OMJFORNIAl

FISH-GAME

"CONSERVATION OF WILDLIFE THROUGH EDUCATION"

California Fish and Game Is a journal devoted to the conservation of wild-
life. If its contents are reproduced elsewhere, the authors and the California
Department of Fish and Game would appreciate being acknowledged.

Subscriptions may be obtained at the rate of $5 per year by placing an
order with the California Department of Fish and Game, 1416 Ninth Street,
Sacramento, California 95814. Money orders and checks should be made out
to California Department of Fish and Game. Inquiries regarding paid sub-
scriptions should be directed to the Editor.

Complimentary subscriptions are granted, on a limited basis, to libraries,
scientific and educational institutions, conservation agencies, and on exchange.
Complimentary subscriptions must be renewed annually by returning the post-
card enclosed with each October issue.

Please direct correspondence to:

Perry L. HerrgeseiJ, Ph.D., Editor
California Fish and Game
1416 Ninth Street
Sacromento, California 95814

1

J

0

V

VOLUME 70

APRIL 1984

NUMBER 2

Published Quarterly by

STATE OF CALIFORNIA

THE RESOURCES AGENCY

DEPARTMENT OF FISH AND GAME

—LDA—

66 CALIFORNIA FISH AND CAME

STATE OF CALIFORNIA
GEORGE DEUKMEJIAN, Governor

THE RESOURCES AGENCY
GORDON VAN VLECK, Secretary for Resources

FISH AND GAME COMMISSION

WILLIAM A. BURKE, Ed.D., President

Brentwood

BRIAN J. KAHN, Vice President ABEL C. GALLETTI, Member

Santa Rosa Los Angeles

NORMAN B. LIVERMORE, JR., Member ALBERT C. TAUCHER, Member

San Rafael Long Beach

DEPARTMENT OF FISH AND GAME
HOWARD D. CARPER, Director

1416 9th Street
Sacramento 95814

CALIFORNIA FISH AND GAME

Editorial Staff

Editorial staff for this issue consisted of the following:

Wildlife William E. Grenfell, Jr.

Marine Resources Robert N. Lea, Ph.D.

Inland Fisheries Jack Hanson

Editor-in-Chief Perry L. Herrgesell, Ph.D.

67
CONTENTS

Page
Variation in Trophic State Indicators in Two Northern Califor-
nia Reservoirs K. R. Gina Rothe 68

Comparative Condition of Black-Tailed Deer, Odocoileus
hemionus columbianus, in Two Herds in Trinity County,

Califonria John G. Kie, Timothy S. Burton, and John W. Menke 78

Pup Dependency Period and Length of Reproductive Cycle:
Estimates from Observations of Tagged Sea Otters, Enhydra

lutris, in California Frederich E. Wendell, Jack A. Ames,

and Robert A. Hardy 89

A Review of Selected Remote Sensing and Computer Tech-
nologies Applied to Wildlife Habitat Inventories .... Kenneth E. Mayer 101

Identification of Salt Marsh Harvest Mice, Reithrodontomys
raviventris, in the Field and with Cranial Characteristics

Howard S. Shellhammer 113

A Second Record for California and Additional Morphological
Information on Entosphenus hubbs/\/\adyko\/ and Kott 1976
(Petromyzontidae) Vadim D. Vladykov and Edward Kott 121

Book Reviews 128

ERRATUM

McGriff, Darlene and John Modin. 1983. Thelohania contejeani paras\l\sm
of the crayfish, Pacifastacus leniusculus, in California. Calif. Fish Game,
69(3): 178-183.

Page 182. Second sentence of second paragraph should read:
"However, O'Keeffee and Reynolds (in press), using microscopic examina-
tion, found that up to about 4.0% of Austropotamobius pallipes infected with
Thelohania in Ireland eluded detection by macroscopic examination."

68 CALIFORNIA FISH AND CAME

Calif. Fish and Came 70 ( 2 ) ; 68-77 1 984

VARIATION IN TROPHIC STATE INDICATORS IN
TWO NORTHERN CALIFORNIA RESERVOIRS ^

K. R. GINA ROTHE

Dept. of Geology and Physical Sciences

California State University

Chico, CA 95929

Monitoring of certain limnological parameters has been conducted in Paradise and
Magalia Reservoirs since 1974 to determine if enlargement of Paradise Reservoir or
increased development in the watershed would lead to a deterioration of water
quality. Increased levels of nitrogen and phosphorus compounds and phytoplankton
in 1976 suggested enrichment was occurring. However, from 1976-1980, these
parameters returned to their 1974 levels. This reversal can best be explained by an
increased flushing rate due to abnormally high rainfall.

INTRODUCTION

Paradise and Magalia are two water supply reservoirs in the foothills of the
Sierra Nevada at 780 m and 680 m elevation, respectively (Figure 1). These
reservoirs, owned by the Paradise Irrigation District, are the primary source of
domestic water for the city of Paradise. Magalia and Paradise Dams impound
flow in Little Butte Creek (Figure 1). Magalia Dam, constructed in 1917, is
located 4 km downstream from Paradise Dam and is a 30 m high hydraulic-fill
structure. Paradise Dam, constructed in 1 954 and enlarged in 1 976 is a 52 m high
rolled, earth-fill structure. Annual discharge from Magalia Dam has ranged from
3,430,377 m^ to 27,090,236 m^ in the last 10 years. Development of large tracts
of mountainous land in the Little Butte Creek watershed above the reservoirs and
the concomitant problem of waste disposal have been a cause for concern to
the Butte County FHealth Department and the Paradise Irrigation District. In
1971-73, the California Department of Water Resources (DWR) conducted an
investigation of surface and groundwater in the watershed and identified certain
limnological parameters which needed regular monitoring: concentration of
nitrogen and phosphorus compounds in the surface and bottom waters and
volume of phytoplankton in the two reservoirs during March-May. Since this
would encompass spring turnover, nutrients and phytoplankton populations
would be at their highest levels.

These parameters have been measured biennially since 1974. In addition
dissolved oxygen, temperature and conductivity were recorded. During this
period Paradise Dam was raised approximately 8 m, increasing reservoir capaci-
ty from 7.9 X 10 ^ mMo 14.2 X 10^ m^. The purpose of this study was to
monitor changes in these limnological parameters which could result from the
reservoir enlargement as well as the continued increase in the human population
in the watershed. If significant change occurred. Paradise Irrigation District
would be able to initiate corrective measures to prevent further deterioration.

' Accepted for publication November 1982.

RESERVOIR LIMNOLOCICAL PARAMETERS

69

Note: drawn from USGS Quadrangle
Map - Paradise East, Calif.

FIGURE 1. Location of Paradise and Magalia Reservoirs in northern California and location of
sampling stations.

METHODS

Water samples were collected from the top and bottom metre of the water

column at three stations in Paradise and two stations in Magalia Reservoir

(Figure 1). These were frozen until analysis could be done. Organic nitrogen

was determined by the total Kjeldahl method followed by nesslerization ( APHA

70 CALIFORNIA FISH AND CAME

1975). The nessler method is sensitive to 20 jag/l ammonia nitrogen. Nitrate
nitrogen was determined by the Brucine method (APHA 1975), which is recom-
mended for concentrations from 0.1-1 mg NO3 — N/l. Total nitrogen was ob-
tained by adding the organic nitrogen results and the nitrate-nitrogen results.
Data for 1972 was taken from the DWR reservoir study.

Total phosphorus was determined by a persulfate digestion followed by the
ascorbic acid method (APHA 1975), which is sensitive to concentrations as low
as 10 jag P/l.

Phytoplankton were collected at one station in each reservoir with a Wiscon-
sin net pulled vertically through the water column (1974, 1978, 1980) or with
a Van Dorn water sampler at 5 m intervals (1972, 1976). Samples were pre-
served until they could be counted. Organisms/I were converted to volume of
cells/ml by measuring linear dimensions and using appropriate formulas.

Dissolved oxygen, temperature and electrical conductivity were measured at
one station in each reservoir at 1 m intervals throughout the water column with
a Hydrolab IIB meter.

All sampling was done every 2 weeks during the spring. Station locations
corresponded as closely as possible to those used by DWR in their 1971-73
study.

RESULTS

Total Nitrogen and Phosphorus

Total nitrogen was always less than 0.5 mg/l-N but generally showed an
increase from 1972 to 1976 and then generally decreased from 1976 to 1980
(Figure 2). No trend was discernible between surface and bottom waters or
between the two reservoirs, although the overflow from Paradise runs into
Magalia.

Total phosphorus concentrations were similarly quite low (Figure 3), with
levels increasing to their highest in 1976 (0.1-0.35 mg/l-P) and then decreasing
to less than 0.05 mg/l-P by 1980.

For all years investigated, total nitrogen and phosphorus tended to be depleted
in surface waters and stayed the same or increased slightly in deepest waters in
both reservoirs.

Volume of Phytoplankton

The variation in volume of phytoplankton paralleled the variation in nutrient
concentration (Figure 4). Phytoplankton volume was lowest in 1974 and 1980
and highest in 1976.

Phytoplankton composition varied over the eight years of study. In 1972 the
dominants were flagellates, but the DWR study did not identify them further. In
1974, the phytoplankton of both reservoirs was dominated by Asterionella for-
mosa with Dinobryon sertularia occasionally being numerous. D. sertularia was
most abundant in both reservoirs in 1976, reaching 143 x 10* juVml in Paradise,
and 69 X 10* jaVml in Magalia. In 1978 this phytoplankter often constituted 99%
of the population, but the volumes were reduced from 1976. In 1980, D. sertu-
laria again was predominant, constituting over 50% of the volume. The max-
imum volumes were similar to what they were in 1974.

RESERVOIR LIMNOLOCICAL PARAMETERS

71

Total \itrogcn
(mg/1 - N)
1980

0.10

0.00

0.30

0.20

Total Nitrogen
fing/1 - N)
1978

0. 10

0.00

\^

-- P-B

^>c:rir_r

— M-B

^

— . ^

p- ^

^

^

N.

^ — ^-^^^

^ ^~~"~~-^>^

^^ P-S

^ M-S

, M-B

0.30

Total Nitrogen
(rag/l - N)
1976

0.20

0.10

Total Nitrogen
(mg/1 - N) 0.05
1974

0.00

Total Nitrogen
(mg/1 - N)
1972

0.10

0.00

P-S

.M-B

.P-S, P-B, M-S

M-B

P-S = Paradise Surface
P-B = Paradise Bottom
M-S - Magalia Surface
M-B = Magalia Bottom

P-S

Mar

Apr

May

June

FIGURE 2. Variation in total nitrogen during the spring in surface and bottom water samples from
Paradise and Magalia Reservoirs, 1972-1980.

72

CALIFORNIA FISH AND CAME

Total Phosphorus
(mg/1 - P)
1980

0.10

0.00

M-B~.

Total Phosphorus
(mg/1 - P)
1978

0.05

0.00

Total Phosphorus
(rag/l - P)
1976

.30

.20

.10

0.00

Total Phosphorus
(mg/1 - P)
1974

0.05

0.00

P-S, P-B, M-B

M-S

0.10

Total Phosphorus
(mg/1 - P)
1972

0.00

M-S, M-B
Mar

Apr

Mav

P-S = Paradise Surface
P-B = Paradise Bottom
M-S = Magalia Surface
M-B = M;ir:i1 la Bottom

June

FIGURE 3. Variation in total phosphorus during the spring in surface and bottom water samples
from Paradise and Magalia Reservoirs, 1972-80.

Blue-green algae was totally missing from the phytoplankton in 1978 and 1980,
but had been present in small quantities previously.

Physical Parameters
Variation in temperature, oxygen and conductivity were documented for 1972
and 1978 (Figure 5), and were representative of the entire period of study.

RESERVOIR LIMNOLOCICAL PARAMETERS

73

PtaMlMkl

in

(«V«1)

Shis'

NMilU MMWVtff

Phrta^Unktw

VoluM of
PhvloplBnkton

Phvlopll
I MO

Ayr M»r Jm^

Apr my Jun«

Miy JuB«

FIGURE 4. Variation in the volume of phytoplankton in Paradise and Magalia Reservoirs during
the spring (1972-80).

Both reservoirs were generally isothermal during most of the spring, but
became thermally stratified in May or June. This persisted until August or Sep-
tember.

Paradise Reservoir has a multiple outlet system with releases at the bottom,
6.5 m, and 15 m above the bottom. Average water column temperatures are
generally higher in Paradise Reservoir because colder water is released through
the bottom outlet and flows into Magalia Reservoir.

Oxygen levels were generally at saturation throughout the water column, but
during the two study years when summer conditions were examined (1972 and
1974), levels dropped to less than 50% saturation. Occasionally levels dropped
to < 1 mg/l near the bottom in mid-summer.

Electrical conductivity was consistently low throughout the study period:
50-70 fimhos/cm. It increased slightly during the summer.

74

CALIFORNIA FISH AND GAME

C

00

c

oc

c

3

-o

o
>

01

B

<U

a

3

4J

c

Id

60

n<

F?

.^^

S

H

00
03

■D

C

T3

n3

m

0)
Q.

E

"O

c

ro
C
00

:^

o

D

o

X E

CO

O.

I J

>>

F

X

0)

LU

o

l-

O

o

o

o

r^

rg

at
BO

a

e
a.

u

j:

D.'-

3

E

— u

-D

c

zx

0

o

OO D.

u

U

X E

X 0)

ro

LU

o (-

IJ

o

o o

o

rj f^j

CnI

U

o;

<D

c

c

o

o

O

•^

o

^H

03

•^

^.

fT3

>

s

o

■ —

tfl

E

o

a.

X

CL'--

g

^u

:i

o

r^

C ^^

(_)

X 4)

•w-*

LU

O H

4X

O

o o

tf)

O

rg rj

rj

s

Id

a.

CTl

O

^

o

1— '

'^

•H

t^

o.

■a:

o

o

(U
Q

1

RESERVOIR LIMNOLOGICAL PARAMETERS

75

>!

o

o
o

o

rg

O

CM

o o
o ^

e
J.

c
o

a

) a
. E

i-

(U

on

a

I)

>.

F

X

11

UJ

o

H

o

o

O

o

CM

rg

^r--

E

3. '^

n.
ao

o
o

a a

76

CALIFORNIA FISH AND CAME

DISCUSSION

It has often been observed that newly formed reservoirs begin in an enriched
or highly productive state resulting from the decomposition and subsequent
nutrient release of recently flooded land {Visser1970, Mitchell 1973). However,
this may be offset by increased flushing or shortened retention time. Taylor
(1971 ), in a study of six TVA reservoirs, found reservoirs with a longer retention
time tended to be more productive, possibly because nutrients were changed
to more available forms. Nicola and Borgeson (1970) observed that a high rate
of flushing retarded nutrient accumulation. This is particularly critical to produc-
tivity in reservoirs with outlet structures that draw water from the hypolimnion,
as is the case in Magalia Reservoir.

In Paradise and Magalia Reservoirs, the increased flushing during a period of
high-rainfall years appears to have reversed any trend toward eutrophication.
Rainfall data was collected at Paradise Reservoir by DWR. 1972 and 1976 were
relatively dry years (annual total < 76.2 cm) compared to 1974, 1978 and 1980
(annual total > 127 cm) (Figure 6). The increase in the level of trophic state
indicators, nutrient concentration and phytoplankton volume observed in 1976
was reversed in 1978. By 1980 the reservoirs were approaching their 1974
trophic condition (total nitrogen < 0.10 mg/l-N, total phosphorus < 0.05
mg/l-P, phytoplankton < 2.0 x lOVVml).

Monthly
Rainfall
( Inches)

at
Parad i sc
Rescrvo : r

20.0

10.0

Annual Totals (Inches)

1972

29.24

1974

84.53

1976

26.75

1978

74.79

1980

64 . 57

0.0

Jan

Feb

Mar

Apr May

FICURE 6. DWR Precipitation Record at Paradise Dam for selected years

June

RESERVOIR LIMNOLOGICAL PARAMETERS 11

Long term precipitation data show that the mean value of precipitation falling
on the watershed is approximately 1 17.8 cm/yr. During the eight years of study
there has been considerable variation in total annual rainfall (67.9 cm — 215.3
cm ). Rain in this mediterranean climate occurs almost exclusively from October
-April. The monthly averages for this watershed are listed in Table 1 (Paradise
Irrigation District records for 56 years).

TABLE 1. Average Monthly Rainfall In The Little Butte Creek Watershed (Paradise Irrigation
District — 56 years of record converted to cm)

Month Average rainfall (cm)

Jul 0.18

Aug 0.40

Sep 1.47

Oct 6.98

Nov 15.34

Dec 21.67

)an 24.56

Feb 20.47

Mar 16.71

Apr 10.26

May 3.91

]un 1.65

In these lakes, the amount of rain in the winter months determines the lake
limnological characteristics during the spring months. Abnormally high rainfall
in one month has the effect of bringing in nutrients, which sets the scene for a
large phytoplankton population in spring. If heavy rains occur two months in a
row, the effect is to flush nutrients out of the reservoirs and low concentrations
of nutrients support low phytoplankton populations. This can best be seen in the
1976 and 1980 results.

In 1976, over 15.2 cm of rain fell in February. The amounts of total nitrogen
and phosphorus were high both in surface and bottom waters (Figures 2 and
3). Peak phytoplankton populations occurred in April and May (Figure 4).

In 1980, 37.1 cm of rain fell in January followed by 36.3 cm in February.
Surface nutrient levels were very low and the spring phytoplankton volume was
3-^ orders of magnitude smaller.

In nutrient-poor lakes such as these, and in climates where rainfall varies so
radically, care must be taken in applying general limnologic and trophic state
indicator concepts. It is imperative that study of trophic change be long term.
An important note for future water use planning is that during dry years, when
water demand will be highest, water quality will be lowest.

LITERATURE CITED

American Public Health Association. 1975. Standard methods for the analysis of water and wastewater, 14th
Edition, APHA, Washington, D.C.

California Department of Water Resources. 1973. Magalia Reservoir watershed limnology and water quality study.
Calif. Dept. Water Resources, Northern District.

Mitchell, D.S. 1973. Supply of plant nutrient chemicals in Lake Kariba, p. 165-169 in W.C. Akermann, C.F. White,
E.B. Worthington, Eds., Man-Made Lakes; Their Problems and Environmental Effects. Geophysical Monogr.

17

Nicola, S.J. and D.P. Borgeson. 1970. The limnology and productivity of three California coldwater reservoirs. Calif.

Fish Came, 56 (1): 4-20.
Taylor, Mahlon. 1971. Phytoplankton productivity response to nutrients correlated with certain environmental

factors in six TVA reservoirs, p. 209-217 //vC.E. Hall, Ed., Reservoir Fisheries and Limnology, Amer. Fish. Soc.

Special Publ. No. 8.

Visser, S.A. 1973. Pre-impoundment features of the Kainji Area and their possible influence on the ecology of the
newly formed lake, p. 590-595 in W.C. Akermann, C.F. White, E.B. Worthington, Eds., in Man-Made Lakes:
Their Problems and Environmental Effects. Geophysical Mongr. 17.

78 CALIFORNIA FISH AND CAME

Calif. Fish and Came 70 ( 2 ) : 7&-88 1 984

COMPARATIVE CONDITION OF BLACK-TAILED DEER,

ODOCOILEUS HEM ION US COLUMBIANUS,
IN TWO HERDS IN TRINITY COUNTY, CALIFORNIA ^

JOHN C. KIE

USDA Forest Service

Pacific Southwest Forest and Range Experiment Station

2081 E. Sierra Ave., Fresno CA 93710

TIMOTHY S. BURTON

California Department of Fish and Game

PO Box 1480, Redding, CA 96001

AND

JOHN W. MENKE

Department of Agronomy and Range Science

University of California, Davis, CA 95616

Measures of condition, health, and reproductive status were compared in black-
tailed deer from the Weaverville and Hayfork herds in Trinity County, California, in
winter 1979-60. Weaverville deer showed significantly lower bled and eviscerated
carcass weights, significantly lower reproductive synchrony, apparently lower repro-
ductive rates, and apparently higher frequency and abundance of parasites. No
differences in kidney fat, femur marrow fat, mean date of conception, and fetal
weight gain were detected between herds. The data suggest Weaverville deer are
in poorer condition than Hayfork deer.

INTRODUCTION

Black-tailed deer, Odocoileus hem/onus columbianus, is an important wildlife
resource in Trinity County ( Figure 1 ) . The Weaverville herd, of which 80 to 90%
is migratory (U.S. Fish and Wildlife Service 1975), occupies about 362,500 ha.
Summer and transitional range used by the Weaverville herd occupies about
90% of the total area, and winter range about 10% (Burton, unpublished).
Vegetation consists of coniferous forest, predominantly Douglas-Fir, Pseudot-
suga menziesii, and ponderosa pine, Pinus ponderosa. Digger pines, P. sabiniana,
occur at lower elevations on drier sites. Interspersed within the coniferous forest
are patches of hardwoods such as California black oak, Quercus kelloggii; Ore-
gon white oak, Q. garryana; interior live oak, Q. wislizenii; and madrone, Ar-
butus menziesii, as well as shrub species such as wedgeleaf ceanothus,
Ceanotfius cuneatus; lemon ceanothus, C. lemonii; deerbrush, C integerrimus;
manzanita, Arctostaphylos spp.) silk-tassel, Carrya fremontii; and mountain ma-
hogany, Cercocarpus betuloides. The Hayfork herd is also migratory (Figure 1 ).
It occupies about 270,000 ha, of which about 84% is summer and transitional
range and 16% is winter range (Dunaway 1966). The vegetation on the seasonal
ranges used by the FHayfork herd is similar to that on Weaverville deer ranges,
with the addition of chamise, Adenostoma fasciculatum, as a winter range shrub.

This research was supported by the Pacific Southwest Region and Shasta-Trinity National Forest, USDA Forest
Service, and the Trinity River Basin Fish and Wildlife Task Force. Work was done while senior author was with
the Department of Agronomy and Range Science, University of California, Davis. Accepted for publication
December 1982.

TRINITY COUNTY DEER CONDITION

79

Trinify
and Lewisfon
Reservoirs

FIGURE 1. Weaverville and Hayfork deer herds in Trinity County, California.

Between 1960 and 1963, about 6,980 ha in Trinity County were inundated by
the filling of Trinity (Clair Engle) and Lewiston Reservoirs. Much of this area was
deer winter range used by the Weaverville herd. The result was a decline of
between 4,000 and 6,000 deer (U.S. Fish and Wildlife Service 1975, Kie et al.
1982). As part of an ongoing study, aspects of deer population dynamics (Kie
et al. 1982), potential for mitigating deer habitat loss (Kie et al. 1980), and deer

80 CALIFORNIA FISH AND GAME

response to habitat improvement programs (Kie and Menke 1980) have been
previously reported. The purpose of this research was to compare measures of
condition, health, and reproductive status betv^een the Weaverville deer herd
and the Hayfork deer herd. Only the former was directly affected by reservoir
construction.

METHODS

To assess condition, health, and reproductive status of deer, a systematic
collection program was undertaken during winter 1979-80. Ten Weaverville
deer and 9 Hayfork deer were collected between 16 December 1979 and 5
February 1980, and 11 Weaverville deer and 10 Hayfork deer were collected
between 21 March 1980 and 22 April 1980. Deer, mostly adult does, were shot
with a rifle from a vehicle. Date, time, location, habitat type and browse species
present were recorded. Each day when the collection was finished, all deer were
processed. Bled carcass weight (BCW) was measured first. An examination for
external parasites was made before the body cavity was opened. Eviscerated
carcass weight (ECW) was measured after the internal organs had been
removed, but with the head and hide still attached.

Both kidneys were removed and weighed with and without their capsule of
connective tissue and the attached fat. From these weights, the kidney fat index
(KFI) was calculated (Anderson and Medin 1965, after Riney 1955). One femur
was removed and later processed for femur marrow fat (FMF) by the oven-dry
method (Neiland 1970). Adrenal glands were fixed in 10% formalin, trimmed
of excess connective tissue, and weighed to determine paired adrenal gland
weight (ADRWT).

Presence and abundance of internal parasites were noted. Levels of both
external and internal parasites were recorded as negative, light, medium, or
heavy (Salwasser and Jessup, unpublished).

The number of fetuses, and the weight and hind foot length (HFL) of each,
were recorded. Fetal sex was recorded when it could be determined on a
morphological basis. The following regression of age against HFL, developed for
mule deer in the central Sierra Nevada (Salwasser and Holl 1979), was used to
determine the age of each fetus:

Estimated age in days = 68 + 0.59 (HFL in mm)

Ovaries were fixed in 10% formalin, and sectioned with a razor blade. Corpora
lutea of pregnancy were counted (Cheatum 1949a).

A 1-pint rumen sample was removed and frozen. Results of rumen analyses
are reported elsewhere (Kie etal. In Press). A primary incisor was removed and
sent to Matson's Microtechnique (Milltown, Montana) for age determination
based on cementum annuli. (Trade names and commercial enterprises or
products are mentioned for information only. No endorsement by the U.S.
Department of Agriculture is implied.)

Deer were classified as fawns between 0 and 1 1 months of age, or as adults
12 months and older. Adults were further classified as yearlings (12-23 months)
or mature deer (24 months and older).

To determine sources of variation, values for BCW, ECW, KFI, FMF, and
ADRWT from 33 does were subjected to an analysis of covariance, using the
BMDP statistical program P2V (Dixon 1977). The analysis considered the main

TRINITY COUNTY DEER CONDITION

81

factor effects of herd (two levels) and collection period (two levels), the effects
of herd X period interactions, and the effects of a covariate (age in years). This
procedure required the assumption of a linear relationship between the covari-
ate and each dependent variable. To satisfy this assumption, log ,o BCW, log lo
ECW, and log 10 age were used in the analysis of covariance of BCW and ECW.
Arcsin transformations were used with KFI and FMF values to ensure uniform
residual variances and additivity of treatment effects (Gilbert 1973).

RESULTS AND DISCUSSION

Age

The age distributions of collected deer were similar between the Weaverville
and Hayfork herds (Figure 2). The difference between herds in the relative
proportion of deer less than 2.5 years old reflects sampling selectivity and is not
believed to represent actual age distributions. In the older age classes, deer 2.5
years old were most common in the Weaverville collection, and deer 3.5 years
old were most common among the FHayfork sample. FHowever, sample sizes
were too small to infer age structure in the populations at large.

10

^ 6
I 4

0

R

In fi

El Weaverville
D Hayfork

In

in i 11

JL

0.5

1.5

2 5 3,5

7 b 8,5

9.5

10 +

4 5 5.5 6.5
Age (years)

FIGURE 2. Age distribution of deer collected from the Weaverville and Hayfork herds.

Body Weight

Both BCW and ECW (Table 1 ) were significantly higher (p < 0.05 and p <
0.10, respectively) among Hayfork deer than among Weaverville deer. BCW
and ECW did not differ between collection periods. Mature Weaverville deer
appeared to lose body weight between collection periods; mature Hayfork deer
did not (Table 1 ); however, this trend was not statistically significant (no herd
X period interaction in BCW and ECW) . As expected, BCW and ECW increased
significantly (p < 0.01) as deer became older.

Dressed weight of white-tailed deer, C virginianus, is a sensitive indicator
of physical condition ( Park and Day 1 942 ) . Deer on a high nutritional plane are
heavier than those on a low nutritional plane (Leopold et al. 1951, Severinghaus
1955, Robinette et al. 1973). The significantly lower BCW and ECW values for
Weaverville deer in this current study are thought to reflect poorer range condi-
tions and a chronically lower nutritional plane than those of Hayfork deer.

2—78005

82

CALIFORNIA FISH AND CAME

■Si

<T3

r\ro oo Ln(Ti cd"^

</>

c
u

5

,—

■a

rn

in

(T'

o

r^

o^

•^

o

rsi

0^

o

^

rsi

N

a.

Si

o o

rO O

T— rsi Tj-

p
rsi rS

C

_o

"o
U

I
I

CO

•a

o

r-)

CO r-i

CO

1-^

■^

CO

sD

O

o

rs|

rsi

■— cT>

rs|

rsi

2i

c

.J

c
c

LU

I/)'
O

G

3

■o

<

_3
>

■o
c

.2P

c

1

<

§

u

CO

~ — '

.^

OD

2

5

00

§

u

op
'(1)

OD >\ DD *^

.E c:

o m -s

(U S

oo

S CO

io ^ 5 -5 ^ 2

^

5

Q
<

DC

T3

C

DD

<U n3 : Ol

_ OJ C : H

DD ^ OD
Q-

J

TRINITY COUNTY DEER CONDITION 83

Fat Values

All KFI values in these collections were less than 25%, except a value of 41%
for a doe 8.5 years old in the early Weaverville collection. This value was
included in the mean (Table 1 ), but excluded as an outlier in the analysis of
covariance results. KFI and FMF did not differ significantly between herds.
Within herds, both fat values decreased between collection periods (p < 0.05
for KFI and p < 0.01 for FMF). There were no herd X period interactions in
KFI and FMF. KFI and FMF values decreased as mature deer became older (p
< 0.05). However, yearling deer tended to have lower KFI values than did
mature deer as a group (Table 1 ).

The KFI, developed for red deer, Cervus elaphus, in New Zealand, is valuable
as an indicator of comparative nutrition between populations (Riney 1955). FMF
decreased in white-tailed deer in New York during starvation, leaving the mar-
row gelatinous (Cheatum 19496). As physical condition in deer declines, kidney
and other visceral fat stores are mobilized before marrow fat is depleted (Harris
1945, Riney 1955, Pojar and Reed 1974). In this current study, the relationship
between KFI and FMF is identical to that for mule deer, O. hemionus, in Colo-
rado (Pojar and Reed 1974). FMF values less than 60% were observed only
when KFI was less than 15% (Figure 3). Connective tissue around the kidneys
may comprise 5 to 10 percentage points of the observed KFI; therefore, KFI
values less than 1 5% represent minimal kidney fat deposition. Most or all kidney
fat was mobilized in deer in declining condition before significant marrow fat
stores were lost.

Adrenal Gland Weight

ADRWT did not differ between herds; however, deer from the late collection
had significantly higher ADRWT values than those collected earlier ( p < 0.05 ) .
Also, ADRWT increased significantly with increasing age (p <0.01), a trend
similar to that reported in mule deer (Anderson, Medin, and Bowden 1974)
(Table 1).

Adrenal gland size increased with an irruption and die-off of sika deer, Cervus
nippon, in Maryland, but only after a very high population density had been
reached (Christian, Flyger, and Davis 1 960) . Adrenal gland weight was inversely
correlated with KFI in black-tailed deer in Tehama County, California (Hughes

and Mall 1958); no such relationship was found in this study.

Reproduction
Reproductive rates of does collected during early winter were based on
corpora lutea of pregnancy. Few embryos were readily observable at that time,
although most does had conceived by late December. Fetal rates were used as
a measure of reproduction during late winter because all pregnant does collected
were well into gestation. Reproductive rates were lower among Weaverville
does than among Hayfork does. Mean reproductive rates were 0.00 fetuses per
doe for two Weaverville yearlings and 0.75 fetuses per doe for four Hayfork
yearlings (Table 2 ) . Three of the Hayfork yearlings were carrying one fetus each.
All mature does collected were pregnant; reproductive rates were 1.39 and 1.50
fetuses per mature doe for Weaverville and Hayfork deer, respectively. Fetal
rates differed between herds because more Hayfork does were carrying twin
fetuses.

84

CALIFORNIA FISH AND CAME

100|-

-^80

c
o

u

k.

a

- 60
o

o

o

3

E

0)

40

20

O

o

o

t o

o •o

0 o

^0 •

0 o
o

8°

#Early collection
O Late collection

10 20 30 40

Kidney fat index (percent)

50

FICURE 3. Relationship between kidney fat index and femur marrow fat for Weaverville and
Hayfork deer.

Reproductive rates for Trinity County deer are similar to those for other
black-tailed deer herds in northern Califo; nia. Fetal rates for mature black-tailed
does from Glenn, Tehama, and Siskiyou counties in California ranged from 1.54
to 1 .73 embryos per doe ( Bischoff 1 958) . In Lake County, 1 .45 fawns per mature
doe were produced in good quality, mixed-shrubland habitat, but only 0.71 ;
fawns per mature doe were produced in dense, chamise chaparral habitat I
(Taberand Dasmann 1958). The lower mean fetal rates among Weaverville deer \
in this study may indicate poorer range condition.

Fetal sex ratios and fertilization rates did not differ between herds. Of the 27 i
fetuses examined from does collected during late winter, 12 (44%) were male.
These 27 fetuses corresponded to 28 corpora lutea of pregnancy, for a 96%
fertilization rate. For other black-tailed deer in northern California, fetal sex ratios
ranged from 44% to 64% males, and fertilization rates from 90% to 94%
(Bischoff 1958, Taber and Dasmann 1958).

TRINITY COUNTY DEER CONDITION 85

TABLE 2. Mean Reproductive Rates by Age of Doe, and Conception Date Information for
All Fetuses (Sample Size in Parentheses).

lVeavervi//e Hayfork

Reproductive rate (fetuses per doe)'

Yearling does 0.00 ( 2) 0,75 ( 4)

Mature does 1.39 (18) 1.50(10)

Conception date

Mean 9 Dec (15 fetuses) 9 Dec (12 fetuses)

Standard Deviation 14.9 days 5.6 days

Range 22 Nov-2 Jan 1 Dec-19 Dec

• Ovulation rates substituted for fetal rates for early collection.

Mean conception date, 9 December, was the same for both the Weaverville
and Hayfork herds (Table 2). This is 9 days earlier than previously reported (Kie
and Menke 1980) because a different equation was used to estimate fetal age.
Most black-tailed does in Siskiyou and Tehama Counties bred in December;
however, farther south in Glenn and Lake Counties, peak breeding occurred in
November (Bischoff 1957, Taber 1953). Reproduction was significantly (p <
0.01) less synchronous in Weaverville deer (Table 2). The greater variability
may be related to poor nutritional intake. Verme (1965) reported that white-
tailed does fed a reduced quantity of rations exhibited delayed, irregular breed-
ing; although no delay in mean date was noted in this current study.

TABLE 3. Frequency of Occurrence {i) and Mean Level* of Abundance {\) of Selected
Parasites.

Weavervi//e Hayfork

(n = 21) (n ^ 19)

Parasite f_(%) X_ n%) X_

Ectoparasites

Ticks 86 1.00 74 0.74

Keds 48 0.57 42 0.42

Other ectoparasites (fleas, lice, mites) 38 0.67 5 0.05

Endoparasites

Tapeworm cysts 81 1.05 68 0.68

Nasal botfly larva 29 0.62 47 0.68

Hydatid cysts (canine tapeworm) 10 0.14 5 0.05

• Levels: 0 = negative, 1 = light, 2 = medium, 3 = heavy for all parasites except hydatid cysts, where level equals

actual number found per deer.

One Weaverville doe in the late collection was estimated to have conceived
as late as 2 January, but all does in the late Hayfork collection had bred by 19
December. Late breeding by Weaverville deer could have negatively biased
estimates of their reproductive rates because the estimates were based, in part,
on does from the early collection. However, all adult does in the early collection
had ovulated, except for one Weaverville yearling. Therefore, small sample sizes
in this study present greater interpretive problems than such biases.

Fetal weight was expressed as a function of fetal HFL to show patterns of
weight gain for Weaverville and Hayfork fetuses (Figure 4). No differences in
fetal weight gain between herds were noted. Therefore, a common allometric

86

CALIFORNIA FISH AND CAME

equation (Y === aX ^) was fitted by converting both weight and HFL to base 10
logarithms, and using linear, least squares techniques (Baskerville 1972). The
estimated values were: a = 0.03649 (SE = + 0.01319, - 0.00969), b = 2.072
(SE = ± 0.071 ), r' = 0.97, and n = 27.

lOOOr-

20 40 60 80 100 120

Fetal hind foot length (mm)

140

FIGURE 4. Relationship between fetal hind foot length and fetal weight for early and late collec-
tions.

Parasites •

Ticks {Dermacentorspp., Ixodes spp.); keds, Lipoptena depressa, and other
ectoparasites, such as fleas, lice, and mites, were all found with greater fre-
quency and in greater abundance among Weaverville deer (Table 3). Tape-
worm cysts ( Taenia spp.) and hydatid cysts (canine tapeworm, Echinococcus

TRINITY COUNTY DEER CONDITION 87

granulosus) were also more frequent and abundant in deer collected from the
Weaverville herd. Only nasal botfly larva, Cephenemia jellisoni, were detected
with greater frequency in Hayfork deer. However, among those Weaverville
deer parasitized by botfly larva, numbers of larva were greater; therefore, mean
levels of abundance were only slightly different between herds. These differ-
ences in parasite occurrence and abundance between herds may be related to
differences in dietary quality (Olson 1974). Despite close inspection, no evi-
dence was found of either liver flukes, Fasciola hepatica, or lungworms {Dic-
tyocaulus spp ) .

CONCLUSIONS

The construction of Trinity and Lewiston Reservoirs resulted in a loss of 6,980
ha of winter range used by the Weaverville herd. Small sample sizes make
interpretation of any single parameter difficult; however, when viewed as a
complex of variables, the data indicate that Weaverville deer are in poorer
condition. Variables either indicated this trend with statistical significance
(BCW, ECW, and reproductive synchrony), indicated the trend but were not
tested for statistical significance (reproductive rates, and frequency and abun-
dance of all parasites except nasal botfly larva), or supported a null hypothesis
of no differences in condition between herds (KFI and FMF). These differences
may be related to lower dietary quantity or quality, or both, among Weaverville
deer. Only the frequency and abundance of nasal botfly larva suggested that
Weaverville deer were in better condition. Taken as a whole, the data support
the hypothesis of poorer condition among Weaverville deer, which may be
related to the loss of winter range. If the loss of deer habitat in Trinity County
is to be mitigated, ongoing Weaverville winter range improvement programs
should be continued, improved, modified, and expanded, where feasible (Kie
and Menke 1980).

ACKNOWLEDGMENTS
We thank the Trinity River Basin Fish and Wildlife Task Force and the USDA
Forest Service for funding and administering this study. We also thank D. Jessup,
P. McLaughlin, G. Monroe, B. Pyshora, T. Ramsey, D. Smith, T. Stone, S. Thomp-
son, and D. Yount of the California Department of Fish and Came; and H. Black,
Jr., P. Brouha, T. Brumley, J. Okula, R. Pedersen and H. Salwasser of the USDA
Forest Service for field assistance and advice. C. Evans and J. Verner of the USDA
Forest Service offered critical comment on an earlier draft of this manuscript.

LITERATURE CITED

Anderson, A. E., and D. E. Medin. 1%5. Two condition indices of the Cache La Poudre nnule deer herd and their

application to management. Colo. Came, Fish, Parks, Came Info. Leaf. 23, 3 pp.
Anderson, A. E., D. E. Medin, and D. C. Bowden. 1974. Growth and nrxjrphometry of the carcass, selected bones,

organs, and glands of mule deer. Wildl. Monogr., 39, 122 pp.
Baskerville, C. L. 1972. Use of logarithmic regression in the estinriation of plant biomass. Can. J. For. Res., 2:49-53.
Bischoff, A. I. 1957. The breeding season of some California deer herds. Calif. Fish Came, 43:91-%.

1958. Productivity in some California deer herds. Calif. Fish Came, 44:253-259.

Cheatum, E. L. 1949d. The use of corpora lutea for determining ovulation incidence and variations in the fertility

of white-Uiled deer. Cornell Vet., 39:282-291 .

19496. Bone marrow as an index of malnutrition in deer. New York Slate Conserv., 3:19-22.

Christian, ). |., V. Flyger, and D. E. Davis. 1960. Factors in the mass mortality of a herd of sika deer, Cervus nippon.

Chesapeake Sci., 1:79-95.

88 CALIFORNIA FISH AND GAME

Dixon, W. |., ed. 1977. BMDP, Biomedical computer programs, P-series. Univ. Calif. Press, Berkeley, 880 pp.
Dunaway, D. 1966. A habitat management plan for the Hayfork deer herd unit. U5DA Forest Service, Shasta-Trinity

National Forest. 16 pp. plus illustr.
Gilbert, N. 1973. BiorDetrical interpreution. Clarendon Press, Oxford, 125 pp.
Harris, D. 1945. Symptoms of malnutrition in deer. J. Wildl. Manage., 9:319-322.

Hughes, E., and R. Mall. 1958. Relation of the adrenal cortex to condition in deer. Calif. Fish Game, 44:191-1%.
Kie, |. G., and J. W. Menke. 1980. Deer response to Trinity River habitat manipulation, a cooperative study. Univ.

of Calif. Davis, USDA Forest Service, Calif. Dept. Fish and Game, Trinity River Basin Fish and Wildlife Task

Force, 1 58 pp.
Kie, J. G., ). W. Menke, |. R. David, and W. M. Longhurst. 1980. Mitigating the effects of reservoir development

on black-tailed deer in Trinity County. Cal-Neva Wildl. Trans., 27-40.

Kie, ). G., T. S. Burton, and J. W. Menke. 1982. Deer populations and reservoir construction in Trinity County,

California. Calif. Fish Game, 68(2):109-117.
Kie, J. G., T. S. Burton, J. W. Menke, and W. E. Crenfell, Jr. 1984. Food habits of black-tailed deer, Odocoileus

hemionus columbianus, in Trinity County, California. Calif. Fish Game, In Press.

Leopold, A. S., T. Riney, R. McCain, and L. Tevis, Jr. 1951. The Jawbone deer herd. Calif. Dept. Fish and Game,
Game Bull (4) 139 pp.

Nieland, K. A. 1970. Weight of dried marrow as indicator of fat in caribou femurs. J. Wildl. Manage., 34:904-907.

Olson, O. W. 1974. Animal parasites. Their life cycles and ecology, 3rd ed. Univ. Park Press, Baltimore, Maryland,

562 pp.
Park, B. C, and B. B. Day. 1942. A simplified method for determining the condition of white-tailed deer herds in

relation to available forage. USDA Tech. Bull. 840, Washington, D.C., 60 pp.

Pojar, T. M,, and D. F. Reed. 1974, The relation of three physical condition indices of mule deer. Colo. Game,

Fish, Parks, Game Info. Leaf. 96, 4 pp.
Riney, T. 1955. Evaluating condition of free-ranging red deer (Cervus elaphus), with special reference to New

Zealand. New Zealand J. Sci. and Tech., 366:429^63.

Robinette, W. L., C. H. Baer, R. E. Pillmore, and C. E. Knittle. 1973. Effects of nutritional change on captive mule
deer. J. Wildl. Manage., 37:312-326.

Salwasser, H., and S. A. Holl. 1979. Estimating fetus age and breeding and fawning periods in the North Kings
deer herd. Calif. Fish Game, 65:159-165.

Severinghaus, C. W. 1955. Deer weights as an index of range conditions on two wilderness areas in the Adiron-
dack region. New York Fish and Game J., 2:154-160.

Taber, R. D. 1953. Studies of black-tailed deer reproduction on three chaparral cover types. Calif. Fish Came,
39:177-186.

Taber, R. D., and R. F. Dasmann. 1958. The black-tailed deer of the chaparral. Calif. Dept. Fish and Came, Game
Bull. (8) 163 pp.

U.S. Fish and Wildlife Service. 1975. Deer loss compensation program resulting from Trinity River Division,
Central Valley Project, California; A Report to the Trinity River Basin Fish and Wildlife Task Force. U.S. Fish
and Wildlife Service, Portland, Oregon, 30 pp.

Verme, L. J. 1965. Reproduction studies on penned white-tailed deer. J. Wildl. Manage., 29: 74-79.

SEA OTTER REPRODUCTION AND PUPS 89

Calif. Fish and Came 70(2) : B9-^00 ]')M

PUP DEPENDENCY PERIOD AND LENGTH

OF REPRODUCTIVE CYCLE: ESTIMATES

FROM OBSERVATIONS OF TAGGED

SEA OTTERS, ENHYDRA LUTRIS,

IN CALIFORNIA^

FREDERICH E. WENDELL

California Department of Fish and Came

213 Beach Street

Morro Bay, CA 93442

JACK A. AMES

California Department of Fish and Came

2201 Carden Road

Monterey, CA 93940

AND

ROBERT A. HARDY

California Department of Fish and Came

213 Beach Street

Morro Bay, CA 93442

Pup dependency periods and length of the reproductive cycle for sea otters,
Enhydra lutris, were estimated from observation of tagged individuals in California.
Pup dependency periods were estimated to range between five and eight months
and reproductive cycle lengths to range between 11 and 14 months. These are shorter
than most previous estimates.

INTRODUCTION

Pup dependency periods and reproductive cycle lengths for sea otters, Enhy-
dra lutris lutris, were estimated from observations of tagged individuals in Cali-
fornia. Concern for the health of the sea otter population and concern over the
loss of shellfish fisheries through sea otter predation have provided the impetus
for considerable research effort. The observation of a large number of tagged
individuals was part of this research effort.

There are only two estimates of the sea otter pup dependency period and
reproductive cycle length based on observations of tagged individuals. Using
data from three sea otters tagged in Prince William Sound, Alaska, Johnson and
Jameson (1979) estimated that for some individuals the reproductive cycle
length was one year and the pup dependency period was six months or longer.
Loughlin, Ames and Vandevere (1981 ) reported estimates from observations of
two tagged females through several cycles. They estimated the pup dependency
period to be 8 to 8.5 months and the reproductive cycle to be approximately
one year. These estimates were considerably shorter than most of those gener-
ated from indirect methods. The sample sizes, however, were very small.

Indirect methods of estimating pup dependency period and length of the
reproductive cycle utilized population growth characteristics or data from
female urogenital tracts. Lensink (1962) felt the wide difference in the propor-
tion of pups in the Amchitka Island and Kanaga Island population could be

' Accepted for publication April 1983.

90 CALIFORNIA FISH AND CAME

accounted for by possible differences in the frequency that females bear young.
He suggested that Kanaga Island females bred every year while Amchitka Island
females gave birth every other year. After examining data obtained from study-
ing female urogenital tracts collected in Alaska, Chapman estimated the length
of the reproductive cycle to be two years (Kenyon 1969). Schneider (1971)
using similar methods also suggested the average reproductive cycle was proba-
bly two years; he assumed that there was a 12-month pup dependency period.
Kenyon (1969) also felt the normal period of dependency for Alaskan sea otter
pups was at least a year. Vandevere (1972) thought the period was less than
one year, based on an observed decline in the ratio of dependent young to adults
in the months prior to the peak in pupping. Barabash-Nikiforov et al. (1968)
reported a dependency period of six to seven months for Enhydra lutris
gracilis.

This report is our estimate of the length of the pup dependency period and
reproductive cycle for sea otters in California based upon observations of tagged
pups and adults. We believe our results, from observing a large number of tagged
animals, represent the best data available for sea otters in California.

MATERIALS AND METHODS

The California Department of Fish and Game has tagged sea otters since
September 1977 under a cooperative agreement with the U.S. Fish and Wildlife
Service. Two hundred and thirty-nine sea otters were captured and tagged with
a highly visible colored plastic tag on each hind foot between August 1977 and
September 1981 (Ames, Hardy and Wendell 1983). The reproductive cycle
length and pup dependency period estimates were derived from observations
of 139 tagged adult females and their pups, including two tagged previously by
Loughlin (1978).

Sightings were made using a variety of observation equipment. Most observa-
tions were made using a 50X-80X Questar telescope. Observations were made
throughout the sea otter's range (Figure 1). Observer effort, however, was
greatest toward both ends of the range.

Estimating Pup Dependency Period

A sea otter pup's period of dependency begins with its birth and ends with
separation from its mother. Weaning was assumed to occur at the time of
separation. Observing the birth and weaning of a pup, however, is essentially
impossible. Estimates of that period were derived from sightings of identifiable
mothers and pups. The length of a pup's period of dependence was inferred
using three methods. The three methods differed by the way the day of birth
was inferred.

By Observation of Mother/ Pup

The last date a female was sighted alone and the subsequent sighting of that
female with a pup brackets the pup's birth date. The date of weaning is bracketed
by a reverse sequence of sightings. An ideal series of sightings would bracket
both the birth and weaning of a pup within narrow time limits. Two dependency
periods (minimum and maximum) were produced for each individual using this
method (Figure 2).

SEA OTTER REPRODUCTION AND PUPS

91

Pt Ano Nuevo

N

Tag

Tagg

ed

Area

M F

Pups

1

20 0

0

2

14 43

6

3

0 4

2

4

1 20

2

5

1 9

0

6

6 5

4

7

7 36

4

8

15 10

1

9

25 4

0

Morro Bay

Pt Sal

FIGURE 1. Number of sea otters tagged by area, August 1977 to September 1981.

By Natal Pelage

At birth the sea otter's fur or natal pelage has a characteristic wooley appear-
ance. The transition from natal to adult pelage, although somewhat variable,
occurs approximately 10 wks after birth (R. Jameson, USFWS, pers. commun.).
Observation data collected on tagged mothers and pups included the observer's
identification of the pup as a small (wooley) or large pup. These observations
were used to approximate the pup's date of birth. Birth was assumed to have
occurred 10 wks prior to the last identification of the pup as a wooley. This
approximation of the date of birth was used to reduce the maximum length of
the pup dependency period. The determination of the minimum length was not
affected.

By Weight at Capture

Growth rates were used to convert weight at capture to age at capture for pups
caught during tagging operations. The period from the estimated birth date to
the date of capture was added to the period from capture to weaning to estimate
the maximum pup dependency period. The growth rates used in this conversion

92

CALIFORNIA FISH AND CAME

were derived from observations of two tagged females and their pups. Observa-
tions bracketed the birth dates for these pups to within a 6- and a 14-d period.
The pups were subsequently captured when they weighted 6.8 and 4.1 kg,
respectively. By using a 1.8 kg birth weight (Kenyon 1969) both pup's growth
rates were approximated ( Figure 3 ) . The most conservative growth rate of 0.058
kg/d for the female and 0.061 kg/d for the male pup from this approximation
was used in converting capture weights to age at capture.

SIGHTING
DATE

o o o

00 CO 00

I I I

iX> <^ '—'

.— I C\J oo

I I I

r-~ r~~ r~-.

O

00

I

1 — I

O-J

I

CO

DEPENDENCY PERIOD
maximum

o

CO

I

I

r

minimum

o

00

I

O

CO

I

CO

o

00

I

"=!■

C\J

I

00

I
I

CO

00

I

00
I

I

00

I

00
I

I

00

CO

I

I
en

00
I

ID
I

o

FEMALE WAS A
A = alone

with small pup

W

sm

AAAAWWWWWAAAA
sm sm sm Ig Ig

ig

= with large pup

minimum
•maximum

W
sm

REPRODUCTIVE CYCLE

FICURE 2. Example of a sighting record showing the minimum and maximum pup dependency
period and a reproductive cycle length.

HI

oc

3

1-

Q.

^^^

<

(A

O

>.

*rf

Q

re

T3

•-i_i-^

UJ

o

<

FICURE 3.

100

80

60

40

20

female 0.058 kg/day

12 3 4 5 6 7
WEIGHT at CAPTURE (kg)

8

Crowth rates of two wild sea otter pups (male and female) based on an assumed 1.8
kg birth weight, a birth date known to within a few days (14 and 6 days) and a
subsequent weight at capture (4.1 kg 23 days after first sighting and 6.8 kg 80 days after
first sighting). The growth rates used to convert weight at capture to age at captu
were the slowest extremes.

re

SEA OTTER REPRODUCTION AND PUPS 93

With each method of estimation, the length of the interval between the
minimum and the maximum dependency period for individual pups varied
considerably. Since the actual period was only known to occur somewhere
between these widely varying minimum and maximum periods a graphical
display was used to summarize the data. This display was based on the assump-
tion that each half-month interval between the minimum and maximum ob-
served lengths of dependency had an equal likelihood of containing the actual
dependency period. Summing comparable half-month intervals for all pups
produced a peak or mode which represents the most common period of de-
pendency for the data group.

Estimating Reproductive Cycle Length

The length of the reproductive cycle for sea otters in California was also
estimated using tag sighted data. The birth of a pup, the most readily identified
event in the reproductive cycle, was used to mark the beginning and end point
of a cycle. The use of tag sighting data allowed only an approximation of the
actual length of the reproductive cycle based on minimum and maximum
lengths. Two data groups, segregated by the method of determining the date of
birth of a pup, were used. These data include only those sightings where there
was a potentially successful weaning of the first pup. To be conservative, five
months was used as the minimum time sufficient to allow for a successful
weaning.

By Observation of Mother/ Pup

The dates of birth of both pups in the cycle were inferred using observations
of the female alone, followed by a subsequent observation of the female with
a pup (Figure 2).

By Weight at Capture

If a known-weight pup was the first pup in the cycle, the estimated interval
between date of birth and capture was added to the period between capture and
the first sighting of the next pup for the maximum length of the reproductive
cycle. The minimum length was determined solely from sighting data.

A graph indicating the frequency distribution of comparable half-month inter-
vals was also used to summarize the data on the length of the reproductive cycle.

RESULTS
Estimating Pup Dependency Period

By Observation of Mother/ Pup

Twenty-nine series of sightings of tagged females had the observations neces-
sary to determine a minimum and maximum period (Table 1 ) . Analysis of these
sighting data indicated a 6.5 to 7-month mode ( Figure 4) . The shortest maximum
length was slightly over four months. The longest minimum length was slightly
over eight months.

Six of the 29 series of sightings were more precise, with minimum observed
lengths close to the maximum (within 60 days). All of the pups had dependency
periods which were somewhere between 3 and 7.5 months in length. Summing

94

CALIFORNIA FISH AND CAME

comparable half-month intervals for this subgroup, produced a broad peak
between four and seven months (Figure 4). The longest minimum was about
seven months while the shortest maximum was almost four months.

By Natal Pelage

Thirteen of the 29 minimum and maximum lengths generated from observa-
tion data were modified using information on the loss of natal pelage (Table 1 ).
The mode produced from an analysis of these minimum and maximum lengths
was between 5.5 and 8 months (Figure 5). The longest minimum was slightly
over eight months while the shortest maximum was 5.5 months.

TABLE 1. Minimum and Maximum Pup Dependency Periods for Sea Otters Derived from
(1) Observations of Tagged Individuals, (2) Modified by Observations of Pelage
or, (3) Observations and Weight at Capture.

Tag

Tag

^

Area

Minimum-Maximum
(months)

Method

#

Area

Minimum-Maximum /
(months)

Vfethc

1*

1

4.7- 5.3

23*

2

3.1- 5.4

1

2

2

5.7- 7.2

3.1- 5.5

.2

3*

2

6.0-10.5

24

2

5.8-14.9

1

4

2

3.3- 4.1

5.8-12.0

2

5

2

6.9- 7.5

25*

2

3.5- 7.2

1

6

2

4.2- 6.0

3.5- 5.6

2

7

4

1.^10.9

26

2

5.8-12.9

1

8

5

2.0- 8.7

5.8- 9.2

2

9'

6

1.4- 9.0

27*

4

1.1-10.8

1

10

7

0.1- 9.8

1.1- 7.8

2

n*

7

1.9- 5.0

28*

7

1.3-12.6

1

12

7

4.1- 7.4

1.3-11.8

2

13*

7

1.3-10.5

29

8

1.0- 7.1

1

14*

7

4.9- 8.2

].0- 6.9

2

15

7

6.3- 6.7

30

2

7.7- 8.9

3

16

7

1.5-12.0

31

2

6.3- 7.2

3

17*

2

5.5- 8.2

32

2

4.7- 5.4

3

5.5- 7.7

33

2

5.5-10.9

3

18*

2

1.8- 8.5

34

2

1.4- 6.3

3

1.8- 5.5

35

2

0.3- 5.7

3

19*

2

6.4-12.9

36

2

5.0- 5.7

3

6.4- 9.3

37

2

6.3- 6.8

3

20

2

7.7-13.0

38

4

4.8- 7.0

3

7.7- 9.1

39

4

7.8-10.3

3

21*

2

5.9-11.1

40

4

7.&- 8.1

3

5.9- 8.3

41

6

6.8- 7.3

3

22*

2

8.2-15.9
8.2- 9.5

* pup born at least 1 year after tagging or second pup if captured with a pup

By Weight at Capture

Twelve of the pups captured and weighed during the tagging effort were also
resighted at the key points necessary to estimate a length of pup dependency

SEA OTTER REPRODUCTION AND PUPS

95

(Table 1). The mode was between five and seven months (Figure 6). The
longest minimum was almost eight months and the shortest maximum was 5.5
months.

n = 29

POTENTIAL MONTHS OF DEPENDENCY

FIGURE 4. Frequency distribution of potential periods of pup dependency within half-month
intervals derived from observations of tagged individuals. Subset (darkened area)
derived from observations where minimum and maximum periods were within 2
months.

96

CALIFORNIA FISH AND CAME

Li.
O

if)

lU

O

z

LU
OC
CC

Z)

CO o

o
o

en

I- Q

13 5 7 9

POTENTIAL MONTHS OF DEPENDENCY

FICURE 5. Frequency distribution of potential periods of pup dependency within half-month
intervals derived from observation of tagged individuals where the maximum period
was reduced by observation of loss of natal pelage.

O

UJ

o

z

LU
CC
S OC

D 3

c/) o
o
o

Q
LU
O

o

<

C/)

-J
<

9

>

Q

5 1
3 -

n-

= 12

1 1 1

1 1

1 -

1

^/^/^V/

, vy//////,///^/';-7/////,y. .

]

13 5 i

max -

^ 9

I]

11

13

1 r T- n

15

ZH

^^^^^^^^^

Si^B^

ff^

^mwipM^^n""^

1

! ' 1 '

— ' — 1

3 5 7 9 11 13

POTENTIAL MONTHS OF DEPENDENCY

15

FIGURE 6. Frequency distribution of potential periods of pup dependency within half-month
intervals derived from observation of tagged individuals where the maximum period
was adjusted by determining a date of birth from weight at capture. Subset (darkened
area) represents portion where tagged pup was known to have been sucessfully
weaned.

SEA OTTER REPRODUCTION AND PUPS

97

Tagged pups, if they were subsequently observed as apparently healthy inde-
pendent otters, were the only individuals for which weaning was known to be
successful. Of the twelve tagged pups, five were known to be successfully
weaned. A mode for this subgroup occurred between 5 and 7.5 months (Figure
6). The longest minimum was slightly over 7.5 months and the shortest max-
imum was almost six months.

Estimating Reproductive Cycle Length

By Observation of Mother/ Pup

Seventeen series of sightings had the observations necessary to determine
minimum and maximum lengths of the reproductive cycle (Table 2). An analysis
of these data indicated a mode at 12.5 to 13 months (Figure 7). The shortest
maximum was slightly over 1 2 months and the longest minimum was almost 20.5
months.

TABLE 2. Minimum and Maximum Reproductive Cycle Lengths for Sea Otters Derived from
(1) Observations of Tagged Individuals or (2) Observations and Weight at Cap-
ture.

Minimum ^

•iaximu

(months)

20.0

24.3

12.2

13.5

10.4

14.0

7.7

12.6

6.8

20.5

7.4

14.7

18.7

20.1

18.3

21.6

12.8

18.8

11.3

13.1

8.1

18.7

8.1

18.3

20.4

26.4

14.4

22.1

12.9

19.7

10.1

12.1

9.8

17.6

8.2

24,4

12.4

26.8

9.5

16.2

11.4

13.9

7.0

14.1

9.2

19.1

6.5

15.5

1.1

13.9

8.1

11.8

Method

By Weight at Capture

Nine additional series of sightings used weight at capture to estimate the date
of birth of the first pup in the cycle (Table 2 ) . The modal estimate of the length
of the reproductive cycle was 11 to 14 months (Figure 8). The shortest max-
imum was just under 12 months and the longest minimum was slightly over 12
months.

98

CALIFORNIA FISH AND GAME

11

en

uj q

O ^
(/) o

o
o

CO

_i

Q <
UJ 3
O Q

H Q

Z

3

n = 17

3

11 13 15 17 19 21 23 25

max 1

1
1

1

^^^

1
1

1

^

i^^m

1

1

I ■ 1 • I • I ■ 1 • 1

„_ J

7 9 11 13

POTENTIAL MONTHS IN REPRODUCTIVE CYCLE

FIGURE 7. Frequency distribution of potential lengths of sea otter reproductive cycle within half-
month intervals derived from observations of tagged individuals.

U)

UJ

o

Li.

z

o

UJ

cc

s

oc

13

■D

(/)

o

o

o

9 1

7
5
3
1

Cfl

_i

D <
UJ 3
O Q

<i

I- Q

V,,

^

"1

n = 9

H

L

1

7

9 1

— m X

-1 .»

1

1

13

■ ■ 1

15

17

19

21

23

25

27

— min —

1
1

^Mllll

■

1

1

1
1

1

■ 1

^^^^^^^^H

1 ' 1 ' f '

T "f

' 1 ' 1

-

1 ' T •

I ' I '

' ( ■

1 ' 1 '

1 ' 1 '

r • I

9 11 13 15 17 19 21 23 25

POTENTIAL MONTHS IN REPRODUCTIVE CYCLE

27

FIGURE 8. Frequency distribution of potential lengths of sea otter reproductive cycle within half-
month intervals derived from observations of tagged individuals where the date of birth
of first pup was determined from weight at capture.

DISCUSSION

Pup Dependency Period
The three methods of estimating pup dependency period produced similar
estimates, all within the range of five to eight months. There were no appreciable
differences in the estimated length of pup dependency for pups captured or born

SEA OTTER REPRODUCTION AND PUPS 99

shortly after capture and that estimated for subsequent pups or pups born at least
one year after capture. This strongly suggests that the capture or tagging opera-
tion did not influence the duration of the mother-pup bond. Since successful
weanings occurred near the extremes of this five to eight months range, pup
dependency periods are at least as variable as the range suggests. Although
shorter than most estimates, our results are similar to two reported estimates
(Johnson and Jameson 1979; Loughlin, Ames and Vandevere 1981 ).

The disappearance of some tagged females with pups, which had shown only
limited movement for five to eight months, suggests that a marked shift in
location may, in some instances, be part of the weaning process. The observa-
tions of one of these, a known age female, also provided the first direct measure
of age at first reproduction. She was slightly over four years old at the birth of
her first pup, indicating she probably became sexually mature at three years of
age. This agrees closely with the suggested age of sexual maturity for some
female otters in Alaska (Kenyon 1969).

Reproductive Cycle

Both estimates of the length of the reproductive cycle fall within the range of
11 to 14 months. This range is significantly shorter than most estimates based on
indirect assessment methods. They are, however, similar to the estimates report-
ed by Johnson and Jameson ( 1 979 ) and Loughlin, Ames and Vandevere ( 1 981 ) .
Four of the minimum lengths were appreciably longer than the 11 to 14 month
mode. These females could be losing their second pup prematurely without
observer detection, thereby adding a gestation period to what was assumed to
be a normal cycle. It is also possible they could have experienced a long delay
in implantation, lengthening the total gestation period.

These results suggest that in California annual reproduction is normal and that
the reproductive potential of the population may be twice as high as generally
accepted. Since direct observational studies in Alaska and studies on E.I. gracilis
in the western Pacific Ocean have also shown annual reproduction occurring,
we would suggest that annual reproduction is the norm throughout the range.
The expression of this reproductive potential in terms of real growth of the
population depends on age specific reproduction and mortality. Based on obser-
vations of identifiable mothers which apparently lost pups the mortality rate for
pups may be quite high. Continued observational effort and additional tagging
of dependent animals could eventually yield composite life tables and an indica-
tion of age specific reproductive rates. Efforts to manage otter populations and
to resolve resource conflicts involving sea otters need to be based on these types
of reliable population dynamics data.

ACKNOWLEDGMENTS

A considerable amount of effort was expended gathering the observational
data utilized in this report. We wish to thank A. Baker, L. Belluomini, S. Benech,
J. Bodkin, E. Faurot, L. Ferm, M. Flippo, B. FHatfield, R. Jameson, D. Miller, C.
Ribic, L. Smith, F. Sorenson, J. Vandevere, B. Van Ness, and C. Woodhouse for
their contribution in this effort.

LITERATURE CITED

Ames, ). A., R. A. Hardy, and F. E. Wendell. 1983. Tagging materials and methods for sea otters, Enhydra lutris.

100 CALIFORNIA FISH AND GAME

Barabash-Nikiforov, I., S. V. Marakov and A. M. Nikolaev 1968. The kalan or sear otter. Nauka Press, Leningrad.

184 p.
Johnson, A. and Jameson, R. 1979. Evidence of annual reproduction among sea otters. Page 31. in Abstr. 3rd

Biennial Conf. Bio. Marine Mammals, 7-11 October 1979.
Kenyon, K. W. 1969. The sea otter in the eastern Pacific Ocean. No. Am. Fauna 68, 352 p.
Lensink, C. 1962. The history and status of sea otters in Alaska. Univ. Microfilms Limited, Ann Arbor, Mich., 188

p.
Loughlin, R. R. 1978. A telemetric and tagging study of sea otter activities near Monterey, California. Nat. Tech.

Info. Serv., PB-289 682. p 1-64.

Loughlin, T. R., J. A. Ames and J. E. Vandevere 1981. Annual reproduction, dependency period, and apparent
gestation period in two California sea otters, Enhydra lutris. Fishery Bull. 79(2); 347-349.

Schneider, K. B. 1 971 . Reproduction in the female sea otter. Pittman-Robertson Job Prog. Rpt. W-1 7-4, 8. 1 1 R. Alaska

Dept. Fish and Game, p 1-38.
Vandevere, J. E. 1972 Behavior of southern sea otter pups. Proc. Ninth Annu. Conf. Biol. Sonar and Diving

Mammals, p 21-35.

WILDLIFE HABITAT REMOTE SENSING 101

Calif. Fish and Game 70 ( 2 ); 1 02-1 1 2 1 984

A REVIEW OF SELECTED REMOTE SENSING AND

COMPUTER TECHNOLOGIES APPLIED TO WILDLIFE

HABITAT INVENTORIES^

KENNETH E. MAYER

Forest Resources Assessment Program

California Department of Forestry

1416 9th Street

Sacramento, CA 95814

Remote sensing and computer technologies, as applied to wildlife habitat invento-
ries, are reviewed. Subjects covered include, aerial photography (film types, scales
and formats), Landsat data (multispectral scanner, return-beam vidicom camera,
and thematic mapper), and geobased computer information systems (technical
prerequisites and applications). Moreover, tables are provided which address the
applicability of remote sensing data for assessment of habitat diversity elements;
application of remote sensing data to various project sizes; and the relationship
between wildlife models, levels of inventory and the appropriate remote sensing
systems.

INTRODUCTION

An important objective of wildlife management is to identify, map, assess, and
manage vegetation assemblages as habitat for wildlife. In the past, these efforts
have been hampered by inaccessible terrain, the lack of relatively inexpensive
inventory tools which can deal with large ecological units, and the absence of
appropriate management systems for data storage, management and retrieval.
Fortunately, techniques have evolved from remote sensing and computer tech-
nologies which offer solutions to these problems (Anderson, Wentz, and Tread-
well 1980).

Rapid advancements in the remote sensing and computer fields have made
it very difficult to keep abreast of the technology. Because of new sensors with
greater spectral sensitivity and resolution, and changing computer philosophy
and sophistication, the wildlife biologist is now offered a variety of options from
which to choose. The purpose of this paper is to provide insight to these various
options and offer a clear picture of some of the more commonly accepted
techniques as well as more sophisticated systems which have application to
wildlife habitat inventories.

TECHNOLOGY REVIEW

Remote sensing is a relatively new name for an old concept. Simply stated,
remote sensing is the process of acquiring information about a subject without
actually coming into contact with it. Remote sensing, as an operational habitat
inventory system, can produce useful information about a variety of habitat
elements at different scales.

Most remote sensing systems require electromagnetic (EM) energy, gener-
ated by the sun. Electromagnetic energy travels in a harmonic, sinusoidal fashion
at the velocity of light in various wavelengths which can be received by special-
ized EM sensors. Thus, many different types of information can be provided by
remote sensing. EM sensors commonly used in natural resource inventories are
operated from airborne and/or spaceborne platforms.

' Accepted for Publication May 1983.

102 CALIFORNIA FISH AND GAME

Aerial Photography

Film Type

Aerial photographs have provided data to resource scientists for more than
100 years. There are many types of films which are sensitive to various wave
lengths of EM energy. By examining different combinations of wavelengths one
can highlight important attributes of a given resource. The most commonly used
film types are as follows:

(i) Black and White Film. Sensitive to ultraviolet radiation (0.3 to 0.4 ja
m) and that part of the EM spectrum visible to humans (0.4 to 0.7 /x
m).
(ii) Blackand White Infrared. Sensitive to ultraviolet radiation (0.3 to 0.4
jam), visible (0.4 to 0.7 jam), and the near-infrared portion of the EM
spectrum (0.7 to 0.9 jam),
(iii) Color Film. Sensitive to the visible spectrum (0.4 to 0.7 jam). The

three primary colors are blue, green, and red.
(iv) Color Infrared (CIR). Sensitive to green, red, and infrared light (0.5
to 0.9 jam). CIR is commonly called false color because during the
photographic process blue images result from objects reflecting pri-
marily green energy, green images result from objects reflecting pri-
marily red energy, and red images result from objects reflecting
primarily in the photographic infrared portion of the spectrum. There-
fore, the shades of red that appear on the photograph depict vegeta-
tion reflecting different intensities of infrared light. Note that infrared
refers to reflected EM energy and not emitted energy as in heat ( Figure

1).

Photo Scale

Resolution and photo scale are important factors in selecting the proper data
type for wildlife habitat inventories. One of the most commonly used photo
scales is 1 :1 5,840. This scale is used extensively by the United States Department
of Agriculture Forest Service (USES). A shift from use of 1:15,840 to 1:24,000 is
currently underway to make the photographic scale more compatible with 7.5"
(1:24,000) quadrangle maps which are employed by many resource agencies.
Small scale photography (large numbered fraction, 1:100,000) covers a large
ground area, usually with low resolution (detail determination ) . Therefore, large
scale photography usually offers higher resolution but covers a smaller ground
area. The selection of an appropriate scale is dependent on information require-
ments. Thus it is economically important to select the smallest scale that will
provide the detail required to complete the job. Furthermore, photo cost is
directly proportional to the number of photos taken, film type and format (trans-
parency and contact print) used. Selecting the proper photo criteria is very
important, as trade-offs between criteria can determine the success of the inven-
tory project. For a complete review of remote sensing principles, films and
formats refer to Avery (1977), or Lillesand and Kiefer (1979).

WILDLIFE HABITAT REMOTE SENSING

103

0.9

en
s:

«V E

0.8

0.7

0.6

0.5

0.4

0.3

B
L

A

C
0

C
K

L
0
R

0

0(

w

I
N

B

H
I
T

F
R

A

L
K

E
I

C
0
L

R
E
D

N

0

&

F

R

R

W

A

H

R

I
T

E

E
D

'

Film tijpo.
FIGURE 1. Approximate wavelength sensitivities of four common aerial films

Landsat Satellites
The Earth Resource Technology Satellite (ERTS) Program, now referred to as
Landsat 1, 2, and 3, has been developed by the National Aeronautics and Space
Administration (NASA) to provide a practical approach to resource manage-
ment. Currently, the National Oceanic and Atmospheric Administration
(NOAA) is responsible for the Landsat Satellites. The Landsat program vi'as
initiated in 1972 with the lauch of ERTS-1, followed by Landsat-2 in 1975,
Landsat-3 in 1978, and Landsat-4 in 1982. Landsat is an earth-viewing satellite
which operates in a sun-synchronous, near-polar orbit, at an approximate aiti-

104 CALIFORNIA FISH AND CAME

tude of 917 km. Each image (scene/frame) encompasses roughly a square,
slightly skewed geographical area approximately 184 km on a side, which covers
a land area of approximately 3.5 million ha. The image is recorded in four
wavelength regions (bands) of the EM spectrum 0.5 to 1.1 jam, two are in the
visible and two are in the near-infrared portion of the light spectrum. Photo-
graphic film is not carried on the satellite. Instead, the satellite records data with
an electro-optical device called a multispectral scanner (MSS). The amount of
light reflected from the earth is recorded numerically in each wavelength band.
Landsat data are basically point sources of information. The MSS scans across
the scene and divides the lines of data collected into rectangular units called
picture elements (pixels). As the forward progress of the satellite sets up the
sensors, another scan is made, element by element, line by line. Each pixel
covers approximately 0.46 ha on the ground, and thus forms the lower size limit
for the smallest feature that can be resolved by the scanner. Morever, the MSS
light reflectance values are averaged across each pixel. Images of the same
portion of the earth's surface are obtained at 18-day intervals; two satellites in
operation allow for a nine-day repeat coverage.

A source of data provided by Landsat that is often overlooked is the return-
beam vidicon (RBV) camera. Landsat-3 carriers a two-camera RBV system. The
cameras produce two side-by-side images with high resolution. The EM energy
is stored on a photosensitive surface of the camera tube. After the shuttering is
complete the image is scanned by an electron beam to produce a video output
signal. The data from each camera are read out sequentially. Each of the images
covers a ground area of approximately 98 x 98 km. One MSS image coincides
with four RBV images.

As a continuation of the satellite program, NASA has recently launched a more
sophisticated, Landsat satellite, Landsat-4. Landsat-4 carries a relatively new EM
sensor, the thematic mapper (TM), which provides more spectral sensitivity
(narrower spectral bands) with greater resolution than the MSS. Like the MSS
the data products are output in a digital form for use within a computer or
displayed in a pictorial format. The TM spectral bands range from 0.45 to 12.5
jutm within the visible, infrared, and thermal parts of the EM spectrum (Table 1 ) .
In addition to the narrower spectral bands, the TM pixels have a higher resolu-
tion of approximately 30 m ^ or 0.09 ha. Currently, data from only two Landsats
are available to the general public.

TABLE 1. Thematic Mapper Spectral Sensitivity and the Application to Vegetation Mapping.

Band

TM 1
TM 2

TM3

TM4

TM 5
TM 6
TM7

Wavelength bands

Application to Mapping Vegetation

fxm

0.45 to

0.52

Sensitive to cholorphyll and carotinoid
concentration

0.52 to

0.60

Slight sensitivity to chlorophyll, plus
green light characteristics

0.63 to

0.69

Sensitive to chlorophyll

0.76 to

0.90

Sensitive to vegetation density and or
biomass

1.55 to

1.75

Sensitive to water in vegetational struc-
tures

2.08 to

2.35

Sensitive to soil minerals and water in
structures and soil minerals

10.5 to

12.5

Emitted thermal properties

WILDLIFE HABITAT REMOTE SENSING 105

Ceobase (Computer) Information Systems
The term Ceobase Information System (CIS) has been used in the past to
describe single-purpose graphic systems, geographic data base description tech-
niques, and information systems that retrieve, reformat, analyze, and display
geographically referenced information. To clarify this point, the term CIS can be
defined as a computerized resource data base that is derived from, and can be
related back to a ground base. The CIS includes a set of procedures for accessing
and modeling data in order to describe and possibly predict resource dynamics.
A system such as described above is composed of two types of information.
These types of data are (i) attribute, and (ii) graphic data. These data can be
accessed together or separately. The key benefit of a CIS is that any geographical
referenced data can be used. Information about habitat, wildlife species, re-
source dynamics, overlay data (soils, climate, erosion, etc.), as well as update
information, is appropriate for the system. A successful merging of ground
information and data derived from remote sensing can be completed with the
CIS context. Some features of a CIS include:

(i) A data base that can be indexed by various geographic projection
reference systems and linked with key words or attributes (e.g., the
location of peregrine falcon eyries by latitute/longitude or Universal
Transverse Mercator (UTM) projections),
(ii) Manipulation and analysis of the data base by (a) aggregating data
into various categories (e.g., aggregation of specific habitat require-
ments for bark gleaning birds); (b) displaying or listing aggregations
(e.g., maps or individual listings of wildlife occurrence information);
(c) tabulation of spatially related data (e.g., quantitative habitat
maps); (d) sorting, merging, filtering or arranging data layers (e.g.,
sorting cavity nesting birds by habitat type); and (e) performing mod-
eling of data by attribute value or by geographic location (e.g., wildlife
species/habitat relationships models linked to habitat inventories to
predict species occurrence).
(Hi) Display output information by geographic location and or selected
attribute(s) with cartographic accuracy (e.g., maps of critical habitat
areas for selected species),
(iv) The ability to quantitatively process spatially-related data as input to

mathematical predictive models (e.g., the prediction of animal num-
bers or probability of occurrence of mule deer in habitats of specified
management areas.

APPLICATION TO HABITAT INVENTORIES
Aerial Photography

Photo Scale

Selection of the proper photo scale for habitat inventories depends on infor-
mation requirements and the size of the project area. The inherent limitation of
each remote sensing system will require preproject planning in order to select
the most accurate (adequate detail) and cost-effective data for the inventory
problem (Tables 2 and 3).

106

CALIFORNIA FISH AND CAME

C 60
O C

•i o

01

>
I

■o

01

■o

»« L-

lA O

6C O

c a

■^ I

Q I

c =

V o

z i

60 fq

'c
c
ja

a.

■o
c

I

^ Ni

^

^ Ji

£5.

0 —

n-t ro

^

1

O "— rO T—

01

>^

V
N

.a c

•I!

a. 4,

.2 fl-

k

m •^

> E

vi: (A

m 01

BO >.

.E ^

£ S

0; O

o 01
£ Z

« c

•^ is

o 3

•j-S >^

" := S

°* 2 15

^ 01 »«

:5-o £

3 ® o

CM

^

.§

D "O

•— ro r^ >—
I

<U

.3

E
3

E

00

IV

-a

00

c

\/l

c

0)
(A

V

s

E
<1>

01
Q.

>-

oJ

0

?i

f^

0

t/1

E

(/»

0

^^

u

u

0

0>

Si

c

a;

t/i

y

n

'0

0

WILDLIFE HABITAT REMOTE SENSING

107

60

>

I

ST

o
o

Q.

V

3
V

o

c
e

u

>i

:2

-a

.5

§

^

•So

b:

C5

f^

e

OO^— <OOrsrs|i-0"^OrMr— .— OOrn

O O O <3 O rsl rsi

(A r>^

c
E

t:?

^s-

<:

^

igh flig
U-2

P

3:

cj

7^

Csir^mrO' — • — ' — ' — Orsirsi

>

z

c
E

I"'

1^

l'^

s?

rvi<*^rvi(-^i^(-»^n-iromr^rsirsjrsji — r<^

m r^ rn rsi rs (^

1 — 1 — r*^mm'^<~oro< —

3

^-5

Q
eo

'17 . .

c >.

0* •-

0:2
E £

« 2
«^ f^
o -o

as

.= 80

>■ o

.- \J

ea
<

0=; 0;

>

QO op

■^ ^ >-

^ 03 ^

^ — 0;

•~ y 9-

a 't: —

>-
n3

O -2

i7> O

E 2P

o 'oi

t

o "
>

OD
O

c ?
o

?^ ■■= -^

N

.1^

■B s

^ t/i (^

' t: "^ '^

5 Q. Q.

5:

03 O

S- Q. Q. -•-

■" 2 E -S

c
E

o aj

>- Q.

C

E

S £

— Q-

3 ■•^

=3 ™

2£

1 !^
5 ifl

— «

•^ 5
-5

S-E

c c

t £

o ^

i I

00 -

c i
n ns

q; I-

108 CALIFORNIA FISH AND CAME

Low-level (the relationship between focal length and flying altitude) aerial
photos (1:10,000) with a resolution of approximately 0.3m are best suited to
provide information on habitat diversity (vertical and horizontal layering) and
special habitat elements (Grenfell, Salwasser, and Laudenslayer 1982) for site
specific areas. Accurate measurements of these habitat parameters are possible
with high resolution, large scale, photography. Furthermore, project areas must
be small (4-41 ha) as the ground area covered by each photograph is small.

Medium-level aerial photos (1:10,000) can also be used at the project level
( <405 ha) to obtain a variety of habitat diversity information. For example, the
common USFS photos (1:15,840) can be used to identify tree species, crown
diameters, crown closure, tree height; and to develop habitat profile assess-
ments. Some data can be obtained about intrapatch variability, but the informa-
tion available is usually not as detailed as compared to information available
from low-level photography. One would expect a decrease in the ability to
identify and measure diversity parameters as you move to a smaller scale. This
does not always hold true. Many times, when the scale becomes smaller, more
of the habitat element becomes visible. This is especially true when patch shape,
size, and juxtaposition are important.

High altitude U-2 photography (1:30,000 to 1:120,000) has a resolution of
approximately three to five metres depending on the scale. The NASA 23 x 46
cm, 23 X 33 cm, and optical bar 13 x 25 cm photography can be used to derive
information about vegetation dominance, series, and plant communities. De-
pending upon the scale and resolution of the photography, high-altitude photos
can provide reliable information on horizontal habitat parameters.

An interesting example of this type of photography is optical bar (Befort,
FHeller, and Ulliman 1981 ). Within the last few years, the USFS has made exten-
sive use of optical bar photography. The U-2 aircraft can be equipped with a
panoramic camera (Itek KA80A) which is operated at approximately 19,760 m.
The image is a swath that covers a ground area approximately 69 km long, at
a view angle of 1 20°. This attribute provides a clear picture of the mosaic patterns
and relationships of habitats and associated landforms. Center photo strips of the
13 x 127 cm film format have a scale of approximately 1:30,000, with high
resolution (1-2 m ) . Photo strips that are on the ends of the image are of smaller
scale. Furthermore, the resource features on the outside portions of the strip
exhibit excessive lean; although, technology is available which will correct this
anomaly. Optical bar photography exists in transparency form, which makes
field use difficult without special equipment. A major asset of optical bar and
other types of high-altitude photography is the low cost-per-ground-area cov-
ered. Serious consideration should be given to this type of aerial data when
"extensive" habitat mapping projects are undertaken.
Film Types

Selection of the proper film type depends on the type of information desired.
In resource management, black-and-white and natural color are generally the
most commonly used film. Color infrared (CIR) can provide detailed informa-
tion about vegetation resources, especially when plant species are important. For
example, there is a similarity on color and black-and-white film between hard-
wood and coniferous trees. However, healthy hardwood types have higher
infrared reflectance than healthy coniferous species. Therefore, CIR film records

WILDLIFE HABITAT REMOTE SENSING 109

distinct differences between these tree groups. Infrared film is also very valuable
for detecting loss of plant vigor caused by a variety of factors. As an example,
plant water loss results in reduced infrared reflectance, while the visible portion
of the EM spectrum is unchanged.

Landsat

The ability of Landsat digital data to provide resource managers with intensive
inventory information has improved dramatically in recent years. Landsat images
are used in two basic forms. First, raw (unaltered) MSS and RBV digital images
are transformed into picture-quality images. Photo-interpretation of these data
can provide gross information about broad vegetation categories. Indentification
of large ecological units and associated ecoclines is the primary use of this form
of data. More recently, RBV and MSS data have been digitally normalized and
merged to create an image with optimum color and detail (Lauer and Todd
1981 ). A left-hand, stereo, conjugate image can be produced by merging topo-
graphic data with the RBV and MSS data sets. The stereo image produced
provides greater sharpness and resolution than MSS or RBV images alone. RBV
data have been available from the beginning of the Landsat program but few
scientists have taken advantage of the high resolution images. Photographic
interpretation of an MSS false-color composite has been shown to be valuable
for providing first-level information about forest and rangeland resources ( Heller
et al. 1975). However, by utilizing the merged MSS-RBV data, smaller features
(i.e., streets, stream channels, and associated riparian vegetation), as well as
different tree and shrub species groups, can be identified.

The most widely used form of Landsat data is the digital information obtained
from the MSS. Through the use of creative computer methodologies, a priori
classification probabilities and the incorporation of terrain data have offered the
biologist a species-specific look at large vegetative units. Digital classifications
are most often approached by three methods:

(i) Supervised approach. The analyst supervises the selection of spectral
classes by delimiting geographical areas (training areas) that represent
resource categories of interest. Multivariate clustering (grouping of
spectral classes) is not performed.
(ii) Unsupervised approach. In this approach, spectral classes are gener-
ated by clustering large numbers of pixels into natural groupings which
represent the spectral characteristics of the image (training areas are
not selected),
(iii) Guided clustering or multicluster fields. In this approach the analyst
selects training areas that represent various resources. Clustering is
performed on each training area to generate spectral classes unique
to the resource categories.

Of the three methods, guided clustering has been shown to be a consistently
reliable (a range from 70 to 90 percent overall accuracy) approach to the
classification process, providing species specific information on forest vegetation
(Fleming, Berkebile, and Hoffer 1975; Fox and Mayer 1979; Caydos and New-
land 1978; Mayer, Fox and Webster 1980; Walsh 1980).

Research by Craighead (1976), Fox, Mayer, and Forbes (1980); Laperriere et
al. (1980); Mayer and Fox (1981 ); Cannon, Knopf, and Pettinger (1982); and
others, has shown that by using the proper classification approach, detailed

110 CALIFORNIA FISH AND CAME

information about habitat types and structure covering large geographical re-
gions can be obtained. Moreover, the information exists as a prestratified, 0.405
ha sampling grid and is computer compatible. Landsat data, using an aggregated
approach, can provide vegetation information at the formation and series levels
(Parker and Matyas 1981, Paysen et al. 1980). It is important to remember that
Landsat is not a panacea. In order for the system to be effective, a multistage
sampling strategy should be used with Landsat providing information at only one
or two levels.

Geobase Information Systems

Wildlife species/habitat relationship models, surface cover, species inventory
data, remote sensing, terrain data, historical data, and other concomitant infor-
mation can be integral parts of a CIS. Table 4 is an example of multilevel,
descriptive models that relate to habitat type and wildlife species. Furthermore,
the remote sensing system applicable to each level is provided.

In order for a CIS to be applicable to wildlife inventories it must have the
ability to accept vector, raster, and tabular/textual data. Vector data refer to the
common digitizing procedure whereby line segments are created from polygon
data. Raster data are in a cellular or point-data format. For example, Landsat
digital data (raster data), tabular/textual information, and digitized wildlife spe-
cies occurrence information (vector data) can be processed together to pro-
duce qualitative and quantitative assessments.

The power of the CIS is in the evaluation of several data planes and informa-
tion levels. For example, a CIS can process soil, vegetation descriptions, remote
sensing (Landsat and/or air photo) data and information about species/ habitat
preference in one function to identify and spatially predict areas of management
concern. (Mead et al. 1981; Brass, Maw, and Peterson 1981). Furthermore,
output products can be listings of areas or attributes by management unit as well
as plotter maps of spatially oriented information with cartographic accuracy.

CONCLUSIONS

Some of the most important problems facing resource managers today deal
with human impacts on wildlands as a result of changing land use and land
management policies. "Sound" decisions about resource management can be
made only when accurate and adequate data are provided. Remote sensing and
CIS technologies are tools, which the biologist can use to examine resources at
various resolutions.

Aerial photographs provide an excellent base for resource inventories. Recent
advancements in computer and space technologies provide the biologist with
sophisticated tools to inventory, monitor, and predict the complex interactions
of ecological systems. Landsat MSS, TM, and the RBV are high-technology
systems that offer a complete view of associated ecosystems. Furthermore, with
the advent of innovative computer methodologies, Landsat digital classifications
can be completed with high accuracy at the pixel level. Lastly, computer soft-
ware and graphics hardware that have been developed to store, manipulate and
analyze georeferenced resource data offer increased accuracy for resource
management decisions.

WILDLIFE HABITAT REMOTE SENSING

111

c

-?

.O
o

.N

^ ,^
.5

.2

<3

T^ — J

_ .Qj

^.2

c
o

(LI

>
dl

Ol

>-
0

n3

i^

=J

, ■

>-

U

i^

w

i_l

0

0

m

u

C

5

$

>

0

0

O

>
Ol

P — >-

U O

o

a;

Ol

00

"O

c

0

ro

E

DC

E

t/>

'ly^

^

g.

C

-c

0

Ic

(U

t/1

<5

i

.0

a;
0

T3

E

0^

E

1

— 1

5

15

E
0

g

(LI
(-1

ro

ro

(_i

l_l

>$

C
<Li
IT)

OJ

Q.

03

5

0

Q.

(13

E

15

ro

03

5

0

5

u

0

0

a;

-0

CL

C

1

"f

u^

ro

1

1

_=>-->-

Ol

>

-a

n3

C
ro

O ro

" i

.E E

o

(LI

(LI
Q.

3 ■ —

-2 ^ E

1 - ^

d c o

\J

■—

^

u

l>^

<b

(LI

S

T3

■ — 1

(T3

"OJ

>

(LI

Ol

(D

ly^

(_) ^

ro »i ->,

3 O S

(-> U S

-5 X ^

O

^ 5

— ID
^ T3

Qj ro

>-
u
ro

>
(L;

Ol
Ol

O O ^

03 U c-sl

(LI o; >-

ro ro o

O) <Ll

o o

ll>

:: E

■"^ ^ !^ ro o

T3

^ OJ

ro

O '~>

>- Q.

ro £
^_Q ro
' O ^

^^

^- ro

CL '•^

ro >^
E o
c -^

o E

ro nj
-o >

Q-
E
-' T3 ro

03 ro
;E ro

O 2
C^ ro

O t=
^ ro

■,-, OJ

^ -^i

ro
X

c
ro

~9 =

(U (_)

5 o

a; -o
E O

0-1

ro

oc o
c ro

E

Ol

g_ (L. ro o o

— '^ S", u -^

a;

Oj T3

<Ll
>

C

3 ro

t- *_- 03

"a o

c ■"

ro (^

o E

ro ro

U^

C ^3

ro p

o

T3
Ol

o -^

^ .y

03 (-1

03 CL

(J i-ri

> ii

ro

ro Oy --

-^ E S

Q. O -^

=J ^ C

2 cb T

00 ^ ^

Oy (LI

(L> ro

Q- (LI

3
>-

ro

E

(LI ro

c
o

^ -t; y a* i' i

ro 42 y ^r y -S

(L>

Q..ti Q.

t/1 ij^ U^ <J^

'^

c
o

(LI

(U

>

(J
ro

\_

u

i-i

E

■LI

in ^

ro (_i ii

DC op >

X X ^

o

(LI

■o

>
dl
I

d)

(U

1/1

ro

E

3

, ■

0;

(-1

t^

(J

0

ro

O

LTl

OJ

o;

>.

ro

ro

0

t

OJ

d

-0

-0

CL>

0

0

>

rrt f^

-5 ^

&5

Q-15

lyi ro

Ol -£=

'^ "^

(LI O

°- 1=
'^ O

01 03
t; OC

vS; DO

ro

O C

•^ ro
>r c

ro -o
9- "H

ro ly^
O ro

ro o

O ro

31

Ol .ti

ro
15

"5 S
E "°

_ai
.0
ro

Q-

ro

o

3 Ol

Q--a

o E
>- ro

^ DO

-Q .£
ro j^

CL =3

O i:

03 L^

(LI

ro Qj

ro o

(LI

>

DO

<y^

ro

ro

ro
ni

-C

ro

ro

o a^ ro

-~ 3 c
_£■ ro • =

y '^ S
ro (^ 5

30 a

11
00

c
ro

ro OJ

^ ^ B

5 -Q ro

o ■> ._

-a E —

o 3 -

Q. E >r

i ^

ro

ro

•=■ 00
-Q =3

^ ^ 0

ro
Cl

ro
(_i

^ X

C O

O CL

oc ™

(LI >-

5b£

00 '^ c

E (LI >-

c 00 t

b ro (_i

-o - o-

C (LI "

- > c

Ol o o

3 >-_f2

:^rE 3

»— ^ n
ro

E cL 5.

■C; ^ Ol

_ lO Q. ILl

ro OJ O Q.

112 CALIFORNIA FISH AND CAME

ACKNOWLEDGMENTS

I would like to thank the members of the California Interagency Wildlife Task
Group for their support which was the driving force in the conceptualization of
this paper. Specifically I wish to thank J. Verner, Research Ecologist with the U.S.
Forest Service, Forest Sciences Laboratory in Fresno, California. His creative
editorial comments were invaluable. H. Salwasser and W. Laudenslayer of the
U.S. Forest Service, Region 5, offered encouragement and editorial support.
Their efforts are gratefully appreciated. Special thanks must be given to B.
Marcot, L. Fox III, and R. Gutierrez for their creative efforts in the development
of this manuscript.

LITERATURE CITED

Anderson, W. H., W. A. Wentz, and B. D. Treadwell. 1980. A guide to remote sensing information for wildlife
biologist. Pages 291-303 In S. D. Schemnitz, ed. Wildlife management techniques manual. 4th ed. Wildlife
Soc, Washington, DC 1-686.

Avery, T. E. 1977. Interpretation of aerial photographs, 3rd ed. Burgess Publ. Co., Minneapolis, MN. 392 p.

Brass, J. A., K. D. Maw, and D. L. Peterson. 1981. Forest policy and managment planning applications of Landsat

data within a geobased information system: The Santa Cruz demonstration. Final Report, NASA — ARC Moffett

Field, CA. 201 p.
Befort, W. A., R. C. Heller, and ). J. Ulliman. 1981. Ground resolution of high altitude photographs. Univ. Idaho.

Forest, Wildlife and Range Experiment Station, ISSN; 0073-4594. 4p.
Craighead, J. 1976. Studying grizzly habitat by satellite. National Geographic Magazine, 150(1) : 148-158.
Cannon, R. W., F. L. Knopf and L. R. Pettinger. 1982. Use of Landsat data to evaluate lesser prairie chicken habitats

in western Oklahoma. J. Wildl. Man.ige., 46(4) : 915-922.

Fleming, M. D., S. S. Berkebile, and R. M Hoffer, 1975. Computer aided analysis of Landsat-1 MSS: a comparison
of three approaches including a modified clustering approach. LARS Information Note: 072475. 9 p.

Fox, L, and K. E. Mayer. 1 979. Using guided clustering techniques to analyze Landsat digital data for mapping forest

land in northern California. Pages 364-367 in Proc. 5th Int. Symp. on Processing of Remotely Sensed Data.

Purdue Univ., LARS, Lafayette, IN., 1-468.
Fox, L., K. E. Mayer, and A. R. Forbes. 1980. Forest resource inventory classification of the McCloud Ranger District,

Mt. Shasta, CA; Using Landsat Digital Data. NASA Document. NSG 2341, 69 p.
Gaydos, L., and W. L. Newland. 1978. Inventory of land use and land cover of the Puget Sound Region using

Landsat digital data. J. Research, USGS, 6 (6): 807-814.

Crenfell, W., H. Salwasser, and W. Laudenslayer. 1982. The California wildlife-fish/habitat relationships system.

Cal-Neva Wildl. Trans. Heller, R. C, tech. coord. 1975. Evaluation of ERTS — data for forest and range surveys.

PSW Forest and Range Experiment Station, Publ. No. 112. 67 p.
Laperriere, A.]., P.C. Lent, W. G. Gassaway, and F. A. Nodler. 1980. Use of Landsat data for moose-habitat analyses

in Alaska. J. Wildl. Manage., 44(4) : 881-887.
Lauer, D. T., and W. J. Todd. 1981. Land cover mapping with merged Landsat RBV and MSS stereoscopic images.

Pages 68-89 in ASP-ACSM Fall Tech. Symp. ISSN 0271-4043, 1-629.
Lillesand, T. M., and R. W. Kiefer. 1979. Remote sensing and image interpretation. John Wiley and Sons Inc., New

York NY. 612 p.

Mayer, K. E., and L. Fox III. 1981. Mapping wildlife habitat on the McCloud Ranger District of the Shasta-Trinity
National Forest using Landsat digital data. PSW, USES, San Francisco, CA. 19 p.

Mayer, K. E., L. Fox III, and ). L. Webster. 1980. Forest Conditon mapping of the Hoopa Valley Indian Reservation
using Landsat data. Pages 217-242 in Proc. of 1st Int. Symp. of remote sensing for natural resources. Univ.
Idaho, Moscow, 1-515.

Mead, R. A., T. L. Sharik, S. P. Prisley, and j. T. Heinen. 1 981 . A computerized spatial analysis system for assessing
wildlife habitat from vegetation maps. Canadian J. Remote Sensing, 7(1); 34—40.

Parker, I., and W. Matyas. 1981. CALVEG: A classification of California vegetation. U.S.D.A., Forest Service
Regional Ecology Group, San Francisco, CA.

Paysen, T. E., ). A. Derby, H. B. Black Jr., V. C. Bleich, and J. W. Mincks. 1980. A vegetation classification system
applied to southern California. Gen. Tech. Rep. PSW-45, 33p.

Walsh, S.J. 1980. Coniferous tree species mapping using Landsat data. J. Remote Sensing of Environ., 9(1); 11-26.

SALT MARSH HARVEST MOUSE IDENTIFICATION 113

Calif. Fish and Came 70 (2 ): 1 1 3-1 20 1 984

IDENTIFICATION OF SALT MARSH HARVEST MICE,

REITHRODONTOMYS RAVIVENTRIS, IN THE FIELD

AND WITH CRANIAL CHARACTERISTICS '

HOWARD S. SHELLHAMMER

Department of Biological Sciences

San Jose State University

San Jose, California 95192

Salt marsh harvest mice can be identified in the field using the four characters of
the tail described by Fisler (1965). These characters are assigned 0, 1, or 2 points and
the total score is indicative of either the salt marsh harvest mouse, Reithrodontomys
raviventris, or the closely related and often sympatric western harvest mouse, R.
megalotis. Tail/body ratios, and to some extent the color of the venter, are useful
in some populations while behavior is useful in all populations. Identification is easist
when numerous traits are used.

Identification of skulls is much more difficult, depending in part on several subjec-
tive observations and on having undamaged skulls to study.

INTRODUCTION

The salt marsh harvest mouse, Reithrodontomys raviventris, with its two
subspecies R. r. raviventris and R. r. halicoetes, is a state and federally-listed
endangered species endemic to the marshes of San Francisco Bay and is similar
to the unprotected and ubiquitous western harvest mouse, R. megalotis (Shell-
hammer 1982, Shellhammer et al. 1982). Most biologists have considerable
difficulty identifying the two species even though Fisler ( 1 965 ) described criteria
for their identification.

The salt marsh harvest mouse is often called the "red-bellied" harvest mouse
because of the intense cinnamon or buff color of the venter of many animals
found in the marshes of the South San Francisco Bay. This a misnomer because
most animals in this species have whitish venters, especially those in San Pablo
and Suisun Bay marshes, where venters often are whiter than those of western
harvest mice.

The most reliable characters, though the most difficult to use, are those
associated with the diameter, pattern, coloration and shape of the tail as well
as the tail/body ratio, although the latter is most usuable in Suisun and South
San Francisco Bay populations (Fisler 1965).

I assign numeric values to these tail characters in an adaptation of the hybrid
index technique of Anderson (1949). Total scores for all tail characters have
been effective in differentiating between the two species, especially when used
in conjunction with tail/body ratios, coloration of the venter and behavior.
Zetterquist (1978) used this general technique and found it effective in South
San Francisco Bay populations.

I present the details of my technique in this paper along with Fisler's (1965)
description of venter types. His classic work is the basis for this and any other
technique of identifying salt marsh harvest mice and I give him full credit. In
addition, my scoring technique, described herein, adds greatly to usefulness of
the features he described for identifying the two species. I have examined 116

' Accepted for publication May 1983.

114 CALIFORNIA FISH AND CAME

skulls of the two species and have discovered a few characters which can be
used to identify these forms using intact, or nearly intact, skulls.

MATERIALS AND METHODS

Body Traits

Sixty-seven mice of both species were live-trapped, under endangered species
permits, in the marshes near Collinsville (Solano County) at the eastern edge
of Suisun Bay in 1 978 and 1 979 ( Envirodyne Engineers 1 978, Biosystems Analysis
1978). These animals were characterized as to four tail characters (diameter at
20 mm from the rump, tail pattern, color of ventral hairs, and shape of tip),
tail/body ratios, color of venters, behavior, and presence or absence of orange
tuffs in front of their ears.

Forty-five mice of both species were live-trapped in South San Francisco Bay
marshes, 0.8 and 3.2 km north of Alviso, Santa Clara County, and in Coyote Hills
Regional Park, Alameda County, during 1977. They were checked for only two
tail traits (pattern of tail and shape of the tip), tail/body ratio, color of venter
and behavior.

The tail traits were scored (Table 1 ) and summed in each case for a possible
score of four in the South San Francisco Bay populations and eight in the Suisun
Bay population. The scores and tail /body ratios were compared with venter
coloration and behavior to establish species identifications.

TABLE 1. Values Assigned to Tail Traits.

Trait

0

Value
/

2

Tail at 20 mm from rump
Tail pattern
Ventral hairs on tail
Tip of tail

2.1 or more
unicolor
tan
blunt

2.0 mm
intermediate
intermediate
intermediate

1.9 mm or less
bicolor

white to grayish-white
pointed

The tip of the tail is seen best by holding the tail so it was backlit and hence
the tip stands out from its surrounding hairs. A salt marsh harvest mouse has a
tail with very little taper and a blunt tip. It looks like it has a missing tip, but
without the swelling and loss of hair that usually accompanies such a condition.
The tail of a western harvest mouse is thinner, more tapered towards the tip and
the tip is decidedly pointed.

Adequate light was necessary to assess the color of the hairs on the venter
of the tail. Salt marsh forms have tan hairs on the ventral side of their tails while
western forms have a considerable number to a majority of white hairs. The
intermediate condition consists of mostly tan hairs with a few white to whitish
hairs. A hand lens was useful in observing this character.

The tail of a western harvest mouse in the Bay region is bicolored, but not
markedly so. The three categories for this trait are best described as "nearly
unicolored" for the salt marsh form, "indistinctly bicolored" for the intermediate
situation and "lightly bicolored" for the western form.

The diameter of the tail was somewhat difficult to measure because of the
hairs surrounding the tail. It was measured using a plastic ruler but a caliper is
more satisfactory. Several measurements were necessary to correctly estimate
the width to the nearest 0.1 mm.

The length of the body and the tail were measured several times to get an

SALT MARSH HARVEST MOUSE IDENTIFICATION

115

accurate measurement. The tail/body ratio was obtained by dividing the tail
length by the body length and multiplying the answer by 100. Hind foot length
or ear length were not recorded, as they add nothing to the identification
process.

The animal's behavior was observed throughout the handling process. I noted
the color and pattern of the venter (belly) of each animal and looked for orange
tuffs of hair in front of the base of the ears.

All of this information was recorded on a separate sheet in my field notebook
for each animal (Figure 1).

Male Female
Location:

Tag tt

Date of capture

Recapture(s)

Venter coloration
after Fisler (1965)

1 . White, greyish
white venter.

2. Cinnamon pectoral
spot

3. Band of color
across chest.

4. Ventral band, 3/4
of \ienXer ijol^'iTe. .

5. Color and white
mixed, 1/2 white.

6. Trace of (1/4) white.

7. All cinnamon, of
varying intensity

Body length Tail length T/B %
HFL EL -WT. - = gm

Testes desc. Vagina per. PG Lact.

T -1 .• ■ X J /'^ interm. () blunt C'^j
Tail tip pointed {__) <^ \-y

Tail bicolor Q interm. Q unicolor O

White hairs Q_) a few (~^ all tan (^
Diameter of tail at 20 mm from body

1.9 mm or < Q 2.0. Q 2.1 or > Q

Venter coloration

12 3 4 5 6 7

Behavior

active interm. docile

Orange tuffs in front of ears: yes no

Vegetation at capture site:

FIGURE 1. Sample sheet from field notebook.

Cranial Features

Numerous skulls of both species were examined for distinguishing features.
The one usable, measureable feature is the position of the posterior foramen in
respect to the length of the palate posterior to the incisive foramen (Figure 2).

116

CALIFORNIA FISH AND CAME

The measurement was made with a Helios caliper to the nearest 0.01 mm. The
distance between the posterior end of the incisive foramen and the anterior end
of the posterior palatine foramen was divided by the distance between the
posterior end of the incisive foramen and the anteriormost part of mesopterygoid
fossa. The result was expressed as a decimal.

.-.vj

FIGURE 2. Palatine measurements. View is of the palate with the incisive foramen at the top and
the posterior platine foramen in the middle.

Two more subjective cranial features were identified. The medial inflection of
the zygoma (the zygomatic arch), as viewed from above, was estimated and
weighted as follows: an inflection greater than one thickness of the zygoma =
0 points (the salt marsh mouse pattern), an inflection of one thickness or less
= 1 point, and no inflection = 2 points (the western mouse pattern) (Figure
3). The other feature was the visibility of the sphenopalatine foramen when
viewed from the side ( Figure 4) . A large visible foramen, extending considerably
above the zygoma (the salt marsh pattern) = 0 points, a foramen with a small
portion visible over the zygoma = 1 point and no foramen visible from a lateral
view (the western pattern ) — 2 points. The scores for these two characters were
totalled for each specimen and compared with the palatine measurement.
Scores ranged from 0 for skulls with the strongest salt marsh harvest mouse
characteristics to 4 for those with strongest western harvest mouse characteris-
tics.

SALT MARSH HARVEST MOUSE IDENTIFICATION

117

FIGURE 3. Medial inflection of the zygoma (Zygomatic arch). No inflection
= b, inflection of one thickness of the zygoma or more = c.

a, intermediate

VjjJM\>~'

ZYGOMA

VjvjOmvAW

FIGURE 4. Sphenopalatine foramen as viewed from the side. Foramen not visible = a, partially
visible = b, and foramen highly visible = c.

Sixty-six salt marsh harvest mice and 50 v^estern harvest mice skulls were
checked for these traits. Fifty-five salt marsh and 40 western harvest mice were
from the Museum of Vertebrate Zoology at the University of California at Berke-
ley; the remainder were from the collection at San Jose State University. Skulls
were compared from five general areas which were (with the salt marsh harvest
mouse locations listed first) : Callinas Creek, Marin County, marsh versus various
locations in Marin County; Sonoma Creek, Sonoma County, versus Sonoma
County; Martinez marsh versus Martinez marsh, both Contra Costa County;
Richmond marsh versus Richmond marsh, both Contra Costa County; and Al-
viso marsh, Santa Clara County, versus Santa Clara and Santa Cruz Counties.

118 CALIFORNIA FISH AND GAME

The first three groups of sites are within the range of R. r. halicoetes which
extends from Pt. San Pedro, Marin County, to Collinsville, Solano County, and
along the Contra Costa County Coast from near Pittsburg to Martinez. The other
two groups of sites are with the range of R. r. raviventris which extends south
from Pt. San Pedro in Marin County and into South San Francisco Bay from the
San Mateo Bridge on the west and the Richmond area on the east.

RESULTS
Body Traits

The pattern of the tail, the shape of its tip and tail /body ratios were sufficient
to differentiate between the two species in the South San Francisco Bay popula-
tions. Tail score totals (ranging from 0 to 4) were plotted against tail/body ratios
and there was a complete separation of the two populations. All 29 of the salt
marsh harvest mice had red bellies (i.e., a score of 5, 6, or 7 on Fisler's scale)
while none of the western harvest mice scored higher than a three. The salt
marsh forms were docile while the western harvest mice were highly active and
bit often.

The situation at Collinsville was similar, but not quite as clear cut. All four tail
traits were used but none of the seven western harvest mice had a score higher
than six out of a possible total of eight. Only two of the salt marsh forms scored
higher than a two (a three and a four) . Only three of the 61 salt marsh harvest
mice had tail /body ratios as low as the longest-tailed western harvest mouse
(i.e., 114%). One animal had a tail /body ratio percentage of 114% and interme-
diate traits for the other characters and could not be assigned to either species.

Cranial Traits
Palatal ratios showed highly significant differences between the means of each
pair of samples tested (p values were all less than 0.001 ). Nineteen of the 116
skulls fell into the zone of overlap (0.44 to 0.49) for this measurement. Only one
of the 50 skulls on the salt marsh harvest mouse side of the zone of overlap had
a "western" score, i.e., a three using the two subjective cranial features. Eleven
of the 41 skulls on the western harvest mouse side of the zone had intermediate
scores of two, but only one skull had a salt marsh harvest mouse skull score of
one.

DISCUSSION

It is easiest to identify salt marsh harvest mice in the northeastern and southern
portions of their range, e.g., in Collinsville and Alviso. Reddish venters are the
rule at Alviso and their overall coloration is very dark. The animals have very
short tails (hence tail/body ratios are less than 100%). Animals at Collinsville
have very long tails (tail/body ratios usually exceeding 120%), low tail scores
and "thick" tails. A number of investigators including myself have noticed that
the tails of animals in the Suisun Bay marshes and along the Contra Costa County
coast, and even in the Napa marshes, have thicker tails than Fisler reported in
1965. Many western harvest mice have tails with diameter of 2.0 or 2.1 mm
instead of 1.9 mm or less, while salt marsh harvest mice have tails which range
from 2.1 mm to as high as 3.0 mm, but are usually 2.3 to 2.5 mm.

The dine in length of tail aids the observer in these two areas. The tail/body
ratio, however, is not usable in the Richmond, Petaluma (generally), Sonoma
Creek, Callinas and Corte Madera areas.

SALT MARSH HARVEST MOUSE IDENTIFICATION 119

Body coloration, especially that of the venter, is usable only in the southern-
most part of the range. In areas where both coloration and tail/body ratios are
unusable, characters associated with the tail provide the only diagnostic traits.
They are not easy to use, at least not at first. Most investigators tend to score
more animals as "ones" when they first start using this set of characters unless
they have trapped western harvest mice previously. Investigators who have seen
representatives of both species find it easier to identify tail characters and
behavior.

Behavior is a moderately valuable trait. Most R. r. raviventris, especially those
from Palo Alto to Newark, are very gentle. They are often torporous. Many R.
r. halicoetes are moderately active and may bite, but are seldom as active as
western harvest mice.

The presence of orange tuffs of hair in front of the base of the ears should be
noted as most salt marsh forms have them and few westerns do. It is an obvious
trait but one of relatively little value, especially by itself.

The various traits discussed in this paper are best used when an investigator
has captured a number of animals in the same marsh and hence can construct
a two dimensional plot incorporating tail traits and tail/body ratios.

Single individuals are easy to identify in certain populations. Animals from
Suisun Bay-Collinsville-Martinez marshes with tail/body ratios of 114% or more
and tail diameters of 2.2 mm or more are salt marsh harvest mice. Animals from
the Palo Alto-Alviso-Newark areas of the South Bay with tail/body ratios of 98%
or less and tail thickness of 2.1 mm or more are the salt marsh form. Almost any
mouse in these areas with a reddish belly is a salt marsh form, although one
should check the other traits.

Fisler did not find type 7 venters on any western harvest mice, hence, such
a feature is diagnostic if an animal rates a seven. Such animals are found relative-
ly more frequently in the San Pablo-Richmond and Corte Madera marshes and
less frequently in the Gallinas and Petaluma marshes and along the Petaluma
River. In these marshes the four tail traits are all important. A white-bellied
mouse with a tail/body ratio of 109%, a tail thickness of 2.1 mm, a nearly
unicolored tail with a blunt tip and only a few white hairs on its ventral side is
a salt marsh harvest mouse. The sum of the tail features (Table 1 ) is usually a
score of three or less. The pattern of the tail and the tip of the tail are often rated
by cautious, novice observers as scores of one but the color of the ventral hairs
of the tail and the diameter of the tail are easier to judge and score correctly.
Western harvest mice generally score a five or a six, sometimes a seven or eight.
Salt marsh harvest mice generally score zero to two. A score of four presents
a problem, especially if the tail/body ratio and color do not provide usable
information.

The identification of the skulls is difficult. The one accurately measureable
feature that I have found requires at least a complete skull back to the middle
of the orbit and a precision measuring device. The two subjective traits require
the presence of an intact zygoma and this fragile arch is easily and often broken.
Studies are under way to find other quantitative cranial features. Based on
current knowledge, a skull with a palatal measurement of 43% (.43) or less is
probably a salt marsh harvest mouse. An inflected zygoma and easily seen
sphenopalatine foramen (viewed directly from the side of the skull) support
such a diagnosis.

120 CALIFORNIA FISH AND CAME

ACKNOWLEDGMENTS
The assistance and support of the following persons is sincerely appreciated:
J. Michaels and J. Gustafson of the California Department of Fish and Game (JM
presently is in the USFWS), R. Jackson of Biosystems Analysis, Inc. of San
Francisco, W. Lidicker and O. Pearson of MVZ at University of California,
Berkeley, and the staff of both the Coyote Hills Regional Park and the San
Francisco Bay National Wildlife Refuge Headquarters at Newark.

LITERATURE CITED

Anderson, E. 1949. Introgressive hybridization. John Wiley and Sons, New York. 109 p.

Biosystems Analysis, Inc. 1979. A report on the 1979 summer trapping program for the endangered salt marsh

harvest mouse at the Montezuma Power Plant site. Prepared for Pacific Gas and Electric Company. 53 p.
Envirodyne Engineers (presently Biosystems Analysis, Inc.). 1978. A report on salt-marsh harvest mouse trapping

and vegetation survey at.the Montezuma Power Plant site. Fossil Fuel 1 and 2 Project. Prepared for Pacific

Gas and Electric Company. 53 p.

Fisler, C. F. 1965. Adaptations and speciation in harvest mice of the marshes of San Francisco Bay. Univ. Calif.
Publ. Zool. 77:1-108.

Shellhammer, H. S. 1982. Reithrodontomys raviventris. Mammalian Species, No.169:1-3. American Society of
Mammalogists.

Shellhammer, H. S., R.Jackson, W. Davilla, A. Cilroy, H. T. Harvey and L. Simons. 1982. Habitat preferences of
salt marsh harvest mice (Reithrodontomys raviventris). Wasmann J., Biol. 40(1-2)102-114.

Zetterquist, D. K. 1978. The salt marsh harvest mouse (Reithrodontomys raviventris raviventris) in marginal
habitats. Wasmann J. Biol., 35(1)68-76.

FRESHWATER LAMPREY OCCURRENCE AND MORPHOLOGY 121

Calif. Fish and Came 70(2); 121-127 1984

A SECOND RECORD FOR CALIFORNIA AND
ADDITIONAL MORPHOLOGICAL INFORMATION ON
ENTOSPHENUS'' HUBBSI \\.\D\KO\ AND KOTT 1976

(PETROMYZONTIDAE) 2

VADIM D. VLADYKOV

Department of Biology

University of Ottawa

Ottawa, Ontario

KIN 6N5

AND

EDWARD KOTT

Department of Biology

Wilfrid Laurier University

Waterloo, Ontario

N2L 3C5

Entosphenus ■• hubbsih a nonparasitic freshwater species of lamprey described by
Vladykov and Kott (1976a). The original 11 newly metamorphosed specimens and
1 ammocoete were obtained on 15 February 1972 by D. P. Christenson from the
Friant-Kern Canal, east of Delano, California. In the present study, we describe
additional material of E. hubbsi consisWn^ of 1 ammocoete, 8 metamorphosing, and
119 metamorphosed specimens, collected by R. R. Menchen from the Merced River,
near Merced Falls, a tributary of the San Joaquin River system in California. This
lamprey was separable from other species of Entosphenus by a low number of trunk
myomeres (50-57), and reduced dentition.

INTRODUCTION

Entosphenus hubbsi\s a nonparasitic lamprey, originally described by Vlady-
kov and Kott (^^76a ), and dedicated to Carl Leavitt Hubbs, a distinguished
friend and keen student of lamprey taxonomy.

The first specimens were collected in the Friant-Kern Canal, east of Delano,
Kern Co., California. This canal, which connects the Kern River system with the
San Joaquin River, is about 85% concrete lined, with a flow of at least 56.6 m^/s.
Thus, this canal cannot be considered as a normal habitat for such a small
nonparasitic lamprey. Vladykov and Kott (1979) suggested that this species may
be present in the Kern River, and possibly in the San Joaquin River system.

MATERIAL AND METHODS
The present study is based on an examination of 1 ammocoete, 8 metamor-
phosing, and 119 metamorphosed individuals of £ /7iyZ?^5/ collected by R. R.
Menchen during February and March, 1977, from the Merced River, near
Merced Falls, a tributary of the San Joaquin River (Figure 1 ). This new material
not only provides additional information on distribution of this lamprey but
allows for a more detailed study of its morphology. .

' Robins eta/. (1980) and Shapovalov, Cordone, and Dill (1981) placed Entosphenus \n the genus Lampetra. We
(Vladykov and Follett 1967; Vladykov and Kott 1976c/and 1979) maintain that Entosphenus is a distinct genus
from Lampetra.

^ Accepted for publication June 1983.

122

CALIFORNIA FISH AND CAME

Oregon

Arizona

Mexico

FIGURE 1 Map showing localities where Entosphenus hubbsi were collected.

Definitions of body proportions follow Vladykov and Follett (1965) and
Vladykov and Kott (1980). Terminology of the teeth is that of Vladykov and
Follett (1967). Information on dentition is derived from a study of 30 specimens
from the Merced River.

WILDLIFE HABITAT REMOTE SENSING 123

Trunk myomeres were counted between the last (7th) gill opening and the
anterior tip of the cloacal slit (Hubbs and Trautman 1937, Vladykov 1949).
Information on velar tentacles has been published by Vladykov and Kott
( 1 9766) , and the taxonomic significance of male urogenital papillae of different
species of lampreys was given by Vladykov and Kott (1982 a).

DESCRIPTION OF SPECIMENS

Metamorphosed Individuals

Ninety-eight males and 18 females were measured (Table 1 ). The total length

of the males ranged from 95 to 139 mm (mean 112.3 mm) while the females

ranged from 81 to 119 mm (mean 102.1 mm). Measurements of different body

proportions are expressed in percentage of the total length. Mean values for

males and females respectively were: disc length 7.0, 6.1; prebranchial length

1 1 .2, 1 0.6; trunk length 47.1 , 51 .3; tail length 28.5, 27.0; disc length in percentage

of branchial length 58.9, 56.1. The number of trunk myomeres (Table 2) ranged

from 50 to 57 (mean 53.5).

TABLE 1. Body Proportions in Percent of Total Length and Disc Length in Percent of Bran-
chial Length of Metamorphosed Specimens of Entosphenus hubbsi from Two
Localities. Data Refer to Means (ranges in Parentheses) for Each Character.

Authority Present Study Vladykov and Kott (1976a )

Locality Merced River Friant-Kern Canal

Sex Male Female Male Female

No. specimens 98 18 8 3

Total length, mm 112.3 102.1 129.1 140.3

(95-139) (81-119) (117-142) (140-141)

Prebranchial length 13.1 11.7 9.4 9.0

(7.8-14.4) (10.1-13.5) (8.8-10.7) (8.5-9.6)

Branchiallength 11.2 10.6 10.5 9.3

(9.7-12.8) (8.9-11.7) (9.5-11.2) (8.5-10.7)

Trunk length 47.1 51.3 54,0 54.5

(44.6-51.1) (46.9-55.0) (51.3-56.9) (53.5-55.0)

Tail length 28.5 27.0 28.5 28.6

(25.5-30.8) (24.3-29.6) (27.3-30.8) (27.5-30.1)

Eye length 1.8 1.9 17 1.8

(0.8-2.8) (1.5-2.5) (1.4-2.1) (1.4-2.0)

Disc length 7.0 6.1 4.3 4.4

(4,3-^.7) (4.2-7.0) (3.7^.7) (3.9-4.6)

Disc length 58.9 56.1 — —

Branchial length (40.0-81.8) (38.5-63.6)

TABLE 2.

Species

Number of Myomeres in Metamorphosed Entosphenus hubbsi and E. tridentatus.
Collected at Same Locality from the Merced River, California, February-March

1977.

No.
Exam-
ined

Number of myomeres

E. hubbsi 119

£ tridentatus 5

50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 Mean

2 6 20 20 32 25 10 4 --------- - 53.8

------------1 --1 2 1 65.2

Because the species is nonparasitic, the teeth (Table 3) were weakly cornified
and often many were hardly visible. As is characteristic of the genus Entosphe-
nus, a single row of posterials was present, consisting of 8 to 1 1 ( mean 9.0) teeth

124 CALIFORNIA FISH AND CAME

that were weakly developed. Among 30 specimens, only 1 had the supraoral
lamina with 3 cusps; all the others had 2 cusps. Cusps on the infraoral lamina
ranged from 4 to 6 ( mean 5.2 ). All but 3 specimens ( Table 3 ) had 8 endolaterals,
as is characteristic of the genus, but the number of their cusps may vary from
1 to 3. Cusps on the anterial field were usually arranged in a single row and were
of uniform size. They ranged from 3 to 7 (mean 4.5). Cusps on the lingual
laminae were not developed, hence no counts could be made.

TABLE 3. Number of Cusps on the Teeth of Entosphenus hubbsiUom Two Localities. Data
Refer to Means and Ranges for Each Character. Number of Counts in Parentheses.

Authority Present study Vladykov and Kott (1976a)

Locality Merced River Friant-Kern Canal

Anterials 4.5 (29)

3-7
Supraorals 2.0(30) 2.1 (11)

2-3 2-3

Endolaterals;

Typical formula M-M (24) 1-1-1-1 (21)

Variant formulae 1-2-2-1 (12)

1-2-1-1 (6)

2-2-1-1 (4)
Formulae observed in two cases each: 1-3-2-1,2-2-2-2

2-2-2-1, 2-1-1-1
Formulae observed in a single case: 0-3-2-1, 0-2-1-1,

1-0-1-1, 1-1-2-1,

1-1-3-1, 2-2-3-1 -

Infraorals 5.2 (30) 5.0 (9)

4-6 5

Posteriais 9.0 (18) 10.3 (6)

8-11 9-12

Velar tentacles in 1 5 specimens ranged from 3 to 5 ( mean 4.7) . Tentacles were
unpigmented and short, the longest ones ranged in length from 0.5 mm to 0.9
mm. The lateral tentacles did not form dorsal wings that are typical of £ triden-
tatus.

The sides and back of specimens preserved in 4-5% formalin were gray-
brown and the lower surface was whitish. Both dorsal and caudal fins had black
specklings.

Sexual dimorphism was manifested by the presence in males of a relatively
longer urogenital papilla { Figure 2 ) . The average length of this papilla in 9 males
was 1.6 mm (ranged from 1.0 to 2.0 mm), and as a percentage of branchial
length was 1 3% ( range 8 to 1 7% ) . Females had a much shorter papilla. Howev-
er, those close to spawning time developed an anal finlike fold (Figure 2) that
was absent in males.

Sexual dimorphism was also evident in the following body proportions, ex-
pressed in percentages of total length and branchial length respectively: in males,
disc lengths were 7.0 and 58.9, while in females, 6.1 and 56.1. In contrast, trunk
length as a percentage of total length was 47.1 in males and 51.3 for females
(Table 1).

FRESHWATER LAMPREY OCCURRENCE AND MORPHOLOGY

125

126 CALIFORNIA FISH AND CAME

Total lengths of males and females of £ hubbsi from the Friant-Kern Canal
were longer than those of specimens from the Merced River (Table 1). The
explanation for this was the different stages of sexual development of metamor-
phosed individuals. Specimens from Friant-Kern Canal were in early metamor-
phosis, prior to shrinkage of the body, hence the two dorsal fins were far apart.
The lampreys from Merced River were more advanced in maturation, and as
shrinkage had already occurred, the two dorsal fins touched each other. For
further details on body shrinkage, a unique feature of lampreys, consult Vlady-
kov and Kott (1978), and Vladykov et al. (1982/?).

Specimens from the two locations also differed in disc length (Table 1 ). In
lampreys from the Merced River, taken close to spawning time, the discs had
attained their full development, but in the Friant-Kern Canal sample the discs
were still small.

Ammocoetes

A single ammocoete from Merced River measured 100 mm in total length and
had 57 myomeres. Body proportions in percent of total length were: prebranchi-
al length 9.0, branchial length 12.0, trunk length 49.0, tail length 29.0. The
precursor of the tongue was strongly pigmented. The bulb had heavy black
pigmentation that was stronger than that in £ folletti (Vladykov and Kott
1976c) and that also extended around the elastic ridge.

COMPARISONS WITH ENTOSPHENUS TRIDENTATUS
Five recently metamorphosed specimens and one ammocoete of the parasitic
species Entosphenus tridentatus were collected from the Merced River at the
same time and place as £ hubbsi. They ranged in length from 115 to 126 mm
( mean 1 1 9.8 mm ) and were of a size comparable to the specimens of £ hubbsi.
The two species are readily distinguishable on myomere counts (Table 2). In £
hubbsi myomeres ranged from 50 to 57, whereas in £ tridentatus they ranged
from 62 to 67. The single ammocoete of £ tridentatus had 70 myomeres.

Average body proportions in the newly metamorphosed £ tridentatus, sexes
combined, expressed as a percentage of total length were: disc length 7.0;
prebranchial length 14.2; eye length 3.3; branchial length 9.0; trunk length 45.4;
tail length 31.3; disc length in percentage of branchial length was 78.5. The eye
length was considerably greater in £ tridentatus than in similarly sized £ hubbsi
A small eye length is typical of nonparasitic species of lampreys in general.

GEOGRAPHICAL DISTRIBUTION

The two samples of £ hubbsi came from south central California. The first 1 1
metamorphosed specimens and 1 ammocoete were obtained from the Friant-
Kern Canal, east of Delano, Kern County, California (Vladykov and Kott
1976a). Because this habitat definitely is not suitable for such a small lamprey
as £ hubbsi, we postulated that it could have been swept into the canal from
the Kern system ( Figure 1 ) . Before construction of the canal, the Kern basin was
an inland drainage system, however, no record of £ hubbsi v/as reported from
the Kern area proper. Further information on the physical aspects of the Friant-
Kern Canal may be obtained from Fact Sheet U.S. Dept. Interior (1974).

At present, this canal connects the Kern River with the San Joaquin River.
Because of the connection between these two river systems, we were almost
certain that £ hubbsi could be discovered in the San Joaquin River system as

FRESHWATER LAMPREY OCCURRENCE AND MORPHOLOGY 127

well. Our expectation was fulfilled by the collection of abundant material of £
hubbsi from the Merced River.

£ hubbsi may also live in the Sacramento River. However, at present, the San
Joaquin River system should be considered as the typical habitat for this species.
Because £ hubbsi is nonparasitic in habit and is limited in its distribution, it
should be considered as an endangered species and should be saved. Other
reasons why any nonparasitic lamprey must be protected were given in detail
by Vladykov (1973).

ACKNOWLEDGMENTS

R. R. Menchen, biologist, California Department of Fish and Came, Sacra-
mento, kindly obtained a well-preserved collection of Entosphenus hubbsi Uova
Merced River, near Merced Falls. D. Hickey, Professor, Department of Biology,
University of Ottawa, read the manuscript and offered some useful suggestions.
The photographs of lampreys were taken by C. Ben-Tchavtchavadze of the
same department. This study was supported by grants A-1736 and A-0575 from
the Natural Sciences and Engineering Research Council, Canada. To the above-
named persons and to NSERC, the authors extend their most sincere thanks.

LITERATURE CITED

Hubbs, C. L. and M. B. Trautman. 1937. A revision of the lamprey genus Ichthyomyzon. Misc. PubL Univ. Mich.
Mus. ZooL, 35: 1-109.

Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1980. A list of
common and scientific names of fishes from the United States and Canada. 4th ed. Am, Fish. Soc. Spec. Publ.
No. 12. pp. 1-174.

Shapovalov, L., A. ). Cordone, and W. A. Dill. 1981. A list of the freshwater and anadromous fishes of California.
Calif. Fish Came, 67(1); 4-38.

U.S. Dept. Interior, 1974. Friant Dam and Millerton Lake. Fact Sheet, Bureau of Reclamation, Sacramento, Califor-
nia, 3pp.

Vladykov, V. D. 1949. Quebec lampreys (Petromyzonidae) 1. List of species and their economical importance.
Contrib. Dep. Fish. Quebec 27: 65 p.

Vladykov, V. D. 1973. North American nonparasitic lampreys of the family Petromyzonidae must be protected.
Can. Field-Nat., 87: 235-239.

Vladykov, V. D. and W. I. Follett. 1965. Lampetra richardsoni, a new nonparasitic species of lamprey (Pe-
tromyzonidae) from western North America. Fish, Res. Board Can., J., 22: 139-158.

1967. The teeth of lampreys (Petromyzonidae); their terminology and use in a key to the

Holarctic genera. Fish. Res. Board Can., )., 24: 1067-1075.

Vladykov, V. D. and E, Kott. 1976a. A new nonparasitic species of lamprey of the genus Entosphenus Gill, 1862
(Petromyzonidae) from south central California, South. Calif. Acad. Sci., Bull., 75: 60-67.

1 9766. The taxonomic significance of velar tentacles in holarctic lampreys ( Petromyzonidae) .

Rev. Trav, Inst. Peches Marit., 40: 787-789,

1976c, A second nonparasitic species of Entosphenus Gill, 1862 (Petromyzonidae) from

Klamath River system, California. Can. J, ZooL, 54: 974-989,

1976c/. Is Okkelbergia Creaser and Hubbs, 1922 (Petromyzonidae) a distinct taxon? Can. J.

ZooL, 54: 421^25,

1978. A new nonparasitic species of the holarctic lamprey genus Lethenteron Creaser and

Hubbs, 1922 (Petromyzonidae) from northwestern North America with notes on other species of the same
genus, Biol. Pap, Univ, Alaska, 19: 74 p,
1979. List of Northern Hemisphere lampreys (Petromyzonidae) and their distribution. Fish.

Mar, Serv. Misc. Spec. PubL, 42: 1-30.

1980, Description and key to metamorphosed specimens and ammocoetes of Petromyzonidae

found in the Great Lakes region. Can, J, Fish, Aquat, Sci,, 37: 1616-1625,
1982a, Correct scientific names for the least brook lamprey and the American brook lamprey

(Petromyzontidae). Can. J. ZooL, 60: 856-864,
Vladykov, V, D„ C. B,, Renaud, E, Kott, and P, S. Economidis, 19826. A new nonparasitic species of Holarctic
lamprey genus Eudontomyzon Regan 1911 (Petromyzontidae), from Greece. Can. J. ZooL, 60: 2897-2915.

128 CALIFORNIA FISH AND GAME

BOOK REVIEW

Fisheries Management,

Edited by Robert T. Lackey and Larry A. Nielsen; John Wiley and Sons, Inc., New York; 1980;
422 pp.; illustrated

This is an excellent introductory textbook. The sixteen chapters are contributed by authors actively
working in the respective specialties. This direct familiarity gives the discussions a sense of relevance
often missing from textbooks w/ritten by a single author who is forced to give a second-hand
treatment of material outside his particular province. The field of fishery management is full of
uncertainties and constraints which are both challenging and frustrating. This is an unusual textbook
in that it conveys these feelings in addition to the usual factual material.

Fishery science is rrecessarily multidisciplinary. The chapters reflect this: the discussions of aquatic
ecology (Charles Warren, William Liss) draw on both field ecology and recently developed theoreti-
cal ecology. The latter holds promise for understanding fishery resources, and fisheries may provide
an excellent opportunity to wed theory and practice. The introduction to fish population dynamics
(Albert Tyler and Vincent Callucci) gives a sense of mathematics and modeling which is further
developed in a lucid chapter on systems principles (Carl Walters). The social sciences are well
represented by discussions of sociology /anthropology (Michael Orbach), planning and policy
analysis (Adam Sokoloskii), and an entertaining but effective chapter on objectives of management
(Peter Larkin). The discussions of actual fishery management cases show the profound difference
between the natural history orientation of inland fisheries (Richard Noble, Edwin Cooper) and the
analytical and political orientation of marine fisheries (J. L. McHugh, John Gulland).

The shortcomings of the book are not serious. The discussion of fishery economics (Frederik Bell)
seemed overly condensed and should have been more fully developed. The chapter on environmen-
tal analysis (John Cairns, Jr.) covered the topic of pollution in detail, but did not make any mention
of the controversial issue of power plant impacts on fish resources. All the chapters would benefit
from expanded lists of references.

A few years ago there was a dearth of texts on fishery management. Now there are more being
published than appears to be justified. This particular volume is a worthwhile contribution, and
should remain a useful reference for many years. — Alec D. MacCall

An Artist's Catch — Watercolors

by Frank Stick, Edited by David Stick. The University of North Carolina Press. 1981. 245 color
plates, 356 pp. $29.95.

This beautiful book is the result of Frank Stick's decision during the latter part of his life to put
together a collection of watercolors of fishes of the southeast coast of the LInited States and the
Caribbean. Frank's brother David both edited the book as well as wrote the introductory biograph-
ical sketch. A short preface describes why the artist began painting fish and the methods used to
obtain and identify the collections. The remainder of the book is devoted to the 245 color plates
of Frank Stick's excellent watercolors.

Unlike the Japanese and some early European artists, few American artists have specialized in
painting fish, and even more our loss that this country has not produced any significant published
collection of color paintings of North American marine or freshwater fishes. This book probably
reflects the most serious American effort to date. The artist has included detailed diagnostic charac-
ters in the water colors, for each species and has illustrated the different color phases for several
species.

This excellent collection should be welcomed by all lovers of fish, be they sportsmen, naturalists,
students or professional scientists. — Daniel W. Cotshall

78005-800 12-83 3,600 LDA

;Cl

o 5: s

5= 5

in -^ SO S

> 3z »
n m ~ 'E

> "• -n o

0 u> 3C _

S "is

*• 2 5
m

M

S""

c

i.i

c/> 2

■D- ^

-♦ (B

Z3.

P ^

K RA
POST-
AID

^ a

> -1
Of"

« r:

m

-♦♦