerulea and Engraulis mordax. U.S. Fish
Wildl. Serv., Fish. Bull. 56:505-533.
SCHOLANDER, P. F., L. Van Dam, C. L. Claff, and J. W.
Kanwisher.
1955. Micro gasometric determination of dissolved oxygen
and nitrogen. Biol. Bull. (Woods Hole) 109:328-334.
UcHiDA, K,, S. Imai, S. Mito, S. Fujita, M. Ueno, Y. Shojima, T.
Senta, M. Tahuku, and Y. Dotu.

1958. Studies on the eggs, larvae and juveniles of Japanese
fishes. [In Jap.] Kyushu Univ., Fac. Agric, Fish. Dep., 2d
Lab. Fish. Biol., Ser. 1, 89 p.
Uotani, I.

1973. Diurnal changes of gas bladder and behavior of
postlarval anchovy and other related species. [In Jap.,
Engl, abstr.] Bull. Jap. Soc. Sci. Fish. 39:867-876.

Wittenberg, J. B.

1958. The secretion of inert gas into the swim-bladder of
fish. J. Gen. Physiol. 41:783-804.

855

CERIANTHARIA, ZOANTHIDEA, CORALLIMORPHARIA, AND

ACTINIARIA FROM THE CONTINENTAL SHELF AND SLOPE

OFF THE EASTERN COAST OF THE UNITED STATES

Bernt Widersten'

ABSTRACT

Specimens were examined from 95 stations located between lat. 37°49'N, long. 75°25'W and lat.
44°4rN, long. 66°14'W and from depths between 9 and 366 m. The material was collected by the Bureau
of Commercial Fisheries, Biological Laboratory, Woods Hole, Mass., in the years 1955-68. The collection,
which will be deposited in the Northeast Fisheries Center, National Marine Fisheries Service, Woods
Hole, comprises two ceriantharian species, Cerianthus borealis and Ceriantkeopsis americanus; one
zoanthid species, Epizoanthus incrustatus; one species of Corallimorpharia, Corynactis delawarei n.
sp., and 19 species of Actiniaria, Edwardsia sulcata, Halcampa duodecimcirrata, Haloclava producta,
Peachia parasitica, Bolcera tuediae, Tealia crassicornis, Actinostola callosa, Stomphia coccinea,
Paranthus rapiformis, Antkoloba perdix, Metridium senile fimbriatum, Haliplanella luciae,
Sagartiogeton verrilii, Hormathia nodosa, Actinauge verrilli, Phelliactis americana n. sp., Am-
phianthus nitidus, Stephenauge nexilis, and Stephenauge (?) spongicola.

The following description of the anthozoan species
from the western North Atlantic is based on
material collected by the Bureau of Commercial
Fisheries, Biological Laboratory, Woods Hole,
Mass., during 1955-68. The collection will be de-
posited in the Northeast Fisheries Center, Na-
tional Marine Fisheries Service, Woods Hole.

Besides the morphological descriptions of
different species, much importance has been at-
tributed to the cnidom of the studied specimens.
The sizes of the nematocyst capsules mentioned in
the description refer to unexploded capsules.

While the fixation and preserving of the mate-
rial in Formalin'-' and alcohol had only slightly
affected the sizes of the nematocysts, the measure-
ments of the column, tentacles, pedal disc, and
other organs are, naturally, not directly compara-
ble with those in living specimens.

The terminology used in this paper follows that
by Stephenson (1935) and Carlgren (1949). The
nomenclature of the nematocysts is the classical
one, founded by Weill (1934) and amplified by
Carlgren (1940a, 1945, 1949).

The sectioned mterial was stained with Heiden-
hain's azan or iron hematoxylin-eosin.

All nematocyst measurements are given in
microns.

'Institute of Zoology, Uppsala University, Uppsala, Sweden.
-Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

A list of the stations with names of the species
collected at each station and with ADP (automatic
data processing) codes for latitude and longitude;
time, day, month, year, and number of collection;
vessel; cruise; station number; gear; water depth;
water temperature; and sediment type is on file at
the Northeast Fisheries Center, Woods Hole.

DESCRIPTIONS

Ceriantharia
Cerianthidae

Cerianthus borealis yerriW 1873

OCCURRENCE.-40°10'N, 71°00'W, 146 m, silty
sand, 1 specimen; 41°00'N, 70°48'W, ? m, 1 spec-
imen; 41°50'N, 67°56'W, 51 m, sand, 2 specimens;
42°41'N, 70°05'W, 114 m, gravel, 2 specimens.

GENERAL CHARACTERISTICS.-The spec-
imens were strongly damaged in their proximal
parts. The morphology of the distal part of the
body as well as the composition of the cnidom and
the sizes of the nematocysts were typical of the
species (cf. Carlgren 1940a). While the specimens
from the two southernmost localities were young
(diameter of the distal part of the body 4-8 mm),
the other individuals were older, the largest of
them being equipped with 150 labial as well as
marginal tentacles.

857

FISHERY BULLETIN: VOL. 74, NO. 4

NEMATOCYSTS (those of the southernmost
specimen within parentheses).— Margina/ tenta-
cles: microbasic 6-mastigophors 21 x 3.2-3.8 - 26 x
3.8, 36 X 4.4 - 41 X 4.9, 54.5 x 9.2 - 61 x 7.1 (25.3 x
3.8 - 28.3 X 4.4); atrichs 33 x 4.4 - 36 x 5.5 (9.8 x
1.6 - 27.5 X 2.7); spirocysts 20 x 3.3 - 56 x 6.5 (22
X 3.8 - 28 X 3.8-4.9). Labial tentacles: microbasic
6-mastigophors (axial filament short and thin, less
than half the length of the capsule) 37 x 3.8-4.4 -
45 X 5.5 (ca. 18.5 x 2.7); microbasic b-
mastigophors (axial filament = about half the
length of the capsule 20.7-35.4 x 3.8 (19.1 x 3.3 -

30.5 X 4.9); microbasic 6-mastigophors (axial
filament tall and coarse; more than half the length
of the capsule) 57 x 8.3 - 63.7 x 7.1 (23.4-30 x
4.9-6); atrichs (not common) 55 x 9.8 - 63.7 x 28.3
(ca. 18 X 3.3); spirocysts 18 x 3.3 - 54 x 6
(14.7-16.8 X 2.7-3.3) jum. Holotrichs were very rare
(in the distal part of the column = 22.3 x 14.7 -

43.6 X 17.4, and in the telocraspedon = 61 x 13.6
ium).

Ceriantheopsis americanus (Verrill 1864)

OCCURRENCE. -42°04'N, 67°30'W, 40 m,
gravelly sand, 1 specimen; 42°25'N, 70°56'W, 13 m,
1 specimen.

GENERAL CHARACTERISTICS.-The prox-
imal parts of the two studied specimens were
missing. The individuals were young; the only
specimen, being preserved with a 12-mm-long
column part, had a diameter of 6 mm. The mar-
ginal tentacles were equipped with stout basal
parts and acute apices. The labial tentacles were
about 70 (69 in one specimen).

NEMATOCYSTS.-Cohmr? (distal part): micro-
basic 6-mastigophors (not common) 19.6 x 3.8 -

32.7 X 5.5; atrichs (very common) 26 x 6.5 - 50 x
10.9-16.4. Marginal tentacles: microbasic 6-
mastigophors 16.3 x 3.8-4.4 - 19 x 4.9; atrichs (?)
ca. 12 X 2.7, 8.7-12.5 x 4.9; spirocysts 12 X 2.7 -
27.3 X 4.9. Labial tentacles: microbasic 6-
mastigophors 16.3 x 3.8 - 32.7 x 6.5; spirocysts
13.6 X 2.7 - 27.3 x 4.9. Actinopharynx: microbasic
6-mastigophors (axial filament more than half the
length of the capsule) 21 x 4.9 - 32 x 6; microbasic
6-mastigophors (axial filament less than half the
length of the capsule) 13.6 x 2.7 - 18 x 3.3, 20.7 x
3.8 - 23 x 4.4. Filaments (orthocraspedon):

microbasic 6-mastigophors 19-21 x 3.3, 26 x 5.4 -
33 X 6.5; spirocysts (very rare) ca. 21 x 3.8 jixm.

Zoanthidea
Epizoanthidae

Epizoanthus incrustatus Diiben and Koren 1847

OCCURRENCE.-40°03'N, 71°16'W, 183 m, 4
specimens; 42°10'N, 65°37'W, 238 m, two colonies
with 10 and 19 specimens respectively, and one
solitary specimen, on a shell fragment.

GENERAL CHARACTERISTICS.-The color
of the column and the coenenchyme is greyish
brown; both are strongly encrusted with sand
grains. The polyps were in the contracted state
about 5 mm tall, with the column diameter about 4
mm. Most of the 17 capitular ridges as well as the
insertions of the 36 mesenteries were indistinct
(because it is heavily encrusted with sand). The
tentacles numbered about 36.

NEMATOCYSTS.-Co??m7?.- holotrichs 22 x 7 -
24 X 8.2; spirocysts 22 x 4.4 - 31 x 5.4. Tentacles:
microbasic p-mastigophors 22-33 x 3.3(35 x 6);
microbasic 6(?)-mastigophors 22-23.4 x 4.4; holo-
trichs 22-24 X 7.6-8.2 (common), 34 x 15.3 - 40 x
17.4 (not common); spirocysts (very common) 10 x
3.8 - 32 X 4.9. Actinopharynx: microbasic p-
mastigophors (not common) ca. 22 x 6; holotrichs
21-25 X 7.6, 38 x 14.2 - 41 x 14.7. Filaments
microbasic p-mastigophors 20 x 5.4 - 28 x 6.5
microbasic 6(?)-mastigophors 11 x 4.9 - 21 x 6
holotrichs 23 x 7.6 - 26 x 8.3 /xm.

Corallimorpharia
Corallixnorphidae

Corynactis delawarei n. sp.

HOLOTYPE.— Deposited as a series of sections
in the Zoological Institute, Uppsala. Syntypes
deposited in the U.S. National Museum, catalog
number USNM 54322. Thirty-two specimens ag-
gregated on a tube fragment of an onuphid poly-
chaete, collected by the vessel Delatvare from the
type-locality on 14 June 1962, with a 1-m Nat-
uralist dredge, in station number 9.

TYPE-LOCALITY.-39°56'N, 69°45'W, 201 m,

858

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

sandy bottom, on a tube fragment of an onuphid
polychaete.

DIAGNOSIS.-Column rather firm, smooth, 15
mm tall, bright red (sometimes whitish) to reddish
brown in color. Tentacles rather short, with well-
limited acrospheres; two or three per endocoele,
the .total number being 90. Sphincter long, ento-
dermal to ento-mesogloeal. Maximum number of
mesenteries 60. At least six pairs of mesenteries
perfect. Retractors diffuse. Cnidom: column —
holotrichs and spirocysts; tentacles (acrospheres)-
holotrichs and microbasic b- and jo-mastigophors;
tentacles (peduncles)- microbasic h- and jo-masti-
gophors, atrichs (?), and spirocysts; actinopharynx-
holotrichs, atrichs, microbasic p-mastigophors, and
spirocysts; ^A7a me/? ^s— holotrichs, microbasic p-
mastigophors, atrichs (?), and spirocysts.

GENERAL CHARACTERISTICS.-The col-
umn is smooth and rather firm. In the contracted
state there are distally a number of transverse
as well as a few longitudinal furrows. The shape of
the column is proximally dependent on the shape
of the substrate. The color of the column and the
pedal disc is bright red to reddish brown. (There
are also, however, some whitish individuals in the
collection, with red mesenterial insertions shim-
mering through the ectoderm.) The longitudinal
muscle sheet of the ectoderm forms a thin, but
distinct layer in the column. The tentacles are
rather short, with cylindrical peduncles and well-
limited, white acrospheres. The entodermal as well
as the ectodermal muscle sheets of the peduncles
are well developed. The inner tentacles are shorter
than the outer ones; the exocoelic tentacles are the
largest. The stichodactyline arrangement of the
tentacles is rather indistinct in the often strongly
contracted specimens of the collection. There are,
however, two or three tentacles per endocoele, the
total number being 90. The sphincter is long,
entodermal to entomesogloeal (Figure lA), and
is only occasionally capable of covering all the
tentacles. The actinopharynx is short, in the
contracted state, with longitudinal as well as
transverse folds. There is only one indistinct
siphonoglyph. The retractors of the maximum 60
mesenteries are diffuse, forming an insignificant
sheet over the edge of the mesentery (Figure IB).
At least six pairs (including the directive pair) of
the mesenteries are perfect. Reproduction is
probably asexual by longitudinal fission. The size

Figure l.-Corynactis delawarei n. sp. A. Section through the
sphincter region of the column. B. Cross section of a perfect
mesentery, ent-entoderm, mgl— mesogloea.

of the normally cylindrical column is of a max-
imum 15 mm, with a proximal diameter of 8 mm.

NEMATOCYSTS.-Column: holotrichs 38 x 8.7
- 53.4 X 10.9; spirocysts 18.5 x 3.3 - 27 x 4.4.
Tentacles (acrospheres): holotrichs 69 x 21.8 - 85
X 10.9-16.4; hoplotelous microbasic p-
mastigophors 33 x 5.5 - 82 x 8.8-9.7; hoplotelous

859

microbasic 6-mastigophors 31-34 x 4.4-4.9; 69-72
X 5.5 Tentacles (peduncles): microbasic b-
mastigophors 19-24 x 3.8-4.9; microbasic p-
mastigophors (very rare) ca. 36 x 7.6; atrichs (?)
ca. 15 X 5.5; spirocysts 19 x 3.8 - 38 x 5.5.
Actinopharynx: holotrichs 28 x 5.5 - 46 x 12;
atrichs (rare) ca. 12 x 4.9; microbasic p-
mastigophors (rare) ca. 22 x 6; spirocysts 17-23 x
3.8-4.4. Filaments: holotrichs 37 x 12 - 78 x 32.7;
hoplotelous microbasic p-mastigophors 24 x 5.5 -
44 X 10.9-12; atrichs (?) 34 x 3.8-35 x 4.4; spi-
rocysts (rare) ca. 31 x 3.8 jum.

The three individuals in the collection having a
whitish color of the column (see above) deviate
from the combination of the cnidom and the
frequency of the nematocysts in some organs.
Tentacles (acrospheres): holotrichs 49 x 16.4 - 65.4
X 35.4; atrichs (?) 22 x 5.5 - 41 x 10.4; microbasic
p-mastigophors 19 x 4.9 - 46 x 6.5. Tentacles
(peduncles): holotrichs ca. 66.5 x 19.6; hoplotelous
microbasic p-mastigophors 37 x 7.1 - 64 x 8.7;
microbasic 6(?)-mastigophors 61 x 6.5 - 88 x 8.7;
spirocysts 23 x 2.7 - 52 x 4.4 jum.

It is probable that the nematocysts charac-
terized as atrichs in the tentacles of the whitish
variety actually are holotrichs, the structure of
which was made unobservable by the fixing agent.
Difficulties in distinguishing between the two
nematocyst types has been pointed out by Carl-
gren (1945) with concern to corallimorpharians.

Until studies on vital material of the whitish
color form have been undertaken, which will
possibly confirm the presence of atrichs in the
acrospheres, I am inclined to consider it as a
member of the species Corynactis delawarei. [In
Corynactis annnlata (Swedish Museum of Natural
History, reference number 1244) collected off
Tristan da Cunha, there is, however, a nemato-
cyst equipment in the acrospheres suggestive of
that described in the whitish color variety: holo-
trichs 46 X 12.5 - 60 X 9.3; microbasic p-
mastigophors (rare) ca. 20 x 4.9; atrichs (?)
19 X 5.4 - 25 X 6.5 jum; microbasic 6-mastigophors
were not found in the specimen studied by me.]

Actiniaria
Edwardsiidae

Edwardsta sulcata (Verrill 1864)

OCCURRENCE.-44°00'N, 68°15'W, 110 m, silt-
860

FISHERY BULLETIN; VOL. 74, NO. 4

clay, 6 specimens, collected from three dredges.

GENERAL CHARACTERISTICS.-The physa
is well developed. The scapus is divided into
longitudinal compartments separated by the
mesenterial insertions of the macrocnemes. The
color of the scapus is yellowish grey. The nemathy-
bomes are numerous and often closely aggre-
gated (Figure 2A). The periderm is strong, but
easily falls off; its color is yellowish brown. The
scapulus is provided with high, longitudinally
oriented ridges in the strongly contracted mate-
rial. The maximum length of the scapus and
scapulus is 40 mm, whilst the largest diameter is 4
mm. The 16 tentacles are conical, without ridges or
nematocyst concentrations. The yellowish-white
actinopharynx has one distinct siphonoglyph. The
retractors of the eight macrocnemes are circum-
scribed and more or less reniform (Figure 2B). The
parietal muscles are strongly developed with 10-12
partly branched muscle (lamellae on each side of
the lamella of the septum (Figure 2C, D). The meso-
gloea of the mesentery is much thinner in the
vicinity of the retractor portion than was de-
scribed by Carlgren (1931) in Edwardsia elegans
(Figure 2B).

NEMAT0CYSTS.-5cajDMs (nemathybomes):
microbasic 6-mastigophors 90 x 5.4 - 110 x 6,
49 X 3.8 - 71 X 4.4 (the smaller nematocyst type
has reached a considerable less degree of
specialization than the larger type, the axial
filament of which shows great conformity with
that in 6-mastigophors of Edwardsia longicornis
Carlgren (cf. Carlgren 1940a). Scapulus: basitrichs
14 X 1.6 - 16 X 2.2. Tentacles: basitrichs 19-
26 X 2.2-2.7; spirocysts 10 x 2.7 - 25 x 3.8. Ac-
tinopharynx: basitrichs 16-26 x 2.2; microbasic
p-mastigophors (rare) ca. 24 x 4.4. Filaments:
basitrichs 19 x 2.2 - 24 x 2.7; microbasic p-
mastigophors (often with somewhat bent capsules;
axial filament = one third to half the length of
the capsule) 21-33 x 3.8-4.4; microbasic p-masti-
gophors (axial filament remarkably thin, and
about three-fourths of the length of the capsule)
23-32 X 4.4-4.8 jum.

There are many morphological similarities
between E. sulcata and E. sipuncnloides. The
6-mastigophors in the nemathybomes of the latter
species are, however, always much smaller (in one
specimen from the U.S. east coast, studied by me,
they were 62 x 4.9 - 72.5 x 5.5, 42-44 x 4.4 /xm;
cf. also Carlgren 1931).

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

ect

0.1 mm

Figure 2.-Edwardsia sulcata. A. Section through three nemathybomes. B. Cross section through a retractor. C, D.
Sections through two parietal muscles, ect-ectoderm, ent-entoderm, mast b-sectioned parts of microbasic 6-
mastigophors, mgl-mesogloea, mgl f-mesogloeal fibril, mgl pm-mesogloeal tract of parietal muscle.

861

FISHERY BULLETIN: VOL. 74, NO. 4

Halcampa duodecimcirrata M. Sars 1851

OCCURRENCE.-43°10'N, 70°25'W, 64 m, till, 1
specimen.

GENERAL CHARACTERISTICS.-They agree
with earlier descriptions of the species (cf. Carl-

gren 1893). The reddish scapus is provided with
tenaculi and distinct mesenterial insertions shim-
mering through the ectoderm. The six pairs of
macrocnemes (Figure 3A) (including two pairs of
directives) have strongly developed, circum-
scribed, and reniform retractors (Figure 3B). The
parietal muscles are provided with a rather small

100

mgl pm

Figure 3.-Hakampa duodecimcirrata. A. Section through a macrocneme. B. Cross section of a retractor. C.
Section through the peripheral part of a mesentery (the ectoderm is omitted in the figure), mgl-mesogloea, mgl
pm— mesogloeal lamella of parietal muscle, ooc— oocyte, pm— parietal muscle, retr ent— entoderm of retractor
muscle.

862

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

number of muscle lamellae (Figure 3C). The short,
conical tentacles were 10 (12?) in number.

NEMATOCYSTS.-Smp?/.s: basitrichs (rare) ca.
22 X 3.3, spirocysts (rare) ca. 28 x 3.8. ScapulUs:
spirocysts 25 x 3.8 - 49 x 4.4; basitrichs were not
found in the very damaged scapular ectoderm.
Tentacles: basitrichs (rare) 11.5 x 1.6-2.2 - 20 x
2.7; spirocysts 14 x 2.7 - 25 x 3.3. Actinopharynx:
microbasic p-mastigophors 24-32 x 3.5; spirocysts
17 x 2.2 - 25 X 4.9. Filaments: microbasic p-
mastigophors 22-23 x 4.4-4.9; basitrichs (?) 12-
13.6 X 3.8 jum.

The alleged differences as to the form of the
tentacles between H. duodecimcirrata and H.
chrysantheUum would argue against my decision
to refer the specimen to the former species.
Considering the extensive contraction of the
specimen, this characteristic must, however, be
regarded as of minor importance. Of greater
importance here is the conformity with H. duo-
decimcirrata of the sizes of the nematocysts in
different organs (cf. Carlgren 1940a). The number
of fertile mesenteries (eight in the studied spec-
imen) is another argument for the individual
being placed in H. duodecimcirrata.

Haloclavidae

Haloclava producta (Stimpson 1856)

OCCURRENCE. -39°00'N, 74°45'W, 15 m,
sandy bottom, 1 specimen.

GENERAL CHARACTERISTICS.-The col-
umn of the strongly contracted and partly dam-
aged specimen is fusiform with the ectoderm in
closely lying, transverse folds. The color is grey.
The scapus has a few sand grains attached to the
ectoderm. The length of the column is 16 mm, with
the greatest diameter (at the middle of the body)
about 8 mm. The retractors of the protomesen-
teries are very strong, circumscribed, and ren-
iform (Figure 4A). The four pairs of metamesen-
teries are weaker than the protomesenteries. The
parietal muscles are rather strong (Figure 4B).
There is no sphincter. The actinopharynx is rather
short with a very deep siphonoglyph. The number
of tentacles was impossible to confirm; as there
were only mesogloeal fragments left of the tenta-
cles, neither the nematocyst types nor their sizes
can be treated. The location of the fragments of
the tentacles favors the belief that there are 20
tentacles in the living animal.

NEMATOCYSTS.- Co/wmr?: basitrichs 20 x 2.7
- 24.5 X 3.3. Actinophanjnx: basitrichs 14 x 2.2 -
17.4 X 2.7, 38-57 x 4.4-4.9; spirocysts (only one
found) 43.1 X 3.8. Filaments: basitrichs 14 x 2.7 -
25 X 3.3, 70-83 x 4.4-5.5, 54.5 x 7 - 75 x 6.5-7.1
jLim.

Peachia parasitica (Agassiz 1859)

OCCURRENCE.-44°16'N, 67°38'W, 91 m, silt-
clay, 1 specimen.

GENERAL CHARACTERISTICS.-The col-
umn of the specimen is strongly contracted, with
the length 24 mm and the largest diameter (at the
middle of the body) 15 mm. The proximal diameter
of the column is 8 mm. The exact arrange-
ment of the extended lobes of the conchula was not
possible to observe in the specimen. There is no
sphincter. The only siphonoglyph is thick-walled
and of the typical Peachia appearance. The
number of mesenteries are 20, six pairs being
perfect, and supplied with strong, diffuse retrac-
tors with rather high muscle lamellae. The four
pairs of imperfect mesenteries are equipped with
rather small, diffuse retractors and are laterally
and ventrolaterally located. The 10 conical ten-
tacles have broad bases.

NEMATOCYSTS.-Co^Mmw: basitrichs 27-34 x
3.8-4.4. Tentacles: basitrichs 27-39 x 3.8-4.4; spi-
rocysts ca. 23 X 3.3 Actinopharynx: basitrichs
40-46 X 5.5; spirocysts 19-23 x 2.2-2.7. Filaments:
basitrichs 27 x 3.8 - 38 x 4.4; basitrichs (?) 39 x
6 - 45 X 7.6; microbasic p-mastigophors (rare) ca.
28 X 3.8 jum.

The filamental nematocysts named "basitrichs
(?)" (above) might be ;)-mastigophors. As I have
had no chance of observing the exploded capsules
and as the axial filament does not show the typical
p-mastigophor structure in the unexploded cap-
sules, I am not now inclined to consider these
nematocysts, which are probably homologous to
the "penicilli-like mastigophors" found by Carl-
gren (1940b), as microbasic p-mastigophors.

Actiniidae

Bolocera tuediae (Johnston 1832)

OCCURRENCE.-41°27'N, 69°02'W, 146 m, 1
specimen; 41°50'N, 69°26'W, 165 m, 1 specimen;
42°15.5'N, 69°59.5'W, ? m, 1 specimen; 42°25'N,

863

FISHERY BULLETIN: VOL. 74, NO. 4

Figure A.-Haloclava producta. A. Section through a retractor muscle. B. Section through the peripheral part of a
mesentery and adjacent parts of the body wall, ect-ectoderm, mgl-mesogloea, mgl pm-mesogloeal tract of parietal
muscle.

864

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

67°06'W, 366 m, 3 specimens; 42°48'N, 69°39'W, 183
m, 1 specimen; 43° 17'N, 70°24'W, 46 m, 1 specimen;
43°19'N, 67°16'W, 201 m, 1 specimen; 43°20'N,
68°45'W, 119 m, 1 specimen.

GENERAL CHARACTERISTICS.-The struc-
ture of the specimens agrees with earlier descrip-
tions of the species (cf. Carlgren 1891:242, 1893:50;
Stephenson 1935:130; Verrill 1922:G 115).

NEMATOCYSTS.-Co??nH»:basitrichsl6 x 2.2
- 21 X 2.7, 33 X 3.3 - 41 x 3.8 (-63 x 5.5).
Tentacles: hsisitrichs 21 x 2.7-36.5 x 3.3-3.8,52 x
3.8 - 87 X 4.9-6.5 (most often 60-70 x 4.5-6);
spirocysts 31-74 x 3.3-5.5; Actinopharynx: basi-
trichs 50 x 4.4 - 79 x 5.5; microbasic p-
mastigophors 23-33 x 5.5 Filaments: basitrichs
20-22 X 2.7-3.8, 50 x 3.8 - 74 x 4.4-5.5; microbasic
/)-mastigophors 19.6 x 4.9 - 35.4 x 5.5 jum.

Tealia crassicornis (Miiller 1776)

OCCURRENCE. -41°02'N, 69°00'W, 80 m,
gravelly sand, 1 specimen; 41°13'N, 68°58'W, 102
m, gravelly sand, 3 specimens; 41°33'N, 69°47'W,
27 m, gravelly sand, 1 specimen; 4r50'N, 67°56'W,
51 m, sand, 3 specimens; 42°11'N, 65°56'W, 229 m,
gravel, 1 specimen; 42°25'N, 66°05'W, 249 m,
gravel, 1 specimen; 42°26'N, 67°02'W, 366 m, 2
specimens; 43°11'N, 66°31'W, 92 m, gravel, 3
specimens; 43°11'N, 67°05'W, 181 m, 1 specimen;
43°12'N, 65°33'W, 73 m, shelly sand, 1 specimen;
43°33'N, 69°35'W, 159 m, 1 specimen; 43°37'N,
68°12'W, 198 m, 1 specimen; 43°49'N, 68°31'W, 95
m, 2 specimens; 43°52'N, 66°42'W, 102 m, 2
specimens; 43°53'N, 68°38'W, 91 m, 1 specimen;
44°26'N, 67°28'W, 73 m, till, 1 specimen; 44°30'N,
66°30'W, 157 m, 1 specimen.

GENERAL CHARACTERISTICS.-The mor-
phology of the studied specimens agrees with
earlier descriptions of the species (cf. Verrill 1867;
Carlgren 1893). The pedal disc is wide, circular
(diameter = 16-114 mm) or oval (16 x 22 - 47 x
63 mm). The rather firm column is in the contract-
ed state, cylindrical to semispherical, 14-38 mm
high. In those cases where the column is provided
with verrucae, these are chiefly spread over the
distal parts of the column. In some specimens
there is a distinct annulus with 48 marginal
verrucae. The number of mesenteries is somewhat
larger proximally than distally (in a specimen with
68 mesenteries only four were limited to the

proximal part of the column). The two outer of the
four to five mesenterial cycles are often not quite
completed. With the exception of the youngest,
proximally located cycle, and the 10 oldest perfect
pairs, the mesenteries are fertile. In the specimens
coming from 43°11'N, 66°31'W, the entodermal
and circumscribed sphincter is obviously asymme-
tric, with one half of it considerably more strongly
developed than the other.

Many of the specimens in the collection are
viviparous with larvae and young stages equipped
with tentacles lying in the proximal part of the
gastrocoele.

NEMATOCYSTS.-Co/hw/c basitrichs 5.5 x 1.1
- 9 X 2.7. 18 X 2.2 - 27 x 2.7-3.3; in larger
specimens found in the deeper localities: 12.5-14 x
2.7, 23-37 X 3.8, 79 x 5.5 - 83.4 x 8.2; spirocysts 22
X 2.7 - 69 x 4.4. Tentacles: basitrichs 10-14 x
2.2-2.7, 20 X 1.6 - 36.5 x 2.7-3.8; spirocysts 17.4 x
2.7 - 71 X 4.9-5.5. Actinopharnyx: basitrichs 49 x
5.5-6 - 91 X 6-7.1, 12 x 1.6 - 26 x 2.7; microbasic
p-mastigophors 23 x 4.9 - 30 x 5.5-6.5; spirocysts
(rare) 28-41 x 3.8. F(7a»ie«^s: basitrichs 11 x 2.2-
34 X 2.7, 49 X 5.5-6 - 68 x 7.1; microbasic
p-mastigophors 20 x 4.9 - 41 x 6.5 jum.

Actinostolidae

Actinostola callosa (Verrill 1882)

OCCURRENCE.-42°10'N, 69°57'W, 142 m, 1
specimen; 42°11'N, 68°16'W, 198 m, 1 specimen;
42°21'N, 68°02'W, 190 m, 3 specimens; 42°26'N,
66°35'W, 302 m, 1 specimen; 42°27'N, 66°08'W, 247
m, gravel, 1 specimen; 42°51'N, 65°12'W, 159 m, 1
specimen; 42°54'N, 69°35'W, 159 m, 2 specimens;
43°21'N, 69°57'W, 155 m, 1 specimen; 44°41'N,
66°14'W, 134 m, till, 1 specimen.

GENERAL CHARACTERISTICS.-The mor-
phology of this species has been carefully de-
scribed by Carlgren (1893:71). The length of the
column varies between 13 and 196 mm, and the
diameter of the pedal disc is 13-48 mm. The
tentacles are arranged in six cycles (6-1-6-1- 12 -(-
24-1-48-1- 96). The mesenteries (in five or six cycles)
are arranged according to the Actinostola rule.
Twenty-four pairs of mesenteries are perfect,
those of the two inner cycles (including the two
directive pairs) being sterile, as well as those of the
outer cycle.

865

FISHERY BULLETIN: VOL. 74, NO. 4

500
/jm

0.5

866

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

Figure 5.-Stomphia coccinea. A. Cross section of the sphincter in a young specimen. B. Section through part of a tentacle
showing the arrangement of the muscle fibrils (ml) in a young specimen. C. Section through a siphonoglyph. D. Section
through a mesentery in a young specimen. E. Section through a retractor and parietobasilar muscle of a young
specimen. F. Section of an oocyte and nurse cells, mgl-mesogloea, mgl r-mesogloea of retractor, nc-nurse cell,
ns-nucleus, ooc-oocyte, pbm-parietobasilar muscle, retr ent-entoderm of retractor muscle, y-yolk.

867

FISHERY BULLETIN: VOL. 74, NO. 4

NEUATOCYSTS.-Column: basitrichs 19 x 2.7
- 35.4 X 2.7-3.3; spirocysts 24.5 x 3.3 - 57 x 5.5-6.
Tentacles: basitrichs 26 x 2.2 - 41 x 2.7 (most
often 30-35 x 2.7); microbasic 6-mastigophors (in
the apex) 42.5 x 7.1 - 56 x 7.1-8.2; spirocysts 27 x
2.7 - 64 X 7.6. Actinophanjnx: basitrichs 21-32 x
2.7; microbasic p-mastigophors 22-28 x 4.9.
Filaments: microbasic p-mastigophors 21-28 x
4.4, 20-29 X 5.5 jum. The cnidom in specimens from
the eastern North Atlantic was described by
Carlgren (1940a).

Stomphia coccinea (Miiller 1776)

OCCURRENCE.-41°20'N, 69°22'W, 49 m, 3
specimens; 41°37'N, 66°16'W, 91 m, sand, 4
specimens; 42°18'N, 65°28'W, 113 m, sandy gravel,
1 specimen; 42°32'N, 65°39'W, 95 m, gravel, 1
specimen; 42°40'N, 65°56'W, 91 m, sandy gravel, 4
specimens; 43°10'N, 66°04'W, 92 m, gravel, 1
specimen; 43°21'N, 66°22'W, 60 m, shelly gravel, 2
specimens; 44°12'N, 66°36'W, 91 m, gravel, 1
specimen; 44°16'N, 66°28'W, 201 m, gravel, 1
specimen; 44°24'N, 67°14'W, 90 m, till, 1 specimen;
44°25'N, 66°25'W, 188 m, till, 3 specimens; 44°26'N,
66°19'W, 174 m, till, 1 specimen.

GENERAL CHARACTERISTICS.-The mor-
phology of the studied specimens agrees with
earlier descriptions made of the species (e.g.,
Carlgren 1893). The height of the contracted
column is 3-28 mm. The pedal disc is wide with a
distinct limbus. The relations between the length
of the column and the diameter of the pedal disc is
in the contracted state 14/23-7/17. The meso-
gloeal, diffuse sphincter is distally very strong
(Figure 5A). The tentacles, conical and longi-
tudinally furrowed with an apical pore, are ar-
ranged in four (sometimes five ?) cycles. The
tentacular muscles are mesogloeal (Figure 5B).
The number of mesenteries varies (in one of the
larger specimens it is equal to 120 in the pro.ximal
part of the body). Most often (with the exception
of the southernmost specimens) there are 16 pairs
of perfect and sterile mesenteries. (In the
specimens from 41°20'N, the number of perfect
and sterile mesenteries is approximately 24-29,
with an organization reminiscent of that in, e.g.,
Parasicyonis.) The long, folded actinopharynx is
equipped with two siphonoglyphs (Figure 5C). All
the imperfect mesenteries excluding those of the
last cycle are often fertile (Figure 5D), the oocytes
being provided with well-developed nurse cells

868

(Figure 5F) during oogenesis. The parietobasilar
muscles form even in very young individuals their
own lobes high up in the column (Figure 5D, E).

NEMATOCYSTS. -Co/i/mn: basitrichs 12-
20 X 2.2-2.7, 30.5-38 x 4.4-5.5. Tentacles: basi-
trichs 14 X 2.2 - 24.5 X 3.3; microbasic 6-masti-
gophors (30.5 x 6.5 -) 39-53 x 6.5-7.1; spirocysts
19 X 3.3 - 50 X 4.4-5.5 (in the young specimen
from 44°24'N, the column of which was 3 mm high
and the number of tentacles equal to 36, there
was a somewhat different size for the tentacular
nematocysts: basitrichs 15 x 2.2 - 22 x 2.7,
microbasic 6-mastigophors 28 x 5.3 - 33 x 7.1,
spirocysts 14-22 x 2.7-3.8 jum). Actinopharynx:
basitrichs 14 x 2.2 - 23 x 3.8; microbasic p-
mastigophors 18 x 3.3 - 27 x 4.9; spirocysts 22 x
3.8-57 X 4.9-5.5. Fi/amenfs: basitrichs 9.5 x 2.2-
22 X 2.7; microbasic p-mastigophors (17 x 3.8 -) 19
X 4.9 - 29 X 5.5, 24 x 3.3 - 29 x 3.3-4.4 /xm.

My placing of the three specimens from the
southernmost station within S. coccinea may be
discussed. In some morphological characteristics,
they resemble Anthosactis as well as Parasicyonis;
apart from the development of the perfect mes-
enteries, the morphological differences between
these, obviously young individuals, and the adult,
typical S. coccinea are, however, not so com-
prehensive as to require description of a new
subspecies.

Paranthus rapif omits (Lesueur 1817)

OCCURRENCE. -37°49'N, 75°25'W, 12 m,
sand-silt-clay, 1 specimen.

GENERAL CHARACTERISTICS.-The col-
umn is smooth, much wider distally than prox-
imally, and with a reddish brown color. The length
of the column is 26 mm, with the proximal
diameter 8 mm; the distal is 23 mm. The numerous
tentacles are arranged in five (six ?) cycles. They
are imperfectly retractile and acuminate. The
sphincter is mesogloeal, of diffuse type, and weak.
The yellowish, longitudinally folded actino-
pharynx is provided with two siphonoglyphs. The
mesenteries are proximally fewer than distally,
where they are arranged in four cycles (6-(-6-i-
12-1-24 pairs). Twelve pairs of mesenteries
(including the two pairs of directives) are perfect.
The mesenterial retractors are of diffuse-restrict-
ed type. The parietobasilar muscles are only

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

slightly developed. The pedal disc is well defined
and excavated. The individual studied was sterile.

NEMATOCYSTS. -Co/wmn (distally): basi-
trichs 17 x 1.6; microbasic p-mastigophors (?) ca.
16 X 3.3; spirocysts 33-60 x 3.8. Column (prox-
imally): basitrichs 21-26 x 2.2, 16 x 2.7 - 26 x
3.8; microbasic p-mastigophors (rare) ca. 33 x
5.5; spirocysts 14 x 2.7 - 53 x 3.3. Tentacles:
basitrichs 22 x 2.2 - 24.5 x 2.7; microbasic
jo-mastigophors 22 x 3.8 - 27 x 4.4; spirocysts 15
X 2.2 - 26 X 3.3. Actinopharynx: basitrichs 25 x
2.7 - 30 X 3.3; microbasic p-mastigophors 18.5 x
4.9 - 24.5 X 6; microbasic jo(?)-mastigophors 27 x
4.9 - 30 X 5.5. Filaments: basitrichs 22 x 3.3 - 32
X 3.8-4.9; microbasic /^-mastigophors (axial
filament = about half the length of the capsule) 20
X 4.4 - 26 X 5.4; microbasic p-mastigophors (axial
filament = almost the length of the capsule) 12.5
X 4.4 - 15 X 4.9; spirocysts 49-57 x 3.8-5.5 jum.

The most obvious difference between the
above-mentioned specimen and the earlier de-
scription of the species (cf. Carlgren and Hedg-
peth 1952:159), besides the different color of the
column and the occurrence of 12 pairs of perfect
mesenteries, is the size of the filamental basi-
trichs. While the filaments in material from Port
Aransas and Port Isabel are provided with basi-
trichs ranging in size from 12.7 to 14 x 2.2 jum, the
above-described specimen has much larger nema-
tocysts of the corresponding type: 22 x 3.3 -
32 X 4.9 jum. The same tendency can be seen also
with regards to the basitrichs of the column.

Antho/oba perdix CVerrill 1882)

OCCURRENCE.-40°06'N, 71°00'W, 179 m, 1
specimen; 40°10'N, 70°00'W, 114 m, silty sand, 1
specimen; 40°10'N, 71°15'W, 110 m, silty sand, 1
specimen.

GENERAL CHARACTERISTICS.-The col-
umn is smooth and, in the material studied, tran-
versely wrinkled. The smaller specimens are
olive-shaped; the larger specimen is cup-shaped
with an expanded distal part. The color is greyish,
with scattered, irregularly shaped, brownish spots.
The oral disc of the larger specimen is greyish
yellow with faintly marked, brown, and radially
directed streaks. An outer lip-shaped fold is here
provided on its outside with a zone, reddish brown
in color. Two parallel ribbons of the same color

divide this fold and the central part of the oral disc
into two halves. The excavated pedal disc is faintly
set off from the column. The length of the column
is 12-21 mm, and its diameter is 16-44 mm. The
pedal disc is maximally 23 mm in diameter. The
tentacles are numerous (in the largest specimen
about 600), short and conical, greyish in color. They
are longitudinally furrowed and provided with an
apical pore; in the smaller specimens they are
sometimes equipped with small, papillar processes.
The tentacles are arranged in five cycles, those of
the outer cycles being much smaller than the inner
ones (even in the largest specimen the outer
tentacles are papillary). The fifth cycle of mesen-
teries is not complete. The number of perfect
mesenteries in the larger specimen is 48. There are
more mesenteries distally than proximally. The
sphincter is alveolar (Figure 6). The retractors are
diffuse and extended in length. The entoderm of
the tentacles and the oral disc is reddish brown. All
the specimens were sterile.

NEMATOCYSTS. -Co/Mmw: basitrichs 20-
28 X 2.7-3.3. Tentacles: basitrichs 15-16 x 1.6-2.2,
23 X 2.2 - 36 X 3.3-3.8; spirocysts (very numerous)

Figure 6.-Antholoba perdix. Sections through the distal parts
of the sphincter in young specimens, mgl-mesogloeal layer.

869

FISHERY BULLETIN: VOL. 74, NO. 4

19 X 3.3 - 46 X 3.8. Actinopharynx: basitrichs
14 X 1.6 - 16 X 2.2 (- 27 x 2.7); microbasic p-
mastigophors 15 x 3.3 - 31 x 4.9. Filaments: ba-
sitrichs 14 X 1.6 - 28 X 2.7; microbasic jo-masti-
gophors (very numerous) 14 x 4.4 - 31 x 4.9-
5.5 jLim.

Metridiidae

Metridium senile fimbriatum [WerriW 1865)

OCCURRENCE. -40°35'N, 67°59'W, 84 m,
gravel, 5 specimens; 40°40'N, 68°01'W, 84 m, sand,
1 specimen; 40°51'N, 68°55'W, 66 m, sand, 4
specimens; 41°04'N, 71°24'W, 42 m, 1 specimen;
42°00'N, 69°56'W, 48 m, gravelly sand, 1 specimen;
42°15'N, 70°12'W, 26 m, 3 specimens; 42°22'N,
70°18'W, 33 m, 2 specimens; 42°42'N, 65°18'W, 91
m, 1 specimen; 42°42'N, 65°40'W, 90 m, 3
specimens; 42°47'N, 66°25'W, 99 m, 1 specimen;
42°54'N, 66°14'W, 166 m, 2 specimens; 43°07'N,
65°57'W, 97 m, 1 specimen; 43°17'N, 65°35'W, 40 m,
gravel, 1 specimen; 43°36'N, 68°50'W, 115 m, 2
specimens; 43°43'N, 66°30'W, 84 m, 1 specimen;
43°44'N, 66°28'W, 75 m, 1 specimen.

GENERAL CHARACTERISTICS.-They agree
with earlier descriptions of the species (cf., e.g.,
Carlgren 1893:102). The height of the column
varies between 5 and 35 mm, and the diameter of
the pedal disc is 7-47 mm. The color is yellow to
yellowish brown in the preserved state. The
specimens from 40°5rN were all very young, the
youngest being equipped with only 12 tentacles.

NEMATOCYSTS (sizes of the above-men-
tioned, small specimens in parentheses).- Co^?/mn:
basitrichs 15 x 2.7 - 19 x 3.3 (10 x 1.6 - 12 x
2.2); microbasic amastigophors 26-28 x 3.8-4.4
(16 X 3.3 - 21 X 3.8-4.4); microbasic p(?)-masti-
gophors 23-31 x 3.8-4.4; spirocysts 22-27 x 3.3-
3.8. Tentacles: basitrichs (11.5 x 1.6 -) 18 x 2.2 - 28
X 2.7-3.3 (17-21 X 2.7-3.3); microbasic amastigo-
phors 13 X 2.7 - 15 X 3.3; spirocysts 21 x 3.3 - 31 x
4.9 (11 X 2.7 - 17 X 4.4). Actinopharynx: basitrichs
26-39 X 3.8 (17 x 2.2 - 27 x 2.7-3.3); microbasic
p-mastigophors 22-23 x 3.8 (17-23 x 3.3-3.8);
microbasic amastigophors (rare) ca. 31 x 4.4.
Filaments: basitrichs (very rare) ca. 14 x 3.8 (12 x
2.7); microbasic p-mastigophors 16-25 x 4.4, 24-32
X 3.8-4.4 (12-14 X 4.4, 21-23 x 3.8-4.4). Acontia:
microbasic 6-mastigophors 51-67 x 3.2-4.9 (40 x

3.8 - 57 X 4.4); microbasic amastigophors (28 x
3.8 -) 49-64 X 4.9-5.5 (36 x 4.4 - 55 x 5.5) jum.

Aiptasiomorphidae

Haliplanella luciae (Verrill 1898)

OCCURRENCE.-39°00''n, 76°22'W, 16 m, silty
clay, 10 specimens.

GENERAL CHARACTERISTICS.-They agree
with earlier descriptions (Stephenson 1925:888,
1935:197; Field 1949:10). The sizes of the nema-
tocyst capsules deviate, however, in some respects
from what has been described earlier (cf . Carlgren
1940a).

NEMATOCYSTS. -Coiwrnw: basitrichs 10-
11 X 1.6, 19 X 3.3 - 23 X 3.8; microbasic p- or
amastigophors 19 x 3.8 - 21 x 4.4. Tentacles: ba-
sitrichs 15 X 1.6-20 X 2.2; microbasic p-
mastigophors 18 x 3.8 - 25 x 4.9; spirocysts ca.
16 X 4.4-4.9. Actinopharynx: basitrichs (?) (the
capsules are slightly bent) 30-32 x 2.7-3.3;
microbasic jo-mastigophors 23-27 x 3.8; micro-
basic p- or amastigophors 21-23 x 2.7-3.3. Fil-
aments: basitrichs 14 x 1.6 - 19 x 2.2; microbasic
p-mastigophors 22-28 x 3.8; microbasic a(?)-
mastigophors 17 x 3.3-28 x 3.8; spirocysts 12.5 x
2.7 - 17 X 5.4. Acontia: basitrichs 15-18 x 1.6;
microbasic p-mastigophors 43 x 5.5 - 56 x
6.5 jLim. It was not possible to determine if there
are any microbasic amastigophors present in the
acontia, as all the mastigophor capsules were
unexploded.

The only difference in the cnidom of the above-
mentioned specimens and earlier descriptions of
the species (cf. Carlgren 1945; Field 1949) besides
the unsettled presence of microbasic amasti-
gophors in the acontia (cf. Hand 1955) are the
occurrence in this sample, of basitrichs in the
actinopharynx, in agreement with the conditions
in Aiptasiomorpha texaensis (cf. Carlgren and
Hedgpeth 1952).

Sagartiidae
Sagartiogeton verrtllt drlgren 1942

OCCURRENCE.-40°32'N, 67°05'W, 338 m, 3
specimens, on fragments of mussel shell; 42°25'N,
66°21'W, 256 m, gravel, 1 specimen.

GENERAL CHARACTERISTICS.-The length

870

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

of the column varies between 8 and 18 mm, whilst
the diameter of the pedal disc is 12-20 mm. The
column is greyish, salmon-colored and divisible
into scapus and scapulus. The scapus is provided
with distinct mesenterial insertions shimmering
through the ectoderm and has small tenaculi. The
tentacles of the studied specimens are about 9 mm
long, conical and acute. They are hexamerously
arranged; in the largest specimen there are five
cycles (6 -I- 6 -I- 12 -1-24 -I- 40). The reddish-brown
actinopharynx is strongly folded and provided
with two siphonoglyphs. The color of the tentacles
is greyish white, the largest specimen with red-
dish-brown pigmentation. The pairs of mesente-
ries are arranged in four to five cycles (proximally
there are 90-100 mesenteries— distally only half
the number are developed). The number of perfect
mesenteries tends to vary. In one of the studied
specimens there are 17 pairs (including the two
directive pairs). The first cycle of mesenteries is
sterile. The retractors of the first cycle of mesen-
teries are strong, of a circumscribed diffuse type
(Figure 7A); in the other mesenteries they are
diffuse. The parietobasilar muscles are rather
weak (Figure 7B). The acontia are numerous and
whitish in color. The mesogloeal, diffuse spincter is
rather long.

NEMAT0CYSTS.-5m/>ws: basitrichs 9-11.5 x
1.6, 16-17 X 2.2-3.3; microbasic amastigophors
30.5-35 X 4.4, ca. 15 x 3.8 (rare). Tentacles: basi-
trichs 13 X 2.2 - 27 X 2.7; microbasic amasti-
gophors (16 X 3.8-) 26 X 4.4 - 44 x 6.5; spirocysts
22 x 3.8 - 36 X 6. Actinopharynx: basitrichs 27-
32 X 3.3; microbasic p-mastigophors ca. 23 x 4.4,
27-31 X 4.4-4.9. Filaments: basitrichs (rare) ca.
27 X 3.8; microbasic p-mastigophors 26-31 x
4.4-4.9; spirocysts 22 x 4.9 - 34 x 6. Acontia:
basitrichs 36.5 x 3.8 - 43 x 4.4-4.9; microbasic
amastigophors 57 x 6 - 64 x 7.1 jum.

Hormathiidae

Hormathia nodosa (Fabricius 1780)

OCCURRENCE.-40°54'N, 66°35'W, 265 m, 4
specimens; 41°30'N, 69°00'W, 146 m, till, 1
specimen; 42°14'N, 69°57'W, 102 m, 1 specimen;
42°26'N, 66°28'W, 265 m, gravel, 8 specimens.

GENERAL CHARACTERISTICS.-The 9-28
mm high scapus is provided with white tubercles
arranged in longitudinally oriented rows. The

etrent

Figure l.-Saqartiogeton rerrilli. A. Section of the retractor of a
perfect mesentery with adjacent parts of the actinopharynx.
B. Section of the peripheral part of a mesentery (from the distal
part of the scapus). ent— entoderm, mgl— mesogloea, retr
ent-entoderm of retractor muscle.

scapular ridges are white and are 12 in number.
The scapus is covered by a thin, greyish-white to
greyish-brown periderm and is equipped with
shallow longitudinal as well as transverse furrows;
on the edge of the scapulus only radiating furrows
are seen. The diameter of the column is 12-33 mm
(diameter of scapus: diameter of the pedal disc =
ca. 3:4). The tentacles are conical, reddish brown,
and in older specimens longitudinally furrowed.
They are 96 in number. The mesenteries have a
maximum of 48 pairs (6-1-6-1-12-1-24), 6 of which
(including the 2 pairs of directives) are perfect and
sterile. The anatomical characters agree with
earlier descriptions (cf. Carlgren 1893, 1933; Verrill
1922).

NEMATOCYSTS.-Scapws: basitrichs 8 x 1.1 -
11 X 1.6, 21-24 X 3.3-3.8; spirocysts 17 x 3.8 -
25 X 4.4. Tentacles: basitrichs 17 x 2.2 - 34 x
3.8; spirocysts 23 x 3.3-4.4 - 44 (56) x 5.4-7.6.
Actinopharynx: basitrichs 16 X 1.6 - 35 X 3.3-
3.8; microbasic p-mastigophors 23-33 x 3.8. Fila-

871

FISHERY BULLETIN: VOL. 74, NO. 4

ments: basitrichs 14 x 1.1 - 16.4 x 1.6, 28 x 3.3 - 31
X 3.8-4.4; microbasicp-mastigophors 21 x 3.3 - 23
X 3.8. Acontia: basitrichs 32-40 x 3.8-4.4 /xm.

Horniathia nodosa (.-') (Fabricius 1780)

OCCURRENCE.-41°34'N, 68°40'W, 128 m,
sandy silt, 1 specimen.

GENERAL CHARACTERISTICS.-The col-
umn is divisible into scapus and scapulus, the
former being provided with a thin periderm and
rather large, acuminated tubercles spread over the
surface. The color of the scapus is proximally dark
greyish brown, distally brown. Bordering upon the
scapulus there are 12 large marginal tubercles.
The pedal disc is not excavated; there are traces of
mussel shell. The tentacles lack bulbous swellings
on the abaxial side. They are arranged in four
cycles. The actinopharynx and the sphincter agree
with those in H. nodosa (cf. Carlgren 1893). The
number of mesenteries is 96 (6 -t- 6 -H 12-1-24 pairs),
the perfect ones being 24 pairs in the distal part of
the column. Immediately above the margin of the
actinopharynx there are 20 pairs of perfect mes-
enteries. Only the six pairs of protomesenteries
are sterile. The morphology of the retractors,
parietobasilar, and basilar muscles agrees with
that in typical H. nodosa. The length of the pre-
served specimen is: scapus 16 mm and scapulus 7
mm. The size of the pedal disc is 36 x 49 mm. The
sizes of the different nematocyst types differ only
slightly from those described in H. nodosa (see
above). The large number of perfect mesenteries
is, however, remarkable.

In view of the many morphological similarities
between this specimen and typical H. nodosa, I
consider it as an aberrant specimen of this species.

Actinauge ferr////' McMurrich 1893

OCCURRENCE. -42°11'N, 65°56'W, 229 m,
gravel, 1 specimen; 42°20'N, 67°28'W, 289 m, sandy
gravel, 1 specimen; 42°50'N, 69°00'W, 187 m,
sand-silt-clay, 1 specimen.

GENERAL CHARACTERISTICS.-The mor-
phology of these specimens agrees with earlier
descriptions of the species (cf. McMurrich 1893;
Carlgren 1933). The scapus is equipped with a
greyish-brown or brown periderm; it has a re-
ticular appearance, arising from transverse as

872

well as longitudinal, rather low, furrows. Distally
there are 12 coronary tubercles. The firm wall of
the scapulus is often whitish and is provided with
24, white, scapular ridges, proximally fusing two
by two into 12. The scapus is cylindrical or dome-
shaped, with the length 29-30 mm. The diameter of
the scapus is proximally 17-30 mm and distally
19-20 mm. The length of the scapulus is 14 mm. The
pedal disc is strongly excavated, often embracing
sand grains. The long and tapering tentacles are
arranged in four to five cycles. The outer tentacles
are basally provided with abaxial swellings, which
give rise to distinct processes. There are four
cycles of mesenteries. Six pairs (including the two
directive pairs) are perfect and sterile.

NEMATOCYSTS.-ScajoMs: basitrichs 8 x 1.6 -
23 X 4.4. Tentacles: basitrichs 12 x 2.2 - 27 x 2.7-
3.3, ca. 40 X 3.8 (rare); microbasic p-mastigophors
24.5 X 3.8-5.2 - 38 x 8.2; spirocysts 19 x 3.3 - 37 x
4.4-6; 46-56 x 5.5-7. Actinopharnyx: basitrichs 13
X 1.6 - 17 X 2.2, 28-50 x 3.3; microbasic p-
mastigophors 22 x 3.8 - 29 x 4.4. Filaments:
basitrichs 11 x 1.1 - 17 x 2.2; 28-30.5 x 3.3,
microbasic p-mastigophors 19 x 3.8-4.9 - 35 x 4.4.
Aco7itia: basitrichs (14 x 2.2 -) 26 x 3.3 - 36.5 x
3.8-4.4 jum.

Phelliactis americana n. sp.

HOLOTYPE.-Specimen collected by the vessel
Delaware from the type-locality (station number
27) on 19 February 1963 with an otter trawl.
Deposited in the U.S. National Museum, catalog
number USNM 54323.

TYPE-LOCALITY.-42°48'N, 63°42'W, 366 m,
temperature -t-1.7°C.

PARATYPE.-Specimen collected by the vessel
Albatross IV from station number 73 (42°17'N,
65°55'W, 238 m, gravel) on 15 August 1968 with a
1-m Naturalist dredge. Deposited in Northeast
Fisheries Center, Woods Hole.

DIAGNOSIS OF HOLOTYPE.-Column firm,
divisible into scapus and scapulus; somewhat
asymmetric. Scapus distally with 48 rows of large,
sometimes acute, tubercles. Scapular ridges about
70. Sphincter mesogloeal, and alveolar, very
strong. Tentacles about 190, conical, and long-
itudinally furrowed with basal, abaxial swellings.
Mesenteries in five cycles, 12 pairs being perfect

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

and sterile. Retractors of diffuse, restricted type.
Parietobasilar muscles weak. Cnidom: scapus
basitrichs; tentacles basitrichs and spirocysts;
actinopharynx basitrichs and microbasic p-
msLStigophors; filaments basitrichs and microbasic
jD-mastigophors; acontia basitrichs.

GENERAL CHARACTERISTICS.-The col-
umn is firm and divisible into scapus and scapulus.
It has a somewhat asymmetric appearance, one
half of the body being larger than the other. The
scapus (18 mm long) is cylindrical in the contracted
state and has a reticular appearance with low
tubercles formed by longitudinally as well as
transversely oriented, low furrows; distally the
scapus is provided with 48 rows of larger, some-
times acute, tubercles. The color of the remaining
traces of periderm is brownish. The proximal part
of the body is pillarlike, with the diameter 30 mm.
About 70 scapular ridges are continued in the
basilar swellings of the outer tentacles. The
sphincter is rather short, but very strong,
especially orally; it is alveolar and vertically
stratified (Figure 8B). The actinopharynx is
equipped with 12 longitudinal folds on each side of
the two symmetrically arranged siphonoglyphs.
The tentacles number about 190; they are rather
short, conical, and longitudinally furrowed and are
basally provided with abaxial swellings. The me-
senteries are arranged in five cycles
(6 -H 6 -I- 12 -(- 25 -I- 50 pairs), 12 pairs (including the 2
pairs of directive mesenteries) being perfect and
sterile. The retractors are of diffuse type, rather
strong, and with their, in some perfect mesente-
ries, rather restricted pennons near to the actino-
pharyngeal wall (Figure 8C, D). The parietobasilar
muscles are weak. The column, being somewhat
wider distally than proximally, lacks cinclides. The
whitish acontia are numerous and often very long.
The mesogloeal layer is very thick in the whole
column as well as in the mesenteries.

In the paratype the distal part of the column is
in some parts severely damaged; the oral part is
also introverted, giving rise to an oral slit, 58 mm
long. The length of the scapus in this specimen is
40-30 mm; it is provided with low tubercles spread
out over the column; distally there are 24 tubercles
bordering the scapular ridges. The tentacles are
arranged in four cycles (there are about 70 in the
outer cycle) and are provided with abaxial swell-
ings (Figure 8A). The mesenteries are hexamer-
ously arranged in five cycles (the last cycle is,
however, not complete in this specimen); prox-

imally there are 75 pairs in total. The number of
perfect and sterile mesenteries was impossible to
determine in the paratype, but there are probably
less than 12 pairs (probably 8). The wide and
peripherally almost membraneous pedal disc is, to
a small extent, excavated; its diameter measures
90 mm.

NEMAT0CYSTS.-5ca/)ws: basitrichs ca. 14 x
1.6-2.2, 24.5-44 x 3.3; spirocysts (not found in the
paratype) 27 x 4.4 - 60 x 5.5. Tentacles: basi-
trichs 17-21 X 2.2 (not common), 34-43 x 3.3-3.8
spirocysts 38 X 4.4-4.9 - 75 x 8.7. Actinopharynx
basitrichs 16 x 2.2 (rare), 37 x 3.3 - 42 x 3.8
microbasic p-mastigophors 30 x 4.4 - 39 x 4.9.
Filaments: basitrichs 12 x 1.6 - 22 x 2.2, 33 x
2.7 - 48 x 3.3; microbasic p-mastigophors 28 x
4.4 - 34 X 4.9. Acontia: basitrichs 16 x 2.2 -
23 x2.5, 32 X 3.3 - 52 x 3.8 Mm.

There are some morphological similarities
between the above described specimens and Phel-
liactis hertwigii Simon as well as Ph. incerta
Carlgren. The retractors of the perfect mesente-
ries are, however, stronger in Ph. americana, and
the number of perfect mesenteries is larger (in the
holotype 12 pairs).

Amphianthus nitidus (Verrill 1899)

OCCURRENCE.-41°27'N, 66°06'W, 128 m, 1
specimen; 41°39'N, 65°50'W, 183 m, 6 specimens;
42°10'N, 65°29'W, 163 m, 1 specimen.

GENERAL CHARACTERISTICS.-The col-
umn is firm, in the contracted state semispherical,
and 9-16 mm high. The color is greyish white with
a blue luster. The scapus is, in one of the studied
specimens, equipped with eight low, extended
tubercles. The diameter of the pedal disc is 12-16
mm. There is a distinct limbus. The tentacles are
hexamerously arranged in four to five cycles
(6 -I- 6 -H 12 -I- 24 -I- a seldom completed fifth cycle),
rather short, conical, and sometimes provided with
an apical pore. The inner tentacles are larger than
the outer ones. There are four to five cycles (57
pairs at most) of hexamerously arranged mesen-
teries, eight to nine pairs of which (including the
two directive pairs) are perfect. All the mesente-
ries, except those of the last cycle and at least one
of the directive pairs, are fertile. The number of
mesenteries is larger proximally than distally. The
acontia are numerous and yellow. The distally very
strong, mesogloeal sphincter, the actinopharynx,

873

FISHERY BULLETIN: VOL. 74, NO. 4

&-nigl

Figure S.—Phelliactis americana. n. sp. A. Section through the basal part of a tentacle. B. Section through a part
of the sphincter, showing the alveolar arrangement of the muscle fibrils (the fibrils are omitted in the figure). C.
Section through a retractor from the third cycle of mesenteries. D. Cross section of a part of a retractor of one of
the directive mesenteries, ent-entoderm, mgl-mesogloea.

874

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

and musculature of the specimens agree with
earher descriptions (cf . Carlgren 1934).

NEMATOCYSTS.- Co/»w??: basitrichs 8 x
1.6 - 12.5 X 2.7, 29-35 x 2.7; microbasic p-masti-
gophors 17 X 4.9 - 27 x 6; spirocysts 29 x 3.3 - 61
X 6. Tentacles: basitrichs 18.5 x 3.8 - 30 x 6;
microbasic p(?)-mastigophors 18.5 x 4.4 - 29 x
4.4-5.5; spirocysts 19 x 3.3 - 47 x 7.6. Actino-
phari/n.r: basitrichs 24-25 x 3.3; microbasic p-
mastigophors 23 x 4.4 - 26 x 5.5 (axial filament =
about half the length of the capsule); ca. 27 x 4.9
(axial filament almost as long as the capsule).
Filaments: basitrichs ca. 9 x 2.2; microbasic
p-mastigophors 22 x 4.4 - 28 x 4.9. Acontia:
basitrichs ca. 14 x 2.2; 42 x 6 - 57 x 6.5 jim.

Stephanauge nexilis (Verrill 1883)

OCCURRENCE.-41°54'N, 65°44'W, 366 m, 2
specimens; on the denuded axis of an octocoral.

GENERAL CHARACTERISTICS. -The yel-
lowish, firm column is strongly elongated in the
sagittal plane. The dimensions of the scapus is
proximally 22 x 4 mm, the height of the column
being 7 mm. In one of the specimens, the scapus is
provided with 26 low, circularly arranged tubercles
bordering 28 vague, radiating scapular ridges. The
mesenterial insertions into the body wall are
distinct. The number of the yellow, short, basally
wide tentacles is not greater than that of the
mesenteries (72 and 78). The sphincter is alveolar
and strong, slowly diminishing in thickness
towards the proximal part of the scapus. The wide
actinopharynx is brownish yellow. It is equipped
with two siphonoglyphs. The mesenteries are
hexamerously arranged, more than six pairs
(including the two pairs of directives) being
perfect. At least some of the perfect mesenteries
are equipped with genital organs. The retractors
are diffuse and rather weak. The strong parieto-
basilar muscles produce distinct muscular lobes
high up in the scapus, approximately at the middle
of the mesenteries. The number of mesenteries is
not greater proximally than distally. No acontia
were found in these specimens (they might have
been few and hidden by the strongly developed
filaments), but basitrichs of probably acontian
origin were measured in one of the specimens. No
cinclides could be found.

NEMATOCYSTS.-Scajoi/s: basitrichs (rare) ca.
12 X 2.2; microbasic p-mastigophors 12 x 4.4-
5.5 - 19 X 5.5. Tentacles: basitrichs 9 x 1.7 - 20 x
2.7; microbasic p-mastigophors (axial filament
almost as long as the capsule; diameter =
1.5 iim) 21 X 5.5 - 23 x 7.1; spirocysts 25 x 3.8 -
49 X 4.4. Actinopharynx: basitrichs 15-18 x 3.3;
microbasic p-mastigophors 17 x 4.4 - 22 x
5.4-6. Filaments: basitrichs 14-16 x 2.7-3.3
(-28 X 2.7); microbasic p-mastigophors 16 x
3.8 (-6) - 27 X 5.5. Acontia{l): basitrichs 30.5-36.5
X 3.8 /xm.

Hormathiidae (?)
Stephanauge (?) spongicola (Verrill 1883)

OCCURRENCE.-39°56'N, 69°45'W, 201 m, 3
specimens; 40°00'N, 69°30'W, 128 m, 3 specimens;
40°02'N, 70°47'W, 161 m, 6 specimens; 40°03'N,
71°16'W, 183 m, 16 specimens. At all the localities
the specimens were found on the outside of the
parchmentlike tubes of onuphid polychaetes.

GENERAL CHARACTERISTICS.-The col-
umn is often smooth but sometimes provided with
a few adhesive warts; it is divisible into scapus and
scapulus. The scapus was reddish brown to greyish
brown in the preserved material, its length being
4-12 mm. The largest diameter is 10-11 mm. The
scapulus is whitish to pale red. The periderm of the
scapus is thin and easily falls off. The tentacles are
conical, acute, and yellowish. They are hexame-
rously arranged in four to six cycles, those of the
inner cycles being distinctly longer than the outer
ones. The sphincter is short, mesogloeal, and
agrees in its structure with that described by
Carlgren (1950). It is not capable of covering all the
tentacles. The actinopharynx is about three-
quarters the length of the column, wide, and
equipped with 18-20 deep, closely lying longi-
tudinal folds (Figure 9A); it is yellowish in color.
One (?) to four siphonoglyphs are present. In a
specimen with two siphonoglyphs there was an
eccentric position for them. There are (5-) 8-12
pairs of perfect mesenteries, the imperfect ones
being 8-16 (-22) pairs. The structure of the retrac-
tors of the perfect mesenteries was in agreement
with that described by Carlgren (1950) and in
many ways reminiscent of those in Phellia gaus-
apata. The number of directive pairs varies, being
two, three, or four. The retractors of the perfect

875

FISHERY BULLETIN: VOL. 74, NO. 4

o
o

actinoph
epith

500/um

Figure 9.—Stephanauge( ?) spongicola. A. Section through part of the actinopharynx. B. Cross section of a retractor of one
of the protomesenteries. C. Cross section of the peripheral part of a perfect mesentery (the section is from the distal part
of the column). D. Section through a part of the body wall. E. Section through a mesenterj' of the last cycle, actinoph
epith— epithelium of actinopharynx, ect— ectoderm, eincl— ectodermal invagination forming an imperforate cinclis,
ent-entoderm, mgl-mesogloea, pbm-parietobasilar muscle, retr ent-entoderm of retractor muscle.

876

WIDERSTEN: ANTHOZOA FROM EASTERN COAST OF UNITED STATES

100/jm

mesenteries are strong, five to eight pairs being
circumscribed, and sometimes reniform (Figure
9B), those of the other perfect mesenteries being
diffuse but with a tendency to become restricted.
The parietobasilar muscles are rather strong,
forming distinct lamellae on the peripheral parts
of the mesenteries (Figure 9C). The imperfect
mesenteries lack retractors as well as filaments
(always ?) (Figure 9E). In those specimens where
genital organs were found, these were always
developed in perfect mesenteries. The acontia are
numerous and provided with basitrichs. Only one
imperforate cinclis (Figure 9D) was found in the
sections of the species. Probably the species re-
produces asexually by laceration. The proximal
part of the column and the often wide pedal disc
are often asymmetrical.

NEMATOCYSTS.-Co/wmn: basitrichs (21 x
2.7 -) 23-27 X 3.3-3.8, 31-39 x 4.4; atrichs 19-20
X 4.4-4.9, 39-45 x 12.5-14.7. Tentacles: basitrichs
11 X 1.6-2.2 - 33 X 3.3-3.8; atrichs (not common;

in many tentacles completely missing) 39 x 13.1 -
49 X 5.5, spirocysts (very numerous, and in
some of the studied specimens with a very small
variation in size) 17 x 2.2 - 34 x 3.8-4.9. Actino-
pharynx: basitrichs 14 x 2.2 - 32 x 3.8; microba-
sic p-mastigophors 17-26 x 3.8-4.4. Filaments:
basitrichs 12-15 x 2.2; microbasic p-mastigophors
13 X 3.3 - 26 X 4.3-5.5. Acontia: basitrichs 13-
16 X 2.2-2.7, 33 x 3.3 - 45 x 3.8 /xm.

In specimens from 40°03'N, there were also
found atrichs in the filaments (12 x 6, 18 x
4.9 - 24 X 5 jLtm) as well as holotrichs (22 x 4.9 -
24 X Sjiim). Both these nematocyst types are
probably residues of intaken food-the specimens
in question were found together with some in-
dividuals of Epizoanthus incrustatus.

This species, first described by Verrill (1883) as
Sagartia spongicola, has been the object of later
investigations by, e.g., McMurrich (1898) and
Carlgren (1950). Carlgren (1950) (on the basis of
acontian armament with basitrichs ?) described
the species as a hormathiid and a member of the

877

FISHERY BULLETIN: VOL. 74, NO. 4

genus Stephayiauge, being aware of the existing
anatomical differences in the development of the
sphincter, the retractors, and the number of perfect
mesenteries, siphonoglyphs, and directive mesen-
teries. To these differences should be added the
occurrence of atrichous haplonemes, not only in
the column ectoderm, but also in, at least, some of
the inner tentacles. The arrangement of the
mesenteries into filament-equipped perfect and
into imperfect ones devoid of filaments as well as
retractors should also be taken into consideration.
The morphology of this species shows so many
differences from other species of the genus Ste-
phanauge that I consider it very doubtful to place
the species in this genus, or, taking into conside-
ration the occurrence of atrichs in the studied
specimens, in any other hormathiid genus.

ACKNOWLEDGMENTS

I thank Roland L. Wigley, National Marine
Fisheries Service, Woods Hole, who kindly placed
the material at my disposal and who also supplied
me with station data.

I also thank Karl-Georg Nyholm, Institute of
Zoology, University of Uppsala, and Tor G.
Karling and Roy Olerod, Swedish Museum of
Natural History, Stockholm, for their kind com-
pliance in placing laboratory facilities and desired
material at my disposal. My thanks are also due to
Bo Molin, Uppsala, for the magnificent technical
assistance he provided by sectioning and staining
some of the studied material and for redrawing
the figures.

LITERATURE CITED

Carlgren, 0.

1891. Beitrage zur Kenntnis der Actinien-Gattung Bolocera
Gosse. Ofvers. K. Vetenskaspakad. Fbrh. 48:241-250.

1893. Studien uber Nordische Aktinien I. K. Sven. Veten-
skaspakad. Handl. 25:1-148.

1931. Zur Kenntnis der Actiniaria Abasilaria. Ark. Zool.
23A(3), 48 p.

1933. The Godthaab Ezpedition 1928. Zoantharia and
Actiniaria. Medd. Gretnland 79(8), 55 p.

1934. Zur Revision der Actiniarien. Ark. Zool. 26A(18),36p.
1940a. A contribution to the knowledge of the structure and

distribution of the cnidom in the Anthozoa. K. Fysiogr.

Sallsk. Handl. Lund, N.F. 51:1-62.
1940b. Actiniaria from Alaska and Arctic waters. J. Wash.

Acad. Sci. 30:21-27.
1945. Further contributions to the knowledge of the cnidom

in the Anthozoa, especially in the Actiniaria. Lunds Univ.

Arsskr., Ny Fobijl 41(9):l-24.

1949. A survey of the Ptychodactiaria, Corallimorpharia,
and Actiniaria. K. Sven. Vetenskapsakad. Handl. 4(1):
1-121.

1950. A revision of some Actiniaria described by A. E.
Verrill. J. Wash. Acad. Sci. 40:22-28.

Carlgren, 0., and J. W. Hedgpeth.

1952. Actiniaria, Zoantharia and Ceriantharia from shallow
water in the northwestern Gulf of Mexico. Publ. Inst.
Mar. Sci. 2(2):141-172.
Field, L. R.

1949. Sea anemones and corals of Beaufort, North
Carolina. Duke Univ. Mar. Stn. Bull. 5:1-39.
Hand,C.

1955. The sea anemones of central California. Part III. The
Acontiarian anemones. Wasmann J. Biol. 13:189-251.
McMURRICH, J. P.

1893. Report on the Actiniae collected by the United States
Fish Commission steamer Albatross during the winter of
1887-1888. Proc. U.S. Natl. Mus. 16:119-216.

1898. Report on the Actiniaria collected by the Bahama
Expedition of the State University of Iowa, 1893. Bull.
Lab. Nat. Hist., State Univ. Iowa 4:225-249.
Stephenson, T. A.

1925. On a new British sea anemone. J. Mar. Biol. Assoc.
U.K. 13:880-890.

1935. The British sea anemones. Vol. II. Ray Soc. (Lond.),
Publ. 121, 426 p.
Verrill, A. E.

1864. Revision of the Polypi of the eastern coast of the
United States. Mem. Boston Soc. Nat. Hist. 1:1-45.

1867. Notes on the Radiata in the Museum of Yale College,
with descriptions of new genera and species. Trans. Conn.
Acad. Arts Sci. 1:247-596.

1883. Report on the results of dredging, under the super-
vision of Alexander Agassiz, on the east coast of the
United States, during the summer of 18S0, by the U.S.
Coast Survey steamer "Blake',' Commander J. R. Bartlett,
U^.N., commanding. XXI. Report on the Anthozoa, and
on some additional species dredged by the "Blake" in
1877-79, and by the U.S. Fish Commission Steamer "Fish
Hawk" in 1880-82. Bull. Mus. Comp. Zool. ll(l):l-72.

1922. The Actiniaria of the Canadian Arctic Expeditions,
with notes on interesting species from Hudson Bay and
other Canadian localities. In A. E. Verrill (editor), Al-
cyonaria and Actinaria, p. 89-164. Rep. Can. Arctic Exped.
1913-18 8(G).
Weill, R.

1934. Contribution a I'etude des cnidaires et de leurs
nematocystes I-II. Trav. Stn. ZooL Wimereux 10-11:1-701.

878

DUAL STRUCTURAL EQUILIBRIUM IN
THE FLORIDA SHRIMP PROCESSING INDUSTRY

Jose Alvarez, Chris 0. Andrew, and Fred J. Prochaska'

ABSTRACT

Stability, entry, exit, and mobility patterns for six size categories of firms in Florida shrimp processing
industry for the 1959-71 period were studied by utilizing Markov Chain analysis. Forecasts over time
predict that a structural equilibrium in the industry will be achieved by 1985. The forecasted changes in
firm distribution suggest that Florida shrimp industry sales will become increasingly concentrated due
to expansion in number of both small and large firms. A dual equilibrium, resulting in fewer
medium-sized firms and more small- and large-sized firms, can be explained by the tendency for small
firms to develop a specialty product and/or services in order to differentiate their markets from those of
the very large firms. Medium-sized firms, then, tend to expand in size, or decline and either move to
specialty products and services or exit from the industry.

Structural characteristics and patterns of Florida shrimp processing firms over the 1959-71 period,
and the forecasts reveal several important structural characteristics of the industry. Entry into the
Florida shrimp processing industry is relatively easy for small firms and more diflScult for large firms.
All firms are likely to move up in size by one only step or size category per time period. Exit from the
industry in one time period is less probable for small and large firms than for medium-sized firms. Large
firms are most likely to maintain their size between any two time periods and also experience less
probability of declining in size than do medium- and small-sized firms.

Shrimp are the most important seafood processed
in Florida. Total value of the shrimp processed in
Florida in 1972 was slightly over $88 million.
Processed shrimp products account for approx-
imately 69% of Florida's total volume of nonin-
dustrial seafood products and 70% of the value of
seafood processed. In 1972, Florida's share of
processed shrimp production in the southeast
region was 28% (the southeast region representing
about 75% of U.S. production). The growth of this
industry was substantial during the last decade;
both Florida's production and share of the U.S.
market increased (Alvarez 1974).

Despite the growth in processing experienced
by this industry, shrimp landings in the State
declined significantly during the 1960-73 period. In
1960, 51 million pounds (23 million kg) of shrimp
were landed; however, by 1973, landings declined
to only 20 million pounds (13 million kg). Current-
ly, the volume of shrimp processed in the State is
three times as large as the volume of landings in
the State, with the deficit being met by imports
and non-Florida U.S. landings (Alvarez 1974).
These comparisons indicate the basis of concern
for the growth potential and nature of competition

'Food and Resource Economics Department, Florida Sea
Grant Program, University of Florida, Gainesville, FL 32611.

within the Florida shrimp processing industry. In
a recent study addressing this concern (Alvarez
1974), emphasis was placed on processor sales
concentration since there was evidence of "market
power" in raw product purchases. The present
study corroborates the findings of that study and
further explains the results.

Predictions regarding future economic rela-
tionships are important to this industry for cur-
rent managerial and investment decisions by
firms and for long-run planning in optimizing firm
size, scale economies, and product lines. Knowl-
edge of the estimated number and size distribu-
tion of firms in the future will also help predict the
character and intensity of competition within the
market. Markov Chain analysis, employed in this
study, is a useful tool for making such predictions.
The analysis is a discrete-time stochastic process
for which the state of the process at any time k
depends only on the state of the process at the
immediately previous time k -\. A Markov Chain
is described by listing the states of the chain, the
initial probabilities of being in various states, and
the probabilities of transition from one state to
another (Bishir and Drewes 1970).

The purpose of this paper is to analyze by size
category the entry and exit patterns of firms in
the Florida shrimp processing industry during the

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

879

FISHERY BULLETIN: VOL. 74, NO. 4

1959-71 period. The prevailing entry and exit
patterns during the 1959-71 period are then used to
forecast firm distribution over time and predict
the equilibrium state of firms within the market.
Results from a 1973 survey (Alvarez 1974) of the
Florida shrimp processing industry are utilized in
discussing the economic and managerial implica-
tions of entry and exit patterns identified in this
analysis.

This study only considers shrimp processing
firms and not handlers who deal exclusively with
raw headless shrimp. Shrimp processors cook, peel
and devein, and bread or prepare specialty shrimp
products.

THEORETICAL CONSIDERATIONS

That market structure of an industry, according
to Bain (1968), embodies the framework or condi-
tioning environment within which specific enter-
prise behavioral characteristics evolve. This
behavior encompasses both the market conduct
and the market performance of firms. These
conditions in turn influence the type of structural
equilibrium achieved within an industry. The
following brief paragraph discusses the market
structure theory relevant to this paper.

Market structure is defined as ". . . those char-
acteristics of the organization of a market which
seem to influence strategically the nature of
competition and pricing within the market" (Bain
1968). The number and size distribution of sellers,
the conditions of entry, exit, and mobility within
the industry are important aspects of market
structure to be considered. The number of sellers
specifies how many firms are competing for the
buyer's dollar. Generally, an increase in the
number of competing firms is indicative of a
movement toward freer competition (Ward and
Smoleny 1973). The size distribution of firms is
generally measured by volume of sales or by the
proportion of total output of the industry supplied
by a firm or a group of firms. Conditions of entry
are defined as the relative easy or difficulty with
which new firms may enter the market, deter-
mined generally by the advantage or control which
established firms exercise over potential entrants
(Bain 1968). Mobility gives an indication of the
ability for firms within an industry to make
adjustments in their size and, therefore, is an
indicator of the degree of structural rigidity
within an industry (Ward and Smoleny 1973).
Structural equilibrium is that point where net

changes are no longer shown in the market struc-
ture. The number and distribution of firms remain
fixed. Firm entry and exit occur at offsetting rates
(Ward and Smoleny 1973).

ENTRY AND EXIT PATTERNS
DURING THE 1959-71 PERIOD

Lack of time series data for total sales by
individual firms necessitated use of employment
data during the 1959-71 period as a measure of
firm size (Florida State Chamber of Commerce
1959-71). A comprehensive research project based
on a 1973 survey conducted by the authors showed
that firm size measured by employment compared
favorably with sales or volume as a measure of
firm size (Alvarez 1974). Productivity per worker
for firms with similar product lines (95% of in-
dustry sales) is quite similar to further corroborate
this conclusion. Thus, employment is a good proxy
for firm size in the shrimp industry.

The Florida shrimp processing industry is com-
posed of several firms, each of a given size. The
measurement of size as well as size categories
(states of nature) are defined in this study as
follows:

Size offi rm

'ate

{employees)

Sales classification, 1972

—

1- 10

small

<$2 million

11- 30

small

<$2 million

31-100

medium

$2-12 million

101-300

medium

$2-12 million

>300

large

>$20 million

Thirty-one firms processed shrimp in Florida
during the 1959-71 period. These firms and their
respective states of nature throughout the entire
period are presented in matrix form (Table 1) in
2-yr intervals because the data are only reported
biannually. Rows in the matrix specify the
different states of nature for each firm during the
period under consideration. Firm number 2, e.g.,
with state 1 in 1959 and 1961 was not in business,
then in 1963 entered the industry in state 6 (firm
size of over 300 employees), maintained that size in
1965 and 1967, and exited from the industry in
1969. From the data contained in the matrix of
Table 1, the transition matrix presented in Table 2
was calculated.

The probabilities on the transition matrix illus-
trate the stability (diagonal), entry (row one), exit
(column one), and mobility (off diagonal) patterns

880

ALVAREZ ET AL.: DUAL STRUCTURAL EQUILIBRIUM

Table 1. -Total number of Florida shrimp processing firms and
their respective states of nature' during the 1959-71 period.

Firm no.

1959

1961

1963

1965

1967

1969

1971

'For a definition of state of nature utilized in this study see text.

Table 2.-Transition matrix of the Florida shrimp processing

industry.

Employees

States of nature

(number)

0.8025

0.0617

0.0370

0.0123

0.0247

1- 10

.1053

.7895

.1053

—

11- 30

.1351

—

.7838

.0811

—

31-100

.1667

.0556

.2222

.5000

.0556

—

101-300

.1667

—

.7222

.1111

>300

.0769

—

.0769

.8462

to delineate the structure of the Florida shrimp
processing industry during the 1959-71 period.
Each entry (P,,) in Table 2 represents the
probability of a firm moving from state i (row) to
state j (column); e.g., P34 (0.0811) is the probability
of a firm increasing in size from state 3 to state 4 in
the next time period, and P42 (0.0556) is the
probability of a firm decreasing in size from state 4
to state 2 in the following time period.

Industry stability, the probability of a firm
maintaining the same size between any two suc-
cessive time periods, is represented by numbers on
the diagonal. The highest probabilities in the
transition matrix are for shrimp processing firms
to maintain the same size between any two time
periods, suggesting that the industry is fairly
stable. Firms of the largest size (state 6) are most

likely to maintain their size. Medium-sized firms in
state 4 are least stable, illustrated by an equal
probability of remaining in the same size category
or changing between any two periods.

Firm entry, specified in row one, is most proba-
ble for the smaller sizes (0.0167 for sizes 2 and 3)
while the probabilities decrease for larger sizes.

Firm exit probabilities, shown in column one,
are lowest for the largest and smallest firms.

Firm mobility, measured by increases or
decreases in firm size, is shown by the off-diagonal
numbers in the transition matrix. Shrimp
processing firms of any size have at least some
probability of moving one state upward at a time
but almost zero probability of increasing in size by
more than one state at a time. Moving downward
in size scale is somewhat different. The largest
firms (state 6) have a small probability of going
from state 6 to 5, and zero of moving more than
one state at a time. The second largest firms (state
5) have zero probability of moving down possibly
because state 4 is not stable for various economic
reasons. There are probabilities of declines by one
or two states for firms of size 4 but a zero
probability of decline from state 3 to state 2.

CHARACTERISTICS OF
THE DUAL EQUILIBRIUM

Several important implications for the structure
of the Florida shrimp processing industry can be
drawn from the above description of the transition
matrix, for the 1959-71 period. A dual equilibrium,
created by instability of medium-sized firms and
greater stability of small and large firms, is
evident in the industry. Medium-sized firms are
least stable as shown by the highest probabilities
for either exiting from the industry or increasing
or decreasing in size, and the highest probabilities
for moving down more than two states in any time
period. The dual equilibrium, with most stability
for firms with less than 30 and for firms with 300 or
more employees, is the result of a special charac-
teristic of the Florida shrimp processing industry.

The largest firms may be able to exert some
"market power" for a number of reasons. To be
competitive, firms desiring to sell a general line of
slirimp products must be sufficiently large to
achieve the economies of scale in purchasing and
processing presently experienced by large firms.
Even though entry into the largest size is difficult,
exit from that size in one time period is very
unlikely. Size characteristics along with the high

881

probability of remaining in the largest state for a
long period of time permit large firms to be more
secure and ultimately more stable than small
firms. Thus, large firms develop greater access to
raw supply sources which are currently scarce, and
greater knowledge of the national market accom-
panied by stability in supplying their customers.

Small firms, being able to enter with relative
ease, find it very difficult to advance in size but
remain in their state without too much difficulty.
These firms are more likely to succeed if they
produce specialty products, sell in isolated mar-
kets, or develop forward integration from shrimp
fishing operations.

Firms of medium size, neither displaying the
characteristics of large nor small firms, either exit
from the industry or make adjustments in their
size and/or product lines. Medium-sized firms tend
to be unstable initially because they apparently
are not organized to successfully enter shrimp
specialty markets yet are too small to compete in
the national major line shrimp markets.

FORECASTING FIRM

DISTRIBUTION AND PREDICTING

A STRUCTUAL EQUILIBRIUM

A forecast- for the 1961-71 period of the number
of shrimp processing firms in each state of nature
was conducted and compared with the actual
number of firms appearing in the data during the
same period (Table 3). The purpose of this
procedure was to evaluate the appropriateness of
the transition matrix for forecasting firm dis-
tribution within the industry.^ When comparing
actual firm numbers to predicted numbers in
states 2 through 6 for 1961 through 1971, 17 of the
30 predictions were accurate and in state 4, which
is least stable, 5 of the 6 predictions were accurate,
giving confidence that the dual equilibrium struc-
ture remains intact with the predicted numbers.

-To forecast firm distribution in the Florida shrimp processing
industry over time requires that the transition matrix be
stationary; that is, the probabilities in the transition matrix do
not change over time. Although, the chi-square "goodness-of-fit"
test was conducted and the results show the transition matrix to
be stationary, predictions should be considered tentative due to
the small number of obsen'ations per cell caused by the low
number of firms in the industry. Forecasted distribution,
however, being very close to that found in the past, indicates that
the transition matrix remains useful for prediction.

^Some of the differences may be due to the small number of
observations or to rounding procedures.

FISHERY BULLETIN: VOL. 74. NO. 4

Forecasting Firm Distribution Over Time

The biannual forecasted distribution of firm size
for the Florida shrimp processing industry from
1973 to equilibrium appears in Table 3. Few
changes in the number of firms in each state are
observed. The smallest sizes (states 2 and 3)
experienced an increase of one firm each while the
remainder (states 4, 5, and 6) show no change.
Thus, there is an increase of two in the total
number of active firms. The number and size of
firms in the industry will attain a structural
equilibrium in a relatively short period of time.

Equilibrium State Within the Market

The equilibrium matrix for the Florida shrimp
processing industry was calculated to show the
final distribution of firms within the industry
under the assumption of a stationary transition
matrix (Derman et al. 1973). In equilibrium, firms
may still enter and exit but neither the number of
firms in each state of nature nor the total number
of firms in the industry changes once the equilib-
rium is reached.

The distribution of firms in the equilibrium
state compared with the distribution of firms

Table 3.-Actual number' of firms in each state of nature in the
Florida shrimp processing industry, compared with the corre-
sponding predicted numbers- using the transition matrix,
1959-71 and forecasting to 1985 and equilibrium.

States

of nature

Total no

Year

active firms

1959

1961

1963

1965

1967

1969

1971

1973

1975

1977

1979

1981

1983

1985

Equilibrium^

'Data from source (Florida State Chamber of Commerce).

^Computed using the transition matrix.

^The equilibrium probabilities of transition in column order for
the six states of nature were one (0.3881), two (0.1319), three
(0.2454), four (0.0685), five (0.0603), and six (0.1058) for each of
the six respective columns.

882

ALVAREZ ET AL.: DUAL STRUCTURAL EQUILIBRIUM

during the 1959-71 period (Table 3) shows that
firms of the smaller sizes (states 2 and 3) increase
in number as the industry reaches the structural
equilibrium, while firms in states 4 and 5 decrease
and firms in the largest size increase in number.
This is a consequence of the industry dual equilib-
rium conditions of entry and exit identified in the
1959-71 period.

At the structural equilibrium, and in support of
the dual equilibrium, the probabilities for firm
entry are highest for firms with less than 30
employees and for those with more than 300
employees. Thus, the least amount of entry ac-
tivity will occur within the medium-sized firms.

Mean Lifetime for Each Size Category

Mean lifetime values for each size category were
calculated (Table 4) and further support the
prevalence of a dual equilibrium in the Florida
shrimp industry. Mean lifetime represents the

Table 4. -Mean lifetime in years for each size category for the
Florida shrimp processing industry.'

Column 1

Column 2

States of

Average^

Perfect*

Column 3

nature^

(yr)

Ratios

9.500

2.304

0.243

9.250

2.650

.286

4.000

2.418

.604

7.200

2.128

.296

13.000

2.236

.172

'Mean lifetime represents the number of years a firm tends to
stay in a given size category. In this case, results were multiplied
by 2 since each time period equals 2 (yr) in the data.

^State 1 is not included because it is an absorbing state.

'Calculated from the transition matrix with the formula

(1/1 - P,,)-

■•Time spent in each state for a perfectly mobile industry as cal-
culated from the equilibrium size distribution.

sColumn 2 h- Column 1.

number of years a firm tends to stay in a given
state of nature. The largest firms tend to maintain
their size for a greater number of years (13) than
firms in any other size category. Firms of sizes 2
and 3 have mean lifetime values of 9 yr, while
firms of size 4 and size 5 tend to remain for an
average of 4 and 7 yr in their respective states.
These findings are the result of the firms'
probabilities of maintaining their size between
any two time periods. Column 2 of Table 4 repre-
sents the number of years spent in each size for an
equilibrium distribution (perfectly mobile in-
dustry); the values are very similar. The data in
Column 3 indicate state rigidity where the smaller
the ratio, the more rigid the state. State 6 is the
most rigid state in the industry, followed by states
2, 3, 5, and 4, respectively.

LITERATURE CITED

Alvarez, J.

1974. The Florida shrimp processing industry: Economic
structure and marketing channels. M.S. Thesis, Univ.
Florida, Gainesville, 168 p.
Bain, J. S.

1968. Industrial organization. John Wiley and Sons, Inc.,
N.Y., 678 p.

BiSHIR, J. W., AND D. W. DrEWES.

1970. Mathematics in the behavioral and social sciences.
Harcourt, Brace & World, Inc., N.Y., 714 p.
Derman, C, J. G. Leon, and 0. Ingram.

1973. A guide to probability and application. Holt, Rine-
hart, and Winston, Inc., N.Y., 750 p.
Florida State Chamber of Commerce.

1959-71. Directory of Florida industries. Florida State
Chamber of Commerce, Jacksonville, Biannual Issues.
Ward, R. W., and C. Smoleny.

1973. The market structure of Florida fresh grapefruit
packers: An application of Markov Chain analyses. Econ.
Res. Dep. 73-1. Fla. Dep. Citrus and Univ. Fla, Gainesville,
91 p.

883

DISTRIBUTION AND ECOLOGY OF PELAGIC FISHES

STUDIED FROM EGGS AND LARVAE IN

AN UPWELLING AREA OFF SPANISH SAHARA

Maurice Blackburn^ and Walter Nellen-

ABSTRACT

Fish eggs and larvae were taken in vertical zooplankton hauls in a small upwelling area off Spanish
Sahara. Series of hauls were made repetitively from March to May 1974, sometimes with accompanying
hydrocasts. About 58% of the eggs and 72% of the larvae belonged to the following pelagic species:
Sardina pilchardus, Engraulis encrasicholus, Trachurus spp., and Maurolicus sp. It was estimated
from contemporaneous current meter data and other information that the eggs of those species were
spawned very close in time and space to where they were collected. Thus adult Sardina and Engraulis
appeared to occur typically on the continental shelf, adult Trachurus at the edge of the shelf, and adult
Maurolicus over the continental slope. These distributions were verified for Sardina and Trachurus
from fishing results of Polish vessels. Acoustically detected concentrations of fish were identified by
species according to those results.

The area of abundance of Sardina was characterized by ma.xima of phytoplankton and small
zooplankton. Abundance of Sardina eggs changed with time, because of variations in the size of the
adult population in the area (acoustically estimated) and in its production of eggs. The major change in
population size coincided with a similar change in the amount of food, especially phytoplankton,
available. Variations in egg production may have been associated with the mean temperature in the
water column, since eggs were scarce when the mean was below 16.5°C even when adults were
abundant.

A multidisciplinary group of U.S. scientists made
an oceanographic study off Spanish Sahara from
March through May 1974. The program is called
Coastal Upwelling Ecosystems Analysis (CUEA)
and is part of the International Decade of Ocean
Exploration (IDOE). The operation off Spanish
Sahara (Figure 1) was called JOINT-I. It made
observations of many kinds over an upwelling area
which was small enough to be studied synoptically
in great detail repetitively under various condi-
tions such as changes in the wind field. Most of the
work was done from the coast to long. 18°00'W,
between lat. 21°30' and 21°50'N. The continental
shelf in this area is bounded by the 100-m isobath,
beyond which there is a steep slope (Figures 2-4).
Pelagic fish are a major component of the animal
biomass in the area. They support large fisheries
conducted by several nations. It was the task of a
small group of CUEA investigators to estimate
biomass of pelagic fishes by species and, if possi-
ble, by trophic levels during JOINT-I; to show the
distributions of these biomasses in space and time;

'Institute of Marine Resources, University of California, La
Jolla, CA 92093.

-Institut fiir Meereskunde, Universitat Kiel, Kiel, West
Germany.

Figure 1.

20"

-Part of northwest Africa showing the principal area
of JOINT-I work.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

885

FISHERY BULLETIN: VOL. 74, NO. 4

and to explain the distributions in terms of envi-
ronmental parameters. Biomass of total pelagic
fish wsLS estimated acoustically (Thorne et al. in
press). Partitioning it by species was to be based
on the following: contemporaneous catches by
fishing or fishery research vessels, samples of fish
taken by the CUEA ships, fish eggs and larvae
from the zooplankton catches of the CUEA ships,
and the literature. In the outcome, only the fish
eggs and larvae (ichthyoplankton) were useful
during the cruise. Good information on fish catches
by other vessels was not received until many
months later, sampling from the CUEA ships was
unproductive for adults of e'pipelagic species, and
the literature did not resolve all questions. The
ichthyoplankton results and the fish catches
agreed as to the principal species present in
different parts of the area. Acoustically detected
concentrations of fish (Thorne et al. in press) were
identified accordingly.

This paper gives the principal results of work on
the eggs and larvae. It then uses the egg distribu-
tions to estimate contemporaneous distributions
of adults of some species and compares those with
data from contemporaneous fish catches and the
literature. Finally the paper attempts to explain
the distributions of an abundant species, Sardina
pilchardus (Walbaum), according to environmen-
tal data collected at the same time as the eggs.

MATERIAL AND METHODS

Zooplankton

The fish eggs and larvae were sorted from the
zooplankton catches made during JOINT-I and
partly identified by Blackburn. Most of the
identifications were made later by Nellen. The
zooplankton catches were made and processed,
apart from the ichthyoplankton, by R. I. Clutter.
Some observations on the zooplankton in general
are relevant in this study. A more complete report
on JOINT-I zooplankton will appear elsewhere.

The net hauls for zooplankton were m^ade ver-
tically from 200 m or the bottom, whichever was
less, to the sea surface. Two cylindro-conical,
nonclosing Bongo plankton nets mounted side by
side were used. Each net had a mouth diameter of
60 cm and a uniform mesh size of 102 /xm. Nets
were lowered at 40 m/min and hauled up at 60
m/min. A calibrated digital flowmeter was
mounted in the mouth of each net. Volume of
water filtered by the two nets ranged from 12 to

158 m-^, depending mainly upon the haul length.
Only one net was used in series 1 and 2 (Table 1).

Processing was as follows, with exceptions
shown in footnotes to Table 1. The catches from
the two nets were immediately combined and
suspended in water. The suspension was shaken
and four V4-aliquots were decanted. Each of two
aliquots was filtered through a series of sieves
(mesh sizes 1,050, 505, 223, and 102 jum) until no
more water dripped. This procedure yielded sub-
samples of zooplankton in four size ranges, ap-
proximately 100 to 200, 200 to 500, 500 to 1,000, and
> 1,000 jum. The subsamples from one aliquot were
scraped from the filters, blotted on paper towels
until no more water appeared, and weighed. The
subsamples from the other aliquot were washed off
the filters and preserved in Formalin.-^ The fish
eggs and larvae were sorted from the preserved
500- to 1,000- and >l,000-jLim samples and com-
bined. The four wet weights per haul were stan-
dardized in grams under 1 m- of sea surface.
Allowance was made trigonometrically for effects

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

23-24 MARCH 1974 (ALONG ~2I''40'N)

SARDINE EGGS/m2

1808

ANCHOVY EGGS/m2 a

_o 0 Q

ZOOPLANKTON 100-500^ g/m2

B ■ B Bfi

ZOOPLANKTON >500/i g/m2

_S.

1000
-800

600

400

200

0
-100
- 50

.^_

CHLOROPHYLL mg/m2

300
-200
- 100

200

W I7''40'

17" 30

I7">20

I7»I0

ir-oo

Figure 2.— Distribution of sardine eggs, anchovy eggs, and
environmental parameters along lat. 21°40'N on 23-24 March
1974 (series 5 in Table 1).

886

BLACKBURN and NELLEN: EGGS AND LARVAE IN AN UPWELLING AREA

Table 1. -Means of variables for the water column at stations from long. 17°08' to 17°25'W, in series of stations
along lat. 21°40'N together with indications of relative abundance of adult sardines explained in Discussion.

Small

Abundance

Date

No. cf

Sardine

Anchovy

Temp.

Chlorophyll

zooplankton

of adult

Series

1974

stations

eggs/m^

eggs/m2

(mg/m^)

(g/m^)

sardines

8- 9 Mar.

16.5

115

285

Low

10-11 Mar.

16.5

290

Low

15-17 Mar.

16.0

193

578

C")

18 Mar.

16.0

527

(')

23-24 Mar.

648

195

16.5

164

High

1- 2 Apr.

16.5

Low

5 Apr.

15.5

187

Medium

12-13 Apr.

17.0

147

Medium

22-23 Apr.

16.0

192

Medium

9-10 May

431

16.5

323

High

'Three stations for eggs and zooplankton.

^Estimated from settled volumes at 1 ml = 0.8 g. Not corrected for phytoplankton contamination.
sQne station for chlorophyll.
■•Two stations for eggs and zooplankton.

sEstimated according to mean ratio of small to total zooplankton at same longitudes in other series, namely
67%. Not corrected for phytoplankton contamination.
'Unknown.
Two stations for chlorophyll.

SARDINE EGGS/m2

ANCHOVY EGGS/m2

ZOOPLANKTON I00-500>i g/m2
a B B

ZOOPLANKTON >500^ g/m2

CHLOROPHYLL mg/m2

1000

800

600

400

200

100

SARDINE EGGS/m2

W I7''40'

Figure 3.— Distribution of sardine eggs, anchovy eggs, and
environmental parameters along lat. 21°40'N on 22-23 April 1974
(series 9 in Table 1).

00000

ANCHOVY EGGS /m2

00 0 0 0 0

ZOOPLANKTON 100-500^
^^"' _ - ■ ■ i

_ilu

J_r

ZOOPLANKTON >500^ g/m2

CHLOROPHYLL mg/m2

- 50

1000
800
600

400
200
0
200

-100

-50
0

300
200
100
0
Om

100
200

W I7''40'

I7''30'

I7»20

I?"!©'

i7'00'

Figure 4.-Distribution of sardine eggs, anchovy eggs, and
environmental parameters along lat. 21°40'N on 9-10 May 1974
(series 10 in Table 1).

of nonzero wire angles on distance covered by the
net. To determine effects of clogging, the expected
flow of water through the net was compared with
that indicated by the flowmeter revolutions.
Counts of various kinds of fish eggs and larvae
from each haul were standardized in numbers
under 1 m^ of sea surface.

In general the spatial distribution of zooplank-
ton biomass was similar for the 100- to 200- and
200- to 500-]um fractions. The two fractions of
larger-sized plankton were also distributed
similarly, but not like the smaller-sized fractions.
Thus we distinguish only zooplankton at 100 to
500 jLim and at >500 nm (Figures 2-4). Most of the

887

FISHERY BULLETIN: VOL. 74, NO. 4

biomasses given here, but not all (see Table 1),
have been corrected for contamination by phyto-
plankton. The correction was made as follows. The
amount of chlorophyll a was determined in a
'/4-aliquot by SCOR methods (UNESCO 1966) and
partitioned among the four subsamples according
to inspection of the preserved samples. The in-
spection indicated approximate relative amounts
of phytoplankton in the samples. The chlorophyll
weight for each subsample was converted to
carbon following Lorenzen (1968) and then to wet
weight according to Gushing et al. (1958). The
correction generally reduced the original biomass
by less than 10% but occasionally up to 30%. All
biomasses shown in Figures 2 to 4 have been
corrected.

The preserved samples of zooplankton <500 jum
were not examined for ichthyoplankton, because
few specimens (except some newly hatched larvae)
were expected to pass through a SOO-jnm sieve. For
eggs of Engraulidae, which are oval and measured
from 500 to 580 jum (mean 570 /xm) in transverse
diameter in our material, our numbers per haul
could have been slightly too low because of losses
through the 500-jum sieve. It is unlikely that these
losses were high. During a later cruise
(AUFTRIEB 1975) in the same area, we counted
engraulid eggs in the catches of two Bongo nets of
uniform mesh sizes, 300 and 500 ixm, but otherwise
identical and hauled side by side in the same net
assembly. Egg numbers were 122 and 145, so the
300-/xm net retained no more than the 500-jLim net.

Temperature and Chlorophyll a

These properties were measured from hydro-
graphic casts which used plastic 5-liter Niskin
bottles with reversing thermometers. Sampling
depths in the upper 200 m were usually 0, 3, 10, 20,
30, 50, 75, 100, 150, and 200 m, depending on the
bathymetry. Concentrations of chlorophyll a were
determined by SCOR methods (UNESCO 1966)
and integrated in milligrams per square meter.
The integration program summed the area of each
depth integral using the area formula of a
trapezoid. Samples for chlorophyll a were gener-
ally not taken below 75 or 100 m, because results
of other casts showed little chlorophyll below those
depths.

Area and Periods of Study

Almost all the zooplankton hauls and hydro-
graphic casts of JOINT-I were made in the area

888

shown in Figures 1 and 5. They were generally
made along an east-west line at about lat. 21°40'N,
where series of hauls and casts (not always
together) were frequently repeated. Figure 5A
shows the positions of all zooplankton hauls made
in the area. Nine other hauls were scattered in
space and time in adjacent areas, and are not used
in this paper. No distinction is made here between
day and night hauls. Hauls on the shelf were made
mostly by day and those on the slope mostly at
night. Eggs are of more interest than larvae in this
study as explained above and should have been
equally available by day and night. Larvae might
have avoided the nets more by day than by night.

The total period of JOINT-I in which zooplank-
ton hauls were made was 8 March to 10 May 1974.
It was divided by port calls into three parts. Legs
1, 2, and 3 (Table 2). The periods of these legs (first
to last zooplankton haul) were 8 to 24 March, 1 to 14
April, and 22 April to 10 May.

Ten series of hauls were made together with
hydrographic casts along lat. 21°40'N, each series
occupying 1 to 3 days. Figures 2 to 4 show data for
some of the series and Table 1 summarizes data for
all of them.

Table 2.— Principal categories of fish eggs and larvae taken on
JOINT-I in the area of Figure 5, showing numbers per square
meter averaged for hauls on each leg of the cruise and summed
for the cruise.

Leg 1

Leg 2

Leg 3

Cruise

total

(41

(22

(38

Calegoiy

hauls)

No.

Eggs:

Sardina

77.7

10.9

75.9

6,308

35.1

Engraulis

19.0

16.1

14.8

1,695

9.4

Maurolicus

6.2

29.3

15.4

1,487

8.3

Soleidae

9.9

15.0

8.8

1,071

6.0

Carangidae

4.8

0.7

18.1

897

5.0

Others

55.3

56.1

79.8

6,531

36.2

Larvae:

Clupeoidei

60.5

82.9

84.7

7,522

69.7

Heterosomata

24.1

16.5

5.0

1,541

14.3

Sparidae

6.6

22.6

9.3

1,120

10.4

Maurolicus

1.2

1.3

2.8

185

1.7

Myctophidae

0.9

0.7

1.3

102

0.9

Carangidae

0.4

2.1

0.4

0.7

Others

3.5

2.3

1.5

251

2.3

IDENTIFICATION AND
ENUMERATION OF
EGGS AND LARVAE

The eggs and larvae from all stations in Figure
5A were identifiable in the categories shown in
Table 2. Most of the identifications were made at
the Institut fur Meereskunde from the large collec-
tions, literature, and experience of northwest

BLACKBURN and NELLEN: EGGS AND LARVAE IN AN UPWELLING AREA

I-

LiJ

UJ
(T.
(T

O
O

SiJJiSiKiSiai

E
>^

O
I

3
CO

I-

3 _

tn t.

■rj I*

O cfl

3 E

cfl 3

S *-

Q. °
o ^

£ s
<•"§

O M

lo be

o OJ

c -a

0) >

^ o
- P.

Z c

— m

a. .

o >>

<^ t

ca ^
en O
bO J=
bo "

«4-H

C )

<i;

-1-'

-Q

<i)

■*->

889

FISHERY BULLETIN: VOL. 74, NO. 4

African ichthyoplankton available there.
Identifications of larvae were more complete than
those of eggs, as is usual in work of this kind.

Among the eggs the following kinds, which are
well known in literature because of conspicuous
characters, were easily identified: Sardina, En-
graiilis, Maurolicus, and Soleidae.

The eggs of Sardina and Sardinops are alike but
the only species of either genus recorded off
northwest Africa is Sardina pilchardus
(Walbaum). Sardina pilchardus occurs off south-
western Europe, in the Mediterranean, and on the
coast of northwest Africa as far south as lat. 20°N
(de Buen 1937; Larraneta 1960; Maurin 1968;
Furnestin and Furnestin 1970). We identify the
eggs as that species, which we later call "sardine."
Egg diameters in our material range from 1.33 to
1.50 mm (mean 1.46 mm), slightly lower than those
of the same species in the Mediterranean (1.40 to
1.70 mm; Larraneta 1960). However they are
considerably larger than those of Sardinella, the
other clupeid genus that might occur, whose eggs
measure 1.1 to 1.3 mm off west Africa (Marchal
1967).

Engraulid eggs were easily recognizable by
their oval shape. Two species of Engraulidae have
been reported off southern Spanish Sahara, En-
graulis encrasicholus (Linnaeus) and Anchoa
guineensis (Rossignol and Blache) (Lozano Cabo
1970; Bravo de Laguna Cabrera and Santaella
Alvarez 1973). No adults were obtained during
JOINT-L so identification has been made from the
eggs. Eggs of E. encrasicholus range from 0.90 to
1.9 mm in length and 0.42 to 1.2 mm in maximum
breadth (Demir 1963); corresponding ranges for A.
guineensis are 1.05 to 1.23 and 0.54 to 0.58 mm,
respectively (Marchal 1966), and for our material
1.33 to 1.50 and 0.50 to 0.58 mm, respectively. Our
eggs could belong to either species as far as
breadth is concerned, but only to E. encrasicholus
on the basis of length. We refer to this species later
as "anchovy." It occurs off western Europe and in
the Mediterranean and Black seas, as well as off
northwest Africa, where its southern limit is not
exactly known (de Buen 1931 and references
above).

The eggs of Maurolicus (family Gonostoma-
tidae) are those of M. muelleri (Gmelin), which has
been recorded off southern Morocco and northern
Mauritania (Maurin et al. 1970). The eggs of
Soleidae could belong to several species recorded
off Spanish Sahara (Maurin et al. 1970; Lozano
Cabo 1970).

The carangid eggs were identified with help
from E. H. Ahlstrom, who noted that some of them
resembled Trachurus. They measured about 0.9 to
1.0 mm, in the size range reported for T. trachurus
(Linnaeus) off northwest Africa (Kiliachenkova
1970). Three other species of Trachurus have been
recorded off northwest Africa, namely T. pictura-
tus (Bowdich), T. trecae Cadenat, and T. mediter-
raneus Steindachner. Trachurus picturatus is not
common and T. mediterraneus may be a sub-
species of T. trachurus (Letaconnoux 1951; Mau-
rin et al. 1970; Witzell 1973). The three mostcommon
carangids in the area of Figure 1 are T. trachurus,
T. trecae, and Caranx rhonchus Geoffroy St.
Hilaire. The first two spawn off Spanish Sahara
from about November to April, and C. rhonchus
from about May to August (Boely et al. 1973).
Aboussouan (1967) and Conand and Franqueville
(1973) described larvae of these species. The
distinctions between larvae of Trachurus and C.
rhonchus are slight and the larvae of the two
Trachurus species cannot be distinguished. Most
of our carangid eggs are probably Trachurus
("horse mackerel"), which was abundant along the
coast of Spanish Sahara between March and June
1974. The most likely species is T. trachurus. All
specimens of Trachurus taken in research
trawling during JOINT-I were that species. We
took 22 post-larval and juvenile Trachurus up to 6
cm long in various hauls of a micronekton net
during JOINT-I. All specimens large enough to be
identified were T. trachurus. We identified caran-
gid eggs conservatively and so may have failed to
count some.

The remaining eggs, 36% of the total, were of
several kinds not readily identifiable by us.
Probably few of them were eggs of pelagic species,
except possibly some carangids as suggested
above. They lacked segmented yolks and thus were
probably not Isospondyli. Scomber japonicus
Houttuyn is a pelagic species that spawns mostly
from December to February in the vicinity of Cap
Blanc (references in Blackburn 1975). If Scomber
eggs occurred in our collections, they were prob-
ably not abundant. We found no Scomber larvae.
Other abundant pelagic species of the JOINT-I
area spawn principally in summer (Blackburn
1975). Thus unidentified eggs probably were
mostly demersal species, as were 25% of larvae, i.e.,
Heterosomata and Sparidae, as shown in Table 2.
Spatial distribution of unidentified eggs resem-
bled that of the demersal larvae (Figure 5G, H).

All larvae were identified to some taxon. Closer

890

BLACKBURN and NELLEN: EGGS AND LARVAE IN AN UPWELLING AREA

identifications could have been made in some cases
but were not needed for this study. Clupeoids
predominated. Many clupeoids were small (about 5
to 10 mm) and had lost part of the intestine,
probably because of the repeated filtering of the
zooplankton. Clupeidae and Engraulidae were not
separately counted, but both families were well
represented. Preanal myomeres were counted in
randomly selected good clupeid specimens. These
counts ranged from 41 to 43, which agree with
Sardina pilchardus (Saville 1964). Comparable
ranges in two other west African clupeids, Sar-
dinella aurita Valenciennes and Sardinella eba
(Valenciennes), are respectively 38 to 41 and 35 to
38 (Conand and Fagetti 1971). These species were
looked for because Maigret (1972) found Sardin-
ella larvae near the area of JOINT-I in May.
Evidently they were absent or scarce in our ma-
terial. They were absent or scarce in the 1974 fish
catches reported to us. We conclude that our
clupeoid larvae were Sardina pilchardus and
Engraulis encrasicholus, like the clupeoid eggs.
Carangid larvae were scarce. Larvae in the last
line of Table 2 ("Others") were Merluccius, Cal-
lionymus, Paralepididae, and Anguilliformes
(leptocephali).

Table 2 shows that Sardina dominated the egg
samples. It shows also that abundance of Sardina
eggs varied greatly during JOINT-I, which is
discussed later.

SPATIAL DISTRIBUTION OF
EGGS AND LARVAE

Figure 5B-H shows distribution and abundance
of the principal kinds of eggs and larvae identified,
during the whole period of cruise JOINT-I. All
positive hauls for each kind were charted and the
observed numbers per square meter were con-
toured without averaging. The purpose of Figure 5
is to show where maxima and minima occurred,
although some of them were more prominent at
those locations on some legs of the cruise than on
others. For example the midshelf maximum of
Sardina eggs was not prominent on Leg 2, when
eggs were scarce everywhere (cf. Tables 1, 2). We
were most interested in the pelagic species and
especially in their eggs, whose distributions should
be close to those of the adults. Furthermore, the
methods employed were more suitable for eggs
than larvae. Some larvae could have avoided the
nets, especially in daytime.

Sardine and anchovy eggs were absent close

inshore, most abundant on the continental shelf
between the 50- and 100-m isobaths, and occasion-
ally found just beyond the shelf edge (Figure 5B,
C). These eggs occur most abundantly in the
uppermost 25 m of the water column (Furnestin
and Furnestin 1959; Larraneta 1960; Demir 1963),
where temperatures on JOINT-I were about 16° to
17°C (Figures 2-4). The eggs take about 3 days to
hatch at such temperatures (Larraneta 1960;
Demir 1963), so their average age should be about
1.5 days.

Six vertical arrays of current meters were
moored during JOINT-I (Figure 5A). No ichthyo-
plankton were collected near array number 6. The
other arrays operated for periods of about 20 days
(number 3) to 60 days (number 2). Means of the
meridional and zonal components of water
movement, v and u, are available for each current
meter during the period of operation (Pillsbury et
al. 1974). The top meter in each array was about 20
m below the surface. At this depth, mean v was
about 20 cm/s on the continental shelf (arrays 1
and 2) and 10 cm/s on the edge and slope (arrays 3,
4, and 5), towards the south. Mean u was about 2
cm/s towards the west, except at array 3 where it
had the same velocity towards the east. Thus, from
where it was spawned by the parent, a sardine or
anchovy egg of average age on the continental
shelf could have drifted about 14 nautical miles to
the south and 1.4 miles to the west. The movement
to the west is negligible for our purpose. The
coastline and isobaths run generally north and
south along this section of the coast, as do isopleths
of surface temperature and surface nitrate con-
centration (Voituriez et al. 1974; D. W. Stuart and
J. J. Walsh, pers. commun.). Thus the parent fish
probably occurred over the same bathymetry and
under the same environmental conditions as the
eggs did, but slightly farther north.

Carangid eggs (Figure 5D) were found on the
outer half of the shelf, especially at the edge.
Kiliachenkova (1970) found eggs of Trachurus
trachurus distributed in exactly the same way in
the same area in November, December, and May.
The literature does not clearly show the vertical
distribution of the eggs of T. trachurus. Kilia-
chenkova (1970) found them abundant at the
surface. The eggs of the related T. symmetricus in
the California Current are most common at the
surface but fairly abundant down to 30 m, with
smaller numbers occurring deeper (Ahlstrom
1959). We, therefore, assume our eggs came mostly
from the top 30 m. Trachurus trachurus eggs

891

FISHERY BULLETIN: VOL, 74, NO. 4

hatch 3 or 4 days after being spawned at temper-
atures from 15° to 19°C (Letaconnoux 1951), so
average age in our material should be 1.5 to 2 days.
Then, taking mean v as 10 cm/s we estimate that a
TrachurHs egg collected near the shelf edge was
probably spawned near the edge about 7 to 10
miles farther north.

Maurolicua eggs (Figure 5E) were most abun-
dant just outside the shelf edge. Adults are meso-
pelagic fish of the continental slope (Maurin et
al. 1970; Hureau and Tortonese 1973) and pre-
sumably spawn there. We frequently found eggs
on the outer one-third of the shelf as well as on the
slope, which suggests some eastward transport.
The current meter data from arrays 3 and 4 show a
mean u about 10 cm/s to the east at 60 m. This
could account for the observed distribution if
MauroiicKs eggs occur at that depth and hatch in a
few days. Eggs of M.japonicus off Japan are most
abundant at 50 to 60 m (Nishimura 1957). This
species is considered synonymous with M. muel-
leri (Hureau and Tortonese 1973).

Clupeoid larvae (Figure 5F) were abundant at
midshelf , on the outer shelf, and over the slope. In
general their distribution extended about 10 to 15
miles west of the eggs. Their average age probably
was 10 to 20 days more than that of the eggs.
Larvae of Sardina pilchardus and Engraulis
encrasicholus occur most commonly in the upper
25 m (Fage 1920). Thus the movement of 20-m
shelf water towards the west at about 0.9 nautical
mile/day generally explains the observed larval
distribution. This water movement is presumably
the Ekman transport, which provides a mechanism
for the coastal upwelling.

Larvae of demersal fish (flatfish and sparids)
occurred mostly on the shelf as expected, but
occasionally on the slope. They were most common
in inshore waters where eggs and larvae of pelagic
species were scarce (Figure 5G, H).

VERIFICATION FROM
COMMERCIAL FISH CATCHES

From egg and larval evidence, the adult pelagic
fishes in the area and period of JOINT-I should
have been predominantly S. pilchardus and E.
encrasicholus, especially the former, on the shelf;
Trachunis, probably T. trachurus, at the shelf
edge; and the mesopelagic M. muelleri, on the
continental slope. Difi'erences in fecundity
between species could affect these findings, how-
ever, and other species could have been present but

892

not spawning. Commercial fish catches provide a
useful check on the results of the studies with eggs
and larvae. Some useful information of that type
was kindly provided by the Sea Fisheries Institute
of Gdynia, Poland.

Polish pelagic (mid-water) trawlers of the Odra
Deep Sea Fishing Company fished just south of the
JOINT-I area at the end of March 1974. They
operated from lat. 20°40' to 21°00'N, between the
coast and shelf edge. Reported catches (tons/day)
of pelagic species were about 3.3 Trachurus spp.,
6.5 Caranx rhonchus, and 0.2 Scomber japonicus.
Caranx rhonchus was the principal species within
the 50-m isobath, Trachurus the principal fish in
more offshore waters. During April, the trawlers
were located far north of the JOINT-I area
between lat. 23° and 27° N, where their catches
were predominantly Sardina pilchardus.

The Polish research vessel Professor Siedlecki,
equipped for large-scale pelagic trawling, made 77
hauls between 13 May and 24 June, starting just
after JOINT-I. The hauls were made between lat.
20°16' and 25°01'N which includes the area of
JOINT-I. Hauls north of lat. 21°00' were all on the
continental shelf between the 35- and 70-m
isobaths and caught almost exclusively Sardina.
Hauls south of lat. 21°00' were made at the shelf
edge (100-m isobath) and caught almost exclu-
sively Trachurus or Sardina, usually Trachurus.
Klimaj (1971, 1973) summarized results of com-
mercial Polish trawling from 1965 to 1971 in a
small area (his area 22) which includes the area of
JOINT-I. The principal pelagic fishes taken from
March to May were Trachurus spp., Caranx
rhonchus, Scomber japonicus, and Pomatomus
saltatrix. Caranx rhonchus was common only in
March and P. saltatrix only in May. The other two
were important in all months, with Trachurus
generally much more abundant than Scomber. The
Trachurus would have been either T. trachurus or
T. trecae, which are not distinguished in the Polish
fishery.

It was noted earlier that the principal spawning
seasons of Caranx and Scomber are respectively
later and earlier than the period of JOINT-I. The
spawning season of Pomatomus is also later
(references in Blackburn 1975). Thus these forms
could have occurred in the area and period of
JOINT-I although we did not recognize them in
the ichthyoplankton. Caranx rhonchus probably
did occur in March, especially inshore, and S.
japonicus may have occurred, although not in
great abundance.

BLACKBURN and NELLEN: EGGS AND LARVAE IN AN UPWELLING AREA

The Polish data support our conclusion that
Trachurus was the principal pelagic fish at the
edge of the shelf. Our conclusion that Sardina
pilchardus was an important species on the shelf
is supported by the results of the Professor Sied-
lecki hauls, but not by those from the commercial
vessels. Commercial fishing for that species is
concentrated farther north, especially between
lat. 24° and 26°N (Chabanne and Elwertowski
1973; Odra Company results given above). Sardine
catches of the Professor Siedlecki were much
higher between lat. 22° and 25°N (mean of 62
hauls, 2.37 tons/h) than between lat. 20° and 22°N
(mean of 15 hauls, 0.17 ton/h). There appears to be
no commercial fishing for Engranlis off Spanish
Sahara.

SPATIAL AND TEMPORAL

DISTRIBUTION OF

SARDINE AND ANCHOVY EGGS

In this section we characterize the area in which
sardine and anchovy eggs occurred on JOINT-I,
and note temporal changes in their abundance.
The findings on areal distribution would apply also
to adult fish in reproductive condition. We have
assembled data on temperature, chlorophyll a,
small zooplankton (<500 jum), large zooplankton
(>500 jum), sardine eggs, and anchovy eggs for the
10 series along lat. 21°40'N. Figures 2 to 4 show the
data for three series, including the two series in
which sardine eggs were most abundant. Anchovy
eggs were most abundant in the 23-24 March
series (Figure 2).

Vertical distributions of temperature and den-
sity varied as shown by Barton (1974) and L. A.
Codispoti (pers. commun.), and are not discussed in
detail. Figure 3 shows typical coastal upwelling
and Figure 4 a relaxation of upwelling conditions.
Figure 2 shows weak coastal upwelling and up-
welling at the shelf edge. Other series showed
similar variations. It is doubtful if upwelling ever
occurred only at the edge.

Chlorophyll a in the water column always
showed a primary or secondary maximum on the
middle or outer part of the shelf, and sometimes
another maximum over the slope. The maximum
over the slope was found when upwelling occurred
at the edge, as in Figure 2, and was probably a
result of it. Maxima of small zooplankton were
distributed like those of chlorophyll. Both chloro-
phyll and small zooplankton were relatively low,
close inshore in all series, and also beyond the shelf

edge in series where second maxima did not occur.
Large zooplankton were relatively scarce on the
shelf in each series. Their biomass increased
sharply at the edge, and generally continued high
as far offshore as we sampled.

Sardine and anchovy eggs were virtually
confined to the middle and outer parts of the shelf
on all series, regardless of their abundance. Their
mean abundance there is given in Table 1,
together with means of temperature, chlorophyll,
and small zooplankton for the water column in the
same area, for each series. Temperature means are
approximate.

DISCUSSION

Sardine eggs were most abundant on the middle
and outer continental shelf during haul series 5
and 10, moderately abundant during series 8, and
scarce on other series (Table 1). Figures 2 to 4 show
the abundance on series 5, 9, and 10. Low numbers
of eggs indicate either a small population of adults
in the vicinity, or one that is spawning little. Mean
biomass of adult fish was estimated acoustically
for the same part of the shelf on the same sam-
pling line, at various dates commencing 31 March
(Thorne et al. in press). This biomass showed an
irregular increase with time. It was about 8 g/m^
on 31 March, 40 g/m- on 6 to 9 April and 22 to 26
April, and 80 g/m^ on 1 to 6 May. These four
periods were close in time to series 6, 7, 9, and 10,
respectively. The predominant species was proba-
bly sardine as stated earlier. The egg numbers
show that adult sardines were probably abundant
on series 5 and moderately so on series 8, but we
have no acoustic estimates of biomass for those
series or for series 1 to 4.

The low mean egg number on series 6 probably
reflected a very small adult population, but it is
unlikely that the low numbers on series 7 and 9 did
so, in view of the biomass estimates just given. It
is more probable that sardine spawning was
inhibited during series 7 and 9. The low mean
temperatures in the water column during those
series, namely, 15.5° and 16.0°C (Table 1), could
have been responsible. Furnestin and Furnestin
(1959, 1970) stated that spawning of Sardina is
absent or feeble below 15.5°C and optimal from
16.0° to 18.0°C, especially over 16.5°C, in Moroccan
waters. Spawning might, therefore, be low at 15.5°
to 16.0°C in waters off Spanish Sahara. The limit-
ing effect of temperature appears to be not on the
spawned eggs, which can develop at 10°C

893

FISHERY BULLETIN: VOL. 74, NO. 4

(Larrafieta 1960), but on the adults, as to whether
or not they release eggs. The adults occur in most
parts of the water column (Furnestin and Fur-
nestin 1970; Thorne et al. in press), which is the
reason for considering mean water temperature
here. Furnestin and Furnestin (1970) make it clear
that spawning depends on the temperatures over
most of the water column, not necessarily on those
in the upper 25 m where most eggs are found.
Thick layers of water below 15.5°C make an area
unsuitable for sardine spawning even if there is
warm water at the surface, according to those
authors. Figure 3 shows such a situation for series
9. From the criteria of Furnestin and Furnestin
and the vertical distributions of temperature in
our 10 series (examples given in Figures 2-4), it
can be said that temperature conditions on series
3, 4, 7, and 9 were unsuitable for sardine spawning
on the middle and outer shelf. Conditions on the
other series were relatively suitable with mean
temperatures for the water column at 16.5° or
17.0°C. It can then be deduced that adult sardines
were scarce on series 1 and 2, because few eggs
were found. We have no information about rela-
tive abundance of adults on series 3 and 4; they
could have been present but not spawning. Rela-
tive abundance of adult sardines on the other
series is given as low, medium, or high in Table 1,
according to indications discussed above.

This succession of changes in abundance of
adults is too irregular to be attributed to growth of
individuals in a stationary population. It must be
due largely to movements into and out of the small
area studied. In the last major change, the biomass
approximately doubled in about 2 wk between
series 9 and 10. No pelagic fish species has such a
high growth rate for adult individuals. It was
noted during April and May that fish on the
continental shelf were more abundant north of the
sampling line (as far as lat. 22°20'N, which was the
limit of the acoustic surveys) than along the
sampling line (Thorne et al. in press). The fishing
results of the Professor Siedlecki also indicated
that sardines were more abundant to the north of
our area than within it. It is therefore very
probable that the biomass increase between series
9 and 10 represented a movement of sardines into
the study area from the north.

It is of interest to consider possible causes of the
sardine movements. A population of sardines
living off the southern part of the coast of Spanish
Sahara would be likely to move into a particular
area, like our study area, when conditions were

894

suitable to them and move out of the area when
conditions became unsuitable. The principal de-
terminants of distribution of pelagic fish are
believed to be temperature and food supply. Tem-
perature conditions in the study area were suit-
able for adult sardines during the whole period of
JOINT-I, since they occur in waters from 14° to
18°C off Morocco (Furnestin and Furnestin 1970).
Changes in abundance of food might however
have caused movements of sardines into and out of
the study area. No studies of the diet of Sardina
pilchardus have been made off Spanish Sahara
except for two fish mentioned later. Elsewhere in
its range, including waters off Morocco, it feeds on
phytoplankton and small zooplankton (Larrafieta
1960; Furnestin and Furnestin 1970). The dis-
tribution of sardines along the sampling line was
like that of phytoplankton and small zooplankton
as shown earlier: all three having maxima on the
middle and outer parts of the continental shelf.
This suggests that relative abundance of one or
both of those kinds of food determines sardine
distribution in a spatial sense and might do so in a
temporal sense.

Comparison of means of zooplankton concen-
tration with data on sardine abundance (Table 1)
shows no relation between them. If means of
chlorophyll concentration are used, there is the
following relation: sardine abundance is low when
chlorophyll values are 115 mg/m- or less, and
medium or high when chlorophyll values are 147
mg/m- or more. This suggests that sardines en-
tered the study area in order to feed on phyto-
plankton when it was relatively abundant and left
the area when phytoplankton was relatively
scarce.

No adult sardines were obtained during
JOINT-I. On cruise AUFTRIEB 1975 we caught
two sardines in the same area in February. M.
Elbrachter kindly identified the contents of their
stomachs: one contained no organisms except
foraminifera, and the other contained phyto-
plankton in good condition, including 15 species of
diatoms, and 2 species of dinoflagellates, and 2
copepods. Thus 5. pilchardus feeds on phyto-
plankton and zooplankton off Spanish Sahara, as it
does off Morocco and in other parts of its range.
Phytoplankton might be an important part of the
diet of the Sahara sardine, suflRciently to cause the
sardine to move in relation to changes in phyto-
plankton abundance as suggested by our data, but
we cannot be certain. More work on the diet of the
sardine off Spanish Sahara is needed. Mauri tanian

BLACKBURN and NELLEN: EGGS AND LARVAE IN AN UPWELLING AREA

sardines have more gillrakers than Moroccan
sardines of the same size (Furnestin 1955). This
could signify that the mean size of organisms in
the diet of sardines decreases from north to south
along the African coast.

Table 1 shows that abundance of anchovy
(Engraulis) eggs does not run parallel in time with
that of sardine eggs. There is a large difference
betvi^een the ratio of the mean numbers of the two
kinds of eggs on series 5 and 10, for instance,
although temperatures were about the same
(Figures 2, 4). We are unable to draw any conclu-
sions about changes in anchovy abundance and
their causes, even in the tentative ways attempted
here for the sardine.

The concentration of Trachurus at the shelf
edge may indicate a feeding aggregation on large
zooplankton, such as euphausiids and large
copepods, which are more abundant there than on
the shelf (Figures 2-4). The high abundance of
large zooplankton sometimes extends farther
offshore than Trachurus, however. Some other
factor must help to determine abundance of Tra-
churus. The diet of T. trachurus and T. trecae off
northwest Africa is about 80% euphausiids, 10%
copepods, and 10% small fish such as anchovy
(Boely et al. 1973). Phytoplankton is sometimes a
minor constituent of Trachurus stomach contents,
however (Letaconnoux 1951; Overko 1964; S. Schulz
pers. commun.).

ACKNOWLEDGMENTS

The valuable assistance of several people is
noted in the text. Part of the senior author's work
was done at the Institut fiir Meereskunde, whose
generous hospitality is acknowledged. This study
was part of the activity of the CUBA program,
supported by the International Decade of Ocean
Exploration of the U.S. National Science Founda-
tion, Grant Number GX-33502, and of the Institut
fiir Meereskunde.

LITERATURE CITED

Aboussouan, a.

1967. Oeufs et larves de teleosteens de I'ouest africain. 5.
Caranx rkonchus GeoflFr. St.-Hil. (Carangidae). Affinites
avec Trachurus trecae Cadenat. Bull. Inst. Fondam. Afr.
Noire, Ser. A, Sci. Nat. 29:1039-1050.
Ahlstrom, E. H.

1959. Vertical distribution of pelagic fish eggs and larvae off
California and Baja California. U.S. Fish Wildl. Serv.,
Fish. Bull. 60:107-146.

Barton, D.

1974. Upwelling events during JOINT I. CUBA (Coastal
Upwelling Ecosystems Anal.) Newsl. 3(5):21-27.

Blackburn, M.

1975. Summary of existing information on nekton of Span-
ish Sahara and adjacent regions, northwest Africa.
CUEA (Coastal Upwelling Ecosystem Anal.) Tech. Rep. 8,
49 p.

BoELY, T., A. Wysokinski, and J. Elwertowskl

1973. Les chinchards des cotes senegalaises et
mauritaniennes. Doc. Sci. Provis., O.R.S.T.O.M., Dakar,
46, 47 p.
Bravo de Lacuna Cabrera, J., and E. Santaella Alvarez.

1973. Observaciones biologico-pesqueras en el banco pes-
quero sahariano. Bol. Inst. Esp. Oceanogr. 171, 79 p.
Chabanne, J., AND J. Elwertowskl

1973. Cartes des rendements de la peche des poissons
pelagiques sur le plateau continental nord-ouest africain
de ll°a 26°N. Doc. Sci. Provis., O.R.S.T.O.M.. Dakar, 49,
8 p.

CONAND, F., AND E. FaGETTL

1971. Description et distribution saisonierre des larves de
sardinelles des cotes du Senegal et de la Gamble en 1968 et
1969. Cah. O.R.S.T.O.M., Ser. Oceanogr. 9:293-318.

CONAND, F., AND C. FraNQUEVILLE.

1973. Identification et distribution saisonierre de larv'es de
carangides au large du Senegal et de la Gamble. Bull.
Inst. Fondam. Afr. Noire 35, Ser. A, Sci. Nat. 4:951-978.
Gushing, D. H., G. F. Humphrey, K. Banse, and T. Laevastu.

1958. Report of the committee on terms and equivalents.
Cons. Perm. Int. Explor. Mer Rapp. P.-V. Reun. 144:15-16.

de Buen, F.

1931. Clupeides et leur peche (4*' rapport). Rapp. P.-V. Reun.

Comm. Int. Explor. Mer Mediterr. 6:289-336.
1937. Clupeides et leur peche (7'' rapport). Rapp. P.-V. Reun.
Comm. Int. Explor. Mer Mediterr. 10:317-419.
Demir, N.

1963. Synopsis of biological data on anchovy Engraulis
encrasicolus (Linnaeus) 1758 (Mediterranean and ad-
jacent seas). FAO Fish. Synop. 26, 44 p.
Face, L.

1920. Engraulidae, Clupeidae. Rep. Dan. Oceanogr. Exped.
1908-1910, Mediterr. 2, Biol. A9, 140 p.
Furnestin, J.

1955. Remarques sur quelques caracteres de la sardine de
Mauritanie (Sardina pilchardus Walbaum). Cons. Perm.
Int. Explor. Mer Rapp. P.-V. Reun. 137:24-25.
Furnestin, J., and M. L. Furnestin.

1959. La reproduction de la sardine et de I'anchois des cotes
atlantiques du Maroc (Saisons et aires de ponte). Rev.
Trav. Inst. Peches Marit. 23:79-104.

1970. La sardine marocaine et sa peche. Migrations tro-
phiques et genetiques en relation avec I'hydrologie et le
plancton. Cons. Perm. Int. Explor. Mer Rapp. P.-V. R6un.
159:165-175.

HUREAU, J. C, AND E. TORTONESE.

1973. Carangidae. In J. C. Hureau and Th. Monod (editors).
Check-list of the fishes of the north-eastern Atlantic and
of the Mediterranean, Clofnam I, p. 373-384. Unesco,
Paris.
Kiliachenkova, V. A.

1970. Development and distribution of eggs and larvae of
Trachurus trachurus L. Cons. Perm. Int. Explor. Mer
Rapp. P.-V. Reun. 159:194-198.

895

FISHERY BULLETIN: VOL. 74, NO. 4

Klimaj, a.

197L Atlas rybacki (Fishery atlas) szelfu Afryki Polnocno-

zachodniej. Tom. 1. Morsk. Inst. Rybacki, Gdynia, 129 p.
1973. Atlas rybacki (Fishery atlas) szelfu Afryki Polnocno-
zachodniej. Tom. 2. Poiowy Pelagiczne. Morsk. Inst.
Rybacki, Gdynia, 128 p.
LarraReta, M. G.

1960. Synopsis of biological data on Sardina pilchardus of
the Mediterranean and adjacent seas. In H. Rosa, Jr. and
G. Murphy (editors), Proc. World Scientific Meeting on the
Biology of Sardines and Related Species, p. 137-173. Food
Agric. Organ. U.N., Rome.
Letaconnoux, R.

1951. Contribution a I'etudedesespecesdu genre Trachurus
et specialement du Trachurus trachurus (Linne 1758).
Mem. Off. Sci. Tech. Peches Marit. 15, 67 p.
LORENZEN.C. J.

1968. Carbon/chlorophyll relationships in an upwelling
area. Limnol. Oceanogr. 13:202-204.
LozANO Cabo, F.

1970. Caracteristiques zoogeographiques de la faune ichthy-
ologique des cotes des lies Canaries, du Maroc, du Sahara
espagnol et de la Mauritanie avec une etude speciale des
poissons cotiers. Cons. Perm. Int. E.xplor. Mer Rapp. P.-V.
Reun. 159:152-164.

Maigret, J.

1972. Campagne experimentale de peche des sardinelles et
autres especes pelagiques, Juillet 1970-Octobre 1971.
Tome 1. Obser\ations concernant I'oceanographie et la
biologie des especes. S.C.E.T.-International, 148 p.

Marchal, E. G.

1966. Oeufs, lan-es et post-lar\-es de I'anchois de Golfe de
Guinee, Anchoviella guineensis (Blache et Rossignol).
Doc. Sci. Provis., O.R.S.T.O.M., Abidjan 5. 15 p.

1967. Cle provisoire de determination des oeufs et larves des
clupeides et engraulides ouest-africains. Doc. Sci. Provis.,
O.R.S.T.O.M., Abidjan 14, 10 p.

Maurin, C.

1968. Ecologie ichthyologique des fonds chalutables atlan-
tiques (de la bale ibero-marocaine a la Mauritanie) et de la

Mediterranee occidentale. Rev. Trav. Inst. Peches. Marit.
32:1-147.
Maurin, C, F. Lozano Cabo, and M. Bonnet.

1970. Inventaire faunistique des principales especes ichthy-
ologiques frequentant les cotes nord-ouest africaines.
Cons. Perm. Int. Explor. Mer Rapp. P.-V. Reun. 159:15-21.
Nishimura, S.

1957. Vertical distribution of the floating eggs of Mau-
rolicus japonicus Ishikawa, a gonostomatid fish, in the
sea. Rep. Jap. Sea Reg. Fish. Res. Lab. 3:13-22.
OVERKO, S. M.

1964. Biology and fishing of scad near the northwestern
shores of Africa. [In Russ.] Tr. Atl. Nauchno-issled. Inst.
Rybn. Khoz. Okeanogr. 11:45-64.

PiLLSBURY, R. D., J. S. BOTTERO, R. E. StILL, AND E.
MiTTELSTAEDT.

1974. A compilation of observations from moored current
meters. Vol. VIII. Wind, currents and temperature off
northwest Africa along 21°40'N during JOINT-I, Feb-
ruary-April 1974. Sch. Oceanogr. Oreg. State Univ., Data
Rep. 62, Ref. 74-20, 143 p.
Saville, a.

1964. Clupeoidae. Cons. Perm. Int. E.xplor. Mer, Fiches
Identification Oeux Lanes Poissons 1, 5 p.
Thorne, R. E., 0. A. Mathisen, R. J. Trumble, and M.
Blackburn.
In press. Distribution and abundance of pelagic fish off
Spanish Sahara during CUEA Expedition JOINT-I.
Deep-Sea Res.
UNESCO.

1966. Determination of photosynthetic pigments in sea-
water. Monogr. Oceanogr. Methodol. 1, 69 p.
VoiTURiEZ, B., p. Dufour, and R. Le Borgne.

1974. Preliminary results on R. V. CAPRICORNE 7402
Cruise in Cape Blanc. CUEA (Coastal Upwelling Ecosys-
tem Anal.) Newsl. 3(5):l-7.
WiTZELL, W.N.

1973. Gonostomatidae. In J. C. Hureau and Th. Monod
(editors). Check-list of the fishes of the north-eastern
Atlantic and of the Mediterranean, Clofnam I, p. 144-
122. Unesco, Paris.

896

LIFE HISTORY OF COHO SALMON, ONCORHYNCHUS KISUTCH,
IN SASHIN CREEK, SOUTHEASTERN ALASKA

Richard A. Crone' and Carl E. Bond^

ABSTRACT

The freshwater life of coho salmon, Oncorhynchus kisutch, in Sashin Creek, southeastern Alaska, was
studied from the fall of 1963 through the summer of 1968. Additional information on age composition
and fecundity of adults returning to Sashin Creek and a nearby stream was collected through the fall of
1972. Some pre-1963 data on coho salmon entering and leaving Sashin Creek were used. Weir counts and
estimates of numbers of adult salmon determined from spawning ground counts and mean redd life
were poor measures of the total escapement of coho salmon in Sashin Creek; an estimate made from
tagging a jwrtion of the escapement and subsequently determining tagged-to-untagged ratios of
spawners on the riffles proved to be a more reliable measure. The number of spawning coho salmon
varied for the years 1963 through 1967 from 162 to 916; the dominant age group was 43. The salinity of
the surface water of the estuary of Sashin Creek usually is less than 10-15"/oo; bioassays of salinity
tolerance indicated that coho salmon fry can survive in these salinities. In 1964, 44,000 coho salmon fry
migrated to the estuary soon after emergence, although none of the. scales collected from returning
spawners in subsequent years showed less than 1 yr of freshwater residence. Survival curves
constructed from periodic estimates of the stream populations of juvenile coho salmon for the years
1964-67 showed that mortality was highest in midsummer of the first year of life, when 62% to 78% of
the juveniles were lost in a 1-mo period. Most coho salmon smolts migrated from Sashin Creek in late
May or early June. In the spring of 1968, 1,440 smolts left Sashin Creek-37% were yearlings, 59% were
2-yr-olds, and 4% were 3-yr-olds. The average fork lengths were 83 mm for yearlings, 105 mm for
2-yr-olds, and 104 mm for 3-yr-olds.

Coho salmon, Oncorhynchus kisutch (Walbaum),
occur over a broad geographic range in the North
Pacific Ocean and Bering Sea. They spawn in
coastal streams from northern California to
northwestern Alaska and from northern Hok-
kaido, Japan, to the Anadyr River, USSR (Figure
1). The young usually remain in fresh water for 1
to 3 yr before migrating to sea as smolts; they are
sexually mature after about 14 to 18 mo in the sea.
In some systems some fry emigrate to salt water
in their first spring or summer of life, but they
apparently do not contribute significantly to the
adult return (Chamberlain 1907; Gilbert 1913;
Pritchard 1940; Wickett 1951; Foerster 1955).

Among the numerous populations of coho salm-
on, there are differences in freshwater life history
that appear to be related to latitude. In the
southern one-third of their range, coho salmon
typically remain in fresh water about 1 yr before

'Northwest Fisheries Center Auke Bay Fisheries Laboratory,
National Marine Fisheries Service, NOAA, Auke Bay, AK 99821.
Permanent address: 1211 Snipes Street West, The Dalles, OR
97058.

-Department of Fisheries and Wildlife, Oregon State Univer-
sity, Corvallis, OR 97331.

they migrate to sea in their second year of life-15
to 18 mo from egg deposition (Pritchard 1940;
Briggs 1953; Smoker 1953). Farther north, in
Alaska, coho salmon remain 1, 2, or 3 yr
(occasionally 4) in fresh water after they emerge
from the gravel (International North Pacific
Fisheries Commission 1962; Godfrey 1965; Drucker
1972). In some of the Alaska streams and in
Kamchatka, USSR, coho salmon that remain in
fresh water for 2 yr may represent a larger
percentage of the population than those that
remain for 1 yr (Gilbert 1922; Semko 1954; An-
drews 1962; Logan 1963; Engel 1966; Kubik 1967;
Redick 1968; Armstrong 1970; Drucker 1972).

Most studies of coho salmon behavior and sur-
vival in fresh water have been conducted in the
southern and central parts of the range: Califor-
nia, Oregon, Washington, and British Columbia in
the eastern Pacific (Neave 1948; Wickett 1951;
Briggs 1953; Smoker 1953; Shapovalov and Taft
1954; Foerster 1955; Salo and Bayliff 1958; Chap-
man 1962, 1965; Koski 1966); and Kamchatka in the
western Pacific (Kuznetsov 1928; Gribanov 1948;
Semko 1954). Information on more northerly
stocks is much less detailed.

Manuscript accepted: April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4. 1976.

897

FISHERY BULLETIN: VOL. 74, NO. 4

Figure 1. -Geographic
range of coho salmon in
North Pacific Ocean and
Bering Sea. Dashed line in-
dicates coastline where coho
salmon enter streams for
spawning.

""K/s-^""^!-— X •»•

^WV

/ T -j^~J^(^:^^~^

POINT ^ ^""^ 1 ^\-\
HOPE — /^ 1 ^1

>>. f — ~ ALASKA • \

"""^S;^

V ( tV"»^"^^ ANADYR RIVER^U

"^/) (J?J / ;'.'"- 1 . *iJlO^.--'' SASHIN CREEK -il ,^^

CANADA

AMUR RIVE^^-^„/ / P^ •' /'

/ r yy ^—HOKKAIDO
yh~^~. / <.^ ;/ NORTH

OCEA"
pACVflC SACRAMENTO RIVER

SAN FRANCISCO BAY

UNITED STATES

-A "^

^^^^,

In our studies at Sashin Creek, southeastern
Alaska, we have attempted to determine impor-
tant aspects of the life histories of populations of
the coho salmon near the center of its range
(southeastern Alaska). We have compared our
findings with life histories of populations in other
parts of the range and have emphasized 1) char-
acteristics of adult spawners (including time of
stream entry and time of spawning, longevity on
the spawning grounds, age structure, and fecun-
dity), 2) survival of eggs and alevins in the gravel,
and 3) survival and growth of juveniles up to the
time of seaward migration.

STUDY AREAS

Sashin Creek empties into Chatham Strait in
the inner bay of Little Port Walter on the south-
eastern shore of Baranof Island (Figure 2). The
stream originates in Sashin Lake about 3 km from
tidewater and drains about 10 km^ of forested
watershed-mostly western hemlock. Tsuga het-
erophylla, and Sitka spruce, Picea sitchensis (U.S.
Geological Survey 1972).

The discharge pattern of Sashin Creek is
governed by seasonal rainfall and the rate of
melting of accumulated snow. For the 10-yr period
1963-72, annual precipitation at Little Port Walter
averaged about 587 cm (231 inches).-^ Although

^This average was computed from data from volumes 49-58 of
the U.S. Weather Bureau s "Climatological Data, Alaska, Annual
Summary." However, because precipitation for August 1967 (vol.
53) was reported incorrectly as 6.99 inches, we used the figure
from the original records at Little Port Walter of 19.08 inches for
August 1967 in computing the 10-yr average precipitation.

Sashin Lake intercepts part of the runoff and
tends to even out flows in Sashin Creek, discharge
varies from less than 0.3 mVs in midwinter to as
much as 34 m-^/s after heavy rains in September
and October.

Salmon have access to the 1,100 m of stream
between the weir at the upper limits of salt water
and a high waterfall upstream. Coho salmon rarely
spawn in the 160 m of stream immediately below
the waterfall or in the intertidal stream channel;
both areas have a steep gradient and coarse
bottom material.

The spawning ground is divided into three areas
(upper, middle, and lower) which have different
physical characteristics but in total contain about
13,000 m- of spawning gravels (Table 1). The upper
area contains about 25% of the stream's suitable
spawning gravels and is characterized by a steep
gradient (relative to the other sections) and coarse
bottom materials. The middle area has about 30%
of the spawning gravel and an intermediate
gradient with a higher proportion of smaller
gravel and fines. The lower area is the largest and
contains about 45% of the spawning gravel; it has a
low gradient and a high proportion of fines in the
bottom materials.

Rearing areas of juvenile coho salmon include
the three spawning areas plus pools, backwaters,
and to a limited extent, the 160-m section of
stream in the canyon immediately downstream
from the waterfall. In our investigation of juvenile
coho salmon, the three ecologically distinct study
areas were maintained. An additional 3,473 m^
were included in the study areas to incorporate

898

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Figure 2.-Southeastem Alaska and Little Port Walter region,
site of coho salmon study.

first 215 m flows through a muskeg meadow and
the most upstream 75 m flows through forest.

The fish fauna of Sashin Creek consists of pink
salmon, 0. gorbuscha; coho salmon; chum salmon,
0. keta; rainbow trout, Salmo gairdneri; Dolly
Varden, Salvelinus malma; and coastrange scul-
pin, Cottus aleuticus. A few adult sockeye salmon,
0. nerka, occasionally stray into the stream.

ADULT COHO SALMON STUDIES

In our studies of adult coho salmon we deter-
mined: 1) size of escapement, i.e., the number of
coho salmon spawners that returned to Sashin
Creek; 2) average redd life of females; 3) distribu-
tion and density of spawners in each study section;
4) interspecific competition between coho and pink
salmon; 5) age structure of spawners; 6) fecundity
of females; and 7) egg retention of spent females.
In addition, for comparison with data from Sashin
Creek, we obtained data on the age and fecundity
of adult coho salmon from Nakvassin Creek in Port
Herbert, a 7-km-long fiord about 5 km north of
Little Port Walter (Figure 2). Nakvassin Creek,
about 0.4 km long, is the outlet stream from
30-hectare Nakvassin Lake. Coho and sockeye
salmon, Dolly Varden, rainbow trout, coastrange
sculpin, and threespine stickleback, Gasterosteus
aculeatus, inhabit the lake. These species plus pink
and chum salmon inhabit Nakvassin Creek.

pools and backwaters, for a total of 16,557 m-
(Table 1).

In 1965-67, part of Funny Creek, a small tribu-
tary of Sashin Creek near tidewater, was added to
the study area. Funny Creek is about 1.5 m wide on
the average and slow flowing; the bottom is mostly
mud and detritus but has a few gravel areas that
are used by coho salmon for spawning. The Funny
Creek study area included 441 m^ of stream from
its junction with Sashin Creek upstream 290 m; the

Size of Escapement

Adult coho salmon generally enter Sashin Creek
from early August to early November, but the
greatest numbers enter from late August to
mid-October. Spawning. usually begins early in
October and ends in mid-November.

Adult salmon have been counted in Sashin Creek
since 1934 through a weir at the head of tidewater.
From 1934 to 1969, counts of coho salmon at the

Table 1.- Surf ace area, average gradient, and size composition of bottom materials less than 15.2 cm in diameter in

three study areas of Sashin Creek. ^

Spawning
area (m^)

Total area^

(m2)

Average
gradient

(%)

Percentage

of spawning area

composed of

Study
area

Cobbles
( >12.7 mm)

Pebbles and

granules

(1,68-12.7 mm)

Sands and silts
(<^1.68 mm)

Upper
Middle
Lower

Total

2,945
4,067
6,072

13,084

4,049
4,441
8,067

16,557

0.7
0.3
0.1

0.3

81
61
47

16
26
36

13
17

'Table adapted from McNeil (1966).
2This area includes pools and backwaters.

899

FISHERY BULLETIN: VOL. 74. NO. 4

weir ranged from 0 to 567 (Table 2). The weir
counts are not accurate measures of the number of
coho salmon in the escapements, however, because
the weir was maintained primarily to count pink
salmon and the panels were usually removed at the
end of the pink salmon run near the end of
September. Moreover, coho salmon can jump over
the weir panels and many did so each year and
were therefore not counted.

Because of the problems with weir counts, an
effort was made to obtain accurate estimates of
the coho salmon escapements in 1963-65 and 1967
on the basis of repeated observations of the
number and distribution of salmon in the three
study areas (Table 3). Adults on the spawning
riffles were counted by periodic visual censuses,
and the counts were recorded separately for each
area. In 1963 and 1964, salmon were counted only

when water conditions were most favorable for
observing fish; spawners were not recorded sepa-
rately by sex. In 1965 and 1967, visual surveys were
conducted daily, except for 6 days in 1965 when the
water was too high to make observations; males
and females were recorded separately. Funny
Creek was included in the surveys in 1965 and 1967.
Spawners on the riffle areas were usually count-
ed between 1000 and 1400 h, when light conditions
were most favorable for observing fish. The ob-
server (wearing polarizing glasses to reduce glare
at the water surface) began counting at the
upstream end of the spawning area and continued
downstream. In 1965 and 1967, the observer
recorded the location of individual females with
reference to section markers spaced at 30.5-m
intervals and a baseline running longitudinally
between markers in the stream. The number of

Table 2.-Number of adult coho salmon counted into Sashin Creek at the weir by 2-wk intervals,

1934-69.'

Two-week

period

1-14

15-28

29 Aug. -

12-25

26 Sept.-

10-23

24 Oct.-

Year

Aug.

11 Sept.

Sept.

9 Oct.

Oct.

7 Nov.

Total

1934

—

1935

—

1936

—

236

—

1937

—

1938

—

(')

—

1939

—

{')

—

122

1940

—

1941

—

1942

—

1943

—

212

—

1944

—

249

262

—

328

1945

—

219

2232

—

567

1946

—

222

—

111

1947

—

250

—

111

1948

—

138

—

208

1949

—

170

—

202

1950

—

1951

253

—

135

1952

—

138

(2)

—

212

1953

235

—

115

1954

—

108

(2)

—

154

1955

210

—

168

1956

—

'(2)

—

1957

—

236

—

1958

—

219

—

179

1959

2133

—

271

1960

—

(2)

109

1961

(2)

1962

—

(2)

1963

—

202

2107

—

311

1964

—

(2)

1965

—

100

223

124

1966

—

1967

—

1968

—

100

270

—

184

1969

—

(2)

—

Total

249

1,246

1,274

961

456

4,255

Percent of all

fish counted

1.6

5.9

29.3

29.9

22.6

10.7

—

'Dally counts for 1934-63 from Olson and McNeil (1967).
2Welr discontinued during this period.

900

CRONE and BOND: LIFE HISTORY OF COHO SALMON

males near each female and the number of males
on the riffles but not with females also were
recorded.

The estimates of the total number of spawners
based on the periodic counts on the spawning
riffles were obtained in the following manner. The
counts of both sexes were plotted against time
(Figure 3). In the figure each point for 1965 and
1967 represents the average of three successive
daily counts of spawners, and each point for 1963
and 1964 represents a daily count. A curve was
drawn by eye through each set of points, and the
resulting area under the curve represents the

5 10

NOVEMBER

SURVEY DATE

Figure 3.-Stream survey estimates of numbers of spawning
echo salmon in Sashin Creek, based on periodic counts on
spawning riffles, 1963-65 and 1967. Area under each curve is
spawning effort, expressed in fish-days.

spawning effort in fish-days (see Table 3 and
section on Redd Life). The estimates of the total
number of spawners were then derived by divid-
ing the total number of fish-days by the average
redd life (the number of days a female spends on
the spawning site or redd). The method was
modified from McNeil (1966). The average of the
mean redd life computed for coho salmon females
in 1965 and 1967 was used to calculate total
number of spawners in 1963 and 1964.

As indicated in the following tabulation, es-
timates of the total number of spawners derived
from the stream survey data were much higher
than the counts at the weir, except for 1965.

Year
1963

Counted

at weir

311

Derived from stream .survey

(spawning effort-redd life)

458

1964

1965

124

1967

209

We also estimated the size of the escapement in
1965 and 1967 by conducting a mark-recapture
experiment using the Bailey modification of the
Petersen formula as given by Ricker (1958). In
1965, 46 adult coho salmon (32 females and 14
males) were tagged before spawning; in 1967, 73
unspawned coho salmon (28 females and 45 males)
were tagged. The tags used were plastic Petersen
disks. Marked-to-unmarked ratios were obtained
from observations made during the visual cen-
suses and these were used to estimate the popula-
tions. Based on the marked-to-unmarked ratios,
the estimated number of coho salmoh spawners
(both sexes) in Sashin Creek was 221 in 1965 and
370 in 1967 (Table 4).

The estimates of escapement size in 1965 and
1967 based on marked-to-unmarked ratios were
much higher than either the counts at the weir or
the estimates based on spawning effort and redd
life (Table 4). Several possible sources of error
existed in estimating numbers of spawners from
spawning effort and redd life: 1) The levels of
spawning activity were lower at low streamflows,
when visibility was good, and higher at high
streamflows (Figure 4), when visibility was re-
stricted. (In other words, the least accurate counts
of spawners occurred when the greatest numbers
were spawning.) 2) Some redds were occupied only
at night (indicated from our limited observations).
3) The assumption that the mean spawning life of
females was equal to that of males could be invalid.

901

FISHERY BULLETIN: VOL. 74, NO. 4

Table 3.-Distribution and density of spawning coho salmon in three areas of Sashin Creeit in 1963-65 and 1967.

Distribution

% of total salmon observed

Spawning effort (fish-days)

Density of spawning
(fish-days per square meter)

Brood

Upper

Middle

Lower

Upper

Middle

Lower

Upper

Middle

Lower

year

area

Total

area

Total

1963'

553

2,652

2,289

5,494

0.19

0.65

0.38

0.42

1964

297

674

976

<0.01

0.07

0.11

0.07

1965

607

543

1,224

0.03

0.15

0.09

1967

320

1,151

828

2,299

0.11

0.28

0.14

0.18

'W. J. McNeil, unpublished notes on 1963 coho studies. On file at Auke Bay Fisheries Laboratory, Auke Bay,
AK 99821.

Table 4.— Estimates of coho salmon escapements to Sashin
Creek, 1963-65 and 1967, based on three methods of estimation.

Numb

er of coho salmon

each year

Method of estimation

1963

1964

1965

1967

Weir count

Spawning effort and redd life
Marked to unmarked ratios
(95% confidence interval)

Spawner escapement
assumed in this report

311

458

'916

23
81

'162

124

221

197-250

221

83
209

370
342-403

370

'Based on observations in 1965 and 1967 that spawning ground
counts and redd life estimates were about one-half the estimates
based on marked-to-unmarked ratios of spawners.

1>0

120

</)

K
UJ

I/I
u.
O
K

CD 75

Z
60

SPAWNERS

STREAM DISCHARGE

DATE

5 10 IS

NOVEMBER

Figure 4.— Spawning ground counts of coho salmon in relation to
stream discharge, Sashin Creek, 1967.

We believe that the best estimates of abundance
of spawners are those based on tagging and
observing marked-to-unmarked ratios, but such
estimates are not available for 1963 and 1964.
Therefore, because the estimates derived from
spawning effort and redd life in 1965 and 1967

902

were approximately one-half the estimate from
marked-to-unmarked ratios, the assumption was
made that in 1963 and 1964, only 50% of the
spawners in Sashin Creek were estimated from
spawning effort and redd life. Our best estimates
of the numbers of coho salmon spawners are 916 in
1963. 162 in 1964, 221 in 1965, and 370 in 1967 (Table
4). These values are used in the remainder of this
report.

The population estimates of coho salmon
spawners in Sashin Creek do not include jack coho
salmon (precocious males of various freshwater
ages but only one summer of marine life) because
none were tagged; without tags, their presence on
the spawning riffles would have been difficult to
detect. Apparently, only a few jack coho salmon
enter Sashin Creek; none were seen during the
1965 surveys of the spawning grounds, and only
five were seen during underwater observations in
1967.

Redd Life

The estimates of mean redd life used in our
calculations of escapement size were based on
experiments with marked females in Sashin Creek
in 1965 and 1967. Many untagged females that
could be identified from natural markings, such as
wounds, fungused areas, color, and size, were used
along with the tagged females. A female had to be
observed at the same location on two consecutive
days before she was considered to have selected a
permanent spawning site. One day was added to
the observed redd life for females, on the assump-
tion that they began to construct a redd an aver-
age of one-half day before first being observed and
remained on the redd for an average of one-half
day after last being observed.

The mean redd life of female coho salmon varied
between tagged and untagged (identified from
natural markings) females in both 1965 and 1967.
In 1965 the mean redd life for 56 females (18

CRONE and BOND: LIFE HISTORY OF COHO SALMON

tagged and 38 untagged) was 13 days (range of 6
to 21 days). In 1967, 151 females (21 tagged and 130
untagged) remained an average of 11 days on the
spawning riffles (range of 3 to 24 days). Tagged
females had shorter mean redd lives than un-
tagged females- 12 versus 13 days in 1965 and 9
versus 11 days in 1967. The difference between
tagged and untagged fish may have been due to
difficulty in positively identifying untagged
females with short redd lives, thus biasing the
mean toward a greater value. Also, handling and
tagging might have resulted in a shorter mean
redd life for tagged females.

Tagged male coho salmon in Sashin Creek and
both sexes in Funny Creek had shorter spawning
lives than tagged females in Sashin Creek.
Because males in Sashin Creek in 1965 were all
tagged with the same color and could not be
recognized individually on the spawning grounds
as they moved from female to female, their
spawning life could not be calculated. In 1967,
however, the males in Sashin Creek had a mean
spawning life of 9 days. The spawning life of coho
salmon (males and females combined) was from 3
to 7 days in Funny Creek in 1965 and 1967. Rapidly
rising and falling water levels, which caused some
spawners to leave the area at low flows, and
predation by brown bears, Ursus arctos, probably
contributed to the shorter spawning life in Funny
Creek.

In Sashin Creek, stream life and spawning life
are not the same length because many coho salmon
enter the stream and mature in pools for a month
or more before they spawn. In Oregon, the mean
length of time that female coho salmon spent in a
tributary of Wilson River, Spring Creek, before
death was 11 days (Willis 1954). Adults migrating
into Spring Creek frequently begin spawning as
soon as they enter the tributary; apparently they
stay in the larger Wilson River until they are ripe
(Willis 1954). Koski (1966) reported mean stream
lives of 13.7 and 13.1 days for coho salmon that
spawned in two tributaries to Drift Creek, Oreg.

Distribution and Density of Spawners

The distribution and density of spawners on the
Sashin Creek spawning grounds were observed in
1963, 1964, 1965, and 1967 (Table 3). Distribution in
each study area is expressed as the percentage of
the total number of salmon observed spawning
and as total spawning effort (fish-days) observed
in each area. Density is the observed spawning

effort divided by the square meters of spawning
area.

In each of the 4 yr, the density of spawning coho
salmon was higher in the middle and lower study
areas than in the upper. In 1963, 1965, and 1967, the
middle area had the highest density of spawners;
in 1964 the lower area had the highest.

In Funny Creek a few coho salmon were seen
spawning in 1965 and 1967. In those 2 yr about 4%
of the estimated escapement of coho salmon to
Sashin Creek spawned in Funny Creek.

The distribution and escapement of pink and
coho salmon in Sashin Creek are shown in Table 5.
Pink salmon usually were distributed more evenly
throughout the creek than coho salmon. Merrell
(1962) and McNeil (1966) reported that spawning
pink salmon used the upper area extensively only
in years when spawning escapements were large;
when pink salmon escapements were small,
spawning was concentrated in the lower area. The
fact that relatively few coho salmon used Sashin
Creek may explain why such a small proportion
spawned in the upper area. In addition, ecological
features of that area such as steep gradient and
coarse bottom materials may limit its usefulness
for spawning.

Table 5.-Distribution and escapement of spawning coho and
pink salmon in three study areas of Sashin Creek, 1963-67.

Percentage

of total salmon

observed

Year and

species of

Escape-

Upper

Middle

Lower

salmon

ment

area

1963

Coho

916

Pink

16,757

1964

Coho

162

Pink

2,193

1965

Coho

22l"

Pink

14,813

1966

Coho

(')

—

Pink

5,761

1967

Coho

370

Pink

38,067

'The weir count of coho salmon was 90 when the weir gates
were removed in md-September.

Interspecific Competition

Pink salmon are the most abundant fish in
Sashin Creek-the number of adult pink salmon
ranged from about 2,000 to 72,000, and their
progeny ranged from about 0.3 to 3.6 million for
the years of this study, 1963-72.

Because pink salmon complete their spawning

903

FISHERY BULLETIN: VOL. 74, NO. 4

in Sashin Creek before coho salmon spawning
begins, spawning by coho salmon could be det-
rimental to pink salmon embryos. In 1965, we
tried in each of the three study areas to assess the
effect of coho salmon superimposing their redds on
those of pink salmon. The densities of live pink
salmon embryos, which were estimated from
routine sampling of the spawning riffles with a
hydraulic sampler prior to coho salmon spawning
(McNeil 1964), were used in conjunction with the
average size of a coho salmon redd to estimate the
total number of pink salmon embryos that could
have been destroyed in gravel disturbed by
spawning coho salmon. At 13 redds throughout the
stream the average area of gravel disturbed by
spawning coho salmon was 2.6 m- per redd.

The possible effect of coho salmon spawning on
pink salmon embryos in October 1965 is shown in
Table 6. The estimated spawning population of 110
female coho salmon would have disturbed a total of
286 m- of spawning gravel. Hydraulic sampling of
the spawning grounds in late September before
the coho salmon spawned indicated an average
density of 680 live pink salmon embryos per square
meter (see footnote 1, Table 6). About 200,000 live
pink salmon embryos resided in areas disturbed by
coho salmon spawners.

In years when the numbers of coho and pink
salmon spawners are similar to those of 1965, it is
doubtful that coho salmon spawning has a
significant detrimental effect on the survival of
pink salmon embryos. Even assuming complete
mortality of pink salmon embryos in gravels
utilized by spawning coho salmon, the impact on
survival of pink salmon in 1965 would have been
slight-about 2% of the viable pink salmon em-
bryos present. Mortality of pink salmon eggs from
redd superimposition by coho salmon could be
significant if the number of coho salmon spawners
were to greatly increase by natural or artificial
processes.

Age Determination

We determined the age structure of samples of
adult coho salmon in Sashin Creek in 1965-67 and
1969 and in Nakvassin Creek in 1966-72 by scale
analysis (Table 7). Most of the salmon had spent
two summers and two winters in fresh water after
emergence from the gravel, had migrated to sea in
the beginning of their third year, and had then
spent two summers and one winter in the ocean
(designated in the Gilbert-Rich system as age 4-i).
A smaller portion of those sampled had spent 1 yr
in fresh water after emergence, had entered the
sea at the beginning of their second year, and had
then remained two summers in the ocean (age 82).
Adults that had migrated to sea at the beginning
of their fourth year of life and spent two summers
in salt water (age 54) usually constituted the
smallest fraction of each year's run.

The presence of a large and dominant brood
year of coho salmon in Nakvassin Creek is in-
dicated by the percentage age distribution of
returning adults. In 1967, 40% of the adults sam-
pled for scales were age 30—1964 brood coho
salmon that had spent 1 yr in fresh water before
migrating as smolts; in 1968, 94% of the adults

Table 7.- Age structure as determined from samples of scales of
adult coho salmon from Sashin Creek, 1965-67 and 1969, and
Nakvassin Creek, 1966-72.

Source and year
of sample

No. of fish
sampled

Percentage age distribution

Sashin Creek:

1965

1966

1967

1969

Nakvassin Creek:

1966

1967

1968

1969

1970

1971

1972

Table 6.-Possible effect of coho salmon spawning on pink salmon embryos in Sashin Creek steambed in October 1965.

Area

Percentage of

observed coho

salmon spawning

effort

Live pink salmon Estimated

Area of gravel embryos/m^ viable pink Estimated

Estimated disturbed by before salmon embryos viable pink

coho salmon coho salmon coho salmon disturbed by salmon embryos

females (m^) spawned' coho salmon in study areas

Percentage of

total pink

salmon embryos

disturbed by

coho salmon

Upper

Middle

Lower

Total

51
43

100

7
56
47

110

18
146
122

286

750

1,200

300

2680

14,000

175,000

37,000

226,000

2,209,000
4,880,000
1,822,000

8,911,000

0.6
3.6
2.0

32.2

'W. J. f^cNeil, Auke Bay Fisheries Laboratory, (pers. commun.).
^Mean density, weighted according to area size.
3|^ean percentage, weighted according to area size.

904

CRONE and BOND: LIFE HISTORY OF COHO SALMON

were 43—1964 brood coho salmon that had spent 2
yr in fresh water; and in 1969, 29% of the adults
were 54—1964 brood coho salmon that had spent 3
yr in fresh water. Another large brood year in-
dicated by the ages of returning adults is the 1967
brood. No similar patterns of a strong brood year
are evident in the 4 yr of data from Sashin Creek
coho salmon (Table 7).

Direct comparison of the many studies on age
composition of coho salmon must be done with
caution because of year-to-year variations and
different sampling techniques, but a general clinal
change in freshwater and total age with latitude is
suggested— southerly populations are predomi-
nantly age 3.2 and northerly populations predomi-
nantly age 43. In British Columbia, Washington,
Oregon, and California, coho salmon (exclusive of
jacks) are almost all age 3_> (Pritchard 1940; Marr
1943; Smoker 1953; Shapovalov and Taft 1954;
International North Pacific Fisheries Commission
1962). Gilbert (1922) reported that about 60% of the
coho salmon of the Yukon River were age 43; the
remainder were age 3^. Coho salmon populations
in most streams studied in the Cook Inlet-Kenai
Peninsula area of Alaska are composed of 60% to
95% age 43 fish (Andrews 1962; Logan 1963; Engel
1966; Kubik 1967; Redick 1968). In the Karluk
River system on Kodiak Island, Alaska, age 43 fish
also are dominant but age 54 fish, rather than age
3^, are the second most abundant (Drucker 1972).
Semko (1954) listed age composition of coho salm-
on from the Bolshaya River, Kamchatka, for 8
yr; in two of the years (1946 and 1947) age 43 adults
outnumbered age 32. The highest percentage of
age 43 fish reported by Semko (1954) was 64.7%.
The age composition of coho salmon from the
commercial fisheries of the Taku and Stikine rivers
in southeastern Alaska in 1955 was 68.0% age 32
and 28.2% age 43 (International North Pacific
Fisheries Commission 1962). A later report on
Stikine River coho salmon caught in 1955 gives age
composition as 45.2% age 32 and 51.9% age 43
(Godfrey 1965). Of several thousand coho salmon
represented by scales collected from the commer-
cial fisheries in southeastern Alaska, about half
spent one winter in fresh water (age 32) and half
spent two winters in fresh water (age 43) (Smoker
1956). Nearly equal numbers of ages 82 and 43 also
were reported for coho salmon at Hood Bay Creek
in southeastern Alaska (Armstrong 1970).

Fecundity

We determined the fecundity of female coho

salmon from Sashin Creek in 1966, 1970, and 1971
and, for comparison, from nearby Nakvassin
Creek in 1966-72 (Table 8). Most of the females
from Sashin Creek were collected at the weir and
the rest were collected with sport fishing gear in
the estuary (a total of 3 to 22 each year). All
samples from Nakvassin Creek were collected with
sport fishing gear in the estuary (6 to 45 females
each year). Ovaries from individual females were
placed in containers of water and boiled until the
eggs hardened and separated from the ovarian
tissues. The mean of the annual fecundity samples
from Sashin Creek was 3,186 eggs per female (33
fish); the fish from Nakvassin Creek were slightly
smaller and the mean of the samples was 2,326
eggs (116 fish).

The relation between number of eggs and fork
length for Sashin Creek and Nakvassin Creek coho
salmon was calculated by the method of least
squares regression. The regressions for Sashin
Creek and Nakvassin Creek are Y = -441.48
-1-51.633X (r = 0.31) and F. = -824.59 -I- 47.686X
(r = 0.37), respectively. Y is the estimated number
of eggs and X is the fork length in centimeters of
females. Log transformations of number of eggs
and fork length did not increase the values of r
significantly.

Average fecundities of coho salmon reported for
other streams range between 1,983 and 5,343
(Table 9). Although these values were derived in
many different ways and therefore are not strictly
comparable, a general trend of increasing fecun-
dity from south to north and east to west does
appear.

Table 8.-Mean and range of fecundity and length of female
coho salmon from Sashin Creek in 1966, 1970, and 1971 (3 to 22
fish) and Nakvassin Creek in 1966-72 (6 to 45 fish).

No. (

Number of eggs

Fork 1

ength (cnfi)

femajf^^

Creek and year

sampl

Mean

Range

Mean

Range

Sashin Creek:

1966

2,868

1,195-4,418

70.5

64.8-73.7

1970

3,472

3,277-3,581

65.6

64.3-68.2

1971

3,217

2,537-4,665

69.6

63.1-79.8

Nakvassin Creei

1966

2,463

1,853-2,931

67.8

64.8-70.0

1967

2,143

1,737-2,565

65.6

61.3-70.8

1968

2,545

2,086-3,301

66.3

60.0-68.6

1969

2,228

1,664-3,120

66.4

56.0-69.5

1970

2,294

1,259-3,127

64.1

57.0-67.5

1971

2,414

2,000-2,816

66.7

63.5-69.9

1972

2,194

1,182-3,574

63.9

56.0-71.0

Retained Eggs

In 1965 and 1967, dying and dead spent female
coho salmon were examined for retained eggs

905

FISHERY BULLETIN: VOL. 74, NO. 4

Table 9.-Summary of available data on fecundity of coho salmon throughout most of the geographic range.' The data are not strictly
comparable among the various published and unpublished sources because of differences in methodology. Localities arranged in
counterclockwise order from California to Sakhalin Island, USSR.

Area

Average
no. of
eggs

No. of

fisfi in

sample

Average

fork length

(cm)

Source of data

California:

Scott Creek
Oregon:

Fall Creek, Alsea River

Big Creek
Washington:

Minter Creek

University of Washington, Seattle
British Columbia:

Cowichan River, Vancouver Island

Oliver Creek (tributary to
Cowichan River)

Beadnell Creek (tributary to
Cowichan River)

Sweltzer Creek

Fraser River

Nile Creek

Namu Cannery

Port John Creek
Alaska:

Sashin Creek, southeastern

Nakvassin Creek, southeastern

Bear Creek, Cook Inlet

Dairy Creek, Cook Inlet

Cottonwood Creek, Cook Inlet

Fish Creek, Cook Inlet

Swanson River, Cook Inlet

Lake Rose Tead, Kodiak Island

Lake Miam, Kodiak Island

Karluk River, Kodiak Island
Union of Soviet Socialist Republics:

Kamchatka River, Kamchatka

Bolyshaya River, Kamchatka

Paratunka River, Kamchatka
Tyml River, Sakhalin Island

22,616

366.3

Shapovalov and Taft (1954)

1,983

266.2

Koski (1966)

3,030

70.2

James R. Graybill (pers. commun. 31 May, '

2,500

1,120

Salo and Bayliff (1958)

3,100

Allen (1958)

2,329

—

Neave (1948)

2,267

—

Foerster (1944)

2,789

___

Foerster (1944)

2,300

—

Foerster and Ricker (1953)

3,152

*65.3

Foerster and Pritchard (1936)

2,310

(*)

—

Wickett (1951)

3,002

♦69.8

Foerster and Pritchard (1936)

2,313

—

Hunter (1948)

3,186

68.6

Present study

2,326

116

65.8

Present study

4,115

193

—

Lawler (1963, 1964)

'3,595

179

66.9

Logan (1968)

4,177

155

72.8

Lawler (1963), Engel (1965) combined

2,346

220

55.1

Andrews (1961), l^cGinnis (1966) combined

2,426

112

—

Calculated from Andrews (1962)

3,448

1,019

62.3

Calculated from Engel (1967)

4,201

—

f^arriott (1968), Van Hulle (1970) combined

4,209

277

—

Van Hulle (1971)

4,706

'62.1

Drucker (1972)

4,883

860.4

Kuznetsov (1928)

4,300-

—

856.5

Semko (1954)

5,343

4,350

—

859.1

Gribanov (1948)

4,570

—

Smirnov (1960)

1973)

'Table adapted from Rounsefell (1957) and Allen (1958).

^Value calculated from regression curve.

^Mean length determined from 338 females.

■•Total length.

^Three to eight specimens per year.

'After introduction of Swanson River coho salmon stocks into Bear Lake.

'Mideye to fork length.

8Lengths given by Gribanov (1948), not from females sampled for fecundity.

during daily stream surveys. Only seven spent
females were examined each year because high
water washed most dying spawners from the
stream. The number of eggs ranged from 0 to 64
and averaged 8 per female for the two seasons.
Koski (1966) examined 30 spent female coho salm-
on in an Oregon stream and found an average of
four eggs per female. In streams of Kamchatka,
Semko (1954) found that coho salmon retained
0.3% of the actual fecundity (about 7 to 16 eggs per
female).

JUVENILE COHO SALMON STUDIES

With anadromous salmon, the result of fresh-
water production is a juvenile migrating to the
ocean-a smolt or fry physiologically adapted to

enter salt water, where most growth takes place.
Our studies were designed to measure the yield of
coho salmon smolts and to determine some of the
factors that bear on this yield. We counted and
sampled the juvenile coho salmon at a weir as they
left Sashin Creek and entered the estuary, and
also sampled juveniles in the stream with seines.
In addition, after determining that many fry of
unknown physiological capabilities entered salt
water, we performed experiments to determine
the ability of these fry to survive the salinities
existing in the estuary. For studies in the stream
coho salmon juveniles were considered as two
groups-fry (age 0) and fingerlings (age I and
older).

Specific topics considered here are: 1) the
numbers of coho salmon smolts and fry entering

906

CRONE and BOND: LIFE HISTORY OF COHO SALMON

the estuary from Sashin Creek each year, 1956-68;
2) the migration of fry to the estuary and their
ability to survive in salt water; 3) age and growth
of juveniles in the stream; 4) survival through
various life stages (potential egg deposition to fry
emergence and as juveniles in the stream); and 5)
mortality in fresh water.

Juveniles Entering the Estuary

In Sashin Creek the emergence of coho salmon
fry from the gravel usually begins in April and is
completed by the end of May, although in
especially cold years emergence may not start
until June or July. Juvenile coho salmon usually
live in Sashin Creek for 1 to 3 yr before migrating
to salt water as smolts, but some migrate to the
estuary during their first spring or summer as fry.
The migration of fry to salt water soon after they
emerge has been reported in several other streams
(Chamberlain 1907; Gilbert 1913; Pritchard 1940;
Wickett 1951; Foerster 1955), but none of these
authors reported a substantial return of adult
salmon from such early-migrating fry. All of the
scale samples from adult coho salmon at Sashin
Creek indicated that the fish had spent at least 1 yr
in fresh water. The absence of adults originating
from early-migrating fry suggests very poor
survival of fry entering the estuary at a small size
(usually <35 mm from Sashin Creek), which could
be the result of heavy predation or some failure to
adapt physiologically to the marine environment.
We have assumed that we can identify adults
derived from early-migrating fry on the basis of
the pattern of circuli on their scales. However, if

fry surviving in the estuary developed scale pat-
terns indistinguishable from those of fry spending
a year or more in fresh water, our assumption that
age 0 emigrants did not contribute to the adult run
could be incorrect.

Numbers of Fry and Smolts

Counts and estimates of the numbers of juvenile
coho salmon migrating from Sashin Creek ranged
from 218 to 44,023 fry and 928 to 2,865 smolts
between 1956 and 1968 (Table 10). In 1964, 44,023
fry left the stream between 19 April and 28
August, and most migrated in a 2-wk period in
mid-June-the greatest migration was 3,528 fry on
15 June. The smolt migration in Sashin Creek
varied about threefold from 1956 through 1968
(excluding 1965-66-Table 10). The relatively low
counts of smolts in 1964 probably resulted from a
change in the trapping procedures at the weir.
Before 1964, all fish migrating from Sashin Creek
were captured. The procedures used in 1964 were
designed to capture a portion of the emigrating
fry and did not retain smolts well. Because of high
water and ice damage to the weir, complete counts
of emigrating fry and smolts were not made for
1965-67. In 1965 and 1967, estimates of fry and
smolts were based on catches in fyke nets that
sampled the migrants at the weir site. No es-
timates were obtained in 1966.

A comparison of the time of smolt migration
from Sashin Creek with the time of migration
from streams and lakes along the eastern Pacific
coast from south-central Alaska to central coastal
California indicates that there is a tendency

Table lO.-Numbers and times of migration of coho salmon frj' and smolts past the Sashin Creek weir and yield of smolts,

1956-68.'

Numbe

r counted

Date of 1;

argest

Date last fish was

Yield of smolts

at '

weir^

Counting period

migration

observed to emigrate
Fry Smolts

per 100 m^ of

Year

Fry

Smolts

Fry

Smolts

rearing area

1956

928

16 Apr.-30 June

15 June

—

20 June

5.5

1957

373

1,961

10 Apr. -29 June

17 June

24 May

27 June

11.5

1958

2.854

1,015

7 Mar. -3 June

4 May

20 May

31 May

2 June

6.0

1959

218

1,587

1 Apr.-21 July

14 July

27 May

.15 July

3 July

9.3

1960

9.923

1,258

17 Mar.-2 July

12 June

10 June

30 June

7.4

1961

2,699

2,489

22 Mar.-19 June

21 May

28 May

17 June

14.6

1962

1,209

2,865

11 Mar.-4 July

14 June

27 May

4 July

3 July

16.9

1963

1,236

1,599

11 Mar.-8 July

30 May

24 May

1 July

3 July

9.4

1964

44,023

3334

15 Mar.-28 Aug.

15 June

24 May

28 Aug.

6 July

1965*

12,000

11 Apr. -30 July

24 June

—

15 July

—

1967*

10,000

1,400

10 Apr. -8 Aug.

28 June

25 May

8 Aug.

5 July

8.2

1968

1,665

1,440

26 Mar.-3 Aug.

5 June

24 May

3 Aug.

5 July

8.5

'The year 1966 is not included because the weir was damaged and substitute sampling was not conducted.

^Daily counts for 1956-64, available from Olson and McNeil (1967).

^Counting procedure changed from total to partial counts; holding facilities were inadequate for retaining all smolts cap-
tured. J I. t H

"Weir not functional; fyke net(s) fished to sample a portion of the spring emigration. Numbers of fry and smolts presenteo
are estimates made from fyke net catches.

907

FISHERY BULLETIN: VOL. 74, NO. 4

toward earlier migration in the soutliern part of
the range (Table 11).

To compare yields of coho salmon smolts
between strearns, we express yield in numbers per
unit area. Estimates of the annual yield of coho
salmon smolts from Sashin Creek for the period
1956-68 (except 1964-66) ranged from 5.5 to
16.9/ 100m- of rearing area (Table 10). The yield of
smolts for a 5-yr period in three streams tributary
to Drift Creek ranged from 18 to 67/100 m-
(Chapman 1965). The much lower yield of smolts
from Sashin Creek probably reflects increased
mortality accompanying the additional 12 mo of
freshwater residence for most smolts from Sashin
Creek. The number of nonmigrant yearling coho
salmon in Sashin Creek in early summer
(determined from population studies) approx-
imates the yield of smolts from Drift Creek
tributaries more closely than does the yield of
smolts (all ages) from Sashin Creek.

Early Emigration and
Salinity Tolerance of Fry

The number of fry entering the estuary is great
enough (Table 10) that the question of their fate in
salt water is important. Many factors such as
predation, failure to find adequate food, failure to
adjust physiologically to salt water, and disease
may act alone or in combination to determine the
survival of fry entering marine waters. We had
opportunity to explore adjustment to salt water as
a factor in survival of migrating fry.

Early-migrating coho salmon fry might have
reentered Sashin Creek undetected, although they
could not have done so while the fry and smolt weir
was in operation. In addition, a low waterfall
immediately downstream from the weir is a bar-
rier to upstream migration of coho salmon fry
except for several days each year when above-
average high tides inundate the falls. Our popula-

Table U. -Timing of seaward migration of coho salmon smolts from streams and lakes in Alaska, British Columbia, Washington,

Oregon, and California.

Location

Migration period

Peak of migration

Source of data

South-central Alaska:

Fire Lake (lat. 61°2rN)

Bear Lake (lat. 60°12'N)

Little Kitoi Lake, Afognak Island
(lat. 58-12'N)

Karluk Lake, Kodiak Island
(lat. 57'^'27'N)

Lake Margaret (lat. 57'46'N)

Lake Genivieve (lat. 57'"46'N)
Souhteastern Alaska:

Taku River (lat. 58°33'N)

Eva Lake (lat. 57"'24'N)

Hood Bay Creek (lat. 57°20'N)

Sashin Creek (lat. 56°23'N)
Central coastal British Columbia:

Port John (Hooknose Creek)
(lat. 52^08'N)
Southern British Columbia:

Cultus Lake (lat. 49°03'N)
West-central Washington:

Minter Creek (lat. 47°22'N)
Northwestern Oregon:

Gnat Creek (lat. 46°12'N)
Northern coastal Oregon:

Spring Creek (lat. 45"36'N)
Central coastal Oregon:

Drift Creek tributaries: Deer,

Flynn. and Needle Branch Creeks

(lat. 44°32'N)
Crooked Creek (lat. 44°25'N)

Southern coastal Oregon:

Sixes River (lat. 42°51'N)
Central coastal California:

Waddell Creek (lat. 37°06'N)

Mid May-early July
Late May-early Aug.

Late May-late July

Mid May-early July
Mid Mar. -early July
Mid May-mid July

Mid Apr-mid June^
Mid May-mid June
Early May-late June
Apr. -early July

Mid Apr. -early June
Apr. -June
Feb. -early June
Apr.-early June
Late Feb. -May

Feb. -May
Feb. -early June

Mar.-June
Apr.-early June

Late May-early June
Early June

Mid June

Late May-early June
Late May-early June
Late May-early June

Mid May-early June
Late May

Mid May-early June
Late May-early June

May

Late May-early June

May

Late Mar.-early May

Late Mar.-early Apr.
Apr. -May

Apr.-May
Late Apr.-May

Wallis (1967, 1968)
Logan (1963)

Parker and Vincent (1956)

Drucker (1972)
Van Hulle (1971)
Van Hulle (1971)

Meehan and Siniff (1962)
Armstrong (1970)
Armstrong (1970)
Table 10, this report

Hunter (1948, 1949)
Foerster and Ricker (1953)
Salo and Bayliff (1958)
Willis'
Willis et al."

Chapman (1962, 1965)
Harry H. Wagner (pers. commun.
9 July 1973)

Reimers'

Shapovalov and Taft (1954)

'Trapping facilities were completed after the beginning of the migration.

^Period when a sampling trap was operated.

'Willis, R. A. 1962. Gnat Creek weir studies. Final Rep., BCF Contract 14-17-0001-469, Fish Comm. Oreg., Res. Div., 71 p.

"Willis, R. A., R. N. Breuser, A. L. Oakley, and R. W. Hasselman. 1959. Coastal Rivers Investigations Prog. Rep., August 1957-June 1958,
Fish Comm. Oreg., 24 p.

sReimers, P. E. 1971. The movement of yearling coho salmon through Sixes River estuary. Coastal Rivers Investigations Prog. Rep. 71-
2, Fish Comm. Oreg., 15 p.

908

CRONE and BOND: LIFE HISTORY OF COHO SALMON

tion studies of juvenile coho salmon in Sashin
Creek suggest that no large-scale reentry of coho
salmon fry occurs.

Coho salmon fry from an Oregon coast stream
adjusted to water of moderately high salinities in
laboratory tests (Conte et al. 1966). Our field
observations, live-box experiments, and bioassays
at Little Port Walter confirm that ability for fry
from Sashin Creek. In July 1964, after about 44,000
coho salmon fry had migrated from Sashin Creek,
schools of fry were seen near the surface of the
inner bay. Most of them appeared to be in water of
low salinity above a density interface about 30 cm
deep, but they retreated to deeper more saline
waters when disturbed.

To study the ability of coho salmon fry to adjust
to the saline conditions in the Little Port Walter
estuary, some fry were confined in live-boxes in
the inner bay during the summer of 1964. Two
sizes of live-boxes were used: six small boxes (86 by
86 by 122 cm deep) were arranged so that the
water depth in the box was 81 cm, and two large
boxes (122 by 122 by 244 cm deep) were suspended
from a floating frame so that water depth was 235
cm. The small boxes were arranged in three sets of
two boxes each, and 60 fry were placed in each box;
the fry were from the weir trap, the inner bay, and
Sashin Creek. Twenty-five fry from the inner bay
were placed in each of the large boxes.

Initially, the Sashin Creek and weir trap fry in
the live-boxes entered high-salinity water for
short periods only, whereas some inner bay fry
remained in the high-salinity water for long
periods. Most of the fry stayed at or near the
density interface close to the top of the box where
the salinity was 14"/oo or less; but some, especially
those from the inner bay, swam for extended

periods near the bottom of the box where salinity
was 28 to 29"/oo.

Survival and growth of the fry seemed to be
related more to the size of the live-box and the
resulting competition for food than to the fry's
ability to adjust to the saline water of the bay. In
the small live-boxes, during the first 5 days 3% of
the fry died and in 30 days 26% had died; in the
large live-boxes in 35 days only 8% of the fry died
(Table 12). The general comparison is true both for
the entire small-box group versus the large-box
group and for the small-box group of fish from the
inner bay versus the large-box group (also from
the inner bay). No supplemental food was provid-
ed, and the fry in the small boxes grew very little
or not at all, whereas those in the large boxes grew
about 5 mm (Table 12).

Tests were conducted in July 1964 to measure
the ability of coho salmon fry to survive abrupt
transfer to higher salinity waters. Salinities were
determined with hydrometers. Plastic buckets
were used as test containers, and as in the live-box
studies, coho salmon fry were taken from Sashin
Creek, the weir trap, and the inner bay. For each
test, 10 fish were abruptly transferred from their
source water to the test water. The fry from all
three sources survived 48 h in salinities up to
23.5/00 (Table 13). In 29"/oo water, the fry from the
inner bay lived for 48 h, but none of those from the
weir trap and only 50% of those from Sashin Creek
survived 48 h. Of seven fry from the weir trap that
survived 96 h in 17.6"/oo salinity, three were trans-
ferred to 31"/oo for 48 h, and one survived; four were
transferred to 23.5"/oo and three survived for 48 h.
Seven fry that survived 96 h in 23.5"/oo salinity
were transferred to 31"/oo-four of these fry were
still alive 48 h later when the experiment was

Table 12.-Mortality and growth of coho salmon fry held in live-boxes in the inner bay of Little

Port Walter in the summer of 1964.

Size of

Average

Average fork

live-box
and no

Source

Total mortality after

initial fork
length of

length at end
of experiment

Of fisti

of fish

5 days

30 days

35 days

fry (mm)

(mm)

Small:

Weir trap

36.8

35.9

Inner bay

38.6

Weir trap

36.8

35.7

Sashin Creek

36.8

Inner bay

38.6

38.9

Sashin Creek

36.8

36.4

Large:

Inner bay

40.0

44.7

Inner bay

(')

40.0

45.4

'Experiment terminated after 30 days.
^No observations until day 35.

909

FISHERY BULLETIN; VOL, 74, NO. 4

Table 13.-Cumulative deaths of echo salmon fry taken from the inner bay of Little Port
Walter, the weir trap at the mouth of Sashin Creek, and Sashin Creek and held in waters of
various salinities, July 1964.

Size range

Cumu

lative deaths at

Salinity
(°/oo)

Number
of fish

Source
of fish

(fork length,
mm)

24 h

48 h

72 h

96 h

144 h

Weir trap

32-38

Inner Bay

37-45

(')

—

Weir trap

35-39

(')

—

Sashin Creek

35-41

—

2.2

Inner Bay

36-47

(')

—

12.5

Inner Bay

35-40

(')

—

12.5

Weir trap

35-42

—

12.5

Sashin Creek

38-45

—

17.6

Weir trap

33-39

19.3

Weir trap

34-39

—

19.3

Sashin Creek

34-63

—

23.5

Weir trap

33-40

23.5

Weir trap

32-40

Inner Bay

31-42

—

Weir trap

29-35

—

Sashin Creek

33-52

(')

—

Weir trap

31-42

—

'Accidentally discontinued before 72 h,

^Of seven survivors in 17 6°/oo at 96 h, three were transferred to 31 °/oo for 48 h, and one
survived; four were transferred to 23.5*^/00, and three survived for 48 h.
'Period from 96 to 144 h at 31°/oo salinity.

ended (Table 13). Otto (1971) found that salinity
tolerance of juvenile coho salmon was increased by
a 35-day exposure to water of lower salinity.

The inner bay has horizontal and vertical salin-
ity gradients (Powers 1963), which are excellent
for acclimation of young salmon to salt water.
Much of the surface of the inner bay is usually less
than 10-15'/iKi salinity. Our bioassays show that fry
able to survive 48 h in 2Z"/w salinity should be able
to acclimate fully to salinities encountered in the
inner bay if they have access to the low salinity
refuge. Our observation of coho salmon fry in
live-boxes confirms this ability.

Juveniles in the Stream

Age Determination

Because length frequencies of different age-
groups can overlap broadly at some times of the
year (Figures 5-7), scales of juvenile coho salmon
were analyzed to assign age-groups. Analysis of
scale samples indicated that an average of 11% of
all fingerlings (age I and older) collected from the
stream were age II. Sizes of the various age-
groups sampled by month for 1964-67 are shown in
Figure 5.

• i

.ll..

■ ACE 0
J 'i' a ACE 1
i ' □ ACE II

JUNE
n=3l

0
0

: it
l_.

II
1 1 n m n

JULY
n=62

hi^ 1 ri

1 II

AUGUST
n=l66

n f!

0 ' "

Tn_

SEPTEMBER
n=IOO

rni

■ ■

_l_

-1-

100

Figure 5. -Length frequencies and
ages of coho salmon juveniles from
Sashin Creek grouped by month of
collection, June to September 1964-67.
Ages were determined by analysis of
scale samples, and lengths were mea-
sured on preserved fish. Arrows in-
dicate location of mean lengths for each
age class.

FORK LENGTH (mm)

910

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Most coho salmon fry (age 0) collected in June
were under 40 mm in fork length and had not
formed scales. Gribanov (1948) found that the
scale covering usually appeared on young coho
salmon from Kamchatka at 40 mm long.

Growth and Age Characteristics

Growth of juvenile coho salmon in Sashin Creek
was determined from fork lengths (measured to
the nearest millimeter) of samples of fry and
fingerlings. Fry were collected periodically during
summer 1964, and fry and fingerlings were cap-
tured during each of several population estimates
of juveniles in Sashin Creek during 1965-67. An
additional 50 fry from each of the three study
areas of Sashin Creek were measured in mid-July
1966 and mid-August 1967. Samples of fingerlings

120

^ RANGE

110

1 l^MEAN

FRY FINCERLINCS

100

1.0

■

n=l.

2.7

n=U2

n=J2t

n=l

L 1

1 -'"

n=227

•0

•

■

•

SEPTEMBER

Figure 6.-Mean and range of fork lengths of fry (age 0) and
fingerling (ages I and II) coho salmon, Sashin Creek, 1964-68.
Lengths were measured from live fish.

from the three study areas were measured in early
July, early August, and mid-September 1968.

There was no consistent difference in the mean
fork lengths of corresponding age groups of
juvenile coho salmon captured in the upper, mid-
dle, or lower areas in any sampling period (Table
14). Juveniles from Funny Creek were usually
slightly smaller than those from Sashin Creek
during a corresponding period.

The length data for juvenile coho salmon sam-
pled in Sashin Creek for 1964-68 have been com-
bined by month for fry and fingerlings (Figure 6).
The difference between the fork length of fry and
fingerlings was pronounced in early summer, but
by July the lengths of fast-growing fry and
slow-growing fingerlings overlapped (Figure 6).
Occasionally it was difficult to assign the proper
age-group to juveniles in the overlapping sizes,
although they could usually be separated by the
brightly colored and proportionally longer fins and
smaller eyes of the fry.

The average fork length of coho salmon fry in
Sashin Creek in October is about 60 mm. The
average length of those that do not become smolts
the following spring but remain in the stream a
second year is usually 65-75 mm by July. In the
1968 migration, age I smolts averaged 83 mm, age
II smolts 105 mm, and age III smolts 104 mm
(Figure 7). In comparison, in 1968 coho salmon
from Hood Bay Creek in southeastern Alaska
averaged 83 mm fork length as age I smolts and 96
mm as age II smolts; in 1969 age I smolts averaged
79 mm and age II smolts 91 mm (Armstrong 1970).
For the years 1956, 1965, and 1968, coho salmon
smolts migrating from Karluk Lake, Kodiak Is-
land, Alaska, averaged HI mm as age I smolts, 139

10
8
6

ACE 1

MEAN LENGTH 83mm

N=95

■ ■■

ACE II

u^O

MEAN LENCTH 105 mm

O 8
S 6

i '

10
8
6

■

N=I5I

ACE III

MEAN LENCTH 10«mm

N=10

2
0

—

■

^ ■

.X.

_l_

_i_

-L

100

_L.

-L.

FORK LENCTH (mm )

110

120

.i-

Figure 7.- Length frequencies of ages
I, II, and III coho salmon smolts, Sashin
Creek, 1968. Total sample was 256 fish,
of which 37% were age 1, 56% age II, and
4% age III. Ages were determined by
analysis of scale samples, and lengths
were measured on live fish.

911

FISHERY BULLETIN: VOL. 74, NO. 4

Table 14.-Fork length (mm) of coho salmon fry and fingerlings captured in three study areas of Sashin Creek and Funny Creek,

1964-68.

Sashin Creek

Date and Upper area Middle area Lower area Total Funny Creek

type of 11

sample Range Mean No. Range Mean No. Range Mean No. Range Mean No. Range Mean No.

1964:

7 July:

Fry ___ ___ ___ 31-46 36.8 43 _ _ _

15 July:

Fry ___ ___ ___ 33-49 39.8 92 _ _ _

21 July:

Fry ___ ___ ___ 32-46 37.0 34 _ _ _

28 July:

Fry ___ ___ ___ 34-50 39.2 91 _ _ _

31 July:

Fry ___ ___ ___ 33-52 38.9 39 _ _ _

18 Aug.

Fry ___ ___ ___ 37-59 45.1 56 _ _ _

1965:
17 July:

Fry 34-44 38 5 37 34-44 38.0 43 35-45 38.1 103 34-45 38.2 183 _ _ _

Fingerlings 55-80 67.7 71 53-90 68.8 99 51-86 69.7 147 51-90 69.0 317 _ _ _

1 Aug.:

Fry ___ ___ ___ ___ 37-43 39.1 29

Fingerlings — — _ ___ ___ ___ 47-8O 62,1 28

11 Aug.:

Fry 39-47 41.9 17 37-46 40.5 12 35-49 40.3 99 35-49 40.5 128 35-46 39.1 84

Fingerlings 56-84 73.4 24 56-90 70.6 96 54-91 70.2 102 54-91 70.7 222 49-91 63.8 43

30 Sept.:

Fry 45-74 62.2 73 59-69 64.3 7 49-69 60.4 7 45-74 62.2 87 39-70 53.4 61

Fingerlings 76-104 88.0 1 1 84 84 1 75-97 86.0 2 75-104 87.4 14 71-105 85.8 44

1966:

27 June:

Fry 35-40 37.0 34 34-43 37.4 133 35-44 39.3 60 34-44 37.9 227 34-39 36.6 61

Fingerlings 60-89 71.9 24 53-85 67.6 69 56-91 71.8 135 53-91 70.6 228 50-98 72.1 78

8 July:

Fry 35-43 37.7 24 33-47 37.7 95 34-50 40.3 101 33-50 39.0 220 34-50 38.0 102

Fingerlings 62-93 77.9 48 58-95 75.4 45 56-104 77.5 56 56-104 77.0 149 50-102 70.6 105

14 July:

Fry 34-44 37.8 50 33-46 38.9 50 36-47 39.5 50 33-47 38.7 150 36-47 39.9 50

29 July:

Fry 34-56 46.4 50 36-53 42.4 53 36-55 43.0 51 34-56 43.9 154 33-56 40.8 54

Fingerlings 55-98 76.3 63 65-105 79.1 58 68-102 81.6 45 55-105 78.7 166 61-105 75.5 50

14 Aug.:

Fry 38-58 48.1 100 36-59 46.2 100 37-60 46.2 100 36-60 46.8 300 32-63 44.7 100
Fingerlings 60-102 82.0 89 63-107 84.2 100 65-106 84.6 100 60-107 83.7 289 56-119 81.1 100

8 Sept.:

Fry 44-68 57.3 50 39-67 49.6 100 46-68 58.0 100 39-68 54.5 250 40-68 52.1 100

Fingerlings 67-102 86.7 50 63-95 81.5 10 69-113 87.8 36 63-113 86.6 96 61-108 85.3 100

1967:

23 July:

Fry 33-41 36.5 50 34-45 37.9 71 33-52 38.8 120 33-52 38.0 241 36-42 37.9 50

Fingerlings 52-104 70.8 100 53-93 69.5 105 53-92 72.8 112 52-104 71.1 317 52-103 77.1 107

4 Aug.:

Fry ——____ 34.49 38.6 116 34-49 38.6 116 35-50 40.2 75

Fingerlings — _ _ _ _ _ 57-104 76.2 228 57-104 76.2 228 55-104 74.3 50

17 Aug.:

Fry 36-46 40.7 50 35-50 40.9 54 37-55 41.1 50 35-55 40.9 154 _ _ _

5 Sept.:

Fry ___ ___ ___ ___ 38-64 46.3 38

Fingerlings — — — — — — — — — — — — 72-90 81.0 3

17 Oct.:

Fry ___ ___ ___ ___ 45-69 59.5 159

Fingerlings — __ ___ ___ ___ 70.97 79.8 27

1968:

2 July:

Fingerlings 56-96 74.6 110 59-103 80.0 98 59-107 83.2 104 56-107 79.2 312 57-103 77.9 74

1 Aug.:

Fingerlings 59-101 76.9 68 57-104 82.2 80 66-120 81.3 85 57-120 80.3 233 59-107 82.4 62
20 Sept.:

Fingerlings 79-100 88.7 22 — — — _ _ _ 79-100 88.7 22 _ _ _

'Dates given for 1965, 1966, 1967, and 1966 measurements are middates of the measuring period.

mm as age II smelts, 151 mm as age III smelts, and Sashin Creek produce 40- to 70-mm coho salmon
175 mm as age IV smolts (Drucker 1972). Kam- fry by September (Gribanov 1948), 85-mm age I
chatka streams at about the same latitude as smolts, and 130-mm age II smolts (Semko 1954). In

912

CRONE and BOND: LIFE HISTORY OF COHO SALMON

California (Shapovalov and Taft 1954) and British
Columbia (Foerster and Ricker 1953), the mean
lengths of coho salmon smolts (mostly age I)
usually ranged from 110 to 120 mm.

In coho salmon, attaining the smolt stage is
apparently more a function of size than age. Data
on lengths and numbers of juvenile coho salmon in
Sashin Creek during September and early summer
suggest that most require two summers of fresh-
water residence to reach smolt size. Coho salmon
can grow much faster; some juveniles in a brackish
pond in Oregon grew from about 40 mm (890 to the
pound) to about 120 mm and became smolts in only
3 mo instead of the usual 1 yr (Garrison 1965).

The growth of juvenile coho salmon in Sashin
Creek varies from year to year. During summer
1966, for instance, fry were larger than in 1964,
1965, and 1967 (Figure 8). In the summers of 1966
and 1968, fingerlings (mainly age I) were larger
than in 1965 and 1967 (mainly age I). The number
of fingerlings in 1966 and 1968^ (*3,000 on 1 July)
was less than in either 1965 (=5,000 on 1 July) or
1967 (*3,500 on 1 July), and less competition for
food would be expected and could account for the
larger size of the fingerlings in 1966. Also, the
presence of fewer fingerlings in summer 1966 may
have allowed the fry to reach a larger size because
of less competition for food or space. Food abun-
dance, controlled by factors other than coho salm-
on population size, may have an important
influence on coho salmon growth in Sashin Creek.
We have no information on possible year-to-year
differences in food supply independent of fish
populations which could result in differences in
growth of juvenile coho salmon.

"•An estimate of 2,960 fingerlings in Sashin Creek was made on
2 July 1968.

Survival from Potential Egg Deposition
to Emergence

The estimated potential egg depositions for
brood years 1963, 1964, and 1965 were determined
by multiplying the mean fecundity (determined
from 1966, 1970, and 1971 samples) by the es-
timated number of females (one-half of the es-
timated population of spawners). These estimates
are considered to be only rough approximations:
1,460,000 for 1963; 260,000 for 1964; and 350,000 for
1965.

Estimates of the numbers of preemerged salm-
on alevins in the streambed were obtained in the
early spring by hydraulic streambed sampling
(McNeil 1964). In Sashin Creek this sampling is
done to estimate the number of pink salmon
alevins, but after relatively large escapements of
coho salmon reliable estimates of the number of
coho salmon alevins in the streambed also can be
made. No coho salmon alevins were found during
the hydraulic streambed sampling in the spring of
1966, so the estimate of the alevin population was
zero. Because many age 0 fry were in the stream in
the summer of 1966, we have estimated the
number of alevins that were in the gravel that
spring by interpolation of the survivorship curves.

The numbers of preemerged coho salmon
alevins for 1964-66 estimated from the results of
hydraulic sampling or interpolation of survivor-
ship cur\^es are: 214,000 in spring 1964 (1963 brood
year), 58,000 in spring 1965 (1964 brood year), and
100,000 in spring 1966 (1965 brood year). From
these figures and estimates of potential egg depo-
sition, we calculated survival from potential egg
deposition to just before fry emergence to be 15%,
22%, and 26% for the 1963, 1964, and 1965 brood
years, respectively.

Figure &-Mean fork lengths of coho
salmon fry measured several times each
summer, 1964-67, and resulting
fingerlings (*90% age I and 10% age II)
the next summer, 1965-68, Sashin
Creek.

JUNE

913

FISHERY BULLETIN: VOL. 74, NO. 4

Survival of Juveniles in Sashin Creek

We estimated the population periodically during
the summers from 1964 through 1967 to establish
curves depicting changes in the number of
juvenile coho salmon by brood year during their
freshwater life. In 1964 the numbers of fry were
estimated in July and August. In 1965 the
numbers of fingerlings (predominantly age I, the
balance age II) and fry were estimated in July and
August. In 1966 the numbers of fingerlings and fry
were estimated in June, July, and September. In
1967 estimates were made of coho salmon
fingerlings and fry in Sashin Creek in July and in
Funny Creek in July and August.

Juvenile salmon in the stream were captured by
a combination of baiting and seining. A
homogenized mixture of salmon eggs, ovarian
tissue, and water was prepared with an electric
blender and injected into the stream at the seining
site (Figure 9). Underwater observations indicated
that several squirts of the egg solution from a
plastic squeeze bottle were adequate to attract
rainbow trout, Dolly Varden, coastrange sculpins,
and coho salmon fingerlings and large fry from at
least 30 m downstream. The downstream sides of
gravel bars, logs, and rocks were chosen as col-
lecting sites because these obstructions formed
slow-water areas in which the bait would linger for
several minutes. In some instances it was neces-
sary to construct a rock barrier to divert the
current and create a suitable site. During early
summer, when coho salmon fry are quite small,
they congregate along the shallow edges of the
stream and in backwaters. These small fry will not
travel far in response to bait, and we often had to
seine for them along the stream edges and back-
waters near the baiting site.

Captured fish were anesthetized with MS-222
Sandoz"' and marked by removing part of one fin. A
different fin clip was used for each marking date
within a summer. When they recovered from the
anesthetic, the marked fish were released at the
collection site.

To allow the marked juveniles to become redis-
tributed, we did not begin to recapture them until
several days after they were marked. To reduce
bias in the population estimate, we selected ran-
dom points as seining sites during the recapture
portion of the e.xperiment. Random numbers
between 0 and 99 were chosen from a table of

^Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

random numbers (Snedecor 1956) for each of the
30.5-m (100- foot) sections of stream. The numbers
chosen represented the distance in feet down-
stream from the origin of each section to the sites
that would be baited. These distances were paced
off, and one or several places across the width of
the stream at the site were baited. A site was
repeatedly baited and seined until only a few fish
could be taken in each seine haul. All fish captured
at a site were anesthetized and examined for
marks. When they recovered from the anesthetic,
the fish were released. The numbers of unmarked
and marked coho salmon juveniles were recorded
for each site.

The Bailey-Petersen mark-and-recapture
method (Ricker 1958) was used to make all
population estimates, e.xcept in August 1964 when
a Schnable multiple mark-and-recapture method
(Ricker 1958) was used for fry. In 1966 and 1967 the
numbers of juveniles to be marked and recaptured
were predetermined to obtain preassigned levels
of accuracy and precision of population estimates
(Robson and Regier 1964). We tried to mark and
recapture enough fish to be 95'>? certain that the
error in estimating the population was not more
than lO'^r (Table 15). Confidence limits to popula-
tion estimates were obtained by methods given by
Ricker (1958) and Robson and Regier (1968).

The number of coho salmon fry decreased
greatly between the first and second estimate
(Tables 15, 16). In the month between estimates,
the population dropped by 71^^ in 1964, 78*^ in 1965,
and 62*^ in 1966. Weir counts of emigrant coho
salmon fry, which were continued until mid-
August in 1964, showed that only about 2,000 fry
(4*^ of the first population estimate) left Sashin
Creek between mid-July and mid-August popula-
tion estimates. Fyke net catches indicated that
even fewer fry migrated from the stream in 1965
and 1966 than in 1964. Therefore, we attribute the
large decrease in number of fry each year to
mortality rather than emigration. Observation of
the activities of fish and avian predators led us to
believe that predation probably accounted for the
major portion of the mortality. The number of
fingerling coho salmon also decreased as the
summer progressed, although not as rapidly as the
number of fry (Table 17).

The fry population was greater in 1964 than in
1965, 1966, or 1967 as a result of the large number
of spawners entering Sashin Creek in the fall of
1963. The population of fingerlings in 1965 was also
greater than in 1966 or 1967; the fingerlings in 1965

914

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Figure 9.-Underwater views of technique used to capture juvenile coho salmon, Sashin Creek. Bait is a blended

mixture of salmon eggs, ovarian tissue, and water.

915

FISHERY BULLETIN: VOL, 74, NO. 4

Table 15.- Population estimates of juvenile coho salmon from mark-and-recapture experiments in Sashin Creek and Funny Creek,

1964-67.

Fin

Census

Number of

Middate of

clip

Number

sample

marked fish

Population

95% confidence

estimate

Location

used'

Age-group

marked

size

recaptured

estimate

interval

1964:

12 July

Sashin Creek

LV&RV

Fry

1,454

4,421

123

51,852

43,939-62,216

12 Aug.

Sashin Creel<

ULC

Fry

1,475

1,929

174

15,185

12,530-17,840

1965:

17 July

Sashin Creek

Fry

1,801

485

20,355

14,849-28,039

17 July

Sashin Creek

Fingerlings

510

520

4,581

3,659- 5,960

1 Aug.

Funny Creek

LLC

Fry

276

221

1,201

900- 1,601

1 Aug.

Funny Creek

LLC

Fingerlings

213

107

852

565- 1,283

11 Aug.

Sashin Creek

Fry

847

804

149

4,546

3,965- 5,294

11 Aug.

Sashin Creek

Fingerlings

949

581

143

3,836

3,356- 4,459

11 Aug.

Funny Creek

Fry

221

244

984

762- 1,300

11 Aug.

Funny Creek

Fingerlings

106

141

557

370- 841

1966:

27 June

Sashin Creek

ULC

Fry

2,263

2,541

163

35,077

30,436-40,951

27 June

Sashin Creek

ULC

Fingerlings

332

520

2,883

2,312- 3,731

27 June

Funny Creek

ULC

Fry

716

509

225

1,616

1,474- 1,793

27 June

Funny Creek

ULC

Fingerlings

160

251

187- 328

29 July

Sashin Creek

LLC

Fry

3,002

2,957

660

13,434

12,584-14,394

29 July

Sashin Creek

LLC

Fingerlings

816

817

420

1,585

1,488- 1,701

29 July

Funny Creek

LLC

Fry

208

338

1,037

816- 1,300

29 July

Funny Creek

LLC

Fingerlings

223

257

147

389

354- 442

8 Sept.

Area U^

Anal

Fry

227

378

915

757- 1,081

8 Sept.

Area U^

Anal

Fingerlings

155

468

293- 755

8 Sept.

Funny Creek

Anal

Fry

287

314

146

615

552- 700

8 Sept.

Funny Creek

Anal

Fingerlings

110

100

317

221- 451

1967:

23 July

Sashin Creek

ULC

Fry

1.604

1,015

131

12,346

10,616-14,604

23 July

Sashin Creek

Fingerlings

1,431

890

418

3,043

2,848- 3,274

23 July

Funny Creek

ULC

Fry

202

213

801

622- 1,092

23 July

Funny Creek

Fingerlings

289

359

206

503

459- 551

17 Aug.

Sashin Creek

Anal

Fry

996

—

17 Aug.

Sashin Creek

Anal

Fingerlings

807

—

17 Aug.

Funny Creek

Anal

Fry

249

158

2,084

1,383- 3,381

17 Aug.

Funny Creek

Anal

Fingerlings

354

1,239

745- 3,218

'LV, RV, ULC,

LLC, and Anal refer

to left pelvic.

right pelvic, up[

ser lobe of caudal, lowe

r lobe of Cauda

, and anal fin cl

ps, respectively.

'Estimates of population size in the whole of Sashin Creek were not made.

were mainly progeny of the abundant 1963
spawners.

Variations in average annual streamflow have
been shown to affect significantly the number of
juvenile coho salmon in Washington streams
(Smoker 1953), but in Sashin Creek, other factors
such as parent escapement, original number of
coho salmon fry, and competition probably have
more influence on determining the number of
juvenile coho salmon. Sashin Creek is located in an
area of heavy rainfall that has small variations in
the annual total precipitation and annual average
stream discharge. From 1964 through 1967, annual
precipitation ranged from 546 to 643 cm. Greater
variations in average stream discharge for a
specific month occur from year to year. Annual
variations in stream discharge during the 1-mo
period in midsummer when populations of juvenile
coho salmon decreased most rapidly do not appear
to be correlated with the rates of population
decline (Table 17).

In 1965, 1966, and 1967, when estimates of
juvenile coho salmon populations were made in
each study area, the highest densities of coho

salmon in Sashin Creek usually occurred in the
lower study area, which is characterized by slow
water. Densities of coho salmon fry and
fingerlings were even higher in Funny Creek,
another slow-water habitat (Table 18).

Funny Creek was unique in our study areas in
that the populations of juvenile coho salmon
sometimes increased during the summer. The
estimated number of coho salmon fingerlings
increased from 251 to 389 between late June and
late July in 1966 (1964 brood year) and from 503 to
1,239 between mid-July and mid-August in 1967
(1965 brood year); the number of fry increased
from 801 to 2,084 between mid-July and mid-
August in 1967 (1966 brood year) (Table 16). The
95*% confidence interval estimates (Table 15) in-
dicate that the populations did increase, and the
additional coho salmon juveniles must have immi-
grated to this area from Sashin Creek. On all other
occasions, in both streams the populations of fry
and fingerlings decreased between estimates. The
movement of juvenile coho salmon from Sashin
Creek into Funny Creek during midsummer sug-
gests the use of this small tributary stream as a

916

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Table 16.-Population estimates of juvenile coho salmon of brood years 1963-66' in Sashin Creek and Funny Creek in the summers of

1964-67. Separate estimates are included for three areas of Sashin Creek.

Number

of fish (by

brood year)

Brood year 12 July

12 Aug.

17 July

1 Aug.

11 Aug.

27 June

29 July

8 Sept.

23 July

17 Aug.

and area

1964

1965

1966

1967

1963:

Sashin Creek:

Upper

—

668

—

593

Middle

—

1,216

—

1,115

Lower

—

2,533

—

2,079

Stream estimate

51,852

15,185

4,581

—

3,836

Funny Creek

—

852

557

1964:

Sashin Creek:

Upper

2,979

—

254

402

509

468

Middle

2,195

—

951

690

555

—

Lower

14,738

—

3,477

1,562

555

—

Stream estimate

20,355

—

4,546

2,883

1,585

21,350

Funny Creek

—

1,201

984

251

389

317

1965:

Sashin Creek:

Upper

1,192

1,497

915

810

—

Middle

7,759

5 091

—

801

—

Lower

26,662

7;851

—

1,453

—

Stream estimate

35,077

13,434

28,000

3,043

—

Funny Creek

1.616

1,037

615

503

1,239

1966:

Sashin Creek:

Upper

1,828

—

Middle

1,862

—

Lower

8,627

—

Stream estimate

12,346

—

Funny Creek

801

2,084

iThp pQtimatprI r

inniilation

s of fish of a

brood vear a

it aae 1 (seco

nd summer

of life) inc

lude an ave

rage of 11%

age II fish

from the

preceding brood year.

^Estimate of population from expansion of estimated populations in upper area and Funny Creek.

Table 17.-Mean stream discharge and percentage decrease in
numbers of juvenile coho salmon between a first estimate (late
June to late July) and a second estimate (mid-July to mid-
August) in Sashin Creek, 1964-67.

Mean stream

discharge

(mVs)

Decrease m

population size

Year

Fry

Fingerlings

1964

'1.70

71%
(51,852 to 15,185)

(2)

1965

0.99

78%

16%

(20,355 to 4,546)

(4,581 to 3,836)

1966

0.82

62%

45%

(35,077 to 13,434)

(2,883 to 1,585)

1967

1.42

{')

(12,346)

(3,043)

'Stream discharge data for August 1964 not measured. An esti-
mate of mean stream discharge for the period was calculated
from July 1964 stream discharge and rainfall data in conjunction
with the August 1964 rainfall pattern.

^Size of fingerling population not estimated in 1964.

^First population estimate; second population estimate was not
completed.

feeding area or refuge from undesirable condi-
tions in Sashin Creek, such as competition or
predation. Fall migration of juvenile coho salmon
into small tributary streams in Oregon has been
reported (Skeesick 1970).

Estimates of the number of coho salmon fry and
fingerlings were used to construct curves depict-

ing the changes in the sizes of the populations of
three of the brood years studied (Figures 10, 11).
Estimates of the total number of fry and
fingerlings in Sashin Creek in early September
1966 are projected from estimates of population
size obtained in the upper area and in Funny
Creek. In these two study areas in early Sep-
tember the number of fry averaged 60% and the
number of fingerlings 85% of their populations in
late July. We assumed that these percentages
pertained also to the lower and middle areas of
Sashin Creek.

Survival and Instantaneous Mortality Rates

We compared survival and instantaneous mor-
tality rates of juvenile coho salmon of three brood
years by dividing their freshwater lives into the
following five periods between the time of egg
deposition and late in the second summer of life:

Period Time covered

1 Egg deposition to just before emergence
(mid-October to late March or early
April).

917

FISHERY BULLETIN: VOL. 74, NO. 4

Table 18.-Densities of juvenile coho salmon by brood year (1963-66)' on dates of population estimates in Sashin Creek and Funny

Creek. Separate estimates are included for three areas of Sashin Creek.

Density of fish (per

square meter)

Brood year

12 July

12 Aug.

17 July

1 Aug.

11 Aug.

27 June

29 July

8 Sept.

23 July

17 Aug.

and area

1964

1965

1966

1967

1963:

Sashin Creek:

Upper

—

0.16

—

0.15

Middle

—

0.27

—

0.25

Lower

—

0.31

—

0.26

Stream estimate

3.13

0.92

0.28

—

0.23

Funny Creek

—

1.93

1.26

1964:

Sashin Creek:

Upper

0.74

—

0.06

0.10

0.13

0.12

Middle

0.49

—

0.21

0.16

0.12

—

Lower

1.83

—

0.43

0.19

0.07

—

Stream estimate

1.23

0.27

0.17

0.10

'0.08

Funny Creek

—

2.72

2.23

0.57

088

0.72

1965:

Sashin Creek:

Upper

0.29

0.37

0.23

0.20

—

Middle

1.75

1.15

—

0.18

—

Lower

3.31

0.97

—

0.18

—

Stream estimate

2.12

0.81

20.48

0.18

—

Funny Creek

3.66

2.35

1.39

1.14

2.81

1966:

Sashin Creek:

Upper

0.45

—

Middle

0.42

—

Lower

1.07

—

Stream estimate

0.75

—

Funny Creek

1.82

4.73

'The estimated populations of fish of a brood year at age I (second summer of life) include an average of 11% age II fish from the pre-
ceding brood year.

^Estimate of density calculated from population obtained from expansion of estimated populations in upper area and Funny Creek.

2 Just before emergence to first estimate of
fry population (end of period 1 to late
June or mid-July).

3 First to second estimate of fry population
during first summer (end of period 2 to
late July or mid-August).

4 Second estimate of fry population to first
estimate of population as yearlings (end
of period 3 to late June or mid-July of the
following year).

5 First to second estimate of yearling
population (end of period 4 to late July or
mid-August).

Although the lengths of the corresponding periods
for the three brood years are similar, they varied
according to vv^hen the population estimates were
made.

We compared the estimates of the population at
the end of each of the five periods with the original
population (potential egg deposition) to obtain
percentage survival during nearly 2 yr of their
freshwater life for the brood years 1963-65 (Table
19). Survival from potential egg deposition to just
before fry emergence (period 1) was estimated to

Table 19.— Survival through five periods' in the freshwater life
of three brood years of coho salmon in Sashin Creek, expressed as
a percentage of potential egg deposition.

Percentage

survival th

rough period

year

1963
1964
1965

14.66
22.31
25.71

3.55

7.83

10.02

1.04
1.75
3.84

0.28
0.99
0.77

0.23
0.54

See text for explanation of time covered in each period.

be 15%, 22%, and 26% (mean of 21%) for the 1963,
1964, and 1965 broods, respectively. Other inves-
tigators have found similar survival to emergence
for coho salmon. A range in survival to emergence
in terms of counted fry of 11.8% to 40.0% (mean of
21.0% and 26.5%) is reported for two tributaries of
the Cowichan River, British Columbia (Pritchard
1947). Koski (1966) obtained survival values to fry
emergence of 0% to 78% (mean of 27.1%) for
individual redds of coho salmon in three streams
tributary to Drift Creek. For Karymaisky Spring
on Kamchatka, Semko (1954) reports survival to
emergence of 0.8% to 21.4% (mean 10.6%).

Because the lengths of the five periods were not
equal and a specific period was not the same length

918

CRONE and BOND: LIFE HISTORY OF COHO SALMON

1,500 -

l.fOO

1,300

1,200

1,100

1,000

100 -

POTENTIAL ECC DEPOSITION

T — I—I — I — I — I — I — I — I — I 1 — I — I — r

ASONDJFMAMJJAS

PREEMERCED ALEVINS

FRY

YEARLINGS

1961 BROOD ^o.

1 — I — I — I — I — I — I — I — r"T —

~r I I I I I I I I I I I I I I I I I I r
OCT. NDJ FMAMJJA SONDJ FMAMJ JAS

MONTH

Figure lO.-Estimated populations of juvenile coho salmon of
three brood years, Sashin Creek, from potential egg deposition to
late summer of second year. (Arithmetic plot.)

for each of the three brood years, instantaneous
mortality coefficients (Ricker 1958) were computed
to compare mortality for the periods and years
(Table 20). The equation for determining the
instantaneous mortality coefficient,

M,, =

-In (^„ )

follows the notation of McNeil (1966), where t, the
interval of time, is in months (one unit is equal to 1
mo); the symbol In represents the natural loga-
rithm; j is the brood year; and n is the study period.

2,000 -

1,000
800

600
500

too

, 300

= 200

I-
<

a.
O

I-

100
80

60
50
W
30

, POTENTIAL ECC DEPOSITION

1965
'\ ^'^°°°\ PREEMERCED ALEVINS

1961 \

2 -

YEARLINCS

-I I I L

SEPT

ONDJ FMAMJ J ASONDJ FMAMJ J AS
MONTH

Figure 11. -Estimated populations of juvenile coho salmon of
three brood years, Sashin Creek, from potential egg deposition to
late summer of second year. (Semilogarithmic plot to indicate
mortality rate.)

Table 20.-Instantaneous mortality coefficients during five
periods^ in the freshwater life of three brood years of coho
salmon in Sashin Creek.

Brood
year

Instantaneous mortality coefficient in period

1963
1964
1965

0.37
0.27
0.25

0,38
0.30
0.32

1.23
1.88
0.87

0.12 0.25

0.05 0.55

0.14 —

'See text for explanation of time covered in each period.

S,, is the survival within the nth period and is
calculated from the formula,

Si-52* . . -'^ = S,

Sn =

Si'S2' . . .'S(n - U

where S is survival through n study periods ex-
pressed as a percentage of potential egg deposi-
tion (Table 19).

Instantaneous mortality was higher through

919

FISHERY BULLETIN: VOL. 74, NO. 4

periods 1 and 2 for the 1963 brood than for the 1964
and 1965 broods. We estimated that eggs of the
1963 brood were over four times as abundant as
those of the 1965 brood and over five times as
abundant as those of the 1964 brood. Resulting
density-dependent factors such as superimposi-
tion of redds, selection of inferior redd sites, and
emigration of fry from the stream because of lack
of living space could be the cause of the higher
initial mortality of the 1963 brood.

For all three broods the highest instantaneous
mortality occurred in period 3-between the first
and second population estimates of the first sum-
mer of life-during July and the first half of
August (Table 20). Predation from fishes (both
intraspecific and interspecific) is thought to be a
major cause of this high mortality. In period 2 the
fry live in the backwater and shallow edges of the
stream where larger piscivorous fish do not
regularly occur. During period 3 the fry move into
deeper parts of the main channel where current is
still relatively slow, but here larger fish occur and
the fry may be more subject to predation.

Instantaneous mortality during the winter
(period 4) was much less than that of the first
summer. Predation probably was less during this
period for two reasons: 1) in winter the feeding
rate of cold-blooded predators is slowed, and 2)
restricted access because of ice and snow and
lowered activity and availability of the juvenile
coho salmon combine to lessen the hunting success
of warm-blooded predators.

Mortality increased during the second summer
(period 5) but only to a third or less of the corres-
ponding part of the first summer (period 3).

Some of the estimated mortality of fry and
fingerlings might have been due to undetected
emigration from the creek. When the fry weir or
fyke nets were fished in summer (periods 3 and 5),
however, only a few fry and fingerlings emigrated
and the low mortality rate in period 4 also suggests
that only a few fry emigrated in fall and winter.
Some age I smolts probably migrated from the
stream in the spring of period 4 in each of the 3 yr
studied. The drop in population of a brood year due
to age I smolt emigration is included as part of
period 4 mortality. Age composition of smolts in
1965, 1966, and 1967 was not determined. The age
composition of returning adults in 1966 and 1967
(25% and 29% age So, respectively) indicates that
some age I smolts emigrated in the spring of 1965
and 1966. In 1968 the smolts were sampled for age
composition; about 500 yearling smolts migrated.

920

Scale samples for age analysis were not collected
from adults in 1968.

SUMMARY

The number of adult coho salmon that enter
Sashin Creek varies from year to year. Coho
salmon have been counted at the weir as they
enter Sashin Creek each year since 1934, but this
count has usually been incomplete.

Several methods were used to estimate coho
salmon escapements to Sashin Creek for the years
1963-65 and 1967. These included weir counts,
adults on spawning riflfles, mean redd life, and
marked-to-unmarked ratios of spawners. The last
system produced the most accurate estimates,
resulting in 916, 162, 221, and 370 salmon for the
respective study years.

In the 4 yr that spawning ground studies were
made, the density of coho salmon on the spawning
grounds in Sashin Creek tended to be greater in
the middle and lower study areas than in the upper
area.

The effect of coho salmon spawning on the
survival of pink salmon embryos was insignificant
in 1965 relative to the population ratios of coho and
pink salmon present. Significant numbers of pink
salmon embryos might be killed if relatively large
numbers of coho salmon utilized Sashin Creek for
spawning.

The 4.3 age-group of coho salmon made up 78%,
59%, 64%, and 62% of the adults that returned to
Sashin Creek in 1965, 1966, 1967, and 1969-higher
percentages of this age group than reported for
most other streams. In California, Oregon, Wash-
ington, and southern British Columbia, adult coho
salmon are almost exclusively age 82. Studies of
growth and scales of fry, fingerlings, and smolts
and estimates of the population sizes of juveniles
indicate that most coho salmon remain in Sashin
Creek for two summers and winters.

In some years, substantial numbers of coho
salmon fry enter the estuary of Sashin Creek
shortly after emergence. These fry were tested
and found to be able to survive in salinities
encountered in the inner bay of the Little Port
Walter estuary. However, analysis of scales of
adult coho salmon returning to Sashin Creek
revealed none that had migrated to the estuary at
the fry stage, suggesting no fry (or at best very
few) that migrate to the ocean survive to return as
adults. This agrees with studies of other stocks of
coho salmon.

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Estimates of populations of fry in the early
summer for the 4 yr studied ranged from about
12,000 to 52,000, and apparently varied directly
with potential egg deposition of the brood year.
However, by early in the second summer of fresh-
water life, the three broods studied had been
reduced to a relatively narrow range of 3,000 to
4,500. Weir counts indicate 1,000 to 3,000 coho
salmon smolts migrate from Sashin Creek each
year.

The survival of coho salmon from potential egg
deposition to just before the emergence of fry in
Sashin Creek averaged 21%; this percentage is
similar to survival reported for stocks from other
areas in the eastern Pacific. Mortality of embryos
and alevins was highest for the large 1963 brood,
which suggests that some of the mortality before
emergence was due to compensatory factors such
as selection of inferior redd sites and superimpo-
sition of redds.

Highest mortality during the freshwater life of
coho salmon from Sashin Creek occurred in July
and early August of the first summer in all three
broods studied. The lowest mortality occurred over
winter.

ACKNOWLEDGMENTS

We thank William J. McNeil, Robert J. Ellis, and
William R. Heard, supervisors of the Little Port
Walter research station during the course of this
study, for their direct assistance in many forms
and for making manpower available for conduct-
ing the field work. Sincere appreciation is ex-
pressed to the permanent and temporary staff at
Little Port Walter for assistance in the field.

LITERATURE CITED

Allen, G. H.

1958. Notes on the fecundity of silver salmon
(Oncorhynchus kisutch). Prog. Fish-Cult. 20:163-169.
Andrews, R. E.

1961. Inventory and cataloging of the sport fish waters in
the Cook Inlet and Copper River drainages. Alaska Dep.
Fish Game, Sport Fish Div., 1960-61 Dingell-Johnson Proj.
Rep. 2:47-55.

1962. Inventory and cataloging of sport fish and sport fish
waters of the Cook Inlet drainages. Alaska Dep. Fish
Game, Sport Fish Div., 1961-62 Dingell-Johnson Proj. Rep.
3:221-225.

Armstrong, R. H.

1970. Age, food, and migration of Dolly Varden smolts in
southeastern Alaska. J. Fish. Res. Board Can. 27:991-1004.
Briggs, J. C.

1953. The behavior and reproduction of salmonid fishes in a

small coastal stream. Calif. Dep. Fish Game, Fish Bull. 94,
62 p.
Chamberlain, F. M.

1907. Some observations on salmon and trout in Alaska.
[U.S.] Bur. Fish., Rep. Comm. Fish. 1906. Doc. 627, 112 p.
Chapman, D. W.

1962. Aggressive behavior in juvenile coho salmon as a cause
of emigration. J. Fish. Res. Board Can. 19:1047-1080.

1965. Net production of juvenile coho salmon in three
Oregon streams. Trans. Am. Fish. Sec. 94:40-52.

CoNTE, F. P., H. H. Wagner, J. Fessler, and C. Gnose.

1966. Development of osmotic and ionic regulation in
juvenile coho salmon Oncorhynchus kisutch. Comp. Bio-
chem. Physiol. 18:1-15.

Drucker, B.

1972. Some life history characteristics of coho salmon of the
Karluk River system, Kodiak Island, Alaska. Fish. Bull.,
U.S. 70:79-94.
Engel, L.J.

1965. Egg take investigations in Cook Inlet drainage and
Prince William Sound. Alaska Dep. Fish Game, Sport
Fish Div., 1964-65 Dingell-Johnson Proj. Rep. 6:155-163.

1966. Egg-take investigation in Cook Inlet drainage and
Prince William Sound. Alaska Dep. Fish Game, Sport
Fish Div., Fed. Aid Fish Restoration, 1965-66 Prog. Rep.
7:109-116.

1967. Egg take investigation in Cook Inlet drainage and
Prince William Sound. Alaska Dep. Fish Game, Sport
Fish Div., Fed. Aid Fish Restoration, 1966-67 Prog. Rep.
8:111-116.

Foerster, R. E.

1944. Appendix IV. Report for 1943 of the Pacific Biological
Station, Nanaimo, B.C. Annu. Rep. Fish. Res. Board Can.
1943:22-26.
1955. The Pacific salmon (genus Oncorhynchus) of the
Canadian Pacific coast, with particular reference to their
occurrence in or near fresh water. Int. North Pac. Fish.
Comm., Bull. 1:1-56.
Foerster, R. E., and A. L. Pritchard.

1936. The egg content of Pacific salmon. Biol. Board Can.,
Prog. Rep. Pac. Biol. Stn. Pac. Fish. Exper. Stn. 28:3-5.
Foerster, R. E., and W. E. Ricker.

1953. The coho salmon of Cultus Lake and Sweltzer Creek.
J. Fish. Res. Board Can. 10:293-319.
Garrison, R. L.

1965. Coho salmon smolts in ninety days. Prog. Fish-Cult.
27:219-220.
Gilbert, C. H.

1913. Age at maturity of the Pacific coast salmon of the
genus Oncorhynchus. Bull. [U.S.] Bur. Fish. 32:1-22. (Doc.
767.)
1922. The salmon of the Yukon River. Bull. [U.S.] Bur. Fish.
38:317-332. (Doc. 928.)
Godfrey, H.

1965. Salmon of the North Pacific Ocean-Part IX, Coho,
Chinook and masu salmon in offshore waters. 1. Coho
salmon in offshore waters. Int. North Pac. Fish. Comm.,
Bull. 16:1-39.
Gribanov, V. I.

1948. The coho salmon {Oncorhynchus kisutch Walbaum)-a
biological sketch. Izv. Tikhookean. Nauchno-issled. Inst.
Rybn. Khoz. Okeanogr. 28:43-101. (Fish. Res. Board Can.
Transl. Ser. 370, 83 p.)
Hunter, J. G.

1948. Natural propagation of salmon in the central coastal

921

FISHERY BULLETIN; VOL. 74, NO. 4

area of British Columbia. Fish. Res. Board Can., Prog.
Rep. Pac. Coast Stn. 77:105-106.
1949. Natural propagation of salmon in the central coastal
area of British Columbia. II. The 1948 run. Fish. Res.
Board Can., Prog. Rep. Pac. Coast Stn. 79:33-34.
International North Pacific Fisheries Commission.

1962. The exploitation, scientific investigation and man-
agement of salmon (genus Oncorhynchns) stocks on the
Pacific coast of the United States in relation to the
abstention provisions of the North Pacific Fisheries
Convention. Int. North Pac. Fish. Comm., Bull. 10. 160 p.

KosKi,K V.

1966. The survival of coho salmon (Oncorhiinchiix kixutch)
from egg deposition to emergence in three Oregon coastal
streams. M.S. Thesis, Oregon State Univ., Corvallis, 84 p.

KUBIK, S.

1967. Population studies of anadromous species with em-
phasis on upper Cook Inlet drainage. Alaska Dep. Fish
Game, Sport Fish Div., Fed. Aid Fish Restoration, 1966-67
Prog. Rep. 8:117-128.

Kuznetsov, 1. 1.

1928. Some observations on spawning of the Amur and

Kamchatka salmons. Izv. Tikhookean. Nauchno-issled.

Inst. Rybn. Khoz. Okeanogr. 2. (Fish. Res. Board Can.

Transl. Ser. 22, 9 p.)
Lawler, R. E.

1963. Silver salmon egg taking investigations in Cook Inlet
drainage. Alaska Dep. Fish Game, Sport Fish. Div.,
1962-63 Dingell-Johnson Proj. Rep. 4:161-173.

1964. Egg take investigation in Cook Inlet drainage and
Prince William Sound. Alaska Dep. Fish Game, Sport
Fish Div., 1963-64 Dingell-Johnson Proj. Rep. 5:123-132.

Logan, S. M.

1963. Silver salmon studies in the Resurrection Bay area.
Alaska Dep. Fish Game, Sport Fish Div., 1962-63 Dingell-
Johnson Proj. Rep. 4:175-194.

1968. Silver salmon studies in the Resurrection Bay area.
Alaska Dep. Fish Game, Sport Fish Div., Fed. Aid Fish
Restoration, 1967-68 Prog. Rep. 9:117-134.

Marr, J.C.

1943. Age, length, and weight studies of three species of
Columbia River salmon (Oiicorhynchus keta, O. gorhuscha,
0. ki.^utch). Stanford Ichthyol. Bull. 2:157-197.

Marriott, R. A.

1968. Inventory and cataloging of the sport fish and sport
fish waters in southwest Alaska. Alaska Dep. Fish Game,
Sport Fish Div., Fed. Aid Fish Restoration, 1967-68 Prog.
Rep. 9:81-93.

McGinnis, D.

1966. Inventory, cataloging and population sampling of the
sport fish and sport fish waters of the Cook Inlet
drainage. Alaska Dep. Fish Game, Sport Fish Div., Fed.
Aid Fish Restoration, 1965-66 Prog. Rep. 7:155-169.

McNeill, W. J.

1964. A method of measuring mortality of pink salmon eggs
and larvae. U.S. Fish. Wildl. Serv., Fish. Bull. 63:575-588.

1966. Distribution of spawning pink salmon in Sashin
Creek, southeastern Alaska, and survival of their
progeny. U.S. Fish Wildl. Ser\-., Spec. Sci. Rep. Fish. 538,
12 p.
Meehan, W. R., and D. B. Siniff.

1962. A study of the downstream migrations of anadromous
fishes in the Taku River, Alaska. Trans. Am. Fish. See.
91:399-407.

Merrell, T. R., Jr.

1962. Freshwater survival of pink salmon at Sashin Creek,
Alaska. In N. J. Wilimovsky (editor). Symposium on pink
salmon, p. 59-72. Univ. B.C., H. R. MacMillan Lect. Fish.,
Vancouver.

Neave, F.

1948. Fecundity and mortality in Pacific salmon. Trans. R.
Soc. Can., Ser. 3, Sect. 5, 42:97-105.
Olson, J. M., and W. J. McNeil.

1967. Research on pink salmon at Little Port Walter, Alaska,
1934-64. U.S. Fish Wildl. Serv., Data Rep. 17, 301 p. on 5
microfiche.

Otto, R. G.

1971. Effects of salinity on the survival and growth of
pre-smolt coho salmon (Oncorhynchns kisutch). J. Fish.
Res. Board Can. 28:343-349.
Parker, R. R., and R. E. Vincent.

[1956]. Progress report on research studies at the Kitoi Bay
Research Station. Alaska Fish. Board and Alaska Dep.
Fish. 19.55 Annu Rep. 1957:25-67.
Powers, C. F.

1963. Some aspects of the oceanography of Little Port
Walter estuary, Baranof Island, Alaska. U.S. Fish Wildl.
Sen-., Fish. Bull. 63:143-164.

Pritchard, a. L.

1940. Studies on the age of the coho salmon (Oncorhynchns
kisutch) and the spring salmon (Oncorhynchns tshawyts-
cha) in British Columbia. Trans. R. Soc. Can., Ser. 3, Sec. 5,
34:99-120.
1947. Efficiency of natural propagation of Pacific salmon.
Can. Fish Cult. 1:22-26.
Redick, R. R.

1968. Population studies of anadromous fish popula-
tions—southwestern Kenai Peninsula and Kachemak Bay.
Alaska Dep. Fish Game, Sport Fish Div., Fed. Aid Fish
Restoration, 1967-68 Prog. Rep. 9:135-155.

RiCKER, W. E.

1958. Handbook of computations for biological statistics of
fish populations. Fish. Res. Board Can., Bull. 119, 300 p.
RoBSON, D. S., AND H. A. Regier.

1964. Sample size in Petersen mark-recapture
experiments. Trans. Am. Fish. Soc. 93:21.5-226.

1968. Estimation of population number and mortality
rates. In W. E. Ricker (editor). Methods for assessment of
fish production in fresh waters, p. 124-158. IBP (Int. Biol.
Programme)., Handb. 3.
Rounsefell, G. a.

1957. Fecundity of North American Salmonidae. U.S. Fish
Wildl. Ser\-., Fish. Bull. 57:451-468.

Salo, E. 0., AND W. H. Bayliff.

1958. Artificial and natural production of silver salmon,
Oncorhynchns kisutch, at Minter Creek, Washington.
Wash. Dep. Fish., Res. Bull. 4, 76 p.

Semko, R. S.

1954. The stocks of west Kamchatka salmon and their
commercial utilization. Izv. Tikhookean. Nauchno-issled.
Inst. Rybn. Khoz. Okeanogr. 41:3-109. (Fish. Res. Board
Can. Transl. Ser. 288, 131 p.)

Shapovalov, L., and a. C. Taft.

1954. The life histories of the steelhead rainbow trout
(Salmo gairdneri gairdneri) and silver salmon
(Oncorhynchns kisutch) with special reference to Waddell
Creek, California, and recommendations regarding their
management. Calif. Dep. Fish Game, Fish Bull. 98, 375 p.

922

CRONE and BOND: LIFE HISTORY OF COHO SALMON

Skeesick, D. G.

1970. The fall immigration of juvenile coho salmon into a
small tributary. Oreg. Fish Comm. Res. Rep. 2:90-95.
Smirnov, A. I.

1960. The characteristics of the biology of reproduction and
development of the coho Oncorhynchus kisutch
(Walbaum). Vestn. Mosk. Univ., Ser. VI, Biol. Pochv.
1960(1):9-19. (Fish. Res. Board Can. Trans). Ser. 287.)
Smoker, W. A.

1953. Stream flow and silver salmon production in western

Washington. Wash. Dep. Fish. Fish. Res. Pap. 1(1):5-12.
[1957.] Silver salmon research. Alaska Fish. Board and
Alaska Dep. Fish. 1956 Annu. Rep. 8:42-46.
Snedecor, G. W.

1956. Statistical methods. 5th ed. Iowa State Univ. Press,
Ames, 534 p.
U.S. Geological Survey.

1972. Water resources data for Alaska, 1968. Part 1. Surface
water records. U.S. Dep. Inter. Geol. Surv., Water Res.
Div., 155 p.
Van Hulle, F. D.

1970. Inventory and cataloging of the sport fish and sport
fish waters in southwest Alaska. Alaska Dep. Fish Game,

Sport Fish Div., Fed. Aid Fish Restoration, 1969-70 Prog.
Rep. 11:47-63.
1971. Inventory and cataloging of the sport fish and sport
fish waters in southwest Alaska. Alaska Dep. Fish Game,
Sport Fish Div., Fed. Aid Fish Restoration. 1970-71 Prog.
Rep. 12:29-51.
Wallis, J.

1967. Salmonid rearing and migration study- Fire Lake
system and Fort Richardson cooling pond. Alaska Dep.
Fish Game, Sport Fish Div., Fed. Aid Fish Restoration,
1966-67 Prog. Rep. 8:129-152.

1968. Salmonid rearing and migration study- Fire Lake
system and Ft. Richardson and Elmendorf rearing
ponds. Alaska Dep. Fish Game, Sport Fish Div., Fed. Aid
Fish Restoration, 1967-68 Prog. Rep. 9:173-191.

WiCKETT, W. P.

1951. The coho salmon population of Nile Creek. Fish. Res.
Board Can., Prog. Rep. Pac. Coast Stn. 89:88-89.

Willis, R. A.

1954. The length of time that silver salmon spent before
death on the spawning grounds at Spring Creek, Wilson
River, in 1951-52. Oreg. Fish Comm. Res. Briefs 5:27-31.

923

VERTICAL DISTRIBUTION AND DIEL MIGRATION OF EUPHAUSIIDS
IN THE CENTRAL REGION OF THE CALIFORNIA CURRENT

Marsh J. Youngbluth^

ABSTRACT

The density, vertical range, and diel movement of total zooplankton and euphausiid populations in the
central region of the California Current were determined during a period of coastal upwelling,
July-August 1970. Collections were made along four transects with opening-closing Bongo nets towed
through 50- to 100-m intervals in the upper 800 m. Four- to nine-depth intervals at 13 day-night stations
were sampled. Twenty euphausiid species from seven genera were identified from 124 hauls.

Zooplankton assemblages in the nearshore regions differed from those farther offshore in having a
larger biomass as well as a smaller number and higher density of several species. Diel vertical
movement among euphausiid populations, particularly Euphansia pacifica, tended to be more
pronounced in offshore waters. This behavior suggests that, although assemblages of zooplankton are
strongly structured by physical factors, some species alter their vertical distribution and diel migration,
presumably in response to the prevailing food supply.

Since 1949 an intensive plankton sampling pro-
gram has been conducted in the California Cur-
rent under the auspices of the California Coop-
erative Oceanic Fisheries Investigations
(CalCOFI). These surveys, concentrating on the
distribution and density of pelagic organisms in
the upper 150 m, have revealed abundance and
dispersion patterns of zooplankton which are
related to annual and seasonal changes in hydro-
graphic conditions (Brinton 1960, 1962a; Thrailkill
1963; Fleminger 1964, 1967; Alvarino 1965;
McGowan 1967; Berner 1967; Isaacs et al. 1969).
Within a given year, varying proportions of zoo-
plankton assemblages typical of any one of several
water masses are likely to be present (Berner 1957;
Bieri 1959; Brinton 1962b; Johnson and Brinton
1963; Cushing 1971). Seasonal hydrographic
fluctuations near the coastal boundary of the
current act to further transform the numbers and
types of pelagic species that develop. For example,
the eutrophic environment produced by coastal
upwelling during the spring and summer months
is characterized by a much higher biomass and
lower species diversity than the more oligotrophic,
off"shore portion of the current (Frolander 1962;
Hebard 1966; Laurs 1967; Longhurst 1967; Pieper
1967).
Considerably less attention has been given to

'Hopkins Marine Station of Stanford University, Pacific
Grove, CA 93950; present address: Harbor Branch Foundation
Laboratory, RFD 1, Box 196, Fort Pierce, FL 33450.

Manuscript accepted March 1976.

FISHERY BULLETIN: VOL. 74. NO. 4, 1976.

the vertical distribution of zooplankton in this
current principally because it is difficult, time
consuming, and costly to repeatedly sample dis-
crete depths. The scope of this study was to
describe the vertical distribution and diel migra-
tion of zooplankton, particularly euphausiids, in
nearshore and offshore oceanic regions of the
central California Current during a period when
coastal upwelling was well developed. The samples
were collected in the summer of 1970 on two
cruises, Stanford Oceanographic Expedition
(SOE) cruise 22 and CalCOFI cruise 7008.

DESCRIPTION OF
THE ENVIRONMENT

The California Current is a blend of water
masses (Subarctic, North Pacific Drift, Central,
and Equatorial) and is therefore an extremely
variable environment (Reid et al. 1958; McGowan
1971). It flows southward throughout the year with
an average velocity of less than 0.5 knot. The
boundaries of this transitional water are between
lat. 48° and 23°N and extend to 700 km (long.
130° W) from the coast. Between depths of 200 and
400 m, a subsurface countercurrent flows north-
ward at about 0.5 knot from Baja California to
Cape Mendocino (Kin'dyushov 1970).

Near the coast, hydrographic fluctuations in this
current have been separated into seasonal periods
of divergence (upwelling), relaxation (oceanic),
and convergence (downwelling) (Bolin and Abbott

925

FISHERY BULLETIN; VOL. 74, NO. 4

1963; Dodimead et al. 1963). From March to
August, the force of the prevailing northerly
winds and the earth's rotation cause the southerly
flowing surface waters, within 100 km of the shore,
to move away from the coast (Yoshida 1955). This
displaced water is replaced by cooler, more saline,
nutrient-rich water upwelled from deeper regions,
providing favorable conditions for high rates of
primary production. In the central part of this
current, the upwelling period is considered to
begin and end with the shifting of the 9°C iso-
therm above and below the 100-m level (Barham
1957).

The northerly winds subside from September to
November and surface water temperatures in-
crease, resulting in the formation of a strong
thermocline in the upper 50 m. During this sur-
face-warmed period, in the absence of upwelling,
tongues of offshore, oceanic water of the Califor-
nia Current may reach the coast. Where this occurs
there is probably considerable mixing of oceanic
and neritic planktonic communities (e.g., see
Longhurst 1967).

When southerly winds prevail, in the period
from December to February, a northerly flowing,
coastal countercurrent (Davidson Current) may
develop. Surface water converges toward the coast
and disrupts the stratification characteristic of the
surface-warmed water period. Vertical eddy cir-
culation results, promoting the overturn, mixing,
and downwelling of warm, lower salinity, nu-
trient-poor surface waters. This mixed water
period can be characterized by a temperature
gradient of less than 1°C in the upper 50 m.

The environments sampled at the shoreward
and seaward stations in the summer of 1970
differed in several ways. Physical and chemical
features relating to phytoplankton studies during
the SOE cruise are presented in Malone (1971).
These and other hydrographic features in the
upper 800 m at each station are tabulated and
discussed in Youngbluth (1973). By way of sum-
mary, it was clear from the low temperatures and
high salinities and the shoreward elevation of
nitrate isopleths that upwelling conditions
prevailed near the coast. Chlorophyll-a values in
the upper 150 m decreased with increasing dis-
tances from shore, 2.1-0.5 mg/m\ The photic zone
was usually deeper at the seaward stations, rang-
ing from 55 m in coastal regions to 105 m at the
western edge of the transects. The depth of the
thermocline was shallower nearshore and deeper
offshore, ranging from 5 to 40 m, respectively. The

largest temperature difference between the ther-
mocline and 150 m was about 4°C. At depths below
150 m, temperatures differed by 2°C or less among
stations.

Temperature-salinity (T-S) curves from each
station were compared to two different schemes
(Youngbluth 1973). First, the data, when plotted
with T-S relationships that characterize the per-
cent mixing between waters near the northern
and southern limits of the California Current
(Okutani and McGowan 1969), indicated that
between 150 and 800 m 70-100% northern water
was present. The small percentage of southern
water was most noticeable at the intermediate and
nearshore stations of the southern transect.
Second, the data, when contrasted with T-S curves
that distinguish water masses, revealed that
samples below 250 m were collected in North
Pacific Intermediate water.

MATERIALS AND METHODS

Zooplankton samples were collected with open-
ing-closing Bongo nets of 0.333-mm mesh and cod
ends with 0.222-mm mesh (McGowan and Brown
1966). At nearly all stations, shallow and mid-
water casts were made, within 3 h of midday and
midnight at nearly the same location (Table 1).
Shallow tows were taken with a single frame
(SOE) or with four frames (CalCOFI) in the upper
150-200 m. Three (SOE) or four (CalCOFI) frames
were used on mid-water hauls between 200 and 600

Table l.-The date and position of Bongo net tows.

Date

Position

Cruise

1970

Station

Lat. N

Long. W

SOE 22

27 July

28 July

37^09'

124'24'

29 July

30 July

37-^15'

128°45'

31 July
1 Aug.

36 39'

130°53'

3 Aug.

4 Aug.

39^54'

129 58'

5 Aug.

6 Aug.

39=53'

127=48'

7 Aug.

8 Aug.

39 53'

125 = 48'

16 Aug.

18 Aug.

19 Aug.

20 Aug.

68
74

43''49'
43°55'

43°32'

125=49'
128 03'

129=58'

CalCOFI 7008

27 Aug.

28 Aug.

20 Aug.

21 Aug.

70.75
50.80

35^23'
38=40'

123=27'
126 21'

18 Aug.

19 Aug.

16 Aug.

17 Aug.

50.110
50.140

37^40'
36°40'

128=33'
130=44'

926

YOUNGBLUTH: VERTICAL DISTRIBUTION OF EUPHAUSIIDS

m. Depth intervals of about 100 m were sampled. A
single frame (CalCOFI) was employed at depths
from 600 to 800 m. The nets were hauled along a
single oblique path (all CalCOFI and shallow SOE
casts) or undulated obliquely through the depth
intervals sampled (all mid-water SOE casts). Each
point on the graphs representing these data is the
middepth of the water column sampled.

The strata sampled were recorded with a
Benthos depth-time device attached a few meters
below the bottom frame. Vessel speed during the
tows ranged between 2 and 2.5 knots (3.7 and 4.6
km/h) and was regulated to maintain a wire angle
of approximately 50°. Mean volumes of 619 m-^
(SOE shallow tows), 957 m'' (SOE mid-water
tows), and 546 m-^ (all CalCOFI tows) were filtered.
All data were standardized to a volume of 1,000 m^,
assuming 100% filtration efficiency. Clogging of
net apertures was observed only in the uppermost
nets at the nearshore stations on the CalCOFI
cruise.

The samples were preserved in 5% Formalin-
solution buffered to pH 7.6. All organisms longer
than 2 cm were removed from the sample and wet
weights were determined after draining the
remaining portion on a 0.222-mm mesh screen and
blotting it on absorbent paper for 20 min. Du-
plicate estimates varied by an average of 6%.

The larvae (furcilia), juveniles (postlarvae and
immatures), and adults (sexually mature) of all
euphausiid species were studied. All individuals of
the less abundant species were identified and
counted. The densities of the more numerous
species were determined from subsamples made
with a modified Folsom Plankton Splitter. The
average number of specimens examined in the
subsamples was about 300. Duplicate counts were
compared with each other by calculating a Percent
Similarity Index (Whittaker 1952).

If the index indicated at least 80% agreement
between the first two replicates, no other counts
were made. Occasionally a third count was
necessary.

The taxonomy of adult euphausiids follows
Boden et al. (1955). Identification of certain
difficult groups, e.g., Nematoscelis spp., Thysan-
oessa spp., and all larvae were verified by E.
Brinton, T. Antezana, and K. Gopalakrishnan at
the Scripps Institution of Oceanography. When
specimens lacked some of the usual key characters.

general body form and eye size, shape, and color
were used to distinguish the species.

RESULTS

Sampling Variability Between Cruises

Samples were collected along four transects.
The stations ranged from 130 to 693 km off the
coast (Figure 1). During the CalCOFI cruise, a
smaller average volume of water was filtered by
each net. Presumably this smaller volume could
have introduced some bias by reducing species
diversity and abundance estimates. Comparisons
of the results from each cruise indicate that,
except for three rarely caught species: 1) the
number of euphausiid species collected was iden-
tical and 2) the order of species abundances was
quite similar on each cruise. Biomass values of
total zooplankton tended to be larger at the
seaward stations during the CalCOFI cruise. This
difference is most likely related to the greater
number and narrower, vertical width of the tows
taken during this cruise, and, to some extent,
growth and development of each life stage as well

130

125^

120"

115^

I—

O SOE 22 27 JUL- 20 AUG 1970

• CALCOFI 7008 '6-27 Auo 1970

■ ■ I ■ ... I I l__l 1 i I 1 1 1 1 L-

-Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

130 125" 120 115

W. LONGITUDE

Figure 1. -Positions of the day-night stations in the central
portion of the California Current.

927

FISHERY BULLETIN: VOL. 74, NO. 4

as offshore transport of the more numerous species
observed in nearshore waters.

Distribution of Zooplankton Biomass

The standing stock of zooplankton was highest
at the nearshore stations and decreased with
distances seaward. The largest and smallest
biomass values occurred along the southern tran-
sect and corresponded with high and low phyto-
plankton stocks, respectively (Malone 1971).

Zooplankton were concentrated in the upper
100-150 m at most stations, particularly in the
mixed layer (ca. 0-40 m). Densities ranged from 10
to 580 g wet wt/ 1,000 m ' (Figure 2). Below 100-150
m, the amount of zooplankton at approximately
100-m intervals was generally between 25 and 150
g wet wt/ 1,000 m-'. Diel fluctuations in biomass
were greatest in the surface water (0-150 m)

LiJ

ZOOPLANKTON

(GM WET WT PER
50 100 50 100 150

BIOMASS

1000 M^)

50 100 150

STATION 81

100

300

■

STATION 74

500

STATION 68

STATION^ 39

— )

ST>^TION, 56 ,

100
300
500

STATION^ 47 ^

5
i

STATION 25
1 «

STATION 16

100
300

500

=>ST^TlON 9

100
300
500
100
300
500
100
300
500

(a)

ZOOPLANKTON BIOMASS

(GM wer WT PEO lOOO m3)

50 2^

450 . 550

1/1
CE

U
»-
UJ

I
t-

Q.
llJ

•oo

S'ATlON '0-^

300

500

(b)

Figure 2.-The vertical distribution of zooplankton biomass.
Clear bars indicate day samples; dark bars designate night
samples, (a) SOE cruise 22 and (b) CalCOFI cruise 7008.

increasing at night by factors of 1.2-8. At two
nearshore stations, CalCOFI 50.80 and 70.75, large
quantities of phytoplankton clogged net meshes in
the upper 50 m and prevented any quantitative
comparison of day and night catches. Below 150 m,
a small but consistent increase in biomass was
usually observed at night {P = 0.20, Sign Test).
During the day, at some intervals between 250 and
400 m, biomass was equal to or slightly greater
than concentrations in the upper 150 m.

Diversity, Density, and
Distribution of Euphausiid Species

Twenty species of euphausiids distributed
among seven genera were identified. Thirteen
species occurred frequently enough and in
sufficient numbers to allow descriptions of their
vertical distribution. In the upper 150 m, 8 species
formed 50% of the total abundance and 12 made up
90%. Nine species were found at more than half the
stations. In the total water column sampled (ca.
0-700 m), 6 and 11 species composed 50 and 90% of
the total species abundance. At most stations one
or two species were numerically dominant.

The distributions of euphausiids at midday and
midnight are discussed in the following para-
graphs. Only examples of a few species are illus-
trated to represent the major patterns observed
since a large number of profiles were derived from
the data for all the species collected at each station
(Youngbluth 1973). In many cases, diurnal changes
in vertical distributions were obscured either by
patchiness or avoidance or incomplete sampling
due to gear failure or foul weather. This account is
thus a composite description of the data from all
stations.

Euphausia

Four species of Euphausia were takeni-E".
pacifica, E. recurva, E. gibboides, and E. mutica.
With the exception of E. pacifica, these species
were only abundant at the offshore stations along
the southern transects (SOE 16, 25; CalCOFI
50.110). In this region densities of each species
usually ranged between 10 and 200/1,000 m^.
Juveniles and larvae were often more than twice
as numerous as adults. The daytime habitat of E.
mutica larvae was between 100 and 400 m.
Juveniles of this species were found only in one
haul which sampled from 400 to 500 m (SOE 16).
Euphausia gibboides and E. recu rva were collected

928

YOUNGBLUTH: VERTICAL DISTRIBUTION OF EUPHAUSIIDS

somewhat deeper, 250-350 m and 400-600 m, re-
spectively. At night all stages of these species
migrated into the upper 100 m. Euphausia mutica
and E. recurva appeared in the upper 50 m,
whereas E. gibboides was more widely distributed
with most of the population between 50 and 250 m
(e.g., CalCOFI 50.140).

The relative abundance, vertical distribution,
and diel migration of E. pacifica varied with
distance from the coast (Figure 3). Data from all
stations are illustrated to show- the number of
patterns exhibited by this species. Larvae and
juveniles tended to occupy a much wider vertical
range in nearshore waters. The bulk of the larvae
was usually found in the upper 150 m day and
night. The single exception to this pattern was
observed at CalCOFI 70.75 where the larvae were
abundant at 250 m throughout the day and in very
large numbers in the upper 100 m at night.
Juveniles were numerous in the surface waters as
well as at depths to 450 m. The adult phase was
frequently most abundant between 200 and 400 m
during the daytime. Offshore, during the day,
densities of this species were reduced, adults were
rarely collected, and populations occurred at
deeper, narrower intervals. At night, both near-
shore and offshore, only some members of each
stage migrated to the surface waters from depths
of 250-450 m. The general features of the geo-
graphical distribution of this species in the central
regions of the California Current agree with
observations by Brinton (1962b, 1967); the vertical
dimensions are more detailed.

Tessa rab rachio n

The only species of this genus, T. oculatus was
frequently found in small numbers, i.e., 10-
20/1,000 m^. Juveniles and adults of this charac-
teristic subarctic species were common and oc-
curred between 70 and 500 m. Somewhat greater
numbers were collected at night. The larvae tend-
ed to remain closer to the surface, i.e., from 70 to
200 m (SOE 74; CalCOFI 50:110), than juveniles
and adults which were usually found between 200
and 400 m. Thus, this species inhabits a wide depth
interval below the thermocline regardless of the
time of day.

Thysanopoda

Three species of Thysanopoda were collected- T.
aequalis, T. acutifrons, and T. egregia. Very few

specimens of these species were taken. Larvae of
T. aequalis, a species typical of central water
masses, were found in the upper 200 m at the
offshore stations of the southernmost transect.
Adults, found only at night at the same stations,
were collected above 300 m. Larvae of T. acuti-
frons were not observed. One juvenile and one
adult were taken during the day between 400 and
500 m at different offshore stations. At night a
total of 11 adults and 2 juveniles were caught
between 200 and 500 m (SOE 16, 74; CalCOFI
50.110, 50.140). One to four larvae of T. egregia
were collected between 50 and 450 m at nearly all
but the most northern stations.

Thysanoessa

Three species of Thysanoessa were found- T".
spinifera, T. gregaria, and T. longipes. Thysan-
oessa spinifera was only collected near the coast,
most frequently in the upper 150 m. Small densi-
ties of juveniles, the most abundant stage, were
present in tows from 150 to 350 m (CalCOFI 50.80,
70.75). Adults were not collected. The preponder-
ance of T. spinifera in the neritic environment
has been noticed previously (Brinton 1962a;
Hebard 1966). Diel changes in the vertical dis-
tribution of juveniles indicate that perhaps some
members of this phase migrated into the upper 100
m at night (Figure 4a). These data support other
studies that have suggested this species is a diel
migrant (Regan 1968; Day 1971; Alton and Black-
burn 1972).

Thysanoessa gregaria occurred at all but one
location (SOE 68). This species was found most
often in the upper 150 m, although it ranged to 300
m. Juvenile phases dominated the catches during
the SOE cruise. All stages were abundant among
the CalCOFI samples gathered 2 wk later. Densi-
ties were greater along the southern transects.
From 50 to 500 individuals/ 1,000 m"* were recorded
within depth intervals where the largest concen-
trations occurred. Larvae usually resided in the
upper 50 m. Juveniles and adults were numerous
between 50 and 200 m and often 3-10 times more
numerous in the night tows. These data suggest
that the older stages probably avoided the sam-
pling gear during the day. At one station
(CalCOFI 70.75), all stages of T. gregaria were
observed only in the upper 20 m during the day. At
night this species ranged to 400 m with the largest
densities occurring between 60 and 100 m and no
specimens were collected in the upper 30 m. These

929

FISHERY BULLETIN: VOL. 74, NO. 4

NUMBER PER 1000 CUBIC METERS

-2.^3. 4

_2._3

_2 _3

^2.^3.

2._3.

_2 -3

-2 3 _4

„2. 3

,2,„3

0 10 10 10 10 0 10 10 10 0 1,0 1,0 1,0 0 10 10 10 1,0 0 1,0 10 1.0 0 1.0 10 10 0 10 10 101,0 0 10 10 10 0 10 10 10'

I-
U

100
300
500
700

■f}

-I — I — I — I — rr

LAPVAE
-t 1 1

JUVENILES

STATION 81

' ' '

J»

',•

■

' 5

•

■

_3o :

■

LARVAE

■juveniles ■

STATION 74
1 1 1

5 ,,W >

t: J.

LARVAE
I I 1 1-

.• > •

JUVENILES
— I t 1 —

STATION 39
ADULTS

»' O t' *'

LARVAE " "JUVENILES '
- 1 1 1 1 L 1 ) t

STATION 47

UJ
Q

n-'--',. '

^AK ,.'AE

-H 1 I I

JUVENILES

-t *~ ■ I

JUVENILES
1 1 1 1 1

• LARVAE ■
1 1 1

STATION 9
■ ADULTS ■

o o DAY

• •NIGHT

(a)

NUMBER PER 1000 CUBIC METERS

0 10 10 10 10^0 10 10 10 0 10 10 10 0 10 10 10 10^0 10 10 10 0 10 10 10 0 10 1010 10 0 10 10 10 0 10 10^10''

100
CO

en 300-
u

500

700

.!.

-I — I — 1 — t-

JUVENILES

.r..--

STATION 50,140

• , , , I 1

w- .* «^ * *

> m

LARVAE

-i 1 1 h-

.•A

JUVENILES
— I 1 » —

STATION 50.110

• , , , I

. b'"

"liZro

LARVAE
I t 1—

JUVENILES
— t » 1 —

STATION 5080

ADULTS I

1 1 1 1

O O DAY

IE • •NIGHT

I-

100
300
500-
700-

1°'

.'•

o ,'

v.-

JUVENILES
-* 1 1 ( —

7- — ' — ' — r

LARVAE
-♦ »— • —

'-.

STATION 7075
ADULTS t
— I ( 1 1

(b)

Figure 3.- The vertical distribution of Euphausia pacifica according to stage of development, (a) SOE cruise 22 and (b) CalCOFI cruise

7008.

observations indicate this species can be con-
tagiously dispersed.

All phases of T. longipes (unspined form) were
collected in the upper 150 m. Juveniles and adults
were also abundant between 200 and 800 m. Por-

tions of these older populations appeared to mi-
grate toward the surface at night at several
stations (SOE 9, 74, 81; CalCOFI 50.80, 50.110)
(Figure 4b). The vertical range of this species
agrees with observations by Brinton (1962b) and

930

YOUNGBLUTH: VERTICAL DISTRIBUTION OF EUPHAUSIIDS

q:
hi

I-

z
I

Q.
LU
Q

100-^ ■"

300-

SOO-j^

700-

10 10^ 10^
I I I

NUMBER PER 1000 CUBIC METERS
O 1.0 1.0^ 10^ 0

-ODAY

--•NIGHT
I I —

LARVAE

STATION 70.7 5

JUVENILES
-I 1 I

■i

1,0 ip^ 1,0'^

STATION 50.80

JUVENILES
I I I

(a)

I-

U
Q

100-

300-

500-

70O-

NUMBER PER 1000 CUBIC METERS

10 10^ lO-' 0
I I I

NONE

-ODAY

LARVAE

• •NIGHT

I t

1,0 ^p^ ^p^ 0

JUVENILES
I I I

'P 'P' '."'

•O,

o«

I ;

STATION 50.110

ADULTS

_— .0-»
"l I

(b)

Figure 4.-Examples for the vertical distribution
of (a) Thysanoessa spinifera, (b) Thijsanoessa
longipes, and (c) Stylocheiron longicorne accord-
ing to stage of development.

liJ

I-
liJ

UJ
Q

NUMBER PER 1000 CUBIC METERS
o 10 i,o' io' o 1,0 i,o' i,o' 0 1,0 i,o' 1.0^

100-

30O

500-

7oa

o,«

O— ODAY LARVAE

• ©NIGHT

JUVENILES
1 I

O *

STATION 60.110

ADULTS
H 1 •—

(C)

931

FISHERY BULLETIN: VOL. 74, NO. 4

Ponomareva (1963). The abundance in subsurface
waters and the possible migratory behavior of this
unspined form have not been documented
previously.

Nematoscelis

Two species of Nematoscelis were taken-A'^.
tenella and A^. difficilis. Nematoscelis tenella was
collected at only one day-night station (SOE 16).
The few adults and juveniles caught, 2-13/1,000
m', were found between 400 and 500 m during the
day and 0 and 250 m at night. Nematoscelis
difficilis occurred between the surface and 450 m
at all but one station (SOE 25). This species was
more abundant near the coast along the CalCOFI
transect. Densities ranging in the hundreds per
1,000 m' at nearshore stations were an order of
magnitude larger than concentrations among
samples from waters farther offshore. Larvae and
most juveniles were taken only in the upper 100 m.
Adults were more abundant between 100 and 300
m, particularly at night.

Stylocheiron

Five species of this genus were found-S. affine,
S. longicorne, S. maximum, S. elegatum, and S.
abbreviatum. These species occupied similar depth
intervals day and night although each species
tended to inhabit a separate portion of the water
column. Stylocheiron affine occurred only along the
southern transect. All stages were collected
between 40 and 135 m and primarily at the most
offshore stations where densities of 60-150/1,000
m^ were recorded (SOE 25; CalCOFI 50.140). Each
stage was often more abundant in the night
samples. Stylocheiron longicorne, the most abun-
dant species of this genus, ranged between 70 and
350 m, but the bulk of the populations were within
the 150- to 250-m interval (Figure 4c). More
specimens were usually caught at night. Stylo-
cheiron maximum occurred in low densities at
every station, i.e., 5-40/1,000 m^. Larvae and
juveniles of this species were found most often
between 70 and 200 m. Adults were generally
deeper, ranging from 200 to 400 m. Differences
between day and night distributions indicate that
this species migrated less than 100 m, if at all.
Very small densities of S. elongatum, 1-27/1,000
m^, were observed at the offshore stations between
200 and 600 m. A few adults of S. abbreviatum
were found along the southern transect of the SOE

932

cruise. Four individuals were collected in the upper
150 m at stations well offshore (SOE 25, 39) and one
between 300 and 400 m nearshore (SOE 9).

Nematob rachion

Two species of Nemafobrachion were found— A'^.
boopis and N.Jlexipes. One to two individuals of A^.
boopis, mostly juveniles, were taken between 300
and 500 m and only during the day at a few,
southern stations (SOE 9; CalCOFI 70.75, 50.140).
Nemafobrachion fiexipes occurred at all but one
station (SOE 68). Small concentrations, usually
1-30/1,000 m-^ but ranging up to 69/1,000 m'', were
found regardless of the time of day. Juveniles
were often the most numerous stage. This species
was frequently encountered at 200-500 m during
the day. At night specimens were collected from
450 m to the surface with most of a population in
the upper 150 m.

PATTERNS OF ABUNDANCE,

VERTICAL DISTRIBUTION,

AND DIEL MIGRATION

The abundance and vertical distribution of the
more numerous euphausiid species in the upper
500-700 m differed in relation to distance from
shore, longitudinal position in the area sampled,
and vertical ranges occupied during a given day.
The largest densities of euphausiids occurred near
the coast (Table 2). Among the nearshore stations
(ca. 100-150 km from the coast) E. pacifica was the
numerically dominant euphausiid day and night,
composing 75-90% of all species observed. At
intermediate distances from the coast (ca. 300 km),
E. pacifica was less abundant, making up 36-60%
of the species collected, but still ranked first except
in the north (SOE 74) where Thysanoessa longipes
formed 69% of the day catch. Other euphausiid
species constituting 15-30% of the total number
included S. longicorne, E. gibboides, T. gregaria,
and T. longipes. At stations farthest offshore (ca.
600-700 km) along the southern transects, T.
gregaria and S. longicorne were the most abun-
dant species, forming 75% of the total during the
day. At night, larger numbers of E. gibboides, E.
mutica, and E. reciirva were collected such that
these populations also ranked among 75% of the
euphausiids collected. To the north, T. longipes and
S. longicorne were the abundant species, compos-
ing 70-80% of all euphausiids. These changes in
species composition and dominance represent the

YOUNGBLUTH: VERTICAL DISTRIBUTION OF EUPHAUSIIDS

c
bo

_3
Q.
CS

3
O

t«

>-,

c«

J= =«

'-2

•«->

3
n!
J=
Q.
3
0)

c«

J=
H

.J
■<

■^

+
(O CO r^ h-

Qj CD

o o o

£ "-

CVJ T- PJ

CD at

VI CD
1"°

Q.
CO

CD ■<t

CO CO

1- CO eg
T- r- ^

CO 00 CO CO

1- >- <D ^

ell T- •4 CM

O CM

CM «0 T-

r^ CM TT

CO CM eji

t o
eg eg

eg CO CO

eg CD

CD
CM

O
O (D

a> T-

■t .-CM

CM eg CO ■'t

CO eg 00 t- -.j-

CM CO CM eg rr

CD eg

IT)

•r-

"^ 1

CM '

+ +

.- CO C35 05

IT) CO

eg o>

CM 1-

CO CD
CO <J)

CO CO '- (^ eg .-
CO "-

CM m

CD C7> CD

CM y-

r-- .-

+
CO >n

00 CO '

If I I

•>r CM

CO CO

CM CM

I I

+
CO ■^ eg

Q.
O
T3

c
o

+
^ +

CM h.
C35

+ +

eg 1-

T- e5 eg

eg
o

o *o

°&

C Q)
O CO

■c (0
o -o

CO o

- CO

o o

■Q 3

'5.S

«>

£
5.

CO
(0

'5.
o
o

■Q

- C <0 CO

W i^ "0

5 cc

CJ) o
0) c

D> CO

-c

2 S

<U CO
O S

UjUjUjUjl~l— l~l— l~t-l--

CO ■
O i;

9)
2 2

c
S°

2 .2 CO

CD > ^

0) CO c ■— z
CO (o to CO 2 2

° 2

C:) ><

C CD

5 E

Qi
T3

O Q,

1) <B

Q.
CO
U
CO

■5

c
■o

933

FISHERY BULLETIN: VOL. 74, NO. 4

complex interaction of: 1) the recruitment and
mixing of species characteristic of the water
masses that compose the California Current, 2) the
daily vertical movements of euphausiids, 3) the
ability of most species to avoid the sampling gear,
and, to some extent, 4) their contagious dispersion.
Consequently only the more obvious patterns have
been noted.

The vertical distributions of adults and
juveniles in the upper 500-700 m are summarized
in Table 3 and compared with data from the
southern part of the California Current. As
previously mentioned, 7 of the 20 species collected
appear to be diel migrants. Distances of 300 m or
more were traversed by four species of Euphau-
sia. Portions of other populations such as Nema-
tobrachion flexipei^, T. longipes, and T. spinifera
may migrate up to 200 m.

The larval phases of most species live in the
upper 150 m. Tessarabrachion oculatus and Stylo-
cheiron spp. larvae were found more often below
the thermocline. The young of Euphausia spp.
tended to occupy and migrate through the same
depths as the older stages. In nearly all instances,
differences in density between day and night
catches of larvae were small.

The nonmigrating species included Thysanoessa
gregaria, Tessarabrachion oc id at us, S. maximum,
S. affine, and S. longicorne. The first three species
were usually scattered throughout a broad vertical
range. The other two species, S. affine and S.

longicorne, were vertically segregated and oc-
curred within much narrower depth intervals. The
different strata occupied by these two nonmigrat-
ing species was also observed in other regions by
Brinton (1967), Baker (1970), and Youngbluth
(1975).

DISCUSSION

Differences in the distribution patterns of many
species of zooplankton have been associated with
their response to environmental gradients, par-
ticularly temperature and illumination (Harris
1953; Lewis 1954; Banse 1964; Boden and Kampa
1967). In this study, the causative factors
influencing vertical and horizontal distributions
are difficult to elucidate. It is clear, however, that
the thermocline was an upper distribution bound-
ary for several species, e.g., T. oculatus, E.
gibboides, S. affine, S. longicorne, and 5. maximum.
In the southern part of the California Current, the
upper range of these species was also restricted by
the thermocline (Brinton 1967). Studies on the
tolerance of E. pacifica to changes in temperature
and salinity suggest that other unknown factors
probably regulate its distribution in the California
Current (Gilfillan 1972a, b).

Recently Isaacs et al. (1974) have proposed that
"by responding to light intensity, most vertically
migrating marine creatures are directed to food.
... In areas of low standing crops of phytoplank-

Table 3.-Comparisons of diel changes in the vertical distributions of adult and
juvenile euphausiids. Depth ranges (m) are 10% and 90% levels.

Central California

Southern

California

Current

Current (8

rinton 1967)

Species

Day

Night

Day

Night

Euphausia pacilica

20-500

0-450

150-425

0-150

E. recurve

300-600

0-50

180-550

0-150

E. gibboides

300-600

10-150

'300-500

40-120

E. mutica

2370-470

0-150

—

Tessarabrachion oculatus

70-450

—

Thysanoessa spinilera

125-300

0-150

—

T. gregaria

0-200

0-350

20-180

0-250

T. longipes

0-800

—

Nematoscelis dillicilis

0-400

250-200

0-275

Stylicheiron alline

40-100

35-100

550-200

15-250

S. longicorne

100-350

125-300

S. maximum

70-450

M30-200

Nematobrachion llexipes

200-500

0-130

100-450

100-350

Sampling range (m)

0-800

0-600

Sampling interval

50-150

25-

100

Gear employed

Bongo nets

Leavitt nets

Mesh opening

0.333 mm

0.550

'Adults only.

'Based on seven specimens

from one station.

'Mostly juveniles.

^Maximum concentration of

juveniles.

934

YOUNGBLUTH: VERTICAL DISTRIBUTION OF EUPHAUSIIDS

ton, daylight penetrates further into the ocean
causing the migrating animals to descend deeper.
In the turbid water associated with high standing
crop, the migrating forms remain closer to the
surface." Observations on the vertical distribution
and daily movements of one euphausiid species in
this study lend support to this hypothesis. In more
turbid, upwelled water near the coast where
standing stocks of phytoplankton were greater
(e.g., CalCOFI 50.80), populations of juvenile E.
paciftca were larger and extended over wider
vertical ranges but their diel vertical migrations
were not pronounced. In clearer, more oligotrophic
waters farther offshore (e.g., CalCOFI 50.110; SOE
16, 47, 74), populations were reduced in size,
occupied deeper, usually narrower depth intervals,
and daily vertical movements were more obvious.
From these few observations it appears that
density levels and migration intensities of this
species may be coupled with the standing stock of
phytoplankton in surface waters.

The persistence of nonmigrating forms, e.g.,
Stylocheiron spp., within the same, relatively
narrow depths day and night in waters of varying
origin and the recurrence of the finding in this and
other studies (Brinton 1967; Youngbluth 1975) that
only a portion of a population categorized as a
migrating form, e.g., Euphausia spp., may ac-
tually make daily vertical movements to surface
waters, suggest that factors in addition to tem-
perature and light act to regulate the distributions
recorded. These observations indicate that more
attention should be directed toward sampling
those horizons where zooplankton populations are
concentrated to determine how distributional and
behavioral patterns are structured by the physical
and biological fluctuations within their preferred
habitats.

ACKNOWLEDGMENTS

I thank Malvern Gilmartin, John H. Martin, and
Donald P. Abbott for their suggestions and crit-
icisms. Pete Davoll, Dan Essin, Mel Malkoff, Tom
Malone, and Robie Robison helped set and recover
the Bongo nets aboard the RV Proteus. Support
for ship time, gear, and data analyses was provid-
ed in part by NSF grants GB 8374, GB 8404, GD
27254, and GA 28306 and in part by an NSF
predoctoral dissertation grant GA 29056. CalCOFI
samples were obtained through the program at the
Scripps Institution of Oceanography.

LITERATURE CITED

Alton, M. S., and C. J. Blackburn.

1972. Diel changes in the vertical distribution of the eu-

phausiids, Thyxanoessa spinifera Holmes and Euphausia

pacifica Hansen in coastal waters of Washington. Calif.

Fish Game 58:179-190.
Alvarino, a.

1965. Distributional atlas of Chaetognatha in the California

Current region. Calif. Coop. Oceanic Fish. Invest., Atlas 3,

291 p.
Baker, A. de C.

1970. The vertical distribution of euphausiids near Fuer-
teventura, Canary Islands ('Discovery' SOND Cruise,
1965). J. Mar. Biol. Assoc. U.K. 50:301-342.

Banse, K.

1964. On the vertical distribution of zooplankton in the
sea. Prog. Oceanogr. 2:53-125.
Barham, E. G.

1957. The ecology of sonic scattering layers in the Monterey
Bay area. Stanford Univ., Hopkins Mar. Stn. Tech. Rep. 1,
182 p.
Berner, L. D., Jr.

1957. Studies on the Thaliacea of the temperate northeast
Pacific Ocean. Ph.D Thesis, Scripps Inst. Oceanogr., Univ.
Calif., San Diego, 144 p.
1967. Distributional atlas of Thaliacea in the California
Current region. Calif. Coop. Oceanic Fish. Invest., Atlas 8,
322 p.

BlERI.R.

1959. The distribution of the planktonic Chaetognatha in
the Pacific and their relationship to the water masses.
Limnol. Oceanogr. 4:1-28.

Boden, B. p., M. W. Johnson, and E. Brinton.

1955. The Euphausiacea (Crustacea) of the North Pacific.
Bull. Scripps Inst. Oceanogr., Univ. Calif. 6:287-400.

Boden, B. P. and E. M. Kampa.

1967. The influence of natural light on the vertical migra-
tions of an animal community in the sea. Symp. Zool. Soc.
Lond. 19:15-26.

BoLiN, R. L. and D. p. Abbott.

1963. Studies on the marine climate and phytoplankton of
the central coastal area of California, 1954-1960. Calif.
Coop. Oceanic Fish. Invest., Rep. 9:23-45.

Brinton, E.

1960. Changes in the distribution of euphausiid crustaceans
in the region of the California Current. Calif. Coop.
Oceanic Fish. Invest., Rep. 7:137-146.

1962a. The distribution of Pacific euphausiids. Bull. Scripps

Inst. Oceanogr., Univ. Calif. 8:51-269.
1962b. Variable factors affecting the apparent range and
estimated concentration of euphausiids in the North
Pacific. Pac. Sci. 16:374-408.
1967. Vertical migration and avoidance capability of eu-
phausiids in the California Current. Limnol. Oceanogr.
12:451-483.
Gushing, D. H.

1971. Upwelling and the production of fish. Adv. Mar. Biol.

9:255-334.
Day, D. S.

1971. Macrozooplankton and small nekton in the coastal
waters off Vancouver Island (Canada) and Washington,

935

FISHERY BULLETIN: VOL. 74, NO. 4

spring and fall of 1963. U.S. Dep. Commer., Natl. Mar.
Fish. Serv., Spec. Sci. Rep. Fish. 619, 94 p.
DoDiMEAD, A. J., F. Favorite, and T. Hirano.

1963. Salmon of the North Pacific Ocean. Part II. Review of
the oceanography of the subarctic Pacific region. Int.
North Pac. Fish. Comm. Bull. 13, 195 p.

Fleminger, a.

1964. Distributional atlas of calanoid copepods in the
California Current region, part I. Calif. Coop. Oceanic
Fish. Invest., Atlas 2, 313 p.

1967. Distributional atlas of calanoid copepods in the
California Current region, part II. Calif. Coop. Oceanic
Fish. Invest., Atlas 7, 213 p.

Frolander, H. F.

1962. Quantitative estimations of temporal variations of
zooplankton oflf the coast of Washington and British
Columbia. J. Fish. Res. Board Can. 19:657-675.

GiLFILLAN, E.

1972a. Seasonal and latitudinal effects on the responses of
Euphaiisia pacijica Hansen (Crustacea) to experimental
changes of temperature and salinity. In A. Y. Takenouti
(editor). Biological oceanography of the northern North
Pacific Ocean, p. 443-463. Idemitsu Shoten, Tokyo, Jap.

1972b. Reactions of Euphauiiia pacijica Hansen (Crustacea)
from oceanic, mi.xed oceanic-coastal and coastal waters of
British Columbia to experimental changes in temperature
and salinity. J. Exp. Mar. Biol. Ecol. 10:29-40.

Harris, J. E.

1953. Physical factors involved in vertical migration of
plankton. Q. J. Microsc. Sci. 94:537-550.

Hebard, J. F.

1966. Distribution of Euphausiacea and Copepoda off
Oregon in relation to oceanographic conditions. Ph.D.
Thesis, Oregon State Univ., Corvallis, 94 p.

Isaacs, J. D., A. Fleminger, and J. K. Miller.

1969. Distributional atlas of zooplankton biomass in the
California Current region: Spring and fall 1955-1959.
Calif. Coop. Oceanic Fish. Invest., Atlas 10, 252 p.

IssACS, J. D., S. A. ToNT, and G. L. Wick.

1974. Deep Scattering Layers: vertical migration as a tactic
for finding food. Deep-Sea Res. 21:651-656.
Johnson, M. W., and E. Brinton.

1963. Biological species, water-masses and currents. In M.
N. Hill (editor), The Sea, Vol. 2, p. 381-414. Interscience,
N.Y.

Kin'dyushov, v. I.

1970. Seasonal changes of water masses in the California
region of the Pacific Ocean. [In Russ., Eng. summ.]
Okeanologia. 10:596-606.

Laurs, R. M.

1967. Coastal upwelling and the ecology of lower trophic
levels. Ph.D. Thesis, Oregon State Univ., Corvallis, 121 p.

Lewis, J. B.

1954. The occurrence and vertical distribution of the Eu-
phausiacea of the Florida Current. Bull. Mar. Sci. Gulf
Caribb. 4:265-301.

Longhurst, a. R.

1967. Diversity and trophic structure of zooplankton com-
munities in the California Current. Deep-Sea Res.
14:393-408.
McGowan, J. A.

1967. Distributional atlas of pelagic molluscs in the Califor-
nia Current region. Calif. Coop. Oceanic Fish. Invest.,
Atlas 6, 218 p.
1971. Oceanic biogeography of the Pacific. /» B. M. Funnell
and W. R. Riedel (editors). The micropaleontology of
oceans, p. 3-74. Camb. Univ. Press., Lond.
McGowAN, J. A. AND D. M. Brown.

1966. A new opening-closing paired zooplankton net.
Scripps Inst. Oceanogr. Ref. 66-23, 56 p.

Malone.T. C.

1971. The relative importance of nannoplankton and net-
plankton as primary producers in the California Current
system. Fish. Bull., U.S. 69:799-820.

Okutani, T. and J. A. McGowan.

1969. Systematics, distribution, and abundance of the
epiplanktonic squid (Cephalopoda, Decapoda) larvae of the
California Current, April, 1954-March, 1957. Bull.
Scripps Inst. Oceanogr., Univ. Calif. 14, 90 p.

PlEPER, R. E.

1967. Mesopelagic faunal discontinuities in the eastern
North Pacific. AC Electronic-Def. Res. Lab. Rep. TR67-19,
38 p.

Ponomareva, L. a.

1963. Euphausiids of the North Pacific, their distribution
and ecology. Akad. Nauk SSSR Inst. Okeanol. (Translated
by Israel Program Sci. Transl., Jerusalem 1966, IPST cat.
no. 1368, 154 p.)

Regan, L.

1968. Euphausia pacijica and other euphausiids in the
coastal waters of British Columbia: Relationships to
temperature, salinity and other properties in the field and
laboratory. Ph.D. Thesis, Univ. British Columbia, Van-
couver, 235 p.

Reid, J. L., Jr., G. L. Roden, and J. G. Wyllie.

1958. Studies of the California Current system. Calif.
Coop. Oceanic Fish. Invest., Prog. Rep., 1 July to 1 Jan.
1958, p. 27-57.
Thrailkill, J. R.

1963. Zooplankton volumes off the Pacific Coast, 1959. U.S.
Fish Wildl. Serv., Spec. Sci. Rep. Fish. 414, 77 p.
Whittaker, R. H.

1952. A study of summer foliage insect communities in the
Great Smokey mountains. Ecol. Monogr. 22:1-44.
Yoshida, K.

1955. Coastal upwelling off the California coast. Rec.
Oceanogr. Works Jap., New Ser. 2(2):8-20.
Youngbluth.M. J.

1973. The vertical distribution, diel migration, and com-
munity structure of euphausiids. Ph.D. Thesis, Stanford
Univ., Palo Alto, Calif., 298 p.
1975. The vertical distribution and diel migrations of
euphausiids in the central waters of the eastern South
Pacific. Deep-Sea Res. 22:519-536.

936

THE APPLICATION OF SYSTEMATIC SAMPLING TO A STUDY OF
INFAUNA VARIATION IN A SOFT SUBSTRATE ENVIRONMENT^

Stephen Scherba, Jr.^ and Vincent F. Gallucci^
ABSTRACT

Stratified systematic sampling was applied to an intertidal macrofauna sediment study. A stratified
systematic sampling plan retains the advantages of the more common fixed level transect sample, and
possesses additional advantages which recommend it for use in some, intertidal studies. The field data
collected in this study demonstrated the effectiveness of stratified systematic sampling for quantifying
both sediment and population characteristics along a sediment gradient, and for the testing of
biological hypotheses.

Intraarea, interarea, and interseason hypotheses about sediment composition were tested in terms of
particle size distributions. Populations of bivalves and polychaetes were simultaneously sampled and
hypotheses concerning spatial and seasonal variations in an intertidal mud flat were tested.
Experimental results using stratified systematic sampling suggest that Newell's hypothesis can be
extended to encompass temporal variation. Fine sediment grades (silty areas) may act to insulate
infauna against the extremes of seasonal stresses.

Sediment composition, as measured by the average percentage composition by weight of various
grain sizes, was not sufficient to predict macrofaunal presence.

The study of the complex relationship existing
between macrofauna (e.g., bivalves and poly-
chaetes) and their soft substrate environment is of
wide interest in marine biology. Soft sediments
are both a shelter from predators and a food source
for deposit feeders. The particle size distribution
of the sediment influences such factors as food
availability, the depth of the aerobic layer, water
content, pH differentials, and growth rates. De-
trital content and particle size distribution of the
sediment are largely determined by the hy-
drodynamics of currents. However, Rhoads (1967)
demonstrated that macrofauna modify sediment
stability, composition, and water content by ac-
tivities such as building tubes, ingesting sediment
and detritus (to remove bacteria from sediment
particles), depositing feces, etc. The particle size
distribution of the sediment is, therefore, one
measure of certain types of biological activity
(Newell 1965).

Studies in soft substrate environments usually
involve sediment samples which contain large
numbers of macrofaunal species in different den-
sities as well as different particle size distribu-

'Contribution No. 430 of the College of Fisheries, University of
Washington, Seattle, WA 98195.

-The Department of Physiology and Biophysics, University of
Washington, Seattle, WA 98195.

^College of Fisheries and The Center for Quantitative Science,
University of Washington, Seattle, WA 98195. Send reprint
requests to this address.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

tions. Then, it may be necessary to make compar-
isons between samples which may call for the use
of statistical methods as found in standard text-
books (e.g., Sokal and Rolf 1969). The validity of
tests of comparisons, however, must rest upon the
application of valid sampling plans in the field, but
most valid sampling plans do not meet the needs
of the ecologist. This paper reports on the use of
stratified systematic sampling which, to our
knowledge, is heretofore unused in the marine
literature. Stratified systematic sampling seems
to meet the needs of most studies that would have
used transecting methods which, generally, are
statistically unacceptable. The sampling method is
applied to a study of animal and sediment
gradients' in a basically marine embayment sub-
ject to seasonal variation in density of animals and
algae. Hypotheses comparing areas of different
sediment composition in the winter, spring, and
summer are tested using animal-presence and
sediment-particle-size data collected using a
stratified systematic sampling plan. The bay is
shown to contain a sediment gradient from fine
silt to coarse sand, with an associated polychaete
distribution expressed by both the number of
animals per species and by the number of species
found.

A variety of sampling methods are described in
the literature. For example, the works of Skellam
(1958) and Sen et al. (1974) presented different

937

FISHERY BULLETIN: VOL. 74, NO. 4

types of transect methods, and Saila and Gaucher
(1966), Russell (1972), and Loesch (1974) described
subtidal stratified random sampling methods.

Transecting methods used in the soft substrate
of the intertidal zone usually involve the choice of
a narrow belt of one or two sampling units in
width, placed perpendicular to the water line.
Samples are collected at fixed and predetermined
tidal levels (e.g., every 10 m) to correspond to
changes in such things as algal and sediment
composition (Matthiessen 1960; Vassallo 1969;
Warme 1971; Bloom et al. 1972). This method will
be referred to as fixed level transect sampling
(FLTS). A common denominator in this type of
work is that no probability model is used in
selecting the location of the sampling units. An
alternative to FLTS is simple or stratified random
sampling where the discrete uniform probability
distribution underlies the selection of sample sites.
The disadvantage of random sampling is that
there is no guarantee that sample sites will be in
those areas where experimental interest is
focused. However, without an underlying
probability model, valid statistics may not be
estimated (Cochran 1963) because the sample sites
may not be independently located and subsequent
statistical tests may be invalid. These points are
often overlooked.

Stratified systematic sampling (SSS) is
proposed as an alternative to the FLTS method
currently popular in intertidal fieldwork. The
usefulness of SSS is demonstrated by applying it
to a study of spatial and temporal variation in a
macrofauna-sediment relationship. This field
study was conducted at Garrison Bay (Figure 1), a
small embayment on San Juan Island, Wash. (lat.
48°35'N, long. 128°08'W).

MATERIALS AND METHODS

Applications of systematic sampling are found
in the forestry literature (Osborne 1942; Finney
1948; Matern 1960; Faber 1972). Mathematical
details are found in sampling texts such as Coch-
ran (1963), Raj (1968), or Sukhatme and Sukhatme
(1970) and in many theoretical papers.^ Systematic
sampling assumes that the sampling units in the
area to be studied are numbered consecutively.
The attractiveness of the method is increased
by the relatively sessile nature of many intertidal

^Scherba, S., Jr., and V. F. Gallucci. 1976. Quantification of
species-presence gradients by stratified systematic sampling
and the autocovariance function. Unpubi. manuscr., 15 p.

Figure 1. -Shoreline of Garrison Bay showing the location of the
four study areas. Insert shows the representative arrangement
of the subareas (A and B) and the strata (1 and 2) in these areas.

organisms. From A^ sampling units numbered 1, 2,
. . ., N; n sampling units are selected, all evenly
spaced at a distance of K sampling units apart.
Thus, N = riK. The location of the first unit to
actually be sampled is randomly chosen by select-
ing a number between 1 and K from a table of
random numbers. Hence, systematic sampling is
based upon a uniform probability distribution (Raj
1968). SSS is a variation in which the region is
divided into strata, e.g., at the locations of the
fixed levels where samples would have been col-
lected using FLTS. Each stratum is independently
sampled in the manner described above.

Four regions with different sediment composi-
tions were a priori defined in the intertidal area of
Garrison Bay (Figure 1). Representative areas
within these regions were sampled in the winter
(January and February), spring (May and June),
and summer (July and August) 1974. A north-
south sediment gradient exists because fine
sediment is deposited at the closed end of the bay
where the water is less turbulent. Thus, the south
(closed) end of the bay consists principally of fine
grades of sediment, while the north (open) end
consists mainly of coarser grades. Visual exami-
nation indicates that perpendicular to the water,
there is a sediment gradient as well as a zonation
of intertidal animals. However, the statistical
comparisons of the data from strata, which were

938

SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA

placed parallel to the water, do not show the
gradient. This is probably a consequence of the
short distance between strata.

Field Procedures

Within each region (Figure 1), a rectangular
study area was defined, measuring 95 m in length
parallel to the waterline, and approximately 7 m
wide, perpendicular to the waterline. Two parallel
strata, approximately 2.5 m apart, were placed
within each area, parallel to the waterline. The
stratum at the highest tidal level was designated
stratum 1 while the lower stratum was designated
stratum 2. Stratum 1 within the areas was located
at -1.4, -1.2, -1.1, and -1.2 feet in areas 1, 2, 3, and 4
respectively; while stratum 2 in those same areas
was at -1.5, -1.7, -1.5, and -1.6 feet."' The study
areas were numbered one (1) to four (4) (south to
north) and defined by stakes marked with
fluorescent tape for night identification.

It is necessary to test the homogeneity of
sediment composition within a region if the areas
are to be considered representative. This test was
accomplished by dividing each area into two
subareas,*' separated by 5 m, and denoted as A (for
the northmost subarea) and B (for the southmost).
Each subarea contained about 448 sampling units.
Two samples were then collected on each stratum,
from each subarea, using a systematic sampling
plan.

Each subarea was considered to contain sepa-
rate populations, and the two population Kol-
mogorov-Smirnov procedure with n = 4 (Conover
1971) was applied to the data collected. This use of
both subareas was carried out only for the winter
sampling. Winter sampling of the subareas was
done on: 8 January 1974 (lA, IB), 9 January 1974
(3A, 3B), 2 February 1975 (2A, 2B), and 3 February
1974 (4A, 4B).

Spring and summer sampling was conducted
only in subareas IB, 2A, 3A, and 4A as follows: 24
May 1974 (IB, 3A), 21 June 1974 (2A, 4A), 19 July
1974 (IB, 4A), and 16 August 1974 (2A, 3A). Each
stratum in these four subareas was independently
sampled during these two seasons with n = 4 on
each stratum.

All samples were collected using a thick-walled,
cylindrical corer made of polyvinyl chloride pipe,

10 cm inside diameter and 18 cm long. The corer
was pressed into the sediment to 18 cm, and its
contents removed by hand, placed in a labeled
plastic bag, and taken to the laboratory. Each
sample was passed through a 1-mm sieve, and the
contents retained by the sieve were sorted twice
by eye to remove all bivalves and polychaetes (the
only members of the macrofauna identified).
These organisms were placed in 80% ethanol and
8% formaldehyde, respectively, for later
identification. Only the common bivalves and
polychaetes were identified to genus and species.
The sediment portion of each sample was dried at
100°C for approximately 4 h. The method used to
quantify the particulate properties of the
sediment was the percentage composition by
weight of selected sediment grain sizes. A me-
chanical shaker was used to pass the sediment
portion of each sample through a series of
Wentworth sieves (1.981, 0.495, 0.246, 0.124, 0.063
mm). The contents of each sieve were weighed and
recorded as percentage of the total weight of that
sample.

Statistical Procedures

Estimates of the variances of the sample means,
obtained from SSS were approximated by the
estimate of the variance of the sample mean from
a simple random sample (see Cochran 1963), i.e., by
using

var

where

*Tidal heights are reported in feet to conform with U.S. Coast
and Geodetic Survey Tide Tables.
®We thank A. R. Sen for this suggestion.

The rational for this approximation is discussed

later.

Two statistical tests were used to quantify the
sampling results. The K sample Kolmogorov-
Smirnov(K-S)testwitha = 0.10, using the Tg test
statistic (Birnbaum and Hall 1960; Conover 1971),
was used to test hypotheses about variation in
sediment composition. The chi-square test for
several multinomials with a = 0.05 (Conover 1971)
was used to test hypotheses about variation in
bivalve and polychaete community structure.

In the within-area sediment homogeneity test
empirical distribution functions were constructed
for each subarea. The K-S test (a = 0.10) was then
used to test the null hypothesis (Ho) of equality of

939

FISHERY BULLETIN: VOL. 74. NO. 4

these distribution functions. Using winter sam-
ples only, the test failed to reject Hq; thus, the data
from each A and B subarea pair were combined
and considered to be one subarea for comparison to
the subareas sampled in the spring and summer.
Hence, all data were analyzed as if they had been
collected from four equal sized subareas, of
dimensions 45 m by 7 m, during each season, using
a sample of size four on each stratum.

The empirical cumulative distribution functions
were constructed from the data by defining a
random variable X as the sum of the percent of the
total sediment weights retained in the sieve sizes
<0.063, 0.063, and 0.124 mm. The random variable
X takes a value of each sample, in each subarea, on
each stratum. Thus, the empirical distribution
functions constructed from this data characterized
the sum of the weights of three finest sediment
grades (and by subtraction from 100%, the three
coarsest grades as well) for each stratum in each
subarea. These three sieve sizes were grouped
together because they constitute what may be
called the finer grades of sediment and they
probably have the greatest biological impact
(Newell 1965). If the grain size which is of prin-
cipal importance to the organisms is known, then
the random variable could be chosen accordingly.
There is much evidence that grain size is impor-
tant to the organisms (e.g., see Loosanoff and
Tommers 1948; Sanders 1958; Wieser 1959; Gray
1974). Subject to this limitation of comparing only
the finer sediment groups, the sediment data may
be statistically compared stratum to stratum in
any one subarea, between subareas, or in combi-
nations of these, both within or between seasons.

In each case, the null hypothesis for the K-S test
on sediment was

species types with entries in the expected value
table which were either greater than unity, or at
least, not far below unity. All species identified are
listed, but, in certain cases, some species were
grouped into families for the analysis; these are
noted in the tables of data. Grouping of data is
often advisable on statistical or biological grounds
depending upon the objectives of the study. When
data were grouped in this study, the grouping was
dictated by sample sizes and was consistent with
biological facts such as where the organisms occur
in Garrison Bay, their modes of feeding, and their
taxonomy.

The data were organized into contingency
tables for a multinomial distribution. We denote
the probability of a randomly selected value from
the /th population as being classified in the jth
class by P,,. The columns of the table represent
species (classes) while the rows represent popula-
tions, i.e., a particular stratum in a given subarea
during a specific season. The null hypothesis may
be stated as:

Ho: Pi, =P^,= ... = P,, for alii; i =

l,2,...,c, (2)

and the alternative

H^: there is at least one P^j "/ P^j for some
j and pair i, k where r equals the number of rows
and c equals the number of columns. Under Hq,

P 11 — P2I = . . • = Prl = Pi

Plr — P2C —

= P = P

Ro.FAx) = FAx) = ... = F,{x) (1)

and the alternative

H^: there is at least one inequality where
Fj (x) is the cumulative distribution function of the
random variable A" corresponding to area J.

The statistical analysis of the distribution of
animal populations was based upon standard
chi-square procedures (Conover 1971). Let the
random variable Z have a multinomial distribution
where the number of classes corresponds to the
number of species types used, and the number of
trials is the total number of individuals of all
species. The chi-square test was applied to those

where Pj = Q /N; Q = sum of observations in
column j; N = total number of observations from
all samples; and P, estimates P^. When a row or
column of a particular contingency table equalled
zero, it was not possible to reach a decision about
the chi-square null hypothesis. To maintain con-
sistent comparisons, no alteration of the contin-
gency tables was made in such cases. The results of
some of these tests of homogeneity are summa-
rized in the next section.

RESULTS

The sampling data and the estimates of the
variances of the sample means appear in Tables

940

SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA

1-3. It was necessary to pool the winter samples
that were used to test for homogeneity (see Sta-
tistical Procedures) within each area. The samples
in each area were considered to be from one
subarea to correspond to the subareas used for the
spring and summer sampling. The spring and
summer samples were collected exclusively from
subareas IB, 2A, 3A, and 4A. The data in Tables 2
and 3 are for the species of bivalves and poly-
chaetes which could be identified from the
samples.

Sediment

The sediment data (Table 1) and statistical
analyses confirm the existence of a particle size
gradient from the closed (south) end to the open
(north) end of the bay. Subareas IB and 4A appear
similar in Table 1, but subarea IB is located in the
closed end of Garrison Bay (Figure 1) which is
much muddier with more fine grained particles
and has poorer drainage than subarea 4A. Because
the data suggest that subareas IB and 4A contain
approximately the same proportion of coarse grain
particles (i.e., ^1.981 mm), a qualitative descrip-
tion was used to supplement the quantitative
analysis based on grain size composition by per-
centage weight: the 1.981-mm sieve in samples
from subarea IB contained large amounts of shell
fragments, which will remain in suspension longer
due to their flattened shape, while the same sieve
size in subarea 4A contained mostly round pebbles,
which settle more rapidly, and few shell frag-
ments. Thus, despite their heavy weight, shell
fragments were carried into the quiet part of the
bay.

Samples from subarea 2A often had added
weight in the 1.981-mm sieve in the form of rocks
of about 5 cm across. This is probably the result of
the activities of early settlers or of recent an-
thropoligical investigations.

The K-S procedures (a = 0.10) confirmed the
existence of a north-south sediment gradient
between similarly numbered strata in all seasons
between all four subareas. All six of the null
hypotheses (1) of equality were rejected.

The sources of this gradient were located by
using the K-S procedure (a = 0.10) to compare all
combinations of subarea pairs and seasons for
similarly numbered strata. This resulted in the
testing of 36 null hypotheses (1) of no difference, of
which 25 were rejected and the remainder accept-
ed. Thus, a gradient may be said to be the result of

C\J

(Z3

tic

C oj

^ E

■^

„ "^

II a

« '"

<U v..

•c °

o a;

M y

a> c

■H c«

cs -r-

S >

— 0)

tn j::

0) *J

i^-s

.2 £

m s

<» in

O 0)

.C 0)

o t.

rt («

0) m

.E S

£ 9

■^ u

T3 ca

<u a.

^.S

Z e

j: 0)

tm j2

^ i

rt OJ

■li. -c

2 H

«4-l

-4->

c«

f— 1

E-i

<B?

o E

't:"r

03 N

«l

f? S" CVI lO o ,-^

C\J ,- O ,- OJ T-

O^ O O o O C>J

■^ 0> C^ 'J- CO 00

>o m" '- eg CO T-'

^ r- ■»- in

T- o o> IT) eg
o o o o ^
(O in o f^ cvj
Tt s. oi ^ cvi

CM r^ O) CO CO 3^

^ »- ;-^ O O c^

csj o <i3 OJ r^ cNj
O) CO O 0> OJ o

1- 1- m

. 00

T-^ '- T|^ o CO C\i

C7) eg CO in eg Tj;

^^ 00 *- o CO 00

"- ^ T- TJ-

cn —r- c^ — ■

O) eg t^ CO eg ^

00 »- CO lO 00 GO

CD S.' ^' ^' •» 00

'°. "^ o CO •*. oT
OJ CO ^- CO '— ^^
■* (5) CO X ej o

O K O O) CO Oi

>- CO _^,^,^

CO ^ r^ IT) CM

o o eg (D t 1-
o> ■^^ r^ u^ o> ID

CD 0> to (b CO CO
»- t- T- CO

3,^ m o "- CO

in »-. ■V CJ) 1- C3>

o> o^ O) ir> in o

1- •.- i- n

CD -^ o o r^

C3)

O^ ^ CD CO CNJ in

CO <D (3> O) O CO

■^' in t-' o in eg

T- T- T- m

-.^■^ .^ — oo .—
o i~- in CO in Tf

r-' O T-' in O 00

tC^ ■* ID in in CO
CD O) in iri 00 T

T- T- I- CO

00 ..^ ,^ .-^ -.^
eg CO 00 CO ■* ~

hT eg C7) in CO eg

o '- o in in CD
•.- »- eg T- ^ c\j

— r^ — ■v — ~
o t-: CO o o C3

'- S- ^ o '-^ CD

rC m r^ r^ eg eg

CO CO Tt CD in T-
CD (D eg -tf- a> GO
O p ^ CM Tf o>

o ci o o o T^

<s> E
o E

O) -^ .— > ^ *-

t CO CO CO CO ~

c> o ^ 3, O- ^

CO CO <D c£> eg TT

Tt CD oj o o o

cvj •- OJ CO

<D _ — . m
OJ ^-incq-

o T- »- in o CO

00 00 CO ■^ CO <D

CO o) CM d O) •^"

C\J »- r- CO

'- CO C3) m eg — ,

t- CO »" ^ f^ CD

CD o O O O '»

O 1^ T- 'J- CD ^

CO (V.' O) V I- y-

T- (O

05 „ CO CD

1- <D p CO in j;::-

O T- T- O O CO

in oT CD <D CO CO

'J- cji C3> <b in Tt

r- »" ^

cj) ^ ^,^ in in

01 CM CD O) CM £-

C3 T-' Tf C> O »-

<D In CM T- CJ> CM

r«- o '^ t^ ^ — '^

O CO 00 CM CO 00

o c> c^ g.2ico

'l- 05 CM '»■ C> CM

CO (jj CD CO in oi
r- 1- in

eg <3) r-- eg cj> <D

c. '^ *" si. i ifi

CM CO CD C7) in CJ)
0> K C> '- CM f-.'

T- CO 1- »- '-

1- (O '-'-'" in"
c> eg '- c3 o in
of o p CD in o>

^ CM CD C3 eg (D

T- CM eg ^ 1- 1-

CVJ CO p PT PT

CO <D K '- CO T—

o oT eg p in t
CO K ^ ^' C3 o

»- Tf CO o> in CO

r-; oTp P P p

K (O iri ^ CM in

?- CO ■■- T- t-

p Pp p p G^

T- \ft CD O^ O ^

O) CO *- ^. "^ P
O O CD CO S: CD

»- eg CO »-

. "^ C7> • C>J CO

eg t-. CM ^^co

o <D o p CM r«-

(O ^' S^ CM CO CD

CO CO "■ CO in T-

cD <D eg Tf O) 00

o o ■^ CM ^ p

o o o o o »-

941

FISHERY BULLETIN: VOL. 74, NO. 4

different subareas being dissimilar in a season. In
particular, it was found that: 1) in the spring,
subareas IB and 2A were significantly different on
both strata, while only stratum 2 in those subareas
was significantly different in the summer; 2) both
strata in subareas IB and 3A and subareas 3A and
4A were significantly different throughout all
seasons; and 3) stratum 1 of subareas 2A and 4A
were significantly different only in the spring,
while stratum 2 in these subareas was different in
each season.

The K-S procedure (« = 0.10) was used to test
sediment composition homogeneity both between
the strata of a given subarea and among the three
seasons for a single stratum. Over half of these
null hypotheses were accepted. Therefore,
sediment composition of the strata remained
largely stable throughout the three seasons and
apparently lacked a consistent zonation perpen-
dicular to the water.

Polychaetes

Table 2 shows that the dominant polychaete
species vary according to season and sediment
type. These species were found to be: Lumbrineris
bicirrata, Dorvillea japonica, Scohplos pugetten-
sis, Cirratidiis cirratus, and Capitella capitata. In
this study the dominant species is the species with
the largest number of individuals.

Spatial and temporal dominance patterns may
be seen. In subarea 2A, the dominant organism is
generally D. japonica (in all seasons on stratum 2
and in the winter and spring on stratum 1).
Capitella capitata is usually the dominant species
in subarea 3A (S. pugettensis being dominant
there only in the winter on stratum 1). The in-
crease in this species during the summer, as
compared to the spring, on both strata of subarea
3A may have been influenced by the presence of a
dense algal mat of Enteromorpha sp. which
covered large intertidal areas. Subarea 4A has the
greatest fluctuation with respect to the dominant
species. On stratum 2 of subarea 4A, C. capitata is
dominant in the spring and summer, replacing L.
bicirrata, the winter dominant. Capitella capitata
is dominant only in the spring on stratum 1 of
subarea 4A; C. cirratus is dominant in both winter
and summer. Subarea IB shows the smallest
seasonal fluctuation of any subarea in both total
polychaete assemblage and dominant species.
Cirratuliis cirratus is dominant on both strata in
the spring and summer, replacing the winter

dominants S. pugettensis (on stratum 1) and L.
bicirrata (on stratum 2).

No simple seasonal pattern is discernible on the
strata of the various subareas (see Table 2).
Stratum 1 in both subareas IB and 3A shows a
steady increase in total number of individuals
between spring and summer. In the cases of
subarea IB (stratum 2), subarea 2A (strata 1 and
2), and subarea 4A (stratum 2), the largest number
of individuals is present in the spring. Subarea 3A
(stratum 2) and subarea 4A (stratum 1) have the
largest number of individuals in the winter, due to
Cirratnlus capitata and Capitella cirratus, re-
spectively. However, there is insuflRcient data to
conclude that stratum 2 is uniformly sustaining
the greatest total numbers of individuals season-
ally (perhaps due to the small horizontal distance
separating the strata in each subarea).

Table 2 shows that it is possible to rank the
subareas, in descending order, with regard to
number of species present: subareas IB, 4A, 2A,
and 3A; as well as with respect to the total number
of individuals: subareas IB, 2A, 4A, and 3A. There
are occasional seasonal reordering of these ranks.

Statistical analysis using the chi-square
procedures (a = 0.05, 33 df) confirmed the exis-
tence of a within season polychaete distribution
(for the 12 groups used in the analysis) on iden-
tically numbered strata, between the four sub-
areas in five of these six comparisons. The one
exception was the comparison of stratum 1,
between the four subareas, during the winter. In
that instance, a "no decision" result was reached.

To investigate the sources of this difference in
distribution, the polychaete assemblage on
similarly numbered strata, all combinations of
subarea pairs and season were examined using
chi-square tests (a = 0.05, 11 df). Eleven of these
36 null hypotheses (2) resulted in a "no decision"
conclusion while the remaining 25 were rejected
using this analysis. In the case of stratum 2, the
null hypotheses comparing subareas IB and 2A, IB
and 3A, IB and 4A, 3A and 4A, and 2A and 4A were
rejected in all seasons. The fluctuation of this
biotic distribution in time (season) and space
(sample area) is apparent.

The homogeneity (2) of the polychaete assem-
blage between the three seasons for a single
stratum was examined using chi-square tests
(a = 0.05, 22 df). Of the eight null hypotheses of
homogeneity (2), six were rejected (i.e., both strata
1 and 2 in both subareas IB and 3A, and stratum 2
in both subareas 2A and 4A). The two remaining

942

SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA

E
E

bib

Q.
W2

■^ 01

II E
!^ a.

03 0)

.2^

c o

J= .5

o <"

o ..

E I*

E «i I

to c

_M C

3 ^

3
C

O^ P <^. .,^_>„S'~'-'. "^. ,-^'^^. '-'^

00 0<NOO00-'-<D0C)C\i0 000<b

•■-ocga)00ooin<Di-in^ococNjT-co
■»• lo eg 1- T-

<D <0 N- O)

O O ■>- o>

f_- -^ — CO ^^ ID

oooooodoc\i'-'*cocoosiooc\j
co'-'j-^oo'^o^inoot^intD'^oa)

ID (D OJ CO eg CO ■>-

. — .(D,— ^^COCDCOCD CO

^^ lO ^- ^^ O O CM O o

O0 000000'--'-T-'cJC3O0C)00

h- CDCOOO^-^ ^-^ <DO>

OOOOOOOOC0CMCN4i^i-O-^OO_O_
■<tOO'-'-OCMOCSiCOCV4-*'-0<DO'-lO

eg ■<t m T- CD (D

„ — ^ — — ^^OlCDlD'

CD CO

o CO

oooooooocococgo^mT-o_oo^
^oo'ooO'-'-mcDcocg'<s-cDCDOT-'^

in CJ) ^ r- CM ^ CO

(D O — — ^ — .'^CO^ — CO

o in^^Tfgj-cjcO'fcg^^^ .,-^".

OOO^OO^OOOC0^C0C33OOCNJOOO
r^O^OO^o'o^OO)COir)lOCDCOOOO^

CO h-. T- eg CM

coo <D — ^r^ococgcDCD^-^-— .r^

^ „ ^. ". ^ „ °. ^ "" CO "^ 9 CD 9 eg "^ '". h~:
ooooooooin'-ooioo-^^^co

oocoinoOT-oegco^cMOi-cDinoeg
t^ CO '-co in

1- •* -^t

(DcjjiDcn cDco^ — cncDr^co

o^'j-oco^^^^qcriing-cnogcggj

OOOCDOOOO^COOOr-^O^^O^CO

^•^cot^ooegcjininooincocoegr^'^a)

f^ CO CO CO '"

-^^,— ->CD >-
1^ 1- •* m o ■'f

^ eg in

OOOOOOOOCOi- '- ocooo^^o^

O N. CO CD

— -^ CO

— ^ -^ ■ -^ CD

,co — cr-co

rvi o 2 <N ;

CO eg CO

ooegCMooooegcD^eg-r-o-^oo'-

oooooocoo^'-eg

T- 1- (D 00 ^

1^ r^ CO in eg r^ Tt

CM CD CO

CO in cj) ^ — ^ in
o CM ■^ eg eg

CD - — . - — - in ^ ^ r^" CO - — ,

„>" „<^. c;r<^ "^ S ^ CO -^p <^

oooegcbo^CDO-^egcocDT-Oi— ooco

1— CNj^cDegocoO'— o<DCM^<D*<rOegh*
CO m eg eg o> cm

CO CDIDCO t^ ■ — .^-v^-> OCO,-^

o ooo '- oir^in coco^

oooooopoco▼-^oo_Olno^o^£^i
CM o T^ »- eg o

OGOOh^OK^COOOCO

1- eg -r-

y> 2

° eg

0) I

■2 <o

o ^
5^

c
CO oj

J o "

CO D)
« '^.

■5- S -

~ s ■

o
■2 o

2 <D O

2^5

O rv (0 CO

^ ^ 05 (^

OJ (1) a <0

Q--S
OO

<5 (0

n 5 C

O 0)

CD '-'•

CD —

5; S i: C ~ Q.

■^ fl^ CH r\ -— O

0) <0 ^

•2^■2 I ,

CD Q ^ Q W 2 O

~ O 0) -i;

•~ O ■= t
o o w ■-:

^ ^ (D
C

r,. " o

« (D O

■ i^ CD ^

c: 0) -c:

t CO X

"t o «?:

i- Q.
O ,«

CO o
5 CO

c: 03

P
O i~

— CDinCDCO— ~CM, . CO

r^inegocDcocMCM co

o_o_oo^ooooegoooo'^0'-oo
'-ococoino'-oinmeg'-^coco<73oco

t- CM t

.(D

00 CO h*> 03 eg coco. .. ^ . .

C)^_^q^^T-^^co^^eg~cooi-coi-t^

OOOOO^OOoOf-OO^'^T-CsioO

CMOcoocooinoco^cOT-^egrj-h^ino
T- 1- eg r- o <-

in
in

in
eg

■<»■

CO
00

■0-

■<r

in
eg

coococMOocvjoo^cOinio*^cor-^CM^
•<j- 1- eg --

CO mincD CO ,-^c73 — ^ ^^id — ^oo O)
CM egcMO o cocor^inegocsjco co
oocDOOooooiO'-oego'-CDoc)
coocMc^i^-oego^^^^03co^T-^-cooln

T- C73 ^ T-

03 03 CD — ^

<yi 00 o "-

CO ,^,-~CJ> -~. CO ^ — — . —.

.^^ . ^-.'-'. <3>03e3^^CMOT-iD eg

oooo'-oooT-egoo'-cs'j-^oeg

^oinT-cooegocof^inocOT-cnTj-o5r
>- 1- '- eg in CM T-

CD in
~. ^'-'. ^. ,-. -^ ^~. .-, . . ^^ .-, .

oooooooo g,:f.2,2,2-2.J2.2.£2.

ooT-cMOooor^03inoN.eginoO'-
r- ■'t CD CO

»- -.^ — ^ CO in
O) eg ^ o eg

— CO CO CO

cj) eg CM o

OOOOOOOOOCOCMOOO'-OOO

03 V^ o 'co' ^-v — — ^ in Id
^inh^cMin-^T-cM m

CD CO

P ,^ "-! ,^ —. -^ .— -.-.•-■ -. ■. . -. ^ -:

OOCDOOOOOOOOOCsi-T-^OOO

»-oegoootO'>tcO'-incj3t-~egcMoin
eg CO m

CD (D CO CO

^_^ ^^ o ^_^ C3 o o ^^ ^
ooooctiocbooo-'TOo

00 -.^ CO CJ3 CO —

CO in o o) o ^

o »- o o o

ooc0O'-egCMOomin'-Trc0'<tO'<j-o
in o

in CD - — .T-or^-— V ,-vCDf^

eg o CM03r^T-^ inor-

Oo'oOO'oOOCMCDOCDf-O'-OO^O

ooegoooT-0(D<D03'^coor^^^o
eg eg en

^ Tt T- T- ,-> CD ^

C33 in Tf ^ CM o in

in CO

CM CM

OOOOOOOOOOOOCMOOOOO

oo^ooo^o<Dcoeo<DCJ3cviT-oegco

1— T- t- CO

. — ^o o in <D 1- -^
in r^ ** CM o ^ r^

ooooooooejooooooj_o_oo_
oooooo^O'-ococMi-cDcnofgo

(0 2

£1

CD 3

CO 93

_ 03
(0 O)

13 S

i: o
.2.1

Q.-C
is Q.
COO

03 Q.
CD (0
O O

£ ^5
C 5 03

2 03 -^

CO 5

CD O

Q -J

<D <^ (0 CO

' ' " '5. '

i^-J c Co <^

O 03 a <o

." t: o 2

O °>"D t
Q. 3 CD -i

— O) O- (fl

(W T- CO

2 "
c

CO
0) Q.

— Co rsi
O " M
* '~ 3

Its I

~ o 03 ■;;

^ o ■- "-

O L3 to

3-2

•o =:

C 03
k- (D

«i: o

o E _
■o o .«

CO (J

25 CO

^.2 a

i: c 03

■2 SJffi

X s c

Ociaj — CiDxsc
Qco20«t0'^0l-

2 w

« U)

JC CD
♦' c

E ^
cira

3 o
o 3

C33 ~
03 <0

c to
° o

S£

r> c

1-3

943

FISHERY BULLETIN: VOL. 74, NO. 4

null hypotheses resulted in no decision. The data
from Table 2 indicate that the apparent variation
does occur in these two cases (stratum 1 in both
subareas 2A and 4A) as well.

The homogeneity (2) of the polychaete assem-
blage between strata, in a given subarea, in a given
season was also examined using chi-square tests
(a = 0.05, 12 df). Five of the 12 null hypotheses
were rejected (i.e., subarea 4A in the winter,
subareas IB, 2A, and 3A in the spring, and subarea
IB in the summer). A no decision result was
reached in the remaining cases.

Bivalves

The sampling data collected on the bivalve
populations in Garrison Bay are given in Table 3.
The data are organized as follows: 1) Protothaca
staminea, Venerupis japonica, and Saxidomus
giganfeus were grouped as one into the Veneridae;
2) Macoma incom^picita, M. irus, and M. nasuta
were grouped as one into the Tellinidae; 3) Tran-
sennella tantilla, CUnocardium nuttalli, Mya
arenaria, and Mysella fumida were considered
individually; and 4) Macoma secfa was considered
apart from the Tellinidae because of its usual
occurrence in clean sandy environments.

The size and the number of sampling units in
this study were generally inadequate for sampling
most mature bivalves. As a consequence, hypoth-
eses for small bivalves, such as T. tanfilla and M.
tumida, are best represented by the data in this
study. Indeed, large densities of T. tantilla were
found in all four subareas, with the largest
numbers in subarea 3A, and M. tumida was found
in large numbers only in subarea 4A.

The north-south bivalve distribution, as con-
structed from these data, is somewhat different
from that found in the polychaetes. The data in
Table 3 show that the subareas may be ranked in
descending order with respect to the total
numbers of individuals as follows: subareas 3A
and IB and subareas 2A and 4A are about the
same. However, occasional seasonal reorderings
do occur. The high densities in subarea 3A are
probably due to the presence of large numbers of
T. tantilla. In terms of the number of species
present, subarea IB generally ranks highest and
the remaining three subareas are almost
indistinguishable.

Differences in the bivalve distributions within a
season, on like-numbered strata, and between
subareas were examined using chi-square tests

944

<u
E

3
C«

bio
c
°C
a.

^ Ji

B.
^ I

c« o

CO ^

3 C
O e«
^ ._

C c«

I I
c< ««

> o

S m

u. ^

3 m

O 0)

c g

<u .E

o ^

-^ 2
'>

-5

3
C

00 O 00

^-i ■*■ ° — ,-, ...^ — ,

o o o o o o o"

cNj r^ CM o o o o

m m —^

(>i est -^ ^__ ^^

2.2.C.S ££2.
CO CO CD o o o o

o o

CM O

1^ 00
r- O
O O

i S2.2.S
o> o o o o

■>»■ in

s. eg ^_^ ^^

o o o o o

O CM o o o

3
Z

(0
0)

o
o>
a.
CO

in <r> ^^ (O
c\j o in o

o o ^o o^o o

c\j in <o 1- o o o

00 — ■«»•
o oj r^

o T-^ o o o o o

CM 00 CD O O O O

CO 1^ CO CD CO

P '" P ° R „
o o o o o^o^o
■<r CO CD ?^ '- o o

— .CD —

^ o CD ^ ,^ ,..^

^ cj in o o o o

r~ CO >- o o o o

CD ^-^ CO 00 CD CD

O t^ CO o o o

o t£o o o o o

CO h- 00 CM ■•- »- O

CO CO

eg CM

o o o o o o o
CO CO o Cm o T- o

»- CD . — ID in

■^ O CM O CM

O O CO o o o o

CD CO CO ^CM O O

CD ^ — ^ O) CO CD

C> t C3) CO O

O CM CD CD C3 O O

»- O ■* IT >- O O

10 ~

S 2 1.^

S ■ ==> CO '

0 o S OJ -^

»- CM

03 ."^

? <B 2 W ^ 2 3

^ — h- C *"

^^ *^ r* en — .;:

5 £ « o o ■» ^

C — E U C CD »

ai CD «- ,
> I- K .

a.
CO

<3) CM •»

00 r~. ^..^ _^ ^^ ^^ f~-.

C) o o o o o cb

in h- o o o o CD

O O CD — ~

in in o 1-

o o o o o o t£

Tf CM 7- O O O CM

CO CD

o o ^^ ^^^^^^^^

C3 C) O O O O O

"- »- o o o o o

1- ^..00 r^

'^ '^ p -^ — ,— '".

C3 CM CD O O O O
CD T- CM O O O 'J-

CD CO CO CO

O 00 CM CM

0000 000

■^ CM CO O O O b-

0000000
CD o) in o o o o

— T o -^ -.^ „ „

O O CO o o o o

o o m o o o o

T- CO

00 ^ CO
O t^ CM

o o 0^0 o o o
CM O CO o^ >- o o

CO CM ^-~ (D

<3 r- in o

CJ ci >- CD o o o

"-: ^"1 cu „ „ „ -^

O O '- o o o o

7- CD

o o o o d o o
0000^00

CD CD ^ — CO

0000^^^^^^

CD CD CD CD O O O

5 ~

? _ -2 1^

CO CO
■O T3

I 3 CO
2| ■

0.3

2 B ? <o

5 E S o o " »
c — S CJ c <o w

5 o 2 «^.^^

SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA

(a = 0.05, 6 df). All six of the null hypotheses (2)
were rejected. Thus the bivalve distributions
between subareas are different. To investigate the
sources of this difference in distribution, the
bivalve assemblage on similarly numbered strata,
in all combinations of subarea pairs and season
were examined using chi-square tests (a = 0.05, 2
df). Nineteen of these 36 null hypotheses of
homogeneity (2) were rejected. Thus, the bivalve
distribution is not consistent in either time
(season) or space (sample subareas).

The homogeneity (2) of the bivalve assemblage
between the three seasons for a single stratum
was examined using chi-square tests (a = 0.05, 4
df). Four of the eight null hypotheses were reject-
ed. The homogeneity (2) of the bivalve assemblage
between strata, in a given subarea, in a given
season was also examined using chi-square tests
(a = 0.05, 2 df ). Two of the 12 null hypotheses were
rejected. Thus, a definitive statement about the
dependence between bivalve presence and season
cannot be made. Furthermore, the differences
between the strata of a single subarea are appar-
ently minimal.

DISCUSSION

The sediment and macrofauna data collected in
the Garrison Bay study were analyzed under the
assumption of intrasample independence within
each subarea (i.e., the contents of one sampling
unit neither predicts nor influences the contents of
any other unit). The assumption is based upon the
homogeneity of macrofauna and sediment compo-
sition within study subareas. Macrofaunal
homogeneity is defined here as meaning that all
members of a given species on a given stratum are
described by the same spatial probability dis-
tribution. Although specific probability distribu-
tions were not fit to the data, chi-square and
Kolmogorov-Smirnov tests are legitimately ap-
plied to the sample data.

There are K different systematic samples, each
of size n, that could be chosen (recall N = nK); one
of these is selected at random. The sample mean of
the rth such systematic samples, y,, and the
population mean, Y, are defined respectively as:

^i =( i 2/ij)/nand

i" = ( 2 i yuVnK
i = 1 y = 1

where y^, is the attribute of interest in the sample

(e.g., the number of individuals of a given species
in the jth sample). Since systematic sampling is a
probability sampling scheme, a valid expression
for the variance of the sample mean is

var(^) = ( V {% -Yf)/K

i = 1

(Sukhatme and
Sukhatme 1970).

Alternative expressions of this quantity have been
derived (Cochran 1963). No difficulties arise in
using any of these forms of var {ij, ) in theoretical
studies, but in applications of systematic sam-
pling, no reliable estimate of var {y) is known
from taking only one sample of size n from an
area. This is a disadvantage of SSS. In practice,
approximations to var {y, ) are used as estimators
of this statistic. The texts by Cochran (1963:224-
227) and Sukhatme and Sukhatme (1970:369-370)
present several methods to approximate var {%)
from a single systematic sample. However, if m
(^2) independent systematic samples (each of size
n) are taken on the same stratum at the same time,
an exact (as opposed to an approximate) estimate
of var {%) is possible. Letting % represent the
sample mean from one of the m systematic sam-
ples, then

var {y) = ^ | j {y^ - yf/m{m - 1)

where ^ = ( i '% )/m

i = 1

(Sukhatme and
Sukhatme 1970).

In this study the estimate of the variance of the
sample means was approximated by the variance
calculated for a simple random sample (Cochran
1963). This is reasonable because of the within-
area homogeneity of the sediment and
macrofauna in each of the four study areas. Of
course, it is preferable to take at least two in-
dependent systematic samples, each of size n.

Cochran (1963) discussed the difference in
precision between random and systematic sam-
pling based on the results of these methods upon
certain types of population data. Special attention
should be given to data which is either inherently
periodic or subject to a periodic input, e.g., tidal
forces. Under these circumstances, K must be
carefully selected. Periodic variation in the north-
south direction in Garrison Bay is considered to be

unlikely.

The use of SSS allows strata to be placed at tidal
heights where experimental interest is focused.
Thus, samples may be taken at fixed tidal levels as

945

FISHERY BULLETIN; VOL. 74, NO. 4

in FLTS and statistically valid estimates of means
and variances on a stratum found. Furthermore,
no greater physical effort is required in SSS than
in FLTS. SSS also provides a method to quantify
species-presence gradients. Hence, SSS is free of
some of the disadvantages of FLTS while main-
taining the advantages often ascribed to FLTS
plans.

There is a sediment gradient in the bay in the
sense of a gradual increase in coarseness (silt to
sand) south to north over the four areas for all
seasons. Over all seasons, subarea IB generally
contained the largest number of bivalve species
and, were it not for the abundance of Transen-
nella tantilla (which is discussed later), subarea IB
would have the largest number of bivalves also. In
addition, in almost all seasons subarea IB con-
tained the largest number of individuals and
species of polychaetes. Thus, there is a distribution
in bivalve and polychaete presence, from high
density and species numbers to low as the
sediment becomes more coarse. The sediment
composition, as measured by average percentage
composition by weight of various grain sizes, is a
necessary factor to consider in predicting
macrofauna population dynamics, but it is not a
sufficient predictor by itself. This viewpoint is
based on the necessity of employing qualitative
information concerning the types of material
retained by the L981-mm sieve (see Results sec-
tion), and the role we attribute to the algae
Enteromorpha sp. in the population dynamics of T.
tantilla (see later discussion).

Newell (1965) found a higher number of
microorganisms in areas composed of finer grades
of sediment and an associated higher number of
the deposit feeders {Hydrohia ulvae and Macoma
balthica). He concluded that the large number of
microorganisms was a result of the greater sur-
face area of fine sediment grades which is related
to the amount of organic nitrogen (protein)
available to deposit feeders. The polychaete data
from Garrison Bay, and subsequent statistical
analyses, suggest that Newell's (1965) hypothesis
can be extended to incorporate a statement about
the biological effects of different sediment compo-
sitions in the presence of temporal heterogeneity.
Recall that the sediment data show that subarea
4A, the most exposed subarea, experiences greater
interseason fluctuations than does subarea IB, the
most sheltered subarea. Furthermore, the poly-
chaete assemblage in subarea IB shows the small-
est seasonal fluctuation with regard to both total

numbers of individuals and species as compared
with subarea 4A. Subareas 2A and 3A also show
smaller seasonal variations in both polychaete
assemblage and sediment composition than does
subarea 4A. All of this suggests that mixed fine
sediment grades (silty areas) may act as insulators
for certain infauna against seasonal stresses. That
is, fine sediments with their larger total surface
area to volume ratio retain larger quantities of
nutrients (organic nitrogen) and hold more inter-
stitial water. If the areas composed chiefly of fine
sediment grades occur in the cul de sac of an
embayment, where wave action is minimal, then
these areas are more likely to retain larger
numbers of individuals and species than other
areas within the embayment. Thus, despite the
periodic fluctuations in many environmental pa-
rameters of the intertidal zone, a constant
sediment particle composition contributes to a
high degree of environmental predictability.
Slobodkin and Sanders (1969), Levinton (1972),
and Gray (1974) considered aspects of the con-
sequences of temporal predictability for deposit
and suspension feeders.

The bivalves and polychaetes listed in Tables 2
and 3 represent both suspension and deposit
feeders. Rhoads and Young (1970) advanced the
hypothesis that animals of one trophic level
modify the environment and affect the dynamics
of members of another trophic level, and called it
trophic group amensalism. They found suspension
feeders in the subtidal to be generally restricted to
sandy or firm mud bottoms, and deposit feeders to
be more numerous in soft silty substrates. The
polychaete results generally support this
hypothesis. An exception noted by Young and
Rhoads (1971) was the case in which it was
hypothesized that tube-building polychaetes (both
suspension and deposit feeders) make it possible
for higher densities of bivalve and polychaete
suspension feeders to coexist with deposit feeders
in silty sediments because of their ability to bind
particles together and thereby stabilize
sediments. This hypothesis may be useful in
explaining why suspension feeders, e.g., the tube
builder Oiceniafusiformis and the members of the
Veneridae, are so numerous in subarea IB, as well
as why the tube building terebellid Thelepus
cripus, a surface level deposit feeder, reaches its
maximum density in subarea IB. The combination
of tube building coupled with the feeding behavior
of suspension feeders may provide these organ-
isms a survival advantage in this otherwise soft

946

SCHERBA and GALLUCCI: SYSTEMATIC SAMPLING OF INFAUNA

silty area. Further studies are being conducted to
develop hypotheses for Garrison Bay.

Newell's (1965) hypothesis does not appear to
explain the abundance and apparent sediment
preferences of T. tantilla. Maurer (1969) found T.
tantilla to be ubiquitous in a bay with a sediment
gradient similar to that of Garrison Bay, while
attaining its greatest numbers in a region com-
posed principally of finer sediment particle sizes.
Excluding subarea 3A, similar results follow for T.
tantilla in Garrison Bay. The increased abundance
of this bivalve in the summer on both strata of
subarea 3A indicates that the principal response of
T. tantilla may be to something other than just
sediment composition. The extensive covering of
subarea 3A by a dense algal mat of Enteromorpha
sp. is probably involved in the population explo-
sion. Transennella tantilla would gain protection
from some physiological stresses such as elevated
temperatures and increased water evaporation by
the sun and wind. Similar dense mats of Enter-
omorpha sp. were not found in the other three
areas at the sampling times.

The polychaete assemblage in Garrison Bay is
described by a distribution which is apparently
sediment and season dependent. The limited data
on the distribution of bivalves does not have the
same patterns. Preliminary analyses from an
investigation (Gallucci)^ involving the collection of
large numbers of bivalves in Garrison Bay sub-
stantiates the lack of a simple gradient relation-
ship for bivalves. Life in a calcium carbonate shell
seems to allow for greater independence from
environmental fluctuations than life near the
sediment surface without such a shell.

Although the effects of seasonal and sediment
type variations are often evident, causal links
must be established by the examination of specific
factors, e.g., competition, predation, food
availability and selection, salinity, and tempera-
ture. Toward this end, Hylleberg and Gallucci
(1975) and Gallucci and Hylleberg (1976) have
examined the role of food availability and
sediment composition upon the growth of the
deposit feeder Macoma nasuta in Garrison Bay.
Garrison Bay daylight summer surface water
temperatures are about 1°C higher in the closed
end than in the open end (Gallucci, unpubl. data),
and short stretches of intertidal areas sustain a

^Gallucci, V. F. 1976. Bivalve community relationships as
determined from age composition and growth rates. Unpubl.
manuscr., 30 p.

subsurface freshwater runoff.

In this paper we have developed an appropriate
sampling method for marine studies and the
statistical machinery for testing certain relevant
hypotheses. We have applied these methods in an
intertidal study. The biological results pertain to
sediment and animal gradients under seasonal
change. Conclusions are based upon statistical
comparisons in which the null hypothesis was
rejected, tempered by extensive biological studies.

The data and results of the Garrison Bay study
have obvious significance for shellfish culture.
Factors such as the selection of sediment type in
which to establish seed beds, interspecies associa-
tions, the season in which to make population
assessments, and the sampling techniques should
all be considered if sound management decisions
are to be made.

ACKNOWLEDGMENTS

We are grateful to A. 0. D. Willows, director of
the Friday Harbor Marine Laboratories, for
providing excellent research facilities. The coop-
eration of C. E. Lindsay, Washington Department
of Fisheries, and S. J. Zachwieja, National Park
Service, in establishing the research areas is
gratefully acknowledged. The research was sup-
ported by the National Institutes of Health
Biometry Training Grant (#67-0488) and by
WashingtonSeaGrantFunds(SG61-8227),fromthe
National Oceanographic and Atmospheric Ad-
ministration. This paper is based upon portions of
a thesis by S. Scherba, Jr. accepted by the
Biomathematics faculty, University of Washing-
ton, in partial fulfillment of the requirements for
the Master of Science degree. The untiring field
and typing assistance of Elaine Scherba is grate-
fully acknowledged.

LITERATURE CITED

BiRNBAUM, Z. W., AND R. A. HaLL.

1960. Small sample distributions for multi-sample statistics
of the Smirnov type. Ann. Math. Stat. 31:710-720.
Bloom, S. A., J. L. Simon, and V. D. Hunter.

1972. Animal-sediment relations and community analysis of
a Florida estuary. Mar. Biol. 13:43-56.
Cochran, W. G.

1963. Sampling techniques. 2d ed. John Wiley & Sons, Inc.,

N.Y., 413 p.
Conover, W. J.

1971. Practical nonparametric statistics. John Wiley &
Sons,N.Y.,462p.

947

FISHERY BULLETIN: VOL. 74, NO. 4

Faber.J. A.J.

1972. Precision of sampling by dots for proportions of land
use classes. Inst. Stat. (Raleigh, N.C.), Mimeogr. Ser. 773,
153 p.
Finney, D. J.

1948. Random and systematic sampling in timber surveys.
Forestry 22:64-99.
Gallucci, V. F., AND J. Hylleberg.

1976. A quantification of some aspects of growth in the
bottom-feeding bivalve Macoma nasuta. Veliger
19:59-67.
Gray, J. S.

1974. Animal-sediment relationships. Oceanogr. Mar. Biol.
Annu. Rev. 12:223-262.

Hylleberg, J., and V. F. Gallucci.

1975. Selectivity in feeding by the deposit-feeding bivalve
Macoma nasuta. Mar. Biol. (Berl.) 32:167-178.

Levinton, J.

1972. Stability and trophic structure in deposit-feeding and
suspension-feeding communities. Am. Nat. 106:472-486.
LOESCH, J. G.

1974. A sequential sampling plan for hard clams in lower
Chesapeake Bay. Chesapeake Sci. 15:134-139.
Loosanoff, V. L., and F. D. Tommers.

1948. Effect of suspended silt and other substances on rate
of feeding of oysters. Science (Wash., D.C.) 107:69-70.
Matern, B.

1960. Spatial variation: Stochastic models and their ap-
plication to some problems in forest surveys and other
sampling investigations. Medd. Statens Skogsforkning-
sinst. 49:1-144.
Matthiessen, G. C.

1960. Intertidal zonation in populations of Mya arenaria.
Limnol. Oceanogr. 5:381-388.
Maurer, D.

1969. Pelecypod-sediment association in Tomales Bay,
California. Veliger 11:243-249.
Newell, R.

1965. The role of detritus in the nutrition of two marine
deposit feeders, the prosobranch Hydrobia ulvae and the
bivalve Macoma balthica. Proc. Zool. Soc. Lond. 144:25-45.
Osborne, J. G.

1942. Sampling errors of systematic and random surveys of
cover-type areas. JASA (J. Am. Stat. Assoc.) 37:256-264.
Raj, D.

1968. Sampling theory. McGraw-Hill Book Co., N.Y., 302 p.
Rhoads, D. C.

1967. Biogenic reworking of intertidal and subtidal

sediments in Barnstable Harbor and Buzzards Bay,
Massachusetts. J. Geol. 75:461-476.
Rhoads, D. C, and D. K. Young.

1970. The influence of deposit feeding organisms on
sediment stability and community trophic structure. J.
Mar. Res. 28:150-178.
Russell, H. J., Jr.

1972. Use of a commercial dredge to estimate a hardshell
clam population by stratified random sampling. J. Fish.
Res. Board Can. 29:1731-1735.
Saila, S. B., and T. a. Gaucher.

1966. Estimation of the sampling distribution and numer-
ical abundance of some mollusks in a Rhode Island salt
pond. Proc. Natl. Shellfish. Assoc. 56:73-80.
Sanders, H. L.

1958. Benthic studies in Buzzards Bay. I. Animal-sediment
relationships. Limnol. Oceanogr. 3:245-258.
Sen, a. R., J. Tourigny, and G. E. J. Smith.

1974. On the line transect sampling method. Biometrics
30:329-340.
Skellam, J. G.

1958. The mathematical foundations underlying the use of
line transects in animal ecology. Biometrics 14:358-400.

Slobodkin, L. B., and H. L. Sanders.

1969. On the contribution of environmental predictability to
species diversity. Brookhaven Symp. Biol. 22:82-95.

SOKAL, R. R., AND F. J. ROHLF.

1969. Biometry: The principles and practice of statistics in
biological research. W. H. Freeman & Co., San Franc,
776 p.

SUKHATME, P. v., AND V. B. SUKHATME.

1970. Sampling theory of surveys with applications. Iowa
State Univ. Press, Ames, 452 p.

Vassallo, M. T.

1969. The ecology of Macoma inconspicua (Broderip &
Sowerby, 1829) in central San Francisco Bay. Part I. The
vertical distribution of the Macoma community. Veliger
11:223-234.

Warme, J. E.

1971. Paleoecological aspects of a modern coastal lagoon.
Univ. Calif. Publ. Geol. Sci. 87, 131 p.

WiESER, W.

1959. The effect of grain size on the distribution of small
invertebrates inhabiting the beaches of Puget Sound.
Limnol. Oceanogr. 4:181-194.

Young, D. K., and D. C. Rhoads.

1971. Animal-sediment relations in Cape Cod Bay, Mass. I.
A transect study. Mar. Biol. (Berl.) 11:242-254.

948

SIZE COMPOSITION AND GROWTH OF YOUNG ROCK CRAB,
CANCER IRRORATUS, ON A ROCKY BEACH IN MAINE^

Jay S Krouse^

ABSTRACT

Monthly hand collections of small rock crab, Cancer irroratus, were made from an intertidal zone in
East Boothbay, Maine, from June 1972 through April 1975. An analysis of size and sex frequencies
indicated: 1) young-of-the-year crabs (<5 mm carapace width) entered the intertidal area in late
summer-early fall and remained there through the second fall with a resultant width range between 15
and 40 mm; 2) a deceleration and/or cessation of growth in winter; 3) an emigration of crabs >40 mm
carapace width from the intertidal area associated with declining winter temperatures and/or
behavioral changes; 4) sex ratios approximated a 1:1 relationship; and 5) small male and female rock
crabs (<60 mm carapace width) had a common growth rate.

While searcing beneath the rocky substrate of an
intertidal zone for juvenile American lobster,
Homarus americanus Milne Edwards, whose early
distribution and abundance is generally unknown,
I discovered numerous small rock crab, Cancer
irroratus Say, burrowed under the rubble.
Because rock crab is a valuable commercial species
as well as an important food source of lobsters
(Ennis 1973), I believe it important to describe the
distribution of young crabs in their natural envi-
ronment along with other life history information
(size structure, sex ratio, and growth).

METHODS

Rock crabs were carefully hand collected about
once a month during extreme low slack tides from
the intertidal zone of Grimes Cove, East Boothbay,
Maine (Figure 1). The rocky substrate of this
unsheltered seaward cove consists of rocks of
assorted sizes intermingled amongst areas of
bedrock, sand, and pulverized shells. By using
large boulders as landmarks, it was possible to
sample consistently the same general area near
the low water mark. Unfortunately, for various
reasons, samples could not be obtained for all
months of the study.

After two biologists concurrently expended 1 h
gathering crabs, their catches were immediately
returned to the laboratory where sex and carapace

'This study was conducted in cooperation with the National
Marine Fisheries Service, NOAA, Department of Commerce,
under Public Law 88-309, as amended, Commercial Fisheries
Research and Development Act, Project 3-153-R.

'Maine Department of Marine Resources, West Boothbay
Harbor, ME 04575.

Manuscript accepted April 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

width (distance between the two most posterior
notches on the anterolateral border) to the nearest
millimeter were recorded. The sex of crabs < 10 mm
carapace width (CW) was determined under a
dissecting microscope.

Width-frequency histograms were compiled by
2-mm increments for rock crabs caught each
month from June 1972 through April 1975.

RESULTS AND DISCUSSION
Size Composition and Seasonal Distribution

Since there were no discernible differences in
size distribution between male and female crabs,
the data for sexes were combined in monthly
width-frequency histograms (Figure 2). This
similarity in size composition of male and female
crabs <60 mm CW suggested a common growth
rate for both sexes up to this size, unlike the
marked size disparity of larger male and female
rock crabs (>60 mm CW) caught in commercial
lobster traps which was primarily attributed to a
decrease in the growth rate of females after the
onset of sexual maturity (Krouse 1972).

Modal groups, which most likely represented
one or, perhaps, more molt classes, were quite
conspicuous in each of the monthly histograms.
However, due to extensive overlapping of modes I
was unable to quantitatively follow these modal
groupings from month to month for purposes of
estimating mortality rates.

Inspection of monthly histograms revealed that
young-of-the-year crabs (recently metamorphosed
from megalops to first crab, <5 mm CW) initially

949

Of^f-^K^

FISHERY BULLETIN: VOL. 74, NO. 4
69 'SS"

Figure 1. -Chart of Boothbay region
with a seaward view at low tide of the
intertidal beach of Grimes Cove, East
Boothbay, Maine.

appeared during September 1972 and late August
1973 and 1974. This seasonal appearance of young
crabs agreed with earlier observations of female
rock crabs hatching their eggs in late spring and
early summer in Maine waters (Krouse 1972) and
the culture work of Sastry (1970) which demon-
strated that 40 - 60 days are required for rock crabs
to develop through the pelagic larval stages to the
first crab stage at 15°C and a salinity of 307oo.

Histograms showed a gradual upward progres-
sion of the first modal grouping (comprised of

950

young-of-the-year crabs, <10 mm CW) from
August through December 1974, while distribu-
tions from January through April 1975 revealed
relatively little change (Figure 2). This apparent
cessation of growth was further supported by
sighting very few, if any, cast exoskeletons and/or
soft-shelled crabs while sampling during the
winter. At other times of the year, when crabs
were growing, numerous recently cast shells
and/or shedders were readily observed. In spring
when growth resumed, the percentages of in-

KROUSE: SIZE AND GROWTH OF ROCK CRABS

>-
O
Z 3

Figure 2.-Width-frequency distribu-
tions for rock crabs collected monthly
by hand at an intertidal area in East
Boothbay, Maine, 1972-75.

1972 ■

- 74

HAND COLLECTED

i:S TRAP COLLECTED

X =23.93

X 78.84

" =2426

" '5480

»;=Q25

= 0.18

CARAPACE WIDTH. MM

dividuals <10 mm CW began to diminish progres-
sively until late summer when young-of-the-year
crabs once again settled to the bottom (Table 1).

Rock crabs >40 mm CW were decidedly less
abundant during the fall and winter (Table 1). In
fact, not a single crab >50 mm CW was captured
from January through April (Figure 2). This
seasonal shift in size distribution suggested that
crabs >40 mm CW moved seaward from the inter-
tidal zone with declining temperatures. Jeffries
(1966) reported that C. irroratus moved from
Narragansett Bay, R.I., in winter to the deeper
and warmer ocean waters. Conversely, in southern
waters during the late fall and winter, rock crabs
moved into Delaware Bay (Winget et al. 1974) and
inshore waters of Virginia (Shotton 1973; Terretta
1973) as the water temperatures fell within a
preferred range.

Aside from the apparent thermal effects and/or

Table 1.- Percentage of rock crabs of two carapace widths in
monthly samples for 1972-75.

S10 mm

s 40 mm

S10 mm

2 40 mm

Month

(%)

Month

(%)

Jan.

39.6

Aug.

1.6

16.6

Mar.

47.4

2.6

Sept.

13.7

12.7

Apr.

34.8

2.7

Oct.

31.4

6.7

May

32.8

2.9

Nov.

40.9

8.0

June

7.8

8.5

Dec.

40.7

3.7

July

2.4

8.8

behavioral changes on the seasonal displacement
of these large crabs from the intertidal zone, this
movement may also be associated with the larger
crabs' physical ability to emigrate with ease from
an area of low temperature. In addition, the size of
these crabs may inhibit their ability to find suit-
able burrows in the littoral zone necessary to
afford protection from the often tempestuous
winter sea. Jeffries (1966) stated that C. irroratus
was not well suited for burrowing into coarse
bottom.

Sex Ratio

Ratios of males to females for each of the
monthly samples ranged from 0.60:1 to 1.57:1
(Table 2). The chi-square test revealed that only
sex ratios of catches of July and August 1973
deviated significantly (P = 0.05) from a 1:1 rela-
tionship. Thus I concluded that sex ratios of the
intertidal catches approximated a 1:1 relationship
(1,353 males: 1,376 females); whereas, rock crabs
larger than 50 mm CW collected in traps near
Boothbay Harbor, Maine, showed disproportion-
ate sex ratios which varied by season and locality
(Krouse 1972). It appears that these disparate sex
ratios were primarily a function of the onset of
sexual maturity which subsequently altered the
growth rate and seasonal distribution of male and
female crabs.

951

FISHERY BULLETIN: VOL. 74, NO. 4

Table 2.-Sex ratios of the monthly collections of rock crabs
taken intertidally in East Boothbay, Maine, 1972-75. Sex ratios
that deviated significantly {P = 0.05) from 1:1 are marked '.

Mo.

1972 1973 1974 1975
M:F M;F M:F M:F

Mo.

1972 1973
M:F M:F

1974

M:F

1975
M:F

Jan.

1.27:1 — 0.96:1 Sept. 0.89:1

1.57:1

Mar. — —
Apr. — 0.60
May — 1.07
June 1.02:1 0.95
July 1.25:1 1.55
Aug. 1.01:1 0.71
0.87

0.76
1
1.03
0.82
0.86

1.23:1
1 1.08:1
1 —
1 —
1 —
1 —

0.71:1
0.68:1 —

Oct. 0.68:1 — 1.23:1 —
Nov. — — 0.96:1 —

Dec. — — 0.93:1 —

Total 0.98:1 0.99:1 0.95:1 1.08:1

Growth

Carapace width prior to shedding was plotted
against the new carapace width after shedding for
45 crabs that molted while captive in the labora-
tory. This relationship was fitted by the method of
least squares using the simple linear equation
Y = Si + bX, where 7 = postmolt CW,
X = premolt CW, and a and b were constants.
Analysis of covariance, which was used to test
homogeneity of regression coefficients, revealed
no statistical differences between growth in-
crements of males and females, so all data were
pooled. The calculated equation for crabs ranging
from 9 to 48 mm CW was Y = 0.566 + 1.247X.
This relation was similar to Cleaver's (1949) con-
stants (a = 0.57; b = 1.23) calculated for Dunge-
ness crab, C. magister, juvenWes (5-91 mm CW).

Based on the relationship between premolt vs.
postmolt and measurements of cultured post-lar-
val crabs (stages I-V) obtained from Herbert C.
Perkins, formerly of the National Marine Fish-
eries Service, West Boothbay Harbor, Maine, I
estimated sizes for instars I-XIII (Table 3). Sizes
for instars above XIII were not computed because
of the inherent uncertainties of extrapolating
beyond the data range. If we assume that Maine
rock crabs begin to attain maturity about 60 mm
CW (Krouse 1972; Scarratt and Lowe 1972) and if
as suggested by Butler's (1961) work with C.
magister the premolt vs. postmolt relationship
changes with the onset of sexual maturity, then
sizes for instars beyond XIII (53 mm CW) are
inadequately described by the aforementioned
regression.

Because the increments of growth (24.3-28.3%)
for post-larval crabs (instars III-V) cultured in the
laboratory were appreciably less than those
growth increments (29.2 to 30.6%) for instars
VI-VIII of the captive wild crabs, widths for
instars II-XIII were estimated by the empirical

value (2.6 mm CW) for stage I and then the
subsequent stages were calculated with the linear
regression (Table 3). Instar sizes calculated by this
procedure were larger (about one instar size
greater) than those sizes based on empirical data
for stages I-V and predicted by regression for
instars VI-XIII, e.g., instar V (estimated by
regression) = 9.5 mm and instar VI (other
method) = 9.6 mm. For purposes of this study,
those instar sizes calculated with the empirical
post-larval data were favored.

Table 3. -Comparison of instar sizes of rock crabs. For one
group, instars I-V represent actual measurements and instars
VI-XIII are calculated by the relationship Y = 0.566 + 1.247Z;
for the other group, instar I is an actual measurement and the
remaining instars are estimated from the aforementioned
equation.

l-V: Actual measurements

VI-XIII: reg

ression values

Regression values

Carapace

Increase

Carapace

Increase

Instar

width (mm)

(%)

width (mm)

(%)

2.6

3.7

42.3

3.8

46.4

III

4.6

24.3

5.3

39.5

5.9

28.3

7.2

35.3

7.4

25.4

9.5

32.5

9.6

30.3

12.4

30.6

VII

12.5

30.6

16.1

29.2

VIII

16.2

29.2

20.6

28.2

20.8

28.2

26.3

27.4

26.4

27.4

33.3

26.8

33.5

26.8

42.1

26.4

XII

42.4

26.4

53.0

26.0

XIII

53.4

26.0

66.7

25.7

An attempt was made to objectively assign size
with age by correlating instar size with the
monthly width-frequencies (Figure 2). As men-
tioned previously, post-larval crabs (2-5 mm CW)
first entered the sampling area in August or
September after having hatched in late spring or
summer and having developed through the larval
stages during the remainder of the summer. Size
distributions for April 1974 through April 1975
revealed first entry of young-of-the-year crabs in
August followed by subsequent growth of young
crabs until about January when growth ceased and
this modal group stabilized at about 4-20 mm CW
(instars III-IX). This wide size range is best
explained by varying hatching and settling dates
whereby some crabs entered the population per-
haps 1 to 2 mo later than the rest. These crabs
settled to the bottom when temperatures were
likely to be declining; thus these individuals
experienced little growth until the following
spring.

952

KROUSE: SIZE AND GROWTH OF ROCK CRABS

-.n

APRI
22.5
30

(NIK C

■

jrsir^ —
(SiK e «

m
at

-S «
otOrMin

< . ..

. mix e

s§£§

. r. M c •

. lO IK c •

"''.OS

ZoduoS

. OIK C i"

§Ss§

2 ' "«

COlK C «

-I L.

J l_

-I ■— I 1 ■ ■ ■

c
o

'Si

u
>,

o
o
PQ

■a

3
O"

£2 « ^

<N CO ■<»•

M 00 *

<N 00 *

(NOO* CMOO^ <N0Ot tNOO't (NOO^

A3Nano3ad iNjDiijd

953

FISHERY BULLETIN; VOL. 74, NO. 4

When growth resumed in late spring and sum-
mer, only a few juveniles (previous year's young-
of-the-year) were <10 mm CW (stage VI). Unfor-
tunately, the overlapping of modes prevented any
objective determination of the upper limit of this
second modal grouping that represented the
juvenile crabs. Nevertheless, several of the
monthly distributions (September-October 1972
and August 1973) exhibited an upper limit around
40 mm CW for this second modal grouping (Figure
2).

Although it was not possible to make precise
determinations of age and growth from informa-
tion of this study, the data suggest that young-
of-the-year crabs ranged in size from about 4 to 20
mm CW (instars III-IX) and the juveniles in their
second fall ranged from approximately 15 to 40
mm CW (instars VIII-XII).

Hand -Collected Vs.
Trap-Caught Crabs

Prior to sampling small crabs in the intertidal
zone, our rock crab work was based on incidental
catches of crabs with wire lobster traps (25.4-
x25.4-mm mesh) in the Boothbay Harbor area
(Krouse 1972). Histograms plotted by 1-mm in-
crements for 2,426 hand-collected and 5,480 trap-
caught crabs (1972-74) graphically revealed
marked differences in size composition of the
catches for these two methods of capture (Figure
3). Average width for hand-collected crabs was
23.9 mm and 78.8 mm for trapped crabs. Even
though these two complementary modes of cap-
ture sampled a broad range of sizes (2-133 mm
CW), many crabs between 40 and 60 mm CW
eluded either type of collection. This scarcity of
crabs between 40 and 60 mm CW can be attributed
to: 1) selectivity of traps against sizes <60 mm CW
(based on Figure 3, crabs <70 mm CW were not
fully vulnerable to the gear), and 2) movement of
crabs >40 mm CW from the intertidal zone in
association with low winter temperatures and
possible behavioral changes with size. Scarratt and
Lowe (1972) reported that small rock crabs (<65
mm CW) in the Northumberland Strait, Gulf of St.
Lawrence, inhabited rocky areas, whereas larger
crabs left the rocky substrate for sand and mud

bottoms. Jeffries (1966) noted that C. irroratus
dwelled chiefly on sand in the Narragansett
Bay- the type of bottom this species is adapted to
because of its well developed walking and bur-
rowing abilities.

ACKNOWLEDGMENTS

I thank A. Dolloff, C. Crosby, and D. Libby for
their assistance with field collections and data
compilations. I am also grateful to J. C. Thomas
for his review of this paper.

LITERATURE CITED

Butler, T. H.

1961. Growth and age determination of the Pacific edible
crab Cancer magister Dana. J. Fish. Res. Board Can.
18:873-891.
Cleaver, F. C.

1949. Preliminary results of the coastal crab (Cancer
magister) investigation. Wash. Dep. Fish., Biol. Rep.
49A:47-82.
Ennis, G. p.

1973. Food, feeding, and condition of lobsters, Homarus
americanus, throughout the seasonal cycle in Bonavista
Bay, Newfoundland. J. Fish. Res. Board Can. 30:1905-1909.
Jeffries, H. P.

1966. Partitioning of the estuarine environment by two
species of Cancer. Ecology 47:477-481.
Krouse, J. S.
1972. Some life history aspects of the rock crab. Cancer
irroratus, in the Gulf of Maine. J. Fish. Res. Board Can.
29:1479-1482.
Sastry, a. N.

1970. Culture of brachyuran crab larvae using a re-circulat-
ing sea water system in the laboratory. Helgolander wiss.
Meeresunters. 20:406-416.
Scarratt, D. J., and R. Lowe.

1972. Biology of rock crab (Cancer irroratus) in Northum-
berland Strait. J. Fish. Res. Board Can. 29:161-166.

Shotton, L. R.

1973. Biology of the rock crab. Cancer irroratus Say, in the
coastal waters of Virginia. M. A. Thesis, Univ. Virginia,
Charlottesville, 72 p.

Terretta, R. T.

1973. Relative growth, reproduction and distribution of the
rock crab. Cancer irroratus, in Chesapeake Bay during the
winter. M.A. Thesis, College of William and Mary, Wil-
liamsburg, Va., 104 p.

WiNGET, R. R., D. Maurer, and H. Seymour.

1974. Occurrence, size composition and sex ratio of the rock
crab. Cancer irroratus Say and the spider crab, Lihinia
emarginata Leach in Delaware Bay. J. Nat. Hist.
8:199-205.

954

MINIMUM SWIMMING SPEED OF ALBACORE, THUNNUS ALALUNGA

Ronald C. Dotson'

ABSTRACT

Measurements of density and pectoral lifting area of albacore, Thunnus alalunga, were made and
compared with those previously described for yellowfin tuna, Thunnus albacares; bigeye tuna,
Thunnus obesus; and skipjack tuna, Katsuwonus pelamis. Albacore have densities within the range of
yellowfin tuna of similar size. The pectoral lifting area of albacore was always greater than skipjack
tuna but similar to yellowfin tuna and bigeye tuna for fish less than 70 cm long. Larger albacore had
increasingly larger fins than did the other species.

Minimum speed necessary for hydrostatic equilibrium of albacore was calculated and compared at 50
and 80 cm fork lengths to values calculated for the species above. Albacore minimum speeds were slower
than those for skipjack tuna, similar to those of yellowfin tuna, and greater than those of bigeye tuna.
Density variations of albacore, attributed to fat content and gas bladder volume, significantly affected
estimates of minimum speed. Calculated speeds were slower than those estimated for albacore tracked
at sea or estimated from tag returns.

Albacore tuna, Thunnus alalunga (Bonnaterre),
being negatively buoyant in seawater, must swim
continuously to maintain their position in the
water column. The albacore's long pectoral fins
help to compensate for their negative buoyancy by
providing lift, thus lowering the swimming speed
necessary to maintain hydrostatic equilibrium.

A model developed by Magnuson (1970) proposes
that the minimum swimming speed of a scombrid
fish is set by the necessity to maintain hydrostatic
equilibrium rather than to provide adequate gill
ventilation. When the lift provided by the pectoral
fins necessary to compensate for the weight of the
fish in water is estimated, the corresponding
swimming speed can be considered the minimum
necessary for the maintenance of hydrostatic
equilibrium. This model was used by Magnuson
(1973) to compare minimum speeds of several
species of scombrid fishes that diff'ered in pectoral
lifting area, body shape, body density, and the
presence or absence of a gas bladder.

The purpose of this paper is to 1) estimate the
minimum swimming speed of albacore; 2) compare
the minimum swimming speed of albacore with
those for other scombrids; and 3) compare cal-
culated minimum swimming speeds of albacore
with swimming speeds estimated from sonic
tracking of albacore at sea and from long distance
tag returns.

'Southwest Fisheries Center La Jolla Laboratory, National
Marine Fisheries Service, NOAA, La Jolla, CA 92038.

MATERIALS AND METHODS

To compute the minimum swimming speed with
Magnuson's (1970) model, it is necessary to deter-
mine the mass of the fish, the lifting area, the
density of the seawater, and the density of the
fish. As the peduncle keels probably provide neg-
ligible lift (Magnuson 1973), they are excluded in
the computation of minimum speeds.

To determine the mass of albacore, 477
specimens caught between long. 130° and 140°W
and lat. 30° and 40°N during June 1974 were
weighed to the nearest gram on a magnetically
dampened pan balance and their fork lengths
recorded to the nearest millimeter. Specimens
were weighed and measured within 15 min after
capture.

A regression In M = In a + 6(ln L), where M is
mass in grams and L is fork length in millimeters,
was fitted to the length-mass data. The resultant
equation was

M = 4.514 X lO-^L^ ^^-^^

(1)

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

with 95% confidence limits on the exponent from
2.8245 to 2.9246.

The total pectoral lifting area (A) is equal to the
projected surface area of the pectoral fins plus the
projected body area between them, due to their
analogy to wings in which the pressure distribu-
tion set up by the wings extends across the fu-
selage (Magnuson 1970). The pectoral lifting area
was determined by tracing the outline of the

955

FISHERY BULLETIN: VOL. 74, NO. 4

detached left pectoral fin on a piece of paper
having a thickness of 0.25 mm. The outlined form
was cut out and w^eighed to the nearest 0.01 g on a
balance. Projected area was calculated from a
ratio of the paper weights to a standard, and
doubled to account for the other pectoral fin.

Thirty-three 100 cm- pieces of the paper were
measured with a micrometer and weighed to
determine the affect of variations in paper thick-
ness and cutting accuracy on the calculations. The
thickness of the paper varied less than 1% around
the mean, and cutting accuracy varied by ± 2%.
The affect on calculations of pectoral lifting area
was, therefore, assumed to be negligible.

The projected body area between the pectoral
fins was determined by multiplying the body
width at the pectorals by the width of the pectoral
fin at its point of attachment to the body as
measured on the fresh fish. Pectoral fin lifting
areas were determined for 42 fish caught in the
area described above and for 8 larger fish caught
off Oregon in October 1974. The following rela-
tionship was established between the lifting area
{A in square centimeters) and the fork length (L in
millimeters):

4.7351 X 10- ''L-*''".

(2)

Albacore observed cruising under the baitboat
kept their pectoral fins extended continuously at a
sweepback angle of approximately 45°. The tips of
an albacore 's fins are also not rigid, and the effect
of this on the lifting capacity of the fin has been
ignored.

A water density (7)^) of 1.025 g/ml was deter-
mined from temperature and salinity data from
the offshore region described above. This also
equalled the mean water density within the near-
shore albacore fishery.

Fish densities were determined for three groups
of fish: group 1— seven fish caught during June
1974 in the offshore region; group 2—14 fish caught
60 miles south of San Diego on 23 July 1975,
presumably 2 wk after they appeared off the coast;
and group 3-37 fish caught on 13 September 1975
in the same region as group 2 but assumed to have
been near the coast for 2 mo.

The group 1 fish were frozen immediately after
capture and, when returned to the laboratory,
thawed and weighed on a spring balance while
suspended in seawater to determine the density of
the fish in seawater (Df).

Fish from groups 2 and 3 were weighed in
seawater on a pan balance immediately after
capture and their densities in seawater
determined.

VARIATIONS IN DENSITY

The density of group 1 fish (Figure 1) is well
within the range of those determined for fresh fish
of similar size indicating that freezing and thaw-
ing probably had negligible affect on density
determinations. All specimens were caught on or
near the surface by jigline or rod and reel, and
there was no difference in density attributable to
one method of capture over the other.

Rough estimates of the development of the gas
bladders of 21 fish in group 3 were made im-
mediately after other measurements were com-
pleted. In specimens less than 56 cm FL (fork
length), the bladder was small (approximately 1
cm wide and 8 or 9 cm long) and contained little or
no gas. In specimens 60 to 70 cm FL, the bladder
was approximately 5 cm wide and 16 cm long and
filled with gas to a depth of 4 or 5 cm. Fish over 80
cm FL had bladders approximately 30 cm long and
10 cm in diameter which occupied a large volume of

I.IIO-

I.IOO -

^ 1.0901-

^ 1.0801-

(/)

1.070 -

S 1.060 -
1.0501-

« X

45 50 55 60 65 70 75

FORK LENGTH (cm)

Figure 1. -Computed densities for
three groups of albacore: group 1
(triangles); group 2 (crosses); and
group 3 (dots). See text for explana-
tion of groups.

956

DOTSON: SWIMMING SPEED OF ALBACORE

the gut cavity. All developed gas bladders ap-
peared full with two exceptions, and these may
have been damaged during capture or dissection.

Seven albacore caught in September with fork
lengths of 63 to 68 cm were examined to determine
the effect of the gas bladder on density. Gas was
removed from the bladder by a cannula (inside
diameter = 1 mm) which was inserted through the
ventral surface of the body while the fish was
submerged in seawater, and the fish was then
weighed while still submerged. The mean density
increase with gas extraction was 0.007 g/ml (Table
1). Although this is probably a conservative es-
timate, the difference in density calculated before
and after gas removal is used as the effect of the
gas bladder on fish density. In an albacore less
than 56 cm FL, the small gas bladder is not
expected to affect density whereas the large and
fully developed gas bladder of albacore greater
than 80 cm FL should reduce density to a greater
extent than was measured on the smaller fish
above.

Densities of group 2 fish were considerably
higher than those of similar size fish in groups 1
and 3 (Figure 1). Seasonal variations in density
due to changes in fat content have been described
for other pelagic species by Aleev (1963). Mass
estimates were calculated from the length for each
fish in all three groups using Equation (1), and
compared with the observed values. The mean of
the observed values for group 2 fell 403 g below the
estimate from the regression line, ranging from
172 g greater to 999 g less. Because fish in group 2
had apparently just migrated into the area of
capture, presumably from the central or western
Pacific, the loss in mass was assumed to have been
caused by the utilization of fat during migration.
Group 1 would not yet have utilized this fat, and
group 3 is assumed to have added fat by feeding in
the rich coastal waters.

The densities in group 2 were recomputed on the
assumption that the mass difference between the
individuals and the regression curve is attributed
to fat loss. An equation was developed by Magnu-
son (1970) relating the density {Df) of a scombrid
without a gas bladder to the percentage (P) of the
total body weight that is fat. The equation

Df = 1.100- 0.0017 P

(3)

was used to recompute densities for the fish in
group 2. The effect of the gas bladder on density
was assumed to be 0.007 g/ml because fish in group

2 were in the same size range as the above fish for
which gas bladder measurements were taken. This
value was added to the observed density and the
percentage body weight in fat calculated. The
difference in mass (assumed to be fat loss) of each
individual was then added and new densities
determined with the increased percentage of body
fat. The density effect of the bladder was sub-
tracted from this value to yield a density adjusted
for fat loss. When determining fat content in the
fish, the density effect of the bladder was taken
into account, except for those fish with measured
densities greater than 1.100 g/ml, which is the
level Magnuson (1970) chose as the density for a
scombrid without a gas bladder. Fish with densi-
ties greater than 1.100 were assumed to have
empty or damaged gas bladders, and the density
difference due to the gas bladder was subtracted
from the recomputed density.

Recomputed densities of group 2 are plotted in
Figure 2 with the measured densities of groups 1
and 3. The close fit of the recomputed densities
appears to support the assumption that fat con-
tent and gas bladder volume can account for the
disparity in densities observed for group 2 in the
original data. Density values are, therefore, ex-
pected to vary considerably depending on the
development and condition of the gas bladder and
the fat content of the fish when it is caught.

DETERMINATION OF
MINIMUM SPEED

To estimate the minimum speed for hydrostatic
equilibrium, it is necessary to calculate the amount
of lift a fish must produce. The lift (Zy) required by
a scombrid to attain hydrostatic equilibrium,
expressed in dynes, is determined from the rela-
tion (Magnuson 1970)

Lf = M

^l-^)980cm/s2 .

(4)

When the lift is assumed to be provided solely by
the pectoral fins, and the coefl^cient of lift for the
pectorals is assumed to be 1.0, then the equation
for minimum swimming speed becomes
(Magnuson 1970)

V =

i^i

/2{A)j

(5)

Calculations of minimum swimming speed from

957

FISHERY BULLETIN: VOL. 74, NO. 4

Table 1. -Density changes in albacore following gas bladder deflation and resultant eflFect on estimation of minimum speed.

Albacore no.

Characteristic

Mean

627

643

646

648

653

679

654

1.058

1.063

1.062

1.055

1.061

1.060

1.058

1.059

1.063

1.069

1.068

1.058

1.067

1.071

1.066

0.005

0.006

0.003

0.006

0.007

0.013

0.007

45.0

48.6

48.7

44.5

49.8

47.3

47.9

47.4

47.9

52.2

52.4

46.5

53.9

51.7

56.3

51.6

2.9

3.6

3.7

2.0

4.1

4.4

8.4

4.2

Fork length (mm)

Density with gas bladder (g/ml)

Density without gas bladder (g/ml)

Change in density due to gas bladder (g/ml)

Minimum speed with gas bladder (cm/s)

Minimum speed without gas bladder (cm/s)

Change in minimum speed due to gas bladder (cm/s)

Figure 2.-Computed densities for
three groups of albacore with group 2
densities (crosses) recomputed after
correction for fat loss. Recomputation
of group 2 densities is explained in the
text.

>-
H
if)
Z
UJ

I.IIO-

i.ioo -

1.090

1080

1.070

1.0601-

1.050

^■^

1 1

" i

•

•
•
•

• •

•

•
•

•

« " . .

•

•«

•
•

a ,. 1

•

■

this equation assume 100'^ extension of the pecto-
ral fins.

The mass of the fish [M) and the lifting area (,4)
can be calculated using Equations (1) and (2),
respectively. The density of the environment {D^)
is 1.025 g/'ml. If we use M = 2,540 g, D, = 1.082
g/ml, A = 77.4 cm- for a 50-cm fish, the calculated
minimum speed T' is 54 cm/s.

Density variations due to fat content and gas
bladder volume can affect the minimum swimming
speed necessary to maintain hydrostatic equilib-
rium. For a 65-cm albacore, a loss of 10*^ of its body
weight in fat would result in a 10*^ increase in
minimum speed. Loss or emptying of the gas
bladder results in an 8% increase in minimum
speed. Minimum speeds calculated from data on
fish with full gas bladders and in good condition
are therefore considered to be the minimum ob-
tainable while retaining hydrostatic equilibrium.

COMPARISON OF MINIMUM
SPEEDS OF FOUR SCOMBRIDS

Minimum speeds were calculated for albacore;
yellowfin tuna, Thunnus albacares;higeye iunsL, T.
obesus; and skipjack tuna, Katsunvnus pelamis, at
fork lengths of 50 and 80 cm. The speeds are given
in Table 2 with the density, mass, and pectoral
lifting area used in the computations.

The minimum swimming speed of albacore

55 60 65 70 75

FORK LENGTH (cm)

T.\BLE 2.-Estimated minimum speeds of four species of scom-
brids at fork lengths of 50 and 80 cm. The mass of the fish (M),
pectoral lifting area (A), and density of the fish (D) used in the
computations are also given.

Fork

length

Species

(cm)

(9)

(cm2)

(g/ml)

(cm/s)

Thunnus

2,588

77.40

1.082

alalunga

9,992

271.04

1.056

Thunnus

2,429

96.63

1.047

obesus^

10,825

233.80

M.030

Thunnus

2,501

91.56

1.087

albacares*

10,338

220.50

1.050

Katsuwonus

2,539

47.88

1.090

pelamis*

12,567

137.20

1.096

107

Data from present paper.
^M, A, and D, from Magnuson (1973).
^Extrapolated value.
*M calculated from Chatwin (1959),
(1973).

A and 0, from Magnuson

decreases from 57 cm/s when they are 50 cm FL to
45 cm/s at 80 cm FL. The decrease is a direct result
of the allometric growth of the pectoral fins
(Yoshida 1968) and the gas bladder (Gibbs and
Collette 1966). The gas bladder of albacore does
not have significant development when the fish is
less than 55 cm long, but has considerable volume
at a fish length of 65 cm, and apparent complete
development when the fish has reached 80 cm FL
(data, this paper). Combined with the increasing
length of the pectoral fins, the result is a relatively

958

DOTSON: SWIMMING SPEED OF ALBACORE

abrupt drop in minimum speed between 60 and 70
cm FL (Figure 3).

Albacore and yellowfin tuna have very similar
densities (Table 2), but the pectoral fins of albacore
are smaller in young fish (Gibbs and Collette 1966),
increasing very rapidly in size as the fish mature
(Figure 4). Thus, small albacore have a faster
minimum swimming speed than small yellowfin

120 r

^ no -

o 100

Gj 90
UJ

w 80

^ 60
(/)

2
z>

t-
w

.jr D^lomit

T gibccorts

T.oiMiua

40 50 60 70

FORK LENGTH (cm)

Figure S.-Ttie estimated minimum swimming speed of four
scombrids using Magnuson's (1970) model for hydrostatic
equilibrium.

400

< 300

p 200

-J

<
tt.

t-
o

100

T otelifvo

T.ot*tvt
T altccarii

JH omicmis

0L_L
40

50 60 TO 80

FORK LENGTH (cm)

Figure 4.-Regression cunes for pectoral lifting area (A) versus
fork length. Curves for Thunnus obesus, T. albacares. and
Katsuurjnus pelamis are from Magnuson (1973). The cune for T.
alalunga is from Equation (2) in the text.

tuna, and albacore over 65 cm have a slower
minimum swimming speed than the same size
yellowfin (Table 2, Figure 3) assuming similar fat
content and gas bladder development.

Because bigeye tuna have a larger gas bladder
than albacore and also have large pectoral fins,
both of which grow allometrically (Gibbs and
Collette 1966), their estimated minimum swim-
ming speed is only half that of albacore at both 50
and 80 cm in length (Table 2).

The minimum swimming speed necessary for
hydrostatic equilibrium of 50-cm albacore is 70%
that of 50-cm skipjack and only 40% when each is
80 cm long (Table 2). Unlike albacore, skipjack
have no gas bladder and always have small, short
pectoral fins; therefore, skipjack tuna must swim
faster as their mass increases in order to maintain
hydrostatic equilibrium (Figure 3).

In Table 2 and Figure 3, density values for
bigeye tuna were extrapolated beyond observed
values and those of albacore were chosen from
"fat" fish; therefore, actual values shown may not
be exact, but the gross relationships among species
are expected to hold true.

FIELD ESTIMATES OF
ALBACORE S^TMMIXG SPEEDS

During August 1972, the National Marine Fish-
eries Sen-ice in cooperation with the American
Fishermen's Research Foundation tagged six
albacore with sonic tags and tracked their
movements off the coast of Monterey Bay, Calif.
(Laurs et al. 1972).2

Mean speeds observed during sonic tracking of
three fish near 85 cm fork length were 95 cm/s
during daylight hours and 62 cm/s during the
night. These speeds are higher than the calculated
minimum of 42 cm/s for a fish this size.

Each of two tagged albacore approximately 80
cm long, which were caught after a trans-Pacific
migration, had a computed minimum or straight
line speed (based on great circle route and time
free) of 26 nautical miles/day or 55 cm/s (Japanese
Fisheries Agency 1975). The calculated minimum
speed of 45 cm/s is remarkably close to the es-
timated minimum migration speed of these two

2Laurs, R. M., H. S. H. Yuen, and J. H. Johnson. 1972. Study of
the smaU-scale movements of albacore using ultrasonic tracking
techniques. In Report of Joint National Marine Fisheries
Ser%-ice-American Fishermen's Research Foundation Albacore
Studies Conducted during 1971 and 1972, p. 54-72. Unpubl. Rep.
SWFC,NOAA,LaJolla.

959

fish but could be an artifact of many interacting
processes and events.

ACKNOWLEDGMENTS

I thank G. D. Sharp, R. M. Laurs, and L. C. Chen
for their support and assistance. G. D. Stauffer
provided the programs for regression analysis. J.
J. Magnuson and W. H. Neill reviewed the manu-
script and made many helpful suggestions.

LITERATURE CITED

Aleev, Yu. G.

1963. Function and gross morphology in fish. Izd. Akad.
Nauk SSSR, Mosc, 245 p. (Translated from Russian by the
Israel Program Sci. Trans!., 1969, 268 p.; available U.S. Dep.
Commer, Natl. Tech. Inf. Serv-., Springfield, Va., as TT
67-51391.)

FISHERY BULLETIN: VOL. 74, NO. 4

Chatwin, B. M.

1959. The relationships between length and weight of
yellowfin tuna {Neothunnus macropterus) and skipjack
tuna (Katstiironus pelamis) from the Eastern Tropical
Pacific Ocean. [In Engl, and Span.] Inter-Am. Trop. Tuna
Comm., Bull. 3:305-352.
GiBBS, R. H., Jr., and B. B. Collette.

1966. Comparative anatomy and systematics of the tunas,
genus Thunnus. U.S. Fish. Wild!. Serv., Fish. Bull.
66:65-130.
Japanese Fisheries Agency.

1975. Report of tuna tagging for 1974. [In Jap.] Pelagic Res.
Sec., Far Seas Fish. Res. Lab., 18 p.
Magnuson, J. J.

1970. Hydrostatic equilibrium of Euthijnnitf; affinis, a

pelagic teleost without a gas bladder. Copeia 1970:56-85.
1973. Comparative study of adaptations for continuous
swimming and hydrostatic equilibrium of scombroid and
.xiphoid fishes. Fish. Bull., U.S. 71:337-356.
YOSHIDA, H. 0.

1968. Pectoral fin length of juvenile albacore. Copeia
1968:625-626.

960

PRODUCTION OF FRY AND ADULTS OF THE 1972 BROOD OF

PINK SALMON, ONCORHYNCHUS GORBUSCHA, FROM GRAVEL

INCUBATORS AND NATURAL SPAWNING AT AUKE CREEK, ALASKA

Jack E. Bailey, Jerome J. Pella, and Sidney G. Taylor*

ABSTRACT

Production of fry and adults of the 1972 brood of pink salmon, Oncorhynchus gorbuscka, at Auke Creek,
Alaska, was compared between a grave! incubator hatchery and natural spawning. Natural production
in the creek above the hatchery weir (estimated from hydraulic sampling) was 73,900 fry (SE: 32,800)
from an estimated initial seeding of 934,065 eggs (SE: 42,81 1) for a survival rate of 0.079 (SE: 0.035). An
estimated total of 579,000 unfed fry (SE: 25,296) were released from the hatchery for a comparable
survival rate of 0.743 (SE: 0.047). Exactly 84,000 of the hatchery fry and 5,500 of the creek fry were
released after being marked by clipping fins. All adults returning to the weir were examined for marks,
and some additional marks were recovered from sport and commercial fishermen; 667 marked hatchery
fish and 74 marked creek fish were recovered. Estimated survival of hatchery fry to returning adult was
only 0.0079 (SE: 0.0003) equal to 0.59 (SE: 0.071) the corresponding estimate of 0.0135 (SE: 0.0016) for
creek fry, which suggests that hatchery fry were inferior to creek fry in the marine environment;
however, hatchery fry emigrated seaward 2 wk earlier than creek fry and may have encountered less
favorable marine conditions. Survival from eggs to returning adult stage was 5.50 times (SE: 2.59)
higher for hatchery fry than for creek frj' because of much greater survival from egg to fry in the
hatchery; the difference is not statistically significant. Hatchery fry were generally shorter but heavier
than creek fry and emigrated seaward at a slightly earlier stage of development. No differences in size
or time of return of adults could be traced to the incubation environment from which they came.

The level of harvest of pink salmon, Oncorhynchus
gorbuscha, in Alaska in the 1970's (Seibel and
Meacham 1975) has been about one-ninth the level
of the 1930's (Kasahara 1963). This decline, in view
of recent advances in salmon hatchery systems
(Bams 1972), might be countered by large-scale
artificial propagation of salmon fry to supplement
natural spawning. As a first step toward develop-
ing systems for enhancing or rehabilitating the
depleted stocks, the National Marine Fisheries
Service, Northwest Fisheries Center Auke Bay
Fisheries Laboratory and the Alaska Department
of Fish and Game agreed in August 1971 to begin
testing a gravel incubator hatchery on Auke Creek
near Juneau in southeastern Alaska.

Auke Creek was selected because it is accessible
and has a fish weir and a dependable water supply
from nearby Auke Lake. Lake water is especially
desirable for hatcheries in Alaska because the
water temperature generally remains above
freezing (3°-4°C). However, lake water has at least
one disadvantage-it has a different seasonal

'Northwest Fisheries Center Auke Bay Fisheries Laboratory,
National Marine Fisheries Service, NOAA, P.O. Box 155, Auke
Bay, AK 99821.

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

temperature pattern than most of the streambed
waters where pink salmon eggs normally incubate.
Bams (1972) avoided the problem of temperature
differences by collecting hatchery water from
beneath the streambed, but this is not always
feasible in Alaska because of the severe freezing
conditions encountered at many potential hatch-
ery sites. This report on the 1972 brood pink
salmon at Auke Creek compares hatchery produc-
tion and natural production in regard to 1) survival
from eggs to emergent fry, fry to returning
adults, and eggs to returning adults; 2) size, stage
of development, and emergence timing of fry; and
3) size and time of return of adults returning to
Auke Creek from hatchery and creek fry.

MATERIALS AND METHODS

A heated building (7.3 by 13.4 m) provided space
for a water filter and ultraviolet purifier; incuba-
tors; instruments for measuring temperature and
oxygen; equipment for censusing, sampling, and
marking fry; and instruments for measuring and
counting adult salmon. The building was located
on Auke Creek near a fish-counting weir at the
head of tide where eggs could be collected from

961

FISHERY BULLETIN: VOL. 74, NO. 4

returning adult salmon. The hatchery water sup-
ply came from nearby Auke Lake. The eggs were
incubated to the eyed stage in Heath'^ incubators
and then transferred to gravel incubators to
complete development.

Water Filter and Purifier

The water filter and ultraviolet purifier system
supplied treated water to one-half of the hatchery
incubators; the rest were supplied with untreated
water. The filter was rated to remove particles 10
jtim in diameter or larger. The purifier was de-
signed to give a minimum dosage of 35,000
iiiW-s/cm^' at 2,537 A. The water treatment had no
apparent beneficial effect.

Natural Spawning

From 4 August to 21 September 1972, 1,768
adult pink salmon entered the fish counting weir.
About 55%, 459 females and 527 males, were
released to spawn above the weir. The rest were
kept for fecundity counts and hatchery spawn
source. Ten females from which we obtained
fecundity counts were treated as a simple random
sample in later analysis, although no serious effort
was made to assure randomness of selection.
Average fecundity in this sample was 2,035
eggs/female (SE: 93.27). This estimate agreed
closely with 2,023 eggs/ female from an inventory
of eggs obtained from the 386 females used as the
hatchery spawn source after a rough correction for
eggs retained. Most pink salmon released above
the weir spawned in a 297-m section of stream
between the weir and Auke Lake. Fewer than 20
adults spawned in Lake creek above Auke Lake.
The alevin population of Auke Creek was es-
timated 20-21 March 1973 with a hydraulic pump
census (McNeil 1964).

Collection and Eyeing of Eggs

Eggs for seeding incubators were obtained from
the Auke Creek pink salmon run 8 August through
22 September 1972. These dates cover nearly the
entire run, thereby assuring representation of all
parts of the run in the next generation. Eggs were
collected from 386 females (about 45% of the
females in the spawning run) in the manner

described by Bailey and Taylor (1974). Malachite
green treatments, 15 ppm. for 1 h, were used at
weekly intervals between 17 August and 19 Oc-
tober to control fungus growth until eyed eggs
were removed from the Heath trays.

Raising Eyed Eggs to Fry Stage

The eyed eggs were raised to the fry stage in
four gravel incubators (Bams 1970) designated A,
B, C, D (Table 1). The incubators measured 1.2 by
1.2 by 1.2 m and used a system of perforated pipes
and horizontal layers of graded gravel to achieve
uniformity of upwelling flow through the eggs and
gravel. Flow to A, B, and C was initially set at 75
liters/min and to D at 79 liters/min. Incubators A,
B, and C were loaded with an estimated 150,000
eggs (SE: 1,030) each and incubator D with an
estimated 158,000 eggs (SE: 1,085) (Table 1).
Therefore each incubator initially contained 2,000
eggs per liter/min.

Iron bacteria sheaths and a flocculent iron
precipitate accumulated in the incubators. The
material seemed to accumulate as rapidly in
incubators receiving filtered and irradiated water
as in those receiving untreated water. The in-
tended water flow through the incubators receiv-
ing treated water could not be maintained. Flow
through incubator C had dropped from the desired
1.26 liters/s to 0.88 liter/s 18 December 1972, and
flow through incubator B had dropped to 0.95
liter/s 3 January 1973. Flow through these in-
cubators was maintained at 0.63-1.07 liters/s for
the rest of the incubation period. The full 1.26
liters/s was maintained at all times in the two
incubators receiving untreated water, probably
because the hydraulic head on the untreated water
supply was about twice the head on the treated
water.

The estimates of numbers of eggs seeded in each
incubator were determined by the method of
Burrows (1951). Through an oversight, records of

Table L-Operating conditions in four gravel incubators seeded
with eyed pink salmon eggs, Auke Creek, 1972. For each
incubator the volume of substrate and eggs was 1.246 m-'.

-Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Number

Water flow

Type of

Incubator

eggs

(lit.

ers/min)

water treatment

150,000

Untreated

150,000

Filtered and ultra-
violet treated

150,000

Filtered and ultra-
violet treated

158,000

Untreated

962

BAILEY ET AL.: PRODUCTION OF FRY AND ADULTS AT AUKE CREEK

o 12

5,0

I-

* 6

JKE CREEK

INCUBATORS

I I I I I I
AUG.

I I I I I I I

5 15 25' 5
SEFT.

15 25
OCT.

I' I 'I I I I I I T I I I I I I I I

5 15 25' 5 15 25 ' 5 15 25 ' 5
NOV. DEC. JAN.

I I r I I 1 I I I I I I I I I I I )
15 25' 5 15 25 ' 5 15 25' 5 15 25

Figure 1. -Temperatures in
gravel incubators and in surface
water of Auke Creel<, 8 August
1972 through 17 May 1973.

FEB.

MARCH

APRIL

the procedure were of insufficient detail to es-
timate the precision of the initial seedings. Vari-
ances of these initial seedings were estimated
from data obtained in recent years, 1974 and 1975.
This source of error was determined to be negligi-
ble in later calculations.

Eggs were fertilized on the following schedule:
incubator A, 4-31 August; B, 4 August to 8 Sep-
tember; C, 11-17 September; and D, 17-21
September.

Water Temperatures

We measured temperatures daily with a mer-
cury thermometer (to the nearest 0.1°C) in Auke
Creek and in the incubators from the time the first
eggs were collected until the fry left the creek.
While eggs were being collected (8 August to 22
September 1972), water in Auke Creek was
warmer than water in the incubators (Figure 1).
The creek water was cooler than the incubator
water from 9 October throughout the rest of the
incubation period, which ended when the fry
emerged.

Oxygen Levels

Oxygen concentrations in the water supply to
the hatchery and in effluents from the incubators
were measured to the nearest 0.01 mg/liter by the
Winkler method. Oxygen was measured at weekly
intervals from shortly after eyed eggs were seeded
(9 November 1972) until the fry began to emerge
(23 March 1973). Oxygen content of the water
supplied to the incubators decreased steadily-
from 9.6 mg/liter (73% saturation) on 22
November 1972 to 7.8 mg/liter (59% saturation) on
23 March 1973 (Figure 2). Oxygen in effluents from
gravel incubators decreased from 9.3 mg/liter
(71% saturation) to 6.7 mg/liter (51% saturation)
during the same period.

>- *

Q 5

O
(/t

5 3

SOURCE WATER

EFFLUENT WATER

I -

ol—

NOV

DEC. I JAN

100

z
o

D
<

FEB

MAR

Figure 2.-Dissolv'ed oxygen levels in source water and effluent
water of gravel incubators at Auke Creek Hatchery, 9 November
1972 through 23 March 1973.

Counting and Processing Fry

We collected emigrating creek fry and hatchery
fry to measure and mark, to determine time of
migration, and to estimate abundance of hatchery
fry. Two 0.91- by 0.91-m fyke nets with floating
live-boxes were used to index the daily emigration
of creek fry and to collect fry for a mark and
recovery experiment. The daily counting of fry as
they emerged from gravel incubators and the
collection of fry for fin clipping and measuring
was expedited by passing the incubator effluents
over a cone-shaped sampling device-^ and then
through a second sampling device consisting of a
set of five parallel troughs. The first device provid-
ed small subsamples of fry from which total
numbers emerging could be estimated; the second

3A blueprint for the cone-shaped fish sampler was supplied by
the Washington Department of Fisheries.

963

FISHERY BULLETIN: VOL. 74, NO. 4

device separated a larger subsample from the total
numbers for marking.

We calibrated the cone device as a sampler with
which to estimate total numbers of emerging fry
from the incubators. Inspection of the relationship
of total fry emigrating from an incubator (y)
plotted against fry retained by the sampler (x) on
24 occasions indicated a constant ratio (straight
line through the origin) with increasing variation
at higher subsampler counts (Figure 3). Con-
sequently, the average of the 24 ratios (y/x)
available from the calibration study is taken as the
slope estimate (Snedecor 1956: 153-156) and was
calculated as 24.537 (SE: 1.072). The major portion
of the fry passed the cone sampler and were then
routed through the parallel troughs, one of which
emptied into a holding tank and four of which
emptied into the hatchery drain and then into
Auke Creek. With these two devices we captured
about one-fourth of the gravel incubator fry each
day without impeding the seward migration of the
other three-fourths.

Twice weekly, samples of 50 fry from each
gravel incubator and the fyke nets were preserved
in 5% Formalin. The preserved fry were allowed to
stand for 6 wk before lengths were measured to
the nearest millimeter and wet weights to the
nearest milligram. An inde.x to stage of develop-
ment (Bams 1970) of the fry was computed from
the formula

^ 10 ^ weight in milligrams

^D — .

length in millimeters

This index is used only on unfed fry to indicate the
relative yolk content. It is not a condition factor.

Weighted means and variances of pooled data
were computed on the basis of the fraction of the
migrant fry represented by each sample. Statis-
tical comparisons were made of lengths, wet
weights, and developmental index as follows:

where Y,,.
W.

= weighted mean

= observed mean measurement in ith

period
= proportionofrunleavingin/th period

from index sampling, and

V{Y,y)= J^ WfViY^)

100 200 300 400

NUMBER OF FRY IN SUBSAMPLER (x)

Figure 3.- Relation of number of fry migrating to number of fry
in subsampler. Each point represents one sample.

where F(r„)
V{Y,)

n =

variance estimate of weighted

mean

sample variance of estimated

mean in ('th period

number of periods sampled.

i = 1

Technicians marked hatchery fry by clipping
the adipose and left ventral fins and creek fry by
clipping the adipose and right ventral fins. Before
marking, the fry were anesthetized in a solution of
1:7,500 MS-222 (Tricaine methanesulfonate)
buffered with sodium bicarbonate to pH 6.1-6.4;
the solution was kept cool in a water table and
recirculated to keep the oxygen content high.
Surgical iris scissors were used to excise fins under
a 3x magnifying lens. Technicians marked an
average of about 200 fry/h on this study, whereas
technicians marked about 400 fry/h on a similar
study in Canada (R. A. Bams pers. commun.).
Samples of fry from each technician were ex-
amined several times daily to ensure that the
correct fins were excised as close to the body as
possible. All marked fry were released at 1130 h
the same day they were marked; most of the
unmarked fry that had left the incubator or the
stream at the same time had migrated seaward 24
h earlier. Dead fry remaining in the release tank
were counted each morning. The immediate mor-
tality from marking was less than 0.1% for both
hatchery and creek fry. Totals of marked fry
released were 84,000 from the hatchery and 5,500

964

BAILEY ET AL.: PRODUCTION OF FRY AND ADULTS AT AUKE CREEK

from the creek. The daily numbers of marked fry
released from the hatchery and the creek were
roughly proportional to the respective migrations
of fry from these two sources. There was a slight
bias toward marking too few fry during the first
half of the migration, but the bias was in the same
direction and magnitude on both types of fry.

Less than 1% of the creek fry died in the fyke net
and floating live-box, indicating slightly greater
physical abuse for marked creek fry than marked
hatchery fry.

Recovery of Marked Adults

Returning 1972 brood adults were counted at the
weir in Auke Creek in the summer of 1974; some
adult salmon were anesthetized and measured.
Mideye-to-tail-fork lengths were measured to the
nearest millimeter and weights to the nearest
0.01 kg.

Analysis of Survival

Survival probabilities from egg to fry and fry to
returning adult are estimated from estimates of
initial number of eggs, fry produced, and return-
ing adults. Ratios of these survival estimates are
used to compare survival of hatchery and creek
salmon. Variances of survival and estimates of
ratios of these survival estimates were approx-
imated by the delta method (Deming 1943; Paulik
and Robson 1969). Finite population correction
factors were ignored in variance calculations;
changes in variance estimates would have been
insignificant.

Estimation of survival from marking requires
special argument. The expected total unmarked
returns from hatchery and creek fry combined is

T = Us + U's'

where U and U' are initial numbers of unmarked
fry from the creek and hatchery respectively, and s
and s' are the probabilities of survival of the two
unmarked groups at sea. Marking increases mor-
tality. If the probability of survival from marking
is T and identical for both groups, the probabilities
of survival from both causes are st and s't for creek
and hatchery fry respectively. The expected total
return of the unmarked fry had they been marked,
r, is

r = UsT + U's't.

The ratio of T to T is t. Therefore, we estimate
survival from marking from estimates of Tand T"
as

f = f'/f.

The expectation T is estimated by the total un-
marked recoveries to the weir. The expectation T
is estimated from appropriate combinations of
estimates of numbers of unmarked creek and
hatchery fry and estimates of marine survival of
marked fry of both groups.

Total variation among incubators in estimated
survival from egg to fry is divided into three
sources: 1) Underlying variation due to hetero-
geneity of genetic composition of pink salmon and
environmental conditions among incubators, 2)
binomial variation within incubators, and 3) sam-
pling error in estimation of numbers of eggs and
fry. We imagine an unobserved universe of sur-
vival probabilities s with mean s has been sampled
randomly by our study; four members were drawn,
each applying to one of our incubators. Actual
survival within an incubator varies from its as-
sociated probability of survival due to binomial
variation; instead of a fraction s surviving, the
actual fraction is s. This actual rate was not
observed; rather, we estimated s by s, the ratio of
estimated fry to estimated eggs.

Total variance of estimated survival among
incubators, a^, is defined by

af=f(i-s)2

where E denotes the expectation operation over
the three sources of variation.
This expression may be rewritten as

o2 = Eli's -s) + (s-s) + is- s)f.

After completing the square and evaluating the
expectations of the terms, we find

of = a? + ai + a§

where al = E (s - sf, the variance of underlying
survival probabilities among incubators; <^2 = -^
(s - sf, the average binomial variance; <J3=E
(s - sf, the average variance due to errors in es-
timates of fry and eggs ; £" denotes the expectation
operation over the first source of variation; and^
denotes the expectaton operation over the first
two sources of variation.

965

FISHERY BULLETIN: VOL. 74, NO. 4

Our experiment provides four unobserved
selections from the underlying probabilities of
survival, four estimates of binomial variation, and
four estimates of variance in estimated survival
due to errors in estimates of eggs and fry. Aver-
ages of the four estimates for the second and third
sources are used to estimate oj and aj. The sample
variance of s is used to estimate of. The estimate of
of is obtained by subtraction.

RESULTS

First, we evaluate the effectiveness of the Auke
Creek Hatchery by comparing survival of hatchery
and creek fishes at different life stages: egg to
emergent fry, fry to returning adult, and egg to
returning adult. Next we estimate survival from
marking. Then we compare size, stage of develop-
ment, and emergence timing of hatchery fry with
creek fry. Finally, we compare size and time of
return of hatchery fish and creek fish as adults.

Survival from Egg to Fry

We estimated survival from potential egg
deposition to fry for creek fry as the ratio of an
estimate of the alevins in the spawning area of the
creek above the weir in the spring of 1973 (just
before emergence) to an estimate of the potential
egg deposition. Because 459 females spawned in
the stream above the weir, we estimate potential
egg deposition as (459)(2,035) = 934,065 [SE:
(459)(93.27) = 42,811 eggs].

On 20 and 21 March 1973, we determined the
number of live alevins in each of 86 O.l-m^ units of
a simple random sample from the 8,600 such units
making up the spawning area above the weir. The
average number of alevins per unit was 8.593 (SE:
3.814). Hence, total live alevins in the spawning
area was estimated to be (8,600)(8.593) = 73,900
alevins [SE: (8,600)(3.814) = 32,800 alevins].

Survival to time of sampling is estimated as the
ratio of estimated total alevins to estimated
potential egg deposition or 73,900/934,065 = 0.079
(SE: 0.035).

In the gravel incubators, estimated survival
from live egg to fry was calculated as the ratio of
estimated total emigrants to initial numbers of
live eggs. The sums of the daily numbers of fry in
subsamples from the four incubators were as
follows: (A) 5,960; (B) 5,792; (C) 5,153; and (D) 6,692.
Total emigrations from the incubators and corre-
sponding standard errors were estimated using

966

the calibration results: (A) (5,960)(24.537) =
146,240 fry [SE: (5,960)( 1.072) = 6,389 fry]; (B)
142,118 (SE: 6,209); (C) 126,439 (SE: 5,524); and (D)
164,202 (SE: 7,174). The grand total of fry emigrat-
ing was 579,000 (SE: 25,296).

Estimates of survival from live eyed eggs to fry
and the standard errors of these estimates were as
follows: (A) 146,240/150,000 = 0.975 (SE: 0.043);
(B) 0.947 (SE: 0.041); (C) 0.843 (SE: 0.037); and (D)
1.039 (SE: 0.045). The estimate for incubator D is
not feasible, but since it lies within a standard
error of the feasible range, we do not suspect
errors in data recording or calculations. The mean
of the survival estmates was 0.951, and the sample
variance of the estimates was 0.00667.

This variance estimate is divided into three
components-af, 65, a§-representing variation in
underlying survival probabilities, binomial varia-
tion, and variation due to errors in estimating
eggs and fry respectively. The estimates are as
follows: a\ = 0.00493, oi is negligible, and
03 = 0.00174. Therefore, most of the total variance
of survival estimates among the four incubators
seems due to variation in underlying survival
within the incubators rather than binomial varia-
tion or variation in egg or fry counts.

The incubator survival rates are from live eyed
egg to fry. The creek survival rate is from poten-
tial egg deposition to fry. To make the survival
rates comparable, we adjust the incubator survival
to that from potential egg deposition to fry. The
proportion of potential egg deposition which
develops to the eyed stage in the hatchery is
estimated as the ratio of total estimated eyed eggs
obtained from the 386 females artificially spawned
to estimated potential egg deposition by the
females, or 614,000/785,510 = 0.782 (SE: 0.036).
The adjusted incubator survival rate from poten-
tial egg deposition to fry is (0.782)(0.951) = 0.743
(SE: 0.047).

Survival from Fry to Returning Adult

Although most of the marked returning adults
were recovered at the weir in Auke Creek below
their point of origin, some were recovered from
sport and commercial fishermen and from the
intertidal spawning area of Auke Creek (Table 2).
Total recoveries from all sources were used to
estimate relative survival from fry to returning
adult: 667 of the marked hatchery fish and 74 of the
marked creek fish were recovered. Estimated
survival of hatchery fry to returning adults is

BAILEY ET AL.: PRODUCTION OF FRY AND ADULTS AT AUKE CREEK

667/84,000 = 0.0079 (SE: 0.0003). Estimated sur-
vival of creek fry for the same period is
74/5,500 - 0.0135 (SE: 0.0016). Therefore, our
estimate of relative survival of hatchery fish as
compared to creek fish is 0.0079/0.0135 = 0.59 (SE:
0.071).

although the precision of that estimate is ex-
tremely low, as indicated by the standard error-a
rough 95% confidence interval would include the
possibility that survival from potential egg depo-
sition to adult was smaller for hatchery operations
than for natural spawning.

Table 2.-Source of recoveries ofmarked pink salmon adults
originating from fry marked at Auke Creek in 1973.

gin

mat

Source of recovery

Hatchery

Creek

Commercial fishery

Sport fishery

Intertidal area of Auke Creek

Auke Creek weir

Total

644

667

Survival from Egg to Returning Adult

While hatchery fry suffered greater losses than
creek fry in the marine environment, their in-
creased survival under the artificial conditions
during incubation was compensating. Overall
relative survival from potential egg deposition to
returning adult can be estimated as the ratio of
the products of survival from potential egg depo-
sition to alevin and from fry to returning adult for
hatchery and creek fry. The survival of hatchery
fish relative to creek fish is

(0.743)(0.0079)/(0.079)(0.0135) = 5.50 (SE: 2.59).

Production of adults by the hatchery is estimated
to be 5 to 6 times that of natural production,

Survival from Marking Effects

Estimates of the initial numbers of unmarked
creek and hatchery fry are 68,400 and 495,000,
respectively. Unmarked recoveries to the weir
totaled 5,545. Survival of marked fry to return at
the weir is estimated by the ratios of marked
recoveries at the weir (Table 2) to numbers of
marked fry released, or 71/5,500 = 0.01291 for
creek fry and 644/84,000 = 0.00767 for hatchery
fry. Then survival from marking is estimated to be

[(68,400)(0.01291) + (495,000)
(0.00767)]/5,545 = 0.84.

Determination of the precision of the marking
mortality estimate was not attempted because of
the apparent complexity of the problem.

Fry Size and Developmental Index

Most of the fry from gravel incubators were
shorter (Figure 4) but heavier (Figure 5) than
creek fry, although there were two exceptions: fry
from incubator A had an average weight of 260.0
mg, which was not significantly different from the
average weight for creek fry-260.2 mg (Table 3);

CREEK

INCUBATOR A
fNCU BATOR B
INCUBATOR D
INCUBATOR C

31.1 31.5 31.6 31 7 31.8

CREEK

INCUBATOR A
INCUBATOR B
INCUBATOR C
INCUBATOR D

24S

31.9 32.0 32.1 32.2
LENGTH OF FRY (MM)

32.3 32.1 32.5 32.6 32.7

Figure 4. -Weighted means and 95%
confidence intervals for these means of
lengths of preserved fry from Auke Creek
and four gravel incubators.

250

260 265

WEIGHT OF FRY (MG)

270

275

Figure 5. -Weighted means and 95%
confidence intervals for these means of
weights of preserved fry from Auke
Creek and four gravel incubators.

967

FISHERY BULLETIN: VOL. 74, NO. 4

Table 3.-Pooled means and variances of means for lengths, weights, and development
index, AT,,, of pink salmon fry (50 fry /sample) at Auke Creek in spring of 1973.

Source

Creek
Incubator:

Number of
samples

Length (mm)
Mean Variance

Weight (mg)
Mean Variance

/<P index

Mean

8
8
4
5

32.45

31.57
32.17
32.21
32.29

0.00272

0.00252
0.00276
0.00412
0.00483

260.2

1.630

260.0 1.917

269 9 1.856

273.2 3.352

268.6 2.714

1.964

2.008
2.009
2.012
1.987

Variance

5.36 X 10-4

5.50 X 10-4

5.54 X 10-4

9.61 X 10-4

9.92 X 10-4

CREEK

INCUBATOR A
INCUBATOR B
INCUBATOR C
INCUBATOR D

Figure 6. -Weighted means and 95%
confidence intervals of these means of
indices of development, Kd, of preserved
fry from Auke Creek and four gravel
incubators.

and fry from incubator D had an average length of
32.29 mm, which was not significantly different
from the average length for creek fry-32.45 mm.
Indices of development were higher for fry from
all the gravel incubators than for creek fry (Figure
6). The mean indices of development for gravel
incubator fry ranged from 1.987 to 2.012, whereas
the mean for creek fry was only 1.964 (Table 3). In
an earlier test (Bailey and Taylor 1974) the aver-
age Kp index decreased about 0.005 unit/day in
the final stages of alevin development. Since the
average /Q> index for incubator fry was 0.016 unit
higher than the index for creek fry, incubator fry
apparently emerged about 3 days earlier in their
development.

Time of Emergence and Seaward Migration

Fry of the 1972 brood from the gravel incubators
migrated voluntarily between 15 March and 23
May 1973; the median date was 14 April (Figure 7).
Creek fry emigrated between 16 March and 15
May; the median date was 27 April (Figure 7).

Size of Returning Adults

Length measurements of adults from the 1972
brood that returned to the weir in 1974 are
classified by sex, origin (whether creek or hatch-
ery), and time of return (either early or late run).
Mean lengths and sample sizes (Table 4) were used
as basic observations with which to perform an
analysis of variance (Scheffe 1959: 362-363) to
search for differences in size among the clas-

100

z
o

I 80

H
O
\- 60

O SO

t-
z

O 40

I-

D
U

HATCHERY FRY

CREEK FRY

/ I I I jj , MIDDATE OF

EMERGENCE

15 20 25
MARCH

T — I — I — I — I — I — I — I — I — I — r-

5 10 IS 20 2S ' 5 10 IS 20 2S 30

APRIL

MAY

Figure 7.-Daily cumulative percentage of pink salmon fry
migrations of creek fry and hatchery fry from Auke Creek in
1973; solid lines represent fin-marked fry and dashed lines
represent total fry in the respective migrations.

sifications. Analyses were performed separately
for each sex because underlying variances of
hatchery fish differed significantly between sexes.
Spawning males typically vary more in length
than spawning females. The corresponding tests
for creek fish did not indicate inequality of vari-

968

BAILEY ET AL.: PRODUCTION OF FRY AND ADULTS AT AUKE CREEK

Table 4.-Average lengths of adult pink salmon returning to
Auke Creek weir, early and late runs; the figures in parentheses
represent the number of fish in the samples.

Average lengths (mm)

Early run

Late run

Mark

Origin

Male Female

Unmarked

Hatchery

493.1

495.2

512.4

500.8

and creek

(117)

(92)

(58)

(137)

Ad-LV

Hatchery

500.2

499.6

510.0

495.6

(126)

(70)

(44)

Ad-RV2

Creek

505.6

515.9

515.7

497.3

(19)

(7)

(3)

'Ad-LV = adipose and left ventral fins.
^Ad-RV = adipose and right ventral fins.

ances, probably because of the small sample sizes.
Differences in length due to origin, time of
return, or interaction were not detectable at the
95% level of testing for either sex (Table 5). Only
time of return for females approached statistical
significance (the test would have been significant
at the 90% level). Mean lengths of samples of creek
fish exceeded those of hatchery fish in all cases
(Table 4). While our data suggest that creek fish
were larger than hatchery fish upon return, the
observed differences could be due to chance when
samples were drawn. Larger samples would have
been needed to resolve the issue.

Table 5.-Analysis of variance of size of returning adult male
and female pink salmon classified by origin (creek or hatchery)
and time of return (early or late).

Source

Deg

rees

of freedom

ean square

Males:

Origin, A

30.8025

Timing, B

99.0025

0.0225

Error

188

161.543

—

Females:

Origin, A

81.000

2.09

Timing, B

127.690

3.30

53.290

1.38

Error

120

38.739

—

Timing of Adult Return

Marked hatchery fish entered the weir between
6 August and 25 September and marked creek fish
entered between 16 August and 20 September
(Figure 8). For 644 marked hatchery fish the
median date of return was 13 September 1974; for
71 marked creek fish the median date was 10
September.

DISCUSSION

Gravel incubation of eggs and release of unfed

T — I — I — r

10 15 20 25
AUG.

T — I — I — r

10 15 20 25

SEPT.

Figure 8.-Daily cumulative percentage recovery of marked
adult pink salmon at Auke Creek weir, 1974.

fry increased the survival from potential egg
deposition to returning adult an estimated 5 to 6
times over natural spawning for 1972 brood year
pink salmon at Auke Creek. The estimate lacks
precision, however, and a rough 95% confidence
statement includes the possibility that egg-to-re-
turning-adult survival was less for incubator fry
than for naturally produced fry. Further, the
estimate of relative survival is potentially biased
unless marine mortality due to marking and
fishing was similar for both groups of marked fry.
The similarity of timing of adult returns from
both groups gives no reason to suspect differential
fishing mortality. The low mortality of creek fry in
the fyke net and live-box suggests only slightly
greater physical abuse occurred to marked creek
fry than hatchery fry. The difference in survival
from potential egg deposition to returning adult,
if real, was accomplished in spite of certain
deficiences in the quality of environment provided
for eggs and alevins in the hatchery and in spite of

969

FISHERY BULLETIN; VOL. 74, NO. 4

a lower ocean survival for hatchery fry than for
creek fry.

There is a notable difference in survival from
marking in the tests at Auke Creek and the tests
by Bams (1972, 1974) at Headquarters Creek,
Vancouver Island. The estimate of survival from
marking at Auke Creek, 84%, is much greater than
the 17% and 36% survival we estimated from
Bams' data on Headquarters Creek. Intertidal
alevin production in Auke Creek below the weir
was estimated by hydraulic pump survey to be 16%
of that above the weir. Possible straying of these
intertidal fish above the weir upon return would
only bias our estimate below actual survival from
marking. The slower rate at which our technicians
clipped fins may be the cause of better survival
from marking at Auke Creek.

Our estimates of fry releases and survivals
imply that an increase in numbers of returning
spawners at Auke Creek in 1974 was largely due to
operation of the hatchery. If this is true, then
hatcheries can be built on lake-water sources with
a reasonable expectation of successfully enhanc-
ing salmon numbers. Projections of our data must
be considered tentative because of the lack of
precision. However, the magnitude of the Auke
Creek escapement in relation to escapements to
other streams in northern southeastern Alaska
supports our conclusion that operation of the Auke
Creek Hatchery did in fact enhance the return of
adult salmon. For example, marked hatchery fry
had a recovery rate of 0.767%. Survival from
marking was 84%. The release of 579,000 hatchery
fry would project to (579,000)(0.00767)/0.84 =
5,287 adults. The projected return of creek fry
would be (84,000)(0.01291)/0.84 = 1,291 adults. In
1974, 6,260 adults returned to the Auke Creek weir
from a parent escapement of 1,768 adults. This
3.5-fold increase occurred in the face of a general
scarcity of pink salmon in this part of Alaska.
According to Kingsbury (1975) the lowest es-
capement for pink salmon streams of northern
southeastern Alaska since 1960 occurred in 1974.

The yolk content of fry when they leave the
incubating bed, either natural or artificial, bears
directly on the survival of the fry in the wild. Fry
with a large amount of yolk have not attained
their maximum potential size, are relatively poor
swimmers, may not be able to osmoregulate in
seawater, and are more vulnerable to predators.
On the other hand, fry that have little or no yolk
are losing weight and soon become weakened and
emaciated and again are more vulnerable to

970

predators. Naturally produced fry emerge voli-
tionally from the stream gravel, presumably at the
stage of development that ensures maximum
survival. Our analysis of the developmental index
showed our gravel incubator fry emerged prema-
turely in comparison to creek fry.

Earlier (in the temporal sense) emergence of fry
produced in gravel incubators at Auke Creek also
suggests that the Auke Creek Hatchery environ-
ment was inferior to the natural streambed envi-
ronment. Hatchery fry emerged and migrated
seaward 2 wk earlier than creek fry. This could
place them in the estuary before the spring bloom
of zooplankton on which they feed and before
spring warming of estuarine surface water. The
resulting slow growth rate could mean an exces-
sively long period of high vulnerability to preda-
tors. Experiments by others (Levanidov 1964;
Bams 1967; Kanid'yev et al. 1970; Parker 1971)
show that small juvenile salmon suffer higher
mortality from predation than large juvenile
salmon.

The earlier time of migration and size of hat-
chery fry at Auke Creek were probably caused by
one or more of the following: the higher average
winter temperature of Auke Lake water (4°C) as
compared to the temperature in natural redds in
Auke Creek (0°-2°C); the low oxygen content of
60-70% saturation in lake water supplied to in-
cubators; and the brown organic material from
iron bacteria which accumulated in the gravel
incubators and impeded the flow of water.

ACKNOWLEDGMENTS

We thank Joyce Gnagy, biologist at the Auke
Bay Fisheries Laboratory, for identifying the
organic growth in the gravel incubators. We also
thank personnel of the Alaska Department of Fish
and Game who reported the recovery of marked
fish in the sport and commercial fisheries.

LITERATURE CITED

Bailey, J. E., and S. G. Taylor.

1974. Salmon fry production in a gravel incubator hatchery,
Auke Creek, Alaska, 1971-72. U.S. Dep. Commer., NOAA
Tech. Memo. NMFS ABFL-3, 13 p.
Bams, R. A.

1967. Differences in performance of naturally and
artificially propagated sockeye salmon migrant fry, as
measured with swimming and predation tests. J. Fish.
Res. Board Can. 24:1117-1153.

BAILEY ET AL.: PRODUCTION OF FRY AND ADULTS AT AUKE CREEK

1970. Evaluation of a revised hatchery method tested on
pink and chum salmon fry. J. Fish. Res. Board Can.
27:1429-1452.

1972. A quantitative evaluation of survival to the adult
stage and other characteristics of pink salmon
{Oncorhynchns gorhuscha) produced by a revised hatchery
method which simulates optimal natural conditions. J.
Fish. Res. Board Can. 29:1151-1167.

1974. Gravel incubators: a second evaluation on pink salmon,
Oncorhynch us gorhuscha, including adult returns. J. Fish.
Res. Board Can. 31:1379-1385.

Burrows, R. E.

1951. Method for enumeration of salmon and trout eggs by
displacement. Prog. Fish-Cult. 13:25-30.
Deming, W. E.

1943. Statistical adjustment of data. John Wiley & Sons,
Inc.,N.Y.,261p.
Kanid'yev, a. N., G. M. Kostyunin, and S. A. Salmin.

1970. Hatchery propagation of the pink and chum salmons
as a means of increasing the salmon stocks of Sakhalin. J.
Ichthyol. 10:249-259.
Kasahara, H.

1963. Catch statistics for North Pacific salmon. Int. North
Pac. Fish. Comm., Bull. 12:7-82.
Kingsbury, A. P.

1975. Salmon strategy, tight management for southeastern
pinks. Alaska Fish Tales and Game Trails 9(6):1, 14-15.

Levanidov, V. Ya.

1964. 0 zavisimosti mezhdu razmerami mal'kov Amurskoi
osennei kety Oncorhi/nchus keta infrasp. aiifumnalis
Berg i ikh vyzhivaemost'yu. Vopr. Ikhtiol. 4:658-663.

McNeil, W.J.

1964. A method of measuring mortality of pink .salmon eggs
and larvae. U.S. Fish Wildl. Serv., Fish. Bull. 63:575-588.

Parker, R. R.

1971. Size selective predation among juvenile salmonid
fishes in a British Columbia inlet. J. Fish. Res. Board Can.
28:1503-1510.

Paulik, G. J., and D. S. Robson.

1969. Statistical calculations for change-in-ratio estimators
of population parameters. J. Wildl. Manage. 33:1-27.

SCHEFFfe, H.

1959. The analysis of variance. John Wiley & Sons, Inc.,
N.Y.,477p.
Seibel, M. C, and C. P. Meacham (editors).

1975. A summary of preliminary 1975 forecasts for Alaskan
salmon fisheries. Alaska Dep. Fish Game, Inform. Leafl.
167, 55 p.
Snedecor, G. W.

1956. Statistical methods applied to experiments in agricul-
ture and biology. 5th ed. Iowa State Coll. Press, Ames,
534 p.

971

COMPARISON OF THE MOST SUCCESSFUL AND LEAST
SUCCESSFUL WEST COAST ALBACORE TROLL FISHERMEN

Donald F. Keene^ and William G. Pearcy^

ABSTRACT

Catch data for albacore troll boats were collected from fishermen's logbooks and from dockside
interviews during the 1968, 1969, and 1970 seasons. Fishing powers of these boats were calculated and
used to determine the 10 most successful and 10 least successful fishermen (highliners and lowliners,
respectively) who fished off Oregon and Washington. Characteristics of these two groups of fishermen
were then compared. In general, highliners had longer boats and fished nearer the fleet center and along
the offshore margin of the fleet. Lowliners tended to have smaller boats and fished along the trailing
(south) inshore margin of the fleet. Both groups responded to changes in apparent albacore abundance
by aggregating on days of high apparent abundance, although this response was less pronounced in
1969 and 1970. Highliners caught significantly smaller (but more) fish than the lowliners.

The west coast albacore troll-boat fleet consists of
many types and sizes of vessels (Clemens 1955).
Troll boats range in length from about 10.7 m (35
feet) to over 22.9 m (75 feet) with a displacement
of about 15 tons. Part of this fleet begins fishing
for albacore off the coast of Baja California in
early summer. During the peak of the season (July,
August, September) boats may be found from
Mexico to the Gulf of Alaska. However, the most
productive area usually lies between central Baja
California and the Columbia River (Clemens 1961).
Many boats, particularly those from Oregon and
Washington, fish for other species (salmon, crab,
shrimp) during part of the year (Roberts 1972) and
occasionally during the albacore season when
albacore fishing is slow.

Fishermen in the albacore fleet exhibit a large
range of fishing success. Fishing success has been
related to strictly physical parameters of the
vessel, such as boat length (Fox^). Abramson
(1963) suggested that fishing success is related to
the skill and experience of the captain and crew, as
well as the physical parameters of the boat. Little
is known, however, about how fishing success is
related to the activities of individual albacore
fisherman and the activities of the surrounding
fleet. (The fleet is considered to be an assemblage

'School of Oceanography, Oregon State University, Corvallis,
Oreg.; present address: Bureau of Land Management, Pacific
OCS Office, 300 North Los Angeles Street, Los Angeles, C A 90012.

^School of Oceanography, Oregon State University, Corvallis,
OR 97331.

■Tox, W. W. "Fishing power of U.S. vessels participating in the
Pacific coast albacore fishery 1961-1970." Paper presented at the
24th Tuna Conference, Lake Arrowhead, Calif., Oct. 1973.

of fishing boats within an area of arbitrarily chosen
size.) The objective of this paper is to describe and
compare the characteristics and movements of the
most successful with those of the least successful
albacore fishermen during the 1968, 1969, and 1970
seasons.

METHODS
Sources and Treatment of Data

Information on number of fish caught per day
by troll boats, location of the catch, boat length,
and number of lines (1970 only) was collected from
three sources for the 1968, 1969, and 1970 albacore
seasons: 1) logbooks distributed by Oregon State
University (1969 and 1970), 2) logbooks distributed
by California Department of Fish and Game to
fishermen who volunteered to submit daily infor-
mation, and 3) interviews obtained bypersonnel of
the Oregon Fish Commission at dockside during
unloading of the albacore. Careful screening
avoided duplication of logbook records since ves-
sels often submitted records to more than one
source. Only catch locations between lat. 42° and
49°N were used.

The number of reporting boats varied consider-
ably between years. In 1968, 205 boats reported
their daily catches and locations. In 1969 and 1970,
70 and 113 boats, respectively, reported. The total
number of boats fishing during the 3 yr is un-
known but is estimated to have been between 750
(Panshin 1971) and 1,000.

Data from the logbooks and interview sheets

Manuscript accepted May 1976.

FISHERY BULLETIN: VOL. 74, NO. 4, 1976.

973

FISHERY BULLETIN: VOL. 74, NO. 4

were punched on computer cards. Each card con-
tained three pieces of information: the boat
number, an area-data code (signifying the 1°
latitude by 1° longitude rectangle and the calen-
dar day), and the boat's catch of the day. There
were approximately 3,300 observations in 1968,
1,500 in 1969, and 1,000 in 1970.

A particular boat was chosen to represent the
standard unit of effort. Criteria for the standard
boat choice included the following: it fished 1)
during all three seasons; 2) in area-date strata
concurrently with a majority of the fleet; 3) most
of each season; and 4) consistently to provide a
standard, nonvarying reference for the other
boats.

Estimates of fishing power"* of all boats in the
fleet were initially determined relative to the
standard boat. This was accomplished using a
computer program called FPOW (Berude and
Abramson 1972). FPOW utilizes Robson's (1966)
linear two-factor analysis model for estimating
the relative fishing power of fishing vessels. The
estimates of fishing power derived from the model
are logrithms. FPOW provides an approximate
correction for this bias using a Taylor series
expansion of the estimate about its true value. The
method and assumptions used in FPOW are de-
scribed in Robson (1966) and Abramson and Tom-
linson (1972:1022-1023). The program's storage
capacity was limited to 2,000 catch observations
from a combined total of not more than 200
distinct boats and area-date strata. Data for each
year were broken up into time segments short
enough to satisfy this limitation. Ten segments
were required in 1968, five in 1969, and three in
1970. Each segment was run independently and
provided estimates of each boat's relative fishing
power during the time segment.

Considerable within-season variation occurred
in the average fishing power of the fleet (Table 1),
suggesting that the standard boat fished inconsis-
tently relative to the fleet. An examination of the
logbooks showed that the standard boat occasion-
ally experienced periods of very low catches (10 to

^Fishing power is defined (Beverton and Holt 1957:172) as the
ratio of the catch per unit of fishing time of a particular vessel to
that of another vessel designated as the standard. It is assumed
that both boats must have fished on the same density of fish
during the same time interval and within the same fishing area
when the ratio is determined. Fishing success, on the other hand,
is related to fishing power but is more descriptive. It includes
parameters difficult to quantify. For example, fishing success
may include crew motivation, attitude, and access to useful
information. Together with fishing power, these parameters are
determinants of fishing success.

15 fish per day) while the majority of the fleet in
the immediate area was catching 100 to 200 fish
per boat. This was particularly obvious during
segment 1 of the 1969 season.

As a result of the standard boat's inconsistent
fishing, values of standardized catch per boat day
were also inconsistent between data segments.
For example, an average boat had fishing powers
of 3.70 and 1.01 on 25 July and 26 July 1969,
respectively (Table 1). If the average boat caught
100 fish on 25 July and 100 on 26 July 1969, values of
standardized catch per boat day (100 fish/average
fishing power) would be 27 and 99, respectively, for
these 2 days. Therefore a serial examination of
apparent abundance could not be performed
without normalizing fishing power estimates of
each boat in each data segment.

Fishing power estimates were normalized by
subtracting the appropriate segment's average
fishing power from each boat's fishing power and
adding unity. (By definition the standard unit of
eff'ort is 1.0.) Each boat's fishing power estimate
was now relative to the average fishing power of
all boats fishing during the data segment. This
procedure required the assumption that the fleet
fished consistently relative to the standard boat
throughout each season.

Daily standardized catch per boat within each
area-date stratum was determined by summing
the fish catches and dividing by the summation of
fishing power in that area-data stratum. The
standardized catch per boat day is an index of

Table l.-Data segments for the 1968, 1969, and 1970 albacore

seasons.

No.

No. of

Average

area-

fishing

Segment

Dates

obs.

boats

dates

power

1968:

6-16 July

242

0.69

17-21 July

320

1.14

21-31 July

410

0.99

1-4 Aug.

357

109

0.91

5-7 Aug.

290

108

0.70

8-11 Aug.

310

100

0.88

12-18 Aug.

420

0.82

19-24 Aug.

373

1.03

25-30 Aug.

235

0.53

31 Aug.-IO Sept.

385

113

0.99

1969;

15-25 July

305

3.70

26 July-3 Aug.

374

1.01

4-11 Aug.

326

1.15

12-18 Aug.

212

1.47

19 Aug. -11 Sept.

296

111

1.16

1970:

15-22 July

160

0.35

23-28 July

470

0.91

29 July-2 Sept.

262

0.67

974

KEENE and PEARCY: COMPARISON OF SUCCESS OF ALBACORE TROLL FISHERMEN

apparent abundance, the latter being a function of
the accessibility of the albacore to the boats, the
vulnerability of the fish to the lures (Marr 1951),
and the true abundance of albacore.

The 10 most successful and 10 least successful
fishermen (highliners and lowliners, respectively)
of each season were selected according to their
boats' average fishing power estimates through-
out the entire season. Highliners and lowliners
selected had fished for at least 15 days in 1968 and
1969 and 8 days in 1970. Thus fishermen who fished
exceptionally well or poorly for only a few days in
a season were not considered.

Area-Date Stratum of Apparent Abundance

Small-scale time and space information of
catches and boat positions allowed a departure
from the traditional time-area stratum of 1 mo
and 1° latitude-longitude rectangle (Ayers and
Meehan 1963; Clemens and Craig 1965). A mobile
stratum was conceived to allow comparisons of
apparent abundance and effort regardless of
where the fleet moved, and without the problems
of fixed geographic boundaries.

The new stratum was a circular area, the center
being the daily medial location of the fleet. This
medial point was determined such that the fleet
was equally divided in the north-south and east-
west planes. Criteria for the radius of the circular
area were that it should be 1) as small as possible to
include a homogeneous distribution of fish, but 2)
large enough to accommodate a sufficient number
of boats fishing on a given day so that catch and
effort could be reliably estimated, and 3) large
enough to give reasonable assurance that boats
within the area remained in the area the entire
day. Because of the lack of knowledge of small-
scale albacore distributions, there was little basis
for satisfying the first criterion.

Consecutively larger concentric circles were
drawn around the medial point while noting the
ratio of boats within each circle to the number of
boats in the entire fleet. (Danils (1952) has pre-
sented theoretical considerations of sample point
distributions within such circles.) During much of
each season, over half the boats could be found
within 25 miles of the fleet's center. Exceptions
occurred in each season when the fleet was highly
dispersed or split into two distinct groups. Two
distinct groups of boats occurred on 2, 3, and 4
August 1968 and also 1, 2, and 8 August 1969.
During these days the northernmost center was

chosen to represent the fleet center because it
always contained more boats.

The third criterion suggested a radius of at least
31 miles to insure that vessels remained within the
area the entire day. This radius was determined on
the basis of distances traveled daily by albacore
boats. (This is reported later in this study.) A circle
with a radius of 31 miles was therefore used as the
area size. Figure 1 shows the percentage of boats
that provided catch data within 31 miles of the
fleet center each day during the 1968, 1969, and
1970 seasons. Only the time periods within the
vertical lines in Figure 1 will be considered for this
study. On days outside these periods few boats
reported their catch, or the fleet was small and
highly dispersed. The average daily percentage of
those boats reporting within 31 miles of the fleet
center was 46%, 57%, and 65% for the 1968, 1969,
and 1970 seasons, respectively. The differences
between the 1968 average and the 1969 and 1970
averages were highly significant (f-test, P<0.01),
indicating that the 1968 fleet was more dispersed
in general than the 1969 and 1970 fleets. (This was
not a result of a greater number of boats reporting
in 1968 because the number of boats reporting per
day was often greater in 1969 and 1970 than in
1968.) There was a tendency in both 1968 and 1969

100%

50%

100%

1969

15 19 23 27 31 4 8 12 16 20 24 28 32
JUL. AUG.

Figure l.-Daily percentage of boats within 31 miles of the
albacore fleet center; 1968, 1969, and 1970. Vertical lines on plots
indicate the time periods considered in detail in this study.

975

FISHERY BULLETIN: VOL. 74. NO. 4

for the fleet to become more dispersed as the
season progressed.

Aggregation of the Boats

The index of aggregation used in this study was
the mean separation distance of boats within a
specified area. The index was determined by sum-
ming separation distances between all boats in the
area and dividing this sum by the number of
separation distances. This calculation required
converting LORAN coordinates (given as the 2100
h PDT boat positions) to latitude-longitude coor-
dinates. Accuracy of the iterative technique used
to compute the coordinates has been estimated at
10 m (Thomas 1965:7-9, 38-52), although the absolute
position accuracy varied considerably due to the
precision of the LORAN operator and the distance
from the LORAN transmitters. Boat positions
reported at 2100 h within 200 miles of the coast are
estimated to be within 3 miles of the absolute
positions.

Hunter (1966) stated that mean separation
distance is preferred for measuring relative
changes in spacing, but for comparison of samples
containing different numbers of individuals, mean
distance to nearest neighbor (Clark and Evans
1954) should be used. We did not use mean distance
to nearest neighbor because most fishermen fish
together with one or more companion boats. Mean
distance to nearest neighbor would thus represent
the average distance separating the same groups
of boats and would give little if any information on
actual compactness of the fleet within a specified
area.

RESULTS AND DISCUSSION

Fishing Power Versus Boat Length and
Number of Lines

Sixty-six area-date strata (1° latitude by 1°
longitude rectangles and 1-day periods) were
selected to examine the relationship between the
fishing power of a boat and its length and number
of lines trolled. All strata had at least 20 boats
reporting within them. (The new mobile stratum
was not used here because the intent was to
partition the fishery area into a number of equal
quadrats, the size and location of the quadrat
being of no consequence. Daily boat positions had
been assigned to 1° longitude rectangles by
FPOW, so this stratum was used for convenience.)

Fishing power estimates were then regressed on
boat length and number of lines. (Data on number
of lines were available only for the 1970 season.) In
none of the strata, in any season, was a significant
regression (F-test, P <0.05) found. This indicated
that no significant relationship existed between a
vessel's fishing power and its length or reported
number of lines trolled within a given 1° by 1°
rectangle during any given day.

Because of the scatter of data for small-scale
time and area strata, the above conclusion did not
rule out the possibility of a significant relationship
between fishing power and boat length or number
of lines. Therefore, a larger stratum was chosen
which included all data for each year. Fishing
power estimates were again regressed on boat
length (1968, 1969, 1970) and number of lines
(1970). The results are shown in Table 2.

Boat length was significantly related (P<0.05)
to fishing power of albacore boats in a time-area
stratum of one season and the entire fishery,
particularly in 1968. The significance of boat

Table 2.-Regression equations and analysis of variance data for
boat length (in meters) and number of lines (1970) versus boat
fishing power.

1968

Fishing power = 0.238
FP (12.2-m boat) = 0.798
FP (18.3-m boat) = 1.078

+ 0.046 (boat length)

Source

df Sum of squares

(viean square

F value

Total

Regression

Residual

810 185.459

1 13.835

809 171.624

0.229

13.835

0.212

65.23**

1969

Fishing power = 0.263
FP (12.2-m boat) = 0.863
FP (18.3-m boat) = 1.163

+ 0.049 (boat length)

Source

df Sum of squares

Mean square

F value

Total

Regression

Residual

271 165.265

1 3.214

270 162.051

0.610
3.214
0.600

5.35*

1970

Fishing power = 0.636
FP (12.2-m boat) = 0.916
FP (18.3-m boat) = 1.056

+ 0.022 (boat length)

Source

df Sum of squares

Mean square

F value

Total

Regression

Residual

200 24.777

1 0.698

199 24.079

0.129
0.698
0.121

5.76*

Fishing power = 0.816 + 0.018 (number
FP (8 lines) = 0.960
FP (12 lines) = 1.032

of lines)

Source

df Sum of squares

Mean square

F value

Total

Regression

Residual

200 24.777

1 0.110

199 24.667

0.139
0.110
0.124

0.89 ns

** significant at the 0.01 level.
• significant at the 0.05 level,
ns nonsignificant.

976

KEENE and PEARCY: COMPARISON OF SUCCESS OF ALBACORE TROLL FISHERMEN

length as it related to fishing power was consider-
ably less in 1969 and 1970 than in 1968, although
the 1968 and 1969 regression equations were
nearly identical.

Fox (see footnote 3) reported that fishing power
of albacore troll boats was related to boat length in
a curvilinear manner for the 1961-70 period, with
boats of the length class 12.2 to 14.9 m exhibiting
the highest estimates of fishing power. There was
no clear indication of a curvilinear relationship in

1968, 1969, or 1970, although several very long
boats (>22.9 m) generally did not have as large
fishing powers as the linear relationship predicted,
thus supporting Fox's conclusions. The sample of
boats used by Fox was considerably larger (10 yr)
and therefore had many more observations of
longer boats than used in this study.

Large boats, moreover, make up a minor portion
of the albacore fleet. The average length (and
standard deviation) of the sample of boats in 1968,

1969, and 1970 was 14.9 m (2.7), 14.9 m (2.1), and
15.2 m (2.7), respectively. Some fishermen feel that
larger boats are more successful because of their
increased seaworthiness and endurance, resulting
in fewer trips to port and permitting more time on
the fishing grounds. Fishermen also feel that
larger boats fish the lures better in rough weather.
Whereas smaller boats tend to jerk the lures as the
waves hit the boats, larger boats push smoothly
through the waves with less jerking of the lures.

The reported number of lines trolled in 1970 was
not significantly related to fishing power. The
number of lines reported varied from 6 to 14, with
10 being the mean and mode. The standard devia-
tion was 1.0. The number of trolling lines reported
on log sheets bears little resemblance to the
number of lines used during varying periods of
fishing activity, according to fishermen. When
fishing activity increases, only two or possibly
three lines are pulled by each man. During periods
of intense activity, each man may only handle one
line, although periods of intense activity are
usually of very limited duration. When the catch
rate increases, the longest lines are pulled on
board first and only the short lines are fished. One
fisherman stated that the number of lines used
was determined primarily by the ability of the
crew in avoiding tangling of lines. However, over
90% of the 1968 logbooks (in which crew size was
recorded) indicated a crew size of two. It would
appear that the possible increase in catch as a
result of a larger crew size during the infrequent
periods of intense fishing activity are offset by the

increase in financial cost of a larger crew size. This
is even more apparent considering that a daily
catch of 180 fish (i.e., about 5 fish per hour per man
for a two-man crew) is considered a very good
catch by an albacore fisherman.

Comparison of Highliners and Lowliners

Some comparisons of highliner and lowliner
boats are given in Table 3. Both groups fished
approximately the same number of days and in the
same period each season. The difference in boat
length was highly significant in all years, par-
ticularly in 1968 when highliner boats averaged 4.9
m longer than lowliner boats. In 1969 and 1970 only
1.5 m separated the average length of highliner
and lowliner boats. Seven of the 1968 highliner
boats were over 15.5 m, whereas none of the 1969
and only one of the 1970 highliner boats were over
15.5 m. Essentially the same proportions of 15.5 m
and longer boats made up the fleet samples in each
season. Lowliner boat lengths were consistently
short, between 14.0 and 15.2 m.

Lowliners often fished along the trailing margin
of the fleet during all years as the fleet moved to
the north. Highliners were more centrally located
in the fleet and along the offshore or leading
margin, as shown in Table 3. In 1968 lowliners
were removed from the main body of the fleet,
generally located far to the south and inshore of
the fleet, whereas highliners tended to be slightly
to the south but offshore of the main fleet center.
In 1969 and 1970 both groups were located closer to
the fleet center, although the lowliners were still
three to four times farther away from the fleet
center than were highliners. Lowliners fished
consistently south of the center in all 3 yr.

A detailed description of the location of high-
liners and lowliners is presented in Figures 2-4.

Table 3.-Comparison of highliners with lowliners, west coast
albacore trollers.

Item

1968

1969

1970

Average boat length (m):

Highliners

Lowliners
Average distance to fleet center
(miles):

Highliners

Lowliners
Average daily travel (miles);

Highliners

Lowliners
Average relative fishing power:

Highliners

Lowliners

19.2
14.3**

30 SW
104 SSE

21
31**

1.61
0.65

15.5
14.0**

5 W
22 SW

26
29 ns

1.57
0.46

16.2
14.6**

8N
25 8

27
28 ns

1.24
0.85

** significant at the 0.01 level, Mest.
ns nonsignificant.

977

<
a:
o
<

o
o

400-1

300-

■^200-

^100

^-
o

0
40

80
>I00

I <«

H ■?

>I00

40-

80-

>100

■EFFORT

I \ I V A

1968

SEPARATION
TANCE

FISHERY BULLETIN: VOL. 74, NO. 4
r40

-30

— J—
26

■"1—
30

— r~
15

— I—
23

JUL

AUG

125

cc o

10 tt <

^t; ' ' 1

(D 1

0 d) 0 oio

0 ^
0

•

(D 0 1

1
1

0)0 ^J 2.

0d) 00
%0 ' -

f
1 1

2 1

1 1

00 '
1

1
1 1

1 1 ? 1 1

t- -

>I00

JUL

AUG

Figure 2.— Locations of 1968 highliners and lowliners relative to the center (medial) of the fleet. The top graph indicates the
corresponding levels of apparent abundance of albacore, fishing effort, and boat separation distance within 31 miles of the fleet center.
The lower four plots show the distance of highliners (circled numbers) and lowliners (noncircled numbers) from the medial fleet center.

978

KEENE and PEARCY: COMPARISON OF SUCCESS OF ALBACORE TROLL FISHERMEN

1969

^300

SEPARATION
DISTANCE ~^^

4o^

>I00

80 ^

>I00

i3 Z

Ll. <

UJ (T

UJ
CO

^^??,

, t«»,

2 1

'^ 1 1

, ^

CD (D

<)>

O 1

0 0(D 1 CD

i S
o

>I00

13 — (zrrp — onj — cp — ^
® III

(2)

JUL

77 ^^Y7~T"^r~T7T

(S6)0(D . 2 ® -^ II?

— r-
28

-r
5

-r
9

AUG

Figure 3.-Locations of 1969 highliners and lowliners relative to the center (medial) of the albacore fleet. See Figure 2 for explanation

of plots.

979

FISHERY BULLETIN: VOL. 74, NO. 4

o
<

I-
u

500

400

c300

o
o

tzoo

100

\ EFFORT

, SEPARATION
't DISTANCE

20 24

JUL

>I00

CO
UJ

30 -5

tn
*- 111

. <

— Q

u. t-

li. <

u cr

>I00

1 5>^f ^"S)^ ^

id) 5® '

1 1

Figure 4.-Locations of 1970 high-
liners and lowliners relative to the
center (medial) of the albacore fleet.
See Figure 2 for explanation of
plots.

H-

<
Ul

• •>
X *

o
to

® '

.dip ,

S ' ' '

>I00

_ ...

1 1 1 '

The plots show where these two groups fished with
respect to the fleet center during periods of vari-
able levels of albacore abundance, fishing effort, and

boat separation distance (shown at the top of the
figures).

A very obvious separation of highliners and

980

KEENE and PEARCY: COMPARISON OF SUCCESS OF ALBACORE TROLL FISHERMEN

lowliners occurred in 1968 (Figure 2). Highliners
fished almost exclusively to the northwest and
southwest of the fleet center. When abundance
was low and effort high (26 July-3 August), high-
liners moved far from the fleet center, as seen in
the southwest quadrant. During 5 and 6 August,
when high catches coincided with high levels of
effort, highliners were found close to the fleet
center, but not as close as during periods of low
effort. Lowliners fished mainly to the south and
away from the fleet center during all levels of
abundance. When abundance was high (5-8
August), lowliners in the southeast quadrant
moved closer to the fleet center. Later as catches
declined, the lowliners moved away from the
center (southeast quadrant, 9-15 August).

There was no obvious separation of highliners
and lowliners in 1969 (Figure 3) comparable to
1968. Highliners fished in all quadrants, as did
lowliners. Some highliners fished away from the
fleet center during periods of low abundance (31
July-2 August; 5-12 August), particularly in the
northwest and southwest quadrants when effort
was high (10-12 August). Lowliners again fished
more in the southern quadrants than did high-
liners but not exclusively so and not as far from
the fleet center as in 1968. In fact, most lowliners
were located near the fleet center until all catches
began decreasing after 5 August. Then, some
lowliners moved away from the fleet (southwest,
northeast; 10-11 August) but the majority
remained near the fleet center.

The short 1970 season provided little informa-
tion on the responses of highliners and lowliners
(Figure 4). As the season began (19-21 July)
highliners were fishing at some distance from the
fleet center. During the period of very high catches
(22-29 July) both highliners and lowliners fished
within 40 miles of the fleet center. No boat
reported a location farther than 80 miles from the
center during this time. There was no indication
that either group dispersed in response to the high
levels of effort and aggregation of boats which
occurred. On 22 July, when separation distance
was lowest and on 26 July when effort was highest,
most highliners were fishing within 20 miles of the
fleet center.

Most highliners did not fish Oregon waters after
30 July, the day catches dropped precipitously. The
lowliners that stayed were northwest of the fleet
center. Catches never returned to their original
high levels, and on 4 August the season was
essentially over for the troll boats.

Some albacore fishermen believe that large
numbers of small fish are located in the offshore
fishing area and that highliners are able to exploit
these fish to a greater degree because of their
greater endurance and seaworthiness. To test this
hypothesis, the average weight of each fish per trip
reported by highliners during July and August
was compared with the average fish weight per
trip for lowliners. The results, given in Table 4,
show that highliners caught significantly smaller
fish than lowliners. This supports the fishermen's
belief that smaller fish are found along the
offshore margins of the fishery where highliners
often fish, while larger fish are found along the
inshore margins of the fishery where lowliners
expend more effort.

The difference between average daily net travel
of highliners and lowliners, based on 2100 h PDT
positions, changed significantly within the 3 yr.
Highliners in 1968 moved 10 miles less per day
than did lowliners (Table 3). In 1969 and 1970 there
was no statistical difference between the average
distance traveled by the two groups. Travel dis-
tances in Table 3 can be compared with the daily
travel of the fleet center (Figure 5). The fleet
center moved an average of 14 miles per day in
1968, 29 miles per day in 1969, and 29 miles per day
in 1970. Highliners moved in a much closer rela-
tionship with the fleet in 1968 than did lowliners.
Lowliners in 1968 traveled twice as far as the
general fleet, yet lagged behind the fleet's north-
erly movement. This was much less apparent in
1969 and 1970.

A comparison of average relative fishing powers
showed that highliners of 1968 and 1969 were
about three times more successful than lowliners in
catching fish (Table 3). Lowliner fishing power
decreased in 1969, even though lowliner and high-
liner boat lengths and daily distances traveled
were similar. In 1970 lowliner and highliner char-
acteristics were quite simlar to those of 1969,
except for calculated fishing power. In 1970 fishing
power of lowliners increased while that of high-
liners decreased. This was probably due to the

Table 4.-Average weight (kilograms) of individual albacore per
trip taken by highliners and lowliners during July and August
1968, 1969, and 1970.

Year Highliners

Lowliners

1968 5.7

1969 °-°

1970 6-'

6.2*

6.4**

6.8*

* significant at the 0.05 level.
** significant at the 0.01 level.

981

FISHERY BULLETIN; VOL. 74, NO. 4

15 20 25 30 4 9 14
JUL. AUG.

19 24 29 3

SEPT.

Figure 5.-Net daily movement of albacore fleet centers- 1968,
1969, and 1970.

extremely short season on highly vulnerable fish, a
situation v^^hich did not provide highliners the
opportunity to utilize their capabilities and fully
develop their tactics and strategies.

This study has shown that the most successful
and least successful fishermen can be character-
ized by their activities as well as by the physical
parameters of their vessels. Success is not assured
by many years of experience, or by a large vessel,
although these characteristics are often associated
with the most successful fishermen. We agree with
Abramson's (1963) suggestion that the fishing
power of individual albacore boats is related to
intrinsic factors of the captain and crew, in addi-
tion to the boat's physical parameters.

ACKNOWLEDGMENTS

This research was supported by U.S. Naval
Oceanographic Oflice Contract N62306-70-C-0414;
U.S. Bureau of Commercial Fisheries (now Na-
tional Marine Fisheries Service) Contract 14-17-
0002-333; and Oregon State University Sea Grant
College Program, supported by NOAA, Oflfice of
Sea Grant, Grant No. 04-3-158-4. R. Michael Laurs

and William Fox of the Southwest Fisheries
Center, National Marine Fisheries Service,
NOAA, La JoUa, Calif., and Larry Hreha of the
Fish Commission of Oregon made possible the
analysis of the logbook data.

LITERATURE CITED

Abramson, N. J.

1963. A method for computing estimates and variances of
relative log fishing powers of California albacore
vessels. In H. Rosa, Jr. (editor), Proceedings of the World
Scientific Meeting on the Biology of Tunas and Related
Species, p. 1209-1215. FAO Fish. Rep. 6.
Abramson, N. J., and P. K. Tomlinson.

1972. An application of yield models to a California ocean
shrimp population. Fish. Bull., U.S. 70:1021-1041.
Ayers, R. J., AND J. M. Meehan.

1963. Catch locality, fishing effort, and length frequency
data for albacore tuna landed in Oregon, 1951-60. Oreg.
Fish. Comm., Invest. Rep. 2, 180 p.
Berude, C. L., and N. J. Abramson.

1972. Relative fishing power, CDC 6600, FORTRAN IV.
Trans. Am. Fish. Soc. 101:133.
Beverton, R. J. H., AND S. J. Holt.

1957. On the dynamics of exploited fish populations. Fish.
Invest. Minist. Agric. Fisl\. Food (G.B.), Ser. II, 19, 533 p.
Clark, P. J., and F. C. Evans.

1954. Distance to neighbor as a measure of spatial relation-
ships in populations. Ecology 35:445-453.

Clemens, H. B.

1955. Catch localities for Pacific albacore (Thunnus germo)
landed in California, 1951 through 1953. Calif. Dep. Fish
Game, Fish Bull. 100, 28 p.

1961. The migration, age, and growth of Pacific albacore
(Thunnus germo), 1951-1958. Calif. Dep. Fish Game, Fish
Bull. 155, 128 p.
Clemens, H. B., and W. L. Craig.

1965. An analysis of California's albacore fishery. Calif.
Dep. Fish Game, Fish Bull. 128, 301 p.

Daniels, H. E.

1952. The covering circle of a sample from a circular normal
distribution. Biometrika 39:137-143.
Hunter, J. R.

1966. Procedure for analysis of schooling behavior. J. Fish.
Res. Board Can. 23:547-562.

Marr,J. C.

1951. On the use of the terms abundance, availability and
apparent abundance in fishery biology. Copeia
1951:163-169.
Panshin, D. a.

1971. Albacore tuna catches in the northeast Pacific during
summer 1969 as related to selected ocean conditions.
Ph.D. Thesis., Oregon State Univ., Corvallis, 110 p.

Roberts, K.

1972. Diversity - characteristic of Oregon's year 'round
fishery. Oreg. State Univ. Sea Grant, Ext. Mar. Advis.
Program 15, 2 p.

Robson, D. S.

1966. Estimation of the relative fishing power of individual
ships. Int. Comm. Northwest Atl. Fish., Res. Bull. 3:5-14.
Thomas, P. D.

1965. Mathematical models for navigational systems. U5.
Nav. Oceanog. Off. Tech. Rep. TR-182, 151 p.

982

NOTES

SEASONAL SPAWNING CYCLES OF THE

SCIAENID FISHES GENYONEMUS LINEATUS

AND SERIPHUS POLITUS

The white croaker, Genyonemus lineatus (Ay res),
and queenfish, Seriphus politus Ayres, are two of
the common inshore fishes occurring along the
southern California coast (Miller and Lea 1972).
Detailed reproductive data are not available for
these species. The purpose of this note is to provide
information on their seasonal spawning cycles.

Materials and Methods

Monthly samples are from November 1974 to
October 1975. Most specimens were collected by
hook and line from the Santa Monica Pier, Los
Angeles County, Calif. Remaining fishes were
obtained about 4.2 km south of Santa Monica at
the Scattergood Steam Plant, El Segundo, Los
Angeles County. Scattergood fishes had been
exposed to temperatures between 23° and 41°C.
Histological comparisons of these fishes with
freshly caught specimens showed the ovaries were
not altered by this treatment. Specimens are
deposited in the ichthyology collection of the Los
Angeles County Museum of Natural History.

Fishes were immediately slit and placed in 10%
Formalin.^ Gonads were embedded in paraffin and
histological sections cut at 8 jum. Slides were
stained using iron hematoxylin followed by eosin
counterstain. Seasonal occurrences of oocytes
(Tables 1, 2) were calculated by randomly selecting
areas of slides from each monthly representative
and classifying oocytes as to their category (Type
1, 2, or 3). Areas of a slide were surveyed until at
least 100 oocytes were classified.

Results and Discussion

Three classes of oocytes are present in the
ovaries of G. lineatus (Table 1) and S. politus
(Table 2). Type 1 is the most abundant class and
varies from those recently derived from oogonia to
those approaching Type 2 oocytes. Type 2 oocytes
have diameters between 100 and 270 jum and differ

from Type 1 oocytes in the presence of a zona
pellucida and zona granulosa. Small quantities of
yolk granules may be found on the periphery of
larger representatives of this class. The diameter
of yolk filled mature Type 3 oocytes is greater than
270 jum. The smallest fishes to contain Type 3
oocytes measured 143 mm standard length (SL)
for G. lineatus and 148 mm SL for S. politus.

As shown in Tables 1 and 2 there are several
differences in seasonal distribution of oocytes
reflecting the spawning cycles of G. lineatus and
S. politus. The major difference is in abundance of
Type 3 oocytes indicating G. lineatus comes into
spawning condition in October and spawns inter-
mittently into April. Seriphus politus enters
spawning condition in April and spawns into
August. These data support the findings of
Skogsberg (1939) who reported that S. politus
spawns throughout summer and G. lineatus
spawns from November through May off
California.

Table 1. -Monthly distribution of Genyonemus lineatus oocytes
with mean standard length (mm) ± standard error, November
1974-October 1975.

Total

Type 1

Type 2

Type 3

Month

oocytes

(%)

SL±SE

Nov.

1,369

203.3 ± 8.7

Dec.

1,579

228.8 ± 4.6

Jan.

1,316

217.3 ±5.9

Feb.

1,478

202.1 ± 3.9

Mar,

1,717

200.7 ± 9.8

Apr.

1,631

204.2 ± 4.4

May

2,138

218.2 It 4.7

June

1,251

218.0 ±2.9

July

2,103

212.0 ±4.0

Aug.

1,606

239.8 ± 3.5

Sept.

1,589

243.0 ± 3.8

Oct.

1,340

234.5 ± 6.2

Table 2.-Monthly distribution of Seriphus politus oocytes with
mean standard length (mm) ± standard error, November 1974-
October 1975.

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

Total

Type 1

Type 2

Type 3

Month

N
14

oocytes
1,531

(%)

SL±SE

Nov.

100

192.8 ±2.6

Dec.

1,379

100

215.2 ±4.6

Jan.

1,607

100

203.7 ± 2.4

Mar

1,563

215.2 ± 5.1

Apr.
May
June

1,604

200.2 ± 5.5

1,729

214.6 ±3.1

1,736

217.5 ±3.9

July

1,864

225.8 ± 3.4

Aug.
Sept.
Oct.

1,536

202.9 ± 4.0

1,499

212.9 ± 3.0

1,574

207.6 ± 3.4

983

In both species oocyte maturation is a contin-
uous process that occurs throughout the reproduc-
tive period (Tables 1, 2) with multiple spawnings
occurring. Depleted ovaries containing mainly
Type 1 oocytes were not observ-ed until conclusion
of the spawning season. The presence of various
groups of developing oocytes as occurs in G.
lineatus and S. politus was termed asynchronism
by Yamamoto and Yamazaki (1961) who found this
condition common in fishes with long breeding
seasons and multiple spawnings.

Another difference (Table 1) was the persistence
of small quantities of Types 2 and 3 oocytes in G.
lineatui^ after the conclusion of spawning in April
which persist throughout summer. It is more
typical for remaining vitellogenic oocytes to
undergo atresia at the end of the spawning season
as occurs in S. politus whose inactive ovaries
contained only Type 1 oocytes (Table 2) from
November to January. These low frequencies of
mature summer G. lineatus oocytes may suggest
spawning continued at a reduced frequency dur-
ing this period. A more plausible explanation
might be that these oocytes will ovulate early in
the next spawning season. It thus appears that
some early ovulating G. lineatus oocytes initiated
yolk deposition late in the previous spawning
season and remained over summer. It may be
energetically advantageous for these yolk filled
eggs to remain over summer as opposed to re-
sorbing them.

As G. lineatus ranges from Baja California to
British Columbia and S. politus from Baja
California to Oregon (Miller and Lea 1972), my
data may be useful for subsequent investigations
to determine geographic variation in reproduction
for these species.

Acknowledgments

I thank Camm G. Swift (Los Angeles County
Museum of Natural History) for his helpful sug-
gestions. Several personnel, Kenneth L. Bosworth,
George R. Spencer, and George Thomas, of the
Scattergood Steam Plant were both accommodat-
ing and courteous. Lester Neiper helped in the
collection of specimens from Santa Monica Pier.

Literature Cited

Miller, D. J., and R. N. Lea.

1972. Guide to the coastal marine fishesof California. Calif.
Dep. Fish Game, Fish Bull. 157, 235 p.

984

Skogsberg, T.

1939. The fishes of the family Sciaenidae (croakers) of
California. Calif. Dep. Fish Game, Fish Bull. 54, 62 p.
Yamamoto, K., and F. Yamazaki.

1961. Rhythm of development in the oocyte of the goldfish,
Carassius auratus. Bull. Fac. Fish., Hokkaido Univ.
12:93-110.

Stephen R. Goldberg

Department of Biology
Whittier College
Whittier, CA 90608

FOOD OF FIVE SPECIES OF

COOCCURRING FLATFISHES ON

OREGON'S CONTINENTAL SHELF

The purpose of this paper is to describe and to
compare the food of five flatfish species that
actually cooccurred at one specific time and place
on the central Oregon continental shelf: English
sole, Parophrys vetulus Girard; rex sole, Glyptoce-
phalus zachirus Lockington; rock sole, Lepidop-
setta bilineata (Ayres); petrale sole, Eopsetta
jordani (Lockington); and Pacific sanddab, Citha-
richthys sordidus (Girard). These demersal fishes
are common along the west coast of North Amer-
ica, their ranges overlapping between southern
California and the Gulf of Alaska (Hart 1973).
Parophrys vetulus, C. sordidi(s, and L. bilineata
occur mainly on the inner continental shelf. Eop-
setta jordani is fished commercially on its feeding
grounds (73-128 m), and in deep water (311-457 m)
where spawning occurs (Forrester 1969). Glyp-
tocephalus zachirus has a broad bathymetric
range— it is common off Oregon and Washington
from 90 to 550 m (Alverson et al. 1964). Off Oregon
it was the second most numerous member of a
species association ranging from 119 to 199 m, on
an average sediment type of 69^ sand, 19*^ silt,
and 12% clay (Day and Pearcy 1968). In that same
study, C. sordidus and P. vetulus composed 80.3%
of a species association of fishes in shallower water
(42-73 m) on a sandy bottom. According to Alver-
son (1960), L. bilineata is common on sandy or
gravel bottom. The five flatfish species attain
maximum sizes ranging from 410 mm for C.
sordidus to 700 mm for E. jordani (Hart 1973).

Pearcy and Vanderploeg (1973) listed major food
items— combined from several locations, seasons,
and years-for most of the above species. That

study provided generalized information on food
habits, but little insight into possible intra- or
interspecific differences in diets resulting from
actual interaction among cooccurring fishes. Our
study is based on a single collection minimizing
temporal and spatial variations associated with
sampling. Food items were identified to species
whenever possible. Thus, a detailed comparison of
food taxa is arrived at with minimal geographic
and no seasonal eff'ects.

A trawl haul of 75 min total duration was made
beginning at 1345 h Pacific daylight time, on 13
April 1975 with an Atlantic- Western trawl (24-m
footrope) from the Betty- A, a commercial dragger,
at approximately lat. 44°42'N, long. 124°24'W.
Depth of water was 95-106 m. The sediment was
sand (Byrne and Panshin 1968). Stomachs of fishes
were removed and preserved in Formalin^ at sea
(Table 1). Food items were identified and enumer-
ated in laboratories ashore.

Table 1. -Fishes captured in an Atlantic-Western trawl on 13

April 1975,

Total

length

No. with

No.

range

No.

stomach

Species caught

(mm)

examined

contents

Citharichthys

sordidus

181

90- 377

Parophrys vetulus

230- 450

Glyptocephalus

zachirus

240- 360

Eopsetta jordani

240- 510

Lepidopsetta

bilineata

247- 474

Raja binoculata

940-1,460

Raja rhina

800- 890

Raja kincaidi

560- 570

Pleuronichthys

verticalis

251- 254

Ophiodon elongatus

850

Squalus acanthias

1,000

—

Forage Organisms

All the food items identified from five species of
flatfishes are listed in Table 2, and the major food
taxa (taxa having a frequency of occurrence of
10^ or more) are listed for individual fish of three
species of flounders in Table 3.

Parophrys vetulus had a diverse diet, feeding
primarily on polychaetes and amphipods. Mol-
lusks, ophiuroids, and Crustacea were also repre-
sented. The amphipod Ampelisca macrocephala,
the most numerous single prey species, occurred in
60% of fish. The diversity of the diet of P. vetulus is

•Reference to trade names does not imply endorsement by the
National Marine Fisheries Service, NOAA.

due to the many different types of food consumed
by individual fish (represented by the vertical
columns in Table 3) rather than by different fish
feeding on different prey. Parophrys vetulus
appears to be an opportunistic feeder. Forrester
(1969) reported polychaetes, clams, and ophiuroids
as primary food organisms of P. vetulus, with
incidental occurrences of sandlance, crab, am-
phipods, shrimp, squid, and small fish. Pearcy and
Vanderploeg (1973) found polychaetes, amphipods,
and pelecypods were important prey of P. vetulus
off Oregon.

Glyptocephalus zachirus fed primarily on four
species of amphipods and secondarily on poly-
chaetes. Amphipods occurred in all but one
stomach, polychaetes in 11% of the stomachs with
food. Nematodes were encountered in 38% of the
stomachs but were probably parasitic (Robert
Olson, pers. commun.). Pearcy and Vanderploeg
(1973) also found polychaetes and amphipods to be
the major food of G. zcu-hirus off Oregon.

The principal food oi Lepidopsetta bilineata was
ophiuroids. All but one individual had been feed-
ing on Ophiura, which constituted the bulk of the
stomach contents. A few polychaetes and mollusks
were also present. According to Shubnikov and
Lisovenko (1964), the basic items of its diet are
polychaetes, mollusks, shrimps, and other crus-
taceans. Fishes (sandlance) and echinoderms were
occasionally found in stomachs. Food items
reported for L. bilineata in Hecate Strait, British
Columbia, by Forrester and Thomson (1969) were
clams, polychaetes, crabs, shrimps, sandlance,
herring, echinoderms, and amphipods.

Eopsetta jordani preyed on fishes and decapod
crustaceans. Polychaetes and amphipods were not
present in its diet. Ketchen and Forrester (1966)
found euphausiids, herring, sandlance, and shrimp
as major food items in stomachs of E. jordani.
Pearcy and Vanderploeg (1973) reported shrimps,
pelagic fishes, and euphausiids as major food
items, indicating that this species feeds largely on

pelagic prey.

Citharichthys sordidus had been feeding inten-
sively on the northern anchovy, Engraulis mor-
dax. Anchovy were noted in nearly all the sanddab
when stomachs were removed, and all intact
preserved stomachs contained them. Since an-
chovy were not caught in the otter trawl, feeding
in the net is thought to be unlikely. According to
Pearcy and Vanderploeg (1973), euphausiids,
shrimps, amphipods, and crab larvae were common
in C. sordidus stomachs.

985

Table 2.-Taxa identified from stomach contents of five species of Pacific Northwest flatfishes.*

Parophrys

Glyptocephalus

Lepidopsetta

Eopsetta

Citharichthys

Taxa

vetulus

zachirus

bilineata

jordani

sordidus

POLYCHAETA

94.6

71.4

40.0

Aphroditidae

Aphrodita negligens

Capitellidae

^rCapitellidae spp.
\_Notoniastus spp.

10.8

Chaetopteridae spp.

Cirratulidae

-Cirratulidae spp.

40.5

Chaetozone setosa

Chaetozone spp.

^Tharyx spp.

Goniadidae

Goniadidae spp.

Glycinde pi eta (?)

13.5

Lumbrineriidae

Lumbrineris spp.

18.9

Magelonidae

Magelona spp.

Maldanidae spp.

Nephtyidae

'Nephtys caecoides
.Nephtys spp.

43.2

Onuphidae

Nothria geophilHormis (?)

Nothria iridescens (?)

10.8

14.3

13.3

Nothria spp.

19.0

Opheliidae

Opheliidae spp.

9.5

Ammotrypane aulogaster

8.1

Orbiniidae

Haploscoloplos spp.

Oweniidae

Myriochele oculata

16.2

Myriochele spp.

10.8

Owenia spp.

16.2

Paraonidae spp.

21.6

Pectinariidae

Pectinaria spp.

Phyllodocidae

Eteone longa

Polynoidae spp.

13.3

Sigalionidae

"Sigalionidae spp.
.Thalenessa spinosa

18.9

Spionidae spp.

18.9

Terebellidae

Terebellidae spp.
.Polycirrus spp.

35.1

Unidentified

GASTROPODA

13.5

11.1

Cylichna attonsa

10.8

Mitrella gouldii

Mitrella spp. (?)

11.1

PELECYPODA

27.0

9.5

13.3

Acila castrensis

Axinopsida serricata

Cardiomya oldroydi

Nucula tenuis

16.2

Macoma spp.

Tellina carpenter! (?)

8.1

13.3

Unidentified

■

SCAPHOPODA

16.2

4.8

Dentalium sp.

13.5

Unidentified fragment

Scaphopoda(?)

CEPHALOPODA

11.1

Octopoda

11.1

Beak of Loligo spp. (?)

11.1

CRUSTACEA

91.9

100.0

6.7

28.6

33.3

Cypris larvae (?)

Copepoda (calanoid)

10.8

Mysidacea

Neomysis spp.

11.1

Unidentified

14.3

Ostracoda (?)

Cumacea

10.8

anaidacea

986

Table 2.-Continued.

Taxa

Parophrys
vetulus

Glyptocephalus
zachirus

Lepidopsetta
bilineata

Eopsetta
jordani

Citharichthys
sordidus

Euphausiacea

Euphausia pacitica
Decapoda
Natantia

Crangon spp.
Nectocrangon spp.
Unidentified shrimp
Reptantia

Pagurus samuelis
Mursia spp.
Crab leg
Amphipoda
Ampeliscidae

Ampelisca cristata
Ampelisca macrocephala
Ampelisca spp.
Amphilochidae spp. (?)
Aoridae

Lembos spp.
Argissldae

Argissa hamatipes
Isaeidae
Photis brevipes
Protomedeia spp.
Lysianassidae

Lysianassidae spp.
Acidostoma spp.
Anonyx anivae
Hipomedon wecomus
Oedicerotidae

Monoculodes emarginatus
Monoculodes sp. #1
Synchelidium shoemaker!
Westwoodilla caecula
Phoxocephalidae
Paraphoxus abronius
Paraphoxus daboius (?)
Paraphoxus epistomus (?)
Paraphoxus fatigans
Paraphoxus lucubrans
Paraphoxus milleri
Paraphoxus obtusidens
Paraphoxus variatus
Paraphoxus spp.
Pieustidae

Pleustidae spp.
Pleusymtes coquilla
OPHIUROIDEA

Amphiodia periercta
Amphiodia urtica
Amphiuridae spp.
Ophiura lutkeni
Ophiura sarsii
Ophiura spp.
PISCES
Agonidae
Engraulis mordax
Glyptocephalus zachirus
Radulinus spp.
Unidentified
Fish scale
NEMERTINEA
NEMATODA
SIPUNCULIDA
ECHIURIDA
ACANTHOCEPHALA
Miscellaneous
Gastropod egg case
Egg mass
Lenses
Unidentified remains

83.8
59.5

X
X

18.9

X
X
X
X

X
X

21.6
10.8

16.2

10.8

83.8
10.8
10.8
10.8
35.1

8.1

13.5

2.7

14.3

22.2

11.1
X

95.2

33.3

28.6

X
X

33.3
4.8

33.3

9.5

X
X

93.3

73.3
20.0

100.0
14.3
14.3
14.3
14.3

100.0
100.0

13.3

.Frequency of occurrence is given as a percentage for food taxa whenever these '^It^/;^^' !^"J°J:,°
of the five species of fishes. Opheliidae were significant on a weight basis "^J^'^^^''"'- ^"^ *^^^
Ammotrypane aulogaster for calculation of similarity. An "x" denotes any other occurrence.

2Taxa enclosed within brackets were treated as a single group m Table 3.

r greater of any
combined with

987

Table 3.-Numbers of food items of major' taxa in contents of individual fish stomachs. Each vertical column represents one stomach.

Total
no.

Fish species and taxa

Number of food items

Parophrys vetulus
POLYCHAETA:

Capitellidae

Cirratulidae

Glycinde picta (?)

Lumbrineris spp.

Nephtys spp.

Nothria iridescens (?)
^rMyriochele oculata
LMyriochele spp.

Owenia spp.

Paraonidae

Sigalionidae

Spionidae

Terebellidae
AMPHIPODA:

Ampelisca macrocephala

Hippomedon wecomus

Paraphoxus epistomus (?)

Paraphoxus latigans

Paraphoxus obtusidens

Paraphoxus spp.^
GASTROPODA:

Cylichna attonsa
PELECYPODA:

Nucula tenuis
SCAPHOPODA:

Dentallum sp.
COPEPODA (calanoid)
CUMACEA
OPHIUROIDEA:

tAmphiodia periercta
Amphiodia urtica
Amphiuridae
Ophiura lutkeni
NEMERTINEA
Glyptocephalus zachirus
POLYCHAETA:

Nothria iridescens (?)
Nothria spp.^
Opheliidae
AMPHIPODA:
Ampelisca macrocephala
Hippomedon wecomus
Paraphoxus epistomus (?)
Paraphoxus obtusidens

Lepidopsetta bilineata
POLYCHAETA:

Polynoidae

Nothria iridescens (?)
SIPUNCULIDA (?)
PELECYPODA:

Tellina carpenteri (?)
OPHIUROIDEA:

Ophiura lutl^eni

Ophiura spp.^

1
3

7
2

*
1

*
*

3
3

1
1

10
2

1
2

2
1

1
1

8
2

1
1
4

1
1

12
1
2

1
2
1

7
2
1
1

1
1

1
2

1 1 1

2 1

1 1
1

1 1

4 5 4

1 1

* *

1 47
6

8
4

13
8

4 1

1 64
8

13
4

7
4

5
1 5

18
5

6
2

1
1
1

10
8
9

4
2

•

'Taxa having a frequency of occurrence of at least 10%.

^Taxa within brackets were treated as a single group for calculation of similarity.

'Taxa not used to determine similarity.

•Fragment.

Discussion

The flatfishes examined in this study comprised
two distinct feeding types based on the species
composition of prey and the frequency of occur-
rence of major food items. Parophrys vetulus, G.
zachirus, and L. bilineata were benthophagous,
feeding on benthic infaunal and epifaunal in-
vertebrates, mainly polychaetes, amphipods, and

ophiuroids. Eopsetta jordani and C. sordidus were
piscivorous and fed more on pelagic animals,
consuming mainly fishes in addition to shrimp,
mysids, euphausiids, and cephalopods. Fishes did
not occur in the stomachs of the benthic inverte-
brate feeders, except for two fishes found in G.
zachirus.

Differences were sometimes obvious in the food
habits of fishes within each feeding type. The

988

similarity among the food habits of the three
fishes that preyed on benthic invertebrates was
calculated using commonly occurring prey (Table
3) and Horn's (1966) measure of niche overlap. The
overlap was largest between P. vetuliis and G.
zachirus (Cx = 0.40). This is because both fishes
fed on the same species of amphipods and the
polychaete Nothria iridescens. Parophrys vefulus
preyed on a very diverse array of invertebrate
taxa, while G. zachiriAs appeared to be more
selective in its feeding.

The amount of food overlap among the other
species pairs was low (0.19 between P. vefulus: and
L. hilineaia and only 0.03 between G. zach irus and
L. hilineaia). These low values are explained by
the high occurrence of Ophiura lufkeni only in L.
bilineata. Also, L. bilineafa fed on members of two
scaleworm families, Aphroditidae and Polynoidae,
neither of which is represented in the other
flatfishes.

The food habits of the flatfishes that we found to
be mainly piscivorous were also different. Eopsetta
jordani preyed on various fishes, including benthic
agonids, pleuronectids, and cottids, as well as a
benthic shrimp and crab, whereas C sordidus fed
almost exclusively on the pelagic Engraidis
mordax.

Partitioning of the food resources among the
five flatfish species is obvious from our data-the
different, syntopic species fed upon different
organisms. According to MacArthur and Pianka
(1966), a more productive environment should lead
to a more restricted diet in terms of different
species eaten, but in a patchy environment this
does not apply to predators that spend most of
their time searching. If the bottom occupied by P.
retulus is inhabited by patches of invertebrates,
then this species might be such a scavenging
"generalist" predator. Rae (1969) documented a
food interaction similar to the one in this study
between the lemon sole, Microi^fomus kitt, and
witch, Glyptocephalu^ cynoglossus, off Scotland.
The witch, restricted to muddy bottoms, fed on a
more restricted fauna than the lemon sole, whose
diet included the hard-bottom species typical of its
habitat in addition to species from muddy-bottom
types.

Differences in time of feeding could also account
for differences in the species composition of prey.
Diel changes in the habits of prey can serve to
increase or decrease their exposure to predators,
and hence their availability as food (Hobson 1965;

Jones et al. 1973). More so than the other species,
the stomach contents of G. zachina^ were in a late
stage of digestion, suggesting that they had fed a
longer time before capture than other species.

The diet of fishes is related not only to their
feeding behavior but also to their digestive mor-
phology and mouth structure. The size of the
mouth relative to body length correlated with the
size of food organisms for bothid flounders in
Georgia coastal waters (Stickney et al. 1974).
Symmetry of the jaws plays an important role in
the mode of feeding, as species with symmetrical
jaws generally take free-swimming food, while
those with asymmetrical jaws are mainly bottom
feeders (Yazdani 1969). Flatfishes that feed on
polychaetes and mollusks typically have smaller
stomachs, larger intestines, and smaller gill rakers
with fewer teeth than flatfishes that feed on other
fishes (DeGroot 1971; Tyler 1973). The mouths of P.
vetuh(^<, G. zachinii^, and L. bilineafa are small,'-'
the jaws and dentition are better developed on the
blind side (i.e., asymmetrical), the teeth are inci-
sorlike (bluntly conical in L. bilineafa), and the gill
rakers are without teeth. These morphological
adaptations correlate with the preponderance of
benthic invertebrates in their diets. The pisci-
vores, E. jordani and C. sordidus, on the other
hand, have larger mouths,'' nearly symmetrical
jaws with sharp teeth, and long gill rakers with
teeth.

Acknowledgments

We thank William Colgate, Allan Fukuyama,
and Valerie Hironaka for identifying amphipods,
mollusks, and ophiuroids, respectively. Discussions
with Michael Richardson and Robert Carney were
useful. This work is a result of research sponsored
[in part] by the Oregon State University Sea
Grant College Program, supported by NOAA
Oflice of Sea Grant, U.S. Department of Com-
merce, under Grant #04-5-158-2.

Literature Cited

Alverson, D. L.

1960. A study of annual and seasonal bathymetric catch

^Length of maxillary into head on ocular side is 4' 4-4y5, 4^/z-5%,
and 3'^-4'/5. respectively (Norman 1934); also see Norman for line
drawings depicting the relative mouth size of flatfishes discussed

in this paper. , „,, , , o\

^Length of maxillary into head is about ^/s and 2% (nearly S),
respectively (Norman 1934).

989

patterns for commercially important groundfishes of the
Pacific Northwest coast of North America. Pac. Mar. Fish.
Comm. Bull. 4:1-66.
Alverson, D. L., a. T. Pruter, and L. L. Ronholt.

1964. A study of demersal fishes and fisheries of the north-
eastern Pacific Ocean. H. R. MacMillan Lectures in
Fisheries, Inst. Fish., Univ. B.C., 190 p.

Byrne, J. V., and D. A. Panshin.

1968. Continental shelf sediments off Oregon. Oreg. State
Univ. Sea Grant E.\t. Mar. Advis. Program 8, 4 p.
Day, D. S., and W. G. Pearcy.

1968. Species associations of benthic fishes on the continen-
tal shelf and slope off Oregon. J. Fish. Res. Board Can.
25:2665-2675.

DeGroot, S. J.

1971. On the interrelationships between morphology of the
alimentary tract, food and feeding behaviour in flatfishes
(Pisces: Pleuronectiformes). Neth. J. Sea Res. 5:121-196.

Forrester, C. R.

1969. Life history information on some groundfish species.
Fish. Res. Board Can., Tech. Rep. 105, [17 p.]

Forrester, C. R., and J. A. Thomson.

1969. Population studies on the rock sole (Lepidupaetta
bilineata) of northern Hecate Strait, British Columbia.
Fish. Res. Board Can., Tech. Rep. 108, 104 p.
Hart, J. L.

1973. Pacific fishes of Canada. Fish. Res. Board Can.. Bull.
180, 740 p.
HOBSON, E. S.

1965. Diurnal-nocturnal activity of some inshore fishes in
the Gulf of California. Copeia 1965:291-302.

HORN,H.S.

1966. Measurement of "overlap" in comparative ecological
studies. Am. Nat. 100:419-424.

Jones, D. A., N. Peacock, and 0. F. M. Phillips.

1973. Studies on the migration of Tritaeta gihbosa, a
subtidal benthic amphipod. Neth. J. Sea Res. 7:135-149.
Ketchen, K. S., and C. R. Forrester.

1966. Population dynamics of the petrale sole, Eopaetta
Jordan i, in the waters of western Canada. Fish. Res.
Board Can., Bull. 153, 195 p.
MacArthur, R. H., and E. R. Pianka.

1966. On optimal use of a patchy environment. Am. Nat.
100:603-609.
Norman, J. R.

1934. A systematic monograph of the flatfishes
(Heterosomata). Vol. I. Psettodidae, Bothidae,
Pleuronectidae. Br. Mus. (Nat. Hist.), Lond., 459 p.
Pearcy, W. G., and H. A. Vanderploeg.

1973. Radioecology of benthic fishes off Oregon. In
Radioactive contamination of the marine environment, p.
245-261. Int. At. Energy Agency, Vienna.
Rae, B. B.

1969. The food of the witch. Mar. Res. Dep. Agric. Fish.
Scotl. 2, 23 p.
Shubnikov, D. a., and L. A. Lisovenko.

1964. Data on the biology of rock sole of the southeastern
Bering Sea. Tr. Vses. Nauchno-issled. Inst. Morsk. Rybn.
Khoz. Okeanogr. 49 (Izv. Tikhookean. Nauchno-issled.
Inst. Morsk. Rybn. Khoz. Okeanogr. 51): 209-214. (Transl.
in Soviet Fisheries Investigations in the Northeast Pacific,
Part II, p. 220-226, by Israel Program Sci. Transl., 1968,
available Natl. Tech. Inf. Serv., Springfield, VA, as TT
67-51204.)

Stickney, R. R., G. L. Taylor, and R. W. Heard III.

1974. Food habits of Georgia estuarine fishes. I. Four
species of flounders (Pleuronectiformes: Bothidae). Fish.
Bull., U.S. 72:515-525.
Tyler, A. V.

1973. Alimentary tract morphology of selected North At-
lantic fishes in relation to food habits. Fish. Res. Board
Can., Tech. Rep. 361, 23 p.
Yazdani.G. M.

1969. Adaptations in the jaws of flatfish (Pleuro-
nectiformes). J. Zool. (Lond.) 159:181-222.

Michael J. Kravitz

William G. Pearcy

M. P. Guin

School of Oceanography

Oregon State University

CorvaUis. OR 97331

AGE DETERMINATION OF A TROPICAL

REEF BUTTERFLYFISH UTILIZING
DAILY GROWTH RINGS OF OTOLITHS

The recent economic expansion of the aquarium
fish industry in Hawaii has raised questions con-
cerning the judicious exploitation of reef resources
(Pellegrin 1973; Randall 1973; Reese 1973). How-
ever, appropriate management strategies cannot
be implemented until sufficient biological data
have been gathered, allowing a characterization of
exploited populations of fishes. The relative
paucity of such information concerning the vast
majority of reef species underscores the need for
future research.

Studies pertaining to the age and growth of
fishes are especially useful in the analysis of
exploited stocks. Unfortunately, efforts to age
tropical fishes in the past have proved to be largely
unsuccessful and/or involve considerable expen-
ditures in time and effort (Pannella 1974). How-
ever, the recent studies of Pannella (1971, 1974)
have initiated the development of a technique for
determining the age of tropical fishes without
having to resort to more elaborate approaches such
as the Peterson method of ageing. Panella has
provided evidence that many species of both tem-
perate and tropical fishes deposit lamellae on their
otoliths with a diel periodicity. These lamellae are
visible as rings or circuli after the otolith has been
properly prepared. In the absence of annuli, these
rings may be used to age fish. A recent investiga-
tion by Struhsaker and Uchiyama (1976) using this
technique was successful in ageing the Hawaiian

990

anchovy {Stolephoriis purpureus) and in showing
the daily nature of these lamellae.

This paper reports on studies of the age and
growth of the Hawaiian endemic millet-seed
butterflyfish, Chaetodon miliaris Quoy and
Gaimard (Perciforms: Chaetodontidae), using this
approach. Butterflyfishes are exceptionally at-
tractive and are heavily exploited by the aquarium
industry in Hawaii. This study was initiated in
order to obtain information useful to state
regulatory agencies in the management of reef
fish stocks.

DORSAL DOME

NUCIEAR AREA

EXCISURA

ROSTRUM

SULCUS

Methods

All fish were collected by spearing around the
island of Oahu, Hawaii, during 1974 and were
measured to the nearest millimeter standard
length (SL) while still fresh. Next, the otoliths
were extracted by means of a horizontal section
through the cranium above the eyes. Of the
three otoliths on each side, only the largest, the
sagitta, was studied. Figure 1 depicts a left sagitta
of a 94-mm C. miliaris viewed medially. After
both sagittae were removed, all membranes and
endolymph were carefully teased way under a
dissecting microscope. The otoliths were then
rinsed in water and placed in a 2% aqueous solution
of HCl for several minutes of etching. They were

Figure l.-Schematic representation of the left sagitta of a
94-nim Chaetodon miliaris viewed medially.

then rinsed again, thoroughly dried, and finally
mounted in depressions of glass slides where they
were immersed in euparal (an aromatic oil which
acts as a clearing agent) and covered with glass
cover slips. After clearing for 2 wk, the otoliths
were ready for reading. Otoliths were read from
the nucleus outward along their long axis with a
compound binocular microscope utilizing trans-
mitted light at a magnification of 400 x (Figure 2).
The rings in each sagitta were counted twice,
using a hand counter, and the average of the four
readings obtained from each specimen was used to
estimate the age of the fish in days.

¥vr^* •

'& V H

Figure 2.-Intemal ring structure of
the otolith of Chaetodon miliaris
specimen number 11. Not all the rings
are visible in this photograph.

991

Results

The counts of rings within otoliths are summa-
rized in Table 1. The average number of rings for
each fish has been rounded to the nearest integer.

On the assumption that one ring is equal to one
day's growth (Pannella 1971, 1974; Struhsaker and
Uchiyama 1976), the data were fitted to the von
Bertalanffy growth equation employing the tech-
niques of Allen (1966). This model states:

/, -L^(l-e

Kit

''')

(1)

where /, = length at time ^

Loo = the average length of a group of fish

grown for an infinite period of time
K =a. growth parameter which describes
the rate at which /, is approaching L^
^0 =the back calculated X intercept or the
time at which size was zero.

120-

I 100

60-

•o
c
• 40

r"
0

I, a 127(1- •

-00311*301

200

T 1 r

400
Days

600

800

Year*

FiGL'RE 3. -The von Bertalanffy growth curve in length fitted to
15 individuals of Chaetadon niiliaris aged by means of otoliths.

The data are plotted along the calculated von
Bertalanffy growth curve in Figure 3. The cal-
culated growth equation for the data in this report
is:

/, = 127(1 - e

,-0.0031(/+30)

)

(2)

when size is expressed as SL in millimeters and
time is expressed in days. Alternatively, when
time is expressed in years, the equation becomes:

/, = 127(1 -g-'l^<'+0.08-'))_

(3)

The estimated asymptotic size of 127 mm SL is a
reasonable figure. Of 345 C. miliaris examined in
another study (Ralston 1975), 4 were larger than
this size. Of those, three were 131 mm SL or less.

Table l.-The number of rings counted in the otoliths of
Chaetodon miliaria collected around Oahu, Hawaii, 1974.

Date of

Standard

Mean number

Range of

Specimen

capture

length (mm)

of rings

counts

7 June

32-38

18 June

65-74

7 June

48-52

11 July

108

99-115

11 July

133

124-138

1 Oct.

118

110-122

5 Oct.

115

107-121

1 Oct.

138

134-141

5 Oct.

147

141-153

5 Oct.

169

162-178

5 Oct.

227

215-238

5 Oct.

228

219-235

5 Oct.

221

216-227

1 Dec.

322

307-333

1 Dec.

375

362-391

while the fourth was 137 mm SL. Because L^ can
be thought of as an average, if sampling is inten-
sive enough, one would expect to find individuals
of a larger size. Of all the fish sampled in this
earlier study, only 1.2% were larger than the
estimated growth ceiling of the von Bertalanffy
model as determined from the otoliths of the 15
individuals reported on here.

The growth of C. miliaris is very fast. The
estimated growth parameter, K, of the von
Bertalanffy equation describes how quickly
growth proceeds. Large values of K are associated
with rapid growth. Beverton and Holt (1959)
presented values of K for 57 species of fishes and of
those, only 6 species have K values exceeding that
of C. miliaris.

It should also be noted that only fish which were
less than 90 mm SL are reported on here. It was
found that the otoliths of larger fish became
increasingly difficult to read. Not only do the
otoliths become thicker, but the peripheral ring
increments become smaller with growth. For these
reasons, larger fish could not be reliably aged in
this study.

Discussion

On 2 August 1966, Wass (1967) defaunated a
small patch reef in Kaneohe Bay, Hawaii, while
studying the repopulation rates of various species
of fishes. In so doing, he sampled 476 C. miliaris in
1 day. He gave a size-frequency distribution,

992

suitable for the Peterson method of ageing, which
is reproduced in Figure 4.

The first mode centered on 7 cm total length
(TL) could well represent a recently recruited
cohort. A size of 70 mm TL corresponds to a length
of 58 mm SL for C. miliaris (Ralston 1975).
Spawning in this species is known to occur
between December and April but peaks around the
end of February or the beginning of March
(Ralston 1975). Consequently, about 155 days
elapsed between the time of peak spawning for
this species and the date of capture of these 476
specimens. Assuming growth according to Figure
3, after 155 days of growth, juvenile C. miliaris
are estimated to be 55 mm SL. This size corre-
sponds closely with the first mode of Wass' size-
frequency distribution (58 mm SL or 70 mm TL),
thus corroborating Figure 3.

Further evidence in support of the von
Bertalanffy growth curv^e and therefore, the in-
terpretation of otolith ring patterns, comes from
examining the size at which C. miliaris first
reproduce. Ralston (in press) reported that both
male and female C. miliaris reached reproductive
maturity at a size of about 90 mm SL. Referring to
Figure 3, fish of this size are about 1 yr old. If
spawning is periodic, as it is in C. miliaris
(Ralston 1975), one expects the onset of reproduc-
tive maturity to occur after some multiple of the

200-1

180-

•

80-

^ 60

40-

20-

^^TL

3 4 5 6 7 8 9 10 11 12 13 14 15
Total Length (cm)

Figure 4.- Size-frequency distribution of Chaetodon miliaris
collected by Wass (1967) in Kaneohe Bay. (Redrawn from his
figure 7.)

interval between spawning periods has elapsed.
One year is one such interval and C. miliaris
becomes reproductive during the first spawning
season after birth.

Evidence presented here in the form of inter-
pretation of the data of Wass (1967) and exami-
nation of age at maturity substantiate the growth
of C. miliaris as described by the von Bertalanffy
curve of Figure 3. These in turn confirm the
accuracy and utility of employing the diel lamellae
in the otoliths of fishes as growth chronometers.
Although a new and as yet somewhat untried
technique, Pannella's method of age determina-
tion offers the potential to age fishes in situations
where this was not feasible in the past.

Acknowledgments

This research was supported by the Hawaii
Cooperative Fishery Research Unit of the U.S.
Fish and Wildlife Service and by NOAA Oflfice of
Sea Grant, U.S. Department of Commerce, under
grant number 04-5-158-17. I thank Leighton
Taylor for providing the impetus to this study and
Paul Struhsaker for bringing Pannella's work to
my attention. Additional thanks are due Robert
MuUer, Robert Moffitt, James Uchiyama, Ivan Gill,
and Sharon Honda for their efforts extended in my
behalf. This paper is based on a portion of a thesis
submitted in partial fulfillment of requirements
for the M.S. degree at the University of Hawaii,
Department of Zoology.

Literature Cited

Allen, K. R.

1966. A method of fitting growth curves of the von
Bertalanffy type to observed data. J. Fish. Res. Board Can.
23:163-179.
Beverton. R. J. H., AND S. J. Holt.

1959. A review of the lifespans and mortality rates of fish in
nature, and their relation to growth and other phys-
iological characteristics. In G. E. W. Wolstenholme and M.
O'Connor (editors), Ciba Foundation Colloquia on Ageing
5:142-177. J. & A. Churchill Ltd., Lond.
Pannella. G.

1971. Fish otoliths: daily growth layers and periodical
patterns. Science (Wash.. D.C.) 173:1124-1127.

1974. Otolith growth patterns: an aid in age determination
in temperate and tropical fishes. In T. B. Bagenal (editor),
Proceedings of an International Symposium on the Age-
ing of Fish, p. 28-39. Unwin Brothers Ltd., Surrey, Engl.

Pellegrin, D.

1973. Curbs urged in collecting tank fish. Honolulu Adver-
tiser, Sept. 6.
Ralston, S.

1975. Aspects of the age and growth, reproduction, and diet

993

of the millet-seed butterflyfish, Chaetodon miliaris
(Pisces: Chaetodontidae), a Hawaiian endemic. M.S.
Thesis, Univ. Hawaii, Honolulu.
In press. Anomalous growth and reproductive patterns in
populations of Chaetodon miliaria (Pisces: Chaetodon-
tidae) from Kaneohe Bay, Oahu. Pac. Sci.
Randall, J. E.

1973. Marine parks seen as key to reef beauty. Honolulu
Advertiser, Sept. 5.
Reese, E.

1973. Collectors as a threat to reef fishes. Honolulu Star-
Bulletin, May 15.

StRUHSAKER, P., AND J. H. UCHIYAMA.

1976. Age and growth of the nehu, Stolephonta purpureas
(Pisces: Engraulidae) from the Hawaiian Islands as
indicated by daily growth increments of sagittae. Fish.
Bull., U.S. 74:9-17*
Wass, R.

1967. Removal and repopulation of the fishes on an isolated
patch coral reef in Kaneohe Bay, Hawaii. M. S. Thesis,
Univ. Hawaii, Honolulu.

Stephen Ralston

Zmlogij Department
University of Hawaii
Honolulu, HI 96S22
Present address:
College of Fisheries
Unii'ersity of Washington
Seattle, WA 9S195

AN EPIBENTHIC SAMPLER USED TO

STUDY THE ONTOGENY OF VERTICAL

MIGRATION OF RANDALL SJORDANI

(DECAPODA, CARIDEA)'

Pandalus jordani Rathbun, like many other
species of pandalid slirimps, undergo regular diel
changes in their vertical distribution (Tegelberg
and Smith 1957; Alverson et al. 1960; Pearcy 1970,
1972; Robinson in press). Little is known, however,
about the vertical distribution and diel migrations
of larval and juvenile shrimp, or at what stage of
the life history vertical migration and benthic
existence are initiated.

Berkeley (1930) found that size or age of larval
P. danae increased with increasing depth in a
semienclosed embayment in British Columbia.
Pearcy (1972) published the only information on
day/night differences in benthic occurrence of
juvenile P. jordani. Using a plankton net mounted
on a beam trawl, he collected more juveniles (<7.0

'This research was supported in part by Grant No. 04-5-158-2,
Office of Sea Grant, National Oceanic and Atmospheric Admin-
istration, U.S. Department of Commerce.

mm in carapace length) near the bottom during
day than night.

In order to sample the water column completely,
it was necessary to supplement plankton tows
with a discrete, quantitative sample on or just off
the bottom. Various methods have been used for
this purpose but we thought that all of them were
inadequate for the present study. Many epibenthic
samplers do not have an opening/closing device
and therefore are subject to contamination from
the water column above (Russell 1928; Frolander
and Pratt 1962; Pearcy 1972; Beardsley 1973).
Others are only capable of collecting small sam-
ples, in relatively shallow water (Clutter 1965;
Macer 1967). In others the opening/closing device
seems inefficient or overly complex (Bossanyi 1951;
Wickstead 1953; Macer 1967; Hesthagen 1970).
Design criteria for the sampler used in this study
were: a simple, substrate activated, opening/
closing device capable of quantitatively sampling
in depths greater than 150 m and sampling at least
500 m' of water with no loss of filtration efficiency.

Epibenthic Sampler Design

The epibenthic sampler consists of a sled and a
box, to which are attached a plankton net and a
substrate-actuated opening/closing device
(Figure 1). The frame of the sled was welded from
flat steel strap (5.1 x 0.6 cm). The runners
(23 X 0.6 cm mild steel plate) are joined across the
front by a piece of the same steel bent to conform
to the front of the sled. This serves to carry the
sled over small obstructions on the seabed and
further protect the door of the box when it is in the
open position. A bumper bar (5.1 x 0.6 cm) was
also fitted across the front of the sled to prevent
large obstacles from entering the mouth of the
sampler. Two brackets on either side of the sled
serve as attachment points for the box. The six
different positions allow the box to be positioned
from 2.5 to 22.9 cm off the bottom. Two pieces of
strap (5.1 X 0.6 cm) were welded along the top of
the frame with nine holes to provide various
attachment points for the towing bridle. In addi-
tion, four pairs of towing points were placed
around the front of the frame.

The box (106.7 x 45.7 x 53.3 cm), made of 3.2-
mm mild steel plate, is reinforced in front by steel
strap (2.5 x 0.32 cm), forming a lip around the
mouth of the box (Figure IB). The box is further
reinforced by L stock (2.5 x 0.32 cm) placed
around the box 10 cm from the rear edge. Attach-

994

10 cm

Figure l.-Opening/closing epibenthic sampler: A) sled frame;
B) box with door closed; C) detail, side view of shoe, hinge, and
shoe adjustment device; D) box with door open showing
flowmeter and springs for closing door; E) schematic net
attachment, solid line is box wall, two cross hatched lines are
collars of coarse mesh liner (inner) and plankton net (outer), open
bars are stainless steel straps with bolts; F) safety collar insert
with rings for cable attachment protruding through collar and
PVC cod end with threaded teflon plug; G) schematic lateral view
showing sled, box, net, and canvas chafing gear.

ment points for affixing the box to the sled were
made from 3.8-cm round stock, tapped to 9.5 mm
and reinforced with 5.1 x 0.48 cm flat stock. The
box is fastened to the sled by four stainless steel
bolts (0.95 X 3.8 cm).

The door of the box was made from mild steel
plate (109.2 x 48.3 x 0.48 cm) and is hinged with a
6.4-mm stainless steel rod at five points along the
bottom. The shoe which opens the door upon
contact with the sea floor is triangular shaped
(33.0 X 50.8 X 0.48 cm) and is hinged to allow
adjustment, depending on the distance the box is
set off the bottom (Figure IC). Four large springs
(5.1 X 22.9 cm), attached internally, pull the door

shut when the sled leaves the sea floor (Figure ID).
The door-to-shoe surface area ratio is about 5:1, so
that water pressure effectively holds the door shut
on descent and ascent (Figure IB). A TSK
flowmeter^ is mounted in the middle of the mouth
by a brace (1.9 x 0.48 cm). The nets are attached to
the rear of the box by sandwiching them between
stainless steel straps (5.1 x 0.48 cm) bolted
together at 7.6-cm intervals (Figure IE). The inner
strap has 6.4-mm stainless steel bolts welded to it,
while the outer strap has holes drilled to corre-
spond to the bolts in the inner strap, as well as the
holes in the box and net collars. The entire sled,
except for the springs and stainless steel fittings,
was hot dipped galvanized.

The plankton net was made of 571-jum mesh
nylon monofilament. The filtering area to mouth
area ratio is 9:1. The "cylinder'V'cone" net had a
total mesh area of 7.7 m^, with 2.6 m^ in the cone
and 5.1 m- in the cylinder. The collars were made
of plastic coated nylon webbing. The cod end is a
30.5-cm piece of 10.2 cm outside diameter schedule
80 polyvinyl chloride (PVC) pipe, with a threaded
teflon plug for removing the sample. There is also
a stainless steel insert above the cod end fitted
with two rings protruding through slits in the
collar, for attachment of safety wires from the
sled frame to the cod end, in the event a large
amount of sediment was retained (Figure IF).
Overall length of the net including collars and cod
end is 5.1 m (Figure IG). A small coarse mesh net 1
m deep (2.5-cm stretched mesh) was mounted
inside the plankton net (see Figure IE) to catch
any large animals or objects and prevent them
from damaging the plankton net or the sample in
the cod end. A piece of heavy canvas (1.2 x 3.7 m,
no. 4 duck) was attached to the rear of the sled by
shackles, to protect the plankton net from chafing
on the sea floor (Figure IG).

Epibenthic Sampler Operation

Because of its size and weight (ca. 150 kg in air)
the epibenthic sampler can only be used from a
vessel with a suitable trawl winch; in the present
study the 24.4-m RV Cayuse with a 9.5-mm
diameter trawl wire was used. The sled was fas-
tened to the trawl wire with a ball bearing swivel
and a 3-m bridle of 9.5-mm wire attached to the

2Tsurumi-Seiki Kosakusho. Reference to trade names does not
imply endorsement by the National Marine Fishenes Service,
NOAA.

995

second set of attachment points from the front of
the sled. From these towing points the sled proved
to be stable, never landing on the bottom upside
down. It could be launched and recovered by two
people in moderate-to-rough seas. The sled was
launched while underway at 2 knots. A 20-min tow
(bottom time) at 2 knots was calculated to be
sufficient to filter 500 m' of water. Presence on the
bottom was detected by the winch potentiometer.

Filtering efficiency of the sampler, based on the
degree of clogging, was never markedly reduced
over this time interval. However on soft muddy-
sand bottoms, characteristic of P. Jordan i grounds,
bottom times were reduced to 10 min because of
the amount of fine sediment and meiofauna
stirred up and retained in the cod end. Large
organisms, including adult P. jordani, as well as
the fragile urchin, Alloceutrotus fragilis, and
slender sole, Lyopsetfa exilis, were effectively
retained in the coarse mesh liner and prevented
from reaching the sample in the cod end. Flow-
meter readings indicated that at no time did the
number of animals retained in the liner seriously
occlude the mouth of the plankton net and affect
its efficiency. On coarse sand bottoms, the samples
were very clean, with little sediment and
meiofauna retained, even when the net was only
5-8 cm off the bottom.

Though the sled was never observed firsthand
while on the bottom, evidence from skid marks on
the runners and shoe, behavior of the poten-
tiometer while the sled was on the bottom, and the
relationship between flowmeter readings and
bottom time indicated that the epibenthic sampler
was stable and not prone to dig in or bounce off the
bottom while being towed.

Vertical Distribution of
Larval Pandalus jordani

On 30 and 31 May 1972 the epibenthic sampler
was used to sample near-bottom fauna and open
bongo nets were used to obtain a series of quasi-
vertically stratified plankton samples 10 nautical
miles off Cascade Head, on the central Oregon
coast (lat. 45°04.0'N, long. 124°15.rW). The 0.7-m
diameter bongo frames had paired cylidner/cone
571-/xm Nitex nets, 5.1 m in length with an
effective filtering area to mouth area ratio of 8:1. A
scope to depth ratio of 2:1 was maintained by using
a 40-kg multiplane kite otter as a wire depressor
(Colton 1959). All nets contained TSK flowmeters.
A time-depth recorder was fixed to the wire just

above the bongo nets. Tows were made at four
strata (0-10, 11-50, 51-100, 101-150 m) with the
open bongo nets, and a bottom sample was taken
with the epibenthic sampler at 160 m. Replicate
tows were taken at each depth interval, both day
(1200-1930 h) and night (2105-0400 h). Contamina-
tion in the open bongo net was minimized by
lowering to the depth interval as fast as possible,
doing a stepped oblique tow through the horizon,
and then raising the net as quickly as possible.
Towing time at depth was long enough to keep the
period of contamination below 20% of the total
sampling time for the deepest tows.

The vertical distribution of P. jordani larvae
and juveniles is summarized in Figure 2. During

' , , w , w

300-

200-

100-

DAY

NIGHT

0 lOm

n-50m

51-IOOm

101 150m

' I 1 I I I I

BOTTOM (160m)

Q TOW 1
Q TOW 2

VI VIII X XII JUV VI VIII X XII JUV

LARVAL STAGE

Figure 2.-Vertical distribution of lar\ae and early juvenile
Pandalus jordani, during one day and one night period. All tows
were replicated.

996

the day, larvae were distributed throughout the
water column and were most abundant in the 0- to
10-m depth interval. A trend of increasing age
with depth was evident. Early juveniles were
present in low numbers in the 51- to 100-m and
101- to 150-m intervals. The sled tows revealed a
very high concentration of early juveniles (284 and
290/1,000 m^) on the bottom during midday.

At night larval shrimp were still distributed
throughout the entire water column. The younger
stages (V and VI), found in some abundance in the
0- to 10-m interval during the day, were not
collected at night. Furthermore, an age gradient
with depth was no longer present. This was due, in
part, to the presence of late larvae at all depths in
the water column. The most dramatic feature of
the night distribution was the vertical migration
of the early juveniles as indicated by their virtual
absence on the bottom (0 and 4/1,000 m-') in the
sled samples. Juveniles were again present in the
lower portion of the water column (101-150 m) and
had migrated into the upper 100 m, including the
top 10 m. There was no evidence that larvae
younger than Stage XIII migrated to any extent.
Vertical migratory behavior starts late in the
larval phase, before the molt to juvenile and
recruitment to the bottom.

Acknowledgments

We thank R. Mesecar for suggestions on the
design and T. Nolan for the fabrication of the
epibenthic sampling device. C. B. Miller gave
advice on aspects of the research and critically
read an early draft of the manuscript. W. T.
Peterson, D. 0. Elvin, B. Sullivan, the captain and
crew of the RV Cayuse were patient and helpful
during the sea trials and sampling.

Literature Cited

Alverson, D. L., R. L. McNeely, and H. C. Johnson.

1960. Results of exploratory shrimp fishing off Washington
and Oregon (1958). Commer. Fish. Rev. 22(1):1-11.
Beardsley, A. J.

1973. Design and evaluation of a sampler for measuring the
near-bottom vertical distribution of pink shrimp, Pan-
dalusjordani. Fish. Bull., U.S. 71:243-253.
Berkeley, A. A.

1930. The post-embryonic development of the common
pandalids of British Columbia. Contrib. Can. Biol.
6:79-163.

BOSSANYI,J.

1951. An apparatus for the collection of plankton in the

immediate vicinity of the sea-bottom. J. Mar. Biol. Assoc.
U.K. 30:265-270.
Clutter, R. I.

1965. Self-closing device for sampling plankton near the sea
bottom. Limnol. Oceanogr. 10:293-296.
CoLTON, J. B., Jr.

1959. The multiplane kite-otter as a depressor for high-
speed plankton samples. J. Cons. 25:29-35.
Frolander, H. F., AND I. Pratt.

1962. A bottom skimmer. Limnol. Oceanogr. 7:104-106.
Hesthagen, I. H.

1970. On the near-bottom plankton and benthic inverte-
brate fauna of the Josephine Seamount and the Great
Meteor Seamount. METEOR Forschungsergeb., Reihe D.
8:61-70.
Macer, C. T.

1967. A new bottom-plankton sampler. J. Cons. 31:158-163.
Pearcy. W. G.

1970. Vertical migration of the ocean pink shrimp Pandalus
Jordan i: A feeding and dispersal mechanism. Calif. Fish
Game 56:125-129.
1972. Distribution and diel changes in the behavior of pink
shrimp, Pandah<s jurdani, off Oregon. Proc. Natl.
Shellfish. Assoc. 62:15-20.
Robinson, J. G.

In press. The vertical distribution and diel migration of
pink shrimp {Pandalux jordani) off Oregon. Oreg. Fish
Comm. Invest. Rep.
Russell, F. S.

1928. A net for catching plankton near the bottom. J. Mar.
Biol. Assoc. U.K. 15:105-108.
Tegelberg, H. C, and J. M. Smith.

1957. Observations on the distribution and biology of the
pink shrimp (Pandalun jordani) off the Washington
coast. Wash. Dep. Fish., Fish. Res. Pap. 2(l):25-35.

WlCKSTEAD, J.

1953. A new apparatus for the collection of bottom
plankton. J. Mar. Biol. Assoc. U.K. 32:347-355.

Peter C. Rothlisberg

Marine Science Center

Oregon State University

Newport. OR 97365

Present address:

CSIRO

Division of Fisheries and Oceanography

North Eastern Regional Laboratory

Cleveland, Queensland, il63, Australia

School of Oceanography
Oregon State University
Corvallis, OR 97331

William G. Pearcy

997

HARNESS FOR ATTACHMENT OF

AN ULTRASONIC TRANSMITTER TO

THE RED DRUM, SCIAENOPS OCELLATA

The use of small ultrasonic transmitters for
studying the movement and behavior of fish in the
field is becoming very popular (Stasko 1971). As a
result various methods have been devised for
attaching transmitters to fish either externally or
internally. These methods involve hooking into the
dorsal musculature or insertion into the stomach
(Henderson et al. 1966), surgical implantation into
the peritoneal cavity (Hart and Summerfelt 1975),
and others (Ohsumi 1969). The suitability of a
procedure is dependent on the species of fish and
on the particular objective of the study. For
studies we are initiating on movements of the red
drum, Sciaenopii ocellafa, none of the existent
•procedures were found to be entirely satisfactory.
This note describes a simple inelastic harness we
have developed for the external attachment of an
ultrasonic transmitter to the caudal peduncle. This
attachment method is markedly superior to other
methods we have tried with the red drum. We
believe this procedure will be of immediate value
to many workers involved in tracking studies and
therefore we are describing it now rather than
awaiting the completion of our investigation of
migratory movements of the red drum.

Materials and Methods

The inelastic harness for attaching an ultrasonic
transmitter to the caudal peduncle is shown in
Figure 1. The components of the harness are as
follows:

1. An inelastic plastic pull-tie (5 x 190 mm) of
sufficient length to encircle the caudal peduncle;

2. Sections of soft Tygon' tubing (6-mm OD)
and soft rubber tubing (12-mm OD, 1.5-mm wall
thickness) threaded over the pull-tie to provide a
soft flat cushion that minimizes abrasions and
chafing to the fish when the pull-tie is attached
and tightened;

3. Small plastic pull-ties to firmly affix the
transmitter to the large pull-tie and tubing de-
scribed above.

When attaching the harness, the large pull-tie is
tightened just enough such that it fits snugly
around the caudal peduncle and cannot slip over
the tail (Figure 1 inset). Care must be taken not to
tighten the tie so tightly that it compresses the
peduncle. If the latter occurs, the tie must be cut
off with scissors and replaced. These ties can only
be tightened. The final position of the transmitter
itself should be on the dorsal surface of the
peduncle with the axis of the transmitter situated
at a right angle to the longitudinal axis of the fish.
After attachment the overlapping section of the
pull-tie is cut off.

Harnesses are preconstructed prior to the time
of use such that in the field the only modifications
required are the addition or removal of small
sections of Tygon and rubber tubing to provide a
cushion of the exact size for a particular fish. A
preconstructed harness can be attached to a red
drum in less than 5 min. Plastic pull-ties of various
lengths and widths are available at most hardware
stores that stock materials used bv electricians.

'Reference to trade names does not imply endorsement by the
National Marine Fisheries Ser\Mce, NOAA.

Figure 1. -Inelastic harness for at-
tachment of ultrasonic transmitter to
caudal peduncle of red drum.
Transmitter = Smith Root SR 69.
Total length of harness = 190 mm.
Inset: Red drum (3.2 kg) with harness
and transmitter attached.

998

The red drum used for testing the harness were
caught in the Matanzas Inlet, Fla., by hook and
line and maintained in captivity for approximate-
ly 2 mo prior to testing.

Results and Discussion

Observations of the suitability of the inelastic
harness were conducted in a 3.3-m diameter fiber
glass tank, in an enclosed half-acre pond (max
depth 2.5 m), and in the Intracoastal Waterway
near the Whitney Marine Laboratory. In the fiber
glass tank, two red drum (2.5, 3.5 kg) with har-
nesses and "dummy" transmitters attached swam
normally as soon as released and accepted food of
shrimp and mullet within 30 min. A third red drum
(ca. 3 kg) with harness and active transmitter
(Smith Root SR 69) attached was released in the
half-acre pond. During the 3-wk lifetime of the
batteries in the transmitter, the movements of the
fish were monitored almost daily with a receiver
and hydrophone. The red drum moved actively
about the pond, ate readily, and schooled with
other fish. Mangrove roots, pilings, and other
obstacles in the pond were not snagged by either
the harness or the transmitter. More than 2 mo
after the fish was initially released, the harness
and inactive transmitter remained in place, and
the fish continued to feed and behave normally.

A fourth red drum (3.2 kg) with harness and
active transmitter attached was released into the
Intracoastal Waterway on 12 January 1976 and
tracked continuously for 7 h from a boat with a
74-kHz receiver and hydrophone. The position of
the fish with respect to charted channel markers
was recorded frequently to provide the summary
described below. During the first 1.5 h after
release, the fish moved approximately 1.6 km to
the south of the release point. This movement was
against the direction of the tidal flow. During the
remaining time the fish moved 1.2 km to the north,
again against the direction of the tidal flow.
During this excursion, the fish entered the mouth
of almost every creek encountered. At nightfall
the fish had moved into a deep hole approximately
140 m up a small creek situated 400 m from the
original release point. The fish was not located on
the second day but on the third day was located at
the edge of the main channel of the Intracoastal
Waterway approximately 2 km to the south of the
release point. Tracking of the fish had to be
discontinued due to a malfunction in the receiver.

For the studies we are initiating on migratory
movements of the red drum, the method selected
for the attachment of the transmitter was ex-
tremely important, and we spent considerable
time trying alternative methods. These methods
included the hooking of saddles into either the
dorsal or ventral musculature, surgical attachment
to the pectoral girdle or to the lower jaw bone,
surgical implantation into the peritoneal cavity,
and insertion into the stomach. Utilization of the
inelastic harness provided the following advan-
tages over the other methods we tried.

1. The attachment procedure is simple and
quick enough such that only a few minutes elapse
between the time the fish is caught, tagged, and
released.

2. The procedure results in no bleeding and
causes minimal trauma, damage, or weakening of
the fish.

3. The attachment is secure and assures that the
transmitter remains attached to the red drum for
the lifetime of the transmitters we are using
(Smith Root SR 69 and SR 69A, lifetimes of 20 and
45 days).

Ichihara (1971) described a "saddle type" meth-
od for affixing a transmitter to a fish. This method
employed an elastic strap of neoprene rubber that
encircled the fish anterior to the dorsal fin. The
author noted that fish with elastic harnesses of the
saddle type died within 9 to 30 days. Regarding the
above observations, we have also found that rub-
ber elastic harnesses encircling the caudal pedun-
cle are unsatisfactory because they constantly
compress the peduncle and result in a progressive
deterioration of the entire tail region. However,
our inelastic harness causes no such deleterious
effects. Although we have experimented with the
inelastic harness on the red drum only, we are
certain that it can be used with any large fish
having a fairly rigid tail that is markedly broader
than the caudal peduncle.

Acknowledgments

We are grateful to Jack R. Smith, Department
of Electrical Engineering, University of Florida,
for providing the hydrophone and 74-kHz receiver
used in the current study. We also thank Marine-
land, Inc., for permitting us to use their saltwater
pond and other facilities.

999

Literature Cited

Hart, L. G., and R. C. Summerfelt.

1975. Surgical procedures for implanting ultrasonic trans-
mitters into flathead catfish (Puloclicfis oliraris). Trans.
Am. Fish. See. 104:56-59.
Henderson, H. F., A. D. Hasler, and G. G. Chipman.

1966. An ultrasonic transmitter for use in studies of
movements of fishes. Trans. Am. Fish. Soc. 95:350-356.
Ichihara, T.

1971. Ultrasonic, radio tags and various problems in fixing
them to marine animal body. Suisancho. Suisan Kenkyu-
sho Gyogyo Shigen Kenkyu Kaigido. 12:29-44 (Trans-
lated by Transl. Bur, Foreign Lang. Div., Fish. Res. Board
Can., St. Andrews, N.B., Transl. 1981, 38 p.)
Ohsumi, S.

1969. How to attach the telemetry equipment to marine
life. Kaiyoseibutsu Telem. Kenkyu Kaiho 2:32-36.
Translated by Transl. Bur., Foreign Lang. Div., Fish. Res.
Board Can., St. Andrews, N.B., Transl. 1929, 11 p.)
Strasko, a. B.

1971. Review of field studies on fi.sh orientation. Ann. N.Y.
Acad. Sci. 188:12-29.

William E. S. Carr
Thomas B. Chaney

Whitnei/ Marine Lahoratorj/

Un i re rsit II of Florida at Marineland

Route 1. Box 121

St. Auqustitie. FL -UdSj,

1000

INDEX

Fishery Bulletin Vol. 74, No. 1-4, 1976

ABLE, K. W., and J. A. MUSICK, "Life history, ecology,
and behavior of Liparin inquilinus (Pisces: Cyclopter-
idae) associated with the sea scallop, Placopecten magel-
lanicus"

"Abundance of macrocrustaceans in a natural marsh and
a marsh altered by dredging, bulkheading, and filling," by
Lee Trent, Edward J. PuUen, and Raphael Proctor

ACKER, WILLIAM C.-see LORD et al.

"(An) acoustic method for the high-seas assessment of
migrating salmon," by Gary Lord, William C. Acker,
Allan C. Hartt, and Brian J. Rothschild

Actiniaria
continental shelf and slope, U.S. east coast

Actinauge verrilli

Acfinoniola callosa

Amphianthus nitidus

Antholoha perdix

Bolocera tuediae

Edwardsia sulcata

Halcampa duodecimcirrata

Halipla nella luciae

Haloclava producta

Hormathia nodosa

Hormathia nodosai?)

Metridium senile Jimbriatum

Paranthus rapiformis

Pea<:hia parasitica

Phelliactis americana

Sagartiogeton verrilli

Stephanauge nexilis

StephanaugeC) spongicola

Stomphia coccinea

Tealia crassicorn is

"Age and growth of the nehu, Stolephorus purpureus
(Pisces: Engraulidae), from the Hawaiian Islands as
indicated by daily growth increments of sagittae," by
Paul Struhsaker and James H. Uchiyama

Age determination

butterflyfish, tropical reef

otoliths, daily growth rings

larva! and adult fishes

daily growth increments in otoliths

nehu, Hawaiian Islands

daily growth increments of sagittae

"Age determination of a tropical reef butterflyfish
utilizing daily growth rings of otoliths," by Stephen
Ralston

AHLSTROM, ELBERT H.-see BUTLER and AHLSTROM

Alaska, southeastern
Sashin Creek
salmon, coho, life history

409

195

104

872
865
873
869
863
860
862
870
863
871
872
870
868
863
872
870
875
875
868
865

990
1
9

990

Albacore-see Tuna, albacore

Alopias superciliosus-see Shark, bigeye thresher

Alosa sapidissima—see Shad, American

ALVAREZ, JOSE, CHRIS 0. ANDREW, and FRED J.
PROCHASKA, "Dual structural equilibrium in the
Florida shrimp processing industry"

"Analysis of returns of tagged Gulf menhadenj' by Paul
J. Pristas, Eldon J. Levi, and Robert L. Dryfoos

Anchovy, northern
'^C-benzene

uptake, distribution, and depuration

California Current

larvae, food and feeding

fishery, development, and example application of

simulation model

growth of laboratory-reared in Southern California

comparisons

growth curve

hatching to juvenile

juvenile to adult

weight-length relation

larvae, culture and growth

laboratory cultured foods

survival at metamorphosis

swim bladder inflation, diel changes

adaptive advantages

diel rhythm

larval length

mechanism

relation between sinking speed, swim bladder

volume, and larval length

vertical migration

thermal tolerance and resistance

development and growth

embryos and larvae

juveniles and adults

ANDREW, CHRIS O.-see ALVAREZ et al.

Anopk)pomaJimbria-see Sablefish

Apalachicola Bay, Florida
epibenthic fish and invertebrate populations

long-term fluctuations

"(The) application of systematic sampling to a study of
infauna variation in a soft substrate environment," by
Stephen Scherba, Jr. and Vincent F. Gallucci

ARTHUR DAVID K., "Food and feeding of lar\ae of
three fishes occurring in the California Current, Sardin-
ops sagax, Engraulis mordax, and Trachurus symmet-
ricus" •

897 "Aspects of the reproductive biology of the weakfish,

879
112

545

517

118

277
274
272
273
276

81
85

854
849
850
852

850
853

440
435
434

311

937

517

1001

Cynoscion regalia (Sciaenidae), in North Carolina;' by
John V. Merriner

Atlantic, North
calanoid copepods
caloric values of some

Atlantic, western North

bigeye thresher shark, observations on

fishing effort estimation from aerial search data. . .

Atlantic Ocean
seatrouts
protein taxonomy

Auke Creek, Alaska
salmon, pink
production of fry and adults from gravel incubators
and natural spawning

BAILEY, JACK E., JEROME J. PELLA, and SIDNEY G.
TAYLOR, "Production of fry and adults of the 1972 brood
of pink salmon, Oncnrhpichus gnrhui^cha. from gravel
incubators and natural spawning at Auke Creek, Alaska"

Barnacles, goose
La Jolla, California
on beached flotsam

Bass, striped
'^C-benzene

uptake, distribution, and depuration

benzene eff'ects of

caloric content

fat content

growth

Bathynecfes superbus
epizoites associated with

BELL, JONATHAN-see HANSON and BELL

BENVILLE, PETE, JR.-see KORN et al.

Benzene

bass, striped
effects on growth, fat content, and caloric content .

"Biology of five species of searobins (Pisces, Triglidae)
from the northeastern Gulf of Mexico," by Thomas C.
Lewis and Ralph W. Yerger

BISSON, PETER A., and GERALD E. DAVIS,
"Production of juvenile chinook salmon, Oncorhynchus
tshawytscha, in a heated model stream."

Bivalve variation
application of systematic sampling to study of, in soft
substrate environment

BLACKBURN, MAURICE, and WALTER NELLEN,
"Distribution and ecology of pelagic fishes studied from
eggs and larvae in an upwelling area off Spanish Sahara"

BOND, CARL E.-see CRONE and BOND

BRAY, RICHARD N.-see EBELING and BRAY

BREDER, CHARLES M., JR., "Fish schools as opera-
tional structures"

1002

218

221
503

599

961

212

545

694
694
694

225

694

763

944

885

Brevoortia patronus-see Menhaden, Gulf

Brevoortia tyrannus-see Menhaden, Atlantic

BREWER, GARY D., "Thermal tolerance and resistance
and resistance of the northern anchovy, Engraulis
niordax"

BRINTON, EDWARD, "Population biology of Euphau-
sia pacifica off southern California"

Bristol Bay, Alaska
salmon, sockeye
foods of juvenile in inshore coastal waters, 1966-67.

BROTHERS, EDWARD B., CHROSTOPHER P. MATH-
EWS, AND REUBEN LASKER, "Daily growth in-
crements in otoliths from larval and adult fishes"

BROWN, JOHN C.-see HEWITT et al.

BRUSHER, HAROLD A., and LARRY H. OGREN,
"Distribution, abundance, and size of penaeid shrimps in
the St. Andrew Bay system, Florida"

Buoy, acoustic
used for high-seas assessment of migrating salmon . .

BUTLER, JOHN L., and ELBERT H. AHLSTROM,
"Review of the deep-sea fish genus Scopelengys
(Neoscopelidae) with a description of a new species
Scopelengys clarkei, from the central Pacific"

Butterflyfish, millet-seed
age determination using daily growth rings of otoliths

CAILLOUET, CHARLES W.-see LANSFORD et al.

California, southern

DDT and metabolites in sediments off

Euphauftia pacifica
population biology

California Current
sonar mapping
development and use for pelagic stock assessment .

California Current, central region
zooplankton and euphausiid populations
density, vertical range, and diel movement

471

Callinectes sapidus-see Crab, blue

"Caloric values of some North Atlantic calanoid
copepods," by Geoffrey C. Laurence

Cancer irrorafus-see Crab, rock

Cancer magister—see Crab, Dungeness

CARLSON, H. RICHARD, "Foods of juvenile sockeye
salmon, Oncorhynchus nerka, in the inshore coastal
waters of Bristol Bay, Alaska, 1966-67"

CARR, WILLIAM E. S., and THOMAS B. CHANEY,
"Harness for attachment of an ultrasonic transmitter to
the red drum, Sciaenops ocellata"

CASEY, JOHN G.-see STILLWELL and CASEY

Ceriantharia
continental shelf and slope, U.S. east coast

433

733

458

158
104

142
990

27
733

281

925

218

458

998

Ceriantheopsis americanus

Cerianthus borealis

"Ceriantharia, Zoanthidea,Corallimorpharia, and Actin-
iaria from the continental shelf and slope off the eastern
coast of the United States," by Bernt Widersten

Chaetodon mUiaris-see Butterflyfish, millet-seed

CHANEY, THOMAS B.-see CARR and CHANEY

CHENG, LANNA, and RALPH A. LEWIN, "Goose
barnacles (Cirripedia: Thoracica) on flotsam beached at
La Jolla, California"

Chesapeake Bight
galatheid crustaceans
occurrence of two in

CHESS, JAMES R.-see HOBSON and CHESS

CHITTENDEN, MARK E., JR., "Present and historical
spawning grounds and nurseries of American shad, Alosa
sapidissima, in the Delaware River"

, "Weight loss, mortality, feeding, and dura-
tion of residence of adult American shad, Alosa sapidis-
sima, in fresh water"

Citharichthys sordidussee Sanddab, Pacific

CLARKE. THOMAS A., and PATRICIA J. WAGNER,
"Vertical distribution and other aspects of the ecology of
certain mesopelagic fishes taken near Hawaii"

Clupea harengus pallasi—see Herring, Pacific

COE, JAMES M.-see PERRIN et al.

COLLINS, JEFF, "Oil and grease: A proposed analytical
method for fishery waste effluents"

and RICHARD D. TENNY, "Fishery waste

effluents: A method to determine relationships between
chemical oxygen demand and residue"

"Comparison of the most successful and least successful
west coast albacore troll fishermen" by Donald F. Keene
and William G. Pearcy

"Contribution of the net plankton and nannoplankton to
the standing stocks and primary productivity in Monte-
rey Bay, California during the upwelling season;' by
David L. Garrison

Copepods, calanoid
North Atlantic Ocean
caloric values of some

Corallimorpharia
continental shelf and slope, U.S. east coast

Corynactis delawarei

Cottiis aleuticus
populations of sympatric in four adjacent salmon-
producing coastal streams on Vancouver Island

Coitus asper
populations of sympatric in four adjacent salmon-
producing coastal streams on Vancouver Island

858
857

857

212

462

343

151

635

681

725

973

183

218

858

131

Crab

North American fisheries regulations and their
rationales
procedures for changing laws and regulations 630

Crab, blue
east coast

fishery regulations and their rationales 628

West Bay, Texas
abundance in natural and altered marshes 195

Crab, Dungeness

eggs
mortalities and epibiotic fouling of, from wild
populations 201

larval dynamics off central Oregon coast, 1970-71

abundance 357

climate 357

distribution 357

gut-fullness analysis 364

hydrographic features 353

population analyses, larval 359

sampling variability 356

west coast
fishery regulations and their rationales 626

Crab, fiddler
mercury, cadmium, and lead salts
regeneration and ecdysis, effects on 464

Crab, king
Alaska
fishery regulations and their rationales 624

Crab, rock
size composition and growth of young in Maine

growth 952

sex ratio 951

size composition and seasonal distribution .' 949

Crab, snow
Alaska

fishery regulations and their rationales 625

eastern Canada

fishery regulations and their rationales 627

electrophoretic evidence of hybrid 693

Crab, stone
Florida
fishery regulations and their rationales 630

Crab fisheries

regulations and their rationales

Alaska king ^24

Alaska snow "-^

blue, east coast ^^8

fi?7
Canada, eastern snow

Dungeness, west coast

Florida stone ^^^

CRADDOCK, DONOVAN R., "Effects of increased water
temperature on Daphnia pulex ^"^

Crassostrea gigas-see Oyster, Pacific

CRAWFORD, L., and M. J. KRETSCH, "Effects of cook-
ing in air or in nitrogen on the development of fishy flavor

1003

in the breast meat of turkeys fed tuna oil with and
without a-tocopherol supplement or injection"

Croaker, white
seasonal spawning cycle

CRONE, RICHARD A., and CARL E. BOND, "Life
history of coho salmon, Oncorluinchua kisxtch, in Sashin
Creek, southeastern Alaska"

Crustaceans
restraining living planktonic

"Culture and growth of northern anchovy, Engraulis
mordax, larvae!' by John R. Hunter

Cynxocion mcu;donaldi-see Totoaba

Cynoncion regalis-see Weakfish

DDD

in sediments off southern California

DDE

in sediments off southern California

DDT

and metabolites in sediments off southern California.

"DDT and its metabolites in the sediments off southern
California," by John S. MacGregor

"Daily growth increments in otoliths from larval and
adult fishes," by Edward B. Brothers, Christopher P.
Mathews, and Reuben Lasker

Daphnia pulex
water temperature, effects of increased

discharges of heated water

water passing through cooling systems

DAVIS, GERALD E.-see BISSON and DAVIS

"Day versus night activity of reef fishes in a kelp forest
off Santa Barbara, California!' by Alfred W. Ebeling and
Richard N. Bray

Decision theory
salmon fishery
data acquisition and management

"Decision theory applied to the simulated data acquisi-
tion and management of a salmon fishery!' by Gary E.
Lord

Delaware River
shad, American
present and historical spawning grounds and nurs-
eries

"Description of zoeae of coonstripe shrimp, Pandalns
hypifinotu!^, reared in the laboratory!' by Evan Haynes

"Development and example application of a simulation
model of the northern anchovy fishery!' by Michael F.
Tillman and Donald Stadelman

"Development and use of sonar mapping for pelagic stock
assessment in the California Current area!' by Roger P.
Hewitt, Paul E. Smith, and John C. Brown

1004

89
983

897
220

405
406

703

837

343

323

118

281

Diatoms
freshwater and estuarine, benthic
menhaden, Atlantic, grazing by adult 689

"Diel changes in swim bladder inflation of the larvae of
the northern anchovy, Engraulis mordaj:',' by John R.
Hunter and Carol Sanchez 847

"Distribution, abundance, and size of penaeid shrimps in
the St. Andrew Bay system, Florida!' by Harold A.
Brusher and Larry H. Ogren 158

"Distribution and ecology of pelagic fishes studied from
eggs and larvae in an upwelling area off Spanish Sahara!'
by Maurice Blackburn and Walter Nellen 885

"Distribution, food, and feeding of the threespine stick-
leback, Gastero^teux aculeatus, in Great Central Lake,
Vancouver Island, with comments on competition for
food with juvenile sockeye salmon, Oncorhynchus nerka"
by J. L Manzer 647

DOTSON, RONALD C, "Minimum swimming speed of
albacore, Tkunnus alalunga" 955

Drum, red
ultrasonic transmitter, harness for attachment of . . . 998

DRYFOOS, ROBERT L.-see PRISTAS et al.

"Dual structural equilibrium in the Florida shrimp
processing industry!' by Jose Alvarez, Chris 0. Andrew,
and Fred J. Prochaska 879

EBELING, ALFRED W., and RICHARD N. BRAY,
"Day versus night activity of reef fishes in a kelp forest
off Santa Barbara, California" 703

"Ecology of Hawaiian sergestid shrimps (Penaeidea:
Sergesti'dae)!' by John F. Walters 799

"Economic and financial analysis of increasing costs in
the Gulf shrimp fleet!' by Wade L. Griffin, Newton J.
Wardlaw, and John P. Nichols 301

EDGAR, ROBERT K., and JAMES G. HOFF, "Grazing
of freshwater and estuarine, benthic diatoms by adult
Atlantic menhaden, Brevoortia tyrannus" 689

"Effects of benzene on growth, fat content, and caloric
content of striped bass, Morone saxatilis," by Sid Korn,
Jeannette W. Struhsaker, and Pete Benville, Jr 694

"Effects of cooking in air or in nitrogen on the develop-
ment of fishy flavor in the breast meat of turkeys fed
tuna oil with and without a-tocopherol supplement or
injection!' by L. Crawford and M. J. Kretsch 89

"Effects of increased water temperature on Daphnia
pulex" by Donovan R. Craddock 403

"Effects of mercury, cadmium, and lead salts on regener-
ation and ecdysis in the fiddler crab, Uca pugilator" by
Judith S. Weis 464

"Effects of temperature and salinity on the survival of
winter flounder embryos!' by Carolyn A. Rogers 52

Eflluents - see Fishery waste effluents

Eggs, fish— see Fish eggs

"Electrophoretic evidence of hybrid snow crab, Chio-
nocetes bairdi x opilio" by Allyn G. Johnson 693

"(An) energetics model for the exploited yellowfin tuna,
Thunnus albacares, population in the eastern Pacific
Oceani' by Gary D. Sharp and Robert C. Francis 36

Engraulis mordax-see Anchovy, northern

Eopsetta jordani-see Sole, petrale

"(An) epibenthic sampler used to study the ontogeny of
vertical migration of Pandalus jordani (Decapoda,
Caridea);' by Peter C. Rothlisberg and William G. Pearcy 994

Epizoites
associated with Bathynectes superbus 225

"Epizoites associated with Bathynectes superbus
(Decapoda: Portunidae)," by Elizabeth G. Lewis 225

"Estimates of rates of tag shedding by North Pacific
albacore, Thunnus alalunga',' by R. Michael Laurs,
William H. Lenarz, and Robert N. Nishimoto 675

"Estimation of fishing effort in the western North
Atlantic from aerial search data;' by M. L. Parrack .... 503

Euphausia pacifica
population biology off southern California

annual biomass 752

growth 745

recruitment efficiency and spatial aggregation of

eggs 743

sex ratio 755

southern California eddy 736

spawning and recruitment 738

survivorship 749

temperature relationships of spawners and larvae . 744

EXiphausiids
California Current, central region
density, vertical range, and diel movement 925

Euthyuiixs lineaf US-see Skipjack, black

"Feeding behavior, food consumption, growth, and
respiration of the squid Loligo opalescens raised in the
laboratory," by Ann C. Hurley 176

"Fertilization method quantifying gamete concentra-
tions and maximizing larvae production in Crassostrea
gigas" by William H. Staeger and Howard F. Horton . . 698

"First record of the melon-headed whale, Peponocephala
electra, in the eastern Pacific, with a summary of world
distribution!' by William F. Perrin 457

Fish, adult
daily growth increments in otoliths 1

Fish, epibenthic

long-term fluctuations of populations in Apalachicola

Bay, Florida

distribution 313

physicochemical parameters 313

seasonal fluctuations of dominant species 314

Fish, mesopelagic

vertical distribution and other aspects of ecology near

Hawaii

avoidance 643

migration 641

sex ratio 643

sexual dimorphism 643

Fish, pelagic
eggs and larvae, distribution and ecology off Spanish
Sahara

anchovy eggs, spatial and temporal distributions . . 893

identification and enumeration 888

sardine eggs, spatial and temporal distributions . . . 893

spatial distribution 891

temperature and chlorophyll a 888

zooplankton 886

Fish, trophic interactions

activity patterns of planktivorous fishes 580

collecting fishes 570

collecting zooplankters 569

fishes studied 568

zooplankter activity patterns 571

zooplankter volumes 571

Fish catches
gill nets
relation to frontal periods 449

Fish eggs
Spanish Sahara
distribution and ecology of pelagic in an upwelling
area off 885

Fish fauna

associated with offshore platforms in northeastern

Gulf of Mexico

comparison of two platforms 395

faunal composition 389

habitat occupation and activity patterns 396

winter-summer contrast 395

Fish larvae

anchovy, northern

diel changes in swim bladder inflation 847

California Current

anchovy, northern 517

mackerel, jack 517

sardine. Pacific 517

daily growth increments in otoliths 1

herring. Pacific
predator-prey relationship between, and Hyperoche

medusaru m ^^

northern anchovy, culture and growth 81

Spanish Sahara
distribution and ecology of pelagic in an upwelling
area off ^^

Fish schools
as operational structures

geometrical models ^'^

locomotor problems ^^^

movements

shape ^^"^

siz^ 482

476
spacing

1005

traffic problems 486

turning problems 484

"Fish schools as operational structures;' by Charles M.
Breder, Jr 471

FISHER. WILLIAM S.. and DANIEL E. WICKHAM,
"Mortalities and epibiotic fouling of eggs from wild
populations of the Dungeness crab, Cancer magister" . . 201

Fishery waste effluents
chemical oxygen demand and residue

method to determine relationships between 725

oil and grease

proposed analytical method 681

"Fishery waste effluents: A method to determine rela-
tionships between chemical oxygen demand and residue;'
by Jeff Collins and Richard D. Tenny 725

Fishes, reef
Santa Barbara, California
day versus night activity in a kelp forest 703

Fishing effort estimation
Atlantic, western north

aerial search data 503

FLANAGAN, CHRISTINE A., and JOHN R. HEN-
DRICKSON, "Observations on the commercial fishery
and reproductive biology of the totoaba, Cynoscion
maccbnaldi, in the northern Gulf of California" 531

Flatfish
Oregon, continental shelf
food of five species of cooccurring 984

Florida
shrimp processing industry
dual structural equilibrium 880

Flounder, winter
embryos, effects of temperature and salinity on
survival

embryonic development 56

incubation time and duration of hatching interval . 55

influence on total and viable hatch 53

"Food and feeding of larvae of three fishes occurring in
the California Current, Sardinops sagax, Engraulis
mordax, and Trachurus symmetricus',' by David K.
Arthur 517

"Food of five species of cooccurring flatfishes on Oregon's
continental shelf;' by Michael Kravitz 984

"Foods of juvenile sockeye salmon, Oncorhynchus nerka,
in the inshore coastal waters of Bristol Bay, Alaska,
1966-67;' by H. Richard Carlson 458

FRANCIS, ROBERT C.-see SHARP and FRANCIS

Frontal periods
relation of fish catches in gill nets to 449

FUIMAN, LEE A., "Notes on the early development of

the sea raven, Hemitripterus americanus" 467

"Further observations of the feeding ecology of post-
larval pinfish, Lagodon rhomboides, and spot, Leiostomus

1006

xanthums',' by Martin A. Kjelson and George N. Johnson 423

GADBOIS, D. F., E. M. RAVESI, and R. C. LUND-
STROM, "Occurrence of volatile N-nitrosamines in
Japanese salmon roe" 683

GALLUCCI, VINCENT F.-see SCHERBA and
GALLUCCI

GARRISON, DAVID L., "Contribution of the net plank-
ton and nannoplankton to the standing stocks and
primary productivity in Monterey Bay, California during
the upwelling season" 183

Gasterosteus aculeatus—see Stickleback, threespine

GAUGLITZ, ERICH J., JR.-see HALL et al.

Genyonemus lineatus-see Croaker, white

Glyptocephalus zachirus—see Sole, rex

GOLDBERG, STEPHEN R., "Seasonal spawning cycles
of the sciaenid fishes Genyonemus lineatus and Seriphus
politus" 983

"Goose barnacles (Cirripedia: Thoracica) on flotsam
beached at La Jolla, California;' by Lanna Cheng and
Ralph A. Lewin 212

"Grazing of freshwater and estuarine, benthic diatoms
by adult Atlantic menhaden, Brevoortia tyrannus" by
Robert K. Edgar and James G. HofT 689

Great Central Lake, Vancouver Island
threespine stickleback

distribution, food, and feeding 647

food, with juvenile sockeye salmon, competition for 647

GRIFFIN, WADE L., NEWTON J. WARDLAW, and
JOHN P. NICHOLS, "Economic and financial analysis of
increasing costs in the Gulf shrimp fleet" 301

Growth
daily increments in otoliths

larval and adult fishes 1

daily increments of sagittae

nehu, Hawaiian Islands 9

"Growth and reproduction of the spotted porpoise,
Stenella attenuata, in the offshore eastern tropical
Pacific;' by William F. Perrin, James M. Coe, and James
R. Zweifel 229

Growth model
fish, prehatch and posthatch 609

"Growth of laboratory-reared northern anchovy, En-
graulis mordax, from southern California;' by Gary T.
Sakagawa and Makoto Kimura 271

Gulf of California, northern
totoaba
commercial fishery and reproductive biology 531

Gulf of Mexico
seatrouts

protein taxonomy 599

shrimp fleet

increasing costs, economic and financial analysis of 301

Gulf of Mexico, northeastern
fish fauna

observations on, associated with offshore platforms 387
searobins
biology of five species 93

Gulf of Mexico, northern
Gulf menhaden
analysis of returns of tagged 112

HAEFNER, PAUL A.. JR.-see LAIRD et al.

Halibut, Pacific
Pacific Ocean, northeast
mercury content 783

HALL, ALICE S., FUAD M. TEENY, and ERICH J.
GAUGLITZ, JR., "Mercury in fish and shellfish of the
northeast Pacific. II. Sablefish, Anoplopoma fimbria" . . . 791

LAURA G. LEWIS, WILLIAM

H. HARDMAN. and ERICH J. GAUGLITZ, JR.,
"Mercury in fish and shellfish of the northeast Pacific. I.
Pacific halibut, Hippoglossus stenolepis" 783

HANSON, CHARLES H., and JONATHAN BELL,
"Subtidal and intertidal marine fouling on artificial
substrata in northern Puget Sound, Washington" 377

HARDMAN, WILLIAM H.-see HALL et al.

"Harness for attachment of an ultrasonic transmitter to
the red drum, Sciaenops ocellata',' by William E. Carr and
Thomas B. Chaney 998

HARRELL, LEE W., ANTHONY J. NOVOTNY, MI-
CHAEL H. SCHIEWE, and HAROLD 0. HODGINS,
"Isolation and description of two vibrios pathogenic to
Pacific salmon in Puget Sound, Washington" 447

HART, ALLAN C.-see LORD ET AL.

HASTINGS. ROBERT W., LARRY H. OGREN, and
MICHAEL T. MABRY, "Obser\'ations on the fish fauna
associated with offshore platforms in the northeastern
Gulf of Mexico" 387

HAURY, LOREN R., "Method for restraining living
planktonic crustaceans" 220

Hawaii
mesopelagic fishes taken near
vertical distribution and other aspects of ecology . . 635

sergestid shrimps, ecology of 799

Hawaiian Islands

black skipjack, second record 207

nehu
age and growth as indicated by daily growth in-
crements of sagittae 9

HAYNES, EVAN, "Description of zoeae of coonstripe
shrimp, Pandalus hypsinotus, reared in the laboratory" 323

Hemitripterus americanus—see Sea raven

HENDRICKSON, JOHN R.-see FLANAGAN and
HENDRICKSON

Herring, Pacific
predator-prey relationship between larvae and Hype-
roche rnedusaru m 669

HEWITT, ROGER P, PAUL E. SMITH, and JOHN C.
BROWN, "Development and use of sonar mapping for
pelagic stock assessment in the California Current area" 281

Hippoglossus stenolepis— see Halibut, Pacific

HIRSCH, NINA-see KORN et al.

HOBSON, EDMUND S., and JAMES R. CHESS, "Tropic
interactions among fishes and zooplankters near shore at
Santa Catalina Island, California" 567

HODGINS, HAROLD O.-see HARRELL et al.

HOFF, JAMES G.-see EDGAR and HOFF

Homarus americanus—see Lobster, American

HORTON, HOWARD F.-see STAEGER and HORTON

HUNTER, JOHN R., "Culture and growth of northern
anchovy, Engraulis mordax, larvae" 81

, and CAROL SANCHEZ, "Diel changes in

swim bladder inflation of the larvae of the northern
anchovy, Engraulis mordai" 847

HURLEY, ANN C, "Feeding behavior, food consump-
tion, growth, and respiration of the squid Loligo opales-
cens raised in the laboratory" 176

Hyperoche medusarum
Pacific herring larvae
predator-prey relationship between 669

"Incidence of cull lobsters, Homarus americanus, in
commercial and research catches off the Maine coast!' by
Jay S. Krouse 719

I nfauna variation
soft substrate environment

application of systematic sampling to study of 937

Invertebrates, epibenthic

long-term fluctuations of populations in Apalachicola

Bay, Florida

distribution ^13

dominant species, seasonal fluctuations of 314

physicochemical parameters 313

"Isolation and description of two vibrios pathogenic to
Pacific salmon in Puget Sound, Washington;' by Lee W.
Harrell, Anthony J. Novotny, Michael H. Schiewe, and
Harold 0. Hodgins ^"^

Japan
salmon roe
N-nitrosamines, occurrence of volatile 683

JOHNSON, ALLYN G., "Electrophoretic evidence of
hybrid snow crab, Chionoecetes bairdi x opilio" 693

JOHNSON, GEORGE N.-see KJELSON and JOHNSON

Katsuwonus pelamis-see Tuna, skipjack

1007

KEENE, DONALD F., and WILLIAM G. PEARCY,
"Comparison of the most successful west coast albacore
troll fishermen"

Kelp
Santa Barbara, California
reef fishes, day versus night activity

KIMURA, MAKOTO-see SAKAGAWA and KIMURA

KJELSON, MARTIN A., and GEORGE N. JOHNSON,
"Further observations of the feeding ecology of postlar-
va! pinfish, Lagodan rhomboidefs, and spot, Leiostomus
xa Hthiirus"

KOBYLINSKI, GERARD J. -see LIVINGSTON et al.

KORN, SID, NINA HIRSCH, and JEANNETTE W.
STRUHSAKER, "Uptake, distribution, and depuration
of '^C-benzene in northern anchovy, Engraulis niordax,
and striped bass, Morone saxatilis"

, JEANNETTE W. STRUHSAKER, and

PETE BENVILLE, JR., "Effects of benzene on growth,
fat content, and caloric content of striped bass, Morone
saxatilis"

KRAVITZ, MICHAEL J., "Food of five species of cooc-
curring flatfish on Oregon's continental shelf"

KRETSCH, M. J. -see CRAWFORD and KRETSCH

KROUSE, JAY S., "Incidence of cull lobsters, Homarus
americanus, in commercial and research catches off the
Maine coast"

, "Size composition and growth of young rock

crab. Cancer irrorafus, on a rocky beach in Maine" ....

La JoUa, California
goose barnacles on flotsam beached at

Lagodon rhomboides—see Pinfish

LAIRD. CHAE E., ELIZABETH G. LEWIS, and PAUL
A. HAEFNER.JR.

"Occurrence of two galatheid crustaceans, Muiiida
forceps and Miuiidopsis bermiidezi, in the Chesapeake
Bight of the western North Atlantic Ocean"

LANSFORD, LAWRENCE M., CHARLES W. CAIL-
LOUET, and KENNETH T. MARVIN, "Phospho-
glucomutase polymorphism in two penaeid shrimps,
Penaeiis brasiliensis and Penaeus aztecus subtilis" ....

Larvae
crab, Dungeness
larval dynamics off central Oregon coast, 1970-71 . .

Larvae, fish-see Fish larvae

"Larval dynamics of the Dungeness crab. Cancer magis-
ter, off the central Oregon coast, 1970-71;' by R. Gregory
Lough

LASKER, REUBEN-see BROTHERS et al.

-see ZWEIFEL and LASKER

LAURENCE, GEOFFREY C, "Caloric values of some

1008

973

703

423

545

694
984

719

949

212

462

453

353

North Atlantic calanoid copepods"

LAURS, R. MICHAEL, WILLIAM H. LENARZ, and
ROBERT N. NISHIMOTO, "Estimates of rates of tag
shedding by North Pacific albacore, Thunnus alalunga"

Leiostomus xanthHrussee Spot

LENARZ, WILLIAM H.-see LAURS et al.

Lepidopsetta bilineata—see Sole, rock

LEVI, ELDON J.-see PRISTAS et al.

LEWIN, RALPH A.-see CHENG and LEWIN

LEWIS, ELIZABETH G., "Epizoites associated with
Bathynectes superbus (Decapoda: Portunidae)"

-see LAIRD et al.

LEWIS, FRANK G., Ill-see LIVINGSTON et al.

LEWIS, LAURA G.-see HALL et al.

LEWIS, THOMAS C, and RALPH W. YERGER,
"Biology of five species of searobins (Pisces, Triglidae)
from the northeastern Gulf of Mexico"

"Life history, ecology, and behavior of Liparis inquilin-
Hs (Pisces: Cyclopteridae) associated with the sea scallop,
Placopecfen magellanicus',' by K. W. Able and J. A.
Musick

"Life history of coho salmon, Oncorhynchus kisutch, in
Sashin Creek, southeastern Alaska!' by Richard A. Crone
and Carl E. Bond

Liparis inquiliniis
life history, ecology, and behavior of, associated with
sea scallop

abundance, geographic variation

behavior of

diel rhythm in fish-scallop association

feeding

juveniles

larvae

resting

spawning behavior

LIVINGSTON, ROBERT J., GERARD J. KOBYLIN-
SKI. FRANK G. LEWIS, III. and PETER F. SHERI-
DAN, "Long-term fluctuations of epibenthic fish and
invertebrate populations in Apalachicola Bay, Florida"

Lobster, American
culls off Main coast, incidence of

fishing intensity, effect on

seasonal and size variation in

value loss of catch due to

Loligo opalescens
raised in laboratory
feeding behavior, food consumption, growth, and
respiration

"Long-term fluctuations of epibenthic fish and inverte-
brate populations in Apalachicola Bay, Florida," by
Robert J. Livingston, Gerard J. Kobylinski, Frank G.
Lewis, III, and Peter F. Sheridan

218

675

225

409

897

418
414
416
414
412
412
413
412

311

721
719
723

176

311

LORD, GARY E., "Decision theory applied to the
simulated data acquisition and management of a salmon
fishery"

WILLIAM C. ACKER, ALLAN C. HARTT,

and BRIAN J. ROTHSCHILD, "An acoustic method for
the high-seas assessment of migrating salmon"

LOUGH, R. GREGORY, "Larval dynamics of the
Dungeness crab. Cancer magi>>ter, off the central Oregon
coast, 1970-71"

LUNDSTROM, R. C.-see GADBOIS et al.

MABRY, MICHAEL T.-see HASTINGS et al.

MacGREGOR, JOHN S., "DDT and its metabolites in the
sediments off southern California"

MACHIDORI, S.-see MASON and MACHIDORI

Mackerel, jack
California Current
larvae, food and feeding

McMAHON, ROBERT S.-see ZIMMERMAN and
McMAHON

Macrocrustaceans
abundance in natural and altered marshes

catch between areas, comparison of

day and night catches, comparison of

dissolved oxygen, catch related to

Macrozooplankton
standing stocks off Oregon, seasonal and inshore-
offshore variations

Maine
crab, rock

size composition and growth of young

lobsters, cull

commercial and research catches

Makaira nigricanft-see Marlin, blue

MANOOCH, CHARLES S., Ill, "Reproductive cycle,
fecundity, and sex ratios of the red porgy, Pagrus pagnis
(Pisces: Sparidae), in North Carolina"

MANZER, J. I., "Distribution, food, and feeding of the
threespine stickleback, Gasterosteus aculeatus, in Great
Central Lake, Vancouver Island, with comments on
competition for food with juvenile sockeye salmon,
Oncorhynchus nerka"

Marine fouling
subtidal and intertidal, on artificial substrata in Puget
Sound, Washington

fouling colonization of construction materials

fouling organisms, seasonal distribution

fouling organisms, vertical distribution in subtidal

zone

physicochemical environment

Marlin, blue

Pacific, central North
von Bertalanffy growth curves

837

104

353

517

197
196
197

949
719

775

647

379
381

382
378

553

Marlin, striped
Pacific, central North
von Bertalanffy growth curves 553

MARVIN, KENNETH T.-see LANSFORD et al.

MASON, J.C, "Some features of coho salmon, Oncor-
hynchus kisutch, fry emerging from simulated redds and
concurrent changes in photobehavior" 167

, and S. MACHIDORI, "Populations of sym-

patric sculpins, Cottus aleuficux and Cottus anper, in four
adjacent salmon-producing coastal streams on Van-
couver Island, B.C." 131

MATHEWS, CHRISTOPHER P.-see BROTHERS et al.

MATSUMOTO, WALTER M., "Second record of black
skipjack, Euthynnu!< lineatus, from the Hawaiian Is-
lands" 207

MAY, NELSON, LEE TRENT, AND PAUL J. PRISTAS,
"Relation of fish catches in gill nets to frontal periods". 449

Menhaden, Atlantic
grazing of freshwater and estuarine, benthic diatoms
by adults 689

Menhaden, Gulf
analysis of returns of tagged

autumn releases and recoveries 116

fishing areas 112

recovering tags, methods of 113

tagging, methods of 112

spring releases and recoveries 114

Menidia menidia-see Silverside, Atlantic

"Mercury in fish and shellfish of the northeast Pacific. I.
Pacific halibut, Hippoglossiif^ stenolepis" by Alice S. Hall,
Faud M. Teeny, Laura G. Lewis, William H. Hardman,
and Erich J. Gauglitz, Jr 783

"Mercury in fish and shellfish of the northeast Pacific. II.
Sa.h\efish,Anoplopo7nafimhria',' by Alice S. Hall, Faud M.
Teeny, and Erich J. Gauglitz, Jr 791

MERRINER, JOHN V., "Aspects of the reproductive
biology of the weakfish, Cynoscion regalis (Sciaenidae),
in North Carolina" 18

"Method for restraining living planktonic crustaceans;'

by Loren R. Haury 220

Micronekton
standing stocks off Oregon, seasonal and inshore-
offshore variations '^^

MILLER, R. J., "North American crab fisheries: Regula-
tions and their rationales" 623

"Minimum swimming speed of albacore, Thunnus

alalu nga" by Ronald C. Dotson 955

Monterey Bay, California
net plankton and nannoplankton
standing stocks and primary productivity during
upwelling season, contribution to 183

Morone saxatilis-see Bass, striped

1009

"Mortalities and epibiotic fouling of eggs from wild
populations of the Dungeness crab, Cancer niagister" by
William S. Fisher and Daniel E. Wickham

Muyiida forceps
Chesapeake Bight
occurrence in

Munidopsis berniudezi
Chesapeake Bight
occurrence in

MUSICK, J.A.-see ABLE and MUSICK

N-nitrosamines
volatile, occurrence of
salmon roe, Japanese

Nannoplankton
Monterey Bay, California
standing stocks and primary productivity during
upwelling season, contribution to

Nehu
age and growth, Hawaiian Islands

geographical comparison of growth rates

indicated by daily growth increments of sagittae . .

NELLEN, WALTER-see BLACKBURN and NELLEN

NematobrachioH sp.
California Current, central region
density, vertical range, and diel movement

Nematoscelis sp.
California Current, central region
density, vertical range, and diel movement

Nets, gill
fish catches
relation to frontal periods

NICHOLS, JOHN P.-see GRIFFIN et al.

NISHIMOTO, ROBERT N.-see LAURS et al.

"North American crab fisheries: Regulations and their
rationales;' by R. J. Miller

North Carolina
porgy, red

reproductive cycle, fecundity, and sex ratios

weakfish, reproductive biology

"Notes on the early development of the sea raven,
Hemitripterus americanus" by Lee A. Fuiman

NOVOTNY, ANTHONY J.-see HARRELL et al.

Observations on the bigeye thresher shark, Alopias
superciliosus, in the western North Atlantic;' by Charles
E. Stillwell and John G. Casey

"Observations on the commercial fishery and reproduc-
tive biology of the totoaba, Cynoscion macdonaldi, in the
northern Gulf of California;' by Christine A. Flanagan
and John R. Hendrickson

"Observations on the fish fauna associated with offshore
1010

201

462

683

183

16
9

932

449

623

775
18

467

221

531

platforms in the northeastern Gulf of Mexico;' by Robert

W. Hastings, Larry H. Ogren, and Michael T. Mabry . . . 387

"Occurrence of two galatheid crustaceans, Munida
forceps, and Munidopsis bermudezi, in the Chesapeake
Bight of the western North Atlantic Ocean;' by Chae E.
Laird, Elizabeth G. Lewis, and Paul A. Haefner, Jr 462

"Occurrence of volatile N-nitrosamines in Japanese
salmon roe;' by D. F. Gadbois, E. M. Ravesi, and R. C.
Lundstrom 683

OGREN, LARRY H.-see BRUSHER and OGREN

-see HASTINGS et al.

Oil and grease
fishery waste effluents
proposed analytical method 681

"Oil and grease: A proposed analytical method for fishery
waste effluents;' by Jeff Collins 681

Oncorhynchus gorbuscha—see Salmon, pink

Oncorhynchus kisutck—see Salmon, coho

Oncorhynchus nerka—see Salmon, sockeye

Oncorhynchus tshawytscha—see Salmon, chinook

"Optica! malformations induced by insecticides in em-
br\-os of the Atlantic silverside, Menidia menidia" by
Judith S. Weis and Peddrick Weis 208

Oregon

crab, Dungeness
larval dynamics off central coast, 1970-71 353

micronekton and macrozooplankton
season and inshore-offshore variations in standing
stocks 70

Otoliths
daily growth increments, larval and adult fishes 1

Oyster, Pacific
fertilization method quantifying gamete concentra-
tions and maximizing lanae production 698

Pacific, central North
blue marlin

von Bertalanffy growth cunes 553

Scopelengys clarkei, description of new species 142

striped marlin

von Bertalanffy growth curves 553

Pacific, eastern
whale, melon-headed

first record in, with summary of world distribution 457

yellowfin tuna

energetics model for an exploited population 36

Pacific, eastern tropical
porpoise, spotted
growth and reproduction 229

Pacific, North
albacore
tag shedding, estimates of rates 675

Pacific, northeast
halibut, Pacific

mercury content

sablefish

mercury content

Pagnis pagrus—see Porgy, red

Pandalus hypsinotus— Shrimp, coonstripe

Pandalus Jordan i
sampler, epibenthic, used to study ontogeny of vertical
migration

"Paralytic shellfish poisoning in Tenakee, southeastern
Alaska: A possible cause," by Steven T. Zimmerman and
Robert S. McMahon

Parophrys vetulus—see Sole, English

PARRACK, M. L., "Estimation of fishing effort in the
western North Atlantic from aerial search data"

PEARCY, WILLIAM G., "Seasonal and inshore-offshore
variations in the standing stocks of micronekton and
macrozooplankton off Oregon."

-see KEENE and PEARCY

-see ROTHLISBERG and PEARCY

FELLA, JEROME J.-see BAILEY et al.

Pevaeus aztecus—see Shrimp, brown

Penaeus azfecua sithtilis

phosphoglucomutase polymorphism in

Penaeus hrasiliensis
phosphoglucomutase polymorphism in

Penaeus duorarum—see Shrimp, pink

Penaeus setiferus-see Shrimp, white

Peponocephala electra—see Whale, melon-headed

PERRIN, WILLIAM F., "First record of the melon-
headed whale, Peponocephala electra, in the eastern
Pacific, with a summary of world distribution"

JAMES M. COE, and JAMES R. ZWEIFEL,

"Growth and reproduction of the spotted porpoise,
Stenella attenuata, in the offshore eastern tropical
Pacific"

Petalidium suspiriosum
ecology of Hawaiian

"Phosphoglucomutase polymorphism in two penaeid
shrimps, Penaeus hrasiliensis and Penaeus aztecus
svbtilis" by Lawrence M. Lansford, Charles W. Caillouet,
and Kenneth T. Marvin

Pinfish
feeding ecology, observations of pwstlarval

current speed and feeding intensity

daily rations

evacuation rates

feeding periodicity

783
791

994

679

503

453

457

229

824

453

424
425
424
425

food preferences, size related 428

midday feeding, temporal variation 424

Placopecten magellanicus-see Scallop, sea

Plankton, net
Monterey Bay, California
contribution to standing stocks and primary
productivity during upwelling season 183

Polychaete variation
application of systematic sampling to study of, in soft
substrate environment 942

"Population biology of Euphausia pacifica off southern
California;' by Edward Brinton 733

"Populations of sympatric sculpins, Cottus aleuticus and
Coitus asper, in four adjacent salmon-producing coastal
streams on Vancouver Island, B.C.," by J. C. Mason and S.
Machidori 131

Porgy, red
North Carolina

fecundity 775

reproductive cycle 77,5

sex ratios 775

Porpoise, spotted
growth and reproduction in eastern tropical Pacific

color pattern 242

gestation period 234

growth, fetal 234

growth, postnatal 236

Japanese population, comparison with 266

length at birth 234

length-weight relationships 242

production, gross annual 265

reproduction, female 248

reproduction, male 243

reproduction, seasonality 243

schooling in relation to reproduction 265

sex ratios 264

Porpoises
underwater paint marking 287

"Predator-prey relationship between Pacific herring,
Clupea harengus pallasi, larvae and a predator}' hyperiid
amphipod, Hyperoche medusarum" by Hein von Wes-
temhagen and Harold Rosenthal 669

"Prehatch and posthatch growth of fishes-a general
model;' by James R. Zweifel and Reuben Lasker 609

"Present and historical spawning grounds and nurseries
of American shad, Alosa sapidissima, in the Delaware
River;' by Mark E. Chittenden, Jr 343

PRISTAS, PAUL J.-see MAY et al.

, ELDON J. LEVI, and ROBERT L.

DRYFOOS, "Analysis of returns of tegged Gulf men-
haden" 1^2

PROCHASKA, FRED J.-see ALVAREZ et al.

PROCTOR, RAPHAEL-see TRENT et al.

1011

"Production of fry and adults of the 1972 brood of pink
salmon, Oncorhijnchnsgnrbuficka, from gravel incubators
and natural spawning at Auke Creek, Alaska," by Jack E.
Bailey, Jerome J. Pella, and Sidney G. Taylor

"Production of juvenile chinook salmon, Oncorkynchus
tshain/tscha, in a heated model stream" by Peter A.
Bisson and Gerald E. Davis

"Protein taxonomy of the Gulf of Mexico and Atlantic
Ocean seatrouts, genus Cynoscion" by Michael P. Wein-
stein and Ralph W. Yerger

Pseiidopleuro»ectes americanus-see Flounder, winter

Puget Sound, Washington
marine fouling

subtidal and intertidal, on artificial substrata

Pacific salmon

isolation and description of two vibrios pathogenic to

PULLEN, EDWARD J. -see TRENT et al.

Queenfish
seasonal spawning cycle

RALSTON, STEPHEN, "Age determination of a tropical
reef butterflyfish utilizing daily growth rings of otoliths"

RAVESI, E. M.-see GADBOIS et al.

"Reevaluation of fishing effort and apparent abundance
in the Hawaiian fishery for skipjack tuna, Katsiiwonus
pehniis, 1948-701' by Richard N. Uchida

"Relation of fish catches in gill nets to frontal periodsl' by
Nelson May, Lee Trent, and Paul J. Pristas

"Reproductive cycle, fecundity, and sex ratios of the red
porgy, Pagnis pagni.< (Pisces: Sparidae), in North Ca-
rolina," by Charles S. Manooch III

"Review of the deep-sea fish genus Scopelengys
(Neoscopelidae) with a description of a new species,
Scopelengys clarkei, from the central Pacific!' by John L.
Butler and Elbert H. Ahlstrom

ROGERS, CAROLYN A., "EflTects of temperature and
salinity on the survival of winter flounder embryos" . . .

ROSENTHAL, HARALD-see WESTERNHAGEN and
ROSENTHAL

ROTHLISBERG, PETER C, and WILLIAM G.
PEARCY, "An epibenthic sampler used to study the
ontogeny of vertical migration of Pandaltis jordani
(Decapoda, Caridea)"

ROTHSCHILD, BRIAN J.-see LORD et al.

Sablefish
mercury in, from northeast Pacific

age

geographical location

sex

size

utilization

961

763

599

377

447

983

990

449

775

142

994

795
792
796
793
796

Sagittae
in nehu, daily growth increments
age and growth 9

Sahara, Spanish
pelagic fish eggs and larvae
distribution and ecology in an upwelling area off . . 885

St. Andrew Bay, Florida
penaeid shrimp, distribution, abundance, and size ... 158

SAKAGAWA, GARY T., and MAKOTO KIMURA,
"Growth of laboratory-reared northern anchovy, En-
graulic mordax, from southern California" 271

Salinity
sun'ival, effects on

winter flounder embryos 52

Salmon
migrating, high-seas assessment of

acoustic buoy 104

receiver-decoder system 104

Salmon, chinook
production of juvenile in a heated model stream

associated flora 765

benthos and drift 765

disease 769

food availability 769

periphyton biomass and sedimentation 771

physical characteristics of stream 763

temperature, direct effects on growtn 768

temperature regulation 764

temporal changes in production 767

Salmon, coho
fry, some features of emerging from simulated redds

and concurrent changes in photobehavior 167

life history in Sashin Creek, Alaska

age determination 904

escapement size 899

fecundity 905

fry, emigration and salinity tolerance 908

fry and smolts, numbers 907

growth and age characteristics 911

interspecific competition 903

juveniles, age of in stream 910

juveniles, entering estuary 907

juveniles, survival of 914

redd life 902

retained eggs 905

spawners, distribution and density 903

survival and instantaneous mortality rates 917

sunival from potential egg deposition to emergence 913

Salmon, Pacific

Puget Sound, Washington
isolation and description of two vibrios pathogenic to 447

Salmon, pink

production of fry and adults from gravel incubators

and natural spawning in Alaska

adults, recovery of marked 965

adults, size of returning 968

adults, timing of return 969

egg collection and eyeing 962

1012

fry counting and processing 963

fry size and developmental index 967

natural spawning 962

oxygen levels 963

raising eyed eggs to fry stage 962

survival from egg to fry 966

survival from egg to returning adult 967

survival from fry to returning adult 966

sur\-ival from marking effects 967

time of emergence and seaward migration 968

water filter and purifier 962

water temperatures 963

Salmon, sockeye
Bristol Bay, Alaska, 1966-67

foods of juvenile in inshore coastal waters 458

Great Central Lake, Vancouver Island
competition for food with the threespine stickleback 647

Salmon, fishery
decision theory applied to simulated data acquisition
and management 837

Salmon roe
Japanese

N-nitrosamines, occurrence of volatile 683

Sampling, stratified systematic

infauna variation in soft substrate environment

bivalves 944

polychaetes 942

procedures, field 939

procedures, statistical 939

sediment 941

Santa Barbara, California
reef fishes
day versus night activity in a kelp forest 703

Santa Catalina Island
trophic interactions among fishes and zooplankters . . 567

SANCHEZ, CAROL-see HUNTER and SANCHEZ

Sandab, Pacific
Oregon, continental shelf

food of 984

Sardine, Pacific
California Current
lanae, food and feeding 517

Sardinops sagaj-see Sardine, Pacific

Sashin Creek, Alaska
salmon, coho
life history 897

Scallop, sea
life history, ecology, and behavior of Liparis inquilin-
us associated with 409

SCHERBA, STEPHEN, JR., and VINCENT F. GAL-
LUCCI, "The application of systematic sampling to a
study of infauna variation in a soft substrate environ-
ment" 937

SCHEVILL, WILLIAM E.-see WATKINS and
SCHEVILL

SCHIEWE, MICHAEL H.-see HARRELL et al.

Sciaenops ocellata-see Drum, red

Scopelengijs

review of genus and description of a new species from
central North Pacific 142

Scopelengys darker

new species from the cental North Pacific 142

Sculpins
populations of sympatric on Vancouver Island

age determination 133

age structure 136

annual growth, mortality, and length-weight

relations 138

biomass distribution 137

distribution and relative abundance 133

general life history 133

population estimates 132

population sampling 132

Sea raven
early development, notes on 467

Searobins

biology of five species from northeastern Gulf of

Mexico

bottom type 100

capture, time of 100

depth distribution 99

food habits 102

geographic distribution 99

sexual maturity 101

size-depth relationship 100

spawning season 101

water temperature 100

"Seasonal and inshore-offshore variations in the standing
stocks of micronekton and macrozooplankton off Oregon"
by William G. Pearcy "^O

"Seasonal spawning cycles of the sciaenid fishes Gen-
yonemus lineatus and Seriphus politus" by Stephen R.
Goldberg 983

Seatrouts
protein taxonomy

Atlantic Ocean ^^^

Gulf of Mexico ^99

"Second record of black skipjack, Euthynnus lineatus,

from the Hawaiian Islands!" by Walter M. Matsumoto. . 207

Sergestes armatus

• • 810

ecology of Hawaiian

Sergestes atlanticus
ecology of Hawaiian

Sergestes consobrinus
ecology of Hawaiian

Sergestes cornutus
ecology of Hawaiian

Sergestes erectus
ecology of Hawaiian

1013

Sergestes orient alis
ecology of Hawaiian 811

Sergestes: pectinatmt
ecology of Hawaiian 815

Sergestes sargassi
ecology of Hawaiian 813

Sergestes tanillus
ecology of Hawaiian 812

Sergestes vigilax
ecology of Hawaiian 811

Se rg ia bigem m ea
ecology of Hawaiian 819

Sergia bisulcata
ecology of Hawaiian 822

Sergia fulgens
ecology of Hawaiian 816

Sergia gardineri
ecology of Hawaiian 818

Sergia inequalis
ecology of Hawaiian 821

Sergia laminata
ecology of Hawaiian 824

Sergia maxima
ecology of Hawaiian 823

Sergia scintillans
ecology of Hawaiian 817

Sergia tenuiremis
ecology of Hawaiian 823

Seripkus politus-see Queenfish

Shad, American
spawning grounds and nurseries, Delaware River
adults, areas contributing to successful production of 349

behavior, spawning period 345

spawning period 344

weight loss, mortality, feeding, and duration of res-
idence in fresh water

feeding behavior I53

gonad weight-length relationships prior to spawning 152
somatic weight-length relationships prior to spawn-
ing 152

total and fork length conversion 152

total weight-length relationships prior to spawning 152
upstream mortality 153

Shark, bigeye thresher

western North Atlantic, observations on 221

SHARP, GARY D., and ROBERT C. FRANCIS, "An
energetics model for the exploited yellowfin tuna, Thin-
nus alhacares, population in the eastern Pacific Ocean". 36

Shellfish poisoning, paralytic

Tenakee, southeastern Alaska

possible cause 679

SHERIDAN, PETER F.-see LIVINGSTON et al.
1014

Shrimp
Pandahis jordani
sampler, epibenthic, used to study ontogeny of
vertical migration 994

Shrimp, brown
West Bay, Texas
abundance in natural and altered marshes 195

Shrimp, coonstripe
zoeae
description of, reared in laboratory 323

Shrimp, grass
West Bay, Texas
abundance in natural and altered marshes 195

Shrimp, penaeid
St. Andrew Bay system, Florida, distribution, abun-
dance, and size 158

Shrimp, pink
West Bay, Texas
abundance in natural and altered marshes 195

Shrimp, white
West Bay, Texas
abundance in natural and altered marshes 195

Shrimp fleet
Gulf of Mexico
economic and financial analysis of increasing costs. 302

Shrimp processing
dual structural equilibrium in Ftorida industry

characteristics 881

entry and exit patterns, 1959-71 880

forecasting firm distribution and predicting struc-
tural equilibrium 882

Shrimps
ecology of Hawaiian sergestid
color pattern and daytime vertical distribution: role

of countershading 825

contamination problem in analysis of vertical dis-
tribution data 803

feeding chronology and diet 829

feeding study 803

interspecific relations 831

nighttime vertical distribution and migration 827

Petalidium suspiriosum 824

reproduction and growth 831

sampling area 799

Sergestes armatus 810

Sergestes atlanticus 805

Sergestes consorbrinus 813

Sergestes cornutus 808

Sergestes erectus 809

Sergestes orientalis 811

Sergestes pectinatus 815

Sergestes sargassi 813

Sergestes tantillus 812

Sergestes vigilax 811

Sergia bigemmea 819

Sergia bisulcata 822

Sergia fulgens 816

Sergia gardineri 818

Sergia inequalis 821

Sergia laminata

Sergia maxima

Sergia scintillans

Sergia tenuiremis

vertical distribution

Silverside, Atlantic
embryos, optical malformations induced by insec-
ticides

"Size composition and growth of young rock crab. Cancer
irroratus, on a rocky beach in Mainel' by Jay S. Krouse

SKILLMAN, ROBERT A., and MARIAN Y. Y. YONG,
"Von Bertalanffy growth curves for striped marlin,
Tetrapturus audax, and blue marlin, Makaira nigricans,
in the central North Pacific Ocean"

Skipjack, black
Hawaiian Islands, second record

SMITH, PAUL E.-see HEWITT et al.

Sole, English
Oregon, continental shelf
food of

Sole, petrale
Oregon, continental shelf
food of

Sole, rex
Oregon, continental shelf
food of

Sole, rock
Oregon, continental shelf
food of

"Some features of coho salmon, Oncorhynckus kisutch,
fry emerging from simulated redds and concurrent
changes in photobehavior" by J. C. Mason

Sonar mapping
development and use for pelagic stock assessment in
California current area

automated hydrocoustic data acquisition and

processing system

automated sonar survey

bottom topography

diurnal and seasonal effects

experiment, charter boat

experiment, fish trap

target size

target strength

Spot

feeding ecology, observations of postlarval

current speed and feeding intensity

daily ratios

evacuation rates

feeding periodicity

size related food preferences

temporal variation in midday feeding

Squid

laboratory-reared
feeding behavior, food consumption, growth, and
respiration

824
823
817
823
800

208

949

553

STADELMAN,
STADELMAN

DONALD-see TILLMAN and

984

167

292
294

289
285
291
290
283
285

424
425
424
425
428
424

176

STAEGER, WILLIAM H., and HOWARD F. HORTON,
"Fertilization method quantifying gamete concentra-
tions and maximizing larvae production in Crassostrea
gigas" 698

Stenella attenuata—see Porpoise, spotted

Stickleback, threespine
Great Central Lake, Vancouver Island
competition for food with juvenile sockeye salmon . 647
distribution, food, and feeding 647

STILLWELL, CHARLES E., and JOHN G. CASEY,
"Observations on the bigeye thresher shark, Alopias
superciliosus, in the western North Atlantic"

207 Stolephorus purpureus—see Nehu

STRUHSAKER, JEANNETTE W.-see KORN et al.

STRUHSAKER, PAUL, and JAMES H. UCHIYAMA,
"Age and growth of the nehu, Stolephorus purpureus
(Pisces: Engraulidae), from the Hawaiian Islands as
indicated by daily growth increments of sagittae"

Stylocheiron sp.
California Current, central region
density, vertical range, and diel movement

"Subtidal and intertidal marine fouling on artificial
substrata in northern Puget Sound, Washington," by
Charles H. Hanson and Jonathan Bell

Swim bladder
anchovy, northern
diel changes in inflation

Swimming speed
minimum
tuna, albacore

TAYLOR, SIDNEY G.-see BAILEY et al.

TEENY, FUAD M.-see HALL et al.

Temperature
effects on survival
winter flounder embryos

Tenakee, Alaska
shellfish poisoning, paralytic
possible cause

TENNY, RICHARD D.-see COLLINS and TENNY

Tessarabrachion sp.
California Current, central region
density, vertical range, and diel movement

Tetrapturus audax-see Marlin, striped

"Thermal tolerance and resistance of the northern
anchovy, Engraulis mordax" by Gary D. Brewer

Thunnus alalunga-see Tuna, albacore

Thunnus albacares-see Tuna, yellowfin

221

932

377

847

955

679

929

433

1015

Thysanoessa sp.
California Current, central region
density, vertical range, and diel movement

Thysanopoda sp.
California Current, central region
density, vertical range, and diel movement

TILLMAN, MICHAEL F., and DONALD STADEL-
MAN, "Development and example application of a
simulation model of the northern anchovy fishery"

Totoaba
commercial fishery and reproductive biology in north-
ern Gulf of California

breeding migration

decline of fishery

diet

distribution

habitat

history of fishery

spawning concentration

Trachurus symmetricus-see Mackerel, jack

Transmitter, ultrasonic
harness for attachment of
red drum

TRENT, LEE-see MAY et al.

EDWARD J. PULLEN, and RAPHAEL

PROCTOR, "Abundance of macrocrustaceans in a natu-
ral marsh and a marsh altered by dredging, bulkheading,
and filling"

"Trophic interactions among fishes and zooplankters
near shore at Santa Catalina Island, California," by
Edmund S. Hobson and James R. Chess

Tuna, albacore
Pacific, North

tag shedding, estimates of rates

swimming speed, minimum

comparison among four scombrids

density variations

determining

estimates, field

west coast

troll fishermen, comparison of the most successful

and least successful

Tuna, skipjack
reevaluation of fishing effort and apparent abundance
in Hawaiian fishery, 1948-70

catch reports

comparison of catch per effective trip and catch per

day fished

differences in catch per effective trip between vessel

classes, between areas, and among years

fishing areas

interrelation of total catch, fishing intensity, and

apparent abundance

measures of apparent abundance and fishing inten-
sity

relation between catch per day fished and catch per

effective trip

reporting of zero-catch trips

1016

929

118

534
538
537
537
537
532
535

998

195

567

675

958
956
957
959

973

64
61

64
60

sources of variability in fishing power among vessels 60

standardization of catch per day fished 66

vessels, classes of 61

Tuna, yellowfin
energetics model for an exploited population

food as a population regulator 46

population dynamics 37

spawning survival versus population biomass 47

total drag determination 41

velocity determination 41

Turkey

fed tuna oil with and without a-tocopherol supplement

or injection
effects of cooking in air or nitrogen on development
of fishy flavor in breast meat 89

Uca pugilafor-see Crab, fiddler

UCHIDA, RICHARD N., "Reevaluation of fishing efTort
and apparent abundance in the Hawaiian fishery for
skipjack tuna, KafKuwotniti pelaniis, 1948-70" 59

UCHIYAMA, JAMES H.-see STRUHSAKER and
UCHIYAMA

"Underwater paint marking of porpoises!' by William A.
Watkins and William E. Schevill 687

United States
continental shelf and slope, eastern coast

Ceriantharia, Zoanthidea, Corallimorpharia, and
Actiniaria 857

"Uptake, distribution, and depuration of '^C-benzene in
northern anchovy, Engraulis mordax, and striped bass,
Morone saxafilis," by Sid Korn, Nina Hirsch, and Jean-
nette W. Struhsaker 545

Vancouver Island, British Columbia
Cottus aleuticutt and Cottus aftper
populations of sympatric in four adjacent salmon-
producing coastal streams 131

"Vertical distribution and diel migration of euphausiids
in the central region of the California Current" by Marsh
J. Youngbluth 925

"Vertical distribution and other aspects of the ecology of
certain mesopelagic fishes taken near Hawaii," by
Thomas A. Clarke and Patricia J. Wagner 635

Vibrios
Puget Sound, Washington
isolation and description of two, pathogenic to
Pacific salmon 447

"Von Bertalanffy growth curves for striped marlin,
Tetrapturus audax, and blue marlm, Makaira nigricans,
in the central North Pacific Ocean," by Robert A. Skillman
and Marian Y. Y. Yong 553

WAGNER, PATRICIA J.-see CLARKE and WAGNER

WALTERS, JOHN F., "Ecology of Hawaiian sergestid
shrimps (Penaeidea: Sergestidae)" 799

WARDLAW, NEWTON J. -see GRIFFIN et al.

WATKINS, WILLIAM A., and WILLIAM E.
SCHEVILL, "Underwater paint marking of porpoises"

Weakfish
reproductive biology in North Carolina

"Weight loss, mortality, feeding, and duration of res-
idence of adult American shad, Alosa f:apidissima, in
fresh water;' by Mark E. Chittenden, Jr

WEINSTEIN, MICHAEL P., and RALPH W. YERGER,
"Protein taxonomy of the Gulf of Mexico and Atlantic
Ocean seatrouts, genus Cynof<cion"

WEIS, JUDITH S., "Effects of mercury, cadmium, and
lead salts on regeneration and ecdysis in the fiddler crab,
lira pugilator"

, and PEDDRICK WEIS, "Optical malforma-
tions induced by insecticides in embryos of the Atlantic
silverside, Menidia menidia"

WEIS, PEDDRICK-see WEIS and WEIS

West Bay, Texas

macrocrustaceans, abundance in a natural marsh and a
marsh altered by dredging.bulkheading, and filling. .

WESTERNHAGEN, HEIN VON, and HARALD ROS-
ENTHAL, "Predator-prey relationship between Pacific
herring, Chipea harengxs pallasi, larvae and a predatory
hyperiid amphipod, Hyperoche medm^arum"

Whale, melon-headed
Pacific, eastern
first record in, with summary of world distribution

687

151

599

464

208

195

669

457

WICKHAM, DANIEL E.-see FISHER and WICKHAM

WIDERSTEN, BERNT, "Ceriantharia, Zoanthidea,
Corallimorpharia, and Actiniaria from the continental
shelf and slope off the eastern coast of the United States"

857

YERGER, RALPH W.-see LEWIS and YERGER

-see WEINSTEIN and YERGER

YONG, MARIAN Y. Y.-see SKILLMAN and YONG

YOUNGBLUTH, MARSH J., "Vertical distribution and
diel migration of euphausiids in the central region of the
California Current" 925

ZIMMERMAN, STEVEN T., and ROBERT S.
McMAHON, "Paralytic shellfish poisoning in Tenakee,
southeastern Alaska: A possible cause" 679

Zoanthidea
continental shelf and slope, U.S. east coast

Epiznanthuif incruntatuii 858

Zooplankers
Santa Catalina Island
trophic interactions with fishes 567

Zooplankton
California Current, central region
density, vertical range, and diel movement 925

ZWEIFEL, JAMES R.-see PERRIN et al.

and REUBEN LASKER, "Prehatch and

posthatch growth of fishes-a general model" 609

1017

ERRATA

Fishery Bulletin, Vol. 74, No. 3

Skillman, Robert A., and Marian Y. Y. Yong, "Von Bertalanffy growth curves for striped marlin,
Tetrapturus audax, and blue marlin, Makaira nigricans, in the central North Pacific Ocean," p. 553-566.
1) Page 563, left column, line 8, correct line to read:

all 11 age-groups for females and using 12 and 11 (deleting oldest)

INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN

Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the
following instructions. These are not absolute requirements, of course, but desiderata.

CONTENT OF MANUSCRIPT

The title page should give only the title of the
paper, the author's name, his affiliation, and
mailing address, including Zip code.

The abstract should not exceed one double-
spaced page.

Text footnotes should be typed separately
fi*om the text.

Figures and tables, with their legends and
headings, should be self-explanatory, not requir-
ing reference to the text. Their placement should
be indicated in the right-hand margin of the
manuscript.

Preferably figures should be reduced by photo-
graphy to 5% inches (for single-column figures,
allowing for 50% reduction in printing), or to 12
inches (for double-column figures). The maximvmi
height, for either width, is 14 inches. Photo-
graphs should be printed on glossy paper.

Do not send original drawings to the Scientific
Editor; if they, rather than the photographic re-
ductions, are needed by the printer, the Scientific
Publications Staff will request them.

Each table should start on a separate page.
Consistency in headings and format is desirable.
Vertical rules should be avoided, as they make
the tables more expensive to print. Footnotes in
tables should be numbered sequentially in arabic
numerals. To avoid confusion with powers, they
should be placed to the left of numerals.

Acknowledgments, if included, are placed at
the end of the text.

Literature is cited in the text as: Lynn and Reid
(1968) or (Lynn and Reid 1968). All papers re-
ferred to in the text should be hsted alphabetically
by the senior author's surname under the heading

We prefer that measurements be given in
metric imits; other equivalent units may be given
in parentheses.

FORM OF THE MANUSCRIPT

TITLE PAGE

ABSTRACT

TEXT

LITERATURE CITED

APPENDIX

TEXT FOOTNOTES

TABLES (Each table should be numbered with
an arabic numeral and heading provided)

LIST OF FIGURES (entire figure legends)

FIGURES (Each figure should be numbered
with an arabic numeral; legends are desired)

ADDITIONAL INFORMATION

Send the ribbon copy and two duplicated or
carbon copies of the manuscript to:

Dr. Bruce B. CoUette, Scientific Editor

Fishery Bulletin

National Marine Fisheries Service

Systematics Laboratory

National Museum of Natural History

Washington, DC 20560

Fifty separates will be supplied to an author
free of charge and 100 supplied to his organiza-
tion. No covers will be suppUed.

Contents — continued

SCHERBA, STEPHEN, JR., and VINCENT F. GALLUCCI. The application of
systematic sampling to a study of infauna variation in a soft substrate environ-
ment 937

KROUSE, JAY S. Size composition and growth of young rock crab. Cancer irroratus,
on a rocky beach in Maine 949

DOTSON, RONALD C. Minimum swimming speed of albacore, Thunnus alalunga . . 955

KEENE, DONALD P., and WILLIAM G. PEARCY. Comparison of the most
successful and least successful west coast albacore troll fishermen 973

Notes

GOLDBERG, STEPHEN R. Seasonal spawning cycles of the sciaenid fishes Genyone-
mus lineatus and Seriphus politics 983

KRAVITZ, MICHAEL J., WILLIAM G. PEARCY, and M. P. GUIN. Food of five
species of cooccurring flatfishes on Oregon's continental shelf 984

RALSTON, STEPHEN. Age determination of a tropical reef butterflyfish utilizing
daily growth rings of otoliths 990

ROTHLISBERG, PETER C, and WILLIAM G. PEARCY. An epibenthic sampler used
to study the ontogeny of vertical migration of Pandalus jordani (Decapoda,
Caridea) 994^

CARR, WILLIAM E. S., and THOMAS B. CHANEY. Harness for attachment of an
ultrasonic transmitter to the red drum, Sciaenops ocellata 998

INDEX, VOLUME 74 1001

MBL HMIJI I.IBKARY

iiiH nu

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Fishery bulletin"

See other formats