70 cm (M2); females <70 cm (Fl); fe-
males 71-85 cm (F2); females >86 cm (F3). Size categories
were based on approximate size at maturity and habitat
use. Males mature at about 70 cm, which coincides with
their movement from a mainly pelagic habitat to a deeper
mid-water and demersal habitat. Females undergo a simi-
lar habitat change at 70 cm; however, their size at maturity
is approximately 85 cm. Recoveries by the categories were
then examined by area of recapture.
Results
Suitability of tag
McFarlane and Beamish ( 1986) reported preliminary results
for the modified tag used in the present study compared to
Petersen disc tags and Floy anchor tags. In contrast to the
Petersen disc tag, the Floy anchor tag was quickly abraded
and lost. Of 1688 fish receiving both tags, 49 were recov-
ered from 1978 to 1982. All recovered fish had a Petersen
disc tag; however only 1 1 fish had both tags, and 9 of these
were recaptured during the first 18 months. McFarlane
and Beamish also compared the modified Petersen disc
tag to Petersen disc tags (McFarlane and Beamish, 1986).
Petersen discs attached to spiny dogfish in 1978 and 1979
and recaptured from 1978 to 1980 were compared to modi-
fied Petersen disc tags applied from 1979 to 1982 and stan-
dardized for catch. The standardized recovery percentage
(3.9%) for the modified Petersen disc tag compared to the
Petersen disc tag (2.4%) was significantly higher (P=0.01).
Because the materials used in the tags were similar and
none of the modified tags were returned with one pin miss-
ing, the decrease in percentage of recoveries of Petersen tags
probably was due to mortality caused by tag wounds and
not to disc or pin loss. For a description of the tag wounds
see McFarlane and Beamish (1986). Both titanium pins
remained in the tag in all recovered fish. A metallurgical
stress test indicated that the pins were durable in salt water
and might be expected to last more than 20 years (McFar-
lane and Beamish, 1986). In the present study we report
that fish with the modified Petersen disc tag were recap-
tured with tags intact 20 years after release. After correct-
ing for differences between years in catch (t) for tagged fish
recaptured in the Strait of Georgia in 1988-90, a chi-square
test on the ratios of the numbers released to the numbers
returned indicated no significant level of difference in return
percentages between the standard hard plastic tag and the
more flexible plastic tag used in 1988 (P=0.304).
Tag return rates
Within the Strait of Georgia, tagging took place every year
from 1978 to 1988 with the exception of 1986 (Table 2). Off
the west coast of Vancouver Island, tagging was conducted
in 1984, 1985, and 1987 (Table 2). Tagging was conducted in
northern British Columbia waters in 1980, 1982, and 1987
(Table 2). In total, 70,770 fish were tagged throughout all
McFarlane and King: Migration patterns of Squalus acanthias
361
Table 1
Catch (in metric tons) from 1978 to 2000 of
spiny dogfish (Squalus
acanthias) for areas in w
tiich targeted fisheries operate.
Strait of
West coast of
Washington State
Year
Georgia
Vancouver Island
Puget Sound
coastal waters
1978
2366
271
2647
42
1979
4469
303
3882
129
1980
2133
1874
3004
57
1981
781
312
1808
79
1982
1297
973
1944
38
1983
1281
596
1291
26
1984
1991
460
1445
318
1985
962
1499
971
274
1986
610
1935
746
113
1987
1247
2110
1429
984
1988
1200
3724
1396
200
1989
852
1847
1098
319
1990
820
2353
904
488
1991
667
1958
1303
853
1992
575
1426
931
1044
1993
135
111
758
1245
1994
941
876
958
1392
1995
1494
1076
929
367
1996
3019
938
818
251
1997
1584
531
214
425
1998
1831
953
115
462
1999
1062
1787
111
515
2000
610
2951
62
627
areas: 51,063 fish tagged in the Strait of Georgia, 10,087
fish tagged off the west coast of Vancouver Island, and 9620
fish tagged in northern British Columbia waters (Table 2).
As of 31 December 2000, the total number of tagged fish
recaptured were 2454 (4.8%) for Strait of Georgia released
fish; 297 (2.9%) for the west coast of Vancouver Island
fish; and 190 (2.0%) for fish in northern British Columbia
waters (Table 2).
Recoveries over time
Approximately 93%, 96%, and 93% of the fish released
in the Strait of Georgia, west coast of Vancouver Island,
and northern British Columbia waters, respectively, were
recaptured in the first 10 years at liberty (Fig. 3). Most
recoveries (>70%) were made within <5 years after release
(Fig. 3). In the Strait of Georgia, 81% were recovered <5
years; for the west coast of Vancouver Island 88% were
recovered <4 years; and for northern British Columbia
waters 71% were recovered <5 years. The maximum time
at liberty was 19, 15, and 20 years for the Strait of Geor-
gia, west coast of Vancouver Island, and northern British
Columbia waters, respectively (Table 2).
Movement of tagged fish outside of release areas
Tagged spiny dogfish were recaptured throughout the
North Pacific, from Japan, through Alaska, south to Mexico
(Fig. 4). A large number of recaptured fish were reported
from Puget Sound and Washington State coast (Table 3).
For fish released in the Strait of Georgia, only three fish
were recaptured outside of Canadian or Washington State
waters (Table 3). For fish released in coastal waters (west
coast of Vancouver Island and northern British Columbia),
a large number offish were recaptured in Japan («=18+11)
and United States waters, excluding Washington State
(n=n+5). Two fish tagged off the west coast of Vancouver
Island were recaptured in Mexico (Table 3).
Movement between release areas
Approximately 98% of the total number of recaptured fish
were recaptured in Canadian waters or Washington State
waters (coastal and Puget Sound). Because at least 70%- of
the recoveries occurred within the first 5 years at liberty
(Fig. 3), we compared movement between release areas
by examining the proportion of recaptured fish (percent-
age) at liberty for 5 years or less. For fish released in the
Strait of Georgia, the majority (91%) were recaptured in
the Strait of Georgia and another 5% off the west coast of
Vancouver Island (Table 4). Only 2% were recaptured in
northern British Columbia waters and 1% were recaptured
in Puget Sound or in Washington State coastal waters
(Table 4). For fish released off the west coast ofVancouver
Island, again the majority of recaptures (62%) were in the
area of release. A large percentage were recaptured in the
362
Fishery Bulletin 101(2)
120
r
100
^^
s?
D • ■ °u£..!-'8'"'^'^^°*'^
(/)
E 80
• °^.--*'^- ■
a
°/^ .»
S 60
y^ .■■'''
en
J'
3
/fl i>"'
c
S 40
/'
Q-
j^ y' Strait of Georgia -♦-
/ •' west coast Vancouver Island -n-
20
//' northern British Columbia -♦-
0 1 2 3 4 5 6 7 8 9 10 11 +
Years at liberty
Figure 3
The cumulative percentage of recoveries standardized to catch by release
area and years at hberty.
Table 2
Summary of the
number of released and recaptured spiny
dogfish by release
year and release area.
Year of release
Released
Recaptured
Max
years at liberty
Strait of Georgia
1978
1692
56
12
1979
4563
320
17
1980
7522
478
19
1981
7054
426
17
1982
10646
569
18
1983
1630
61
16
1984
7333
332
17
1985
4124
108
14
1987
3124
39
11
1988
3375
65
12
Total
51063
2454
West coast of Vancouver Island
1984
2066
77
14
1985
7124
198
15
1987
897
22
11
Total
10087
297
Northern British Columbia
1980
1075
34
20
1982
4873
120
16
1987
3672
36
11
Total
9620
190
Strait of Georgia (13%) and in other areas (11%) including
Japan, Alaska, Oregon, California, and Mexico (Fig. 4B).
Only half of the recaptured fish from northern Bristish
(^olumhia releases were recaptured in northern British
Columbia (Table 4). Approximately 13% were recaptured
in the Strait of Georgia, and another 28'^f off the west coast
of Vancouver Island. As with west coast Vancouver Island
releases, a large proportion (7%) of recaptured fish were
found in other areas, namely in waters off Japan, Alaska,
Oregon, and California (Fig. 4C).
McFarlane and King: Migration patterns of Squalus acanthias
363
P^^
Additional Recoveries
Alaska (1)
Japan (1)
Oregon (1)
Additional Recoveries
Alaska (2)
California (5)
Japan (19)
Mexico (2)
Oregon (10)
Figure 4
Recapture locations of tagged spiny dogfish released in (A) the Strait of Georgia; (B) west
coast of Vancouver Island; and (C) northern British Columbia. The number of additional
recoveries of spiny dogfish (by recapture area) that are not plotted on each figure are
listed in the upper right key.
364
Fishery Bulletin 101(2)
Table 3
The number of tagged dogfish recaptured from outside Can
adian waters. SOG =
Strait of Georgia; WCVl =
west coast of Vancouver
Island: NBC = northern British Columbia.
Area of recapl
ure
release Japan Alaska
Puget Sound
Washington
Oregon
California
Mexico
Total
SOG 1 1
28
18
1
0
0
49
WCVl 18 2
U
13
10
5
2
61
NBC 11 1
0
2
2
2
0
18
Total 30 4
39
33
13
7
2
128
Similar to the results for nonstandardized recoveries,
the standardized recaptures for Strait of Georgia, Puget
Sound, west coast of Vancouver Island, and Washington
State coastal waters were based on recaptured fish at
liberty for 5 years or less. For fish released in the Strait
of Georgia, the majority (86%) were recaptured in the
Strait of Georgia (Table 4). A further 7% were recaptured
off the west coast of Vancouver Island, 6% in Washington
State coastal waters, and only 1% in Puget Sound (Table
4). For fish released off the west coast of Vancouver Island,
the majority of standardized recoveries (73%) occurred in
Washington State coastal waters (Table 4). Only 20% were
recaptured in the area of release.
Movement by sex and size
In general, small-size females (<70 cm) released in the
Strait of Georgia and northern British Columbia were of
the sex-size category (Fl) of spiny dogfish that were recap-
tured in the highest proportion in other areas (Table 5).
All sex-size categories of spiny dogfish (except Ml, small
males <70 cm) released off the west coast of Vancouver
Island were recaptured in other areas in high proportions
(Table 5).
Discussion
The low recovery percentage in this study is probably
related to a low reporting rate. Until recently spiny dogfish
catches in Canadian waters were discarded without being
examined. However, this low recovery percentage also
reflects the high abundance of spiny dogfish off the west
coast of North America. Ketchen (1986) reported abun-
dance estimates of 300 000 t for the whole North American
coast and Saunders (1989) estimated 210,000-260,000 t
in Canadian waters. It is clear that spiny dogfish are
common and relatively abundant from northern Oregon
to southeastern Alaska. Knowledge of the movements of
spiny dogfish within eastern waters and between eastern
and western waters is still limited but indicates that this
shark is a key species in those coastal ecosystems. The use
of the more durable modified Petersen tag in other areas
would add to our knowledge of dogfish movement between
these ecosystems.
Table 4
Mean percentage (%) of recaptured fish
(at liberty <5
years) by
area of release and known area of recapture |
including
Puget Sound
(PS), Washington
State coastal
waters (WS), and other
areas. Standardized percentages 1
could be calculated for only Strait of Georgi
a (SOG), west
coast Vancouver Island (WCVl), Puget Sound, and Wash- |
ington State coastal waters and are based
on total stan-
dardized
recaptures for
these areas only.
Standardized
percentages could not be calculated for northern British |
Columbia
(NBC).
Release
Recapture
Percentage of recaptures
Non-
area
area
standardized
Standardized
SOG
SOG
91
86
WCVl
5
7
NBC
2
PS
1
1
WS
1
6
Other
0
WCVl
WCVl
62
20
SOG
13
5
NBC
4
PS
5
2
WS
5
73
Other
11
NBC
NBC
51
SOG
13
WCVl
28
PS
0
WS
1
Other
7
Holland (1957), reporting on tagging studies conducted
in the 1940s, concluded that Puget Sound and the Strait
of Georgia supported indigenous populations. However,
Ketchen ( 1986) reviewing studies by Foerster ( 1942), Fujio-
ka ct al. ( 1974 ), and McFarlane et al. ( 1982) suggested more
movement between the inside populations than reported in
McFarlane and King: Migration patterns of Squalus acanthias
365
Table 5
Percentage (%) of tagged dogfish by size
at release and sex-size category across
recapture areas. (Ml=males <70
cm
M2=males >70
cm. Fl=females <70 cm; F2=females 71
-85
cm; F3=
= females >86 cm.)
Recapture area
West coast
Northern
Release area
Sex
size category
Strait of Georgia
Vancouver Island
British Columbia
Strait of Georgia
Ml
M2
Fl
F2
F3
90
85
53
94
86
6
6
2
6
8
4
9
45
0
6
West coast of Vancouver Island
Ml
M2
Fl
F2
F3
0
37
17
0
39
100
63
61
23
50
0
0
22
77
11
Northern British Columbia
Ml
M2
Fl
F2
F3
0
9
9
0
6
0
9
21
6
8
100
82
70
94
86
earlier studies. Ketchen ( 1986) concluded that populations
in inside waters (i.e. Strait of Geor^a and Puget Sound)
are largely independent of those off the open coast. Our
study supports the idea that the majority of spiny dogfish
in the Strait of Georgia remain in the Strait of Georgia.
However, the low proportion of standardized recoveries in
Puget Sound suggests very little movement between these
two areas. In fact, a higher proportion of standardized re-
coveries of spiny dogfish from the Strait of Georgia were
reported for coastal waters of Washington State than for
Puget Sound.
Until this present study, little tagging had been con-
ducted in open waters off the west coast of North America
(Bonham et al., 1949; Holland, 1957; Ketchen, 1986).
Holland (1957) observed a tendency for fish tagged off
Washington and Vancouver Island to move south in the
fall and winter, and north in spring and summer. The
distance travelled (with a few exceptions) was generally
small. Ketchen (1986) reported a similar pattern (based
on seasonal fishery catches during the liver fishery of the
1940s) but noted that fishing did occur year round from
northern British Columbia to Oregon. He concluded that
some individuals may traverse the full commercial range
of the species (Oregon to northern British Columbia) be-
tween summer and winter, but these instances are more
the exception than the rule.
Seasonal movement aside, it is clear from our study that
male and female spiny dogfish of all size categories in open
coastal areas migrate considerably farther than previous
studies suggest. For example, in the recapture areas where
abundance estimates are similar and standardization to
catch is possible (Strait of Georgia, Puget Sound, west
coast Vancouver Island, Washington State coastal waters),
the percentage of recaptures for west coast Vancouver
Island releases indicates substantial movement south to
Washington State coastal waters. This movement is greatly
underestimated with nonstandardized data. Unfortunately
it is not possible to standardize recapture data for all re-
leases because of the paucity of fishing data or abundance
estimates. Outside of Canadian waters, excluding Washing-
ton State (Puget Sound and coastal waters), there are no
targeted spiny dogfish fisheries (i.e. landings are typically
less than 10 t) and therefore the proportion of recaptured
spiny dogfish in these areas would be expected to be small.
However, it is possible to comment on the long-range move-
ments (up to 7000 km) of these open coastal dogfish.
Tagged spiny dogfish released between 1980 and 1987
in open coastal waters (west coast of Vancouver Island
and northern British Columbia) underwent extensive mi-
grations and approximately 16'7( of recaptured fish came
from outside Canada waters. From earlier studies, a few
recaptured fish indicated that some spiny dogfish at least
are highly mobile. In the early 1940s a large male dogfish
tagged in northern British Columbia waters was recovered
off California 171 days later (Manzer, 1946). Holland ( 1957)
reported that a fish tagged off the west coast of Vancouver
Island was recaptured off Baja California. One trans-Pacif-
ic migration (Washington State to Japan) was reported by
Kauffman ( 1955). In our study, the 30 fish captured off Ja-
pan were all (with one exception) tagged in outside waters.
These recaptures represented 6% of the recaptures from
these release areas. Similarly, nine fish were captured off
366
Fishery Bulletin 101(2)
California and Mexico, all from outside tag release areas.
These rather remarkable recaptures do provide evidence
for the trans-Pacific connection of spiny dogfish stocks but
pose the question of the significance of such east to west
exchanges. One possible significance would be the transfer
of genetic material. The fact that these recaptures occurred
regularly from 1982 to 2001 suggests ongoing migration be-
tween areas. Two recent recoveries off Japan support this
hypothesis. One fish, a 69-cm female was released off the
west coast of Vancouver Island in 1984, and the other, also
a female (72-cm) was released in the Strait of Georgia in
1988. Both fish were recovered in May 2001 off Hokkaido.
Although the evidence is limited, and the magnitude of
the exchange between eastern and western Pacific stock,
and indeed northern Canadian fish and those off southern
California and Mexico, is small, it is clear that the inter-
relationships between these areas needs to be examined if
ecosystem management incorporating these apex preda-
tors is to be developed.
In the eastern North Pacific, the management of spiny
dogfish was recently identified as a priority by the Ameri-
can Fisheries Society (Musick et al., 2000). In addition, the
global decline in many shark populations, and in particular
spiny dogfish in the Atlantic ocean (Stevens et al., 2000),
raises the question: What are the effects of the removal of
large numbers of sharks (spiny dogfish) on marine ecosys-
tems? A recent attempt (Stevens et al., 2000) to examine
this question (albeit a somewhat simplistic one) identified
a number of significant ecological and economic impacts.
The study illustrated that under differing conditions, the
consequences of depleting sharks in certain ecosystems are
complex and could lead to unforeseen consequences that
extend beyond the fished ecosystem. The highly migratory
nature of many shark species complicates management
efforts (Musick et al., 2000), and adding to the complexity
are relationships between distribution, migration, and en-
vironmental conditions, such as those documented for blue,
salmon, and thresher sharks in the northern Pacific (Holts,
1988; McKinnell and Seki, 1998; Bigelow et al., 1999). The
development of effective ecosystem-based management
hinges on understanding the implications of indirect
and direct impacts on ecosystem structure and function
(Fogarty and Murawski, 1998) and will require improved
understanding of 1) species interactions, i.e. what, when,
and where dogfish eat, and what eats dogfish, 2) migration
patterns (both seasonal and long term) from and between
all ecosystems within the range for dogfish (identified in
this report), and 3) changes in migration and distribution
in relation to changing climate and ocean conditions. To
date, the effects of the removal of large numbers of spiny
dogfish remain essentially unexamined, in part because of
the limited information in each of these three areas.
Acknowledgments
Wo thank Bill Andrews, Brad Beailli, Mark Saunders,
Mike Smith, and Maria Surry for conducting field work,
maintaining databases, and preliminary production of
tables and figures.
Literature cited
Bigelow, K. A., C. H. Biggs, and X. He.
1999. Environmental effects on swordfish and blue shark
catches in the US North Pacific longline fishery. Fish.
Ocean. 8:178-198.
Bonham, K., F. B. Sanford, W. Clegg and G. C. Brucker.
1949. Biological and vitamin A studies of dogfish (Squalus
acanthias) landed in the State of Washington. Wash. Dep.
Fish. Biol. Rep 49A:83-114.
Foerster, R. E.
1942. Dogfish tagging — preliminary results. Fish. Res.
Board Can., Pac. Progr. Rep. 53:12-13.
Fogarty, M. J., and S. A. Murawski.
1998. Large-scale disturbance and the structure of marine
systems: fishing impacts on Georges Bank. Ecol. Appl. 8
(suppl. l):56-522.
Fujioka. B. P., and G. DiDonato.
1974. Dogfish tagging studies in Washington waters. In
Puget Sound dogfish (Squalus acanthias) studies, 85 p.
Wash. Dep. Fish, Mar. Fish. Invest., Suppl. Progr. Rep.
74-01.
Holden, M. J.
1977. Elasmobranchs. In Fish population dynamics (A.
Gulland, ed.), p. 187-215. J. Wiley and Sons, New York,
NY.
Holland. G. A.
1957. Migration and growth of the dogfish shark, Squalus
acanthias (Linnaeus) of the eastern North Pacific. Wash.
Dep. Fish. Res. Pap. 2(l):43-59.
Holts, D. B.
1988. Review of US west coast commercial shark fisheries.
MarFish. Rev. 50(1): 1-8.
Kauffman, D. E.
1955. Noteworthy recoveries of tagged dogfish. Wash. Dep.
Fish. Res. Pap. l(3):39-40.
Ketchen. K. S.
1986. The spiny dogfish (Squalus acanthias)in the northeast
Pacific and a history of its utilization. Can. Spec. Publ.
Fish. Aquat. Sci. 88. 78 p.
McFarlane, G. A., and R. J. Beamish.
1986. A tag suitable for assessing long-term movements of
spiny dogfish and preliminary results from use of this tag.
N. Am. J. Fish. Manage. 6: 69-76.
1982. Validation of the dorsal spine method of age determi-
nation for spiny dogfish. In Age and growth offish (R. C.
Summerfelt and G. E. Hall, eds.), p. 287-300. Iowa State
Univ. Press, Ames, lA.
McFarlane. G. A., R. J. Beamish, M. S. Smith, V, Egan, and
D. Brown
1982. Results of spiny dogfish (Squalus acanthias) tagging
in the Strait of Georgia, Queen Charlotte Sound, Hecate
Strait and Dixon Entrance, during 1980. Can. Man. Rep.
Fish. Aquat. Sci. 1646, 123 p.
McKinnell, S., and M. P Seki.
1998. Shark bycatch in the Japanese high seas squid drift-
net fishery in the North Pacific ocean. Fish. Res. 39:
127-138.
Manzer, J. I.
1946. Interesting movements as shown by the recovery of
certain species of tagged fish. Fish. Res. Board Can. Pac.
Progr. Rep. 67, 31 p.
Musick, J. A., G. Burgess, G. Cailliet, M. Camhi, and S. Fordham.
2000. Management of sharks and their relatives (Elasmo-
branchiil. Fisheries 25l3):9-13.
McFarlane and King: Migration patterns of Squalus acanthias
367
Saunders, M. W.
1989. Dogfish. In Groundfish stock assessments for the
West Coast of Canada in 1988 and recommended yield
options for 1989 (J. Fargo and A.V. Tyler, eds.), p. 169-176.
Can. Tech. Rep. Fish. Aquat. Sci. 1646.
Saunders, M. W., and G. A. McFarlane.
1993. Age and length at maturity of the female spiny dogfish
(Squalus acanthias) in the Strait of Georgia, British Colum-
bia, Canada. Environ. Biol. Fishes 38:49-57.
Stevens, J. D., R. Bonfil, N. K. Duluy and P. A. Walker
2000. The effects of fishing on sharks, rays, and chimaeras
(chondrichthyans) and implications for marine ecosystems.
ICES J. Mar. Sci. 57:476-494.
Ware, D. M., and G. A. McFarlane.
1995. Climate-induced changes in Packific hake (Merluccius
productus) abundance and pelagic community interactions
in the Vancouver Island upwelling system. In Climate
change and northern fish populations (R. J. Beamish, ed.),
p. 509-521. Can. Spec. Pub. Fish. Aquat. Sci. 121.
Wydoski, R., and L. Emery.
1983. Tagging and marking. In Fisheries techniques (L. A.
Nielson and D. L. Johnson, eds.), p. 215-237. Am. Fish.
Soc, Bethsuda, MD.
368
Abstract— Larval development of the
southern sea garfish (Hyporhamphus
melanochir) and the river garfish (H.
regularis) is described from specimens
from South Australian waters. Larvae
of H. melanochir and H. regularis
have completed notochord flexion at
hatching and are characterized by an
elongate body with distinct rows of
melanophores along the dorsal, lat-
eral, and ventral surfaces; a small to
moderate head; a heavily pigmented
and long straight gut; a persistent pre-
anal finfold; and an extended lower
jaw. Fin formation occurs in the follow-
ing sequence: caudal, dorsal and anal
(almost simultaneously), pectoral, and
pelvic. Despite the similarities between
both species and among hemiramphid
larvae in general, H. melanochir larvae
are distinguishable from H. regularis by
1) having 58-61 vertebrae I vs. 51-54 for
H. regularis); 2) having 12-15 melano-
phore pairs in longitudinal rows along
the dorsal margin between the head
and origin of the dorsal fin (vs. 19-22
fori/, regularis); and 3) the absence of a
large ventral pigment blotch anteriorly
on the gut and isthmus (present in H.
regularis). Both species can be distin-
guished from similar larvae of southern
Australia (other hemiramphids and a
scomberosocid ) by differences in mer-
istic counts and pigmentation.
Larval development of the southern sea garfish
(Hyporhamphus melanochir) and
the river garfish U-l. regularis)
(Beloniformes: Hemiramphidae)
from South Australian waters
Craig J. Noell
Department of Environmental Biology
Adelaide University
Soutti Australia 5005
Present address: SARDI Aquatic Sciences
PO Box 120
Henley Beach
South Australia 5022
E-mail address: noell craigiffisaugovsa au
Manuscript accepted 25 October 2002.
Manuscript received 31 December 2002
at NMFS Scientific Publications Office.
Fish. Bull. 101:368-376(20031.
The beloniform family Hemiramphidae
(garfishes or halfbeaks) are small to
medium-size surface-dwelHng marine,
estuarine, and freshwater fishes. The
family contains 12 genera and 101
species worldwide, and more than one-
third of the species belong to the genus
Hyporhamphus (Froese and PaulyM.
The Hemiramphidae are related to
the Exocoetidae (flyingfishes) and,
more distantly, to the Scomberosocidae
(sauries), Belonidae (needlefishes), and
Adrianichthyidae (ricefishes) (Collette
et al., 1984). Six genera and 17 species
of hemiramphids occur in Australian
waters, where garfishes have long been
considered valuable food and bait fish
(Collette, 1974; Kailola et al., 1993).
Two hemiramphid species inhabit
the waters of South Australia (S.A.),
namely the southern sea garfish Hypo-
rhamphus melanochir (Valenciennes,
1846) and the river garfish H. regularis
(Giinther, 1866). Adults of both are
widely distributed along southern Aus-
tralia from Western Australia (W.A.) to
New South Wales, although H. regula-
ris have not been recorded in Tasmania
(Tas.). They support important com-
mercial and recreational fisheries, par-
ticularly in S.A. (Kailola et al., 1993). H.
melanochir are commonly found in shel-
tered coastal waters, whereas H. regu-
laris are confined to estuaries (Jones
et al., 1996). Juveniles and adults of
both species co-occur in some estuaries
of southern Australia, e.g. Port River-
Barker Inlet of S.A. (34°45'S, 138°31'E)
(Jones et al., 1996) and Peel-Harvey Es-
tuary of W.A. (32°32'S, 115°43'E) (Noell,
unpubl. data).
Despite their widespread distribution
and economic importance, the early
life history of H. melanochir is only
partially described (i.e. reproductive
biology [Ling, 1958]; egg development
(Jordan et al., 1998], and there is no
published information for H. regularis.
Furthermore, although adults are eas-
ily identified with keys and descriptions
provided by Collette (1974), no such in-
formation exists for the larvae. A funda-
mental prerequisite for any larval fish
study is, undoubtedly, their accurate
identification (Neira et al., 1998).
Thus far, at least some larval stages
have been described for 19 hemiram-
phids worldwide (Sudarsan, 1966; Tai-
wan 1967; Hardy 1978; Chen, 1988; Wat-
son, 1996; Prince Jeyaseelan, 1998),
eight of which belong to Hyporham-
phus. The purposes of this paper are
to describe the larval development of
H. melanochir and H. regularis and
to document distinguishing characters
between larvae of these species.
' Froese, R, and D. Pauly 2001. FishBase.
World Wide Web electronic publication.
Accessed 28 Nov 2001. Web site: www.fish
base.org.
Noell: Larval development of Hyporhamphus melanochii and H. regulans
369
Materials and methods
Most larvae were collected with a neuston net in Gulf
St, Vincent (34°29'S, ISS'IS'E) and the Bay of Shoals
(SS^ST'S, 137"37'E) of South Australia. The neuston net
was a square-framed bongo net with a mouth area of 0.5 m-^
fitted with 500-|jm mesh, to which a 30-cm diameter pneu-
matic float was attached to both sides of the frame. This
attachment ensured that, while being towed, the top of the
frame rode steadily above the water surface and that -0.4
m^ of the mouth area was submerged. The net was towed
from the stern of the vessel inside a circular direction for 5
min at speeds of 2-4 knots. Additional larvae were collected
by hand from beneath a wharf in Barker Inlet where they
often school during daylight at mid-flood tide. Transform-
ing larvae and juveniles were collected at night with a
dip net and spotlight at Outer Harbor (34°46'S, 138°28'E)
and Barker Inlet. The term "transforming" is used here
to describe the stage between the end of the larval phase
and the start of the juvenile phase, i.e. after the attain-
ment of all fin rays and before the formation of scales. All
specimens examined in this study were collected between
November and March. Larvae were sorted from plankton
samples immediately after collection based on reference
larval specimens from the South Australian Museum fish
collection that were identified to family. Larvae were fixed
in 10% formalin buffered with sodium j3-glycerophosphate
( 1 g/L) and later preserved in 70% ethanol.
A total of 47 H. melanochir (6.4—48.3 mm body length,
BL) and 49 H. regularis (7.0-46.9 mm BL) larvae through
juveniles were used to describe morphometries, meristics,
and pigmentation. Larvae were identified as hemiram-
phids based on larval and adult characters reported in the
literature (Collette, 1974; Hardy, 1978; Collette et al., 1984;
Chen, 1988; Watson, 1996; Trnski et al., 2000). Develop-
mental series were assembled by using the series method
(Neira et al., 1998), the accuracy of which was verified by
a molecular technique (Noell et al., 2001). Terminology of
early life history stages follows that of Kendall et al. ( 1984).
Representative series for both species are deposited with
the I.S.R. Munro Fish Collection (CSIRO, Hobart, Tas.).
(Registration numbers: H. melanochir (n = 13), CSIRO L
3072-01, 3073-01 to -08, 3074-01 to -02, 3075-01 to -02; H.
regulans (n=12), CSIRO L 3076-01 to -07, 3077-01 to -02,
3078-01 to -03.)
Larvae were examined with a Wild M3Z stereomicro-
scope at 6.5-40x magnifications by using various com-
binations of incident and transmitted light. Body mea-
surements were taken with SigmaScan Pro® 4.01 image
measurement software (SPSS Inc., 1999) and are accurate
to less than 0.05 mm. This method was particularly useful
for measuring cumulative distances of bent larvae. Abbre-
viations and definitions of routinely taken body measure-
ments follow Leis and Carson-Ewart (2000). Lower jaw
length (LJ) is defined as the horizontal distance from the
tip of the lower jaw to the anterior margin of the pigmented
region of the eye. Lower jaw extension (LJx) is defined as
the horizontal distance from the tip of the lower jaw to
the tip of the snout. Eye diameter was measured along
both horizontal (EDh) and vertical midlines (EDv) of its
pigmented region. Body depth was measured at two points:
at the pectoral base (BDp) and at the anus (BDa). Other
measurements taken were snout length (SnL), head length
(HL), pre dorsal-fin length (PDL) and preanal length (PAL).
All measurements are expressed as a percentage of BL.
Pigment refers to melanin. Drawings were prepared with
the aid of a camera lucida.
Selected specimens were cleared and stained with alcian
blue and alizarin red-S, following the method of Potthoff
( 1984), in order to count fin rays and vertebrae. Myomeres
were difficult to count reliably at either end and thus
vertebral counts (which include the urostyle) of stained
larvae were taken instead. For small larvae that had un-
formed centra, corresponding neural or haemal spines were
counted to obtain the number of vertebrae.
Results
Southern sea garfish (Hyporhamphus melanochir
Valenciennes, 1846) (Fig. 1 )
Description of larvae The smallest H. melanochir larva
examined was a 6.4-mm newly hatched, laboratory-reared,
postflexion-stage specimen. Some yolk remained, although
yolk absorption was complete in the smallest field-collected
larva (6.9 mm).
Larvae are elongate to very elongate (BDp=7-13% BL),
and have a body depth slightly tapered towards the anus
(BDa= 7-9% BL). Relative body depth at the pectoral base
decreases slightly during larval development (Table 1).
Larvae have 58-61 vertebrae (Table 2). The gut is relatively
thick, long, straight, and nonstriated. PDL and PAL remain
in the ranges of 70-75% and 71-76% BL, respectively (ex-
cept for the 17.0-mm larva, which had a PDL and PAL of
62% BL). The first dorsal-fin ray is slightly anterior to or di-
rectly above the corresponding anal-fin ray. There is no gap
between the anus and the anal fin. A long preanal finfold,
initially the same length as the gut, persists through to
the transformation stage before it disappears. There is no
head spination. The small to moderate head (HL= 16-24%
BL) decreases in size in relation to BL with larval growth
(Table D.The longer lower jaw protrudes beyond the snout
(LJx) by 4% BL at 11.0-11.5 mm, increasing to a maximum
of 34% BL in the 29. 3-mm juvenile. The mouth is oblique
and reaches to the level of the center of the eye in newly
hatched larvae. The maxilla subsequently moves forward
in relation to the eye and by 12.1-14.4 mm it does not reach
the eye. Very small villiform teeth are present on both the
premaxilla and dentary in newly hatched larvae. The mod-
erate to large eye (EDh=6-109} BL or 33-42% HL) is elon-
gate (EDv=78-88% EDh) and decreases in size in relation
to BL. A single rudimentary nasal papilla first appears as
a small fleshy lump in the olfactory pit by 17.0 mm. Scales
first appear between 20.4 and 29.3 mm laterally on the tail,
anterior to the caudal peduncle.
Development of fins Completion of fin development in H.
melanochir occurs in the following sequence: C — ► D — ► A
-* Pj, P2 (Table 2). All principal rays of the caudal fin (7-1-8)
370
Fishery Bulletin 101(2)
E
Figure 1
Larval, transforming larval, and juvenile Hyporhamphus melanochir . (A) 6.4-mm reared yolksac larva; newly
hatched (redrawn from Jordan et al., 1998) (L 3072-01). (B) 9.3-mm larva (L 3073-01). (C) 13.3-mm larva
(composite drawing of two damaged larvae of same BL) (L 3073-02 and -03). (D) 20.4-mm transforming larva
(L 3074-01). (E) 29.3-mm juvenile (L 3074-02). Myomeres omitted in (D) and (E).
and several incipient dorsal- and anal-fin rays are pres-
ent in newly hatched larvae. A full complement of 15-18
dorsal-fin and 17-20 anal-fin rays is attained at 11.4 and
12.1 mm, respectively. The pectoral base and finfold form
prior to hatching, and incipient rays appear shortly after
(by 7.2 mm); all 11-13 rays are formed by 19.6 mm. The
pelvic fin buds appear by 13.3 mm, and all six pelvic-fin
rays are formed by 19.6 mm.
Pigmentation tiy/xirpliiimphiis mi'laiUH-hir larvae are
moderately to heavily pigmented. Head pigmentation
consists of melanophores on the tip of the lower jaw, snout,
olfactory pit, and opercula, and a patch of several large
melanophores on the midbrain. The extended lower jaw
is heavily pigmented throughout development and mela-
nophores extend laterally along the dentary. The eye is
partially pigmented in the newly hatched larva, but fully
pigmented by 6.9 mm. The gut is heavily and uniformly
pigmented dorsally and laterally along the entire length,
and melanophores are often coalesced, but pigmentation
becomes obscured as the overlying musculature develops.
Dorsal pigmentation initially consists of 12-15 large mela-
nophore pairs in longitudinal rows between the head and
origin of the dorsal fin (Fig. 2A), and a continuous band
along either side of the dorsal-fin base. Dorsal pigmenta-
tion gradually decreases in intensity thereafter. Three
Noell: Larval development of Hyporhamphus melanochtr and W regularis
371
Table 1
Morphometries of larval, transforming larval, and juvenile Hyporhamphus melanochir (expressed as % of BL). Mean ±SD is given
when sample size n >1. Dashed lines differentiate larvae, transforming larvae, and juveniles in descending order.
BL(mm)
n
SnL
LJ
LJx
EDh
EDv
HL
PDL
PAL
BDp
BDa
6.4'
1
2.1
2.7
0.6
9.9
8.7
24.4
74.6
75.5
16.32
8.0
6.9
1
3.0
4.0
1.0
9.2
7.5
23.5
69.7
71.9
12.7
8.6
7.0-7.5
9
2.8 ±0.8
3.9 ±
1.1
1.1 ±0.4
9.1 ±0.3
7.3 ±0.3
22.0 ±0.8
70.8 ±0.9
72.6 ±1.0
11.9 ±0.3
8.9 ±0.3
7.5-8.0
7
3.6 ±0.9
4.6 ±
1.2
1.0 ±0.4
9.1 ±0.5
7.1 ±0.3
22.8 ±
1.6
71.3 ±1.3
73.0 ±0.9
11.8 ±0.6
8.9 ±0.7
8.0-8.5
9
3.6 ±0.5
4.9 ±0.8
1.3 ±0.4
8.8 ±0.3
7.1 ±0.3
21.9 ±
1.0
70.9 ±0.7
72.7 ±0.9
11.7 ±0.7
8.9 ±0.7
8.5-9.0
3
3.5 ±0.8
5.0 ±
1.1
1.5 ±0.3
8.6 ±0.2
7.0 ±0.3
21.1 ±0.2
71.5 ±0.2
72.5 ±0.4
11.3 ±0.5
9.1 ±0.5
9.0-9.5
4
3.4 ±0.5
5.0 +
0.7
1.6 ±0.3
8.0 ±0.3
6.5 ±0.2
20.9 ±0.8
71.7+0.7
72.8 ±0.6
11.4 ±0.7
8.4 ±0.3
11.0-11.5
4
3.7 ±0.7
7.2 ±
1.5
3.5 ±1.2
7.5 ±0.4
6.3 ±0.2
20.3 ±
1.3
71.6 ±0.6
72.3 ±0.5
10.6 ±0.7
8.5 ±0.6
12.1
4.1
9.0
4.9
7.6
6.4
19.7
72.6
72.6
10.2
8.6
14.4
3.7
12.3
8.6
6.9
5.8
19.2
70.2
71.4
9.4
7.8
17.0
2.9
10.3
7.4
5.6
4.7
15.9
61.7
61.7
7.3
6.6
19.6
4.9
28.7
23.8
6.0
4.9
18.6
70.4
71.2
8.2
7.4
20.4
4.0
24.2
20.2
5.7
5.0
17.5
72.5
71.6
8.9
7.6
29.3
4.4
38.4
34.0
5.3
4.8
17.0
69.9
71.0
8.2
7.2
.3.3.3
4.9
38.3
33.4
5.3
4.6
17.9
71.7
72.7
8.5
7.6
41.3
5.2
36.2
31.1
5.5
5.1
18.6
74.1
74.1
9.2
7.8
48.3
5.6
36.9
31.3
5.2
4.8
17.8
74.0
74.0
9.7
8.1
' Yolksac larva.
-' Includes yo
k sac.
Table 2
Meristic counts of larval, transforming larval,
and juvenile
Hyporhamphus melanochir
Numbers
in bold indicate the BL at which a
full complement of rays is
first attained
Dashed lines differentiate larvae
, transforming larvae, and juveniles in
descending order.
D = dorsal; A =
anal
Pi =
pectoral
?.,=
pelvii
; C = caudal.
BL(mm)
-
Fin rays
Branchiostegal rays
Vertebrae
D
A
Pi
P2
C
6.4'
8
9
base
0+7+8+0
3
38+21
7.2
8
8
1
0+7+8+0
3
39+20
7.3
9
10
1
0+7+8+0
3
39+19
7.6
11
11
1
0+7+8+0
3
38+20
7.9
10
11
1
0+7+8+0
3
40+20
8.3
11
11
2
0+7+8+0
4
39+20
8.4
13
14
2
1+7+8+1
5
39+21
9.4
14
16
4
1+7+8+1
5
40+21
11.4
15
16
6
2+7+8+1
7
38+20
12.1
16
17
7
2+7+8+2
7
39+20
14.4
16
19
9
bud
2+7+8+2
9
39+20
19.6
17
19
11
6
4+7+8+4
12
38+20
20.4
16
17
12
6
4+7+8+4
12
39+19
29.3
17
18
11
6
4+7+8+4
13
38+20
33.3
17
18
12
6
5+7+8+5
12
38+20
41.3
16
19
11
6
4+7+8+5
12
40+19
48.3
16
19
11
6
4+7+8+5
12
39+19
' Yolksac larva.
372
Fishery Bulletin 101(2)
Figure 2
Pigmentation of an 8.5-mni Hyporhamphus melanochir larva. (A)
Dorsal view; arrows indicate the margins of the 12-15 melano-
phore pairs in longitudinal rows. (B) Ventral view.
Table 3
Morphometries of larval, transforming larval, and juvenile Hyporhamphus regularis (expressed as % of BL). Mean ±SD is given
when sample size n > 1. Dashed lines differentiate larvae, a transforming larva, and juveniles in descending order
BL(mm)
7.0'
7.5-8.0
8.0-8.5
8.5-9.0
9.0-9.5
9.5-10.0
10.0-10.5
13.1
SnL
LJ
LJx
EDh
EDv
HL
PDL
PAL
BDp
1
9
12
10
5
3
3
1
3.2
2.8 ±0.3
2.8 ±0.4
2.8 ±0.2
2.8 ±0.1
2.8+0.1
3.0 ±0.3
4.0
4.4
4.4 ±0.4
4.3 ±0.5
4.2 ±0.2
4.4 ±0.4
4.4 ±0.3
4.8 ±0.9
7.6
1.2
1.7 ±0.3
1.5 ±0.3
1.5 ±0.2
1.6 ±0.4
1.7 ±0.2
1.8 ±0.6
3.7
8.2
7.6 ±0.1
7.5 ±0.3
7.2 ±0.2
7.0 ±0.1
6.9 ±0.4
6.8 ±0.2
6.9
6.5
6.3 ±0.1
6.2 ±0.2
5.9 ±0.1
6.0 ±0.2
5.8 ±0.2
5.8 ±0.2
5.4
20.4
19.9 ±0.6
19.6 ±0.9
19.2 ±0.3
19.1 ±0.4
18.7 ±0.5
18.9 ±0.6
19.6
73.
73,
72,
72,
72,
72,
72
73
2
1 ±0.6
8 ±0.7
9 ±0.9
3 ±0.7
1±1.6
2 ±0.6
2
71.6
71.8 ±0,
71.6 ±0,
71.8 ±0
71.5 ±0,
71.2 ±1
71.3 ±0
71.9
11. 6-'
11.2 ±0.2
11.0 ±0.5
10.6 ±0.2
10.6 ±0.2
10.6 ±0.4
10.2 ±0.3
9.1
BDa
7.4
8.2 ±1.2
7.7 ±0.3
7.4 ±0.3
7.6 ±0.3
7.3 ±0.2
7.6 ±0.3
7.8
18.1
1 4.5
18.6
14.1
6.3
5.4
19.9
73.8
72.6
9.5
8.1
24.7
31.5
33.8
46.9
1 5.6
1 6.2
1 6.4
1 7.3
27.7
30.0
27.7
22.1
23.9
21.3
6.3
6.0
6,1
damaged damaged 5.6
5.5
5.5
5.6
4.7
20.8
21.0
21.6
21.8
72.
74,
74.
75,
73.6
75.5
74.3
75.4
9.7
10.7
10.9
11.2
8.1
8.5
8.9
9.8
' Yolksac larva.
^ Includes yolk sac.
distinct lines of pigment appear along the dorsal margin
in juveniles (by 29.3 mm) and remain to adult stage. A
series of melanophores form a dashed, sometimes continu-
ous, midlateral line. Melanophores appear laterally on the
caudal peduncle by 14.4 mm and then proliferate anteri-
orly to form a broad medial stripe that remains, forming a
silver stripe from the caudal peduncle to the operculum of
adults. Ventral pigmentation consists of continuous bands
of melanophores either side of the anal-fin base (Fig. 2B).
Fins are unpigmented, except the caudal fin, which has
small melanophores on the ray bases.
River garfish {Hyporhamphus regularis
Gunther, 1886) (Fig. 3)
Description of larvae The smallest H. regularis larva ex-
amined (7.0 mm) had completed notochord flexion and had
a yolk sac. Yolk absorption was complete by 7.6 mm.
Larval H. regularis closely resemble larval H. melanochir
morphologically (see Tables 1 and 3), but differ somewhat
in relative length of the lower jaw, relative positions of the
dorsal- and anal-fin origins, and in number of vertebrae.
The longer lower jaw protrudes beyond the snout (LJx) by
Noell: Laival development of Hyporhamphus melanochir and H. regulans
373
Figure 3
Larval and juvenile /fvpor/iomp/i us regularis. (A) 7.1-mm yolksac larva (L 3076-01 ). (B) 9.4-mm larva
(L 3076-02). (C) 12.3-mm larva (L 3076-03). (D) 15.5-mm larva (L 3076-04). (E) 24.7-mm juvenile (L
3077-01). Myomeres omitted in (D).
4% BL at 13.1 mm and increases to a maximum of 24% BL
in the 31.5 mm juvenile. The first dorsal-fin ray is shghtly
posterior to or directly above the corresponding anal-fin
ray. Larvae have 51-54 vertebrae. Scales first appear be-
tween 18.1 and 24.7 mm laterally on the tail, anterior to
the caudal peduncle.
Development of fins Completion of fin development in H.
regularis occurs in the following sequence: C — ► D — ► A — >
Pj, Pg (Table 4). Development of the caudal fin is incomplete
at birth; 6-1-7 principal rays are present in the 7.0-mm
yolksac larva, and the full complement (7-t-8) shortly after,
by 7.7 mm. Distinct anal-fin bases are visible at 7.0 mm. A
full complement of 14-17 dorsal and 15-19 anal-fin rays
is attained at 10.1 and 10.5 mm, respectively. The pectoral
base and finfold are present at birth, and incipient rays first
appear by 8.1 mm; all 11-12 rays are formed by 18.1 mm.
The pelvic fin buds appear by 13. 1 mm, and all six pelvic-fin
rays are formed by 18.1 mm.
Pigmentation Pigmentation of H. regularis larvae is
similar to that of H. melanochir larvae except along the
dorsal and ventral margins. Dorsal pigmentation consists
of 19-22 melanophore pairs in longitudinal rows between
the head and dorsal fin origin (Fig. 4A). A large pigment
blotch is present ventrally on the isthmus and anteriorly
on the gut.
Discussion
This study provides the first descriptions of larval develop-
ment of hemiramphids endemic to marine (H. melanochir)
and estuarine (//. regularis) waters of Australia.
374
Fishery Bulletin 101(2)
Table 4
Meristic counts of larval, transforming
arval,
and juvenile Hyporhamphus regularis
Nun
bers
in bold indicate the BL at which a |
full complement of rays is first attained
Dashed lines differentiate larvae
, a transforming 1
arva
and juveniles in
descending order.
D = dorsal; A =
anal; P, = pectoral
,P2 =
pelvic
; C = caudal.
BL(mm)
Fin rays
Branchiostegal rays
Vertebrae
D
A
Pi
P2
C
7.0'
anlage
bases
base
0+6+7+0
2
35+19
7.7
4
6
base
0+7+8+0
3
34+19
7.8
6
7
base
0+7+8+0
4
34+18
8.1
5
7
1
0+7+8+0
4
34+19
8.3
8
9
2
0+7+8+0
4
33+20
8.6
9
11
2
0+7+8+0
5
34+19
8.9
11
11
2
0+7+8+1
5
34+20
9.3
11
12
3
1+7+8+1
5
33+18
9.6
10
11
3
0+7+8+1
5
35+19
10.1
14
14
4
1+7+8+1
6
33+20
10.5
13
15
5
1+7+8+1
6
35+19
13.1
14
16
7
bud
2+7+8+2
8
35+19
18.1
14
17
11
6
4+7+8+4
12
35+18
24.7
16
17
12
6
4+7+8+4
11
34+19
31.5
15
17
12
6
4+7+8+4
11
34+18
33.8
15
18
11
6
4+7+8+4
11
33+19
46.9
16
17
11
6
4+7+8+4
11
35+18
' Yolksac larva.
Both H. melanochir and H. regularis share charac-
ters common to other described hemiramphid larvae.
They are generally characterized by their lack of
head or fin spines; elongate body; long straight gut;
extended lower jaw; a main pigmentation pattern con-
sisting of rows of melanophores on the dorsal, lateral,
and ventral surfaces of the body; and advanced state
of development at hatching (Collette et al., 1984;
Watson, 1996; Trnski et al., 2000). Although the size
at which fins develop varies slightly between H. mela-
nochir and H. regularis, the sequence of development
for both species is the same as that for most hemiram-
phids, i.e. C -> D. A -► P, -► P.^ (Collette et al., 1984).
Hyporhamphus melanochir larvae are distinguish-
able from H. regularis by 1) having 58-61 vertebrae
(vs. 51-54 for H. regularis); 2) having 12-15 mela-
nophore pairs in longitudinal rows along the dorsal
margin between the head and origin of the dorsal
fin (vs. 19-22 for H. regularis); and 3) the absence
of a large ventral pigment blotch anteriorly on the
gut and isthmus which is present in H. regularis.
Despite the difficulty in counting myomeres, either
the number of vertebrae in cleared and stained specimens
or the number of myomeres between the pectoral-fin base
and anus (usually three less than the number of precaudal
vertebrae; see Tables 2 and 4) revealed a consistent differ-
ence between both species.
Figure 4
Pigmentation of an 8.7-mm Hyporhamphus regularis larva. (Al
Dorsal view; arrows indicate the margins of the 19-22 melano-
phore pairs in longitudinal rows. ( B ) Ventral view; arrow indicates
the ventral pigment blotch.
The geographic distributions of larval H. melanochir
and H. regularis were separate in most samples; only three
H. melanochir were found among H. regularis from Barker
Inlet, whereas no H. regularis were among H. melanochir
from the Bay of Shoals or Gulf St. Vincent. Larvae of other
Noel): Larval development of Hyporhamphus melanochir and H regularis
375
hemiramphid species may overlap in
distribution with those of H. melanochir
and H. regularis outside South Australian
waters. Meristic characters (summarized
in Table 5) can often distinguish H. mela-
nochir and H. regularis from the other
species, except the eastern sea garfish (//.
australis), which has overlapping meristic
counts and currently undescribed larvae.
The storm garfish iHemiramphus robus-
tus) has fewer anal-fin rays (11-14) and
develops both a dark blotch below the
dorsal fin and a pigmented pelvic fin in
the juvenile stage (Collette, 1974; Collette
et al., 1984). The long-finned garfish (Eu-
leptorhamphus viridis), an oceanic species
that rarely frequents nearshore waters, is
strikingly different from other hemiram-
phids, being much more elongate and
slender, and having divergent meristic
counts, including more dorsal- and anal-
fin rays (21-25 and 20-24, respectively),
more vertebrae (69-73), fewer pectoral-fin
rays (7-9), and fewer gill rakers (25-33)
(Collette, 1974; Hardy 1978; Chen, 1988;
Trnski et al., 2000).
Larvae of the saury {Scomberosox sau-
rus) (family Scomberosocidae) also occur
in southern Australia and are the only
other species that could be confused with
hemiramphids. These are distinguishable
from hemiramphid larvae by their higher
myomere count (62-70), greater number
of principal caudal-fin rays (16-17), pres-
ence of dorsal and anal finlets, and much
heavier pigmentation (Bruce and Sutton,
1998; Trnski etal, 2000).
Acknowledgments
I am grateful to D. Short, L. Triantafillos,
and the crew of the RV Ngerin for assist-
ing in the field collection of specimens.
The South Australian Museum allowed
access to catalogued hemiramphid speci-
mens, and A. Jordan donated a newly
hatched H. melanochir larva and assisted
with the examination of distinguishing
characters. B. Bruce and F. J. Neira pro-
vided tips on larval drawing techniques.
I also thank B. Bruce, S. Donnellan, A.
Fowler, A. Jordan, F. J. Neira, T. Trnski,
and T. Ward for kindly reviewing the
manuscript. This project was supported
by a Fisheries Research and Develop-
ment Corporation grant 97/133 and was
undertaken while receiving an Austra-
lian Postgraduate Award (Industry) at
Adelaide University.
c ^
Q) CO
■Q a!
■2 -a
CO
CO
CD
1
1
c^.
c^-
C^-
CM
in
3 ™
10
II
II
II
II
II
s
CO
oi
[^
O 0;
C<1
CO
CM
IM
cc CO
t/3
U
0
CO
"- k
(N
1
1
1
1
<" -s
CO
0
CO
.—I
II
CO "^
1
IM
(M
(N
ei
00
i-H
+
+
+
+
b
+
CO
m
10
CD
from
uded
.dal.
0:
CO
1
CO
1
CO
1
CO
1
1
t^
,_<
c--
0
OJ CJ 2
10
IM
CO
c^
CO
^.s s
■ — "
■ — ■
won
t- m ^^
T3 -3 "
econ
are
ilvic;
UJ
-.0
CO
00 ^
M >. a
t^
in
00
10 10
^ n 11
1
10
1
0
10
1
s ^
1 1
Otnoted.
n this sti
oral; Pj :
0
(M
CO
to 00
r-
10
10
10 10
10 0
(fl
II
II
5
II
II
0
II II
II II
0 0
IM
IM
eg (N
CM CM
^^11
€
1
CO
1
1
CD
1
00 _^
1 1
00 OS
1 1
00 00
t., i^ a,
^
C^
S -2 II
>
+
+
1 — 1
^ c^.
+ +
+ -t-
■H +
CO CD
t^
CO
1
CD 0
-H 0
00 m
cep
H.r
nal;
1> Tt
1 1
CO
1
CO Tl>
1 1
1 1
CO 00
1 1
X _ CO
03 Tf
lO
CO
t> 00
CD 00
CO CO
to 2;
2
CO
s s
2 2
2 2
s :!<
OS •~ ^-
•^ -c: CO
~ <-> aa
CD O I,
03
.2 c o
• b
(0
' — ^ "il II
■^ 2
2
%
h
« g a; a)
c 15
CO be
%
7
eq
1
0
c^-
,-H
,-H
rH
'^
i O b -D
0
Ta
ated fr
unts fo
availa
CO
=; P c
U5"
a CO
ral c
atio
4
(0
'0
'0
+
'O
ia. Dat
verteb:
inform
c^.
lO
^
00
^
+
+
+
+
+
00
c»
00
I>
X
tj
-t-
+
+
+
+
f-
t^
C^
10
r-
"m he o
^ C C
+
+
+
1
-I-
m '? II
em Au
inguis
wer. ?
0-'
&
cb
CD
CD
■5 CO --^
en
>,
O '-^ +
CO
in so
The
pper
^
05 cJl
CO
CO CO
CO
M
c
„
,-H ,-H
•— •
'-H
eu
1 1
00 t^
1
(M
1 1
^ 0
1
1
T3 - 3
•—1
^H r-l
•— '
^H
c 2: t«
iS »
^ --H C
c/j ~ Oj
Tf "tJ-
Tf
0
0
CTi
-a oi >
CN OI
CM
CM
^ S '^
<
1 1
1-H 0
1
1
1
0
1
a. -1= 0*
Cq CM
6 o 55
2 " "
■- -c ir
ll-S
10
10
f-
00
c~
Ji S CO
C<1
■5 -^ ■"
Q
1
7
T
1
1
1 11
CO
in
10
^
(M
CO QJ — •
■J-j
^ c 2
0 )-. T3
3
0 gen
'r>
«
X CT> 0 -D
*i CO CO
"q.
CO
3
O) C^ CJ _^
Meristic coun
if not in total
precaudal + c
C3
3
V.
3
S- 0:
5 2 ^ « If
■c g 1 1 ? .»
CO
OJ
"u
0)
If
1
Zj
3
C3 0
■5 i
° 3
0 ~
■£.a
II
2 -2
"t "3
0 Oc
Cfi
tS
a:
a;'
:i:'
a;'
376
Fishery Bulletin 101(2)
Literature cited
Bruce, B. D., and C. A. Sutton.
1998. Scomberosocidae: sauries. In Larvae of temperate
Australian fishes: laboratory guide for larval fish identi-
fication (F. J. Neira. A. G. Miskiewicz, and T. Trnski, eds. ),
p. 98-101. Univ. Western Australia Press, Perth, Western
Australia.
Chen, C. H.
1988. Hemiramphidae. In An atlas of the early stage fishes
in Japan (M. Okiyama, ed.), p. 265-275. Tokai Univ. Press,
Tokyo. |In Japanese.]
Collette, B. B.
1974. The garfishes (Hemiramphidae) of Australia and New
Zealand. Rec. Aust. Mus. 29:11-105.
Collette, B. B., G. E. McGowen, N. V. Parin, and S. Mito.
1984. Beloniformes: development and relationships. In
Ontogeny and systematics of fishes (H. G. Moser, W. J.
Richards, D. M. Cohen, M. P Fahay, A. W. Kendall Jr, and
S. L. Richardson, eds.), p. 335-354. Am. Soc. Ichthyol.
Herpetol., Spec. Publ. 1.
Gomon, M. F, C. J. M. Glover, and R. H. Kuiter
1994. The fishes of Australia's south coast, 992 p. State
Print, Adelaide, Australia.
Hardy Jr., J. D.
1978. Development of fishes of the mid-Atlantic Bight: an
atlas of egg, larval and juvenile stages. Vol. 11. Anguillidae
through Syngnathidae. U.S. Fish Wildl. Serv. FWS/OBS-
78/12, 458 p.
Jones, G. K., J. L. Baker, K. Edyvane, and G. J. Wright.
1996. Nearshore fish community of the Port River-Barker
Inlet Estuary, South Australia. I. Effect of thermal efflu-
ent on the fish community structure, and distribution and
growth of economically important fish species. Mar Fresh-
water Res. 47:785-800.
Jordan, A. R., D. M. Mills, G. Ewing, and J. M. Lyle.
1998. Assessment of inshore habitats around Tasmania for
life-history stages of commercial finfish species. Fishing
Research and Development Corporation, Final Report
Project 94/037, 176 p.
Kailola, P J., M. J. Williams, P C. Stewart, R. E. Reichelt,
A. McNee, and C. Grieve.
1993. Australian fisheries resources, 422 p. Bureau of
Resource Sciences and the Fisheries Research and Devel-
opment Corporation, Canberra, Australia.
Kendall, A. W., Jr., E. H. Ahlstrom, and H. G. Moser.
1984. Early life history stages of fishes and their characters.
In Ontogeny and systematics of fishes (H. G. Moser, W. J.
Richards, D. M. Cohen, M. P Fahay, A. W. Kendall Jr , and S.
L.Richardson, eds.), p. 11-22. Am. Soc. Ichthyol. Herpetol.,
Spec. Publ. 1.
Leis, J. M., and B. M. Carson-Ewart.
2000. The larvae of Indo-Pacific coastal fishes: an identifica-
tion guide to marine fish larvae. Fauna Malesiana Hand-
books, 2, 850 p. Brill, Leiden.
Ling, J. K.
1958. The sea garfish, Reporhamphus melanochir (Cuvier
and Valenciennes) (Hemiramphidae), in South Australia:
breeding, age determination, and growth rate. Aust. J.
Mar Freshwater Res. 9:60-110.
Neira, F. J., A. G. Miskiewicz, and T. Trnski.
1998. Larvae of temperate Australian fishes: laboratory
guide for larval fish identification, 474 p. Univ. Western
Australia Press, Perth, Western Australia.
Noell, C. J., S. Donnellan, R. Foster, and L. Haigh.
2001. Molecular discrimination of garfish Hyporhamphus
( Beloniformes ) larvae in southern Australian waters. Mar.
Biotechnol. 3:509-514.
Parin, N. V, B. B. Collette, and Y. N. Shcherbachev
1980. Preliminary review of the marine halfbeaks (Hemir-
amphidae, Beloniformes) of the tropical Indo- West-Pacific.
Trudy Inst. Okeanol. Akad. NAUK SSSR 97:7-173. [In
Russian.]
Potthoff, T
1984. Clearing and staining techniques. In Ontogeny and
systematics of fishes (H. G. Moser, W. J. Richards, D. M.
Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson,
eds.), p. 35-37. Am. Soc. Ichthyol. Herpetol., Spec. Publ. 1.
Prince Jeyaseelan, M. J.
1998. Manual offish eggs and larvae from Asian mangrove
waters, 193 p. UNESCO, Paris.
SPSS Inc.
1999. SigmaScan® Pro 5.0. Chicago, II.
Sudarsan, D.
1966. Eggs and larvae of a hemirhamphid fish from Manda-
pam. J. Mar. Biol. Assoc. India 8:342-346.
Talwar, P K.
1967. Studies on the biology of Hemiramphus marginatus
(Forsskal) (Hemirhamphidae-Pisces). J. Mar Biol. Assoc.
India 9:61-69.
Trnski, T, J. M. Leis, and B. M. Carson-Ewart.
2000. Hemiramphidae. In The larvae of Indo-Pacific coastal
fishes: an identification guide to marine fish larvae (J. M.
Leis, and B. M. Carson-Ewart. eds.), p. 154-158. Fauna
Malesiana Handbooks, 2. Brill, Leiden.
Watson, W.
1996. Hemiramphidae: halfbeaks. In The early stages of
fishes in the California Current region (H. G. Moser, ed.),
p. 634-641. Calif Coop. Oceanic Fish. Invest. Atlas 33.
377
Abstract-Teeth of 71 estuarine dol-
phins iSotalia guianensis) inciden-
tally caught on the coast of Parana
State, southern Brazil, were used to
estimate age. The oldest male and
female dolphins were 29 and 30 years,
respectively. The mean distance from
the neonatal line to the end of the first
growth layer group (GLG) was 622.4
±19.1 pm (n=48). One or two accessory
layers were observed between the neo-
natal line and the end of the first GLG.
One of the accessory layers, which was
not always present, was located at a
mean of 248.9 ±32.6 pm (^=25) from
the neonatal line, and its interpreta-
tion remains uncertain. The other layer,
located at a mean of 419.6 ±44.6 pm
(n=54) from the neonatal line, was
always present and was first observed
between 6.7 and 10.3 months of age.
This accessory layer could be a record
of weaning in this dolphin. Although
no differences in age estimates were
observed between teeth sectioned in
the anterior-posterior and buccal-lin-
gual planes, we recommend sectioning
the teeth in the buccal-Ungual plane
in order to obtain on-center sections
more easily. We also recommend not
using teeth from the most anterior
part of the mandibles for age estima-
tion. The number of GLGs counted
in those teeth was 50% less than the
number of GLGs counted in the teeth
from the median part of the mandible
of the same animal. Although no sig-
nificant difference (P>0.05) was found
between the total lengths of adult male
and female estuarine dolphins, we
observed that males exhibited a second
growth spurt around five years of age.
This growth spurt would require that
separate growth curves be calculated
for the sexes. The asymptotic length
(TL^), k, and ^,j obtained by the von Ber-
talanffy growth model were 177.3 cm,
0.66, and -1.23, respectively, for fe-
males and 159.6 cm, 2.02, and -0.38,
respectively, for males up to five years,
and 186.4 cm, 0.53 and -1.40, respec-
tively, for males older than five years.
The total weight (TW)/total length (TL)
equations obtained for male and female
estuarine dolphins were TW = 3.156 x
10-6 X TL 3 2836 (r=0.96), and rW = 8.974 x
10-5 X XL 2 '5182 (r=0.95), respectively.
Manuscript accepted 28 October 2002.
Manuscript received 31 December 2002
at NMFS Scientific Publicaitions Office.
Fish. Bull. 101:377-383 (2003).
Age and growth of the estuarine dolphin
(Sotalia guianensis) (Cetacea^ Delphinidae)
on the Parana coast, southern Brazil
Fernando Cesar Weber Rosas
Institute Nacional de Pesquisas da Amazonia (INPA)
Laboratono de Mamileros Aquaticos
Caixa Postal 478
Manaus, AM, 69011 970, Brazil
E-mail address; frosas@inpa gov br
Andre Silva Barreto
Universidade do Vale do Itaiai (UNIVALI)
CTTMar, Caixa Postal 360
Itaiai, SC, 88302 202, Brazil
Emygdio Leite de Araujo Monteiro-Filho
Universidade Federal do Parana (UFPR)
Departamento de Zoologia
Caixa Postal 19020
Curitiba, PR, 81531-970, Brazil
and
Institute de Pesquisas Cananeia (IPeO
Rua Joao Salim, Lote 26-Quadra Y
Parque Xangrila
13098-106, Campinas, Sao Paulo, Brazil
Until recently the genus Sotalia was
monospecific (S. fluviatilis) and had a
marine and a riverine ecotype (da Silva
and Best, 1996). Using tridimensional
morphometric analyses, Monteiro-Filho
et al. (2002) were able to separate it into
two distinct species: Sotalia fluviatilis,
which lives in freshwater, and Sotalia
guianensis. which lives in the marine
environment. Because tucuxi is the
vernacular name used for the freshwa-
ter species, Rosas and Monteiro-Filho
(2002) suggested "estuarine dolphin"
as the vernacular name for S. guianen-
sis, as previously mentioned by Watson
(1988).
Age is important in characterizing
population dynamics of mammals.
Growth layer groups (GLGs) observed
in teeth of mammals have been used
to estimate ages, and the greatest
progress in this area has occurred with
studies carried out on marine mammals
(Klevezal, 1980; Hohn et al., 1989). The
method consists of counting GLGs found
in the dentine and cement of the ani-
mals' teeth, which are deposited every
year in most species (Klevezal, 1996).
Calibrating age estimates and identify-
ing accessory layers (not annual) are
essential for reliable age determination
(Hohn, 1990). Some population param-
eters are extremely sensitive to errors
and age estimate deviations, and the
absence of or an inadequate calibration,
could lead to incorrect interpretations
(Hohnetal., 1989).
Because there is no sexual dimorphism
in the body proportions of adult Sotalia
guianensis, all previous growth studies
analyzed both sexes together (Borobia,
1989; Schmiegelow, 1990; Ramos et al.,
2000). However, there is evidence of dif-
ferentiated growth between male and
female estuarine dolphins around pu-
berty (Rosas and Monteiro-Filho, 2002),
thereby making it necessary to analyze
growth separately for the sexes.
The objectives of this paper were 1)
to estimate the ages of S. guianensis
378
Fishery Bulletin 101(2)
caught incidentally or stranded on the Parana coast, Brazil;
2) to give some guidelines to promote reliable age estimates
in this species; 3) to describe the growth in body length (cm)
according to the ages (years) of male and female estuarine
dolphins, by using classical mathematical growth models;
and 4) to describe the body-weight-body-length relation-
ship for both sexes of this dolphin.
Materials and methods
Teeth from 71 individuals of S. guianensis (34 males, 28
females and 9 of undetermined sex), incidentally caught
or found stranded on the Parana coast, southern Brazil
(25°18'S; 48°05'W-25°5'S8; 48°35'W), from January 1997
to July 1999, were used to estimate age. The total body
weight (kg) and standard measurements of individuals
were made in accordance to Norris (1961). Total length
(cm) was measured in a straight line from the tip of the
beak to the central notch of the tail, in an axial projec-
tion. The skulls and teeth were collected, prepared, and
deposited in the collection of the Instituto de Pesquisas
Cananeia (IPeC).
Preparation of the teeth, from the decalcification to the
mounting of the slides, was carried out in the Laboratory
of Marine Mammals and Marine Turtles of the Depart-
ment of Oceanography of the Funda^ao Universidade do
Rio Grande (FURG). The method of Hohn et al. (1989) was
used, with the following adaptations: 1) decalcification time
varied from one hour for newborn or young individuals,
up to a m2iximum of 12.5 hours for old adults, 2) Harris's
hematoxilin was used for staining, according to Molina and
Oporto (1993), and immersion times of the sections varied
from three to six minutes.
Because the absence of a pre-established age estimation
model for S. guianensis, we tested both anterior-posterior
and buccal-lingual planes for cutting teeth. Age estimation
was performed by counting GLGs in the dentine. GLGs
were defined as being the sequence of a thin nonstained
layer, a thick stained layer, and a very thin layer that is
strongly stained (very dark). Each complete GLG was as-
sumed to represent one year (Ramos et al., 2000).
Teeth were selected from the middle of the tooth rows.
However, to check for differences in age estimation among
those positioned along the tooth row, we compared the
number of GLGs in teeth from the middle of the tooth row
with the number of GLGs in those from the most anterior
part of the tooth row of the same animal.
The senior author read teeth slides at least three times
during a minimum period of three weeks. Estimated age
was taken as the last reading, assuming that reading accu-
racy improves with practice (Pinedo and Hohn, 2000). Age
was estimated without access to biometric and biological
data, thereby avoiding reader bias.
By using only central sections or those close by, in which
at least 80'7r of the pulp cavity was exposed (Fig. 1), we
obtained the following measurements with an ocular mi-
crometer in a com[)ound microscope: 1) distance (in pm)
from the neonatal line up to the end of the first GLG in
the dentine; 2) distance from the neonatal line to the first
Figure 1
Central section of a Sotalia guianensis tooth cut
in the buccal-lingual plane showing the position of
the neonatal line (NL), the always-present acces-
sory layer (AL), first GLG (1) and second GLG (2).
Magnification: .30x.
accessory layer in the dentine; and 3) distance from the
neonatal line to the second accessory layer in the dentine,
if present. All measurements were made perpendicular to
the external margin and at the neck of the tooth (an area
located between the crown and root of the tooth).
Ages of individuals less than one year were estimated
in months, by using as a base the percentage proportion of
the mean distance between the neonatal line and the end
of the GLG of the first year (Ramos, 1997).
Several models have been created over the years to
describe growth, including the von Bertalanffy, Gompertz,
logistical, and Richards models. Schnute's generic growth
model helps to choose the model which is best adapted to
the length and age data of the species studied. Schnute's
model (1981) is defined as:
y'(/) =
v:'+(v^-vf)-
where Y(t) represents a measurement (length, weight,
volume) at age t; variables r, and T., are ages of young and
old specimens, respectively, andyj andy.^ are sizes at these
ages. These sizes, together with a and 6, are the parameters
Rosas et al.: Age and growth of Sotalia guianensis
379
3 4
o
E 2
I Females (n=28)
I Males (n=34)
11
li
m Mill III
111 I II T"
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Age (years)
28 30
Figure 2
Age distribution of female and male Sotalia guianensis on the Parana coast, southern
Brazil, recorded between 1997 and 1999.
to be estimated. To define the growth model that would best
fit the length and age data of S. guianensis, the Schnute
model was applied to the length-at-age data.
Growth equations were calculated separately for the
sexes, with 34 males and 28 females. Growth model ad-
justment to the data was made by using the nonlinear
iterative Quasi-Newton method, minimizing the residual
sum of squares.
Total-weight to total-length relationships were estab-
lished by using 42 individuals of S. guianensis (23 males
and 19 females) with the equation
7W = 0 X rL»
(Santos, 1978),
where TW = total weight in kg;
TL = total length in cm;
0.05). However,
among the animals with an age equal to or greater than
24 years (7!=7), 85.7% were females. The oldest male was 29
years old and the oldest female was 30 years old (Fig. 2).
Age estimates for S. guianensis individuals varied from
0 to 30 years. Although the age mode was in the 0 and
1 year classes (Fig. 2), 53.5% of all animals whose ages
were estimated were seven years or more. This proportion
remained relatively constant between the sexes; 55% of
the males and 50% of the females were equal to or greater
than seven years.
Growth
When applied to the present data, Schnute's model indi-
cated that the von Bertalanffy growth equation fitted the
length and age data of S. guianensis better (Table 1, a>0
and b>l; Schnute, 1981). Even though the predictive power
of Schnute's model is greater than von Bertalanffy 's (see
"Explained variance" in Tables 1 and 2), the latter is justi-
fied by its historical use, and therefore has a greater value
for populational comparison, and incorporates a better
understanding of the biological meaning of its variables.
Comparing the total lengths (TL) of the estuarine dol-
phins of six years or more, we found no significant differ-
380
Fishery Bulletin 101(2)
ence between the sexes (^test, P>0.05). However, it was ob-
served that males possibly exhibit a discontinuity in growth
around five years. The existence of a secondary growth
spurt around this age was considered to be due to the onset
of puberty in this species (Rosas and Monteiro-Filho, 2002),
and necessitated the calculation of separate growth curves
for each sex. It should be noted that sexual maturity of
the dolphins here analyzed was determined by Rosas and
Monteiro-Filho (2002) to occur at seven years in males. In
order to estimate the fit of a two-step model, the sample was
divided into two groups: 1) up to five years (prepuberty) and
2) older than five years (subadult and adults).
The growth parameters obtained for males and females
are given in Table 2. The results obtained by Borobia ( 1989)
and Schmiegelow ( 1990) using the von Bertalanffy growth
model are also indicated in Table 2 for comparison. The
growth parameters obtained by the analyses of males up
to five years of age and those older than five are presented
Table 1
Schnute growth
model parameters applied
to Sotalia
guianensis on the
Parana coast.
southern Brazil, "r," and
"r2" are predetermined ages in
years; "vj" and "y^" ^""^
estimated sizes at
ages r, and T.,,
in cm; "a" and
"b" are adi-
mensional parameters. "SQ" represents the residual sum
of squares, and "Expl. var" represents the variance of the
data explained by the model.
Parameters
Females
Males
All
I^i
0
0
0
y,
93.11
86.04
89.53
h
28
28
28
^2
181.72
190.45
185.70
a
0.14
0.07
0.13
h
7.64
9.45
7.86
SQ
1242.35
1898.75
4375.56
Expl. var. (%)
93.47
92.17
92.47
n
28
34
71
in Table 3. By dividing the sample in two, the fit of the von
Bertalanffy model improved considerably (Tables 2 and
3). The growth curves of S. guianensis males and females
obtained by the von Bertalanffy model are presented in
Figure 3.
The t-test applied to parameters a and b of the weight/
length regression equations for males and females revealed
a significant difference (t=2,25; df=38; P<0.05). Therefore,
this relationship was analyzed for the sexes separately and
the equations obtained were
7W = 3. 156 X 10-6 X XL ^ 2836
TW = 8.974 X 10-5 X tL ^ "82
(males) (r=0.96)
(females) (r=0.95).
Discussion
Age estimation
Although there was no difference in the age estimation
between teeth orientated in the buccal-lingual and ante-
rior-posterior planes, we recommend the buccal-lingual
plane to obtain easier on-center or close-to-center sections,
which are essential for accurate age estimates.
The differences found in counting GLGs in teeth from
the anterior extremity and the median region of the tooth
row of the same animal corroborate the results obtained by
Hui ( 1980) for Tkirsiops truncatus. Therefore, we also do not
recommend using teeth from the most anterior part of the
mandible for age estimation in S. guianensis.
The mean distance between the neonatal line and the end
of the first GLG obtained in the present study (622.4 pm)
was approximately double that obtained by Ramos (1997)
(297.8 pm) for estuarine dolphins on the coast of Rio de Ja-
neiro. The differences, however, must be analyzed carefully:
the measurements carried out in our study were always
made in the neck of the teeth, whereas those made by Ra-
mos ( 1997 ) were from the base of the neonatal line. However,
the differences may be related to the interpretation of the
position of the first annual layer The accessory layers (no-
nannual), observed between the neonatal line and the end
Table 2
Von Bertalanff}
growth model parameters
applied to Sotalia
guianensis on the Parana coast, southern
Brazil, and parameters from
the literature. "
TL,
"= asymptotic length (
cm), "k"= growth constant and
'<„"= theoretical age at which the length of the animal is
zero. "SQ" represen
ts the residual sum of
squares, and '
Expl
var" represents the
variance of the data
explained by the model.
Parameters
Our study
Borobia (1989)
Schmiegelow (1990)
Females
Males
All
TL^
177.31
179.10
179.53
187.21
182.6
k
0.66
1.00
0.79
0.20
0.41
'o
-1.23
0.72
0.95
-4.05
-1.57
SQ
1944.25
3732.30
6942.93
—
—
Expl. var (%)
89.78
84.61
88.06
—
—
n
28
34
71
24
22
Rosas et al.: Age and growth of Sotalia gutanensis
381
of the GLG of the first year, frequently
appear in a very conspicuous manner,
especially in the tip of the tooth, and can
be easily confused with annual layers.
The assumption that accessory layers
are annual could result in a duplication
of the real age of young animals up to two
years old, with significant consequences
in the interpretations of populational bio-
logical parameters ( Hohn, 1990 ). The ideal
situation would be that a GLG deposition
model already existed for the species being
studied, thereby avoiding counting acces-
sory layers as being annual (Hohn et al.,
1989). In most odontocete species, includ-
ing S. guianensis, accessory layers do not
continue up to the end of the root of the
tooth, in contrast to annual layers, which
can be seen from the tip to the base of the
root of the tooth. However, to identify ac-
cessory layers it is necessary that the sec-
tions selected for age determination are
central, or close to the center of the pulp
cavity (Pinedo and Hohn, 2000). Off-center
sections can be used for age estimation,
but reading errors increase markedly and
consequently induce unreliable age esti-
mates (Pinedo and Hohn, 2000).
The reasons for GLG deposition in teeth
are unknown (Hohn et al., 1989). Howev-
er, several reasons have been suggested,
including seasonal variations in growth
rate, genetic physiological cycles, dietary
changes, hormonal influences, and intrin-
sic factors on the metabolism in general
(Boyde, 1980; Klevezal, 1980; Scheflfer and
Myrick, 1980). Although all these factors
could be influential, variations in the diet
certainly play a significant role. According
to Klevezal ( 1996), a descriptive record of
the dietary changes of an animal during the year should
initially be looked for in structures that have a large degree
of sensitivity, such as teeth. It is known that dentine reacts
to the introduction of fluoride, calciferol and a series of other
components in the organism, forming layers with different
degrees of mineralization (Klevezal, 1996), which is known
as a calcium-traumatic reaction of dentine. Therefore, it is
possible to find a record of dietary changes in the dentine,
starting from weaning (Klevezal, 1996).
We believe that the accessory layer in the dentine found
at approximately 419.6 pm from the neonatal line, could be
a record of the end of weaning in the estuarine dolphin. It
was observed in all the teeth of individuals older than 6.7
months and could be a hypomineralized layer caused by a
reduction of calcium in the body due to the absence of milk
in the diet (Klevezal, 1996). The other accessory layer found
closer to the neonatal line (mean of 248.9 pm) was not
observed in all animals and the interpretation of this layer
remains uncertain. It may be related to the beginning of
weaning, as has been suggested for the bottlenose dolphin
200-
180-
A . . fi
^^-^"A A A A
160-
^^-b-tr-n
140 -
i
120-
f
100-
o oU T-
£ 0
CJ)
c
"en
° 200
180
160-
5 10 15 20 25 30
B
\ ° 0 oo oo
^ o o o o
°p
140 -
/
120-
/
100-
:
0 5 10 15 20 25 30
Age (years)
Figure 3
Growth curve of male (A) and female (B) Sotalia guianensis on the Parana
coast, southern Brazil.
Table 3
Von BertalanfTy growth model parameters for male
Sotalia guianensis on the Parana coast, southern Brazil.
"TL,"= asymptotic length (cm), "*"= growth constant and
'7,j"= theoretical age at which the length of the animal is
zero. "SQ" represents the residual sum of squares, and
"Expl. var" represents the variance of the data explained
by the model.
Parameters
Up to 5 years More than 5 years
k
SQ
Expl.var (%)
n
159.64
2.02
-0.38
510.99
94.20
15
186.41
0.53
-1.40
1013.98
50.90
19
382
Fishery Bulletin 101(2)
(T. truncatus) (Hohn*). This hypothesis still needs to be
confirmed. However, all the S. guianensis individuals that
were still nursing, but which already had remains of solid
food in their stomachs («=5), had only an accessory layer
that is closer to the neonatal line — they did not have the
layer that we are assuming marks the end of weaning.
According to Rosas (20001, there was no significant
difference in incidental catches between mature and im-
mature individuals of S. guianensis caught on the coast of
Parana, suggesting a similar vulnerability of young and
adult estuarine dolphins to fisheries. Because the animals
analyzed in our study were the same ones used by Rosas
(2000), this lack of significant difference between mature
and immature individuals can suggest a representative age
distribution of the individuals analyzed.
Because the maximum estimated age in our study was
30 years, and because the dolphins here analyzed were
incidentally caught in fishing nets, it seems reasonable to
assume that the longevity of the estuarine dolphin may
be 30-35 years. This hypothesis is also corroborated by
the study carried out by Ramos (1997) with S. guianensis
on the coast of Rio de Janeiro State (southeastern Brazil).
Although the age of the oldest male observed in our study
was 29 years, the frequency of males older than 21 years
was less than 3%, which is extremely low when compared
with the frequency of 21.5% for females older than 21 years.
These results suggest a greater life expectancy for females,
which is also corroborated by a study carried out by Ramos
(1997) in Rio de Janeiro.
Growth
The use of Schnute's model is helpful in deciding which
growth model should be used. Even though the researcher
can usually decide which model is most appropriate by
looking at the data, subtle differences in data distribution
could cause one or another model to be more adequate. Use
of a generic model allows this choice without intervention
of the researcher and avoids any unconscious bias towards
or against any model.
The discontinuity of growth in male S. guianensis in our
study could have been due to the small sample size or may
have been due to a second growth spurt, which has already
been observed in the total length oiStenella attenuata (Per-
rin et al., 1976), Lissodelphis borealis (Ferrero and Walker,
1993), and Phocoenoides dalli (Ferrero and Walker, 1999),
and in the weight of male Tursiops truncatus (Cockroft and
Ross, 1990). The k value obtained for male S.guianensis
up to five years was very high, meaning that asymptotic
length in this phase of life was reached quickly. The ces-
sation of growth exhibited by the model for males up to
5 years probably is not true in the biological sense but
could be an artifact created by the model and the small
sample size. Most probably there is a marked reduction in
growth with the start of sexual maturation and a greater
investment in the weight or reproductive apparatus (or
' Hohn, A. A. 1999. Personal comimiii licMiiDirt l.ahoralory,
Sc)ulli(>ast Fishiories Scienct' ('enter. National Marine Fi.slieries
Ser\'ice, 101 I'river.s Island Road. Bcaulort, NC 28516-9722.
both). The hypothesis of a greater investment in weight is
supported by the observed difference in the weight-length
coefficient between males and females. Additionally, sexual
investment of male estuarine dolphins is very high — testes
of adult males can reach up to 32 cm in length and weigh
up to 3.3% of the total body weight (Rosas and Monteiro-
Filho, 2002).
After the secondary growth spurt in males, the final
asymptotic length did not differ very much from that in
females. Previous growth studies carried out by Borobia
(1989), Schmiegelow (1990), and Ramos et al. (2000) with
the estuarine dolphin did not mention the existence of a
second growth spurt in males, possibly because the authors
did not analyze the growth of males and females separately.
According to Ramos et al. (2000), male and female data
were combined because of the absence of sexual dimor-
phism in the body size of adults of this species.
Borobia (1989) and Schmiegelow (1990), who also used
the von Bertalanfiy model, obtained different values for the
growth equation parameters (Table 1). The sample used by
Borobia ( 1989 ) did not have many individuals in ages 1 and 2,
and none in the 0 age class. The absence of animals that "an-
chor" the beginning of the curve could result in low estimates
of ^ and ?Q. Additionally, Borobia ( 1989) examined individuals
from different locations along the distribution of the species
and thus did not take into consideration possible geographi-
cal variations. The results obtained by Schmiegelow (1990)
are similar to those of our study, probably because both of
them used animals from the same region.
Ramos et al. (2000) analyzed the growth of S. guian-
ensis using the Gompertz growth model and obtained an
asymptotic length (191.7 cm) which was much greater
than that obtained in our study and in previous studies
(Borobia, 1989; Schmiegelow, 1990) (Table 1). This differ-
ence could be due to 1) the small number of individuals
older than 12 years (/!=3) in their sample; or 2) a difference
in asymptotic lengths between southeastern and southern
Brazil populations. Similar differences have been observed
between asymptotic lengths of Pontoporia hlainvillei from
Rio de Janeiro (southeastern Brazil) and Sao Paulo and
Parana (same area of the present study), where larger in-
dividuals were found in Rio de Janeiro (Ramos et al., 2000;
Rosas, 2000). Therefore, it is possible that environmental
variables could be responsible for larger sizes in the area
studied by Ramos et al. (2000), both for S. guianensis and
for f! hlainvillei.
Although no significant difference was observed in the
asymptotic length between adult males and females, the
differentiated growth in time between the two sexes is
probably responsible for the difference observed in the
weight-length relationship.
In most species, the length exponent (0) of the weight-
length relationship is usually close to 3 (Santos, 1978). The
estimated values of this exponent for the estuarine dolphin
(3.2 for males and 2.6 for females) suggest that the longi-
tudinal and transversal body growth in this species follows
a similar pattern.
Our results suggest that it is important to study growth
by analyzing the sexes separately, because there may be dif-
ferential growth between the sexes before the adult age.
Rosas et aL: Age and growth of Sotalia guianensis
383
Acknowledgments
We sincerely thank the fishermen of Vila da Barra do
Superagtii and Ilha das Pe^as (Parana coast) for the infor-
mation they provided and their help in collecting the inci-
dentally caught dolphins. We thank Fundagao O Boticario
de Prote^ao a Natureza and the MacArthur Foundation for
financial support, and IBAMA/PR, especially Guadalupe
Vivekananda, head of the National Park of Superagtii. We
also thank Maria Cristina Pinedo, who allowed us to use her
laboratory and equipment for the age estimations, and Kesa
K. Lehti, who translated the manuscript from Portuguese
into English. Renata Ramos and an anonymous referee
provided critical and insightful comments on the manu-
script. Coordenagao de Pessoal de Nivel Superior (CAPES)
provided a fellowship to the senior author This study is part
of a dissertation presented by Fernando C. Weber Rosas, sub-
mitted in partial fulfillment for a Ph.D. degree in Zoology at
the Universidade Federal do Parana, Curitiba, Brazil.
Literature cited
Borobia, M.
1989. Distribution and morphometries of South American
dolphins of the genus Sofa/(a. MSc. thesis, 81 p. McDon-
ald College, McGill University. Montreal, Quebec, Canada.
Boyde, A.
1980. Histological studies of dental tissues of Odontocetes.
In Age determination of toothed whales and sirenians (W.
F. Perrin and A. C. Myrick, eds.), p. 65-87 Rep. Int. Whal.
Comm., Special Issue 3.
Cockcroft, V. G., and G. J. Ross.
1990. Age, growth, and reproduction of bottlenose dolphins
Tursiops truncatus from the east coast of Southern Africa.
Fish. Bull. 88:289-302.
da Silva, V. M. F, and R. C. Best.
1996. Sotalia fluviatilis. Mammalian Species 527:1-7.
Ferrero, R. C, and W. A. Walker
1993. Growth and reproduction of the northern right whale
dolphin, Lissodelphis borealis, in the offshore water of the
North Pacific Ocean. Can. J. Zool. 71:2335-2344.
1999. Age, growth and reproductive patterns of Dall's por-
poise (Phocoenoides dalli) in the Central North Pacific Ocean.
Mar Mamm. Sci. 15(2):273-313.
Hohn, A. A.
1990. Reading between the lines: Analysis of age estimation
in dolphins. In The Bottlenose dolphin (S. Leatherwood
and R. R. Reeves, eds.), p. 575-585. Academic Press, New
York, NY.
Hohn, A. A.; M. D. Scott, R. S. Wells, J. C. Sweeney, and
A. B. Irvine.
1989. Growth layers in teeth from known-age, free-ranging
bottlenose dolphins. Mar Mamm. Sci. 5 (41:315-342.
Hui,C.A.
1980. Variability of dentine deposition in Tursiops truncatus.
Can. J. Fish. Aquat. Sci. 37:712-716.
Klevezal, G. A.
1980. Layers in the hard tissues of mammals as a record of
growth rhythms of individuals. In Age determination of
toothed whales and sirenians ( W F. Perrin and A. C. Myrick,
eds.), p. 89-94. Rep. Int. Whal. Comm. Special Issue 3.
1996. Recording structures of mammals. Determination of
age and reconstruction of life history, 274 p. A.A. Balkema,
Rotterdam, Netherlands.
Molina, D. M., and J. A. Oporto.
1993. Comparative study of dentine staining techniques
to estimate age in the Chilean Dolphin, Cephalorhynchus
eutropia (Gray 1846). Aquat. Mamm. 19 ( 1 ):45-48.
Monteiro-Filho, E. L. A., L. R. Monteiro, and S. F dos Reis.
2002. Skull shape and size divergence in the dolphins of the
genus Sotalia : a tridimensional morphometric analysis. J.
Mamm. 83(1):125-134.
Norris, K. S.
1961. Standardized methods for measuring and recording
data on the smaller cetaceans. J. Mamm. 42 (4):471-476.
Perrin, W. F, J. M. Coe, and J. R. Zweifel.
1976. Growth and reproduction of the spotted porpoise,
Stenella attenuata, in the offshore eastern tropical Pacific.
Fish. Bull. 74(2 ):229-269.
Pinedo, M. C, and A. A. Hohn.
2000. Growth layer patterns in teeth from the franciscana,
Pontoporia blainvillei: developing a model for precision in
age estimation. Mar Mamm. Sci. 16 (l):l-27.
Ramos, R. M.A.
1997. Determina?ao de idade e biologia reprodutiva de Pon-
toporia blainvillei e da forma marinha de Sotalia fluviatilis
(Cetacea:Pontoporiidae e Delphinidae) no norte do Rio de
Janeiro. M.Sc. thesis, 95 p. Universidade Estadual
do Norte Fluminense. Campos dos Goytacazes, Rio de
Janeiro, Brasil.
Ramos, R. M. A., A. P M. Di Beneditto, and N. R. W. Lima.
2000. Growth parameters of Pontoporia blainvillei and Sota-
lia fluviatilis (Cetacea) in northern Rio de Janeiro, Brazil.
Aquat. Mamm. 26 (l):65-75.
Rosas, F C. W.
2000. Interagoes com a pesca, mortalidade, idade, reprodu9ao
e crescimento de Sotalia guianensis e Pontoporia blainvil-
lei (Cetacea, Delphinidae e Pontoporiidae) no literal sul do
Estado de Sao Paulo e literal do Estado do Parana, Brasil.
Ph.D. diss., 145 p. Universidade Federal do Parana. Curi-
tiba, PR, Brasil.
Rosas, F C. W, and E. L. A. Monteiro-Filho.
2002. Reproduction of the estuarine dolphin (Sotalia guia-
nensis) on the coast of Parana, southern Brazil. J. Mamm.
83(2):507-515.
Santos, E. P.
1978. Dinamica de popula?6es aplicada a pesca e piscicul-
tura, 129 p. Editora de Humanismo, Ciencia e Tecnologia
"HUCITEC" Ltda, Sao Paulo.
Scheffer, V. B., and A. C. Myrick.
1980. A review of studies to 1970 of growth layers in the
teeth of marine mammals. In Age determination of toothed
whales and sirenians (W. F Perrin and A. C. Myrick, eds.), p.
51-63. Rep. Int. Whal. Comm., Special Issue 3.
Schmiegelow, J. M. M.
1990. Estudo sobre cetaceos odontocetes encontrados em
praias da regiao entre Iguape (SP) e Baia de Paranagua (PR)
(24°42'S-25°28'S) com especial referenda a Sotalia fluviatilis
(Gervais, 1853) (Delphinidae). M.Sc. thesis, 149 p. Univer-
sidade de Sao Paulo, Institute Oceanografico, Sao Paulo.
Schnute, J.
1981. A versatile growth model with statistically stable
parameters. Can. J Fish. Aquat. Sci. 38:1128-1140.
Watson, L.
1988. Whales of the world. A handbook and field guide to all
the living species of whales, dolphins and porpoises, 302 p.
Hutchinson, London.
384
Abstract— Offshore winter-spawned
fishes dominate the nekton of south-
eastern United States estuaries. Their
juveniles reside for several months in
shallow, soft bottom estuarine creeks
and bays called primary nursery areas.
Despite similarity in many nursery char-
acteristics, there is, between and within
species, variability in the occupation of
these habitats. Whether all occupied
habitats are equally valuable to indi-
viduals of the same species or whether
most recruiting juveniles end up in the
best habitats is not known. If nursery
quality varies, then factors controlling
variation in pre-settlement fish distribu-
tion are important to year-class success.
If nursery areas have similar values,
interannual variation in distribution
across nursery creeks should have less
effect on population sizes or production.
I used early nursery period age-specific
growth and mortality rates of spot {Leios-
tomus xanthurus) and Atlantic croaker
iMicropogonias undulatus) — two domi-
nant estuarine fishes — to assess relative
habitat quality across a wide variety of
nursery conditions, assuming that fish
growth and mortality rates were direct
reflections of overall physical and biologi-
cal conditions in the nurseries. I tested
the hypothesis that habitat quality varies
for these fishes by comparing growth and
mortality rates and distribution patterns
across a wide range of typical nursery
habitats at extreme ends of two systems.
Juvenile spot and Atlantic croaker were
collected from 10 creeks in the Cape
Fear River estuary and from 18 creeks
in the Pamlico Sound system. North
Carolina, during the 1987 recruitment
season (mid-March-mid-June). Sampled
creeks were similar in size, depth, and
substrates but varied in salinities, tidal
regimes, and distances from inlets. Spot
was widely distributed among all the
estuarine creeks, but was least abundant
in the creeks in middle reaches of both
systems. Atlantic croaker occurred in the
greatest abundance in oligohaline creeks
of both systems. Instantaneous growth
rates derived from daily otolith ages were
generally similar for all creeks and for
both species, except that spot exhibited
a short-term growth depression in the
upriver Pamlico system creeks — perhaps
the result of the long migration distance
of this species to this area. Spot and
Atlantic croaker from upriver oligohaline
creeks exhibited lower mortality rates
than fish from downstream polyhaline
creeks. These results indicated that even
though growth was similar at the ends
of the estuaries, the upstream habitats
provided conditions that may optimize
fitness through improved survival.
Manuscript accepted 25 October 2002.
Manusript received 31 December 2002
at NMFS Scientific Publications Office.
Fish. Bull. 101:384-404 (2003).
The relative value of different estuarine
nursery areas in North Carolina for
transient juvenile marine fishes
Steve W. Ross
NC National Estuarine Research Reserve
5600 Marvin Moss Ln.
Wilmington, North Carolina 28409
E-mail address: rosss(guncwil.edij
Offshore winter-spawned (OWS) fishes
are a major component of the nekton of
southeastern United States and Gulf of
Mexico estuaries. Their larvae migrate
across the shelf, enter estuaries, and the
majority of juveniles reside for several
months in shallow, soft bottom estuarine
creeks and bays called primary nursery
areas (PNAs). Very high concentrations
of fishes in these PNAs suggest that
they are valuable habitats, perhaps
because they are good sources of food
and shelter (Boesch and Turner, 1984;
Mclvor and Odum, 1988; Miltner et al.,
1995). Despite similarity in some PNA
physical characteristics, there is vari-
ability in habitats occupied (between
and within species), especially with
regard to salinity, tidal influence, acces-
sibility (i.e. distance from inlets), and,
perhaps, food and predator regimes
(Weinstein, 1979; Ross and Epperly,
1985). Assessing the relative value of
all PNA habitats to individuals of the
same species is increasingly important
(Weinstein, 1982; Sogard, 1992; Guin-
don and Miller, 1995; Beck et al., 2001).
If PNA value varies, do most of the
recruiting juveniles end up in the best
habitats (Thresher, 1985)? Understand-
ing variation in habitat quality during
a major early life history phase should
yield insight into causes of variability in
year-class strength, particularly if juve-
nile fish distributions vary interannu-
ally. If PNA quality varies, then factors
controlling variation in presettlement
distribution are important to year class
success because animals could be trans-
ported to habitats of unpredictable
quality. If nursery areas have similar
value, interannual variations in dis-
tribution across nursery creeks should
have less effect on ultimate population
sizes or production.
General estuarine distributions of
two dominant OWS fishes, spot (Leio-
stomus xanthurus) and A.i\a.ntic croaker
(Micropogonias undulatus), exhibit con-
sistent patterns throughout their rang-
es. Juvenile Atlantic croaker routinely
concentrate in oligohaline creeks or
bays (Weinstein, 1979; Mercer, 1987a)
— a pattern that suggests that the
upstream regions are most valuable
to this species. Spot, however, are more
ubiquitously and variably distributed
through the shallow PNAs (Ross and
Epperly, 1985; Mercer, 1987b), perhaps
indicating less dependence on a particu-
lar estuarine region. Despite these gen-
eralities, both species can be present in
large numbers in almost any estuarine
creek or bay over the full salinity range
(e.g. Nelson et al., 1991). In general, ju-
veniles of both species seem to avoid (or
are unsuccessful in) more open water
areas of estuaries during the early part
of the nursery period.
The main purpose of this paper is to
assess relative habitat value for two
dominant members of the OWS fish
group, spot and Atlantic croaker, across
a wide variety of North Carolina PNA
conditions. I assumed that fish growth
and mortality rates were direct reflec-
tions (integrators) of overall physical
and biological conditions in PNA habi-
tats. Therefore, 1 used early nursery
period age-specific growth and mortal-
ity rates of spot and Atlantic croaker, in
addition to distribution data, to assess
relative habitat quality, testing the hy-
pothesis that habitat quality varies for
these fishes across a broad range of typ-
ical PNAs in two very different estua-
rine systems. Growth and mortality can
be influenced by fish density; however,
Ross (1992) found similar growth and
mortality rates for spot and Atlantic
Ross: Relative value of different estuarine nursery areas for juvenile marine fishes in North Carolina
385
76' '30'
Upper
TS'SO"
35-30'
PAMLICO f
SOUND y
35°00:
78°0(t
77"5a'
Figure 1
Pamlico Sound (A) and Cape Fear Estuary (B) in coastal North Carolina. Gen-
eral areas (e.g. upper, middle, lower) and sampling locations during March-June
1987 (solid dots) are labeled in the enlargements. Numbers correspond to station
descriptions in Table 1.
croaker across wide ranges of densities in these systems.
The similarity in these rates imply that PNAs were below
carrying capacities. Thus, I did not consider density as a
variable affecting growth or mortality for the following
comparisons of PNA quahty.
Methods
Study area
To encompass the greatest variability possible in estuarine
habitats, I sampled nursery creeks in two widely separated,
geophysically different North Carolina estuarine systems:
1) Pamlico Sound and River and 2) the Cape Fear River
( Fig. 1 ). Each system was partitioned into general areas (e.g.
upper, middle, lower) and stations were selected to repre-
sent these areas. Stations were located in creeks through-
out both systems that previous sampling (Weinstein, 1979;
Ross and Epperly, 1985; NC Division of Marine Fisheries^
indicated were consistently productive for juvenile marine
fishes during the spring-summer season. All creeks were
similar in depth, size, and sediment type. The greatest phys-
ical differences between stations were the salinities, tidal
regimes, and distances from the nearest inlets (Table 1).
Pamlico Sound is a shallow lagoon estuary whose
hydrography is controlled by wind (Giese et al., 1979;
Pietrafesa et al., 1986a; Pietrafesa and Janowitz, 1988).
NC (North Carolina) Division of Marine Fisheries. Unpubl.
data. Program 120 Nursery Area Survey, P.O. Box 769, More-
head City, NC 28557.
386
Fishery Bulletin 101(2)
Table 1
Distances to nearest inlets (D in km),
salinity (%<:) ranges
and means, mean depths (m).
and sediments of stations
sampled from
mid-March through mid-June 1987 in
the Pamlico Sound and Cape Fear systems. Sediment symbols are m = mud.
fs = fine sand.
g = grass, and fs-m = fine sand and mud.
Area
D
Salinity range (mean)
Mean depth
Sediment
Pamlico Sound
Lower
"Coastguard" Creek
3
15.5-23.0(18.9)
0.9
m
"Doctors" Creek
4
15.0-24.5(18.6)
0.7
m
"Tom Bragg Slough"
7
13.0-22.0(17.4)
0.9
fs-m
Royal Point Bay
8
17.0-20.4(18.6)
0.7
fs
Mid-Lower
Oyster Creek
26
11.0-15.0(13.5)
0.9
m
Southwest Prong
33
13.5-23.0(17.8)
1.0
m'
Merkle Hammock Creek
41
11.0-14.5(15.1)
0.9
fs-g'
Codduggen Creek
46
9.0-14.5(12.3)
0.9
fs-m'
Middle
Caffee Creek
46
10.8-18.5(13.4)
1.8
m
Oyster Creek
48
10.8-17.8(13.2)
1.2
m'
unnamed Creek
51
10.9-18.0(13.7)
1.6
fs-m
"Swan" Creek
60
10.0-15.9(12.4)
1.1
m
Tooley Cr
61
10.2-17.1(12.5)
1.3
m^
Head Rose Bay
64
10.0-13.3(11.1)
2.5
m
Upper
Mallard Creek
92
1.3-7.2(3.5)
1.2
m
Flatty Creek
99
0.9-4.0(2.6)
1.0
m
Broad Creek
101
0.3-8.7 (2.4)
1.4
m
Little Creek
104
0.6-3.9(2.3)
1.0
m
Cape Fear estuary
Lower
Molasses Creek
4
15.4-29.0(21.5)
1.0
m
Piney Pt. Creek
6
15.0-29.9(22.5)
0.6
m
Dennis Creek
7
11.8-31.0(21.7)
1.0
m-fs
Dutchman Creek
7
12.6-24.6(17.7)
0.6
m
Middle
Town Creek
29
0.0-10.2(3.7)
2.0
m
Mott Creek
29
0.0-13.4(5.6)
1.0
m
Upper
Jackeys Creek
36
0.0-6.9(1.7)
0.5
m
Toomers Creek
44
0.0-2.3(0.3)
3.0
m
Horseshoe Bend
43
0.0-2.3(0.4)
0.6
m
Smith Creek
44
0.0-4.8(1.5)
1.3
111
' See Ross and Epperly (1985) for additiona
sediment data.
Eighteen stations were located in creeks in four areas along
an approximately 100-km transect from Oeracoke Inlet to
the upper Pamlico River (Fig. 1). Four polyhaline stations
were located on Portsmouth Island (lower area). Water
depths there were largely controlled by semidiurnal lunar
tides (range usually <0.7 m, Giese et al., 1979); however,
on one occasion I observed that northerly winds (>37 km/h)
moved large quantities of water into these creeks. The mid-
lower area consisted of four stations on Cedar Island, which
exhibited less depth variation (dampened lunar tides) than
creeks on Portsmouth Island. Six creeks were sampled in
the middle area: three each in Rose and Swanquarter
bays. Tidal influence was negligible here (Pietrafesa et al.,
1986a). The creeks in the above three areas were largely
surrounded by Spartino and Juricus marsh grasses. The
upper area was represented by four creeks (oligohaline or
freshwater) surrounded by a mixture of woodlands and
patches of marsh. Water levels and currents here were
almost entirely controlled by winds or river flow (or both)
(Hobbie, 1970; Pietrafesa et al., 1986a).
The Cape Fear River is more typical (compared to the
Pamlico Sound system) of United States East and Gulf
Ross: Relative value of different estuanne nursery areas for luvenile marine fishes in Nortfi Carolina
387
coasts estuaries (i.e. a drowned river valley). Diurnal lu-
nar tides (average range about 1.5 m) were a dominant
feature throughout the study area (Welch and Parker, 1979;
Pietrafesa and Janowitz, 1988). Ten stations were located
along a 50-km transect of the Cape Fear system (Fig. 1).
In the lower estuary, four polyhaline creeks were sampled
on the west side of the inlet (Oak Island). Two creeks were
sampled on opposite sides of the middle of the estuary and
four oligohaline creeks were sampled in the upper estuary
near Wilmington. All stations in this system were sur-
rounded by Spartina marshes.
Field sampling
All stations were sampled during daylight with two one-
minute tows (68.6 m each) of a small-mesh trawl (3.2-m
headrope length, 6.4-mm bar mesh wings and body, 3.2-mm
tail bag mesh). Catches from the two tows were combined
for the station sample. Surface and bottom salinities (near-
est %c) and water temperatures ( nearest °C ) were recorded
after each sample. Mean salinities and temperatures for
each area were analyzed for differences by using <-tests for
all possible combinations of area pairs.
Sampling was designed to provide biological data during
the time of early residency in the nursery creeks, but before
significant emigration. Most recruitment of young juvenile
fishes into these creeks has ended by late-April, and some
fishes begin to emigrate by June— July (Weinstein, 1979;
Ross and Epperly, 1985; author's pers. obs.). To minimize
the influence of emigration on the calculation of growth and
mortality rates, sampling occurred during seven periods,
every other week from mid-March through mid-June 1987
(about 14 d between samples). Synoptic samples over this
large region were generated by assigning areas to four
crews for trawling during the same period of each sample
week.
Newly recruiting OWS juvenile fishes of the 1987 year
class were sorted from the catches and preserved in the
field in 100% ethyl alcohol. About one month after col-
lection the fishes were identified, counted, and standard
lengths (SL) were measured to the nearest mm. Analyses
were limited to spot and Atlantic croaker, the two most
abundant species. Catch per unit of effort (CPUE) was
calculated by dividing the total number of individuals of
a species captured by the number of trawl tows in a time
period or area. Subsamples of these fishes representing
several collection dates and all areas were measured for
SL, blotted, and weighed to the nearest 0.01 g and were
used to develop a weight-length relationship using linear
regression. Differences in weight-length relationships be-
tween areas were assessed by using analysis of covariance
(covariate=logSL) in a general linear model procedure
(SAS Institute, 1988).
Otolith aging
Subsamples for aging were randomly selected from early
(early and mid- April) and late (mid and late May) dates
and from downstream (lower) and upstream (upper) areas
in each system. Sagittae were removed from these fishes,
mounted on microscope slides with thermoplastic cement,
and polished (often on both sides) until thin sections were
obtained. Otoliths were viewed with oil immersion and
transmitted polarized light at magnifications between 500
and 625x, and images were projected through a video system
to a screen. Rings, presumed to be daily, were counted. The
formation of daily rings has either been validated (Peters
et al., 1978; Baldevarona, 1987; Siegfried and Weinstein,
1989) or assumed (Warien and Chester, 1985; Cowan, 1988)
for the two species in the size ranges used here. Even so, the
counted rings need not be deposited daily for growth rate
comparisons, nor is it necessary to know their periodicity. It
is required that groups offish being compared exhibit the
same ring formation periodicity over the time and space
of the comparison. Spot and Atlantic croaker do not form
growth rings until after yolksac absorption, about four to
five days after spawning (Peters etal., 1978; Warien, 1980).
Therefore, to estimate actual ages for mortality calcula-
tions, five days were added to the ring counts.
Although it seems reasonable to assume that juvenile
spot and Atlantic croaker form daily sagittae rings, the
precision (repeatability) of ring counts and the ability to
identify daily rings needs addressing. Before aging the
samples used in this study, I examined several hundred
spot and Atlantic croaker otoliths. Counts by myself and
2-3 other otolith readers were compared. Our ring identi-
fications were compared to samples of known age spot and
Atlantic croaker provided by the National Marine Fisheries
Service (Warien^). These preliminary samples were used
as a training device to ensure that daily rings were ac-
curately identified and were not confused with shadows or
subdaily rings. Subdaily rings may not even be resolvable
at the magnifications (<625x) used in the present study
(Campana et al., 1987; Isely^). After aging the samples used
in this study, I re-aged a random selection of spot (without
knowledge of previous age assignments) and obtained a
mean difference in counts of 2.85 (SD=2.03,n=27). Because
Atlantic croaker otolith rings were usually easier to count,
I assumed that the above count difference was generally
similar for this species. I assumed that the ages reported
here had a count precision of ±3 days. I also assumed that
any aging errors were randomly distributed throughout
the samples and were not spatially or temporally biased.
Growth and mortality
Linear regression of the form Log,gSL = 6 -i- miage) was
used to model growth. The slope of this line, m, is the
instantaneous daily growth rate. Differences in growth
rates between areas were assessed by using analysis of
covariance (covariate=age) in a general linear model pro-
cedure (SAS Institute, 1988). Absolute and relative daily
growth rates were calculated by using values predicted
with the age-SL regression equation (Ricker, 1975). For
each sampling date, mean SLs were compared between all
2 Warien, S. M. 1989. Personal commun. National Marine
Fisheries Service Beaufort Lab, Beaufort, NC 28516.
' Isely, J. 1989. Personal commun. Zoology Dept., NC State
Univ., Raleigh, NC 27695.
388
Fishery Bulletin 101(2)
30
25
i. 20
2-
c
0.05) between areas. Mean tempera-
tures throughout the Pamlico system were not significantly
different (P>0.05), except that the lower area was cooler
than the others (P<0.05). Comparisons between systems
revealed no significant differences (P>0.05) between middle
or upper area temperatures. Lower Cape Fear creeks were
Ross: Relative value of different estuanne nursery areas for luvenile marine fishes in Nortfi Carolina
389
o
PAMLICO SOUND
upper
CAPE FEAR
upper
Figure 3
Mean catch per unit of effort for spot in general areas of the Cape Fear and
PamHco Sound, NC, estuaries by sampling week, March^une 1987. Vertical
bars represent plus and minus one standard error of the mean. The -V-axis
abbreviations (from left to right) stand for March, April, April, April, May.
May, June.
significantly (P<0.0.5) warmer than those of the lower
Pamlico. The cooler temperatures in the lower Pamlico
may have resulted from sampling there at earlier times of
the day or on different days of the sampling week.
Salinity was more variable than temperature, particu-
larly within the mesohaline and polyhaline areas (Fig. 2).
Within both the Pamlico and Cape Fear systems all areas
exhibited significantly different «-test, P<0.05) bottom
salinities from each other. As expected, the Cape Fear
estuary, with its larger, more channeled river flow and
obvious tidal effects, was a more variable system than the
Pamlico system. During this study mean salinities within
the lower and middle areas of the Cape Fear system varied
over a range of 11.8%f and 14.2%f, respectively, whereas
mean salinities in all other areas (including the Pamlico)
varied over a range less than 9%f . The lower Cape Fear
creeks had significantly higher (P<0.05) salinities than
those of the lower Pamlico; however, the middle and upper
Pamlico areas had significantly higher (P<0.05) salinities
than their counterparts in the Cape Fear, even though the
upper Pamlico area was twice as far from an inlet as the
upper Cape Fear. Salinities declined rapidly in the Cape
Fear with increasing distance from the inlet; however, this
relationship was more variable in the Pamlico System
(Fig. 2). In both systems, overall mean salinity (S) was ac-
curately predicted by distance (D, in km) from the inlet:
Cape Fear: S=23.5 - 0.55(D), r2=0.95, n=10 and Pamlico:
S=20.4 - 0.17(D), r-=0.92, n = l8.
Leiostomus xanthurus
Distribution Spot was the more widely distributed of the
two species (Figs. 3-5). At the earliest sample date, small
numbers of spot had accumulated in all areas of the Pam-
lico system (Fig. 3). Peak abundance was observed in the
lower and mid-lower areas by the second sample date and
in the middle and upper areas by the third sample date.
Although more spot were collected in the lower Pamlico,
numerical differences between areas were not extreme
(Figs. 3 and 4). Overall, the least numbers of spot occurred
in the middle region (Figs. 3 and 4).
In the Cape Fear system, overall spot abundance was
similar between the upper and lower regions, and, as in the
Pamlico, the least numbers were collected from the middle
390
Fishery Bulletin 101(2)
20
10
LOWER
Ul
18 Mar
n=123
MID-LOWER
Li
n=392
~1 1 \ 1 1 1 1 r
n=785J
MIDDLE
UPPER
n = 1040
Standard length (mm)
Figure 4
Spot length frequencies in general areas of the Pamlico Sound, NC, estuary by sampling week,
March-.June 1987, Solid dots represent mean SL.
Ross: Relative value of different estuarine nursery areas for luvenile marine fishes in North Carolina
391
20
10-
0
30
20-
10-
0
30
_ 20
-.O
O
~ 10
o
a) 0-
Z 20-
c
Q)
y 10-
«)
a.
0
20-1
10-
0
20
10-
0
20-
10-
0
LOWER
18Mar
n=544
1 Apr
n.ll05
15 Apr
n-3515
29ADr
n=1337
CAPE FEAR ESTUARY
MIDDLE
—I r -T
soj. J. J.
- T . . r T . , ,
n=2
13May
n=1052
27May
n=1734
n=970
n=520
n=230
n=253
n=50
i n=18
Am .
UPPER
I
n=1
— 1 1 1 r
i^
X
n=1938
n=2705
Jl.
n=109
n = 250
n=424
• n=2012
0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60
Standard length (mm)
Figure 5
Spot length frequencies in general areas of the Cape Fear, NC. estuary by sampling week, March-June 1987. Solid dots represent
mean SL.
area creeks (Figs. 3 and 5). Initial recruitment was high
in the lower Cape Fear area, whereas it lagged behind in
the middle and upper areas (Fig. 3). Nevertheless, all Cape
Fear areas reached peak numbers by the second or third
sampling dates (Fig. 3).
Size distributions The bulk of the spot year class recruited
to all creeks from mid-March to early April; however, some
small fish (<25 mm) continued to enter the creeks through
the end of May (Figs. 4 and 5). The smallest spot captured
in both systems were always between 13 and 15 mm SL
(Figs. 4 and 5). Because the trawl can collect spot and
Atlantic croaker at least to 10 mm SL , 13-15 mm prob-
ably represented the smallest recruitment sizes of spot to
nursery habitats.
Differences in mean spot SLs between areas steadily in-
creased during the study from <1 mm (through mid- April)
to 4.3 mm in the Cape Fear and from 1 to 6.2 mm in the
Pamlico. Although mean spot SLs exhibited the small-
est variations between areas in the Cape Fear system
(Fig. 5), they were significantly different (paired ^tests,
P<0.05) between areas on most sampling dates. No area
in the Cape Fear had consistently larger or smaller mean
spot SLs. After mid-April mean SLs of spot from all four
Pamlico areas were larger than those from the Cape Fear.
Within the Pamlico system (Fig. 4) spot in the mid-lower
area had significantly larger (P<0.05) mean SL (except
sampling weeks four and six), and fish in the middle area
were always significantly smaller (P<0.05) than those in
the other three areas.
Growth Estimated spot ages ranged from 61 to 157 days
(17-35 mm SL, /!=379). All of the age-SL relationships
used to assess growth rates (Fig. 6) were highly significant
(P<0.0001). Regression residuals were evenly distributed
around zero, indicating that the exponential growth model
392
Fishery Bulletin 101(2)
40
30-
(1)
/
/
/
PAMLICO
/
Upper
/ /
'' / ''
/ /
y
y
y
y
y y
y /
' / ^
/ y
/ y
y y
y
y
y
-^,.'' Lo9,oSl=0.935
+ 0.004(Age)
T 1 1 1 1 —
r2=0.93.
n = 96
r 1 1 r
Loa,QSL=0.875 + 0.005(A8e)
r2 = 0.91,n.96
-I 1 1 1 1 f—
40-
30
20
10-
CAPE FEAR
Upper
" Lo9|„Sl = 0.856»0.005(A9o)
r^ = 0.94,n=91
CAPE FEAR
Lower
log,QSL=0.829+0.005(Agel
r2 = 0.90,n = 96
■l 1 r—
-I 1 1 f-
60 80 100 120 140 160 60 80 100 120 140 160
Age (days)
Figure 6
Spot growth rate curves (solid lines) based on otolith daily ages for upper and lower areas
of the Pamlico Sound and Cape Fear, NC estuaries. Dotted lines represent 95 percent
confidence intervals.
(logj^SL = 6 -t- mkage)) was appropriate. Instantaneous
daily growth rates (slopes of the regressions) were similar
(analysis of covariance, P>0.05) between upper and lower
areas in the Cape Fear and the lower Pamlico (Fig. 6).
Therefore, a combined age-SL regression for all spot in the
upper and lower Cape Fear estuary and the lower Pamlico
area was developed: logipSL = 0.861 + 0.0048(a^e), r2=0.90,
n=2S3. Analysis of covariance indicated that spot from
the upper Pamlico region exhibited significantly slower
(P>0.05) overall growth rates (Fig. 6) than fish from the
other three areas.
Age-specific absolute and relative growth of spot was
predicted from the age-SL regression for the upper and
lower ("ape Fear and lower Pamlico combined and the up-
per Pamlico (Table 2). Predicted absolute growth rates in
the Cape Fear and lower Pamlico areas increased from 0.16
mm/d between 60 and 65 days of age to 0.43 mm/d between
ages 150 and 155 days of age, and the largest increase in
absolute growth occurred between ages 95 and 105 days of
age (Table 2). Relative growth remained constant around
1.13-1.14 %/d SL over the whole age range examined
(Table 2). Although predicted sizes at ages were larger in
the upper Pamlico area than those of the other three areas,
the absolute growth rates were lower, increasing from 0.16
mm/d between ages 60 and 65 d to 0.39 mm/d between ages
150 and 155 d (Table 2). Absolute growth rates in this area
also exhibited the largest increases around 100-105 days.
Relative growth rates in the upper Pamlico were lower
than in the other areas and averaged 1.01 9t/d SL (Table 2).
The ages when absolute growth in all areas was greatest
(95-105 d) translated to SL ranges around 21-23 mm. This
SL range dominated the length frequencies in all areas
during the first two weeks of April (Figs. 4 and 5). Water
temperatures were steadily increasing in all areas prior to
mid-April (Fig. 2).
(Jrowth was also compared by using weight-length rela-
tionships. These relationships for spot were highly signifi-
cant (analysis of covariance, P<0.0001) and took the usual
Ross: Relative value of different estuarine nursery areas for |uvenile marine fisfies in Nortfi Carolina
393
Table 2
Predicted age-specific
mean standard lengths (SL), absolute (mm/d), and relative (%/d SL) grow
th rates for spot from the upper and
lower Cape Fear and lower Pamlico combined (CFR
f LPAM land the
upper Pamlico area (UPPAM).
Age
CFR + LPAM
UPPAM
Mean
Absolute
Relative
Mean
Absolute
Relative
(days)
SL
growth rate
growth rate
SL
growth rate
growth rate
60
14.09
15.60
65
14.89
0.16
1.14
16.39
0.16
1.01
70
15.74
17.22
75
16.63
0.18
1.13
18.09
0.17
1.01
80
17.58
19.01
85
18.58
0.20
1.14
19.98
0.19
0.99
90
19.63
20.99
95
20.75
0.22
1.14
22.05
0.21
1.00
100
21.93
23.17
105
23.17
0.25
1.13
24.35
0.24
1.04
110
24.49
25.59
115
25.88
0.28
1.14
26.88
0.26
1.02
120
27.35
28.24
125
28.91
0.31
1.13
29.68
0.29
1.03
130
30.55
31.19
135
32.28
0.35
1.15
32.77
0.32
0.99
140
34.12
34.43
145
36.06
0.39
1.14
36.18
0.35
1.02
150
38.11
38.02
155
40.27
0.43
1.13
39.95
0.39
1.03
Table 3
Weight-standard length (W-SL) relationships
for spot and Atlantic croaker from all areas of the Cape Fear and Pamlico systems,
March-June 1987.
Area
Formula
r-
n
Spot
Cape Fear Upper
W= 10-« 2.5,5^4 02)
0.96
100
Cape Fear Middle
W = 10-» 90(SL3 05)
0.97
134
Cape Fear Lower
W= 10-5 00(SL3 12)
0.97
212
Pamlico Upper
W = 10-6 "^CSZ-^-S')
0.98
456
Pamlico Middle
W = 10-5 'stSL^es)
0.97
252
Pamlico Mid-Lower
W = 10-« 12,Si,3.89)
0.98
246
Pamlico Lower
W = 10-« 22(S/,3 95)
0.97
250
Atlantic croaker
Cape Fear Upper
W = 10-5 80(SL3,64)
0.97
138
Cape Fear Middle
W = 10-5 30(SZ,3 31)
0.86
118
Cape Fear Lower
W = 10-5 64(SL3«)
0.93
98
Pamlico Upper
W = 10-5 55(SL3 «)
0.96
188
Pamlico Lower
W = 10-»85{SL2 ^SS)
0.88
46
curvilinear form (Table 3, Fig. 7). In the Pamlico system dif-
ferences between areas were not large; however, middle area
spot had significantly (P<0.05) lower weights per length, es-
pecially in larger individuals (Fig. 7). In the Cape Fear sys-
tem, spot in the upper area had significantly (P<0.05) larger
weights per length than those from the other two areas.
Mortality Spot mortality rates were based on individuals
aged >85 days. All regression slopes describing the declin-
ing numbers of spot with increasing ages were significantly
different from zero (P<0.0001). Instantaneous mortality
rates over this time period ranged from 0.037 to 0.066
(Fig. 8). Analysis of covariance indicated that within each
394
Fishery Bulletin 101(2)
3
PAMLICO ,
2-
//
] .
X'Middl*
ai
2 0-
CD
CO 4-
1
■
CAPE FEAR /
3
Upp«r /
2-
1-
/ y
/ y
0-
-^^^^^""^
15 25 35 45
Standard lengtti (mm)
Figure 7
Spot weight-length relationships for the Pamlico
and Cape Fear, NC, estuaries, March-June 1987.
The line for the area that exhibited a significant
difference was labeled (i.e. Pamlico Middle area.
dashed line, was different and Cape Fear Upper
area, solid line was different); the remaining line
in each panel represented the relationship for all
other areas combined.
system, upper area spot displayed significantly (P<0.05)
lower instantaneous mortality rates than did fish from
the lower estuaries, especially in Cape Fear Lower Cape
Fear spot exhibited a statistically similar (P>0.05) instan-
taneous mortality rate to those in the lower and upper
Pamlico creeks. Daily mortality rates during the present
study were 4.97 %/d in the upper Pamlico, 6.39 %/d in the
lower Pamlico, 3.63 %IA in the upper Cape Fear, and 6.01
'7r/d in the lower Cape Fear
Micropogonias undulatus
Distribution Atlantic croaker distributions in both sys-
tems were skewed toward upstream, oligohaline creeks
(Fig. 9-11). In the Pamlico system almost no Atlantic
croaker were collected in the lower or middle areas (Fig. 9).
Atlantic croaker recruitment in the Pamlico lagged behind
the ("ape Fear in abundance and timing (Fig. 9), and peak
densities occurred throughout the Pamlico near the end of
the sampling.
Patterns of Atlantic croaker recruitment were like those
of spot in the upper and middle Cape Fear. Like spot, most
of the Atlantic croaker year class had recruited to these ar-
eas by mid-April, although small Atlantic croaker (<20 mm)
continued to colonize these creeks through mid-late May
(Fig. 10). Also, peak abundance was reached in the middle
and upper Cape Fear during the same weeks as those for
spot (2nd and 3rd, respectively) (Fig. 9). Except for the
larger CPUE in mid-March, Atlantic croaker recruitment
in the lower Cape Fear was similar to that of most Pamlico
system creeks (Fig. 9).
Atlantic croaker entering PNAs from mid-March through
late April in both systems appeared to bypass lower and
middle area nursery creeks (unlike spot) to a greater extent
than fish recruiting after April (Fig. 9). Late recruitment
of small Atlantic croaker was especially apparent in the
mid-lower Pamlico (Fig. 11).
Size distribution Atlantic croaker initially collected in the
Cape Fear creeks through the first week of April spanned
a size range of 11-23 mm (Fig. 10, 15.4 mm SL mean), and
mean sizes were not significantly different (paired /-test,
P>0.05) between areas in the first or second sampling
week. After this time, Atlantic croaker from the middle
Ross: Relative value of different estuanne nursery areas for juvenile manne fishes in North Carolina
395
PAMLICO
PAMLICO
10-
Upper
"-v^ Lower
8-
6-
4-
"^^^^%^--~ i
^^\^ ^*^
2-
LoggN=12.40-0.051(Age) ""-~- ^^^-^^ ""
r2 = 0.86,n=55 ""~--,^ ^^^
LoggN=14.44-0.066(AgeP-^ ^\^^ "-^
r2 = 0.87,n=50 ""^^ ^^-.^^
E 0-
e J
C
"^^^ -
^"~-^.
^ 10-
^
CAPE FEAR
CAPE FEAR
8-
"~-^^^ Upper
^^^ ^•-^^ Lower
6-
4-
""^^^S^
~~\$C\-
2-
LoggN=9,79-0.037(Age) ""--^^
LogeN = 13.34-0.062(AgeP-.
0-
r2 = 0.60,n=42
r2 = 0.76,n=35
60 80 100 120 140 160 180 200 220 60 80 100 120 140 160 180 200 220
Age (days)
Figure 8
Spot mortality rate curves (solid lines) based on otolith daily ages for upper and lower areas of the Pamlico and
Cape Fear, NC, estuaries. Dotted lines represent 95 percent confidence intervals.
Cape Fear area were usually significantly shorter (P<0.05)
than those from other areas, and Atlantic croaker from the
lower Cape Fear were significantly longer (P<0.05) than
those of other areas.
Atlantic croaker collected from the upper Pamlico creeks
were significantly larger {P<0.05) on all sample dates than
those occupying any area of the Cape Fear during the same
weeks. Within the Pamlico, upper and mid-lower mean SLs
were the same (P>0.05) except in the last week when mid-
lower mean SL was significantly larger (P<0.05).
Growth Atlantic croaker ( 12-35 mm SL, n =383 ) ages esti-
mated from otoliths ranged from 62 to 234 days, and all
age-SL relationships (Fig. 12) were significant (P<0.0001).
Residuals of these regressions exhibited no pattern: there-
fore, the growth models appeared to be appropriate. The
instantaneous daily growth rates within each system
were not significantly different (analysis of covariance,
P<0.05) between upper and lower areas (Fig. 12). Between
systems, upper Pamlico Atlantic croaker grew more slowly
than those from the upper Cape Fear (P<0.05). Overall age-
length relationships for upper and lower Cape Fear com-
bined and for the upper and lower Pamlico Atlantic croaker
combined were the following: for Cape Fear — log, gSL =
0.915 -I- 0.0027{age ), r2=0.87, n=229: for Pamlico— logj^SL =
0.970 -I- 0.0024(age), r2=0.87, n=158.
The above combined equations for each system were
used to calculate age-specific absolute and relative Atlan-
tic croaker growth rates (Table 4). Early absolute Atlantic
croaker growth rates in the Cape Fear system increased
most rapidly in ages <105 days, averaging 0.085 mm/d
(Table 4). After this age. Cape Fear growth rates increased
at a steady, slow rate, reaching 0. 19 mm/d by age 215 days.
Relative Atlantic croaker growth rates in the Cape Fear
were constant over the whole age range at about 0.63 %/d
SL. The larger Pamlico system Atlantic croaker exhibited
similar absolute growth rates to Cape Fear fish and these
increased rapidly from 0.077 mm/d between ages 60 and 65
d to 0.106 mm/d between ages 120 and 125 d to 0.175 mm/d
between ages 210 and 215 d (Table 4). Relative growth
rates were less than those from the Cape Fear and were
constant around 0.56 %/d SL.
Weight-length relationships for Atlantic croaker were
highly significant in all areas (P<0.G001 ) (Table 3). In both
systems fish from upper area creeks exhibited significantly
larger (P<0.05) weights per length than those from other
areas, particularly at the larger sizes (Fig. 13). Slopes of
middle and lower Cape Fear weight-length relationships
were not significantly different from each other (P>0.05).
Mortality Catch curves used to estimate Atlantic croaker
mortality rates were calculated by using ages >125 days.
All regression slopes were significantly different from zero
(P<0.0001), although the relationship was more variable
for the mid-lower Pamlico area because of the small sample
size. Instantaneous mortality rates for Atlantic croaker in
the nursery creeks ranged from 0.008 to 0.038 (Fig. 14).
Atlantic croaker in the upper and mid-lower Pamlico
396
Fishery Bulletin 101(2)
PAMLICO SOUND
upper
Figure 9
Mean catch per unit of effort for Atlantic croaker in general areas of the
Cape Fear and Pamlico Sound, NC, estuaries by sampling week, March-
June 1987. Vertical bars represent plus and minus one standard error of
the mean. The .v-axis labels (from left to right) stand for March, April. April,
April, May, May, June.
areas had similar instantaneous mortality rates (analysis
of covariance, P>0.05). Upper Cape Fear Atlantic croaker
exhibited significantly lower mortality rates than those in
the lower Cape Fear (P<0.0,5). All Atlantic croaker mortal-
ity rates in the Cape Fear were significantly higher than
those in the Pamlico. Daily mortality rates for Atlantic
croaker in the upper and lower Cape Fear were 2.96 %/d
and 3.73 %/d, respectively and in the upper and mid-lower
Pamlico were 0.90 and 0.80 ^r/d, respectively.
Discussion
Primary nursery area habitats in two different estuaries
were not equally valuable for spot and Atlantic croaker.
Considered together, growth, mortality, and distribution
data indicated that upstream oligohaline creeks provided
the best environment, followed closely by downstream
polyhaline areas. In all regards, the middle reaches of
the estuary appeared to be less valuable (or at least less
used). These consistent results for both species in the two
separate estuarine systems lend support to their general
applicability. Other studies were marginally useful in
evaluating these results because of their lack of synoptic
comparisons across a wide variety of habitats and because
of limitations in or lack of growth and mortality data for
estuarine juveniles of these species.
The main evidence that oligohaline habitats provided
better environments than polyhaline areas was that spot
(both systems) and Atlantic croaker (in Cape Fear) exhib-
ited significantly lower mortality in the freshwater PNAs.
Miller et al. ( 1985) reported lower mortality for these spe-
cies in mesohalino areas compared to high salinity areas
of Pamlico Sound. In the Cape Fear River, Weinstein and
Ross: Relative value of different estuanne nursery areas for juvenile marine fisfies in North Carolina
397
30
20
10-
Li
50
20-
10-
^ 30
Sl 20
>^
^ 10-
3
n- 0
0)
I 20-
£ 10
20-
10
0
20-1
1 . t.
1.
18 Mar
n.30
lApr
na6
15 Apr
n=6l
29 Apr
n^ 190
l3May
n=573
27May
n=3a09
-1 — r— r-
10 Jun
-1"^
CAPE FEAR ESTUARY
MIDDLE
1
i
< 1 «i r-»r-
X
4-
n=689
i^
^* 1 T
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Standard length (mm)
Figure 10
Atlantic croaker length frequencies in general areas of the Cape Fear, NC, estuary by sampling week, March-June 1987. Solid dots
represent mean SL.
Walters (1981) found consistently high mortality for spot
in polyhaline creeks during two years, but variable mortal-
ity between years (one year higher, one year lower) in low
salinity regions. Mortality rates reported in the present
study may not be affected by fish density (Ross, 1992), and
it is unlikely that starvation (Currin et al., 1984) played a
major mortality role. Predation may cause most of the PNA
natural mortality; it was previously proposed that preda-
tion rates were lowest in oligohaline habitats because these
areas contained relatively fewer predators (Weinstein and
Walters, 1981; Currin et al., 1984; Miller et al., 1985). This
hypothesis continues to lack direct, convincing evidence.
Predators in oligohaline habitats (e.g. southern flounder,
catfishes, gar, striped bass, etc.) may, in fact, be just as
numerous near the upriver nurseries (author's pers. obs.;
Patrick and Moser, 2001; Moser"*) as marine predators are
around polyhaline creeks. Also, because water levels in the
upriver creeks, especially in the Pamlico, do not vary as
much as in polyhaline areas, predators may have more op-
portunity to use these creeks (Currin et al., 1984).
* Moser, M. L. 1998. Personal commun. NW Fisheries Science
Center, NMFS, 2725 Montlake Blvd., Seattle, WA 98112 .
One alternative explanation for lower mortality esti-
mates in upriver PNAs is that mortality could be related,
perhaps indirectly, to ambient salinity. Although freshwa-
ter conditions probably do not increase mortalities of these
fishes (Moser and Hettler, 1989), there may be negative
effects of high salinity on survival that have not been in-
vestigated. Moser and Hettler (1989) reported that spot
exhibited the highest respiration rates in high salinity
conditions, which suggest increased stress.
Another potential explanation is that fishes may leave
high salinity areas more rapidly than freshwater areas.
Although 1 attempted to minimize effects of emigration on
mortality estimates by limiting the analyses to the period
before mid-June, the mortality rates 1 calculated could have
contained an unknown effect of emigration. Other studies
(Weinstein, 1983; Weinstein and O'Neil, 1986; Miller and
Able, 2002; author's pers. obs. ) supported my assumption
that emigration of spot and Atlantic croaker from PNAs
was negligible at least through June. Such early habitat
fidelity seems to be a common trait among juvenile fishes
(Rountree and Able, 1992; Ross and Lancaster, 2002). Many
individuals of OWS juvenile fishes leave PNAs by July
(Ross, 1988; NC Division of Marine Fisheries^); therefore,
398
Fishery Bulletin 101 (2)
PAMLICO
UPPER
[ [[ 18Mar
n= 3
^-U
MID-LOWER
*^,*
n=78
J^-M
~i 1 1 1 1 r
-+-
n=70
n = 47
0 10 20 30 40 50 60 70
Standard length (mm)
Figure 11
Atlantic croaker length frequencies in general areas of the Pamlico Sound, NC, estuary by sampling week, March-
June 1987. Solid dots represent mean SL.
previous mortality estimates are likely confounded by emi-
gration because measures of declining fish numbers were
extended longer into the nursery season (through August,
Weinstein and Walters, 1981; through October, Currin et
al., 1984; through July, Miller et al., 1985).
Growth in weight (weight-length relationships) also in-
dicated advantages of oligohaline habitats for these fishes.
Higher weights per length have been equated with greater
fitness (Friedland et al, 1988; Bolger and Connolly, 1989).
Improved fitness was suggested by a consistent trend for
individuals of both spot and Atlantic croaker in both sys-
tems to be heavier per length in the oligohaline creeks.
Laboratory experiments on spot (Moser, 1987) resulted in
heavier fish per length in freshwater, and the weight differ-
ence was attributed to a higher feeding rate in freshwater,
rather than water absorption because of osmotic imbal-
ance. Spot from oligohahne areas of the James River, VA,
were heavier per length compared to those from several
other estuaries (McCambridge and Alden, 1984), but the
role of salinity in these differences was unclear. Peterson
et al. (1999) indicated that reduced salinity itself caused
higher growth rates (in weight) for Atlantic croaker in oli-
gohaline conditions.
Growth (in length) rates and size distributions indi-
cated that PNA habitats at extreme ends of estuaries were
equally valuable to both species (with one exception). The
exception — depressed spot growth rates in the upper Pam-
lico area — did not appear to be correlated with lower sa-
linities or temperatures because spot from other areas with
low salinity and similar or lower temperatures exhibited
higher growth rates. The most obvious difference between
upper Pamlico creeks and all other areas was the extremely
long (often >100 km) estuarine migration required to reach
them. Potential costs involved in such migrations should
be examined as should the degree to which the lower spot
growth rates persisted into later life. General lack of growth
rate variation between oligohaline and polyhaline habitats
suggested that salinity (and probably tidal influence) did
not affect growth to a degree detectable in the present study.
This conclusion is supported by previous studies (Moser and
Gerry, 1989; Moser and Hettler, 1989; Miller et al., 2000)
despite a general prediction that fish growth rates should
be higher in brackish waters (Boeuf and Payan, 2001).
The lack of evidence for negative effects of fish density on
growth (Ross, 1992) indicated that resources in oligohaline
or polyhaline PNAs may not limit these fishes. Currin et al.
( 1984) also suggested that food resources did not limit spot
production in middle areas of Pamlico Sound.
Lack of spatial variation in early estuarine growth rates
was also found in the few relevant studies available. Wein-
Ross: Relative value of different estuarine nursery areas for juvenile marine fisfies in North Carolina
399
40
PAMLICO
PAMIICO
Upper /
Lower /
/
/
/
/
/
/ ,
/
/ /
/ /
/ /
./ /
/ /
/ /
30-
y
• X y
^ X ^
20-
yy/
^■y--''
' X y
yy
^y
>'
,^-' Log,pSL=0.979+0.002(Ag«)
Loa,QSL=0.98U0.002(A9e)
r2 = 0.85,n=121
r2 = 0.91,n = 35
o> 10-
c
' — r — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — r — 1 — 1 — r — i — i — i — i — r— i — r— i i i i i i i -i— i t i t r i i i i • ■ . i
-5? 1
"D
C2
■a
c
CAPE FEAR ,'
Upper /
CAPE FEAR
^ ^0-
Lower ^
/
/
/
/ /
/
/ /
/
/ /
/
/ /
/ /
30-
/ / / ■
//
f / '
/ /
/ / /
/ /
/ / /
' / '
20-
/y
.'#^
y / •
^ y y
■'y>-"
■■::^"
-^^' Loa,QSl = 0.907+0.003(Age)
^^''Log,(jSL=0.936+0.003(Age)
10-
'" r'' = 0.87,n=127
'' r2-0.87,n = 100
50 70 90 110 130 150 170 190 210 230 50 70 90 110 130 150 170 190 210 230
Age (days)
Figure 12
Atlantic croaker growth rate curves (solid lines) based on otolith daily ages for upper and
lower areas of the Cape Fear and Pamlico Sound, NC. estuaries. Dotted lines represent
95 percent confidence intervals.
stein and Walters (1981) and O'Neil and Weinstein (1987)
reported no consistent differences in spot growth rates be-
tween oligohaline and polyhaline creeks in the Cape Fear
River and York River, VA, estuaries, respectively. Miller et
al. (1985) indicated that spot and Atlantic croaker growth
rates were probably not different between Pamlico Sound
mesohaline and polyhaline areas. Similarly, Beckman and
Dean (1984) found no significant differences in spot growth
rates among localities in a small, polyhaline South Caro-
lina estuary. Necaise (2000) failed to find growth differ-
ences among juvenile summer flounder caged (and fed ad
libitum) over a wide range of abiotic habitats in southern
North Carolina. Guindon and Miller (1995), however, did
find growth rate differences among caged (not fed) south-
ern flounder across abiotically similar oligohaline habi-
tats in the Pamlico River. Differences in fish growth rates
among estuarine habitats (e.g. Sogard, 1992) indicate that
there are different species-specific responses to habitats,
responses related to zoogeography, or responses related to
habitat structure or food availability. My data and most of
the above studies, covering different years and a variety of
estuaries, suggest that variation in growth rates, especially
for spot, between PNAs is generally lacking or at least dif-
ficult to detect. Such results are consistent with the view
that these fishes are hardy, omnivorous, opportunistic
colonizers of an undersaturated environment.
Increasing evidence suggests that oligohaline or fresh-
water habitats in the southeastern United States are im-
portant nurseries for the OWS juvenile fishes (Rogers et
al, 1984; Rozas and Hackney, 1984;Moser and Gerry, 1989;
Moser and Hettler, 1989; Peterson and Ross, 1991). In fact,
they may be the most valuable habitats, particularly for
maximizing survival of some species. The nursery creeks
I sampled supported similar growth rates for two species;
however, fitness may be most improved upriver, where both
growth (in weight) and survival are optimized. Anderson
(1988) predicted that juvenile temperate fishes gener-
ally choose to maximize growth over reducing mortality.
400
Fishery Bulletin 101(2)
Table 4
Predicted
the upper
age-specific mean standard lengths (SL), absolute (mni/d), and relative (%/d SL) growth rates for Atlantic croaker from
and lower Cape Fear combined (CFR) and the upper and mid-lower Pamlico combined (PAM).
Age
(days)
CFR
PAM
Mean
SL
Absolute
growth rate
Relative
growth rate
Mean
SL
Absolute
growth rate
Relative
growth rate
60
11.94
13.61
65
12.32
0.08
0.63
14.00
0.08
0.57
70
12.71
14.39
75
13.11
0.08
0.62
14.79
0.08
0.56
80
13.52
15.21
85
13.95
0.09
0.63
15.63
0.08
0.55
90
14.39
16.07
95
14.84
0.09
0.63
16.52
0.09
0.56
100
15.31
16.98
105
15.79
0.10
0.63
17.45
0.10
0.56
110
16.29
17.95
115
16.81
0.01
0.64
18.45
0.10
0.56
120
17.34
18.97
125
17.89
0.11
0.63
19.50
0.11
0.56
130
18.45
20.04
135
19.03
0.12
0.63
20.61
0.11
0.57
140
19.63
21.18
145
20.25
0.12
0.64
21.78
0.12
0.56
150
20.89
22.39
155
21.55
0.13
0.63
23.01
0.13
0.56
160
22.23
23.66
165
22.94
0.14
0.63
24.32
0.13
0.56
170
23.66
25.00
175
24.41
0.15
0.63
27.70
0.14
0.56
180
25.18
26.42
185
25.97
0.16
0.63
27.16
0.15
0.56
190
26.79
27.93
195
27.64
0.17
0.63
28.71
0.16
0.56
200
28.51
29.51
205
29.41
0.18
0.63
30.34
0.17
0.56
210
30.34
31.19
215
31.30
0.19
0.63
32.06
0.18
0.56
although the two are intimately related (Werner and Gil-
liam, 1984). Selecting for optimized growth, however, ap-
pears not to be an issue for these two estuarine generalists.
If upstream PNAs are better nurseries (i.e. provide better
conditions for survival and perhaps growth), delayed PNA
recruitment (longer estuarine migrations), especially for
Atlantic croaker, may maximize ultimate fitness (Miller
et al., 1985; Shapiro, 1987). Factors affecting transport of
young to upstream areas may, therefore, be an important
determinant of population fitness.
Unexpected patterns of recruitment into middle region
creeks suggested that their function or recruitment po-
tential as fish nursery areas may difTer significantly from
other regions. Even though these creeks were physically
similar to creeks on either end of the estuarine transects,
lower abundances of spot and Atlantic croaker in middle
areas suggested that they either avoided (bypassed) middle
areas or endured higher initial mortalities there. Higher
initial mortality in middle regions seems unlikely because
catches were generally low throughout the sampling pe-
riod. Relatively poor habitat quality could explain the low
densities of fishes in these creeks. This hypothesis was sup-
ported by the fact that most fishes settling in middle re-
gions of both systems exhibited significantly smaller mean
lengths and were lighter per length. The same pattern was
observed for Atlantic menhaden in these systems (Ross,
1992). Szedlmeyer (1991) also found lower abundances
and species richness in middle reaches of a Florida estuary
and suggested that either less diverse habitat or greater
salinity variation (or both) influenced this result. Ross
and Kpperly (1985) found stations close to the periphery
of Pamlico Sound (including the middle area of this study)
Ross: Relative value of different estuanne nursery areas for juvenile manne fishes in Nortfi Carolina
401
PAMLICO
2
Uppsr /
1-
y^ Lower
Weight (g)
o
^.^^'''^
CAPE FEAR .. /Mid
Upper / /
2-
/^
J/
1-
Jt'/ Lower
0-
-.^..^.^^^'^''^
10 20 30 40 50
Standard length (mm)
Figure 13
Atlantic croaker weight-length relationships by area in
the Pamlico and Cape Fear, NC, estuaries, March-June
1987.
to be the most productive, but their study lacked stations
near the inlets and in freshwater areas. Weinstein et al.'s
(1980) uppermost Cape Fear stations were the same as
my middle area and generally produced lower densities of
spot than polyhaline areas, but lacking upriver stations,
the meaning of this in the present context is inconclusive.
These fishes seem to opt either for rapid settlement in poly-
haline environments or delayed settlement in oligohaline
areas — mesohaline settlement being less "preferred."
The conclusion that PNAs were not equally valuable
and the observation that spot were not most abundant
in the best habitats, indicated that variation in estuarine
distribution could control or at least regulate (fine tune)
year-class strength. If movement to general regions of
the estuary is largely passive (Pietrafesa et al., 1986b; Pi-
etrafesa and Janowitz, 1988), then my results predict that
year-class strength of these species would be decreased
when transport conditions force the majority of the recruits
toward middle or lower region PNAs. Alternatively, year
class strength would be enhanced by conditions favoring
greater upstream transport, assuming carrying capacities
of the habitats were not exceeded. Ross (1992) proposed
that these systems were recruitment limited, that post-
settlement mortality was less important in controlling
year-class strength than early life history events prior to
settlement. If true, factors affecting variation in estuarine
distribution may indirectly adjust year-class strength, not
control it. Additional data on mortality rate variation in
relation to density during the estuarine and oceanic early
life history is required to validate this hypothesis.
Acknowledgments
I thank John M. Miller, G, T. Barthalmus, L. B. Crowder,
and L. J. Pietrafesa for their support during this study. I
thank K. H. Pollock for statistical advise. Field sampling
required the efforts of many people. The NC Division of
Marine Fisheries (Washington and Wilmington offices)
played a large role in sampling, and I especially thank
Fred Rohde, John Schoolfield, Otto Rutten, Morris Allison,
Greg Judy, Lele Tison, and Jess Hawkins of that organiza-
tion. B. M. "Mac" Currin was important throughout the
study, and I thank him for his contributions in the field,
laboratory, and in reviewing manuscripts. John S. Burke
also provided help in the field. I thank the Beaufort Labo-
402
Fishery Bulletin 101(2)
I °
^ 10-
o
_I
8
PAMLICO
Upper
Lo9gN=5.48-0.009(Age)
2] r2 = 0.73,n = 31
-I — I — I — I — 1 — I — I — I — I — I — I — I — 1 — I — r-
CAPE FEAR
Upper
LogaN = 11.44-0.030(Age
r2 = 0.90,nO6
log^N=3.53-0.008(Age)
r2 = 0.28,n = 27
PAMLICO
Lower
T — 1 — I — r — 1 — r — I — 1 — I — 1 — I — I — I — r — i — i — i — i — i — r
CAPE FEAR
Lower
LoggN=12.61-0.038(Age
r2=0.94,n = 30
T — I—I — I — I — I — I — I — 1 — I — I — I — 1—1 — I — 1 — I — 1 — I — 1—1 — I — 1 — 1 — I— - — I — I — I — I — I — 1 — I — 1 — I — I — I — 1—1 — 1 — I — I — I — I — I — I — I — I — I — I — r-
70 90 110 130 150 170 190 210 230250 270 290 70 90 110 130 150 170 190 210 230250 270 290
Age (days)
Figure 14
Atlantic croaker mortality rate curves (solid lines) based on otolith daily ages for upper and lower areas of the
Pamlico and Cape Fear, NC, estuaries. Dotted lines represent 95 percent confidence intervals.
ratory (National Marine Fisheries Service) and the Biol-
ogy Laboratory of Carolina Power and Light Company for
providing space. David Colby made valuable contributions
to this research. I appreciate Jeff Isely's advice and help
in analyzing fish otoliths. I thank Ernie Aschenbach for
sorting samples and mounting otoliths. And lastly I thank
Mary L. Moser for help and support during all stages of
this work from field sampling to reading numerous manu-
script drafts.
Literature cited
Anderson, J. T.
1988. A review of size dependent survival during pre-recruit
stages of fishes in relation to recruitment. J. Northwest
Atl. Fish. Sci. 8:55-66.
Baldevarona, R. B.
1987. Effects of feeding and stocking density on growth and
survival of spot, LeioHtDtnus xanthurus. Ph.D. diss., 117 p.
Univ. South Carolina, Columbia, SC.
Beck, M. W., K. L. Heck Jr. K. W. Able, D. L. Childers,
D. B. Eggleston, B. M. Gillanders, B.. Halpern, C. G. Hays,
K. Hoshino, T. J. Minello, R. J. Orth, P F Sheridan,
M. P. Weinstein.
2001. The identification, con.servation. and management of
estuarine and marine nurseries for fish and invertebrates.
Bioscience 51:6.'i.'i-641.
Beckman, D. W., and J. M. Dean.
1984. The age and growth of young-of-the-year spot, Leios-
tomus xanthurus Lacepode, in South Carolina. Estuaries
7:487-496.
Boeuf, G., and P. Payan.
2001. How should salinity influence fish growth? Comp.
Biochem. Physiol, (part C) 130:411-423.
Boesch, D. F, and R. E. Turner
1984. Dependence of fishery species on salt marshes: the role
of food and refuge. Estuaries 7:460- 468.
Bolger, T, and R L. Connolly.
1989. The selection of suitable indices for the measurement
and analysis offish condition. J. Fish. Biol. 34:171- 182.
Campana, S. E., J. A. Gagn, and J. Munro.
1987. Otolith microstructure of larval herring iClupea
harengus): image or reality? Can. J. Fish. Aquat. Sci. 44;
1922-1929.
Cowan, J. H., Jr.
1988. Age and growth of Atlantic croaker, Micropogonias
undulatus, larvae collected in the coastal waters of the
northern Gulf of Mexico as determined by increments in
saccular otoliths. Bull. Mar Sci. 42:349-357.
Crecco, V, T. Savoy, and L. Gunn.
1983. Daily mortality rates of larval and juvenile American
shad (A/o.sa sapidissima) in the Connecticut River with
changes in year-class strength. Can. J. Fish. Aquat. Sci.
40:1719-1728.
Currin, B. M., J. P. Reed, and J. M. Miller.
1984. Growth, production, food consumption and mortality
of juvenile spot and croaker: a comparison of tidal and non-
tidal nursery areas. Estuaries 7:451-459.
Essig, R. J., and C. F. Cole.
1986. Methods of estimating larval fish mortality from daily
increments in otoliths. Trans. Am. Fish. Soc. 115:34-40.
Friedland, K. D., G. C. Carman, A. J. Bejda, A. L. Studholme,
and B. 011a.
1988. Interannual variation in diet and condition in juvenile
Ross: Relative value of different estuanne nursery areas for luvenile manne fishies in Nortfi Carolina
403
bluefish during estuanne residency. Trans. Am. Fish. Soc.
117:474-479.
Giese, G. L., H. B. Wilder, and G. G. Parker Jr.
1979. Hydrology of major estuaries and sounds of North
Carolina. U.S. Geological Survey Water Resources Invest.
79-46, 175 p.
Guindon, K. Y., and J. M. Miller
1995. Growth potential of juvenile southern flounder, Parali-
chthys lethostigma, in low salinity nursery areas of Pamlico
Sound, North Carolina, USA. Netherlands J. Sea Res. 34:
89-100.
Hobbie, J. E.
1970. Hydrography of the Pamlico River Estuary, N.C.
Water Resources Res. Inst. Rep. 39, 69 p. Water Resources
Research Institute, Raleigh, NC.
McCambridge, J. T., Jr, and R. W. Alden III.
1984. Growth of juvenile spot, Leiostomus xanthurus Lace-
pede, in the nursery region of the James River, Virginia.
Estuaries 7:478-486.
Mclvor, C. C, and W E. Odum.
1988. Food, predation risk, and microhabitat selection in a
marsh fish assemblage. Ecol. 69: 1341-1351.
Mercer, L. P.
1987a. Fishery management plan for Atlantic croaker {Mi-
cropogonias undualtus). Atlantic States Mar Fish. Comm.
Fish. Management Rep. 10, 90 p. Atlantic States Marine
Fisheries Commission, Washington, DC.
1987b. Fishery management plan for spot (.Leiostomus
xanthurus). Atlantic States Mar. Fish. Comm. Fish. Man-
agement Rep. 11, 81 p. Atlantic States Marine Fisheries
Commission, Washington, DC.
Miller, J. M., L. B. Crowder, and M. L. Moser
1985. Migration and utilization of estuarine nurseries by
juvenile fishes: an evolutionary perspective. In Migration:
mechanisms and adaptive significance (M. A. Rankin, ed.),
p. 338-352. Contrib. Mar Sci. (suppl. 27.
Miller, J. M., W. H. Neill, K. A. Duchon, and S.W. Ross.
2000. Ecophysiological determinants of secondary produc-
tion in salt marshes: a simulation study. In Concepts and
controversies in tidal marsh ecology (M. P. Weinstein and
D. A. Kreeger, eds.), p. 315-331. Kluwer Academic Press,
Dordrecht, NL.
Miller, M. J. and K. W. Able.
2002. Movements and growth of tagged young-of-the-year
Atlantic croaker (Micropogonias undulatus L.) in restored
and reference marsh creeks in Delaware Bay, USA. J. Exp.
MarBiol. Ecol. 267:15-33.
Miltner, R. J., S. W. Ross, and M. H. Posey
1995. Influence of food and predation on the depth distri-
bution of juvenile spot [Leiostomus xanthurus) in tidal
nurseries. Can. J. Fish. Aquat. Sci. 52:971-982.
Moser, M. L.
1987. Effects of salinity fluctuation on juvenile estuarine
fish. Ph.D. diss., 150 p. North Carolina State Univ.,
Raleigh, NC.
Moser, M. L., and L. R. Gerry.
1989. Differential effects of salinity changes on two estua-
rine fishes, Leiostomus xanthurus and Micropogonias
undulatus. Estuaries 12:35-41.
Moser, M. L., and W. F. Hettler.
1989. Routine metabolism of juvenile spot, Leiostomus xan-
thurus (Lacepede), as a function of temperature, salinity
and weight. J. Fish Biol. 35:703-707.
Necaise, A. M.
2000. Habitat evaluation as measured through the growth
of juvenile red drum, Sciaenops ocetlatus, and summer
(lounder,Paralichthysdentatus. M.S. thesis, 49 p. North
Carolina State Univ. Raleigh, NC.
Nelson, D..M., M..E. Monaco, E..A. Iriandi, L..R. Settle, and
L. Coston-Clements.
1991. Distribution and abundance of fishes and inverte-
brates in southeast estuaries. ELMR (Estuarine Living
Marine Resources) Rep. 9, 167 p. NOAA/NOS Strategic
Environmental Assessment Division, Rockville, MD.
Neter, J., W. Wasserman, and M. H. Kutner
1985. Applied linear statistical models. Regression, analy-
sis of variance, and experimental design, 1127 p. Irwin.
Homewood, IL.
O'Neil, S. P., and M. P. Weinstein.
1987. Feeding habitats of spot, Leiostomus xanthurus, in
polyhaline versus meso-oligohaline tidal creeks and shoals.
Fish. Bull. 85:785-796.
Patrick, W. S., and M. L. Moser
2001. Potential competition between hybrid striped bass
(Morone saxatilis x M. americana) and striped bass (M.
saxatilis) in the Cape Fear River estuary, North Carolina.
Estuaries 24:425-429.
Peters, D. S., J. C. DeVane Jr., M. T. Boyd, L. C. Clements, and
A. B. Powell.
1978. Preliminary observations on feeding, growth, and
energy budget of larval spot {Leiostomus xanthurus). In
Ann. Rep. Southeast Fish. Cent., Beaufort Lab. to U.S. Dep.
Energy, p. 377-379. Beaufort Laboratory, National Marine
Fisheries Service, Beaufort, NC.
Peterson, M. S., and S. T. Ross.
1991. Dynamics of littoral fishes and decapods along a
coastal river-estuarine gradient. Est. Coast. Shelf Sci.
33: 467^83.
Peterson, M. S., B. H. Comyns, C. F. Rakocinski, and
G. L. Fulling.
1999. Does salinity affect somatic growth in early juvenile
Atlantic croaker, Micropogonias undulatus (L.)? J. Exp.
Man Biol. Ecol. 238:199-207.
Pietrafesa, P. J., and G. S. Janowitz.
1988. Physical oceanographic processes affecting larval
transport around and through North Carolina inlets. Am.
Fish. Soc. Symp. 3:34-50.
Pietrafesa, L. J., G. S. Janowitz, T. Chao, R. H. Wiesberg,
F. Askari and E. Noble.
1986a. The physical oceanography of Pamlico Sound.
Univ North Carolina Sea Grant Publ. UNC-WP-86-5, 125
p. Univ. North Carolina Sea Grant Program, Raleigh, NC.
Pietrafesa, L. J., G. S. Janowitz, J. M. Miller, E. B. Noble,
S. W. Ross, and S. P Epperly
1986b. Abiotic factors influencing the spatial and tempo-
ral variability of juvenile fish in Pamlico Sound, North
Carolina. In Estuarine variability (D. A. Wolfe, ed.), p.
341-353. Academic Press, New York, NY.
Ricker,W. E.
1975. Computation and interpretation of biological statis-
tics of fish populations. Bull. Fish. Res. Board Can. 191,
382 p.
Rogers, S. G., T E. Targett, and S. B. Van Sant.
1984. Fish-nursery use in Georgia salt-marsh estuaries: the
influence of springtime freshwater conditions. Trans. Am.
Fish. Soc. 113:595-606.
Ross. S. W.
1988. Age, growth and mortality of Atlantic croaker in North
Carolina, with comments on population dynamics. Trans.
Am. Fish. Soc. 117:461-473.
1992. Comparisons of population dynamics of juvenile spot
{Leiostomus .xanthurus), Atlantic croaker {Micropogonias
404
Fishery Bulletin 101(2)
undulatus), and Atlantic menhaden {Brevoortia tyrannus)
among diverse North Carolina estuarine nursery areas.
Ph.D. diss.. 144 p. North Carolina State Univ., Raleigh,
NO.
Ross, S. W., and S. P. Epperly.
1985. Utilization of shallow estuarine nursery areas by
fishes in Pamlico Sound and adjacent tributaries. North
Carolina. In Fish community ecology in estuaries and
coastal lagoons (A. Yanez-Arancibia, ed.), Ch. 10, 207-232.
UNAM (Universidad National Autonoma de Mexico) Press,
Mexico.
Ross, S. W., and J. E . Lancaster
2002. Movements and site fidelity of two juvenile fish spe-
cies using surf zone nursery habitats along the southeastern
North Carolina coast. Environ. Biol. Fishes 63:161-172.
Rountree, R. A., and K. W. Able.
1992. Foraging habitats, growth, and temporal patterns
of salt-marsh creek habitat use by young-of-the-year
summer flounder in New Jersey. Trans. Am. Fish. Soc.
121:765-776.
Rozas, L. R and C. T. Hackney
1984. Use of oligohaline marshes by fishes and macrofaunal
crustaceans in North Carolina. Estuaries 7: 213-224.
SAS Institute, Inc.
1988. SAS/STAT user's guide, release 6.03 edition. SAS
Inst., Inc. Cary, NC.
Shapiro, D. Y.
1987. Inferring larval recruitment strategies from the dis-
tributional ecology of settled individuals of a coral reef fish.
Bull. Mar. Sci. 41:289-295.
Siegfried, R. C, II, and M. R Weinstein.
1989. Validation of daily increment deposition in the otoliths
of spoHLeiostomus xanthiiruii). Estuaries 12:180-185.
Sogard, S. M.
1992. Variability in growth rates of juvenile fishes in differ-
ent estuarine habitats. Mar Ecol. Prog. Ser. 85:35-53.
Szedlmayer, S. T.
1991. Distribution and abundance of nearshore fishes in the
Anclote River estuary, west-central Florida. Northeast
Gulf Sci. 12:75-82.
Thresher, R. E.
1985. Distribution, abundance, and reproductive success in
the coral reeffishAcanthochromispolyacanthus. Ecol. 66:
1139-1150.
Warlen, S. M.
1980. Age and growth of larvae and spawning time of Atlan-
tic croaker in North Carolina. Proc. Ann. Conf Southeast
Assoc. Fish. Wildl. Agencies 34:204-214.
Warlen, S. M., and A. J. Chester.
1985. Age, growth, and distribution of larval spot, Leiosto-
mus xanthurus, off North Carolina. Fish. Bull. 83:587-
599.
Weinstein, M. P.
1979. Shallow marsh habitats as primary nurseries for fishes
and shellfish. Cape Fear River, North Carolina. Fish. Bull.
77:339-357.
1982. Commentary: a need for more experimental work in
estuarine fisheries ecology. Northeast Gulf Sci. 5:59-64.
1983. Population dynamics of an estuarine-dependent fish,
the spot (Leiostomus xanthurus), along a tidal creek-sea-
grass meadow coenocline. Can. J. Fish. Aquat. Sci. 40:
1633-1638.
Weinstein, M. P, and S. R O'Neil.
1986. Exchange of marked juvenile spots between adjacent
tidal creeks in the York river estuary, Virginia. Trans. Am.
Fish. Soc. 115:93-97.
Weinstein, M. P., and M. P. Walters.
1981. Growth, survival and production in young-of-the-year
populations of Leiostomus xanthurus Lacepede residing in
tidal creeks. Estuaries 4:185-197.
Weinstein, M. P., S. L. Weiss, and M. F. Walters.
1980. Multiple determinants of community structure in
shallow marsh habitats. Cape Fear River estuary. North
Carolina, USA. Mar Biol. 58:227-243.
Welch, J. M., and B. B. Parker.
1979. Circulation and hydrodynamics of the lower Cape
Fear River, North Carolina. U.S. Dep. Comnier., NOAA
Tech. Rep. NOS 80, 108 p.
Werner, E. E., and J. F Gilliam.
1984. The ontogenetic niche and species interactions in size-
structured populations. Ann. Rev. Ecol. Syst. 15:393-425.
Zar, J. H.
1984. Biostatistical analysis, 2nd ed., 718 p. Prentice Hall,
Inc., Englewood Cliffs, NJ.
405
Abstract— Age and growth estimates
for the winter skate (Leucoraja ocel-
lata) were estimated from vertebral
band counts on 209 fish ranging in size
from 145 to 940 mm total length (TL).
An index of average percent error
(lAPE) of 5.8% suggests that our aging
method represents a precise approach
to the age assessment of L. ocellata.
Marginal increments were significantly
different between months (Kruskal-
Wallis P<0.001) and a distinct trend of
increasing monthly increment growth
began in July. Estimates of von Berta-
lanffy growth parameters suggest that
females attain a slightly larger asymp-
totic TL (L„=1374 mm) than males
(L_=1218 mm) and grow more slowly
(*=0.059 and 0.074, respectively). The
oldest ages obtained for the winter
skate were 19 years for males and 18
years for females, which corresponded
to total lengths of 932 mm and 940 mm,
respectively. The results indicate that
the winter skate exhibits the charac-
teristics that have made other elasmo-
branch populations highly susceptible
to exploitation by commercial fisheries.
Age and growth estimates of the winter skate
(Leucoraja ocellata) In the western Gulf of Maine
James A. Sulikowski
Michael D. Morin
Seung H. Suk
W. Huntting Howell
Zoology Department, Spaulding Hall
University of New Hampshire
46 College Road
Durham, New Hampshire 03824
E mail address (for J A, Sulikowski) |sulikow@hotmail com
Manuscript accepted 17 December 2002.
Manuscript received 31 December 2002
at NMFS Scientific Publications Office.
Fish. Bull. 101:405-413 (2003).
Little is known about the biology of
many elasmobranchs, including impor-
tant parameters such as validated age,
growth, age at maturity, reproductive
cycles and annual fecundity (Frisk et al.,
2001). Difficulty in obtaining samples,
the large size of specimens, their high
mobility, and minor commercial value
are just a few of the problems that make
such studies complicated and in some
respects impractical (Cailliet et al.,
1983; Cailliet et al., 1986). The recent
intensification in commercial fishing
of elasmobranchs (Cailliet et al., 1983;
Brown and Gruber, 1988; Kusher et al.,
1992; Dulvey et al., 2000) has made the
collection of their life history informa-
tion essential to the realistic manage-
ment of their populations (Cailliet et al.,
1983; Ryland and Ajayi, 1984; Dulvey
et al., 2000). Historically, batoids have
been of minimal commercial value
(Otwell and Lanier, 1979; Sosebee,
1998); hence the majority of research
on elasmobranchs has focused on com-
mercially valuable sharks (e.g. Holden,
1977; Natanson et al., 1995; Walmsley-
Hart et al., 1999). According to the
characteristics outlined by Winemiller
and Rose (1992) and the comparative
analyses of Frisk et al. (2001), skates,
like other elasmobranchs, fall into the
category of equilibrium strategists and
as such reach sexual maturity at a late
age, have a low fecundity, and are rela-
tively long-lived. These characteristics,
coupled with fisheries that select for the
removal of large individuals (especially
those over 100 cm total length), make
these particular fish highly susceptible
to overfishing (Hoenig and Gruber,
1990; Dulvey et al., 2000; Frisk et al.,
2001).
Traditionally, skates caught by
ground fishing operations were dis-
carded (Martin and Zorzi, 1993; Jun-
quera and Paz, 1998; Sosebee, 1998).
New and expanding markets for skate
wings have made retention of these fish
commercially more lucrative in recent
years (Sosebee, 1998; New England
Fishery Management Council^. Skate
harvests in the U.S. portion of the
western North Atlantic are currently
unregulated. Moreover, biological infor-
mation on skate life histories is almost
nonexistent (Frisk, 2000). This combi-
nation of factors is believed to have led
to a depletion of common skates (Raja
batis) in the Irish sea (Brander, 1981).
The winter skate (Leucoraja ocellata)
is a large species (total length over 100
cm) of skate of the family Rajidae (Big-
elow and Schroeder, 1953; Robins and
Ray, 1986; New England Fishery Man-
agement Council'). It is endemic to the
inshore waters of the western Atlantic,
from the Newfoundland Banks and the
southern Gulf of St. Lawrence in Can-
ada to North Carolina in the United
States (Bigelow and Schroeder, 1953).
Despite this wide range, little direct
biological data is available for this spe-
cies (Simon and Frank, 1996; Casey and
' New England Fishery Management Coun-
cil. 2001. 2000 Stock assessment and
fishery evaluation (SAFE) report for the
northeast skate complex, 179 p. New
England Fishery Management Council,
50 Water Street, Mill 2 Newburyport, MA
01950.
406
Fishery Bulletin 101(2)
Myers, 1998, Frisk, 2000). Recent assessment studies in
the northeast U.S. (Northeast Fisheries Science Center^),
suggest that the biomass of the winter skate may be below
threshold levels mandated by the Sustainable Fisheries
Act (SFA). To add insight into the life history of this species
and the status of the stock (Simpfendorfer, 1993; Frisk et
al., 2001 ), we estimated age and growth rates of L. ocellata
by interpreting annular counts and marginal increments
on vertebral centra from specimens collected in the western
Gulf of Maine.
Materials and methods
Sampling
A total of 304 winter skates were captured by otter trawl
between November 1999 and May 2001 at locations that
ranged from 1.6 to 32 km off the coast of New Hampshire.
Approximate depths at these locations ranged between 9
and 107 m. Skates were maintained alive on board the
vessel until transport to the University of New Hampshire's
Coastal Marine Laboratory (CML). There, individual fish
were euthanized (0.05g/L bath of MS222). We measured
total length (TL in mm) as a straight line distance from the
tip of the rostrum to the end of the tail, and disc width (DW
in mm) as a straight line distance between the tips of the
widest portion of pectoral fins. Total wet weight (kg) was
also recorded. In order to differentiate between the small,
immature specimens of little skates (Leucoraja erinacea,
a congener species also found in the Gulf of Maine) and
winter skates, rows of teeth in the upper jaw were counted.
Skates whose number of teeth ranged between 72 and 110
per row were identified as L. ocellata and skates whose
number of teeth ranged between 38 and 64 per row were
identified as L. erinacea (Bigelow and Schroeder, 1953). To
reduce any uncertainty in species identification, skates
having between 38 and 71 teeth per row were not used in
this study.
Preparation of vertebral samples
Vertebral samples, taken from above the abdominal cavity,
were removed from 132 females and 98 males, labeled,
and stored frozen. After defrosting, three centra from each
specimen were freed from the vertebral column, stripped
of excess tissue and air dried. Large centra were cut sagit-
tally, while held within a vise, with a DremeF"^' tool fitted
with a mini-saw attachment. Smaller centra were sanded
with a DremeF'^' tool to replicate a sagittal cut. Processed
vertebrae were mounted horizontally on glass microscope
slides and ground with successively finer grits (#180,
#400, #600), of wet-dry sandpaper. Each vertebra was then
remounted and the other side ground to produce a thin (300
micrometer) "hourglass" section.
Counts of annuli
Vertebral sections were viewed through a compound micro-
scope (25-40x) with reflected light (Fig. 1). A growth ring
(annulus) was defined as an opaque and translucent band
pair that traversed the intermedialia and that clearly
extended into the corpus calcareum (Casey et al., 1985;
Brown and Gruber, 1988). The birth mark (age zero) was
defined as the first distinct mark distal to the focus that
coincided with a change in the angle of the corpus calca-
reum (Casey et al, 1985; Wintner and Cliff, 1996).
Three nonconsecutive counts of annuli were made for
the three vertebral sections from each specimen without
prior knowledge of the length of the skate or previous
counts. If the variability between readings was more than
two years, that particular specimen was eliminated from
further analyses. Count reproducibility was estimated by
using the index of average percent error ( lAPE ) described
by Beamish and Fournier (1981):
M/^£=1/A'^(|//?^(|a-,, -X,|/X,)jxl00,
where N = the number of skates aged;
R = the number of readings;
X = the ith age determination of thejth fish; and
X = the average calculated for thejth fish.
An upper limit for the lAPE was arbitrarily set at 15'7( for
each vertebra. Vertebrae with statistically acceptable lAPE
indexes were used for estimation of asymptotic growth
rates (Brown and Gruber, 1988; Cailliet and Tanaka,
1990). The average of the mean counts for all three centra
defined the age estimate for each specimen (Casey et al.,
1985; Wintner and Cliff, 1996).
A von Bertalanffy growth function (VBGF) was fitted
to the data with the following equation (von Bertalanffy,
1938):
where L,
K
to
total length at time t (age in years);
theoretical asymptotic length;
Brody growth constant; and
theoretical age at zero length.
^ Northeast Fisheries Science Center. 1999. 30th northeast
regional stock assessment workshop, 477 p. Northeast Fisher-
ies Science Center, 166 Water Street Woods Hole, MA 02543-
1026.
Growth in length data were analyzed by using FISHPARM,
a computer program for parameter estimation of nonlinear
models with Marquardt's ( 1963) algorithm for least-square
estimation of nonlinear parameters (Prager et al., 1987).
Marginal increment analyses
The annual periodicity of band pair formation was inves-
tigated by using marginal increment analyses (MIA).
Because the annuli in older specimens were closer together,
marginal increments were calculated from five specimens
per month whose centra contained either four or five
annuli. For MIA determination, the distance of the final
opaque band and the penultimate opaque band from the
centrum edge were measured with an ocular micrometer.
Sulikowski et al.: Age and growth estimates of Leucorqa ocellata
407
^m
WM
4 "\
J
w\
/ . '
h_
-Jm
J^
#:
1
OSS-section of a vertebral centn.
= birth mark; Arrows represent
u
r
ale caught
in July and estimated to
Longitudinal ci
be 5 years. BM
w
Figure 1
im from a 422-mm-TL m
age in years.
The marginal increment was calculated as the ratio of the
distance between the last and penultimate bands (Brans-
tetter and Musick, 1994; Cailliet, 1990; Simpfendorfer,
1993; Simpfendorfer, 2000). Mean average increments by
month of capture were plotted to identify trends in band
formation by using a Kruskal-Wallis one-way analysis of
variance on ranks. (Simpfendorfer, 1993; 2000).
Results
Morphological measurements
A total of 230 specimens were used for this study. Males
(;!=98) ranged between 147-932 mm TL, 82-601 mm
DW, and 0.015-6.2 kg. Females (/i = 132) ranged between
145-940 mm TL, 82-635 mm DW, and 0.015-7.5 kg. A
linear relationship existed between the total length, disk
width, and mass relationships for male, female, and the
sexes combined (all r- values were greater than 0.85). Two
skates (one male: TL=147 mm, DW=82 mm, weight=0.015
kg; and one female TL= 145 mm,DW=82 mm, weight=0.015
kg) hatched from egg cases during May 2001 in the CML
after gestating 18 months. One wild male specimen (age-0,
TL=175 mm, DW=100 mm, weight=0.027 kg) was also cap-
tured and incorporated into the results of this study.
the centra) were easily distinguished from complete bands.
Of the 230 processed vertebrae, 209 (91%) were readable.
These 209 vertebrae (males=88;females=121) had annular
count estimates that agreed within two years, resulting in
an TAPE of 5.8%. Mean total length and disk width at age
for male, female, and sexes combined are given in Table 1.
The relationship between TL and centrum diameter was
linear {r~=Q.92\ P<0.05; Fig. 2) and there were no signifi-
cant differences (ANCOVA, P<0.05) between males and
females. Because no significant difference existed for TL
and centrum diameter between the sexes, the data were
combined (Fig. 2).
Marginal increments were averaged from five speci-
mens for each month, except June when skates belonging
to the 4 and 5 year age classes were unavailable. Marginal
increments were significantly different between months
(Kruskal-Wallis P<0.001) and a distinct trend of increas-
ing monthly increment growth began in July (Fig. 3). Maxi-
mum marginal increment measurement occurred in May.
Minimum marginal increment measurement occurred in
July Two recently hatched males (one from the laboratory
(147 mm TL) and one from field collections (175 mm TL))
had opaque zones on the distal edge of their vertebral cen-
tra. Reviewing this information, we suspect that a single
opaque band may be formed annually on the vertebral
centra during June-July in the winter skate.
Vertebral analyses
No difficulty was encountered in estimating the age of L.
ocellata. False bands (bands that do not completely encircle
Age and growth estimates
We assumed that opaque-translucent band pairs were
formed annually, and we fitted von Bertalanffy growth
408
Fishery Bulletin 101 (2)
1000 -
Inlerccpi -0.63 • f^"""^ *
Slope 0.01 , Jt»V*^
800 -
209 - • f%^ •
r 0.92 • •"TC^^« '•
e"
IXT-** '
E. 600 -
• 1 -iP^'
c
*4 *'^* •
IS 400-
jtf'^V^
o
5^r*
1-
^^^9
.^""""^f ^
^""^ %^
200 -
2 4 6 8 10
Centrum diameter (mm)
Figure 2
The relationship of total length (mm) to centrum diameter (mm) for combined
sexes of winter skate.
Table 1
Average total length, TL, and d
ISC width, DW,
at age for winter skates (L
ocellafa) by sex
and combined
sexes. Mean ±1 SEM;
sample sizes
(no. offish in samp
e) are given in
parentheses.
Age
Male TL
Female TL
Sexes combined
Male DW
Female DW
Sexes combined
0
161 (2) ±14
145(1)
156 (3) ±10
93 ±7
81
89 ±7
1
228(1)
—
228(1)
139
—
139
2
264 (5) ±14
268 (4) ±21
266 (9) ±11
153 ±5
158 ±8
155 ±5
3
340 (4) ±20
317 (9) ±9
324 (13) ±9
198 ±6
188 ±6
191 ±6
4
379 (12) ±8
392 (25) ±8
388 (37) ±6
223 ±8
233 ±5
230 ±4
5
435 (4) ±19
429 (16) ±13
430 (20) ±6
264 ±11
259 ±13
260 ±6
6
536 (5) ±13
501 (7) ±16
516(12)±12
338 ±8
310 ±15
322 ±11
7
609(1)
551 (12) ±6
556 (13) ±7
392
342 ±6
346 ±6
8
651(1)
565 (11) ±13
570(12) ±13
401
352 ±10
356 ±10
9
658 (9) ±24
632 (9) ±20
645 (18) ±15
420 ±18
403 ±16
411 ±11
10
690 (12) ±20
704 (8) ±18
696 (20) ±14
441 ±22
447 ±17
444 ±11
11
735 (10) ±17
761 (5) ±22
744 (15) ±14
479 ±24
498 ±20
485 ±14
12
743 (5) ±24
763 (5) ±19
753 (10) ±15
488 ±12
501 ±15
494 ±8
13
830 (3) ±7
772 (3) ±16
801 (6) ±15
495 ±6
506 ±8
500 ±5
14
838 (4) ±10
803 (3) ±23
821 (7) ±13
530 ±9
527 ±24
529 ±10
15
841 (4) ±12
—
841 (4) ±12
541 ±19
—
541 ±13
16
860 (4) ±4
842(1)±0
857 (5) ±5
565 ±16
542
560 ±10
17
921(1)
—
921(1)
579
—
579
18
—
940 (2) ±0
940 (2) ±0
—
623 ±13
623 ±13
19
932(1)
—
932(1)
601
—
601
curves ( VBGC ) to total length-at-age data ( Fig. 4 ). The VBGC
provided a good fit with a low .standard error for males,
females, and both sexes combined (Table 2). The /„ values
(-1.4 to -1.6) compared favorably with gestation rates for
the two skates hatched in captivity (1.5 years) (Table 2). The
von Bertalanffy growth parameters for males, females, and
the sexes combined were similar but k values were higher
for males and sexes combined, than for females.
Sulikowski et al : Age and growth estimates of Leucora/a ocellata
409
1 0-
09-
i
Average marginal increment
o o o o o
1
Y"
/
r-^
1
03-
1
Jan Feb Mar Apr May Jun Jul Aug Sept Oct Hov Dec
Month
Figure 3
Mean monthly marginal increments of opaque bands for L. ocellata
from the Gulf of Maine. Marginal increments were calculated from
five specimens per month whose centra contained either 4 or 5 annuli.
Error bars represent ±1 SEM.
Discussion
The relationship between TL and centrum diameter was
linear and significant, indicating that the centra grew pro-
portionally to skate length for all size classes, and thus this
structure was useful for age analyses (Kusher et al., 1992).
The 5.8% lAPE index suggests that our aging method rep-
resents a precise approach to the age assessment of L. ocel-
lata. Minimal width of the marginal increment for winter
skates captured in May supports the hypothesis of annual
band formation in this species. Moreover, these results
compare favorably to growth cycles in marginal incre-
ments for other skates found in temperate waters whose
vertebral bands are formed annually (Holden and Vince,
1973; Waring, 1984; Natanson, 1993).
Von Bertalanffy parameters, as determined by our study,
suggest that females attain a slightly larger asymptotic TL^
(1374 mm) than males (1218 mm) and grow more slowly
(/f=0.059 and 0.074, respectively). This trend follows a com-
mon pattern in batiods. Holden ( 1977), Waring ( 1984), Ry-
land and Ajayi ( 1984), Brander ( 1981 ), and Walmsley-Hart
et al. (1999) found similar tendencies in several species of
skates, and Martin and Cailliet (1988) found comparable
results in the bat ray (Myliohatis californica).
Our estimates of L„ exceeded those of the largest speci-
mens in our field collections (940 mm for females and 932
mm for males). Nevertheless, data from extensive trawl
surveys in the western Gulf of Maine and the Mid-Atlantic
offshore region spring and autumn bottom trawl surveys
from 1967 to 2000 (Northeast Fisheries Science Center-)
indicated that mean TL did not exceed 1000 mm. Thus,
we suspect that our von Bertalanffy equation produces an
accurate estimation of L,^ for winter skate. Walmsley-Hart
Table 2
Calculated von
Bertalanffy parameters
for male, female.
and combined
sexes of L. ocellata. r~ is
the coefficient of
determination.
Parameter
Male
Female
Combined sexes
LjmmTL)
1218
1374
1314
k (/year )
0.074
0.059
0.064
/,, (year!
-1.418
-1.609
-1.531
r-
0.946
0.939
0.946
SE
0.01
0.01
0.001
n
88
121
209
et al. (1999) overestimated L^ fori?. puUopunctata and sug-
gested that small sample size and rareness of large indi-
viduals were most likely responsible. Because fishing gear
was not biased towards a specific marketable skate size
and because all size classes of L. ocellata were represented,
it is quite possible that the rareness of large individuals led
to the augmented L.^ in combined and individual sexes in
our study. Possibly, a larger sample size of winter skates
would produce significant and divergent results with re-
gard to von Bertalanffy parameters. However, the close fit
of the data to the VBGC for L. ocellata indicates the VBGC
is an appropriate model for this species.
Preliminary estimates of age and growth parameters are
available for winter skate in Canadian waters ( eastern Sco-
tian Shelf) from Simon and Frank ( 1996), who reported the
results of a study conducted at St. Mary's University by R.
Nearing. Combined sexes of winter skates (/2=242) with TL
410
Fishery Bulletin 101(2)
1000 -
A
800 -
,,v>H''^^
▼ f L'^
f >^^
-^^ » ^
600 -
JX » .
400 -
200 ; , y^
I -
1 r 1 r 1 1 1 1 1 1
£ 0 2 4 6 8 10 12 14 16 18 20
c
1 1000-1
B
800-
600-
i-P
400-
^^
200-
(
y^
0 -
0 2 4 6 8 10 12 14 16 18 20
Age (years)
Figures 4
Von BertalanfTy growth curves generated from vertebral data for (A) male, (B)
female, and (C) combined sexes of winter skate (L. oceltata) from the western
Gulf of Maine. Individual VBGC parameters are given in Table .3.
ranging from 120 to 1060 mm and ages from 0 to 16 years
provided von Bertalanffy parameters of L. = 114.1 cm, k -
0.14405, and /„ = 0.00315. However, these data should be
viewed with caution because no lAPE values nor valida-
tion of the annual nature exist for these estimates, and it
is likely that the older specimens had been under-aged by
four or more years (Simon ').
/C values (an estimation of how quickly an animal grows
to LJ were similar for both sexes of winter skate. These
growth rates are commensurate with other skate species of
^ Simon, J. 2001. Personal commun. Bedford Institute of
Oceanography, P.O. Box 1006, Dartmouth, Nova Scotia, Canada
B2Y 4A2.
similar size, but slower than skate species of smaller size
(Table 3). The oldest ages obtained for the winter skate
were 19 and 18 years for males and females, respectively.
These data are in agreement with the assumption that
larger batoids, such as L. ocellata and R. pullopunctata
(Walmsley-Hart et al., 1999) are longer lived and grow
more slowly than smaller species, such as R. erinacea,
which has been aged to 8 years with a k value of 0.352
(Johnson, 1979; Waring, 1984).
Accurate stock assessment data for skates is difficult to
collect in the northeast United States because species are
rarely differentiated in landings information (New England
Fishery Management Council'). Because of this lack of dif-
ferentiation of species in landings, fluctuations in stock size
Sulikowski et al.: Age and growth estimates of Leucora/a ocellata
411
1200 ■
c
1000 •
sngth (mm)
0 i J Jrf '
J^^g i O O Female
- 600 ■
S
o
1-
Q B'^ » Male
400 ■
ll"
200 :
^^
1
'
0 2
4 6,8 10 12 14 16 18
Age (years)
Figures 4 (continued)
20
Table 3
Comparisor
of von Bertalanffy growth parameters
for several skate species.
Scientific name
Sex
L„ (mm)
k t,
(years)
Max age (yr)
Source
Raja rhina
90-
1047 (TL)
0.17
-0.16
13
Zeiner and Wolf 1993
Raja microoceltata
90-
1370 (TL)
0.086
-3.009
9
Ryland and Ajayi, 1984
Raja montagui
90-
978 (TL)
0.152
-1.719
7
Ryland and Ajayi, 1984
Raja erinacea
90-
527 (TL)
0.352
-0.449
8
Waring, 1984
Raja wallacei
90-
422 (DW)
0.26
-0.17
15
Walmsley-Hart et al., 1999
Raja clauata
90-
1050 (TL)
0.215
0.045
10
Brander and Palmer 1985
Raja pullopunctata
o-
771 (DW)
0.05
-2.20
18
Walmsley-Hart et al., 1999
Raja pullopunctata
9
1327 (DW)
0.08
-1.95
14
Walmsley-Hart et al., 1999
Leucoraja ocellata
90-
1314 (TL)
0.064
-1.531
19
This study
will be difficult to detect and successful implementation of
fisheries management plans will remain problematic. Our
study provides some basic age and growth parameters for
the winter skate and it supports the hypothesis that L. ocel-
lata, like other elasmobranchs, require conservative man-
agement because they grow slowly and are susceptible to
overexploitation ( Brander, 1981; Kusher et al., 1992; Zeiner
and Wolf, 1993; Frisk et al., 2001).
Acknowledgments
We thank Captain Joe Jurek of the FV Mystique Lady
for the collection of skates. We also thank Noel Carlson
for maintenance of the fish at the U.N.H. Coastal Marine
Laboratory and Charles Walker for use of his equipment.
This project was supported by a University of New Hamp-
shire Hubbard Endowment Fund and the U.N.H. Center
for Marine Biology.
Literature cited
Beamish, R. J., and D. A. Fournier
1981. A method for comparing the precision of a set of age
determinations. Can. J. Fish. Aquat. Sci. 38:982-983.
Bigelow, H. B., and W. C. Schroeder
1953. Fishes of the Gulf of Maine. Fish. Bull. 53:63-65.
Brander, K.
1981. Disappearance of common skate Rata batis from Irish
Sea. Nature 290 (5801):48-49.
Brander, K.. and D. Palmer
1985. Growth rate of Raja clavata in the Northeast Irish
Sea. J. Cons. Ciem.42(2):125-128.
Branstetter, S., and J. A. Musick.
1984. Age and growth estimates for the sand tiger in the north-
western Atlantic ocean. Trans. Am. Fish. Soc. 123:242-
254.
Brown, C. B., and S. H. Gruber
1988. Age assessment of the lemon shark, Negaprion bre-
virostris, using tetracycline validated vertebral centra.
Copeia 3:747^753.
412
Fishery Bulletin 101(2)
Cailliet, G. M.
1990. Elasmobranch age determination and verification; an
updated review. In Elasmobranchs as living resources:
advances in the biology, ecology, systematics and the status
of the fisheries (H. L. Pratt Jr, S. H. Gruber, and T. Tanuichi
eds), p. 157-165. U.S. Dep. Commer., NOAA Technical
Report, NMFS 90.
Cailliet, G. M., L. K. Martin, D. Kusher. P. Wolf, and B. A. Weldon.
1983. Techniques for enhancing vertebral bands in age esti-
mation of California elasmobranchs. U.S. Dep. Commer.,
NOAA Tech. Report NMFS 8:157-165.
Cailliet, G. M., R. L. Radtke and B. A. Weldon.
1986. Elasmobranch age determination and verification. In
Indo-Pacific fish biology: proceedings of the second interna-
tional conference on Indo-Pacific fishes (T. Uyeno, R. Aral,
T. Taniuchi and K. Matsuura, eds.), p. 345-360. Icthyol.
Soc. Jpn., Tokyo.
Cailliet, G. M., and S. Tanaka.
1990. Recommendations for research needed to better
understand the age and growth of elasmobranches. In
Elasmobranchs as living resources: advances in the biol-
ogy, ecology, systematics and the status of the fisheries (H.
L. Pratt Jr., S. H. Gruber and T. Tanuichi, eds), p. 505-508.
U.S. Dep. Commer., NOAA Technical Report NMFS 90.
Casey, J. G., H. L. Pratt, and C. E. Stillwell.
1985. Age and growth of the sandbar shark (Carcharhinus
plumbeus ) from the western North Atlantic. Can. J. Fish.
Aquat. Sci. 42{5):963-975.
Casey, J. M., and R. A. Myers.
1998. Near extinction of a large widely distributed fish.
Science 28:690-692.
Dulvey N. K., J. D. MetCalfe, J. Glanville, M. G. Pawson, and
J. D. Reynolds.
2000. Fishery stability, local extinctions, and shifts in com-
munity structure in skates. Cons. Biol. 14( 1 ):283-293
Frisk, M. G.
2000. Estimation and analysis of biological parameters in
elasmobranch fishes and the population dynamics of the
little skate Raja erinacea, winter skate R. ocellata and
Barndoor skate R. laevis. Masters thesis, 167 p. Center
for Environmental Science, Univ. Maryland, College Park,
MD.
Frisk, M. G., T. J. Miller, and M. J. Fogarty
2001. Estimation and analysis of biological parameters in
elasmobranch fishes: a comparative life history study. Can.
J. Fish. Aquat. Sci. 58(5);969-981.
Hoenig, J. M., and S. H. Gruber.
1990. Life history patterns in the elasmobranchs: implica-
tions for fisheries management. In Elasmobranchs as
living resources: advances in the biology, ecology, system-
atics and the status of the fisheries (H. L. Pratt Jr, S. H.
Gruber and T. Tanuichi eds), p. 1-16. U.S. Dep. Commer.,
NOAA Technical Report, NMFS 90.
Holden, M. J., and M. R. Vince.
1973. Age validation studies on the centra of Raja clauata
using tetratcycline. J. Cons. Int. Explor Mer 35:13-17.
Holden, M. J.
1977. Elasmobranchs. In Fish population dynamics (J. A.
GuUand, od.i, p. 187-214, J. Wiley and Sons, London.
Johnson. G. F.
1979. The biology of the little skate. Raja erinacea, in Block
Island Sound, Rhode Island. Masters thesis, 119 p. Univ.
Rhode Island. Kingston, RI.
Junquera, S., and X. Paz.
1998. Non-traditional resources: skate fishery and survey
results in Division 3 NO. Sci. Counc. Res. Doc. NAFO, no.
98/26, 6 p. Northwest Atlantic Fisheries Organization,
Dartmouth, Nova Scotia [Canada].
Kusher, D. I., S. E. Smith, and G. M. Cailliet.
1992. Validated age and growth of the leopard shark, T^iakis
semifasciata, with comments on reproduction. Environ.
Biol. Fish. 35:187-203.
Martin L. K., and G. M. Cailliet.
1988. Age and growth determination of the bat ray, Mylioba-
tis californica, in central California. Copeia 3:762-763.
Martin, L., and G. D. Zorzi.
1993. Status and review of the California skate fishery. In
Conservation biology of elasmobranchs (S. Branstetter, ed.),
p. 39-52. U.S. Dep. Commer., NOAA Tech. Report NMFS 115.
Marquardt. D. W.
1963. An algorithm for the least squares estimation of non-
linear parameters. J. Soc. Ind. Appl. Math. 2:431-441.
Natanson, L. J.
1993. Effect of temperature on band deposition in the little
skate. Raja erinacea. Copeia 19931 1 ):199-206.
Natanson, L. J., G. C. Casey, and N. E. Kohler
1995. Age and growth estimates for the dusky shark, Car-
charhinus obscurus. in the western North Atlantic Ocean.
Fish. Bull. 93(1):116-126.
Otwell, W S., and T. C. Lanier.
1979. Utilization of North Carolina skates and rays. Compl.
Rep. N.C. Div. Mar Fish. February, 49 p. North Carolina
Department of Natural Resources and Community Devel-
opment, Morehead City, NC.
Prager, M. H., S. B. Saila, and C. W. Recksiek.
1987. Fishparm: a microcomputer program for parameter
estimation of nonlinear models in fishery science. Tech.
Rep. (87-10):l-37. Old Dominion University Norfolk VA.
Robins, C. R., and G. C. Ray
1986. A field guide to Atlantic coast fishes of North America,
354 p. Houghton Mifflin Company, Boston, MA.
Ryland, J. S., and T. O Ajayi.
1984. Growth and population dynamics of three Raja species
(Batoidei) in Carmarthen Bay, British Isles. J. Cons. Int.
Explor Mer 41:111-120.
Simon, J. E., and K. T. Frank.
1996. Assessment ofthe Division 4VsW Skate Fishery. DFO
Atl. Fish. Res. Doc. 96/105, 6 p. Department of Fisheries
and Oceans, Dartmouth. Canada.
Simpfendorfer, C. A.
1993. Age and growth of the Australian sharpnose shark.
Rhizoprionodon taylori. from north Queensland. Australia.
Environ. Biol. Fi.sh. 36(3):233-241.
2000. Age and growth of the whiskery shark. Furgaleus
macki. from southwestern Australia. Environ. Biol. Fish.
58:335-343.
Sosebee, K.
1998. Skates. Status of fishery resources off the northeast-
ern United States. U.S. Dep. Commer, NOAA Tech. Memo.
NMFS-NE. 115:114-115.
von Bertalanffy, L.
1938. A quantitative theory of organic growth (inquires of
growth laws II). Human Biology 10:181-183.
Walmsley-Hari, S. A., W. H. H. Sauer, and C. D. Buxton.
1999. The biology ofthe skates Raja wallacei and R. puUo-
punctata (Batoidea: Rajidae) on the Agulhas Bank, South
Africa. S. Afr. J. Mar. Sci. 21:165-179.
Waring, G. T.
1984. Age, gi-owth and mortality ofthe little skate off the
northeast coast of the United States. Trans. Am. Fish,
Soc. 113:314-321.
Sulikowski et al : Age and growth estimates of Leucora/a ocellata
413
Winemiller, K. O., and K. A. Rose.
1992. Patterns of life history diversification in North Ameri-
can fishes: imphcation for population regulation. Can. J.
Fish. Aquat. Sci. 49:2196-2218.
Wintner, S. P. and G. Cliff.
1996. Age and growth determination of the blacktip shark,
Carcharhinus limbatus, from the east coast of South Africa.
Fish. Bull. 94:135-144.
Zeiner, S. J., and P G. Wolf
1993. Growth characteristics and estimates of age at matu-
rity of two species of skates iRaja binoculata and Raja
rhina) from Monterey Bay, California. In Conservation
biology of elasmobranches (S. Branstetter, ed.) p. 87-90.
U.S. Dep. Commer, NOAA Technical Report NMFS 115.
414
Abstract— Large (>458 mm) striped
bass {Morone saxatilis) are dominant
predators in Chesapeake Bay. In recent
years, the Chesapeake Bay stock of
striped bass has increased dramatically,
raising concerns about their predatory
impact and their forage requirements.
In response to these concerns and the
need for more recent ecological stud-
ies, this investigation was conducted
to characterize feeding habits of large
striped bass in Chesapeake Bay. Stom-
ach contents from 1225 striped bass
from 458 to 1151 mm TL were exam-
ined in the spring and fall of 1997 and
1998. Striped bass consumed 52 differ-
ent species of vertebrates and inverte-
brates; however, only a few species of
clupeoid and sciaenid fishes dominated
diets across both the seasons and size
ranges of striped bass examined. Of
finfish species, menhaden \Brevoortia
tyrannus) was the dominant prey in
most areas and gizzard shad (Doro-
soma cepedianum) replaced menhaden
in importance in lower salinity waters.
Spot [Leiostomus xanthurus) and
other sciaenid fishes and anadromous
herrings t.Alosa spp.) also contibuted
large percentages of striped bass diet.
Although pelagic schooling fishes
formed the majority of the diet, benthic
fishes contributed a higher percentage
to the diet than in previous studies of
striped bass diet composition.
Diet composition of large striped bass
iMorone saxatilis) in Chesapeake Bay*
John F. Walter III
Herbert M. Austin
Virginia Institute of Marine Science
School of Marine Science
The College of William and Mary
PO Box 1346, Gloucester Point, Virginia 23062
E mail address (for J, F Walter): ifwalterfiivims edu
Manuscript accepted 22 October 2002.
Manuscript received 9 January 2003
at NMFS .Scientific Publications Office.
Fish. Bull. 101:414-423 (2003).
Along the Atlantic coast of North Am-
erica, the striped bass is one of the most
important commercial and recreational
fishes (Richards and Rago, 1999). In
the face of intense overfishing, the
Atlantic Coast population of striped
bass experienced drastic declines in the
1970s (Field, 1997; Richards and Rago,
1999). During these periods of intense
harvesting, smaller fish dominated the
stock composition and the fishery (Koo,
1970). With the relaxation of fishing
pressure and the implementation of
regulations designed to protect older
age classes, populations rebounded
to the point where, currently, large,
older fish comprise a high percentage
of the population (Richards and Rago,
1999). The increased abundance of
large striped bass has raised concerns
over both the predatory impact and
prey needs of this large population of
seasonally abundant species in Chesa-
peake Bay.
Within Chesapeake Bay, historically
a center of striped bass abundance and
one of the largest sources of juvenile
production for the Atlantic coast (Mer-
riman, 1941; Berggren and Lieberman,
1978; Kohlenstein, 1981), striped
bass are seasonally abundant upper
trophic level predators. Chesapeake
Bay striped bass are partitioned into
a resident, primarily male or juvenile,
group of fish found year-round and a
migratory group consisting of older,
larger (>711 mm total length) and of-
ten primarily female fish found in the
spring and fall (Chapman, 1987). The
Atlantic States Marine Fisheries Com-
mission manages fish greater than 711
mm (28 inches) total length as migra-
tory (ASMFC) because the majority of
these fish leave Chesapeake Bay and
migrate throughout the Atlantic coast.
Striped bass within Chesapeake Bay
migrate during the spring when mature
fish ascend tidal freshwater tributaries
to spawn (Chapoton and Sykes, 1961;
Dorazio et al., 1994). After spawning,
these fish leave Chesapeake Bay and
migrate northward along the Atlantic
coast, returning to Chesapeake Bay in
large numbers during the fall. With a
major peak in March-April and a minor
peak in October-November, the histori-
cal landings data reflect the migratory
behavior and seasonal abundance of
larger fish (Koo, 1970).
Diet studies represent the first step
in determining the magnitude and di-
rection of trophic interactions and are
essential data for the management of
both predators and prey (Livingston,
1985). For the management of multi-
species fisheries, detailed information
on fish food habits is required in order
to account for the temporal, spatial,
and ontogenetical nature of trophic
interactions (Walters et al., 1999; Hol-
lowed et al., 2000; Wliipple et al, 2000).
Although the feeding habits of resident
juvenile and early adult striped bass
have received considerable study in
Chesapeake Bay (Hollis, 1952; Markle
* Contribution 2507 of the Virginia Insti-
tute of Marine Science, School of Marine
Science, The College of William and Mary,
Gloucester Point, VA 23062.
• ASMFC (Atlantic States Marine Fisheries
Commission I. 2000. Public informa-
tion document for Amendment 6 to the
Interstate Fishery Management Plan for
striped bass, 17'p. ASMFC, 1444 Eye
Street NW. Washington, DC 20005. http:
//www.jcaa.org/I^ID.htm (March 2001).
Walter and Austin: Diet composition of Morone saxalilis in Chesapeake Bay
415
and Grant, 1970; Setzler et al., 1980;
Boynton et al., 1981; Limburg et al.,
1997; Hartman and Brandt, 1995a)
and in other locations (Schaefer,
1970; Manooch, 1973; Rulifson and
McKenna, 1987), no studies have in-
cluded enough specimens larger than
600 mm total length to adequately
characterize the diet of migratory
fish. The absence of dietary informa-
tion for these larger striped bass may
have been due to the difficulty in
sampling larger striped bass and also
to the relative scarcity of large striped
bass in Chesapeake Bay during times
of severe overfishing (Koo, 1970).
Nevertheless, the absence of diet data
represents a gap in our knowledge of
the trophic dynamics of large striped
bass that form the major portion of the
spawning stock, are prized fisheries
targets and, through successful fish-
eries management, have emerged as
a significant seasonal predatory force
within Chesapeake Bay. We specifi-
cally address the diet composition of
large (458-1151 mm) striped bass in
Chesapeake Bay to determine the im-
portant species in their diet during the
spring and fall periods of abundance.
James River '-.j.^ yi^' ^'
ii^V-V ir, AflanBc Ocean
'4-
100
7S'0O-
100
74'.S0-
200 Kilometers
Figure 1
Map of Chesapeake Bay showing spatial distribution of striped bass samples
from March 1997 to May 1998.
Methods
From March 1997 to May 1998, 1225 striped bass were
collected from various localities in Chesapeake Bay, its
Virginia tributaries, and the Chesapeake Bay mouth
(Fig. 1). Fish were collected from recreational fishermen,
charterboat captains, and seafood dealers, as well as from
scientific monitoring programs in the spring (48.5%) and
fall (51.5%), corresponding to seasonal migration patterns
and fishing seasons. Fish ranged in size from 458 to 1151
mm TL (mean 653.7mm) and were 0.91-17.6 kg in weight
(mean 3.69 kg). Hook-and-line gear, gill nets, fyke nets,
and otter trawls were used to capture fish. Fish captured
in pound nets were excluded from this analysis because
of complications introduced by the confinement of the
fish in pound nets. Fish captured by hook and line were
recorded as such in order that the bait and chum used with
this gear could be excluded from the diet analyses. Total
length (±1.0 mm), sex, and weight (±0.001 kg wet weight)
were recorded for each fish, as well as location, date, and
method of capture. Stomachs were removed by cutting the
alimentary canal anterior to the stomach and posterior to
the pylorus and the contents were frozen until processed.
In some cases, stomachs of fish donated by charterboat
captains and recreational fishermen were removed by the
fisherman.
Fish stomachs were thawed and emptied, and their
contents were blotted dry and weighed. Contents were
sorted and identified to the lowest possible taxon, weighed,
counted, and measured. Diet composition was analyzed by
using three measures described in Hyslop (1980): percent
frequency of occurrence, percent weight, and percent
number These values were combined to give an index
of relative importance (Pinkas et al., 1971). The index of
relative importance for a particular prey category i (IRI,)
is expressed as
IRl^ = (%N+ %W) X %F,
where %N = the percentage of a prey species by number;
%W - the percentage of a prey species by weight;
and
%FO = the percent frequency of occurrence of a prey
species.
IRI values were calculated as percent IRI values (Cortes,
1997 ). In calculating IRI values, we excluded several items
appearing in the stomachs, such as chum (ground men-
haden), bait, trash and plant material because they were
deemed to be non-naturally occurring food items. Several
prey species were combined either because of difficulties
in identification of partially digested prey to species or
because of ecological or taxonomic similarity. Both bay
anchovy iAnchoa mitchilU) and striped anchovy (Anchoa
416
Fishery Bulletin 101(2)
hepsetus) were combined into a single-prey category. In
addition, gizzard and threadfin shad (Dorosoma cepedia-
num and D. petenense), and blueback and alewife herring
(Alosa aestivalis and A. pseudoharengus) were treated as
single-prey categories. Unidentified prey consisted pri-
marily of unidentified fish remains and were recorded as
such.
Striped bass were categorized by fish length and month
of capture. Fish were partitioned into two size classes cor-
responding to mixed resident and migratory fish (458-710
mm total length) and coastal migrant fish (711-1255 mm
total length) based on the Atlantic States Marine Fisher-
ies Service classification of fish 711 mm and above as
fully recruited to the coastal migratory stock. For spatial
analysis of feeding habits, each fish was placed into one of
two salinity regimes; tidal freshwater (0-5 ppt) or meso-
haline waters (6-28 ppt). Tidal freshwater- waters include
the upper reaches of the James, York, Rappahannock, and
Potomac rivers. Mesohaline waters include the open waters
of Chesapeake Bay and the lower reaches of most rivers.
No fish were collected in the fall from tidal freshwater For
both monthly and spatial analyses, diet was quantified by
weight only.
To measure intensity of feeding, a stomach fullness index
(SFI) was calculated according to Hureau ( 1969);
„^, Stomach coiuen! wt'iiilu
Sri = X 10.
Fish weight
SFI values were calculated for all fish regardless of the
presence or absence of stomach contents.
A regression of striped bass total length versus prey total
length was fitted by least-squares linear regression of the
untransformed values. Prey lengths were reconstructed
from partially digested backbones by using regressions of
backbone length on total length obtained from samples col-
lected in 1998 by the authors and those given by Hartman
and Brandt (1995a).
Results
Of the 1225 striped bass examined, 688 (56%) contained
items in the stomachs (Table 1). Thirty-four different spe-
cies offish and 18 species of invertebrates were observed in
the diet. Overall, clupeid fishes dominated the diet and men-
haden, in particular, accounted for 44';'( of the weight and
occurred in 18% of all stomachs (Table 2). Menhaden ranged
in length from 103 to 360 mm total length. A '/r IRI value of
58.3 for menhaden was higher than that for all other species
combined. Anchovies were numerically the most abundant
(229^ ) of all prey items and were equal to spot (Leiostomus
xanthiirus) in '7( IRI, both sharing a value of 12.3. Other prey
in order of decreasing '^r IRI were gizzard shad (genus Doro-
soma ) with a ''/i IRI of 6.7, and blue crab (C'a//(>iectes saptdus}
with %IRI values of 3.4. Atlantic croaker (Micropogonius
undulatus) and summer flounder iParaUchthys dentatus)
had '''( IRl values of 1.1 and 1.0, respectively.
All other prey categories had 'MRI values <1 and ap-
peared relatively unimportant in the overall diet of striped
bass, although some increased in relative importance at
certain times and locations. Invertebrates were relatively
minor constituents of the overall diet of large striped bass,
providing only 4.4% of the total IRI. In contrast, clupeid
fishes contributed 65% of the IRI and both sciaenid and
engraulid fishes combined contributed over 25% of the
total IRI.
Clear seasonal and spatial patterns in diet corresponded
with the migratory behavior of large striped bass. Striped
bass in both sizes classes, 458-710 mm and 711 mm and
above, migrated into tidal freshwater to spawn in the
months of March, April, and May. Striped bass fed in the
tidal freshwater region, although at a reduced intensity
as evidenced by the lower stomach fullness values and
the lower percentages of nonempty stomachs compared
to those at other times and locations (Table 1). Gizzard
shad, white perch (Morone americana), and anadromous
herrings (Alosa pseudoharengus and Alosa aestivalis) were
the main constituents of the diet of both sizes of striped
bass in the tidal freshwater region (Table 3, Fig. 2).
During spring, striped bass also pass through the meso-
haline waters of Chesapeake Bay prior to and after spawn-
ing, during which time they feed fairly heavily as indicated
by higher than average stomach fullness values and per-
centages of nonempty stomachs (Table 1). Approximately
83% of the striped bass sampled from mesohaline waters
during this time had food in the stomachs indicating active
feeding during the pre- and postspawning migration. Men-
haden dominated the diets by weight of both size classes of
striped bass from mesohaline waters in the spring. Striped
bass of both size classes also consumed croaker, blue crab,
and white perch (Table 3, Fig. 2); however, the size classes
differed in that smaller fish consumed bay anchovy and
juvenile spotted hake ( Urophycis regia ) and larger striped
bass consumed anadromous herrings.
Large striped bass are generally absent from Chesapeake
Bay in significant numbers in the summer and return in
the fall to mesohaline waters of Chesapeake Bay and its
lower tributaries. The fall return is essentially a feeding
migration and the high stomach fullness values and high
percentages of nonempty stomachs (Table 1 ) indicate active
feeding. Striped bass of both size classes fed predominantly
upon menhaden, which had percent weight values between
53% and .58 Ve (Fig. 3). Sciaenid fishes, including spot, At-
lantic croaker, and weakfish iCynoscion regalis) combined
provided between 23"^^ and '3l"i of the diet by weight for
both size classes of fish. Notable differences occurred in
the high percentage of summer flounder (Paralichthys
dentatus) found in the diets of larger striped bass (\b%
by weight) and in the high percentage of both butterfish
(Peprilus triacanthus, 4%) and gizzard shad (11%) found
in the diets of smaller fish (Fig. 3). The only invertebrates
found in abundance in the diets during this time were blue
crabs, which contributed 70"^; of the diet by weight for the
smaller size class of striped bass in September (Table 3).
The greatest number of species occurred in the diet in fall
with forty-four different species of prey items observed,
although many were isolated occurrences of rare prey and
only a few species contributed to the overall diet at this
time.
Walter and Austin: Diet composition of Morone saxatilis in Chesapeake Bay
417
Table 1
Distribution of striped bass collections by month
with location, capture method.
percentage of nonempty (% full) stomachs, and |
stomach f
uUness index.
%
Stomach
Standard
Month
Location
Method
Total
full
fullness index
deviation
Striped bass, 458-710 mm total 1
ength
Feb
Potomac River
gill net
14
64.3'7f
1.13
1.73
Mar
York, Rappahannock,
James River
gill net, fyke net
116
47.47r
0.36
0.79
Apr
York, Rappahannock
James River
gill net, fyke net
159
25.2%
0.38
1.52
May
Upper York River
electroshock
28
71.4%
1.15
1.80
Jun
Middle Bay
gill net, hook and line
77
93.5%
4.85
3.87
Sep
Middle Bay
hook and line, gill net
74
27.0%
0.30
0.62
Oct
Lower Bay
hook and line, gill net
245
58.4%
1.06
1.93
Nov
Lower Bay
hook and line, gill net
114
74.6%
2.08
3.08
Dec
Lower Bay
hook and line, gill net, trawl
12
91.7%
1.48
1.24
Striped bass 711-1255 mm total length
Mar
York, Rappahannock
James River
gill net, fyke net
12
50.0%
0.31
0.69
Apr
York, Rappahannock,
James River
gill net, fyke net
85
31.8%
0.60
1.44
Mav
Upper York River
electroshock
7
85.7%
0.82
1.74
Jun
Middle Bay
hook and line
66
81.8%
2.75
2.14
Sep
Middle Bay
hook and line, gill net
20
25.0%
0.21
0.24
Oct
Lower Bay
hook and line, gill net
45
42.2%
0.71
1.52
Nov
Lower Bay
hook and line, gill net
95
74.7%
1.69
2.80
Dec
Lower Bay
hook and line, gill net
56
80.4%
1.23
1.66
Total
all
all
1225
56.1%
1.00
2.03
Table 2
Stomach contents of striped bass from Chesapeake Bay, 1997-98 (n
=688, total number of stomachs with quantified contents).
%
frequency
%by
Weight
%by
Prey
Occurrences
of
occurrence
Number
number
in grams
mass
%IR1
Class Osteichthyes
Clupeidae
Brevoortia tvrannus
132
20.63
319
18.11
14757.03
44.40
58.34
Alosa spp.
7
1.09
20
1.14
977.38
2.94
0.20
Dorosoma spp.
43
6.72
142
8.06
4623.73
13.91
6.68
Unknown clupeid
18
2.81
21
1.19
134.56
0.40
0.20
Moronidae
Morone saxatilis
1
0.16
1
0.06
19.46
0.06
0.00
Morone americana
19
2.97
24
1.36
750.09
2.26
0.49
Sciaenidae
Leiostomus xanthurus
86
13.44
179
10.16
3315.84
9.98
12.25
Bairdielta chrvsoura
13
2.03
17
0.97
244.61
0.74
0.16
Cynoscion regalis
15
2.34
19
1.08
835.62
2.51
0.38
Micropogonias undulatiis
20
3.13
21
1.19
2123.82
6.39
1.07
Unknown scieanid
14
2.19
21
1.19
61.41
0.18
0.14
continued
A significant relationship between striped bass total
length and prey total length (P<0.05, r'-=0.26) was ob-
served which indicated that larger and older striped bass
ate larger prey (Fig. 41. The fit of the regression was poor,
indicating that, although larger striped bass did consume
larger prey, they also consumed smaller prey.
418
Fishery Bulletin 101(2)
Table 2 (continued)
Prey
Occurrences
'7f frequency % by Weight % by
of occurrence Number number in grams mass
%m\
Engraulidae
Anchoa spp.
Other fish
Paralichthys dentatus
Membras martinica
Menidia menidia
Anguilla rostrata
Symphurus plagiusa
Pepritus triacanthus
Urophycis regia
Notropis spp.
Trinectes maculatus
Pomatomus saltatrix
Eucinostomus argenteus
Gobiosoma bosc
Synodus foetens
Strongylura marina
Scophthalmus aquosus
Mugil curema
Sphoeroides maculatus
Hypsoblennius hentzi
Fundulus heteroclitus
Unidentified fish remains
Class Crustacea
Callinectes sapidus
Neomysis americana
Squtlla empusa
Ovalipes ocellatus
Lironeca ovalis
Callinectes spp.
Penaeus setiferus
Crangon septemspinosa
Paleamonetes pugio
Cancer irroratus
Upogebia affinis
Class Bivalvia
Mytilus edulis
Crossostrea virginica
Class Gastropoda
All gastropods
Class Polychaeta
All polychaetes
Class Hydrozoa
All hydroids
Phylum Porifera
All sponges
Miscellaneous items
Chum (ground menhaden)
Bail I menhaden, spot, etc)
74
17
1
12
10
9
6
3
3
5
3
3
1
2
56
55
13
23
13
6
4
5
3
4
1
1
11.56
2.66
0.16
1.88
1.56
1.41
0.94
0.47
0.47
0.78
0.47
0.47
0.16
0.31
0.16
0.16
0.16
0.16
0.16
0.16
8.75
8.59
2.03
3.59
2.03
0.94
0.63
0.78
0.47
0.63
0.16
0.16
0.16
0.63
0.31
0.16
399
30
15
25
21
40
12
26
5
5
3
3
1
2
3
129
90
35
15
6
7
5
5
9
1
1
22.66
1.70
0.85
1.42
1.19
2.27
0.68
1.48
0.28
0.28
0.17
0.17
0.06
0.11
0.17
0.06
0.06
0.06
0.06
0.06
4.03
7.33
5.11
1.99
0.85
0.34
0.40
0.28
0.28
0.51
0.06
0.06
0.06
0.23
0.11
0.06
256.29
0.77
0.39
0.00
7.92
0.02
0.00
0.00
2.29
0.01
12.26
2256.59
6.79
1.02
26.17
0.08
0.01
27.00
0.08
0.13
544.48
1.64
0.20
111.59
0.34
0.17
385.88
1.16
0.08
400.00
1.20
0.06
8.45
0.03
0.01
23.56
0.07
0.01
184.21
0.55
0.02
39.92
0.12
0.01
0.10
0.00
0.00
68.54
0.21
0.00
67.96
0.20
0.00
14.42
0.04
0.00
36.08
0.11
0.00
4.80
0.01
0.00
4.15
0.01
0.00
3.39
0.01
0.00
128.37
0.39
1.75
439.81
1.32
3.36
11.09
0.03
0.47
174.26
0.52
0.41
103.68
0.31
0.11
0.54
0.00
0.01
28.83
0.09
0.01
13.00
0.04
0.01
1.34
0.00
0.01
2.24
0.01
0.01
7.73
0.02
0.00
0.59
0.00
0.00
2.00
«
**
2.00
*
**
159
28
0.00
0.01
0.00
0.00
continued
Walter and Austin: Diet composition of Morone saxatilis in Chesapeake Bay
419
Tidal Freshwater, 458-7 10 mm
(« =43)
menhaden
unk clupeid ,„,
2% ' '»
Tidal Freshwater. 711 mm and above
(n=39)
white perch
Mesohaline, 458-710 mm
(H=70)
gizzard shad
89%
Mesohaline, 711 mm and above
("=12)
river herring
blue crab 40/^
11%
croaker /
19% /
spotted hake \';////,';,';,z:';'
white perch ^s'W4<~^~^^^~^^^~^~^~^^^-^^^>^^^
Figure 2
Percentage by weight of prey in the diets of striped bass captured in the spring (February-June). Note that only
stomachs with contents other than bait were used in the construction of these figures.
Table 2 (continued)
% frequency
% bv
Weight
'Xf by
Prey
Occurrences
of occurrence
Number
number
in grams
mass
%IRI
Miscellaneous items (cont.)
Plant material
11
*
*
*
«
*
**
Woody material
6
*
*
*
«
*
**
Plastic trash
1
*
*
*
*
*
**
Cigarette butts
2
*
*
*
*
*
**
Stones, gravel
2
*
*
*
*
*
**
Feathers
2
*
+
*
*
*
**
* Not quantified.
** Not included in IRI ca
culations.
420
Fishery Bulletin 101(2)
tfl
Hi
3
"35
o
o
O
o
o
O
in
c^
•*
O
o
o
o
O
CO
rH tC
d
d
d
d
d
in
d
d
d
d
d
d
d
d
■^
CM d
_>
1-H
CM
M
„
oj
-i;
d.
o
in
in
o
o
00
en
1-H
i>
o
o
o
o
o
t^
CO CO
a
o
a
d
t^
d
d
d
CO
d
d
d
d
d
d
d
d
in
CO d
cfl
c
w
fN
CM
CO
0)
'=C
C
0;
k^
a.
0)
c
4
1-
r3
CO
[/:
OJ
QJ
N
;-.
CO
u
—
QJ
o
d
oi
q
CN
O
d
o
d
1— (
CO
d
o
o
d
o
d
o
d
o
d
O
d
o
d
CO
CD
CM rH
rH d
a
S
o
CU
>
CO
c
M
00
02
1
u
o
X
CO
o
d
CO
irj
o
o
d
o
d
eg
in
00
q
CO
CO
d
CD
d
o
d
o
d
CO
d
02 in
LO CM
I— '
o
■—I
c^
•— t
^H
-^
>^
CO
oa
o
o
CD
o
o
o
o
q
Tt;
c^
o
o
o
o
o
[r^
00 00
a
d
r-i
d
d
d
d
,-i
t>^
CO
d
d
d
d
d
iri
CD d
OJ
M
(N
CM
CO
a
a
w
-^ ^
to
tt; u
.2 i
a
o
d
o
I— 1
CD
I— 1
1—1
■*
■*
in
o
d
o
d
l-H
d
o
d
o
d
CO
CO
d
CM
CM O
d d
c:
:^
CO
tH
CO
•-H
«
a
03
O
-Q
-a
o
a
C3
'C
«
d.
o
o
CTl
t^
o
o
o
o
o
o
ir;
1-H
t^
o
o
o o
X
O
Q.
d
d
CC
CO
d
d
d
d
d
d
CO
C-^
CO
d
d
d d
^
CO
o
t^
o
c:
^j
X
tx
o
JIX
.ti
o
C35
T)"
o
CD
^
»-H
o
o
•— '
1-H
o
•-*
CM
o
o o
^
^
Q;
d
CO
t^
d
d
d
d
d
d
d
lO
d
i—i
CO
d
d d
^
a
lO
•"*
C^
.— t
CM
c
a*
o
bi
OJ
a
o
^
s
o
d
?— 1
CO
■^
o
o
o
o
CO
o
in
"*
CM
o
O
o
CM O
a
m
c-^
CO
d
d
d
oq
1-H
d
CO
d
CM
d
d
d
CM d
i.
CO
■^
CO
C^i
CM
CJJ
o
CM
'x
c3
o
X
J3
ft
c
c
0}
^
c
_2
0;
■a
CO
CO
■3
ts
o
f-*
o
CT>
!>J
-*
o
05
^
o
o
o
O
(J)
in
t^
rH CO
c
JS
o
S
g
d
(N
d
d
00
d
00
iri
■^
■*-*
d
d
d
1-^
CO
■^
CO CO
s
t^
CO
in
^
6
6
CO
r^
CM
Tf f-
B
a>
o
Ui
!/l
1
i2
rH
r-4
1
i^
rH
I>
CA
CO
s
?
in
o
o
55
s
CO
00
i-H
^H
CO
J3
£
CM
CO
5-
in
5
5
T-H
rH .^5
"V
-a
c
o
a
<
c
a
-4->
O
z
a
■c
CO
a
<
CO
s
c
3
■-3
a
o
> O
♦J
s
55
M
Discussion
Our study addresses the diet of striped bass
above 458 mm total length in Chesapeake Bay.
In previous studies of striped bass diet (Hollis,
1952; Hartman and Brandt, 1995a) in Chesa-
peake Bay and adjacent waters (Manooch,
1973), few fish above 458 mm were sampled.
The comprehensive work by Hartman and
Brandt ( 1995a) did not include fish above age
6. The current study focuses specifically on
the diet of larger striped bass that previously
were undersampled or were rare during peri-
ods of severe overfishing (Koo, 1970).
Throughout the two size ranges of striped
bass sampled and in both seasons and loca-
tions, schooling fishes dominated the diets in
Chesapeake Bay. In particular, clupeid fishes
(menhaden, gizzard shad) and the closely
related anchovies exceeded all other prey spe-
cies in frequency of occurrence, number, and
biomass. Among other fishes, only spot rivaled
the clupeids and anchovies in overall impor-
tance; however, white perch, croaker, weakfish,
and summer flounder contributed important
percentages of the diet in certain seasons. Hol-
lis (1952), Manooch (1973) and Hartman and
Brandt ( 1995a ) and Overton ( 2002 ) also found
that schooling clupeoid fishes formed the ma-
jority of the diets of striped bass from Chesa-
peake Bay and nearby Albemarle Sound.
There was a shift in the relative importance
of smaller schooling fishes (anchovies) in
striped bass 458-710 mm to larger schooling
fishes (menhaden, gizzard shad) in striped
bass 711-1151 mm. Although there was a
tendency for larger striped bass to consume
larger prey, this relationship should more
accurately be described as one where larger
striped bass have a greater size range of prey
to consume (Fig. 4). The largest striped bass
consumed prey ranging from several millime-
ters up to 400 mm in total length, correspond-
ing to 40'^( of their total length and equaling
the ratio of mean maximum forage length to
striped bass length found by Manooch ( 1973).
Similarly, smaller striped bass consumed prey
that approached 40'~f of their total length;
however, most prey consumed by all sizes of
striped bass were smaller, young-of-the-year
fishes — a finding corroborated by Overton
(2002), who predicted an optimal prey size to
be 21% of the striped bass length.
The predominance of fish in adult striped
bass diets attests to the piscivorous nature of
larger striped bass and corroborates the find-
ings of other studies (Hollis, 1952; Manooch,
1973; Overton, 2002). Hartman and Brandt
(1995a) and Gardinier and Hoff (1982) ob-
served an ontogenetic shift at 200 mm TL
Walter and Austin: Diet composition of Morone saxatilis in Chesapeake Bay
421
Mesohaline, 458-710 mm
(n=2.')9)
croakerweakfisti
butterfisri 2% 2%
4%
Mesohaline, 711-1151 mm
(H=14())
eel gizzard shad
weakfish ,%
6%
menhaden
58%
flounder
15%
Figure 3
Percent by weight of prey in the diet of striped bass captured in the fall (September-December). Note that only
stomachs with contents other than bait were used in the construction of these figures.
450 -\
M)0 -
Ql 150
from invertebrate to vertebrate prey in the
diet of smaller striped bass. In the present
study, we sampled size ranges above 458
mm and found no clear ontogenetic dietary
shift between vertebrate and invertebrate
prey. Invertebrates, primarily blue crab,
constituted a minor percentage of the over-
all diet and were significant in the diet only
in May and September in mesohaline wa-
ters of Chesapeake Bay. This is in contrast
to the high percentages of invertebrates
found in the diets of large striped bass in
New England waters and likely represents
latitudinal differences in the availability of
fish prey (Nelson et al.^).
The seasonal and spatial differences in
the diet of striped bass correspond to the
behavioral and seasonal migration pat-
terns of the fish and reflect changes in the
community composition at the location and
time of capture. The major seasonal trend
is spring feeding on gizzard shad, anad-
romous herrings, and white perch, corre-
sponding to spawning migrations of both striped bass and
their prey into tidal freshwater Many spring samples came
from upper river sites where gizzard shad and white perch
are year-round residents and herrings are anadromous mi-
grants (Murdy et al., 1997). This pattern of spring feeding
P<0.05, 1^^0.26
Plot
250 500 750 1000
Striped bass total length (mm)
Figure 4
of prey total length against total length for striped bass.
1250
on anadromous herrings and gizzard shads was also found
by Trent and Hassler 1966) in the Roanoke River, NC.
Striped bass captured in spring from the lower, more sa-
line sections of the rivers exhibited high levels of feeding
intensity and consumed primarily menhaden, sciaenids,
anchovies, and blue (VIMS^) crabs. In the spring, Manooch
2 Nelson, G. A., B. C. Chase and J. Stockwell. 2002. Feeding
habits of striped bass {Morone saxatilis) from coastal waters
of Massachusetts, 29 p. Massachusetts Department of Marine
Fisheries Annisquam River Marine Fisheries Field Station 30
Emerson Ave. Gloucester, MA 019.30.
' VIMS (Virginia Institute of Marine Science). 2002. Juvenile
fish and blue crab trawl survey. VIMS, P. O. Box 1346 Glouces-
ter Point. VA 23062. http://www.fisheries.vims.edu/vimstrawl
data/. (March 20011
422
Fishery Bulletin 101(2)
(1973) found menhaden and anadromous herrings to be
predominant (Homer and Boynton'') foods in brackish wa-
ters of Albemarle Sound and HoUis (1952) found menhaden
as well as anchovies and blue crabs to be predominant food
of striped bass in brackish waters of Chesapeake Bay. The
predatory impact of migratory striped bass depends upon
their residence time in these waters, as well as on striped
bass population size and feeding rates. Carmichael et al.
(1998) estimated that striped bass spend approximately
one week in their upstream and one week in their down-
stream transit of the Roanoke River There are no estimates
of residence time in the open waters of Chesapeake Bay or
Albemarle Sound; however, striped bass larger than 711
mm are captured in recreational fisheries in Chesapeake
Bay into June, suggesting that they are present in Chesa-
peake Bay from March through June.
After leaving Chesapeake Bay and summering in New
England waters, large striped bass return to the bay in fall
(Dorazio et al., 1994) and fed primarily upon menhaden,
spot, and anchovies. At this time, most fish were taken from
open waters of Chesapeake Bay. In the lower bay during
fall, large numbers of transient young-of-the-year (YOY)
marine fishes (menhaden, spot, croaker, flounder, and
weakfish) congregate in open waters of Chesapeake Bay
prior to the fall out-migration, thus making them acces-
sible prey for returning striped bass. Striped bass exhibited
higher stomach fullness values and higher percentages of
nonempty stomachs in November and December than in
all other months, with the exception of June. This finding,
in conjunction with observations of striped bass aggres-
sively pursuing baitfishes in surface waters during the fall
(Mollis, 1952, this study), indicates high feeding intensity.
In bioenergetic simulations, striped bass growth potential
and prey density peaked in October (Brandt and Kirsch,
1993). Because much of the annual growth (Hartman and
Brandt, 1995a, 1995b) and gonadal development (Berlin-
sky and Specker, 1991) occur in the fall, this period is of
primary importance both for the accumulation of body
mass for overwintering and for the initial development of
gonadal products.
Although pelagic fishes, notably anchovy and menhaden,
provided the bulk of the diet for large striped bass, this study
differs from the diet study of Hartman and Brandt (1995a)
and the network analysis of Baird and Ulanowicz (1989) in
that benthic fishes also contributed significantly to the diets.
Baird and Ulanowicz ( 1989) estimated that striped bass ob-
tained 91-100''^( of their diet from pelagic trophic pathways
and Hartman and Brandt (1995a) estimated that 68-75%
of the diet of age-2 to age-6 striped bass came from pelagic
sources. These estimates contrast with the high percentages
of benthic spot, croaker, summer flounder, and gizzard shad
observed in this study and indicate that larger striped bass
either prey to a greater extent upon benthic fishes or the
overall diet has shifted towards benthic prey Menhaden
and bay anchovy juvenile abundance indices have declined
' Homer. M, and W R. Boynton. 1978. Stomach analysis offish
collected in the (lalvert (Miffs region, Chesapeake Bay — 1977.
Rep. UMCEES 78-154 CBL, ,!6,i p. Chesapeake Biological
Laboratory, Univ. Maryland, Solomons, MD.
over the past 10 years (VIMS^) suggesting that a dietary
shift towards benthic prey may have occurred since the
collections of Hartman and Brandt (1995a) and the stud-
ies cited in the Baird and Ulanowicz ( 1989) model. Without
comprehensive and systematic annual diet sampling, it is
difficult to separate dietary shifts from differences in the
sizes of fish sampled or the sampling locations. Baird and
Ulanowicz (1989) incorporated diet composition data from
Hollis (1952), Gardinier and Hoff (1982), Manooch (1973),
and Homer and Boynton'* that included very few striped
bass larger than >600 mm and their model included no link-
ages between striped bass and gizzard shad, spot, croaker,
or summer flounder Furthermore, the absence of gizzard
shad in the Baird and Ulanowicz ( 1989) model represents a
missing pathway that might link benthic detritus directly
to piscivore production as occurs in freshwater impound-
ments where gizzard shad are the major prey of striped
bass (Mathews et al., 1988) and play a pivotal role in the
freshwater ecosystem (Stein et al., 1995).
Acknowledgments
This work represents part of a thesis presented to the Col-
lege of William and Mary (School of Marine Science) by
the first author. We would like to thank the first author's
committee members, David Evans, Robert Diaz, John
Hoenig, and Thomas Munroe, for reviewing the thesis and
this manuscript. We would like to acknowledge the many
seafood dealers and recreational and commercial fisher-
man who provided fish samples. This research was funded
by the Virginia Recreational Fishing Advisory Board and
the Virginia Commercial Advisory Board (grant numbers
RF-97-08 and CF-97-08).
Literature cited
Baird, D., and R. E. Ulanowicz.
1989. The seasonal dynamics of the Chesapeake Bay eco-
system. Eeol. Mono. 59(4):329-364.
Berggren, T. J., and J. T. Lieberman.
1978. Relative contribution of Hudson, Chesapeake and
Roanoke striped bass, Morone saxatilis, stocks to the Atlan-
tic Coast fishery Fish. Bull. 76:335-342.
Berlinsky, D. L. .and J. L. Specker.
1991. Changes in gonadal hormones during oocyte develop-
ment in the striped bass, Morone saxatilis. Fish Phys.
Biochem. 9:51-62.
Boynton, W. R., T. T. Polgar, and H. H. Zion.
1981. Importance of juvenile striped bass food habits in the
Potomac estuary Trans. Am. Fish. Soc. 110:56-63.
Brandt, S. B., and J. Kirsch.
1993. Spatially explicit models of striped bass growth
potential in Chesapeake Bay. Trans. Am. Fish. Soc. 122:
845-869.
Carmichael, J. T, S. L. Haeseker, and J. E. Hightower,
1998. Spawning migration of telemetered striped bass in
the Roanoke River, North Carolina. Trans. Am. Fish. Soc.
127:286-297
Chapman, R. W.
1987. Changes in the population structure of male striped
Walter and Austin: Diet composition of Morone saxatills in Chesapeake Bay
423
bass spawning in three areas of the Chesapeake Bay from
1984 to 1986. Fish. Bull. 85:167-170.
Chapoton, R. B., and J. E. Sykes.
1961. Atlantic coast migration of large striped bass as evi-
denced by fisheries and tagging. Trans. Am. Fish. Soc. 90:
13-20.
Cortes, E.
1997. A critical review of methods of studying fish feed-
ing based on analysis of stomach contents: an applica-
tion to elasmobranch fishes. Can. J. Fish Aquat. Sci. 54:
726-738.
Dorazio R. M., K. A. Hattala, C. B. McCollough, and J. E. Skjeveland.
1994. Tag recovery estimates of migration of striped bass
from spawning areas of the Chesapeake Bay. Trans. Am.
Fish. Soc. 123:950-963.
Field, J. D.
1997. Atlantic striped bass management: Where did we go
right? Fisheries 22:6-9.
Gardinier, M., and T Hoff.
1982. Diet of striped bass in the Hudson River Estuary.
N.Y. Fish Game J. 29:152-165.
Hartman, K. and S. Brandt.
1995a. Trophic resource partitioning, diets and growth of
sympatric estuarine predators. Trans. Am. Fish. Soc. 124:
520-537.
1995b. Predatory demand and impact of striped bass, blue-
fish and weakfish in the Chesapeake Bay: applications
of bioenergetics models. Can. J. Fish. Aquat. Sci. 52:
1667-1687.
Mollis, E. H.
1952. Variations in the feeding habits of the striped bass,
Roccus saxatitis, in Chesapeake Bay. Bull. Bingh. Ocean.
Coll. 14(1):111-131.
Hollowed, A. B., N. Bax, R. Beamish, J. Collie, M. Fogarty,
P. Livingston, J. Pope, and J. C. Rice.
2000. Application of multispecies models in assessment of
impacts of commercial fishing. Symposium proceedings
in the ecosystem effects of fishing. ICES J. Mar Sci. 57:
707-719.
Hureau, J. C.
1969. Biologie comparee de quelques poissons anarctiques
(Nototheniidae). Bull. Inst. Oceangr. Monaco 68:1-44.
Hyslop. E. J.
1980. Stomach content analysis- a review of methods and
their application. J. Fish Biol. 17:411-429.
Kohlenstein, L. C.
1981. On the proportion of the Chesapeake Bay stock of
striped bass that migrates in the coastal fishery. Trans.
Am. Fish. Soc. 110:168-179.
Koo, T S. Y.
1970. The striped bass fishery in the Atlantic states. Ches.
Sci. 11:73-93.
Livingston. P. A.
1985. An ecosystem model evaluation: the importance offish
food habits data. Mar. Fish. Rev. 47:9-12.
Limburg, K. E., M. L Pace, and D. Fischer.
1997. Consumption, selectivity and use of zooplankton by
larval striped bass and white perch in a seasonally pulsed
estuary. Trans. Am. Fish. Soc. 126:607-621.
Manooch, C. S., III.
1973. Food habits of yearling and adult striped bass, Morone
saxatilis. from Albemarle Sound, North Carolina. Ches.
Sci. 14:73-86.
Markle, D., and G. Grant.
1970. The summer food habits of young-of-the-year striped
bass in three Virginia rivers. Ches. Sci. 11:50-54.
Mathews, W. J., L. G. Hill, D. R. Edds, J. J. Hoover, and
T. G. Heger.
1988. Trophic ecology of striped bass, Morone saxatilis, in a
freshwater reservoir (Lake Texoma, USA). J. Fish Biol.:
33:273-288.
Merriman, D.
1941. Studies on the striped bass (Roccus saxatilis) of the
Atlantic Coast. Fish. Bull. 50: 1-77.
Murdy, E. O., R. Birdsong, and J. A. Musick.
1997. Fishes of the Chesapeake Bay, 324 p. Smithsonian
Institute Press, Washington, DC.
Overton, A. S.
2002. Striped bass predator-prey interactions in Chesa-
peake Bay and along the Atlantic Coast. Ph.D. diss., 226 p.
Univ. Maryland Eastern Shore, Princess Anne, MD.
Pinkas, L.
1971. Food habits study. /« Food habits of albacore, bluefin
tuna and bonita in California waters (L. Pinkas, M. S. Oliph-
ant, and I. L. K. Iverson, eds.), p. 47-63. Calif Dep. Fish
Game, Fish. Bull. 152.
Richards, R. A., and P. J. Rago.
1999. A case history of effective fishery management:
Chesapeake Bay striped bass. N. Am. J. Fish. Manage.
19:356-375.
Rulifson, R. A., and S. A. McKenna.
1987. Food of striped bass in the upper Bay of Fundy,
Canada. Trans. Am. Fish. Soc. 116:119-122.
Schaefer. R.
1970. Feeding habits of striped bass from the surf waters of
Long Island. N.Y. Fish Game J. 17:1-17.
Setzler, E. M., W. R. Boynton, K. V. Wood, H. H Zion, L Lubbers,
N. K. Mountford, P. Frere. L. Tucker, J. A. Mihursky.
1980. Synopsis of biological data on striped bass, Morone
saxatilis (Walbaum). U.S. Dep. Commer., NOAA Techni-
cal Report NMFS Circ. 433, 69 p.
Stein, R. A., D. V. DeVries, and J. M. Dettmers.
1995. Food-web regulation by a planktivore: exploring the
generality of the trophic cascade hypothesis. Can. J. Fish
Aquat. Sci. 52: 2518-2526.
Trent, L., and W Hassler
1966. Feeding habits of adult striped bass, Roccus saxati-
lis, in relation to stages of sexual maturity. Ches. Sci. 7:
189-192.
Walters, C, D. Pauly. and V. Christensen.
1999. Ecospace: prediction of mesoscale spatial patterns in
trophic relationships of exploited ecosystems, with empha-
sis on the impacts of marine protected areas. Ecosystems
2:539-554.
Whipple, S. J., J. S. Link, L. P. Garrison, and M. J. Fogarty.
2000. Models of predation and fishing mortality in aquatic
ecosystems. Fish Fisheries l(l):22-40.
424
Abstract— The tautog, Tautoga onitis
(Linnaeus), ranges from Nova Scotia
to South Carolina and has become a
popular target for recreational and com-
mercial fisheries. Although tautog are a
multiple spawning species, reproductive
potential, measured as annual fecun-
dity, has not been estimated previously
with methods (batch fecundity, spawn-
ing frequency) necessary for a species
with indeterminate annual fecundity.
A total of 960 tautog were collected from
the mouth of the Rappahannock River
in the lower Chesapeake Bay to 45 km
offshore of Virginia's coastline to inves-
tigate tautog reproductive biology in the
southern portion of the species range.
Tautog did not exhibit a 1:1 sex ratio;
56'7f were females. Male tautog reached
50^1 maturity at 218 mm TL, females
at 224 mm TL. Tautog spawned from 7
April 1995 to 15 June 1995. at locations
from the York River to 45 km offshore.
Batch fecundity estimates ranged from
2800 to 181,200 eggs per spawning for
female tautog age 3-9, total length 259-
516 mm. Mean batch fecundity ±SEM
for female tautog ages 4-6 was 54,243
±2472 eggs and 106,256 ±3837 eggs for
females ages 7-9. Spawning frequency
was estimated at 1.2 days, resulting in
58 spawning days per female in 1995.
Estimates of potential annual fecun-
dity for tautog ages 3-9 ranged from
160,000 to 10,510,000 eggs.
Reproductive seasonality, fecundity,
and spawning frequency of tautog
{Tautoga onitis) in the lower Chesapeake Bay
and coastal waters of Virginia*
Geoffrey G. White
School of Marine Science
Virginia Institute ol Manne Science
College of William and fVlary
PO. Box 1346
Gloucester Point, Virginia 23062
Present address: Atlantic States Marine Fistienes Commission
1444 Eye Street, NW, 6"^ Floor
Washiington, DC. 20005
E-mail address: gwhiteigasmfc org
Thomas A. Munroe
National Marine Fisheries Service
National Systematics Laboratory. NMFS/NOAA
Smithsonian Institution
Post Office Box 37012
NHB, WC 57, MRC-153
Washington, DC 20013-7012
Herbert M. Austin
School of Manne Science
Virginia Institute of Manne Science
College of William and Mary
PO Box 1346
Gloucester Point, Virginia 23062
Manuscript accepted 22 October 2002.
Manuscript received 31 December at
NMFS Scientific Publications Office.
Fish. Bull.:424-424 (2003).
The tautog, Tautoga onitis (Linnaeus),
ranges from Nova Scotia (Bleakney,
1963; Scott and Scott, 1988) to South
Carolina (Sedberry and Beatty, 1989;
BeardenM, although it is most abun-
dant between Cape Cod and New
Jersey (Bigelow and Schroeder, 1953).
In Virginia, tautog occur within the
Chesapeake Bay from Gwynn's Island
(mouth of Rappahannock River) and
Sandy Point (Eastern Shore) southward
to the mouth of the bay (Hildebrand
and Schroeder, 1928), and in coastal
Atlantic waters out to 65 km offshore
(Richards and Castagna, 1970; Musick,
1972; Hostetter and Munroe, 1993). The
major habitat requirement for this spe-
cies is hard-bottom structure that fish
can remain under, within, or alongside
(011a et al., 1974). Adult tautog inhabit
hard-bottom environments including
natural reefs and rock outcroppings,
as well as man-made structures such
as jetties, bridge-tunnel networks, arti-
ficial reefs, and shipwrecks. Near the
southern terminus of the species range
suitable hard-bottom habitat to support
tautog populations becomes less abun-
dant and may limit population size
(Eklund and Targett, 1990; Hostetter
and Munroe, 1993).
Tautog are a long-lived, slow-growing
species with a maximum recorded age of
34 years in Rhode Island (Cooper, 1967)
Contribution 2505 of the Virginia Institute
of Marine Science, Gloucester Point, VA
23062.
Bearden, C. M. 1961. Common marine
fishes of South Carolina. Bears Bluff
Lab. Conlr, vol 34, 47 p. IDeposited at
South Carolina Department of Natural
Resources, Marine Resources Library,
217 Fort Johnson Road, PO. Box 12559,
Charleston, SC 29422.1
White et al : Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
425
and 31 years in Virginia (White, 1996). Growth parameters
offish between northern and southern regions (Hostetter
and Munroe, 1993) are comparable, except that Virginia
tautog have exhibited almost twice the growth increments
in young-of-the-year and age 13+ fish. Likewise, growth
relationships are similar for tautog from New York (Briggs,
1977) and Virginia (Hostetter and Munroe, 1993).
Within preferred habitats, juvenile and adult tautog
develop home sites (011a et al., 1979). Tagging studies in-
dicate seasonal movements between inshore and offshore
habitats, but minimal north-south movement (Cooper,
1966; Briggs, 1977; Lynch-; Bain and Lucy''). During win-
ter, adult tautog located offshore of Virginia are active at
temperatures above 6.1°C (Adams, 1993). Likewise, tautog
at inshore locations within Chesapeake Bay (Arendt et al.,
2001a, 2001b) remain active at water temperatures of 5°C
or above. In northern parts of its range, adult tautog move
inshore and spawn when water temperatures increase in
the springtime (Chenoweth, 1963; Cooper, 1966; Stolgitis,
1970; 011a et al, 1974, 1979), although some portion of the
population remains offshore year-round (011a and Samet,
1977; Hostetter and Munroe, 1993). Very little is known
about the reproductive biology of tautog in Virginia. Tau-
tog begin spawning when water temperatures reach about
11°C (Chenoweth, 1963; 011a et al., 1974, 1980; Eklund
and Targett, 1990; Hostetter and Munroe, 1993); thus the
spawning season begins later in the spring at higher lati-
tudes. Spawning season extends from mid-April through
June in Virginia (Hostetter and Munroe, 1993), mid-May
through early August in Massachusetts (Stolgitis, 1970),
and from late May to early June in Rhode Island (Che-
noweth, 1963). Macroscopic gonad analyses and gonadoso-
matic indices have indicated that male tautog mature by
age 3 and females by age 3-4 throughout the species range
(Chenoweth, 1963; Cooper, 1967; Stolgitis, 1970; Briggs,
1977; Hostetter and Munroe, 1993); however, sample sizes
of young (age 2-3) fish were small in these studies, and
earlier maturation has been noted (011a and Samet, 1977;
Hostetter and Munroe, 1993). To date there has been no
histological examination of the reproductive biology of this
species in Virginia, or elsewhere.
In laboratory aquaria, tautog have been observed to be
a multiple spawning species, spawning as discrete pairs
and as groups (011a and Samet, 1977; OUa et al., 1977).
Although hermaphroditism is common among labrids
(Warner and Robertson, 1978), tautog are thought to be
strictly gonochoristic (Olla and Samet, 1977). However,
two color patterns of males exist in samples from Virginia
waters; approximately 85% of males exhibit strong dimor-
^ Lynch, T. R. 1996. Marine finfish investigation, sport fish
population survey in Rhode Island marine waters: tautog stud-
ies, 1987-1993. " Rhode Island Division of Fish and Wildlife
performance report project: F-54-R-1, study I-I. reference docu-
ment TT-396, 55 p. Rhode Island Division of Fish and Wildlife,
Marine Fisheries Section, 3 Fort Wetherill Road, Jamestown, RI
02835.
^ Bain.C. M.and J. A. Lucy. 1996. Virginia Game Fish Tagging
Program annual report, 1995. Virginia Mar Res. Rep. 96-2, 10 p.
Virginia Marine Resources Commission, 2600 Washington Ave,
Newport News, VA 23067.
phism and 15% of smaller males (<550 mm) show external
coloration similar to females (Hostetter and Monroe, 1993,
this study). Further, histological analysis of 379 male tau-
tog revealed no evidence of hermaphroditism (Munroe and
White, unpubl. data)
The reproductive potential of tautog, measured as annu-
al fecundity, has not been addressed to date. To accurately
estimate potential annual fecundity for multiple spawning
species, batch fecundity must be multiplied by the number
of spawnings per year, i.e. spawning frequency multiplied
by spawning season length (Hunter and Macewicz, 1985).
Chenoweth (1963) and Stolgitis (1970) estimated batch
fecundity and length of spawning season but did not mea-
sure spawning frequency. The only estimate of number of
spawnings per female per year were those of Olla et al.
(1977), who observed tautog spawning daily for 68-96
consecutive days in laboratory aquaria; however, results
obtained in aquaria studies may not directly apply to natu-
ral habitats and have not been used to estimate potential
annual fecundity.
Our study, the first to investigate cellular aspects of
the reproductive biology of tautog in natural habitats of
the lower Chesapeake Bay, was necessary 1) to describe
this species as a determinate or indeterminate spawner;
2) to describe annual and spawning season ovarian cycles
at the cellular level; 3) to assess age at maturity based on
histological sections of gonad tissue; 4) to estimate batch
fecundity for females from the southern portion of the spe-
cies range; 5) to estimate spawning frequency; and 6) to
estimate potential annual fecundity.
Materials and methods
A total of 960 tautog (>150 mm total length [TL]) were
collected opportunistically between April 1994 and Sep-
tember 1995 from commercial and recreational fishermen
(ca. 909c ) as well as from research projects (ca. 10% ) at Vir-
ginia Institute of Marine Science (VIMS). A minimum size
of 150 mm TL was selected based on maturity information
presented in Hostetter and Munroe ( 1993). Collection loca-
tions ranged from Gwynn's Island at the mouth of the Rap-
pahannock River to 45 km offshore of Virginia's coastline,
at depths of 1-35 m (Fig. 1). Approximately one quarter of
the fish were taken from within the Chesapeake Bay, one
quarter from the Chesapeake Bay Bridge Tunnel (CBBT,
depth=5-15 m), and one half from around the Chesapeake
Bay Light Tower (24 km offshore, depth= 17-20 m).
For each fish, total length (mm) and total weight (TW. g)
were measured. Sex was assessed externally from sev-
eral dimorphic characters previously described by Cooper
(1967), Olla and Samet (1977), and Hostetter and Munroe
(1993). Males were distinguished by their pronounced
white chin, blunt forehead, solid black or gray coloration
on the upper half of the body and white underneath, and
a small white spot (about 15 mm diameter) mid-laterally,
below the middle of the dorsal fin. Female tautog have a
less pronounced chin, sloped forehead, and a mottled brown
coloration. After determination of sex (external examina-
tion), gonads were excised, staged macroscopically, and
426
Fishery Bulletin 101(2)
Figure 1
Map of lower Chesapeake Bay and nearby coastal waters of Virginia. Closed circles indicate
sites where tautog were collected; open circles indicate collection sites of tautog in spawning
condition. CBBT = Chesapeake Bay Bridge Tunnel. CELT = Chesapeake Bay Light Tower.
weighed to the nearest 0.01 g (GW). Maturity classification
was assigned as outlined in Table 1, based on eight macro-
scopic stages, modified from Lowerre-Barbieri et al. ( 1996)
for multiple spawning species. One gonad was randomly
chosen by coin toss for histological processing and placed in
Davidson's fixative. For females staged macroscopically as
spawning females, the remaining ovary was placed in 10%
neutrally buffered formalin for batch fecundity counts.
Whole unsectioned opercle bones are the accepted
method to age tautog (Cooper, 1967; Simpson, 1989;Hostet-
ter and Munroe, 1993). Opercle bones were removed and
processed to examine age at maturity and age-related
fecundity. Opercles were boiled for 1-3 minutes to remove
flesh, scrubbed under warm flowing water, dried for two
days, and read with transmitted light. Age of each fish
was determined from two readings of both opercles (when
possible). An annulus was defined as the transition from
a translucent zone to an opaque zone. Annulus formation
was previously validated by Hostetter and Munroe ( 1993).
1 April was used as a birth date to allow maximum growth
within the biological year (April to March), and to avoid
overlap with fish in the next year class.
Gonads selected for histological processing were placed
in Davidson's fixative for two days before transverse sec-
tions of anterior, middle, and posterior ovarian tissue (or
anterior and posterior sections of testes) were taken and
placed in tissue cassettes. Variation between left and right
gonads was accounted for by random selection of one go-
nad for fixation. Tissue samples were then rinsed overnight
with flowing tap water and placed in TO'^r EtOH. Standard
histological processing (tissue embedded in paraffin, sec-
tioned at 5-7 pm, and stained with Harris's hematoxylin
and eosin-Y) (Luna, 1968) was performed for all samples.
Male gonads were classified microscopically into two
stages: sexually mature or immature. Female microscopic
gonad stages were assigned based on the occurrence and
relative abundance of seven oocyte developmental stages
(Wallace and Selman, 1981; West, 1990; Hunter et al.,
1992): primary growth, cortical alveoli, partially yolked,
advanced yolked, germinal vesicle migration, germinal ves-
icle breakdown, and hydrated oocytes. Final oocyte matura-
tion (FOM) comprises germinal vesicle migration, germinal
vesicle breakdown, and hydrated oocyte stages (Wallace
and Selman, 1981; West, 1990). Fully developed ovaries
were distinguished from partially spent/redeveloping
ovaries by the presence of postovulatory follicles (POFs).
Microscopic gonad stages are described in "Description of
microscopic gonad stages," ("Results" section), summarized
While et al : Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
All
Table 1
Description of macroscopic and microscopic gonad stages (modified from Lowerre-Barbieri et al.,1996) for female tautog. Mac-
roscopic criteria refer to whole fresh ovaries. Gonad stages 4, 5, and 3a comprise the inner spawning cycle. FOM = final oocyte
maturation. GSI = gonadosomatic index. GVBD = germinal vesicle breakdown. POF = postovulatory follicle. MA = macrophage
aggregate.
Gonad stage
Macroscopic criteria
Microscopic criteria
1 Immature
2 Developing
3 Fully
developed
4 Hydrated
5 Running ripe
3a
Partially
spent/
redeveloping
Spent/
regressing
7 Resting
Ovaries very small, tubular in shape, white to light
pink in color; no oocytes visible (mean GSI=0.50)
Ovaries small to medium, tubular shape, dark yel-
low to light orange in color; yolked (opaque) oocytes
begin to appear (mean GSI=2.25 )
Ovaries medium to large, appear slightly grainy,
pale mustard in color; yolked oocytes are abundant
(mean GSI=3.25 )
Ovaries large to very large, pink to orange in color;
firm, yolked oocytes interspersed with large trans-
parent (hydrated) oocytes (mean GSI=11.74 )
Ovaries large to very large; few transparent oocytes
in ovarian tissue, transparent oocytes have been
ovulated into expanded lumen, and are easily ex-
truded when gonad is excised (mean GSI=10.12)
Ovaries somewhat flaccid, large, slightly more pink
than in hydrated stage; lumen has collapsed, occa-
sionally a few remnant hydrated oocytes extruded
from excised ovary (mean GSI=8.84); similar to
stage 3.
Ovaries flaccid, small to medium, red to purple;
some tissue devoid of yolked oocytes at anterior
end of ovary, yolked oocytes visible but less abun-
dant (mean GSI=1. 37)
Ovaries small, purple-opaque to maroon in color;
few or no yolked (opaque) oocytes visible (mean
GSI=1.50)
Oogonia and primary growth oocytes present; high pro-
portion of connective tissue, no atresia or MAs, ovarian
membrane thinner than in resting stage.
Primary growth, cortical alveoli, and some partially
yolked oocytes present.
Primary growth to advanced yolked oocytes present; no
FOM stages, POFs, or remnant HOs.
Primary growth to germinal vesicle migration (GVMl
and hydrated oocytes present, hydrated oocytes are
unovulated; 1-day POFs may be present.
Primary growth through GVM, and ovulated hydrated
oocytes and fresh POFs present; lumen usually seen as
separation of ovigerous folds.
Primary growth through GVBD oocytes present, no
unovulated hydrated oocytes, few remnant ovulated
hydrated oocytes; lumen collapsed, POFs abundant.
Primary growth through advanced yolked oocytes pres-
ent; major atresia of all stages except primary growth
oocytes.
Primary growth and cortical alveoli oocytes present,
occasional atretic oocytes; MAs abundant, more oogonia
tissue, less connective tissue, and thicker ovarian mem-
brane than in immature stage.
in Table 1, and shown in Figure 2, A-H. Percent agreement
between macroscopic and microscopic female gonad stages
was calculated to evaluate the accuracy of macroscopic
staging (used in all previous studies of tautog reproduc-
tive biology). Microscopic stages were assumed to be more
accurate because histologic examination provides evidence
of differences in cellular development.
Chi-square analysis (n=489 fish) was used to test for
significant deviations from an expected 1:1 sex ratio for all
fish. Deviations from a 1:1 sex ratio among 100-mm length
intervals were also analyzed by chi-square to determine if
size or age had a significant effect on sex ratio.
Length and age at maturity were analyzed for fish col-
lected from April to rnid-June to reduce the possibility of
classifying resting, mature fish as immature. Females were
considered mature if classified into microscopic stages 2-7
(Table 1). Males were considered mature if spermatoc3ftes
or spermatozoa were present in histological sections.
Length at maturity was based on 110 females and 79
males (150-350 mm TL). A logistic regression curve was
fitted to the data, to estimate length at 50% maturity (Lgg).
Age at maturity was based on 135 females and 104 males
(ages 1-6).
To determine the annual spawning season, a gonadoso-
matic index (GSl=(gonad weight I somatic weight) x 100)
was calculated by using somatic weight [S'W=TW-GW)
for each sex. A more precise estimate of tautog spawning
season was determined from microscopic gonad stages. The
spawning season was defined by the first and last day that
female tautog. were collected with ovaries staged as either
hydrated, running ripe, or partially spent/redeveloping.
Spawning locations were detected by the presence of hy-
428
Fishery Bulletin 101(2)
Figure 2, A-D.
Histologic stages of tautog ovaries. (A) immature (stage 1). (B) developing (stage 2). (C) fully
developed (stage 3). (Dl hydrated (stage 4). PG = primary growth oocyte. CA = cortical alveoli
oocyte. PY = partially yolked oocyte. AY = advanced yolked oocyte. HO = hydrated oocytes.
Magnification = 50x.
drated and running ripe fish because those stages have
high abundances of hydrated oocytes and fresh POFs,
which indicate cither imminent or recent spawning activ-
ity (Hunter and Macewicz, 1985).
Oocyte development patterns (synchronous, group syn-
chronous, or asynchronous) and type of annual fecundity
(indeterminate or determinate) were assessed by oocyte
size-frequency distributions (Hunter and Macewicz, 1985)
and histology (Hunter and Macewicz, 1985; West, 1990).
Six fish were selected for analysis of oocyte size-frequency
distributions; three fish (TL=300, 400, 450 mm [±10 mm))
in April and another throe in June, representing gonad de-
velopment early and late in the spawning season. For each
fish, oocytes were hydraulically separated from the ovar-
ian membrane and each other, collected in a 0.1-mm sieve,
and preserved in 2% formalin following the method of Low-
erre-Barbieri and Barbieri ( 1993). Preserved samples were
stirred to reduce bias due to differential settling of different
White et al.: Reproductive seasonality, fecundity, and spawning frequency of Toutoga onills
429
i
fift^
fe^/^L.'^'.
Figure 2, E-H.
Histological stages of tautog ovaries (continued). (E) running ripe (stage 5), frame 1 shows
obvious lumen as distinguishing character, frame 2 shows typical field of view without lumen.
Hydrated oocytes have been ovulated (OHO) and now reside in the lumen (L). (F) partially
spent/redeveloping (stage 3a). (G) spent (stage 6). (H) resting (stage 7). PG = primary growth
oocyte. CA = cortical alveoli oocyte. AY = advanced yolked oocyte. GVM = germinal vesicle migra-
tion oocyte. GVBD = germinal vesicle breakdown oocyte. POF = postovulatory follicle. AO =
atretic oocyte. MA = macrophage aggregate. Magnification = 50x.
stage oocytes, and a 5-mL aliquot was removed and placed
in a gridded petri dish. Grids were selected for counting
by using a random number table, and maximum diam-
eters of the first 500 oocytes encountered were measured
to the nearest 0.001 mm with a Biosonic Optical Pattern
Recognition System.® Data were grouped in 0.05-mm size
classes for presentation (0.1 mm group=0.075 to 0.124 mm
oocytes).
Batch fecundity was determined gravimetrically by
using a modification of the hydrated oocyte method (Hunt-
er etal., 1985). The method calls for both ovaries to be fixed
in IC/f formalin, but we had a formalin wet weight for only
430
Fishery Bulletin 101(2)
one ovarian lobe. Therefore, we conducted a calibration
experiment to determine the percent change in ovarian
weight between fresh- and formalin-fixed ovaries. On 25
April 1996, 18 female tautog were collected, fresh ovarian
weight was measured to the nearest 0.01 g, and both ova-
ries were placed in 10% neutrally buffered formalin. For-
malin-fixed wet weight was measured to the nearest 0.01 g
six times over 30 days to determine when weight stabilized
after fixation. Percent change in weight was calculated for
each specimen, and regressed against fresh weight of the
ovary, thus percent change in weight between formalin-
fixed wet weight and fresh ovary weight was calculated
with the negative exponential relationship:
Percent weight change = 21.452 e'" oie^cwi
where GW = fresh gonad weight.
=0.671;
Calibrated (formalin fixed) gonad weight (CGW) was cal-
culated as
CGW = percent weight change x GW.
Then, batch fecundity was estimated by using the formula
Y = iy/x)CGW,
where Y= batch fecundity;
y = number of hydrated oocjrtes in the tissue sample;
X = formalin wet weight of tissue sample; and
CGW = calibrated formalin-fixed wet weight of ovaries.
Assumptions of the hydrated oocyte method which must
be met include 1) all eggs in the most advanced mode are
spawned; 2) fecundity is directly proportional to ovary
weight; and 3) no bias exists in the estimation of egg
abundance within the most advanced mode, in the selection
of mature females for analysis, or in the position within
and between ovaries from which subsamples were taken
(Hunter and Goldberg, 1980; Hunter et al., 1985). The use
of hydrated oocytes, which are much larger than the next
largest cell size class and are formed only when spawning
is imminent, supported acceptance of these assumptions.
Following the methods of Hunter et al. (1985), we selected
ovaries from 29 females for batch fecundity analysis. These
were the only females that had stage-4 (hydrated. Table
1) ovaries without postovulatory follicles as confirmed
through histological analysis. If postovulatory follicles
were found in histological sections, then that fish was
excluded from fecundity analysis.
To test for differential oocyte development between ante-
rior, middle, and posterior sections of ovarian tissue, point
counting analyses ( Weibel et. al., 1966) were performed on
histological sections to determine the relative volume of
seven coll types and POFs in the ovary. The relative vol-
ume of each cell type was calculated by using the number
of points within a grid (121 point.s/grid) overlying each cell
type:
V=FJF
where V^, = relative volume of one cell type;
P„ = number ofpoints overlying a specific cell type;
and
P,„i= number of points in grid.
To ensure that fields of view were chosen randomly,
each ovarian section was divided into 5x5 mm areas with
an overlay grid. Three areas per section were chosen with
a random number table to ensure that counting fields of
view did not overlap. Within each 5x5 mm area, point
counts were made through a gridded reticule ( 121 points)
at 4x magnification. Average relative volume of each cell
class was calculated from the three areas as P,/363, and
compared between anterior, middle, and posterior ovarian
sections with multiple analysis of variance (MAN OVA)
(Minitab, 1995). Response variables (8) were the average
relative volume of each cell class. Differences between
fish (10) were removed by blocking on fish. After no posi-
tional effects were detected ( Wilk's test value 0.25, F=1.35,
df= 16,22, P=0.25), it was concluded that oocyte develop-
ment was evenly distributed throughout the ovaries of
tautog. Blocking by fish proved beneficial and effective in
removing any artifact caused by differences in kill time
between fish, and in increasing the quality of the test by
increasing sample size. Nonsignificant positional effects in
ovarian development allowed estimation of batch fecundity
from only the middle ovarian section. All hydrated oocytes
were counted from three subsamples of approximately 0.3 g
from the middle of the formalin-fixed ovary.
Simple linear regressions were used to describe relation-
ships between batch fecundity and TL, TW, and age. Rela-
tive fecundity was calculated as batch fecundity divided by
GW, and regressed against TL, TW. and age.
Diel spawning periodicity estimates for tautog at the
mouth of the Chesapeake Bay indicated that spawning
occurs during daylight hours but that spawning windows
shift with ebb tidal currents (White, unpubl. data). To esti-
mate spawning frequency by the hydrated oocyte method
(DeMartini and Fountain, 1981; Hunter and Macewicz,
1985), samples with known kill times must be collected just
prior to, and during, the spawning window. Most samples
were collected at dockside; thus kill time for individual fish
was unknown, and the hydrated oocyte method could not be
performed. Therefore, spawning frequency was estimated
by the POF method (Hunter and Goldberg, 1980; Hunter
and Macewicz, 1985) by using descriptions of fresh (day 0,
0-12 h) and degenerating (day 1, 12-24 h) POFs of tautog
(White, unpubl. data). Fresh POFs in tautog ovaries can be
identified as a clearly defined, loosely folded ribbons of the-
cal and granulosa cells that contain visible luniina. similar
to "0 day" POFs in anchovies ( Hunter and Macewicz, 1985 ).
One-day-old tautog POFs have deteriorated such that in-
dividual cell walls are no longer apparent in thecal and
granulosa cells, and the structure appears less organized
and has a small or indistinguishable lumen similar to that
of 24-48 h anchovy POFs (Hunter and Macewicz, 1985). A
full description of POF degeneration in tautog ovaries is
presented elsewhere (White, unpubl. manscr).
Annual fecundity was estimated as the number of
spawnings per female multiplied by batch fecundity for
White et aL: Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
431
Table 2
Percent agreement between
microscopic and
macroscopic
ovarian stages assigned to tautog
(n=484)
captured from June
1994
through Septen
iber 1995. See Table 1 for gonad stage descriptions. Data are
expressed as number of ovaries
staged.
Microscopic
Macroscop
ic gonad stage
stage
1
2
3
4
5
3a
6
7
1
34
—
—
—
—
3
2
3
15
—
—
—
—
—
40
3
4
5
3
18
11
3
31
38
1
8
2
—
5
1
3a
1
—
4
55
3
11
1
—
6
1
—
—
—
—
1
6
—
7
12
3
—
—
—
—
38
132
Agreement
637f
42'y,
69%
24%
67%
79%
13%
75';;
each fish. Number of spawnings per female was calculated
by dividing the number of days in the spawning season by
estimated annual spawning frequency. The relationship
between mean annual fecundity per 50-mm length inter-
val and total length was analyzed with both linear and
exponential regression.
Results
Description of microscopic gonad stages
Tautog ovarian development was described by eight micro-
scopic gonad stages (Table 1) characteristic of multiple
spawning species. Each stage can be differentiated by
a unique suite of histological characteristics. Immature
ovaries (Fig. 2A) are characterized by the presence of only
oogonia and primary growth oocytes within a thin ovar-
ian membrane and a relatively high volume of connective
tissue. Developing stage ovaries (Fig. 2B ) are characterized
by the presence of primary growth, cortical alveoli, and
partially yolked oocytes. The fully developed ovary (Fig.
2C ) is characterized by the presence of primary growth to
advanced yolked ooc3rtes and the absence of oocytes in final
oocyte maturation (FOM) classes, POFs, or remnant HOs.
Hydrated ovaries (Fig. 2D) are distinguished by the promi-
nence of hydrated oocytes still inside the ovarian follicles
and may also contain degenerating POFs from an earlier
spawning, but they noticeably lack oocytes in the germinal
vesicle breakdown state (GVBD). The running ripe stage
(Fig. 2E) is classified by the presence of an expanded ovar-
ian lumen, ovulated hydrated oocytes free in the lumen
(although hydrated oocytes are frequently washed out of
the sample during the staining procedure), a large number
of fresh POFs, and germinal vessicle migration (GVM)
oocytes that tend to be the most advanced stage present
within ovigerous folds. Partially spent/redeveloping ovaries
(Fig. 2F) are classified by the lack of an ovarian lumen and
presence of occasional remnant hydrated oocytes, primary
growth to GVBD oocytes, and an abundance of POFs. The
spent stage (Fig. 2G) is characterized by resorption of
yolked oocytes (atresia), and sometimes the presence of
macrophage aggregates (MAs), which are groups of cells
containing the pigments lipofuscin, ceroid, and melanin
(Wolke, 1992). These cells appear to be a collection of
scavenging cells that remove cellular debris and foreign
substances by phagocytosis when stimulated by excessive
degenerating tissue (Wolke, 1992). In tautog ovaries, MAs
are assumed to be associated with the resorption of yolked
oocytes after the spawning season. Resting stage ovaries
(Fig. 2H) contain primary growth and cortical alveoli
oocytes, and atresia may be present. Resting ovaries can be
distinguished from immature stage ovaries by a thickened
ovarian membrane, relatively more oogonia than connec-
tive tissue and the presence of MAs.
The reliability of macroscopic staging to predict actual
reproductive stage as detected by microscopic analysis was
examined for 484 females. We considered any level of agree-
ment above 80% to be acceptable. The ability of macroscopic
staging to predict actual microscopic stage varied consider-
ably, from 13Vr to 79% agreement for individual stages (Table
2), indicating that macroscopic staging was generally unreli-
able for estimating actual reproductive stage for many ovar-
ian stages. Percent agreement between macroscopic and mi-
croscopic staging was only 51% overall. The best agreement
was for the fully developed, partially spent/redeveloping,
and resting ovarian stages (agreement 69%, 79%, 75%,
respectively). Intermediate values were obtained for im-
mature, and running ripe stages, (agreement 63% and 67%,
respectively). The poorest agreement occurred in assigning
developing and spent stages (42%, 13%, respectively).
Sex ratios
Of the 938 tautog sexed, 522 (56%) were females and 416
were males. Overall, sex ratios varied significantly from an
expected 1:1 ratio, with more females than males (1.25:1,
^2=11.98, P<0.01). At lengths <400 mm, females were
432
Fishery Bulletin 101(2)
A Male tautog (n=79)
(10 -
B Female tautog (n=100)
•
•^*—
MO ^
^r
fn -
y= 100/(l+(38987e'""'""'"''"-') r = 0,88
«
f
70
7
60
/
50
t
'
40
/
30 ■
/
20 -
I /
10 ■
^y
H — ' — ^
' 1
0 -
' ' ' ' 1 ' ' ' ' 1 ' ' ' ' 1
50
150 200 250
Total length (mm)
350
Figure 3
Length at maturity for (A) male and (B) female tautog collected during the 1995
spawning season in lower Chesapeake Bay and nearby coastal waters of Virginia.
Curves represent the logistic regression line.
more abundant than males, whereas the sex ratios were
not significantly different from 1:1 at fish lengths >400 mm
(Table 3).
Length and age at maturity
Tautog length at 509f maturity (^r,o' was 218 mm for males
(/!=79) and 224 mm for females (« = 110, Fig. 3). All males
and females were mature at 300 mm. No females less than
227 mm had hydrated oocytes or POFs that would have
indicated spawning activity.
Age at first maturity was defined as the age at which
at least 50% of the fish are mature. Mature gonads were
present in Wc of males at age 1, 38% at age 2, and 93% at
age 3. Zero percent of females were mature at age 2, 78%
at age 3, and >97% at age 4.
Spawning season and location
GSI values indicated that tautog spawned from April
through June and that peak values occurred in April for the
1995 spawning season (Fig. 4). The 1995 spawning season
was more precisely defined as 7 April-15 June based on
the presence of females in spawning condition (i.e. staged
histologically as hydrated, running ripe). At the beginning
of the spawning .season, females progressed into spawning
condition over approximately two weeks. The end of the
spawning season was determined conservatively, based
Wh
Ite et al : Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
433
lu
o
-Female {n=476)
■Male In =^79)
Jun
Month
Figure 4
Gonadosomatic index based on somatic weight (total weight-gonad weight) for tautog col-
lected in lower Chesapeake Bay during 1994-95. ?i=sample size; error bars indicate sample
standard deviation.
Table 3
Sex ratio of tautog by 100-mm length intervals
at P=0.01. NS = nonsignificant.
and chi-square va
ues
of tests for a 1:1 ratio.
*Significance atP=
=0.05.
**Significance
Total length (mm)
No.
of males
No
of femal
es
No
expected (50%)
% Females
Chi-square
101-00
12
24
18
67
4.00*
201-300
57
88
73
61
6.63*
301-400
147
216
182
60
13.12**
401-00
140
125
133
47
0.85 NS
501-600
38
53
46
58
2.47 NS
601-700
11
14
13
56
0.36 NS
Total
405
520
on the last day female tautog were collected in spawning
condition (15 June) instead of the first collection date of a
spent female (27 June).
Tautog were collected (Fig. 1) in spawning condition
within the Chesapeake Bay (York River, Buckroe Beach),
at the mouth of the Chesapeake Bay (Chesapeake Bay
Bridge Tunnel, Cape Henry wrecks, Anglo- African wreck),
and at offshore locations (Chesapeake Bay Light Tower,
one site 45 km offshore) and there was no apparent trend
in spawning season by location.
Ovarian developmental pattern and type of fecundity
Ovarian stages defined for tautog (Table 1) are typical of
multiple spawning species. Tautog hydrate and spawn
only a small fraction of the yolked oocytes in the ovary
for any one spawning event. Macroscopically, hydrated
ovaries appear speckled because of the intermittent occur-
rence of large, clear hydrated oocytes among the dominant
numbers of opaque, yolked oocytes. Further, the lumen of
running ripe ovaries was full of ovulated hydrated oocytes,
yet there was still a large volume of tissue with maturing
yolked oocytes. The occurrence of spawning stage ovaries
over a protracted period also suggested a multiple spawn-
ing pattern. Tautog were collected in spawning condition
(hydrated and running ripe stages) and the partially spent/
redeveloping stage throughout the April-.June spawning
period, but no spent or resting fish were collected until late
June, suggesting that individual females were spawning
repeatedly during the spawning season.
434
Fishery Bulletin 101(2)
1(K)
80
60
40
20
0
Total length = 300 mm
D April 1995
■ June 1995
-J4--t-
Total length = 400 mm
D April 1995
■ June 1995
Oocyte diameter (mm)
Figure 5
Oocyte size frequency distributions (0.05-mm intervals) in tautog ovaries. Six fish are
represented, two in each length class: (A) 300 mm, (B) 400 mm, (C) 450 mm. Collection
dates were 26 April 1995 and 2 June 1995. Fish were staged macroscopically as hydrated
or running ripe, and oocytes were hydraulically separated and the diameter of 500 oocytes
was measured to the nearest 0.001 mm.
Oocyte size-frequency distributions measured at the begin-
ning and end of the spawning season were also used to classi-
fy tautog annual fecundity as determinate or indeterminate.
Primary growth and cortical alveoli oocytes were continu-
ously yolked, matured, and spawned throughout the spawn-
ing season, evidenced by 1 ) lack of hiatus between advanced
yolked oocytes and less mature oocytes, and 2) abundance
of yolked oocytes (size range 0.30-0.55 mm) not decreasing
over the spawning season (Fig. 5). This type of development
defines tautog as having indeterminate fecundity.
Two patterns of oocyte development are common among
multiple spawning fishes: group synchronous and asyn-
chronous oocyte development. Tautog oocyte size-frequency
distributions (Fig. 5) show no distinct gaps in develop-
White
et al ; Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
435
ment, or modes of oocjrtes, except for
hydrated oocytes (0.95 to 1.25 mm) — a
finding that indicates asynchronous
oocyte development. The presence of
primary growth through advanced
yolked oocytes, as well as oocytes in
FOM and POFs in histologic sections
of fully developed and partially spent
ovaries (Fig. 2, C and F), is evidence
that tautog exhibit a multiple spawn-
ing pattern.
Ovarian cycle
( 1 ) Immature
1)
Sexual Maturity
f7^ (7) Rest
ing
(2) Developing
(6) Spent/
Regressing
Egg maturation in tautog is a complex
process comprising both seasonal and
annual components. This complex pat-
tern of multiple spawning exhibited by
an inner spawning cycle (made up of
hydration, ovulation, spawning, and
redevelopment) within the annual
ovarian developmental cycle, is sum-
marized for tautog in Virginia in Figure
6. In the spring, fully developed ovaries
contain primary growth to advanced
yolked oocytes, but lack POFs. Fish enter the spawning cycle
by hydration and ovulation of the first batch of oocytes. After
the first spawning event, partially spent^redeveloping ova-
ries contain fresh POFs (indicating recent spawning during
the previous 24 hours) and another batch of oocytes in FOM
in preparation for the next spawning event. Thereafter, the
inner spawning cycle is repeated throughout the spawning
season. Histological examination of hydrated ovaries during
the spawning season revealed the co-occurrence of hydrated
oocytes (indicating an imminent spawn) and degenerating
POFs, suggesting that some tautog are capable of repeating
the inner spawning cycle on a daily basis. At the end of the
spawning season, ovaries progress to the spent-regressing
stage, where, through the process of oocyte atresia, the
remaining stock of yolked oocytes are resorbed before the
ovary enters the resting stage.
Batch fecundity
Batch fecundity was determined for 29 female tautog rang-
ing in total length from 260 to 520 mm, total weight 475 to
3,500 g, and ages 3-9 (Fig. 7). Although there was a high
degree of variation in batch fecundity between individual
fish, significant relationships were found between batch
fecundity and fish length, weight, and age. Batch fecundity
was more closely related to total length and total weight
than to age. Batch fecundity (BF) increased significantly
with total length (ANOVA, n=29, F=16.92, P<0.0005,
power=0.97), following the regression equation (Fig. 7A)
f'^:^ (3a) Partially spent/
Redeveloping
(3) Fully
Developed
FOM
Ovulation
(5) Running Ripe
(4) Hydrated
^g
Figure 6
Annual and spawning season cycles of ovarian development in tautog. FOM = final
oocyte maturation.
BF = 56,066 Ln(TW) - 322,091
=0.501.
Batch fecundity also increased significantly with age
(ANOVA, n=29, F=10.22, P<0.004, power=0.88), following
the regression equation ( Fig. 7C )
BF = 15,73 1(AG£) - 20,599
[r2=0.271.
BF = 425.76(rL)- 84,534
|r2=0.39].
Batch fecundity increased significantly with total weight
(ANOVA, n=29, F=16.80, P<0.0005, power=0.99), the
regression equation being (Fig. 7B)
Tautog relative fecundity (BF/GW) did not increase sig-
nificantly with fish length (ANOVA, n=29, F=1.98, P=0.17)
or age (ANOVA, n=29, F=1.72, P=0.20), but there was a
significant increase in relative fecundity with total fish
weight (ANOVA, n=29, P=4.46, P=0.044).
Spawning frequency
Histological examination of 169 tautog collected from 7
April 1995 to 15 June 1995 revealed some variation in the
abundance of three reproductive states that were indica-
tive of imminent or recent spawning. Forty-four percent of
female tautog had HOs, 32'y^ had fresh POFs without any
HOs, and 84'2 of females collected had 1-day-old POFs
(Table 4). Tautog spawning frequency was calculated as
1.2 days based on the percentage of fish with 1-day-old
POFs following the methods of Hunter and Goldberg
(1980). Number of spawnings per female tautog in 1995
was calculated as the spawning season (70 days) divided
by spawning frequency (1.2 days/spawning), yielding 58
spawnings per female.
Individual tautog in natural habitats were capable of
spawning daily after entering the spawning season. Evi-
dence of daily spawning was provided by the rapid ovarian
development observed in histological sections: 1) 70 fish
with HOs and degenerating POFs: 2) 90 fish with both
fresh and degenerating POFs: and 3) partially spent/
436
Fishery Bulletin 101(2)
^uo.uou
:^
L
•
150.000
HKi.om
1
•
• • •
^-<^
50.000
•
••
••
H — ' —
•
' — . — 1 — . — • — . — . — (-
•
•
BF = 425 76(TLl- 84.534 r'
= 0.39
0 -
_j — , — , — , — 1 — , — , — , — . — j — , — ^
— ' — ' — 1
250
300
350 400 450
Total length (mm)
550
..'i)(l.illK.I -
F B
•
150.000
•
•
•«
100.000
• •
•
50.(KI0
4
0^
•
•
1 ' ' '
■-1-^
•
•
•
BF =
= 56.066Ln(TW)- 322.091
r'
= 0,50
() -
1 — ^
' ' 1
J — , — , — . — i — . — , — , — , — 1 — ^ — , — ^_
^-1-^
' ' ' 1
0 50(.) ia)0 I5(K) :0(H) 2500 .3000 3500 400(1
Total weight (g)
00.(J00 ^
[C
1
•
50.000 -
(W.OOO
•
•
•
• •
50.000
•
— \
•
•
\
•
•
1
•
s
1-
•
BF =
15.73l(Agel- 20.599 r'
= 0 27
0 J
— 1 1 1 —
1
23456789 10
Age (years)
Figure 7
Tautog batch fecundity (no. of eggs) versus (A) total length (mm), (B) total
weight (g) and (C) age (years) for 29 tautog collected in lower Chesapeake
Bay and nearby coastal waters of Virginia in 1995. (Complete data for indi-
vidual fish are available in White, 1996).
redeveloping females with both GVBD oocytes and fresh
POFs (Fig. 2F).
Potential annual fecundity
Annual fecundity was calculated as 58 spawnings/female
multiplied by batch fecundity. Annual fecundity varied
from 160,000 eggs (259 mm, age-3 fish), to 10,510,000 eggs
(511 mm, age-9 fish). Mean annual fecundity increased
significantly (ANOVA, ;;=5, F=16.69, P=0.015) with fish
size. A linear regression of mean annual fecundity on fish
length (50 mm size classes, Fig. 8), was described by the
following relationship:
Mean AF = 23,480(TL) - (5x106)
[r2 = 0.811.
Discussion
Macroscopic and microscopic gonad staging
Macroscopic ovarian staging of multiple spawning fishes
can be difficult because subtle differences at the cellular
level may not be detectable macroscopically (Parrish et
al., 1986). However, macroscopic analysis does provide a
rapid estimate of maturity and results in a general descrip-
tion of spawning seasons at a reduced cost compared to
White et al.: Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
437
0,00E+00
300
350 400 450
Total lengtfi (mm)
500
550
Figure 8
Linear relationship between mean annual fecundity (MAF) and body length in 50-
mm length classes for tautog. Sample sizes are noted above each data point. Error
bars indicate standard error of the mean.
Table 4
Reproductive state of female tautog collected during 1995 spawning
season. HO =
= hydrated oocyte.
POF =
= postovulatory follicle.
No. with fresh
No. with
No. mature,
Total no. of
Date (1995)
No.
with HOs
POFs, no HOs
1-day POFs
not spawning
mature females
7 April
11
0
11
1
13
8 April
11
4
12
13
29
15 April
1
0
1
0
1
22 April
9
2
10
4
14
26 April
12
23
47
1
49
9 May
15
16
33
2
35
24 May
0
0
1
0
1
31 May
5
6
11
0
11
1 June
10
33
14
0
14
2 June
1
0
1
0
1
15 June
0
0
1
0
1
Total
75
54
142
21
169
Total %
44.4
32.0
84.0
—
—
time-consuming histological methods. West (1990) noted
that there have been few attempts to assess the accuracy
of macroscopic gonad staging with histological analysis.
Most scientists attempting to assess reproductive stage of
female tautog will most likely use macroscopic criteria.
Given that all eight microscopic stages cannot be identi-
fied in a macroscopic context, a revised macroscopic gonad
staging was developed (Table 5) with validated agreement
against microscopic analysis. The validated stages (Table 5 )
generally agree with previous studies of tautog reproduc-
tive biology (Chenoweth, 1963; Stolgitis, 1970; Hostetter
and Munroe, 1993). However, even with this revised, sim-
plified staging criteria, we caution others that agreement
between the new criteria and microscopic staging is still
low for some ovarian stages, rendering this method less
reliable than microscopic analysis.
Despite these limitations, macroscopic staging errors were
usually offby only one developmental stage (Table 2 ). Errors
in macroscopic staging were most likely due to the rapid de-
velopment of ovarian tissue required to sustain daily spawn-
ing events. The low percent agreement (SI** overall I between
macroscopic and microscopic classifications of tautog ovar-
438
Fishery Bulletin 101(2)
Table 5
Revised macroscopic gonad stages for future research
on tautog. These revised macroscopic stages are based on the gonad stages
seen in Table 1.
Gonad stage (Table 1)
Revised gonad stage
Macroscopic appearance
Immature
Immature
Ovaries very small, tubular in shape, white to light pink in color, no oocytes
visible. (Same as "immature stage" from Table 1.)
Developing
Developing
Ovaries medium to large with slightly grainy appearance, pale mustard
and
in color, yolked (opaque) oocytes present, no hydrated (transparent)
Fully developed
oocytes visible through ovarian membrane. (This stage is a combination of
"developing" and "fully developed" stages from Table 1.)
Hydrated
Spawning
Ovaries large to very large, pink to orange in color, may be dotted with
and
transparent oocytes, yolked oocytes interspersed with large transparent
Partially spent/
(hydrated) oocytes, occasionally a few remnant hydrated oocytes. (This
redeveloping
stage is a combination of "hydrated" and "partially spent/redeveloping"
stages from Table 1.)
Running ripe
Running ripe
Ovaries large to very large, pink to orange in color, hydrated oocytes have
been ovulated, expand lumen of ovary and are easily extruded from excised
ovary; few hydrated oocytes in ovarian tissue. (Same as "running ripe"
stage from Table 1 . )
Spent and Resting
Spent
Ovaries flaccid, small to medium, red to purple in color, few yolked (opaque)
oocytes visible, some or all of ovary having no oocytes visible. (This stage is
a combination of "spent" and "resting" stages from Table 1.)
ian stages is similar to that of studies on Lutjanus vittus,
which found the accuracy of macroscopic staging for ripe
gonads to be only 61% (CSIRO data, cited in West, 1990).
Macroscopic staging of tautog ovaries functioned to de-
scribe the annual gonad cycle, yet it did not separate the
fully developed (stage-3) and partially spent (stage-3a)
ovaries. Macroscopic analysis does not yield proof of POFs,
atretic oocytes, and macrophage aggregates — cellular
structures that help distinguish fully developed, partially
spent/redeveloping, spent, and resting females. Thus, mac-
roscopic analysis could not provide evidence of multiple
spawning in tautog. Histological techniques used in this
study were necessary to accurately describe the annual
cycle and the inner (multiple spawning) spawning cycle
of ovarian development for tautog. Histological staging
also permitted identification of fully hydrated ovaries that
could then be used for batch fecundity estimation. Further,
histology slides were used for point-counting analyses to
test for positional differences in development between an-
terior, middle, and posterior regions within the ovary.
Sex ratios
Sex ratios vary greatly among published studies on tautog
life history. This variability may be due to true differences
in the composition of local populations, or it may be an arti-
fact of sampling strategies rooted in collection seasons or
gear biases. In our study, sex ratios were skewed towards
females for fish under 400 mm and did not differ signifi-
cantly from a l.l ratio for tautog greater than 400 mm.
Collections were primarily made by book-and-line angling
throughout the year, although sample sizes were low
between July and September. Hostetter and Munroe ( 1993)
found no significant difference in sex ratios for fish less than
200 mm, but significantly more males than females for fish
between 201-500 mm in Virginia. Their sampling occurred
over a period of seven years, and fish were collected primar-
ily with fish traps and hook and line. Eklund and Targett
(1990) found a female-to-male sex ratio of 0.86:1 in the
trap fishery between April and December 1987. Chenoweth
(1963) collected more females than males with an otter
trawl at three stations in Narragansett Bay, RI, between
May and September 1961. Factors that affect sex ratios of
tautog from fishery-dependent collections are still unknown
and provide an opportunity for further research into the sex
ratios and reproductive success of this species.
Length and age at maturity
Published reports of tautog length and age at maturity
(from studies with macroscopic techniques and GSI) are
similar for the entire species range. Estimates of tautog
lengths and ages at maturity in our study were similar
to results reported by Hostetter and Munroe (1993) for
tautog captured off Virginia. Hostetter and Munroe ( 1993)
reported that both sexes show evidence of gonadal matu-
ration at age 3 in Virginia. Likewise, age and length at
maturity for tautog collected in Massachusetts (Stolgitis,
1970) are also similar; 40% of age-2 (149-175 mm) males
and 87% of age-3 (171-239 mm) males were mature, and
females attained 71% maturity at age 3 (187-206 mm) and
100% maturity at age 4. In Rhode Island, Cooper (1966)
White el al.: Reproductive seasonality, fecundity, and spawning frequency of Tautoga onttis
439
found that males matured at 200 mm (age 3) and females
at 190 mm (age 3).
For a small number of fish sampled in northern areas,
Hostetter and Munroe (1993) suggested that precocious
development may be occurring in tautog as a response to
fishing pressure. The smallest females collected in spawning
condition have been 227 mm in Virginia (this study), 261
mm in Massachusetts (Stolgitis, 1970), 216 mm (Chenoweth,
1963) and 180 mm (Hostetter and Munroe, 1993), in Rhode
Island. A definitive answer on precocious development is not
possible at this time because data on small fish are limited
in all studies. Detailed histological analysis should be per-
formed on tautog from 100 to 250 mm TL to discern maturity
schedules for specimens in this size range.
Spawning season and location
Tautog spawn over at least a two-month period throughout
the species range, and the initiation of spawning activity
occurs later in the spring to early summer in more north-
ern regions (Chenoweth, 1963; Stolgitis, 1970; Briggs,
1977; Hostetter and Munroe, 1993). In our study, spawn-
ing occurred from 7 April through 15 June 1995 (70 days),
similar to the time interval reported by Hostetter and
Munroe ( 1993) for tautog in Virginia. In New York waters,
tautog have been recorded to spawn for four months (early
May through early September; Austin, 1973). In Rhode
Island, tautog spawn from early June through late July
(Chenoweth, 1963), and spawning seasons as long as
three months (mid-May through early August) have been
reported for fish in Massachusetts (Stolgitis, 1970). Abun-
dance of tautog eggs in plankton collections also shows that
the spawning season occurs progressively later in more
northern regions (Sogard et al., 1992). The earlier spawn-
ing season in Virginia has been attributed to differences in
water temperature (Hostetter and Munroe, 1993). Increas-
ing water temperature during springtime is a major cue to
initiate spawning, but termination of spawning activity has
not been related to environmental cues. However, Austin
(1973) suggested that the effective spawning season may
be shorter than the season of egg release for this species,
based on a decrease in larval abundance as water tempera-
ture exceeded 21.0°C in Long Island Sound.
Tautog were collected in spawning condition within the
Chesapeake Bay and as far as 56 km offshore in this study
and by Hostetter and Munroe (1993). Eklund and Targett
(1990) sampled tautog in spawning condition 22-37 km
off the coast of Maryland and Virginia. Field observations
of daily movements showed that tautog exhibit fidelity
to a home site which they return to each night (011a et
al., 1974, 1975), suggesting that tautog remain at one
location throughout the spawning season. Arendt et al.
(2001b) found that tautog tended to move between sites
during the winter and early spring as the spawning sea-
son began and remained at a single site throughout the
summer. Tagging studies indicate that discrete spawning
groups exist at sites in Narragansett Bay (Cooper, 1966);
however, movements between sites were not quantified.
Sufficient data are not available to determine if tautog ex-
hibit spawning-site fidelity throughout the spawning sea-
son, or if multiple spawning sites are used within general
inshore and offshore classifications. It is generally believed
that most tautog migrate inshore in the spring to spawn
(Cooper, 1966) and some portion of the population remains
offshore year round (Eklund and Targett, 1990; Hostetter
and Munroe, 1993). Although we have documented adult
spawning activity at both inshore and offshore locations,
spawning success in these areas, as well as larval drift and
recruitment patterns, are unknown at this time.
Spawning pattern and type of fecundity
Histological analysis of ovarian tissue supports the clas-
sification of tautog as a multiple spawning species with
a complex reproductive cycle. The complexity of ovarian
maturation (Fig. 6) in this species has not been recognized
in previous studies on its reproductive biology. The typical
cycle of female development for multiple spawning species
is defined by eight microscopic gonad stages (Lowerre-Bar-
bieri et al., 1996) which include an annual cycle (5 stages)
and an inner spawning cycle (3 stages). Although tautog
have been observed to be multiple spawners in laboratory
aquaria (Olla and Samet, 1977; Olla et al, 1977), we define
oocyte development and type of fecundity using recently
improved methods (Lowerre-Barbieri and Barbieri, 1993)
and histological techniques on fish taken from natural envi-
ronments; therefore, they are directly comparable to other
studies of reproductive biology without artifacts associated
with aquarium conditions. Analysis of oocyte size-frequency
distributions and histological sections of ovarian tissue
indicates that tautog have asynchronous oocyte develop-
ment and indeterminate annual fecundity. Therefore,
counting the number of oocytes in the ovary prior to the
spawning season is inadequate to measure potential annual
fecundity because new batches of eggs continuously mature
from primary growth oocytes through hydrated oocytes and
are released during spawning events (Hunter et al., 1985).
Chenoweth ( 1963) analyzed oocyte size-frequency distribu-
tions of three tautog collected over the course of the spawn-
ing season in Rhode Island and noted that the number of
mature yolked oocytes did not decline through the spawn-
ing season. He suggested that not all yolked oocvtes were
spawned and that some portion remained in the ovary and
were resorbed after the spawning season. This observation
is consistent with asjoichronous oocyte development.
Fecundity
This is the first study on tautog reproduction for which
potential annual fecundity has been estimated by multiply-
ing batch fecundity by spawning frequency. Batch fecundity
was more closely related to total length and total weight
than to age. This result makes sense when one considers
the extreme variability in length at age exhibited by tautog
(Cooper, 1967; Hostetter and Munroe, 1993). Batch fecun-
dity ranged from 2800 eggs to 181,200 eggs in 29 females
age 3-9 (Fig. 7). The oldest tautogs collected in this study
were a 31-year-old male and a 17-year-old female. After
reaching maturity, individual females may spawn up to 58
times a year for at least 14 years.
440
Fishery Bulletin 101(2)
Table 6
Comparison of batch fecundity estimates for tautog from three different studies. Mean batch fecundity was calculated for two age
groups, 4-6 and 7-9, because all females were mature by age 4 and over 90% of all tautog sampled in White (1996) were less than
10 years old.
Mean batch fecundity ±SEM
Study
Age 4-6 n
Age 7-9 n
Chenoweth (1963)' Rhode Island
Stolgitis ( 1970)' Massachusetts
White (1996)' Virginia
49,967 ±1032 29
46,833 ±4500 6
54,243 ±2472 18
103,214 ±4005 14
117,478 ±2488 23
106,256 ±3837 10
' Mean batch fecundity for age groups 4—6 and 7-9 was calculated from the raw data presented in
the reference.
Previous estimates of tautog fecundity by Chenoweth
(1963) and Stolgitis (1970) were not annual fecundity es-
timates (Table 6). They counted mature, transparent eggs
in the ovary, currently referred to as hydrated oocytes,
but they did not distinguish tautog as having indetermi-
nate annual fecundity and had no measure of spawning
frequency. By counting only the hydrated oocytes, these
investigators actually estimated batch fecundity. However,
it is interesting to note the similarity of batch fecundity
estimates (Table 6) over the period of 30 years between
studies and wide geographic areas, i.e. from Chesapeake
Bay to Narragansett Bay (550 km).
Spawning frequency had not been previously calculated
for tautog with methods developed by Hunter and Macewicz
( 1985). Although the hydrated oocyte method is less expen-
sive, it requires collection of females just prior to spawning.
With the hook-and-line collection method, it is difficult to
collect sufficient samples in a short period of time. Therefore
spawning frequency was estimated in our study by using
the POF method to read histologic preparations of ovarian
tissue. A female spawning every 1.2 days over the 70-day
spawning window would spawn on an estimated 58 days in
1995. Under artificial conditions, Olla et al. ( 1977) observed
tautog spawning on 68-96 consecutive days in laboratory
aquaria. Therefore, an estimate of 58 spawning days in nat-
ural habitats is not unrealistic. Chenoweth (1963) raised,
but could not answer, the question of whether individual
tautog spawn throughout the entire spawning season. The
spawning-frequency estimate presented here, and observa-
tions of tautog spawning on 68-96 consecutive days in labo-
ratory aquaria (Olla et al, 1977), indicate that tautog are
capable of spawning daily throughout the spawning season
in natural habitats under appropriate environmental condi-
tions (temperature, day length, etc.).
Estimates of potential annual fecundity for Virginia
tautog age 3-9 ranged from 160,000 to 10,510,000 eggs.
However, net annual fecundity may be lower because of
remnant hydrated oocytes, atresia, nutritional status
of adult females, or environmental conditions (McEvoy
and McEvoy, 1992). Based on our samples (females age
3-9), a linear regression provided the most predictive
power (r2=0.81) to estimate mean annual fecundity (Fig.
8). Although female tautog live to be \7+ years old, it is
estimated that 90% of tautog in Virginia waters are age
10 or younger (Hostetter and Munroe, 1993). Therefore, as
the data range in this study is similar to the age structure
of the resource, we suggest that the regression equation
(Mean AF=23,480(TL) - (SxlQi^)) is the most appropriate
formula for use by fishery managers for estimating annual
fecundity of tautog in the southern portion of its range.
We have made a theoretical comparison of potential an-
nual fecundity for tautog (ages 4-9) between the northern
and southern areas by combining results from several stud-
ies in the northern range of tautog. Although commonly
cited as representing annual fecundity estimates, the
methods of Chenoweth (1963) and Stolgitis (1970) clearly
show that their results are batch fecundity estimates. For
our comparison, we selected the lowest value for age-4
and the highest for age-9 tautog as the sampled range
of batch fecundity estimates. We averaged northern data
from Chenoweth (1963: age 4, 265 mm TL, 6000 BF and
age 9, 401 mm TL, 224,000 BF) and Stolgitis (1970: age 4,
261 mm TL, 7000 BF and age 9, 486 mm TL, 260,000 BF)
to create a batch fecundity range of 6500-242,000 eggs.
This range was multiplied by the 68-day "spawning season"
observed in laboratory aquaria by Olla et al. ( 1977) to cal-
culate a range for potential annual fecundity of 442,000 to
16,456,000 eggs per female in northern areas. Our samples
from the southern range (age 4: 275 mm TL, 5000 BF and
age 9: 511 mm TL, 181,200 BF) multiplied by 58 spawn-
ing events in 1995 results in potential annual fecundity of
290,000 to 10,510,000 eggs per female. Differences in these
estimates of potential annual fecundity are primarily due
to the number of spawnings per year and are questionable
because the spawning frequency estimate based on aquari-
um studies may not apply for naturally spawning fish. This
comparison indicates that we still lack adequate informa-
tion on the spawning frequency and annual fecundity for
fish from the northern part of the species range.
Although batch fecundity estimates appear similar
between southern and Northern portions of the tautog's
range, previous batch fecundity estimates from northern
populations are 30 years old. Reported spawning seasons
between areas also vary in length from two to four months,
which could greatly affect potential annual fecundity esti-
mates with this method. Estimates of annual fecundity in
White et al : Reproductive seasonality, fecundity, and spawning frequency of Tautoga onitis
441
the northern regions of the species range should be pur-
sued to determine if tautog annual fecundity varies with
latitude. Evidence of different growth rates (Cooper, 1965,
1966, 1967; Stolgitis, 1970; Hostetter and Munroe,1993;
White, 1996), seasonality of occurrence in coastal waters,
and winter activity cycles between tautog in southern
versus northern regions (011a et al., 1974; Hostetter and
Munroe, 1993; Arendt et al., 2001a, 2001b) strongly point
to considering latitudinal effects when analyzing and com-
paring any biological features of this species. Even if batch
fecundity and spawning frequency remain relatively con-
stant over latitude, size structure of the stock will dictate
estimates of total egg production: thus continued research
is necessary to monitor size structure and abundance of
tautog resources throughout the species range. Additional
data on larger, older females is necessary to evaluate the
relative contribution of older females to population fecun-
dity and egg production. Because many aspects of tautog
life history affect recruitment, further investigation is re-
quired on egg dispersal, egg mortality, larval drift, larval
mortality, hatching success, first feeding success, pre- and
postsettlement mortality, juvenile mortality, recruitment,
stock structure, and spawning stock biomass (ASMFC"*).
Historically, tautog have supported a predominantly
(90%) recreational fishery throughout their range
(ASMFC^). Over the past 15 years, this popular food and
sport fish has increased substantially in value as a com-
mercially targeted species. As popularity and fishing effort
increased, landings peaked in 1993 but have declined more
recently, prompting the Atlantic States Marine Fisheries
Commission (ASMFC"*) to pass a coastwide management
plan for tautog in April 1996.
Tautog annual fecundity is a key piece of data neces-
sary for egg production models and estimates of spawning
stock biomass. and there are no reliable estimates of tautog
spawning stock biomass to date (ASMFC''). In April 1998,
the ASMFC imposed a 14-inch (350-mm) minimum size
limit, effective for tautog caught from Massachusetts to
Virginia. The benefits of instituting a size limit for tautog
are well supported by data from this study. A minimum size
limit allows tautog in the southern regions of the species'
distribution to have at least one spawning season, and
most likely two, thereby affording the opportunity for each
female to contribute on average 3.22 million eggs (calcu-
lated from the linear regression equation. Fig. 8) towards
the annual population fecundity.
Acknowledgments
The authors gratefully acknowledge suggestions and criti-
cisms of three anonymous reviewers that greatly improved
the manuscript. We thank all of those who made this
research possible. This study represents part of a M.S.
thesis (by G. G. Wliite) School of Marine Science, Virginia
" ASMFC (Atlantic States Marine Fisheries Commission). 1996.
Fishery management plan for tautog. Fish. Manage. Rep. 25,
56 p. Atlantic States Marine Fisheries Commission, 1444 Eye
Street, NW, 6"> Floor, Washington, DC 20005.
Institute of Marine Science (VIMS), College of William and
Mary. VIMS volunteers who assisted with our sampling
efforts included D. Estes, R. Holmquist, D. Seaver, and M.
Wagner, members of the Juvenile Finfish Trawl Survey,
and T. Holden, who collected specimens. We appreciate the
interest and cooperation of recreational and commercial
fishermen, especially "Old Joe," Clark and Chester Stultz,
who provided fish, For help processing samples, we thank J.
Brust, W. Coles, C. Cooksey, S. Gaichas, J. Harding, and M.
Wagner (taug circles). We thank C. Bonzek and R. Harris
for computing assistance, D. Evans, R. Diaz, and L. Gar-
rison for statistical help, and J. Harding for critical edito-
rial reviews of an earlier draft of the manuscript. We also
thank R Blake for training in histology procedures, and W.
Vogelbein for aid with interpretation of histology sections.
Finally G. White thanks his parents, who had the foresight
to get him "'hooked on" life in, on, or under water while very
young. This project was funded by grant numbers RF-94-5
and RF-95-3 from the Virginia Marine Resources Commis-
sion, Recreational Saltwater License Fees.
Literature cited
Adams, A. J.
1993. Dynamics of fish assemblages associated with an
offshore artificial reef in the Southern Mid-Atlantic Bight.
Unpubl. M.S. thesis, 97 p. College of William and Mary,
Williamsburg, VA.
Arendt, M. D., J. A. Lucy, and D. A. Evans.
2001a. Diel and seasonal activity patterns of adult tautog,
Tautoga onitis, in lower Chesapeake Bay. inferred from
ultrasonic telemetry. Environ. Biol. Fishes 62: .379-391.
Arendt, M. D., J. A. Lucy, and T. A. Munroe.
2001b. Seasonal occurrence and site utilization patterns of
adult tautog, Tautoga onitis, (Labridae), at manmade and
natural structures in lower Chesapeake Bay. Fish. Bull.
99:519-527.
Austin, H. M.
1973. Distribution and abundance of ichthyoplankton in the
New York Bight during the fall in 1971. N.Y. Fish Game
J. 23:58-72.
Bigelow, H. B., and W. C. Schroeder.
1953. Fishes of the Gulf of Maine. U.S. Fish Wild!. Serv.,
Fish. Bull. .53:1-577.
Bleakney, J. S.
1963. Notes on the distribution and reproduction of the fish
Tautoga onitis in Nova Scotia. Can. Field-Nat. 77:64-65.
Briggs, P. T
1977. Status of tautog populations at artificial reefs in New
York waters and effect of fishing. N.Y. Fish Game J. 24:
154-167.
Chenoweth, S. B.
1963. Spawning and fecundity of tautog, Tautoga onitis
(L.). Unpubl. M.S. thesis, 60 p. Univ. Rhode Island, N.
Kingston. RI.
Cooper, R. A.
1965. Life history of the tautog, Tautoga onitis (Linnaeus),
from Rhode Island. Unpubl. Ph.D. diss., 153 p. Univ.
Rhode Island, N. Kingston. RI.
1966. Migration and population estimation of the tautog,
Tautoga onitis (Linneaus), from Rhode Island. Trans. Am.
Fish. Soc. 95:239-247.
442
Fishery Bulletin 101(2)
1967. Age and growth of the tautog, Tautoga onitis (Linnaeus),
from Rhode Island. Trans. Am. Fish. Soc. 96:134-142.
DeMartini, E. E., and R. K. Fountain.
1981. Ovarian cycling frequency and batch fecundity in the
queenfish. Seriphus politus: attributes representative of
serial spawning fishes. Fish. Bull. 79:547-60.
Eklund, A. M., and T. E. Targett.
1990. Reproductive seasonality of fishes inhabiting hard
bottom areas in the Middle Atlantic Bight. Copeia 1990:
1180-1184.
Hildebrand, S. F., and W. C. Schroeder.
1928. Fishes of the Chesapeake Bay Bull. U.S. Bur Fish.
43 (part 1): 1-366.
Hostetter, E. B., and T. A. Munroe.
1993. Age, growth, and reproduction of tautog Tautoga onitis
(Labridae: Perciformes) from coastal waters of Virginia.
Fish. Bull. 91:45-64.
Hunter, J. R., and S. R. Goldberg
1980. Spawning incidence and batch fecundity in northern
anchovy, Engraulis mordax. Fish. Bull. 77:641-652.
Hunter, J. R., and B. J. Macewicz.
1985. Measurement of spawning frequency in multiple
spawning fishes. In An egg production method for esti-
mating spawning biomass of pelagic fish: application to
the northern anchovy, Engraulis mordax (R. Lasker, ed.), p.
79-94. U.S. Dep. Commer , NOAA Tech. Rep. NMFS 36.
Hunter, J. R., N. C. H. Lo, and R. J. H. Leong.
1985. Batch fecundity in multiple spawning fishes. In An
egg production method for estimating spawning biomass of
pelagic fish: application to the northern anchovy, Engraulis
mordax (R. Lasker, ed.), p. 67-77. Dep. Commer, NOAA
Tech. Rep. NMFS 36.
Hunter J. R., B. J. Macewicz, N. C. H. Lo, and C. A. Kimbrell.
1992. Fecundity, spawning, and maturity of female Dover
sole Microstomus pacificus, with an evaluation of assump-
tions and precision. Fish. Bull. 90:101-128.
Lowerre-Barbieri, S. K., and L. R. Barbieri.
1993. A new method of oocyte separation and preservation
for fish reproduction studies. Fish. Bull. 91: 165-170.
Lowerre-Barbieri, S. K., M. E. Chittenden Jr, and L. R. Barbieri.
1996. The multiple spawning pattern of weakfish in the
Chesapeake Bay and Middle Atlantic Bight. J. Fish Biol.
48:1139-1163.
Luna, L. G. (ed.)
1968. Manual of histologic staining methods of the Armed
Forces Institute of Pathology. American Registry of Path-
ology, Z"^ ed., 258 p. McGraw-Hill Book Co., New York, NY.
McEvoy, L. A., and J. McEvoy
1992. Multiple spawning in several commercial fish species
and its consequences for fisheries management, cultivation
and experimentation. J. Fish Biol. 41(suppl. B):125-136.
Minitab.
1995. Minitab reference manual, release vlOXtra, 542 p.
Minitab Inc., State College, PA.
Musick, J. A.
1972. Fishes of Chesapeake Bay and the adjacent coastal
plain. In A checklist of the biota of the lower Chesapeake
BaylM.L.Wass.ed), p. 175-212. Virg. In.st. Mar Sci. Spec.
Sci. Rep. 65.
Olla, B. L., A. J. Bejda, and A D. Martin.
1974. Daily activity, movements, feeding, and seasonal occur-
rence in the tautog, Tautoga onitis. Fish. Bull. 72:27-35.
1975. Activity, movements, and feeding behavior of the
cunner, Tautogolabrus adspersus, and comparison of food
habits with young tautog, Tautoga onitis, off Long Island,
New York. Fish. Bull. 73:895-900.
1979. Seasonal dispersal and habitat selection of cunner, Tau-
togolabrus adspersus, and young tautog, Tautoga onitis, in
Fire Island Inlet, Long Island, New York. Fish. Bull. 77:255-
261.
Olla, B. L., A. J. Bejda, and A. L. Studholme.
1977. Social behavior as related to environmental factors
in the tautog, Tautoga onitis. In The behavior of marine
organisms: social behavior and communication; navigation;
and development of behavior, p. 47-99. Proc. Annu. North-
east Reg. Meet. Animal Behav. Soc. Plenary Papers. Mar Sci.
Res. Lab. Tech. Rep. 20, Memorial Univ, St. John's, Nfld.
Olla, B. L., and C. Samet.
1977. Courtship and spawning behavior of the tautog, Tau-
toga onitis (Pisces: Labridae), under laboratory conditions.
Fish. Bull. 75:585-599.
Olla, B. L., A. L. Studholme, A. J. Bejda, and C. Samet.
1980. Role of temperature in triggering migratory behav-
ior of the adult tautog Tautoga onitis under laboratory
conditions. Mar Biol. (Beri.) 59:23-30.
Parrish, R. H., D. L. Mallicoate, and R. A. Klingbeil.
1986. Age dependent fecundity, number of spawnings per
year, sex ratio, and maturation stages in northern anchovy,
Engraulis mordax. Fish. Bull. 84:503-517.
Richards, C. E., and M. Castagna.
1970. Marine fishes of Virginia's Eastern Shore (inlet, marsh,
and seaside waters). Chesapeake Sci. 11:235-248.
Scott, W. B., and M. G. Scott.
1988. Atlantic fishes of Canada. Can. Bull. Fish. Aquat.
Sci. 219:1-731.
Sedberry, G. R., and H. R. Beatty.
1989. A visual census of fishes on a jetty at Murrells Inlet,
South Carolina. J. Elisha Mitchell Sci. Soc. 105:59-74.
Simpson, D. G.
1989. Population dynamics of the tautog, Tautoga onitis, in
Long Island Sound. Unpubl. M.S. thesis, 65 p. Southern
Connecticut State Univ., New Haven, CT.
Sogard, S. M., K. W. Able, and M. R Fahay
1992. Early life history of the tautog Tautoga onitis in the
Mid-Atlantic Bight. Fish. Bull. 90:529-539.
Stolgitis, J. A.
1970. Some aspects of the biology of the tautog, Tautoga
onitis (Linnaeus), from the Weweantic River Estuary, Mas-
sachusetts, 1966. Unpubl. M.S. thesis, 48 p. Univ. Mass.,
Amherst, MA.
Wallace, R. A., and K. Selman.
1981. Cellular and dynamic aspects of oocyte growth in
teleosts. Am. Zool. 21:325-343.
Warner, R. R., and D. R. Robertson.
1978. Sexual patterns in the labroid fishes of the Western
Caribbean, I: the wrasses (Labridae). Smithson. Contrib.
Zool. 254:1-27.
Weibel, E. R., G. S.Kistler, and W. F Scherie.
1966. Practical stereological methods for morphometric
cytology. J. Cell Biol. 30:23-28.
West, G.
1990. Methods of assessing ovarian development in fishes: a
review. Aust. J. Mar Freshwater Res. 41: 199-222.
White, G. G.
1996. Reproductive biology of tautog. Tautoga onitis, in
the lower Chesapeake Bay and coastal waters of Virginia.
Unpubl. M.S. thesis, 100 p. College of William and Mary,
Williamsburg, VA.
Wolke, R. E.
1992. Piscine macrophage aggregates: a review. Ann. Rev,
Fish Dis. 1992:91-108.
443
Investigation of congeneric hybridization in and
stock structure of weakfish (Cynoscion regalis)
inferred from analyses of nuclear and
mitochondrial DNA loci*
Jan F. Cordes
John E. Graves
School of Marine Science
Virginia Institute of fv\anne Science
College of William and Mary
Gloucester Point, Virginia 23062
Present address (for J. F, Cordes): Department of Animal Science
University of California
Davis, California 95616
E-mail address (for J F Cordes): |fcordes@ucdavis edu
The weakfish (Cynoscion regalis) is
distributed along the east coast of the
United States from Massachusetts to
eastern Florida and is most abundant
from New York to North Carolina (Big-
elow and Schroeder, 1953). Histori-
cally there has been some question as
to the taxonomic relationship between
weakfish and sand seatrout (Cynosc/on
arenarius); some suggest they may be
separate populations of a single species
(Moshin, 1973; Weinstein and Yerger,
1976; Cowan, 1985; Ditty, 1989), and
others treat them as separate species
(Schlossman and Chittenden, 1981).
Weakfish support substantial com-
mercial and recreational fisheries
throughout their range. Precipitous
drops in total annual catches between
1980 and 1994 led to a temporary ban
on commercial fishing in federal waters
in 1995 (Anonymous, 1995), and there
is concern that bycatch of juvenile
weakfish by shrimp trawlers at the
southern end of the species' range
is adversely impacting abundance
(Vaughan et al.').
As water temperatures warm in
the spring, weakfish move north and
inshore into estuaries to spawn. When
inshore temperatures cool in the fall,
juveniles move south to overwinter
off the coast of North Carolina, and
older fish are thought to migrate
south and offshore (Wilk-). Traditional
studies based on tag and recapture
data (Nesbit, 1954), scale structure
(Perlmutter et al., 1956), morphomet-
ric data (Scoles, 1990), and various
life history characters (Shepherd and
Grimes, 1983, 1984) suggest two or
more independent stocks of weakfish.
These characters may be influenced
by environmental differences, however
(Shepherd and Grimes, 1983), and may
not reflect genetically distinct (repro-
ductively isolated) stocks. Genetic
analyses of weakfish stock structure
in the mid Atlantic Bight employing
allozyme analysis (Crawford et al.,
1989) and restriction fragment length
polymorphism (RFLP) analysis of
whole molecule mitochondrial (mt)
DNA (Graves et al, 1992) were unable
to reject the null hypothesis that weak-
fish along the U.S. east coast comprise a
single, genetically homogeneous stock.
However, the power of both the analy-
ses was limited by low overall genetic
variation.
Recent analyses of new molecular
markers, including microsatellite
DNA loci and nuclear gene intron
regions, have revealed elevated levels
of genetic variation in relation to tra-
ditional methods, such as allozymes
or RFLP analysis of mtDNA (Miller
and Kapuscinski, 1996; Brunner et
al., 1998). Although higher levels of
genetic variation do not necessarily
provide greater stock resolution (Seeb
et al., 1998), microsatellite loci have
revealed stock structure for some spe-
cies, where more traditional molecular
markers have not (Bentzen et al., 1996;
Ruzzante et al., 1996; Patton et al..
1997). Similarly, analyses of variable
gene intron regions have revealed stock
structure within several marine fishes
(Palumbi and Baker, 1994; Moran et
al., 1997; Leclerc et al., 1996; Chow and
Takeyama, 2000). In this study we em-
ployed analyses of nuclear and mtDNA
markers to evaluate stock structure in
weakfish along the east coast of the
United States and to investigate pos-
sible hybridization between weakfish
and other Cynoscion species.
Materials and methods
Sample collections were restricted to
young-of the-year (YOY) fish (less than
140 mm SL) that are reported to remain
in their natal estuaries during the first
several months of growth (Rowe and
Epifanio, 1994). YOY were collected
in the summers of 1996 and 1997 from
five sites along the east coast of the
United States (Fig. 1), maintained on
ice after capture, transported to the
laboratory, and frozen at -80°C. Muscle
tissue was excised from each sample
and either stored at -80°C or placed
in DMSO buffer (25 mM EDTA, 20%
DMSO, saturated NaCl) and stored at
room temperature. Genomic DNA was
isolated by following the protocol of
Sambrook et al. (1989), as modified in
Cordes (2000).
Specific identification of individuals
was determined by using a molecular
key for 16 species of Chesapeake Bay
* Contribution 2532 of the Virginia Institute
of Marine Science. College of William and
Mary, Gloucester Point, VA 23062.
' Vaughan, D. S.. R. J. Seagraves. and K.
West. 1991. An assessment of the
Atlantic weakfish stock, 1982-1988. Atl.
States Mar. Fish. Comm. Spec. Rep. 21.
29 p. + tables. Atlantic States Marine
Fisheries Commission, 1444 Eve St., NW
6"' Floor, Washington, D.C. 20005.
- Wilk, S. J. 1976. The weakfish— a wide
ranging species. Atl. States Mar. Fish.
Comm, Mar. Resourc. Atl. Coast Fish.
Leaflet 18, 4 p. Atlantic States Marine
Fisheries Commission. 1444 Eye St., NW
6'*' Floor, Washington. D.C. 20005.
Manuscript accepted 21 October 2002.
Manuscript received 9 January 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:443-450 (2003).
444
Fisher/ Bulletin 101(2)
45"
40"
35"
30°
Figure 1
Sampling locations for young-of-the-ycar (YOY) weakfish
iCynoscion regalis) in the summers of 1996 and 1997. Sites
are Peconic Bay, New York (NY); Delaware Bay, Delaware
(DB); Chesapeake Bay, Virginia (CB); Pamlico Sound, North
Carolina (NO; and Doboy Sound, Georgia (GA).
sportfishes (including eight species of sciaenids) based on
a 12S/16S mtDNA gene region digested with Rsa I (Cordes
et al., 2001). Additional 12S/16S mtDNA/Rsa I patterns
were generated for silver seatrout iCynsoscion nothus) and
sand seatrout (C. arenarius) from the Gulf of Mexico, as
well as for banded drum {Larinius fasciatus), gulf kingfish
(Menticirrhus littoralis), and star drum iStellifer lanceola-
tus) from the South Atlantic Bight following procedures in
Cordes etal. (2001).
The following microsatellite primers (Table 1 ) developed
for red drum (Sciaenops ocellatus) and spotted seatrout
{Cynoscion nehulosus) loci were used to amplify weakfish
DNA: SOC050 and SOC044 (Turner et al., 19981, SOC014
(Chapman'), and CNE612 (Chapman et al., 1999). Ampli-
fications of all microsatellite loci were carried out in 10 pL
^ Chapman, R. W. 1998. Unpubl. data. Marine Resources Re-
search Institute, Department of Natural Resources, Charleston,
SC 29422.
reactions containing 8.30 pL sterile dHgO, 1.0 pL lOx PCR
buffer with 15 niM MgCU, 0.20 pL 10 niM dNTP mixture,
0.05 pL forward primer ( 100 pm/pL) labeled with a fluores-
cent dye (Licor), 0.20 pL reverse primer (100 pm/pL), 0.05
pL Taq I polymerase (5 U/pL), and 0.20 pL weakfish DNA.
Samples were first denatured for 4 min at 95"C, followed by
32 cycles of PCR amplification performed under the follow-
ing conditions: 1 min. at 94°C, 1 min. at 50°C, and 1 min.
at 72°C. Reactions were given a final 7 min. extension at
72°C. PCR product alleles were separated electrophoreti-
cally on a 6% Long Ranger™' polyacrylamide gel with a
model 4000 automated DNA infrared sequencer from Li-
Cor (Lincoln, NE).
Universal actin gene primers developed by G. Warr and
M.Wilson (cited in Reece et al., 1997) were used to identify
and refine an approximately 800-bp actin intron region lo-
cus (CRESIAl) in weakfish (Cordes, 2000). Weakfish PCR
amplification products obtained with S7 ribosomal protein
intron 2 (RP2) primers originally developed from swordfish
Xiphius gladius (Chow and Hazama, 1998) were cloned
and sequenced as described in Cordes (2000) and checked
against sequences published in Genbank to confirm their
identity. The original RP2 primers were then used without
modification for amplification of all samples. Both the CRE-
SIAl and RP2 amplifications were carried out under the
same conditions outlined above for the microsatellite loci,
with the exception that the annealing temperature was
lowered to 45°C. CRESIAl and RP2 amplification products
from a subset of each weakfish collection were screened
for polymorphisms with a panel of restriction endonucle-
ases and the resulting digestions were separated on 2.5%
agarose gels with 1% NuSieve and 1.5% agarose in IxTBE
buffer (Cordes, 2000). Gels were stained in IxTBE buffer
containing 30 pL (5 mg/niL) ethidium bromide (EtBri, vi-
sualized on a Spectroline model TR-302 transilluminator,
and photographed with a Polaroid CU-5 land camera. Dra I
was the only enzyme that revealed polymorphic restriction
sites within CRESIAl, and only Hinf I revealed reliably
scored polymorphisms in RP2. All YOY weakfish samples
were subsequently screened for variation at CRESIAI/Z)ra
IandRP2///in/'I.
Microsatellite gel images and restriction enzyme diges-
tion patterns for CRESIAl and RP2 were analyzed by
using the software program RFLPScan Plus 3.0 (CSPI-
Scanalytics, 1996). Statistical analyses for all loci were
performed with the Arlequin 1.1 software program of
Schneider et al. (1997). Nonparametric, exact-significance
tests (exact 0 significance tests and exact probability tests)
were used to evaluate sample genotype distributions for
departures from Hardy-Weinberg expectations. Unbiased
estimators of exact significance probabilities for the Hardy-
Weinberg equilibrium tests were calculated by using the
Markov chain algorithm of Guo and Thompson (1992)
with a Markov chain length of 100,000 steps. Patterns
of genetic diversity and divergence within and between
populations were evaluated by using the analysis of mo-
lecular variance (AMOVA) of Excoffier et al. (1992), which
generates F-statistics analogous to the Svaluesof Weir and
Cockerham (1984). Significance of /"'-statistics was evalu-
ated with exact F permutation procedures (Excoffier et al.,
NOTE Cordes and Graves: Hyridization and stock structure in Cynosaon regalis
445
1992).Type-I error was controlled
for all multiple testing with the
sequential Bonferroni method of
Rice (1989).
Results and discussion
Inclusion of nontarget
species in weakfish samples
Initial analysis of the SOC050 mi-
crosatellite data revealed a sig-
nificant departure of genotypic
frequencies from expectations
of Hardy-Weinberg equilibrium
for the Georgia 1997 sample,
even after correction for mul-
tiple tests (a=0.005). Similarly,
initial SOC050 AMOVA results
indicated a significant within-
population variance (P=0.031),
and exact F permutation tests of
population pairwise F^,^ values
resulted in a number of near-
significant corrected P values,
all involving the Georgia 1997
sample. Inspection of the Georgia
1997 SOC050 alleles revealed a
bimodal size distribution due to
the presence of several unusually
small alleles less than 187 bp in
size. It was suspected that alleles
in the smaller mode might be the
result of misidentified individuals,
hybridization, or introgression.
Analysis of putative weakfish
with small SOC0.50 alleles with
the 12S/16S marker of Cordes
et al. (2001) resulted in three
distinct restriction digestion
patterns. One pattern matched
that reported for weakfish, and
the other two did not match any
of the 16 species surveyed by
Cordes et al. ( 2001 ). To determine
the identity of the unknown pat-
terns, voucher samples of five ad-
ditional sciaenid species (listed
in the "Materials and methods"
section above) were analyzed
with the 12S/16S mitochondrial
marker. The two unknown pat-
terns matched those of silver
seatrout (Cynoscion nothus) and
sand seatrout (C arenarius)
(Table 2). The SOC050 locus was
subsequently amplified for all
silver seatrout (/! = 13) and sand
seatrout («=15) samples, produ-
o
So
o
c
c
_ XI
JV a
•c
c
■&
•c
o
o t"
<= ^
*J to
CO QJ
a ^
,
T3
-a
C
3
CO
u:
is
o
x;
U
CO
w
o
c
cc
!N
CU
Qi
x
446
Fishery Bulletin 101(2)
Table 2
Restriction digestion patterns
of the 12S/16S mitochondrial DNA
region for putative weakfish (Cynoscion regc
lis) individuals in
the Georgia 1997 sample, sand seatrout (C.
arenarius) and
silver seatrout (C. nothus) digested
with the enzyme
Rsa I. n = number
of individuals exhibiting the adjacent pattern. Apparent total size
differences may
be gel artifacts due to unresolved bands <100 |
bp in one or more of the species.
Species
n
Restriction
Fragment
Sizes
(bp)
Total size (bp)
Georgia 1997 Sample
weakfish
3
461
300
200
167
1128
unknown A
7
413
300
200
167
1080
unknown B
5
461
300
256
167
1184
Known standards
Cynoscion arenarius
15
461
300
256
167
1184
Cynoscion nothus
13
413
300
200
167
1080
cing allele sizes of 175-181 bp for silver seatrout and
175-193 bp for sand seatrout. Amplification of the silver
seatrout and sand seatrout samples with the remaining
three microsatellite and two intron loci did not provide fur-
ther evidence of hybridization. SOC044 and CNE612 allele
size ranges for both species fell within the range exhibited
by the weakfish samples, and the SOC014 and both intron
loci did not amplify in either the silver seatrout or sand
seatrout samples.
Individuals with unusually small SOC050 alleles from
the Georgia 1997 sample fell into one of four general
classes. Seven individuals had silver seatrout mtDNA and
two small SOC050 alleles and were presumably pure silver
seatrout. The inclusion of these individuals in the collection
may not be surprising because both weakfish and silver
seatrout are common in the South Atlantic Bight (Bigelow
and Schroeder, 1953; Hildebrand, 1955) and are difficult to
distinguish during their early life history stages. Although
the latter species is known to inhabit deeper waters as
adults (Ginsburg, 1931), both species are inshore summer
spawners (Devries and Chittenden, 1982; Shepherd and
Grimes, 1984).
Three individuals possessed sand seatrout mtDNA and
two small SOC050 alleles and were presumably pure sand
seatrout. Some researchers have suggested that weak-
fish and sand seatrout represent separate populations
of a single species (Moshin, 1973; Weinstein and Yerger,
1976; Cowan, 1985; Ditty, 1989), and others treat them as
separate species (Schlossman and Chittenden, 1981) with
distributions confined to the western Atlantic (weakfish)
and the Gulf of Mexico (sand seatrout). Paschall (1986)
was unable to distinguish between the two species using
allozyme electrophoresis. In contrast, results presented
here are consistent with the existence of two distinct spe-
cies, with weakfish and sand seatrout co-occurring off the
east coast of the United States at least as far north as
Doboy Sound, Georgia. This distribution pattern is consis-
tent with the phylogeographic patterns of 19 freshwater,
coastal, and marine species distributed along the U.S. East
Coast and the Gulf of Mexico that exhibited geographically
concordant forks in their intra- or interspecific mtDNA
phylogenies (or in both phylogenies) (Avise, 1992). In the
present situation, apparently distinct Gulf (sand seatrout)
and Atlantic (weakfish) species may have reestablished
contact in a hybrid zone (see below) through movement of
the Gulf species into the Atlantic.
Three individuals had weakfish mtDNA and a single
small SOC050 allele and were presumably hybrids of
weakfish and sand seatrout or silver seatrout (with female
weakfish parentage). In addition, two individuals pos-
sessed sand seatrout mtDNA and a single small SOC050
allele and were presumably hybrids of weakfish and sand
seatrout with female sand seatrout parentage. These data
suggest that hybridization occurs between weakfish and
sand seatrout and that the genetic exchange is not gender
restricted. Because of the overlap in microsatellite allele
sizes seen between silver seatrout and sand seatrout, hy-
bridization between weakfish and silver seatrout could not
be excluded. The lack of suspected hybrids with silver seat-
rout mtDNA, however, suggests that hybridization did not
involve this species. The possibility exists that the putative
hybrids are in fact weakfish with rare mtDNA haplotypes
common to the three Cynoscion species studied here. This
seems unlikely because only one 12S/16S mtDNA//?,sn
I pattern was noted among 40 weakfish in the species
identification study of Cordes et al. (2001). Furthermore,
analysis of 20 weakfish taken from each of the four loca-
tions outside of Georgia with the 12S/16S marker revealed
no new mtDNA patterns. Also, the mtDNA haplotypes seen
in sand seatrout and silver seatrout seem to vary in size
and can not be clearly related to the weakfish haplotype
by the addition or deletion of presumed restriction sites.
This condition is more in keeping with mtDNA of different
species, although the apparent size differences may be gel
artifacts due to unresolved bands <100 bp in one or more
of the species.
Re-evaluation of the remaining 1996-97 SOC050 data re-
vealed occasional occurrences of small alleles in individual
fish m all but the New York samples (Table 3). Examina-
tion of the 12S/16S mtDNA region of these individuals
identified a single silver perch (Bairdiella chrysoiiro) in
the Chesapeake Bay 1997 sample (silver perch mtDNA
NOTE Cordes and Graves: Hyridization and stock structure in Cynosaon regalis
447
Table 3
Frequencies of unusual alle
es in
four geographical sam-
pies of weakfish (Cynoscion
regc
ilis) taken in 1996 and
1997. The number of individuals
with anoma
lous alleles
that were subsequently eliminated from the
population
structure analysis is given in
parentheses after the sample |
names.
Sample
Allele (bp)
Frequency
Georgia 1996 (4)
175
0.018
177
0.009
179
0.009
North Carolina 1996(1)
177
0.010
Chesapeake Bay 1996(1)
177
Delaware Bay 1996 (2)
179
0.011
181
0.011
Georgia 1997(15)
171
0.021
173
0.010
175
0.125
177
0.094
179
0.010
North Carohna 1997(1)
177
0.009
Chesapeake Bay 1997 (3)
171
0.009
177
0.009
179
0.009
Delaware Bay 1997 (5)
177
0.023
181
0.034
and two alleles 171 bp in size). All other individuals were
putative hybrids with weakfish mtDNA and a single small
SOC050 allele characteristic of silver and sand seatrout. As
mentioned previously, subsequent analysis of 20 weakfish
taken from each of the four locations outside of Georgia
with the 12S/16S marker revealed only weakfish mtDNA.
If the small SOC050 alleles found in the more northern
populations are not simply rare weakfish alleles shared in
common with the other two Cynoscion species, they may
indicate that introgressive hybridization is responsible for
the migration of the smaller alleles into northern waters
(although the northward movement of hybrid fish out of
the contact zone cannot be excluded). As a result of these
findings, all individuals in the 1996 and 1997 collections
exhibiting at least one small SOC050 allele less than 183
bp in length were eliminated from the population structure
analyses.
Stock structure analysis
All four microsatellite loci were polymorphic in all sampled
locations in both years. Allele frequency distributions for
each locus are available from the authors upon request.
Sample sizes in), number of alleles (A''), expected hetero-
zygosities (gene diversities), and significance test results
for Hardy-Weinberg equilibrium are provided in Table 4.
Levels of variation differed greatly among the four mic-
rosatellite loci. The number of alleles ranged from two
(SOC014) to 37 (CNE612), and average expected heterozy-
gosities ranged from 0.085 (SOC014) to 0.928 (CNE612).
These values are consistent with heterozygosity ranges
reported in other multilocus microsatellite studies on spe-
cies including Atlantic cod (Bentzen et al., 1996), northern
pike (Miller and Kapuscinski, 1996), pink and sockeye
salmon (Seeb et al., 1998), and Arctic char (Brunner at
al., 1998). In contrast, Crawford et al. (1989) and Graves
et al. (19921 found very low levels of genetic variation in
an analysis of weakfish populations using allozymes and
mtDNA restriction fragment-length polymorphism (RFLP)
analyses, respectively. None of the genotypic distributions
for any of the four microsatellite loci at any of the collection
locations in either year differed significantly from Hardy-
Weinberg expectations after correcting for multiple tests
(Table 4).
Digestion of actin intron (CRESIAl) amplifications with
the restriction endonuclease Rsa I revealed a single poly-
morphic restriction site that produced two alleles. Expected
heterozygosities ranged from 0.000 for the monomorphic
Georgia 1997 sample to 0.096 for the Chesapeake Bay
1996 sample. Digestion of the RP2 amplifications with
the restriction endonuclease Hinf I also resulted in two
alleles. Expected heterozygosities ranged from 0.194 in the
Delaware Bay 1997 sample to 0.370 in the Georgia 1997
sample. Levels of genetic variation within the two nuclear
gene intron regions were low in relation to three of the four
microsatellite loci and were more similar to those found in
the polymorphic allozyme loci of Crawford et al. ( 1989). In
another study where nuclear intron RFLP analysis was
used, similar levels of heterozygosity in Pacific salmon
were found (Moran et al., 1997), as in RFLP studies of
anonymous single copy nuclear (ascn) DNA loci in Atlantic
cod {Gadus morhua) (Pogson et al., 1995) and blue marlin
{Makaira nigricans) (Buonaccorsi et al., 1999). In con-
trast, higher heterozygosities (44-58%) were reported in
an ascnDNA/RFLP analysis of striped bass {Morone saxa-
tilis) by Leclerc et al. (1996). None of the sample genotype
distributions for either locus differed significantly from
Hardy-Weinberg expectations after correcting for multiple
tests (Table 4).
To test for population structure, microsatellite loci were
analyzed individually and as a combined data set. AMOVA
results did not reveal significant differences between sam-
ple locations or years for any of the four loci or for the com-
bined data (all P>0.05). Single-locus population pairwise
Fgy values were relatively low, and mean Fj^.j. values ranged
from 0.002 (SOC050 and CNE612) to 0.018'(SOC044). Exact
F permutation tests were not significant for any of the four
loci or the combined data set after correction for multiple
testing.
AMOVA results for both the actin and RP2 loci indicated
no significant differences between sample locations or years
(all P>0.05). Single-locus population pairwise F^^ values
for the actin locus were consistently low (mean=0.005),
ranging from Fgj < 0.000 for most of the comparisons to
an Fg-p of 0.035 between Georgia 1996 and Georgia 1997
and between Chesapeake Bay 1996 and Georgia 1997.
A single exact F permutation test, between Delaware 1996
448
Fishery Bulletin 101(2)
Table 4
Sample sizes
in), number of alleles (A^), expected heterozygosities ^H )
and P values for tests of Hardy-Weinberg equilibrium for
four microsatellite loci
the actin intron (CRESIAl), and the ribosomal protein 2 intron (RP2)
gene regions
GA = Georgia, NC =
North Carolina, CB = Chesapeake Bay, DB = Del
aware Bay,
NTY = New York. NT =
monomorph
c sample not tested.
GA 1996
NC 1996
CB 1996
DB 1996
NY 1996
GA 1997
NC 1997
CB 1997
DB 1997
NY 1997
SOC050
n
51
49
64
46
46
33
52
55
42
54
N
7
6
6
5
7
6
6
6
8
7
«exp
0.741
0.702
0.694
0.731
0.724
0.758
0.712
0.740
0.737
0.722
P'
0.067
0.542
0.507
0.130
0.959
0.566
0.577
0.721
0.349
0.174
SOC044
n
47
46
63
55
55
36
60
56
52
56
N
2
3
2
2
2
2
2
2
2
2
«e,p
0.362
0.434
0.374
0.251
0.416
0.407
0.302
0.350
0.203
0.419
P'
1.000
0.019
0.496
0.303
0.512
0.010
0.669
0.116
0.0.512
0.198
SOC014
n
43
48
64
52
54
39
56
55
52
57
N
2
2
2
2
2
2
2
2
2
2
^exp
0.090
0.081
0.046
0.075
0.170
0.144
0.053
0.088
0.038
0.068
P'
1.000
1.000
1.000
1.000
1.000
1.000
0.027
1.000
1.000
1.000
CNE612
n
39
43
62
50
46
33
54
52
50
56
N
17
23
20
20
22
19
25
23
22
23
^exp
0.916
0.943
0.928
0.912
0.916
0.935
0.934
0.934
0.923
0.936
P'
0.050
0.113
0.898
0.530
0.238
0.752
0.522
0.419
0.060
0.290
CRESIAl
n
40
42
40
42
40
36
51
54
45
55
N
2
2
2
2
2
1
2
2
2
2
«exp
0.096
0.089
0.031
0.055
0.096
0.000
0.025
0.053
0.047
0.020
P'
0.076
0.091
1.000
0.036
0.078
NT
1.000
1.000
1.000
1.000
RP2
n
48
45
45
42
41
29
48
49
42
41
N
2
2
2
2
2
2
2
2
2
2
^exp
0.237
0.200
0.217
0.230
0.253
0.373
0.237
0.201
0.194
0.253
P'
0.184
0.432
0.104
0.120
0.180
0.298
0.189
0.465
0.052
0.179
' None of the samples differed significar
tly from Hardy-Weinberg expectations after sequential Bonferroni corrections (a=0.005l.
and Georgia 1997, was significant after correction for mul-
tiple testing (a<0.001). Single-locus population pairwise
Ffj.]. values for the RP2 locus were also low (mean=0.006),
ranging from F^-^ < 0.000 for most of the comparisons to
a high of 0.050 between Georgia 1997 and Delaware Bay
1997. None of the exact F permutation tests were signifi-
cant after correction for multiple testing.
From our results we were unable to reject the null hy-
pothesis that weakfish comprise a single, genetically ho-
mogeneous stock. These results are consistent with those
based on allozymes (Crawford et al., 1989) and RFLP
analysis of mtDNA (Graves et al., 1992) and illustrate the
point that increased genetic variability in microsatellites
in relation to more traditional markers will not always
provide greater stock resolution (Seeb et al., 1998). The
amount of genetic exchange necessary to prevent the ac-
cumulation of significant genetic divergence between fish
from different locations may be as little as a few individu-
als per generation (Allendorf and Phelps, 1981). Weakfish
tagging data indicate that low levels of exchange occur
between geographically distant populations of weakfish
(Bain et al., 1998). Estimates of natal homing in yearling
weakfish, calculated by Thorrold et al. (2001) using geo-
chemical signatures in the otoliths of the same weakfish
used in the present study, indicated spawning-site fidelity
ranging from SI'S to 81%, suggesting exchange rates suf-
ficient to prohibit genetic divergence between locations.
The inclusion of nontarget species in our weakfish sam-
ples illustrates the advantages in using multiple marker
systems. If only a single microsatellite locus had been
used, or if the study had been restricted to nuclear intron
markers alone, it is very likely that the sand seatrout and
NOTE Cordes and Graves: Hyridization and stock structure in Cynosdon regalis
449
silver seatrout specimens would have gone unnoticed. This
could easily have resulted in a type-II error. Likewise, if
nongenetic markers such as otolith microchemistry had
been used exclusively, the analyses ofThorrold et al. (2001)
would have been based on a mixed-species sample. Instead,
it was possible not only to recognize the individuals as
anomalous but also to identify them to species and pro-
vide evidence of hybridization between at least two of the
Cyrioscion congeners. It is hoped that further refinement
of the inter- and intraspecific molecular markers developed
here and in other studies will eventually be helpful in fur-
ther clarifying the taxonomic status, population structure,
and possible hybridization within the genus Cynoscion.
Acknowledgments
We would like to thank all those who supplied us with weak-
fish samples, including Louis Barbieri. Susan Lowerre-
Barbieri, Christina Grahn, Patrick Geer, Mike Greene, and
Simon Thorrold. We also appreciate the samples of banded
drum, gulf kingfish, and star drum provided by Trey Knott
and the silver seatrout and sand seatrout specimens pro-
vided by Bill Karel. We gratefully acknowledge Kim Reece,
John Gold, Linda Richardson, and Robert Chapman for gen-
erously providing us with their published and unpublished
primer sequences. Funding for this study was provided
through the Virginia Marine Resources Commission.
Literature cited
Allendort; F. W., and S. R. Phelps.
1981. Use of allelic frequencies to describe population
structure. Can. J. Fish. Aquat. Sci. 38(12):1507-1514.
Anonymous.
1995. Overfished weakfish stock forces closure of federal
waters. Fisheries 21:46-47.
Avise, J. C.
1992. Molecular population structure and the biogeographic
history of a regional fauna: a case history with lessons for
conservation biology. Oikos 63:62-76.
Bain C, J. Lucy, and M. Arendt.
1998. Virginia game fish tagging program annual report
1997. Virginia Marine Resource Report 98-3 (VSG-98-
01), 22 p. Virginia Institute of Marine Science. Virginia
Sea Grant Marine Advisory Program, College of William
and Mary, Gloucester Point, VA.
Bentzen P., C. T. Taggart, D. E. Ruzzante, and D. Cook.
1996. Microsatellite polymorphism and the population
structure of Atlantic cod tGadus morhua) in the northwest
Atlantic. Can. J. Fish. Aquat. Sci. .53 ( 121:2706-2721.
Bigelow H., and W. Schroeder.
1953. Fishes of the Gulf of Maine. U.S. Fish and Wildlife
Service Fish. Bull. 53, 577 p.
Brunner P. C, M. R. Douglas, and L. Bernatchez.
1998. Microsatellite and mitochondrial DNA assessment of
population structure and stocking effects in Arctic charr
Salvelmux alpinus (Teleostei: Salmonidael from central
Alpine lakes. Mol. Ecol. 7 (2): 209-223.
Buonaccorsi V. P., K. S. Reece, L. W. Morgan, and J. E. Graves.
1999. Geographic distribution of molecular variance within
the blue marlin {Makaira nigricans): a hierarchical analysis
of allozyme, single-copy nuclear DNA, and mitochondrial
DNA markers. Evolution 53 ( 2 1:568-579.
Chapman R. W., G. R. Sedberry, J. C. McGovern, B. A. Wiley, and
J. A. Musick.
1999. The genetic consequences of reproductive variance:
studies of species with different longevities. In Life in the
slow lane: ecology and conservation of long-lived marine
animals. Am. Fish. Soc. Symp. 23:169-181.
Chow S., and K. Hazama.
1998. Universal PCR primers for 87 ribosomal protein gene
introns in fish. Mol. Ecol. 7:1247-1263.
Chow, S., and H. Takeyama.
2000. Intron length variation observed in the creatine
kinase and ribosomal protein genes of the swordfishX;p/zJas
gladius. Fish. Sci. 64(3):397-402.
Cordes, J. F
2000. Application of genetic markers to provide species iden-
tification and define stock structure: analyses of selected
marine fishes of the mid-Atlantic bight. Ph.D. diss., 142 p.
Virginia Institute of Marine Science, College of William and
Mary, Gloucester Point, VA.
Cordes, J. F., S. L. Armknecht, E. A. Starkey, and J. E. Graves.
2001. Forensic identification of sixteen species of Chesa-
peake Bay sportfishes using mitochondrial DNA restriction
fragment-length polymorphism (RFLP) analysis. Estuar-
ies 24(1 ):49-58.
Cowan, J. H., Jn
1985. The distribution, transport and age structure of drums
(family Sciaenidae) spawned in the winter and early spring
in the continental shelf waters off west Louisiana. Ph.D.
diss., 182 p. Louisiana State University, Baton Rouge, LA.
Crawford, M. K., C. B. Grimes, and N. E. Buroker
1989. Stock identification of weakfish, Cynoscion regalis, in
the middle Atlantic region. Fish. Bull. 87: 205-211.
CSPI-Scanalytics.
1996. RFLPScan Plus, version 3.0. SCPI-Scanalytics, Bil-
lerica, MA.
DeVries, D. A., and M. E. Chittenden Jr
1982. Spawning, age determination, longevity, and mortal-
ity of the silver seatrout. Cynoscion nothus. in the Gulf of
Mexico. Fish. Bull. 80:487-500.
Ditty J. G.
1989. Separating early larvae of sciaenids from the western
North Atlantic: a review and comparison of larvae from the
northern Gulf of Mexico off Louisiana and Atlantic coast of
the U.S. Bull. Mar Sci. 44:1083-1105.
Excoffier L., P. Smouse, and J. Quattro.
1992. Analysis of molecular variance inferred from metric dis-
tances among DNA haplotypes: application to human mito-
chondrial DNA restriction data. Genetics 131:479^91.
Ginsburg, I.
1931. On the differences in the habitat and size of Cynoscion
arenarius and Cynoscion nothus. Copeia 1931:144.
Guo, S., and E. Thompson.
1992. Performing the exact test of Hardy-Weinberg propor-
tion for multiple alleles. Biometrics 48:361-372.
Graves, J. E., J. R. McDowell, and M. L. Jones.
1992. A genetic analysis of weakfish Cynoscion regalis
stock structure along the mid-Atlantic coast. Fish. Bull.
90:469-475.
Hildebrand, H. H.
1955. A study of the fauna of the pink shrimp (Penaeus duor-
arum Burkenroad) grounds in the Gulf of Campeche. Publ.
Inst. Mar Sci. 4:169-232.
Leclerc, G. M., M. Diaz, and B. Ely
1996. Use of PCR-RFLP assays to detect genetic variation at
450
Fishery Bulletin 101(2)
single-copy nuclear loci in striped bass (Morone saxatilis).
Mol. Mar. Biol. Biotechnol. 5:138-144.
Miller, L. M.. and A. R. Kapuscinski.
1996. Microsatellite DNA markers reveal new levels of
genetic variation in northern pike. Trans. Am. Fish. Soc.
125(6):971-977.
Moran, P., D. A. Dightman, R. S. Waples, and L. K. Park.
1997. PCR-RFLP analysis reveals substantial population-
level variation in the introns of Pacific salmon (Oncorhyn-
chus spp.). Mol. Mar Biol. Biotechnol. 6(4);315-327.
Moshin, A. K. Mohammad.
1973. Comparative osteology of the weakfishes (Cynoscion )
of the Atlantic and Gulf coasts of the United States ( Pisces-
Sciaenidae). Ph.D. diss., 148 p. Texas A&M Univ., College
Station, TX.
Nesbit, R. A.
1954. Weakfish migration in relation to its conservation.
U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 115:81.
Palumbi, S. R., and C. S. Baker
1994. Contrasting population structure from nuclear intron
sequences and mtDNA of humpback whales. Mol. Biol.
Evol. 11:426-435.
Paschall, R. L.
1986. Biochemical systematics of the seatrouts of the
western Atlantic genus Cynoscion. Master's thesis, 100
p. Univ. New Orleans, New Orleans, LA.
Patton, J. C, B. J. Galloway, R. G. Fechhelm, and M. A. Cronin.
1997. Genetic variation of microsatellite and mitochondrial
DNA markers in broad whitefish (Coregonus nasus) in the
Colville and Sagavanirktok rivers in northern Alaska.
Can. J. Fish. Aquat. Sci. 54: 1548-1556.
Perlmutter, A., S. W. Miller, and J. C. Poole.
1956. The weakfish (Cynoscion regalis) in New York waters.
N.Y. Fish Game 3:1^3.
Pogson, G. H., K. A. Mesa, and R. G. Boutilier
1995. Genetic population structure and gene flow in the
Atlantic cod ^adus morhua: a comparison of allozyme and
nuclear RFLP loci. Genetics 139:375-385.
Reece, K. S., M. E. Siddall, E. M. Burreson, and J. E. Graves.
1997. Phylogenetic analysis ofPerkinsus based on actin gene
sequences. J. Parasitol. 83(3):417-423.
Rice, W. R.
1989. Analyzing tables of statistical tests. Evolution 43:
223-225.
Rowe, P. M., and C. E. Epifanio.
1994. Tidal stream transport of weakfish larvae in Delaware
Bay, USA. Mar Ecol. Prog. Ser 110:105-114.
Ruzzante, D. E., C. T. Taggart, and D. Cook.
1996. Spatial and temporal variation in the genetic compo-
sition of a larval cod (Gadus morhua) aggregation: cohort
contribution and genetic stability. Can. J. Fish. Aquat. Sci.
53:2695-2705.
Sambrook, J., E. F Fritsch, and T Maniatis.
1989. Molecular cloning: a laboratory manual, 3 vols. Cold
Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Schlossman, P. A., and M. E. Chittenden Jr.
1981. Reproduction, movements, and population dynamics
of the sand seatrout, Cynoscion arenarius. Fish. Bull. 79:
649-669.
Schneider, S., J.-M. Kueffer, D. Roessli, and L. Excoffier
1997. Arlequin: a software for population genetics data
analysis (vers. 2.0), 111 p. Univ. Geneva, Switzerland.
Scoles, D. R.
1990. Stock identification of weakfish, Cynoscion regalis, by
discriminant function analysis of morphometric characters.
Masters thesis, 55 p. Virginia Institute of Marine Science,
College of William and Mary, Gloucester Point, VA.
Seeb, J. E., C. Habicht, J. B. Olsen, R Bentzen, J. B. Shaklee,
and L. W. Seeb.
1998. Allozyme, mtDNA, and microsatellite variants de-
scribe structure of populations of pink and sockeye salmon
in Alaska. Bull. NPAFC 1:300-319.
Shepherd, G. R., and C. B. Grimes.
1983. Geographic and historic variations in growth of weak-
fish, Cynoscion regalis, in the middle Atlantic Bight. Fish.
Bull. 81:803-813.
1984. Reproduction of weakfish, Cynoscion regalis, in the
New York Bight and evidence for geographically specific
life history characteristics. Fish. Bull. 82:501-511.
Thorrold, S. R., C. L. Latkoczy, P. K. Swart, and C. M. Jones.
2001. Natal homing in a marine fish metapopulation. Sci-
ence 291:297-299.
Turner, T. F, L. R. Richardson, and J. R. Gold.
1998. Polymorphic microsatellite DNA markers in red drum
[Sciaenops ocellatus). Mol. Ecol. 7( 12):1771-1788.
Weinstein, M. P., and R. W. Yerger
1976. Protein taxonomy of the Gulf of Mexico and Atlantic
Ocean seatrouts, genus Cynoscion. Fish. Bull. 74:599-
607.
Weir, B. S., and C. C. Cockerham.
1984. Estimating F-statistics for the analysis of population
structure. Evolution 38:1358-1370.
451
Dynamic age-length keys
Are Salthaug
Institute of Marine Research
Nordnesgaten 50
PO Box 1870
N-5817 Bergen, Norway
E-mail address arese'imrno
Information about age composition is
important when analyzing fish popu-
lation dynamics. Age determination
of individual fish is more difficult and
time consuming than the recording
of length measurements but by using
age-length keys, age distributions can
be estimated without much difficulty
from length distributions (Fridrikson,
1934). Knowledge of the age-length
composition in the population or in
a given subgroup of the population
is required for constructing adequate
age-length keys. Various methods
for construction and evaluation of
age-length keys are described in the
literature (see e.g. Fridrikson, 1934;
Macdonald and Pitcher, 1979; Schnute
and Fournier, 1980; Kimura and Chi-
kuni, 1987; Hayes, 1993; Terceiro and
Ross, 1993; Goodyear, 1997). Because
of individual variation in growth rates
and the variation in mortality rates at
different ages and sizes, the age and
length composition of a fish stock are
constantly changing. With sufficient
information about a fish stock, the
change in the age-length composition
can be modeled and theoretical age-
length keys can be constructed for
specific time periods. Age distributions
can then be estimated from length
distributions taken at different times
of the season. In this work, a simple
but useful modeling approach for con-
structing dynamic age-length keys is
described and applied to data from the
Atlantic cod [Gadus morhua) stock in
the Barents Sea.
probability of an individual being a
certain length (/) within an age group
(a ) at a given time is assumed to follow
a normal probability density function
(Fig. lA), Niji^, fj^), with expectation
s^ and standard deviation o^. When
lengths of individual fish are recorded,
they are normally classified as discrete
length groups (e.g. 1-cm or 5-cm length
intervals). The probability (P) for an
individual in age group a to belong in a
discrete length group, s, at a given time
is then given by
P.S ] N{^i,,s
■ age=8
• \,f''^^
^/ o
• age=9
01
^^
1^
» age=10+
0 0
or'-'
n n
0.0
01 0.2 03 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
1986 1990
^ 08
r
0.5
^^
». /^
0.4
/^ ° 06
• X^
o/d
. \X
0.3
^,
^y^
ay^ o, • 0.4
^^
0.2
A
i^
01
•
^^^..^ ^ 0.2
1/^'
• m^
l^""^
-jTa! ■
- 0 0
^
0 0
t^*>^ ^
t
o 0.0
0.1 0.2 0.3 0.4 0.5 0-0 01 0.2 0.3 0.4 0,5 0.6 0-7 0,8
Q.
O
Q.
"O
0)
1987 1991
.S 0.5 r
. 0,8
o
y^ 0
■ ^y^
0.4
"^t^o
■ y^
A 5^ Op
■ ^^^
A o^ o
0.3
.,^^r 0^
■ -^
0.2
J*o ■
■ ^y^
.^'^ • 02
° /^
0.1
-A
m^^^ ■■ •
10^
.^^•'
00
J^
0 0
LgjClr^
0.0
0.1 0.2 0.3 04 05 00 0.1 0.2 0.3 0.4 0.5 0.6 07 08
1988 1992
0.6
^^ 06
♦ ^/
0.5
^< 05
• j^
0.4
^V^
* ° y^^
0.3
A ^^ °^
^y"^
0.2
o^o
uy^
♦ 5^i" ■
01
X m ■ 01
A \< "* •
L^
k# ■
d^^ *
0 0
K0
0 n
00
0 1 0.2 0 3 0.4 0.5 0.6 0 0 0 1 0.2 0.3 0.4 0.5 0.6
Predicted proportion
Figure 3
Observed anc
predicted proportions of different age groups in monthly samples (f!>300) from commercial
catche.s in the
period 1985-2000. Each age group has its own symbol (see plot for 1985 and 1993). The diago-
nal is shown,
which is where the points should lie. Note that the range on the axes varies between years
according to the maximum values
NOTE Salthaug; Dynamic age-length keys
455
o
■c
1993
1994
00 01 02 03 04
1995
1996
1998
1999
2000
0 5 0 6 0 0 0 1
Predicted proportion
Figure 3 (continued)
456
Fishery Bulletin 101(2)
Literature cited
Foumier, D. A., J. R. Sibert, J. Majkowski, and J. Hampton.
1990. MULTIFAN a likelihood-based method for estimat-
ing growth-parameters and age composition from multiple
length frequency data sets illustrated using data for south-
em bluefin tuna (Thunnus maccoyii). Can. J. Fish. Aquat.
Sci. 47:301-317.
Fridrikson, A.
1934. On the calculation of age distribution within a stock of
cod by means of relatively few age determinations as a key
to measurements on a large scale. Rapp. P. V. Reun. Cons.
Int. ExplorMer 86:1-14.
Goodyear, C. P.
1997. Fish age determined from length: an evaluation of
three methods using simulated red snapper data. Fish.
Bull. 95:39-46.
Hayes, B.
1 993 . A statistical method for evaluating differences between
age-length keys with application to Georges Bank haddock,
Melanogrammus aeglefinus. Fish. Bull. 91:550-557.
Jorgensen, T.
1992. Long-term changes in growth of North-east Arctic
cod (Gadus morhua) and some environmental influences.
ICES J. Mar Sci. 49:263-277.
Kimura, D. K., and S. Chikuni.
1987. Mixtures of empirical distributions: an iterative appli-
cation of the age-length key. Biometrics 43:23-35.
Macdonald, P D. M., and T. J. Pitcher
1979. Age-groups from length-frequency data: a versatile
and efficient method of analyzing distribution mixtures. J.
Fish. Res. Board Can. 36:987-1001.
Pennington, M., L-M. Burmeister, and V. Hjellvik.
2002. Assessing trawl-survey estimates of frequency dis-
tributions. Fish. Bull. 100:74-80.
Schnute, J., and D. Foumier.
1980. A new approach to length-frequency analysis: growth
structure. Can. J. Fish. Aquat. Sci. 37:1337-1351.
Terceiro, M., and J. L. Ross.
1993. A comparison of alternative methods for the estima-
tion of age from length data for Atlantic coast bluefish
iPomatomus saltatrix). Fish. Bull. 91:534-549.
Tuljapurkar, S., and H. Caswell.
1997. Structured-population models in marine, terres-
trial, and freshwater systems, 570 p. Chapman & Hall,
London.
457
Effects of blood extraction on horseshoe crabs
iLimulus polyphemus)
Elizabeth A. Walls
Department of Fisheries and Wildlife Sciences
Virginia Polytecfinic Institute and State University
Blacksburg, Virginia 24061-0321
Present address: Center for Environmental Studies
Virginia Commonwealth University
1000 West Cary Street, Box 843050
Richmond, Virginia, 23284
Jim Berkson
Department of Fisheries and Wildlife Sciences
Virginia Polytechnic Institute and State University
Blacksburg, Virginia 24061-0321
E-mail address (for J Berkson. contact author); |berkson@vt.edu
the United States, mandated that all
biomedical companies actively bleed-
ing horseshoe crabs estimate mortal-
ity rates resulting from their bleeding
process (Schrading et al.'^). Because of
the unique methods of the different
biomedical companies, each company
was required to quantify its own rate
of mortality.
BioWhittaker, a CAMBREX com-
pany, is the largest producer of LAL.
In response to the ASMFC mandate,
BioWhittaker requested that Virginia
Tech conduct the mortality study for
their company. Our objective was to
determine horseshoe crab mortality
for a two-week period following the
bleeding process.
Methods
Horseshoe crabs ILimulus polyphemus)
are caught by commercial fishermen for
use as bait in eel and whelk fisheries
( Berkson and Shuster, 1999) — fisheries
with an annual economic value of $13
to $17 million (Manion et al.M. Horse-
shoe crabs are ecologically important,
as well (Walls et al., 2002). Migratory
shorebirds rely on horseshoe crab eggs
for food as they journey from South
American wintering grounds to Arctic
breeding grounds (Clark, 1996). Horse-
shoe crabs are also essential for public
health (Berkson and Shuster, 1999).
Biomedical companies bleed horse-
shoe crabs to extract a chemical used
to detect the presence of endotoxins
pathogenic to humans in injectable and
implantable medical devices (Novitsky,
1984; Mikkelsen, 1988). Bled horseshoe
crabs are returned to the wild, subject to
the possibility of postbleeding mortal-
ity. Recent concerns of overharvesting
have led to conflicts among commercial
fishermen, environmentalists acting on
behalf of the shorebirds, and biomedi-
cal companies (Berkson and Shuster,
1999; Walls etal., 2002).
In order to create an effective, sus-
tainable management policy for the
horseshoe crab resource, the comple-
tion of a stock assessment that incor-
porates human-induced mortalities is
necessary. A stock assessment is not
currently available because of a lack of
critical information on the horseshoe
crab population (Berkson and Shuster,
1999). One critical piece of information
needed is an estimate of the mortali-
ties involved in the biomedical bleed-
ing process. With an estimated 260,000
horseshoe crabs bled in 1997 (HCTC2),
the last year with data available, mor-
talities may not be negligible.
Five biomedical companies on the At-
lantic coast of the United States bleed
horseshoe crabs in the laboratory for
the production of Limulus Ameobocyte
Lysate (LAL). The horseshoe crabs are
caught by fishermen under contract to
biomedical companies, bled, then re-
turned to their point of capture.
The LAL test used to detect endo-
toxins in humans is derived from the
blue, copper-based blood of the horse-
shoe crab. Although alternate tests
exist for the detection of endotoxin, the
LAL test is the most effective because
it is capable of detecting as little as
one millionth of a billionth of a gram of
endotoxin (Mikkelsen, 1988). The LAL
test is now a standard test used to pro-
tect human health around the world,
and horseshoe crabs are the sole source
of LAL.
Each biomedical company maintains
its own procedures for harvesting
horseshoe crabs, extracting the horse-
shoe crabs' blood, releasing the bled
horseshoe crabs, and developing the
LAL substance. In 1998, the Atlantic
States Marine Fisheries Commission
(ASMFC), the Commission responsible
for horseshoe crab management in
We compared mortality rates between
horseshoe crabs that underwent the
bleeding process (bled) and horseshoe
crabs that were suitable to undergo
the bleeding process but were not
bled (unbled). Throughout the 1999,
2000, and 2001 bleeding seasons
(June through August), BioWhittaker
obtained horseshoe crabs by trawling
in the Atlantic Ocean off the coasts of
Chincoteague, Virginia, or Ocean City,
Maryland (or off both coasts). After
capture, the horseshoe crabs were
brought to BioWhittaker's bleeding
' Manion, M. M., R. A. West, and R. E, Uns-
worth. 2000. Economic assessment of
the Atlantic coast horseshoe crab fishery,
71 p. Division of Economics, U.S. Fish
and Wildlife Service, Arlington, VA.
2 HCTC (Horseshoe Crab Technical
Committee). 1998. Status of the horse-
shoe crab {Limulus polyphemus) popula-
tion of the Atlantic coast, 9 p -t- figures and
tables. Horseshoe Crab Technical Com-
mittee, Atlantic States Marine Fisheries
Commission. Washington, D.C.
3 Schrading, E., T. O'Connell, S. Michels, and
P. Perra. 1998. Interstate management
plan for horseshoe crab, 59 p. Atlantic
States Marine Fisheries Commission,
Washington D.C.
Manuscript accepted 6 November 2002.
Manuscript received 9 January 2003 at
NMFS Scientific Pubhcations Office.
Fish. Bull. 101:457-459 (2003).
458
Fishery Bulletin 101(2)
Table 1
Comparison of mortality rates between bled and unbled groups
of horseshoe crabs captured near Chincoteague, Virginia, and
Ocean City, Maryland,
1999-2001.
Dates
Unbled horseshoe crabs
Bled horseshoe crabs
No. of crabs
No. of crabs
% dead at
No. of crabs
No. of crabs
% dead at
monitored
monitored
that died
study end
monitored
that died
study end
8-22 Jul 99
10
0
0%
10
0
0%
22 Jul 99-5 Aug 99
10
0
0%
10
3
30%
19 Jun 00-3 Jul 00
30
0
0%
30
0
0%
7-21 Jul 00
30
0
0%
30
0
0%
1-15 Aug 00
30
1
3.3%
30
6
20%
6-20 Jun 01
30
0
0%
30
0
0%
20 Jun 01-04 Jul 01
30
0
0%
30
2
6.7%
15-29 Aug 01
30
0
0%
30
5
16.7%
Total
200
1
0.5'J
200
16
SV,
facility in Chincoteague, Virginia. At the bleeding facil-
ity, we randomly selected a predetermined number ( 10 in
1999, 30 in 2000 and 2001) of newly matured male horse-
shoe crabs (identified by pristine shell condition and the
presence of boxing-glove lower claws [Shuster*] ) from all
of the horseshoe crabs obtained in that day's trawls. We
selected newly matured male horseshoe crabs to minimize
covariance in our study. These horseshoe crabs were not
bled and served as a control in the experiment. They were
packed in coolers labeled "unbled," and set aside. The same
number of newly matured male horseshoe crabs were then
randomly selected from the remaining horseshoe crabs and
underwent BioWhittaker's normal bleeding process. Upon
completion of the bleeding process, the horseshoe crabs
were packed in coolers labeled "bled."
All coolers containing horseshoe crabs, both bled and
unbled, were immediately packed in an air-conditioned ve-
hicle and transported to the Virginia Seafood Agricultural
Research and Extension Center in Hampton, Virginia. The
horseshoe crabs were removed from the coolers and the
unbled horseshoe crabs were marked with external tags
to distinguish them from the bled horseshoe crabs. These
markings were unobtrusive and did not cause any undue
stress to the unbled horseshoe crabs. All of the horseshoe
crabs were placed in four replicated, flow-through holding
tanks, and equal numbers of bled and unbled horseshoe
crabs were held in each tank. The horseshoe crabs remained
in the tank system at Hampton for two weeks. Horseshoe
crabs were maintained in appropriate conditions (Brown
and Clapper, 1981), and monitored daily. Horseshoe crabs
that died during the two-week period were removed and
returned to the ocean at the time of their death.
At the conclusion of each two-week period, the status
of each horseshoe crab (dead or alive) was recorded. All
•* Shuster, C. N., Jr 1999. Managing the horseshoe crab re-
source: it's the adult age that counts, 32 p. Virginia Institute
of Marine Science, College of William and Marv, (Gloucester
Point, VA.
surviving horseshoe crabs were removed from the tank,
placed in coolers, packed in an air-conditioned vehicle, re-
turned to BioWhittaker's bleeding facility in Chincoteague,
Virginia, and returned to the Atlantic Ocean in accordance
with BioWhittaker's standard operating procedures. This
procedure was repeated eight times during summers 1999,
2000, and 2001. The results from each of the replicates
were combined, and the overall percentage mortality was
calculated for the bled and unbled groups.
Using Fisher's exact test, we evaluated statistical sig-
nificance of differences in mortality between the bled and
unbled horseshoe crabs (Mehta and Patel, 1999). We then
calculated a 957? confidence interval for average differen-
tial mortality using the common odds ratio in the statistical
program StatXact (Mehta and Patel, 1999).
Results
A Fisher's exact test for statistical significance showed
differences between mortality rates in bled and unbled
horseshoe crabs (P=2.085E-04). Bled horseshoe crabs
(n=200) had an overall mortality rate of S'J't compared
to the 0.5% mortality rate of the unbled horseshoe crabs
(«=200; P<0.001) (Table 1). Thus, this study estimates
average differential mortality between bled and unbled
horseshoe crabs to be 7.5'y. The 95^^ confidence interval
for this average differential mortality ranges from 0.14%
to 38.1% as calculated with the common odds ratio (Mehta
and Patel, 1999).
Discussion
Our results indicate that horseshoe crab mortality due to
bleeding is relatively low. Two small-scale studies had pre-
viously estimated postbleedmg mortality. Rudloe (1983),
observing bled and unbled horseshoe crabs in a penned
cove in Florida, found that bleeding increased mortality by
NOTE Walls and Berkson: Effects of blood extraction on Limulus polyphemus
459
10% during the first year after bleeding, and 11% during
the second year. Thompson (1998) estimated that mortal-
ity associated with LAL processing was 15% during the
first week following blood extraction by observing bled and
unbled horseshoe crabs in tanks in South Carolina.
Each LAL producer has a unique bleeding method, meth-
od of capture, distance and method of travel to the bleed-
ing laboratory, a different holding time and conditions, and
method of return of the bled crab that is most appropriate
to that company's setting and situation. The results found
in this study reflect those of BioWhittaker and may not be
reflective of other companies' procedures.
We examined the survival of the horseshoe crabs in a
controlled environment (tank), as opposed to their natural
environment. Our survival rate for horseshoe crabs may
not reflect the survival rate of horseshoe crabs returned
to the wild. Transfer and holding processes induce stress
on the horseshoe crabs. Thus, the survival of the bled
horseshoe crabs could be compromised by translocation
and confinement in tanks. However, the tank environment
may provide protection for horseshoe crabs when they are
in a weakened state and are more susceptible to predation
following blood-extraction.
Further, this study looked only at newly matured male
horseshoe crabs in an attempt to minimize variation of ex-
ternal influences, so that the only difference between the
two groups was whether or not they underwent the blood
extraction process. Additional studies should examine dif-
ferences in mortality in other age and sex classes.
The Food and Drug Administration estimates that
260,000 horseshoe crabs were caught, bled, and returned
by biomedical companies when last reported in 1997
(HCTC^). Assuming the 7.5% mortality rate found in
our study is applicable to each biomedical company, and
assuming that the number harvested for the biomedical
companies has stayed relatively constant, we estimate
that approximately 18,750 horseshoe crabs die yearly as
a result of the biomedical procedure. In comparison, the
commercial fishery reported landings of 5,543,000 pounds
in 1999 and 3,756,000 pounds reported in 2000, all with a
100% mortality rate (NMFS, 2002). In the overall picture
of the magnitude of horseshoe crabs caught and the associ-
ated mortality rates, it is evident that the bleeding process
has a substantially smaller impact than the commercial
fishery on the horseshoe crab population. However, infor-
mation on both biomedical and commercial fishery-induced
mortality are necessary to determine the total harvest mor-
tality of horseshoe crabs.
The information presented in this study provides an
estimate of the postbleeding mortality rate, an element
^ HCTC (Horseshoe Crab Technical Committee). 1998. Status
of the horseshoe crab (Limulus polyphemus) population of the
Atlantic coast, 9 p. + figures and tables. Horseshoe Crab Tech-
nical Committee, Atlantic States Marine Fisheries Commission.
Washington, D.C.
of human-induced mortality on horseshoe crabs. This is
one critical piece of information required to conduct a
stock assessment and to develop an effective management
strategy.
Acknowledgments
The authors wish to thank Carl N. Shuster Jr. and William
McCormick for their helpful advice on the design of this
study. Michael Schwarz, Ryan Cool, and Michael Jahnke
of the Virginia Seafood Agricultural Research and Exten-
sion Center, a unit of Virginia Tech, provided the facilities
for holding the horseshoe crabs and maintained them.
Funding for this study was provided by BioWhittaker,
a CAMBREX company. We especially thank Tammy
Newcomb and Michael Vaughan for their helpful advice
throughout all stages of this study.
Literature cited
Berkson, J., and C. N. Shuster Jr
1999. The horseshoe crab: the battle over a true multiple use
resource. Fisheries 24:6-10.
Brown, G. G., and D. L. Clapper.
1981. Procedures for maintaining adults, collecting gametes,
and culturing embryos and juveniles of the horseshoe crab,
Limulus polyphemus. In Laboratory animal management:
marine invertebrates, p. 268-290. National Academy
Press, Washington, D.C.
Clark, K.
1996. Horseshoecrabsandtheshorebirdconnection. //! Pro-
ceedings of the horseshoe crab forum: status of the resource.
(J. Farrell and C. Martin, eds.), p. 23-25. Univ. Delaware
Sea Grant College Program, Lewes, DE.
Mehta, C, and N. Patel.
1999. StatXact 4 Windows. CYTEL Software Corporation,
Cambridge, MA.
Mikkelsen, T.
1988. The secret in the blue blood, 125 p. Science Press,
Beijing, China.
NMFS (National Marine Fisheries Service).
2002. Annual commercial landing statistics. Available on-
line at www.nmfs.gov. Accessed 03/01/02.
Novitsky, T. J.
1984. Discovery to commercialization: the blood of the horse-
shoe crab. Oceanus 27:13-18.
Rudloe, A.
1983. The effect of heavy bleeding on mortality of the horse-
shoe crab, Limulus polyphemus, in the natural environ-
ment. J. Invert. Path. 42:167-176.
Thompson. M.
1998. Assessments of the population biology and critical
habitat for the horseshoe crab, Limulus polyphemus. in the
South Atlantic Bight. M.Sc. thesis, 50 p. Univ. Charles-
ton, Charleston, SC.
Walls, E. A., J. Berkson, and S. A. Smith.
2002. The horseshoe crab, Limulus polyphemus: 200 million
years of existence, 100 years of study. Rev. Fish. Sci. 10(1):
39-73.
460
Erratum
Fishery Bulletin 98(1):127-138 (2000).
Seyoum, Seifu, Michael D. Tringali, Theresa M. Bert,
Doug McElroy, and Rod Stokes
An analysis of genetic population structure in red drum
(Sciaenops ocellatus) based on mtDNA control region
sequences
On page 128 (right column, second paragraph) the authors wrote
■". . . state agencies in Alabama. Florida, South Carolina, and
Texas studied the feasibility of stock enhancement as a means of
supplementing wild populations." Later in the same paragraph,
they also stated that ". . . because broodstock for large-scale
enhancement programs along the Atlantic seaboard have been
obtained from Mosquito Lagoon and nearby esturaries. there
is a potential for artificial genetic exchange between putatively
separate gene pools (e.g. those of Mosquito Lagoon and the
Carolinas)."
The authors would like to clarify that the South Carolina
Department of Natural Resources u.ses only locally obtained red
drum for broodstock in its stocking programs. It was culturisls
for private facilities in South Carolina who used broodstock from
Florida Mosquito Lagoon to produce red drum for worldwide
distribution. The risk posited in our Introduction, as it related
to the Carolinas, pertained to escapement or mishandling of the
imported broodstock and their progeny.
Fishery Bulletin 101(1)
461
Superintendent of Documents Publications Order Form
*5178
I I YES, please send me the following publications:
Subscriptions to Fishery Bulletin
for $55.00 per year ($68.75 foreign)
The total cost of my order is $ .
. Prices include regular domestic
postage and handling and are subject to change.
(Company or Personal Name)
(Please type or print)
(Additional address/attention line)
(Street address)
(City, State, ZIP Code)
(Daytime phone including area code)
(Purchase Order No.)
Please Choose Method of Payment:
I I Check Payable to the Superintendent of Documents
□ GPO Deposit Account | | | | | I I |— H
I I VISA or MasterCard Account
(Credit card expiration date)
To fax
your orders
(202) 512-2250
(Authorizing Signature)
Mail To: Superintendent of Documents
PO. Box 371954, Pittsburgh, PA 15250-7954
Thank you for
your order!
^50
Fishery Bulletin
Guidelines for contributors
Content of papers
Articles
Articles are reports of 10 to 30 pages (double
spaced) that describe original research in one or
a combination of the following fields of marine
science: taxonomy, biology, genetics, mathematics
(including modeling), statistics, engineering, eco-
nomics, and ecologj'.
Notes
Notes are reports of 5 to 10 pages without an
abstract that describe methods and results not
supported by a large body of data. Although all
contributions are subject to peer review, responsi-
bility for the contents of articles and notes rests
upon the authors and not upon the editor or the
publisher It is therefore important that authors
consider the contents of their manuscripts care-
fully. Submission of an article is un-derstood to
imply that the article is original and is not being
considered for publication elsewhere. Manuscripts
must be written in English. Authors whose native
language is not EngHsh are strongly advised to
have their manuscripts checked for fluency by
English-speaking colleagues prior to submission.
Preparation of papers
Text
Title page should include authors' full names and
mailing addresses (street address required) and
the senior author's telephone, fax number, e-mail
address, as well as a list of key words to describe the
contents of the manuscript. Abstract must be less
than one typed page (double spaced) and must not
contain any citations. It should state the main scope
of the research but emphasize the author's con-
clusions and relevant findings. Because abstracts
are circulated by abstracting agencies, it is impor-
tant that they represent the research clearly and
concisely. General text must be typed in double-
spaced format. A brief introduction should state the
broad significance of the paper; the remainder of
the paper should be divided into the following sec-
tions: Materials and methods, Results, Discussion
(or Conclusions), and Acknowledgments. Headings
within each section must be short, reflect a logical
sequence, and follow the rules of multiple subdi-
vision (i.e. there can be no subdivision without at
least two subheadings). The entire text should be
intelligible to interdisciplinar\' readers; therefore,
all acronyms and abbreviations should be written
out and all lesser-known technical terms should be
deflned the first time they are mentioned. The
scientific names of species must be written out the
first time they are mentioned; subsequent mention
of scientific names may be abbreviated. Follow Sci-
entific style and format: CBE manual for authors,
editors, and publishers (6th ed.) for editorial style
and the most current issue o( the American Fish-
eries Society's common and scientific names of
fishes from the United States and Canada for
fish nomenclature. Dates should be written as fol-
lows: 11 November 1991. Measurements should be
expressed in metric units, e.g. metric tons (t). The
numeral one (1) should be typed as a one, not as a
lower-case el (1).
Footnotes
Use footnotes to add editorial comments regarding
claims made in the text and to document unpub-
lished works or works with local circulation. Foot-
notes should be numbered with Arabic numerals
and inserted in 10-point font at the bottom of the
first page on which they are cited. Footnotes should
be formatted in the same manner as citations.
If a manuscript is unpublished, in the process
of review, or if the information provided in the
footnote has been conveyed verbally, please state
this information as "unpubl. data," "manuscript
in review," and "personal commun.," respectively.
Authors are advised wherever possible to avoid ref-
erences to nonstandard literature (unpublished lit-
erature that is difficult to obtain, such as internal
reports, processed reports, administrative reports,
ICES council minutes, IWC minutes or working
papers, any "research" or "working" documents,
laboratory reports, contract reports, and manu-
scripts in review). If these references are used,
please indicate whether they are available from
NTIS (National Technical Information Service) or
from some other public depository. Footnote format:
author (last name, followed by first-name initials);
year; title of report or manuscript; type of report
and its administrative or serial number; name and
address of agency or institution where the report is
filed.
Literature cited
The literature cited section comprises works that
have been published and those accepted for pub-
lication (works in press) in peer-reviewed jour-
nals and books. Follow the name and year system
for citation format. In the text, write "Smith and
Jones (1977) reported" but if the citation takes
the form of parenthetical matter, write "(Smith
and Jones. 1977)." In the literature cited section,
list citations alphabetically by last name of senior
author: For example. Alston, 1952; Mannly. 1988;
Smith, 1932; Smith, 1947; Stalinsky and Jones.
1985- Abbreviations of journals should conform
to the abbreviations given in the Serial sources
for the BIOSIS previews database. Authors are
responsible for the accuracy and completeness of
all citations. Literature citation format: author
(last name, followed by first-name initials); year;
title of report or article; abbreviated title of the
journal in which the article was published, volume
number, page numbers. For books, please provide
publisher, city, and state.
Tables
Tables should not be excessive in size and must be
cited in numerical order in the text. Headings in
tables should be short but ample enough to allow
the table to be intelligible on its own. All unusual
symbols must be explained in the table legend.
Other incidental comments may be footnoted (use
italic arable numerals for footnote markers). Use
asterisks only to indicate probability in statistical
data. Place table legends on the same page as the
table data. We accept tables saved in most spread-
sheet software programs (e.g. Microsoft Excel).
Please note the following:
• Use a comma in numbers of five digits or more
(e.g. 13,000 but 3000).
• Use zeros before all decimal points for values
less than one (e.g. 0.31).
Figures
Figures include line illustrations, computer-gener-
ated line graphs, and photographs (or slides). They
must be cited in numerical order in the text. Line
illustrations are best submitted as original draw-
ings. Computer-generated line graphs should be
printed on laser-quality paper Photographs should
be submitted on glossy paper with good contrast.
All figures are to be labeled with senior author's
name and the number of the figure (e.g. Smith,
Fig. 4). Use Helvetica or Arial font to label ana-
tomical parts (line drawings) or variables (graphs)
within figures; use Times Roman bold font to label
the different sections of a figure (e.g. A, B, C).
Figure legends should explain all symbols and
abbreviations seen within the figure and should be
typed in double-spaced format on a separate page
at the end of the manuscript. We advise authors to
peruse a recent issue ot Fishery Bulletin for stan-
dard formats. Please note the following:
• Capitalize the first letter of the first word of
axis labels.
• Do not use overly large font sizes to label axes
or parts within figures.
• Do not use boldface fonts within figures.
• Do not create outline rules around graphs.
• Do not use horizontal lines through graphs.
• Do not use large font sizes to label degrees of
longitude and latitude on maps.
• Indicate direction of degrees longitude and
latitude on maps (e.g. 170°E).
•Avoid placing labels on a vertical plane
(except ony axis).
•Avoid odd (nonstandard) patterns to mark
sections of bar graphs and pie charts.
Copyright law
Fishery Bulletin, a U.S. government publication, is
not subject to copyright law. If an author wishes to
reproduce any part oi Fishery Bulletin in his or her
work, he or she is obliged, however, to acknowledge
the source of the extracted literature.
Submission of papers
Send four printed copies (one original plus three
copies) — clipped, no/ stapled — to the Scientific Edi-
tor, at the address shown below. Send photocopies
of figures with initial submission of manuscript.
Original figures will be requested later when the
manuscript has been accepted for publication.
Do not send your manuscript on diskette until
requested to do so.
Dr Norman Bartoo
National Marine Fisheries Service, NOAA
8604 La Jolia Shores Drive
LaJolla, CA 92037
Once the manuscript has been accepted for pub-
lication, you will be asked to submit a software
copy of your manuscript. The software copy should
be submitted in WordPerfect or Word format (in
Word, save as Rich Text Format). Please note that
we do not accept ASCII text files.
Reprints
Copies of published articles and notes are avail-
able free of charge to the senior author (50 copies)
and to his or her laboratory (50 copies). Additional
copies may be purchased in lots of 100 when the
author receives page proofs.
I
mbl
n •>
U.S. Department
of Commerce
Volume 101
Number 3
July 2003
T •>
Fishery
Bulletin
U.S. Department
of Commerce
Donald L Evans
Seaetary
National Oceanic
and Atmospheric
Administration
Vice Admiral
Conrad C. Lautenbacher Jr.,
USN (ret)
Under Secretary for
Oceans and Atmosphere
National Marine
Fisheries Service
William T Hogarth
Assistant Administrator
for Fisheries
The Fishery Bulletin (ISSN 0090-0656)
is published quarterly by the Scientific
Publications Office, National Marine Fish-
cries Service, NOAA, 7600 Sand Point Way
NE, BIN C15700, Seattle, WA 98115-0070.
Periodicals postage is paid at Seattle, WA,
and at additional mailing offices. POST-
MASTER: Send address changes for sub-
scriptions to Fishery Bulletin, Superin-
tendent of Documents, Attn.: Chief, Mail
List Branch, Mail Stop SSOM, Washing-
ton, DC 20402-9373.
Although the contents of this publica-
tion have not been copyrighted and may
be reprinted entirely, reference to source
is appreciated.
The Secretary of Commerce has deter-
mined that the publication of this peri-
odical is necessary according to law for
the transaction of public business of this
Department. Use of funds for printing of
this periodical has been approved by the
Director of the Office of Management and
Budget.
For sale by the Superintendent of
Documents, U.S. Government Printing
Office, Washington, DC 20402. Subscrip-
tion price per year: $45.00 domestic and
$56.25 foreign. Cost per single issue:
$28.00 domestic and $35.00 foreign. See
back for order form.
Scientific Editor
Dr. Norman Bartoo
Editorial Assistant
Sarah Shoffler
National Marine Fisheries Service, NOAA
8604 La Jolla Shores Drive
La Jolla, California 92037
Managing Editor
Sharyn Matriotti
National Marine Fisheries Service
Scientific Publications Office
7600 Sand Point Way NE, BIN C15700
Seattle, Washington 98115-0070
Editorial Committee
Dr Harlyn O. Halvorson
Dr. Ronald W. Hardy
Dr. Richard D. Methot
Dr Theodore W. Pietsch
Dr Joseph E. Pow/ers
Dr Harald Rosenthal
Dr Fredric M. Serchuk
Dr George Watters
University of Massachusetts, Boston
University of Idaho, Hagerman
National Marine Fisheries Service
University of Washington, Seattle
National Mahne Fisheries Service
Universitat Kiel, Germany
National Marine Fisheries Service
National Marine Fisheries Service
Fishery Bulletin web site: fishbull.noaa.gov
The Fishery Bulletin carries original research reports and technical notes on investigations in
fishery science, engineering, and economics. It began as the Bulletin of the United States Fish
Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery
Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through
volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing
through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system
began in 1963 with volume 63 in which papers are bound together in a single issue of the
bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a
periodical, issued quarterly. In this form, it is available by subscription from the Superintendent
of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in
limited numbers to libraries, research institutions. State and Federal agencies, and in exchange
for other scientific publications.
U.S. Department
of Commerce
Seattle, Washington
Volume 101
Number 3
July 2003
Fishery
Bulletin
Contents
The conclusions and opinions expressed
in Fishery Bulletin are solely those of the
authors and do not represent the official
position of the National Marine Fisher-
ies Service (NOAA) or any other agency
or institution.
The National Marine Fisheries Service
(NMFS) does not approve, recommend, or
endorse any proprietary product or pro-
prietary material mentioned in this pub-
lication. No reference shall be made to
NMFS, or to this publication furnished by
NMFS, in any advertising or sales pro-
motion which would indicate or imply
that NMFS approves, recommends, or
endorses any proprietary product or pro-
prietary material mentioned herein, or
which has as its purpose an intent to
cause directly or indirectly the advertised
product to be used or purchased because
of this NMFS pubUcation.
Articles
463-475 Brule, Thierry, Ximena Renan, Teresa Colas-Marrufo,
Yazmin Hauyon, Armin N. Tuz-Sulub,
and Christian Deniel
Reproduction in the protogynous black grouper
(Mycteroperca bonaci (Poey)) from the southern Gulf of Mexico
476—483 Comeau, Michel, and Manon Mallet
The effect of timing of tagging on streamer-tag recapture rates
for American lobster (Homarus americanus)
484—500 Diamond, Sandra L.
Estimation of bycatch in shrimp trawl fisheries: a comparison
of estimation methods using field data and simulated data
501 -513 Hanseiman, Dana H., Terrance J. Quinn II, Chris Lunsford,
Jonathan Heifetz, and David Clausen
Applications in adaptive cluster sampling of Gulf of Alaska
rockfish
Companion articles
514-534 Itoh, Tomoyuki, Sachiko Tsuji, and Akira Nitta
Migration pattems of young Pacific bluefin tuna
(Thunnus orientalis) determined with archival tags
535-544 Itoh, Tomoyuki, Sachiko Tsuji, and Akira Nitta
Swimming depth, ambient water temperature preference, and
feeding frequency of young Pacific bluefin tuna
{Thunnus orientalis) determined with archival tags
545-565 Jagielo, Thomas, Annette Hoffmann, Jack Tagart, and
Mark Zimmermann
Demersal groundfish densities in trawlable and untrawlable
habitats off Washington: implications for the estimation of
habitat bias in trawl surveys
Wootfi Holt, CceaoograpriiC '^
Libra'»
AUG 5 2003
Fishery Bulletin 101(2)
566-582 Loughlin, Thomas R., Jeremy T. Sterling, Richard L. Merrick, John L. Sease, and Anne E. York
Diving behavior of immature Steller sea lions (Eumetopias jubatus)
583-589 McBride, Richard S., Justin R. Styer, and Rob Hudson
Spawning cycles and habitats for ballyhoo (Hemiramphus brasiliensis) and balao (H. balao) in south Florida
590—602 Morato, Telmo, Encarnacion Sola, Maria P. Gros, and Gui Menezes
Diets of thomback ray (Raia clavata) and tope shark (Galeorhinus galeus) In the bottom longline fishery
of the Azores, northeastern Atlantic
603-613 Mullin, Keith, D., and Gregory L. Fulling
Abundance of cetatceans in the southern U.S. North Atlantic Ocean during summer 1998
614—626 Rogers-Bennett, Laura, Donald W. Rogers, William A. Bennett, and Thomas A. Ebert
Modeling red sea urchin (Strongylocentrotus franciscanus) growth using six growth functions
627—639 Skomal, Gregory B., and Lisa J. Natanson
Age and growth of the blue shark (Prionace glauca) in the North Atlantic Ocean
640—652 Teel, David J., Donald M. Van Doornik, David R. Kuligowski, and W. Stewart Grant
Genetic analysis of juvenile coho salmon (.Oncorhynchus kisutch) off Oregon and Washington reveals few
Columbia River wild fish
653—672 Terceiro, Mark
The statistical properties of recreational catch rate data for some fish stocks off the northeast U.S. coast
673—683 Williams, Ashley J., Campbell R. Davies, Bruce D. Mapstone, and Garry R. Russ
Scales of spatial variation in demography of a large coral-reef fish— an exception to the typical model?
Notes
684—692 Klimley, A. Peter, Salvador J. Jorgensen, Arturo Muhlia-Melo, and Sallie C. Beavers
The occurrence of yellowfin tuna (Thunnus albacares) at Espiritu Santo Seamount in the Gulf of Mexico
693—697 Landaeta, Mauricio R, Francisco J. Neira, and Leonardo R. Castro
Larvae of Dactylopsaron dimorphicum (Perciformes: Percophidae) from oceanic islands in the southeast Pacific
698-703 Pooler, Penelope S., David R. Smith, Robert E. Loveland, Mark L. Botton,
and Stewart F. Michels
Assessment of sampling methods to estimate horseshoe crab (Limulus polyphemus L.) egg density
in Delaware Bay
704-711 Powell, Allyn B.
Larval abundance, distribution, and spawning habits of spotted seatrout (Cynoscion nebulosus) in Florida Bay,
Everglades National Park, Florida
712-718 Zimmerman, Christian E., and Roger L. Nielsen
Effect of analytical conditions in wavelength dispersive electron microprobe analysis on the measurement of
strontium-to-calcium (Sr/Ca) ratios in otoliths of anadromous salmonids
719
Subscription form
463
Abstract— An analysis was made of
sexual pattern, spawning season, sizes
at sexual maturation, and sex change
in black grouper {Mycteroperca bonaci )
from the southern Gulf of Mexico.
Samples were taken between 1996 and
2000, from industrial and small-craft
commercial fisheries, in offshore and
inshore waters of the continental shelf
of the Yucatan Peninsula iCampeche
Bank), including the shallow waters of
National Marine Park Alacranes Reef
For all collected specimens (/! = 1229),
sex and maturation condition were
determined by histological analysis of
the gonads. The offshore sample con-
sisted of 75.1% females, 24.3% males,
and 0.6% transitional-stage fish. All
individuals collected from inshore
waters were females. Gonadal structure
and population structure characteris-
tics for Campeche Bank black grouper
were consistent with the characteristics
of monandric protogynous hermaphro-
dism for a serranid fish. Sexually active
males and females were observed year-
round, although ripening females, with
stage-Ill, -IV, and -V vitellogenic oocytes
in the ovaries, dominated in samples
taken between December and March.
In addition, peak occurrence of ripe-
running females with hyaline oocytes
or postovulatory follicles (or both) in
the ovaries was recorded in January
and February. A few precocious females
began spawning in October and Novem-
ber, and others were still in spawning
condition in May and June. Fifty per-
cent maturity of females was attained
at 72.1 cm fork length (FL). Median size
at sexual inversion was 103.3 cm FL,
and 50% of the females measuring 111.4
cm FL had transformed into males. The
southern Gulf of Mexico grouper fishery
was considered deteriorated and lacked
a well-defined management strategy.
Results of the present study provide
helpful information on black grouper
reproduction in this area and could help
Mexican authorities choose appropriate
management strategies for this fishery,
such as minimum size limit, closed fish-
ing season, and protection of spawning
aggregations.
Reproduction in the protogynous black grouper
(Mycteroperca bonaci (Poey)) from the
southern Gulf of Mexico
Thierry Brule
Ximena Renan
Teresa Colas-Marrufo
Yazmin Hauyon
Armin N.Tuz-Sulub
Centra de Investigacion y de Estudios Avanzados del IPN Unidad Merida
Antigua Carretera a Progreso km. 6
Apartado postal 73 Cordemex
Codigo Postal 97310 Merida
Yucatan, Mexico
E-mail address (lor T. Brule), ibrulecfimda.cinveslav.mx
Christian Deniel
Institut Universitaire Europeen de la Mer, Ressources Halieutiques-Poissons Manns
Unlversite de Bretagne Occidentale
Place Nicolas Copernic
Technopole Brest Iroise
29820 Plouzane, France
Manuscript approved for publication
11 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:463^75 (2003).
The black grouper {Mycteroperca bon-
aci) is one of the 20 most commonly
sought serranid fishes in the tropical
western Atlantic region (Sadovy, 1994).
The species ranges from Massachusetts
and Bermuda to southeastern Brazil
(Bohlke and Chaplin, 1993; Fischer,
1978; Bullock and Smith, 1991; Begossi
and Figueiredo, 1995). It is found on
irregular bottoms such as coral reefs,
drop-off walls, and rocky ledges, in
depths from 10 to at least 100 m (Roe,
1977; Manooch and Mason, 1987; Bull-
ock and Smith, 1991; Heemstra and
Randall, 1993; Huntsman et al., 1994).
According to Shapiro (1987), the sa-
lient feature of grouper reproduction
is protogynous hermaphroditism. The
first reasonable evidence of protogyny
in M. bonaci was published by Smith
(1959), although there have been other
occasional reports on black grouper re-
production (Erdman, 1956; Smith, 1961,
1971, 1972; Naranjo in Garcia-Cagide
et al., 1994). Systematic study of sexual
pattern and sexual maturation in the
species has only been carried out by
Garcfa-Cagide and Garcia (1996) in
Cuban waters and by Crabtree and
Bullock (1998) in Florida waters. This
grouper has been reported to form
spawning aggregations in the Gulf of
Mexico and Caribbean Sea (Fine, 1990;
Carter and Perrine, 1994; Domeier and
Colin, 1997; Eklund et al., 2000).
Black grouper is an important com-
mercial and recreational fin fish re-
source in Bermuda, southern Florida,
Cuba, the southern Gulf of Mexico, and
Venezuela (Manooch and Mason, 1987;
Cervigon, 1991; Heemstra and Randall,
1993; Claro et al., 1994). In the southern
Gulf of Mexico between 1989 and 1999,
groupers accounted for 18-307< of the
total offshore commercial marine re-
sources harvested from the Campeche
Bank (the continental shelf surround-
ing the northern coast of the Yucatan
Peninsula) and resources landed in
inshore waters off the state of Yucatan
(SEMARNAP, 2000a). At least 18 grou-
per species are commercially exploited
in this region — the most important of
these by catch number and weight are
red grouper {Epinephelus morio), fol-
lowed by black grouper and gag {Myc-
teroperca microlepis) (Colas-Marrufo et
al., 1998). Because grouper landings in
464
Fishery Bulletin 101(3)
Figure 1
Map of the Campeche Bank, Mexico, showing the geographic distribution of sampling locations (•) for black
grouper (Mycteroperca bonaci) observed during the period 1996-2000. Samping locations marked (O) are
where ripe-running female black grouper were caught. Sample locations: 1 = April 1996; 2 = April and May
1996; 3,4 = May 1996; 5 = November 1996; 6, 7 = December 1996; 8 = Januai-y 1997; 9 = January and Febru-
ary 1997; 11 = Febi-uary 1997; 12 = May 1997; 13 = June 1997; 14 = June and July 1997; 15 = July 1997; 16
= July and August 1997; 17, 18 = August 1997; 19 = September 1997; 20 = September and October 1997;
21 = October and November 1997; 22 = November 1997; 23 = December 1997; 24,25 = January 1998; 26 =
January and February 1998; 27 = February and March 1998; 29 = March 1998; 30,31 = June 1998; 32 = July
1998; 33 = August 1998; 34 = August and September 1998; 35 = September 1998; 36 = April 1999; 37,38 =
May 1999; Alacranes Reef was sampled November and December 1999, January, February, and August to
November 2000. No black giouper were caught in sample locations 10 (22''33'-88°24'W; February 1997) and
28 (22°30'N-89°30W; March 1998).
the Campeche Bank decreased between 1991 and 1997,
the Mexican government proposed management measures
to protect the grouper resource, but without considering
the biological characteristics and fishery aspects of each
exploited species (SEMARNAP, 2000a, 2000b). Given that
sustainable resource management is founded on stock as-
sessments and knowledge of the biology of exploited species
(Sadovy, 1997 ), more information on the biology of the most
abundant groupers from the southern Gulf of Mexico in
general, including Campeche Bank, is necessary to imple-
ment and refine management strategies.
This lack of knowledge is especially acute for Campeche
Bank black grouper. For example, although growth, feed-
ing, and reproduction of the Yucatan red grouper are well
documented, none of this information is available for the
black grouper in this region (Brule and Deniel, 1994; Brule
et al., 1994, 1999). This lack of information is alarming be-
cause M. bonaci can account for 40% of the grouper catch
by weight for some commercial vessels, and if this species
is not included in stock monitoring and reproduction stud-
ies, effective overall management of the southern Gulf of
Mexico grouper fishery could be seriously undermined
(Colas-Marrufo et al., 1998).
With the final aim of defining more accurate and effi-
cient management practices for the Campeche Bank gi-ou-
per fishery, we present analyses of sexual status, sexual
cycle, spawning season, size at sexual maturation, and
sex change for black gi-ouper from the southern Gulf of
Mexico.
Materials and methods
Black grouper were collected from commercial catches
taken from rocky bottoms in both offshore and inshore
waters of the Campeche Bank and in the shallow waters
of the Alacranes Reef complex. Alacranes Reef is the most
important complex of coral reefs located on the Yucatan
continental shelf Because of its high scientific and eco-
nomic potential, the Mexican government declared this reef
a National Marine Park in June 1994 (Fig. 1). In offshore
waters, black grouper (/i =880) were caught by the long-line
industrial fleet from 38 locations mainly situated in the
northeastern part of the Campeche Bank, at depths rang-
ing from 40 to 210 m, between April 1996 and May 1999.
In inshore waters, some specimens (n=39) were obtained
Brule et a\ : Reproduction in Mycteroperca bonaa
465
from the small craft-fleet, whose crew captured them using
spear guns at depths ranging from 4 to 20 m, in areas close
to the port of Progreso, between November 1998 and May
1999. On Alacranes Reef, black groupers («=206) were cap-
tured with spear guns by small-craft fishermen at depths of
10-12 m between November 1999 and February 2000.
Fork and standard lengths (FL, SL), whole and gutted
weights (WW, GW), and weight of gonads (gW), were re-
corded for all collected fish. All lengths reported in the pres-
ent study are fork length and all weights are gutted weight.
In discussion, total length data for Cuba (Garcia-Cagide
and Garcia, 1996) and Florida (Crabtree and Bullock,
1998) populations were converted to fork lengths by using
the fork-length to total-length relationship calculated by
Crabtree and Bullock ( 1998).
Criteria presented by Sadovy and Shapiro (1987) were
used to diagnose sequential hermaphroditism in black grou-
per of Campeche Bank. Sex by size-frequency distributions
for black groupers were compared by using the Kolmogorov-
Smirnov nonparametric test, and differences between male
and female mean fork lengths were analyzed by using a
one-tailed 2-test (if n>30) or a one— tailed <-test (if n<30). The
male-to-female ratio (M:F), excluding transitional-stage
fish (referred to as "transitional fish" in this article), was
calculated and Pearson chi-square or Yates's corrected chi-
square goodness-of-fit tests were carried out to determine if
sex ratio differed significantly from unity (Scherrer, 1984).
Significance level, «. was 0.05 in all instances.
Sex-dependent change in fin pigmentation, as described
by Crabtree and Bullock (1998), was examined in a sub-
set of fish (« = 104) caught from Alacranes Reef between
August and November 2000. The colors of the pectoral,
anal, dorsal, and caudal fins were recorded and histologi-
cal sections of the gonads were prepared and assessed for
sex identification.
For all the fish sampled in all locations, sex and sexual
development were determined by examination of the micro-
scopic structure of the gonads. These were preserved in
Bouin's fluid, embedded in Paraplast and sectioned to 6 ftm
thickness. Ovaries and testes sections were stained in Gabe
and Martoja's triple stain for light microscopy (Gabe, 1968).
Fish were identified as female, male, or as transitional.
Based on red grouper microscopic features for oogenesis
(Brule et al., 1999) and for spermatogenesis (Moe, 1969),
six descriptive stages were recognized in black grouper
ovaries and five in the testes. Histological sections of the
ovaries were also scanned for the presence of postovulatory
follicles and atretic oocjrtes in alpha or beta stages (Lam-
bert, 1970). Using the criteria defined by Smith ( 1959) and
Sadovy and Shapiro ( 1987 1, we considered individuals with
gonads containing primarily ovarian tissue, degenerating
or not, with few clusters of spermatocytes, spermatids, or
spermatozoa to be undergoing sexual inversion. According
to the sexual classes defined by Brule et al. ( 1999) for red
grouper, female and male black grouper were classified as
resting, ripening, ripe-running, or spent, and fish in the
process of sexual inversion were classified as transitional.
Using the histological features considered by Shapiro et al.
(1993) as sign of prior spawning activity for the red hind
(Epinephelus guttatus) we were able to distinguish resting
mature females from immature (virgin) females that had
never spawned.
Reproduction periodicity was evaluated for both sexes
by examining seasonal variations in the gonadosomatic
index (GSI=100 xgW/GW) and in the relative proportion of
individuals in each sexual class. Specimens from offshore
waters taken during different years were pooled by month,
and mean GSI values and percent frequencies of sexual
classes were generated monthly for a single year Immature
individuals were discarded from this analysis.
Size at which SO'X of females were sexually mature (L,;q)
was determined by using a binary logistic regression (SYS-
TAT statistical computer package for Windows, version 8.0,
SPSS Inc., Chicago, ID. For our analysis, resting mature,
ripening, ripe-running, and spent females were considered
as sexually mature individuals. Moreover, the minimum
size at which females become sexually mature (L^,;„) was
recorded, and the percentage of females of maximum
length at first maturity, ^^iJ^^nax with L^^^ ~ maximum
length of females recorded in samples, was determined
(Grimes, 1987). Sexual transition was analyzed by using a
binary logistic regression to estimate the length at which
50% of the females transformed to males (P^q) according to
Crabtree and Bullock (1998). The size range and median
size at which sex inversion occurs were estimated by fol-
lowing the procedures of Shapiro ( 1984). Furthermore, the
variation in size at sex change was analyzed by using two
ratios defined by Shapiro (1987): ratio 1, size range of tran-
sitional fish divided by maximum size of fish in samples;
and ratio 2, range of overlap in size of males and females
divided by maximum size of fish in samples.
Results
Size-frequency distributions
All individuals collected from inshore waters were females
ranging from 25.6 to 58.0 cm in length (Fig. 2). The off-
shore fish sample, which did not include the Alacranes Reef
sample, was composed of 75.1% females, 24.3% males, and
0.6% transitional fish. Females ranged in size from 57.0 to
123.5 cm, males from 86.0 to 132.0 cm, and transitional fish
from 99.0 to 121.5 cm. The Alacranes Reef sample was com-
posed of 95% females ranging in size from 46.0 to 100.0 cm
and 5% males from 97.0 to 135.0 cm. In the offshore sample,
the male size range differed significantly from that of the
females (Kolmogorov-Smirnov; 7!=875; P<0.05), and male
mean fork length (114.6 ±7.1 cm; mean ±SD) was greater
than female mean fork length (96.6 ±12.1 cm; one-tailed
2-test, n=875; P<0.05). Similar results for male and female
size ranges (Kolmogorov-Smirnov; n=206;P<0.05) and male
(115.7 ±12.9 cm) and female (67.6 ±11.2 cm) mean fork
lengths (one-tailed t-test, «=206; P<0.05) were observed for
black grouper from the Alacranes Reef sample.
Sex ratio
The male-to-female ratios were calculated for each 5-cm
size class from 25.1 to 135.0 cm length (Table 1). The
466
Fishery Bulletin 101(3)
140
120
100 •
80
60 ■
40
20
E
D Female (inshore waters)
D Female (offshore waters)
^tt^aie
■ Transitional
-P-P-
P r fn ,11 ,l,i ,1,1 ,H ,11 JL.ik.lb^
20 30 40 50 60 70 80 90 100 110 120 130
B
40 1
n Female
30 •
0Male
20-
10-
n
n •
n
,n,R], ,
,n,n, ,
JL,^
20
50 60 70 80 90
Fork length (cm)
100 110 120 130
Figure 2
Size-frequency distribution for females, transitional individuals, and males for
black grouper {Mycteroperca bonaci) caught between April 1996 and February
2000 in offshore and inshore waters of the Campeche Bank (A), and in shallow
waters of Alacranes Reef (B), Mexico.
sex ratios were female-biased in size classes less than
110.1 cm, did not differ significantly from a 1;1 sex ratio
in the 110.1-115.0 cm size class, and were male-biased in
size classes larger than 115.0 cm. The overall black grouper
sex ratio was 1:4, which differed significantly from unity
(Xf-=400.8,P<0.05).
Fin pigmentation
Of the 104 black grouper analyzed to detect gender-
associated color changes, 98 were females (size range
47.0-99.0 cm), five were males (99.0-115.0 cm), and one,
which presented previtellogenic oocytes and nests of sper-
matocytes and spermatozoa in its gonads, was classified as
transitional (99 cm). All males, as well as the transitional
specimen, displayed the male color phase with jet black
pigmentation on pectoral, anal, and caudal fins. Only 59f of
the females (size range 50.0-100.0 cm, «=5) had jet black
pigments on their fins.
Gonadal structure
All ovaries presented a central cavity with a germinal epi-
thelium forming the surface layer of a series of projecting
ovigerous folds or lamellae of the tunica albuginea.
Of the 225 males assessed histologically, 769f (;! = 170)
presented a membrane-lined central cavity in the testes.
This lumen remained unused in the transport of spermato-
zoa, and sperm ducts or sinuses within the gonadal capsule
were observed in ZT7< of the specimens (/;=84) (Fig. 3, A and
B). Previtellogenic oocytes (stages I and II ) remained in
the testes of 13''i of the males (?)=30), and only one of these
(107.0 cm) presented previtellogenic oocytes in degenera-
tion within lamellae in a fully developed testis dominated
by crypts of spermatocytes, spermatids, and spermatozoa.
Yellow bodies w(>re oi)sorved in the testes of 96'r of the
males.
Internal gonadal structure for the five black gi-oupers
classified as transitional was very similar to that of im-
Brule et al : Reproduction in Mycteroperca bonaci
467
Table 1
Number of sampled fish; proportion of females, males, and transitional-stage fish (transitional fish); and sex ratio by length class
for black gi-ouper {Mycteroperca bonaci) collected in the inshore and offshore waters of the Campeche Bank and shallow waters of
Alacranes Reef, Mexico, between April 1996 and February 2000. Alacranes collection = collection at the Alacranes Reef
Fork length
class (cm)
Females
Total
collected
Transitional
fish
Males
Sex ratio
(male:
female)
Offshore
collection
Inshore
collection
Offshore
collection
Alacranes
collection
Offshore
collection
Alacranes
collection
n
(%)
for total
collected
n
(%)
n
(%)
25.1-30.0
2
0
0
2
100.0
0
0
0
0
30.1-35.0
4
0
0
4
100.0
0
0
0
0
35.1-40.0
7
0
0
7
100.0
0
0
0
0
40.1-45.0
19
0
0
19
100.0
0
0
0
0
45.1-50.0
4
0
5
9
100.0
0
0
0
0
50.1-55.0
1
0
22
23
100.0
0
0
0
0
55.1-60.0
2
2
34
38
100.0
0
0
0
0
60.1-65.0
0
9
34
43
100.0
0
0
0
0
65.1-70.0
0
10
27
37
100.0
0
0
0
0
70.1-75.0
0
22
28
50
100.0
0
0
0
0
75.1-80.0
0
32
21
53
100.0
0
0
0
0
80.1-85.0
0
33
8
41
100.0
0
0
0
0
85.1-90.0
0
62
9
71
98.6
0
1
0
1
1.4
1:71
90.1-95.0
0
101
3
104
99.0
0
1
0
1
1.0
1:104
95.1-100.0
0
108
4
112
93.3
2
1.7
3
3
6
5.0
1:18.67
100.1-105.0
0
130
0
130
91.5
1
0.7
11
0
11
7.7
1:11.82
105.1-110.0
0
81
0
81
66.9
0
40
0
40
33.1
1:2.03
110.1-115.0
0
49
0
49
40.8
1
0.8
69
1
70
58.3
1:0.70*
115.1-120.0
0
15
0
15
22.7
0
47
4
51
77.3
1:0.29
120.1-125.0
0
7
0
7
19.4
1
2.8
28
0
28
77.8
1:0.25
125.1-130.0
0
0
0
0
0
12
2
14
100.0
130.1-135.0
0
0
0
0
0
2
1
3
100.0
Total
39
661
195
895
79.6
5
0.4
214
11
225
20.0
1:3.98
* Value did not differs signifi
cantly from 1:1
sex ratio (Xi^;
P>0.05).
mature or resting females. Stage! anii -II oocytes, yellow
bodies, and sometimes bundles of muscle and connective
tissue were present within the lamellae. Intermixed with
the female tissue, these gonads contained a few nests of
spermatogonia, spermatocytes, or spermatozoa, although
degeneration of female germinal tissue was not observed
(Fig. 3C). These transitional specimens were captured in
September, November, and December 1997 and in January
and March 1998.
Sexual cycle
Females captured from inshore waters during January,
February, March, May, November, and December were
immature and had low individual GSI values (GSI range
0.01-0.18%), and only oogonia and previtellogenic oocytes
were observed in their ovaries.
Mean GSI for mature females caught in offshore waters
began to increase in December (0.5%), reached a maximum
value in February (2.2%), and declined to a near minimum
level in March (0.7%) and April (0.6%) (Fig. 4). Highest in-
dividual GSI values for females were observed in October
(4.9%), December (6.0%), January (6.7%), and February
(9.6%). Mean GSI for males caught in offshore waters in-
creased in December (0.13% ) and January (0.14% ) — reach-
ing a maximum value in February (0.22%) and declining
from March (0.12%) to August (0.11%) (Fig. 4). Highest in-
dividual GSI values for males were observed in September
(0.43%) and February (0.39%).
Ripening females, with stage-Ill, -IV, and -V vitellogenic
oocytes in their ovaries, were observed year-round, but
dominated in collections made between December and
March (42-56% of females) (Fig. 5). Advanced vitellogenic
oocytes undergoing final oocyte maturation were noted on-
ly for some females captured between January and March
(Fig. 6A). Ripe-running females, with hyaline oocytes or
postovulatory follicles (or with both) in their ovaries, were
recorded between October and June, and peaked in occur-
468
Fishery Bulletin 101(3)
rence in January (28%) and February (52%)(Fig. 6B). The
gonads of 50 ripe-running females caught between Janu-
ary and April, and during June and November, contained
both postovulatory follicles and stages III-V vitellogenic
oocytes without sign of degeneration (Fig. 6C). Spent fe-
Figure 3
Photomicrographs of histological sections from male and
transitional black grouper iMycteroperca honati) gonads
collected from Campeche Bank, Mexico. (A) Section from a
1 14-cm-FL ripening male captured in July 1997, showing
lamellae, lumina, spermatogonia, and spermatocyte cysts,
and lamellae sinuses full of spermatozoa. (B) Section from a
r2'2-cm-FL ripe-running male captured in September 1997,
with sperm sinus full of spermatozoa in gonadal capsule.
(C) Section from a 113-cm-FL transitional fish captured in
January 1998, showing previtellogenic oocytes (stages I)
and scattered spermatocyte cysts. CX' = gonadal capsule; L =
lumen; LA = lamellae; PO = previtellogenic oocyte; SPC =
spermatocyte; SPG = spermatogonia; SPZ = spermatozoa;
SS = sperm sinus. Scale bars = 200 microns.
males, with atretic and remaining vitellogenic oocytes in
their gonads, were caught between January and August
{3-297f ). Resting mature females, with stages -I and -II oo-
cytes, bundles of muscle, and yellow bodies in their ovaries
were abundant in samples taken from May to November
(54-98%). Ripening or ripe-running males were recorded
year-round and spent males were observed in November
(4%), from January to March (10-40%), and from May to
July (8-22%).
Various females from the Alacranes Reef were ripening
in November (GSI range: 0.03-4.44%, n=l2), December
(GSI range: 0.22-7.18%, n=9) and February (GSI range:
0.10-6.61%, n=25). In February, some of them were ripe-
running, with postovulatory follicles in ovaries (GSI=1. 77%
and 1.91%, n=2) and others were spent (GSI=0.74% and
0.88%, n=2). Alacranes Reef males were ripening or ripe-
running in November, December, and February (GSI range:
0.03-0.44%, «=11).
Location and timing of spawning
Between April 1996 and February 2000, 61 ripe-running
females were caught at 11 offshore fishing locations situ-
ated in the northeastern part of the Campeche Bank (depth
range: 51-68 m), and from shallow waters of the Alacranes
Reef (8-10 m) (Fig. 1). All had vitellogenic oocytes (stages-
III-V) with hyaline oocytes or postovulatory follicles (or
with both) in their ovaries. Most of these females were
caught during, or close to, the new moon phase (Table 2).
Sizes at maturity and at sexual transition
The smallest mature female (L^j|jj=58.0 cm) was caught
in shallow waters of the Alacranes Reef and had stage-Ill
oocytes in its ovaries. Fifty-percent maturity of females
was attained at 72.1 cm in size and all females larger than
95.1 cm were mature (Fig. 7). Because the largest female
observed in samples was 123.5 cm (L^,^^), the percentage
of females at maximum length at first maturity was l^„^,J
Females changed sex between 85.5 and 125.0 cm in
length (the overlap zone between male and female sizes)
and the median size of sexual inversion was 103.3 cm. By
the time they attained a length of 111.4 cm, 50'7f of the
females in the sample had transformed into males (Fig. 8).
Size range of transitional fish (99.0-121.5 cm) was 17% of
maximum fish size ( 135.0 cm) (ratio 1, see "Materials and
methods"'section), and sex change occurred over 29'7f of the
maximum size observed for the species (ratio 2). Immature
males were not observed during the study.
Discussion
Sexual pattern
Previous research strongly suggests that sex reversal
occurs in M. bonaci (Smith, 1959, 1961; Garcia-Cagide and
Garcia, 1996; Crabtree and Bullock, 1998). Observations on
gonadal and population structure characteristics for black
Brule et al. : Reproduction in Mycteroperca bonaci
469
10-
Female
•
8-
•
c
o
o
•
a. 6-
•
•
•
• •
« 4-
t
■D
■/•.I
>
T3
•
IT)
u^m^
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
Female
I I I I I I I I I \ — I — r
0.5-1
0.4-
0.3-
0.2-
0.1 -
0.0
0.3-
0.2-
0.1
0.0-
Male
I I \ I 1 i I 1 — I — ( — I — I
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
Male
I I I — I — I — i — I — I — I — I — I — I — I
c,e Qp- ^d^ Qe'^ -^^ ^eP ^«J ^'' ^^-^ x,-:?^ ^,^'' ^^'^
Month
Figure 4
Individual and mean gonadosomatic index (GSI) values for female and male black grouper, (Mycteroperca bonaci) collected
in offshore waters of the Campeche Bank, Mexico, between April 1996 and May 1999. All calculated GSI for this study were
grouped to show monthly variations over a single year Vertical bars denote standard error. Numbers above the bars denote
the sample sizes.
grouper from the Campeche Bank are consistent with the
description of monandric protogynous hermaphroditism
previously used for this species. During the present study,
three of the five criteria suggested by Sadovy and Shapiro
(1987) for identifying protogyny in hermaphroditic fishes
were identified in black grouper: membrane-lined central
cavities in testes; sperm sinuses in the gonadal wall; and
transitional individuals. The five black grouper specimens
considered as transitional individuals (0.6% of sampled
offshore fish ) did not have degenerating ovarian tissue in
their gonads. For protogynous species, the degeneration of
ovarian tissue should logically accompany proliferation of
testicular tissue for the specimen to be termed "transi-
tional." However, according to Sadovy and Shapiro (1987),
the paucity of reported cases providing such descriptions
raises doubts as to whether transitional gonads display
this kind of histological profile. These observations also
may be a function of the fact that the incidence of tran-
sitional fish in field collections is generally relatively low,
as shown by the single transitional specimen (0.1% offish
collection) identified by Crabtree and Bullock (1998) in a
sample of the Florida black grouper population. Precocious
spermatocyte or sperm cysts in immature or functional
ovaries as observed by Smith (1964; 1965) and Bullock et
al. (1996) in coney (Cephalopholis fulva), graysby (Cepha-
lopholis cruentata ), and yellowedge grouper iEpinephelus
flavolimbatus) were not found in black groupers ovaries
during our study.
Other aspects of population structure indicating mo-
nandric protogynous hermaphrodism were also seen in M.
bonaci from the Campeche Bank. These included bimodal
size-frequency distributions, where males were larger than
females, female-biased sex ratios in size classes less than
1 10. 1 cm, and male-biased ratios in size classes larger than
115.0 cm. No male smaller than 86.0 cm was identified in
any of the samples. Despite these biases, the overall male-
to-female sex ratio calculated in our study (1:4) was less
skewed towards females than those reported by Garci'a-
Cagide and Garcia (1996) (1:30.3) and Crabtree and Bull-
ock (1998) (1:15.4). Notwithstanding, the sex ratio for black
470
Fishery Bulletin 101(3)
n Resting 0 Ripening ■ Ripe-running H Spent
100
80
60
40
20
62 101 88 66 64 50 29
i
ii
rnif
18 54 59 35
i!
^1
i
I
B
100
80
60
40
20
0
Sep Oct Nov Dec Jan Feb Mar Apr l^ay Jun Jul Aug
D Resting Q Ripening ■ Ripe-running Ca Spent
23 34 28 26 10 9 5 1 12 26 27 13
I
ill
8 1
III
I
u
I
Sep Oct Nov Dec Jan Feb Mar
Montti
Apr May Jun Jul Aug
Figure 5
Monthly percent frequencies of female (A) and male (B)
black grouper iMycteroperca honaci) collected in offshore
waters of the Campeche Bank, Mexico (April 1996-May
1999), sorted according to sexual classes over one year.
Number offish sampled is given for each month.
grouper sampled from Florida waters may not resemble
that for the entire population. Crabtree and Bullock ( 1998)
stated that they probably underestimated the number of
males in their sample because the large black grouper ex-
amined were eviscerated and could not be sexed. .
As observed for the first time by Crabtree and Bullock
( 1998) for black grouper from Florida waters, sexual dimor-
phism was displayed by the Campeche Bank population.
Notwithstanding, a low proportion of females possessed fin
pigmentation. Furthermore, it is also possible that indi-
viduals undergoing transition from female to male display
the male color phase. However, conclu.sions based on differ-
ences in fin pigmentation in male, female, and transitional
black grouper from the Campeche Bank are limited by the
small number of specimens examined for this purpose.
Spawning season
Sexually active black groupers from the Campeche Bank
(ripening females and ripening or ripe-running males)
were observed year-round. The monthly relative propor-
Figure 6
Photomicrographs of histological sections from female
black grouper iMycteroperca bonaci) gonads collected from
Campeche Bank, Mexico. (A) Section from a 97-cm-FL
ripe-running female captured in January 1998, showing
late vitellogcnic oocytes (stage V) undergoing final oocyte
maturation, showing yolk and oil globule fusion and germi-
nal vesicle migration; note presence of postovulatory follicle.
(B) Section from a 79-cm-FL ripe-running female captured
in February 1998, with hyaline oocyte (stage VI) and some
early and late vitellogenic oocytes (stages III and IV). (C)
Section from a 92-cm-FL ripe-running female captured in
November 1997, with early and late vitellogenic oocytes
(stages III and V) and postovulatory follicles. EVO = early
vitellogenic oocyte; GV = germinal vesicle; HO = hyaline
oocyte; LVO = late vitellogenic oocyte; OG = oil globule; PO
= prcvitellogenic oocyte; PDF = postovulatory follicle; YG =
yolk globule. Scale bars = 200 microns.
tion of individuals in each sex class and mean gonadoso-
matic indices showed a more consistent annual cycle for
Brule et al. : Reproduction in Mycteroperca bonaa
471
Table 2
Description of samples of ripe- running female black grouper {Mycteroperca 6onac/) collected in the offshore waters of the Campeche
Bank and shallow waters of Alacranes Reef, Mexico, between April 1996 and February 2000. Numbers refer to map locations in
Figure 1. HO = hyaline oocjrte; POF = postovulatory follicle. WW = whole weight.
Sampling dates
Fishing
location
Depth of
capture
(m)
Date of
full and new
moon
n
Histological
feature of
ovaries
Size range
if females
FL(cm)
WW (kg)
31
Oct 1997
21
51-55
15
1, 31 Oct
1
HO
101.0
17.2
21,22,27 Nov 1997
22
68
14
29
Nov
3
POF
75.0-91.5
5.3-11.2
6-12
Jan 1997
8
—
23
8
Jan
6
HO or POF
85.0-108.5
8.6-16.5
8-21
Jan 1998
24
59
12
28
Jan
10
POF or HO and POF
91.0-105.5
11.8-17.3
29,30
Jan 1998
25
59
12
28
Jan
2
POF
94.0; 101.0
12.4;14.8
1-5
Feb 1997
9
—
22
7
Feb
9
HO or POF
88.0-110.0
11.0-20.2
22-28
Feb 1998
27
55
11
26
Feb
17
HO and POF
76.0-117.0
6.4-25.1
3,6
Feb 2000
Alacranes reef
8-10
19
5
Feb
2
HO and/or POF
89.0; 88.0
8.0; 10.0
2,8
Mar 1998
27
55
12
27
Mar
2
POF
76.0; 84.0
6.2; 8,0
25
Mar 1998
29
—
12
27
Mar
2
POF
78.0; 94.0
5.4; 9.1
24
Apr 1996
2
55
17
3
Apr
1
POF
75.0
6.3
16
May 1999
37
67
30
15
May
1
POF
90.0
10.2
4-11
Jun 1997
13
64
20
5
Jun
5
OH or POF
69.5-97.0
4.7-13.9
100
« 75
50
25-
112 81 15
n = 895
1-50 = 72.1 cm 50
53 /•
_L
25 50 75 100
Fork length (cm)
125
Figure 7
Percent of mature females at length for black grou-
per (Mycteroperca bonaci) from inshore and offshore
waters of the Campeche Bank and shallow waters of
Alacranes Reef, Mexico (April 1996-February 2000).
Proportion of sexually mature females within each
size class is plotted with a binary logistic regression.
Line indicates length at SO'/c maturity (Lj,,). Number
of fish sampled is given for each size class.
100
2 4 7 9 38 37 53 72
1
19 23 43 50 41\
118%14
75
\
n = l120 Y^i
P50 = 1 1 1 .4 cm \
50
,
I119
25
-
Vee
%35
0
1 ... 1 ... 1 ... 1
V3
14««—
, 1 1 , i 1
20 40 60 80 100 120 140
Fork length (cm)
Figure 8
Percentage of female black grouper (Mycteroperca
bonaci) as a function of fork length. Samples were
taken between April 1996 and February 2000 in
inshore and offshore waters of the Campeche Bank
and shallow waters of Alacranes Reef Mexico. Pjo is
the length predicted by binary logistic regression at
which 50% of the sampled black grouper are female.
Number offish sampled is given for each size class.
females than for males. According to Sadovy ( 1996), ovaries
best reflect duration of fish spawning activity. Under this
assumption, the length of the black grouper spawning
season was evaluated by using Sadovy 's criteria (1996) as
defined to assess the duration of annual spawning in reef
fish species. Spawning season for black grouper from the
Campeche Bank was shown to extend from December to
March — a spawning month being defined as one in which
472
Fishery Bulletin 101(3)
50% or more of the sampled females have yolked oocytes.
Peak spawning occurred in February when 50"^ or more of
the sampled females had hyaline oocytes or postovulatory
follicles (or both) in their ovaries. In the case of spawning
seasonality evaluated by GSI data, peak spawning activity
was assigned to the months of January and February in
which the mean GSI of female attained was SOVr or more of
the maximum mean female GSI recorded during the study
(2.20% in February). A very few precocious females started
to spawn in October (1% of sampled fish) and November
(3'/i) and some were still in spawning condition during
May (4%) and June (8%). These results are consistent
with previous reports of the M. bonaci spawning season
in Puerto Rico, the Bahamas, Cuba, and Florida (Erdman,
1956; Smith, 1961, 1971, 1972; Garcia-Gagide et. al., 1994;
Garcia-Gagide and Garcia, 1996; Crabtree and Bullock,
1998). Spawning in Bermuda appears to be anomalous,
with reproductive activity extending from about early May
to early August ( Smith, 1971).
Spawning pattern
The authors did not observe spawning aggregations for
black grouper from the Campeche Bank as defined by
Domeier and Colin (1997) for tropical reef fishes. Black
grouper is reported to form spawning aggregations
between January and February in Belize, Honduras, and
Florida (Carter, 1989; Fine, 1990; Carter and Perrine, 1994;
Eklund et al., 2000). Only indirect methods of establishing
spawning occurrence, based on female gonadal condition,
were used in the present study. Ripe-running females were
caught at 11 offshore locations on the Campeche Bank and
in the shallow waters of National Marine Park Alacranes
Reef, during eight months of the year. The species seemed
to spawn preferentially at or around the new moon phase.
However, more information, such as direct observations of
spawning behavior and data on fish densities at aggrega-
tion sites during nonreproductive and reproductive periods,
is necessary to know precisely where and when adult M.
bonaci gather for spawning in the southern Gulf of Mexico.
The presence of oocytes at various stages of vitellogenesis
in the ovaries of ripening females and the co-existence of
postovulatory follicles and vitellogenic oocytes in those of
ripe-running females suggest that some individuals may
spawn more than once during the spawning season.
Sexual maturity and sex ciiange
The size at which 50% of females were sexually mature
(Lr,Q) was lower for black grouper from the Campeche Bank
(72.1 cm I than for those from Florida (82.6 cm) or Cuban
(84.4-108.7 cm) waters. The L„„„/L„,,,^ ratio indicated
that the females from the Campeche Bank reached first
maturity at a higher proportion of maximum length (47%)
than those from Florida waters (40% I. This spatial varia-
tion in size at first sexual maturity has also been observed
in female red grouper from southern and eastern Gulf of
Mexico (Brule et al., 1999).
The size at which 50% of the females transformed to
males (P5Q) was lower for black grouper from the Campeche
Bank (111.4 cm) than for those from Florida (119.9 cm).
Notwithstanding, the size range in which males overlapped
with females and the ratio-2 results were almost identical
for black grouper from Campeche Bank (39.5 cm and 29%)
and south Florida (39 cm and 26%) waters (Crabtree and
Bullock, 1998). According to Shapiro ( 1987), these data are
more consistent with a mechanism for behavioral induction
of sex change than with the idea that this process occurs at a
characteristic size or age for all members of a population.
Fisliery ciiaracteristics and fisliery management
As members of the warm-temperate and tropical reef
fish complexes, groupers have consistently proven highly
vulnerable to anything other than light levels of fishing
pressure (Sadovy, 1997). Because of their biological char-
acteristics, these species must be conservatively managed
to avoid rapid overfishing and stock collapse ( Sadovy, 1997;
Coleman et al., 2000). Some groupers from the western
Atlantic are even considered endangered and threatened
species (lUCN/SSCi). Assessment of Morris et al. (2000) and
Musick et al. (2000) led to the classification of black grouper
as a vulnerable species, that is, not critically endangered,
endangered, or threatened severely, but facing a high risk
of extinction in the wild in the medium-term future.
The trend in grouper catches in the state of Yucatan
has been one of progressive decline from an historical
maximum of 13,993 metric tons (t) in 1991, to 8556 t in
1997, followed by an increase to 11,045 metric tons in
2000 (SAGARPA, 2001). According to Monroy-Garcia et
al. (2001), catch increase reflects an increase in fishing
effort during the last three years. Recent assessments of red
grouper population from the Campeche Bank indicate that
the current biomass of exploited stock is well below that
of maximum biological productivity and that the fishery is
considered deteriorated (SEMARNAP, 2000a, 2000b). In
response to the multiple threats facing groupers in the Gulf
of Mexico, the U.S. and Mexican governments have imple-
mented regulations designed to either reduce or contain
effective fishing effort (input controls), or to restrict total
catch (output controls) to predefined limits. Notwithstand-
ing, grouper fishery regulations used in Mexican waters are
less restrictive than those imposed in U.S. waters. The U.S.
regulations currently consist of the following: an annual
commercial quota of 4445 t for the shallow-water grouper
complex, which includes the black grouper; commercial and
recreational minimum size limits of 61.0 cmTL and 55.9 cm
TL, respectively; a seasonal closure on commercial harvest
and prohibition on sale of this species from 15 February to
15 March; and a recreational aggregate daily bag limit of five
groupers per person (Gulf of Mexico Fishery Management
Council^). The Mexican regulations include license limita-
' 2001. lUCN/SSC (International Union for the Conservation of
Nature and Natural Resources/Species Sur\'ival Commission).
SSC Red Li.st ProRramme lUCN/SSC UK Office, 219c, Hunting-
don Road, Caml)rid},'e CB3 ODL, United Kingdom. Web page:
http://www.rodlist.org
'■^2001. GulfofMcxico Fishery Management Council. The Com-
mons at Rivcrgate, 3018 U.S. Hwy. 301 N., Suite 1000 Tampa,
Florida 33619-2266. Web page: http://www.gulfcouncil.org
Brule et al, : Reproduction in Mycteroperca bonaci
473
tion and a minimum legal total length of 30 cm for the Mexi-
can fleet, and an annual catch quota of 3900 t for the Cuban
fleet (SEMARNAP, 2000b). As can be seen, commercial and
recreational exploitation of the Campeche Bank grouper re-
source still lacks a well-deflned management strategy. Many
of the obstacles noted by Huntsman and Waters (1987) in
developing snapper-grouper management plans for the Gulf
of Mexico and U.S. South Atlantic are seen in the southern
Gulf of Mexico. For instance, grouper landings are not identi-
fied to the species level in Mexican fisheries statistic, but all
species are reported in the "mero " (grouper) category, which
includes eighteen species of the genera Cephalopholis, Epi-
nephelus, and Mycteroperca (Colas-Marrufo et al., 1998).
Information on grouper recreational catches is lacking, auid
biological data on the Campeche Bank species, especially on
reproduction, are either scarce or nonexistent. Moreover, the
30-cm-TL minimum size limit was applied to prevent the
marketing offish considered too small and was only related
to the growth overfishing problem.
Although landing trends by species for the southern Gulf
of Mexico grouper fishery remain undetermined, black
grouper along with red grouper and gag appear to be the
most abundant serranid fishes off the northern coast of the
Yucatan Peninsula. Colas-Marrufo et al. (1998) reported
black grouper to be second to red grouper in total number
(12%) and weight (40%) of grouper catches taken from the
Campeche Bank by some commercial fishing boats between
1996 and 1998. If a decrease in commercial grouper land-
ings from Mexican waters is confirmed in the near future,
protection measures such as regulation of specific catch
will be required for each of these three grouper species.
Results from the present study may aid in better esti-
mating and thus maintaining reproductive output in the
black grouper population from the Campeche Bank. With
these data, fishing regulations can be based on reproduc-
tive aspects. This information will make it possible to
propose a minimum size limit for this species near the size
at which 50% of females are sexually mature (72 cm FL)
(output control) and a closed season during peak spawn-
ing in February (input control) — both of which would help
prevent recruitment overfishing. Furthermore, if it is con-
firmed that black grouper from the Campeche Bank spawn
in the shallow waters of the Alacranes Reef complex, it
should be easy to temporarily ban fishing in the spawning
area(s) through enforcement of the regulations protecting
this National Marine Park.
Acknowledgments
This research was supported by grant 2184P-B9.507 from
the Consejo Nacional de Ciencia y Tecnologi'a (CONACYT);
grant C-1-99/062 from the Fondo Mexicano para la Conser-
vacion de la Naturaleza (FMCN); and the SEMARNAP/
S.S.S. "24 de Febrero'VCINVESTAV agreement for use of
the fishing vessel Unicap VII. For their assistance during
this study, we would like to thank R. Robles de Benito,
V. Alcantar-Cardenas, and M. Garduno-Andrade from
SEMARNAP/IPN-CRIPY (Merida/Yucalpeten); J. Peraza-
Menendez and M. Castillo-Martinez from CECADESU/
CREDES (Yucalpeten); J. Rodriguez-Felix from S.S.S.
"24 de Febrero" (Progreso); A.M. Pech from CONYUC fish
house (Progreso); J.L. Carrillo-Galaz and F Alvarez-Car-
rillo from the fishing cooperative SCPP "Pescadores de
Sisal" (Progreso); and L. Contreras-Garcia, Port Captain
of Progreso. For their assistance in all aspects of the
field collection, we are grateful to M. Sanchez-Crespo;
V. Durarte-Gracia, S. Mena-Gonzalez, C. Ureiia-Chio,
J. Hernandez-Viguegas, T. Ramirez-Hernandez, and P.
Mina-Coello. We also wish to thank L. Gus-Peltinovich for
assistance with photography.
Literature cited
Begossi, A., and J. L. de Figueiredo.
1995. Ethnoichthyology of southern coastal fishermen: cases
from Biizios island and Sepetiba bay (Brazil). Bull. Mar.
Sci. 56:710-717.
Bohlke, J. E., and C. G. H. Chaplin.
1993. Fishes of the Bahamas and adjacent tropical waters,
2""' ed., 771 p. Univ. Texas Press, Austin, TX.
Brule, T., and C.Deniel.
1994. Expose synoptique des donnees biologiques sur le
merou rouge Epinephelus morio (Valenciennes, 1828) du
golfe du Mexique. FAO Synopsis sur les peches 155,
39 p. FAO, Rome.
Brule, T., C. Deniel, T. Colas-Marrufo, and M. Sanchez-Crespo.
1999. Red grouper reproduction in the southern Gulf of
Mexico. Trans. Am. Fish. Soc. 128:385-402.
Brule, T, D. Ordaz Avila, M. Sanchez Crespo, and C. Deniel.
1994. Seasonal and diel changes in diet composition of juve-
nile red grouper (Epinephelus morio) from Campeche Bank.
Bull. Mar. Sci. 55:255-262.
Bullock, L. H., M. F Godcharles, and R. E. Crabtree.
1996. Reproduction of yellowedge grouper, Epinephelus
flavolimbatus, from the eastern Gulf of Mexico. Bull. Mar.
Sci. 59:216-224.
Bullock, L. H., and G. B. Smith.
1991. Seabasses (Pisces: Serranidae). Mem. Hourglass
Cruises, vol. .8, part 2, 243 p.
Carter, J.
1989. Grouper sex in Behze. Nat. Hist. 10:61-68.
Carter, J., and D. Perrine.
1994. A spawning aggregation of dog snapper, Lutjanusjocu
(Pisces: Lutjanidae) in Belize, Central America. Bull. Mar.
Sci. 55:228-234.
Cervigon, F.
1991. Los peces marines de Venezuela, vol. 1, 2nd ed., 423 p.
Fimdacion Cientifica Los Roques, Caracas.
Claro, R., J. Baisre, and J. P. Garcia-Arteaga.
1994. Evolucion y manejo de los recursos pesqueros. In
Ecologia de los peces marinos de Cuba (R. Claro, ed.),
p. 435-456. Centro de Investigaciones de Quintana Roo
(CIQRO), Chetumal, Quintana Roo.
Colas-Marrufo, T, T. Brule, and C. Deniel.
1998. Anahsis preliminar de las capturas de meros realizadas
a traves de unidades de la flota mayor en el sureste del Golfo
de Mexico. Proc. Gulf Caribb. Fish. Inst. 50:780-803.
Coleman, F. C, C. C. Koenig, G. R. Huntsman, J. A. Musick,
A. M. Eklund, J. C. McGovem, R. W. Chapman, G. R. Sedberry,
and C. B. Grimes.
2000. Long-lived reef fishes: the grouper-snapper complex.
Fisheries 25 (3):14-21.
474
Fishery Bulletin 101(3)
Crabtree, R. E., and L. H. Bullock.
1998. Age, growth, and reproduction of black grouper, Mycte-
roperca boitaci, in Florida waters. Fish. Bull. 96:735-753.
Domeier, M. L., and P. L. Colin.
1997. Tropical reef fish spawning aggregations: defined and
reviewed. Bull. Mar. Sci. 60:698-726.
Eklund, A. M., D. B. McClellan, and D. E. Harper.
2000. Black grouper aggregation in relation to protected
areas within the Florida Keys National Marine Sanctuary.
Bull. Mar. Sci. 66:721-728.
Erdman, D. S.
1956. Recent fish records from Puerto Rico. Bull. Mar. Sci.
Gulf Caribb. 6:315-340.
Fine, J. C.
1990. Groupers in love: spawning aggregations of Nassau
grouper in Honduras. The Explorers Journal, Fall 1990,
131-134.
Fischer, W.
1978. FAO species identification sheets for fisheries pur-
poses. Western Central Atlantic ( fishing area 31), vols. 1-7,
van pag. FAO, Rome.
Gabe, M.
1968. Techniques histologiques, 1113 p. Masson, Paris.
Garcia-Cagide, A.. R. Claro, and B. V. Koshelev.
1994. Reproduccion. In Ecologia de los peces marinos de
Cuba(R. Claro, ed.), p. 187-261. Centro de Investigaciones
de Quintana Roo (CIQRO), Chetumal, Quintana Roo.
Garcia-Cagide, A., and T. Garcia.
1996. Reproduccion de Mycteroperca bonaci y Mycteroperca
venenosa (Pisces: Serranidae) en la plataforma cubana.
Rev Biol. Trop. 44:771-780.
Grimes, C. B.
1987. Reproductive biology of the Lutjanidae: a review.
In Tropical snappers and groupers: biology and fisheries
management (J. J. Polovina and S. Ralston, eds.), p. 239-294.
Westview Press, Boulder, CO.
Heemstra, P. C, and J. E. Randall.
1993. FAO species catalogue, vol. 16: Groupers of the world
(Family Serranidae, Subfamily Epinephelinae). An anno-
tated and illustrated catalogue of the grouper, rockcod, hind,
coral grouper and lyretail species known to date. FAO
Fisheries Synopsis 125, 382 p. FAO, Rome.
Huntsman, G. R., J. Potts, and R. W. Mays.
1994. A preliminary assessment of the populations of seven
species of grouper (Serranidae, Epinephelinae) in the West-
ern Atlantic Ocean from Cape Hatteras, North Carolina to
the Dry Tortugas, Florida. Proc. Gulf Caribb. Fish. Instit.
43: 19.3-213.
Huntsman, G. R., and J. R. Waters.
1987. Development of management plans for reef fishes-
Gulf of Mexico and U.S. South Atlantic. In Tropical snap-
pers and groupers: biology and fisheries management (J.
J. Polovina and S. Ralston, eds.), p. 533-560. Westview
Press, Boulder, CO.
Lambert, J. G. D.
1970. The ovary of the guppy, Poecilia reticulata. The
atretic follicle, a corpus atrcticum or a corpus luteum prae-
ovulationis. Z. Zellforsch. Mikrosk. Anat. 107:54-67.
Manooch, C. S., and D. L. Mason.
1987. Age and growth of the Warsaw grouper and black
grouper from the southeast region of the United States.
Northeast Gulf Sci. 9:65-75.
Moe, M. A.
1969. Biology of the red grouper Epinephelus morio (Valen-
ciennes) from the eastern Gulf of Mexico. Fla. Dep. Nat.
Resour. Mar Res. Lab. Prof Pap. Sen 10, 95 p.
Monroy-Garcia, C, R. Burgos-Rosas, V. Moreno-Garcia, and
E. Gimenez-Hurtado.
2001. Informe de investigaciones conjuntas Mexico-Cuba
sobre el mero {Epinephelus morio, Valenciennes, 1828) en
el Banco de Campeche. Convenio de pesca Mxico-Cuba, 42 p.
Secretaria de Agricultura, Ganaderia, Desarrollo Rural,
Pesca y Alimentacion (Mexico) y Ministerio de la Industria
Pesquera (Cuba ). Institute Nacional de la Pesca, Progreso,
Yucatan.
Morris, A. V, C. M. Roberts, and J. P. Hawkins.
2000. The threatened status of groupers (Epinephelinae).
Biodivers. Conserv. 9:919-942.
Musick, J. A., M. M. Harbin, S. A. Berkeley, G. H. Burgess,
A. M. Eklund, L. Findley, R. G. Gilmore, J. T Golden,
D. S. Ha, G. R. Huntsman, J. C. McGovem, S. J. Parken
S. G. Poss, E. Sala, T. W. Schmidt, G. R. Sedberry, H. Weeks,
and S. G. Wright.
2000. Endangered Species-Marine, estuarine, and diadro-
mous fish stocks at risk of extinction in North america ( exclu-
sive of pacific salmonids). Fisheries (Bethesda) 25(11):
6-30.
Roe, R. B.
1977. Distribution of snappers and groupers in the Gulf of
Mexico and Caribbean Sea as determined from exploratory
fishing data. In Proceedings of the CoUoquim on snapper-
grouper fisheries ressources of the western Central Atlan-
tic Ocean (H. R. Bullis and A. C. Jones, eds.), p. 129-164.
Florida Sea Grant Program, Report 17, State University
System of Florida Sea Grant Program, Gainesville, FL.
Sadovy,Y.
1994. Grouper stocks of the western central Atlantic: the
need for management and management needs. Proc. Gulf
Caribb. Fish. Inst. 43:43-64.
1996. Reproduction of reef fishery species. In Reef fish-
eries (N. V. C. Polunin and C. M. Roberts, eds.), p.l5-
59. Chapman and Hall. London.
1997. Problems ofsustainability in grouper fisheries. Proc.
Fourth Asian Fish. Forum, p. 321-324. China Ocean Press,
Beijing.
Sadovy, Y., and D. Y. Shapiro.
1987. Criteria for the diagnosis of hermaphroditism in
fishes. Copeia 1987:136-156.
Scherrer, B.
1984. Biostatistique, 850 p. Gaetan Morin Editeun Bou-
cherville, Quebec.
SAGARPA (Secretaria de Agricultura, Ganaderia, Desarrollo
Rural, Pesca y Alimentacion).
200 1 . Anuario estadistico de pesca 2000, 268 p. SAGARPA,
Mexico.
SEMARNAP (Secretaria de Medio Ambiente, Recursos
Naturales y Pesca).
2000a. Anuario estadistico de pesca 1999, 271 p. SEMAR-
NAP, Mexico.
2000b. Sustentablilidad y pesca responsable en Mexico.
Evaluacion y manejo 1997-1998, 691 p. SEMARNAP-
IPN, Mexico.
Shapiro, D. Y
1984. Sex reversal and sociodemografic processes in coral
reef fishes. In Fish reproduction: strategies and tactics,
3"' ed. (G. W. Potts and R. J. Wootton, eds). p. 103-118.
Academic Press, London.
1987. Reproduction in groupers. In Tropical snappers and
groupers: biology and fisheries management (J. J. Polovina
and S. Ralston, eds.), p. 295-327. Westview Press, Boulder,
CO.
Brule et al. ; Reproduction in Mycteroperca bonaa
475
Shapiro, D. Y., Y. Sadovy, and M. A. McGehee.
1993. Periodicity of sex change and reproduction in the red
hind, Epinephelus guttatus, a protogynous grouper. Bull.
Mar. Sci. 53:1151-1162.
Smith, C. L.
1959. Hermaphroditism in some serranid fishes from Ber-
muda. Pap. Mich. Acad. Sci. 44:111-119.
1961. Synopsis of biological data on groupers {Epinephelus
and allied genera) of the Western North Atlantic. FAO
Fisheries Biology Synopsis 23, 61 p. FAO, Rome.
1964. Hermaphroditism in Bahamas groupers. Bull. Mus.
Nat. Hist., N.W. 73:42-47.
1965. The patterns of sexuality and the classification of ser-
ranid fishes. Am. Mus. Novit. 2207:1-20.
1971. A revision of the American groupers: Epinephelus and
allied genera. Bull. Am. Mus. Nat. Hist. 146:67-242.
1972. A spawning aggregation of Nassau grouper, Epineph-
elus strmtus (Bloch). Trans. Am. Fish. Soc. 101:257-261.
476
Abstract — Streamer tags are com-
monly used to study the ecology and
population biology of the American lob-
ster (Homarus americanus). Aquarium
observations suggest that streamer
tag loss, either through tag-induced
mortality or tag shedding, is related to
the molt stage of the lobster at the time
of tagging, and the molting event itself.
Tag-induced mortality, where lobsters
did not molt, occurred within eleven
and sixteen days following tagging for
lobsters tagged in postmolt (4%) and
late premolt (10%) stages, respectively;
whereas no lobsters tagged in early
premolt or intermolt stages died. Tag-
induced mortality at time of molting
was observed for lobsters tagged in
late premolt stage (11%), and tag shed-
ding was observed for lobsters tagged
both in early (25%) and late premolt
(11%) stages, but was significantly
higher (P=0.014) for lobsters tagged
in early premolt stages. Autopsies
revealed that lobsters died mainly of
organ perforations (hepato-pancreas
and pericardial sac) following the tag-
ging process, and rupture of the dorsal
thoraco-abdominal membrane during
the molting process. The total tag loss
was estimated at 4% for lobsters tagged
after molting, and 27%^ and 31% for lob-
sters tagged in early and late premolt
stages, respectively. There was no tag
loss for lobsters tagged in the intermolt
stage during four months of labora-
tory observations (July-October). To
minimize streamer tag loss, lobsters
should be tagged during the intermolt
or postmolt stage. Based on field stud-
ies, recapture rates for lobsters tagged
in premolt stage are always lower than
those of lobsters tagged in postmolt
stage. Furthermore, recapture rates
during the second year, for lobsters
that molt in the year following tagging,
were drastically reduced, and no lobster
was recaptured after four years at large.
Finally, to account for tag loss during
the first year at large, a minimal adjust-
ment of 24.9% (SD 2.9%) and 4.47, (SD
1.6%) for the recapture rate of lobsters
tagged immediately before and after the
molting season, respectively, is recom-
mended. Adjustments beyond one year
at large are not recommended for the
American lobster at this time.
Manuscript approved for publication
19 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:476-483 (2003).
The effect of timing of tagging on streamer-tag
recapture rates for American lobster
(Homarus americanus)
Michel Comeau
Manon Mallet
Department of Fisheries and Oceans
343 University Ave
Monaon, New Brunswick, Canada E1C 9B6
E-mail-address (for M Comeau) comeaumra'dfo-mpo gc ca
Tagging methods to study the move-
ment, growth, and exploitation rate
for the American lobster {Homarus
americanus) have improved over the
last 70 years. One major improvement
in the mid 1960s was the introduction
of an insertion tag called the "sphyrion
tag" that is anchored to muscle tissue
(Scarratt and Elson, 1965), instead of
body tags (Templeman, 1935) or cara-
pace-piercing tags (Wilder, 1953) used
earlier. By the late 1980s, the sphyrion
tag was replaced by another insertion
tag called the polyethylene "streamer
tag" (Landsburg, 1991; Moriyasu et al.,
1995) initially developed for shrimps
(Penaeus spp.; Marullo et al., 1976).
Insertion tags have the advantage
of being retained through a series of
molts, thus providing information on
long-term movement and more accurate
data on growth.
Tag loss could greatly bias the esti-
mate of population characteristics and
fishery parameters (Ricker, 1975). For
obtaining population estimates from
mark-recapture data, tag loss gener-
ally refers to the reduction of the initial
number of tagged animals by means
other than fishing. In a series of tagging
studies, Comeau et al. (1999) noticed a
constant pattern of lower tag recovery
rates for lobsters tagged in the premolt
stage than for lobsters tagged in the
postmolt stage. They indicated that the
level of fishermen participation (recov-
ery rate) could be a possible cause of
tag loss because they noticed a steady
decline of recapture rates where mul-
tiyear tagging studies were conducted.
However, because the same type of
streamer tag was used and each lobster
was handled individually, the fishermen
participation could not explain the dif-
ference between recapture rates for
lobsters tagged in premolt and postmolt
stages for a given year; hence possible
tag loss at molting was suspected as
the cause of the lower recapture rates
(Comeau et al., 1999). Furthermore, in
an attempt to estimate mortality rates,
Comeau and Mallet (2001) used a
mark-recapture model and simulations
to evaluate the best estimator. They
concluded that the level of tag loss is
high and could be a serious problem for
estimating fishery parameters for the
American lobster if information on tag
loss is not available.
Moriyasu et al. (1995) showed that
sphyrion tag loss for lobsters held in
aquaria varies between 3% and 23%
depending on the molt stage at tag-
ging. They also mentioned that lobsters
tagged with sphyrion tags showed a
significantly lower return rate (19%)
than lobsters tagged with streamer tags
(44%) in a recapture study in the field
and suggested a possible lower level of
tag loss among lobsters tagged with
streamer tags. However, they did not es-
timate the tag loss for streamer-tagged
lobsters. Recently, Rowe and Haedrich
(2001) showed that the shedding rate
for streamer tags in the field could
reach 18% (40% for molted animals and
11% for nonmolting animals) after 8-12
months based on double tagging with
a secondary carapace marking. They
also found that streamer tag shedding
was not related to sex or size, but they
did not study the level of tag-induced
mortality.
Various causes can reduce the initial
number of tagged animals, mainly tag
shedding, tag-induced mortality, and
Comedu and Mdllet Effect of timing of tagging on tag recapture rates for Homarus americanus
477
death from natural causes (Beverton and Holt, 1957). In
the present article, "tag shedding" refers to the physical
detachment of the tag from a lobster and "tag-induced
mortality" refers to the actual death of a lobster caused by
the tagging process.
In the southwestern Gulf of St. Lawrence, lobsters are
harvested either in the spring prior to the July-August
molting season, or in late summer and early fall (early-
August to early October, partially during and shortly after
molting) (Comeau and Savoie, 2001). The purpose of our
study was to estimate the level of streamer tag loss for
the American lobster tagged before and after the summer
(July- August) molting season by using aquarium observa-
tions. From the results of our aquarium study, we deter-
mined adjustments of the recapture rate in relation to the
molt stage at the time of tagging.
Materials and methods
Aquarium observations
Two experiments were carried out at the "Aquarium et centre
marin" (New Brunswick Department of Agriculture, Fisher-
ies and Aquaculture) in Shippagan, New Bi-unswick, with
lobsters captured in Baie des Chaleurs 147°52'N; 64°52'W).
Because the main focus of our study was to investigate tag
loss in relation to the molting stage, only males were con-
sidered because they have a higher probability of molting
annually compared to sexually mature females (Comeau
and Savoie, 2001). All lobsters caught were brought to the
laboratory where carapace length (CD and shell rigidity
were recorded (the latter with a durometer) (Comeau and
Savoie, 2001 ). In both experiments, lobsters were tagged by
the same person to avoid variability in tagging procedure
following the technique described by Moriyasu et al. (1995).
Streamer tags manufactured by Hallprint Pty. Ltd. (15
Crozier Rd, Victor Harbor, South Australia, 5211 Australia)
were used. As is routinely done in our tagging studies in
the field (Comeau et al., 1998, 1999), tagged lobsters were
kept in a holding tank for a minimum of 30 min following
tagging, and dead lobsters were removed from the experi-
ment. Tagged lobsters were then transferred to large tanks
partitioned with 25 x 25 cm individual compartments. These
tanks were supplied with running seawater at ambient
temperature. Lobsters were fed rainbow smelt (Osmerus
mordax) twice a week. Lobsters were examined three times
a day in July and August, and on a daily basis for the rest
of the experiment. The date of tag shedding, of molting, or
of death was recorded for each lobster.
Limited aquarium space prevented the use of a control
group of untagged lobsters. However, autopsies were per-
formed on all lobsters that died in the course of the experi-
ments in order to identify the cause of death. The following
tagging traumas, causing death, were identified (Krouse
and Nutting, 1990): 1) perforation of vital organs, such as
the pericardial sac and the hepato-pancreas; 2) rupture
of the thoraco-abdominal membrane; and 3) necrosis or
infection of lobster tissue at both the point of entry emd
exit of the tag.
The first experiment began on 23 June 1998 before the
summer molting season with the tagging of 229 hard-shell
male lobsters ranging between 66 and 78 mm CL. To avoid
unnecessary manipulation of the lobsters, the molt stage
at tagging was estimated by the number of days between
tagging and molting. Individuals that molted within 30
days of and 30 days after tagging were considered lobsters
tagged in late premolt and early premolt stages (Aiken,
1980), respectively. The molt stage of lobsters that died dur-
ing the experiment was determined by observations of the
pleopods (Aiken, 1980). A total of 191 male lobsters were
tagged in premolt stage (56 in early and 135 in late premolt
stage) and 38 in the intermolt stage (lobsters that did not
molt over the entire experiments). Observations of premolt
tagged lobsters that molted in June and July ended on 8
September 1998. Observations of the remaining lobsters
tagged on 23 June 1998 ended on 30 October 1998. Before
releasing the lobsters in the water, their wounds from the
tag insertion were examined for infection.
The second experiment began 9 September 1998 with the
tagging of 187 soft-shell male lobsters in the postmolt stage
based on shell condition criteria (Aiken, 1980; Comeau and
Savoie, 2001). They ranged in size between 66 and 83 mm
CL. Observations of postmolt tagged lobsters ended on 30
October 1998.
Field studies
Comparison of recapture rates from field studies Six
tagging studies were carried out after the fishing season
(May-June) in Caraquet, New Brunswick, between 1994
and 1996. Male and female lobsters were captured, tagged,
and released on the commercial fishing grounds. For each
year, tagging was done before (in early July with hard-
shell lobsters) and after (in mid-September with soft-shell
lobsters) the molting season. In 1995, the July tagging was
delayed one week because the lobster fishing season ended
on 7 July instead of 30 June. In 1996, tagging was carried
out in early October instead of mid-September because of
bad weather. As with the aquarium experiments, the same
person performed the tagging procedure. Tagged lobsters
were kept in a holding tank for a minimum of 30 min
following tagging and any dead lobsters were removed.
Finally, an awareness campaign described in Comeau et
al. (1998) was conducted to maximize the participation
of fishermen in reporting tagged lobsters as tag recovery
took place during the fishing seasons following the year of
tagging.
Adjustment of the recapture rate Recapture rates ob-
served during the fishing season following tagging were ad-
justed by using the information from the aquarium obser-
vations. In contrast to tag misreporting that biases the
estimated number of recaptured lobsters, tag loss affects
the number of tagged animals (AT) in the population avail-
able to the fishery. To account for possible tag loss in our
field studies, AT was adjusted as follows. Letp represent the
tag-retention rate parameter estimate based on aquarium
observations with variance p(7-p;n"', where n is the total
number of animals observed in each aquarium experiment.
478
Fishery Bulletin 101(3)
Number and rate (in percentage) of tag s
in postmolt, intermolt, and premolt stages
into early and late premolt divisions, and
Table 1
ledding and tag induced mortality for American lobsters (Homarus americanus) tagged
in the aquarium experiments. Lobsters tagged in the premolt stage have been separated
n is the number of lobster used to calculate the adjusted recapture rate.
Premolt
Intermolt
Postmolt
Early
Late
Total
Initial number' [n)
56
135
191
38
183
Tag loss without molting
Tag-induced mortality
Shedding
Total
0
0
0
13(10%)
0
13(10%)
13 (7%)
0
13 (7%)
0
0
0
7 (4%)
1 (<1%)
8 (4%)
Tag loss during or after molting
Tag-induced mortality
Shedding
Total
1 (2%)
14 (25%)
15 (27%)
14 (10%)
15(11%)
29(21%)
15 (8%)
29(15%)
44 (23%)
Total tag loss
15 (27%)
42 (31%)
57 (30%)
0 (0%)
8 (4%)
' Number used in the experiment after accounting for lobsters that died within 30 minutes of tagging.
The adjusted number of tagged animals released during the
field studies effectively available to the fishery is equal to
with variance
V(N^^j) = Np(\-p).
The field study recapture rate (t) is defined as
t = 'n(Kd?-\
where m = the number of tags returned.
(1)
(2)
(3)
The variance of < can be calculated by using conditional
theory and an approximate variance of (iV^j.)"' (Seber,
1982). In this study, we used Monte-Carlo simulations to
obtain the 90% confidence interval (CI) for /. The simula-
tions were carried out in two steps to include variability
associated with tag loss from the initial N animals tagged
and released, and variability in recapture rate. First, as-
suming that the number of tags retained (A^,,,;,) follows a
binomial distribution with parameters N andp, a random
number (N^j.) was selected from a Binomial(N,p), where N
was set equal to the number of animal tagged during a par-
ticular study, andp equal to the proportion of tag retention
estimated from the aquarium study. With this simulated
^adj ^ conditional value of < was derived as
100 mm) differently
(with the tag inserted in only one abdominal muscle) than
did Moriyasu et al. (1995) for smaller lobsters. The differ-
ence could also be explained by the artificial conditions of
our aquarium experiment. Under natural conditions tagged
lobsters could shed their tags through interspecific interac-
tions (Rowe and Haedrich, 2001), intraspecific interactions,
and by being dislodged by obstacles in their habitat (Ennis,
1986; Krouse and Nutting, 1990). Streamer tag loss related
to inter- and intraspecific interactions and the habitat has
already been reported for the brown shrimp (Howe and
Hoyt, 1982; P. aztecus) and the tiger prawn (Hill and Was-
senberg, 1985;P. esculentus). However, more research would
be needed to identify the cause of tag shedding in nature
and assess its variability in relation to different lobster
habitat before the recapture rate could be adjusted based
on inter- and intraspecific interactions and the habitat.
The overall level of streamer tag loss compared to sphy-
rion tag loss seems to be lower, but also depends upon the
molt stage of the lobster at tagging and molting. In their
study, Moriyasu et al. (1995), suggested that sphyrion tag
loss mainly occurs within days after tagging or during
molting and is related to the lobster molt stage at tagging.
We observed lower levels of tag loss compared to those of
Moriyasu et al. ( 1995), except for the tag shedding during
molting for early premolt lobsters and tag-induced mor-
tality for lobsters tagged in late premolt. They observed
3% and 11% of tag shedding without molting for lobsters
tagged in early and late premolt stages, respectively, com-
pared to none in our study. Furthermore, the most striking
difference is the level of tag loss that reached 10% and
30% for lobsters tagged in intermolt and postmolt stages
compared to 0% and <1%, respectively, in our study. The
difference in tag loss for lobsters tagged in the postmolt
stage could be explained by the physical nature of the
tags themselves and the tagging techniques. Compared to
the streamer tag that is threaded through two abdominal
muscles, the sphyrion tag is anchored to only one muscle by
means of a hypodermic needle (Moriyasu et al., 1995). Be-
cause the muscles of postmolt lobsters (in the early stages)
are not well formed, it is difficult to firmly embed an object,
such as a tag, and the probability of tag loss for a tag em-
bedded into only one thin muscle is greater than that for
a tag treaded through two muscles. Hence, it seems that
streamer tags are more effective in terms of tag retention
compared to sphyrion tags for lobsters tagged in intermolt
and postmolt stages, but equally so for lobsters tagged in
the premolt stage.
In field tagging studies, streamer tags yielded a good re-
capture rate within the first year following tagging for lob-
sters tagged immediately before or after the molting sea-
son. The efficiency of streamer tags compared to sphyrion
tags had already been established (.Moriyasu et al., 1995;
Comeau et al., 1999). Moriyasu et al. (1995) reported that
there was a significantly greater recapture rate for lob-
sters tagged with streamer tags (44%) compared to those
tagged with sphyrion tags (19%). Based on the results of
Comeau et al. (1999), the recapture rate of lobsters tagged
with sphyrion tags is 22% and 16% for lobsters tagged
in premolt and postmolt stages, respectively, compared
to 33% and 45% for lobsters tagged with streamer tags.
These recapture rates corroborate aquarium observations
by Moriyasu et al. (1995) on the tag retention of sphyrion
tags and ours on the tag retention of streamer tags for
lobster tagged at various molt stages.
Knowledge of the level of tag loss is paramount for ad-
justing the recovery rate to estimate population character-
istics and fishery parameters for the American lobster. We
observed that the recapture rate dropped significantly in
the second and third years at large; this finding suggests
a high level of tag loss. A similar multiple-years recapture
rate pattern was observed for six other sites within the
southwestern Gulf of St. Lawrence (Comeau and Savoie,
2002). Rowe and Haedrich (2001) indicated that the
streamer tag shedding level for lobsters that molted almost
a year later reached 40%. This high level of tag shedding,
probably related to the streamer tag remaining firmly
attached to the old dorsal thoraco-abdominal membrane
during molting, might explain the drastic decrease of tag
recaptures observed between the first and the second tag-
recapture periods in our field study. We believe that the
adjustment of the recapture rate due to tag loss should
be limited to lobsters recaptured within the first year at
large and prior to the molting season for lobsters tagged
in intermolt and postmolt stages.
The multiple-years recapture pattern of a high recapture
rate within the first year at large followed by low recapture
rates in subsequent years could have a significant impact
on multiyear tagging models. These models that were pro-
posed by Sober ( 1970) could be used to estimate population
characteristics and fishery parameters if underlying as-
sumptions are followed. Based on these multiyear models
originally developed for birds (Seber, 1970; Brownie et al.,
1985), a suite of models adapted for fishery data and ad-
justments, mainly by reparameterization, were proposed
to address underlying assumption violations (Pollock et
al., 1991, 2001; Heam et al., 1998, 1999; Hoenig et al.,
1998a, 1998b; Frusher and Hoenig, 2001; Latour et al.,
2001a, 2001b). Some of these models were developed to
take into account fishing effort, incomplete mixing, and
tag recovery rate. The latter is a composite parameter
involving tag retention and tag-induced mortality (tag
loss), exploitation rate, and tag reporting rate. There is no
argument that the participation of fishermen in returning
tags (tag reporting rate) is very important; however, for
crustacean fisheries, underlying assumptions dealing with
tag loss (see assumptions 2 and 3 in Pollock, 1991, 2001)
are equally important and have to be addressed. In general.
Pollock et al. (2001) indicated that the assumptions of no
tag loss could be violated in two ways; by tag loss in the first
482
Fishery Bulletin 101(3)
few days after tagging or by chronic tag loss spread over
an extended period of time, the latter being more difficult
to model. Furthermore, Seber (1970) mentioned that the
usefulness of the multiyear models depends not only on
the validity of the underlying assumptions but also on the
number of recaptures (i.e. parameter estimates based on a
small number of tag recaptures would be biased). From our
aquarium observations and field studies that show a small
number of lobsters caught during the second and third re-
capture periods, we conclude that a significant chronic tag
loss does occur for the American lobster due to molting (i.e.
the molt stage of the lobster at tagging and molting itself).
Chronic tag loss impedes the effectiveness of multiyear re-
capture models currently used (Heam et al, 1998; Hoenig
et al., 1998a; Frusher and Hoenig, 2001) because it is not
taken into account. We believe that assuming only a con-
stant short-term tag loss for lobsters tagged with streamer
tags is inadequate and can only bias estimates of survival
and exploitation rate. Correcting for chronic tag loss after
the first year at large for the American lobster, however,
requires further knowledge, and more studies would be
required to fully understand long-term tag loss.
In conclusion, a high level of streamer tag loss is a major
obstacle for using tagging studies to estimate natural mor-
tality or to apply multiyear models for the American lobster.
Because streamer tag loss is related to molting, adjustment
is difficult because the molting frequency is size, sex, and
environment dependent (Comeau and Savoie, 2001). In our
attempt to estimate mortality at molt, it was found that
differences in recapture rates of lobsters tagged in premolt
and postmolt stages for a given molting period were not sta-
tistically significEmt, thus suggesting a low level of natural
mortality during the molt. The recapture rate for the 1996
tagging, for instance, was even higher for lobsters tagged in
the premolt stage. Hence, tagging with streamer tags to es-
tablish the level of natural mortality during the molt, or any
other mortality that could be low, for the American lobster
is not recommended. The alternative would be to develop
another insertion tag with better retention through the
molting process. Nevertheless, the streamer tag remains an
adequate choice for studying lobster ecology and population
biology. Streamer tags could be used to tag intermolt and
postmolt lobsters during single recapture tagging studies
to estimate the exploitation rate (Xiao et al., 1999). Based
on our observations, a minimum adjustment of 24.9% (SD
2.9%) and 4.4% (SD 1.6%) is suggested for lobsters tagged
in premolt and inter- or postmolt stages, respectively, and
recaptured during the first recovery period.
Acknowledgments
The authors wish to thank all fishermen from Caraquet
and the adjacent wharves that returned lobster tags and T
Brideau, B. Comeau, J. Roussel, and F. Savoie for their tech-
nical assistance in the field and during the tag collection.
We especially thank A. Godin and his staff at the Aquarium
et centre marin in Shippagan, New Brunswick, for their
professional help during the aquarium experiment. We also
want to thank J. M. Hanson and M. Moriyasu for critically
reviewing the manuscript, and three anonymous reviews
for thoughtful suggestions that improved the quality of
this manuscript.
Literature cited
Aiken, D. E.
1980. Molting and growth. In The biology and manage-
ment of lobsters. Vol. I: Physiology and behavior (J. S. Cobb
and B. F. Phillips, eds.), p. 91-163. Academic Press, New
York, NY.
Beverton, R. J. H., and S. H. Holt.
1957. On the dynamics of exploited fish populations. U.K.
Minist. Agric. Fish., Fish. Invest. 19:1-533.
Brownie, C, D. R. Anderson, K. P. Burhnam, and D. S. Robson.
1985. Statistical inference from band recovery data: a
handbook, 2nd ed. U.S. Fish Wildl. Serv. Resour. Publ.
156, 305 p.
Comeau, M., W. Landsburg, M. Lanteigne, M. Mallet, P. Mallet,
G. Robichaud, and F. Savoie.
1998. Lobster (Homarus americanus) tagging project in
Caraquet (1993)— tag return from 1994 to 1997. Can.
Tech. Rep. Fish. Aquat. Sci. 2216, 35 p.
Comeau, M., M. Lanteigne, G. Robichaud, and F. Savoie.
1999. Lobster (Homarus americanus) movement in the
southern Gulf of St. Lawrence — summary sheets of tagging
projects conducted between 1980 and 1997. Can. Ind. Rep.
Fish. Aquat. Sci. 249, HI p.
Comeau, M., and M. Mallet.
2001. Estimating mortality rates by capture-recapture,
catch-effort and change-in-ratio models for a spring Ameri-
can lobster (Homarus americanus) fishery (.LFA 23). Can.
Tech. Rep. Fish. Aquat. Sci. 2373, 20 p.
Comeau, M., and F. Savoie.
200 1 . Growth increment and molt frequency of the American
lobster (Homarus aiiiericanus) in the southwestern Gulf of
St. Lawrence. J. Crust. Biol. 21:923-936.
2002. Movement of American lobster (Homon/,s americanus)
in the southwestern Gulf of St. Lawrence. Fish. Bull. 100:
181-192.
Ennis, G. P
1986. Sphyrion tag loss from the American lobster Homarus
americanus. Trans. Am. Fish. Soc. 115:914-917.
Frusher, S. D., and J. M. Hoenig.
2001. Estimating natural and fishing mortality and tag
reporting rate of southern rock lobster iJasus cdwardsii)
from a multiyear tagging model. Can. J. Fish. Aquat. Sci.
58:2490-2.501.
Heam, W. S., K. M. Pollock, and E. N. Brooks.
1998. Pre- and post-season tagging models: estimation of
reporting rate and fishing and natural mortality rates.
Can. J. Fish. Aquat. Sci. 55: 199-205.
Heam, W. S., T Polacheck, K. H. Pollock, and W. Whitelaw.
1999. Estimation of tag reporting rates in age-structured
multicomponent fisheries where one component has
observers. Can. J. Fish. Aquat. Sci. 56:1255-1265.
Hoenig, J. M., N. J. Barrowman, W. S. Hearn. and K. H. Pollock.
1998a. Multiyear tagging studies incorporating fishing
effort data. Can. J. Fish. Aquat. Sci. 55:1466-1476.
Hoenig, J. M., N. J. Barrowman, K. H. Pollock. E. N. Brooks, and
W. S. Heam.
1998b. Models for tagging data that allow for incomplete
mixing of newly tagged animals. Can. J. Fish. Aquat. Sci.
55:1477-1483.
Comeau and Mallet: Effect of timing of tagging on tag recapture rates for Homarus amencanus
483
Hill, B. J., and T. J. Wassenberg.
1985. A laboratory study of the effect of streamer tags
on mortality, growth, moulting and duration of noctur-
nal emergence of the tiger prawn Penaeus esculentus
(Haswell). Fish. Res. 3:223-235.
Howe, N. R.. and R R. Hoyt.
1982. Mortality of juvenile brown shrimp Penaeus aztecus
associated with streamer tags. Trans. Am. Fish. Soc. Ill:
317-325.
Krouse, J. S., and G. E. Nutting.
1990. Effectiveness of the Australian western rock lobster
tag for marking juvenile American lobsters along the Maine
coast. Am. Fish. Soc. Symp. 7:94-100.
Landsburg, A. W.
1991. A field comparison of recapture rates of polyethylene
streamer and modified sphyrion tags through molting of
lobster (Homarus americanus). J. Shellfish. Res. 10, 225
P-
Latour, R. J., J. M. Hoenig, J. E. Olney, and K. H. Pollock.
2001a. Diagnostics for niultiyear tagging models with appli-
cation to Atlantic striped bass (Morone saxatilis). Can. J.
Fish. Aquat. Sci. 58:1716-1726.
2001b. A simple test for nonmixing in multi-year tagging
studies: application to striped bass tagged in the Rappahan-
nock River. Trans. Am. Fish. Soc. 130:848-856.
MaruUo, R, D. A. Emiiani, C. W. Caillouet, and S. H. Clark.
1976. A vinyl tag for shrimp (Panaeus spp.). Trans. Am.
Fish. Soc. 105:658-663.
Moriyasu M.. W. Landsburg, and G.Y. Conan.
1995. Sphyrion tag shedding and tag induced mortality
of the American lobster, Homarus americanus H. Milne
Edwards, 1837 (Decapoda, Nephropidae). Crustaceana
68:184-192.
Pollock, K. H., J. M. Hoenig, W. S. Heam, and B. Calingaert.
2001. Tag reporting estimation: an evaluation of the reward
tagging method. N. Am. J. Fish. Manage. 21:521-532.
Pollock. K.H., J. M. Hoenig, and C. M. Jones.
1991. Estimating of fishing and natural mortality when
a tagging study is combined with a creel survey or port
sampling. Am. Fish. Soc. Symp. 12:423-434.
Ricker, W. E.
1975. Computation and interpretation of biological statistics
offish populations. Bull. Res. Board Can. 191, 382 p.
Rowe, S., and R. L. Haedrich.
2001. Streamer tag loss from American lobsters. Trans.
Am. Fish. Soc. 130:516-518.
Scarratt, D. J., and P. F. Elson.
1965. Preliminary trials of a tag for salmon and lobsters. J.
Fish. Res. Board Can. 22:421-423.
Seber, G. A. F.
1970. Estimating time-specific survival and reporting
rates for adult birds from band returns. Biometrika 57:
313-318
1982. The estimation of animal abundance and related
parameters, 2nd ed., 654 p. Griffin, London.
Templeman, W.
1935. Lobster tagging in the Gulf of St. Lawrence. J. Biol.
Board Can. 1:269-278.
Wilder, D. G.
1953. The growth of the American lobster {Homarus
americanus). J. Fish. Res. Board Can. 10:371-403.
Xiao, Y., J. D. Stevens, and G. J. West.
1999. Estimation of fishing and natural mortalities from tag
experiments with exact or grouped times at liberty. Can.
J. Fish. Aquat. Sci. 56:868-874.
484
Abstract— Bycatch. or the incidental
catch of nontarget organisms during
fishing operations, is a major issue in
U.S. shrimp trawl fisheries. Because
bycatch is typically discarded at sea,
total bycatch is usually estimated by
extrapolating from an observed bycatch
sample to the entire fleet with either
mean-per-unit or ratio estimators.
Using both field observations of com-
mercial shrimp trawlers and computer
simulations, I compared five methods
for generating bycatch estimates that
were used in past studies, a mean-
per-unit estimator and four forms of
the ratio estimator, respectively: 1)
the mean fish catch per unit of effort,
where unit effort was a proxy for sample
size, 2) the mean of the individual fish
to shrimp ratios, 3) the ratio of mean
fish catch to mean shrimp catch, 4) the
mean of the ratios offish catch per time
fished (a variable measure of effort),
and 5) the ratio of mean fish catch per
mean time fished. For field data, differ-
ent methods used to estimate bycatch
of Atlantic croaker, spot, and weakfish
jaelded extremely different results, with
no discernible pattern in the estimates
by method, geographic region, or spe-
cies. Simulated fishing fleets were used
to compare bycatch estimated by the
five methods with "actual" (simulated)
bycatch. Simulations were conducted by
using both normal and delta lognormal
distributions of fish and shrimp and
employed a range of values for several
parameters, including mean catches
of fish and shrimp, variability in the
catches of fish and shrimp, variability
in fishing effort, number of observa-
tions, and correlations between fish and
shrimp catches. Results indicated that
only the mean per unit estimators pro-
vided statistically unbiased estimates,
while all other methods overestimated
bycatch. The mean of the individual fish
to shrimp ratios, the method used m the
South Atlantic Bight before the 1990s,
gave the most biased estimates. Because
of the statistically significant two- and
3-way interactions among parameters,
it is unlikely that estimates generated
by one method can be converted or cor-
rected to estimates made by another
method: therefore bycatch estimates
obtained with different methods should
not be compared directly.
Manuscript approved for publication
17 December 2002 by Scientific Editor
Manuscript received 3 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:484:-500(2003).
Estimation of bycatch in shrimp trawl fisheries:
a comparison of estimation methods
using field data and simulated data
Sandra L. Diamond
Department of Biology
Box 3131
Texas Tech University
Lubbock, Texas 79409
E-mail address: Sandra Diamond'fflttuedu
Bycatch, as used in the present study,
is the incidental catch of nontarget
organisms that occurs to some extent in
almost all commercial fisheries (Alver-
son, 1994). Some of these incidentally
caught organisms may be protected spe-
cies— such as marine mammals, marine
turtles, and seabirds — or they may be
fish or invertebrates that are either
harvested as target species by other
fisheries, or species that fishermen call
"trash fish" because they have little or
no economic value. Bycatch in most
commercial fisheries has only been a
major issue since the 1980s — primarily
because individuals caught as bycatch
have historically been discarded at
sea, leaving fishery managers and the
general public unaware of the extent of
bycatch mortality. For many organisms,
bycatch may be a significant source of
mortality, and inclusion of bycatch mor-
tality in stock assessments or manage-
ment plans may be critical for effective
management.
Because bycatch species are not usu-
ally landed, quantifying bycatch poses
a very different problem from that of
quantifying the catch of a target spe-
cies. Several methods of quantifying
bycatch have been tried, including the
requirement that fishermen record
catch and bycatch in logbooks (Walsh
and Kleiber, 2001), use of research ves-
sel surveys to model commercial fishing
(Nichols et al.'), and the placement of
observers aboard fishing vessels (Julian
and Beeson, 1998). Although direct ob-
servation is the most accurate method,
unless observer coverage of the fleet is
complete, estimation of bycatch from
observation data requires sampling of
the fleet and then extrapolating from
the samples (the observations) to the
entire fleet using statistical estimators.
Two types of statistical estimators are
used: mean-per-unit estimators and ra-
tio estimators. In both types of estima-
tors, the observed catch of the bycatch
species iy) is linked to an auxiliary vari-
able (x) for which the population total
is known (Cochran, 1977). In mean-per-
unit estimators, the auxiliary variable
is a measure of fishing effort such as
tow, day, trip, etc., where each unit of
effort is the same as one observation.
In ratio estimators, the auxiliary vari-
able is a variable that is correlated with
the catch of the bycatch species, such
as the catch of the target species or the
number of hours fished (Cochran. 1977).
The major difference between these two
types of estimators is that the auxiliary
variable in the mean-per-unit estimator
is a substitute for the number of obser-
vations rather than a mean value with
a variance, while the auxiliary variable
in the ratio estimator is the mean value
of a quantity that varies from sample to
sample. Although the statistical proper-
ties of these two types of estimators are
well known, the choice of which estima-
tor to use in bycatch research is often
based on the ease of collecting fleet
information on the auxiliary variable,
and not on any inherent properties of
the estimators themselves or on any
specific information about the relation-
ship between the catch of bycatch spe-
cies and the auxiliary variable.
' Nichols, S., A. Shah, G. Pellegrin Jr, and
K. Mullin. 1990. Updated estimates of
shrimp fleet bycatch in the offshore waters
of the US Gulf of Mexico, 22 p. Pascagoula
Laboratory, Southeast Fisheries Science
Center, NMFS, PO Drawer 1207, Pasca-
goula, MS 39568-1207.
Diamond: Estimation of shrimp trawl bycatch
485
Bycatch is a major issue in the shrimp trawl fisheries
of the Gulf of Mexico and the South Atlantic Bight. These
fisheries are the most valuable fisheries in the southeast-
em United States; almost 136,000 metric tons of shrimp,
worth over $700 million, were landed in 2000 (NMFS^). It
is estimated that 60-80% of the catch by weight in these
fisheries is bycatch. Over 150 species have been reported in
shrimp trawl bycatch, including marine turtles (Grouse et
al., 1987) and juveniles of species that are highly valued as
adults in other fisheries, such as weakfish iCynoscion rega-
lis rVaughan et al.-'] ) and red snapper (Lutjanus campecha-
nus [Goodyear'*! ).
Both types of statistical estimators have been used
to estimate bycatch in shrimp trawl fisheries. In the
South Atlantic, biologists have periodically partici-
pated as observers aboard commercial shrimp trawlers
since at least the 1950s to characterize bycatch and
estimate its magnitude (Fahy, 1966; Latham, ^ Lunz et
al.,6 Fahy,'' Fahy« Fahy,^ Wolff,"* Keiser.n Knowltoni^).
For most of the studies conducted between the 1950s and
2 NMFS (National Marine Fisheries Service). 2002. Unpubl.
data Website: http://www.st.nmfs.gov/stl/comniercial/index.
html.
^ Vaughan, D. S., R. J. Seagraves, and K. West. 1991. As assess-
ment of the status of the Atlantic weakfish stock, 1982-1988.
Special Report 21, 29 p. Atlantic States Marine Fisheries
Commission, 1444 Eye Street, N.W., Sixth Floor, Washington,
DC 20005.
^ Goodyear, C. P. 1995. Redsnapper in US watersofthe Gulfof
Mexico. Contribution MlA-95/96-05, 171 p. Miami Laboratory,
Southeast Fisheries Science Center, NMFS,75 Virginia Beach
Drive, Miami. Florida 33149-1099.
^ Latham, F F. 1951. Evidence of fish loss due to shrimping
in Pamlico Sound. Appendix B in The destruction of small
fish by the shrimp trawlers in Pamlico Sound, North Carolina
(G. R. Lunz, J. L.,McHugh, E. W. Roelofs, R. E. Tiller, and C. E.
Atkinson), p. 17-24. Committee Report to the Atlantic States
Marine Fisheries Commission, 1 November 1951. Atlantic
States Marine Fisheries Commission, 1444 Eye Street, N.W,
Sixth Floor, Washington. DC 20005.
« Lunz. G. R.. J. L. McHugh, E. W. Roelofs, R. E. Tiller, and C. E.
Atkinson. 1951. The destruction of small fish by the shrimp
trawlers in Pamlico Sound, North Carolina. Committee Report
to Atlantic States Marine Fisheries Commission, 1 November
1951, 34 p. Atlantic States Marine Fisheries Commission,
1444 Eye Street, N.W., Sixth Floor, Washington, DC 20005.
"Fahy, W. E. 1965a. Report of trash fish study in North Caro-
lina in 1962. Division of Commercial and Sports Fisheries,
NC Department of Conservation and Development, Special
Scientific Report 5. 20 p. NC Division of Marine Fisheries,
3441 Arendell St., Morehead City, NC 28557.
8 Fahy, W. E. 1965b. Report of trash fish study in North Caro-
lina in 1964. Division of Commercial and Sports Fisheries,
NC Department of Conservation and Development, Special
Scientific Report 7, 13 p. NC Division of Marine Fisheries,
3441 Arendell St., Morehead City. NC 28557.
3 FahyW.E. Unpubl.datacitedinBrown, J.,andE.McCoy 1969.
A review of the North Carolina scrap fishery. Division of Com-
mercial and Sports Fisheries. NC Department of Conservation
and Development. Information Series 1, 12 p. NC Division of
Marine Fisheries, 3441 Arendell St., Morehead City, NC 28557.
1" Wolff, M. 1972. A study ofNorth Carolina scrap fishery. NC
Department of Natural and Economic Resources, Special Sci-
entific Report 20, 29 p. NC Division of Marine Fisheries, 3441
Arendell St., Morehead City, NC 28557.
the 1980s, fisheries bycatch was estimated by using a ratio
estimator, that is to say by calculating the observed ratio of
fish (F) bycatch to shrimp (S) by weight and then multiply-
ing by the total pounds of shrimp landed by the fleet (the
F:S ratio estimator). The catch of shrimp was used as the
auxiliary variable primarily because better records were
kept of shrimp landings than of any measure of fleet effort.
By the late 1980s, the problem of shrimp trawl bycatch in
the United States was considered to be of such magnitude
that in 1990 the Magnuson Fishery Conservation and
Management Act (Magnuson Act) was amended to include
bycatch research. Beginning in 1992, observers trained
by the National Marine Fisheries Service (NMFS) to use
a standardized sampling protocol (NMFS'-*) rode aboard
paid volunteer commercial vessels in the South Atlantic
and Gulf of Mexico. The 1992-94 observation data collected
in the South Atlantic were used to estimate bycatch by spe-
cies with a mean-per-unit estimator, which was the weight
or number of fish caught per observed trip multiplied by
the total number of trips taken by the fleet (the CPUE-
mean-per-unit estimator). Trips were used as the auxiliary
variable because fleet effort data were available at the trip
level and this method was thought to be less variable than
the F:S ratio method (SEAMAP'").
To date, there have been no detailed studies on how
these different techniques compare to each other, or how
accurately they estimate bycatch. Vaughan and Nance''' in
a draft paper compared the estimated bycatch of mackerels
(Scomberomorus spp.) and cobia {Rachycentron canadum)
using both methods and found much higher estimates with
the F:S ratio estimator than with the CPUE-mean-per-unit
estimator. Because of the wide range of estimation meth-
ods used over the years, the discrepancy in the estimates
generated by the different methods, and the increasing im-
portance of bycatch estimation for shrimp trawl fisheries
" Reiser, R. K. 1977. The incidental catch from commercial
shrimp trawlers of the South Atlantic states. Technical Report
26, 38 p. South Carolina Wildlife and Marine Resources Depart-
ment, South Carolina Department of Natural Resources, Rembert
C. Dennis Building, 1000 Assembly Street, Columbia, SC 29201.
12 Knowlton, C. J. 1972. Fishes taken during commercial
shrimping in Georgia's close inshore ocean waters. Con-
tributed Series 21, 42 p. Georgia Department of Natural
Resources, Coastal Resources Division, One Conservation Way,
Suite 300, Brunswick, GA 31520.
1-^ NMFS (National Marine Fisheries Service). 1992. Shrimp
trawl bycatch characterization. Sampling Protocol Manual
for Data Collection, 62 p. Galveston Laboratory, Southeast
Fisheries Science Center, NMFS, 4700 Avenue U, Galveston,
TX 77551-5997.
''• SEAMAP (Southeast Area Monitoring and Assessment Pro-
gram ). 1996. Estimates of finfish bycatch in the South Atlan-
tic Shrimp Fishery, July 24, 1995 (R. Peuser, ed.), 64 p. Final
report of the Southeast Area Monitoring and Assessment Pro-
gram (SEAMAP), SEAMAP-South Atlantic Committee, Shrimp
Bycatch Work Group. Atlantic States Marine Fisheries Com-
mission, 1444 Eye Street, N.W., Sixth Floor, Washington, DC
20005.
'^ Vaughan, D. and J. Nance. 1998. Estimates of bycatch of
mackerel and cobia in US South Atlantic shrimp trawls. Re-
port for Gulf of Mexico and South Atlantic Fishery Management
Councils, February 16, 1998, 26 p. NMFS -SEFSC, Beaufort
Laboratory, 101 Fivers Island Road, Beaufort NC 28516.
486
Fishery Bulletin 101(3)
Sound
36°00'
35°00'
Figure 1
Map of North Carolina waters. Shrimping operations were observed in northern Pamlico
Sound (between the mouth of the Pamlico River and southern Roanoke Sound) and the lower
third of the Cape Fear River. For total bycatch, fleet shrimp landings and fleet shrimp effort,
the northern region includes Pamlico Sound and its tributaries, and the southern region
includes from the Cape Fear River to the New River.
and other fisheries, fishery biologists need clear guidance
on which method to use to estimate bycatch and they need
a definitive knowledge of which methods are best under the
varying conditions that might be found in a field observer
study.
In this article, I use both field data and computer simula-
tions to compare the methods of bycatch estimation used
in past studies. First, using field observations of Atlantic
croaker {Micropogonias undulatus), spot (Leiostomas xan-
thurus), and weakfish bycatch from shrimp trawlers in
North Carolina, I compare bycatch estimates generated
by the CPUE-mean-per-unit estimator with two different
forms of the F:S ratio estimator, the mean of the individual
fish to shrimp ratios and the ratio of the mean catch offish
to the mean catch of shrimp. I then simulate fishing fleets
with different catches of fish and shrimp, and estimate
bycatch using the following five different estimators, a
mean-per-unit estimator and four forms of the ratio esti-
mator, respectively: 1 ) the mean fish catch per unit of effort,
where unit effort is a proxy for sample size, 2) the mean
of the individual fish to shrimp ratios, 3) the ratio of mean
fish catch to mean shrimp catch, 4) the mean of the ratios
offish catch per time fished (a variable measure of effort),
and 5) the ratio of mean fish catch per mean time fished.
The simulations employ different mean catches offish and
shrimp, different levels of variability around the catches
offish and shrimp and around the variable measure of ef-
fort in the ratio estimator, and different levels of observer
coverage, or the number of observations. I also investigate
the effects on the bycatch estimates of different underlying
distributions offish and shrimp, including normal distribu-
tions offish and shrimp with different levels of correlation
between the catches of fish and shrimp, and delta lognor-
mal distributions of both fish and shrimp, with differing
probabilities of catching fish or shrimp.
Materials and methods
Field sampling
To compare the methods described in the literature using
field data, I observed shrimping operations aboard com-
mercial shrimp boats from July through October 1995 in
Pamlico Sound, North Carolina, and from August through
October 1995, in the Cape Fear River, North Carolina
(Fig. 1). These two areas have different levels of fishing
effort, different fish-to-shrimp ratios, and different prob-
abilities of catching fish and shrimp. All fishermen coopera-
tors were unpaid volunteers, and I did not direct them in
any way regarding where or how to fish. Although sampled
boats were not randomly chosen, the fishermen appeared
to use gear and fishing methods similar to those of other
shrimpers, and other shrimpers were often seen fishing in
the area near the sampled boats.
Sampled shrimp boats towed one or two nets, and all
nets contained some form of turtle excluder device (TED)
and bycatch reduction device (BRD) required by regulation.
Diamond: Estimation of shrimp trawl bycatch
487
To sample the catch, I used the NMFS bycatch samphng
protocol as described below. If the boat carried two nets and
no try net (the small net towed in front of the main nets
which is used to survey the catch at short time intervals),
I randomly picked one net (the "selected net") by flipping
a coin. If the boat had a try net, I picked the opposite net.
I weighed the total catch of the selected net on a flat agri-
cultural scale by emptying the net into a plastic tub placed
on a scale. After having been weighed, the catch of the se-
lected net was dumped onto the deck or into a culling tray
that was divided so that the catch of the selected net was
separated from the catch of the unselected net. Following
the NMFS protocol, I mixed the selected net contents thor-
oughly with a shovel, then took a random sample and set it
aside until after the rest of the net contents had been sort-
ed. To sort the net contents, marketable shrimp, which are
pink shrimp (Faifantepenaeus duorarum), brown shrimp
{Farfantepenaeus aztecus), and white shrimp (Litopenaeus
setiferus) larger than would comprise about a 70-80 count
(i.e. 70-80 shrimp per pound) were separated from the rest
of the contents of the selected net, weighed, and then re-
turned to the fisherman. The unsampled bycatch from the
selected net was discarded overboard. The random sample
taken from the selected net was then weighed. Market
shrimp in the sample were taken out, weighed and counted
by species, and returned to the fisherman. The bycatch por-
tion of the sample, including undersized market shrimp,
mantis shrimp, and all other fish and invertebrates, was
packaged in plastic bags and placed on ice for the remain-
der of the trip. Bycatch samples were brought back to the
laboratory and frozen. Samples, including market shrimp,
averaged 12% by weight of the total catch of the selected
net. and ranged from 5% to 37% by weight.
Expansion of observed bycatch to the entire tow In the
laboratory, I thawed and rehydrated the bycatch sample in
water. I sorted each sample by species and weighed each
species as a group. All individuals of each species were then
weighed and measured separately. To account for differ-
ences between the scales used on the boat and those used
in the laboratory, and for weight loss due to freezing, I cor-
rected the weight of the total catch of each net measured on
the boat by the ratio of the sample weight from the labora-
tory to the sample weight from the boat as follows:
Corrected total weight ^ =
lab sample weight
boat total weighty x
(1)
boat sample weighty
where corrected total weighty = the corrected weight of the
f-^ selected net;
boat total iveightj = the weight of the entire
catch of the/^" selected net
measured on the boat;
lab sample iveightj = the weight of the bycatch
sample of the/'' net mea-
sured in the laboratory
plus the shrimp sample
weight from the boat; and
boat sample iveightj = the weight of the entire
sample (including shrimp)
from the 7^'' net weighed
on the boat.
This correction averaged less than 5% across all selected
nets. To expand the catch in weight of each bycatch spe-
cies from the sample to the entire selected net (called the
"species net weight"), the total corrected weight of each
selected net was multiplied by the fraction of the sample
from the selected net that consisted of the bycatch species,
as follows:
corrected total weight, x
Species net weighty ^ =
species sample weighty ^
total sample weight j
(2)
where species net weighty : = the estimated catch in
weight of the i^^ species in
the/'' net;
corrected total weight = the corrected weight of the
total catch of the/'' net from
Equation 1;
species sample iveightj , = the weight of the j"' species
in the sample from the /''
net; and
total sample iveightj = the weight of the bycatch
sample from the/'' net mea-
sured in the laboratory plus
the weight of the market
shrimp in that sample mea-
sured on the boat.
Because the net contents were thoroughly mixed before
sampling, I assumed, following the NMFS protocol, that
there would be minimal variance among samples if more
than one were taken.
Expanding the catch in numbers of each bycatch spe-
cies from the sample to the entire selected net (called the
"species net number") could not be done in the same way
as the expansion for the species net weight because there
were often organisms like sea lettuce or pieces of fish or
crabs that were weighed but that could not be counted. The
species net number was therefore calculated by dividing
the species net weight by the average weight per whole
individual:
species net weight, ^
Species net number^ - =
species sample weight, ^
species sample number,^
(3)
where species net number
species net weight,
the estimated number of
individuals of the i''' spe-
cies in the/'' net;
the estimated total weight
of the i*^ species in the /•■
net from Equation 2;
488
Fishery Bulletin 101(3)
species sample weight
species sample number
the weight of the t"^ species
in the bycatch sample from
the/*^ net; and
the number of whole indi-
viduals of the /''' species in
the bycatch sample from
the./"^ net.
To expand the observed bycatch from selected net to the
entire tow, I multiplied either the species net weight or the
species net number from each net by the number of nets
towed concurrently.
Bycatch estimation To compare the methods of bycatch
estimation used in past studies, I estimated the bycatch
of Atlantic croaker, spot, and weakfish (three of the most
commonly caught bycatch species) using two categories
of statistical estimators: a mean-per-unit estimator using
the mean observed bycatch per day expanded by the total
number of days fished (the CPUE-mean-per-unit method)
and a ratio estimator using the observed ratio of fish to
shrimp expanded by the total shrimp landings (the F:S ratio
method). Because my purpose was to compare bycatch esti-
mation methods and not to generate bycatch estimates that
could be used for management purposes, I estimated total
bycatch of these three species only for certain months and
geographic regions within North Carolina corresponding to
the times and areas that I observed shrimp trawling. The
term "shrimp fleet" in the following paragraphs therefore
refers only to shrimpers operating in those times and areas.
In the calculations, I used bycatch per day instead of the
bycatch per tow or bycatch per trip. I could not use tow as
the unit of effort because there was no information on the
number of tows made by the fleet to use as an expansion
factor Although information on the number of trips made
by the fleet was available, I could not use trip as the unit of
effort because, although trips can last several days, all of the
trips that I sampled were one-day trips. If my observations
had also included a random sample of multiday trips, the
unit of effort would have been trips instead of days.
The CPUE mean per unit method was based on the fol-
lowing equations:
Mean ohsened hyaiich^
day
X"^-
(4)
where mean observed _ the observed average bycatch in
bycatch per day weight or number of the i^^ spe-
cies on the d^^ day;
n = the number of days observed; and
F^ J = the sum of the expanded weight
or number of the (''' bycatch spe-
cies observed in all tows made
on the d"' day; and
mean obser\'ed bycatch'
Total bycalcli, .-/.yf = ■ x
clay
tiiiiil trips X
mean days
trip
(5)
where total bycatch ^(.p^,^. = the total fleet bycatch
of the (''^ species esti-
mated by the CPUE
method;
mean observed bycatch ^ per day = the observed average
bycatch of the i^^ spe-
cies per day from Equa-
tion 4;
total trips = the total number of trips
taken by the shrimp
fleet; and
mean days per trip = the average number of
days that each fishing
trip lasted based on the
fleet.
The total trips and mean days per trip were calculated from
the North Carolina Division of Marine Fisheries (NCDMF)
trip ticket database, as follows. To obtain the total number
of trips, I first collapsed the trip ticket database so that each
fisherman could have only one ticket for shrimp on a single
day. In the database, each trip ticket does not represent one
trip, but the sale to one dealer Fishermen could obtain more
than one trip ticket per day by selling different size catego-
ries of shrimp (each size category commands a different price,
and generates a separate trip ticket ), or by selling their catch
to more than one dealer I then calculated the time (in days)
between the first and last trips for each fisherman whose
trips occun-ed between 1 July and 31 October in Pamlico
Sound and its tributaries (called the northern region) and
between 1 August and 31 October in the Cape Fear River
and nearby waters (the southern region). Because inshore
waters were closed to shrimping on weekends, I multiplied
all time spans greater than 7 days by 5/7 (0.714) to obtain
the number of days fished. The number of days fished was
summed and then divided by the number of trips for each
region to obtain the mean days per trip.
The F:S ratio estimator method was initially undertaken
in two ways: by using the mean of the fish to shrimp ratios,
called the mean of the ratios or the "basic" F:S ratio esti-
mator method (Equation 6), and by using the ratio of the
average catch offish to the average catch of shrimp, called
the ratio of the means or the "grand" F:S ratio estimator
method (Equation 7). The two methods are shown math-
ematically as follows:
Total h\catch, , ^„ = — > -^ x total shrimp landed.
(6)
Total hvcatcli ,v,. =-^ x total slniinp landed, (7)
rf=l
where total bycatch, y,,i, = the total fleet bycatch of the ;"'
species estimated by the basic
F:S method;
total bycatch, ^^g = the total fleet bycatch of the /"'
species estimated by the grand
F:S method;
Diamond: Estimation of shnmp trawl bycatch
489
F, J = the sum of the expanded
weight or number of the ;'"'
species observed in all tows
made on the d^^ day;
Sj = the sum of the expanded weight
of market shrimp observed in
all tows made on the c/* day; and
n = the number of days observed.
Because of the small number of days observed in each
area, I also used the basic F:S ratio estimator with the
Hartley-Ross correction for biases caused by small sample
size (Cochran, 1977):
Toliil hvccilcli^ f^if = total hxcatch, ^ j„ H (\\-rx), (8)
)i - 1
where total bycatch^ ^jf^ = the total fleet bycatch of the ;'"'
species estimated by the bias-
corrected F:S ratio estimator;
total bycatch^ p,.g = the total fleet bycatch of the ("^
species estimated by the basic
F:S ratio using Equation 6;
n = the number of days observed;
N = the total number of days fished
from the trip ticket database;
y, = the mean bycatch of of the ;"^
species observed per day in
weight or numbers from Equa-
tion 4;
F = the mean of the F:S ratios from
Equation 6; and
.V = the mean catch of market
shrimp observed per day in
weight or numbers.
Total shrimp landings used in Equations 6 and 7 were
obtained from the NCDMF trip ticket database for the
northern region from July to October and for the southern
region from August to October. In the trip ticket database,
some shrimp weights were reported as "heads-on" and
others as "heads-ofT'; therefore I converted heads-off weight
to heads-on weight with a conversion factor of 1.583, taken
from the average of pink, brown, and white shrimp conver-
sion information used by the National Marine Fisheries
Service (Fisheries Statistics of the United States, 1977).
Bycatch simulations
For the bycatch simulations, I created different fishing
fleets of 1000 "boats" in Matlab 5.0 (The Mathworks,
Natick, MA). For the normally distributed catch data, the
catch of fish, the catch of shrimp, and the hours fished for
each boat in a fleet were generated by using multivariate
random normal distributions with a mean and variance
that was specific to that fleet. I simulated observer data
for each fleet by taking a random sample of boats from
the fleet, resampling the sample 1000 times, then using
the mean of the bootstrapped observer data in the equa-
tions described below to estimate fleet bycatch. In the dif-
ferent fleets, the mean catches of fish and shrimp ranged
from 0.01 to 1000, giving fish to shrimp ratios of 0.001 to
100,000. In some fleets, the catches offish and shrimp were
correlated, with correlation coefficients ranging from 0.5 to
-0.5 (Table 1). Coefficients of variation (CVs) for fish catch
and hours fished ranged from 20% to 80%, CVs for shrimp
catch ranged from 20% to 120%, and the number of obser-
vations ranged from 20 to 500, giving observer coverages of
2% to 50%. of the fleet. Although the range of mean catches
I used in the simulations may seem fairly broad, they are
within the range of the field data, depending on whether
these were the mean catches per tow, per day, or per trip.
The ranges of CVs for fish and shrimp catches were fairly
narrow compared to those from the field data because CVs
vary up to several hundred percent, particularly for patchy
species. Observer coverage in the field is usually much less
than 50%>, but I picked 50% as the upper limit of the range
to see if greater observer coverage (i.e., a greater sample
size of observations per fleet) increased the accuracy of the
bycatch estimates.
Bycatch estimates were calculated by using a mean per
unit estimator and four forms of the ratio estimator, as
described below. The CPUE mean per unit estimator was
calculated by using the following equations, which are
more general versions of Equations 4 and 5:
Total bycatch, ,
Mean obsen'ed bycatch, I
UE
Mean hvcatch
(9)
UE
-xtntal fleet effort. (10)
where mean observed the observed average bycatch of
bycatch^ per UE = the i"" species per unit of effort
(tow, day, or trip);
n = the number of observed tows,
days, or trips;
F, ug = the expanded weight or number
of the i'*' bycatch species observed
on the [/£"' tow, day, or trip;
total bycatch-,(,piji' = the total fleet bycatch estimated
by the CPUE method; and
total fleet effort = the total number of tows, days,
or trips fished by the fleet.
The four ratio estimators were as follows: 1) the mean
of the individual F:S ratios, called the "basic F:S'' ratio
estimator (Eq. 11), 2) the ratio of the F:S means, called
the "grand F:S" ratio estimator (Eq. 12), 3) the mean of the
individual catch per effort ratios using a variable measure
of effort such as hours fished as the auxiliary variable,
called the "basic CPE" ratio estimator (Eq. 13), and 4) the
ratio of the mean catch per mean effort using a variable
measure of effort such as hours fished as the auxiliary
variable, called the "grand CPE" ratio estimator (Eq. 14).
Both F:S ratio estimators (Eqs. 11 and 12) are similar to
the ones used in the field study (Eqs. 6 and 7), except that
the observations could be from a tow, day, trip, or other
measure of unit effort, rather than one day, as used in the
field study.
490
Fishery Bulletin 101 (3)
1 V- F.
Total bycatch- ^^^ = — > — ^ x total shrimp landed (11)
Total bycatch
"-^.
■■=1
- X total shrimp landed ( 12)
where total bycatch^ pgg = the total fleet bycatch of the i"^
species estimated by the basic
F:S ratio estimator;
total bycatch^ pgQ = the total fleet bycatch of the i'^
species estimated by the grand
F:S ratio estimator;
F, ^, = the expanded weight or number
of the /''^ bycatch species ob-
served in the e'*' tow, day, or trip;
S^ = the expanded weight of market
shrimp observed in the e'*^ tow,
day, or trip;
n = the number of tows, days, or
trips observed; and
total shrimp landed = the sum of the total weight of
shrimp landed by the fleet.
1 -v-i F
Tdtal b\calch^^p^g = —y —'^^x total hours fished (13)
Total bycatch ^ p^^; = -^ x total hours fished (14 )
where total bycatch, cpi^g = the total fleet bycatch of the f*
species estimated by the basic
CPE ratio estimator;
total bycatch, (-.p^-Q = the total fleet bycatch of the
/"' species estimated by the
grand CPE ratio estimator;
F, ^ = the expanded weight or
number of the i*-^ bycatch
species observed in the e""
tow, day, or trip;
//^, = the hours fished in the e''' tow,
day, or trip;
n = the number of observed tows,
days, or trips; and
total hours fished = the sum of all hours fished by
the fleet.
To avoid confusion, it is important to note how fishing
effort is used in the five estimators. All five estimators use
a unit measure of fishing effort, such as a tow, day, or trip,
as one sample, and the sample size for a fleet is the number
of tows, days, or trips observed. In the CPUE mean-per-unit
estimator, the estimate of total bycatch is a simple expan-
sion of the observed bycatch per sample to the whole fleet.
In the F:S ratio estimators, the unit effort appears in the
calculations because the ratios of fish to shrimp are the
amounts caught per tow, day, or trip (i.e. per sample). In
the CPE ratio estimators, two measures of effort are used.
As before, one measure of effort is the unit effort, such as
a tow, day, or trip, that is equivalent to a sample, and the
second measure of effort is the variable measure of effort,
such as the hours fished, the distance towed, or the area
covered, that is used as the auxiliary variable. The CPE
ratio estimator is thus based on the amount offish caught
per hour fished (for example) in each tow, day, or trip.
The delta lognormal simulations were very similar to
the normal simulations, except that I simulated the catch
of fish and shrimp by using probabilities of catching fish
or shrimp ranging from 0.05 to 0.95, multiplied by average
catches offish or shrimp generated from random lognormal
distributions with means ranging from 0.01 to 1000. Log-
normal functions have parameters of// and a'^. which are
the mean and variance of the normally distributed variable
before logarithmic transformation. To obtain values of;/
and a^ from a lognormal distribution with a given mean
and variance, I used an iterative procedure (the Solver
procedure in Microsoft Excel, vers. 2000, Microsoft Corpo-
ration, Redmond, WA) to estimate fj and o^ based on the
following equations:
mean = e
[-^)
(15)
(16)
where mean = the mean of the lognormal distribution of
the catch offish or shrimp;
var - the variance of the lognormal distribution
of the catch offish or shrimp;
H = the mean of the normally distributed catch
of fish or shrimp before logarithmic trans-
formation; and
o^ = the variance of the normally distributed
catch of fish or shrimp before logarithmic
transformation.
Levels of observer coverage and CVs for fish catch,
shrimp catch, and effort and were the same as in the
normally distributed data. In these simulations, sampled
shrimp catch could be zero if the probability of catching
shrimp was low and the number of observations was small,
leading to F:S ratios of infinity. In these cases, for the basic
F:S ratio estimator, the fish-to-shrimp ratio was the catch
offish divided by the expected catch of shrimp (probability
of catching shrimp times the mean catch ). For the grand F:
S ratio estimator, if the average bootstrapped sample catch
of shrimp was zero, Matlab substituted a value of 65535
to avoid division by zero. To avoid biases, these grand F;
S simulations were left out of the subsequent analyses. In
field saini)liiig, tows that caught no shrimp at all were rare,
but tows that caught only small unmarketable shrimp that
were discarded as bycatch occurred occasionally early in
the season and after big rainstorms.
Diamond: Estimation of shrimp trawl bycatch
491
Table 1
Parameters and their
values used in the bycatch simulations for normal and delta lognormal distributions of fish and shrimp.
Abbreviations for the parameters are shown in parentheses
Distribution
Parameter
Values
Normal
Mean fish catch (AvgF)
0.01, 0.1, 1, 10, 100, 1000, 10,000
Fish CV (FCV)
20%, 50%, 80%, 120%
Mean shrimp catch (AvgS)
0.01, 0.1, 1, 10, 100, 1000, 10,000
Shrimp CV(SCV)
20%, 50%, 80%
Mean hours fished
1.0
Hours fished CV (ECV)
20%, 50%, 80%
F:S ratio
0.001, 0.01, 0.1, 1, 10, 100, 1000, 10,000, 100,000
Correlation coefficient (r)
-0.5,-0.25,0,0.25,0.5
Number of observations (n)
20,50,100,500
Observer coverage
2%, 5%, 10%, 50%
Delta lognormal
Probability of catching fish (P(F))
0.05,0.2,0.5,0,8,0.95
Mean fish catch (AvgF)
0.01, 0.1, 1, 10, 100, 1000, 10,000
Fish CV (FCV)
20%, 50%, 80%
Probabihty of catching shrimp (P(S))
0.05,0.2,0.5,0.8,0.95
Mean shrimp catch (AvgS)
0.01, 0.1, 1, 10, 100, 1000, 10,000
Shrimp CV(SCV)
20%, 50%, 80%, 120%
Mean hours fished
1.0
Hours fished CV (ECV)
20%, 50%:, 80%
F:S ratio
0.001, 0.01, 0.1, 1, 10, 100, 1000, 10,000, 100,000
Number of observations ( n )
20,50,100,500
Observer coverage
2%, 5%, 10%, 50%
To statistically analyze the overall biases shown by
each estimator regardless of fishing conditions (i.e. mean
catches of fish or shrimp, CV, etc.), I first used paired
sample /-tests (SAS v. 8, The SAS Institute, Gary NC) to
separately compare each of the five estimates of fleet by-
catch with the "actual" bycatch for that fleet, based on the
following equation:
where % bias
estimated bycatch^ ^^^
"actual" bycatch ,
estimated bycatch, ,„ ^ - "actual" bycatch, ,, x 1 00 ( j^ 7 )
"actual" bycatch,^
„ I, = the bias in the m"^ estimator for
the 6'h fleet;
,„ f, = the bycatch of the i^^ species by
the w'*' estimator for the b^^ fleet;
and
the simulated actual bycatch of
the ;"^ species by the 6"^ fleet.
For these statistical tests, all fleets with normal distribu-
tions offish and shrimp were combined and analyzed sepa-
rately from the fleets with delta lognormal distributions of
fish and shrimp, giving sample sizes of 21,600 fleets for the
normal distribution and 118,810 fleets for the delta lognor-
mal distribution (Table 1). To look for significant factors
influencing the bycatch estimates for each of the five esti-
mation methods, I used ANOVAs on all main effects and all
2-way and 3-way interactions of the main effects for each
estimator. Although 7-way interactions were possible in the
normally distributed simulations and 8-way interactions
were possible in the delta lognormal simulations (Table 1 ),
I stopped the analysis at 3-way interactions because of the
difficulty in interpreting higher level interactions. Main
effects were the following: mean catches offish and shrimp,
CVs of fish catches, CVs of shrimp catches, CVs of hours
fished, number of observations or observer coverage, cor-
relation coefficient in the normally distributed simula-
tions, and the probabilities of catching fish and shrimp
in the delta lognormal simulations. In these ANOVAs, the
response variable was the percent bias for each method, as
calculated above.
Results
Field sampling
I observed 16 tows from five trips in Pamlico Sound
between July and October 1995 and 24 tows from five trips
in the Cape Fear River between August and October 1995
(Table 2). According to the 1995 trip tickets, these months
comprised the peak of the summer brown shrimp and fall
white-pink shrimp seasons; 77% of the total shrimp catch
and 75% of the total trips in the northern region and 63%
of the total shrimp catch and 54% of the total trips in the
southern region occurred during those months. All observed
tows were daytime tows, which is when fishing generally
occurs in these areas. All nets sampled in July were 2-seam
or 4-seam flat trawls, designed to catch brown shrimp, and
492
Fishery Bulletin 101(3)
Table 2
Characteristics of boats and fishing operations observed in North Carolina waters in 1995
trawlers operating using their standard operating procedures. Each enti-y represents one c
fishing trip lasted one day. Avg. h/tow = mean hours per tow.
All boats were commercial shrimp
bserved fishing trip. Each observed
Area
Month
Boat name
Boat length (m
) No.
of tows
No
of nets
Headrope length (m
) Avg. h/tow
Pamlico Sound
Jul
Last Toy
8.3
5
2
9.8
1.2
Aug
Islander
8.9
2
18.4
1.1
Aug
Last Toy
8.3
2
13.1
1.3
Sep
Islander
8.9
5
18.4
1.1
Oct
Islander
8.9
2
18.4
1.2
Average
8.7
3.2
1.2
15.6
1.2
Cape Fear River
Aug
Cajiin Lady
16.0
9
2
18.0
0.9
Aug
Cajiin Lady
16.0
2
2
18.0
0.7
Sep
Sea Mullet
14.2
4
2
14.8
1.3
Sep
Cajun Lady
16.0
5
2
18.0
1.0
Oct
Dorothy Glen
11.4
4
1
19.7
1.9
Average
14.7
5
1.8
17.7
1.2
Table 3
Landed and observed shrimp catch (heads-on kg landed), effort (total number of trips and days/trip), and catch per unit of effort
(CPUE, kg /trip and kg/day) for regions within North Carolina. Information on fleet totals was obtained from the North Carolina
Department of Marine Fisheries trip ticket database for vessels fishing during July through October 1995 in the Northern region
(Pamlico Sound and tributaries), and during August through October 1995 in the Southern region (the Cape Fear River and nearby
waters). Observations were conducted during these same months in Pamlico Sound and the Cape Fear River.
Shrimp catch
No. of
Mean
CPUE
Days
CPUE
Region or location
(kg landed)
trips
days/trip
(kg/trip)
fished'
(kg/day)
Fleet totals
Northern
2,018,612
3196
3.64
631.6
11,633
173.5
Southern
122,893
1716
3.48
71.5
5972
20.6
Observations
Pamlico Sound
278
5
1
55.6
5
55.6
Cape Fear River
867
5
1
173.4
5
173.4
^ This value represents the maximum days fished because the calculations are
between landings.
based on the assumption that fishing takes place every allowable day
both tongue trawls and flat trawls were sampled in August
through October. Tongue trawls are modified mongoose
trawls that have a higher vertical profile for catching
white shrimp. In addition, the tongue trawls had a greater
headrope length than the flat trawls; therefore many of the
fishermen switched from pulling two flat trawls to pulling
one larger tongue trawl. Tows typically lasted around one
hour The observed catch of shrimp per day in the Cape
Fear River was almost three times higher than the observed
catch of shrimp per day in Pamlico Sound (Table 3).
Total shrimj) landings and total shrimp trips during the
observed months from the 1995 trip ticket database were
used as the expansion factors in the estimates. Over half
of the total .shrimp landings, or 2,018,622 kg, were caught
in the northern region between July and October and only
122,893 kg came from the southern region between August
and October, the months that corresponded to the observa-
tions. Although there were about twice the number of trips
and days fished the northern region, the average catch per
trip (kg/trip) from the northern region was almost nine
times higher than the catch per trip from the southern
region (Table 3).
The different estimation methods made a tremendous
difference in the estimates of bycatch, but the differences
were exactly opposite in the two geographic regions and
varied somewhat by species. Total bycatch estimates de-
rived with the basic F:S ratio estimator (mean of the ratios)
by both weight and number were two to seven times higher
than those based on the CPUE-mean-per-unit method for
all species in the northern region, and about two to five
times lower by both weight and number for all species in
the southern region (Table 4). For Atlantic croaker and
Diamond: Estimation of shrimp trawl bycatch
493
Table 4
Total bycatch
in weight and numbers estimated from observation data
using different estimation methods. The CPUE-mean-
per-unit estimator (CPUE=
catch per unit of effort), w
hich is based on the catch per day.
uses day as a proxy for sample size. The
basic F:S ratio estimator is
the mean of indi
vidual fish (F) to shrimp (S)
ratios, and the grand F:S ratio estimator
is the ratio of
the mean catch of fish to the mean catch of s
hrimp. The Hartley-Ross method is the basic F:S ratio estimator corrected for small |
sample sizes. AC = Atlantic croaker, SP = spo
t, and WF = weakfish. The northern region includes Pamlico Sound and
ts tributaries.
and the southern region includes from the C
ape Fear River to the New River. Estimates
are for July through October 1995 in the
northern region and Angus
t through Octobei
1995 in
the southern region
. See text for calculations. Equations for the 95% CL are |
from Cochran
(1977), Equations 2.24, 6.12, and 6.14.
Bycatch estimate by weight (millions of kg)
F:S ratio estimator
CPUE-mean-per-unit
estimator
Region and
Basic
Grand
Hartley-Ross
species
Total wt.
±95% CL
Total wt
±95% CL
Total wt.
±95% CL
Total wt.
±95% CL
Northern
AC
0.6
0.5
2.1
4.1
1.7
0.1
2.6
0.1
SP
0.5
0.8
2.9
5.6
1.5
0.3
2.3
0.3
WF
0.6
1.2
1.5
1.5
1.9
0.4
1.7
0.4
Southern
AC
0.1
0.2
0.03
0.05
0.02
0.09
N/Ai
N/A
SP
0.02
0.05
0.006
0.01
0.003
0.03
N/A
N/A
WF
0.2
0.2
0.03
0.04
0.2
0.1
N/A
N/A
Bycatch estimate by number (millions)
F:S ratio estimator
CPUE-mean-per-unit
Region and
estimator
Basic
Grand
Hartley-Ross
species
Total no.
±95% CL
Total no
±95% CL
Total no.
±95% CL
Total no.
±95% CL
Northern
AC
28.1
23.7
186.0
263.8
84.8
8.4
144.1
8.4
SP
18.7
27.2
146.5
305.5
56.5
11.7
109.2
11.7
WF
11.8
20.8
36.2
34.8
35.7
6.4
35.9
6.4
Southern
AC
13.7
20.9
4.1
7.1
1.7
12.7
N/A
N/A
SP
1.5
3.4
0.4
0.6
0.2
1.8
N/A
N/A
WF
19.1
29.2
3.5
4.9
2.3
16.3
N/A
N/A
' The estimator
gave negative estimates for bycatch.
spot, the grand F:S ratio estimate (ratio of the means) was
intermediate between the basic F:S ratio estimate and the
CPUE-mean-per-unit estimate in the northern region, but
was lower than either of the other estimates in the south-
em region. The grand F:S ratio estimate for weakfish was
larger by both weight and number than either of the other
two estimates in the northern region and was the smallest
in the southern region. The Hartley-Ross bias-corrected
F:S ratio estimator gave estimates for the northern region
that fell between the basic and grand F:S methods but gave
negative estimates for the southern region (Table 4). CVs of
the catch rates were generally larger for the basic F:S ratio
estimator method than for CPUE-mean-per-unit estima-
tor for spot in the northern region and Atlantic croaker in
both regions, and smaller for spot in the southern region
and weakfish in both regions. CVs estimated by the grand
F:S ratio method were much smaller than those for either
of the other methods for the northern region and much
larger than the others for the southern region (Table 5).
The variance of the catch rates with both methods was
usually much larger than the mean (sometimes by an or-
der of magnitude), indicating that catches were aggregated
(Table 5). The confidence intervals around the bycatch esti-
mates were huge regardless of method because of the small
number of observed fishing days and the large variability
in catch rates (Table 4).
494
Fishery Bulletin 101(3)
Table 5
Observed catch rates in weight and numhers for selected species obtained with different estimation methods from field data. Obser-
vations in Pamlico Sound took place in July through October 1995 and observations in the Cape Fear River took place in August
through October 1995. The CPUE-mean-per-unit estimator (CPUE=catch per unit of effort), which is based on the catch per day,
uses day as a substitute for sample size. The basic F:S ratio estimator is the mean of the individual fish (F) to shrimp (S) ratios,
and the grand F:S ratio estimator is the ratio of the mean catch offish to mean catch of shrimp. AC = Atlantic croaker, SP = spot,
and WF = weakfish. See text for calculations. Equations for the standard deviations are taken from Cochran (1977), Equations 2.20
and 2.45.
Observed catch rate by
weight
F:S ratio estimator
Basic
Grand
CPUE-mean
per-unit estimator
kg fish/kg
shrimp
SD
CV(%)
kg fish/kg
shrimp
SD
CV(%)
species Avg. kg/day
SD CV (%)
Pamlico Sound
AC 49.64
34.79 70
1.62
1.05
154
0.86
0.05
6
SP 41.57
52.40 126
1.46
2.21
152
0.72
0.17
16
WF 53.04
79.90 151
0.75
0.60
80
0.92
0.15
16
Cape Fear River
AC 24.01
21.56 90
0.27
0.31
116
0.14
0.61
438
SP 4.14
6.71 162
0.05
0.06
129
0.02
0.18
757
WF 31.80
30.10 95
0.25
0.24
94
0.18
0.81
443
Observed catch rate by
number
CPUE-mean
per-unit estimator
F:S ratio estimator
Basic
Grand
no. fish/kg
shrimp
SD
CV(%)
no. fish/kg
shrimp
SD
CV(%)
species Avg. no./day
SD CV (%)
Pamlico Sound
AC 2418
1639 68
92.15
105.29
114
42.02
3.37
8
SP 1610
1885 117
72.56
121.92
168
27.97
4.69
17
WF 1014
1438 142
17.92
13.89
77
17.62
2.54
14
Cape Fear River
AC 2287
2826 124
33.70
46.90
139
13.20
83.35
632
SP 257
464 181
3.05
4.16
136
1.48
12.10
816
WF 3198
3941 123
28.09
32.14
114
18.45
106.75
579
Bycatch simulations
For the normally distributed data, the CPUE-mean-per-
unit estimator was the only estimator whose estimated
bycatch was not significantly different than the actual sim-
ulated bycatch {% bias=0.006, P=0.94). All four of the ratio
estimators significantly overestimated bycatch (Table 6),
although the average bias was less than a 1% overestimate
for the grand F:S and grand CPE ratio estimators. The
basic F:S ratio estimator and the basic CPE ratio estimator
both overestimated bycatch by 300-4007^ (Table 6). Using
a model that included all main effects and all 2-way and 3-
way interactions in the ANOVA, I found that the CV of the
auxiliary variable (either shrimp catch or hours fished) was
a significant main effect for all four of the ratio estimators,
but there were no significant main effects for the CPUE
mean-per-unit estimator (Table 7). Observer coverage was
also a significant main effect for the F:S and CPE grand
ratio estimators, but was not significant for the basic F:
S or CPE ratio methods. The grand F:S ratio estimator,
the grand CPE ratio estimator, and the basic F:S ratio
estimator all showed several significant 2-way and 3-way
interactions (Fig. 2), whereas the basic CPE ratio estimator
had no significant 2-way or 3-way interactions. The CPUE-
mean-per-unit estimator showed only two significant 3-
way interactions among variables, and observer coverage
occurred in both. The correlation between fish catches and
shrimp catches was a significant main effect for the basic
Diamond: Estimation of shrimp trawl bycatch
495
Table 6
Mean percent bias of each of the estimators with normal
and delta lognormal distributions offish (F) and shrimp (S)
from simulated data. Percent bias (Eq. 17) was calculated
separately for each simulated fleet. N = the number of
fleets in each analysis. The * indicates that the mean
estimated bycatch is significantly different than the mean
actual bycatch in a paired sample t-test (P<0.05). The
CPUE-mean-per-unit estimator (CPUE=catch per unit
of effort) uses unit effort as a proxy for sample size. The
basic F;S ratio estimator is the mean of the individual fish
to shrimp ratios, and the grand F:S ratio estimator is the
ratio of the mean catch offish to the mean catch of shrimp.
The basic CPE ratio estimator is the mean of the ratios of
catch per effort, where effort is a variable measure such as
hours fished, and the grand CPE estimator is the ratio of
the mean catch offish to the mean estimate of effort.
Estimator
CPUE mean-per-unit
Basic F;S ratio
Grand F:S ratio
Basic CPE ratio
Grand CPE ratio
Mean % bias
Normal
distribution
TV =21,600
Delta lognormal
distribution
N= 118,810
0,006
427.80*
0.65*
336.13*
0.46*
0.09
9.98*
12.23*
30.75*
0.47*
F:S estimator, and showed significant interactions with
other parameters in both the grand F:S and grand CPE
ratio estimators.
For the delta lognormally distributed data, the CPUE-
mean-per-unit estimator was the only estimator whose
estimated bycatch was not significantly different than
the actual simulated bycatch (% bias=0.087%, P=0,64),
All four of the ratio estimators significantly overestimated
bycatch (Table 6), with estimates ranging from a less than
1% overestimate using the grand CPE ratio estimator to a
30% overestimate with the basic CPE ratio estimator (Table
6). Using all 2-way and 3-way interactions in the ANOVA,
I found that significant main effects for both the basic and
grand F:S ratio estimators were the probability of catching
shrimp and the CV of the shrimp catch. The CV of the fish
catch and observer coverage were also main effects in the
grand F:S ratio method. The probability of catching fish was
an additional main effect in the basic F:S ratio method. The
only significant main effect in both CPE ratio estimators was
the CV of effort, and the only significant main effect in the
CPUE-mean-per-unit method was the CV of the fish catch.
All five methods exhibited several statistically significant
2-way and 3-way interactions (Table 7),
Discussion
The differences in bycatch estimates generated from the
field data show how confusing bycatch estimation can be
SV\n(^P
\ 400
%' 300
',, which is the catch of the bycatch species, and.*,, which is
the auxiliary variable, is a straight line through the origin
(indicating that the ratio of bycatch to shrimp or the catch
of bycatch per hour fished is constant over all observations)
and if the variance ofy, about this line is proportional to
X,. In practice, these conditions rarely hold true. The ratio
of fish to shrimp catches and the bycatch per hour fished
from field data often vary considerably among observa-
tions because of the patchy spatial distributions offish and
shrimp, seasonal differences in the relative abundances of
fish and shrimp, movements associated with development
through different life stages, and environmental factors. In
addition, the bias of a ratio estimator is on the order of 1/n,
indicating that the bias will be small if « is large (Cochran,
1977). In practice, n, or the number of onboard bycatch
observations, is often very small, particularly if the data
are stratified by season or area, leading to large biases in
ratio estimators.
The Hartley-Ross ratio estimator, which is a form of the
basic ratio estimator method, may in some cases be an
unbiased or less biased ratio estimator for small samples
(Cochran, 1977). However, the Hartley-Ross method was
not effective for the field data in the present study, giving
nonsensical negative estimates of bycatch for all species
in the southern region, although the estimates in the
northern region were generally (but not always) some-
where between the basic and grand F;S ratio methods.
The problems with the Hartley-Ross ratio estimator in the
southern region may have been due to two factors: 1) the
very low value for total shrimp landings from trip tickets
in the southern region, and 2) discrepancies between the
observed average catch of shrimp per day and the fleet
shrimp catch per day from the trip ticket database. The
Hartley-Ross equation starts with the mean of the indi-
vidual fish to shrimp ratios expanded by the total shrimp
landings (the basic F:S ratio estimator) and corrects the
estimate based on the sampling fraction multiplied by
a quantity that includes the average observed catch of
shrimp per day (Eq. 8). The total shrimp landings recorded
on trip tickets for the southern region were extremely low,
about 16 times lower than the total shrimp landings in the
northern region, although the number of days fished was
about half as many in the southern region. In addition, the
average shrimp catch per day on vessels that I observed in
the southern region was much greater than the average
reported on trip tickets (173.4 kg per day observed vs. 20.6
kg per day from trip tickets), whereas the average shrimp
catch per day of shrimp that I observed in the northern
region was much lower than the catch per day shown on
trip tickets (55.6 kg per day observed vs. 173.5 kg per day
from trip tickets). The result of this combination of factors
was that the estimated total bycatch before correction in
the southern region was very small due to the low amount
of total shrimp landings, whereas the correction factor was
very large because of the high observed average catch of
shrimp, leading to negative estimates of total bycatch.
These problems did not occur in the northern region. Low
shrimp landings in the southern region compared to the
northern region may have been due to an actual difference
in the abundance of shrimp or differences in fishing habits
such as a smaller number of nets per boat, tows per day,
or tow times per tow in the southern region. However, it is
also possible that more fishermen in the southern region
than the northern region keep their catch or sell part
of their catch independently without generating a trip
ticket, which would reduce the total landings of shrimp
in the trip ticket database. The differences in the average
observed catch of shrimp per day were probably due to a
combination of factors, most of them based on the prob-
lem of nonrandom or nonrepresentative sampling of boats.
Because I depended on volunteer fishermen, the observed
shrimp boats and captains were not randomly selected. In
addition, because no records are kept of the boat size, gear
used, fishing habits, or effort history of fishermen in the
fleet, sampled boats could not be compared to unsampled
boats for these factors. However, most of the fishermen
whose boats I observed in the Cape Fear River (the south-
ern region) owned large boats and made an average of 5
tows per day, whereas the fishermen I observed in Pamlico
Sound (the northern region) generally had smaller boats
and made an average of 3.2 tows per day. If the fishermen
whose boats I observed in the Cape Fear River fished more
than the average number of tows per day and the observed
fishermen in Pamlico Sound fished fewer than the aver-
age number of tows per day, then the catch per day values
would show these discrepancies. Other factors could have
been differences between observed boats and average
boats in the number of nets per boat, or tow times.
All of the methods that I used for bycatch estimation for
the field data were based on the summed catches over all
tows on a single day, because in this study the variance of
catches among tows within days was much less than the
variance among days. The use of tows as the basic unit of
effort would therefore have underestimated the total vari-
ance. Sampling only day-trips probably contributed to the
covariance among tows because tows spread over several
days (and probably several locations) in a multiday trip
would probably vary more among tows within a trip than
tows in a single day. For randomly sampled multiday trips,
estimation methods based on tows rather than days or trips
may be preferred to those based on a trip as the unit of ef-
fort because the sample size of tows increases faster than
the sample size of days or trips, which would tighten the
confidence intervals around the estimates. However, the
use of tows as the unit of effort could be considered pseu-
Diamond: Estimation of shrimp trawl bycatch
499
doreplication (Hurlbert, 1984) and could lead to erroneous
variance estimates if the tows in a trip are not independent
samples (Cochran, 1977). The choice of whether to use trips
or tows as the unit of effort is dependent on two factors: 1)
whether there is a high degree of covariance among tows
in a trip, and 2) whether there is an independent estimate
of the average number of tows per trip to use as an expan-
sion factor.
Confidence intervals around the bycatch estimates are
not symmetrical, although they are shown in Table 4 as
symmetrical to allow for easier comparisons between the
methods in the field study. Because of the small numbers
of observations, most of the confidence intervals in the
field study were larger than the means, with the general
exceptions of the grand ratio estimators for all species in
the northern region, which were surprisingly small. Most
grand ratio estimators underestimate the true catch rate
and are positively skewed unless the sample size is greater
than 30 and the CVs of both the observed fish catch and
the auxiliary variable are less than 10% (Cochran, 1977).
As seen in Table 5, CVs of the observed fish catch from field
data are rarely as low as 10%, and many are over 100%.
The very small confidence intervals for all species in the
northern region, and the very large confidence intervals for
all species in the southern region generated by the grand
ratio estimators are due to the nonrandom sampling of
boats for the average catch of shrimp in both areas. This
nonrandom sampling affects the confidence intervals be-
cause the average catch per day is a term in the denomi-
nator of the equation used to estimate the variance of the
grand ratio estimator (Eq. 2.45 in Cochran, 1977). A very
large value for the average catch per day from trip tickets
compared to the value from observations as in the north-
ern region causes an underestimate in the variance and
reduces the confidence intervals, whereas a small value
for the average catch per day from trip tickets compared
to the value from observations as in the southern region
causes an overestimate in the variance and increases the
confidence intervals.
The field data shown here indicate some of the problems
that are peculiar to observing and estimating bycatch in
shrimp trawl fisheries in comparison to other fisheries.
First, there are several hierarchical levels of variability
that are ignored because of the logistical difficulties of sam-
pling shrimp trawls. If the National Marine Fisheries Ser-
vice (NMFS) protocol for shrimp trawl bycatch is followed,
only one sample of the catch is taken from a net because
of the large numbers of organisms caught in a typical tow.
The NMFS protocol depends on the observer thoroughly
mixing the catch so that a single sample characterizes
the entire catch without variance, but mixing the catch to
obtain a random sample is sometimes difficult because of
the weight of the catch, the position of the culling tray, the
size of the boat, or weather conditions. In addition, some
species such as crabs may redistribute themselves after
the catch is mixed by simply walking away. Stender and
Barans (19941 found differences in fish-to-shrimp ratios
when sampling the net compared with enumerating ev-
erything in the net. However this source of variability is
not measured when following the NMFS protocol and not
included in the bycatch estimates. Second, only one net
is generally sampled per tow, although the boat may tow
two, four, or more nets. There is therefore an expansion
from the sampled net to the number of nets per tow so that
variance among nets is ignored, and this process also adds
error. Third, the expansion term, regardless of whether the
total shrimp landings or the total shrimp effort is used, is
assumed known without error. To include the error in the
expansion term further widens the confidence intervals
around the final estimates (Diamond and Hanan'").
One of the most interesting findings from the simulations
was that all the methods tended to overestimate bycatch.
None of the overall bycatch estimates, and relatively few of
the individual fleet simulations, generated underestimates
of the actual values. Although the mean-per-unit and grand
ratio estimates overestimated bycatch by less than 19f , if
the bycatch is large enough, these estimators could errone-
ously add hundreds of thousands offish to the catch-at-age
matrices used in stock assessments. Inaccurate stock as-
sessments could have consequences for the management
of fisheries, particularly for species like red snapper that
are managed by quotas on the directed fisheries that are
based on the level of bycatch or that have target levels set
for rebuilding fish stocks. One method that might be used
to "correct" bycatch estimates for the mean-per-unit esti-
mator would be to use the estimator to calculate the catch
of the target species from the observations, and then to
compare the estimated target species catch with the total
landings. Although this correction method assumes that
the total landings of the target species are accurate (which
is rarely a valid assumption), comparison of the estimated
total catch of the target species to the actual landed catch
might help to pinpoint biases and to adjust the estimated
bycatch.
Because of the differences in estimates generated by
the different methods of estimating bycatch, interpreta-
tions of bycatch estimates and comparisons of bycatch
studies should be made very cautiously. It is often difficult
to tell in past studies whether estimates were generated
by basic F:S or grand F:S methods, but basic F:S methods
overestimate bycatch to a much greater degree. Because of
the statistically significant 2-way and 3-way interactions
among parameters, it is unlikely that estimates generated
by one method can be converted or corrected to other meth-
ods, so bycatch estimates made over time using different
methods should not be directly compared. In addition,
any bycatch estimate should include some indication of
the variance around either the estimate or the catch rate,
although variance estimates can be misleading if samples
are not random. Finally, estimates of the weight or num-
ber of species taken as bycatch, no matter how large or
small, are meaningless without an estimate of population
abundance. Small populations could be harmed by rela-
tively small amounts of bycatch, whereas large populations
'6 Diamond, S. L., and D. Hanan. 1986. An estimate of harbor
porpoise mortality in California set net fisheries April 1, 1983
through March 31, 1984. National Marine Fisheries Service
Admin. Report SWR-86-15, 40 p.
500
Fishery Bulletin 101(3)
could be able to withstand even large amounts of bycatch.
For this reason, the consequences of bycatch can only be
evaluated if examined in conjunction with some estimate
of stock size.
Acknowledgments
My sincerest gratitude goes to the fishermen who allowed
me on their boats: Allan Hines, Bud George, Pete Dixon,
Ben Ingraham, H.O. Golden, Tommy Peters, Al Gillikin,
and Brad Styron. Bud George also provided many helpful
suggestions on how to weigh the catch. I also appreciate the
help given to me by Oliver and Tina Lewis, Bimbo Melton,
Tony Cahoun, Gracie Golden, Jim Bahen, John Schoolfield,
Beth Bums, Bob Hines, and Jim Murray. Trish Murphey,
Mike Street, and Dee Lupton from NCDMF provided the
shrimp trip ticket data. Peter Lamb, Tyler Stanton, Sue
Zwicker, Pam Robinson, Walter Mayo, Amy Makepeace,
Martin Gallagher, Dawn O'Harra, and Jim Armstrong
helped to sort and identify the bycatch species. Jim Rice,
Larry Crowder, Joe Hightower, Ken Pollock, and Doug
Vaughan provided valuable input on earlier drafts of this
manuscript. Thanks also to Rich Strauss and Richard Ste-
vens for their help in using Matlab software. This manu-
script was significantly improved by the comments of Scott
Nichols and two anonymous reviewers. This research was
supported by a National Science Foundation Pre-doctoral
Fellowship, the J. Francis Allen Scholarship from the
American Fisheries Society, the Joseph L. Fisher Disserta-
tion Award from Resources for the Future, and MARFIN
Grant no. NA57FF0299.
Literature cited
Alverson, D. L. 1994.
1994. A global assessment of fisheries bycatch and discards,
233 p. FAO (Food and Agriculture Organization) of the
United Nations, Rome, Italy. [ISBN 92-5-103555-5.]
Cochran, W. G.
1977. Sampling techniques, 428 p. John Wiley and Sons,
New York, NY.
Grouse, D. T., L. B. Crowder, and H. Caswell.
1987. A stage-based population model for loggerhead sea
turtles and implications for conservation. Ecology 68(5):
1412-1423.
Fahy, W. E.
1966. Species composition of the North Carolina industrial
fish fishery. Comm. Fish. Rev. 28(7):l-8.
Fisheries Statistics of the United States.
1977. U.S. Fish and Wildlife Service, Bureau of Commercial
Fisheries, Statistical Digest 71, 407 p.
Hurlbert, S. H.
1984. Pseudoreplication and the design of ecological field
studies. Ecol. Monogr. 54 (2):187-211.
Julian, F, and M. Beeson.
1998. Estimates of marine mammal, turtle, and seabird
mortality for two California gillnet fisheries: 1990-1995.
Fish Bull. 96:271-284.
Stender, B. W., and C. A. Barans.
1994. Comparison of the catch from tongue and two-seam
shrimp nets off South Carolina. North Am. J. Fish.
Manage. 14:178-195.
Walsh, W.A., and P. Kleiber
2001. Generalized additive model and regression tree
analysis of blue shark iPrionace glauca) catch rates by
the Hawaii-based commercial longline fishery. Fish. Res.
52(2):115-131.
501
Abstract — Adaptive cluster sampling
(ACS) has been the subject of many
publications about sampling aggregated
populations. Choosing the criterion
value that invokes ACS remains prob-
lematic. We address this problem using
data from a June 1999 ACS survey
for rockfish, specifically for Pacific
ocean perch {Sebastes aliitus). and for
shortraker (S. borealis) and rougheye
(S. aleutianus) rockfish combined. Our
hypotheses were that ACS would out-
perform simple random sampling ( SRS )
for S. aliitiis and would be more appli-
cable for S. alutiis than for S. borealis
and S. aleutianus combined because
populations of S. alutus are thought
to be more aggregated. Three alterna-
tives for choosing a criterion value were
investigated. We chose the strategy that
yielded the lowest criterion value and
simulated the higher criterion values
with the data after the survey. System-
atic random sampling was conducted
across the whole area to determine the
lowest criterion value, and then a new
systematic random sample was taken
with adaptive sampling around each
tow that exceeded the fixed criterion
value. ACS jaelded gains in precision
(SE) over SRS. Bootstrapping showed
that the distribution of an ACS estima-
tor is approximately normal, whereas
the SRS sampling distribution is
skewed and bimodal. Simulation
showed that a higher criterion value
results in substantially less adaptive
sampling with little tradeoff in preci-
sion. When time-efficiency was exam-
ined, ACS quickly added more samples,
but sampling edge units caused this
efficiency to be lessened, and the gain in
efficiency did not measurably affect our
conclusions. ACS for S. alutus should
be incorporated with a fixed criterion
value equal to the top quartile of previ-
ously collected survey data. The second
hypothesis was confirmed because ACS
did not prove to be more effective for S.
borealis-S. aleutianus. Overall, our ACS
results were not as optimistic as those
previously published in the literature,
and indicate the need for further study
of this sampling method.
Applications in adaptive cluster sampling
of Gulf of Alaska rockfish
Dana H. Hanselman
Terrance J. Quinn II
School of Fisheries and Ocean Sciences
University of Alaska Fairbanks
11275 Glacier Hwy.
Juneau, Alaska 99801
E-mail address (for D. H. Hanselman); ftdhh@uaf.edu
Chris Lunsford
Jonathan Helfetz
David Clausen
Auke Bay Laboratory
Alaska Fisheries Science Center
National Manne Fishenes Service
11305 Glacier Hwy.
Juneau, Alaska 99801
Manuscript approved for publication
•30 January 2003 by Scientific Editor.
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:501-.513 (20031.
In nature, populations are sometimes
distributed in a patchy, rare, or aggre-
gated manner. Conventional sampling
designs such as simple random sam-
pling (SRS) do not take advantage of
this spatial differentiation. Thompson
(1990) introduced a sampling design
called adaptive cluster sampling (ACS)
to survey these types of distributions.
Adaptive cluster sampling, in theory,
can be much more precise for a given
amount of effort than conventional
sampling designs (Thompson, 1990).
In practice, however, this is not always
the case. In some cases, the variance
is greatly reduced, but bias is induced
from stopping rules and criterion values
that are sometimes changed mid-survey
(Lo et al., 1997). In 1998, we conducted
a survey on Gulf of Alaska rockfish in
which ACS was efficient and successful,
but the gains in precision, if any, were
small compared to those of a SRS of the
same size (Quinn et al., 1999; Hansel-
man et al., 2001).
Recently papers about ACS have in-
cluded efficiency comparisons (Christ-
man, 1997), restricted ACSs (Lo et al.,
1997; Brown and Manly, 1998), boot-
strap confidence intervals (Christman
and Pontius, 2000), and bias estimates
(Su and Quinn, 2003). However, little
work has been done on determining
the criterion value that, when exceeded,
invokes additional sampling. In the fol-
lowing study, we examine the details for
choosing this criterion value by using
data from a 1999 field survey for Gulf
of Alaska rockfish. We then simulate
the outcome of the experiment with dif-
ferent criterion values after the survey.
We also compare the efficiency of ACS
to SRS.
In the basic adaptive cluster sam-
pling (ACS) design, a simple random
sample (SRS) of size n is taken; if _y
(the variable of interest) exceeds c (a
criterion value), then neighborhood
units are added (e.g. units above, be-
low, left, and right in a cross pattern.
Fig. 1) to the sample. These are called
network units. If any network unit has
y>c, then its neighborhood is added.
Units that do not exceed the criterion
are called edge units, and sampling
does not continue around them. This
process continues until no units are
added or until the boundary of the area
is reached (Thompson and Seber, 1996).
Neighborhoods can be defined in any
general way. The only condition is that
if unit i is in the neighborhood of^, then
unit j is in the neighborhood of i. The
"unbiasedness" of the estimators relies
on all neighborhood units of >'>c being
sampled. If logistics cause the sampling
to be curtailed before the sampling is
complete, then biased estimators can
502
Fishery Bulletin 101(3)
Cross Pattern
A3
A3
A2
A3
A3
A2
Al
A 2
A3
A3
A 2
Al
R
Al
A 2
A3
A3
A 2
Al
A 2
A3
A3
A 2
A3
A3
Line
ar Pattern
A4
A3
A2
Al
R
Al
A2
A3
A4
Figure 1
Diagram of a basic cluster sampling design, showing the
maximum possible number of adaptive hauls for the cross
(S=3) and the linear (S=4) patterns with the imposition of a
stopping rule. The initial random tow is denoted as "R," anc
the adaptive tows as "A" and their respective level number
result. For our study, all samples were called "tows" because
our study was a trawl survey.
When little information is available to preset a fixed
criterion value, order statistics are often used to choose
a criterion value (Thompson and Seber, 1996). The basic
idea is that an initial random sample is conducted. Next,
the values of the random tows are ordered, and ACS is
conducted around the top r stations. The variable r is de-
cided by the experimenter and depends on the amount of
resources available and the suspected aggregation of the
population. The criterion value is then set at the value of
the next highest tow ( ;+ 1 ). This was the design used in the
1998 adaptive cluster sampling survey for rockfish (Quinn
et al., 1999, Hanselman et al., 2001). The use of order
statistics has several limitations, however. First, initial
random samples must be taken before the adaptive phase
can begin. This procedure can be inefficient, because the
experiment may have to move a large distance back to the
previous tows that exceeded the criterion, by which time
the aggregation may have moved or dispersed. In some
cases, this procedure may result in a very small criterion
value that leads to an overwhelming amount of adaptive
sampling around some tows. Second, the process of achiev-
ing simple unbiased estimates of abundance is more com-
plicated with order statistics because the criterion value is
dependent on the sampling.
In our study, we address methods to avoid these limita-
tions and illustrate these methods with a 1999 ACS survey
for Gulf of Alaska rockfish. The primary target of the sur-
vey was Pacific ocean perch iSebastes alutus [POP] ). These
fish have extremely uncertain biomass estimates in the
Gulf of Alaska (Heifetz et al.'). The estimates are based in
part on a standardized stratified random survey conducted
by the National Marine Fisheries Service every three years
(every two since 2000). This uncertainty is likely due to
their highly clustered distribution (Lunsford, 1999) and
has led to two independent surveys ( 1998, 1999) to test the
benefits of ACS in sampling POP. Shortraker (S. borealis)
and rougheye (S. aleutianus) rockfish combined (SR-RE)
are also tested to compare the results of a population that
is considered highly clustered (POP) versus one that is
considered more uniformly distributed (SR-RE). SR-RE
are combined because they co-occur in identical habitat
and are managed as a complex.
Materials and methods
In June 1999, ACS was carried out between 140° and 144°
west longitude near Yakutat in the Gulf of Alaska (Fig. 2).
Approximately 75% of sampling was directed toward the
POP depth stratum ( 180-300 m) and 25% directed toward
SR-RE depths (300^50 m). A 182-ft. factory trawler, the
Unimak, was chartered to conduct trawl samples. Fish-
ing and field operations are described in Clausen et al.^
Duration of all trawl hauls was 15 (POP) and 30 (SR-RE)
minutes on the bottom. SR-RE tows were made parallel to
the depth contours in a linear pattern (Fig. 1) because the
slope that SR-RE inhabit is too steep for perpendicular
tows. Travel time between all tows was recorded to exam-
ine time efficiency.
Initially, a set of systematic random tows was conducted
from west to east across the entire study area to determine
the criterion value. Samples were chosen systematically by
longitude and distributed randomly by depth within each
longitudinal strip. This procedure was a necessary proxy for
simple random sampling because of poorly known bathym-
etry in the area. The use of simple random latitudes and
longitudes often results in the selection of sites that are well
out of the sampling depth interval. After random sampling
was completed, we compiled and examined the data to set
the criterion value. Criterion values were chosen based on
a hierarchy of three alternatives described below. Next, we
conducted a new set of random tows from east to west across
the area, in which any tows exceeding the criterion value
were adaptively sampled. A distance of 0.19 km (0.1 nmi)
was used between all adaptive tows and the initial random
tow to avoid depletion effects on the catches.
' Heifetz, J., D. L. Courtney, D. M. Clausen, J. T. Fujioka, and J. N.
lanclli. 2001. Slope rockfish. In Stock as.scssment and fish-
ery evaluation for the groundfish resources of the Gulf of Alaska,
72 p. North Pacific Fishery Management Council. 605 W. 4"'
Ave. Suite 306, Anchorage, AK 99501.
2 Clausen, D.M.,D.H. Hanselman, C. Lunsford, T. Quinn II, and J.
Heifetz. 1999. fZ/i/niaA' enterprise cruise 98-01 rockfish adap-
tive sampling experiment in the central Gulf of Alaska 1998,
49 p. Auke Bav Lab, NMFS, NOAA, 1 1305 Glacier Hwy, Auke
Bay, Alaska, 99801.
Hanselman et al : Applications in adaptive cluster sampling of Gulf of Alaska rockfisfi
503
145°W
144"W
I43°W
Ma-w
14 I '-'W
14()°W
bU N
59°N
Figure 2
Map of sampling area in the Gulf of Alaska on the Uniinak 99-01 adaptive sampling cruise. "R" symbols
are the initial random tows for the criterion phase, "r" symbols are random stations in the survey phase,
"A" symbols are adaptive cluster samples.
Three methods were formulated for determining a fixed
criterion value c of POP catch-per-unit-of-effort ( CPUE). ( 1 )
We combined and calibrated past survey and fishing data
to provide the anticipated distribution of CPUE in the 1999
survey. Then we calculated the SO"^*^ percentile of that dis-
tribution as the criterion value. Our rationale was that this
value would correspond to that obtained from order statis-
tics. (Three networks were sampled in 1998; therefore the
criterion value was set to the 4'*^ highest of the ordered 15
initial tows, which corresponded approximately to the 80'^
percentile. ) ( 2 ) We used the mean CPUE of past survey and
fishery data because when we compared the 80"^ percentile
criterion against the 1998 ACS survey's data, the sampling
would have resulted in primarily edge units. (3) After a
representative random sample was taken across the entire
area in 1999, we would use the initial mean CPUE for the
criterion value for the return trip. The rationale for using
mean CPUE above is that in an aggregated population,
the majority of the tows would be less than the mean. The
actual values of the criterion chosen under each alternative
are described in the results.
We chose the SR-RE criterion to be the mean CPUE of
initial tows. We assumed this was a reasonable criterion
value because if the population of SR-RE were somewhat
uniform, a lower value would result in too much ACS, but
mean CPUE would still be low enough to allow higher cri-
terion values to be examined. Although we concentrated on
evaluating criterion alternatives for POP, we present the
SR-RE data to illustrate that different levels of aggregation
could affect how much can be gained with ACS in terms of
precision and efficiency.
A major problem in applying adaptive sampling is that
sampling may continue indefinitely because of a low crite-
rion value. To limit the amount of adaptive sampling, an
arbitrary stopping rule of S levels was imposed. For those
strata where the cross pattern of adaptive sampUng was
used (POP), the stopping rule was S = 3 levels, allowing for
a maximum of 24 adaptive tows around each high-CPUE
random tow (Fig. 1). For the strata with the hnear pattern
of adaptive sampling (SR-RE), the stopping rule was S = 4
levels, for a maximum of eight adaptive tows around each
high-CPUE random tow. This stopping rule differs from
that of the previous year in which we used a stopping rule of
six because we believed that the possible 30-km difference
between the ends of the networks was too large for efficient
sampling (Clausen^). In addition, no adaptive sampling ex-
tended beyond a stratum boundary. The result of adaptive
sampling around each high-CPUE tow was a network of
tows that extended over and, in some cases, delineated the
geographic boundaries of a rockfish aggregation.
504
Fishery Bulletin 101(3)
Statistical analysis of the results was based on adap-
tive cluster sampling (Thompson and Seber, 1996). First,
we estimated the abundance (kg/km) for the targeted
rockfish species from the n initial random tows using the
standard simple random sampling (SRS) estimator Then,
two adaptive estimators of abundance, a Hansen-Hurwitz
estimator (HH) and a Horvitz-Thompson estimator (HT),
were calculated. We computed standard error (SE) as a
measure of precision. The unbiased HH estimator for the
ACS mean is
1 " T "_
(1)
where w^ and y* = the mean and total (respectively) of
the X, observations in the network that
intersects sample unit /.
The HH estimator essentially replaces tows around which
adaptive sampling occurred with the mean of the network
of adaptive tows that exceeded the criterion CPUE.
The unbiased HT estimator for the ACS mean is
1 ^ •'
A'-f-i'or,
(2)
where y^ = the sum of the y-values for the kth network;
K = the number of distinct networks in a sample;
rtf. = the probability that network k is included in
the sample; and
N = the total number of sampling units.
If there are .r^ units in the kth. network, then
N-.v,
(3)
where N = the total number of sampling units;
n = the initial random sample; and
X)^ = the number of units in the network.
The HT estimator is based on the probability of sampling
a network given the initial tows sampled and involves the
number of distinct networks sampled (in contrast to the
HH estimator which is based only on the initial tows). The
HT estimator often outperforms other estimators as seen
in simulation studies (Su and Quinn, 2003). Both estima-
tors use the network samples and initial random samples,
but not the edge units. This sample size is referred to as v'
(convention established by Thompson (1990) and used in
Thompson and Seber (1996)). To include edge units into
the estimates Thompson and Seber (1996) and Salehi
(1999) used the Kao-Blackwcll theorem, which is a com-
plex method that could theoretically result in more precise
estimates. However, it had little effect for the 1998 survey
data (<1% improvement, Hanselman, 2000); therefore
these calculations were not used in our study.
When a stopping rule is used, the theoretical basis for
the adaptive sampling design changes. It may result in
incomplete networks that overlap and are not fixed in rela-
tion to a specified criterion — changing with the pattern of
the population. In contrast, the nonstopping-rule scheme
has disjoint networks that form a unique partition of the
population for a specified criterion. This partitioning is
the theoretical basis for the unbiasedness of pi^^ and fifjj.
Thus with a stopping rule, some bias may be introduced.
Recent simulation studies (Su and Quinn, 2003) have
estimated the bias induced by using a stopping rule on each
estimator with order statistics, but not with a fixed crite-
rion. Because the use of a fixed criterion is design unbiased,
its estimate should be less biased by the stopping rule than
a sample with order statistics. Therefore, we can use the
Su-Quinn simulation results to approximate the maximum
bias induced by the stopping rule. With a stopping rule of
three and the HH estimator, the maximum positive bias is
17% for a highly aggregated simulated population. With
a stopping rule of three and the HT estimator, the maxi-
mum bias is approximately 12%. Considering our design,
we accepted the tradeoff of relatively small bias for gains
in precision and logistical efficiency.
Additionally, nonparametric bootstrap methods were
adapted from Christman and Pontius (2000) and we used
the HH version of the estimates to examine bias from our
survey. Five thousand resamples were performed by using
n for the SRS bootstrap, and the sample size from the origi-
nal criterion value of 220 kg/km ( v) was used for the ACS
bootstrap. Bootstrap distributions of the data were exam-
ined for SRS and ACS designs to examine the capability of
each design to clearly demonstrate a central tendency.
We evaluated two hypotheses: 1) Adaptive sampling
would be more effective in providing precise estimates of
POP biomass than would a simple random survey design;
and 2) Assessment of POP abundance would benefit more
from an adaptive sampling design than would SR-RE be-
cause POP are believed to be more clustered in their dis-
tribution than SR-RE. SRS estimates were obtained from
the initial random tows, and variance estimates were cal-
culated for the initial sample size in ) and for the equivalent
sample size that included the adaptive tows but not the
edge units (\''). This procedure makes the theoretical com-
parison fair because each estimate is based on the same
number of samples. Total sample size including edge units
(v) was not used in the theoretical precision comparison
but was considered when efficiency issues were examined
later These hypotheses were assessed by comparing the
standard errors (SEs) of ACS to those of SRS. Substantial
reductions in SE with ACS for POP would support the
first hypothesis, whereas no reductions of SE using ACS
for SR-RE would support the second hypothesis. This com-
parison is qualitative because relevant significance tests
are unavailable and the two methods are different in terms
of efficiency.
To evaluate different alternatives and criterion values,
each network was reconstructed as if the higher criterion
values had been used in the field. We also examined the
tradeoff between amounts of additional sampling com-
pared with the gains in precision. A comparison was made
oftiu' SRS results by using sample sizes constructed with
the number of possible samples with the time-per-sample
Hanselman et a\: Applications in adaptive cluster sampling of Gulf of Alaska rockfish
505
Table 1
Data used to determine criterion values c for the 1999 adaptive cluster sampling (ACS) survey. Data from a 1998 ACS survey from
a different area is divided by the National Marine Fisheries Service triennial survey data and fishery data from the same area
to obtain gear efficiency values. The mean of these gear efficiencies are then multiplied against triennial and fishery data from
the new area to yield gear-calibrated CPUEs for the new area. Only numbers in bold were used in calculations, n = the number of
observations of that data set; 80% = the 80"^ percentile catch of that data set.
Data source
Year
Mean CPUE (kg/km)
80%
ACS results from different area and year
(divided by)
CPUEs of corresponding previous area from
triennial and fishery data
(equals)
Gear efficiency of the Unimak
(multiplied by)
Prior CPUE data from area for
1999 ACS survey
(equals)
Calibrated CPUE data for
1999 ACS survey
Criterion value c
1998
284.94
Triennial 1993
38.36
1996
46.64
1993-96
42.54
Fishery 1996-98
30.64
1993
7.44
1996
6.12
1993-96
6.71
1996-98
9.32
Mean
7.63
Triennial 1993
40.32
1996
26.50
1993-96
33.92
Fishery 1996-98
19.61
Triennial 1993
307.52
1996
202.06
1993-96
258.69
Fishery 1996-98
149.57
Mean
219.71
223.92
57
7.89
50
27.33
51
18.79
101
14.03
434
28.18
8.14
11.84
15.85
17.39
X
46.74
29
33.50
25
38.85
54
30.47
190
812.67
29
582.52
25
675.63
54
529.90
137
641.69
data we collected. In this comparison we used three new
sample sizes: 1) v,, the number of samples that could have
been taken in the same amount of time as that for a SRS
if sampling time for edge units was neghgible; 2) v^, in
which the edge units had taken the same amount of time
as non-edge units; and 3) v^, in which the average distance
between each tow type was used as effort instead of time
(with edge units included).
survey data (1993, 1996) and fishery data (1996-98) from
the observer program for the same area. This gear coef-
ficient was then multiplied by the same data for the new
area to establish the expected catches. The data used and
the calculations are shown in Table 1. To implement alter-
native 3, we conducted 13 initial POP and 10 initial SR-RE
random tows across the entire area. Catches from these
initial tows gave us the following results for each criterion
alternative:
Results Alternative 1
Formulation of criterion alternatives
A total of 164 tows were conducted for the ACS experiment.
Nearly all tows were made successfully; only a few excep-
tions were deemed untrawlable and moved to the nearest Alternative 2
trawlable bottom. We determined the POP criterion value
for alternatives 1 and 2 (see below) before the survey by
looking at the 1998 ACS results from a different geographic Alternative 3
area, as well as prior survey and fishery data in our study
area. We obtained the criterion value by calculating a gear
efficiency coefficient for the 1998 sui-vey by using NMFS
For alternative 1, the mean of the 80"' per-
centile of the data from Table 1 is 641.69
kg/km. We rounded this downward to c =
540 kg/km ( 1000 kg/nmi) for ease of opera-
tion in the field (the design was originally
in kg/nmi units).
The mean calibrated CPUE for the area
from Table 1 yielded a criterion value c of
220 kg/km (rounded).
In this alternative, the mean CPUEs from
the initial sample in 1999 yielded criterion
values of c = 250 kg/km for POP and c = 418
kg/km for SR-RE.
506
Fishery Bulletin 101 (3)
Table 2
Summary of density estimates (/J) and standard errors (SE) for the 1999 adaptive cluster sampling experiment for the Sebastes
atutus and the S. borealis-S. aleutianus complex, c is the criterion value, r is the number of adaptive networks, n is the initial
sample size, v' is the adaptive sampling size (excluding edge units). SRS = simple random sampling estimator, HH = Hansen-Hur-
witz adaptive estimator, and HT = Horvitz-Thompson adaptive estimator. Alt. = criterion alternative.
Sebastes alutus
Sebastes borealis and S. aleutianus
Alt. 2
Alt. 3
Alt. 1
—
Alt. 3
—
c (kg/km)
>220
>250
>540
>1080
>418
>540
r
6
6
5
3
5
3
n
25
25
25
25
9
9
v'
74
73
55
48
30
14
''sRS
904
904
904
904
447
447
SE„
496
496
496
496
115
115
se;
288
290
334
358
63
92
^'HH
498
501
566
526
511
486
SE
166
167
192
197
128
141
i'HT
471
472
567
527
511
486
SE
167
167
192
197
128
141
The second phase of the experiment began with random
tows in an east to west direction. Complete location and
CPUE data for both species are located in Appendix I. In
order to analyze all alternatives, the lowest alternative
was used in the field for adaptive sampling during the
second phase, which resulted in the 220 kg/km criterion
value for POP from alternative 2. For SR-RE, the criterion
value was the mean CPUE of 418 kg/km from alternative
3. The remaining alternatives were simulated following the
completion of the survey.
POP results
After the initial tows, 25 random tows were selected for the
return trip across the area. All 25 were completed, of which
six became networks of more than one unit. A total of 106
tows were completed in the POP stratum. At one of the
tows that exceeded the criterion value, the captain deemed
that further adaptive sampling was not feasible because of
the presence of coral. Of the six networks, two overlapped,
resulting in five distinct networks. In these networks, 81
adaptive samples were taken, of which 49 exceeded the
criterion and 32 did not and were therefore edge units and
not included in the sample estimates.
We compared the results of the original adaptive sample
( alternative 2 ) with the simulated results of higher criterion
values (Table 2). The precision of simple random sample
estimates with both n (number of random samples) and v'
(number of random samples plus the number of adaptive
network samples, not edge units) was contra.sted with that
of the adaptive estimators described above. As the criterion
value increased, n remained the same, whereas v' and r (the
number of networks) decreased. At the 220 kg/km criterion
value (alt. 2), there were substantial reductions in SE over
the SRS estimators by using ACS estimators for both the ;;
and v' sample sizes. The 250 kg/km criterion value (alt. 3)
resulted in a nearly identical sample to that of the 220 kg/
km (alt. 2) criterion value and the loss of only one network
sample. Hence, the estimates were nearly identical. The HT
mean estimates were slightly lower than the HH estimates
for the two lowest criterion values (alts. 2 and 3) because
two networks overlapped. These networks became separate
at the next higher criterion value, which aligned the estima-
tors. The next highest criterion value of 540 kg/km (alt. 1)
showed that even though the sample size was reduced by 19
tows from the original criterion value, the ACS estimators
performed nearly as well, yielding just slightly larger SEs.
When the criterion was arbitrarily doubled to 1080 kg/km,
the sample size was further reduced by seven, and had
similar SEs to the 540 kg/km criterion value.
The SRS and ACS bootstraps for POP resulted in very
different distributions. Five thousand replications showed
that the SRS distribution was bimodal and right skewed
(Fig. 3). The SRS mean fell on the second mode, which is
more than twice the ACS mean. This bimodal distribution
is driven by the presence of the very large random catch
(tow no. 60). If that haul is present in a bootstrap repli-
cate, then the SRS estimate tends to be high, leading to the
second mode in the bootstrap distribution. The ACS boot-
strap distribution was symmetric and closely resembled a
normal distribution (Fig. 3). The average estimates of bias
showed that the bias of HH was +4^( and the bias of HT
was -I7c. The standard error had an estimated bias of -t-S'/c
for HH and HT
The results from this POP study and the previous 1998
study were both greatly affected by one or two very large
catches, as we expected for a highly clustered population.
Of interest is what happened when the largest catch was
changed to a nominal catch that still exceeded the criterion
value. Appendix II shows the results of changing haul no.
60 from 12,000 kg/km to 540 kg/km. In the comparison at
v', SRS outperforms ACS in terms of SE. However, it also
Hanselman et al : Applications in adaptive cluster sampling of Gulf of Alaska rockfish
507
Table 3
Comparisons of time per travel (TPT) and time per sample (TPS) of adaptive sampling against simple random sampling for Pacific
ocean perch (S. alutus) and for shortraker tSehastes borealis) and rougheye (S. aleutianus) rockfish combined, on a 1999 adaptive
sampling cruise. TPT is the travel time between tows in hours; TPS is the travel time plus haul time in hours. "Distance between"
is the average travel distance (km) between two adaptive stations and between two random stations. "Adjusted distance" is the
distance if the random sample size was increased to 106.
S. alutus
S. borealis and S. aleutianus
Random Adaptive
Random Adaptive
Time(h) 10.4 11.4
4.4 12.0
No. of hauls 23 72
9 24
TPT 0.45 0.16
0.49 0.50
TPS 0.95 0.66
1.49 1.50
Distance between 20.2 3.22
Adjusted distance 4.73 3.22
shows that the mean of ACS is stable because it
changes little by removing a high catch, whereas
the SRS mean is reduced by half.
SR-RE results
At every third POP random tow, a tow was made in
the SR-RE depth stratum. A total of 35 tows were
made in the SR-RE stratum. Nine random tows
jrielded five distinct networks with 21 network tows
and five edge units. The stopping rule was invoked
for three of the five networks.
At the mean CPUE criterion (418 kg/km, alt. 3),
the adaptive estimators performed approximately
the same in terms of SE compared to the SRS esti-
mator using 71 (Table 2). With v', the SRS estimator
yielded a lower SE than both adaptive estimators.
When the criterion value increased to an arbitrarily
higher value (540 kg/km), the adaptive estimators
performed worse than SRS estimates for both /;
and v'.
Time efficiency
We recorded and compared travel time between
adaptive tows and simple random tows for 149
of the tows (Table 31. Not all the tows were used
because of mechanical failure or because the factory
capacity was reached. In the survey, 38 hours out
of 10 days were spent in transit between sampling tows,
which for a short survey was a substantial amount of the
available time. For POP, substantial gains in travel-time
efficiency were achieved with ACS. Average travel time
for simple random tows (0.45 h) was nearly triple that of
adaptive tows (0.16 h) for POP, which indicated that ACS
can maximize sampling tows for POP when time is limited.
In the SR-RE sampling, travel time for adaptive sampling
(0.5 h) was about the same as simple random sampling
(0.49 h), which was due to long linear samples that are not
as close together as POP tows (Fig. 1). Also, determina-
500 1000 1500 2000
Mean abundance
2500
3000
Figure 3
Bootstrap distributions for the 1999 adaptive sampling survey
(25,000 replicates). Dotted line is the sampling estimate of mean
abundance (kg/km) from the survey. Top graph is the distribution
of mean abundance estimates for simple random sampling. Bottom
graph is the distribution of mean abundance estimates for adaptive
cluster sampling (obtained with the Hansen-Hurwitz estimator).
tion of CPUE required processing of the catch, which took
various amounts of time after the completion of the tow.
Because of this delay, we went to the opposite tow on the
other side of the random tow when sampling SR-RE with
the linear pattern, whereas there were many nearby tows
when sampling POP with the cross pattern.
The travel time was added to the average tow time irom
gear deployment to full retrieval of 0.5 h for POP and 1.0 h
for SR-RE to obtain total sampling time (per sample).
Travel time was reduced by 31% with adaptive sampling
(0.66 h/sample) in relation to simple random sampling
508
Fishery Bulletin 101(3)
(0.95 h/sample) for POP. Sampling time efficiency for SR-
RE was approximately the same for adaptive sampling (1.5
h/sample) and simple random sampling (1.49 h/sample) for
SR-RE. These results are confounded by the fact that the
random tows are spread apart because of the lesser effort
applied to them. The average distance between random
tows (20.2 km) was adjusted to a distance of 4.73 km as
if there were 106 random tows distributed throughout the
area. This distance is still larger than the average distance
between tows in adaptive sampling (3.22 km).
From these time and distance data, we re-estimated the
precision of SRS under three new sample sizes in order to
further compare the relative efficiency of ACS. We denoted
the sample size that could have been taken under SRS, using
the same amount of time as was used during the adaptive
sampling including edge units, as v.. An alternative sample
size V, was the equivalent SRS sample size if the amount of
time to sample edge units in ACS was negligible. This sta-
tistic would be useful if edge units could be determined (i.e.
hydroacoustically or visually [presence or absence] ) without
actually trawling them. A third alternative was to find the
equivalent SRS sample size v^^ that would result from apply-
ing the total distance traveled in the ACS design on random
stations instead. For \'^., more random POP samples would
have been taken than were included in the adaptive estima-
tors (Table 4). The SEs of ACS were still much lower across
all criterion values (Table 2). When we used v, (Table 4), SRS
was much less precise than ACS (Table 2). Finally, when we
used distance instead of time (\'^), the results were almost
exactly the same as those for \\, (Table 4).
Discussion
Our two hypotheses were that ACS would be more precise
than SRS for POP and no more precise for SR-RE com-
bined. The results from the 1999 field study showed that
the SEs for the adaptive POP estimates were smaller than
both SRS estimates, with n and \'', and thus support the
first hjrpothesis. One curious result is that in both 1998
and 1999, the SRS estimate of density was substantially
larger than the ACS estimate, even though, on average,
they were both essentially unbiased. We attributed this
curiosity to the more variable and skewed SRS distribution
in which large sampling error on the high side is possible
more often than in the ACS estimation. Of course we fully
expected that both estimates would average to be the same
value if the experiment could be repeated many times. ACS
reduced the influence of one large CPUE in the relatively
small initial sample, as illustrated by the symmetric and
near-normal shape of the ACS bootstrap distribution. Con-
sequently, we concluded that ACS is a more robust estima-
tor of density than SRS for aggregated populations. One
caveat is that the precision of the estimates, if measured
in terms of coefficient of variation, is similar between the
two methods because of the nuich larger moan estimate for
the SRS estimate. Monte Carlo simulations would be useful
to examine the properties of the estimators under different
criterion values and population densities along the lines of
Su and Quinn (2003).
Table 4
Comparison of simple
random sampling (SRS)
precision
estimates with the inclusion of time and distance informa-
tion, c is the criterion value, v' is the original adaptive clus-
ter sampling adjusted
sample size, v,.
is the time
adjusted
sample size, including
edge units, v,
is the time
adjusted
sample size with edge
unit cost set to zero, v^^ is the dis- |
tance-adjusted sample
size including
edge units
ii is the
mean SRS density estimate, SE is the standard
error for
that sample size.
c (kg/km)
>220
>250
>540
>1080
A' 904
904
904
904
v' 74
73
55
48
SE 294
296
341
365
y, 81
80
67
55
SE 281
283
309
341
^', 59
58
46
41
SE 329
332
373
395
\;, 80
79
67
54
SE 283
285
309
344
The SR-RE adaptive estimates all have higher SEs than
the SRS estimates, and this finding supports the second hy-
pothesis. More than twice as many samples were directed
toward POP than SR-RE, yet the POP density estimates
are much more variable than those for SR-RE. This much
larger variability for POP was indicative of the clustering
that we expected.
This experiment showed that for POP, ACS with a fixed
criterion has some distinct advantages over simple random
sampling and over adaptive cluster sampling with order
statistics, which was used in the previous 1998 survey.
Lower SEs were obtained, at one third less effort than if
we just added an equivalent number of random samples.
Sampling over a broader area yielded better results than
the tightly stratified 1998 design. Our study also assumed
stationary aggregations of fish. This assumption may
have been better satisfied with a fixed criterion because
the adaptive sampling was conducted immediately after a
sample exceeded the criterion value.
Although the fixed criterion eliminates bias induced by
a variable criterion value, we still used stopping rules. If
bootstrapping is a good indicator of bias, then the bias in-
duced by stopping rules is negligible. Additionally, we have
shown that a relatively high criterion value could be used
to help minimize the use of these stopping rules.
Our study showed that ACS is a fast and efficient way
to gain a large number of samples. However, if edge units
do not contribute to a better estimate and they have a sim-
ilar cost or time expense as included samples, then little
is gained. This deficiency shows the need for some method
of determining edge units without actually sampling them.
In fisheries surveys, this use might be a double sampling
design with hydroacoustics as an auxiliary variable
Hanselman et al : Applications in adaptive cluster sampling of Gulf of Alaska rockfisfi
509
(Fujioka^) or a design called TAPAS that hydroacoustically
delineates clusters ( Everson et al., 1996). In other surveys,
it might be possible to detect the presence of the item of
interest without actually surveying the unit (as in aerial
surveys. )
An ACS design should not be attempted without some
prior knowledge of the population distribution. Populations
for which the design would be useful should have an aggre-
gated distribution that can be described by correlated varia-
tion with distance, not just a large variance in relation to the
mean. One way to examine the data is to fit variograms to
examine spatial autocorrelation (Hanselman et al., 2001). If
no prior data exist, it would not make sense to attempt ACS
as an initial sampling design. We have shown that a wide
range of criterion values can be used without considerable
differences in the results. Therefore, only enough prior data
are needed so that an adequate range of population density
can be estimated. If the criterion value chosen resulted in too
many or too few samples, the criterion could be adjusted, and
then the design stratified into two different areas.
Most commercial fish species have survey data that can be
used to determine a fixed criterion. If possible, criterion val-
ues should be determined prior to the survey, so that maxi-
mum efficiency can be attained. We have shown that it may
be appropriate to choose a relatively high sampling criterion
such as the 80'^ percentile of past CPUE without sacrificing
estimation capabilities. This high sampling criterion has sev-
eral practical advantages. First, the design is attractive for
commercial boats to perform the adaptive phase at no-cost
because only large catches are sampled. The current design
does not use the fish sampled during the survey, which, in
the case of deepwater rockfish, would cause certain mortality.
Under an adaptive design, a commercial boat would take the
larger catches and could put them to use. Second, fewer over-
all networks would be sampled because the higher criterion
would evoke less adaptive sampling, which may mean less
overall sajnpling in the survey. Finally, precision would be
gained at a minimal cost and effort. Stopping rules would be
unnecessary, ensuring an unbiased estimate. However, clus-
ter sampling is most effective when the cluster samples are
as heterogeneous as possible. Therefore, caution is required
not to set the criterion too high, or the resulting clusters
will be either too homogeneous or contain only edge units,
leading to no improvement in the estimators. Similarly, if
there are large changes in density from year to year, a fixed
criterion may not be appropriate. In conclusion, adaptive
cluster sampling is appropriate for surveys of highly clus-
tered species with low temporal fluctuations, for which a
fixed criterion can be determined beforehand.
Acknowledgments
We thank the crew of the FV Unimak, in particular Cap-
tain Paul Ison and Production Manager Rob Elzig, for their
3 Fujioka, J. 2001. Unpubl. manuscr. Using hydroacoustics
and double sampling to improve rockfish abundance estimation,
8 p. Auke Bay Laboratory, National Marine Fisheries Service,
NOAA, 11305 Glacier Hwy, Auke Bay, AK 99801.
excellent cooperation in this study. We also acknowledge the
hard work of the scientists that participated in the cruise
and the NMFS personnel who prepared for the charter.
We greatly appreciate the helpful comments from three
anonymous reviewers that helped us refine the paper.
This publication is the result of research sponsored by
Alaska Sea Grant with funds from the National Oceanic
and Atmospheric Administration, Office of Sea Grant, De-
partment of Commerce, under grant no. NA90AA-D-
SG066, project number R/31-04N, from the University
of Alaska with funds appropriated by the state. Further
support was provided by the Auke Bay Laboratory, Alaska
Fisheries Science Center, National Marine Fisheries Ser-
vice and by a Population Dynamics Fellowship to Hansel-
man through a cooperative program funded by Sea Grant
and NMFS.
Literature cited
Brown, J. A., and B. J. F. Manly.
1998. Restricted adaptive cluster sampling. Environ. Ecol.
Stat. 5:49-63.
Christman, M. C.
1997. Efficiency of some sampling designs for spatially clus-
tered populations. Environmetrics 8:145-166.
Christman, M. C, and J. S. Pontius.
2000. Bootstrap confidence intervals for adaptive cluster
sampling. Biometrics 56:503-510.
Everson, I., M. Bravington, and C. Goss.
1996. A combined acoustic and trawl survey for efficiently
estimating fish abundance. Fish. Res. 26:75-91.
Hanselman, D. H.
2000. Adaptive sampling of Gulf of Alaska rockfish. M.S.
thesis, 72 p. Univ. Alaska, Fairbanks, AK.
Hanselman, D. H., T. J. Quinn, C. Lunsford, J. Heifetz, and
D. M. Clausen.
2001. Spatial inferences of adaptive cluster sampling on
Gulf of Alaska rockfish. In Proceedings of the 17th Lowell-
Wakefield symposium: spatial processes and management of
marine populations, p. 303-325. Univ. Alaska Sea Grant
Program, Fairbanks, AK.
Lo, N., D. Griffith, and J. R. Hunter
1997. Using a restricted adaptive cluster sampling to esti-
mate Pacific hake larval abundance. Calif Coop. Oceanic
Fish. Invest. Rep. 38:103-113.
Lunsford, C.
1999. Distribution patterns and reproductive aspects of Pa-
cific ocean perch {Sebastes alutus) in the Gulf of Alaska. M.S.
thesis, 154 p. Univ. of Alaska Fairbanks, Fairbanks, AK
Quinn II, T. J., D. H. Hanselman, D. M. Clausen, J. Heifetz, and
C. Lunsford.
1999. Adaptive cluster sampling of rockfish populations.
Proceedings of the American Statistical Association 1999
Joint Statistical Meetings, Biometrics Section, 11-20. Am.
Statist. Assoc, Baltimore, MD.
Salehi, M. M.
1999. Rao-Blackwell versions of the Horvitz-Thompson and
Hansen-Hurwitz in adaptive cluster sampling. Environ.
Ecol. Stat. 6:83-195.
Su, Z., and Quinn, T. J., II.
2003. Estimator bias and efficiency for adaptive cluster sam-
pling with order statistics and a stopping rule. Environ.
Ecol. Stat. 10, pp. 17-41.
510
Fishery Bulletin 101 (3)
Thompson, S. K.
1990. Adaptive cluster sampling.
1050-1059.
J. Am. Stat. Assoc. 412;
Thompson, S. K., and G. A. F. Seber.
1996. Adaptive sampling, 265 p.
Wiley, New York, NY.
Appendix 1
CPUE (kg/km) data from the 1999 adaptive cluster sampling survey. CPUE
is giver
in kg/km. The format of
"Adaptive 26-1"
corresponds to the first adaptive tow around haul no. 26. POP = Pacific ocean
perch; SR-RE = shortraker and rougheye rockfish |
combined.
Summary table
Tow type
Initial random 2""'
phase random Adaptive network
Adaptive edge unit
Total 1
POP
13
25
49
32
106(119)
SR-RE
10
9
21
5
35(45)
Total
23
34
70
37
141 (164)
' Values in
parenthesis include
initial random tows that are not included in estimation results.
Criterion determining random tows
Tow
Latitude
Longitude
Tow type
POP CPUE
SR-RE CPUE
3
59.59
-143.81
POP random
39.3
43.7
4
59.54
-143.55
POP random
49.2
13.7
5
59.51
-143.55
SR-RE random
3.4
870.9
6
59.58
-143.28
POP random
174.8
112.0
7
59.56
-143.28
SR-RE random
17.7
582.3
8
59.67
-143.01
POP random
72.7
21.0
9
59.69
-142.75
POP random
21.3
6.1
10
59.64
-142.75
SR-RE random
6.3
6.3
11
59.60
-142.49
POP random
9.6
36.2
12
59.59
-142.48
SR-RE random
3.8
608.0
13
59.40
-142.22
POP random
20.7
113.0
14
59.28
-141.96
POP random
25.3
394.4
15
59.27
-141.96
SR-RE random
19.1
713.1
16
59.17
-141.68
POP random
185.4
68.5
17
59.16
-141.68
SR-RE random
24.9
48.5
18
59.04
-141.41
SR-RE random
1.7
450.4
19
59.03
-141.41
POP random
196.5
21.9
20
59.01
-141.14
SR-RE random
30.0
676.9
21
58.78
-140.88
POP random
2271.6
0.0
22
58.75
-140.88
SR-RE random
65.9
80.6
23
58.67
-140.61
POP random
80.6
101.1
24
58.66
-140.35
POP random
98.2
55.0
25
58.66
-140.35
SR-RE random
21.2
140.5
Beginning of adaptive random tows
26
58.70
-140.64
POP random
576.7
0.0
27
58.68
-140.65
SR-RE random
16.3
115.8
28
58.73
-140.71
POP adaptive 26-1
138.1
12.0
29
58.72
-140.65
POP adaptive 26-2
138.4
9.7
30
58.69
-140.62
POP adaptive 26-3
2294.2
0.0
31
58.70
-140.64
POP adaptive 26-4
290.1
0.4
32
58.70
-140.63
POP adaptive 26-8
334.8
0.0
33
58.69
-140.62
POP adaptive 26-9
56.5
21.2
34
58.69
-140.63
POP adaptive 26-10
16.4
1.9
35
58.71
-140.67
POP adaptive 26-11
20.7
3.7
36
58.72
-140.67
POP adaptive 26-12
30.2
1.0
continued
Hanselman et al.: Applications in adaptive cluster sampling of Gulf of Alaska rockfish
511
Appendix 1 (continued)
Criterion determining random tows
Tow
Latitude
Longitude
Tow type
POP CPUE
SR-RE CPUE
37
58.69
-140.61
POP adaptive 26-18
1299.4
1.2
38
58.69
-140.61
POP adaptive 26-17
965.0
55.9
39
58.70
-140.75
POP random
62.0
148.0
40
58.76
-140.85
POP Random
3591.0
58.4
41
58.79
-140.89
POP adaptive 40-1
5934.1
0.0
42
58.77
-140.86
POP adaptive 40-2
4521.0
0.0
43
58.74
-140.83
POP adaptive 40-3
515.7
9.1
44
58.76
-140.86
POP adaptive 40-4
4453.7
37.3
45
58.79
-140.90
POP adaptive 40-5
1338.8
0.0
46
58.79
-140.88
POP adaptive 40-6
393.9
0.0
47
58.77
-140.86
POP adaptive 40-7
109.4
0.0
48
58.75
-140.82
POP adaptive 40-8
85.0
0.0
49
58.73
-140.80
POP adaptive 40-9
67.9
0.1
50
58.74
-140.83
POP adaptive 40-10
128.0
17.6
51
58.76
-140.86
POP adaptive 40-11
1597.3
0.0
52
58.78
-140.89
POP adaptive 40-12
268.5
3.8
53
58.80
-140.90
POP adaptive 40-24
1282.9
0.0
54
58.81
-140.92
POP adaptive 40-13
2304.4
0.0
55
58.80
-140.90
POP adaptive 40-14
776.2
0.0
56
58.79
-140.88
POP adaptive 40-15
882.6
0.0
57
58.75
-140.86
POP adaptive 40-22
168.1
2.7
58
58.78
-140.89
POP Adaptive 40-23
253.9
0.2
59
58.83
-140.95
SR-RE random
24.1
290.2
60
58.88
-140.95
POP random
12001.5
0.0
61
58.87
-140.96
POP adaptive 60-4
10659.3
0.0
62
58.91
-140.97
POP adaptive 60-1
1179.0
0.0
63
58.89
-140.95
POP adaptive 60-2
3050.4
0.0
64
58.86
-140.95
POP adaptive 60-3
2984.7
0.0
65
58.86
-140.95
POP adaptive 60-10
3590.4
0.0
66
58.88
-140.96
POP adaptive 60-11
1086.9
0.0
67
58.91
-140.98
POP adaptive 60-12
1311.7
8.7
68
58.92
-140.98
POP adaptive 60-5
1581.0
0.0
69
58.91
-140.96
POP adaptive 60-6
4148.4
0.0
70
58.89
-140.95
POP adaptive 60-7
1297.4
0.0
71
58.86
-140.94
POP adaptive 60-8
214.1
0.0
72
58.84
-140.94
POP adaptive 60-9
2190.3
0.0
73
58.84
-140.94
POP adaptive 60-20
1502.2
0.0
74
58.83
-140.93
POP adaptive 60-19
2828.9
0.0
75
58.84
-140.93
POP adaptive 60-18
102.9
0.0
76
58.86
-140.94
POP adaptive 60-17
46.6
0.0
77
58.89
-140.95
POP adaptive 60-16
27.8
0.0
78
58.89
-140.95
POP adaptive 60-15
53.4
0.0
79
58.92
-140.97
POP adaptive 60-14
495.7
0.0
80
58.93
-140.98
POP adaptive 60-13
1323.4
0.0
81
59.05
-141.05
POP random
1448.8
0.4
82
Coral encountered
N/A
N/A
83
59.03
-141.08
POP random
560.6
102.8
84
59.03
-141.19
POP random
283.6
298.5
85
59.04
-141.19
POP adaptive 83-1
1119.7
101.3
86
59.04
-141.26
POP adaptive 83-2
1407.0
21.7
87
59.02
-141.22
POP adaptive 83-3
398.1
29.2
continued
512
Fishery Bulletin 101 (3)
Appendix 1 (continued)
Criterion determining random tows
Tow
Latitude
Longitude
Tow type
POP CPUE
SR-RE CPUE
88
59.03
-141.16
POP adaptive 83-4
264.6
87.0
89
59.05
-141.20
POP adaptive 83-5
416.6
47.3
90
59.04
-141.29
POP adaptive 83-6
2186.1
7.0
91
59.04
-141.25
POP adaptive 83-7
482.0
8.7
92
59.03
-141.22
POP adaptive 83-8
115.2
36.6
93
59.02
-141.19
POP adaptive 83-9
182.5
36.4
94
59.02
-141.13
POP adaptive 83-10
41.4
45.5
95
59.02
-141.16
POP adaptive 83-11
29.2
41.1
96
59.04
-141.20
POP adaptive 83-12
261.4
80.6
97
59.04
-141.25
POP adaptive 83-24
109.3
32.0
98
59.04
-141.29
POP adaptive 83-23
62.0
69.4
99
59.05
-141.26
POP adaptive 83-13
186.4
56.2
100
59.05
-141.32
POP adaptive 83-14
443.8
4.5
101
59.04
-141.29
POP adaptive 83-15
1497.1
5.4
102
59.04
-141.25
POP adaptive 83-16
892.0
21.4
103
59.03
-141.22
POP adaptive 83-17
604.8
26.1
104
59.03
-141.16
POP adaptive 84-3
123.5
91.4
105
59.03
-141.22
POP adaptive 84-4
129.3
285.3
106
59.04
-141.26
POP adaptive 84-1
231.2
602.5
107
59.02
-141.32
SR-RE random
49.3
721.9
108
59.05
-141.26
POP adaptive 84-5
214.6
1408.9
109
59.04
-141.35
POP adaptive 84-6
215.0
123.6
110
59.04
-141.31
POP adaptive 84-12
61.5
664.5
111
59.04
-141.32
SR-RE adaptive 107-1
57.5
758.1
112
59.02
-141.37
SR-RE adaptive 107-2
0.0
490.7
113
59.05
-141.20
SR-RE adaptive 107-3
0.0
408.6
114
59.01
-141.42
SR-RE adaptive 107-4
0.0
669.1
115
59.00
-141.14
SR-RE adaptive 107-6
0.0
760.8
116
58.97
-141.09
SR-RE adaptive 107-8
0.0
1540.6
117
58.11
-141.06
SR-RE random
0.0
443.2
118
59.14
-141.60
SR-RE adaptive 117-1
0.0
1052.8
119
59.09
-141.64
SR-RE adaptive 117-2
0.0
1042.0
120
59.16
-141.50
SR-RE adaptive 117-3
51.3
621.6
121
59.07
-141.69
SR-RE adaptive 117-4
25.7
2096.7
122
59.05
-141.46
SR-RE adaptive 117-6
68.4
480.5
123
59.19
-141.40
SR-RE adaptive 117-5
41.2
924.3
124
59.21
-141.73
SR-RE adaptive 117-7
189.0
731.9
125
59.04
-141.78
SR-RE adaptive 117-8
82.3
772.2
126
59.14
-141,34
POP random
61.9
4.8
127
59.15
-141.60
POP random
82.6
55.8
128
59.21
-141.65
POP random
68.5
8.1
129
59.29
-141.75
POP random
84.6
0.0
130
59.23
-141.85
SR-RE random
6.1
1024.1
131
59.27
-141.85
SR-RE adaptive 130-1
2.6
626.9
132
59.21
-141.94
SR-RE adaptive 130-2
1.5
451.9
133
59.27
-141.81
SR-RE adaptive 130-3
4.2
2208.3
134
59.28
-142.00
SR-RE adaptive 130-5
7.4
1605.6
135
59.31
-142.06
SR-RE adaptive 130-7
5.0
1305.2
136
59.19
-142.11
SR-RE adaptive 130-4
0.0
432.4
137
59.17
-141.75
SR-RE adaptive 130-6
1.6
457.4
138
59.39
-141.70
POP random
181.8
25.9
139
59.36
-142.05
POP random
62.9
12.2
continued
Hanselman et al.: Applications in adaptive cluster sampling of Gulf of Alaska rockfish
513
Appendix 1 (continued)
Criterion determining random tows
Tow
Latitude
Longitude
Tow type
POP CPUE
SR-RE CPUE
140
59.40
-142.15
SR-RE random
3.7
772.3
141
59.45
-142.25
SRRE adaptive 140-1
1.1
222.7
142
59.38
-142.31
SRRE adaptive 140-2
0.0
209.0
143
59.42
-142.22
POP random
177.2
36.0
144
59.67
-142.25
POP random
45.4
33.5
145
59.60
-142.35
POP random
8.3
117.8
146
59.71
-142.45
POP random
4.3
32.0
147
59.67
-142.65
SR-RE random
2.0
47.0
148
59.64
-142.65
POP random
18.0
50.8
149
59.67
-142.95
POP random
34.2
3.4
150
59.61
-142.85
POP random
125.0
18.8
151
59.57
-143.05
SR-RE random
3.6
530.5
152
59.59
-143.05
POP random
139.0
39.7
153
59.56
-143.15
SR-RE adaptive 151-1
5.1
555.2
154
59.59
-143.16
SR-RE adaptive 151-2
2.6
255.5
155
59.55
-143.00
SR-RE adaptive 151-3
0.0
314.5
156
59.56
-143.22
POP random
23.5
567.4
157
59.57
-143.25
POP random
43.3
399.3
158
59.54
-143.35
SR-RE random
9.3
82.2
159
59.58
-143.36
POP random
74.9
493.0
160
59.55
-143.45
POP random
2838.5
1.8
161
59.57
-143.65
POP adaptive 160-1
1674.5
54.5
162
59.53
-143.69
POP adaptive 160-2
2912.8
1.8
163
59.55
-143.63
POP adaptive 160-3
196.5
0.0
164
59.52
-143.65
POP adaptive 160-4
148.2
0.5
165
59.52
-143.60
POP adaptive 160-5
75.6
21.0
166
59.58
-143.63
POP adaptive 160-6
863.1
9.4
167
59.56
-143.69
POP adaptive 160-7
41.3
0.0
Appendix il
Results of estimation with haul no. 60 changed from 12000 kg/km to 540 kg/km. c is the criterion value (kg/km), /i is the mesm
Pacific ocean perch density (kg/km) for each estimator, n is the random sample size, v' is the adaptive sample size without edge
units. SE is the standard error of the mean.
c
(kg/km
)
c (kg/km)
>220
>250
>540
>1080
>220
>250
>540
>1080
^'srs^">
445
445
445
445
SE
148
149
175
158
SE
179
179
179
179
^'HT
442
443
536
413
SE(V)
104
104
104
104
SE
149
149
175
158
i'HH
470
473
535
412
514
Abstract — We investigated the migra-
tion and behavior of young Pacific
bluefin tuna (Thunnus orientalis) us-
ing archival tags that measure envi-
ronmental variables, record them in
memor>-, and estimate daily geographi-
cal locations using measured light
levels. Swimming depth, ambient water
temperature, and feeding are described
in a companion paper. Errors of the tag
location estimates that could be checked
were -0.54° ±0.75° (mean ±SD) in lon-
gitude and -0.12° ±3.06° in latitude.
Latitude, estimated automatically by
the tag, was problematic, but latitude,
estimated by comparing recorded sea-
surface temperatures with a map of
sea-surface temperature, was satisfac-
tory. We concluded that the archival tag
is a reliable tool for estimating location
on a scale of about one degree, which
is sufficient for a bluefin tuna migra-
tion study. After release, tagged fish
showed a normal swimming behavioral
pattern within one day and normal
feeding frequency within one month.
In addition, fish with an archival tag
maintained weight-at-length similar to
that of wild fish; however, their growth
rate was less than that of wild fish. Of
166 fish released in the East China Sea
with implanted archival tags, 30 were
recovered, including one that migrated
across the Pacific Ocean. Migration
of young Pacific bluefin tuna appears
to consist of two phases: a residency
phase comprising more than 80% of all
days, and a traveling phase. An indi-
vidual young Pacific bluefin tuna was
observed to cover 7600 km in one trav-
eling phase that lasted more than two
months (part of this phase was a trans-
Pacific migration completed within two
months). Many features of behavior
in the traveling phase were similar to
those in the residency phase; however
the temperature difference between
viscera and ambient temperature was
larger, feeding was slightly more fre-
quent, and dives to deeper water were
more frequent.
Migration patterns of young Pacific bluefin tuna
(Thunnus orientalis) determined with archival tags
Tomoyuki Itoh
Sachiko Tsuji
National Research Institute of Far Seas Fisheries
5-7-1 Shimizu-Orido, Shizuoka
Shizuoka, 424-8633, Japan
E-mail address (for T Itoh): itou@fra.affrcgo.|p
Akira Nitta
Japan NUS Co., Ltd.
Loop-X BIdg., 3-9-15 Kaigan, Minato
Tokyo, 108-0022, Japan
Manuscript approved for publication
22 October 2002 by Scientific Editor.
Manuscript received 3 January 2003
at NMFS Scientific Publications Office.
Fish. Bull. 101:514-534 (2003).
Pacific bluefin tuna (Thunnus orien-
talis), a highly migratory species, is
mainly distributed in the temperate
zone of the northern Pacific Ocean
(Yamanaka, 1982; Bayliff, 1994) in con-
trast to T. thynnus, which inhabits the
Atlantic Ocean (Collette, 1999).
Current knowledge on the migration
of Pacific bluefin tuna is summarized
in the following studies: Aikawa ( 1949);
Bell ( 1963a ); Okachi ( 1963 ); Orange and
Fink ( 1963 ); Nakamura ( 1965 ); Clemens
and Flittner ( 1969); Shingu et al. ( 1974);
Yorita( 1976); Bayliff( 1980); Yamanaka
(1982); Yonemori (1989); and Bayliff et
al. (1991). The majority of bluefin tuna
spawn in the northwest Pacific Ocean
in an area from the Philippines past
Taiwan to Okinawa from April to June,
and small numbers spawn off southern
Honshu in the Pacific Ocean in July and
in the Sea of Japan in August (Yabe et
al., 1966; Ueyanagi, 1969; Okiyama,
1974; Yonemori, 1989; Kitagawa et al.,
1995). Carried by the Kuroshio Cur-
rent, juveniles arrive near the coast of
Japan, move northward during sum-
mer and early autumn, and then most
turn around and move back southward
during late autumn and winter along
the Japanese coast. During the first
few years of their lives, the majority of
young fish repeat a similar north-.south
seasonal migration. However a small
fraction, increasing each year, moves
away from the Japanese coast and often
reaches the eastern side of the Pacific
Ocean, off the United States and Mexico.
These fish stay in the eastern Pacific
Ocean for 1-3 years. Some time later,
as mature fish, they gather in the north-
west Pacific Ocean to spawn and then
disperse after the spawning season.
This information on Pacific bluefin
tuna movements has been accumu-
lated through analyses of fishery catch
data and tag-recapture data. Fishery
catch data are based on different indi-
viduals from limited areas where, and
particular seasons when, fishing took
place. Conventional tagging data pro-
vide migration information regarding
only two points: release and recapture.
Acoustic tracking, another method used
for investigating behavior and migra-
tion of individuals, can collect detailed
information on fish movements and
behavior on a time scale of seconds, but
the duration of the tracking period of
each individual has usually been less
than several days in studies of Pacific
bluefin tuna (Marcinek et al., 2001;
Hisada et al.') as well as in other stud-
ies of Thunnus species (e.g. Carey and
Olson, 1982; Holland et al., 1990; Cayre,
1991; Cayre and Marsac, 1993; Block
1 Hisada, K., H. Kono, and T Nagai. 1984.
Behavior of young bluefin tuna during
migration. In Progress report of the
marine ranching project 4, p. 1-7. Nat.
Res. Inst. Far Seas Fish. Pelagic Fish
Resource Division, 5-7-1 Shimizu-Orido,
Shizuoka, Shizuoka, 424-8633, Japan. [In
Japanese, the title was translated by
authors.!
Itoh et al,: Migration patterns of Thunnus orienta/is determined with archival tags
515
et al., 1997). These methods have not yielded detailed
information regarding migration, behavior, and their rela-
tion to environmental factors for Pacific bluefin tuna over
a long period.
An archival tag is an electronic device that measures
environmental variables and records data in its memory.
When attached to an animal, it allows direct examination
of the relationship between an animal's behavior and
physiological condition, or the ambient environment. One
type of "archival" tag merely stores data; however, another
type not only stores data but also provides daily geographi-
cal locations of the fish by processing the measured envi-
ronmental data. This type of archival tag was anticipated
since the 1980s as a tool that could collect detailed infor-
mation on individual fish behavior (Hunter et al., 1986;
Anonymous, 1994). Metcalfe and Arnold (1997) estimated
the geographical locations and tracks of plaice, a demersal
species, by comparing tidal depth variations with the time
series depth data recorded by archival tags attached to the
fish. However, this method is not suitable for pelagic fish,
which change swimming depth freely. A type of archival tag
that can estimate geographical locations based on change
of light levels during a day — a method more suitable for
pelagic species — has been commercially available since the
early 1990s. So far, archival tags of this type have been used
in several tagging projects (Arnold and Dewar, 2001). The
results published in a few reports on southern bluefin tuna
(T. maccoyii) (Gunn and Block, 2001), and Atlantic bluefin
tuna (T! thynnus) (Block et al., 2001), show the remarkable
value of archival tag data.
As archival tags have come into wide use, results of sev-
eral experiments conducted to evaluate the reliability of
its geolocation estimates have been published (Welch and
Eveson, 1999; Musyl et al., 2001; Gunn et al.^). However,
several points remain to be tested: tag reliability when a
number of tags are deployed for long duration, reliability
of sensors for variables other than light, and the effects of
attaching the tag to fish.
After two preliminary experiments with tags in 1994,
the first with tags placed at a known outdoor location on
land and in air, and the second with tags attached to young
Pacific bluefin tuna held in pens, we applied archival tags
to wild young Pacific bluefin tuna to investigate their be-
havior and migration. In the present study we report the
characteristics of migration for this species based on data
on daily geographical location, as well as the reliability
of archival tag data and the effect of attachment of the
tag to fish. Analyses for swimming depth, ambient water
temperature, and feeding frequency of the species are
undertaken in other papers (Kitagawa et al., 2000; Itoh
et al.,2003).
2 Gunn, J., T. Polacheck, T. Davis, M. Sherlock and A. Betlehem.
1994. The development and use of archival tags for studying
the migration, behavior and physiology of southern bluefin tuna
with an assessment of the potential for transfer of the technol-
ogy to groundfish research. Proc. ICES mini-symposium on
migration, St. Johns, Newfoundland. ICES CM. Mini:2.1, 23 p.
International Council for the Exploration of the Sea, Palaegade
2-4, DK-1261 Copenhagen K, Denmark.
Materials and methods
Outline of the archival tag used in this study
The archival tag used in this study (Northwest Marine
Technology, Inc. Shaw Island, WA ) had a cylindrical stain-
less-steel body ( 16 mm in diameter and 100 mm long, and
weighing 52 g) that was implanted in the animal. A flexible
sensor stalk 2.2 mm in diameter and 150 mm long extended
from the tag through the skin of the animal into the water.
The end of the stalk housed an external temperature sensor
and a light capture region. Light was led from the capture
region by optical fiber to a photodiode sensor in the body
of the tag, which also housed sensors for pressure, internal
temperature, and light. Response times for the tempera-
ture sensor were three seconds for the external sensor and
20 seconds for the internal sensor, and temperature resolu-
tion was 0.2°C for both sensors. Resolution of the pressure
sensor record was 1 m at shallow depths up to 126 m, then
changed to 3 m from that depth to the scale limit of 510 m.
Clock drift was less than 30 seconds per year. The tag had
a data measurement interval of 128 seconds, a 256-kByte
data memory, and an operating life exceeding seven years.
Data were downloaded from recovered tags by using a per-
sonal computer and a fiber-optic connector.
Two types of data files were created within the tag
memory. One data file stored daily records containing date,
estimated times of sunrise and sunset, water temperatures
at 0 m plus two other selectable depths (we selected 60 m
and 120 m), and other information required for, or produced
in, the course of location estimates for each day. This file
is referred as the "summary file" in the "Results" section,
and it stored data for all days after the memory was last
cleared. The times of sunrise and sunset were estimated
within the tag from sea-surface light intensities, which
were inferred from measured depth and measured light
intensity at depth and a water opacity factor determined
from the measured data each Universal Time (UT) day. The
time of midday was determined as the midpoint between
sunrise and sunset times, and longitude was calculated
from the difference between the midday time and 1200h
UT, at a rate of 15 degrees longitude per hour, corrected
for astronomical effects. Latitude was estimated from the
duration of daylight (Hill, 1994).
The second data file contained unprocessed time series
data records taken at 128-second intervals. The tag could
record at any integer multiple of its 128-second measure-
ment interval and a multiple of one was chosen. Each
record consisted of external temperature, internal tem-
perature, pressure, and light intensity, and corresponded
to a known time. This is referred to as the "detail file" in
the "Results" section. It could hold about 54,000 records,
or about 80 days of steady recording at the high data rate
chosen — a small fraction of the tag's overall lifetime. The
time-series memory was divided into two sections, and the
size allocations for the two sections were determined by the
user The first section filled first and did not change there-
after. The second section filled next, but once full, it was
continually overwritten by new data. Thus the first section
always contained the earliest data retrieved from a tag; the
516
Fishery Bulletin 101(3)
second always contained the latest data. We divided the file
into two 40-day sections for releases in 1995 and 1996, and
into 20- and 60-day sections for releases in 1997.
Reliability and calibration of archival tags in air
To examine the reliability of location estimates made
by archival tags, 117 archival tags were left outdoors
(34°59'N; 138°59'E) where they were not affected by artifi-
cial light during July-September 1996 (55 days, five tags),
May-August 1997 (86 days, 14 tags), and October 1997
(five days, 100 tags). Two of the tags were used in two of
the experiments.
Calibration tests of internal and external temperature
sensors were conducted for all tags before being implanted
in fish that were released, and the sensors were recalibrat-
ed for nine tags after they were recovered. Temperature
calibration was done by immersing tags into a series of
water tanks that were set to temperatures ranging from
5.0° to 30.0°C by 5°C intervals. Calibration tests of pres-
sure sensors were also conducted for all tags before release
and on 27 tags after being recovered. Tags were placed in
a pressure chamber with a resolution of 0.1 bar and exam-
ined up to 20 bar. The tags were left at least five minutes
at each temperature or pressure to obtain at least two
measurements at the 128-second recording interval.
Experiment with pen-held fish
Archival tags were attached to three pen-held young Pacific
bluefin tuna of 93-97 cm fork length (FL) at Kasasa in
Kagoshima Prefecture (31°25'N; 130°11'E) in November
1994. The fish had been reared in a net pen (40 m x 25 m
with 12 m depth) for more than two years and were accli-
mated to the environment at the time of the experiment.
Archival tags were inserted into the abdominal cavities
of two of the three fish by the following method. A fish
caught by hook and line was put into a styrofoam box, and
its eyes were covered with a black polyethylene bag. The
belly of the fish was cut with a scalpel about 4 cm anterior
to the anus, 3—4 mL of antibiotic (artificial penicillin, Doil,
Tanabe Seiyaku Co., Ltd., Osaka, Japan) was injected
into abdominal cavity of the fish, and an archival tag was
inserted there with the stalk extending through the inci-
sion. A stitch was made in the middle of the incision with
an absorbable suture (Coated Vicryl, type J583G, Ethicon
Inc., Cornelia, GA ), and the fish was released back into the
pen. All tools and tags were disinfected with 100% ethanol.
No anesthetic was used because with their eyes covered,
the fish remained quiet during the surgery. This simplified
procedure (from making the incision to releasing the fish)
could be completed in less than 90 seconds, thus minimiz-
ing total stress on the animal and, in later experiments
on wild fish, providing the best chance for the animal to
rejoin its original school. In this pen study, the third fish
was tagged externally instead of internally, the tag being
connected by a thin wire rope to a small metal arrowhead
inserted in a muscle near the second dorsal fin base.
During the pen-held fish experiment, none of the fish
were observed to die as a result of tagging. The tag that
had been attached externally came loose from the fish
and was retrieved from the bottom of the pen four days
after tagging. One tagged fish escaped when the pen was
broken. The remaining tag was recovered 453 days later
when the fish was caught from the pen as part of a com-
mercial catch.
Experiments with wild fish
Tag and release experiments on wild young Pacific bluefin
tuna were conducted near Tsushima, at the northeastern
end of the East China Sea, by using chartered commercial
trolling vessels, every November and December from 1995
to 1997. A total of 166 fish, ranging from 43 to 78 cm FL
(age 0 or 1), were internally tagged as described above and
released immediately. Two dart-type conventional tags
were also attached to the second dorsal-fin base of each fish
in the 1997 experiment as visual markers in an attempt to
improve the recovery rate.
Thirty archival tags (18.1%) were recovered. The dura-
tions at sea were 50 days or less for 13 fish, 96-211 days
for 13 fish, and 359-375 days for three fish, all recaptured
around Japan. One additional fish was recaptured off the
west coast of Mexico, on the east side of the Pacific Ocean,
at 610 days after release. Data could not be downloaded
from one archival tag released in 1995 and recovered 30
days after release; all other tags returned data.
Results
Reliability of location estimates
The tag recovered from a fish penned in a known location
for 453 days yielded a record of positions automatically
estimated during that time. Figure 1 plots the errors in
those estimates and the date when each was made. This
tag provided the only position sequence of long duration
obtained from a captive fish. Unfortunately it was discov-
ered later, after the experiment was completed and after
this particular tag was no longer available for further test-
ing, that the light sensitivity of this tag, as well as that of
the tag that yielded data for four days in the captive fish
experiment, was at least a factor of ten lower than that of
other tags. This discrepancy in light sensitivity could be
seen in the daily noon-light intensity data in the summary
file, both during the in-water experiment (when compared
with typical values for tags in wild fish) and when tested
in air (compared with other tags of the group tested in air).
On dark days there was an unusual pattern of early sunset
times and late sunrise times that the tag manufacturer
interpreted as being associated with the low light sensi-
tivity. Thus, although the general trends of error size with
season can be expected to be representative, the absolute
size of the errors was likely inflated in this, the only long-
term record obtained from a captive fish.
Longitude error showed no change with season, but
latitude error increased dramatically near the equinoxes
as expected because day length does not vary significantly
with latitude at that time, and therefore carries little in-
Itoh et al.: Migration patterns of Thunnus orientalis determined with archival tags
517
-1 1 n
10Nov94 18Feb95 29 May 95 6 Sep 95 15Dec95
_ 40 -,
10 Nov 94 18 Feb 95 29 May 95 6 Sep 95 15 Dec 95
Date
Figure 1
Seasonal change of errors in location estimate. Upper and lower panels
show longitudinal and latitudinal errors, respectively. Triangles show
equinoxes. Data are from an archival tag implanted in a fish held in a
pen for 453 days.
formation about latitude. The tag did not provide a latitude
estimate for 18 days around and at the vernal (autumn)
equinox and had large errors for one month before (or after)
as well as 10 days after (or before) that period, respectively.
The same pattern was observed in the test with archival
tags that were left in air. In addition, the latitude estimates
were biased toward south in summer and toward north in
winter, that is, toward erroneously short day lengths.
Occasional large deviations were observed in both lati-
tude and longitude estimates. These were easily identified
as outliers in our analyses of data obtained from wild fish
by comparing them with estimated locations for adjacent
days. When evaluating the accuracy of location estimates
for practical use in analyses of wild fish movements, we
excluded longitude or latitude estimates that differed more
than 10° from the real location and the latitude estimates
not provided by the tag near the equinoxes. These account-
ed for 2.8% of longitude estimates and 8.9% of latitude
estimates obtained in the tests in air, as well as 4.8% of
longitude data and 47.5% of latitude data obtained in the
tests of pen-held fish.
Figure 2 shows the position estimates and error bars
corresponding to one standard deviation for 1 17 tags tested
in air — most of them for a .5-day period, five for 82 days,
and twelve others for various intermediate durations. The
aggregate of all observations in air yields an error estimate
(mean ±standard deviation) of -0.54° ±0.75° for longitude,
and -0.12° ±3.06° for latitude.
When individual tags tested in air were examined
separately, 96%- of tags (112/117) showed average posi-
tion errors within a range of ±1.5° in longitude. Among
these 112 tags with small longitude errors, 95 had been
manufactured within the last half year and had an av-
erage and standard deviation of position error equal to
-0.50° ±0.19°, and the other 17 tags were more than
one year old and had an average position error of -0.51
±0.75°. The average is not significantly different (ANOVA
F=0.01,P>0.05) and the younger tags had a smaller stan-
dard deviation (F=412, P<0.01). No significant difference
of accuracy was observed among the 17 older tags that
could be related to their history, i.e. among four tags kept
in air without release and 13 tags released with fish and
recovered (F=1.01 for average and F=2.81 for standard
deviation, both P>0.05).
For the two tags attached to fish in pens, one tag mea-
sured only five positions with a resulting error estimate of
-2.38 ±0.39 for longitude, and -1.82° ±1.58° for latitude.
The other measured 432 positions, with a resulting error
estimate of -0.53° ±2.46° for longitude and 1.26° ±5.33°
for latitude. This is the data series presented earlier in
Figure 1. The large standard deviation in longitude error —
much larger than that obtained in other tests — initially
raised questions regarding the effect of water on the posi-
tioning techniques. However as mentioned earlier, the low
light sensitivity of both tags used in captive fish was identi-
fied as the likely cause of these large errors.
518
Fishery Bulletin 101 (3)
10.0 1
-10.0
-10.0
0.0
Longitudinal error (degree)
10.0
Figure 2
Average errors of location estimates from each archival tag in the
benchmark tests. Smaller circles are the average errors of the tags
tested in air and the larger circles are the average errors for archival
tags implanted in pen-held fish. Bar shows a standard deviation.
A more useful measure of in-water accuracy was pro-
vided by comparison between actual recapture locations
of 18 tags and the locations that those tags estimated one
or two days prior to capture (thus avoiding the disturbed
light data on the final day). Average differences of the 18
tags were -0.1 ±0.8° (range: -2.0 ±1.7°) in longitude and
-1.6 ±1.8° (range: -5.7 ±0,6°) in latitude.
Because the tag's latitude estimate based on day length
was found to have limited reliability, we estimated latitude
using sea-surface temperature (SST) as recorded in the
summary file for each day. The temperature reference field
used was the SST map published by Japan Fisheries In-
formation Service Center, which gave average SST weekly
for the western Pacific Ocean (west of 160°E), and every
10 days for the eastern Pacific Ocean (east of 160°E). The
longitude value determined automatically by the tag was
used to choose a longitude on the SST map. Along that lon-
gitude line a point was sought where the map SST matched
the SST value recorded by the tag. If multiple points were
found to satisfy this criterion, the point that gave the most
plausible movement was selected, based on fish locations
on several adjacent days. If a location still could not be
determined, it was interpolated as a midpoint between the
adjacent two days' locations.
One example of location re-estimation is shown in Figure 3.
After consulting with the SST maps, we used 1.4% of the
locations estimated automatically from 29 recaptured tags.
and 79.7% of latitudes were changed by -(-0.3 (±2.8°) on av-
erage wdth the SST method. The remaining 18.9% of days did
not provide any reasonable location estimates for various rea-
sons, including anomalous longitudinal estimates, no match
points of SST along the estimated longitudinal line, or the
existence of a wide latitudinal area showing the same SST.
Reliability of temperature and pressure sensors
One hundred tags calibrated within half a year of manufac-
ture showed average errors of 0. 1 ±0. 1°C for both internal
and external temperature sensors. Nine tags recovered from
fish and tested more than one year after manufacture showed
average errors of 0.0 ±0.1°C for both sensors. It thus appears
that no deterioration of the temperature sensors occurred
because of release-recapture or the passage of time.
No large error in pressure sensors was observed dur-
ing calibration of tags before release. However, 20 of 27
tags recovered from tagged and released fish were found
on recalibration to record substantially lower than actual
pressure. One example is shown in Figure 4. No further
deterioration of pressure sensors was observed when these
tags were kept in air for an additional half year There
was no way to know exactly when the sensor deteriora-
tion had occurred during the time the fish were in water
However, the frequency of records showing swimming at
0 m depth was remarkably higher in the second part of
Itoh et al ; Migration patterns of Thunnus orientalis determined with archival tags
519
130°E MO-E
f 1
40°N
"-l\ll
^/
^^
^-M^y^
^S^/
^>
30°N
^1
w
'■
• V
\ A
-■■J
f
' J^
^^A
.
\ \
I y^
4\
20°N
A
130E
140E
40°N
30°N
20°N
Figure 3
Locations estimated by an archival tag with a young Pacific bluefin
tuna before (upper panel) and after (lower panel) replacement of the
original latitudinal estimate based on day length by one using sea-
surface temperature. Locations out of the range of the figure and
those for which latitude was not estimated were not drawn in the
upper panel. Estimated locations for all days are shown in the lower
panel. Open circles in the lower panel are interpolated locations.
the detailed file (i.e. just before recapture) when compared
to the first part of the detailed file (i.e. just after release).
The tag manufacturer analyzed this deterioration in the
pressure sensors, and expected the sensor characteristics
to remain constant after an initial change (if one occurred),
and agreed that the early and late pressure data should be
treated separately. We assumed that the deterioration oc-
curred sometime during the middle period of the time the
fish was free, when no record was being kept in the detail
file. Recorded depths in the second part of the detail file for
eight tags with relatively large deterioration detected were
corrected by using two regression lines joined at around 30
m in real depth for each tag (Fig. 4).
Effect of the archival tag on fish
The effect of both the implantation process and the pres-
ence of the implanted tag in the fish was investigated by
macroscopic observation of recovered fish. Further infor-
mation was obtained by comparing the weight at length
520
Fishery Bulletin 101(3)
Real depth (m)
Figure 4
An example of observed deterioration in a pressure sensor
of an archival tag in a postdeployment recalibration. The
horizontal axis shows the test pressure, vertical axis is pres-
sure recorded by the tag. Dots are observed data. The solid
line bent at 25 m of real depth is formed from two regres-
sion lines, one fitted to data below and one to data above
25 m depth. This approximation to the deteriorated sensor
characteristic was used to correct pressure data for this tag.
Pressure values are converted to depth in meters. A broken
line is that of observed depth equal to real depth.
o
m
and the monthly average growth rates of tagged fish with
wild fish, and also by evidence of feeding to be found in the
records returned in the tags.
The bodies of two fish among 30 recoveries were avail-
able for observation. One fish recaptured 27 days after
release still had a scar on its skin but no trace of the tag
insertion surgery was found in its belly muscle. Another
fish recaptured 200 days after release had no trace of sur-
gery either on its skin or in its belly muscle. Surface skin
around tag stalks was ulcerated in both fish. The stalks
were immobilized in the belly muscle. The cylindrical bod-
ies of both tags were covered with membrane and located
between the stomach and the pyloric caeca. No infection
or necrosis was observed in the visceral organs around tag
bodies or in the muscle around tag stalks.
Body weights of all recaptured fish that were measured
(^=8, 17-203 days after release) were within the range of
those of wild fish of the same fork length (Fig. 5). An average
growth rate of recaptured fish was 1.4 ±0.5 cm per month
(n=6, three fish recaptured at short durations of liberty that
showed no or negative growth were excluded). A subgroup of
four fish recaptured after more than 5 months from release,
i.e. fish at liberty during the summer when growth might
be expected to be faster, had an average growth rate of 1.3
±0.6 cm — similar to that from all durations.
The average number of daily feeding events, which were
found by specific changes of visceral temperature ( Itoh et
al., 2003), increased linearly from no feeding on the day of
release up to a steady rate beginning about 30 days after
release. Thus, it appeared that fish did not feed normally
during this initial period (Fig. 6).
9 -1
/O203
8 -
/
7 -
6 -
y027
5 -
/
4 -
3 -
? -
^/^"l7
-£1^ , ,
1
1
40
50
60
Fork length (cm)
70
80
Figure 5
Comparison of weight at length of young Pacific bluefin tuna between
recaptured fish tagged with archival tags and wild (untagged) fish.
Numerals show days at liberty. An average (thick solid line) and upper
and lower 95'7f confidence limits (thin solid lines) are derived from
11,777 wild fish from 40 to 80 cm in fork length caught in 1995 and
1996 around Japan. Equations for average is H''=2.844x 10''xL-'"",
upper 95% limit is W=.3.028x lO^'x/-^^™, and lower 95% limit is
W=2.745xl0'^xL2 906 where L = fork length in cm and VV = body
weight in kg.
Horizontal movement
Estimated tracks of all fish that traveled out
of the East China Sea along with one fish that
remained in the East China Sea are shown in
Figure 7 and Figure 8. All of these fish were
released off Tsushima in November or Decem-
ber and remained in the East China Sea at least
90 days. After that, four fish entered the Sea of
Japan and moved northward from April to July
(Fig. 7, A-D). Two of them moved southward in
November one year after release (Fig. 7, C and D),
and one of the two fish returned to the region off
Tsushima where the fish were released (Fig. 7C).
One fish remained and was recaptured within
the East China Sea in November, one year after
release, although it had moved to the east coast
of the Korea Peninsula for a period in August and
September (Fig. 7E). Ten fish remained within
the East China Sea for more than five months
and were recaptured from May to June, five to
seven months after release (Fig. 7F).
Two fish moved to the Pacific Ocean (Fig. 7G
and Fig. 8). One of these fish entered the Pacific
Ocean on 7 March 1996, and then traveled east-
ward straight from a position off the south coast
of Kyushu (31°N, 131°E) to one off the east coast
of Choshi (36°N, 142°E), then stayed for a while
Itoh et al.: Migration patterns of Thunnus orientalis determined with archival tags
521
H 4-1
2 -
1 -
cP
O qO
o (HDtDD cm
«„„ "^X) On
«DO o % O
coxm canxoDODD cramaDari C3D0
ojjoo
5^
o o ■ ■
aD c^^o 0(5b°
° oOo(5c£
o
ID 00
— I—l
50
100
Days after release
150
200
Figure 6
The average number of daily feeding events plotted against days after release.
in an area of 32-37°N, 143-147°E (Fig. 7G). This fish was
recaptured by purse seine on 7 June 1996.
The other fish traveled from the western Pacific Ocean
to the eastern Pacific Ocean as follows (Fig. 8). It was re-
leased off Tsushima on 29 November 1996 at 55 cm FL and
remained for a period within the East China Sea. It moved
to the Pacific Ocean on 1 May 1997 and then traveled east-
ward straight from a position off the south coast of Kyushu
to one off the east coast of Choshi then stayed for a while
in an area of 34-39°N, 143-150°E. It moved northeastward
from 30 July to 18 August 1997, then stayed in the area
40-44°N, 152-163°E. It began the trans-Pacific migration
on 11 November 1997 at 41°N, 163°E, and traveled straight
to northern California, U.S.A. (36°N, 127°W) arriving on 15
January 1998.
After arriving in the eastern Pacific Ocean, this fish
initially stayed in an area of 33-40°N, 122-128°W, then
moved southward from 25 February to 3 March, then again
stayed in an area of 25-29°N, 116-119°W. It started mov-
ing northward on 9 May and reached 40°N, 127°W on 25
May, but without staying there moved again southward
and reached an area of 25-29°N, 116-120''W on 12 June,
close to the place from which it had departed. The fish was
recaptured by a recreational fishing vessel on 1 August
1998, 610 days after release, off Baja California, Mexico
(31°48'N, 117n8'W), at 87.6 cm FL
The track of this fish consisted of apparently separable
segments, five traveling periods and six resident ones. All
of the fish that moved out of the East China Sea showed
the same type of pattern, staying resident in an area for a
relatively long period and then traveling continuously for
at least several days in a stable direction.
The terms "traveling phase" and "residency phase" are
used in the following description. If a fish moved continu-
ously for more than three days in a stable direction covering
more than 700 km in total distance, it was considered to be in
a traveling phase — at all other times in a residency phase. A
few movements for short periods or short distances (or both)
were also observed during periods of a residency phase: a fish
resident off the east coast of Hokkaido (40-44°N, 152-163°E)
shifted eastward gradually within the area during a period
of two months (Fig. 8). Another fish resident in the northern
area of the East China Sea moved rapidly to the southern
area of the East China Sea at the end of December and came
back rapidly to the northern area in early May (Fig. 9). In-
dividual movements were completed within a few days and
the total distances moved were far shorter (380 and 310 km,
respectively) than those seen in traveling phases.
A total of 12 traveling phases were identified in records
of six fish (Table 1). The direction of travel stayed constant
within each phase, except in one case where a fish com-
pletely turned around in the middle of traveling (in the east-
em Pacific Ocean in May and June 1998). Daily distances
moved during those traveling phases were calculated. To
reduce the infiuence of scatter in the estimated locations,
three-day running averages of latitude and longitude were
used for calculation. Excluding the one trans-Pacific migra-
tion phase of exceptional length, 7636 km (66 days), the
total distance per traveling phase ranged from about 730
to 3406 km (average: 1430 km). The duration of a traveling
phase was four to 35 days (average: 17 days) and the dis-
tance traveled ranged from 59 to 182 km (average: 104 km).
Six residency phases which occurred between two clearly
identified traveling phases lasted from 40 to 125 days (aver-
age: 81 days). In total, 83% of days were in a residency phase
and 17% of days were in a traveling phase. If residency
phases for which the beginning or end could not be defined
because offish release or recapture were also included, the
average duration of residency phases increased to 110 days,
and the proportion days belonging to each phase became
87% in residency and 13% in traveling phases.
Comparison of fish behavior and ambient water
temperature between traveling and residency phases
Several points regarding fish behavior, described in detail
in Itoh et al. (2003), were compared between all days in a
traveling phase and ten days in the residency phase that
for four fish immediately preceded the traveling phase.
In the case of one fish (no. 241) where data for preceding
522
Fishery Bulletin 101 (3)
130°E
140°E
40°N
30°N
\
.'. Red 20596
130°E
140°E
40°N
30°N
Rec300497
Figure 7
Tracks of young Pacific bluefin tuna estimated by archival tags.
"Rel" and "Rec" mean release and recapture, respectively. Numerals
are dates in ddmmyy. Each panel shows the track of one fish with
an archival tag.
days were not available, ten days from the residency phase
immediately following were used instead. Because errors
in determining geographical positions introduced scatter in
the sequence of estimated positions, the onset (or end) of a
traveling phase was not always easy to define. In response
to this situation, three days of the residency phase nearest
Itoh et al; Migration patterns of Thunnus orientolis determined with archival tags
523
130°E
140°E
40°N
\
'■ 270898
30°N
130°E
140°E
D
/ a1\
1
* fj^\
y
'^S
°N
x2 ■
080898
\
\
AmJ^ 111198
A_^. 'V,
\ 140798 f
^t^
V/ /
?^Rel231197 "'
fl^^
^
■'.. Rec291198
°N
fj
Br ' '-'o'
'
Figure 7 (continued)
to the traveling phase were not included, that is to say, ten
days between the fourth and thirteenth days preceding
(or following) the onset (or end) of a traveling phase were
used.
Among 12 features investigated, three differed between
the two phases (Table 2). The temperature difference be-
tween fish viscera and ambient water (thermal excess) was
more than 1.0°C higher during the traveling phase for four
out of five fish. In the one remaining fish (no. 241), data
used for analysis were those from days at the end of the
traveling phase. The larger thermal excess during a trav-
eling phase was observed at both daytime and nighttime
(Fig. 10). The second significant feature was that all fish
dived to water deeper than 150 m more frequently during
the traveling phase. Except for one fish (no. 241), which
spent a long time in water deeper than 150 m, almost all
524
Fishery Bulletin 101(3)
130°E
140°E
E
f
^'.^
N
A
/
'*TV
J '
' r^
>'V~^^
;^ /^
"jl
170898
■^c
40°N
Rel211197__ ;
m
■"150998
J"
■J
30°N
^f.4
Rec301198
,
ar •
130°E
140°E
40°N
30°N
Figure 7 (continued)
records of excursions below 150 m were due to spikes of
deep diving of short duration, less than 10 minutes. Fi-
nally the frequency of daily feeding events, detected by a
change of visceral temperature, was slightly higher dur-
ing a traveling phase (1.6 ±0.6) than a residency phase
(1.1 ±0.4).
Records of surface temperature from the summary file
were examined to answer three questions regarding the re-
lation of water temperature to traveling. The first question
was whether any water temperature change, an increase in
spring and summer or a decrease in winter, was observed
several days prior to the onset of the traveling phase. Such
a temperature change was observed in 10 out of 12 cases
( Table 3, Fig. 11). In those 10 cases, the water temperature
increased in spring and summer to 19-26''C (average of
22°C), and decreased in winter to 15-17°C (16°C).
Itoh et al.: Migration patterns of Thunnus orientalis determined with archival tags
525
130°E
140° E
40°N
30°N
Figure 7 (continued)
10898
Figure 8
Track of a young Pacific bluefin tuna that traversed the Pacific Ocean, estimated with an archival tag. "Rel" and "Rec" mean release and
recapture, respectively. Numerals are dates in ddnimyy.
The second question is whether the water temperature
at the onset or end of traveling was within the temperature
range of 14— 20°C, considered to be in the temperature range
preferred by young Pacific bluefin tuna (Itoh et al., 2003). If
the act of traveling was simply a reaction to the water tem-
perature, fish would be expected to travel from water with a
temperature out of that range to one within that range. How-
ever that was observed in only two of 12 cases (Table 3).
The third question was whether the temperature at the
end of traveling was a temperature that the fish encoun-
tered for the first time since the onset of traveling. We
found, however, that there was no specific trend in water
temperatures during traveling phases (Fig. 11). Six of 12
temperatures at the ends of traveling phases were not the
first one that the fish had experienced during the phase
(Table 3).
526
Fishery Bulletin 101 (3)
125°E
130°E
35° N
30° N •
rr«.».-^
35° N
30° N ■'
Figure 9
Rapid movements of a fish with an archival tag in the residency
phase in the East China Sea. The fish moved rapidly south (upper
panel; open circles are in December 1995 and solid circles are in
January 1996) and north (lower panel; open circles are in April
1996 and solid circles are in May 1996).
Discussion
Reliability of archival tag data
The reliability of archival tag data for geolocation estimates
based on measured light inten.sity ha.s been examined by
implanting the tags in pen-held fish or attaching the tags to
a stationary subsurface mooring (Welch and Eveson, 1999;
Musyl et al., 2001; Gunn et al.-». About one degree of reli-
ability for both longitude and latitude were the results. Our
study included further tests: for a large number of tags; for
sensors other than light sensors; for reliability of tags over
time; for tags manufactured by Northwest Marine Technol-
ogy that were applied to wild young blucfin tuna and not
fully examined in previous studies; and finally for the effect
of tag attachment on Pacific bluefin tuna.
The benchmark test in this study showed that longitude
estimated by archival tags had an error (mean ±standard
deviation) of -0.54° ±0.75°, which differed by only -0.1°
±0.8° from a comparison of in-water tag position results
with actual recapture locations. In the on-land benchmark
test, the mean error did not change with tag age, although
the standard deviation slightly increased. Ninety-six
percent of all tags tested were considered to have suffi-
cient reliability in longitudinal estimation. We concluded,
therefore, that the archival tag is a reliable tool to estimate
longitude on a scale of about one degree.
Latitudes estimated automatically from day length car-
ried larger errors than estimations of longitude, and the
accuracy of estimation changed with season as well as with
the latitude itself (Hill, 1994; Hill and Braun, 2001). Smith
and Goodman (1986) recommended estimating latitude by
Itoh et al.: Migration patterns of Thunnus orientalts determined with archival tags
527
Table 1
Information
on traveling phases of young bluefin tuna as recorded by archival tags. Daily travel distances
were estimated from a |
three-day ru
nning average of latitude and longitude.
ID offish
and number
Onset of
End of
Duration of
Total
Daily
(e.g. Ml)
traveling phase
traveling phase
traveling
distance
distance
of traveling
phase
traveled
traveled
phase
Area
Date
Location
Date
Location
(days)
(km)
(km/day)
241 Ml
western Pacific
1 May 1997
31N 132E
15 May 1997
36N 143E
15
1,341
89.4
241 M2
western Pacific
30 Jul 1997
40N 147E
18 Aug 1997
42N 152E
20
1,161
58.1
241 M3'
central Pacific
11 Nov 1997
41N 163E
15 Jan 1998
36N 127W
66
7,636
115.7
241 M4
eastern Pacific
25 Feb 1998
32N 126W
3 Mar 1998
29N 121W
7
1,043
148.9
241 M5-
eastern Pacific
9 May 1998
29N 119W
12 Jun 1998
27N 120W
35
3,406
97.3
209 Ml
western Pacific
7 Mar 1996
30N 131E
6 Apr 1996
35N 142E
31
1,860
60.0
164 Ml-
Sea of Japan
4 May 1996
36N 134E
12 May 1996
38N 138E
9
770
85.6
319 Ml-
Sea of Japan
18 Apr 1997
35N 132E
30 Apr 1997
38N 138E
13
1,346
103.5
688 Ml
Sea of Japan
23 Jun 1998
37N 132E
30 Jun 1998
39N 138E
8
877
109.7
688 M2-'
Sea of Japan
3 Nov 1998
44N 140E
14 Nov 1998
34N 128E
12
1,611
134.2
760 Ml
Sea of Japan
13 Jul 1998
35N 129E
8 Aug 1998
40N 137E
27
1,590
58.9
760 M2-'
Sea of Japan
11 Nov 1998
44N 14 IE
14 Nov 1998
38N 136E
4
727
181.7
Average
16.5-^
1430^
103.6
' Trans-Pacific migration.
- Detail file during the traveling phase exists and was
used for analyses
in Table 2.
* The trans-Pacific migration, which
was too long in time and distance, was not included in
the calculation.
comparing measured water temperatures at three depths
to the water temperature maps at each depth. However, it
is quite difficult to obtain water temperature maps for the
whole range of times and areas where tuna migrate other
than those for SST. Therefore we decided for the latitudinal
estimation to rely on the SST maps and on longitude esti-
mated by the tags. One difficulty with the SST method for
latitude is that water of the observed surface temperature
might occur at two different latitudes, thus not implying
a single unique position. However because the latitude of
about 80% of all days could be uniquely determined from
SST, we considered the adjustment method taken here to
be acceptable for the purpose of the present study. Although
it was not possible to check the accuracy of the latitudinal
estimation independently, judging from the accuracy of the
longitude values we used to locate the appropriate North-
South stripe on the SST maps and from the rapidity of
temperature variations found along those stripes, which in
most cases tightly constrained our estimates, we expected
the accuracy of latitudinal estimation to be around one de-
gree, which would be sufficient for a study of Pacific bluefin
tuna migration.
Some deterioration was observed in pressure sensors.
The need for recalibration of sensors after recovery should
be emphasized.
Effect of the tag on fish
Fish in this study were much smaller than those in other
archival tag studies of southern bluefin tuna and Atlantic
bluefin tuna (Block et al., 2001; Gunn and Block, 2001). The
tagging success achieved confirms that the type of archival
tag we used can be applied to fish at least down to 43 cm
FL. No fish died because of the attachment of the tag during
the experiment on pen-held fish. The recovery rate ( 18.1% )
for fish tagged with archival tags was similar to the rate
(19.1%) for those in the conventional tagging experiment
conducted in the 1980s off Nagasaki Prefecture, including
Tsushima, for the same size fish of the species (Bayliff et
al., 1991). This comparison should be made cautiously for
the following reasons. The unusual appearance of an archi-
val tag body would attract the attention of the finder who
gutted the fish and might lead to a higher reporting rate.
Increased fishing effort for young Pacific bluefin tuna in the
northern part of the East China Sea in the 1990s compared
to that in the 1980s might lead to a higher recapture rate.
The inconspicuous stalk of an archival tag which was the
only externally detectable sign of its existence might lead
to a low discovery rate. Indeed, because many recoveries of
archival tags were made by consumers while gutting the
fish, archival tags implanted in the body of the fish must
have been overlooked by fishermen and by sellers at fish
markets. However, judging not only by the similar but also
high recovery rates, it seems that damage and stress of
handling at implantation and that due to the archival tags
being carried by the fish did not have much more effect on
fish survival than did the conventional tags.
Macroscopic observations of two wild fish recovered with
archival tags showed that the surgical injuries that occurred
during archival tag implantations healed after one month
528
Fishery Bulletin 101(3)
Table 2
Comparison of various averaged environmental, physiological, and behavioral values between a
phase for young Pacific bluefin tuna.
raveling phase and
a residency
Subject
Phase
ID offish
Average of
difference
164
241
319
688
760
Number of days
Traveling
Residency
9
10
2
10
13
10
12
10
4
10
Swimming depth (m)
Traveling
Residency
Difference
8.7
14.0
-5.3
22.7
15.0
7.7
12.0
45.3
-33.3
22.8
10.2
12.5
7.5
17.4
-9.9
-5.6
Ambient water temperature (°C)
Traveling
Residency
Difference
12.9
17.2
-4.3
17.4
17.9
0.5
12.8
15.6
2.8
19.8
18.7
1.1
16.3
15.3
0.9
-1.1
Temperature of viscera (°C)
Traveling
Residency
Difference
17.8
20.7
-2.9
21.1
21.8
-0.7
17.5
19.1
-1.6
25.9
23.2
2.7
22.7
20.8
1.9
-0.1
Temperature difference between ambient water and fish viscera (°C)
Traveling
Residency
Difference
4.9
3.6
1.4
3.7
3.9
-0.2
4.7
3.6
1.2
6.1
4.5
1.6
6.4
5.5
1.0
1.0
The number of depth records deeper than 150 m per day
Traveling
Residency
Difference
5.4
1.3
4.1
36.0
13.7
22.3
1.6
0.0
1.6
3.7
0.3
3.4
4.3
0.2
4.1
7.1
The number of feeding events per day
Traveling
Residency
Difference
2.6
1.7
0.9
1.0
0.6
0.4
1.6
1.2
0.4
1.7
0.9
0.8
1.3
1.3
0.1
0.5
Percentage of days when a rapid ascent at dawn was observed
Traveling
Residency
Difference
11%
70%
-59%
100%
70%
30%
8%
50%
-A2%
67%
40%
27%
0%
0%
0%
-9%
Percentage of days when a rapid descent at dusk was observed
Traveling
Residency
Difference
11%
80%
-69%
50%
100%
50%
8%
50%
-42%
27%
10%
17%
50%
40%
10%
-27%
Percentage of days when swimming depth was significantly
deeper during daytime than at nighttime
Traveling
Residency
Difference
89%
80%
9%
100%
70%
30%
50%.
90%.
-40%
75%
80%
-5%
25%
100%
-75%
-16%
Percentage of days when ambient water temperature was
significantly lower during daytime than at nighttime
Traveling
Residency
Difference
56% 100%
70% 100%
-14% 0%
58%
30%
28%
67%
60%
7%
50%
70%
-20%
0%
Percentage of days when temperature offish viscera was
significantly higher during daytime than at nighttime
Traveling
Residency
Difference
89%.
80%
9%
0%.
20%
20%
92%
100%
-8%
75%.
90%
-15%
100%
100%
0%
-7%
Accumulated swimming depth change per day (m)
Traveling
Residency
Difference
5097
5639
-542
6031
4533
1498
6224
6123
101
7884
5266
2618
4576
6386
-1810
373
and there was no scar after a half year. No damage to vis-
ceral organs was observed for the two fish. The finding is con-
sistent with that for southern bluefin tuna (Gunn et al.'^).
Fish tagged with archival tags usually showed similar
behavioral patterns, such as diurnal change of swimming
depth and vertical excursions at dawn and dusk (Itoh et al.,
Itoh et al.: Migration patterns of Thunnus onentalis determined with archival tags
529
Table 3
Water temperature changes associated with the travehng phase of Pacific bluefin tuna.
Test 1:
Circles mark the occurrence of a sea-surface temperature change before the onset of a traveling phase (increasing in
spring-summer, decreasing in autumn-winter). T-change (temperature change) shows the maximum (in spring-summer)
and minimum (in autumn-winter) water temperature (in the temperature change before traveling.
Test 2:
Comparison of temperatures at the onset and end of traveling to a temperature range of 14-20°C, which is thought to be
a preferred temperature range for young Pacific bluefin tuna. Temperature values within the range of 14-20°C are under-
lined.
Test 3:
Circles mark cases where the temperature at the end of the traveling was a temperature that the fish encountered for the
first time since the onset of the traveling Pacific.
Traveling phase
number
Test 1
Test 2
Season
ID offish
T-change
T-onset
T-end
Tests
Spring
209
164
Ml
Ml
O
O
20
19
20.7
14,7
17.0
12.4
O
319
Ml
—
16
15.6
13.0
—
241
Ml
o
23
21.7
19.1
O
Summer
241
M2
o
23
21.9
17.4
o
241
M5
o
20
18.2
19.3
—
688
Ml
o
21
19.5
20.7
—
760
Ml
o
26
24.0
23.6
—
Autumn-winter
241
M3
o
15
14.7
14.3
—
241
M4
—
13
16.0
17.4
o
688
M2
o
17
19.9
22.3
o
760
M2
o
15
14.5
20.1
o
2003 ) from the second day after release, and their feeding
frequency reached a constant level one month after release.
Also, the fish maintained weight-at-length similar to that
of wild fish. The average growth rate of fish tagged with
archival tags observed in our study ( 1.3 cm/month) was less
than the growth rate of wild fish observed between ages one
and twoin previous studies (1.7-3.3 cm/month, Aikawa and
Kato, 1938; Yokota et al., 1961; Yukinawa and Yabuta, 1967;
Bayliffet al., 1991; Bayliff, 1993; Foreman, 1996), except for
the result of Bell (1963b) (1.3 cm/month). Because similar
growth rates were observed for fish recaptured more than
a half year after release that spent the summer at large,
it appears that the lesser growth rate in the present study
is not due to the fact that some fish spent only winter at
liberty, when the growth rate is less than that in summer
(Yukinawa and Yabuta, 1967; Bayliff, 1993). Judging from
these facts, we suggest that the effect of archival tags on
fish behavior and physiology seems to be minor, although
there is a possibility that carrying an archival tag caused
a reduction in growth rate of the fish.
Horizontal movement
Archival tags revealed the movement pattern of young
Pacific bluefin tuna individuals, which could be divided into
the two clearly-separable phases of traveling and residency.
These two phases were observed for all individuals that
moved out from the East China Sea.
T3
0)
8
12
16
20
24
Time of day (fi)
Figure 10
Hourly averages of temperature difference between
ambient water and fish viscera in both traveling
(circle) and residency phase (triangle) of young
Pacific bluefin tuna with archival tags. Data from
five individuals are all combined. Bar indicates the
standard deviation.
The residency phase is considered a normal condition for
young bluefin tuna, comprising 83-87% of their time. Fish
with archival tags tended to stay in the areas of the East
530
Fishery Bulletin 101 (3)
1 Oct 1997 20 Nov 1997 9 Jan 1998 28 Feb 1998 19 Apr 1998 8 Jun 1998 28 Jul 1998
Figure 11
Changes in ambient water temperature recorded by an archival tag in a young Pacific bluefin
tuna. Data are for the fish that traversed the Pacific Ocean. The symbols from Ml to M5 designate
from the first to fifth traveling phase of the fish. Each dot is an ambient water temperature at 0 m
depth recorded in the summary file of the archival tag each day. Large black and brighter circles
are temperatures at the onset and end of traveling, respectively. Arrows indicate the observed
ambient water temperature changes several days before the onset of the travel phase.
China Sea, off the east coast of Choshi, off the east coast
of Hokkaido in the western Pacific Ocean, off Southern
California and Baja CaHfornia, off northern California in
the eastern Pacific Ocean, and off the west coast between
Akita and Hokkaido in the northern Sea of Japan. The first
four areas correspond to the major known fishing grounds
of young Pacific bluefin tuna.
The last two areas do not correspond to previously known
fishing areas for young Pacific bluefin tuna. In the eastern
Pacific, young bluefin tuna are usually caught in an area
from 23 to 34°N, off California to Baja California, from May
to October by purse seine (Calkins, 1982). Catch records
in the northern area around 40°N were scarce, and all of
them were for catches from summer to autumn (Radovich,
1961; BaylifT, 1994). It was not expected that young Pacific
bluefin tuna were to be found around 40°N in winter, but
the archival tag records showed fish staying in an area of
33-40°N, off northern California, from winter to spring. In
the northern Sea of Japan, young Pacific bluefin tuna are
usually caught by set nets in coastal areas, and are not
caught in the offshore area. The archival tag records again
showed fish staying in this area in summer and moving
southward without being captured. These cases clearly in-
dicate the ability and advantage of archival tags to provide
information on fish distribution and migration when and
where fishing has not been conducted.
An archival tag demonstrated that an individual young
Pacific bluefin tuna was able to travel more than 7000 km
without pause and to travel for more than two months. The
daily moving distance during the traveling phase ranged
from 59 to 182 km, and averaged 104 km. Assuming a con-
stant swimming speed, the daily average swimming speed
was estimated as a range from 1.3 to 4.1 knots (average of
2.3 knots).
Itoh et a\ Migration patterns of Thunnus onentalis determined with archival tags
531
These horizontal swimming speeds (1-4 knots) are com-
parable to those of larger young Pacific bluefin tuna and
same-size fish of other Thunnus species, namely yellowfin
tuna ( T. albacares ), bigeye tuna ( T. obesus ), and albacore (T!
alalunga), determined from acoustic tracking experiments
(Laurs et al., 1977; Carey and Olson, 1982; Holland et al.,
1990; Block et al., 1997; Marcinek et al., 2001 ). Sustainable
swimming speeds based on oxygen demand and supply for
yellowfin tuna and skipjack tuna (Katsuwonus pelamis) of
1.5-2 kg in body weight were estimated to be 2-4 times FL
per second (Brill, 1996;Korsmeyeret al., 1996), correspond-
ing to 2.3-4.7 knots for fish of 60 cm FL. Applying to young
bluefin tuna the same rule (2 -4 times FL) used by those
workers as a summary of their data, the expected range
of sustainable swimming speeds that do not accumulate
an oxygen debt would cover the range of estimated aver-
age swimming speeds during traveling phases. Of course,
the swimming speed of young Pacific bluefin tuna based
on a constant moving speed between two successive daily
locations obviously carries some errors. First, a fish might
not maintain a constant swimming speed all day long.
For example, the daytime swimming speed of albacore
observed in an acoustic tracking experiment was reported
to be 1.3-2.1 times as great as that at night (Laurs et al.,
1977). Second, the influence of water current which should
be taken into consideration (Brill, 1996) was completely
ignored. Third, assuming straight-line travel between two
daily positions would lead to underestimation of actual
daily distances traveled, even though the direction during
traveling phases could be approximated as a straight line.
Even if these errors had been large and the true swimming
speed had been twice as large as that which was estimated,
these estimated average swimming speeds were still with-
in the range of sustainable speeds.
Although the horizontal movement clearly differed, many
features regarding vertical movement were the same for
both residency and traveling phases. One parameter that
did differ was that of temperature, where the difference
between water and fish viscera was 1.0°C larger during the
traveling phase than during the residency phase. Feeding
causes an increase of visceral temperature in tuna (Carey
et al., 1984; Gunn et al., 2001; Itoh et al., 2003). However,
the slightly more frequent feeding in traveling phases was
not enough to explain the large thermal excess. In addition,
the larger temperature difference was observed not only at
daytime when the visceral temperature was usually higher
because of more frequent feeding, but also at night when
the visceral temperature was usually lower (Itoh et al.,
2003). The visceral temperature seemed to be raised by
high muscle temperature during traveling phases. If this
is indeed the case, this would lead to an increase in the
delivery rate of oxygen to muscle, which would make the
fish less tired and more able to travel (Stevens and Carey,
1981; Brill, 1996). The distinct traveling phase might be
one of the tactics adopted by young Pacific bluefin tuna to
use energy most efficiently for long distance travel.
During the traveling phase, the frequency of feeding in-
creased slightly and the fish dived to water deeper than 150 m
depth more frequently. The fish would feed and seek food at
least as aggressively as in the residency phase.
Ambient water temperature is one of the most important
environmental factors for young Pacific bluefin tuna (Sund
et al., 1981; Koido and Mizuno, 1989; Ogawa and Ishida,
1989; Itoh et al., 2003). The onset of most traveling phases
were preceded by specific water temperature changes that
reached the upper or lower limit of the preferable water
temperature of 14— 20°C for young Pacific bluefin tuna (Itoh
et al., 2003). Changes in ambient water temperature ap-
pears to be a possible trigger for a fish to move. Because no
remarkable change in frequency of feeding was observed
within several days before or after traveling began, the
possibility that shortage of prey is a trigger for migration
does not seem to be plausible.
If the impulse to travel in young Pacific bluefin tuna is
regulated only by the search for the preferred water tem-
perature range and the aim of traveling is to reach the pre-
ferred water temperature range, the ambient temperature
would be expected to be out of the preferred range at the
onset of travel and within that range at the end. However
this did not occur in the fish studied. In addition, half of
the observed traveling phases were continued after the fish
encountered along the way the same temperature that was
present at the end of traveling. According to these results,
it appears that the preferred water temperature is neither
the sole regulator of traveling in young Pacific bluefin tuna
nor is it the sole aim of traveling.
Data from tagged fish released around Japan in the
1980s revealed that some fish released from Nagasaki
prefecture migrated to the Sea of Japan and others to
the Pacific Ocean (Bayliff et al., 1991). Archival tag data
showed that fish released in the same season and same
area migrated in various patterns involving different onset
times and different destinations when traveling from the
East China Sea. In addition, some fish continued to remain
in the East China Sea. The migration scenario seems not
to be fixed or limited for age 0-1 fish distributed around
the East China Sea. A detailed examination of fish be-
haviors relating to the water temperature has suggested
that although young Pacific bluefin tuna prefer a specific
temperature range, they can still tolerate temperatures
outside of this range (Itoh et al., 2003). This temperature
tolerance would contribute to the diversity of migration
scenarios for the species.
The trans-Pacific migration
The trans-Pacific migration of Pacific bluefin tuna was
originally validated by tagging tuna both from the west-
ern Pacific Ocean to the eastern Pacific Ocean and from the
eastern Pacific Ocean to the western Pacific Ocean (Orange
and Fink, 1963; Clemens and Flittner, 1969). The duration
required for trans-Pacific migration was estimated as 215
days from the shortest interval between the release offish
from one side of the Ocean to the recovery of fish on the
other side (Bayliff et al., 1991). The present study obtained
a full record of daily locations during a trans-Pacific migra-
tion of one fish. The fish took two months to traverse the
whole Pacific Ocean, which was much shorter than expected
from previous records. The starting time for trans-Pacific
migration was estimated by Yamanaka (1982) as May-
532
Fishery Bulletin 101 (3)
August based on fishery information and as autumn and
winter by Bayliff et al. (1991) based on tagging data. The
fish observed in our study started its trans-Pacific migra-
tion in late autumn. Although data concerning distribution
of young Pacific bluefin tuna in the central Pacific Ocean
is limited, a record of young Pacific bluefin tuna catch (age
1-3, age 1 mainly) in the area of 35^5°N, 150°E-140°W
from April to November has been reported (Saito et al.'^).
Moreover, two tagged fish were recaptured in the central
Pacific Ocean at 38°N, 172°E in June and 39°N, 162°W in
June, respectively (Bayliff et al., 1991). Although the sea-
sons differ, the path of the fish tagged with an archival tag
passed near these locations. The limited data available at
present suggest that the trans-Pacific migration route lies
in this area. Together with data which would be obtained
from future recovery of additional fish tagged with archival
tags, we expect that the overall features of trans-Pacific
migration to be revealed in the near future.
Acknowledgments
We thank the staff of the Marino Forum 21 and the
Kagoshima Fisheries Experimental Station for their coop-
eration in the pen-held fish experiment. We also thank troll
fishermen, staff in the Kamiagata Fisheries Cooperative
Association, the Tsushima Fisheries Extension Service,
and the Nagasaki Fisheries Experimental Station for their
cooperation with tag and release procedures. We acknowl-
edge fishermen, consumers, and stafFof the Inter-American
Tropical Tuna Commission for their kindness in returning
archival tags along with pertinent information about the
recapture. We especially thank J. Gunn in CSIRO for giving
us valuable information about implanting the archival tag
in fish. We are grateful to Northwest Marine Technology Inc.
and Tanaka Sanjiro Co., Ltd., for providing us with archi-
val tags. We would like to thank P. Ekstrom of Northwest
Marine Technology Inc. for his critical review and help with
the English text. We thank our staff in Japan NUS Co., Ltd.,
the Suidosya Co., Ltd., and the National Research Institute
of Far Seas Fisheries, and also T Kitagawa in the Ocean
Research Institute of the University of Tokyo, for their help
in implanting the tags in fish. We gratefully acknowledge S.
Kume of Japan NUS Co., Ltd., N. Baba of Fishery Research
Agency, and Z. Suzuki and Y. Uozumi of National Research
Institute of Far Seas Fisheries for their critical review.
Literature cited
Aikawa, H.
1949. Fisheries population ecology. SangyoToshoCo. Ltd.,
Tokyo, 545 p. |In Japanese.]
Saito, S., K. Shimazaki, and T. Sato. 198L Distribution
of bluefin tuna around the area of polar front in the north
Pacific Ocean. In Report of the Maguro Gyogyo Kcnkyu
Kyogikai, in 1980, p. 247-252. Nat. Res. Inst. Far Seas
Fish. Pelagic Fish Resource Division, 5-7-1 Shimizu-Orido,
Shizuoka, Shizuoka, 424-8633, Japan. |In Japanese, the
title was translated by authors.)
Aikawa, H., and M. Kato.
1938. Age determination of fish. I. Bull. Japan. Soc. Sci.
Fish. 7:79-88. [In Japanese.]
Anonymous.
1994. Archival tags 1994: present and future. U.S. Dep.
Commer, NOAA Tech. Memo. NMFS-SEFSC-357, 42 p.
Arnold. G., and H. Dewar.
2001. Electronic tags in marine fisheries research: a 30-
year perspective. In Electronic tagging and tracking in
marine fisheries (J. R. Sibert and J. L. Nielsen, eds.), p.
7-64. Kluwer Academic Publisher, Netherlands. j
Bayliff, W. H. \
1980. Synopsis of biological data on the northern bluefin
tuna, Thunnus thynnus (Linnaeus, 1758), in the Pacific
Ocean. Spec. Rep. lATTC 2:261-293.
1993. Growth and age composition of northern bluefin tuna,
Thunnus thynnus, caught in the eastern Pacific Ocean, as
estimated fi-om length-ft'equency data, with comments on
trans-pacific migrations. Bull. LATTC, 20:503-540.
1994. A review of the biology and fisheries for northern
bluefin tuna, Thunnus thynnus, in the Pacific Ocean. FAO
Fish. Tech. Pap. 336/2:244-295.
Bayliff, W. H., Y. Ishizuka, and R. B. Deriso.
1991. Growth, movement, and attrition of northern bluefin
tuna, Thunnus thynnus, in the Pacific Ocean, as determined
by tagging. Bull. LATTC 20:1-94.
Bell, R. R.
1963a. Synopsis of biological data on California bluefin tuna
Thunnus saliens Jordan and Everman 1962. FAO Fish.
Rep. 6(21:380-421.
1963b. Preliminary age determination of bluefin tuna, Thun-
nus thynnus. Calif Fish. Game 49:307.
Block, B. A., H. Dewar, S. B. Blackwell, T. D. Williams,
E. D. Prince, C. J. Farwell, A. Boustany S. L. H. Teo. A. Seitz,
A. Walli, and D. Fudge.
2001. Migratory movements, depth preferences, and thermal
biology of Atlantic bluefin tuna. Science 293:1310-1314.
Block, B. A., J. E. Keen, B. Castillo. H. Dewar, E. V. Freund,
D. J. Marcinek, R. W. Brill, and C. Farwell.
1997. Environmental preferences of yellowfin tuna {Thun-
nus albacares) at the northern extent of its range. Mar
Biol. 130:119-132.
Brill, R. W.
1996. Selective advantages conferred by the high perfor-
mance physiology of tunas, billfishes, and dolphin fish.
Comp. Biochem. Physiol. 113A:3-15.
Calkins, T. P
1982. Observations on the purse-seine fishery for bluefin
tunaiThunnus thynnus) in the eastern Pacific Ocean. Bull.
LATTC 18:123-225.
Carey, F. G., J. W. Kan wisher, and E. D. Stevens.
1984. Bluefin tuna warm their viscera during digestion.
J Exp. Biol. 109:1-20.
Carey, F. G., and R. J. Olson.
1982. Sonic tracking experiments with tunas. ICCAT Coll.
Vol. Sci. Papers XVII. 2:458-466.
Cayre, P.
1991. Behavior of yellowfin tuna iThunnus albacares) and
skipjack tuna iKatsuwonus pelamis) around fish aggregat-
ing devices ( FADs ) in the Comoros Islands as determined by
ultrasonic tagging. Aquat. Living. Resour 4:1-12.
Cayre, P., and F. Marsac
1993. Modeling the yellowfin tuna iThunnus albacares) ver-
tical distribution using sonic tagging results and local envi-
ronmental parameters. Aquat. Living Resours. 6:1-14.
Itoh et al.: Migration patterns of Thunnus onentalis determined with archival tags
533
Clemens, H. B., and G. A. Flittner.
1969. Bluefin tuna migrate across the Pacific Ocean. Calif.
Fish Game 55:132-135.
CoUette, B. B.
1999. Mackerels, molecules, and morphology. Soc. Fr Ich-
thyol. 25:149-164.
Foreman, T.
1996. Estimates of age and growth, and an assessment of
ageing techniques for northern bluefin tuna, Thunnus thyn-
nus. in the Pacific Ocean, Bull. lATTC 21:75-123.
Gunn, J., and B. Block.
2001. Advances in acoustic, archival, and satellite tagging
of tunas. In Tuna: physiology, ecology, and evolution (B. A.
Block and E. D. Stevens, eds.), p.167-224. Academic Press,
San Diego, CA.
Gunn, J., J. Hartog, and K. Rough
2001. The relationship between food intake and visceral
warming in southern bluefin tuna iThunnus maccoyii).
Can we predict from archival tag data how much a tuna
has eaten? In Electronic tagging and tracking in marine
fisheries (J. R. Sibert and J. L, Nielsen, eds.), p. 109-130.
Kluwer Academic Publisher, Netherlands.
Hill, R. D.
1994. Theory of geolocation by light levels. In Elephant
seals, population ecology, behavior and physiology (B. J.
LeBoeuf and R. M. Laws, eds. ), p. 227-236. Univ Califor-
nia Press, Berkeley, CA.
Hill, R. D.,and M. J. Braun
2001. Geolocation by light level. The next step: latitude. In
Electronic tagging and tracking in marine fisheries (J. R.
Sibert and J. L. Nielsen, eds.), p. 315-330. Kluwer Aca-
demic Publisher, Netherlands.
Holland, K. N., R. W. Brill, and R. K. C. Chang.
1990. Horizontal and vertical movements of yellowfin and
bigeye tuna associated with fish aggregating devices. Fish.
Bull. 88:493-507.
Hunter, J. R., A. W. Argue, W. H. Bayliff, A. E. Dizon,
A. Fonteneau, D. Goodman, and G. R. Sekel.
1986. The dynamics of tuna movements: an evaluation of
past and future research. FAO, Fish. Tech. Pap. 277:vi,
78 p.
Itoh, T, S. Tsuji, and A. Nitta.
2003. Swimming depth, ambient water temperature prefer-
ence, and feeding frequency of young Pacific bluefin tuna
( Thunnus orienlalis) determined with archival tags. Fish.
Bull. 101:535-544.
Kitagawa, Y., Y. Nishikawa, T. Kubota, and M. Okiyama.
1995. Distribution of ichthyoplanktons in the Japan Sea
during summer, 1984, with special reference to scombroid
fishes. Bull. Jpn. Soc. Fish. Oceanogr. 59(2):107-114. (In
Japanese.]
Kitagawa, T, H. Nakata, S. Kimura, T Itoh, S. Tsuji, and
A. Nitta.
2000. Effect of ambient temperature on the vertical distribu-
tion and movement of Pacific bluefin tuna Thunnus thyn-
nus onentalis. Mar Ecol. Prog. Ser. 206:251-260.
Koido, T, and K. Mizuno.
1989. Fluctuation of catch for bluefin tuna {Thunnus thyn-
nus) by trap nets in .Sanriku coast with reference to hydro-
graphic condition. Bull. Jpn. Soc. Fish. Oceanogr. 53:
138-152. [In Japanese.!
Korsmeyer, K. E., H. Dewar, N. C. Lai, and J. B. Graham.
1996. The aerobic capacity of tunas: adaptation for multiple
metabolic demands. Comp. Biochem. Physiol. 113A:17-24.
Laurs, R. M., H. S. H. Yuen, and J. H. Johnson.
1977. Small-scale movements of albacore, Thunnus ala-
lunga, in relation to ocean features as indicated by ultra-
sonic tracking and oceanographic sampling. Fish. Bull.
75:347-355.
Marcinek, D. J., S. B. Blackwell, H. Dewar, E. V. Freund,
C. Farwell, D. Dau, A. C. Seitz, and B. A. Block.
2001. Depth and muscle temperature of Pacific bluefin tuna
examined with acoustic and pop-up satellite archival tags.
Mar Biol. 138:869-885.
Metcalfe, J. D. and G. P Arnold
1997. Tracking fish with electronic tags. Nature 387:665-
666.
Musyl, M. K., R. W. Brill, D. S. Curran, J. Gunn, J. R. Hartog,
R. D. Hill, D. W. Welch, J, P Eveson, C. H. Boggs, and
R. E. Brainard.
2001. Ability of archival tags to provide estimates of geo-
graphical position based on light intensity. In Electronic
tagging and tracking in marine fisheries (J. R. Sibert and J.
L. Nielsen, eds. ), p. 343-367. Kluwer Academic Pubhsher,
Netherlands.
Nakamura, H.
1965. Tuna resources of the world (II), 52 p. Japan Fisher-
ies Resources Conservation Association, Tokyo, Japan. [In
Japanese.]
Ogawa, Y., and T. Ishida
1989. Hydrographic conditions governing fluctuations in the
catch of Thunnus thynnus by set-nets along the Sanriku
coast. Bull. Tohoku Reg. Fish. Res. Lab. 51:23-39. [In
Japanese.]
Okachi, I.
1963. Studies on the distribution and structure of the fish
fauna in the Japan Sea by — catch statistics. II. Supple-
ment of seasonal distribution and fishing condition of the
bluefin tuna. Bull. Jap. Sea Reg. Fish. Res. Lab. 11:9-21.
[In Japanese.]
Okiyama, M.
1974. Occurrence of the postlarvae of bluefin tuna, Thun-
nus thynnus, in the Japan Sea. Bull. Japan Sea Reg. Fish.
Res. Lab. 25:89-97. [In Japanese.]
Orange, C. J, and B. D. Fink.
1963. Migration of a tagged bluefin tuna across the Pacific
Ocean. Calif. Fish Game, 49:307-309.
Radovich, J.
1961. Relationships of some marine organisms of the north-
west Pacific to water temperatures particularly during
1957 through 1959. Fish Bull. Calif Dep. Fish Game 112:
1-62.
Shingu, C, I. Warashina, and N. Matsuzaki.
1974. Distribution of bluefin tuna exploited by longline fish-
ery in the western Pacific Ocean. Bull. Far Seas Fish. Res.
Lab. 10:109-140. [In Japanese.]
Smith, P. and D. Goodman.
1986. Determining fish movements from an "Archival" tag:
precision of geographical positions made from a time series
of swimming temperature and depth. U.S. Dep. Commer.,
NOAA-TM-NMFS-SWFC-60, 13 p.
Stevens, E. D., and F. G. Carey
1981. One why of the warmth of warm-bodied fish. Am. J.
Physiol. 240:R151-R155.
Sund, P N., M. Blackburn, and F Williams.
1981. Tunas and their environment in the Pacific Ocean: a
review. Oceanogr. Mar. Biol. 19:443-512.
Ueyanagi, S.
1969. Observations on the distribution of tuna larvae in the
Indo-Pacific Ocean with emphasis on the delineation of the
spawning areas of albacore, Thunnus atalunga. Bull. Far
Seas Fish. Res. Lab. 2:177-256. [In Japanese.]
534
Fishery Bulletin 101(3)
Welch, D. W., and J. P. Eveson
1999. An assessment of light-based geoposition estimates
from archival tags. Can. J. Fish. Aquat. Sci. 56:1317-1327.
Yabe, H., S. Ueyanagi, and H. Watanabe.
1966. Studies on the early life history of bluefin tuna Thun-
nus thynnus and on the larvae of the southern bluefin tuna
T. maccoyii. Rep. Nankai Reg. Fish. Res. Lab. 23:95-129.
(In Japanese.)
Yamanaka, H.,
1982. Fishery biology of the bluefin tuna resource in the
Pacific Ocean, 140 p. Japan Fisheries Resources Conser-
vation Association, Tokyo, Japan. |In Japanese.]
Yokota, T., M. Toriyama, F. Kanai, and S. Nomura.
1961. Studies on feeding habit of fishes. Rep. Nankai Reg.
Fish. Res. Lab. 14:1-234. (In Japanese.]
Yonemori, T.
1989. To increase the stock level of the highly migrated
pelagic fish. In Marine ranching (Agriculture, Forestry
and Fisheries Research Council secretariat, eds.), p. 9-59.
Koseisha-Koseikaku, Tokyo, Japan. Jin Japanese, the title
was translated by authors.]
Yorita, T.
1976. Bluefin tuna in the Sea of Japan in Hokkaido. Monthly
report of Hokkaido Fishery Experimental Station, 33(3):
2-11. [In Japanese.]
Yukinawa, M., and Y. Yabuta.
1967. Age and growth of bluefin tuna, Thunnus thynnus
(Linnaeus), in the north Pacific Ocean. Rep. Nankai Reg.
Fish. Res. Lab. 25:1-18. [In Japanese.]
535
Abstract — We investigated the migra-
tion and behavior of young Pacific blue-
fin tuna {Thunnus orientalis) using
archival tags. The archival tag mea-
sures environmental variables, records
them in its memory, and estimates
daily geographical locations based on
measured light levels. Of 166 archival
tags implanted in Pacific bluefin tuna
that were released at the northeastern
end of the East China Sea from 1995 to
1997, 30 tags were recovered, including
one from a fish that migrated across the
Pacific. This article describes swimming
depth, ambient water temperature, and
feeding frequency of young Pacific blue-
fin tuna based on retrieved data. Tag
performance, effect of the tag on the
fish, and horizontal movements of the
species are described in another paper.
Young Pacific bluefin tuna swim
mainly in the mixed layer, usually near
the sea surface, and swim in deeper wa-
ter in daytime than at nighttime. They
also exhibit a pattern of depth changes,
corresponding to sunrise and sunset,
apparently to avoid a specific low
light level. The archival tags recorded
temperature changes in viscera that
appear to be caused by feeding, and
those changes indicate that young
Pacific bluefin tuna commonly feed at
dawn and in the daytime, but rarely at
dusk or at night. Water temperature
restricts their distribution, as indicated
by changes in their vertical distribution
with the seasonal change in depth of the
thermocline and by the fact that their
horizontal distribution is in most cases
confined to water in the temperature
range of 14-20°C.
Swimming depth, ambient water temperature
preference, and feeding frequency of young
Pacific bluefin tuna (Thunnus orientalis)
determined with archival tags
Tomoyuki Itoh
Sachiko Tsuji
National Research Institute ol Far Seas Fisheries
5-7-1 Shlmizu-Ondo, Shizuoka
Shizuoka, 424-8633, Japan
Email address (for T Itoh): ltou(3'fra.affrc,go.|p
Akira Nitta
Japan NUS Co, Ltd.
Loop-X BIdg.
3-9-15 Kaigan, Minato
Tokyo, 108-0022, Japan
Manuscript approved for publication
22 October 2002 by Scientific Editor
Manuscript received 3 January 2003
at NMFS Scientific Publications Office.
Fish Bull. 101:535-544 (2003).
Swimming behavior oi Thunnus species
and its relation to various environmen-
tal factors have been examined mainly
by acoustic tracking (e.g. Carey and
Lawson, 1973; Laurs et al., 1977; Carey
and Olson, 1982; Holland et al., 1990b;
Cayre, 1991; Cayre and Marsac, 1993;
Block et al., 1997). Acoustic tracking
has also been applied to young Pacific
bluefin tuna iT. orientalis): to one fish
tracked for three hours around Japan
(Hisada et al.'), and to six fish tracked
for several days each in the eastern
Pacific Ocean (Marcinek et al., 2001).
Acoustic tracking can monitor fish move-
ment, behavior, and even physiological
conditions on a time scale of seconds.
However, the duration of monitoring any
one fish is generally limited to several
days at most because of the short life of
the tracking transmitter This limitation,
together with the high cost of adequate
ship-time, generally makes it difficult to
monitor the behavior of a large number
offish over a long period of time.
An archival tag is an electronic device
that measures environmental variables
and records raw or processed data in its
memory. The archival tag can monitor
animal behavior, its physiological con-
ditions, and the several environmental
factors that the animal is actually ex-
periencing, simultaneously. Data can be
collected for a much longer period than
with acoustic tracking, if the tags are suc-
cessfiilly retrieved. Recently, a type of ar-
chival tag that can estimate geographical
locations has been developed. This type
of tag has been applied to southern blue-
fin tima T. maccoyii (Gunn and Block,
2001) and Atlantic bluefin tuna T. thyn-
nus (Block et al., 1998a, 1998b). These
reports show the remarkable value of
the archival tag data for investigating
fish migration and behavior.
We have implanted archival tags in
young Pacific bluefin tuna since 1994
to investigate their migration and be-
havior This article describes the results
obtained from recovered tags and places
special emphasis on vertical swimming
behavior, preferred water temperature,
and feeding frequency. Some insights
regarding vertical swimming depth
have already been reported in Kitagawa
et al. (2000) who used some of the same
data. We have described in another pa-
per (Itoh et al., 2003) the performance
of the archival tag used in the present
study and the characteristics of young
Pacific bluefin tuna migration based on
data from these same tags.
1 Hisada, K., H. Kono, and T. Nagai.
1984. Behavior of young bluefin tuna
during migration. In Progress report
of the marine ranching project 4, p. 1-7.
Nat. Res. Inst. Far Seas Fish. Pelagic Fish
Resource Division, 5-7-1 Shimizu-Orido,
Shizuoka, 424-8633, Japan. (In Japanese,
the title is translated by the authors.]
536
Fishery Bulletin 101 (3)
Materials and methods
The archival tag used in the present study (Northwest
Marine Technology, Inc., Shaw Island, WA) had four sen-
sors— for external temperature, internal temperature,
pressure, and light intensity. The external and internal
temperature sensors had a 0.2°C resolution and response
times of three seconds and 20 seconds, respectively. Resolu-
tion of the pressure sensor was 1 m of depth between the
surface and 126 m, then 3 m down to the scale limit of 510 m.
The tags measured data every 128 seconds.
Two types of data files were created within the tag
memory. One data file stored daily records containing date,
estimated times of sunrise and sunset, water temperatures
at 0 m, plus two other selectable depths (we selected 60 m
and 120 m), and other information required or produced in
the course of location estimates for each day. This file is re-
ferred to as the "summary file" in this article, and it stored
daily data from the time the memory was last cleared.
The second data file contained unprocessed time series
data records taken at 128-second intervals. The tag could
record at any integer multiple of its measurement interval
and a multiple of one was chosen. Each record consisted of
external temperature, internal temperature, pressure, and
light intensity, and corresponded to a known time. This file
is referred to as the "detail file" in this article. It can hold
about 54,000 records, or about 80 days of steady data at the
high rate chosen — a small fraction of the tags lifetime. The
time-series memory was divided into two sections, and the
size allocation between the two sections was determined
by the user The first section filled first and did not change
thereafter. The second section filled next, but once full, it is
then continually overwrote old data. Thus the first section
always contained the earliest data seen in a tag deploy-
ment and the second always contained the latest data. We
divided the detail file into two 40-day sections for releases
in 1995 and 1996, and into 20- and 60-day sections for
releases in 1997. Most of the analyses in this study were
conducted with the detail file.
Prior to experiments on wild fish, we applied archival
tags to three pen-held Pacific bluefin tuna from 93 to 97
cm in fork length ( FL I at Kasasa in Kagoshima Prefecture
(31°25'N, 130°11'E) in November 1994 to observe the effect
of archival tag attachment and implantation on fish. One of
the fish that had an archival tag inserted in its abdominal
cavity was recovered 453 days after tag implantation, when
the fish was caught for sale in the market.
Experiments on wild young bluefin tuna were conducted
near Tsushima at the northeastern end of the East China
Sea every November and December from 1995 to 1997. A to-
tal of 166 fish, ranging from 43 to 78 cm FL (age 0 or 1 ), were
caught by chartered trolling vessels, tagged on the vessel by
inserting archival tags into their abdominal cavities, and re-
leased immediately. Details of the tag, its performance, and
the manner of tagging are described in Itoh et al. (2003).
Thirty of the 166 archival tags (18.1%) were recovered.
The durations of the tags at sea were 50 days or less for 13
fish, 96-211 days for 13 fish, and 359-375 days for three
fish, all recaptured around Japan. One additional fish was
recaptured off the west coast of Mexico, on the east side of
the Pacific Ocean, 610 days after release. Data could not be
downloaded from one archival tag released in 1995 and re-
covered 30 days after release; all other tags yielded data.
Results
Sample records of swimming depth, water temperature,
and temperature offish viscera as recorded in the detail file
are shown in Figure 1 for three days in winter and three
days in summer The fish changed swimming depth fre-
quently with rapid dives and ascents. The water tempera-
ture changed little in winter, but it changed frequently and
substantially corresponding to dives in summer. Visceral
temperature changed gradually.
Diurnal and seasonal change of swimming depth
Differences between daytime and nighttime swimming
depth, water temperature, visceral temperature, and the
temperature difference between water and fish viscera
(thermal excess) were examined by using the Mann- Whit-
ney Utest (P=0.05) with data in the detail file. Data taken
during the hour before and the hour after both dawn and
dusk (four hours in all) were excluded from the test in
order to distinguish clearly between daytime and night-
time. For this purpose, dawn and dusk were taken as the
times the tag sensed the first or final light of the day. The
average swimming depth was significantly deeper during
daytime for 70% of all recorded days, which accounts for
the additional observation that the water temperature was
significantly lower during daytime for 66% of the days. The
visceral temperature was significantly higher during day-
time for 71% of the days, and thermal excess was signifi-
cantly larger during daytime for 85% of the days.
Fish spent about 40% of their time within a 0-9 m depth
range and the time spent within each depth interval de-
creased as depth increased. This concentration in the 0-9 m
depth range was observed at both daytime and nighttime,
but was more pronounced at night (Fig. 2).
The vertical thermal profiles (Fig. 2) show the change of
depth range of the surface mixed layer, the ocean layer that
lies above the seasonal thermocline, for one year Although
young Pacific bluefin tuna aggregated in the 0-9 m depth
range for almost all months, swimming depth was more
broadly distributed in winter when the depth range of the
mixed layer was greater (e.g. January and March). When
the depth range of the mixed layer became less in summer
(e.g. May and July), fish tended to concentrate near the
sea surface. Then as the depth range of the mixed layer
became greater in autumn (e.g. September and November),
the vertical distribution of the tuna gradually expanded
toward deeper water.
Vertical swimming behavior at dawn and dusk
The fish commonly showed a distinctive vertical movement
at dawn and dusk. At dawn, after a slow and steady descent
for about 40 minutes to reach to a maximum depth of 82
±28 m (average ±SD), fish suddenly and rapidly ascended
Itoh et al.: Swimming depth, water temperature preference, and feeding of Thunnus onentalis
537
12 16
Time of day (h)
Figure 1
Sample records of swimming depth (thick solid line), water temperature ( lower thin solid line), and temperature of viscera
(upper thin solid line) for Pacific bluefin tuna from the detail files of two archival tags. Upper panel: records of a fish from
23 to 25 December 1995 in the East China Sea. Lower panel: records of another fish from 15 to 17 June 1997 in the East
China Sea. Shadows indicate nighttime.
to near the sea surface (upper panel in Fig. 1, and Fig. 3).
At dusk, after a rapid descent from near the sea surface to
the maximum depth of 89 ±34 m, fish slowly and steadily
ascended for about 40 minutes. The time that maximum
depth was recorded was most frequently at four minutes
before (or after) the time when the archival tag sensed the
first (or final) light at dawn (dusk) for 29% (37%) of cases
where this behavior was observed. The end (onset) of rapid
ascents (descents ) occurred most frequently at the time when
the tag sensed the first (final) light, that is to say for 68%
(49%) of all cases. The light detection threshold for the tags
used in our study corresponded to an intensity in the blue-
green transmission window of seawater (470 nanometers
(nm) center, 60 nm width of the filter) of about 3 x 10"'' times
the surface noon solar intensity on a clear day in that same
spectral band. Because the light-sensitive region of the
measurement stalk was located below the animal's body,
we assumed that the eyes encountered a somewhat higher
intensity than did the light sensor. Quick examination of
tags in air in the early October revealed that the time when
the archival tag sensed the first light at dawn or the final
light at dusk was about 40 minutes before sunrise or about
40 minutes after sunset, respectively. Some combination of
these vertical swimming behaviors was observed in 1081
of 1452 days (74%) during which the detail file data were
available for both dawn and dusk. They occurred at both
dawn and dusk in 679 days (47%), only at dawn in 77 days
(5%), and only at dusk in 325 days (22%).
Occurrences of these behaviors differed by area and sea-
son (Table 1). The area was determined by using fish loca-
tions estimated as described in Itoh et al. ( 2003 ). The average
occurrence of these depth excursions at dawn and dusk was
as high as 87-88% in the East China Sea from November
to January but decreased to 15-49% from February to June
and October. In the Sea of Japan, the average occurrence
was 4-39% in April, May, and September to November. In
the Pacific Ocean, the behaviors were observed in 75-90% of
days from May to July, in contrast to the relatively low occur-
rences within the East China Sea in the same season.
Water temperature
Water temperatures recorded in archival tags ranged from
8.3° to 28.4°C at 0 m depth, and 1.4° to 28.4°C when all
depths were combined. A range of water temperatures
that appeared to be preferred by young Pacific bluefin tuna
was estimated. A simple frequency distribution of recorded
water temperatures was inappropriate because fish could
be forced to tolerate water out of their preferred tempera-
ture range because of a lack of water with a more favorable
temperature within the geographical range accessible to the
fish. Instead we compared the range of water temperature
within the geographical area accessible to the fish with the
frequency distribution of actual recorded water tempera-
tures, grouped by year, month, and sea. We assumed that
the accessible areas in each sea were those areas that a
tag reported as its position. These include areas along the
Japanese coast between 35° and 45°N in the Sea of Japan,
30-45°N in the western Pacific Ocean west of 160°E, and
25-45°N in the eastern Pacific Ocean east of 160°E. In
538
Fishery Bulletin 101(3)
B 10
a.
a 90%
9C
%
0% 90
(1
1
|J
1
January
50
100
1
15 20 25 10
0% 90% 90%
0
1 ■
>» 1
50
/
July
00
^
50
100
9
0% 0%
9C
n
1 t
1 1
i
■
/
March
50
' 1
1 *
00
f
15
20
25
90%
September
10 15 20 25 10 15 20 25
Temperature (°C)
90%
50
100
10
90%
10
0%
90%
May
25
90%
0
1
1
1
1
1
1
1'
1
\
^ i
50
1
9 November
00
' 1 '
' 1
15
20
25
Figure 2
Vertical occurrence frequencies of young Pacific bluefin tuna in daytime (clear barl and nighttime (shaded bar) and
vertical thermal profile (line with dots) by month.
120 ->
Time (minutes)
100
120
Figure 3
Average swimming depth of young Pacific bluefin tuna recorded by archival tags at dawn (left panel) and
dusk (right panel). Times marked as 0 are the time just after the end of rapid ascent at dawn and just before
the onset of rapid descent at dusk. All of records in which a rapid ascent or rapid descent occurred were
averaged. Bar shows standard deviation.
the East China Sea, the area considered was enclosed by
four points of 29°N-126°E, 29°N-128°E, 35°N-130°E, and
33°N-126°E, where the majority of estimated tag locations
occurred. The temperature range in each area was derived
from the sea-surface temperature maps published by the
Japan Fisheries Information Service Center.
Itoh et al.: Swimming depth, water tempeiature preference, and feeding of Thunnus orientalis
539
Table 1
The number of days
during which swimming
behaviors with rapid ascent at dawn and rapid descent at dusk were
observed by area
and month.
Number
Dawn
Dusk
Average
Total
Observed
Observed
Total
Observed
Observed
Area
Month
offish
days
days
rate
days
days
rate
observed rate
East China Sea
1
7
87
73
84%
86
77
90%
87%
2
1
23
6
26%
22
16
73%
49%
3
2
36
11
31%
37
12
32%
31%
4
7
66
24
36%
67
41
61%
49%
5
10
220
51
23%
221
128
58%
41%
6
8
182
17
9%
176
37
21%
15%
10
1
24
4
17%
24
6
25%
21%
11
23
132
109
83%
153
142
93%
88%
12
29
487
388
80%
489
457
93%
87%
Sea of Japan
4
1
13
0%
12
1
8%
4%
5
1
9
1
11%
9
1
11%
11%
9
1
12
0%
12
4
33%
17%
10
2
61
10
16%
61
22
36%
26%
11
2
43
15
35%
42
18
43%
39%
Pacific Ocean
5
1
16
11
69%
16
13
81%
75%
6
2
26
19
73%
26
25
96%
85%
7
1
31
25
81%
31
31
100%
90%
Total
1468
764
52%-
1484
1031
69%'
61%
The tag temperature used in our comparison was the
temperature at 0 m depth recorded in the summary file
because that file contained a much larger number of days
than that of the detail file. In support of this, we confirmed
that the average water temperature over all depths for a
day in the detail file had only slight differences of -0.1
±0.7°C in average (range: -4.0-+4.3°C) from the tem-
perature at 0 m recorded for the day in the summary file.
Because the fish swam near the surface, the temperature
recorded to represent 0 m depth also represented well the
temperature at all depths where the fish actually swam.
Frequencies of days were summed from the data for all
individuals in one-degree temperature bins separated by
year, by month, and by area, such as the East China Sea,
the Sea of Japan, and the Pacific Ocean.
The water temperature recorded by archival tags com-
monly ranged from 14° to 20°C (Fig. 4). When water of this
temperature range was located within an accessible area
(e.g. many months in the East China Sea and the Pacific
Ocean, and November in the Sea of Japan), almost all fish
were found in such water. Where water temperature was
higher (e.g. June 1996, June 1997, and between June and
October 1998 in the East China Sea) or lower (e.g. May
1996 and April 1997 in the Sea of Japan), fish tended to
choose water that was close to this range. However there
were a few cases in which fish stayed in water with a
temperature outside of this range (e.g. between July and
September 1998 in the Sea of Japan) even when water of
the 14-20°C range was accessible to them.
Visceral temperature and feeding events
Temperature of the fish viscera ranged from 13.0° to
30.9°C. It was usually higher than the water temperature
and the thermal excess for a given individual ranged from
1.3° to 4.6°C, and averaged 3.0° ±1.0°C. The visceral tem-
perature changed in parallel with the water temperature
all year (Fig. 5).
In preliminary experiments, pen-held fish were fed com-
pletely thawed mackerel of 20-30 cm FL twice a day at
approximately 900 h and 1500 h. The following changes in
visceral temperature before and after feeding were generally
observed (Fig. 6). All figures given below regarding thermal
excess and timing after feeding are average values. The vis-
ceral temperature in a stable state just before feeding (n=5)
was 3.7°C higher than ambient water temperature. This
thermal excess decreased to 2.3°C at 22 minutes sifter feed-
ing, and then increased to 7.6°C at 7.7 hours after feeding.
Then it slowly decreased again and reached a stable thermal
excess of 3.1°C at 21.0 hours after feeding. When the fish
fed again before the visceral temperature stabilized (n=7),
with the thermal excess still as high as 6.8°C, the thermal
excess reached a high of 8.5°C but the increase after feeding
was 1.7°C, much smaller than that observed in the previous
case (3.9°C ). The time required to reach the highest visceral
temperature and the time to change back to a stable state
were similar to the previous case. When the fish were not fed,
because of rough sea conditions, the visceral temperature
stayed stable all day (e.g. 16 January on Fig. 6).
540
Fishery Bulletin 101(3)
Area Year-
Number
Temperature (°C)
month of tish
5 6
718
9
10
11 12 13
14
15
16 17 18
19 20
21
??
23 24 25 26 27 28 29 30
'95-12 7
3 68 49
1
•96-1
4
9
20 29 24
13 7
■96-2
3
1 7
25
19
14 9 8
2 3
■96-3
3
1 , 5
18
23
17 1 3
1
■96-4
2
3
15
14
11 11 5
1
1
■96-5 2
1
3
2 9 4
6 6
3
■96-6 ' 1
' '
2 10
16 il ;
1
■96-11 12
,
! 2i 1
7 14
■96-12 12
1' 17 94
197 17
•97-1 1 8
2
5
17
48 86 67
1
•97-2 1 7
3
17
91
44 32 9
S ■97-3 7
2
82
59 35 35
4
"" ■97-4 7
63
73 22 37
1 1
i ■97-5 5
11 95
32
6 ■97-6 4
24 24
22
23
5 1
1 ^97-11 10
1 14 70
15
i" '97-12 10
1 12 122
122 18
i 1
■98-1 B
1
21
40 SO 70
23 10
4
J
■98-2 7
10
77
57 31 11
4
\ \ 1
■98-3 6
2
39
75 44 23
3
•98-4 6
1
11 37 65
64 2
'98-5 6
,
3 28
82 54
15
4
■98-6 6
14 53
35
14
13 1 , : 1
■98-7 2
1
3
7 15 15 2i ;
■98-8 1
3
3 8 9 3 3 2
■98-9 1
4 12 11 3
■98-10 1
2
3
8
13 5i : 1
■96-11 1
2
2
6
3 4
7
1
3 11 i 1
■96-5 1
. 1 • : 2l 2| 2
2|
. '97-1 1 1
11 21 2 1 3
-2]
1
1
g. '98-6 ' 1
^
5
1
" ■98-7 2
1 1 I
4
12
18
8 4 4 1
o ■98-8 2
1 1 1 1 1
1 26 15 16 3 1
1 ■gs-g 2
1 4
15 9
1 1 13 15 1
■98-10 2
1
5
8 15 10
6 1
13
2
1
■98-11 2
1
4
5
2 3 11
6 6
1 2
■96-3 1
1 ' .
1
2
3 8 4
2 5
■96-4 1
Ill
2
4
5, 2 12
5
■96-5
1
1 : , i 1 1 1
1
4: 17
9 1
■96-6
1
'III 1 1 1
1 ! 1 15
1
1
■97-5 1
'
111
1
12 11
3 2
2 i!
—
•97-6
!
! 1 ■
1 7
12 10
•97-7
2
2 12
6
9
% '97-8
111
2
9 8 6
2 4
S '97-9
13
2
4
8 4 9
;
u '97-10
^ '97-11
-^-
1 1 ' 11 9
8
5
6 1 1
-^ 1' 9 7 2l
2
1
1 7
S. '97-12
3 5
2
1
8 9 3
1
'98-1
2
7,
9
5, 61 2
■98-2
1 14
9
1
i: 3i
1
■98-3
' 20i 10 r
'•
■98-4
8 19
2 1
■98-5
2 7
9
2
1 1 2
4 3
'98-6
1 1 1 1 (
4
5 2 11
7 1
1 i i
'98-7
1 1 1 1 I
1
1 4
10 12
3 III
Figure 4
Frequencies of sea-surface water temperatures for young Pacific bluefin tuna with archival tags.
The number given is the sum (over all individuals) of days in the category based on temperature
at 0 m depth from the summary files. Gray, thick bars indicate the sea-surface temperature range
that appeared to be accessible to the fish in each month.
In the detail files obtained from wild fish, the following
three types of visceral temperature changes were identified
as indicators of feeding ( Fig. 7 1. Type A: a sharp decrease of
1-2°C followed by an increase above the initial temperature.
Type B: an increase above the initial temperature, preceded
by either no decrease or a slow decrease. Type C: a sharp
decrease of 1-2°C followed by a return to the initial temper-
ature. In addition, in only cases where the visceral tempera-
ture changes could not be explained by water temperature
changes were they counted as being caused by feeding. For
example, a slow decrease of visceral temperature when the
fish dived into cold water was not counted as a feeding event.
Cases where the water temperature changed frequently
were also excluded as too difficult to interpret.
Feeding events were observed in 942 days out of 1494 to-
tal recorded days (63% ). Because it was found that fish did
not feed normally for approximately the first 30 days after
release (Itoh et al., 2003), data from the first 60 days after
release were excluded in the following description. After
exclusion of data for the first 60 days, feeding events were
observed in 726 out of the remaining 807 recorded days
(90%). The number of daily feeding events ranged from
zero to ten, with an average of 1.8 ±1.4. Feeding events oc-
curred most frequently in the daytime (69% of all obser\'ed
feeding events more than 60 days after release), followed
by dawn with 27% (Fig. 8). Feeding events rarely occurred
at dusk and at night, accounting for only 1% and 3%^ of the
total feeding events, respectively. For this analysis, dawn
Itoh et al.: Swimming depth, water temperature preference, and feeding of Thunnus onentahs
541
and dusk were defined as explained above, a one-hour pe-
riod centered on the time of first or last detected daylight.
Changes in visceral temperature of "type B" (increase only)
were observed most frequently (51.4% of total observed
feedings), followed by "type C" (decrease only, 45.2%). Bi-
polar events (type A) were as few as 3.5%.
When averaged over individual months, the number of
feeding events per day per individual ranged from 0.9 in
January to 2.2 in June and averaged 1.5. Feeding events
were observed all year, although they were slightly more
frequent in May and June than in other months (Fig. 9).
Discussion
Diurnal and seasonal change of swimming depth
Young Pacific bluefin tuna were previously assumed to
swim near the sea surface based on the fact that most of the
catch was made by surface fishing gear and fish
schools were observed at the sea surface (Yabe et
al., 1953). However, details of their vertical swim-
ming behavior and relationships between their
behaviors and oceanic structures have not been
well investigated. Recently, Marcinek et al. (2001)
observed during an acoustic tracking experiment
over several days that Pacific bluefin tuna in the
eastern Pacific Ocean spent the majority of their
time in the top portion of the water column. Our
archival tag data showed that young bluefin
tuna in the western Pacific Ocean also ordinarily
stayed within the surface mixed layer and most
frequently near sea surface, regardless of the
time of day or the season. The vertical distribu-
tion of fish changed according to the seasonal
change in depth range of the surface mixed layer
and appeared to be controlled by the depth of the
thermocline. Restriction by the thermocline was
also observed for yellowfin tuna (T. albacares) and
bigeye tuna (T! obesus) (Carey and Olson, 1982;
Holland et al., 1990b; Cayre and Marsac, 1993;
Block et al., 1997). Occasionally young Pacific
bluefin tuna dived through the thermocline into
deep, cooler water, but they returned to the surface
mixed layer after a short period.
Diurnal change in swimming depth, i.e. deeper
swimming depth during daytime, was reported
by acoustic tracking studies not only for Thunnus
species, such as yellowfin tuna (Carey and Olson,
1982;Hollandetal.,1990b;Cayre,1991;Yonemori-)
and bigeye tuna (Holland et al., 1990b), but also
for other large pelagic species, such as skipjack
tuna, Katsuwonus pelamis (Yuen, 1970; Dizon et
al., 1978); swordfish, Xiphias gladius (Carey and
30
o 25
20
15
10
3 4 5 6
9 10 11 12
Month
Figure 5
Monthly change of average water temperature (O) and
average visceral temperature (•) in young Pacific bluefin
tuna with archival tags. Average values of each individual
were averaged.
^Yonemori.T. 1982. Swimming behavior of tunas by
the use of sonic tags — a study particularly of swim-
ming depth. Far Seas Fish. Res. Lab. Newsletter
44:1-5. Pelagic Fish Resource Division, 5-7-1 Shi-
mizu-Orido, Shizuoka, Shizuoka, 424-8633, Japan.
[In Japanese.]
29 -
27 -
9 25-
i 23-
1 21-
CL
£ 19 -
^ 17-
15 -
0
Viscera
of a pe
Triang
l/n/YV
[
k A
h 12h Oh 12h Oh 12h Oh 12h Oh 12h Oh 12h Oh
12 Jan 13 Jan 14 Jan 15 Jan 16 Jan 17 Jan
Figure 6
il temperature (thick line) and water temperature (thin line)
n-hold young Pacific bluefin tuna recorded by an archival tag.
les indicate the time of feeding.
o
25
20 -
E 15 -
10
C A C C C C
T^'^^^^^rtT-
Oh 2h 4h 6h 8h lOh 12h 14h 16h 18h 20h 22h
Time of day (h)
Figure 7
An example of visceral temperature change in a wild young Pacific
bluefin tuna recorded by an archival tag. Visceral temperature (thick
line) and water temperature (thin line) are shown. Shadows indicate
nighttime. Data are from a fish in the Sea of Japan on 2 May 1996. A,
B, and C indicate the types of visceral temperature changes described
on page 540.
542
Fishery Bulletin 101(3)
Robison, 1981); blue marlin, Makaira nigricans
(Holland et al., 1990a); mako shark, Isiirus oxy-
rinchus; and blue shark, Prionace glauca (Carey
and Scharold, 1990). However some reports did
not note any difference in swimming depth be-
tween day and night, such as that of Block et al.
(19971 for yellowfin tuna, Cayre (1991) for skip-
jack tuna, and Brill et al. (1993) for striped mar-
lin, Tetrapturus audax. The swimming depth of
young bluefin tuna recorded by the archival tags
in the present study was deeper during the day
for lQ'7i of recorded days. This finding agrees with
the speculation made by Carey and Olson (1982)
that a deeper swimming depth in the daytime is a
common feature for large pelagic fish.
Vertical swimming behavior at dawn
and dusk
A characteristic vertical movement pattern was
found in young Pacific bluefin tuna. They dived
gradually and constantly for about 40 minutes
and then rapidly ascended to the sea surface at
dawn. Inverse behavior were observed at dusk.
The same behavior was reported in larger size
Pacific bluefin tuna in the eastern Pacific Ocean
and in yellowfin tuna ( Block et al., 1997; Marcinek
et al.. 2001). The percentage of days when this
behavior was observed varied according to the
season and area. This variation was commonly
observed in individuals as well as in the group
as a whole.
Two potential reasons for this behavior were
considered. The first one was avoidance of a spe-
cific range of light intensities. The times of onset
and end of the behavior are apparently related
to the time of sunrise and sunset. Assuming that
young Pacific bluefin tuna dislike a specific light
intensity range, we describe their vertical move-
ments at dawn and at dusk as follows. About 80
minutes before sunrise when the light intensity
at the sea surface reaches a specific value near the lower
boundary of the avoided range, fish begin descending into
water with lower light intensity. About 40 minutes before
sunrise when the light intensity in deep water reaches that
avoided range, the fish rapidly ascend almost to sea sur-
face, and as the light brightens further, gradually expand
into their normal distribution pattern while staying within
the range of water depths where light intensities exceed
the avoided low-intensity range. A possible reason for
avoiding a specific intensity range might be an increased
risk of predation at intensities where tuna see less well
than some predator that hunts visually.
The other potential reason for the characteristic vertical
movements is feeding. It is well known that some small
animals show diurnal vertical migration, i.e. they descend
gradually as the light level increases toward dawn and rise
again at dusk. Young Pacific bluefin tuna following these
species to feed on them would show similar behavior How-
ever, young Pacific bluefin tuna were observed to feed only
0%
10% 20%
Frequency
30% 40%
50% 60% 70%
Dawn
Daytime
Dusk
Nigtittime
DA
□ B
■ C
Figure 8
Frequency of feeding events of young Pacific bluefin tuna by period
within a day. Only data taken more than 60 days after release were
used. A, B, and C indicate the types of temperature changes of viscera
described on page 540.
2 -
1 -
1 2 8 12 10 1
3 3
() <>
1
C) () <>
() <> " O
6 7
Month
10 11 12
Figure 9
Average frequency of feeding events by month for young Pacific blue-
fin tuna. Only data taken more than 60 days after release were used.
The number above each point indicates the number of individuals
contributing to the average. Bars show standard deviations.
at dawn, not at dusk, although the characteristic vertical
behavior was observed at both dawn and dusk. In addition,
rapid ascents and descents at a specific time with respect
to sunrise and sunset could not be explained by the verti-
cal migration behavior of bait species. Therefore, feeding
seems not to be a primary cause of the vertical migration
in young Pacific bluefin tuna.
Generally speaking, fishermen consider dawn and dusk
to be good times for catching Pacific bluefin tuna. The ar-
chival tag records showed that young Pacific bluefin tuna
did not usually feed at dusk, although tag records showed
that fish aggregated very close to the sea surface after
their rapid ascent at dawn and before their rapid descent
at dusk. Judging from this behavior, good fishing seemed
to be caused by a concentration offish near the sea surface
rather than by the feeding activities Moreover, the low
light level at these times would make it difficult for fish
to distinguish between artificial bait with a hook and live
prey.
Itoh et al.: Swimming depth, water temperature preference, and feeding of Thunnus orientalis
543
Preferred water temperature
Water temperature is thought to be one of the most impor-
tant environmental factors controlhng the distribution of
young Pacific bluefin tuna (Sund et al., 1981; Koido and
Mizuno, 1989; Ogawa and Ishida, 1989). Kitagawa et al.
(2000) attached importance to the gradient of water tem-
perature; however Uda (1957) emphasized the absolute
value of temperature, although his study was presumably
for large-size fish. Data from the archival tags indicated
that young Pacific bluefin tuna seemed to prefer to remain
in water of 14-20°C. When there was no accessible water
within this temperature range, the fish tended to stay in
water of a temperature as close as possible to this range.
In addition, archival tags showed that the vertical distri-
bution of young Pacific bluefin tuna was restricted by the
thermocline, even when the temperature below the ther-
mocline was in the tunas preferred temperature range
(14-20°C). These observations support the importance
of water temperature as shown in previous studies and
suggest that both the absolute value and the gradient of
water temperature are important as environmental factors
controlling the distribution of young Pacific bluefin tuna.
Feeding
Visceral temperature of pen-held Pacific bluefin tuna with
archival tags changed in a specific way during feeding.
Stomach temperature changes have also been observed
in pen-held giant bluefin tuna in the Atlantic and in pen-
held southern bluefin tuna (Carey et al., 1984; Gunn et al.,
2001). The cycle of visceral temperature change for young
Pacific bluefin tuna was completed in 21 hours (shorter
than that observed in previous studies of 1.5 to 2 days)
probably due to the smaller size of the fish. Similar vis-
ceral temperature changes were also noted in the records
of archival tags recovered from wild fish, ranging from a
distinct pattern the same as that observed in pen-held fish
(type A) to less distinct ones such as type B or type C, which
were observed more frequently. All of these changes could
be distinguished quite easily from gradual decreases of vis-
ceral temperature when fish dived into cold water There-
fore, we assumed that these three types of temperature
changes were caused by feeding. Temperature changes of
type A could be expected if fish consumed a large amount of
food at one time as they do when fed in a pen. However, wild
fish may seldom have an opportunity for such large meals,
and the apparently more frequent small meals would be
expected to cause the less dramatic visceral temperature
changes of types B or C.
In the present study, a visceral temperature change was
taken to indicate feeding only when that change could not
be explained by a change in water temperature. Also, when
the water temperature changed very frequently, it was dif-
ficult to decide whether water temperature could account
for a feeding event and it was not counted as such. Finally
it is possible that feeding might not cause a recognizable
change in visceral temperature. As a result of these three
factors the feeding frequency estimated in our study might
have been underestimated.
Frequencies of feeding events did not change much over
the year, although there was a slightly higher frequency
in early summer Growth in length of young Pacific bluefin
tuna is known to become slow in winter (Yukinawa and
Yabuta, 1967; Bayliff, 1993). Because fish weight at a
length was constant throughout the year for wild young
Pacific bluefin tuna (Itoh, 2001), food consumption in win-
ter appears not to be used for increasing weight at a length.
We did not reach a conclusion on this question and further
investigation of seasonal change in food items and of the
physiology of tuna is needed.
Acknowledgments
We thank the staff of Marino Forum 21 and the Kagoshima
Fisheries Experimental Station for their cooperation in the
pe-held fish experiment. We also thank troll fishermen,
staff in the Kamiagata Fisheries Cooperative Association,
the Tsushima Fisheries Extension Service, and the Naga-
saki Fisheries Experimental Station for their cooperation
with the experiment on wild fish. We greatly acknowledge
fishermen, consumers, and staff at the Inter-American
Tropical Tuna Commission for their kindness in returning
recovered archival tags bearing the information necessary
for our study. We especially thank J. Gunn at CSIRO for
giving us valuable information about implanting the archi-
val tag in fish. We are also grateful to the staff of Northwest
Marine Technology Inc. and Tanaka Sanjiro Co., Ltd., for
providing us with tags. We would like to thank P. Ekstrom
of Northwest Marine Technology Inc. for his critical review
and help with the English text. We thank our staff in Japan
NUS Co., Ltd., the Suidosya Co., Ltd., and the National
Research Institute of Far Seas Fisheries, and also T.
Kitagawa in the Ocean Research Institute of the Univer-
sity of Tokyo, for their efforts in regard to implantation
of the tags in fish. We gratefully acknowledge S. Kume of
Japan NUS Co., Ltd., N. Baba of Fishery Research Agency,
Z. Suzuki, and Y. Uozumi of National Research Institute of
Far Seas Fisheries for their critical review.
Literature cited
Bayliff, W. H.
1993. Growth and age composition of northern bluefin tuna,
Thunnus thynnus, caught in the eastern Pacific Ocean, as
estimated from length-frequency data, with comments on
trans-pacific migrations. Bull. lATTC 20:503-540.
Block, B. A., H. Dewar, E. V. Freund, C. Farwell, and E. D. Prince.
1998a. A new satellite technology for tracking the move-
ments of Atlantic bluefin tuna. Proc. Natl. Acad. Sci. USA
95:9384-9389.
Block, B. A., H. Dewar, T. Williams, E. D. Prince, C. Farwell, and
D. Fudge.
1998b. Archival tagging of Atlantic bluefin tima {Thunnus
thynnus thynnus). Mar. Tech. Soc. J. 32: 37-46.
Block, B. A., J. E. Keen, B. Castillo, H. Dewar, E. V. Freund,
D. J. Marcinek, R. W. Brill, and C. Farwell.
1997. Environmental preferences of yellowfin tuna {Thun-
nus albacares) at the northern extent of its range. Mar.
Biol. 130:119-132.
544
Fishery Bulletin 101(3)
Brill, R. W., D. B. Holts, R. K. C. Chang, S. Sullivan, H. Dewar,
and F. G. Carey.
1993. Vertical and horizontal movements of striped marlin
(Telrapturus audax) near the Hawaiian Islands, determined
by ultrasonic telemetry, with simultaneous measurement of
oceanic cuiTents. Mar. Biol. 117:567-574.
Carey, F. G., J. W. Kanwisher, and E. D. Stevens.
1984. Bluefin tuna warm their viscera during digestion. J.
Exp. Biol. 109:1-20.
Carey, F. G., and K. D. Lawson.
1973. Temperature regulation in free-swimming bluefin
tuna. Comp. Biochem. Physiol. 44A:375-392.
Carey, F G., and R. J. Olson
1982. Sonic tracking experiments with tunas. ICCAT Col-
lective Volume of Scientific Papers XVII. 2:458-466.
Carey, F. G., and B. H. Robison
1981. Daily patterns in the activities of swordfish, JVT/p/i /as
gladius, observed by acoustic telemetry. Fish. Bull. 79:
277-292.
Carey, F. G., and J. V. Scharold
1990. Movements of blue sharks (Prionace glauca ) in depth
and course. Mar. Biol. 106:329-342.
Cayre, P.
1991. Behavior of yellowfin tuna (Thunnus albacares) and
skipjack tuna (Katsuwonus pelamis) around fish aggregat-
ing devices ( FADs ) in the Comoros Islands as determined by
ultrasonic tagging. Aquat. Living. Resour. 4:1-12.
Cayre, P. and F. Marsac.
1993. Modeling the yellowfin tuna (Thunnus albacares) ver-
tical distribution using sonic tagging results and local envi-
ronmental parameters. Aquat. Living Resours. 6:1-14.
Dizon, A. E., R. W. Brill, and H. S. H. Yuen.
1978. Correlations between environment, physiology, and activ-
ity and the effects on thermoregulation in skipjack tuna. In
The pysiological ecology of tunas (G. D. Sharp and A. E. Dizon,
eds. ), p. 233-259. Academic Press, New York, NY.
Gunn, J., and B. Block.
2001. Advances in acoustic, archival, and satellite tagging
of tunas. In Tuna: physiology, ecology, and evolution (B.
A. Block and E. D. Stevens, eds.), p. 167-224. Academic
Press, San Diego, CA.
Gunn, J., J. Hartog, and K. Rough.
2001. The relationship between food intake and visceral
warming in southern bluefin tuna (Thunnus maccoyii).
Can we predict from archival tag data how much a tuna
has eaten? hi Electronic tagging and tracking in marine
fisheries (J. R. Sibert and J. L. Nielsen, eds.), p. 109-130.
Kluwer Academic Publisher, Netherlands.
Holland, K. N., R. W. Brill, and R. K. C. Chang.
1990a. Horizontal and vertical movements of Pacific
blue marlin captured and released using sportfishing
gear Fish. Bull. 88:397-402.
1990b. Horizontal and vertical movements of yellowfin and
bigeye tuna associated with fish aggregating devices. Fish.
Bull. 88:493-507.
Itoh, T
2001. Estimation of total catch in weight and catch-at-age
in number of bluefin tuna Thunnus orientalis in the whole
Pacific Ocean. Bull. Nat. Res. Inst. Far Seas Fish. 38: 83-
111. |In Japanese.]
Itoh, T, S. Tsuji, and A. Nitta.
2003. Migration patterns of young Pacific bluefin tuna
(Thunnus orientalis) determined with archival tags. Fish.
Bull. 101: 514-534.
Kitagawa, T, H. Nakata, S. Kimura, T. Itoh. S. Tsuji, and A. Nitta.
2000. Effect of ambient temperature on the vertical distribu-
tion and movement of Pacific bluefin tuna Thunnus thynnus
orientalis. Mar Ecol. Prog. Ser. 206:251-260.
Koido, T, and K. Mizuno.
1989. Fluctuation of catch for bluefin tuna (Thunnus
thynnus) by trap nets in Sanriku coast with reference to
hydrographic condition. Bull. Jpn. Soc. Fish. Oceanogr. 53:
138-152. [In Japanese.]
Laurs, R. M., H. S. H. Yuen, and J. H. Johnson
1977. Small-scale movements of albacore, Thunnus ala-
lunga, in relation to ocean features as indicated by ultra-
sonic tracking and oceanographic sampling. Fish. Bull.
75: 347-355.
Marcinek, D. J., S. B. Blackwell, H. Dewar, E. V. Freund,
C. Fai-well, D. Dau, A. C. Seitz and B. A. Block.
2001. Depth and muscle temperature of Pacific bluefin tuna
examined with acoustic and pop-up satellite archival tags.
Mar Biol. 138:869-885.
Ogawa, Y, and T. Ishida.
1989. Hydrogi"aphic conditions governing fluctuations in the
catch ofThunnus thynnus by set-nets along the Sanriku coast.
Bull. Tohoku Reg. Fish. Res. Lab. 51:23-39. [In Japanese.)
Sund, P. N., M. Blackburn and F. Williams.
1981. Tunas and their environment in the Pacific Ocean: a
review. Oceanogr. Mar Biol. 19:443-512.
Uda, M.
1957. A consideration of the long years trend of the fisheries
fluctuation in relation to sea condition. Bull. Jap. Soc. Sci.
Fish. 23:368-372
Yabe, H., N. Anraku, and T Mori.
1953. Scombroid youngs found in the coastal seas of Abu-
ratu, Kyusyu. in summer. Contribution of Nankai Reg.
Fish. Res. Lab. 1:1-10. [In Japanese.]
Yuen, H. S. H.
1970. Behavior of skipjack tuna, Katsuwonus pelaniis, as
determined by tracking with ultrasonic devices. J. Fish.
Res. Board Canada 27:2071-2079.
Yukinawa, M., and Y. Yabuta.
1967. Age and growth of bluefin tuna, Thunnus thynnus
(Linnaeus), in the north Pacific Ocean. Rep. Nankai Reg.
Fish. Res. Lab. 25:1-18. ]In Japanese.]
545
Abstract — Demersal groundfish den-
sities were estimated by conducting a
visual strip-transect survey via manned
submersible on the continental shelf off
Cape Flattery, Washington. The purpose
of this study was to evaluate the statis-
tical sampling power of the submersible
survey as a tool to discriminate density
differences between trawlable and
untrawlable habitats.
A geophysical map of the study area
was prepared with side-scan sonar
imagery, multibeam bathymetry data,
and known locations of historical
NMFS trawl survey events. Submers-
ible transects were completed at ran-
domly selected dive sites located in each
habitat type. Significant differences in
density between habitats were observed
for lingcod (Ophiodon elongatus), yel-
loweye rockfish (Sebastes ruberrimus),
and tiger rockfish (S. nigrocinctus)
individually, and for "all rockfish" and
"all flatfish" in the aggregate. Flatfish
were more than ten times as abundant
in the trawlable habitat samples than
in the untrawlable samples, whereas
rockfish as a group were over three
times as abundant in the untrawlable
habitat samples.
Guidelines for sample sizes and
implications for the estimation of the
continental shelf trawl-survey habitat-
bias are considered. We demonstrate an
approach that can be used to establish
sample size guidelines for future work
by illustrating the interplay between
statistical sampling power and 1)
habitat specific-density differences, 2)
variance of density differences, and 3)
the proportion of untrawlable area in
a habitat.
Manuscript approved for publication
12 February 2003 by Scientific Editor.
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:545-565 (2003).
Demersal groundfish densities in trawlable and
untrawlable habitats off Washington: implications
for the estimation of habitat bias in trawl surveys
Thomas Jagielo
Annette Hoffmann
Jack Tagart
Washington Department of Fish and Wildlife
600 Capitol Way N.
Olympia, Washington 98501-1091
E-mail address (for T Jagielo) lagiethim'dfw wa gov
Mark Zimmermann
National Marine Fisheries Service
7600 Sandpoint Way NE
Seattle, Washington 98115-0070
Despite their utility, trawl surveys C£ui-
not obtain quantitative samples from
rough, rocky habitats, and thus have
a limited ability to sample all habitats
representatively (Uzmann et al., 1977;
Kulbicki and Wantiez, 1990; Krieger,
1993; Gregory et al., 1997). Since 1977,
triennial bottom trawl surveys have
been used to estimate the abundance
of commercially and recreationally
exploited groundfish species in the
continental shelf waters off Washing-
ton, Oregon, and California (Shaw et
al., 2000). The data generated from
these NMFS surveys are often a key
component of groundfish stock assess-
ments which are used to set levels of
acceptable biological catch (ABC) for
selected species (PFMC, 2001). Clearly,
proper interpretation of these survey
data with respect to fish habitat prefer-
ences is an important part of developing
unbiased stock assessments for fisher-
ies management.
In trawl survey methodology, popula-
tion biomass is related to CPUE by the
following equation (Dark and Wilkins,
1994):
S, =-L CPUE.x-
a. \ q
where i
B.
area-depth stratum;
estimated biomass in the ith
area-depth stratum;
A, = total area in the ith stratum;
a, = total area swept during a
standard trawl haul in stra-
tum i;
CPUE, = mean catch per unit of effort
in the ith stratum; and
q = the catchability coefficient of
the sampling trawl.
For this model to be an unbiased esti-
mator of abundance, it is necessary to
assume that the area sampled by the
trawl is representative of the entire
area-depth stratum of interest (i.e.
a, is representative of A,). Validating
this assumption becomes particularly
important where untrawlable habitat
comprises a significant proportion of
the total area assessed, and where
species composition and density vary
between habitats. We shall refer to
error in trawl survey estimates of abun-
dance due to differences in groundfish
density between habitat types as the
trawl-survey habitat-bias.
The trawl-survey habitat-bias may
be substantial on the west coast conti-
nental shelf because of the considerable
spatial extent of untrawlable habitat
in some management regions (Shaw et
al., 2000). It is also widely recognized
that demersal groundfish species com-
position and density can vary consider-
ably by bottom type (Richards, 1986;
546
Fishery Bulletin 101(3)
Matthews and Richards, 1991; Stein et al.,1992;
O'Connell and Carlile, 1993; Gregory et al., 1997;
Krieger and Ito, 1999; Nasby, 2000; Yoklavich et al.,
2000). Thus, there is considerable interest in evalu-
ating alternative survey tools. 48°30'
One alternative to trawl surveys that has gained
increased attention in recent years is the method of
direct observation of the seafloor, typically conducted
with a remotely operated vehicle (ROV) or with an
occupied submersible (Auster et al., 1989; Krieger, 48°00'
1993; Caimi et al., 1993; Adams et al., 1995; Gregory
et al., 1997; Nasby, 2000). We evaluated the sampling
power of the benthic video-strip transect method, us-
ing videotapes of the sea floor collected in situ with
an occupied submersible. Our goal was to judge the
feasibility of using this approach to provide mean-
ingful comparisons of demersal groundfish densities
between trawlable and untrawlable habitats on spa-
tial scales large enough to be useful for west coast
fisheries management.
We prepared a geophysical map of the bottom and
conducted a submersible survey at a study site located on
the continental shelf off Cape Flattery, Washington. Our
objective was to provide guidelines on sample sizes (num-
ber of submersible transects) that would be needed to
characterize differences in density between the two habitat
types, and specifically, sample sizes that would be needed
to estimate the trawl survey habitat bias in subsequent
studies designed to cover wider geographic areas. The
study was structured to answer the following questions: 1)
what species occupy trawlable and untrawlable habitats off
Washington; 2) what magnitude of density differences can
be expected between trawlable and untrawlable habitats;
3) what is the variability offish density within each habitat
type; and 4) what sample sizes are required to estimate
density differences between habitats, and the trawl survey
habitat bias, in a statistically reliable manner. Our focus
was on the benthic species and species groups that could
be assessed reliably with our submersible survey method;
primarily rockfish (Sebastes spp.), lingcod (Ophiodon elon-
gatus), and flatfish (Pleuronectiformes).
Materials and methods
Study site
Selection of the study site was aided by examination of
historical NMFS trawl survey records and Washington
Department of Fish and Wildlife (WDFW) trawl fishery
logbook data. We chose a rectangular area west of the
Point of Arches, Washington, which extends from the Juan
de Fuca Canyon in the east (125°17'W) to Nitinat Canyon
in the west ( 125°37'W) and ranges from 48°13' in the south
to48°16' in the north (Fig. 1). We selected this area because
1) this portion of the Washington coast has been the site of
a productive groundfish fishery since the 1940s (Alverson
1951), 2) this location has been surveyed tri-annually since
1977 as part of the NMFS west coast shelf survey, 3) the
area has demersal groundfish species of interest, and 4 )
126°00'
125°00'
Stu dy jSit e ;l /
Figure 1
Location of the study area (marked "study site" on map) on the
continental shelf off Washington State.
the area contains both trawlable and untrawlable habitats.
The seafloor of this area was sculpted and shaped by ice
movements during the late Pleistocene period (approxi-
mately 18-20 thousand years ago) and is characterized by
boulder fields resulting from glacial deposition that cover
substantial portions of the area (GoldfingerM. Planning
for the submersible survey required geodetically precise
knowledge of the seafloor characteristics of the study area.
This was facilitated by conducting geophysical surveys and
by preparing a detailed map, which was instrumental to
the submersible survey design.
Geophysical surveys and map preparation
Geophysical surveys of the study site were conducted by
collecting side-scan sonar and multibeam bathymetry
data simultaneously during a five-day effort on board the
USN Agate Passage (YP-697) in May 1998. Slant-range-
corrected side scan sonar data were collected by using a
Waverly widescan 100-kHz system, with a swath width of
800 m. Eighteen parallel track lines were conducted with
100% overlap. The resulting imagery was assembled into
a mosaic map of the bottom relief for a rectangular area
measuring approximately 5.6 by 24.8 km (13,888 hect-
ares). Bathymetric data, with resolution on the order of
±0.4 m were collected with a Reson Model 8101 multibeam
bathymetry system. The multibeam bathymetry data were
processed to produce a detailed map of the bottom topogra-
phy with 1-m depth contour intervals.
Map overlays were prepared that showed the locations
of trawl survey events and trawl fishery tows. Detailed
NMFS records were used to identify the location of various
events associated with historical surveys of the area. The
NMFS survey event types included good hauls, bad hauls,
short hauls (tows ended early because of rough bottom),
' Goldfinger, C. 2001. Personal commun. Department of Geol-
ogy, Oregon State University, Corvalis, OR 97331.
Jagielo et al.: Demersal groundfish densities in trawlable and untrawlable habitats off Washington
547
•*<
J
•
Jt^'
V
■^
4
^
i
(
i
1
V
)
I
^t
-^ s
'^hsn
Figure 2
Geophysical map of the study area with associated overlays. Area outlined in bold is the submersible survey area. The grid squares
are 800 by 800 m sample units. (Top): Mosaic of side-scan imagery of the seafloor (Middle): Depth contours (1-m isobaths) obtained
from multibeam bathymetry. (Bottom): Locations of historical NMFS shelf trawl survey events. Hatched lines = chain drags, stars =
chain snags, unhatched lines = successful tows.
skipped hauls, chain drags, and chain snags. Interviews
with knowledgeable fishermen were also conducted to
establish the locations of known trawling sites within
the area. The resulting geophysical map, with overlays,
provided a geographically accurate reference of the study
area that allowed a priori classification of the bottom into
trawlable and untrawlable habitat types (Fig. 2). The final
map consisted of the following layers: 1) a mosaic of side-
scan imagery of the bottom (Fig. 2, top); high-resolution
depth contours (1-m isobaths) obtained from multibeam
bathymetry (Fig. 2, middle); and 3) locations of historical
NMFS trawl survey events (Fig. 2, bottom).
Experimental design
Our survey design process made use of the detailed map
of the study area for 1) definition of the sampling unit, 2)
classification of all sampling units as trawlable or untraw-
lable habitats, and 3) specification of the in situ survey
area. A sample of units to be surveyed by submersible was
selected from each habitat type by using computer-gener-
ated pseudo-random numbers.
In defining the size of the sampling unit, we sought to
strike a balance between a spatial scale that was small
enough to have homogeneity but large enough to have
meaning as a trawlable or untrawlable space. We chose
square sample units of 800 by 800 m in size. This size
was smaller than the standard NMFS tow length of about
3,000 m and was well within the order of resolution of the
multibeam bathymetry and side-scan imagery used for
discerning rock outcrops. A grid consisting of the 800 by
800 m sampling units was prepared and overlaid onto the
map of the study area (Fig. 2).
Classification of the sampling units into "trawlable" and
"untrawlable" habitats was facilitated by examination of
the geophysical map of habitat features, together with an
evaluation of historical NMFS trawl survey records. The
survey map layer helped us to interpret the appearance
of trawlable and untrawlable habitat on the bathymetric
and side-scan geophysical map layers. Trawlable bottom
548
Fishery Bulletin 101(3)
4
5
3
2
1
14
13
9
11
15
12
10
8
6
7
16
Figure 3
Schematic diagram of the study area divided into 800 m by 800 m sample units and
classified a priori as untrawlable (shaded) and trawlable (unshaded) habitat types.
The numbered sites represent the eight sample units selected at random from each
habitat type, which were numbered sequentially for the cruise plan.
was inferred from locations with good hauls and unevent-
ful chain drags; untrawlable bottom was inferred from bad
hauls, short hauls, skipped hauls, and chain snags. On the
side-scan mosaic layer, untrawlable locations were typi-
cally darker than surrounding areas, indicating boulder
fields or hard, rocky bottom. Such areas often showed high
bottom relief as evidenced by shadows on the mosaic, and
bathymetric contours that indicated abrupt topographic
features, such as sharp ridges or pinnacles. A sample unit
was classified as untrawlable habitat when 1 ) NMFS sur-
vey events within the unit indicated rough bottom, or 2) the
mosaic or bathymetric layers of the unit resembled other
units that were classified as untrawlable, or 3) a sample
unit of unknown habitat type was completely surrounded
by untrawlable habitat. A sample unit was classified as
trawlable habitat when 1 ) NMFS survey events indicated
successful trawl tows in the unit or 2) when the mosaic or
bathymetric layers of the unit resembled other units that
were classified as trawlable. Our trawlable and untraw-
lable habitat assignments agreed well with information
obtained from knowledgeable fishermen. Each sampling
unit in the entire mapped area was examined visually in
detail according to the above procedure and was classified
accordingly as trawlable or untrawlable habitat.
We selected the eastern portion of the mapped area
for the submersible survey (Figs. 2 and 3t. Our focus was
restricted to this section to minimize the difference in bot-
tom depths between the trawlable and untrawlable areas
as a factor, and for logistical convenience to complete the
most submersible dives possible within our survey budget.
Because the 800 m by 800 m sampling units were too large
to be surveyed in their entirety, we sampled using the strip
transect method at each location. Logistically, this was
accomplished by conducting 2-3 ncmoverlapping passes
across the sampling unit and by pooling these segments
together to form a single transect for analysis.
Submersible survey
We used the submersible Delta to conduct the fish survey
with the support vessel FV Auriga in July of 1998. The
Delta is 4.7 m long, accommodates one observer and one
pilot, and has a maximum operating depth of 365 m. An
acoustic Trak- Point system was used with differential GPS
and WinFrog navigational software (Thales GeoSolutions
(Pacific), San Diego, CA) to track and log the position of
the submersible from the support vessel. The Delta was
equipped with halogen lights, external video cameras, an
external Photosea 35-mm camera with strobe, and a Pisces
Box data-logging system that recorded 1 ) the time of day, 2)
depth of the submersible, 3) its distance from the bottom,
and 4) sea temperature at 5-second intervals. Strip tran-
sects were conducted 1-2 m off bottom at a cruising speed
of approximately 2.5 km/h. All dives were made during
daylight hours.
To quantify fish density, each strip transect was docu-
mented with a high 8-mm video camera mounted exter-
nally on the bow of the Delta, and pointed forward. The
camera was equipped with two parallel lasers, spaced 20
cm apart, which were used for estimating the area that was
swept. The scientific observer onboard the Delta verbally
annotated the videotape record with observations taken
through the submersible viewing ports, to help identify fish
and interpret the videotapes during subsequent analysis.
The high 8-mm tapes were copied to S-VHS format to
facilitate videotape analysis. The transect area that was
swept (m-) was estimated as the product of average area
swept per second (m-/min) and the total transect duration
in minutes (see Appendix I for details). The average area
that was swept per second (m'^/min) was determined from
a set of 30-second samples randomly selected from the
transect. On average, approximately 29'7r of each transect
was subsarapled in this manner. Bottom habitat type was
also visually characterized for the transect subsamples.
Following the method of Stein et al. (1992) and using the
classification criteria developed by Greene et al. ( 19991, we
categorized bottom microhabitat type (mud, pebble, cobble,
boulders, and rock ridge) as primary (at least 50% of the
area viewed) or as secondary (>20% of the area viewed).
The bottom-type measurements observed directly in the
transect subsamples were expanded to estimate microhabi-
tat coverage for each transect.
Jagielo et al.: Demersal groundflsh densities in trawlable and untrawlable habitats off Washington
549
Fish were enumerated by identifying and counting only
those fish observed in the lower portion of the video moni-
tor screen (counting area), below the imaginary line con-
necting the laser spots. Lighting and visibility was greatest
in this zone, and we assumed that the probability of observ-
ing and counting fish in this portion of the video image was
1009f (i.e. g=l). Afish was counted if any portion of the fish
was visible in the counting area. The distance obsei-ved be-
tween the two laser spots was used as a reference to classify
fish into two size categories; large (>20 cm) and small (<20
cm). Fish were identified to the lowest taxonomic level pos-
sible. We recognized that fish detection and identification
were subject to observer error The variability describing
that error was obtained by conducting a repeat counting
of a sample of transects by the same observer. Additional
validation checks were made between multiple observers.
where Zj^ = the percentile of the unit normal which gives
power;
Zj^ = the percentile of the unit normal for the
significance criterion; for a two-tailed test,
a = a^2/2;
d = the standardized effect size index for the
two-tailed <-test calculated as
L.
^
CD
^
0)
a
H
II
(N
JJ
TJ
41
Oh
s
01
-a
3
CO
e
H
c
S
tn
3
II
W
C
S
a
+j
c
^ OJ
CO ^
'C '>
■ 2 CO
i~ a)
« c-
"S
" b
aj Si
5- c
3 M
= g
CO S
V, '5
o -S
c
•2 «
bo
PQ
01 o
rn r^
00 i> cx> CO CD o t^
O r- CD o -^ -^ 1-H
lo ic -^ CD -^ in lo
t~ O) CO
[> in r^
000(NOOC0005t>
OOOOpOlNO'^O
cbooocio-HO'H
C^OOO'^i-HC-) i-H(N
oooooooooooot-ooio
pOptDOOOOOOOOOO(MO
oooooocDooodocJooo
OOOOOOOOOOOOlO-^t>lO
ooooooooopoooopp
OOOCJcdocDOOOOOOcDOO
OOOOOOOOr-i
ppOOOOOOC>5
OCJOOOOOO
O'HOOO— lOO 0(N
OOOOOOOOlOCOOOO-^CDt-H
pppppOOOOCDOOOOOOIO
OOC?OOOOOOOOOCDOCD»-<
ooooooooooooooooo
pooooooooooooooo
OCDOOcicDOOOc
o o
O 00
o o o o o o o
O -H
00 in
O 'H o ,-i <3 CO
ooooooooiO'^cooinincoco
pppooooooo'^coiO'^a^mt>j
CDCDOOOOOO'-ioC
OOOC^OOCDOOOO
oooooo-^oooo
o o d c> CD d c
lOlO'^COTflOCO'^COCSIlO-^CO'-ICO'*
OOOOOOOO'-HOOOOOOO
pooooooooooooooo
^ CD
d T-i
o o o o o
oooooooooooooooo
pOppppppppTOppOOp
oinoooooo-^ooooooo
pcopppooooocoooooo
d'^ddSciddicidcidicicicidi
IM O O O O
■* o o o o
•^ d <6 d d
CO O ■*
CO p p
ddcoc^Tfi-icD^co"^
O O O O O O "-I
o o o o o o in
d d d d d
Oi-lOOOOOOOOO -HO
o
d
o
o
d
o
d
o
d
o
d
c?
o
d
OOOOOOOO OO i-l
05
o
O O £^
O CO o
d d d
CO
in
in
in
^
^
in
o
o
§ 8
00
CO
o
CD
o
o
o
o
o
o
o
o
o
00
o
o
o
o
o
o
o
o
o
in
00
in
in o
>*
CD
Tf
in
(M
^
o
o
o
CO
o
o
o
o
CO
CO
00
HHHHHHHHDDDDDDDD h^ h
TriotCiO'-i(>icom.-HOjcct^
552
Fishery Bulletin 101(3)
A summary of counts for large (>20 cm) and small (<20 cm)
fish is shown in Table 4. Small flatfish and rockfish were
very difficult to count, often becoming indistinguishable
from the background when the videotape was paused,
and their counts are most likely underestimated. Among
the large fish, "total rockfish" as a group was the most
abundant numerically followed by "total flatfish" as a
group. Of the large rockfish identified to species (Table
5), rosethorn rockfish were the most abundant followed
in order by yellowtail, greenstriped, yelloweye, tiger, and
redstripe rockfish. Unidentified rockfish represented 30%
of the total large rockfish enumerated. Of the large flatfish
identified to species (Table 6), Dover sole were most abun-
dant followed in order by arrowtooth flounder and Pacific
halibut. Unidentified flatfish represented 78% of the total
large flatfish counted. Other individual fish species and
groups identified below the generic classification level were
dominated by eelpout (Zoarcidae), raffish , skates and rays
(Raja), and greenling {Hexagramrnos spp.) (Table 7).
Species composition differed considerably between habi-
tats. The number of individually identified species was 15
in the trawlable habitat, and 18 in the untrawlable habitat
(Table 8). Flatfish dominated in the trawlable habitat, and
rockfish in the untrawlable habitat. Yelloweye, redstripe,
silvergray, and quillback rockfish, as well as greenling
and wolf-eel were observed in the untrawlable habitat
but not in the trawlable habitat. Spiny dogfish (Squalus
acanthias), Pacific cod (Gadus macrocephalus), and salmon
(Oncorhynchus spp) were observed in the trawlable habitat
but not in the untrawlable habitat.
Comparisons of fish densities and variances between
habitat types were made only for fish >20 cm in length
and in taxonomic units where reliable identification and
enumeration could be assured (Table 9). Thus, density com-
parisons were performed at the species level for distinctive
species (i.e. lingcod, yelloweye rockfish, and tiger rockfish),
but were made at the group level for "all rockfish" and "all
flatfish" bwcause of the presence of fish that could not be
identified to individual species within each of these groups.
For all comparisions, tests of homogeneity of variance of
fish density between habitats (//(,: s'^i=s^^) were rejected
using Cochran's test (Winer, 1971 ) ( «=0.05, k=2, df=7), indi-
cating heteroscedastisity (Table 9). Significant differences
in densities between habitats were found for each of the
species and group comparisons using the Mann-Whitney
two-sample test on ranks (Winer, 1971) (a=0.05, 2) (Table
9). Densities were higher in the untrawlable habitat for the
"all rockfish" group, tiger rockfish, yelloweye rockfish, and
lingcod; densities were higher in the trawlable habitat for
the "all flatfish" group.
Statistical power analysis
The validity of our approach for analyzing the statistical
sampling power of the submersible survey depends upon,
among other things, fidelity to the assumptions of the
two-sample t-test of means. The <-test requires that 1) the
two sample means are estimated from random samples
drawn from normally distributed populations, and that
2) the variance of the two populations are equal. Because
Table 3
Common and scientific names
of fishes observed at 16 sub-
mersible dive sites off Cape Flattery, Washington.
Common name
Scientific name
Canary rockfish
Sebastes pinniger
Greenstriped rockfish
Sebastes elongatus
Quillback rockfish
Sebastes maliger
Redstripe rockfish
Sebastes proriger
Rosethorn rockfish
Sebastes
helvomaculatus
Silvergray rockfish
Sebastes brevispinis
Tiger rockfish
Sebastes mgrocinctus
Yelloweye rockfish
Sebastes ruberrimus
Yellowtail rockfish
Sebastes flavidus
Greenling
Hexagramrnos spp.
Lingcod
Ophiodon elongatus
Pacific cod
Gadus macrocephalus
Arrowtooth flounder
Atheresthes stomias
Dover sole
Microstomus pacificus
Pacific halibut
Hippoglossus stenolepis
Spotted ratfish
Hydrolagus colliei
Spiny dogfish
Squalus acanthias
Longnose skate
Raja rhina
Big skate
Raja binoculata
Salmon
Oncorhynchus spp.
Wolf-eel
Anarrhichthys ocellatus
Eelpout
Zoarcidae
Poacher
Agonidae
Generic group classifications
Unidentified rockfish
Sebastes spp.
Unidentified flatfish
Pleuronectiformes
Unidentified roundfish
Osteichthyes
our estimates of variance differed considerably between
habitats (Table 9), we examined the properties of our data
in more detail to confirm the reliability of using the t-test
for our statistical power analysis. We conducted a bootstrap
simulation experiment, in which we compared estimates
of empirical power derived from our study (n=8) with the
estimates of power obtained with Equation 1, under the
assumption of asymptotic normality. The results of this
comparison indicated that estimates of statistical power
obtained from Equation 1 were generally conservative
(indicated lower power) in relation to the empirical esti-
mates of power for simulated known differences in density
(Fig. 4). Given this result, we proceeded with our power
analysis based on the <-test, under the assumption that,
based on our observations, this approach will tend to err in
the conservative direction; that is, it will tend to understate
statistical power
It is evident that, as it becomes necessary to detect small-
er effect sizes, the required sample size increases accord-
ingly. The relationship between sample size (;!=the number
of sample units Isubmersible dive sites] in each habitat
t3TJe) and the effect size-index (d) for density comparisons
Jagielo et al.; Demersal groundfish densities in trawlable and untrawlable habitats off Washington
553
Table 4
Summary offish counts for large (>20 cm) and small (<20 cm) fish for major fish groups. Site type
T = trawlable, U = untrawlable.
Site
Site type
Number of
large fish (>
20 cm)
Number of small fish (< 20 cm
Rockfish
Lingcod
Flatfish
Other Total
Rockfish
Flatfish
Other
Total
4
T
0
1
77
8 86
0
95
48
143
5
T
8
0
54
17 79
0
94
15
109
6
T
2
0
76
12 90
0
68
63
131
10
T
7
3
29
5 44
0
26
107
133
11
T
0
1
35
10 46
0
8
101
109
12
T
0
0
46
5 51
0
6
63
69
13
T
39
1
119
19 178
0
77
37
114
15
T
0
0
31
2 33
0
70
64
134
1
U
115
1
6
16 138
43
0
10
53
2
U
128
14
12
28 182
348
3
52
403
3
U
9
2
28
10 49
41
9
58
108
7
U
43
9
13
22 87
0
21
46
67
8
U
32
3
4
9 48
40
2
12
54
9
U
206
5
6
14 231
339
0
11
350
14
u
30
3
8
11 52
38
7
27
72
16
u
111
5
30
9 155
28
4
17
49
Totals
T
56
6
467
78 607
0
444
498
942
u
674
42
107
119 942
877
46
233
1156
All
730
48
574
197 1549
877
490
731
2098
Table 5
Summary offish counts by site for large rockfish (>20
cm). Site
type: T =
= trawlable, U = untrawlable.
Site
Number offish (>20 cm)
Rose-
Yellow-
Silver-
Green-
Quill-
Red-
Yellow
Site
type
thorn
tail
gray
striped
Canary
back
stripe
Tiger
eye
Unidentified
Total
4
T
0
0
0
0
0
0
0
0
0
0
0
5
T
0
0
0
8
0
0
0
0
0
0
8
6
T
0
0
0
2
0
0
0
0
0
0
2
10
T
0
0
0
0
2
0
0
0
0
5
7
11
T
0
0
0
0
0
0
0
0
0
0
0
12
T
0
0
0
0
0
0
0
0
0
0
0
13
T
2
1
0
14
0
0
0
1
0
21
39
15
T
0
0
0
0
0
0
0
0
0
0
0
1
U
31
1
0
9
0
1
0
0
8
65
115
2
U
88
3
0
0
0
0
0
7
12
18
128
3
U
8
0
0
0
0
0
0
0
1
0
9
7
U
16
14
0
1
0
0
0
2
3
7
43
8
U
25
2
0
1
2
0
0
1
0
1
32
9
u
121
1
1
3
0
0
16
6
5
53
206
14
u
15
10
0
1
0
0
0
1
0
3
30
16
u
34
17
3
0
0
0
0
2
7
48
111
Totals
T
2
1
0
24
2
0
0
1
0
26
56
U
338
48
4
15
2
1
16
19
36
195
674
All
340
49
4
39
4
1
16
20
36
221
730
554
Fishery Bulletin 101(3)
Table 6
Summary of fish
counts by site for large
flatfish (>20 cm)
Site type.
T = trawlable, U =
= untrawlable.
Site
Site type
Number of fish ( >20 cm )
Arrowtooth
flounder
Dover sole
Pacific halibut
Unidentified
Total
4
T
3
6
2
66
77
5
T
3
8
1
42
54
6
T
15
2
6
53
76
10
T
0
3
3
23
29
11
T
1
3
3
28
35
12
T
5
2
5
34
46
13
T
10
13
7
89
119
15
T
0
2
1
28
31
1
U
0
4
0
2
6
2
u
0
2
0
10
12
3
u
0
4
2
22
28
7
u
0
0
5
8
13
8
u
0
0
0
4
4
9
u
0
1
1
4
6
14
u
0
1
0
7
8
16
u
1
0
0
29
30
Totals
T
37
39
28
363
467
All
38
51
36
449
574
Table 7
Summary offish counts by site for other large (>20 cm) fish.
Site type: T
= trawlable, U = untrawlable.
Number offish (>20 cm)
Site
Site
type
Greenling
Pacific cod
Ratfish
Spiny dogfish
Skates/Rays
Eelpout
Salmon
Unidentified
Total
4
T
0
0
0
0
0
8
0
0
8
5
T
0
0
6
1
0
10
0
0
17
6
T
0
0
0
0
1
11
0
0
12
10
T
0
0
0
0
4
1
0
0
5
11
T
0
0
0
0
1
8
1
0
10
12
T
0
2
0
0
1
2
0
0
5
13
T
0
1
1
6
5
5
0
1
19
15
T
0
0
0
0
0
1
0
1
2
1
U
2
0
1
0
0
12
0
1
16
2
U
1
0
1
0
0
26
0
0
28
3
u
1
0
1
0
1
6
0
1
10
7
u
3
0
0
0
4
15
0
0
22
8
u
2
0
2
0
0
5
0
0
9
9
u
2
0
6
0
0
5
0
1
14
14
u
0
0
2
0
0
9
0
0
11
16
u
1
0
5
0
0
3
0
0
9
Totals
T
0
3
7
7
12
46
1
2
78
U
12
0
18
0
5
81
0
3
119
All
12
3
25
7
17
127
1
5
197
between trawlable and untrawlable habitats is shown in
Figure 5. To achieve power of 80% (r<=0.05), the required
number of dives ranges from n = 5 (.d=2.0) to n = 17 (d=1.0);
similarly, to obtain 90'/r power would require 8 to 27 dives.
Empirical estimates of c/ from our study ranged from 1.1
for tiger rockfish to 2.0 for flatfish. This result suggests
Jagielo et al.: Demersal groundfish densities in trawlable and untrawlable habitats off Washington
555
Table 8
Composition offish densities
in trawl ab
e and untrawlable sites by species (>20 cm), ranked in descending order of observed abun-
dance (avg. no./hectare).
Italicized species were not found in the other habitat type.
Tra
wlable sites
Untrawlable sites
Species or group
Avg. no./hectare
Species or group
Avg. no./hectare
Eelpout
11.46
Rosethom rockfish
77.78
Dover sole
9.33
Eelpout
19.26
Arrowtooth flounder
9.25
Yellowtail rockfish
10.70
Pacific halibut
6.88
Lingcod
9.78
Greenstriped rockfish
5.65
Yelloweye rockfish
8.65
Skate
2.81
Tiger rockfish
4.40
Spiny dogfish
1.67
Spotted ratfish
3.90
Spotted ratfish
1.54
Greenstriped rockfish
3.76
Lingcod
1.39
Redstripe rockfish
3.39
Pacific cod
0.70
Dover sole
3.00
Rosethom rockfish
0.48
Greenling
2.67
Canary rockfish
0.41
Pacific halibut
1.77
Salmon
0.28
Skate
1.11
Yellowtail rockfish
0.24
Silvergray rockfish
0.79
Tiger rockfish
0.24
Wolf-eel
Canary rockfish
Quillback rockfish
Arrowtooth flounder
0.49
0.36
0.27
0.19
Generic group
All flatfish
114.29
All flatfish
23.90
All rockfish
13.14
All rockfish
155.63
All fish
146.65
All fish
211.70
that it is relatively more difficult (i.e. more dive sites are
required) to detect density differences between habitats for
tiger rockfish, as compared to flatfish. The associated power
curves for these two values of cf are illustrated in Figure 6.
Figure 6 suggests that, given our observations (for values
of c? as low as 1.1), a sample size guideline of approximately
15 submersible dive sites in each habitat type would yield
approximately an 80% chance of detecting a difference in
mean density at least as large as the random noise esti-
mated in the data for a similarly designed study.
Our statistical power analysis also indicated that, when
the relative proportions of untrawlable and trawlable habi-
tat, as well as the variability in the trawl survey estimates
of abundance, are taken into consideration, the problem of
estimating the trawl survey habitat bias can require sub-
stantially more samples than would be required simply
to compare the density differences between two habitat
types. Values of the trawl-survey habitat-bias effect size-
index (c?^), calculated for a range of untrawlable habitat
proportions with empirical trawl and submersible survey
data, are given in Table 10 and are plotted for rockfish and
flatfish in Figure 7. Using the calculated values of d^ from
Table 10, we derived sample size guidelines for rockfish and
fiatfish (at power=0.80, a=0.05). The resulting relationship
between the sample size required to estimate the trawl
survey habitat bias (the «=number of submersible dive
sites in each habitat type) and the proportion of untraw-
lable habitat in a management area (A^/A) is illustrated
in Figure 8. If, for example, the area of untrawlable habitat
represented 20% of a management unit. Figure 8 indicates
that the sample size required to estimate the trawl survey
habitat bias would be « = 31 for rockfish (d^=0.73), and n =
9 for flatfish (c/^=1.41). Sample sizes for lingcod were much
higher (n>100), owing to the comparatively small detectible
effect size required (d(,=0.13).
Discussion
Our study successfully obtained a first look at the variabil-
ity in groundfish densities in trawlable and untrawlable
habitats for a study area off Washington. We also developed
a framework to use these types of observations to derive
sample size guidelines for designing larger-scale studies
to estimate the trawl survey habitat bias. The limited geo-
graphic scope of our study precludes extrapolation of our
specific results to the west coast at large. However, we dem-
onstrated an approach that can be used to establish sample
size guidelines for future work by illustrating the interplay
between statistical sampling power and 1) habitat-specific
density differences, 2 ) variance of density estimates, and 3)
the proportion of untrawlable area in a habitat.
In our study area, we observed striking differences in
species composition and fish density between the traw-
556
Fishery Bulletin 101(3)
Table 9
Summary of estimated fish densities (no./hectare) and summary statistics for selected fish
U = untrawlable.
groups (>20
cm). Site type: T =
trawlable,
Site
Site type
Estimated fish density (
number/10^ m^)
Rockfish
Flatfish
Lingcod
Yelloweye
rockfish
Tiger rockfish
Mean
SE
Mean
SE
Mean
SE
Mean
SE
Mean
SE
4
T
0.00
0.00
15.16
8.55
0.20
0.28
0.00
0.00
0.00
0.00
5
T
1.39
0.10
9.36
1.26
0.00
0.00
0.00
0.00
0.00
0.00
6
T
0.43
0.63
16.26
3.70
0.00
0.00
0.00
0.00
0.00
0.00
10
T
1.15
1.96
4.78
0.45
0.49
0.02
0.00
0.00
0.00
0.00
11
T
0.00
0.00
7.84
3.54
0.22
0.36
0.00
0.00
0.00
0.00
12
T
0.00
0.00
8.52
1.12
0.00
0.00
0.00
0.00
0.00
0.00
13
T
7.54
9.23
23.01
8.97
0.19
0.40
0.00
0.00
0.19
0.40
15
T
0.00
0.00
6.49
3.13
0.00
0.00
0.00
0.00
0.00
0.00
1
U
25.07
29.36
1.31
0.54
0.22
0.34
1.74
1.79
0.00
0.00
2
U
27.05
19.23
2.54
1.61
2.96
2.49
2.54
2.00
1.48
1.24
3
U
1.61
2.68
5.02
4.63
0.36
0.28
0.18
0.30
0.00
0.00
7
U
7.60
8.83
2.30
1.29
1.59
2.65
0.53
0.88
0.35
0.59
8
U
4.62
0.97
0,58
0.55
0.43
0.06
0.00
0.00
0.14
0.27
9
U
34.92
15.63
1.02
0.12
0.85
0.47
0.85
0.93
1.02
1.19
14
U
6.43
7.41
1.71
1.62
0.64
0.80
0.00
0.00
0.21
0.27
16
U
17.20
7.37
4.65
2.58
0.77
0.51
1.08
0.93
0.31
0.28
Summary statistics
1.31
11.43
0.14
0.00
0.02
s2,
6.64
37.92
0.03
0.00
0.00
"J,,
15.56
2.39
0.98
0.87
0.44
^\
151.58
2.69
0.82
0.81
0.28
Cochran
s test for homogeneity of variance (Winer 1971); C^^„ = 0.83
C 0.96 0.93
0.96
1.00
0.98
Mann Whitney test for equality offish densities (Winer 1971); [/,,^„ =
U 61 63
51
60
56
51
Statistics to calculate effect size index id) for submersible survey power analysis
|m,-mj 14.25 9.04 0.84
0.87
0.42
Sp
8.894
4.51
0.65
0.64
0.38
d
1.6
2.0
1.3
1.4
1.1
labia and untrawlable habitats. Flatfish were more than
ten times as abundant in the trawlable habitat samples,
whereas rockfish as a group were over three times as
abundant in the untrawlable habitat samples. Silvergray,
quillback, redstripe, and yelloweye rockfish were observed
in the untrawlable habitat but not in any of the trawlable
habitat samples.
We know of no visual-transect data comparable to
that presented here for fish abundances off Washington.
However, previous habitat specific studies in other areas
have also reported differences in species composition and
fish densities between low relief (trawlable) and highly
rugose (untrawlable) habitats. Richards (1986) conducted
a submersible study in the Strait of Georgia, British Co-
lumbia (21-140 m), and observed that the distribution of
greenstriped, quillback, and yelloweye rockfish varied by
depth and bottom type. Greenstriped rockfish were most
abundant in fine sediment habitats, such as mud and
cobble terrain. Quillback rockfish were most abundant
in complex habitats, and yelloweye rockfish had higher
densities in wall and complex habitats than in fine sedi-
ment habitats. In the coastal fjord of Saanich Inlet, British
Columbia (21-150 m), Murie et al. ( 1994) also reported that
quillback rockfish density was higher in areas of complex
or wall habitat, compared to areas of sand-mud habitat.
Additionally, tiger, copper (S. caurinus), yellowtail, and
yelloweye rockfish were observed only over complex or
wall habitats, and gi-eenstriped rockfish were seen mostly
over sand-mud habitat. Using sunken gill nets to sample
trawlable and untrawlable habitats off Vancouver Island,
B.C. (198-311 m in depth), Matthews and Richards (1991)
reported difTerences in species composition between traw-
lable and untrawlable areas and higher species diversity
in trawlable habitat. Major species on trawlable bottom
Jagielo et al ; Demersal groundfish densities in trawlable and untrawlable habitats off Washington
557
were Pacific ocean perch (S. alutus), splitnose rockfish
(S. diploproa), greenstriped rockfish, and bocaccio [S.
paucispinis). Major species on untrawlable bottom were
sharpchin (S. zacentrus) and redbanded rockfish (S. bab-
cocki). In a submersible study conducted off Southeast-
ern Alaska (188-290 m), Krieger (1993) compared the
fish densities of 4 untrawlable sites with 16 trawlable or
marginally trawlable sites, and reported that densities
of large (>25 cm) rockfish (a category that included Pa-
cific ocean perch, sharpchin rockfish, redstripe rockfish,
and harlequin rockfish (S. variegatus) were highest at
trawlable sites. In a study of shortraker (S. boreal is) and
rougheye (S. aleutianus ) rockfish conducted on the upper
continental slope off southeastern Alaska (262-365 m),
Krieger and Ito (1999) reported that soft substrates of
sand or mud usually had the greatest densities; hard
substrates of bedrock, cobble, or pebble had the least
densities; and habitats containing steep slopes and
numerous boulders had greater densities of rockfish
than habitats with gradual slopes and few boulders.
O'Connell and Carlile (1993) conducted a submersible
survey off southeastern Alaska in two depth strata;
shallow {<108 m) and deep (a 108 m). Yelloweye rock-
fish were observed in cobble, continuous rock, broken
rock and boulder habitats but were most abundant in
broken rock and boulder habitats of the deep stratum.
Habitat-specific studies in Oregon and California have
used finer scales of habitat classification to characterize
fish-habitat associations than our comparatively coarse
trawlable or untrawlable classification. In Oregon wa-
ters. Stein et al. ( 1992) reported estimates offish density
by habitat-type from a submersible study of six stations
at Heceta Bank in waters ranging from 60 to 340 m
in depth. Rockfishes, particularly pygmy (S. wilsoni),
sharpchin, rosethorn, and yellowtail, dominated all
substrates except mud, where Dover sole and black-
belly eelpouts (Lycodes pacificus) were most abundant.
In California waters, Yoklavich et al. (2000) conducted
a submersible study at Soquel canyon (94-305 m) in
Monterey Bay. Cluster analysis grouped fish densities
into six habitat guilds; most distinct were 1 ) guild I (fish
associated with uniform mud bottom of flat or low relief,
dominated by stripetail rockfish (S. saxicola)) and guild
VI (fish associated with rock-boulder habitat of low to
high relief dominated by pygmy rockfish).
To contrast our results in Washington with findings
from Oregon and California, we summarized the fish
density estimates reported by Stein et al. ( 1992 ) and Yokla-
vich et al. (2000) into a format roughly comparable to our
data. Differences in the objectives and methods of their
studies precluded a rigorous quantitative comparison with
our results, particularly because of differences in habitat
classification and survey design (random sampling in our
study, purposive sampling in the other two studies). How-
ever, some interesting similarities are apparent if the most
highly rugose habitats of these two studies are treated as a
proxy for untrawlable habitat and if the low bottom relief
habitats are treated as a proxy for trawlable habitat (Table
11). Seven species (italicized in Table 1 1 ) co-occurred in all
three studies. For all three studies, greenstriped rockfish
Roc)^
• - - ■Lingcod ^i^
S 125-
S 100-
^^^^
f 75-
S 50-
^^^-^^^
25 •
^
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Proportion untrawlable
Figure 8
Sample size guidelines for estimating the trawl survey habitat bias
(A!=the number of submersible dives in each habitat type) as a function
of the proportion of untrawlable habitat (A^/A) in a management area.
for power = BO'/f and o=0.05 for three categories offish.
Fig. 8). However, because our analysis aggregated flatfish
as a group, these results do not address the estimation of
a bias correction factor for individual species, which is a
requirement for any correction factor to be useful for stock
assessment purposes.
As for any survey method, the visual transect survey
method has an array of advantages and disadvantages,
which have been well chronicled elsewhere (Uzmann et
al., 1977; Ralston et al., 1986; Butler et al., 1991; Adams et
al.,1995; Starr et al., 1996; CaiUiet et al., 1999). Some of the
disadvantages include 1) difficulties in fish identification,
particularly for small fish or fish with subtle coloration,
2) the potential for attraction or repulsion of fish from the
560
Fishery Bulletin 101(3)
Table 11
Comparison offish density estimates (average number offish/hectare) in trawlable (D,) and untrawlable (D„) habitats from sub-
mersible studies in Washington, Oregon, and California. Densities for italicized species were reported in all three studies.
Washington (present study)
Oregon '
Species
D,
D„
D.
D..
California-
D,
D..
Rockfish
Bank rockfish
Bocaccio
Canary rockfish 0.41
Cowcod
Darkblotched rockfish
Greenblotched rockfish
Greenspotted rockfish
Greenspotted and greenblotched rockfish
Greenstriped rockfish 5.65
Halfbanded rockfish
Pygmy rockfish
Quillback rockfish
Redstripe rockfish
Rosethorn rockfish 0.48
Sharpchin rockfish
Shortspine thomyhead
Stripetail rockfish
Tiger rockfish 0.24
Widow rockfish
Yelloweye rockfish
Yetlowtail rockfish 0.24
Flatfish
Arrowtooth flounder 9.25
Dover sole 9-33
Pacific halibut 6.88
Other Fish
Eelpout 11-46
Greenling
Lingcod '■■'?-9
Pacific cod 0.70
Pacific hagfish
Pacific hake
Poachers
Spotted ratfish 1-54
Salmon 0.28
Skate 2.81
Spiny dogfish 167
Wolf-eel
0.36
3.76
0.27
3.39
77.78
4.40
3.90
1.11
0.49
165.00
510.00
479.50
119.50
8.65
10.70
33.50
0.19
3.00
249.50
1.77
19.26
2.67
9.78
33.50
93.00
0.00
105.00
6.33
586.00
120.00
0.00
148.00
4.33
152.67
86.33
52.00
1.33
36.33
162.33
237.67
1.67
16.33
39.50
218.67
46.00
220.00
85.67
892.50
126.33
734.33
574.50
40.33
175.33
96.50
138.50
41.33
5.33
304.67
63.67
0.33
33.67
13.50
0.67
78.67
95.50
2.67
28.00
7.50
15.00
9.00
58.00
43.67
3.00
91.67
25.67
4.00
14.67
14.00
138.00
22.67
' Oregon data source: Table ,3 of Stein et al. (1992). Categories "mud" and mud-cobble" were averaged and used as a proxy for trawlable habitat;
categories "flat rock" and "rock ridge" were averaged and used as a proxy for untrawlable habitat.
-' California data source: Table 2 of Yoklavich et al. (2000) Categories "mud," "cobble-mud" and "mud-pebble" were averaged and used as a proxy for
trawlable habitat; categories "rock-mud," "rock ridge," and "rock boulder" were averaged and used as a proxy for untrawlable habitat.
submersible, 3) variation in countability due to habitat type;
for example, due to reduced visibility when the submersible
maneuvered ofTbottom to avoid large boulders, or the failure
to detect fish hiding behind boulders, and 4) the limitation
of the technique to quantifying the density of benthic spe-
cies found in close proximity to the bottom. The advantages
of the visual transect survey method include the ability to
1) sample in habitats that are inaccessible to other survey
methods, 2) observe /;; situ fish behavior, and 3) observe the
distribution of fish and fish-habitat associations on a fine
Jagielo et al.: Demersal groundfish densities in trawlable and untrawlable habitats off Washington 561
scale. Although our study was subject to the limitations of
the visual transect method, we assumed that the method
could reliably estimate (with a catchability of 9=1.0) the true
density of selected demersal bottomfish in both trawlable
and untrawlable habitats for evaluation of the habitat bias
present in the trawl-survey approach (which does not allow
for sampling in untrawlable habitat). We do not feel that
this assumption was severely violated, although we have
no objective measure of the potential biases of the method,
and thus we cannot estimate the consequences of assump-
tion failure. We did recognize clearly that difficulties in fish
identification limited the number of species that we could
quantitatively sample with this technique. Technological
improvements in underwater videography and image rec-
ognition software are likely to enhance the capabilities of
the visual transect survey technique in the future.
In conclusion, it is clear that relatively large-scale sur-
veys are needed to assess bottomfish densities in habitats
that are not accessible to trawl survey gear. For some spe-
cies, it may be possible to derive an area-specific trawl-sur-
vey bias correction factor, but for many other species it is
likely that there will be no substitute for direct estimation
of densities in untrawlable habitat on a routine and synop-
tic basis. In either case, stratified random sampling designs
should be employed with sample sizes sufficient to ensure
acceptable levels of statistical power At present, the in situ
visual transect submersible survey method appears to be a
useful tool for this purpose, and the utility of this method
will likely improve further with technological advances.
Acknowledgments
We would like to thank Farron Wallace and Brian Culver
for help during the submersible dive survey and with fish
identification on the videotapes; Cindy Knudsen for video-
tape area-swept data collection; Kevin Redman and Colin
Stewart (Williamson and Associates) for geophysical data
analysis and mapping, and Mike Famam, and Brian Bunge
(USN) for geophysical data acquisition; the captains and
crews of the USN Agate Passage and FW Auriga for excel-
lent support vessel services; D. Slater, C. Ijames, and J. Lilly
of Delta Oceanographies for safe and efficient use of the
Delta submersible; Victoria O'Connell and Waldo Wakefield
for advice on field logistics and data collection; and Marion
Larkin (FV Larkin), for his insights regarding trawlable
and untrawlable habitat obtained from many years of fish-
ing experience in the study area. This study was supported
by the NOAA National Undersea Research Program, West
Coast and Polar Regions Undersea Research Center, Uni-
versity of Alaska Fairbanks (grant no. UAF 98-0045), the
Washington Department of Fish and Wildlife, and the
National Marine Fisheries Service.
Literature cited
Adams, P. B., J. L. Butler, C. H. Baxter, T. E. Laidig, K. A. Dahlin,
and W. W, Wakefield.
1995, Population estimates of Pacific coast groundfishes
from video transects and swept-area trawls. Fish. Bull.
93:446-455.
Alverson, D. L.
1951. Deep water trawling survey off the coast of Washing-
ton (August 27-October 19, 1951) Commercial Fisheries
Review 13:11. U.S. Dep. Fish. Wild. Serv., Sep. 292.
Auster, P. J., L. L. Stewart, and H. Sprunk.
1989. Scientific imaging with ROVs: tools and techniques.
Mar. Technol. Soc. J. 23(3):16-20.
Butler, J, L., W. W Wakefield, P B. Adams, B. H. Robison, and
C. H. Baxter
1991. Application of line transect methods to surveying
demersal communities with ROVs and manned sub-
mersibles. Proceedings of the Oceans 91 Conference,
Honolulu, Hawaii, 1-3 October 1991, p. 689-696. Marine
Technology Soc, Columbia, MD.
Cailliet, G. M., A. H. Andrews, W W. Wakefield, G. Moreno, and
K. L. Rhodes.
1999. Fish faunal and habitat analyses using trawls, camera
sleds and submersibles in benthic deep-sea habitats off cen-
tral California. Oceanol. Acta 22(6):579-592.
Caimi, F. M., J. H. Blatt, B. G. Grossman, D. Smith, J. Hooker,
D. M. Kocak, and F. Gonzalez.
1993. Advanced underwater laser systems for ranging, size
estimation, and profiling. Mar. Technol. Soc. J. 27(1):31^1.
Cohen, J.
1988. Statistical power analysis for the behavioral sciences,
2"'' ed., 567 p. L. Erlbaum Associates, Hillsdale, NJ.
Dark, T A., and M. E. Wilkins.
1994. Distribution, abundance, and biological characteris-
tics of groundfish off the coast of Washington, Oregon, and
California, 1977-1986. U.S. Dep. Commer Nat. Mar. Fish.
Serv., NOAA Tech. Rep. NMFS 117, 73 p.
Davis, D. L., and R. F. Tusting.
1991. Quantitative benthic photography using laser calibra-
tions, 5 p. Undersea World, San Diego, CA.
Dixon, W. F., and F. J. Massey.
1957. Introduction to statistical analysis, 2"'' ed., p. 244—255.
McGraw-Hill, New York, NY.
Greene, H. G., M. M. Yoklavich, R. M. Starr, V. M. O'Connell, W.
W. Wakefield, D. E. Sullivan, J. E. McRea, Jr, and
G. M. Cailliet.
1999. A classification scheme for deep seafloor habitats.
Oceanol. Acta 22(6):663-678.
Gregory, R. S., J. T. Anderson, and E.L. Dalley.
1997. Distribution of juvenile Atlantic cod Gadus mor-
hua relative to available habitat in Placentia Bay, New-
foundland. Northwest Atl. Fish. Organ. Sci. Counc. Stud.
29:3-12.
Krieger, K. J.
1993. Distribution and abundance of rockfish determined
from a submersible and by bottom trawling. Fish. Bull.
91:87-96.
Krieger, K. J., and D. H. Ito.
1999. Distribution and abundance of shortraker rockfish,
Sebastes borealis, and rougheye rockfish, S. aleutianus,
determined from a manned submersible. Fish. Bull. 97:
264-272.
Kulbicki, M., and L. Wantiez.
1990. Comparison between fish bycatch from shrimp trawl-
net and visual censuses in St.Vincent Bay, New Caledonia.
Fish. Bull, 88:667-675.
Matthews, K. R., and L. J. Richards.
1991. Rockfish (Scorpaenidae) assemblages of trawlable
and untrawlable habitats off Vancouver Island, British
Columbia. N. Am. J. Fish. Manage. 11:312-318.
562
Fishery Bulletin 101 (3)
Murie, D. J., D. C. Parkyn, B. G. Clapp, and G. G. Krause.
1994. Observations on the distribution and activities of rock-
fish, Se6as;es spp., in Sannich Inlet, British Columbia, from
the Pisces IV submersible. Fish. Bull. 92:313-323.
Nasby, N. M.
2000. Integration of submersible transect data and high-
resolution sonar imagery for a habitat-based groundfish
assessment of Heceta Bank, Oregon. M.S. thesis, 50 p.
Marine Resource Management Program, College of Oceanic
and Atmospheric Science, Oregon State Univ., Corvallis,
OR.
O'Connell, V. M., and D. W. Cariile.
1993. Habitat-specific density of adult yelloweye rockfish
Sebastes ruberrimus in the eastern Gulf of Alaska. Fish.
Bull. 91:304-309.
PFMC (Pacific Fishery Management Council).
2001. Status of the Pacific Coast groundfish fishery through
2001 and recommended biological catches for 2002: stock
assessment and fishery evaluation, 26 p. [Document pre-
pared for the Council and its advisory entities). Pacific
Fishery Management Council, Portland, OR.
Peterman, R. M.
1990. Statistical power analysis can improve fisheries re-
search and management. Can. J. Fish. Aquat. Sci. 47:2-15.
Ralston, S., R. M. Gooding, and G. M. Ludwig.
1986. An ecological survey and comparison of bottom fish
resources assessments (submersible versus handline fish-
ing) at Johnston Atoll. Fish. Bull. 84:141-155.
Richards, L. J.
1986. Depth and habitat distributions of three species of
rockfish (Sebastes) in British Columbia: observations from
the submersible Pisces IV. Environ. Biol. Fish. 17(1):
13-21.
Shaw, F. R., M. E. Wilkins, K. L. Weinberg, M. Zimmermann, and
R. R. Lauth.
2000. The 1998 Pacific West Coast bottom trawl survey of
groundfish resources: estimates of distribution, abundance,
and length and age composition. NOAA Technical Memo-
randum NMFS-AFSC-114, 138 p.
Starr, R. M., D. S. Fox, M. A. Hixon, B. N. Tissot, G. E. Johnson,
and W. H. Barss.
1996. Comparison of submersible-survey and hydroacous-
tic-survey estimates offish density on a rocky bank. Fish.
Bull. 94:113-123.
Stein, D. L., B. N. Tissot, M. A. Hixon, and W. Barss.
1992. Fish-habitat associations on a deep reef at the edge of
the Oregon continental shelf Fish. Bull. 90:540-551.
Uzmann, J. R., R. A. Cooper, R. B. Theroux, and R. L. Wigley.
1977. Synoptic comparison of three sampling techniques for
estimating abundance and distribution of selected mega-
fauna: submersible vs. camera sled vs. otter trawl. Mar.
Fish. Rev 39(12): 11-19.
Winer, B. J.
1971. Statistical principles in experimental design, 2"'' ed.,
907 p. McGraw Hill, New York, NY.
Yoklavich, M. M., H. G. Greene, G. M. Cailliet, D. E. Sullivan, R.
N. Lea, and M. S. Love.
2000. Habitat associations of deep-water rockfishes in a
submarine canyon: an example of a natural refuge. Fish.
Bull. 98:625-641.
Appendix I: Procedure used for estimating
the swept transect area
At each sample unit (submersible (iive site), we estimated
the total swept transect area, where the swept area (m^) =
(average area swept per second [m-^/sec]) x (total elapsed
time [seconds] ). The average area swept per second ( m'^/sec)
was computed for a set of randomly selected thirty second
portions of each transect. Conceptually, we determined the
average area swept per second for the subsampled areas
from a series of adjacent trapezoids (Fig. 1).
For each trapezoid, we determined swept area (A,) by
measuring the width that was swept (/, ) and distance that
was swept (T,), where
A =-(/,+/,.: )?;•
2
The process involved a frame-by-frame analysis of the
video image, which required tracking an object from the
center of the video monitor display to the bottom edge of the
video display for a known time interval ( Fig. 2 ). The elapsed
time for this interval was obtained from the video frame
count, and was used to calculate area swept per second.
Width-swept estimates (/,) were calculated from 1 ) the
distance between the laser spots on the video monitor
display (w,), 2) the width of the video monitor display (V),
and 3) the known distance between the lasers (W) (20 cm),
where
Submersible path
Area 3
Area 2
Area
Figure 1
Schematic representation of adjacent trapezoids.
/; =
VW
(1)
Because the width that was swept varied as the submers-
ible distance olT bottom varied, it was measured for each
block. The following procedure was performed in sequence:
Jagielo et aL: Demersal groundfish densities in trawlable and untrawlable habitats off Washiington
563
^ideo monitor
Object 1
\
Object 2
\
r9i
eft
Object 1
(9r
1
Screen 1: Distance between laser spots, m', , is
Tieasured and object adjacent to lasers is identified. '
>
Screen 2: Frame by frame advance until object
idjacent to lasers in screen 1 rests at bottom of
/ideo screen; then, distance between laser spot;
s
1
kV,, is measured, and the process repeats.
Figure 2
Illustration of the video-monitor display showing the frame-by-frame screen advance procedure used to determine the area swept
per second.
1) «', was measured to the nearest milli-
meter, 2 ) an object on the seafioor adjacent
to the laser spots was identified, 3) the
videotape was advanced until the object
appeared at the bottom of the video moni-
tor display, and 4) w^ was measured again
(Fig. 2). The distance that was swept during
this interval (T) is calculated trigonometri-
cally by using the angle of the camera and
constants estimated with the following pro-
cedures of Davis and Tusting (1991). The
process is illustrated in Figures 3 and 4.
The variables of interest are
T = the geodetic distance between the lo-
cation of the laser spots on the sea-
floor and the bottom edge of the cam-
era's field of view (distance swept);
the height of the video camera above
the sea floor;
o = the angle of the camera lens;
8 = the tilt angle of the camera;
D = the distance between the focal point
of the camera and the reflection of
the laser spots on the seafioor;
D] = the horizontal distance from the
camera to a point on the sea floor
at the center of the camera's field of view;
Dg = the horizontal distance from the camera to a point
on the sea floor at the bottom edge of the field of
view; and
D' = the distance from the camera lens to the refiection
of the laser spots on the seafioor;
H =
Figure 3
Schematic representation of the submarine and camera orientation to the
bottom (the line labeled D' represents the center line of the camera). Defini-
tions for the variables are provided in the left column of text on this page.
w = the distance measured between the laser spots as
they appear on the video monitor display;
W = the known distance (20 cm) between the lasers
mounted in parallel on the camera housing.
Note the following relationships:
564
Fishery Bulletin 101(3)
Laser spots on seafloor
Figure 4
Schematic representation of the relationship between the camera lens, image plane, and laser spots on
the seafloor.
Di = D' cos 0, and // = D' sin 0
H D'sine
A=-
tan 6 +
tan e +
(2)
(3)
In Equation 3, estimation of D^ requires the height of
the camera above the seafloor (H); however, the need for
a direct measurement of H can be eliminated by using
camera parameters that provide an independent estimate
ofDMFig. 4).
Figure 4 shows the relationships between the camera
lens, image plane, and laser spots, where d is a constant
representing the distance from the focal point to the image
plane, and c is a constant representing the distance from
the camera lens to the image plane (note that c may be
positive or negative).
Note that both d and c are specific to the video display
monitor employed, W, 0, and « are fixed, and w is observed.
Therefore,
W
D' = D-d-c. ■dndD = d
w
D' = il\—-\\-(
(4)
(5)
Underwater tests were conducted and the constants c and
d were estimated for Delta s video camera and laser set-up
by following the procedures of Davis and Tusting (1991).
The distance traveled (T) for each area-swept trapezoid
(from the center of the image to the lower edge of camera
field of view), then, is
T=a-D, = D'
cos
180
sin 6
1 180
tani 1 e +
180l T
(6)
Appendix II: Derivation of the trawl-survey
habitat-bias estimator, and the trawl-survey
habitat-bias effect size-index (t/^,)
To estimate the trawl survey habitat bias, we contrasted
1 ) the traditional abundance estimator, which does not dis-
criminate between fish density differences in trawlable and
untrawlable habitats (habitat-biased), with 2) an unbiased
abundance estimator that explicitly allows for density dif-
ferences between trawlable and untrawlable habitats.
Let D, = the true density in the trawlable habitat;
A, = the area of trawlable habitat;
D = the true density in the trawlable habitat;
A = the area of untrawlable habitat;
A = the total area = A^ + A„;
A^ = total abundance; and
A = the difference in true densities = D^-D^.
Then, for the unbiased estimator.
Ar = D,A, + £>„Ar
and for the biased estimator,
N = D^ = D,A, + D,A^.
The habitat bias, then, is the difference of the two estima-
tors, or
Bias = (D/i, + D/iJ -
(1)
The total error in the abundance estimator is a function
of both the bias and the variance ¥(/),) of the fish density
estimator
MSE = Bias'^ -t- (A2)V(D,),
(2)
Jagielo et al : Demersal groundfish densities in trawlable and untrawlable habitats off Washington
565
where V(D,) describes the uncertainty in the abundance
estimator. If the bias is much less than this uncertainty,
then its impact will be minimal. Therefore, we arbitrarily
set
Bias- = (A- mo,),
(3)
and substituting A A^ for bias from Equation 1 into Equa-
tion 3 gives
A^2^2 = (A2)V(Dp.
Solving for A gives
A = — 5£)(D,).
A.
(4)
(5)
where SD{ D, ) = the standard deviation of the trawl survey
density estimator in the trawlable habitat.
Thus, the effect size threshold used for detecting differ-
ences in mean density in the power analysis is a product
of the arbitrary decision for the bias in the abundance
estimator to be equal to its standard error.
For the statistical power analysis, we expressed A (the
difference in densities between habitats) as the standard-
ized effect size index (c/^) for a two-sample f-test (Cohen,
1988); dividing (Eq. 5) lay an estimate of the population
standard deviation, which yields
— SD(D,)/s^
A,
566
Abstract — Understanding the onto-
genetic relationship between juvenile
Steller sea lions iEiimetopiaa jubatus)
and their foraging habitat is key to
understanding their relationship to
available prey and ultimately their
sur\nval. We summarize dive and move-
ment data from 13 young-of-the-year
(YOY) and 12 yearling Steller sea lions
equipped with satellite dive recorders
in the Gulf of Alaska and Aleutian
Islands (n=18), and Washington (/i=7)
from 1994 to 2000. A total of 1413 d of
transmission (.v=56.5 d, range: 14.5-
104.1 d) were received. We recorded
222,073 dives, which had a mean depth
of 18.4 m (range of means: 5.8-67.9 m:
SD= 16.4 ). Alaska YOY dived for shorter
periods and at shallower depths (mean
depth=7.7 m, mean duration=0.8 min,
mean maximum depth=25.7 m, and
maximum depth=252 m) than Alaska
yearlings (.?=16.6 m, 0=1.1 min, .v =
63.4 m, 288 m), whereas Washing-
ton yearlings dived the longest and
deepest (mean depth=39.4 m, mean
duration=1.8 min, mean maximum
depth=144.5 m, and maximum depth=
328 m ). Mean distance for 564 measured
trips was 16.6 km; for sea lions slO
months of age, trip distance (7.0 km)
was significantly less than for those
>10 months of age (24.6 km). Mean trip
duration for 10 of the 25 sea lions was
12.1 h; for sea lions slO months of age,
trip duration was 7.5 h and 18.1 h for
those >10 months of age.
We identified three movements types:
long-range trips (>15 km and >20 h),
short-range trips (<15 km and <20 h)
during which the animals left and
returned to the same site, and transits
to other haul-out sites. Long-range
trips started around 9 months of age
and occurred most frequently around
the assumed time of weaning, whereas
short-range trips happened almost
daily (0.9 trips/day, n=426 trips). Tran-
sits began as early as 7 months of age,
occurred more often after 9 months of
age, and ranged between 6.5 and 454
km. The change in dive characteristics
coincided with the assumed onset of
weaning. These yearling sea lion move-
ment patterns and dive characteristics
suggest that immature Steller sea lions
are as capable of making the same types
of movements as adults.
Manuscript approved for publication
29 October 2002 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull 101:566-582 (2003).
Diving behavior of immature Steller sea lions
iEumetopias jubatus)
Thomas R. Loughlin
Jeremy T. Sterling
National Marine Mammal Laboratory
Alaska Fishenes Science Center, NMFS
7600 Sand Point Way, NE
Seattle, Washington 98115
E-mail address (for T R Loughlin) torn loughlin®noaa gov
Richard L. Merrick,
Northeast Fishenes Science Center, NMFS
166 Water Street
Woods Hole, Massachusetts 02543
John L. Sease
Anne E. York
National Marine Mammal Laboratory
Alaska Fishenes Science Center, NMFS
7600 Sand Point Way, NE
Seattle, Washington 98115
Steller sea lions range throughout the
North Pacific Ocean rim and are declin-
ing in numbers in most of Alaska and
Russia (Loughlin et al., 1992; Loughlin
and York 2000). Studies of mitochon-
drial DNA suggest that at least two
stocks exist: an eastern stock (Califor-
nia through southeast Alaska) and a
western stock (Prince William Sound
and areas west) (Bickham et al., 1996;
Loughlin, 1997). For the western U.S.
stock (west of 144°W), counts of adults
and juveniles have fallen from about
110,000 individuals in the late 1970s
to about 25,000 individuals in 2000— a
decline of almost 80%. Although the
numbers of sea lions that died were
greater from the late 1970s to the
early 1990s than at present, the rate of
decline has remained high. As a result of
this decline the U.S. government desig-
nated the western stock as "endangered"
in 1997 under the U.S. Endangered Spe-
cies Act; the eastern stock is designated
as "threatened." Reasons for the decline
in numbers are unknown but may be
hnkcd to reduced availability of prey
caused indirectly by environmental
changes or commercial fishing activi-
ties, or both (Loughlin and Merrick,
1989; Merrick, 1995). Severe environ-
mental perturbations and commercial
fishing, both resulting in changes in the
abundance or availability of prey, have
been implicated in the alteration of pin-
niped foraging behavior and declines in
pinniped abundance (e.g. Trillmich and
Ono, 1991; Melin, 2002). One method for
studying the effect of reduced prey avail-
ability on pinnipeds is to measure diving
behavior and foraging ecology by using
either a time-depth-recorder ( Kooyman
et al., 1983; Gentry and Kooyman, 1986)
from which dive data are retrieved after
the animal returns from a feeding trip
(e.g. Goebel et al., 1991; Boyd et al.,
1994; Werner and Campagna, 1995), or
by using a satellite-linked time-depth
recorder (SLTDR; the newer version is
called a "satellite dive recorder" SDR),
which transmits dive and transmitter-
status information to orbiting satel-
lites and thus eliminates the need to
recapture the animal (e.g. Merrick et
al., 1994).
Few data are available concerning
the foraging ecology of Steller sea lions.
Merrick et al. (1994) and Merrick and
Loughlin (1997) presented information
on the dive characteristics and foraging
Loughlin et al.: Diving behavior of immature Eumetopias jubatus
567
Table 1
Satelhte transmitter number (PTT number), deployment location, age, sex, and morphometric measures of 25 Steller sea lions
studied for diving behavior in Alaska and Washington. 1994-2000. The ten ST- 10 and ST- 16 SDRs we deployed that transmitted
time-line messages are shown with **. 1 = Washington State area; 2 = Kodiak area; 3 = Shumagin Islands; 4 = Unimak Pass area;
5 = Sequam area, n/d = no data obtained. PTT number is the satellite transmitter identification number. Est. = estimated.
PTT number
Location code
(regional)
Age
(months)
Sex
Deployment
date
Length of
transmission
(d)
Mass
(kg)
Girth
(cm)
Length
(cm)
14073
12
M
6/8/95
51.67
86.26
n/d
151
14084
11
F
5/5/99
50.22
77.18
102
159
14085
19
M
1/5/00
14.50
154.36
133
193
14087
16
F
10/3/97
68.99
122.45
113
171
14089
16
F
10/3/97
55.58
111.11
122
173
21103**
22
M
3/30/00
63.80
139.23
128
192
21106**
22
M
3/30/00
83.28
143.31
119
192
14071
2
6
M
12/2/94
19.61
92.00
n/d
n/d
14074
2
6
M
12/6/94
53.13
79.80
n/d
n/d
14077
2
18
M
12/9/94
39.63
n/d
n/d
n/d
14078
2
18
F
12/7/94
57.99
n/d
n/d
n/d
14079
2
7
F
1/14/96
27.11
94.90
115
155
14080
2
7
M
1/16/96
79.04
106.20
115
156
14170**
2
21
M
3/12/00
93.99 Est.
95-105
n/d
150
21094**
2
9
M
3/12/00
66.40
62.20
90
145
14076
3
21
F
3/1/96
52.21
103.70
111
n/d
14081
3
9
M
3/2/96
45.02
n/d
144
n/d
14072
4
22
F
4/13/95
56.35
116.10
n/d
n/d
14075
4
8
F
2/25/96
31.76
104.00
123
140
14164**
4
9
M
3/8/00
97.61
79.60
n/d
n/d
14167**
4
9
F
3/9/00
29.67
100.20
n/d
155
14111**
5
9
F
2/29/00
61.71
87.00
108.5
151
14114**
5
9
F
2/29/00
52.53
85.80
108
157
14116**
5
9
F
2/29/00
56.61
76.20
102.5
148
14163**
5
9
M
2/29/00
104.14
109.00
113
156
behavior of a small sample of Steller sea lions in Alaska;
Loughlin et al. (19981 provided similar information for
Steller sea lions off the Kuril Islands, Russia. Merrick et
al. (1990) and Brandon (2000) presented information on
female pup-attendance behavior of sea lions with VHF
radio-transmitters off the Kuril Islands and Alaska, re-
spectively. These studies showed that during the breed-
ing season, adult female Steller sea lions generally spent
about half their time at sea on relatively brief (18-20 h)
foraging trips. Dives tended to be shallow (,v=21 m), brief
(.v=1.4 min), and frequent (about 13/h). Observations dur-
ing winter showed that females with suckling yearlings
(17-22 months of age) had feeding trips of about 2.3 days,
whereas those with young-of-the-year (5-10 months of age)
had trips lasting 0.9 of a day; time on shore for lactating
females of both groups averaged 14.2 hours (Porter, 1997).
Baba et al. (2000) were able to follow a yearling Steller sea
lion for 5 months using two location-only satellite trans-
mitters; one was attached to the top of the head and the
other on the back. This animal traveled from Hokkaido
to Sakhalin Island and throughout the southern Okhotsk
Sea. No dive data were obtained.
Our objective is to present a description of the diving
behavior of juvenile Steller sea lions for the western stock
of Steller sea lions in Alaska and the eastern stock in Wash-
ington state. We deployed SDRs on juvenile Steller sea li-
ons over a broader geographical range in Alaska and over
a wider range of dates, providing a more comprehensive
picture of the diving behavior of young Steller sea lions.
Additionally, SDRs are now smaller and of higher quality,
so that more detailed information on diving behavior is
available. We then provide in the "Discussion" section a
comparison of the accounts in the present study to those
we published earlier on adult female diving behavior (e.g.
Merrick and Loughlin, 1997).
Materials and methods
We captured 25 free-ranging Steller sea lions of both sexes
from approximately 6-22 months of age at rookeries and
haul-out sites in the Aleutian Islands and Gulf of Alaska
(Table 1, Fig. 1 ) throughout the year from 1994 to 2000, and
during 1995-2000 at Shilshole Marina in Puget Sound, near
Seattle, Washington. Animal age was estimated by using
568
Fishery Bulletin 101(3)
50°0'N-
-50'0'N
XL
I I. ll
12 3 4 5 6 7
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Deployment Age in Months
I r-
16°0'W 13O0'W
Figure 1
Locations where satellite dive recorders (SDRs) were deployed on 25 Steller sea lions in Alaska
and Washington between 1994-2000. "Deployment age" is the age of the sea lions when satellite
transmitters were attached.
mid-June as the presumed birth date (Pitcher and Calkins,
1981) and pubUshed accounts of mass, standard length,
and girth at age (Calkins et al., 1998). Some juveniles
before 1996 were chemically immobilized with Telazol'"
injected intramuscularly by a dart fired from a pneumatic
gun (Loughlin and Spraker, 1989). Those animals were not
weighed; therefore exact dosage levels were not determined.
However, dosages were most likely between 1.5 and 2.5 mg/
kg. Once a sea lion was immobilized, intramuscular injec-
tion of 3-10 cc of Dopram was administered to stimulate
respiration and facilitate recovery. After 1996, young sea
lions were captured on land with a hoop net and physically
restrained. During all years a SLTDR or SDR was glued
to the pelage on the animal's back with fast-setting cpoxy
resin (Loughlin ot al., 1987), and two plastic cattle ear tags
with the same identification numbers were attached, one to
each front flipper The instruments were not recovered and
were expected to be shed during or before molt.
Instrument description and programming
We used 0.5-watt ST-6 SLTDRs (packaged by Wildlife
Computers, Redmond, WA), which provide dive depth,
dive duration, and transmitter status. Further develop-
ment by Wildlife Computers resulted in 0.25-watt ST- 10
and ST- 16 SDRs which could provide five messages: 1 ) dive
depth, 2 ) dive duration, 3 ) transmitter status, 4) proportion
of time at depth, and 5) a time line. Messages are sent at
prescribed intei-vals; transmission interval at sea is every
43 sec and on land it is every 1 min 28 sec. The number
of transmissions (and thus messages received) while the
sea lion is at sea depends on the length of exposure of the
instrument's salt-water switch at the surface. Location
data are not sent by the transmitter but are calculated by
Service-Argos, Inc. from the received message. Additional
information on these instruments and their capabilities
can be found in Merrick et al. (1994). The satellite track-
ing system (Argos) is described in detail in Fancy et al.
(1988) and Stewart et al. (1989). Additional information
can be obtained from the manufacturer at their web site
(www.wildlifecomputers.com).
The ST-6 SLTDR stored, summarized, and transmitted
dive data as histograms. Individual dives and surface in-
tervals were not provided; therefore sampling frequency for
measuring dive behavior was not a consideration (e.g. Boyd,
1993). Software programming of the SLTDR subdivided
each day into four 6-h periods (2100-0300 h, 0300-0900 h,
0900-1500 h, and 1500-2100 h local time). Frequency histo-
Loughlin et al.: Diving behavior of immature Eumetopias /ubatus
569
grams were summarized separately for dive depth and dive
duration for each of the four time periods. The SLTDRs re-
corded dive depth information in six separate "bins": 4—10 m,
10-20 m, 20-50 m, 50-100 m, 100-250 m, and >250 m. We
used 4 m as the minimum depth for a dive based on ear-
her studies in Alaska (Merrick et al., 1994). Dive-duration
bins were 0-60 sec, 60-120 sec, 120-180 sec, 180-240 sec,
240-360 sec, and >360 sec.
The ST-10 and ST-16 units used the same 6-h periods as
the ST-6. However, the ST-10 and ST-16 SDKs subdivided
dive depth information into 14 bins: 4 m; 4-6 m, 6-10 m,
10-20 m, 20-34 m, 34-50 m, 50-74 m, 74-100 m, 100-
124 m, 124-150 m, 150-174 m, 174-200 m, 200-250 m,
and >250 m. Dive duration also contained 14 bins at one-
minute intervals (e.g. 1-2 min, 2-3 min, 3-4 min, etc.). The
14 time-at-depth bins coincided with dive-depth bins (e.g. 0,
4, 4-6, 6-10, etc. and the last was >200). However, the first
bin was set to zero to determine if an animal was on land
based on the proportion of dry readings of the salt-water
switch during a 6-hour period. Time-at-depth was calcu-
lated as the proportion of time that dives occurred within
a particular depth bin of a 6-h period while the sea lion was
at sea (e.g. if an animal was at sea for 3 hours during a 6-h
period and spent half its dive time in bin 50-74, the value
in bin 50-74 would be 25'7f ).
We deployed ten ST-10 and ST-16 SDKs (Table 1) which
transmitted time-line messages in bins of 20-min periods
(there are 72 periods of 20-min each in a 24-h day). These
messages provide information on whether the instrument
was wet or dry >10 min of a 20-min period for each of the
72 periods. Time-line messages thus allow calculation of
time spent at sea and on land.
Maximum dive depth in a 24-h period, from midnight
GMT to midnight GMT. was provided in the status mes-
sage. This is a separate message that provides information
on transmitter status, including a pressure offset, battery
status, number of transmissions to date, at-surface data,
date, time, ID of message, and a saltwater conductivity
reading. All 25 transmitters that we deployed transmitted
a status message.
The ST-6 SLTDRs were on 24 h/day and transmitted a
maximum of 400 transmissions/day. To save battery power
the instrument had a 6-h haul-out period; that is, it would
turn off only if the transmitter was "dry" for 6 hours, indi-
cating that the animal was on land. The ST-10 and ST-16
SDRs had 3-h haul-out periods; the ST-10 had a maximum
of 250 transmissions/day, and the ST-16 had a maximum of
325/day Both the ST-10 and ST-16 had duty cycles of 4 h on
and 2 h off during a 24-h period to distribute transmissions
during different times of the day and to ensure recording of
information in all bins. All duty cycles started at midnight,
with an offset of + 13 h from GMT for Alaska.
Location data
Locations were estimated by the Service-Argos, Inc. clas-
sification scheme, where location class (LC ) 3 is accurate to
<150 m, LC 2 is accurate 150 m-s350 m, LC 1 is accurate
350 m-1000 m, and LC 0 is accurate >1000 m. LCs A and
B have no assigned accuracy range (Service-Argos, 1984;
Keating, 1994). However, after our analysis, Vincent et
al. (2002) used an algorithm published by McConnell et
al. (1992) to filter satellite locations and found that both
filtered and unfiltered LC A locations were of a similar
accuracy to LC 1 locations for four gray seals iHalichoerus
grypus). Because of the large variance in our samples asso-
ciated with LC A locations, we excluded them (and the LC
Bs) from our analyses. We sorted location data by date and
time line to determine the locations for each trip.
Data analysis
Data analysis followed that of Merrick et al. (1994) and
Merrick and Loughlin (1997). Analysis of the number of
dives was prepared by summing counts of dives from the
histograms. Median depths and durations of dives were
calculated by using the range midpoint of a bin (e.g. 7 m for
a 4—10 m bin) as the depth for all dives in the bin. We rec-
ognize that this approach invokes a possible error for dive
profiles in large increment bins (e.g. 50-100 m) where the
mean dive depth is the same, 75 m, regardless of whether
the animal made most of its dives between 51 and 60 m or if
it made most of its dives between 90 and 100 m. This error
is inherent in the data collection process and could not be
eliminated with the instruments used in our study. We also
recognize that more locations will be recorded when the
animals are at the surface for long periods or when transit-
ing to different locations. However, because of the repetitive
transmission of the histogram data and the usual short
duration of short-range trips, there should be no inherent
behavior-based bias in the dive data reported. Differences
in dive depth and duration between locations were tested
by using the Pearson chi-square tests or analysis of vari-
ance (ANOVA) (F-statistic), and P-value differences less
than 0.05 were considered significant. Analysis of trip
distance and duration were analyzed by using a repeated
measures ANOVA, and because the distances were skewed,
they were log-trEinsformed to examine differences among
groups.
Trips were defined and measured for distance by us-
ing an integrated process of the SDR data. For animals
deployed with ST6 SDRs, which did not contain time line
data sets (/! = 15), trip distances were extracted by using a
combination of the dive histogram, duration histogram,
and land or sea data sets to estimate arrival and departure
times as well as locations calculated at sea or on land.
Once arrival and departure times were estimated, the
location data were examined to confirm that all locations
calculated during that trip were wet locations. We then
had all locations for an individual trip and from those
locations we filtered out all A and B locations and im-
posed a swim speed filter (3 m/s). Finally, we reported the
maximum straight-line distance from the departure site.
For animals with STlO/16 SDRs, we were able to extract
arrival and departure times from the time-line messages.
However, if a day of time-line data was not received, we
referenced the time-at-depth data, depth, and duration
histograms to reconstruct the missing day of data. Once
arrival and departure times were calculated we then fol-
lowed the protocol stated above.
570
Fishery Bulletin 101(3)
■ Alaska YOY(n= 13)
D Alaska yearlings (n=5)
■ Washington yearlings (n=7)
^
£
ji
n
4-10 10-20 20-50 50-100 100-150 >150
Depth bins in meters
Figure 2
Percentage of dives occurring in each dive-depth bin for Alaska young-
of-the-year (YOY), Alaska yearling, and Washington yearling Steller sea
lions. Twenty-five animals are represented from 1994 to 2000 and a total
of 222,073 dives.
Results
We report on SDR data obtained from 25 (13 male, 12
female) young-of-the-year and juvenile (estimated ages
of <2 yr) Steller sea lions from Washington state. Gulf of
Alaska, and Aleutian Islands, Alaska (Table 1). Most (22
of 25) were caught during October-March 1995-2000 and
the remainder during May-July (Table 1). Mean number
of days of transmission received from the SDRs was 56.8 d
(range 14.5-104.1 d).
Dive characteristics
We recorded over 222,073 dives for young-of-the-year and
juvenile Steller sea lions which had a mean dive depth of
18.4 m (range of means: 6.1-67.0 m; SD=16.23). Alaska
young-of-the-year dived to shallower depths and for shorter
periods (mean depth=7.7 m, SD=1.7; mean duration=0.8
min, SD=0.1; mean maximum depth=25.7 m, SD=16.9;
and maximum depth=252 m) than did Alaska yearlings
(mean depth=16.6 m, SD=10.9; mean duration=l.l min,
SD=0.4; mean maximum depth=63.4 m, SD=37.7; and
maximum dcpth=288 m), whereas Washington yearlings
dived the deepest and the longest (mean depth=39.4 m,
SD=14.9; mean duration=1.8 min, SD=0.6; mean maxi-
mum depth=144.5 m, SD=32.6; and maximum depth=328
m ). Alaska animals dived to much shallower depths ( mean
depth= 10.3 ml than animals from Shilshole,WA. There was
no significant difference in the mean dive depths among
locations in Alaska (P=0.8). Alaska animals, in compari-
son to the Washington animals, had a significantly greater
proportion of dives in the 4-10 m depth bin ( 707f , P<0.001)
than in the deeper depth bins.
We compared the proportion of dives in the shallowest
bin (depth bin 4— 10 m ) for animals captured in Washington
state versus Alaska using a generalized linear model with
a binomial link function (McCuUagh and Nelder, 1989).
The proportion of shallow dives was significantly greater
(P<0.00 1 ) among the Alaskan animals ( 8 1 .47f ) than among
the Washington state animals (43.8%). Among the Wash-
ington state animals, the proportions of dives in the 1020
m depth bin (20.4%) and the 20-50 m depth bin (19.4%)
were similar; proportions of dives in the deeper depth bins
were progressively fewer (Fig. 2). Maximum and mean-
maximum dive depth were also greater for young sea lions
from Washington that dived to 141.5 m (SE=11.4) mean-
maximum depth versus 33.8 m (SE=7.2) for Alaska sea
lions (F=63.4, 23 and 24 df; P<0.001) (Table 2). We plotted
the maximum depth for each 24-h period by the number of
days in which the Argos satellite received a status message
(which contains maximum depth for 24 hours) and found
that with one exception ( PTT 14078 ), Washington yearlings
consistently dived deeper than their Alaska counterparts
(Fig. 3, A and B). Two of three Alaska young-of-the-year
were shallow divers and the third dived to 250 m once and
beyond 100 m on numerous occasions late in the track-
ing period (Fig. 3C). The maximum depth for all sea lions
that we studied was 328 m for a juvenile sea lion that was
equipped with a SDR at Shilshole, WA (PTT 21106); the
deepest dive for a yearling Alaska sea lion was 288 m (PTT
14078) (Fig. 3).
Mean dive duration was 1.1 min for all young sea lions
(n =226,497 dives). Dive duration was significantly longer
for Shilshole sea lions (x=1.75 min; range: 0.95-3.10) com-
pared to Alaska sea lions (v=0.85 min; range: 0.61-1.86;
F=24.5, 23 and 24 df; P<0.001). Few dives were greater
Loughlin et al.: Diving behavior of immature Eumetopias lubatus
571
Table 2
Summary of d
ive parameters from satellite dive recorders (SDRs) deployed
on Stellar
sea lions in
Washington and Alaska, 1994- |
•2000. "PTT" is
the satellite transmitter identification number.
Mean max.
Mean max.
Max.
Mean
Mean
Mean
Mean
dive depth
dive depth
depth
depth
depth
duration
duration
PTT
(m)
in)
(m)
(m)
M
(min)
(n)
Washington
14073
77.68
38
168
31.99
10,746
0.96
11,241
14084
154.12
34
288
47.29
5183
1.69
5047
14085
187.67
12
280
67.94
1991
3.10
2025
14087
164.37
53
256
33.21
14,287
1.69
14,572
14089
144.00
46
200
44.76
9682
1.82
9659
21103
124.22
18
256
23.59
6920
1.40
5431
21106
159.09
44
328
26.92
11,839
1.61
11,647
Alaska
14071
10.44
18
12
7.20
1541
0.61
1732
14074
11.07
30
20
7.24
3044
0.72
2954
14077
28.57
21
144
9.67
4745
0.81
5146
14078
125.74
23
288
35.00
4186
1.82
3657
14079
12.80
20
44
7.13
7546
0.71
8482
14080
15.80
59
24
7.60
17,236
0.96
18,056
14170
41.87
47
180
10.23
17,741
0.91
19,447
21094
48.90
31
152
11.67
9745
0.79
9388
14076
51.69
26
144
9.29
11,593
0.86
12,739
14081
11.76
34
20
7.25
8717
0.70
9013
14072
68.98
49
100
18.69
12,597
1.06
11,931
14075
8.00
15
12
7.01
5424
0.67
4869
14164
26.09
44
60
9.46
16,352
0.96
16,426
14167
20.73
11
60
7.01
2985
0.81
3022
14111
17.76
25
40
6.44
10,919
0.72
10,903
14114
13.75
16
16
5.84
5846
0.68
6256
14116
24.24
17
40
6.70
7359
0.73
7739
14163
65.55
49
252
10.94
13,849
0.82
15,204
Mean
62.42
135.36
18.42
1.10
SE
11.50
3.23
3.23
0.11
than 6 min (Fig. 4). There was a significant positive Un-
ear relationship between dive duration and dive depth
(r2=0.89, F=7.06, 1 and 23 df, P<0.001), and a significant
positive relationship between sea lion mass at the time of
capture and mean dive duration (r^=0.46, F=3.86, 1 and
20 df,P<0,001) but not girth (r2=0.10,F=1.62, land 14 df,
P=0.22). The relationship between dive duration and dive
depth for males was not different from that for females
(F=1.16, 2 and 21 df, P=0.33). The positive relationship
between dive duration and mass was likely driven by the
greater mass of the male sea lions because the relation-
ship was not statistically significant when the analysis was
restricted to females.
Dive depth and duration showed an interesting ontoge-
netic trend. Alaska animals 7-10 months old typically had
a mean dive duration of <1 min and a mean dive depth of
about 10 m; by 11-12 months of age both increased, almost
doubling in most cases (Fig. 5). Although sample size was
small, this ontogeny of diving to deeper depths for longer
periods at about 11-12 months of age was evident in the
percentage of time at depth (Fig. 6). There was a higher pro-
portion of time spent in the deeper depth bins during May
and June (at age 11 and 12 months, respectively) than when
younger, and the proportion of time hauled out was reduced
for the older animals. Interestingly, the decrease in dive
depth and dive duration for two Washington animals at 23
months of age (Fig. 5) corresponded with movement from
inside Puget Sound to deeper waters off the Washington
coast.
The greatest proportion of all diving (37%) occurred dur-
ing 2100-0300 h; the least (about 16%) during 0900-1500
h (Fig. 7). There were no periods when young-of-the-year
or juvenile sea lions from any location did not dive. The fre-
quency distribution of dives was similar in all time periods
for all age groups from Alaska and Washington (Fig. 7).
Distance and duration of trips at sea
Mean distance of trips at sea for 564 measured trips of
the 25 study animals was 16.6 km (SD=44.9 km; range:
572
Fishery Bulletin 101(3)
Washington yearlings
21106 max depth
14085 max depth
350
300
250
200-
150-
100
50 •
11 21 31
Days with maximum depth (m) per 24-h reception
21103 max depth
Days with maximum depth (m) per 24-h reception
14087 max depth
350
300
250-
200
150
100
50-
Days with maximum depth (m) per 24-h reception
14089 max depth
11 21 31 41 51
Days with maximum depth (m) per 24-h reception
350
300
250
200
150
100
50
1 11 21 31 ■ll
Days with maximum depth (m) per 24-h reception
Figure 3
Summaries of the maximum depth (m) for each 24-h period (O) m which a status message containing maximum depth informa-
tion was received by the Argos satellite for 13 of the 25 SDR-cquipped Steller sea lions. For example, the Argos satellite received
a status message from SDR 21 106 for about 44 days, yet the SDR was operational for a total of 83 days (A; see Table 1 ). These 13
were chosen to compare yearlings of comparable age in Washington (A), and Alaska (B), and to provide typical examples of Alaska
young-of-the-year (C).
10 months of age, the mean distance of all
tripswas24.6km(«=307;range: '°°
/\
50
0
y.^,ee«ee««ee«/V^^V^^^^
1 11 21 31 41
1 11 21
Days with maximum depth (m) per 24-h reception
Days with maximum depth (m) per 24-h reception
?,
14076 max depth
14078 max depth
a.
350-
360-.
4
300
300-
250
200
2S0'
200
aa a
E^
150
a 150
n / V ^^^"~9^ ^\ / \ y\
E
E
100
50
0
-VV-vV--''^-*"--""""\j :
f\i V 1/
1 11 21
1 11 21
0}
5
350-1
300
250
200
150 ■
100 ■
50
Days with maximum depth (m) per 24-h reception
14170 max depth
Days with maximum depth (m) per 24-h reception
1 11 21 31 41
Days with maximum depth (m) per 24-h reception
Figure 3 (continued)
cant gender (P = 0.6) nor gender x age interaction effects
(P=0.19).
Trip distance increased with age. For example, we cap-
tured a 9-nionth-old male sea lion (bearing transmitter
identification number PTT 21094, Table 1) near Kodiak
Island in March 200. It had short trip distances (<10 km)
which tended to concentrate near the capture site and
nearshore (Fig. 8). As the animal matured through April
and May, trip distance progressively increased until the sea
lion was swimming over 50 km offshore beyond the 100-
m depth contour and had a maximtim dive depth >150 m
(Table 2; Fig. 8).
Trip duration was measured for 10 of 25 animals with
SDRs containing time-line data (it was not possible to cal-
culate trip duration for 15 SDRs with the earlier SLTDRs
that did not transmit time-line data). Mean trip duration
for these 10 animals was 12.1 hours (n=544; SD=23.83 h;
range: 1—344 h; median=7.3 h). For animals slO months
574
Fishery Bulletin 101(3)
350t
300
250
200-
150
100
50
0
Alaskan young of the year
14163 max depth
11 21 31 41
Days witti maxmum depth (m) per 24-h reception
14164 max deptti
1 11 21 31
Days with maxmum depth (m) per 24-h reception
21094 max depth
350
SOO-
TS
o
£ 250
a.
i 200
■T
f 150
a 100
m 50
S
0
-e-^-e-^
11 21 31
Days with maxmum depth (m) per 24-h reception
Figure 3 (continued)
■ Alasl10 months of age, the mean duration of all trips was 18.1
hours («=237; SD=34.2 h; range: 1-344 h; median=10.3 h).
Averaged across individual animals, the mean duration of
trips at sea ranged between 6.2 to 21.4 hours; this range
was 6.2 to 17.2 hours for the younger animals and 10.3 and
21.4 hours for the older animals. The analysis of the re-
peated-measures ANOVA on the logarithm of trip duration
showed that the older sea lions had longer trip durations
(P<0.001). We could not test for gender and gender x age
effects because there were no measured trip durations for
females >10 months. Among the younger animals, there
was no gender difference in mean trip duration (P=0.11).
Types of movement
We identified three types of movements for the sea lions at
sea: long-range trips (>15 km and >20 h), short-range trips
(<15 km and <20 h), and transits to other haul-out sites
(Fig. 9). Long-range trips most likely were foraging trips
and began around 9 months of age. These trips had a mean
of 48.7 km (SD=55.7 km; max=240.8 km) and may coincide
with the assumed onset of weaning; they represented 6% of
all trips to sea. The most numerous trips (S&'Yc ) were short-
range foraging trips (v=3.6 km; SD=0.4; max=21.0 km),
which happened almost daily (0.9 trips/d, n=A2& trips).
Transits were movements from one haul-out site to another
haul-out site; these trips were characterized as the straight
line distance from one haul-out site to another and began
as early as 7 months of age but occurred more often after
9 months of age. Transit trips represented 6% of all trips
at sea and had a mean distance of 66.6 km (SD=83.7 km;
range: 6.5-341.9 km).
Discussion
The differences in diving behavior between young Steller
sea lions in Washington and those off Alaska are intriguing.
Possible reasons for these differences include variable habi-
tat type, prey resources, or morphological or genetic differ-
ences. However, there is no evidence, based on morphology
or genetics, to either support or refute differences in the
diving behavior that we observed. The evidence of genetic
differences between the western and eastern stock of Steller
sea lions is based on mtDNA haplotype differences for a seg-
ment of the mitochondrial D-loop which does not code for any
structural proteins (Bickham et al., 1996; Loughlin, 1997).
One morphological difference between the two stocks
is a progressive increase in mass of Steller sea lion pups
from east to west (Merrick et al., 1995), but whether this
difference in mass continues with increasing age is un-
known. Large animals typically dive deeper and longer
than smaller (and younger) animals (Schreer and Kovacs,
1997). Larger animals have less drag per unit of mass and
generally have more blood than smaller ones and thus are
able to store more oxygen. Larger animals also have lower
576
Fishery Bulletin 101(3)
80%
70% -
60%
50% -
40%
30%
20%
10%
0%
a t c d c ' g h
21 00-2 59
■ a; Haul out
n b 4 meters
0 c 4-6 meters
Hd: 6-10 meters
lae 10-20 meters
Hf: 20-34 meters
Hg: 34-50 meters
Eh: 50-74 meters
H i: 74-100 meters
??aTTi>-.^
a 0 c 6 (■ I g t\ \ atcdefgh
3:00-8:59 900-1459
Period
Figure 6
Percentage of time spent at depth for seven young-of-the-year Steller sea lions approximately (A) 7-10 months of age in
Alaska, and percentage of time spent at depth for three young-of-the-year Steller sea lions approximately (B) 11-12 months
of age (during May and June) in Alaska. This figure suggests that as young sea lions approach one year of age they tend
to spend less time hauled out and that a greater proportion of their dives are deeper.
mass-specific metabolic rates than their smaller counter-
parts and thus expend less energy and use less oxygen
stores (Schreer and Kovacs, 1997). Our sample size of sea
lions of comparable age is small; however, we compared the
mean mass of three Washington sea lions to the mean mass
of three Alaska sea lions of approximately the same age
(Table 1) and found that the Alaska animals had less mass
than those in Washington (108 kg vs. 145 kg). Whether or
not this difference in mass can account for the differences
we saw in diving characteristics for animals of similar age
(Fig. 3, A and B) is unknown.
The differences in diving characteristics between animals
tracked in coastal waters of Puget Sound, Washington, and
those tracked in Alaska waters are most likely linked to
localized differences in prey habitat. The primary prey of
Steller sea lions across their range are fish and ccphalo-
pods, both of which have a broad but predictable range
in temporal, spatial, and seasonal nearshore availability.
Typically, each species makes predictable migrations sea-
sonally from pelagic to nearshore waters where they form
large spawning concentrations. The prey are then further
concentrated by local transition boundaries such as frontal
zones and bathymetric features such as submarine chan-
nels (Sinclair et al., 1994). Steller sea lions appear to have
the foraging flexibility to take advantage of both the pre-
dictable behavioral traits of these prey species, as well as
the localized oceanographic conditions that enhance prey
concentrations (Sinclair and Zeppelin, 2002).
The primary prey of Steller sea lions in Alaska waters
is walleye pollock (Theragra chalcogramma ), which is con-
sumed year-round (Sinclair and Zeppelin, 2002). Walleye
pollock is replaced as a dominant year-round prey item by
Pacific whiting (Mer/(/cc/(/.s productus ) in Pacific Northwest
waters (Gcarin et al., 1999). Both species are semidemersal
and can be found from near surface waters to depths >1200
m, depending on localized conditions (Hart, 1973; Esch-
meyer et al., 1983). The greatest abundances of both species
are available to Steller sea lions in nearshore waters over
the continental shelf and perhaps as the prey become more
available during nighttime diurnal vertical movements.
The physical features of Puget Sound, along with its com-
plex bathymetry and the extensive channels and canyons,
provides extensive microhabitat for both predator and
prey species to express the full extent of their depth range.
In this respect, Puget Sound is comparable to the Gulf of
Alaska where Pacific cod (Gadus macrocephalus) is the
predominant winter prey item for Steller sea lions. Pacific
cod is thought to be consumed during spawning when it ap-
Loughlin et al : Diving behavior of immature Eumetoplas jubatus
577
80%
70%
60%
S 50%
Z 40%
g 30%
20%
10%
0%
B
I- c e f g h
21 00-259
3:00-8:59
■ a
Haul out
Db
4 meters
Be:
4-6 meters
Sd
6-10 meters
De
10-20 meters
Bf:
20-34 meters
Hg
34-50 meters
Dh
50-74 meters
S i
74-100 meters
^iMt^
a b c rt o ' g
9:00-14:59
a D c d p f g
15 00-20:59
Period
Figure 6 (continued)
pears to concentrate in the deep nearshore
channels and gulhes of the Gulf of Alaska
(Sinclair and Zeppelin, 2002).
The differences in dive depths that we re-
port also could be typical of the variability
among individuals. Boveng et al. ( 1996 ) ana-
lyzed TDR data for six dive-related variables
and found that dive duration was the least
variable and vertical distance (dive depth)
was the most variable among individual
Antarctic fur seals (Arctocephalus gazella ).
In our study, there was high individual vari-
ability in both dive depth and maximum
depth and little variability in dive dura-
tion— results similar to those of Boveng et
al.'s (1996) study
A female Steller sea lion nurses her pups
during the day, stays with the pup for the
first week, then goes to sea on foraging trips.
Maternal pup-attendance patterns seem to
vary over the sea lion's geographic range;
the average range of time for foraging trips
during lactation are from about 24 h to 2 d at
the southernmost rookery at Aho Nuevo Is-
land, California (Higgins et al., 1988; Hood and Ono, 1997;
but note that some of this variability may have been the
result of El Nifio conditions during part of the Higgins et
■ Alaska YOV
DAIaska yearling
■ Washington yearling
Period (local time)
Figure 7
Percentage of dives occurring in each time bin for Alaskan young-of-the-year
( YOY), Alaska yearling, and Washington yearling Steller sea lions. Twenty-
five animals are represented from 1994 to 2000.
al. study period), about 25 h at Lowrie Island, 19 h at Fish
Island, 11 h for Chirikof Island, and 7 h in the Aleutian
Islands (Brandon, 2000).
578
Fishery Bulletin 101(3)
153°0'W
152''0'W
58°0'N-
57°0'N'
•58°0'N
■57°0'N
152°0'W
isro'w
Figure 8
Figure showing the progressive increase in distance of locations from
shore for a B-nionth-oId sea lion over time. This animal (identification
number: PTT 21094) was equipped with a satellite transmitter near
Kodiak Island in March 2000. Early trips were <15 km from shore in
water <50 m in depth near the capture site. Trips became progi"essively
greater as the animal matured through May 2000 when it was venturing
over 50 km from shore in water >100 m in depth.
Ontogeny of diving ability has been studied in two other
otariids. Baker and Donohue (2000) used data loggers
(which they termed "time wet recorders") to measure time
spent in the water and diving behavior of northern fur
seal {Callorhinus iiniiriiis) pup.s on St. Paul Island, Alaska.
These pups began spending substantial time in the water
at approximately 40-50 d of age that coincided with growth
of the under fur and increases in sea surface temperature.
Time spent in the water increased up to about 100 d of age;
diving to depth did not occur until they were much older and
about to migrate. Horning and Trillmich (1997) conducted
an extensive study on the ontogeny of diving behavior in
Galapagos fur seals (Arctoccphaliis f^alapagnensis), a spe-
cies that weans no sooner than 2 years of age. They found
that in young the development of diving behavior was close-
ly linked to dependence on the mother and that substantial
diving activity did not occur until one year of age; but even
then the young fur seals were still nutritionally dependent
on their mothers and did not dive as deep, or for as long, as
mature females. The weaning date for Steller sea lions is
unknown but is assumed to be between 4 and 12 months,
and most pups are weaned just before the next breeding
season ( 11-12 months) (Porter, 1997). The change in diving
characteristics that we report is interesting in that it coin-
cides with this period. Prior to weaning these pups forage
in the company of their mother and learn to forage on their
own; the need to dive deep for long periods to acquire food
is compensated by nursing from the mother. Once weaning
Loughlin et al.: Diving behavior of immature Eumetopias lubatus
579
500'N
54<'0'N
74 O'W
Figure 9
The three types of movement exhibited by two immature Steller sea Hons
captured at Turf Point, Seguam Island, Alaska, in 2000. A long-range trip
(solid circles) >200 km is shown for PTT 14163 as it left and returned
to Turf Point. A transit trip (open triangles) for PTT 14111 is shown as
it left Turf Point and remained at the east end of Amlia Island where it
went on numerous short-range trips (shaded squares).
occurs, the yearlings are forced to explore more areas to
acquire food for needed energy. Dives become deeper and
longer as these yearlings forage at different depths within
the water column. Just before their first birthday, many
of these young sea lions are capable of diving to the same
depths and for the same duration as those of many adults;
they also begin to forage at greater distances and for longer
periods. Juveniles that we studied had a mean dive depth
of 18.4 m and dive duration of 1.1 min compared to adult
females in Alaska that had a mean dive depth of 21 m and
dive duration of 1.4 min (Merrick and Loughlin, 1997).
Maximum depth in our study was 328 m for a Washington
juvenile and 288 m for an Alaska juvenile. Maximum depth
information for adult females in Alaska was not provided by
the instruments used by Merrick and Loughlin ( 1997); their
maximum depths were characterized by bin data only. They
showed that about 5% of dives by adult females in winter
were greater than 250 m. In another study, adult females
in Alaska were equipped with early-style SLTDRs that had
features that recorded time-depth information and these
SLTDRs showed that the females frequently dived to 200 m
or more (Merrick et al., 1994).
Schreer and Kovacs (1997) summarized maximum dive
depth and dive duration for air-breathing vertebrates and
developed predictive allometric equations for both param-
eters based on body mass. We fitted our Steller sea lion
body mass data to these equations to estimate maximum
dive depth (27.33M^'''*^), where Afj represents body mass
in kilograms, and maximum dive duration (6.22^/^"'"). We
found that the maximum dive depth equation provided rea-
sonably close estimates but that dive durations were typi-
cally overestimated (Table 3). In some cases measured and
580
Fishery Bulletin 101(3)
Table 3
The recorded mass, recorded maximum dive depth, and recorded maximum dive duration for individual young Steller sea lions in
Alaska and Washington from this study and the estimated maximum dive depth (27.33Mj'''"') and estimate maximum dive dura-
tion 1 6.22 M^" ' 1 based on allometric equations in Schreer and Kovacs ( 1997 ). PTT number is the satellite transmitter identification
number. Est.= estimated, n/a = no data obtained.
PTT number
Mass (kg)
Maximum
depth (m)
Est. maximum
depth (m)
Maximum
duration (min)
Est. maximum
duration (min)
14073
86.26
168
212.38
>6
9.71
14084
77.18
288
201.79
>6
9.61
14085
154.36
280
277.57
>6
10.30
14087
122.45
256
249.52
>6
10.06
14089
111.11
200
238.61
4-6
9.96
21103
139.23
256
264.70
13
10.19
21106
143.31
328
268.24
>14
10.22
14071
92.00
12
218.77
2-3
9.78
14074
79.80
20
204.91
>6
9.64
14077
n/a
144
n/a
>6
n/a
14078
n/a
288
n/a
>6
n/a
14079
94.90
44
221.91
>6
9.81
14080
106.20
24
233.70
>6
9.92
14170
Est
95-105
180
Est. 222.02-232.48
>14
Est
9.81-9.91 1
21094
62.20
152
182.72
>14
9.40
14076
103.70
144
231.15
>6
9.89
14081
n/a
20
n/a
>6
n/a
14072
116.10
100
243.48
>6
10.01
14075
104.00
12
231.46
>6
9.90
14164
79.60
60
204.67
>14
9.64
14167
100.20
60
227.53
3-4
9.86
14111
87.00
40
213.22
8-9
9.72
14114
85.80
16
211.86
7-8
9.71
14116
76.20
40
200.60
5-6
9.59
14163
109.00
252
236.51
>14
9.94
estimated maximum dive depth values differed by large
amounts (e.g. sea lion PTT 14071), perhaps because the
deployment period was brief, before deep dives occurred.
For others (e.g. PTT 14074) the difference may have been
due to the young animal's continued dependence on the
female for nourishment; deeper dives do not occur until
weaning. In addition, we note that our dive duration data
were stored in bins of 1-min intervals (from 1 to 6 min in
the early instruments and from 1 to 14 min in the recent
ones); the exact duration of each dive is unknown.
Movement patterns also suggest that the swimming
ability of juvenile sea lions is comparable to that of adults.
It is not unusual for young sea lions to travel distances as
great as 1784 km from the natal rookery; as they approach
adulthood they generally remain within 500 km of their
natal rookery (Raum-Suryan et al., 2002). In our study
some young sea lions traveled several hundred kilometers
between sites while presumably searching for food or ven-
turing from the natal rookery site.
Further analysis of our SDR data is warranted to more
fully understand sea lion diving behavior and its relation-
ships with oceanographic parameters, daily and season
change, and behavioral features as discussed by Fedak et
al. (2001). The time allocation at depth (TAD) index de-
scribed by them will be a useful method for interpretation
of our SDR (and TDR) data. Further analysis of our SDR
data is needed to determine if such a study is possible.
Acknowledgments
Field assistance was provided by numerous NMML staff;
logistical support was provided by Alaska Helicopters and
the captain and crew of the U.S. Fish and Wildlife Service
research vessel Tiglax. The manuscript was improved by
comments from D. DeMaster, G. Duker.T. Gelatt, R. Hobbs,
M. Lander, J. Lee, R. Ream, E. Sinclair, and two anonymous
reviewers.
Literature cited
Baba. N., H. Nitto, and A. Nitta.
2000. Satellite tracking of young Steller sea lion off the coast
of northern Hokkaido. Fisheries Sci. 66:180-181.
Loughlin et al.: Diving behavior of immature Eumetopias jubatus
581
Baker, J. D., and M. J. Donohue.
2000. Ontogeny of swimming and diving in northern fur seal
iCallorhiniis iirsin us) pups. Can. J. Zool. 78:100-109.
Bickham, J. W., J. C. Patton, and T. R. Loughlin.
1996. High variability for control-region sequences in a
marine mammal: implications for conservation and bio-
geography of Steller sea lions (Eumetopias jubatus). J.
Mammal. 77:95-108.
Boveng, P. L., B. G. Walker, and J. L. Bengtson.
1996. Variability in Antarctic fur seal dive data: implications
for TDK studies. Mar Mamm. Sci. 12:543-554.
Boyd. I. L.
1993. Selecting sampling frequency for measuring diving
behavior. Mar Mamm. Sci. 9:424-430.
Boyd, I. L., J. P Y. Arnould. T. Barton, and J. P. Croxall.
1994. Foraging behaviour of Antarctic fur seals during
periods of contrasting prey abundance. J. Animal Ecol.
63:703-713.
Brandon, E. A. A.
2000. Maternal investment in Steller sea lions in Alaska.
Ph.D. diss., 137 p. Texas A&M University, Galveston, TX.
Calkins, D. G., E. Becker, and K. W. Pitcher
1998. Reduced body size of female Steller sea lions from a
declining population in the Gulf of Alaska. Mar. Mamm.
Sci. 14:232-244.
Eschmeyer, W. N., E. S. Herald, and H. Hammann.
1983. A field guide to Pacific coast fishes of North America,
336 p. Boston Houghton Mifflin Company, Boston, MA.
Fancy, S. G., L. F Pank, D. C. Douglas, C. H. Curby G. W. Gamer,
S. C. Amstrup, and W. L. Regelin.
1988. Satellite telemetry: a new tool for wildlife research
and management. U.S. Fish and Wildl. Serv. Resour Publ.
172:154.
Fedak, M. A., P. Lovell, and S. M. Grant.
2001. Two approaches to compressing and interpreting
time-depth information as collected by time-depth record-
ers and satellite-linked data recorders. Mar. Mammal Sci.
17:94-110.
Gearin, P., S. Jeffries, S. Riemer. L. Lehman, K. Hughes, and
L. Cooke.
1999. Prey of Steller's sea lions, Eumetopiasjubatus, in Wash-
ington state. In Abstracts of the 13th biennial conference on
the biology of marine mammals, Wailea, Hawaii November
28 December 3, p. 65. Soc. Marine Mammalogy, Wailea, HL
Gentry, R. L., and G. L. Kooyman.
1986. Methods and dive analysis. In Fur seals, maternal
strategies on land and at sea (R. L. Gentry and G. L. Kooy-
man eds. ), p 28-40. Princeton Univ. Press, Princeton, NJ.
Goebel, M. E., J. L. Bengtson, R. L. DeLong, R. L. Gentry, and
T. R. Loughlin.
1991. Diving patterns and foraging locations of female
northern fur seals. Fish. Bull. 89:171-179.
Hart, J. L.
1973. Pacific fishes of Canada. Bull. Fish. Res. Board Can.
180, 740 p.
Higgins, L. V., D. P. Costa, A. C. Huntley, and B. J. LeBoeuf
1988. Behavioural and physiological measurements of
maternal investment in the Steller sea lion, Eunietopias
jubatus. Mar Mamm. Sci. 4:44-58.
Hood, W. R., and K. A. Ono.
1997. Variation in maternal attendance patterns and pup
behaviour in a declining population of Steller sea lions
[Eumetopiasjubatus). Can. J. Zool. 75:1241-1246.
Homing, M., and F. Trillmich.
1997. Ontogeny of diving behavior in the Galapagos fiir seal.
Behaviour 134:1211-1257.
Keating, K. A.
1994. An alternative index of satellite telemetry location
error J. Wildl. Manage. 58:414-421.
Kooyman, G. L., J. O. Billups, and D. W. Farwell.
1983. Two recently developed recorders for monitoring
diving activity of marine birds and mammals. In Experi-
mental biology at sea (A. G. Macdonald and L G. Priede,
eds. ) p. 187-214. Academic Press, New York, NY.
Loughlin, T. R.
1997. Using the phylogeographic method to identify Steller
sea lion stocks. In Molecular genetics of marine mammals
(A. Dizon, S. J. Chivers, and W. F. Perrin, eds.), p. 159-171.
Spec. Publ. 3 of the Soc. Mar Mammal.
Loughlin, T. R, J. L. Bengtson, and R. L. Merrick.
1987. Characteristics of feeding trips of female northern fur
seals. Can. J. Zool. 65:2079-2084.
Loughlin, T. R., and R. L. Merrick.
1989. Comparison of commercial harvest of walleye pollock
and northern sea lion abundance in the Bering Sea and Gulf
of Alaska. In Proceedings of the international symposium
on the biology and management of walleye pollock, Nov.
14-16, 1988, Anchorage, Alaska, p. 679-700. Alaska Sea
Grant Rep. 89-01, Univ. Alaska, Fairbanks.
Loughlin, T R., A. S. Pedov, J. D. Baker, S. A. Blokhin, and
A. G. Makhnyr.
1998. Diving behavior of adult female Steller sea lions in the
Kuril Islands, Russia. Biosph. Cons. 1:21-31.
Loughlin, T. R., A. S. Perlov, and V. A. Vladimirov.
1992. Range-wide survey and estimation of total number of
Steller sea lions in 1989. Man Mamm. Sci. 83:220-239.
Loughlin, T. R., and T Spraken
1989. Use of Telazol to immobilize female northern sea
lion (Eunietopias jubatus) in Alaska. J. Wildl. Dis. 25:
353-358.
Loughlin, T R., and A. E. York.
2000. An accounting of the sources of Steller sea lion
mortality Mar. Fish. Rev 62(4):40-45.
McConnell, B. J., C. Chambers, and M. A. Fedak.
1992. Foraging ecology of southern elephant seals in relation
to the bathymetry and productivity of the southern ocean.
Antarctic Science 4:393-398.
McCullagh, P, and J. A. Nelder
1989. Generalized linear models, 2nd ed., 261 p. Chapman
and Hall, London.
Melin, S. R.
2002. The foraging ecology and reproduction of the Califor-
nia sea lion (Zalophus californianus californianus). Ph.D.
diss., 150 p. Univ. Minnesota, St. Paul, MN.
Merrick, R. L.
1995. The relationship of the foraging ecology of Steller sea
lions (Eunietopias jubatus) to their population decline in
Alaska. Ph.D. diss., 171 p. Univ Washington, Seattle, WA.
Merrick, R. L., R. Brown, D. G. Calkins, and T. R. Loughlin.
1995. A comparison of Steller sea lion, Eunietopias jubatus,
pup masses between rookeries with increasing and decreas-
ing populations. Fish. Bull. 94:753-758.
Merrick, R. L., and T R. Loughlin.
1997. Foraging behavior of adult female and young-of-the-
year Steller sea lions in Alaskan waters. Can. J. Zool. 75:
776-786.
Merrick, R. L., T. R. Loughlin, G. A. Antonelis, and R. Hill.
1994. Use of satellite-linked telemetry to study Steller
sea lion and northern fur seal foraging. Polar Res. 13:
105-114.
Merrick, R. L., M. K. Maminov, J. D. Baker, and A. G. Makhnyr.
1990. Results of US.-U.S.S.R. joint marine mammal re-
582
Fishery Bulletin 101(3)
search cruise in the Kuril and Aleutian Islands 6 June-24
July 1989. U.S. Dep. Commer. NOAA Tech. Memo. NMFS
F/NWC-177,63p.
Pitcher, W., and D. G. Calkins.
1981. Reproductive biology of Steller sea lions in the Gulf of
Alaska. J. Mammal. 62:599-605.
Porter, B.
1997. Winter ecology of Steller sea lions iEumetopias juba-
tus) in Alaska. M.S. thesis, 84 p. Univ. British Columbia,
Vancouver, B.C., Canada.
Raum-Suryan, K. L., K. W. Pitcher, D. G. Calkins, J. L. Sease, and
T. R. Loughlin.
2002. Dispersal, rookery fidelity, and metapopulation struc-
ture of Steller sea lions (Eumetopias jubatus) in an increas-
ing and declining population in Alaska. Mar. Mamm. Sci.
18:746-764.
Schreer, J. F., and K. T. Kovacs.
1997. Allometry of diving capacity in air-breathing verte-
brates. Can. J. Zool. 75:339-358.
Service-Argos.
1984. Location and data collection system user's guide, 36 p.
Service-Argos, Toulouse, France.
Sinclair, E. H., T. R. Loughlin, and W. Pearcy.
1994. Prey selection by northern fur seals (Callorhinus ursi-
nus) in the eastern Bering Sea. Fish. Bull. 92:144-156.
Sinclair, E. H., and T. K. Zeppelin.
2002. Seasonal and spatial differences in diet in the western
stock of Steller sea lions {Eumetopias jubatus). J. Mamm.
83:973-990.
Stewart, B. S., S. L. Leatherwood, P. K. Yochem, and
M. P. Heide-Jorgensen.
1989. Harborseal tracking and telemetry by satellite. Mar
Mamm. Sci. 5:361-375.
Trillmich, F, and K. A. Ono eds.
1991. Pinnipeds and El Nifio, responses to environmental
stress. Ecological studies 88, 293 p. Springer-Verlag,
Berlin.
Vincent, C, B. J. McConnell, V. Ridoux, and M. A. Fedak.
2002. Assessment of Argos location accuracy from satellite
tags deployed on captive gray seals. Mar. Mamm. Sci. 18:
156-166.
Werner, R., and C. Campagna.
1995. Diving behaviour of lactating southern sea lions (O/aria
flavescens) in Patagonia. Can. J. Zool. 73:1975-1982.
583
Abstract— Two halfbeak species, bal-
lyhoo {Hemiramphus brasiliensis) and
balao (//. balao), are harvested as bait
in south Florida waters, and recent
changes in fishing effort and regula-
tions prompted this investigation of
the overlap of halfbeak fishing grounds
and spawning grounds. Halfbeaks were
sampled aboard commercial fishing ves-
sels, and during fishery-independent
trips, to determine spatial and tem-
poral spawning patterns of both spe-
cies. Cyclic patterns of gonadosomatic
indices (GSIs) indicated that both
species spawned during spring and
summer months. Histological analysis
demonstrated that specific stages of
oocyte development can be predicted
from GSI values; for example, female
ballyhoo with GSIs >6.0 had hydrated
oocytes that were 2.0-3.5 mm diameter
Diel changes in oocyte diameters and
histological criteria demonstrated that
final oocyte maturation occurred over a
30- to 36-hour period and that ballyhoo
spawned at dusk. Hydration of oocytes
began in the morning, and ovulation
occurred at sunset of that same day;
therefore females with hydrated oocytes
were ready to spawn within hours. We
compared maps of all locations where
fish were collected to maps of locations
where spawning females (i.e. females
with GSIs >6.0) were collected to deter-
mine the degree of overlap of halfbeak
fishing and spawning grounds. We also
used geographic information system
(GIS) data to describe the depth and
bottom type of halfbeak spawning
grounds. Ballyhoo spawned all along
the coral reef tract of the Atlantic
Ocean, inshore of the reef tract, and in
association with bank habitats within
Florida Bay. In the Atlantic Ocean,
balao spawned along the reef tract and
in deeper, more offshore waters than did
ballyhoo; balao were not found inshore
of the coral reef tract or in Florida
Bay. Both halfbeak species, considered
together, spawned throughout the fish-
ing grounds of south Florida.
Spawning cycles and habitats for ballyhoo
{Hemiramphus brasiliensis) and balao (H. balao)
in south Florida
Richard S. McBride
Justin R. Styer
Rob Hudson
Florida Marine Research Institute
Florida Fish and Wildlife Conservation Commission
100 8th Avenue SE
St. Petersburg, Flonda 33701-5095
E-mail address (for R. S. McBride): nchardmcbnde^fwc-Statefl-US
Manuscript approved for publication
30 January 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull 101:583-589 (2003).
Combined landings of two halfbeaks
species, ballyhoo (Hemiramphus brasil-
iensis) and balao (H. balao), constitute a
small but valuable bait fishery in south
Florida (Berkeley et al., 1975; McBride
et al., 1996). Both species occupy coastal
pelagic habitat in association with
coral reefs (Starck, 1968; Nybakken,
1997). During the 1990s two changes
in the halfbeak fishery occurred that
caused concerns regarding the exploi-
tation levels in this fishery (McBride,
2001). First, geographic shifts occurred
when halfbeak fishing expanded from
the Atlantic Ocean into the nearshore
waters north of the Florida Keys, an
area known as Florida Bay. Second,
changes in statewide net fishing regu-
lations^ created concerns that the net
fishermen displaced from other fisheries
might preferentially enter the halfbeak
fishery, thereby increasing halfbeak
fishing effort. These two changes could
have specific consequences on halfbeak
reproductive output. For example,
because some fishermen viewed Florida
Bay as a spawning or nursery ground
for halfbeaks, it was of interest to learn
exactly how concentrated spawning
might be in Florida Bay and whether
spawning occurred outside Florida Bay.
In addition, because halfbeak landings
are dominated by a single species, bal-
lyhoo (Berkeley et al., 1975; McBride
et al., 1996), it could be argued that
these changes in the fishery could dis-
proportionally affect spawning by the
less abundant, and potentially more
vulnerable target species, balao.
Both ballyhoo and balao are distrib-
uted widely in the western and eastern
Atlantic Ocean (Collette, 1965), but
no study has defined their spawning
grounds. Berkeley and Houde (1978)
described both species to be small (<32
cm fork length) summer-spawners that
rarely live past two years, but in terms
of spatial coverage, they collected fish
principally from the Miami, Florida,
area. We reviewed reports on regional
ichthyoplankton collections (e.g. Powell
et al., 1989; Limouzy-Paris et al, 1994)
and found that the numbers of halfbeak
eggs and larvae were too few for char-
acterizing the spawning grounds. More-
over, Berkeley and Houde (1978) sug-
gested that standard ichthyoplankton
survey data would underestimate the
abundance of halfbeak eggs or larvae
for three reasons. First, halfbeak eggs
appear to attach to vegetation; therefore
oblique tows may not target the appro-
priate habitats (i.e. benthic or floating
vegetation) and halfbeak eggs would
be completely lost if pleuston was dis-
carded from ichthyoplankton samples.
Second, halfbeak eggs hatch 8-9 days
after fertilization and may disperse far
away from spawning locations. Third,
halfbeak larvae hatch at 5-7 mm and
have pigmented eyes; therefore they
appear capable of avoiding plankton
nets. Various plankton sampling strate-
■ This referendum (s. 16, Art. X of the Florida
Constitution, enacted July 1, 1995) prohib-
its entangling nets in waters inshore of 3
miles on the Atlantic coast and 9 miles on
the Gulf coast of Florida (including Florida
Bay). It also prohibits non-entangling nets
larger than 500 fl^ (such as those nets used
by commercial halfbeak fishermen), in
waters less than 1 mile of Florida's Atlan-
tic coast and 3 miles of the Gulf coast.
584
Fishery Bulletin 101(3)
gies could be developed to overcome these problems, but
we chose an alternative to plankton sampling as a way to
define halfbeak spawning grounds.
B2 00' 81 30'
&^^~
Atlantic
Ocean
82°00*W 8r30W 8100W 80'30'W SO'OO' W
82°0Q' 8130' 8100' 80''3Q' 80°00'
>00 m Isobath
Atlantic
Ocean
. i"> ■ ^ ^^
8200' W BV'30' W
B2 00' 81 '30'
00' W 80"30' W 80°00' W
iO'3Q' 60 00'
Allanllc
Ocean
B2 00 w erao'W eroo'W aoao'W 80"00'W
In this study, we used collections of adult ballyhoo and
balao to define each species" spawning grounds in south
Florida. Analyses of gonad histological preparations iden-
tified a discrete range of gonadosomatic indices (GSI) for
females that were ready to spawn within hours, and the
locations of these fishes were plotted by using geographic
information system (GIS) software (Arc View, version 3.3.,
Environmental Systems Research Institute, Inc., Redlands,
CA). This synthesis of GIS and GSI data was used to map
the spawning grounds of ballyhoo and balao.
Materials and methods
Sampling occurred throughout the south Florida com-
mercial halfbeak fishing grounds, from Palm Beach to
the Marquesas Keys (Fig. lA). The area immediately sur-
rounding Vaca Key was well sampled, but other sections
of the middle Florida Keys were not because commercial
fishermen used only six fishing ports and their day trips
were of limited range. Few samples were obtained from the
Palm Beach area because net fishing is no longer allowed in
much of this area (McBride, 2001). Halfbeak fishing trips
by commercial fishermen were monitored from November
1995 to April 1999 by an onboard biologist during as many
as four trips per month. A subsample of fish from the first
successful net (a lampara net) set, and occasionally from
later sets within a day, was obtained by filling a 5-gallon
bucket from the catch as it was transferred from the net
to holding boxes. This bucket held 100 to 200 halfbeaks,
and these fishes were kept on ice and brought back to the
laboratory for processing.
Fishery-independent collections were made by using cast
nets and small hooks (sabiki rigs) in the middle Florida
Keys. This sampling was specifically designed to include
inshore areas where lampara net fishermen could not fish
because of regulations associated with Florida's net limita-
tion referendum.' The target number of these fishery-inde-
pendent trips from July 1997 to October 1998 was four per
month, and the target sample size for each trip was twelve
fish. Additional fishery-independent sampling occurred in
the springs of 1997, 1998. and 1999.
In the laboratory, whole body weight was recorded to
the nearest 0.1 gram, and the gonads were removed and
weighed to the nearest 0.01 g. Sex was identified with the
aid of a dissecting binocular microscope (25-50x) when nee-
Figure 1
(A) Sampling area for halfbeaks {Hi'iniramphu.'! spp.) during
1995-99 in the Atlantic Ocean and in Florida Bay. Each symbol
represents an individual sample location where halfbeaks were
caught. Fishery-dependent samples (triangles) were taken from
commercial lampara net vessels. Fishery-independent samples
(squares) were collected in the middle Florida Keys, near Vaca Key
(not labeled because of the density of square symbols). Locations of
ripe female (B) ballyhoo (Hemiramphus brasiliensis) and (C) balao
(H. balao) are plotted separately. Ripe females have hydrated eggs,
and this condition was determined in B and C by a gonadosomatic
index >6.0 (see text for supporting evidence).
McBride et al.: Spawning grounds of Hemiramphus brasiliensis and H. balao
585
essary. Weights offish collected from July 1997 to October
1998 were measured for up to 30 females per species per
trip. Fish body and gonad weights were only occasionally
recorded for other trips during 1995-99, but these data
were included in the mapping of ripe females (i.e. females
with hydrated oocytes in ovigerous lamellae) to increase
overall sample size. In total, weight data were collected
for 2908 halfbeak females from 79 commercial fishing net
sets (63 different fishing days) and 59 fishery-independent
sampling events (50 different sampling days). Commercial
catches contributed 1649 ballyhoo and 757 balao females,
and fishery-independent collections added another 497 bal-
lyhoo and 5 balao females. The gonadosomatic index (GSI)
was calculated as
GSI=(GWnBW-GW)) X 100,
locations were recorded by using a global positioning sys-
tem hand-held unit. The latitude and longitude of fishery-
dependent samples were taken onboard the fishing vessel
once the lampara net enclosed the fish. Location data for
fishery-independent samples were taken from an anchored
position. Depth information was divided into the following
categories: area exposed at low tide, 0-1 m, 1-2 m, 2-4 m,
4-6 m, 6-10 m, and 10-20 m; only one of 159 locations was
without depth information. Substrate information was di-
vided into one of the following categories; platform margin
reefs, patch reefs, other hard bottom, seagrass beds, and
bare substrate; 32 of 159 locations did not have substrate
information.
Results
where GW = gonad weight; and
BW = body weight.
The processes of final oocjrte maturation (FOM) were
examined by comparing GSIs with changes in whole
oocyte size and histological criteria. Oocyte diameters
were measured for 39 ballyhoo collected in April 1998 and
March-April 1999. Fixed ovary tissue was washed, teased
apart, and placed in a solution of 33% glycerin to 67% water
Measurements of at least 300 oocytes per fish were made
to the nearest micron with the aid of a video system and
image-analysis software (Optimas, vers. 100, Media Cyber-
netics, Inc., Silver Spring, MD). A minimum-size cut-off of
0.15 mm was used to exclude debris within the petri dish.
Initially, oocyte diameters from six ballyhoo were measured
from four separate sections of ovaries (left, right, anterior,
posterior), but the modal oocyte diameter within each in-
dividual was the same for all four sections; therefore tissue
from other fish was extracted without regard to location
within the ovary. Berkeley and Houde (1978) performed a
similar test and came to a similar conclusion. Ovaries from
fish {n=930 females) collected during March and May 1997,
July 1997-October 1998, and March-April 1999 were pre-
pared for histological examination. Histological methods
are presented in McBride and Thurman (2003). Here, the
most advanced oocyte stage was recorded (in increasing
order of oocyte maturity) as either perinucleolar, cortical al-
veolar, vitellogenic. or as either of two stages for oocytes in
final maturation; nucleus migration and hydration (West,
1990 ). McBride and Thurman ( 2003 ) have reported the size
at 50% maturity to be >160 mm FL for female balao and
>198 mm FL for female ballyhoo (approximately 31.5 gand
60.9 g, respectively, using length-weight relationships from
Berkeley and Houde [1978; their Fig. 71). Mean GSIs and
95% confidence limits were determined for fish with regard
to their most advanced stage of oocyte development, and
a minimum cut-off value was established for GSI values
indicating ripe females.
The locations of ripe females were plotted to indicate
spawning grounds. Water depth and bottom type of these
spawning locations were determined by using the Marine
Resources Geographic Information System at the Florida
Marine Research Institute (w-ww.floridamarine.org). Point
Spawning cycles
Ballyhoo and balao had prolonged spawning seasons that
peaked in late spring and early summer (Fig. 2). Monthly
average GSIs of mature females increased from a low of
<0.4 for both species to a high of 6.4 for balao and 6.9 for
ballyhoo. The average GSI of individual females with only
primary growth oocytes (i.e. their most advanced oocyte
stages were perinucleolar or cortical alveolar) fell within a
narrow interval of 0.1-0.3 (maximum=0.95, Fig. 3). These
females were either small fishes that were immature or
they were larger fishes that were regressed (i.e. mature
but inactive). Vitellogenesis more than doubled the aver-
age GSI values for both species, but all females whose
most advanced oocyte stage was vitellogenic had GSIs less
than 1.37. Dramatic increases in GSI values also occurred
during FOM, and significant differences were evident in
the sequential FOM steps of nucleus migration and nucleus
breakdown. During nucleus migration, but before hydra-
tion, ballyhoo GSIs averaged 3.4 (3.3-3.6; 95% CD and
balao GSIs averaged 5.2 (4.6-5.9). Females with hydrated
oocytes had GSIs averaging 7.4 (7.0-7.9; 95% CD for
ballyhoo and 8.7 (6.6-10.8) for balao. Individual female
GSIs reached an observed maximum of 13.3 for ballyhoo
and 14.2 for balao. By applying these GSI criteria, which
indicate that females with a GSI greater than about three
had oocytes in FOM, it is evident that the average mature
halfbeak female is actively spawning from at least March
to August.
Final oocyte maturation also followed a diel cycle. For
ballyhoo, FOM began about 30-36 hours before ovulation,
hydration of oocytes began about 8-12 hours before ovula-
tion, and ovulation occurred at sunset. Ballyhoo oocytes de-
veloped in a group-synchronous pattern, and during FOM,
a batch of oocytes increased rapidly in diameter (Fig. 4).
Mature female ballyhoo had a bimodal or trimodal distri-
bution of oocyte diameters when spawning. The smallest
mode (<1.0 mm oocyte diameter) represented a reservoir
of primary growth oocytes and vitellogenic oocytes. Larger
modes, between 1 and 3 mm, represented oocytes in FOM.
The presence of two larger ooctye modes (-1.0-2.0 and >2.0
mm) in all females sampled during the afternoon period in-
dicated that female ballyhoo typically spawn every day dur-
586
Fishery Bulletin 101(3)
ing March-April. In total, the trimodal oocyte
frequency represents a reservoir of oocytes prior
to FOM, one batch of oocytes beginning FOM,
and one batch completing FOM. Hydrated oo-
cytes were not observed in balao prior to 1 100 h.
However, in several ballyhoo collected around
dawn (i.e. at approximately 0600-0700 EST),
the nucleus in oocytes of the advanced batch
was still visible along the chorion but the cy-
toplasm was lightening in color. This suggested
that initiation of hydration at daybreak briefly
preceded nucleus breakdown. During the fol-
lowing 12-hour period, oocytes in this maturing
clutch advanced from late nucleus migration to
nucleus breakdown, and increased in diameter
from 1.5-2.0 mm in the morning to 2.0-3.0 mm
in the afternoon. Modal egg size for each of three
running-ripe ballyhoo (i.e. females with hydrat-
ed, ovulated eggs in the ovarian lumen) was
2.35, 2.60, and 2.80 mm diameter The complete
size range of these ovulated eggs was 2.2-3.4
mm diameter These females were collected at
or just before sunset (time: 1810-1855). Most
efforts to sample across the full 24-hour cycle
failed, apparently because halfbeak do not bite
hooks after sundown. A sample of 12 ballyhoo
was collected one night, however, by randomly
throwing a cast net on dense schools of fish.
These fish, collected between 2200 and 2359
hours during March 1997, all appeared to have
recently spawned. Histological preparations
demonstrated that they had fresh postovulatory
follicles, and they contained a distinct clutch of
oocytes in early nucleus migration. Whole oo-
cytes from these fish collected at night were not
archived in formalin; therefore they were not
measured for comparisons to whole oocytes col-
lected at other times during the diel cycle. These
patterns of diel reproductive periodicity also ap-
peared to apply to balao, but the available data
were not conclusive.
Spawning habitat
o
a
-Ballyhoo, n=1,791
Female GSI
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct
1997 Month 1998
Figure 2
Mean (±95'^ confidence limits) gonadosomatic indices by month for bal-
lyhoo (Hemiramphus brasiUensis) and balao (H. balao) females. Values
are calculated for fish larger than size at 50% maturity, reported by
McBride andThurman (2003) as >198 mm fork length (PL) for ballyhoo
and >160 mm FL for balao. n = number of females.
Perinucleolar Cortical Vitellogenic Nucleus Hydration
154/7 Alveolar 24/8 Migration -|20/7
159/24 360/57
Most advanced ooctye stage
Figure 3
Mean (±95% confidence limits) gonadosomatic indices of female bally-
hoo (Hemiramphus brasiUensis) and balao (H. balao) in various stages
of oocyte development. A two-way ANOVA demonstrated a significant
effect of both developmental stage and species (P<0.0001 ) on gonadoso-
matic index. Numbers indicate the number of fish, by stage, for each
species. Hydration = nucleus breakdown.
In our study it was shown that hydrated oocytes
can be inferred from a threshold criterion of GSI >6.0
(Fig. 3), and ripe females (i.e. with a batch of hydrated
oocytes) will spawn within hours. Ripe ballyhoo females
were distributed throughout the fishing grounds in both
the Atlantic Ocean and Florida Bay (Fig. IB). In the Atlan-
tic, ripe ballyhoo females were caught in water depths from
1 to 20 m (mode: 6-10 m, 36.3'7( of the sets containing ripe
ballyhoo in the Atlantic Ocean). In Florida Bay, ripe bal-
lyhoo females were caught in areas that were exposed at
low tide and out to 6-m deep (mode: 2-4 m; 57.9'7f of the
positive sets in Florida Bay). Ripe ballyhoo females were
mainly associated with hard bottom or vegetated habitats
in both areas. In the Atlantic Ocean, ripe ballyhoo females
were collected above platform reefs in 51.79i of the sets,
above seagrass beds in 37.9% of the sets, near patch reefs
in 5.2% of the sets, and over bare substrate in 5.2% of the
sets. In Florida Bay, these fish were also associated with
hard bottom substrates, specifically with vegetated bank
habitat, in 44.77r of the sets and with seagrass beds in
55.3% of the sets.
Ripe balao females were distributed throughout the
Atlantic fishing grounds but not in Florida Bay (Fig. IC).
In the Atlantic, they tended to occur in deeper water than
did ripe ballyhoo females (range: 2-20 m; mode: 10-20 m,
51.37f of the sets containing ripe balao). The habitat as-
sociations of ripe balao females were similar to those of
ripe ballyhoo females in the Atlantic Ocean, but typically
reflected areas offshore rather than inshore of the reef
In the Atlantic Ocean, ripe balao females were collected
above platform reefs in 58.0% of sets, above seagrass beds
McBride et al.: Spawning grounds of Hemiramphus braslllensis and H. balao
587
n=3
1542- 1547 h
00 05 10 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Oocyte diameter (mm)
Figure 4
Oocyte diameters for ballyhoo [Hemiramphus brasiliensis) collected at
different times of the day (EST). At least 300 oocytes were measured per
female. Sizes <0.25 mm were not measured representatively because of the
minimum size criteria of the image analysis software. Sunset occurred at
1830-1900 EST during this collection period (March-April), n = number
of females.
in 25.8% of sets, over bare substrate in 9.7% of sets, and
above undefined hard bottom in 6.5% of sets.
Discussion
These detailed findings of prolonged summer-spawning sea-
sons, extreme iteroparity, and diel reproductive periodicity
are consistent with other studies of halfbeak reproductive
biology. Graham (1939), Ling (1958), Talwar (1962, 1967),
and Berkeley and Houde ( 1978) noted a protracted spawn-
ing season by hemiramphids during warm months. McBride
and Thurman (2003) examined the frequency of postovula-
tory follicles and reported that both species spawn daily
during late spring and early summer, but also that some
portion of the ballyhoo population spawns year-round. The
present study is the first to follow the diel progression of
FOM within the family Hemiramphidae. Lunar periodicity
was not evident but it may have been confounded by the
highly iteroparous nature of both species.
Spawning halfbeaks were distributed so widely through-
out the fishing grounds that no specific areas were identi-
fied for the protection of spawning individuals. We noted
interspecific differences in spawning areas, but these are
not necessarily related to preferences by spawning females
per se. Instead these differences appeared to be the result
of interspecific distribution patterns of adult halfbeaks in
general (i.e. adult ballyhoo are a more inshore species com-
pared to adult balao [McBride, pers. obs.] ). Because balao
were not found in Florida Bay, fishing in Florida Bay does
not affect this species. Spawning by ballyhoo was evident
in Florida Bay, as predicted by fishing industry partici-
pants, but spawning ballyhoo were also widespread along
south Florida's coral reef tract. Existing, albeit recent,
588
Fishery Bulletin 101 (3)
regulations^ should provide some measure of protection
for spawning ballyhoo in inshore waters.
Our study design was limited to the presence and ab-
sence of spawning females and did not identify concentra-
tions of spawning activity associated with specific habitats.
Presumably submerged vegetation is an important micro-
habitat. Several authors have noted that hemiramphid
eggs, including those of ballyhoo, attach by filaments (of
the chorion) to vegetation such as Syringoditim fill forme
and Sargassum sp. in waters less than approximately 6 m
deep (Graham, 1939; Ling, 1958;Talwar, 1962, 1967; Berke-
ley and Houde, 1978). However, Berkeley and Houde ( 1978)
collected eggs in plankton tows. The specific importance for
halfbeak reproductive success of attached versus floating
vegetation, or no vegetation, has not been identified.
The methods of this study define the macroscale spawn-
ing habitat of halfbeaks based on the distribution of spawn-
ing females. We demonstrate here that GSl values, even for
highly iteroparous species, can distinguish females with
hydrated oocytes from females in a less advanced stage
of oocyte development. The GSI value is simple and inex-
pensive to measure, and by including individual halfbeaks
for which we had GSI values but no histological data, we
more than tripled our sample size with little additional
laboratory cost. We could have instead characterized oo-
cyte development macroscopically and such a modification
is well suited when conditions affect weighing devices.
But macroscopic characterization of oocyte development
usually follows an ordinal scale that may vary between
observers.
The distribution of females with hydrated eggs may be
a better indication of spawning habitat than the distribu-
tion of eggs because hydration occurs for only a few hours
(DeMartini and Fountain, 1981; Hunter and Macewicz,
1985; Brown-Peterson et al., 1988; McBride et al., 2002),
whereas egg dispersal may occur over several days. In
this study we assumed that spawning females move only
limited distances within the few hours of the hydration
process, and although limited movement has not been
documented for either ballyhoo or balao, we believe that
our interpretation of the data supports this assumption.
The size of the study area was approximately 200 km by
250 km, and it seems reasonable that spawning halfbeaks
were not moving extensively within this spatial boundary
on an hourly basis. The approach discussed in the present
study may meet the needs of other investigators wanting
to generate a first approximation of spawning habitats for
management purposes, which was the goal of this study.
Also, this approach has good potential for use in areas were
species identification of halfbeak eggs or larvae is problem-
atic (Noell et al., 2001). Analyses requiring a smaller area
or finer spatial resolution will depend on verification of a
hydration period that is short in relation to expected fish
movements.
The specific example presented in our study was limited
because we collected the fish using commercial fishing
vessels on routine fishing operations. This was cost-effec-
tive, but wo were not able to identify spawning habitat
preference or to define the complete geographic extent of
the spawning grounds within south Florida. Gaps in the
distribution of ripe females, which were particularly evi-
dent in the middle Florida Keys, were typically related to
gaps in sampling coverage. In addition, both species pre-
sumably spawn outside the area we sampled. Still, much
of the reported geographic range of ballyhoo and balao in
the western Atlantic Ocean has been covered in the pres-
ent study. The remaining shortcomings of this specific
example could be resolved by using this approach within
a statistically valid sampling design and estimating size-
specific batch fecundity to map reproductive rates within a
spatial and temporal context. The data resulting from such
a comprehensive sampling design would be well-suited for
identifying essential spawning habitat, for siting habitat-
specific investigations of spawning dynamics, or for validat-
ing dispersal models for early life stages of marine fish.
Acknowledgments
We are grateful to many individuals for assistance in this
research. First, to the fishermen and processors in the
halfbeak fishery, all of whom participated in this survey. T.
Brown, J. Hunt, and R. Moretti provided logistical support
in Marathon. R. Beaver, K. Krumm, E. Robillard, D. Snod-
grass, and J. Whittington assisted in fish collection and
processing. G. Gerdeman, P. Nagle, F Stengard, C. Stevens,
and P. Thurman assisted with tissue processing and repro-
ductive staging. C. Anderson assisted in preparing Figure
1 and GIS habitat analyses. B. Mahmoudi, R. Taylor, M.
Zimmermann, and two anonymous reviewers provided
constructive comments. Editorial assistance was provided
by J. Leiby and J. Quinn. This research was funded in part
by a grant from the National Oceanic and Atmospheric
Administration (NOAA) to the Florida Fish and Wildlife
Conservation Commission (Saltonstall-Kennedy Program,
NOAA award no. NA77FD0069).
Literature cited
Berkeley, S. A. and E. D. Houde.
1978. Biology of two exploited species of halfbeaks, Hemir-
amphus brasiliensis and H. balao from southeast Florida.
Bull. Mar. Sci. 28:624-644.
Berkeley, S. A., E. D. Houde. and F. Williams.
1975. Fishery and biology of ballyhoo on the southeast Flor-
ida coast. In Sea Grant Special Report 4, 1-15 p. Univ.
Miami Sea Grant Program, Coral Gables, FL.
Brown-Peterson, N., P. Thomas, and C. R. Arnold.
1988. Reproductive biology of the spotted seatrout.CvMosf 10?!
nebiihsiis, in south Texas. Fish. Bull. 86:.373-88.
Collette, B. B.
1965. Hemiramphidae (Pisces, Synentognathi) from tropical
west Afinca. Atlantidc Report 8:217-235.
DeMartini, E. E., and R. K. Fountain.
1981. Ovarian cycling frequency and batch fecundity in the
queenfish, Seriphus politiis: attributes representative of
serial spawning fish. Fish. Bull. 79:547-60.
Graham, D. H.
1939. Breeding habits of the fishes of Otago Harbour and
adjacent seas. Trans. Proc. Royal Soc. New Zealand 69:361 -
372.
McBride et al : Spawning grounds of Henvramphus biasiliensis and W. balao
589
Hunter, J. R. and B. J. Macewicz.
1985. Measurement of spawning frequency in multiple
spawning fishes. In An egg production method for esti-
mating spawning biomass of pelagic fish: application to
the northern anchovy {Engraulis mordax) (R. Lasker, ed.),
79-94 p. U.S. Dep. Commer.. NOAA Tech. Rep. NMFS 36.
Limouzy-Paris, C. B., M. F. McGowan, W. J. Richards,
J. P. Umaran, and S. S. Cha.
1994. Diversity of fish larvae in the Florida Keys: results
from SEFCAR. Bull. Mar. Sci. 54:857-870.
Ling, J. K.
1958. The sea garfish, Reporbamphiis melanochir (Cuvier
& Valenciennes) (Hemiramphidae), in South Australia:
breeding, age determination, and growth rate. Aust. J.
Mar. Fresh. Res. 9:60-110.
McBride, R. S.
2001. Landings, value, and fishing effort for halfbeaks,
Hemiramphus spp., in the south Florida lampara net
fishery. Proc. Gulf Carib. Fish. Inst. 52nd Ann. Meeting,
Key West, FL.
McBride, R. S., L. Foushee, and B. Mahmoudi.
1996. Florida's halfbeak, Hemiramphus spp., bait fishery.
Mar Fish. Rev 58:29-38.
McBride. R. S., F. J. Stengard, and B. Mahmoudi.
2002. Maturation and diel reproductive periodicity of round
scad (Carangidae: Decapterus punctatus). Mar. Biol. 140:
713-722.
McBride, R. S., and R E. Thurman.
2003. Reproductive biology of Hemiramphus brasilieiisis
and H. balao (Hemiramphidae): maturation, spawning
frequency, and fecundity. Biol. Bull. 204:57-67.
Noell, C. J., S. Donnellan, R. Foster, and L. Haigh.
2001. Molecular discrimination of garfish Hyporhamphus
(Beloniformes) larvae in southern Austrailian waters.
Mar. Biotechnol 3: 509-514.
Nybakken, J. W.
1997. Marine biology: an ecological approach, 4th ed., 481 p.
Addison Wesley Longman, Inc., Menlo Park, CA.
Powell, A. B., D. E. Hoss, W F Hettler, D. S. Peters, and
S. Wagner.
1989. Abundance and distribution of ichthyoplankton in
Florida Bay and adjacent waters. Bull. Mar. Sci. 44:
35-48.
Starck,W.A.,Jr.
1968. A list of fishes of Alligator Reef, Florida, with com-
ments on the nature of the Florida reef fish fauna. Un-
dersea Biol. 1:4-40.
Talwar, P K.
1962. A contribution to the biology of the halfbeak, Hyporh-
amphus georgu (Cuv. & Val.) (Hemirhamphidae) [s;c|. In-
dian J. Fish. 9:168-196.
1967. Studies on the biology of Hemirhamphus [sic] mar-
ginatus ( Forsskal ) ( Hemirhamphidae-Pisces). J. Mar. Biol.
Assoc. India 9:61-69.
West, G.
1990. Methods of assessing ovarian development in fishes: a
review. Aust. J. Mar. Freshwater Res. 41:199-222.
590
Abstract— Tope shark iGaleorhi-
nus galeus) and thornback ray {Raja
clavata) are the two most captured
elasmobranch species by the Azorean
bottom longline fishery. In order to
better understand the trophic dynam-
ics of these species in the Azores, the
diets of thornback ray and tope shark
caught in this area during 1996 and
1997 were analyzed to describe feed-
ing patterns and to investigate the
effect of sex, size, and depth and area
of capture on diet. Thornback rays fed
mainly upon fishes and reptants, but
also upon polychaetes, mysids, natant
crustaceans, isopods, and cephalopods.
In the Azores, this species preyed more
heavily upon fish compared with the
predation patterns described in other
areas. Differences in the diet may be
due to differences in the environments
(e.g. in the Azores, seamounts and oce-
anic islands are the major topographic
features, whereas in all other studies,
continental shelves have been the
major topographic feature). No differ-
ences were observed in the major prey
consumed between the sexes or between
size classes (49-60, 61-70, 71-80, and
81-93 cm TL). Our study indicates that
rays inhabiting different depths and
areas (coastal or offshore banks) prey
upon different resources. This appears
to be related to the relative abundance
of prey with habitat. Tope sharks were
found to prey almost exclusively upon
teleost fish: small shoaling fish, mainly
boarfish (Capros aper) and snipefish
{Macroramphosus scolopax), were the
most frequent prey. This study illus-
trates that thornback rays and tope
sharks are top predators in waters off
the Azores.
Diets of thornback ray {Raja clavata) and tope shark
iGaleorhinus galeus) in the bottom longline fishery
of the Azores, northeastern Atlantic
Telmo Morato
Encarnacion Sola
Maria P. Gros
Gui Menezes
Departamento de Oceanografia e Pescas
Universidade dos A(;ores
PT-9901-862 Horta, Portugal
E-mail address: telmo@noles horta uac pt
Manuscript approved for publication
19 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:590-602 (2003).
The thornback ray {Raja clavata L.), is
a shallow water bottom-living elasmo-
branch found in the Atlantic from Ice-
land and Norway southwards to South
Africa, including Madeira and Azores
islands. This species is also found in the
Mediterranean, western Black Sea, and
southwestern Indian Ocean (Stehmann
and Burkel, 1984). The thornback ray
is commercially exploited in several
countries. In the Azores it is a bycatch
of the bottom longline fishery directed
toward demersal and deepwater teleost
species. Food and feeding habits of the
thornback ray have been intensively
studied since the end of the 19'^^ century
(e.g. Day, 1880-84) and more recently
(e.g. Smale and Cowley, 1992; Ellis et
al., 1996; Daan et al.'l. However, only
two studies have been conducted on
the thornback ray off Portuguese conti-
nental waters (Marques and Re, 1978;
Cunha et al., 1986), and none exists for
populations inhabiting waters around
the oceanic islands or seamounts in the
northeastern Atlantic.
The tope shark {Galeorhinus galeus
(L.)), is a cosmopolitan species that
can be found from about 70°N to about
55"S. Distribution of this species in-
cludes the Atlantic, Pacific and Indian
Oceans (Compagno, 1984). Tope shark
is also commercially exploited by sev-
eral countries around the world, includ-
ing the Azores, where it is a bycatch of
the bottom longline fishery. Clompagno
(1984) and Olsen (1984) reviewed the
biology of this shark; however, there
have been relatively few studies on
their feeding habits. The diet of tope
shark was described by Ford (1921) for
individuals landed at Plymouth U.K., by
Olsen ( 1954 ) in southeastern Australia,
and by Ellis et al. (1996) in the north-
eastern Atlantic Ocean.
Elasmobranchs are among the top
predators in marine environments ( Ellis
et al., 1996); thus they affect the popu-
lations of both fish and invertebrates
at lower trophic levels. However, feed-
ing studies of elasmobranches in the
Azores have been limited to the blue
shark (Prionace glaiica) (Clarke et al.,
1996). Tope shark and thornback ray
are the two most abundant elasmo-
branch species landed by the Azorean
bottom longline fishery. Information on
the feeding habits of these two species
contributes to a better understanding
of trophic dynamics and food webs — in-
formation which is needed as fisheries
scientists advance ecosystem principles
to fisheries management (Pauly et al.,
2000; Pitcher, 2000; Whipple et al.,
2000). The purpose of this study was
to examine the diet of thornback ray
and tope shark, to describe their feed-
ing patterns and the effect of sex, size,
depth, and location on their diet.
Materials and methods
Thornback rays and tope sharks were
collected between March and May
(spring) of 1996 and 1997 during a
' Daan, N., B. Johnson, J. R. Larsen and
H. Sparholt. 1993. Analysis of the ray
[Raja spec.) samples collected during the
1991 international stomach sampling
project. ICESC.M. 1993/0:15,17 p.
Morato et al.: Diets of Rqa davata and Galeorhinus galeus
591
39° N
3rw
CORVO
P
FLORES
c
39° N
38° N
N
t
28° W
27° W
GRACIOSA
r~ -t&
• ^r^ TERCEIRA
S. JORGE ^"^"^^^^A *
, FAIAL ^^^V^ ^^^^^^^^
^^ ^^k^ ^^^PICO Banco D Jo^o de Ci
9H^ 4^^^C ^ Baixo de S Mateus ■
Banco Pri^sa Alice ,6° W 25"W •
^^^^^ 38° N—^ •—
than 60n m ddoth ^
► Less than 600 m depth
30° W
28" W
37° N
39° N
38° N
Mar da Pra^B f
'4
24° W
Figure 1
Locations of the longline sets made in the Azores during the spring of 1996 (•) and 1997 (D).
study on demersal fisheries in Azorean waters (Fig. 1).
Fishes were caught by longhne onboard the RV Arquipe-
lago. Line setting began before sunrise (approx. 05:00 hi
and hauhng started about two hours after setting. From
the fish sampled, total length (TL, to the nearest cm) was
measured, and sex and maturity were determined by mac-
roscopic examination of gonads and claspers with maturity
scales, as proposed by Stehmann (1987). Stomachs were
removed and classified as either everted, regurgitated,
with bait, empty, or with contents. Individuals falling in
any of the first three categories, as well as those that had
obviously eaten fish hooked on the longline, were excluded
from further analysis. Stomachs with contents were placed
in plastic bags and frozen (within about 2 h of capture)
for subsequent analysis. Stomach contents, which partly
consisted of a turbid suspension, were washed with water
in a nylon net of approximately 0.5-mm mesh size to allow
easier examination. The items were carefully separated,
weighed (after removing the surface water by blotting
them in tissue paper), and identified to the lowest possible
taxonomic level. Individuals of each identified taxon were
counted. Whenever fragments were found, the number of
individuals was taken as the smallest possible number of
individuals from which fragments could have originated.
Precision estimates in diet studies have been advocated
and used by several authors (Ferry and Cailliet, 1996;
Morato et al., 1999). We used the cumulative trophic diver-
sity, measured with the Shannon-Wiener index [as H'=— Z
P,(log^P|), where P, is the proportion of individuals in the
;th species] to measure sample size sufficiency (Hurtubia,
1973). Cumulative numbers of randomly pooled stomachs
were plotted against the cumulative trophic diversity. The
asymptote of the curve indicates the minimum number of
stomachs required. Frequency of occurrence (%0), percent-
age number (%N), and weight (%W) for each prey type were
used to describe the diet of both species (for a review see
Hyslop, 1980; Cortes, 1997). Wet weight was used to de-
termine the latter value. The index of relative importance
592
Fishery Bulletin 101(3)
E t/5
o
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0.0
[IRI={%N + %W) X %0] (Pinkas
et al., 1971) and the 9(IRI (as
%77?7, = 100 X IRIJ1IRI,) were
calculated for each prey category
and used in diet comparisons.
Prey taxa occurring in less than
five stomachs were grouped into
higher taxonomic categories.
Ontogenetic differences in the
diet of thornback rays were
examined by grouping fish into
four size classes (49-60, 61-70,
71-80, and 81-93 cm TL). The
diet of thornback rays was also
analyzed by sex, depth (0-100,
101-200, 201-350 m), and area
of capture (coastal areas and off-
shore banks). No further analy-
ses were performed for tope
shark because their diet was
dominated by only one prey cat-
egory (see "Results" section). To
determine if the most important
preys were similar for different
groups of rays, weighted corre-
lation and concordance analyses
were used (Zar, 1999). These methods were preferred to
conventional rank correlation methods (e.g. Spearman)
because they emphasize the high ranking given to the
most important prey categories. Differences in the rank-
ings of IRI values for prey categories between three or more
groups (e.g. three size classes) were tested for significance
with the top-down concordance method iC\= top-down
concordance coefficient) (Zar, 1999). For paired groups (e.g.
males and females) the top-down correlation method (r,p=
top-down correlation coefficient) was used (Quade and
Salama, 1992; Zar, 1999). Schoener's dietary overlap index
(Schoener, 1970) (as C„= 1-0.5 1 1 P^, - P^, | , where P^- was
the proportion (based on %IRI ) of food category / in the diet
ofx; and P^,, was the proportion of food category i in the diet
ofy) was used to measure the diet overlap between sex, size
classes, depth strata, and area of capture.
Cluster analysis was used to describe geographic simi-
larities in the feeding habits of thornback rays. A preda-
tor-prey matrix was built from published data. When more
than one index was available, the following criteria were
used to choose between indexes: IRI or '/< IRI, ^O, %N, %W,
%Volume. The number of prey categories included was
based on the quality of the description found in the pub-
lished sources. Eleven different categories were obtained.
A distance matrix was then calculated by using Euclidean
distance, and the hierarchical form of analysis was applied
(Clarke and Warwick, 1994 ). The grouping of predators was
based on the "average linkage method," and a dendrogram
was used as a graphic form of representation. Finally, tro-
phic levels (TLd^) were estimated for each of the samples {k)
by using the method proposed by Cortes ( 1999) [as TLVf^=l+
'I^,* X TLu^t, where TLi', is the trophic level of each prey
category as estimated by the author, P,^, is the proportion of
prey category i in sample k]. Mean trophic levels were also
Raja clavala
Galeorhimts gateus
0 10 20 30 40 30 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
Number of stomachs witti contents
Figure 2
Randomized cumulative trophic diversity curves for thornback ray (Raja clavata) and
tope shark (Galeorhinus galeus) .
estimated for groups resulting from the cluster analysis,
and differences between them were tested by using one-
way ANOVA (Zar, 1999).
Results
Thornback rays were caught at depths ranging from 10 to
350 m, but primarily (95%) shallower than 250 m. Out of
237 stomachs examined, the contents of four appeared to
have been regurgitated (1.7%), seven contained bait only
(2.9%), 88 were empty (37.1%), and 138 contained prey
(58.2%). Rays with stomachs containing food measured
from 49.0 to 93.0 cm TL. All tope sharks were caught
between 10 and 150 m depth, except for one individual
taken at 300 m. Out of 365 stomachs examined, 174 (47.7%)
were empty, seven ( 1.9% ) contained fish hooked on the long-
line and 184 stomachs (50.4% ) contained prey. Sharks with
stomachs containing food ranged from 58.0 to 153.0 cm
TL. The cumulative trophic diversity curves of both spe-
cies appeared to reach an asymptote, suggesting that a
sufficient number of stomachs were analyzed for both the
thornback ray and tope shark (Fig. 2).
Thornback ray
The main diet components of thornback rays were fish
(%IRI=81 .6) and crustaceans reptants (%IRI=17.4) (Fig. 3).
Fish occurred in 84.1% of stomachs that contained food,
and represented 78.0% of total prey weight and 50.2% of
total prey number (Table 1). Two benthopelagic species,
the snipefish {Macroramphosus scolopax (%IRI=34.01) and
the boarfish (Capros aper (%IRI=26.8]), were by far the
predominant fish prey items. However, some pelagic fish
Morato et al,: Diets of Raja clavata and Galeorhinus galeus
593
Table 1
Values for percentage by number C/rN), weight C^W), occurrence CXO
, and index of relative importance (IRI and fr
[RD for prey
items observed in stomachs (n = 138) of thomback rays
(Raja clavata)
caught off the Azores during the spring of 1996 and 1997. |
Total values are given in bold font.
Prey items
'7.N
%.w
%0'
IRI
'/rlRI
Algae
0.3
0.0
1.5
0.5
0.0
Bivalvia — Chlamys sp.
0.1
0.0
0.7
0.1
0.0
Total Cephalopoda
1.1
1.1
5.1
11.2
0.1
Octopodoidea unidentified
0.1
0.1
0.7
0.1
0.0
Scaeurgus unicirrhus
0.7
0.8
2.9
4.4
0.1
Cephalopoda unidentified
0.3
0.2
1.5
0.8
0.0
Total Polychaeta
3.4
0.8
9.4
39.5
0.8
Hirudinea
0.1
0.0
0.7
0.1
0.0
Crustacea
Stomatopoda
0.1
0.0
0.7
0.1
0.0
Total Natantia
3.1
1.0
10.1
41.4
0.3
Penaeidea unidentified
1.4
0.3
2.9
4.9
0.1
Solenocera membranacea
0.1
0.1
0.7
0.1
0.0
Solenocera sp.
0.1
0.1
0.7
0.1
0.0
Pandalidae
0.3
0.1
0.7
0.3
0.0
Processa intermedia
0.1
0.0
0.7
0.1
0.0
Processa sp.
0.1
0.0
0.7
0.1
0.0
Caridea unidentified
0.1
0.0
0.7
0.1
0.0
Natantia unidentified
0.9
0.4
2.9
3.8
0.1
Total Reptantia
31.9
17.0
47.1
2303.2
17.4
Anomura unidentified
0.1
0.1
0.7
0.1
0.0
Scyllaridae Scyllarus arctus
4.0
0.8
9.4
45.1
0.9
Diogenidae
1.1
1.7
5.8
16.2
0.3
Paguridea
0.3
0.3
1.5
0.9
0.0
Galatheidae Galathea sp.
0.3
0.1
1.5
0.6
0.0
Homolidae Paromola cuvieri
0.6
0.1
2.9
2.0
0.0
Calappidae Calappa granulata
1.8
1.3
7.3
22.6
0.5
Parthenopidae Parthenope sp.
2.8
0.7
0.7
2.5
0.0
Portunidae
0.1
0.0
0.7
0.1
0.0
Total Liocarcinus spp.
14.9
8.3
16.6
385.1
5.5
Liocarcinus marmoreus
9.8
5.1
9.4
140.1
2.8
Liocarcinus corrugatus
3.8
2.7
6.5
42.3
0.8
Liocarcinus spp.
1.3
0.5
2.2
4.0
0.1
Brachyura
4.1
2.4
11.6
75.4
1.5
Reptantia unidentified
0.6
0.4
2.9
2.9
0.1
Decapoda unidentified
0.1
0.0
0.7
0.1
0.0
Total Mysidacea
6.6
0.7
3.6
26.3
0.5
Isopoda
1.6
0.3
5.1
9.7
0.2
Amphipoda-y/fef/m sp.
0.1
0.0
0.7
0.1
0.0
Crustacea unidentified
1.1
0.8
4.4
8.4
0.2
Total Pisces
50.2
78.0
84.1
10811.2
81.6
Myctophidae
0.6
0.3
2.2
2.0
0.0
Moridae Gadella maraldi
0.1
0.2
0.7
0.2
0.0
Caproidae Capros aper
13.7
24.7
34.8
1336.3
26.8
Macroramphosidae Macroramphosus scolopax
16.7
19.3
47.1
1695.6
34.0
Sparidae Pagellus spp.
1.0
5.4
4.4
28.2
0.6
Mullidae Mullus surmuletus
0.1
3.0
0.7
2.2
0.0
Pomacentridae Chromis limbata
0.1
0.2
0.7
0.2
0.0
Carangidae Trachurus picturatus
0.9
2.6
3.6
12.6
0.3
Scombridae Scorn ber Japon icus
0.4
6.0
2.2
14.1
0.3
Pisces unidentified
16.6
16.3
43.5
1431.2
28.7
Rocks
1.0
0.3
5.1
6.6
0.1
Tissue unidentified
0.4
0.8
2.2
2.6
0.1
' Because the %0 is a nonadditive index (Cortes, 1997) for grouping fish items into higher
taxonomic categories (i.e
Pisces, etc), the %0 value was
recalculated by considering the number of stomachs with the
respective higher
taxonomic category. This recalculation affects both the IRI and %IRI |
values.
594
Fishery Bulletin 101(3)
A
100
Thornback ray, Raja clavata
(n= 138)
Tope shark, Galeorhlnus galeus
(n=184)
Pisces
CJ o
B
100
80
60
40
20
0
20
40
60
80
100
Figure 3
Relative importance of prey categories in the diet of (A) tliornback ray (/?"./o clavala) and (B) tope shark (Galeorhirtiis
galeus) ranked from highest IRI values. Where the areas of the boxes are equal to the IRI value [(%N+'/r W) x %0\. ':', N is
the percent number, 7,W the percent weight, and %0 the frequency of occurrence of the prey category. Each tick mark of
%0 represents 10%.
%0
Morato et al.: Diets of Rqa davata and Galeorhinus galeus
595
Table 2
Percentage of relative importance C/t IRI ) of food categories of Raja davata by sex, total length, depth strata, and areas (coastal and
offshore banks). Prey items occurring in less than five stomachs were grouped into higher taxonomic levels. The null hypothesis
of not feeding upon the same most important prey categories was tested by using the top-down correlation method (being r.^ the
top-down correlation coefficient) and the top-down concordance method (being Cj the top-down concordance coefficient). NS = non
significant, *P<0.01.
Sex
Total length (cm)
Depth (m)
Areas
M
49-60 61-70 71-80 81-93 0-100 101-200 201-350 Banks Coastal
Cephalopoda 0.52 0.03 1.44 0.00 0.38 0.63 0.03
Polychaeta 0.62 1.70 0.21 0.43 0.73 6.44 0.54
Penacidea 0.34 0.62 0.72 1.32 0.18 0.00 0.19
Other Natamia 0.10 015 0.52 O.I I 0.06 0.00 0.48
Diogenidae 0.07 1.58 1.45 0.00 0.88 0.45 0.69
Scyllarus anms 1.54 0.57 1.12 0.25 2.48 0.64 0.76
Cahippa xraimlata 0.73 0.31 0.45 0.30 0.90 0.36 0.00
Liocanimis -.pp. 8.12 0.60 1.64 3.30 9.49 0.19 10.44
Other Reptantia 9.20 8.32 1.43 11.88 22.48 0.00 47.44
Mysidacea 0.68 0.50 0.18 0.62 1.02 0.00 0.00
Isopoda 0.53 0.00 0.00 0.24 0.30 0.35 0.02
Cciproscipci- 41.20 24.16 36.26 38.11 23.39 53.34 20.06
Mmnmimphosus scolopax 35.15 58.88 53.60 41.65 34.91 37.35 15.81
Pagelhis^p. 0.46 1.07 0.00 1.08 1.24 0.00 0.56
Myctophidae 0.04 0.12 0.00 0.14 0.09 0.00 0.03
Tnulumis picnmiliis 0.14 0.57 0.34 0.00 0.87 0.26 1.25
Other Pisces 0.58 0.82 0.64 0.56 0.61 0.00 1.71
0.21
0.40
0.12
0.01
0.21
0.84
1.52
1.43
0.31
1 .00
0.47
38.26
54.84
0.19
0.04
0.01
0.15
5.60
15.13
14.72
0.00
0.00
4.35
0.00
0.00
0.00
2.21
0.00
10.63
36.72
9.99
0.65
0.00
0.00
Stomachs with contents (/I) 89 49 19 47 60 11 47
Ct=0.51NS
78
13
3.48
4.23
1.41
0.08
0.00
0.91
0.10
0.00
0.52
16.79
0.00
35.56
33.50
2.27
1.16
0.00
0.00
0.06
0.57
0.29
0.12
0.53
1.21
0.64
12.30
6.33
0.00
0.31
32.53
42.91
0.43
0.01
0.40
1.36
24
=0.44NS
110
prey were also recorded in the stomachs of thornback rays:
the chub mackerel, {Scomber japonicus [7fIRI=0.3]) and
the blue jack mackerel {Tyachuriis picturatus [%IRI=0.3]).
Some individuals also fed upon mesopelagic myctophids
(%IRI<0. 1 ) and upon shallow water benthic fish such as the
red striped mullet (Mullus surmuletus [%IRI<0.1] ) and the
Azorean chromis (Chrornis linibata [%IRI<0.1|).
Reptants occurred in 47.1% of the stomachs examined
and represented 17.0% by weight and 31.9% by number
of the total prey found (Fig. 3A). Swimming crabs (Liocar-
cinus spp. 1%IRI=5.5]), which include both L. marmoreiis
(%IRI=2.8) and L. corrugatus (%IRI=0.8), were the most
important reptant prey item in the diet of thornback ray
(Table 1). Other important reptants included the lesser
locust lobster (Scyllarus arctus [%IRI=0.9]), the shame-
faced crab (Calappa granulata |%IRI=0.5]), as well as
some unidentified Diogenidae (%IRI=0.3) and brachyura
(%IRI=1.5).
Polychaetes (%IRI=0.8) were the third most important
prey category and occurred in 9.4% of the stomachs with
food (Fig. 3A). Mysids (%IRI=0.5), natants (%IRI=0.3),
isopods (%IRI=0.2), and cephalopods (%IRI=0,1) also
occurred in stomachs of thornback rays sampled in the
Azores (Table 1).
A comparison of thornback ray's diet in relation to sex,
length, depth and area of capture (Table 2) suggests that C.
aper and M. scolopax were by far the most important prey
for all subgroups examined. The diets of both sexes were
significantly correlated (r.p=0.70, P<0.01), indicating a high
degree of similarity in the diets of males and females. Both
sexes fed primarily upon two benthopelagic fish species (M.
scolopax and C. aper) and reptants (Table 2). Schoener's
diet overlap index between males and females was 0.72,
also indicating a high level of similarity between diets.
Significant concordance (C^=0.74,P<0.01) was displayed
among thornback rays of different size classes (49-60,
61-70, 71-80 and 81-93 cm TL). Prey categories had simi-
lar %IRI values for the different size classes (Table 2 ), with
the exception of reptants (both Liocarcinus spp. and "other
reptants"), which were more important in the diet of the
two middle size classes. Schoener's index also suggested
a high degree of overlap (>0.60) among all size classes
(Table 3).
Examination of depth-related differences was lim-
ited by the small sample size of rays from deeper waters
(n2oi_350n,= 13). However, the top-down concordance coef-
ficient suggested that individuals captured at different
depths (0-100, 101-200, and 201-350 m) do not feed upon
the same most important prey categories (Ct.=0. 52, P>0.05).
Reptants (both Liocarcinus spp. and "other reptants") and
the fish species T. picturatus were more important in the
diet of rays captured in shallow waters (0-100 m); whereas
596
Fishery Bulletin 101 (3)
100
80
60
40
20
Group I
Group I
^fr^
to in oi
OD CD 00 in CD CD
y- f- y- Oi
2. -— X C >--
LO CD 3- ^ — -
9? 5 5 5
m
m
S
■c
<■
S
C/3
s. --~ —
<
LU
Z
!ii ^
r r 5
<
05
Group I
V ^
^;^ I
Figure 4
Dendrogram of the cluster analysis (Euclidean distance, average linkage method) for geographic
patterns of feeding habits of Raja clavata. In parentheses is given the authorship of the studies: 1 =
Ellis et al .( 1996 ); 2 = Gibson and Ezzi (1987 ); 3 = Ajayi ( 1982 ); 4 = Quiniou and Andriamirado ( 1979 );
5 = Olaso and Rodriguez-Mann ( 1995); 6 = Cunha et al. (1986); 7 = the present study; 8 = Ebert et
al. (1991); 9 = Smale and Cowley (1992).
polychaetes, cephalopods, penaeids, mysids, seabreams (Pa-
gellus sp.), and myctophids were consumed more by rays
caught in deeper waters (Table 2 ). Schoener's overlap index
for individuals captured at different depth intervals (Table
3) indicated low overlap (=0.50), supporting the results of
the top-down concordance coefficient analysis.
Finally, the diet of rays caught in coastal areas and
offshore banks were not significantly correlated (C.p=0.44,
P>0.05), indicating that thomback rays feed upon differ-
ent prey depending on the environment. The Diogenidae,
Liocarcinufs spp., "other reptants," and "other Pisces" were
more important prey for rays in coastal areas, whereas
polychaetes, penaeids, cephalopods, mysids, seabreams
iPagellus sp.), and myctophids were more important for
rays caught at offshore banks (Table 2). However, Schoen-
er's index showed a high level of overlap (0.69) between the
diets of rays caught in the different locations — most likely
due to the high dominance of two benthopelagic fishes in
their diets (75.4% and 69.1% for coastal areas and offshore
banks, respectively I.
Published information on the diet of Ihornback rays is
summarized in Table 4. Estimations of mean trophic levels
vary from 3.1, for the smallest size class ( South Wales: <25 cm
TL), to 4.2 for the Azorean thomback ray (this study; size
Table 3
Schoener's diet overlap index for thornback rays iRaja cla-
vata) size classes and for different depth strata.
Depth (m)
101-200 201-350
Total length (cm)
61-70 71-80 81-93
0-100 0.40 0.29 49-60 0.83 0.66 0.76
201-350 0.50 61-70 0.77 0.77
71-80 0.62
classes 49-60 and 81-93 cm TL). The arbitrarily chosen
cutoff in the cluster analysis was set at 60% dissimilar-
ity, which divided the dendrogram into three groups with
similar feeding patterns (Fig. 4). Cluster group I grouped
the Azorean poi)ulations (all size classes) and had an esti-
mated trophic level of'4. 14 ( ±0.09 SD). Cluster group II con-
tained all other medium and large size classes (i.e. >40 cm
TL), with the exception of small rays from the Canta-
brian Sea, North Spain (17-49 cm TL), and one small- to
Morato et al : Diets of Rqa davata and Galeorhinus galeus
597
ey cat-
NAT=
(1996);
study;
>
IM
lO
CO
(M
.—1
■*
■*
■*
CO
in
CO
CO
T)i
■*
■*
CD
CD
in
CM
T-H
o
CM
r-
in
m
-J
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
CO
■*
Tf
^'
Tj
co
CO
CO
^- ;t^ ■ -^
t*-;
o
I— 1
1-H
t— 1
IM
CO
CO
CO
CO
CO
CO
^
Tf
-
t-
c-
t~
00
03
03
CO o ~: t.
(ii
§1^ :^
■■3 S H "
CO
o
o
o
O
00
00
00
CO
CO
00
o
o
o
o
o
CM
CO
CM
^
Tf
,-H
in
o
CO
1—1
CD
t^
CO
o
00
CD
03
c-
CO
o
o
o
o
o
o
m
CO
in
00
in
f— t
03
»-H
CM
t:~
> ;^ II ..
^ II CO
E
c4
in
t— 1
d
CM
00
CO
CO
1— i
in
d
d
d
d
d
CD
I>
jA
d
^
^
d
Tl-
00
in
'— '
1— I
CM
C!3
00
CD
03
CO
■S M 00
-2 O 2 2
< H ^ c
a.
o
o
o
o
in
o
^_,
00
f— 1
CM
CM
^
CO
CO
03
^
CO
CO
03
CO
CO
^
T-t
■*
o
-T^ M 3 .
CO
03
CO
1—1
•—1
CO
CM
00
^
r- 1
■*
03
in
CO
CO
CO
CM
O
t>
CM
^.
o
00
00
resentei
idacea;
wing St
ha et al
CD
d
CO
CO
CO
00
Tf
CM
in
l>
CM
CM
d
in
00
Tl<
CO
CO
CO
in
CD
1-H
CO
1>
cm'
'^
CO
c~
^
^
CO
CD
c~
00
CO
tT
r^
i>
in
00
c~
t-H
CO
■*
t^
& ^ ^ c
1
o
o
o
o
in
CM
CO
00
CM
o
I>
CM
i>
CM
o
00
CD
■*
CO
■-t
o
iH
03
CM
g S'Scj
■*
o
1— 1
Oi
^
t— (
CM
CD
CM
o
03
P
CM
CO
00
Tt
CO
CO
CM
^
CM
o
o
in
c-
c
CO
CM
CM
.-H
CM
d
in
I>
03
d
CO
■*
cd'
'S'
i-H
.-H
d
d
^
d
d
"a II i II
in
?— 1
CM
CD
"*
CM
in
CM
CO
00
CM
^
I— 1'
M M *J to
JI! r*^ ■'-' • '
o
o
o
o
o
O
O
O
o
O
o
o
^
(-J
5-)
o
o
o
O
o
o
o
o
in
CJ3
1-H
H
o
o
o
o
O
o
o
o
o
o
o
O
o
o
o
o
o
o
o
o
o
o
t^
CO
t^
M
o
CD
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
CM
iri
ra
— O ^ -
CM
T-H
ei:
t; j= *^ c
w
o
O
j-5
o
o
o
o
o
o
o
CO
^
o
o
CO
o
CO
o
00
CM
CM
o
CO
CO
CO
3
3
P"^
a>
00
o
in
o
o
o
o
o
o
Tf
00
o
o
CM
o
°9
o
CO
O
o
oc>
CO
a
,) following
a; AMP = /
?rence (Ref
Rodriguez
^
ci
C3
d
d
d
d
d
d
d
d
CM
d
d
d
CM
d
i-H
d
d
d
,_;
d
^
d
1-H
c
1
S
a
Oh
o
O
o
o
o
o
o
o
o
o
CM
O
in
CM
o
03
o
o
o
o
o
o
CM
O
1-H
-J
^
c +j to
-2 II c —
o
o
o
o
o
o
o
o
o
o
o
O
o
O
t:^
o
o
Tf
o
in
o
o
o
d
IZ)
C>1
r^
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
CM
CO
CO
o
o
o
II
o o - O
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
d
^
Table 4
of trophi
opoda: IS'
Numbers
979): 5 =
a,
o
o
o
o
o
o
o
o
o
o
o
o
o
.— 1
CM
Tl-
o
00
CO
o
00
1-^
Q
ti^
lO
CO
CO
o
o
o
o
o
o
o
o
o
o
o
o
i>
o
in
'ti-
o
CO
CO
o
o
CO
\
o
Tt
^
d
d
d
d
d
d
d
d
d
d
d
d
d
Tf
d
d
^
d
d
d
d
d
d
i. Estimation
EP = Cephal
yslop, 1980).
riamirado (1
evel.
J=
K
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
1-.
o.
o
o
o
d
CO
d
CD
CO
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
d
o
tions
;a;C
id(H
And
hicl
^
^
o
o
o
o
o
o
o
o
o
o
■*
in
CO
in
Q
^
o
o
o
o
o
o
o
o
o
T*
in
in
CT>
o
o
o
o
o
o
o
CO
t~
in
"*
o
o
o
o
o
o
o
o
o
o
o
t-
2 gx-^ §•
pa
d
CO
(N
d
d
d
d
d
d
d
CM
CO
CO
O-l
d
d
d
d
d
d
d
d
d
d
d
c
OJ
O OJ g g II
2 o •" -2 >
J
o
o
o
o
o
^
o
o
o
CO
o
o
o
o
CD
o
rH
CO
,—1
CO
CO
'^
o
o
1-H
Q. c .S S J
o
<>3
c^
o
^
o
o
00
03
"*
t>
o
o
o
o
CM
o
O
o
CM
-*
I-
Tf
o
o
CM
ogra
Echi
e po:
Qui;
2).T
a,
iri
■^
CO
CM
d
CO
CO
CO
i-H
»-i
d
d
d
d
CM
d
d
d
d
d
d
CD
d
d
d
a* .. X II o>
erent g
ECH =
PMist
82); 4 =
ley ( 19
X
T3
C
S
O-
5= ^ ^ ^ ^ ^
E^ ^ # ^ ^ ^
z
Z
Z
Z
>
>
5^ £^
g
^
>,
in diff
alvia;
isces. '
yi ( 19
dCow
s :2
o
O)
o>
o
in
05
03
03
£73
in
o
o
o
in
C33
03
o
o
o
CO
CO
t^
r-
C
2 " II "^ S
.2 ^
V
CO
'f
V
CM
V
in
CO
00
00
A
CO
in
1
1
2
o
o
in
V
o
in
A
CO
1
03
1
.— 1
00
1
1
.—1
00
1
o
in
V
in
A
Q
■- II M II -^
T}<
m
in
in
CO
in
i>
in
■^
CO
t--
CO
CO
■^ > " r^ '^
^ " 0- " e
w ~
(N
CO
-*
CO
1 " « p ^
■^
.M
.M
4^
-4J)
^
S
F CO '^s 00 II
CO
CO
CO
m
m
CO
*j
*J
J3
^
^
CO
a
a
a
M
a
a
.4-3
*j
X
CO
+-»
CR
CO
o
o
o
O
o
o
CO
CO
to
to
a
a
a
o
o
to
a
t3
3
3
;d diets of th
0L= Polychai
REP = Repta
n and Ezzi (]
etal. (1991);
o
u
_w
o
to
1
3
o
m
O
3
o
3
o
CO
o
1
CO
O
3
o
CO
O
A
3
o
CO
a
o
a a
o o
Z Z
a
o
to
a
o
Z
CO
a
o
O
Z
o
O
CO
1
o
O
to
1
CO
01
1-1
O
N
<
CO
OJ
o
to
0)
;-!
s
<
to
1
o
CO
at
a
a
a
CJ
o
CO
(U
o.
a
O
a
s
o
1
ca
6
Categorizi
egories: P(
Natantia;
2 = Gibsoi
8 = Ebert
c
o
CS
o
2
c
<
H
2
'-3
c
<
z
c
<
Z
-a
c
CO
o
o
CO
Wales
Wales
Wales
Wales
CO
03
1
en
1
a)
C3
1=
a
%■.
u
c
a
;.<
m
CJ
C
2
fa
g
c
a
fa
c
'a
a
CO
c
'a
a
CO
"a
be
3
"a
M
3
X.
"a
cm
3
a
"a
be
"a
be
3
t
South Afri
South Afri
South Afri
en
598
Fishery Bulletin 101(3)
Table 5
Values for percentage by number CSN)
, weight C^fW), occurrence C/rO),
and index of relative importance (IRI and '2 IRI) for prey |
items observed in stomachs of tope shark (;i= 184), Galeor
hinus galeus,
caught off the Azores during the
spring of 1996 and 1997.
Number (No.) and percent occurrence
9rO) offish lenses
, fish remains
, and otoliths found in stomach.
are also presented. Total
values are given in bold font.
Prey items
%N
%W
%0'
IRI
%IRI
Total Crustacea
1.0
1.0
3.3
6.5
0.03
Isopoda
3.6
1.1
2.7
12.8
0.3
Crustacea unidentified
1.2
0.0
1.1
1.3
0.0
Total Cephalopoda
0.8
0.2
3.3
3.2
0.02
Octopodidae
0.6
0.3
0.5
0.5
0.0
Cephalopoda unidentified
3.0
0.0
2.7
8.1
0.2
Total Pisces-
98.2
98.8
100.0
19,700.4
99.95
Stemoptychidae unidentified
0.6
0.2
0.5
0.4
0.0
Synodontidae Synodus sp.
0.6
11.5
0.5
6.5
0.2
Trichiuridae Lepidopus caudatus
0.6
0.0
0.5
0.4
0.0
Macrouridae unidentified
0.6
0.0
0.5
0.3
0.0
Phycidae Phycis phycis
1.2
0.0
1,1
1.4
0.0
Caproidae Capros aper
65.0
25.6
38.6
3494.6
93.2
Macroramphosidae Macroramphosus
scolopax
11.2
2.7
8.2
113.5
3.0
Carangidae Trachurus picturatus
2.4
7.6
2.2
21.6
0.6
Total Sparidae
6.5
32.0
4.4
169.7
4.5
Pagellus acarne
2.4
5.8
1.6
13.3
0.4
Pagellus bogaraveo
2.4
14.4
1.6
27.3
0.7
Pagellus spp.
1.2
11.7
1.1
14.1
0.4
Pagrus pagriis
0.6
0.1
0.5
0.4
0.0
Sparidae unidentified
0.6
0.5
0.5
0.6
0.0
Scombridae Scomber Japonicus
2.4
18.4
1.6
33.8
0.9
No. of
%o
Pairs of fish lenses
493
103
Otoliths unidentified
118
75
Fish remains
3
2
' Because the ^0 is a nonadditive index (Cortes, 1997), when grouping fish items into higher taxonomic categories (i.e. Pisces, etc) the %0 val
ie was recal-
culated considering the number of stomachs
with the respective higher taxonomic category. This recalculation will affect both the IRI and
'7fIRI values.
■-' Including unidentified fish, pairs of lenses.
otoliths, and fish remains.
medium-size class of South Wales (35-45 cm TL ). Cluster
group III grouped small rays from several geographic re-
gions, from South Africa (which also includes some large
individuals) to NE Atlantic. Estimates of trophic levels
were 3.46 (±0.84 SD) for the rays of the cluster group II
(i,e, medium and large), and 3.35 (±0.21 SD) for the rays
composing cluster group III (i.e. small ). The estimated tro-
phic levels for the three cluster groups were significantly
different (P<0.001).
Tope shark
The diet of tope shark consisted almost exclusively offish
(%IRI=99.95), along with a few crustaceans (%IRI=0.03)
and ccphalopods 17r IR1=0.02) (Fig. 3B). Recognizable
prey from 14 different taxa were identified (Table 5).
The boarfish (C. aper) was the most important prey item
(%IRI=93.2), accounting for 65.0% of food by number (%N),
25.6% by weight (%W), and occurred in 38.6% of stomachs
that contained food (%0). The second most important prey
item was the snipefish (M. scolopax (%IRI=3.01), which
represented 11.2% of food by number and 2.7% by weight.
Some commercially important fish species were also found
in the stomachs of tope shark; sparids ('7,IRI=4.5, which
included Pagellus acarne, P. bogaraveo. and Pagrus pagrus),
the chub mackerel iS. japonicus |%IRI=0.9| ), and the blue
jack mackerel (T picturatus [%IRI=0.61l. These species
were more important by weight than by number or occur-
rence. The stomachs of tope sharks also contained 493 pairs
of eye lens and fish that were heavily digested, as well as
unidentifiable otoliths.
Discussion
In general, the percentage of empty stomachs for thornback
rays and tope sharks was relatively high compared to the
percentage from literature reports. The percentage of empty
Morato et al.: Diets of Raia davata and Galeorhinus galeus
599
stomachs for tope shark was 47.7% — much higher than the
4.3% obsen'ed by ElHs et al. ( 1996). The percentage of empty
thornback ray stomachs was high (37.1%) when compared
to values reported for the North Sea (9%, Daan et al.'; and
3.7%, Ellis et al., 1996), Carmarthen Bay, South Wales
(4.5%, Ajayi, 1982), west coast of Southern Africa (4.5%,
Ebert et al., 1991; and 2.6%, Smale and Cowley, 1992) and
the Portuguese mainland coast (2.5%, Cunha et al., 1986).
We attribute the high percentage of empty stomachs found
in our study to the use of longlines to catch the fish in the
Azores (trawls were used in the other studies ). Longlining is
a passive fishing method, which suggests that fish that feed
to satiation have a reduced response to bait odor ( Lokkeborg
et al., 1995), meaning that fish with full stomachs tend not
to eat the bait and be caught. Thus, only those fish with
empty stomachs or partial stomach fullness were caught.
Thornback rays captured by longline in the Azores dur-
ing the spring of 1996 and 1997 fed upon a wide variety
of organisms. Fishes (81.6 %IRI) and reptants (17.4 %IRI)
dominated the diet, which also consisted of polychaetes,
mysids. natants, isopods, and cephalopods. In general,
thornback rays in the Azores preyed more heavily upon
fish in comparison with the predation patterns described
in other studies. Ajayi et al. (1982) reported a predomi-
nance of crustaceans (83%W) for all size classes and a low
importance offish (11.6%W) in the diet of thornback rays
in Carmarthen Bay, Bristol Channel. They also reported
amphipods, polychaetes, and some natants as food items.
Using the points method of Hyslop (1980), Ellis et al.
(1996) reported that thornback rays from the North Sea
fed primarily on crustaceans (78.9% ) compared to mollusks
(10.2%) and fish (7.3%). Several others have also reported
a dominance of crustaceans and low importance of fish in
the diet of thornback ray (Fitzmaurice, 1974; Marques and
Re, 1978; Quiniou and Andriamirado, 1979; Cunha et al.,
1986; Gibson and Ezzi, 1987; Smale and Cowley, 1992;
Olaso and Rodriguez-Marin, 1995; Daan et al.'; Ebeling^).
Polychaetes (Holden and Tucker, 1974; Marques and Re,
1978), bivalves (Quiniou and Andriamirado, 1979), holo-
thurians (Ebeling-), and cephalopods (Holden and Tucker,
1974; Marques and Re, 1978; Smale and Cowley, 1992;
Olaso and Rodriguez-Marin, 1995) that were considered
important prey items in the other studies mentioned were
not recorded or were insignificant in our samples.
Differences in diet composition of several predators may
reflect the geographic peculiarities in fauna composition
(e.g. Smale and Cowley 1992), but when comparing diets
based on higher taxonomic levels (such as fish, reptants,
and natants categories), such geographic differences
should not be so obvious. Our geographic analysis (see
Fig. 4) distinguished three major groups: I) the Azorean
individuals; II ) other large individuals; and III ) other small
individuals. Further, the estimated mean trophic levels for
these three major groups were significantly different: 4.14
(±0.09 SD) for the Azores; 3.46 (±0.84 SD) for other large
rays; and 3.35 (±0.21 SD) for smaller rays. The higher
2 Ebeling, E. 1988. A brief survey of the feeding preferences
of Raja davata in Red Wharf Bay in the Irish Sea. ICES CM.
1988/G:58, 5 p.
trophic level for the Azores is a result of a higher degree
of piscivory in this region and an increased consumption
of decapods and fish by larger rays, compared with small
rays. Notwithstanding the difference in sampling methods
(longline vs. trawl caught), it appears that the Azores can
be considered a separate group. In other studies, predator
size played the major role in controlling feeding patterns.
The diet of the thornback ray in the Azores consists of
a greater proportion of fish than in any other area and
may reveal differences in the function of different environ-
ments, because seamounts and oceanic islands are the ma-
jor topographic feature of the Azores region and the other
studies were conducted on continental shelves. The general
function of oceanic seamount environments is still not
completely understood but they are characterized by sub-
stantial enhancement of primary production due to topo-
graphic effects on local hydrographic conditions (Genin and
Boehlert, 1985). However, evidence for enhanced primary
production leading to concentrations offish over seamounts
is sparse (Rogers, 1994). Additionally, the availability and
relative abundance of the two most important fish prey
items found in our work (the benthopelagic species C. aper
and M. scolopax ) vary considerably both seasonally (Grana-
deiro et al., 1998) and annually. Therefore, the high degree
of piscivory in the Azores may result from environmental
features and exceptional fish prey availability during the
sampled years or seasons.
Thornback rays also fed on pelagic fish, as indicated by
the presence of chub mackerel and jack mackerel in stom-
achs— a finding that confirms previous suggestions (see
Daan et al. ' ; Ebeling- ) that thornback rays are active preda-
tors and able to feed semipelagically. The most important
reptants in the diet, Liocarcinus spp., were also reported as
the main prey item for thornback rays by Ellis et al. ( 1996).
The level of importance of isopods and amphipods, mysids,
cephalopods, and polychaetes in the diet of thornback rays
in the Azores was similar to values reported by other au-
thors (Ellis et al., 1996; Daan et al.'; Ebeling^).
Differences in the dentition of females and males were
reported by Quiniou and Andriamirado ( 1979) but we and
Smale and Cowley (1992) observed no differences in the
major prey consumed between sexes. Therefore, sexual di-
morphism in dentition does not appear to be manifested
in dietary preferences between sexes, as was initially
expected.
Several studies have demonstrated differences in preda-
tion patterns for rays of different size classes — primarily
a decrease in importance of crustaceans and an increase
offish with size (e.g. Smale and Cowley, 1992; Elhs et al.,
1996; Daan et al.'; Ebeling^). Some authors attribute these
differences to the ability of large predators to prey upon
larger prey (Smale and Cowley, 1992); others suggest the
difference is due to a pronounced shift from a benthic to a
benthopelagic feeding behavior (Skjaeraasen and Bergstad,
2000; Ebeling^) or the reverse (Quiniou and Andriamirado,
1979). We found no significant size-related differences in
diet. Quiniou and Andriamirado (1979) reported shifts in
diet at a size of 30 to 40 cm TL but we could not verify these
conclusions because our sample included only rays larger
than 49 cm.
600
Fishery Bulletin 101(3)
There have been few data indicating dietary differences
between thornback rays collected at different depths.
Smale and Cowley ( 1992) reported that bottom type used
by rays varies with depth and predicted that the prey spec-
trum would thus also vary, but no depth-related analyses
of diet composition were preformed in their study. Despite
similarities in size (i.e. no differences in the mean size by
depth strata; Menezes-*), we found that rays inhabiting dif-
ferent depths prey upon different resources. The decreasing
consumption of Liocarcinus spp., "other reptants," and T.
picturatus, and the increasing consumption of penaeids,
seabreams, and myctophids with depth of capture of rays,
appears to be in general agreement with the relative abun-
dance of prey with depth. Therefore, such depth-related
variations in diet may simply reflect differences in prey
availability. It is not clear, however, why Scyllarus arctus,
a species with a known depth distribution of 4 to 50 meters
(e.g. Alvarez, 1968; Castellon and Abello, 1983), appears in
stomachs of thornback rays caught between 201 and 350
meters (see Table 2). There is no evidence of vertical mi-
grations of thornback ray associated with feeding activity;
therefore this prey was likely eaten at deep water. Thus, the
depth distribution range of S. arctus in the Azores may be
significantly greater than what was previously known. The
only study that could corroborate this hypothesis (Fransen,
1991) reported one S. arctus caught between 420 and 700
meters depth in the Canary Islands.
Our comparisons between areas (coastal and offshore
banks) were unable to clearly separate the influence of
depth because nearly all coastal samples were obtained
from shallow waters, and offshore bank samples were
collected from much deeper waters. Hence, we were
incapable of determining whether the high level of poly-
chaetes, penaeids, cephalopods, mysids, seabreams, and
myctophids in the diet of rays caught at offshore banks
reflects the availability of these prey in these areas, or in
deeper waters, or both. Nevertheless, our findings indicate
that coastal rays have different diets from rays taken in
offshore banks.
Tope sharks preyed almost exclusively upon teleosts,
along with very few crustaceans and cephalopods. Previ-
ous observations on the feeding behavior of this species
suggested that fish and cephalopods are the main prey
categories ( Elhs et al., 1996; Olsen, 1954). The diet of tope
shark in the Azores consists of fewer species (mainly small
shoaling fish, mainly boarfish and snipefish) compared to
the diet of tope shark documented in previous studies.
These two fish were also important diet components of
other piscivorous species around the Azores between 1993
and 1997, namely cephalopods (Pierce et al., 1994), elas-
mobranchs (Clarke et al., 1996), fishes (Clarke et al., 1995;
Morato et al., 1999, 2000, 2001) and seabirds (Granadeiro
et al, 1998; Ramos et al., 1998a, 1998b). The role of these
two small shoaling fish in the marine food web of the Azores
is not yet fully understood. The fact that these prey may
exhibit strong variation in abundance, raises the question
^Menezes, G. 1995-97. Unpubl. data. Department of Ocean-
ography and Fishcrie.s, University of the Azores. Cais de Santa
Cruz, PT9901-862 Horta, Portugal.
of how well predators can adapt to extensive changes in
their availability.
Stomach-content data offer a good snapshot of the feed-
ing habits of fish species, but diets may vary substantially
with food availability, depth, location, and season. Caution
is, therefore, required when drawing conclusions about
the trophic ecology of marine predators. The trophic role
of thornback rays and tope sharks in the Azores could be
further clarified by year round sampling and by an analysis
of stable isotopes (Gu et al., 1996; Jennings et al., 1997; Pin-
negar and Polunin, 2000), which could provide a less biased
average estimate of predator trophic level.
Acknowledgments
This work is part of a more comprehensive study supported
by the European Union (Design optimization and imple-
mentation of demersal cruise survey in the Macaronesian
Archipelagos (study contract DG XIV/94/034 and DG XIV/
95/095). We thank Joao Gongalves, Ricardo Serrao Santos,
Filipe Porteiro for help with identification of stomach con-
tents, and Helena Krug for help with the identification of
otoliths. Special thanks are due to the scientific staff and
to the crew of the KV Arquipelago for working overtime at
sea. We are also grateful to Malcolm J. Smale, Pedro Afonso,
Joel Carlin, and Natacha Carvalho for reviewing the manu-
script. Comments and suggestions of anonymous reviewers
greatly improved the quality of the manuscript.
Literature cited
Ajayi, T. O.
1982. Food and feeding habits of Raja species (Batoidei) in
Carmarthen Bay, Bristol Channel. J. Mar Biol. Assoc. U.K.
62:215-223.
Alvarez, R. Z.
1968. Crustaceos decapodos Ibericos. Investigacion Pes-
quera, tomo 32, 510 p. Imprenta Juvenil, Barcelona.
Castellon, A., and P. Abello.
1983. Bathymetric distribution of some Reptantia Decapoda
in the Catalan area (Spain). Rapp. Comm. Int. Mer Medit.
28(3):291-294.
Clarke, M. R., D. C. Clarke, H. R. Martins, and H. M. Silva.
1995. The diet of swordfish {Xiphias gladius) in Azorian
waters. Arquipelago (Life Mar. Sci.) 13A:53-69.
1996. The diet of blue shark (Pnonace glauca. L.) in Azorean
waters. Arquipelago (Life Mar Sci.) 14A:41-56.
Clarke, K. R., and R. M. Warwick.
1994. Change in marine communities: an approach to statis-
tical analysis and interpretation, 144 p. Natural Environ-
ment Research Council, UK.
Compagno, L. J. V.
1984. FAO Species catalogue. Vol. 4, Sharks of the world:
an annotated and illustrated catalogue of sharks species
known to date. Part 2 Carcharhiniformes. FAO (Food
and Agriculture Organization) Fish. Synop. 125(4) Part 2:
251-655.
Cortds, E.
1997. A critical review of methods of studying fish feeding
based on analysis of stomach contents: application to elas-
mobranch fishes. Can. J. Fish. Aquat. Sci. 54:726-738.
Morato et al.: Diets of Rqa c/avata and Galeorhinus galeus
601
1999. Standardized diet compositions and trophic levels of
sharks. ICES (International Council for the Exploration
of the Sea.) J. Mar. Sci. 56:707-717.
Cunha, P., J. Calvario. J. C. Marques, and P. Re.
1986. Estudo comparativo dos regimes alimentares de Raja
brachyura Lafont, 1873, Raja ctavata Linne, 1758, Raja
montagui Fowler, 1910 e Raja naevus Miiller and Henlen,
1841 (Pisces: Rajidae) da costa Portuguesa. Arquivos do
Museu Bocage Serie A III(8):137-154.
Day, F.
1880-84. The fishes of Great Britain and Ireland, vol. 1, 336
p., and vol. 2, 388 p. William and Norgate, London.
Ebert, D. A., P. D. Cowley, and L. J. V. Compagno.
1991. A preliminary investigation of the feeding ecology of
skates (Batoidea: Rajidae) off the west coast of Southern
AfHca. S. Afr J. Mar. Sci. 10:71-81.
ElUs, J. R., M. G. Pawson, and S. E. Shackley
1996. The comparative feeding ecology of six species of shark
and four species of ray (Elasmobranchii) in the North-East
Atlantic. J. Mar. Biol. Assoc. U.K. 76:89-106.
Ferry L. A., and G. M. Cailliet.
1996. Sample size and data analysis: are we characterizing
and comparing diet properly? In GUTSHOP '96. Feeding
ecology and nutrition in fish: symposium proceedings (D.
MacKinley and K. Shearer, eds), p. 71-80. Am. Fish. Soc,
San Francisco, CA.
Fitzmaurice, P.
1974. Size distribution and the food of thornback rays (Raja
clavata L. ) caught on rod and line on the Mayo Coast. Irish
Fish. Invest, (ser. B) 11.
Ford, E.
1921. A contribution to our knowledge of the life-histories of
the dogfish landed at Plymouth. J. Mar Biol. Assoc. U.K.
12:468-505.
Fransen, C. H. J. M.
1991. Preliminary report on Crustacea collected in the
eastern part of the North Atlantic during the CANCAP and
MAURITANIA expeditions of the former Rijksmuseum van
Natuurlijke Historie, Leiden, 200 p. Nationaal Natuurhis-
torisch Museum, Leiden.
Genin, A., and G. W. Boehlert.
1985. Dynamics of temperature and chlorophyll structures
above a seamount: an oceanic experiment. J. Mar. Res. 43:
907-924.
Gibson, R. N., and I. A. Ezzi.
1987. Feeding relationships of a demersal fish assemblage
on the west coast of Scotland. J. Fish Biol. 31:55-69.
Granadeiro, J. P., L. R. Monteiro, and R. W. Furness.
1998. Diet and feeding ecology of Cory's shearwater Calonec-
tris diomedea in the Azores, north-east Atlantic. Mar. Ecftl .
Prog. Ser. 166:267-276.
Gu, B., S. L. Schelske, and M. V. Hoyer.
1996. Stable isotopes of carbon and nitrogen as indicators of
diet and trophic structure of the fish community in a shal-
low hypereutrophic lake. J. Fish Biol. 49:1233-1243.
Holden, M. J., and R. N. Tucker.
1974. The food of Raja clavata Linnaeus 1758, Raja mon-
tagui Fowler 1910. Raja naevus Muller and Henel 1841, and
Raja brachyura Lafont 1873 in British waters. J. Cons. Int.
Explon Mer 35(2):189-193.
Hurtubia, J.
1973. Trophic diversity measurement in sympatric preda-
tory species. Ecology 19:36-58.
Hyslop, E. J.
1980. Stomach contents analysis — a review of methods and
their applications. J. Fish Biol. 17:411-429.
Jennings, S., O. Refiones, B. Morales-Nin, N. V. C. Polunin,
J. Moranta, and J. Coll.
1997. Spatial variation in the '''N and "C stable isotope
composition of plants, invertebrates and fishes on Mediter-
ranean reefs: Implications for the study of trophic pathways.
Mar. Ecol. Prog. Ser 146:109-116.
Lokkeborg, S., B. L. 011a, W. H. Pearson, and M. W. Davis.
1995. Behavioural response in sablefish Anoplopoma fim-
bria, to bait odour J. Fish Biol. 46:142-155.
Marques, V. M., and P. Re.
1978. Regime alimentaire de quelques Rajidae des cotes
Portugaises. Arquivos do Museu Bocage. 20 serie VI(34), 8 p.
Morato, T, E. Sola, M. P. Gros, and G. Menezes.
1999. Diets of forkbeard iPhycis phycis) and conger eel
(Conger conger) off the Azores during spring of 1996 and
1997. Arquipelago {Life Mar Sci.) 17A:51-64.
2001. Feeding habits of two congener species of seabreams,
Pagellus bogaraveo and Pagellus acarne, off the Azores (nor-
theastern Atlantic) during spring of 1996 and 1997. Bull.
Mar. Sci. 69(3):1073-1087.
Morato, T, R. S. Santos, and P. Andrade.
2000. Feeding habits, seasonal and ontogenetic diet shift of
blacktail comber, Serranus atricauda (Pisces: Serranidae),
from the Azores, Northeastern Atlantic. Fish. Res. 49(l):51-60.
Olaso, I., and E. Rodriguez-Marin.
1995. Alimentacion de veinte especies de peces demersales
pertenecientes a la division VIIIc del ICES. Otoiio 1991.
Inf Tec. Inst. Esp. Oceanogr. 157, 56 p.
01sen,A. M.
1954. The biology, migration and growth rate of the school
shark, Galeorhinus australis (Macleay) (Carcharhinidae) in
south-eastern Australian waters. Aust. J. Mar. Freshwat.
Res. 5:353-410.
1984. Synopsis of biological data on the school shark,
Galeorhinus australis (Mcleay 1881). FAG Fish. Synop.
139, 42 p.
Pauly, D., V. Christensen, and C. Walters.
2000. Ecopath, Ecosim, and Ecospace as tools for evaluat-
ing ecosystem impact of fisheries. ICES J. Mar. Sci. 57:
697-706.
Pierce, G. J., P. R. Boyle, L. C. Hastie, and M. B. Santos.
1994. Diets of squid Loligo forbesi and Loligo vulgaris in the
northeast Atlantic. Fish. Res. 21:149-163.
Pinkas, L., M. S. Oliphant, and I. L. K. Iverson.
1971. Food habits of albacore, bluefin tuna and bonito in
California waters. Calif Fish Game 152:1-105.
Pinnegar, J. K., and N. V. C. Polunin.
2000. Contribution of stable-isotope data to elucidating food
webs of Mediterranean rocky littoral fishes. Oecologia 122:
399-409.
Pitcher, T J.
2000. Ecosystems goals can reinvigorate fisheries manage-
ment, help dispute resolution and encourage public support.
Fish Fish. 1:99-103.
Quade, D., and I. Salama.
1992. A survey of weighted rank correlation. In Order sta-
tistics and nonparametrics: theory and applications (P. K.
Sen, and I. Salama, eds.), p. 213-224. Elsevier, New York.
Quiniou, L., and G. R. Andriamirado.
1979. Variations du regime alimentaire de trois especes de
raies de la bale de Douarnenez (Raja montagui Fowler,
1919; Raja brachyura Lafont, 1873; Raja clavata L., 1758).
Cybium 7:27-39.
Ramos, J. R., E. Sola , L. R. Monteiro, and N. Ratcliflfe.
1998a. Prey delivered to roseate tern chicks in the Azores. J.
Field Omithol. 69:419-429.
602
Fishery Bulletin 101(3)
Ramos, J. R., E. Sola, F. M. Porteiro, and L. R. Monteiro.
1998b. Prey of yellow-legged gull, roseate tern and common
tern in the Azores. Seabird 20:31-40.
Rogers, A. D.
1994. The biology of seamounts. Adv. Mar. Biol. 30:
306-350.
Schoener, T. W.
1970. Non-synchronous spatial overlap of lizards in patchy
habitats. Ecology 51:408-418.
Skjaeraasen, J. E., and O. A Bergstad.
2000. Distribution and feeding ecology of Raja radiata in the
northeastern North Sea and Skagerrak (Norwegian Deep).
ICES J. Mar Sci. 57:1249-1260.
Smale, M. J., and P. D. Cowley
1992. The feeding ecology of skates (Batoidea: Rajidae) off
the Cape south coast. South Africa. S. Afr. J. Mar. Sci. 12:
823-834.
Stehmann, M.
1987. Quick and dirty tabulation of stomach contents and
maturity stages for skates (Rajidae), squaloid and other
ovoviviparous and viviparous species of sharks. Am.
Elasmobranch Soc. Newsletter 1987(3):5-9.
Stehmann, M., and D. L. Burkel.
1984. Rajidae. /n Fishes of the North-eastern Atlantic and
the Mediterranean (P. J. P. Whitehead, M. BL. Bauchot, J.
BC. Hureau, J. Nielse, and E. Tortonese, eds.), vol. 1, p.
163-196. UNESCO (United Nations Educational, Scien-
tific, and Cultural Organization), Paris.
Whipple, S. J., J. S. Link, L. P. Garrison, and M. J. FogEirty.
2000. Models of predation and fishing mortality in aquatic
ecosystems. Fish Fish. 1:22-40.
Zar, J. H.
1999. Biostatistical analysis, 4"' ed., 663 p. Prentice Hall
International Editions, Upper Saddle River, NJ.
603
Abstract— The U.S. Marine Mammal
Protection Act requires that the abun-
dance of marine mammals in U.S. waters
be assessed. Because this requirement
had not been met for a large portion of
the North Atlantic Ocean (U.S. waters
south of Maryland), a ship-based, line-
transect survey was conducted with a
68 m research ship between Maryland
(38.00°N) and central Florida (28.00°N)
from the 10-m isobath to the boundary
of the U.S. Exclusive Economic Zone.
The study area (573,000 km-) was sur-
veyed between 8 July and 17 August
1998. Minimum abundance estimates
were based on 4163 km of effort and 217
sightings of at least 13 cetacean species
and other taxonomic categories. The
most commonly sighted species ( number
of groups) were bottlenose dolphins,
Tiirsiops truncatus (38); sperm whales,
Physeter macrocephalus (29); Atlantic
spotted dolphins, Stenella frontalis
(28); and Risso's dolphins. Grampus
griseus (22). The most abundant spe-
cies ( abundance; coefficient of variation )
were Atlantic spotted dolphins ( 14,438;
0.63 ); bottlenose dolphins ( 13,085; 0.40 );
pantropical spotted dolphins, S. attenu-
ata (12,747; 0.56); striped dolphms, S.
coeriileoalba (10,225; 0.91); and Risso's
dolphins (9533; 0.50). The abundance
estimate for the Clymene dolphin, S.
clymene (6086; 0.93). is the first for the
U.S. Atlantic Ocean. Sperm whales were
the most abundant large whale (1181;
0.51). Abundances for other species or
taxonomic categories ranged from 20 to
5109. There were an estimated 77,139
(0.23) cetaceans in the study area.
Bottlenose dolphins and Atlantic spot-
ted dolphins were encountered primar-
ily in continental shelf (<200 m) and
continental slope waters ( 200-2000 m ).
All other species were generally sighted
in oceanic waters (>200 m). The distri-
bution of some species varied north to
south. Striped dolphins, Clymene dol-
phins, and sperm whales were sighted
primarily in the northern part of the
study area; whereas pantropical spot-
ted dolphins were sighted primarily in
the southern portion.
Abundance of cetaceans in the southern
U.S. North Atlantic Ocean during summer 1998
Keith D. Mullin
Gregory L. Fulling
Southeast Fisheries Science Center
National Marine Fisheries Service, NOAA
3209 Frederic Street
Pascagoula, Mississippi 39567
E-mail (for K. D Mullin); Keith DMullin(a'noaa-gov
Manuscript approved for publication
11 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:603-613 (2003).
The U.S. Marine Mammal Protection
Act (MMPA) requires that stocks of
marine mammal species in U.S. waters
be maintained at or above their opti-
mum sustainable population (OSP)
level, defined as the number of animals
that will result in maximum productiv-
ity. The MMPA, as amended in 1994,
requires that the U.S. National Marine
Fisheries Service (NMFS) determine
the potential biological removal (PBR)
of each stock for management pur-
poses. PBR is an estimate of the maxi-
mum number of animals that may be
removed from a stock due to human
activities (e.g. fisheries bycatch) while
allowing the stock to reach or maintain
its OSP. The PBR is calculated by using
the estimated minimum abundance of a
stock, half its maximum net productiv-
ity rate (theoretical; or estimated), and
a recovery factor (Barlow et al., 1995).
For the U.S. Exclusive Economic Zone
(EEZ) adjacent to the Atlantic coast of
the continental U.S., the NMFS cur-
rently defines 27 taxa of cetaceans
as stocks (Waring et al., 2001). These
stocks include 24 one-stock species,
bottlenose dolphins (Tursiops trunca-
tus) that are divided into two stocks,
and one mesoplodont beaked whale
stock. Abundance estimates are avail-
able for most of these stocks from U.S.
waters north of the Virginia-Maryland
border (38.00°N). In 1998, except for
three stocks, abundance estimates
were not available for Atlantic ceta-
cean stocks from U.S. waters south of
Maryland (Waring et al., 1997). Abun-
dance estimates for these three stocks
were based on a small amount of effort
from a 1992 winter ship survey south
of Cape Hatteras (Mullin and Ford').
Other cetacean abundance estimates
from U.S. waters south of Maryland are
for portions of the continental shelf or
continental slope (Blaylock and Hog-
gard, 1994; Blaylock, 1995; CeTAP2;
Fritts et al.-').
To estimate the abundance of ceta-
ceans in U.S. Atlantic waters south of
Maryland, a ship survey was conducted
during summer 1998 and the results are
reported in this study. Abundance esti-
mates from this area are combined with
abundance estimates from surveys of
U.S. waters north of the Virginia-Mary-
land border conducted by the NMFS
Northeast Fisheries Science Center to
obtain overall abundance estimates for
western North Atlantic cetacean stocks
(e.g. Waring et al., 2001).
'Mullin, K.D., and R.Ford. 1992. Report
of NOAA ship Oregon II cruise 92-01
(198) (a cetacean survey of U.S. Atlantic
waters south of Cape Hatteras, winter
1992). Southeast Fisheries Science
Center, P.O. Drawer 1207, Pascagoula,
Mississippi 39568.
- CeTAP (Cetacean and Turtle Assessment
Program). 1982. A characterization of
marine mammals and turtles in the mid-
and north-Atlantic areas of the U.S. outer
continental shelf Final Report of the
Cetacean and Turtle Assessment Program
Bureau of Land Management, contract no.
AA551-CT8-48, 450 p. U.S. Dep. Interior,
Washington DC.
i Fntts, T. H., A. B. Irvine, R. D. Jennings,
L. A. Collum, W. Hoffman, and M. A.
McGehee. 1983. Turtles, birds, and
mammals in the northern Gulf of Mexico
and nearby Atlantic waters. Rep. FWS/
OBS-82/65, 455 p. U.S. Fish and Wildlife
Service, Office of Biological Services, Wash-
ington. D.C.
604
Fishery Bulletin 101(3)
40-
35-
30-
North
Atlantic
Ocean
-70
Figure 1
Survey effort (4163 km; thin lines) in Beaufort sea state s4
in the southern U.S. Atlantic study area (outlined by thick
line) during summer 1998. Blank areas indicate Beaufort sea
states >4 that were not included in the survey effort. The 200-,
500-, 1000-, 2000-, and 3000-m isobaths are shown.
Methods
Study area and survey design
The study area (573,000 km^) was North Atlantic Ocean
waters between central Florida (28.00°N) and Maryland
(38.00°N) from the 10-m isobath to the boundary of the
U.S. EEZ, generally 371 km (200 nmi) from the nearest
U.S. point of land (Fig. 1). The study area has a diverse
bottom topography and includes a very narrow continental
shelf (<200 m) at Cape Hatteras which broadens to form
the mid-Atlantic Bight to the north and the Florida-Hat-
teras Shelf to the south. Beyond the shelf, south of Cape
Hatteras are found the following features: the Florida-Hat-
teras Slope, the Blake Plateau (700-1000 m deep), and the
Blake Escarpment. North of the Blake Plateau, the conti-
nental slope from 200-2000 m deep is steep and most of
the study area has water depths >2000 m. The Gulf Stream
is the dominant oceanographic feature in the study area.
From the south, the Gulf Stream Front generally follows
the upper continental slope northward to Cape Hatteras,
where it flows to the northeast. Seaward of the Gulf Stream
are Sargasso Sea waters. North ol' Cape Hatteras and the
Gulf Stream Front, cooler waters, which largely originate
in the Labrador Sea, drift into the study area from the
north and northeast.
Transects covered the study area uniformly in a saw-
tooth pattern from a random start at the southernmost
inshore point and were surveyed from the 68-m NOAA ship
Relentless (renamed Gordon Gunter in 1999) between 8 Ju-
ly and 17 August 1998 from south to north, and from north
to south. Transects were placed to cross the bathymetry
gradient. The narrow band of U.S. waters between central
Florida and Key West, Florida, were partially surveyed but
were not included in the present report.
Data collection
Data were collected by two teams of three observers from
the ship's flying bridge, located 14.5 m above the surface of
the water, during daylight hours, weather permitting (i.e.
no rain, Beaufort sea state <6). Observers used standard
line-transect survey methods for cetaceans that were simi-
lar to those used from ships in the Pacific Ocean and Gulf
of Mexico (e.g.. Barlow, 1995; Hansen et al.'*). Each team
had at least two members experienced in shipboard line-
transect methods and in the identification of tropical and
temperate cetaceans. Two observers searched for cetaceans
using 25x binoculars and another observer searched using
unaided eye or 7x hand-held binoculars and recorded data.
These three observers constituted the "primary team." From
18 July to 17 August, a fourth observer was added to one
team to act as a conditionally independent observer (CIO,
see below). The area from 90° left and right of the ship's
bow to the horizon was searched by the primary team.
Observers changed position (including the CIO position)
every 30^0 minutes, and each team alternated two-hour
watches throughout daylight hours. The survey speed was
usually 18 km/li but varied with sea conditions.
Data were recorded on a computer interfaced with a glob-
al positioning system (GPS) by a data acquisition program.
Data collected for each cetacean sighting included time,
position, bearing, and reticle (a measure of radial distance)
of the sighting, species, group-size, behavior, bottom depth,
sea surface temperature, and associated animals (e.g. sea-
birds, fish). The bearing and radial distance for sightings
that were close to the ship were estimated. Survey effort
data were automatically recorded every two minutes and
included position, heading, effort status, observer position,
and environmental conditions that could affect the observ-
ers' ability to sight animals (e.g. Beaufort sea state, position
of the sun).
Typically, if a sighting was within a 5.5-km strip on ei-
ther side of the ship, the ship was diverted from the tran-
sect line and approached the group so that observers could
identify species and obtain group-size estimates. For each
sighting, the final group-size was estimated by a consensus
■* Hansen, L. J., K. D. Mullin, T. A. Jefferson, and G. P.
Scott. 1996. Visual surveys aboard ships and aircraft. In
Distribution and abundance of marine mammals in the north-
central and western (lulf of Mexico: final report; vol, II: technical
report (R.W. Davis and G. S. Fargion. eds.), p. 5.5-132. Outer
Contnu-ntal Shelf (OCS) Study MMS 96-0027. U.S. Dep. Inte-
rior, Minerals Mgnit. Service, Gulf of Mexico OCS Region, New
Orleans, LA.
Mullln and Fulling: Abundance of cetaceans in the southern Atlantic Ocean
605
Table 1
Number of on-effort cetacean group sightings of each species or
other taxonomic category during 4163 km of survey effort in the |
southern U.S. Atlantic study area during summer 1998.
Species
are listed
in categories pooled to estimate /(O) (see Table 2). The
number of sightings used for line-transect and strip-transect abundance estimates are indicated for each species.
flO) groupings and species
Line-transect
Strip-transect
Large whales
Fin whale [Balaenoptera physalus)
1
0
Minke whale (B. acutorostrata)
1
0
Sperm whale (Physeter macrocephalus)
29
0
Unidentified large whale
6
0
Cryptic whales
Dwarf and pygmy sperm whale (Kogia spp.)
9
0
Mesoplodon spp.
4
0
Unidentified Ziphiidae
3
0
Unidentified small whale
4
0
Unidentified odontocete
12
0
Small whales and large dolphins
Pilot whale (Globicephala spp.)
10
0
Bottlenose dolphin (.Tursiops truncatus)
35
3
Risso's dolphin (Grampus griseiis)
22
0
"Coastal" Atlantic spotted dolphin (Stenella frontalis)
24
3
Unidentified T. truncatus or S. frontalis
7
1
Rough-toothed dolphin (Steno bredanensis)
1
0
Small dolphins
Pantropical spotted dolphin (Stenella attenuata)
6
0
Striped dolphin [Stenella coeruleoalba)
5
0
Clymene dolphin (Stenella clymene)
2
1
"Offshore" Atlantic spotted dolphin (Stenella frontalis)
1
0
Unidentified dolphins
Unidentified dolphins
26
0
Stenella spp.
1
0
Total
209
8
of the primary team. Mixed-species groups were uncom-
mon (five of 217 sightings) and group-size estimates were
made separately for each species.
Species identification
Cetaceans were identified to the lowest taxonomic level
possible from descriptions in field guides and scientific
Uterature (e.g. Leatherwood and Reeves, 1983; Jefferson
et ah, 1993; Carwardine, 1995) (Table 1). An observer's
ability to make identifications depended on weather and
animal behavior. The study area was potentially inhabited
by short-finned pilot whales (Globicephala inacrorhynchus),
which are thought to occur within the study area from about
Virginia south, and long-finned pilot whales (G. melas),
thought to occur from near Cape Hatteras north (Pa3me and
Heinemann, 1993). Because the two species cannot be reli-
ably distinguished at sea, they were recorded simply as pilot
whales. Two forms of the Atlantic spotted dolphin (Stenella
frontalis) were tentatively identified: the larger, more
coastal form, and the smaller offshore form (Perrin et al..
1994 ). Abundances were estimated for each form and for all
Atlantic spotted dolphins combined because only one stock
is currently designated for U.S. Atlantic waters. Coastal and
offshore forms of bottlenose dolphins (Hersh and Duffield,
1990), which constitute the two stocks, were recorded, but
most sightings could not be clearly categorized; therefore,
all bottlenose dolphin sightings were pooled for one overall
abundance estimate. Bottlenose and Atlantic spotted dol-
phins could not always be distinguished at large distances
and a separate estimate was made for animals that could
not be approached and were identified as ''Tursiops or S.
frontalis." Overall abundances for the genus Kogia and the
genus Mesoplodon were estimated. Dwarf sperm whales (K.
sima) and pygmy sperm whales (K. breviceps) were difficult
to distinguish and stranding records of both species are
numerous from U.S. Atlantic shores (Schmidly''). Based on
5 Schmidly, D. J. 1981. Marine mammals of the southeastern
United States and the Gulf of Mexico. U.S. Dep. Interior, U.S.
Fish and Wildlife Service Biological Services Program FWS/
OBS-80/41, 165 p.
606
Fishery Bulletin 101 (3)
Table 2
Estimate of /lO) for each species group (see Table l).n = number of sightings used for the estimate
of /JO ) before truncation (included
in n is the number of sightings in parentheses that occurred while the ship was in
transit in or near the
study area). Truncation = |
the perpendicular distance, >>, at which groups with a greater y were excluded from the analysis.
ESW =
effective strip
width.
/lO)
CV
Truncation
ESW
Species group n
(/km)
1/tO)]
(m)
(m)
Large whales 38 (1)
0.300
0.12
5500
6666
Cryptic whales 33 (1)
0.561
0.13
3000
3565
Small whales and large dolphins 121 (22)
0.498
0.10
4000
4016
Small dolphins 20 (6)
0.398
0.11
4500
5025
Unidentified dolphin 27 (0)
0.496
0.10
4000
4032
Total 239 (30)
stranding records of mesoplodont whales from U.S. Atlantic
shores, sightings of Mesoplodon were probably True's (M.
mirus), Gervais's (M. europaeus) or Blainville's (M. densiro-
stris) beaked whales (Mead, 1989). In some cases cetaceans
could only be identified as large whales (>7 m long), small
whales (nondolphin, <7 m), dolphins, or odontocetes.
Analytical techniques
For each species or taxonomic category, abundance esti-
mates (N) were made with line-transect methods by using
the software program DISTANCE (Colorado Coop. Fish
and Wildlife Research Unit, Colorado State Univ., Fort
Collins, CO) (Buckland et al., 1993) with the equation
A' =
A n 5/(0)
2Lg{0)
where A = size of the study area;
n = number of on-efTort group sightings;
S = mean group-size estimate;
/!0) = sighting probability density function at per-
pendicular distance zero;
L = total length of transect line; and
giO) = probability of seeing a group on the transect
line.
The log-normal 95% confidence interval was computed
for each abundance estimate because it was a product of
estimates and tends to have a skewed distribution. The
variance of N was estimated as
\-dT(N) = N-
var(») var(5) var[/(0)] var[i,'(0)]
»r "^ S' f(Of f(Of
and the coefficient of variation (CV) was estimated as
CViN):
Vvar(/V)
N
The sampling unit was the length of the transect completed
on-efTort each day when the Beaufort sea state was <5. The
formula used to estimate each component of the variance is
given in Buckland et al. ( 1993). Var(«) was length-weighted
and based on the variation in the number of on-effort group
sightings between sampling units that ranged in length
from 39 to 229 km/day.
Estimation of M)
The perpendicular distance, >>, was estimated by using bear-
ing and reticle measurements. The reticle readings were
converted to radial sighting distances (/?) by the method
of Lerczakand Hobbs (1998), and the formula v = i?sin(6),
where b = angle between the sighting and the transect
line. Estimates of /iO) were made by using a hazard-rate,
uniform, or half-normal model with exact perpendicular
sighting distances. For each species group, outlying values
ofy were truncated to improve the fit of the model (Table 2).
Model selection was determined by using Akaike's informa-
tion criterion (AIC; Buckland et al., 1993).
The number of groups sighted of most species was insuf-
ficient to obtain an estimate of /lO). Therefore, sightings of
species with similar sighting characteristics (i.e. body size,
group-size, surface behavior, blow visibility) were pooled to
estimate /(O) for five categories (Table 1). The abundance
for each species was estimated by using the pooled /1 0) and
var|/10)l for its category. The varl/!0)l was assumed to be
zero for the strip-transect estimates explained below. If
the individual detection functions of all species within a
category are indeed very similar, by pooling, the variance,
CV, and confidence interval of each abundance estimate
was probably underestimated because the variance of /lO)
was based on an artificially high sample size. On the other
hand, if the true detection functions of the species within
a category are highly variable, the variance of /lO) for an
individual species may be overestimated.
During the study, effort was sometimes maintained while
in transit to and from ports or along the border of the study
area, but it usually occurred in a small range of water
depths (e.g. parallel to shore) and was excluded because
it could have biased abundance estimates. However, due
to the small number of sightings for the survey, y from the
"transit" sightings were pooled with the on-effort sightings
for estimates of /lO) (Table 2).
Mullin and Fulling: Abundance of cetaceans in the southern Atlantic Ocean
607
Estimation of mean group-size
The group-sizes for most species tended to be related to
y, because in many cases larger groups are easier to see
than small groups with increasingy. In general, the arith-
metic mean of group-size may be an overestimate of the
true mean group-size and could lead to positively biased
abundance estimates. Therefore, a regression of group-size
by y was used to estimate an "expected mean group-size"
(program DISTANCE). The expected mean group-size was
used in the abundance estimate if it was smaller than
the arithmetic mean group-size. For estimates based on
a small number of sightings, the expected mean group-
size was sometimes greater than the arithmetic mean.
Because group-size estimates were usually made after
the ship approached the group, this was assumed to be an
artifact of the small sample size, and the arithmetic mean
was used in these cases. Var(S) was the analytical vari-
ance for mean group-sizes based on arithmetic means or
was estimated as in Buckland et al. ( 1993:79) for expected
mean group-sizes.
Strip-transect estimates
One requirement for unbiased line-transect estimates of
abundance is that the cetacean group should not move in
response to the ship before it is sighted (Buckland et al.,
1993). If cetaceans are not sighted before they respond to
the ship, in cases of attraction to the ship, /TO) and abun-
dance will be overestimated. In the Gulf of Mexico, five
species appear to be consistently attracted to ships to ride
the bow waves (i.e. bottlenose, Atlantic spotted, spinner
[S. longirostris], Clymene [S. clymene], and pantropical
spotted dolphins [S. attenuata]) (Wiirsig et al., 1998). All
sightings made with 25x binoculars had radial distances
>665 m and were assumed to be made before these species
were attracted to the ship. If sightings of these species were
made at radial distances <665 m, because of the possibility
of attraction, they were not included in the line-transect
abundance estimate, and a separate strip-transect abun-
dance estimate was made with these sightings. For each
species, the width of the strip for strip-transect estimates
was set at the line-transect strip width (l/2/(0)) for that
species (Tables 1 and 2). This procedure yields the same
result as the formula given above with flO) for the spe-
cies-group category. However, f\0) for small dolphins and
for small whales and large dolphins combined was not
positively biased by including sightings of groups that
were probably attracted to the transect line. For each spe-
cies, the line- and strip-transect estimates were summed
for one overall abundance estimate.
Conditionally independent observer
The central assumption for estimating abundance with
line-transect methods is that cetacean groups on the tran-
sect line are detected with certainty (i.e.giO) = 1; Buckland
et al., 1993). However, this assumption is usually not met
during cetacean surveys because of availability and per-
ception bias (i.e.g(O) < 1) (Marsh and Sinclair, 1989). Some
groups on the transect line are missed because they may
not be at the surface during the time the ship is in the area
and are not available to be seen, whereas other groups at
the surface are missed by observers (i.e. not perceived)
because of factors such as observer experience, sea state,
and animal behavior, among others.
An attempt was made to estimate g(0) due to perception
bias with a conditionally independent observer (CIO) by
using methods based on Barlow ( 1995). The CIO was used
when the 4-observer team was on duty and was stationed
at 25x binoculars located on a bridge-wing 2.7 m below
the primary team. One individual switched teams each
day; therefore all seven observers on the ship acted as the
CIO at different times. The CIO searched for cetaceans
near the transect line (from 30° left to 30° right of the
bow) when the primary observers were on-effort. The CIO
and the primary team could not see or hear each other.
Whenever the primary team made a sighting, the data
recorder relayed its bearing and reticle to the CIO. When
the CIO made a sighting, the time, bearing and reticle
were noted by the CIO, and the sighting was monitored
until it was sighted by the primary team or, theoretically,
passed abeam, at which time the CIO was to notify the
primary team to divert the ship to identify the species and
estimate group-size.
Results
Abundance estimates were based on 4163 km of effort
in Beaufort sea states <,4 and 217 on-effort sightings of
cetacean species or other taxonomic categories (Fig. 1 and
Table 1). At least 13 cetacean species were sighted. The
most commonly sighted species (number of sightings)
were bottlenose dolphins (38), sperm whales iPhyseter
macrocephalus) (29), Atlantic spotted dolphins (28), and
Risso's dolphins (Grampus griseus) (22). Thirty sightings
occurred during transit in Beaufort sea states s4 (861 km)
and were used to estimate f(Q). Estimates of /iO) ranged
from 0.300/km for large whales to 0.561/kin for cryptic
whales (Table 2).
Conditionally independent observer
The CIO achieved 1775 km of effort (35% of effort, including
transit, with Beaufort sea state s4) and sighted 21 cetacean
groups. Of these, six groups ranging in size from 1 to 10 ani-
mals were missed by the primary team and included three
unidentified dolphin groups, two unidentified odontocete
groups, and one Mesoplodon sp. Each of these sightings was
observed briefly by the CIO but could not be tracked until
they passed the beam of the ship; however, in each of the six
cases no sightings were made by the primary team during
the time frame it would have been possible to sight them.
To estimate g^(O) following the analytical methods described
by Barlow (1995), a separate estimate of /(O) is therefore
required for CIO sightings missed by the primary team.
Because there were only six of these, g(0) could not be esti-
mated for any /lO) category, andi^(O) = 1 and varlgiO)] = 0
was used in each abundance estimate.
608
Fishery Bulletin 101(3)
Figure 2
Locations of on-effort sightings of bottlenose dolphins (n=38),
"coastal" (n=27) and "offshore" (n = l) Atlantic spotted dol-
phins, pantropical spotted dolphins (n=6), striped dolphins
(n=5), and Clymene dolphins (n=2). The 200-, 500-, 1000-,
2000-, and 3000-m isobaths are shown.
Abundance
The following were the most abundant species (abundance;
coefficient of variation) observed in our study: Atlan-
tic spotted dolphins (14,438; 0.63); bottlenose dolphins
(13,085; 0.40); pantropical spotted dolphins (12,747; 0.56);
and striped dolphins (S. coeruleoalba) (10,225; 0.91); and
Risso's dolphins (9533; 0.50). Sperm whales were the most
abundant large whale (1181; 0.51). Abundances for other
species or taxonomic categories ranged from 20 to 6086.
There were an estimated 77,139 (0.23) cetaceans in the
study area.
Group sizes
Mean group sizes for balaenopterids, physeterids, and
ziphiids were less than three animals per group. Bottle-
nose dolphins, pilot whales, Risso's dolphins, and "coastal"
Atlantic spotted dolphins were in groups that averaged
12-18 animals. The average group sizes of pantropical
spotted, Clymene, and striped dolphins ranged from 75 to
110 individuals (Table 3).
Distribution
Cetaceans were distributed throughout the study area,
but few sightings occurred on the eastern Blake Plateau
Figure 3
Locations of on-effort sightings of Risso's dolphins (7i=22),
pilot whales (n=10), rough-toothed dolphins (n = l). Mesoplo-
don spp.(n=4), and unidentified beaked whales (n=3). The
200-, 500-, 1000-, 2000-, and 3000-m isobaths are shown.
(Fig. 1). The distribution of species varied regionally and
by water depth (Fig. 2-4, Table 4). Bottlenose dolphins and
"coastal" Atlantic spotted dolphins were sighted throughout
the study area but primarily in or near continental shelf
waters. Pilot whales and Risso's dolphins were widely dis-
tributed seaward of the continental shelf Sperm whales,
unidentified large whales, "offshore" Atlantic spotted dol-
phins, striped dolphins, and Clymene dolphins occurred
almost exclusively in oceanic waters (>200 m) from Cape
Hatteras northward. Most pantropical spotted dolphin
sightings were in the southern part of the study area.
Discussion
Abundance
Cetacean abundances for the entire study area have not
been estimated previously. Based on stranding records
and previous surveys within or near the study area, all
the species encountered were expected to be sighted. Pre-
vious abundance estimates for the western North Atlantic
stocks of short-finned pilot whales and of dwarf and pygmy
sperm whales, 749 (0.64) and 420 (0.60), respectively, were
based on a winter 1992 ship survey in U.S. oceanic waters
(>200 m) south of Cape Hatteras (Waring et al., 1997). The
1992 dwarf and pygmy sperm whale estimate is similar to
our estimate (580; 0.57); although the 1998 study area was
Mullin and Fulling: Abundance of cetaceans In the southern Atlantic Ocean
609
Table 3
Group size, density and abundance estimates of cetaceans
n the southern U.S. Atlantic Ocean during summer 1998 (n
= number
of on-effort group sightings after truncation, S =
mean group-size estimate, D = animals/100 km''^
N = abundance estimate, CV =
coefficient of variation, LCI and UCI =
= lower and
upper limits of a log-
normal 95% confidence interval).
Species
/!
S
CV(S)
D
AT
Cy(N)
LCI
UCI
Fin whale
1
2.0
—
0.007
41
1.15
6
270
Minke whale
1
1.0
—
0.004
20
1.29
3
156
Sperm whale
28
2.1
0.12
0.206
1181
0.51
445
3136
Dwarf/pygmy sperm whale
8
1.9
0.16
0.101
580
0.57
197
1708
Mesoplodon spp.
4
2.3
0.28
0.061
348
0.76
88
1376
Unidentified Ziphiidae
3
1.7
0.40
0.034
193
0.71
49
755
Pilot whale
9
16.6
0.19
0.892
5109
0.41
2302
11,341
Bottlenose dolphin
line-transect
31
11.8
0.29
2.194
12,571
0.42
5600
28,222
strip-transect
3
5.0
0.12
0.090
514
0.82
118
2249
sum
2.284
13,085
0.40
6098
28,077
Risso's dolphin
18
15.4
0.26
1.664
9533
0.50
3684
24,671
Atlantic spotted dolphin
"coastar
line-transect
21
17.6
0.25
2.211
12,670
0.71
3471
46,244
strip-transect
3
7.3
0.39
0.132
754
0.64
211
2696
sum
2.343
"offshore"
line-transect
1
37.0
—
0.177
1014
0.85
223
4618
strip-transect
0
sum (coastal and offshore)
2.520
14,438
0.63
4672
44,618
Unid. T. truncatus or S. frontalis
line-transect
7
3.9
0.27
0.162
926
0.73
246
3480
strip-transect
1
1.0
—
0.006
34
0.99
6
189
sum
0.168
960
0.71
276
3334
Rough-toothed dolphin
1
8.0
—
0.048
274
1.03
47
1584
Pantropical spotted dolphin
6
77.5
0.25
2.225
12,747
0.56
4420
36,763
Striped dolphin
5
74.6
0.21
1.785
10,225
0.91
2072
50,449
Clymene dolphin
line-transect
2
110.0
0.37
1.053
6031
0.94
1138
31,963
strip-transect
1
2.0
—
0.010
55
1.15
8
361
sum
1.063
6086
0.93
1293
28,652
Stenella spp.
1
15.0
—
0.089
512
1.15
77
3392
Unidentified large whale
6
1.2
0.14
0.025
143
0.58
48
426
Unidentified small whale
3
2.7
0.63
0.054
309
0.86
53
1796
Unidentified odontocete
11
1.4
0.14
0.101
580
0.36
284
1181
Unidentified dolphin
20
1.2
0.13
0.113
775
0.51
291
2066
Sum (all cetaceans)
13.462
77,139
0.23
49,649
119,850
much larger and many sightings occurred north of Cape
Hatteras(Fig.4).
Abundances have also been estimated for small portions
of the study area, but direct comparisons to our estimates
are difficult. Seasonal abundances were estimated for
about 26 cetacean species or genera encountered in U.S.
continental shelf and slope waters between Cape Hat-
teras and Canada during aerial surveys conducted from
1978 to 1982 (CeTAP2), including at least 11 species or
genera sighted during our survey. Fritts et al."' sighted
12 cetacean species during seasonal aerial surveys of the
continental shelf and southern Blake Plateau off central
610
Fishery Bulletin 101 (3)
North
Atlantic
Ocean
X Fin whale
O Minke whale
+ Sperm whale
n Dwarf/pygmy sperm whale
70
Figure 4
Locations of on-effort sightings of sperm whales (;!=29),
minke whales (n=l), fin whales (n=l), and dwarf and pygmy
sperm whales (n=9). The 200-, 500-, 1000-, 2000-, and 3000-m
isobaths are shown.
Florida from 1980 to 1981; eight of these were sighted
during our survey.
The abundance estimate reported in the present study
for the Clymene dolphin represents the first for this spe-
cies in any portion of the U.S. Atlantic EEZ. However, the
estimate was based on only three sightings and has a large
95% confidence interval (1293-28,652 dolphins). The Cly-
mene dolphin was recognized as a valid species in 1981
and is sympatric with the spinner dolphin in the tropical
Atlantic (Perrin et al., 1981). The two species have similar
color patterns, and in previous studies both were possibly
recognized as S. longirostris and were not distinguished
(CeTAP2; Fritts et al.^). The identifications of Clymene
dolphins were made by observers with experience from
the Gulf of Mexico where both species are relatively com-
mon (Hansen et al.*- Hansen et al.^). There is currently no
stock designation for the Clymene dolphin in U.S. Atlantic
waters (Waring et al., 2001).
Our estimate of bottlenose dolphins is for waters >10
m in depth; however, this estimate does not include their
entire water depth range in U.S. Atlantic waters south of
Maryland. Bottlenose dolphins occur year-round in coastal
waters <10 m in depth (the inshore boundary of the study
area) and in some bays and estuaries from Cape Hatteras
south. North of the Cape they have been found close to
shore in waters <25 m in depth only during warm months
'^ Hansen, L. J., K. D. Mullin, and C. L. Roden. 1995. Estimates
of cetacean abundance in the northern Gulf of Mexico from vessel
surveys, 9 p. Southeast Fisheries Science Center, P.O. Drawer
1207, Pascagoula, Mississippi 39568.
Table 4
Mean water depth and sea surface temperature
of cetacean
species sighted
in the southern
U.S. Atlantic Ocean during summer |
1998 (ra=number of groups sighted
on-effort; SE =
= standard error).
Species
n
Water depth (m)
Sea surface temperature (°C)
Mean
SE
Range
Mean
SE
Range
Fin whale
1
48
—
—
25.1
—
—
Minke whale
1
3475
—
—
29.5
—
—
Sperm whale
29
3252
122
2195-4389
29.0
0.28
22.8-29.9
Dwarf and pygmy sperm whale
9
2586
493
766-4079
29.6
0.37
26.9-30.9
Mesoplndon spp.
4
2699
735
774-4353
27.2
1.65
24.0-31.1
Unidentified Ziphiidae
3
1817
832
878-3475
29.8
0.15
29.6-30.1
Pilot whale
10
1527
387
251^280
28.5
0.73
23.2-31.5
Bottlenose dolphin
38
371
89
12-2561
29.3
0.27
23.2-31.3
Risso's dolphin
22
1300
285
44-4755
28.4
0.62
22.9-31.3
"coastal" Atlantic spotted dolphin
27
216
117
13-2524
29.1
0.26
25.1-31.3
"offshore" Atlantic spotted dolphin
1
4298
—
—
27.9
—
—
Rough-toothed dolphin
1
4353
—
—
27.3
—
—
Pantropical spotted dolphin
6
1498
708
598-5030
30.5
0.61
27.6-31.6
Striped dolphin
5
2736
237
2012-3475
23.9
0.37
22.9-25.1
Clymene dolphin
3
756
,538
139-1829
27.9
0.59
26.8-28.8
Mullin and Fulling: Abundance of cetaceans in the southern Atlantic Ocean
611
and are assumed to have migrated along-shore from the
south (Mead, 1975: Kenney, 1990). Aerial surveys of bottle-
nose dolphins conducted in the past along the U.S. Atlantic
included waters from the shore to 10 m in depth. For waters
typically <75 m deep south of Cape Hatteras, the winter
1992 abundance from an aerial survey was 12,435 (0.18)
(Blaylock and Hoggard, 1994). For waters <25 m deep from
Cape Hatteras to northern New Jersey, the abundance from
a summer 1994 aerial survey was 26,809 (0.40) (Blaylock,
1995). The frequency of bottlenose dolphin sightings during
these surveys increased substantially inshore of the 10-m
isobath boundary of the ship study area and, compared
with the estimate from the ship, may account for the gen-
erally larger aerial survey estimates even though they are
for smaller study areas.
There are currently two genetically distinguishable bot-
tlenose dolphin stocks designated in the U.S. Atlantic: the
coastal stock and the offshore stock (LeDuc and Curry, 1998;
Waring et al., 2001). Using mitochondrial DNA from skin
biopsy samples obtained during the summer 1998 study
and other sampling efforts, Torres et al. (in press) reported
no offshore form was sampled within 6 km of shore and no
coastal from was sampled beyond 39 km from shore or in
waters >34 m deep. Therefore an area of overlap of the two
forms occurs within the 1998 study area but the fraction
of each stock in our estimate (13,085: 0.40) is unknown
because the number of biopsy samples between the two
boundaries was very small in the Torres et al. (in press)
study. However, 20 of the 38 bottlenose groups we used to
estimate abundance were found in waters >50 m deep.
Abundances were estimated for ten species and three
other genera of cetaceans, but other species are known or
expected to occur in the study area. Three of these species,
right whales iEubalaena glacialis), humpback whales
(Megaptera novaeangUae), and harbor porpoises (Phocoena
phocoena), occur in the study area seasonally, primarily
in months other than summer months, and abundances
have been estimated from studies of their primary summer
ranges north of the study area (e.g. Knowlton et al., 1994;
Palka, 1995; Smith et al., 1999).
Additional species expected in at least part of the study
area include Bryde's whale (B. edeni), Cuvier's beaked
whale (Ziphius cauirostris), pygmy killer whale (Feresa at-
tenuata), false killer whale (Pseudorca crassidens), melon-
headed whale iPeponocephala electro), killer whale (Orci-
nus orca), common dolphin (Delphinus delphis), spinner
dolphin, and Fraser's dolphin (Lagenodelphis hosei). Each
of these species is thought to have a tropical to subtropical
or broader distribution worldwide (Jefferson et al., 1993),
and except for the common dolphin, an abundance estimate
for each species is available for the adjacent northern Gulf
of Mexico (Hansen et al.^). However, except for the spin-
ner dolphin, each of these species is relatively uncommon
in the northern Gulf of Mexico and was not encountered
every year during four annual spring surveys with effort
similar to that in our survey (Hansen et al.''). Therefore,
many of these species may also be uncommon in the At-
lantic study area and were simply not encountered during
the 1998 survey. During a late summer 1999 ship survey
of the inner half of the southern Atlantic study area that
targeted bottlenose dolphins, a group of Fraser's dolphins
and melon-headed whales was sighted in water 3000 m
deep east of Cape Hatteras (Roden^).
Some species may also inhabit the study area seasonally.
During the 1992 winter ship survey south of Cape Hatteras
(Mullin and Ford'), five groups of balaenopterid whales
were recorded, three of which were classified as uniden-
tified Bryde's or sei whales. Also during the winter 1992
survey, groups of false killer whales and Cuvier's beaked
whales were sighted twice, and pygmy killer whales once.
Common dolphins were sighted between Cape Hatteras
and Maryland in all seasons, except summer, during the
CeTAP^ study but were sighted once in this area during the
late summer 1999 survey (Roden^). Common dolphins are
expected to occur throughout the area surveyed in 1998 but
they may not. Although there are stranding records south
of Cape Hatteras (Schmidly''), there are no valid stranding
or sighting records of common dolphins in the adjacent
Gulf of Mexico despite extensive seasonal surveys of the
northern Gulf (Jefferson, 1995; Hansen et al."*).
Precision
The precision of the abundance estimates was generally
poor. For species or genera abundances, only the estimate
for bottlenose dolphins (the most commonly sighted spe-
cies), Risso's dolphins, and pilot whales had a CV s 0.50
(Table 3). The abundance estimate for the Atlantic spotted
dolphin, the most abundant species, had a CV = 0.63. In
cases where there is human-caused mortality in a cetacean
stock, abundance estimates with a CV < 0.50 are gener-
ally required to avoid incorrectly classifying a cetacean
stock as "strategic" under the U.S. MMPA (i.e. annual
human-caused mortality > annual PBR) less than 10*^ of
the time (Wade and DeMaster, 1999). For most species, the
variance in the encounter rate, varin ), accounted for more
than 70% of the var(N). The distribution of most species
was not uniform in the study area and precision might
be improved by stratifying estimates by water depth (e.g.
shelf and nonshelf) and by area (e.g. north and south of
Cape Hatteras).
Biases in abundance estimates
The survey was designed to meet the assumptions of line-
transect theory (Buckland et al., 1993). However, the abun-
dance estimates are negatively biased to varying degrees
because the central assumption, that cetacean groups on
the transect line are detected with certainty (i.e. ^(0)=1),
was not met, and data were not available to correct esti-
mates for perception and availability bias. By using the
CIO methods described by Barlow (1995), we attempted
to estimate the fraction of groups missed on the transect
line by the primary observers due to perception bias. How-
' Roden, C. L. 1999. Report of NOAA ship Oregon II cruise
99-05 (236) (a cetacean survey of U.S. Atlantic continental shelf
and slope waters between New Jersey and central Florida.
August-September 1999), 32 p. Southeast Fisheries Science
Center, P.O. Drawer 1207, Pascagoula, Mississippi 39568.
612
Flshen/ Bulletin 101(3)
ever, there were too few sightings to make ^^(0) estimates
because the overall group encounter rate was lower than
anticipated and the CIO was only used for 35% of the total
survey effort. In the future, a CIO should be used whenever
the primary team is on-effort and the CIO should search an
area larger than 30° left and right of the bow. Although the
data in proximity of the transect line are most critical for
estimating ^(0), it is also necessary to have enough data to
estimate fiO) for groups missed by the primary team.
More work is needed to develop methods for estimating
g(0) in relation to perception bias in the southern U.S. At-
lantic. Completely independent observers cannot be used
because the ship has to be diverted from the transect line
to identify species and make group-size estimates. Because
many groups can easily be lost once sighted, the ship must be
diverted well before the group passes abeam. Barlow ( 1995)
used a CIO that searched the same area as the primary team
with unaided eye or 7x binoculars. The 25x binoculars were
used in our study to increase the number of CIO sightings
and avoid attraction bias in/tO). Previous experience in the
Gulf of Mexico has indicated that many unaided-eye sight-
ings would be of small groups of species that are attracted
to the ship to ride the bow waves. Conversely, small groups
are the most difficult for an independent observer to track
with 25x binoculars because the ship is not diverted and the
bearing to the group is constantly changing.
Similar to Barlow's (1995) findings on perception bias,
the majority of groups missed by the primary team were
apparently small groups, although the group-sizes were
not estimated at close range. Barlow (1995) estimated g'(O)
ranging from 0.73 and 0.79 for small groups of delphinids
(<21l and cryptic species (which usually occur in small
groups), andg(O) = 1 for groups of >20 delphinids. In ad-
dition to group-size, the magnitude of perception bias is
dependent on behavior, weather (e.g. Beaufort sea state),
and the observer: active groups are less likely to be missed
than resting groups or species whose behavior does not
produce pronounced cues (e.g. blows, splashes).
Availability bias varies by species because of differences
in individual dive cycles, group diving behavior, and group-
sizes. Long-diving sperm whales and beaked whales will
be at the surface for much less time than will many small
delphinids, which have much shorter dive cycles. Diving syn-
chrony among members of a group also affects availability
bias; if dives are a-synchronous, the probability that at least
one animal will be at the surface increases with group size.
Barlow ( 1999) estimated both availability and perception
bias for long-diving whales during ship surveys using 25x
binoculars in a simulation study and estimated that for
dwarf and pygmy sperm whales, Cuvier's beaked whales,
and Mesoplodon spp., abundance estimates need to be in-
creased 2 to 4 times (i.e.g(0)=0.50 to g( 0=0.25) to account
for these biases. Barlow's (1999) estimates of^(O) for per-
ception or availability bias (or both) are probably represen-
tative of the bias in the southern Atlantic survey because
similar ship survey methods were used. However, it may
not be valid to apply them directly to our abundance esti-
mates because cetacean diving behavior and group sizes
may be temporally and geographically specific, and survey
conditions and observers may vary among surveys.
For the strip-transect estimates (Table 2), use of the
line-transect strip width (2xl//(0)] from the 25x binocular
sightings as the strip width was assumed to be conserva-
tive and somewhat negatively biased. The distance from
which animals will come to the ship to ride the bow is
unknown, and variable, depending on factors such as the
animals' previous behavior, number of opportunities for
riding bow waves, and the type of ship. If the strip width
was too narrow, the strip-transect estimates would overes-
timate abundance.
The geographical bathymetric range of the bottlenose
dolphin was not covered during the survey. Because bottle-
nose dolphins undertake seasonal movements in the study
area, in order to estimate the entire population size, ship
survey estimates need to be combined with same-season
abundance estimates from coastal waters <10 m and in-
shore waters (bays, sounds, and estuaries).
Distribution
Water-depth distributions of cetacean species were for the
most part similar to those in the Gulf of Mexico (Mullin et al.,
1994; Davis et al., 1998). Bottlenose dolphins and Atlantic
spotted dolphins inhabit the continental shelf and shelf-edge
region, whereas most other species have primarily oceanic
distributions. The offshore form of the Atlantic spotted dol-
phin has not been identified in the northern Gulf of Mexico.
The sightings of some species were highly regional (e.g.
sperm whales, striped dolphins, Clymene dolphins, pantropi-
cal spotted dolphins) were probably heavily influenced by
oceanographic features such as the Gulf Stream. Much more
survey effort is needed in summer and other seasons before
conclusions can be drawn about each species' distribution.
Acknowledgments
Many people made significant contributions to the success
of this survey including the officers and crew of NOAA ship
Relentless and C. Roden, the Field Party Chief The Relent-
less was configured for marine mammal surveys through
the dedicated efforts of W. Hoggard. The marine mammal
observers were C. Brown, C. Burks, C. Gates, W. Hoggard,
C. Hubard, T Martinez, K. Maze-Foley, M. Newcomer, S.
Swartz, J. Tobias, and K. Touhey. Environmental and ich-
thyoplankton data were collected by L. Bero, P. Brown, W.
Fambrough, D. Fertl, A. Hamilton, A. Hohn, R. Holmes, E.
Keith, E. LaBrecque, J. Litz, and J. Taylor. W. Irvin and T.
Pusser were seabird observers. The survey was designed
with the help of S. Swartz and the late R. Blaylock. Com-
ments by C. Hubard, K. Maze-Foley, and three anonymous
reviewers were very helpful in completing the manuscript.
Literature cited
Barlow. J.
IBM. The abundance of cetaceans in California waters. Part
I; Ship surveys in summer and fall of 1991. Fish. Bull. 93:
1-14.
Mullin and Fulling: Abundance of cetaceans in the southern Atlantic Ocean
613
1999. Trackline detection probability for long-diving whales.
In Marine mammal survey and assessment methods (G. W.
Garner et al., eds. I, p. 209-221. A.A. Balkema, Rotterdam.
Bariow, J., S. L. Swartz. T. C. Eagle, and P. R. Wade.
1995. U.S. marine mammal stock assessments: guidelines
for preparation, background, and a summary of the 1995
assessments. U.S. Dep. Commer., NOAA Tech. Memo.
NMFS-OPR-6, 73 p.
Blaylock, R. A.
1995. A pilot study to estimate abundance of the U.S.
Atlantic coastal migratory bottlenose dolphin. U.S. Dep.
Commer.. NOAA Tech. Memo. NMFS-SEFSC-362, 9 p.
Blaylock, R. A., and W. Hoggard.
1994. Preliminary estimates of bottlenose dolphin abun-
dance in the southern U.S. Atlantic and Gulf of Mexico
continental shelf waters. U.S. Dep. Commer., NOAA Tech.
Memo. NMFS-SEFSC-356, 10 p.
Buckland, S. T, D. R. Anderson, K. P. Bumham. and J. L. Laake.
1993. Distance sampling: estimating abundance of biological
populations, 446 p. Chapman and Hall, London.
Carwardine, M.
1995. Whales, dolphins and porpoises, 256 p. Dorling
Kindersley, New York. NY.
Davis, R. W, G. S. Fargion, N. May, T. D. Leming,
M. Baunigartner, W E. Evans, L. J. Hansen, and K. D. Mullin.
1998. Physical habitat of cetaceans along the continental
slope in the north-central and western Gulf of Mexico. Mar
Mamm. Sci. 14:490-507.
Hersh, S. L., and D. A. Duffield.
1990. Distinction between northwest Atlantic offshore and
coastal bottlenose dolphins based on hemoglobin profile and
morphometry. In The bottlenose dolphin (S. Leatherwood
and R. R. Reeves (eds.l, p. 129-142. Academic Press, San
Diego, CA.
Jefferson, T.A.
1995. Distribution, abundance, and some aspects of the biol-
ogy of cetaceans in the offshore Gulf of Mexico. Ph.D. diss.,
232 p. Texas A&M University College Station, TX.
Jefferson, T. A., S. Leatherwood, and M. A. Webber.
1993. Marine mammals of the world — FAO species identifi-
cation guide, 320 p. FAO, Rome.
Kenney, R. D.
1990. Bottlenose dolphins off the northeastern United
States. In The bottlenose dolphin ( S. Leatherwood and R. R.
Reeves, eds.) p. 369-386. Academic Press, San Diego, CA.
Knowlton, A. R., S. D. Kraus, and R. D. Kenney.
1994. Reproduction in North Atlantic right whales [Euba-
laena glacialis). Can. J. Zoo.72:1297-1305.
Leatherwood, S., and R. R. Reeves.
1983. The Sierra Club handbook of whales and dolphins, 302
p. Sierra Club Books, San Francisco, CA.
LeDuc, R. G., and B. E. Curry.
1998. Mitochondrial DNA sequence analysis indicates need
for revision of the genus T\irsiops. Rep. Int. Whaling
Comm. 47:393.
Lerczak, J. A., and R. C. Hobbs.
1998. Calculating sighting distances from angular readings
during shipboard, aerial, and shore-based marine mammal
surveys. Mar. Mamm. Sci. 14:590-599.
Marsh, H., and D. F. Sinclair
1989. Correcting for visibility bias in strip transect aerial sur-
veys of aquatic fauna. J. Wildl. Manage. 53:1017-1024.
Mead, J. G.
1975. A preliminary report on the former net fishery for
TYfrsiops truncatus in the western North Atlantic. J. Fish.
Res. Board Can. 32:1155-1162
1989. Beaked whale of the genus Mesoplodon. In Hand-
book of marine mammals. Volume 4: River dolphins and the
larger toothed whales ( S. H. Ridgway and R. Harrison ( eds. ),
p. 349-430. Academic Press, San Diego, CA.
Mullin, K. D., W. Hoggard, C. L. Roden, R. R. Lohoefener,
C. M. Rogers, and B. Taggart.
1994. Cetaceans on the upper continental slope in the
north-central Gulf of Mexico. Fish. Bull. 92:773-786.
Palka, D.
1995. Abundance estimate of the Gulf of Maine harbor
porpoise. In Biology of the phocoenids (A. Bjorge and
G.P. Donovan, eds.), p. 27-50. Rep. Int. Whaling Comm.
(special issue 16).
Payne, P. M., and D. W. Heinemann.
1993. The distribution of pilot whales {Globicephala spp.) in
the shelCshelf-edge and slope waters of the northeastern
United States, 1978-1988. In Biology of northern hemi-
sphere pilot whales (G. P. Donovan, C. H. Lockyer and A. R.
Martin, eds.), p. 51-68. Rep. Int. Whaling Comm. (special
issue 14).
Perrin, W. F, D. K. Caldwell, and M. C. Caldwell.
1994. Atlantic spotted dolphin — Stenella frontalis. In
Handbook of marine mammals. Volume 5: The first book of
dolphins (S.H. Ridgway and R. Harrison, eds.) p. 173-190.
Academic Press, San Diego, CA.
Perrin, W F, E. D. Mitchell, J. G. Mead, D. K. Caldwell and
P. J. H. van Bree.
1981. Stenella clyrnene, a rediscovered tropical dolphin of
the Atlantic. J. Mamm. 62:583-598.
Smith, T D., J. Allen, P. J. Clapman, P. S. Hammond, S. Katona,
F. Larsen, J. Lien, D. Mattila, P. J. Palsboll, J. Sigurjonsson,
R T. Stevick, and N. 0ien.
1999. An ocean-basin-wide mark-recapture study of the
North Atlantic humpback whale (Megaptera novaeangliae).
Mar. Mamm. Sci. 15:1-32.
Torres, L. G., P. E. Rosel, C. D'Agrosa, and A. J. Read.
In press. Improving management of overlapping bottlenose
dolphin ecotvpes through spatial analysis and genetics.
Mar Mamm. Sci.
Wade, P. R., and D. P. DeMaster.
1999. Determining the optimum interval for abundance
surveys. In Marine mammal survey and assessment
methods (G. W Gamer, S. C. Amstrup, J. L. Laake, B. F. J.
Manly, L. L. McDonald, and D. G. Robertson, eds.), p. 53-66.
A.A. Balkema, Rotterdam.
Waring, G. T, D. L. Palka, K. D. Mullin, J. W Hain, L. J. Hansen,
and K. D. Bisack.
1997. U.S. Atlantic and Gulf of Mexico marine mammal
stock assessments — 1996. U.S. Dep. Commer, NOAA
Tech. Memo. NMFS-NE-114, 245 p.
Waring, G. T., J. M. Quintal, and S. L. Swartz (eds).
2001. U.S. Atlantic and Gulf of Mexico marine mammal
stock assessments — 2001. U.S. Dep. Commer., NOAA
Tech. Memo. NMFS-NE-168, 310 p.
Wursig, B., S. K. Lynn, T A. Jefferson, and K. D. Mullin.
1998. Behavior of cetaceans in the northern Gulf of Mexico
relative to survey ships and aircraft. Aquat. Mamm. 24:
41-50.
614
Abstract — The growth of red sea
urchins (Strongylocentrotus francisca-
nus) was modeled by using tag-recap-
ture data from northern Cahfornia.
Red sea urchins (n=211) ranging in
test diameter from 7 to 131 mm were
examined for changes in size over one
year. We used the function J,^, = "^i +
fit/,) to model growth, in which J, is the
jaw size (mm) at tagging, and J,^[ is the
jaw size one year later. The function
RJ,), represents one of six deterministic
models: logistic dose response, Gauss-
ian, Tanaka, Ricker, Richards, and von
Bertalanffy with 3, 3, 3, 2, 3, and 2 min-
imization parameters, respectively. We
found that three measures of goodness
of fit ranked the models similarly, in the
order given. The results from these six
models indicate that red sea urchins
are slow growing animals (mean of
7.2 ±1.3 years to enter the fishery). We
show that poor model selection or data
from a limited range of urchin sizes
(or both) produces erroneous growth-
parameter estimates and years-to-
fishery estimates. Individual variation
in growth dominated spatial variation
at shallow and deep sites (F=0.246,
n=199, P=0.62). We summarize the six
models using a composite growth curve
of jaw size, J, as a function of time, t: J
=A(B - e-'^') + Dt, in which each model is
distinguished by the constants A, B, C,
and D. We suggest that this composite
model has the flexibility of the other six
models and could be broadly applied.
Given the robustness of our results
regarding the number of years to enter
the fishery, this information could be
incorporated into future fishery man-
agement plans for red sea urchins in
northern California.
Modeling red sea urchin
iStrongylocentrotus franciscanus) growth
using six growth functions*
Laura Rogers-Bennett
California Department of Fistn and Game and
University of California, Davis
Bodega Marine Laboratory
2099 Westside Rd
Bodega Bay, California 94923-0247
E-mail address; rogeRbennettis/ucdavis edu
Donald W. Rogers
Chemistry Department
Long Island University
Brooklyn, New York 11201
William A. Bennett
John Muir Institute of the Environment
University of California, Davis
Davis, California 95616
Thomas A. Ebert
Biology Department
San Diego State University
San Diego, California 92182
Manuscript approved for publication
5 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:614-626
Marine invertebrates are being fished
at an increasing pace worldwide (Kees-
ing and Hall, 1998). In California,
invertebrates have a greater exvessel
(wholesale) value than do fin-fish
(Rogers-Bennett, 2001). Invertebrate
fisheries are now experiencing seri-
ous declines as have fin-fish fisheries
(Dugan and Davis, 1993; Safina, 1998;
Jackson et al., 2001). The once prosper-
ous commercial abalone fishery in Cali-
fornia which landed in excess of 2000
metric tons per year in the 1950s and
1960s was closed in 1997 (CDFG Code
5521) following the serial depletion of
stocks over time (Karpov et al., 2000).
Commercial divers now target red sea
urchins and other invertebrates. Red
sea urchin landings in California have
also declined dramatically from a high
of 24 metric tons (t) in 1988 to 6 t in
2002, despite management efforts (Kal-
vass and Hendrix, 1997). The.se declines
have generated interest in exploring the
use of alternative fishery management
policies, such as spatially explicit strat-
egies that would protect large old sea
urchins (Rogers-Bennett et al., 1995).
Sea urchin growth models are criti-
cal in the development of innovative
management strategies to sustain the
fishery because, among other things,
models can be used to predict the time
required for sea urchins to enter the
fishery (referred to as "years to fish-
ery") and the age of the broodstock.
Despite the interest in examining sea
urchin growth, modeling efforts have
been hampered by several factors in-
cluding model selection and a lack of
data from a sufficiently wide range of
urchin sizes. Perhaps as a consequence,
estimates of red sea urchin growth
have varied widely, ranging from 3
to 12 years for urchins to grow into
the fishery (Kato and Schroeter, 1985;
Tegner, 1989; Ebert and Russell, 1992;
Smith et al., 1998). Because of the wide
variation in growth estimates, the num-
ber of models and methods being used,
and the difficulties that these present
* Contribution 2176 from the Bodega Marine
Laboratory, University of Davis, Davis, CA
94923-0247.
Rogers-Bennett et al,: Modeling growth of Strongylocentrotus fianciscanus
615
for management, there is a need to evaluate a number of
growth models with a single data set that encompasses a
large range of urchin sizes.
In our study we report the results from six individual
growth models applied to data from a one-year tag and
recapture study of red sea urchins {Strongylocentrotus
franciscanus) in northern California. We supplemented
the number of juveniles in the field by stocking tagged ju-
veniles. Estimates of the number of years required for ur-
chins to gi-ow to minimum legal size in northern California
are generated by the models. We examine the robustness of
these results to changes in the parameters and the impact
of a limited data set from a small range of urchin sizes on
our results. We determine if there are spatial differences
in growth between shallow and deep sites. Finally, we rank
the models according to quality of fit, present a generic
growth curve that combines the six models, and discuss the
implications of our results for fishery management.
Materials and methods
Study sites
Growth rates were determined for red sea urchins in
the Salt Point (38°33'06"N, 123n9'45"W) and Caspar
(39°21'49"N, 123°49'47"W) urchin harvest reserves in
northern California. Commercial urchin harvesting is pro-
hibited in these reserves. We examined spatial variation
within Salt Point by tagging red sea urchins at one shal-
low site (5 m) south of the southern border of the Gerstle
Cove Reserve and at one deep site (17 m) on the leward
side of a large wash-rock. In addition, laboratory-reared
juvenile red sea urchins were stocked at the two sites in
Salt Point. Both of these sites are relatively isolated, sur-
rounded by sand and seasonally dense kelp (Nereocystis).
At the Caspar Reserve, sea urchins were tagged outside a
small cove with seasonally dense kelp (Nereocystis) at a
single depth (7 m).
Tagging
Sea urchins at the study sites were tagged internally and
recaptured after one year. At Salt Point, wild sea urchins
were tagged with tetracycline injections in situ by using
0.5-1.2 mL of 1 g tetracycline/100 mL of seawater (cf
Ebert, 1982; Ebert and Russell, 1992). Six hundred and
nine red urchins were measured with vernier calipers
(±1-2 mm) and tagged at Salt Point on 19 August 1992.
Urchins were recaptured from the Salt Point sites on 18
September 1993 (;!=374 shallow; n=S52 deep). This data
set was normalized to one year by using the factor 12/13.
Our study was not a longitudinal study examining growth
over many years, but rather for one year only.
Juvenile urchins reared in the laboratory for one year
(mean test diameter=17.6 mm) were tagged and stocked
into the shallow and deep Salt Point sites. Juveniles were
tagged by immersion for 24 hours in a calcein solution 125
mg/L seawater, pH adjusted to 8.0 (Wilson et al., 1987).
After tagging, juveniles were transported to the Salt Point
sites and released. Juveniles were stocked (120 at each
of the two depths) on 31 August 1992 and harvested on
18 September 1993 with the adults (see Rogers-Bennett,
2001).
Urchins at the Caspar Reserve were tagged internally
with personal individual transponder (PIT) tags on 28
August 1996 and recovered 20 August 1997 (Kalvass').
PIT tags are glass coated mini-transponders with unique
individual codes that can be read noninvasively by using
a Destron(S) tag reader. Tags were implanted into the body
cavity of the sea urchins through the peristomial mem-
brane. PIT tags are too large for tagging small urchins
(<40 mm).
Estimates of urchin density were made within a circle
(12 m in radius) at each of the two Salt Point sites at the
time of harvest. Drift algae collections were made along a
2 X 10 m transect (20 m^) at each site. Gut contents were
collected from a subsample of 20 urchins from each site.
Gut contents were fixed in alcohol, sorted on a petri dish,
and the most abundant items were recorded from 5 out
of 25 lO-mm'^ grids (Harrold and Reed, 1985). We used a
conservative definition of optimal foods, defining them as
fleshy red or brown algae (Harrold and Reed, 1985). Sub-
optimal foods included green algae, upright and encrust-
ing coralline algae, detritus (animal, plant, and inorganic),
plants (Phyllospadix), mud, and sand.
Growth measurements
Sea urchins can not be reliably aged by using rings on
test ossicles (Pearse and Pearse, 1975; Ebert 1988; Gage,
1992), therefore growth increments after one year must be
measured directly. For the urchins tagged with fluorescent
dyes (tetracycline and calcein), growth was measured as
the change in urchin jaw length (AJ =J,^j-J,) after one
year (Ebert and Russell, 1993). Urchin jaws were dissected
from Aristotle's lantern, excess tissue was removed with
10% sodium hypochlorite, and the jaws were measured to
the nearest 0. 1 mm. Growth was measured by determining
the width of the calcium deposit one year after tagging.
Tags on jaws are more accurate than tags on test ossicles
because ossicles move toward the oral surface during
growth (Duetler, 1926), requiring matching ossicles at the
time of tagging with ambitus ossicles at the time of collec-
tion (Ebert, 1988).
Fluorescence tagged urchins were identified when ex-
posed to an ultraviolet epi-illuminator (Lite-Mite) on a
dissecting scope. Growth increments were determined by
using the Confocal Microscope (BioRad MRC-600, BioRad
Industries, Hercules, CA) with a BHS fluorescence filter
(blue wavelength) and the COMOS software package
(BioRad Industries, Hercules, CA). Growth was measured
from the fluorescent band (indicating size at tagging) to
the esophageal edge of the jaw (final size). Growth was
also recorded from a second growth zone at the labial tip of
the jaw, represented by a glowing arc when present. Initial
jaw size (J,) equals jaw size after one year (J,^i) minus the
' Kalvass, P. 1997. Personal commun. Calif. Dep. Fish and
Game, 19160 S. Harbor Dr., Fort Bragg, CA. 95437.
616
Fishery Bulletin 101(3)
Table 1
Tests for homogeneity of slopes for In {diameter)
compared with In (Jau
) for shallow and
deep samples of red sea
urchins from Salt
Point, California: SS: sum of squares
df degrees of freedom; MS: mean square. Treatment (depth
1 is significant P=
=0.017 when
adjusted for covariate (test diameter).
SS
df
MS
F-ratio
P
Homogeneity of slopes
Sample
0.013
1
0.013
3.643
0.058
In (jaw)
21.216
1
21.216
5760.0
0.000
Sample x \n(jaw)
0.010
1
0.010
0.723
0.101
Error
0.718
195
0.004
Adjusted means
Sample
0.021
1
0.021
5.747
0.017
In (jaw)
29.333
1
29.333
7894.0
0.000
Error
0.728
196
0.004
sum of the esophageal and labial growth. Urchin jaws do
not wear or erode as teeth do. Calculating test growth from
changes in jaw size may yield a conservative estimate for
sublegal red sea urchins (Kalvass et al., 1998).
In the PIT-tagged urchins, growth was measured as the
change in test diameter after one year. Juvenile urchins
less than 30 mm are too small to survive PIT tag implan-
tation. Large adults (>100 mm) may grow too little in one
year to allow growth in test diameter to be measured.
Standard errors in measures of test diameter with calipers
range from 1-2 mm which may be greater than the growth
increment in adults.
Jaw size versus test size
The relationship between jaw length and test diameter was
determined from a large sample {n=384) of red sea urchins
(sample independent of this study) ranging in size from
newly settled individuals to large adults. From this sample
we obtained an allometric equation relating jaw and test
size. Using this equation we converted all the measures
of growth (from fluorescent and PIT tagged urchins) into
initial and final jaw size (one year later).
Jaw size is a plastic trait that can vary spatially (Ebert,
1980b; Rogers-Bennett et al., 1995). Food availability has
been correlated with the size of Aristotle's lantern (com-
posed of ten jaws and five teeth) such that lanterns are large
when food is scarce. Therefore, we examined the relationship
between jaw size and test size, segregating the data from
the shallow and deep Salt Point sites. To do this we used
an ANCOVA (Table 1) with the natural log of test diameter
as the covariate. Measurements of the jaw size at tagging
from the fluorescent marks allowed for estimates of the test
size at tagging using the allometric relationship (Eq. 1). As
a control to test for bias in the conversion of jaw size to test
diameter with Equation 1, we compared the measured test
size at the time of recapture to the predicted test size at the
time of recapture using the allometric relationship (Eq. 1).
Results indicated that, although there was error in the pre-
dicted test size from jaw size, there was no bias.
Results
Red urchin growth
We present growth data from a total of 211 red sea urchins
that were tagged internally and recaptured after one year
in northern California. Recaptured urchins ranged in test
diameter from 7 to 131 mm at the time of tagging. We recov-
ered 161 out of 609 (26.4%) tetracycline-tagged wild urchins
from the two sites at Salt Point. In addition, 38 of the 240
(15.8%) stocked juvenile urchins tagged with calcien were
also recovered. It is unknown whether untagged urchins
included tagged adults which were not growing and therefore
not taking up the tetracycline stain. In the Caspar Reserve
12 of 53 (22.6%) PIT-tagged wild urchins were recovered.
We examined spatial variation in growth and found that
the change in size (AJ) was not significantly different for
urchins in the shallow, as compared to the deep Salt Point
sites (ANCOVA F=0.246, n = 199, P=0.62) with initial jaw
size (J,) as the covariate (independent variable). Similarly,
growth rates were not significantly different between
juveniles recovered in shallow and deep sites (ANCOVA
F=0.387, 11=38, P=0.54). Richards function parameter esti-
mates (J^, K, n ) generated from the shallow site were sta-
tistically identical to those for the deep site. Size-frequency
distributions of urchins recovered from the shallow site
were not significantly different than those at the deep site
(K-S mean difference=0.162, P=0.67). Therefore, growth
data from the shallow and deep sites were pooled.
Urchin density at the shallow site (4.2 urchins/m'-^) was
greater than at the deep site (0.75 urchins/m^). In addition,
drift algae abundance was twice as great in the shallow
(2.7 g/m^) as in the deep site (1.4 g/m'-) at the time of urchin
harvest ( 18 September 1993 ). This resulted in less algae per
urchin in the shallow site compared with the deep site (0.63
g/urchin and 1.85 g/urchin respectively) for that sampling
date. Guts of urchins from the deep habitats contained
more optimal food (fleshy red and brown algae) than guts
of urchins in the shallow sites «=2.79, df=19, P=0.012). Gut
fullness was generally uniform, roughly 50 mL/urchin.
Rogers-Bennett et al,: Modeling growth of Strongylocentrotus franciscanus
617
Jaw size versus test size
ANCOVA analysis indicated that the slopes of
the natural log of test diameter as a function of
the natural log of the jaw size are homogenous
(P=0.101 ), but that the adjusted means are signifi-
cantly different (P=0.017) — urchins in the deeper
habitat having larger jaws (Table 1). Therefore,
we constructed two allometric equations, one for
urchins from the shallow Salt Point site and a
second for urchins from the deep Salt Point site.
However, the two equations were so similar that
they generated identical test diameters for a given
jaw size; therefore we pooled our data from the
shallow and the deep sites.
We used a larger independent data set of /i=384
from wild and cultured urchins to generate the
allometric equation relating jaw size to test size.
There is a strong relationship (r^=0.989, df=382)
between test diameter (D) and jaw length (J) de-
scribed by
£> = 3.31Jii5 (1)
where D = test diameter (mm); and
J = jaw length (mm).
Equation 1 predicts that urchins of legal size in
northern California (test diameters £89 mm) have
jaw lengths alT.S mm.
A comparison (using the allometric relationship
[Eq. 1|) of the measured test size at the time of
recapture with the predicted test size revealed no
bias in the conversion. Although individual values
of measured and predicted test diameters are not
identical, the sum of the differences between the
two reveals no strong directional bias. The sum
of the differences between the measured and
predicted values equals 41 mm for 139 urchins,
resulting in an average discrepancy of 0.30 mm
per urchin. This discrepancy is smaller than the
initial error in the measurement of test size (see
"Materials and methods" section).
The von Bertalanffy model
For many organisms, annual growth rate decreases as size
(age) increases. This process is frequently modeled by using
the von Bertalanffy equation (von Bertalanffy, 1938)
10-1
A
8 •
\.
6 ■
\^^
4 •
2 -
' \.
•\^»
n
U 1 1 1 1 ]
0 5 10 15 20 25
~3
<1
10 n
B
8 ■
6 -
4 •
•
^""^•v. MS _
• *^*v.« • • •
•• *^\^^^
2 ■
^^^-,^^
(
0 5 10 15 20 25
J
Figure 1
(A) Linear function fitted to the middle portion of the urchin growth
data set. (B) Linear von Bertalanffy function superimposed on the
entire data set. c7=jaw length (mm); 4J = ch£uige in jaw length (mm).
J,^i =J,+ J J 1 - e'^) - J,( 1-e-''')
J = JA1-
-Kl
),
(2)
(2a)
which leads to a linear decrease in growth rate as a func-
tion of size. We make the point here, that J,^j and J, refer
to a discrete data set, whereas J is a smooth, continuous
function of ^
Our data and, quite possibly, much of the data collected
in similar studies, are not well represented by the von
Bertalanffy equation. How is it then that the deficiencies
of this well used equation have not come to light? The
answer, not surprisingly, lies in the cancellation of errors
within data sets that only incompletely cover the critical
growth period.
Our data ( Figs. IB and 2) show three features of sea ur-
chin growth that are inconsistent with the von Bertalanffy
model: 1) annual growth, AJ = J,^, -J, , for juveniles that
is lower than anticipated from the model; 2) a maximum
or plateau in the growth function, AJ = fiJf), for urchins
near jaw size J( = ^ mm (test diameter 20 mm); and 3) an
asymptotic approach of 4 J to zero (Figs. IB and 2), which
may be ascribed to indeterminate growth for adults of all
sizes or to dispersion of final adult urchin sizes (Sainsbury,
1980).
There is a good deal of individual variation in growth
rate as a function of J^ , which prevents unequivocal selec-
618
Fishery Bulletin 101(3)
6 -
•
A
5 -
•
•
• •
\ •
\ •
4 -
3 ■
••
• \
2 -
\ •••
1 -
V4^
•
10
15
20
25 30
6n
■
B
5 -
•■••.:■
4 ■
/ * • *
• \
\ •
\ •
^a
3 -
2 ■
1 ■
n ■
• •
\4^
15
20
25
6 -|
c
5 -
4 -
/ * • •
•
\
• •
•
3 -
/ •*
2
^
\«
1 -
^%
^
i*-r^
10
20
25
Figure 2
Annual fjrowth as a function of jaw size for six models: lopislic dose-
response, Gaussian, Tanaka, Ricker, Richards, and von Birtalanffy
models. J=jaw length (mm); 4J = change in jaw length (mm).
tion of one model. Nevertheless, it is clear that
the von Bertalanffy model does not represent
the data well over the full range of urchin
sizes iJf). To investigate this point further, we
divided our data set into three groups over the
range of of J, . The groups are
1 Juveniles (J, < 8 mm) that do not fall on the
linear descent of 4 J versus J, characteristic
of von Bertalanffy growth.
2 Sublegal, actively growing adults (8 mm<
J,<16 mm) that do follow von Bertalanffy
kinetics.
3 Adults ( 16 mm< J, < 24 mm) that appear to
grow to large J^ but only very slowly, and do
not conform to the von Bertalanffy model.
If the data were fitted to the von Bertalanffy
equation, all three groups should give the same
slope AJ IJf because three segments of the same
straight line all have the same slope. Instead,
group 1 gives a small positive slope, group 2
gives a negative slope that leads to unrealistic
conclusions for early growth rate and time-to-
fishery estimates shown in (Fig. lA), and group
3, excluding growth information from sublegal
urchins, yields a plausible mean final jaw size
of 22.6 mm but gives a growth rate constant
that indicates very slow growth for adults and
many decades for time-to-fishery.
In the present study we fitted a decreasing,
linear von Bertalanffy function only to the
sublegal (group 2) urchins (Fig. lA) which did
conform to von Bertalanffy growth. The von
Bertalanffy function for the partial data set of
actively growing urchins in (Fig. 1) has a slope
of -0.504/yr, a AJ intercept of 8.7 mm/yr, and
a Jf intercept of 17.3 mm. These results lead
one to predict that final grow to 90'7f of their
final size in 3.5 years and that mean final size
will be less than the legal size (89 mm test
diameter), which is obviously false. We also
show the same function superimposed on the
entire data set (Fig. IB) where discrepancies
between the von Bertalanffy function and data
groups 1 and 3 above are evident. For our data
set (Fig. 1) and, we suggest, for urchin growth
in general, the von Bertalanffy curve does not
represent early growth, and a transition curve
or a peaked function reflects actual growth
better For our data set, the von Bertalanffy
model gives an overestimate of the rate of ur-
chin growth and an underestimate of the time
to enter the fishery.
The slopes of these three line segments give
an indication how the von Bertalanffy model,
despite its implausible fit to the complete data
set, can give plausible growth parameters. Er-
rors in fitting a von Bertalanffy curve to a data
set resembling ours lie in opposite directions
Rogers-Bennett et al.: Modeling growth of Strongylocentrotus franciscanus
619
for groups 1 and 3 of the growth; consequently
they cancel, in whole or in part. In fact, all re-
ported data sets have many more observations
falling into group 3 than into group 1, which
is either swamped out by group 3 or does not
appear at all. This leaves gi'oups 2 and most
or all of group 3 to determine the slope of the
von Bertalanffy linear function. The average of
these two erroneous slopes may or may not be
a realistic approximation for urchin growth,
depending on the number of measurements in
each group.
Alternative growth models
Curves that rise to a maximum and then decay
asymptotically are very common in the physical
sciences and have been successfully modeled
for more than a century (e.g. Wien, 1896). Any
rising function multiplied into an exponential
decay, e.g. (.v) exp(-.x:), models such a curve
more or less well. The problem is not in find-
ing a model but in selecting from among many
possibilities. We compared several models in
our study and included a Gaussian model for
this data set because it has a small sum of
squared residuals and because it has well-
defined parameters in the arithmetic mean and
standard deviation. Here the arithmetic mean
merely serves to fix the position of the maxi-
mum on the Jf axis and the standard deviation
from f.1 gives the range, in units of J^, of actively
growing animals. The model is descriptive only;
it does not imply a mechanism of growth.
We present results from six growth models,
the logistic dose-response, Gaussian, Ricker,
Tanaka, Richards, and von Bertalanffy mod-
els, in order of quality of fit (Fig. 2). Each
model is characterized by a different AJ =
fiJf), where f(Jf) is a function of annual
growth AJ versus size at tagging, J^. Equa-
tions 3-8 were input as user-defined functions
into a curve-fitting program (TableCurve,
Jandel Scientific, now SPSS, Chicago, IL),
either as f^Jf) or the equivalent J^ j^-^ - J^.
In certain cases, additive parameters that
make a negligible contribution to the final fit
were dropped. This curve-fitting program uses
the Levenburg-Marquardt procedure for find-
ing the minimum of the squared sum of devia-
tions. During the least-squares minimization,
local minima are occasionally found and must
be discarded in favor of the global minimum.
Matrix inversion is performed by the Gauss-
Jordan method (Carnahan et al., 1969).
We present these models ranked by the fit-
ting criterion of the sum of squared residuals,
called "Error Sum of Squares" in the output
from the TableCurve fitting program, which
we have given the abbreviation RSS. Several
6 -|
•
D
5 -
••.•
4 -
/ * * *
•
• •
3 -
/
\
/ * *
v
2 -
/
\^
1 -
A
/
^:<^i.
U
C
1
5
1 r » ♦ T* ■■•• *~n
10 15 20 25
8 -
E
6 -
\
^ 4-
\ * • .'
V*-'
• >
•
• •
^?.
2 -
>^-
r\
'• wbfcfifc^-- .
U
1
} 5
10 15 20 25
6 -
5 -
•
F
4 -
• • • \.
•
• • •
3 -
• •
x*
2 -
•\^ * •
1 -
•-.Mju* .
0 ■
Tmm^m* • •
1
0 5
10 15 20 25
J
Figure 2 (continued)
620
Fishery Bulletin 101 (3)
Table 2
A comparison of the fitt
ng
criteria for six
red
sea urchin
growth functions.
r2 =
the coefficient of determination; RSS = the error |
sum of squares;
AIC
= Akaike
s information en
terion; SBC
= Schwartz-Bayesian
criterion.
r2
SE
RSS
AIC
SBC No. of parameters
Logistic
0.946
0.392
31.9
-393
-383 3
Gaussian
0.945
0.397
32.8
-387
-377 3
Tanaka
0.933
0.436
39.6
-347
-337 3
Ricker
0.918
0.483
48.7
-305
-299 2
Richards
0.900
0.534
59.4
-262
-251 3
von Bertalanffy
0.895
0.545
62.1
-254
-247 2
other fitting criteria are also given in Table 2. We used both
the AIC information criterion, A/C = KlniRSS)- K\nK+2m,
and the Schwartz-Bayesian criterion BIC = KlniRSS)-
(K-m )\n{K), where K is the number of data points, and
m is the number of parameters in the fitting equation
(Akaike, 1979). These tests of curve-fitting quality were
used to bring out any substantive difference between the
2-parameter and 3-parameter equations. The results show
that differences between the 2- and 3-parameter cases are
swamped out by the data, as might have been anticipated
from the disparity between the number of points (iiL'=211)
and the number of parameters. For the present data set, in
applying either of these criteria, one is essentially seeking
the smallest RSS.
Individual models
Logistic dose-response The logistic dose-response curve
(time-to-fishery estimate: 6.6 yr)
f\Ji^=a/a+{J,/br)
(3)
(Hastings, 1997) fits our data the best of all the models
examined here. The curve fit (a=4.4, 6=12.9, c=6.8) with
RSS = 31.9, is a sigmoidal transition function (TableCurve
Windows, vers. 1.0, Jandel Scientific Corp., SPSS, Chicago,
IL). There is a transition between a fast-growing group
of sea urchins, which maintain a constant growth rate
{f{Ji)=annua\ AJbAA mm/yr up to about J(=8 mm), to
sea urchins growing .slowly at a rate that diminishes as
Jf increases beyond 16 mm. The inflection point is at J^ ^
13 mm. There is considerable individual variation in both
data groups, but more in the fast-growing group than in
the larger slow-growing group.
Gaussian The Gaussian function (time-to-fishery esti-
mate: 6.9 yr), although rarely if ever used in this context.
f(J,) = Ae
-{j,-nyil 4.6 -i- 5.7 =
10.3 mm represents urchins that are one standard devia-
tion (a) above the mean of the growth curve, i. e., 16% of the
growing population. Urchins with J, > 17.0 mm, are greater
than 2(7 above maximal growth, i. e. 2.5% of the growing
population. Therefore, a growth curve fit only to adult ur-
chins with J, > 17.0 mm represents only small subset of the
total growing population and is not representative of the
total population. This demonstrates that data from a lim-
ited size range can generate erroneous growth parameters
and shorten estimates of time to enter the fishery.
Variation in growth
The plateau in growth rate implied by the logistic dose-
response curve or the maximum in growth rate for juve-
nile urchins well after settlement implied by the Gaussian
curve suggests that urchin growth is not at its maximum
when sea urchins first settle. It is realistic to imagine that
a sea urchin will be at its maximum growth rate sometime
after the first year or two.
In this study we found high individual variation in sea
urchin growth. Growth in juveniles was especially variable,
despite the fact that the juvenile urchins that were stocked
were full siblings. We found no evidence for an increase
in dispersion as sea urchins grow larger. Data from many
sources suggest individual variation in juvenile growth is
high. Full sibling red urchins (rt=200) reared in the labora-
tory under identical food, temperature, and light regimes
varied in test diameter from 4 to 44 mm at one year ( Rogers-
Bennett, unpubl. data). Similarly, cultured purple urchins
(S. purpuratus). varied from 10 to 30 mm at one year
(Pearse and Cameron, 1991) — a trend observed in other
commercially cultured marine invertebrates (Beaumont,
1994) and fishes (Allendorf et al. 1987).
Our data contain broad distribution in the region of the
small size classes, which is consistent with high individual
variation in growth iK). Varying the growth constant, K,
e.g. in the Ricker model (cf Sainsbury, 1980), produces dis-
persion at the smaller size classes. Our urchin growth data
also show a wide array of large sizes as well. Models have
been used to examine the impact of this type of individual
growth variation. In the von Bertalanffy model, if final size,
J^, is varied 10%, this results in a broad distribution of the
largest size classes (Botsford et al., 1994). We see a broad
distribution in the largest size classes in our data, with
animals larger and smaller than the estimated final size
J^. Many of the animals smaller than J^ could be at their
final size. The biological interpretation of this broad distri-
bution at the largest sizes is an open question. There may
be a wide array of final sizes because of independent values
of /iTand J, (cf Sainsbury, 1980) and each individual hits
its own final size abruptly or at an asymptotic approach to
final size (cf Beverton, 1992) also known as indeterminate
growth (cf Sebens, 1987).
We suggest that the composite model presented in the
present study (Eq. 9) may be useful for a wide array of
invertebrates and fishes especially those with a broad ar-
ray of final sizes.
Spatial patterns in growth
In our study, we found no evidence for spatial patterns in
growth. To observe spatial patterns this would have to be
detectable above the background of individual variation.
Sea urchins from the shallow and deep sites at Salt Point
had measurable differences in gut contents, food availabil-
ity, and oceanographic conditions; however these did not
translate into significant differences in growth between
624
Fishery Bulletin 101(3)
the depths over the year examined. Similarly, no latitudi-
nal differences in red sea urchin growth were found in a
large-scale growth study at 18 sites ranging from Alaska
to southern California where growth varied between
neighboring sites as between much as distant sites (Ebert
etal., 1999).
Future studies could be longitudinal and examine tem-
poral patterns in sea urchin growth, for example during
and after warm water El Niiio years, as has been examined
for abalone in southern California (Haaker et al., 1998);
however these temporal patterns too would have to be
greater than individual variation to be detectable.
Implications for fishery management
Large old sea urchins (>125 mm test diameter) are fished
in California despite fishermen receiving lower prices
for these sea urchins compared with mid-size animals
(Rudie^). Many of the large, old urchins have high gonadal
weights (>100 g) (Carney, 1991; Rogers-Bennett et al.,
1995), thereby potentially contributing more to repro-
duction than smaller urchins (Tegner and Levin, 1983;
Tegner, 1989; Kalvass and Hendrix, 1997). Similarly, large
coral-reef fish also have the potential to contribute more to
reproduction than smaller fish (Bohnsack, 1993).
In fished areas, size-frequency distributions are heav-
ily skewed to smaller urchins indicating that the larger
size classes are absent (Kalvass and Hendrix, 1997). If
the abundance and density of red sea urchins is decreased
during fishing, this will decrease the chances of fertiliza-
tion success significantly (Levitan et al., 1992). Sufficient
numbers of large broodstock are critical because recruit-
ment does not appear to be successful every year (Ebert,
1983; Pearse and Hines, 1987; Sloan et al., 1987). In addi-
tion, fishing can impact recruitment success because the
spines of large urchins provide canopy shelter for juveniles;
therefore an Allee effect may be present (Tegner and Day-
ton, 1977; Sloan et al., 1987; Rogers-Bennett et al., 1995).
Size-structured red sea urchin models that include variable
recruitment or an Allee effect (positive density dependence)
resulted in a >50'7f decrease in estimated population size
even at low fishing mortality levels (Pfister and Bradbury,
1996).
Harvest experiments conducted in northern California
have shown that management strategies that protect large
urchins (upper size limits and harvest reserves) improve
recovery and recruitment after six years compared with
strategies in which large urchins are harvested (lower size
limits only) (Rogers-Bennett et al., 1998). Upper size limits
and reserves have been used in the management of the
sea urchin fishery in Washington state (Bradbury, 2000)
and are currently being considered for California's red sea
urchin fishery (Taniguchi^).
2 Rudie, n. 1994. Personal commun. Catalina Offshore Prod-
ucts Inc., 5202 Lovelock St., San Diego, CA 92110.
' Taniguchi, I. 2002. Personal commun. Calif. Dep. Fish and
Game, 4665 Lampson Ave., Los Alamitos, CA. 90720.
In conclusion, our work and that of others (Ebert and
Russell, 1992, 1993; Ebert et. al., 1999) suggest that red
sea urchins are slow growing, long-lived animals. Intense
harvest rates may have serious consequences because red
sea urchins require seven years to reach harvestable size
in northern California. Declines in red sea urchin landings
in northern California of more than 80% from the peak of
13,800 t in 1988 (Kalvass, 2000) demonstrate that harvest
rates are high. Our growth results suggest that proposed
alternative management strategies that would protect
large, slow growing broodstock inside reserves or upper
size limits for the fishery could be beneficial, in addition to
existing regulations, for sustaining the fishery.
Acknowledgments
Special thanks to H. C. Fastenau, D. Canestro and the
U. C. Santa Cruz research dive class (1992) for help tag-
ging and measuring red sea urchins. D. Cornelius and
the "Down Under" helped harvest urchins. P. Kalvass
shared his growth data from PIT tagged sea urchins. F.
McLafferty discussed models and "the most probable sea
urchin." W. Clark, S. Wang, and F. Griffin gave access to
and instruction on the confocal microscope. C. Dewees, H.
Blethrow, S. Bennett, and K. Rogers all contributed. This
research was funded in part by the California Department
Fish and Game, the PADI Foundation, UC. Davis Natu-
ral Reserve System, and the Bodega Marine Laboratory.
Comments from M. Lamare and M. Mangel improved the
manuscript.
Literature cited
Akaike, H.
1979. A Bayesian extension of the minimum AIC procedure
of autoregressive model fitting. Biometrika 66:237-242.
Allendorf, F. W„ N. Ryman, N., and F. M. Utter.
1987. Genetics and fishery management: past, present and
fiiture. In Population genetics and fishery management, p.
1-19. Washington Sea Grant, Seattle, WA.
Atkins, P. W.
1994. Physical chemistry, 5* ed., 1031 p. W.H. Freeman,
New York, NY.
Barrante, J. R.
1998. Applied mathematics for physical chemistry, 2"'' ed.,
227 p. Prentice Hall, Upper Saddle River, NJ.
Beverton, R. J. H.
1992. Patterns of reproductive strategy parameters in some
marine teleost fishes. J. Fish Biol. 41:137-160.
Beaumont, A. R.
1 994 . Genetics and aquaculture. In Genetics and evolution
of marine organisms, p. 467-486. Cambridge Univ. Press,
Cambridge, England.
Bohnsack J.A.
1993. Marine reserves: they enhance fisheries, reduce con-
flicts and protect resources. Oceanus 36:63-71.
Botsford, L. W., B. D. Smith, and J. F Quinn.
1994. Bimodality in size distributions: the red sea urchin
{Strongylncentrotiis franciscanus) as an example. Ecol.
Appl. 4:42-50.
Rogers-Bennett et al.: Modeling growth of Strongylocentrotus franciscanus
625
Bradbury, A.
2000. Stock assessment and management of red sea urchins
[Strongylocentrotus franciscanus) inVilashington. J. Shell-
fish Res. 19:618-619.
Brody, S.
1927. Growth rates. Univ. Missouri Agri. Exp. Sta. Bull. 97.
Burham, K. P., and D. R. Anderson
1998. Model selection and inference: a practical informa-
tion theoretic approach, 353 p. Springer Verlag. New
York, NY.
Camahan, B., H. A. Luthier, and J. O. Wilkes.
1969. Applied numerical methods, 604 p. Wiley Press, New
York, NY.
Carney, D.
1991. A comparison of densities, size distribution, gonad and
total-gut indices and the relative movements of red sea ur-
chins Strongylocentrotus franciscanus in two depth regimes.
M.S. thesis, 43 p. Univ California, Santa Cruz, CA.
Dugan, J. E., and G. E. Davis.
1993. Applications of marine refugia to coastal fisheries
management. Can J. Fish. Aquat. Sci. 50:2029-2042.
Duetler, F.
1926. Uber das Wachstum des Seeigelskeletts. Zool. Jb.
(Abt. Anat. Ontag. Tiere) 48:119-200.
Ebert, T. A.
1980a. Estimating parameters in a flexible growth equa-
tion, the Richards function Can. J. Fish. Aquat. Sci. 37:
687-692.
1980b. Relative growth of sea urchin jaws: an example of
plastic resource allocation Bull. Mar. Sci. 30:467-474.
1982. Longevity, life history, and relative body wall size in
sea urchins. Ecol. Monogr. 52:353-394.
1983. Recruitment in echinoderms. Echinoderm Studies
1:169-203.
1988. Calibration of natural gi'owth lines in ossicles of two
sea urchins Strongylocentrotus purpurtaus and Echinome-
tra mathaei. using tetracycline. In Echinoderm biology (R.
D. Burke, P. V. Mladenov, P. Lambert, and R. L. Parsley, eds. );
proceedings of the sixth international echinoderm confer-
ence, p. 435-443. A.A. Balkema, Rotterdam.
1999. Plant and animal populations: methods in demogra-
phy, 312 p. Academic Press, San Diego, CA.
Ebert, T. A., and M. P Russell.
1992. Growth and mortality estimates for red sea urchins
Strongylocentrotus franciscanus from San Nicolas Island,
California. Mar. Ecol. Prog. Ser. 81:31-41.
1993. Growth and mortality of subtidal red sea urchins
Strongylocentrotus franciscanus at San Nicolas Island,
California, USA: problems with models. Mar. Biol. 117:
79-89.
Ebert, T. A., Dixon, J. D., Schroeter, S. C. Kalvass, P. E. Richmond,
N. T. Bradbury W. A. and D. A. Woodby
1999. Growth and mortality of red sea urchin Strongylo-
centrotus franciscanus across a latitudinal gradient. Mar
Ecol. Prog. Ser 190:189-209.
Gage, J. D.
1992. Natural growth bands and growth variability in the
sea urchin Echinus esculentus: results from tetracycline
tagging. Mar Biol. 114:607-616.
Haaker, P. L., D. O. Parker, K. C. Barsky, and C. Chun.
1998. Growth of red abalone, Haliotis rufescens (Swainson),
at Johnson's Lee Santa Rosa Island, California. J. Shell.
Res. 17:747-753.
Harrold, C, and D. C. Reed.
1985. Food availability, sea urchin grazing, and kelp forest
community structure. Ecol. 66:1160-1169.
Hastings, A.
1997. Population biology concepts and models, 227 p.
Springer- Verlag, New York, NY.
Jackson, J. B. C, M. X. Kirby, W. H. Berger, K. A. Bjorndal,
L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke,
J. Eriandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange,
H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck,
M. J. Tegner, and R. R. Warner.
2001. Historical overfishing and the recent collapse of
coastal ecosystems. Sci. 293:629-638.
Kalvass, P. E.
2000. Riding the roUercoaster: Boom and decline in the
California red sea urchin fishery. J. Shellfish Res. 19:
621-622.
Kalvass, P. E., and J. M. Hendrix.
1997. The California red sea urchin, Stronglyocentrotus
fransiscanus. fishery: catch, effort, and management trends.
Mar. Fish. Rev 59:1-17.
Kalvass, P. E.; J. M. Hendrix, and P. M. Law.
1998. Experimental analysis of 3 internal marking methods
for red sea urchins. Cahf Fish Game 84:88-99.
Karpov, K.A., P. L. Haaker, I. K. Taniguchi, and L. Rogers-Bennett.
2000. Serial depletion and the collapse of the California aba-
lone (Haliotis spp.) fishery. Can. Spec. Publ. Fish. Aquat.
Sci. 130:11-24.
Kato, S., and S.C. Schroeter
1985. Biology of the red sea urchin Strongylocentrotus fran-
ciscanus, and its fishery in California. Mar. Fish. Rev. 47:
1-19.
Keesing, J. K., and K. C. Hall.
1998. Review of harvests and status of world's sea urchin
fisheries points to opportunities for aquaculture. J. Shell-
fish Res. 17:1597-1604.
Levitan, D. R. M. A. Sewell, and F-S. Chia.
1992. How distribution and abundance influence fertiliza-
tion success in the sea urchin, Strongylocentrotus francis-
canus. Ecol. 73:248-254.
Pearse, J. S., and R. A. Cameron.
1991. Echinodermata: echinoidea. In Reproduction of
marine invertebrates, vol VI( A. C. Giese, J. S. Pearse and V. B.
Pearse, eds.), p. 513-662. Academic Press, New York, NY.
Pearse, J. S., and A. H. Hines.
1987. Long-term population dynamics of sea urchins in a
central California kelp forest: rare recruitment and rapid
decline. Mar. Ecol. Prog. Ser. 39:275-283.
Pearse, J. S., and V. B. Pearse.
1975. Growth zones in the echinoid skeleton. Am. Zool. 15:
731-753.
Pfister, C. A., and A. Bradbury.
1996. Harvesting red sea urchins: Recent effects and future
predictions. Ecol. Appl. 6:29-310.
Quinn, T. J., and R. B. Deriso.
1999. Quantitative fish dynamics, 542 p. Oxford Univ Press,
Oxford, England.
Richards, F J.
1 959. A flexible growth curve for empirical use. J. Exp. Bot.
10:290-300.
Ricker, W. E.
1954. Stock and recruitment. J. Fish Res. Board Can. 11:
559-623.
Rogers, D. W.
1983. BASIC microcomputing and biostatistics, 274 p. The
Humana Press Inc., Clifton, NJ.
Rogers-Bennett, L.
2001. Evaluating stocking as an enhancement strategy for
the red sea urchin, Strongylocentrotus franciscanus: depth-
626
Fishery Bulletin 101(3)
specific recoveries. In Proceedings of the 10* international
echinoderni conference; Dunedin, New Zealand, p. 527-531.
A.A. Balkema, Rotterdam.
2001. Review of some California fisheries for 2000: Market
squid, sea urchin, prawn, white abalone, groundfishes, ocean
salmon, Pacific sardine. Pacific herring, Pacific mackerel,
nearshore live-fishes, halibut, yellowfin tuna, white seabass,
and kelp. Calif Coop. Oceanic Fish. Invest. Rep. 42:12-28
Rogers-Bennett, L., H. C. Fastenau, and C. M. Dewees.
1998. Recovery of red sea urchin beds following experimen-
tal harvest. In Proceedings of the 9"^ internation echi-
noderm conference, San Francisco, CA, p. 805-809. A.A.
Balkema, Rotterdam.
Rogers-Bennett, L., W. A. Bennett, H. C. Fastenau,
and C. M. Dewees.
1995. Spatial variation in red sea urchin reproduction and
morphology: implications for harvest refugia. Ecological
Applications 5:1171-1180.
Rowley, R. J.
1990. Newly settled sea urchins in a kelp bed and urchin
barren ground: a comparison of growth and mortality.
Mar Ecol. Prog. Sen 62:229-240.
Safina, C.
1998. Song for the blue ocean, 458 p. Henry Holt, New
York, NY.
Sainsbury, K. J.
1980. Effect of individual variability on the von Bertalanffy
growth equation. Can. J. Fish. Aquat. Sci. 37:241-247.
Schnute, J.
1981. A versatile growth model with statistically stable
parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140
Sebens, K. P.
1987. The ecology of indeterminate growth in animals.
Ann. Rev Ecol. Syst. 18:371-407.
Sloan, N. A., C. P. Lauridsen, and R. M. Harbo.
1987. Recruitment characteristics of the commercially har-
vested red sea urchin Strongylocentrotus frauciscanus in
southern British Columbia, Canada. Fish. Res. 5:55-69.
Smith, B. D., L. W. Botsford, and S. R. Wing.
1998. Estimation of growth and mortality parameters from
size frequency distributions lacking age patterns: the
red sea urchins (Strongylocentrotus franciscanus) as an
example. Can J. Fish. Aquatic Sci. 55:1236-1247.
Tanaka, M.
1982. A new growth curve which expresses infinite increase.
Publ. Amakusa Mar. Biol. Lab. 6:167-177.
Tegner, M. J.
1989. The feasibility of enhancing red sea urchin, Strongy-
locentrotus franciscanus, stocks in California: an analysis
of the options. Fish. Bull. 51:1-22.
Tegner, M. J, and P. K. Dayton.
1977. Sea urchin recruitment patterns and implications of
commercial fishing. Science 196:324-326.
Tegner, M. J. and L. A. Levin.
1983. Spiny lobsters and sea urchins: analysis of a predator-
prey interaction. J. Exp. Mar Biol. Ecol. 73:125-150.
Troynikov, V. S., and H. K. Gorflne
1998. Alternative approach for establishing legal minimum
lengths for abalone based on stochastic growth models for
length increment data. J. Shellfish Res. 17:827-831.
von Bertalanffy, L.
1938. A quantitative theory of organic growth (inquires on
growth laws II). Human Biol. 10:181-213.
Walford, L. A.
1946. A new graphic method of describing the growth of
animals. Biol. Bull. 90:141-147.
Wien, W.
1896. Uber die Energieverteilung im Emissionspectrum
eines Schwarzen Korpers. Annalen der Physik 1896:
662-669. In The conceptual development of quan-
tum mechanics (M. Jammer, author), 2"'' ed. 1989, p
8-10. Tomask Publishers, American Institute of Physics,
Woodby New York, NY.
Wilson, C. W., D. W. Beckman, J. D. Dean, and J. Mark.
1987. Calcein as a florescent marker of otoliths of larval and
juvenile fish. Trans, Am. Fish. Soc. 1 16:668-670.
Yamaguchi, G.
1975. Estimating growth parameters from growth rate data.
Problems with marine sedentary invertebrates. Oecologia
20:321-332.
627
Abstract— Age and growth estimates
for the blue shark [Prionace glauca)
were derived from 411 vertebral centra
and 43 tag-recaptured blue sharks col-
lected in the North Atlantic, ranging in
length from 49 to 312 cm fork length
(FL). The vertebrae of two oxytetracy-
cline-injected recaptured blue sharks
support an annual spring deposition
of growth bands in the vertebrae in
sharks up to 192 cm FL. Males and
females were aged to 16 and 15 years,
respectively, and full maturity is
attained by 5 years of age in both sexes.
Both sexes grew similarly to age seven,
when growth rates decreased in males
and remained constant in females.
Growth rates from tag-recaptured
individuals agreed with those derived
from vertebral annuli for smaller
sharks but appeared overestimated for
larger sharks. Von Bertalanffy growth
parameters derived from vertebral
length-at-age data are L^ = 282 cm FL,
A' = 0.18, and t„ = -1.35 for males, and
L, = 310 cm FL, K = 0.13, and ig = -1.77
for females. The species grows faster
and has a shorter life span than previ-
ously reported for these waters.
Age and growth of the blue shark (Prionace glauca)
in the North Atlantic Ocean*
Gregory B. Skomal
Massachusetts Division of Marine Fisheries
Marthas Vineyard Research Station
PO Box 68
Vineyard Haven, Massachusetts 02568
E-mail address: Gregory.SkomaiiSlstale.ma.us
Lisa J. Natanson
National Manne Fisheries Service, NOAA
28 Tarzwell Dr.
Narragansett, Rhode Island 02882
Manuscript approved for publication
5 March 2003 by Scientific Editor
Manuscript received 4 April 2003 at
NMFS Scientific Pubhcations Office.
Fish Bull. 101:627-639 (2003).
The blue shark {Prionace glauca) is
a large pelagic carcharhinid that is
widely distributed in the world's oceans.
Throughout its range, it is considered
the most abundant species of large
shark (McKenzie and Tibbo, 1964;
Casey, 1982). In the Atlantic, the blue
shark is distributed from Newfoimdland
to Argentina in the west and Norway to
South Africa, including the Mediter-
ranean, in the east (Compagno, 1984).
There is strong evidence from tagging
data and catch records that blue sharks
in the North Atlantic constitute a single
stock (Kohler et al., 2002). Moreover,
mitochondrial DNA d-loop sequence and
nuclear microsatellite analyses indicate
no differences between blue sharks from
the eastern and western North Atlantic
(ShiyjiM.
Distribution and movements of the
blue shark are strongly influenced
by seasonal variations in water tem-
perature, reproductive condition, and
availability of prey (Kohler et al.,
2002). Blue sharks make frequent
trans-Atlantic movements between the
western and eastern regions, utilizing
the major North Atlantic current sys-
tems (Stevens, 1976, 1990; Casey, 1982,
1985; Kohler et al., 2002). Temporal and
geographic patterns of size and sexual
segregation have been described for this
species, and mating areas and pupping
areas are reported to be in the western
and eastern regions of the North Atlan-
tic, respectively (Casey, 1982; Kohler et
al., 2002). Pregnant females are rare in
the western North Atlantic, which is
dominated by juveniles of both sexes,
adult males, and subadult females
(Pratt, 1979; Casey, 1982; Kohler et al.,
2002). Catch records from the eastern
North Atlantic largely comprised neo-
nates and juveniles of both sexes and
adult females (Aasen, 1966; Stevens,
1975, 1976; Connett, 1987; Silva et al.,
1996; Kohler et al., 2002).
Although subjected to a number of
fisheries, the blue shark is primarily
taken as bycatch in longline fisheries
throughout the North Atlantic (ICCAT,
2002). Most blue sharks are discarded
or only their fins are harvested because
of the low palatability of their flesh
(Castro et al., 1999). Although incom-
plete, blue shark landings estimates
in the North Atlantic reported to the
International Commission for the
Conservation of Atlantic Tunas were
25.1 and 24.2 thousand metric tons (t)
in 1998 and 1999, respectively (ICCAT,
2002). Domestic longline fisheries in the
western North Atlantic rarely land blue
sharks, but it was estimated that annu-
al dead discards ranged from 2.8 to 29.3
thousand blue sharks (99.0-1136.3 t)
during the period 1987-2000 (Cortes,
2002). The major source of landings in
the U.S. has been the recreational fish-
ery, which landed 6.8 thousand blue
sharks in 2000 (Cortes, 2002).
* Contribution 8 of the Massachusetts Divi-
sion of Marine Fisheries, P.O. Box 68, Vine-
yard Haven, MA 02568.
' Shivji, M. 2002. Personal commun.
Nova Southeastern University, 8000 North
Ocean Dr., Dania Beach, FL 33004.
628
Fishery Bulletin 101(3)
Ecologically, the blue shark is an apex predator of im-
portant teleosts and cephalopods (Stevens, 1973; Tricas,
1978; Kohler, 1987). Historical fisheries have shown that
sharks are intrinsically sensitive to sustained exploitation
(see review by Castro et al., 1999). Slow growth, late ages at
maturity, and low fecundities reflect the life history strate-
gies of /^-selected species; stock size is closely linked to re-
cruitment (Hoenig and Gruber, 1990). Although the current
Fishery Management Plan for Atlantic Tunas, Swordfish,
and Sharks has established limits on the U.S. commercial
and recreational fisheries that impact blue sharks (NMFS,
1999), no international management is currently in place.
Given a single North Atlantic stock for this species, any
fisheries exploitation, regardless of its coastal origin, may
impact the population. Accurate age determinations are
necessary for both the assessment and management of the
blue shark because they form the basis for calculations of
growth and mortality rates, age at maturity, age at recruit-
ment, and estimates of longevity.
Age and growth of the blue shark have been described by
a number of studies to varying degrees. In the North Pa-
cific, Cailliet et al. ( 1983 ) and Tanaka et al. ( 1990) used ver-
tebral growth rings and Nakano ( 1994) used both vertebrae
and length-frequency modes to establish growth curves for
the blue shark. In the North Atlantic, Aasen (1966) aged
the species by assigning ages to length-frequency modes.
Later, Stevens (1975), Silva et al. (1996), and Henderson et
al. (2001) established growth curves from vertebral growth
rings of juvenile blue sharks sampled in the eastern North
Atlantic. Low sample sizes, inadequate size ranges, and the
lack of age validation limit the utility of these studies for
the North Atlantic blue shark population. Skomal (1990)
generated growth curves for the blue shark from vertebral
growth rings, tag-recaptures, and length-frequency data.
In that study, vertebrae from oxytetracycline (OTC) in-
jected recaptured blue sharks were used to validate age
estimates. The purpose of the current study is to augment
the work of Skomal (1990) with additional tag-recapture
data, with corroborative vertebral readings of a different
vertebral processing technique, and with more rigorous
growth analyses.
Materials and methods
Interpretation of vertebrae
Vertebrae were obtained from blue sharks caught on
research cruises, commercial, and recreational fishing ves-
sels, and at sport fishing tournaments between 1966 and
2001. Primary sampling took place between Cape Hatteras,
NC, and the Gulf of Maine (NE coast of the United States).
To adequately represent the entire size range of the species,
small sharks were obtained from the eastern Atlantic from
cooperative fishermen and research scientists. When pos-
sible, the 15'^ through 20''' vertebrae were excised for the
study. When such precision was not possible, this section
of backbone was approximated by cutting at the branchial
region adjacent to the fifth gill arch. Excess muscle and
connective tissue were removed from the vertebrae with a
knife. Vertebrae were stored either frozen or preserved in
10% buffered formalin or 70% ethanol.
Only samples that had measured fork length (FL — tip
of the snout to the fork in the tail, over the body curva-
ture), total length (TL — tip of the snout to a point on the
horizontal axis intersecting a perpendicular line extending
downward from the tip of the upper caudal lobe to form a
right angle), or precaudal length (PCL — tip of the snout
to the precaudal pit, over the body curvature) were used
(Kohler et al., 1995). All lengths reported are in FL unless
otherwise noted. TL can be converted to FL by using the
regression (Kohler et al., 1995):
FL = 0.8313 (TL) + 1.39
[«=572 r-=0.991.
PCL can be converted to FL using the regression (NMFS'^)
PCL = 0.9075 (FL) - 0.3956 [n = 106 r2=0.99].
One vertebra from each sample was removed for pro-
cessing. The centrum was sectioned by using a Ray Tech
Gem Saw with two diamond blades separated by a 0.6-mm
spacer Each centrum was cut through the middle along the
sagittal plane; the resulting bow-tie sections were stored
in individual capsules in 70% ethanol. Each section was
digitally photographed with a MTI CCD 72 video camera
attached to a SZX9 Olympus stereo microscope by using
reflected light. All samples were photographed at a mag-
nification of 4x. Band pairs (consisting of one opaque and
one translucent band) were counted and measured from
the images by using Image Pro 4 software (Media Cyber-
netics, Silver Spring, MD). Measurements were made from
the midpoint of the notochordal remnant of the full bow-tie
to the opaque growth bands at points along the internal
corpus calcareum. The radius of each vertebral centrum
( VR) was measured from the midpoint of the notochordal
remnant to the distal margin of the intermedialia along
the same diagonal as the band measurements. Specimens
previously processed histologically (Skomal, 1990) were
used for counts when whole samples for those specimens
were not available for reprocessing. Because of the differ-
ent processing method, histological sections were not used
for measurements.
The criterion for what constitutes a band pair (annulus)
was based on the contouring of the corpus calcareum in
relation to the strength of the band. A clear indentation
of the corpus calcareum at the position of an opaque band
constituted the consummation of a growth layer within the
vertebra and was considered the annulus (Fig. 1). Each lay-
er was considered a temporal growth zone. The first opaque
band distal to the focus was defined as the birth mark (BR)
and a slight angle change in the corpus calcareum coincided
with this mark. In addition, identification of the birth band
was confirmed with back-calculation and by comparison of
the radius of this band with the radius of vertebrae from
young of the year (YOY) and full-term embryos.
•^ NMFS (National Marine Fisheries Service). 2001. Unpubl.
data. Apex Predators Program, 28 Tarzwell Dr, Narragansett,
RI 02882.
Skomal and Natanson: Age and growth of Pnonace glauca
629
The relationship between VR and FL was
calculated to determine the best method for
back-calculation of size-at-age data and to
confirm the interpretation of the birth band.
Age was calculated for each fish based on a
birth date of June 1 (Pratt, 1979), corrected
for date of capture. Regressions were fitted
to the male and female size-at-age data and
an ANCOVA was used to test for difference
between the two relationships. The relation-
ship between FL and VR was best described
by a polynomial equation; therefore the
data were In-transformed before linear
regression. The Fraser-Lee equation of the
In-transformed data was derived for back
calculation:
liUFLJ = b + (ln[FLJ+b) (InradiusJ (InradwsJ'K
where a = age;
b = intercept from the regression; and
c = capture.
Validation
To evaluate the periodicity of band pair
formation, vertebrae from OTC-injected and
measured tag-recaptured sharks were exam-
ined. Over 350 blue sharks of various sizes were measured,
tagged, and injected with a 25 mg/kg body weight dose of
OTC by scientific personnel aboard research and commer-
cial vessels in the North Atlantic. Upon recapture, vertebrae
were removed from injected specimens and stored in 70%
ethanol or were frozen. Vertebrae from these sharks were
processed, digitally photographed as previously described,
and examined for the OTC mark with reflected UV light.
The number of band pairs distal to the OTC mark was then
compared with the number of years at liberty and expressed
as the proportion of the previous complete growth zone.
Data analysis
Aging bias and precision of bands counts were examined by
using age-bias plots and the coefficient of variation (Cam-
pana et al., 1995). Reader 2 counted 98 sections previously
counted by reader 1 (Skomal, 1990). Pairwise comparisons
were generated from these data.
Von Bertalanffy growth functions (VBGF) were fitted to
length-at-age data by using the following equation (von
Bertalanffy, 1938):
Z., = Z..(l-e-''"-'"'),
where L, = predicted length at time t;
L^= mean asymptotic length;
K = a. growth rate parameter (yrM; and
Iq = the theoretical age at which the fish would
have been zero length.
The VBGF was calculated by using the nonlinear regression
function in Statgraphics (Manugistics, Inc., Rockville, MD).
Figure 1
Photograph of a vertebral section from a male blue shark (Prionace glauca)
estimated to be 14 years old.
Tagging data
From 1963 through 1999, members of the NMFS Coopera-
tive Shark Tagging Program tagged 88,899 and recaptured
4967 blue sharks. Only those sharks reliably measured by
biologists or fishermen at both tagging and recapture were
used in the analyses. All measurements were converted to
FL by using the relationships of Kohler et al. (1995).
The Gulland and Holt ( 1959) and Francis ( 1988a) models
were used to generate VBGFs from the tag-recapture data.
The Gulland and Holt ( 1959) method uses graphical inter-
pretation of the recapture data to produce estimates of L,
and K. Specifically, annualized growth rate (cm/yr ) was plot-
ted against average FL (cm) between tagging and recapture
to calculate linear regression coefficients. The slope of the
line is equal to -K and the x-axis intercept is equal to L^.
The Francis ( 1988a) method (GROTAG) uses maximum
likelihood techniques to estimate growth parameters and
variability from tagging data. A coefficient of variation of
growth variability (v), measurement errors (m and s) and
outlier contamination (p) are estimated, as well as growth
rates at two user selected lengths (« and p). The reference
lengths, a and p, were chosen to lie within the range of
tagged individuals. The form of the von Bertalanffy equa-
tion becomes
AZ.=
Su-gp
1-1 +
The simplest model, a linear fit with minimal parameters
(o and s) was used initially and additional parameters
630
Fishery Bulletin 101 (3)
were added to successively increase the model complexity.
Significant improvement in the model results were deter-
mined by using log likelihood ratio tests in accordance with
Francis {1988a). Bootstrapping was used to calculate the
95% confidence intervals for the final parameter estimates.
The modeling and bootstrapping were carried out by using
a Solver based spreadsheet in MS Excel (Microsoft Corp.,
Redmond, WA) (Simpfendorfer^). The value oft^ cannot be
estimated from tagging data alone, it requires an estimate
of absolute size at age, such as size at birth, and was calcu-
lated with the VBGF by solving for t^ such that
t„=l + {\/ K)[\n\L^-L,/ Lj]
where L, = known length at age (size at birth);
K = the von Bertalanffy growth constant; and
L^ = the theoretical maximum attainable length
from the VBGF.
The 99% ofL^
using the equation
> 99% = 5
(ln2)
Results
Interpretation of vertebrae
Vertebral samples from 411 blue sharks were used in our
study: 287 males, 119 females, and five of unknown sex.
These samples comprised free-living sharks ranging from
49 cm to 312 cm FL. In addition, vertebrae from seven late-
term embryos ranging from 36 cm to 43 cm FL were exam-
ined. Blue shark vertebrae did not have consistent prebirth
marks; thus, the first distinct opaque band was generally
considered the birth mark. The location of the birth band
coincided with a slight angle change (Fig. 1).
The FL-VR relationship was slightly curvilinear and the
In-transformed data provided a better linear fit (Fig. 2),
In(FL) = 0.89*ln(VR) + 3.10
[r!=392r2=0.97].
■* Simpfendorfcr, C. 2000. Personal commun. Mote Marine
Laboratory, 1600 City Island Park, Sarasota, FL 33577.
There was no significant difference between the sexes
(ANCOVA,P>0.10).
Confirmation of the birth band was made by comparison
of the BR of all individuals to the VR of YOY and late-term
embryos ( Fig 2 ). The VR of seven late-term embryos ( mean
VR ±95% CI=2.04 ±0.25) was slightly less than the BR
value of the total sample (mean BR ±95% CI=2.70 ±0.03;
n=351); the mean VR of 11 early YOY was slightly higher
than the BR of the entire sample (49-58 cm FL; mean VR
±95% CI =2.97 ±0.18) (Fig. 2). The location of the birth ring
between the VR of both the late-term embryos and the YOY
indicated that the birth ring was identified correctly.
Validation
OTC-injected recaptured blue sharks provided evidence
for the use of vertebral band pairs as age indicators.
Vertebrae from two OTC-injected sharks were returned
after 0.7 and 1.5 years at liberty (Table 1). OTC injection
produced strong fluorescent marks in the vertebral centra
of both these sharks (Fig. 3) and the number of annuli past
the OTC mark coincided with the number predicted from
time at liberty (Table 1 ). In OTC-injected recaptured shark
(B536), an opaque growth band was deposited just after
tagging in May (Fig. 3). In recaptured shark B 116452, an
opaque growth band was deposited just prior to tagging
in June (Fig. 3). These results suggest an annual spring
deposition of growth zones within the vertebrae. Thus, ver-
tebral annuli were validated in these two sharks, which
were up to 4* years of age; the older of these fish (B536)
corresponded to this age. Beyond this age, bands were
assumed to be annual on the basis of the similar nature
of band deposition.
Comparison of counts between two readers indicated no
appreciable bias (Fig. 4). The coefficient of variation fiuc-
tuated around 15%. This level of precision was considered
acceptable; thus, counts generated by both readers and
preparation methods were combined for the analyses. The
reader maintained quality control by periodically recount-
ing earlier samples and by cross-checking the readings.
Length-at-age data indicated that males and females
grow at roughly the same rate. The overlap in observed
size-at-age data, as well as the graphical representation
of the VBGF curves, indicated that there is little differ-
ence in growth for the sexes (Fig. 5). However, the LOW-
ESS (locally weighted regression smoothing) derived
curves as well as the VBGF parameters indicated that,
theoretically, females grow slower and to a larger overall
size than males (Table 2, Fig. 6). The LOWESS curves
clearly showed minor differences in growth beginning at
approximately seven years of age (Fig. 6), but this was
likely an artifact of low female sample size at older ages.
Subsequent analyses are presented for each sex and for
sexes combined for ease of comparison with previously
published studies.
Skomal and Natanson: Age and growth of Prionace glauca
631
300 1
Mean radius of birth mark (n = 351 )
°° o » o -
250 •
200 ■
d
__^
Q nffflMJflP^
1
E
nxfiiBBH^'
o
^^StP^ ^
s:
^X^ ">
P 150 ■
to
O
U-
■) o oO
O Male WNA(n = 273)
Full term
a Female WNA(n= 112)
100-
embryos
\
A MaleENA(n = 7)
\
■ Female ENA (n ^ 2)
50-
\
f
•
\
Average size at birtti
Smallest free-tiving (n = 11)
0
5 10
15 20
Vertebral radius (mm)
Figure 2
Relationship between vertebral radius and fork length in
the blue shark ^Prionace glauca).
Tag-recapture
recapture.
data for OTC
Table 1
-injected recaptured blue shark (Prionace glauca). TFL
= fork length
at tagging.
RFL = fork length at
Sample
number
Sex
TFL
(cm)
RFL
(cm)
Date
tagged
Date
recaptured
Years at
liberty
Growth
(cm)
No. of bands
after OTC mark
B 116452
B536
F
M
116
172
162'
192
18Jun 1987
9 May 1985
21 Dec 1988
16 Jan 1986
1.5
0.7
33
29
1.20
0.68
' Calculated from precaudel length.
Tagging data
A total of 43 blue sharks was recaptured with sufficient
information for tag-recapture analysis. Data from 18
sharks at liberty >0.9 years were used for Gulland and
Holts (1959) method and all the recaptured sharks were
used for the Francis ( 1988a) method (GROTAG).
The results of the likelihood ratio tests with GROTAG
(Francis, 1988a) showed that the more complex nonlinear
model with all six parameters was the best fit for these data
(Table 3, model 3). The mean annual growth rates are ggg =
44.2 cm/yr and gjg^ = 25.5 cm/yr, corresponding to growth
rates at a FL= 90 cm and 180 cm, respectively (Fig. 7). Von
Bertalanffy estimates from the Gulland and Holt (1959)
and GROTAG (Francis, 1988a) methods produced similar
von Bertalanffy curves (Table 4, Fig. 8A).
Longevity
The maximum age determined from vertebral band pair
counts was 16 and 15 years for males and females, respec-
tively. These ages are likely to be an underestimate of
longevity, given the history of fisheries exploitation of this
species. Using Taylor's (1958) method, we determined that
the age at which 95% and 99% of the L^ is reached was
16.5 and 26.1 years, respectively. Fabens (1965) method for
>99% longevity produced an estimate of 20.7 years.
632
Fishery Bulletin 101(3)
Figure 3
Vertebral sections from two OTC-injected blue sharks [Prionace glauca). Annuli and
birth marks are indicated.
Discussion
Several methods have been employed to vali-
date or verify (or both) age estimates derived
from vertebral banding patterns (Cailliet, 1990).
Although corroborative verification often comes
from the interpretation of length-frequency data,
laboratory and field growth studies, and centrum
edge analyses, direct age validation for sharks is
limited to the interpretation of vertebral banding
patterns in OTC-injected fish.
In his review of elasmobranch age and growth
studies, Cailliet (1990) found validated growth
curves for only six species, which included three
carcharhinids: the lemon (Negaprion brevirostris);
the sandbar (Carcharhinus plutnbeus); and the
Atlantic sharpnose [Rhizoprionodon terraenovae)
sharks. Although more than ten years have trans-
pired since this review, validated growth curves for
sharks are still lacking. In lamnids, direct valida-
tion of annual band deposition with the use of OTC
has been reported in a single species, the porbeagle
shark, Larnna nasus (Natanson et al., 2002). Al-
though age estimates from vertebral banding pat-
terns have been reported for several carcharhinids,
including the oceanic whitetip shark, Carcharhin us
longimanus (Seki et al., 1998; Lessa et al., 1999), the
dusky shark, C. obscurus (Natanson et al., 1995; Natanson
and Kohler, 1996; Simpfendorfer, 2000), the blacktip shark,
C. limbatus (Wintner and Cliff, 1995 ), and the bronze whaler,
C. brachyurus (Walter and Ebcrt, 1991 ), ago interpretations
were not validated and vertebral bands were assumed to be
annual. More recently, Simpfendorfer ot al. (2002) validated
the annual formation of vertebral banding patterns in C.
obscurus from Western Australian waters.
In the current study, we have validated annual band pair
deposition in Prionace glauca up to 4* years in age using
18 -|
CV = 0.15
3
4
1
1 %
16 -
N = 98 ,
3
\
14 •
7 1
Al
y
12 -
13
,1
\
y
f
10 -
8 •
26
15
7
y
V
■k
y\
r^
1
6 ■
k>\
4 -
^
2
y^
0 ■
^
Age (yr) of reader 1
Figure 4
Age bias graph for pair-wise comparison of 98 blue shark i.Prio-
nace glauca) vertebral counts from two independent readers. Each
error bar represents the dSVr confidence interval for the mean age
assigned by reader 2 to all fish assigned a given age by reader 1.
The one-to-one equivalence line is also presented.
vertebrae from two OTC-injected fish. These data indicate
that annulus formation occurs in the spring. This seasonal
formation is further supported by the marginal increment
analysis of Skomal ( 1990 ), which shows that one band pair is
formed annually. However, the low sample size and the lack
of OTC-injected recaptured fish over the entire size range of
the species do not allow for full age and growth validation.
Clearly, the study requires OTC-injected recaptured blue
sharks over a broader size range and greater time at lib-
erty— a requirement that is not atypical of age and growth
Skomal and Natanson: Age and growth of Prionace glauca
633
studies on large highly migratory elasmobranchs.
Wintner and Dudley (2000) used two OTC-injected
recaptured individuals to conclude that growth
band deposition is annual in the tiger shark {Ga-
leocerdo ciivier). Moreover, Natanson et al. (2002)
validated annuli in the porbeagle shark up to 11
years of age by using only two OTC-injected and six
YOY recaptured individuals, although the species
was aged to 25 years.
The processes that govern vertebral growth
have yet to be described in elasmobranchs. The
pattern varies from one ring per year in most
carcharhinids (Cailliet, 1990), and two rings per
year in some lamnids (Parker and Stott, 1965;
Pratt and Casey, 1983) to the complete absence
of periodicity (Natanson, 1984). Some research-
ers feel that temperature plays a major role in
this process (Stevens, 1975; Ferreira and Vooren,
1991). The blue shark, however, remains within a
discrete temperature range year-round (Stevens,
1975; Sciarrota and Nelson, 1977; Casey, 1982).
Moreover, acoustic tracking has shown that blue
sharks experience large changes in body tempera-
ture (up to 7°C) as they routinely pass through
the thermocline in their daily periodic dives from
the surface to depths of 200-600 m (Carey and
Scharold, 1990).
The ecology of this species may provide a more likely
explanation of annulus formation. Kohler (1987) found
a seasonal cycle for energy storage that correlated with
the migratory patterns of the blue shark. In general, blue
shark condition was found to be at an annual low in the
winter and spring. Blue sharks use energy stores during
this time for extensive north-south and trans-Atlantic
migrations (Casey, 1985; Kohler, 1987) and periodic deep
dives (Carey and Scharold, 1990). It is logical that growth
may be depressed during these months, thereby causing a
check or annulus in the vertebrae.
Tag-recapture data provide verification of the growth
curves derived from vertebral banding. Francis (1988b)
suggested that growth curves generated from age-length
and length-increment (tagging) data are not directly
comparable and that the comparison of growth rates at
length was more appropriate. Although VBGF parameters
derived from tagging data are noticeably higher, growth
rates were similar for both methods (Fig. 7). The higher
L^ and K can be attributed to the different derivation of
the VBGF parameters and the absence of older recaptured
sharks in the sample.
Pratt (1979) proposed that maturity in the male blue
shark occurs at 183 cm FL and this would coincide with an
age of 4-5 years based on the results of the present study.
Females enter a distinct subadult phase (Pratt, 1979) at
145 cm FL and 2* years of age. Full maturity in females is
attained at 185 cm FL (Pratt, 1979), which corresponds to
about 5 years of age.
Previous estimates of age and growth of the blue shark in
the Atlantic have been determined from vertebral banding
patterns, and verification has been made from the interpre-
tation of length-frequency and tagging data (Stevens, 1975;
350 -
300 -
•
aogft^-^O
'ork length (cm)
o o o
J^*^^
100 ■
Jr^ O Males
A^ * Females
M —0— Female VBGF |
50 i
f Male VBGF
0 2 4 6 8 10 12 14 16 18
Age (years)
Figure 5
Prionace glauca growth curves and size-at-age data based on verte-
bral band counts. Von Bertalanffy growth function curves have been
fitted to the data by sex.
Table 2
Von Bertalanffy growth function parameters and 95% \
confidence intervals calculatec
by using vertebral and tag-
recapture methods for the blue shark (Prionace glauca). |
n = number of sharks in sample.
Method
L,
K
to n
Vertebral Combined
286.8
0.17
-1.43 411
CI
±7.32
0.01
0.20
Male
282.3
0.18
-1.35 287
CI
±7.15
0.02
0.23
Female
310.8
0.13
-1.77 119
CI
±34.8
0.03
0.50
Tag-recapture
GROTAG Combined
302.4
0.23
-0.69 43
Gulland and Combined
331.7
0.19
-0.77 18
Holt (1959) CI
±80.0
0.12
Silva et al., 1996; Henderson et al., 2001) (Table 5, Fig. 8).
The eastern Atlantic vertebral sample of Stevens (1975)
comprised largely females (89%), ranging from 34 cm to
227 cm FL. The resulting growth curve, therefore, largely
reflects female growth (Fig. 8C). His use of whole silver-
stained centra coupled with the lack of maximum-size fish
allowed for the interpretation of only six annuli. From only
mean back-calculated lengths at ages two through five, Ste-
vens extrapolated growth of the species with a VBGF to an
age of 20 years. Similarly, Silva et al. (1996) and Henderson
et al. (2001) investigated age and growth in this species
with whole vertebrae from sharks sampled in the eastern
North Atlantic. In the former study, vertebral SEunples from
634
Fishery Bulletin 101(3)
350
A
300
250
200
150
r
°V
o o
^^-
100
r r
/>"
50
Lf
E
o
0
-, . . .
_l L
8 10 12 14 16
o
Li.
350
300
r B
o
250
r
o
i^-^-^^^
<>
200
r
oo o
o
150
100
L^
^
50
r '
0
~ 1 . . .
1 ... 1 ... 1
. . . 1
6 8 10 12 14
Age (years)
16
Figure 6
Prionace glauca growth curves and size-at-age data based on vertebral
band counts. LOWESS (locally weighted regression smoothing) curves
have been fitted to the data by sex: (A) males and (B) females.
308 juvenile blue sharks collected in the Azores were used
to model early growth in this species. Silva et al. (1996)
calculated an annual growth rate of 30 cm/yr for the first
five years of life and aged the samples to seven years.
More recently, Henderson et al. (2001) used 159 vertebrae
sampled from blue sharks taken from oceanic waters off
Ireland. Like the previous two studies, the size range of
samples was limited to juvenile fish less than 191 cm FL
and the estimated ages ranged from 1 to 6 years.
Stevens (1975), Silva et al. (1996), and Henderson et al.
(2001) modeled blue shark growth with the VBGF. These
curves are similar to each other (Silva et al., 1996, Hender-
son et al., 200 1 ), yet show slower growth than the current
study (Fig. 8) despite the fact that we used criteria similar
to those of Stevens ( 1975) for vertebral interpretation. This
result is not surprising in light of the fact that these three
studies share common methods and sample biases. All
three of the previous studies were performed on juvenile
sharks from the eastern North Atlantic, the vast majority
of which were between 100 and 184 cm FL. Because of the
lack of samples from very small fish, one study (Silva et al.,
1996) included vertebral readings from full-term embryos
in the growth curve. It is well documented that embryonic
growth is not comparable to postnatal gi'owth (Casey et
al., 1985; Pratt and Casey, 1990) and, therefore, embryos
should not be included in a postnatal growth curve. The
lack of large and small specimens in the calculations of
these growth curves is particularly problematic because
validation of the first growth increment is essential as it
forms the basis of further counts. Moreover, the smallest
and largest of the specimens are the most influential in the
estimation of growth (Campana, 2001).
All three of the previous studies used similar whole
centrum vertebral processing techniques and band count
criteria, which would lead to corroborating counts, yet not
necessarily to accurate counts (Campana, 2001). Wliole
vertebrae simply do not allow for high band resolution in
older slower growing fish. Therefore, counts from whole
Skomal and Natanson: Age and growth of Prionace glauca
635
Log-likelihood function values i
GROTAG (Francis 1988a). For a
by at least 1.92 (Francis 1988a).
Table 3
ind parameter estimates for three growth models fitted to Prionace
significant (P<0.05) improvement in fit, the introduction of one extra
* indicates fixed parameters. Model 3 shows 95% confidence intervals
glauca tagging data using
parameter must increase X
Parameter
Symbol (unit)
Model
1
2
3
Log likelihood
Mean growth rates
Growth variability
Measurement error
Outliers
g90 (cm/yr)
gl80 (cm/yr)
V
s (cm)
»! (cm)
P
-197.29
21.53
10.92
0*
1.06
0*
0.83
-176.91
39.04
21.90
0.46
1.37
0*
0.28
-174.61
44.18 (35.37-54.33)
25.46(19.29-33.41)
0.27 (0.06-0.44)
5.39(2.25-7.40)
-2.03 (-5.37-2.10)
0.18
vertebrae generally underestimate ages in larger indi-
viduals. The counts obtained in the three eastern Atlantic
studies may be accurate because they are from juvenile
sharks where vertebral bands are not compressed. In fact,
juvenile growth from our size-at-age data overlaps the
growth cui-ves from these studies. However, the VBGF
growth curves and resulting estimates of growth rate
and age at maturity from the eastern Atlantic studies are
suspect because of the lack of fish at the lower and upper
end of the curve. The general lack of maximum-size fish in
these studies resulted in the estimation of an artificially
inflated L^ and, therefore, a lower growth rate iK) for this
species (Table 5). Vertebral band deposition was assumed
to be annual in these studies, but low sample sizes, sample
bias, and lack of validation limits the utility of this previ-
ous work. In the current study, the use of sections and the
adequate representation of the entire size range for both
sexes yielded more accurate age estimates of 16 and 15
years for males and females, respectively.
Age and growth estimates of the blue shark in the North
Pacific have been determined by using vertebral bands and
length-frequency data (Cailliet et al., 1983; Tanaka et al.,
1990; Nakano, 1994 ). Although the VBGF was used to mod-
el growth based on vertebral interpretation, the resulting
parameters differed greatly among studies (Table 5). In
general, Cailliet et al. (1983) reported a male growth rate
similar to that in our present study, but a much smaller
L^ (Table 5). For females, the latter holds true, but the
growth coefficient is much higher (0.25) than reported in
our study. Tanaka et al. ( 1990 ) found a similar growth trend
in the western North Pacific with females growing faster
than males, but the VBGF parameters were very differ-
ent with higher L^ and lower K values. When compared
to our study, the VBGF parameters of Tanaka et al. ( 1990)
yield slower growth and a greater maximum size for males
and a similar growth rate and smaller maximum size for
females. Tanaka et al. ( 1990 ) attributed these intra- and in-
ter-oceanic differences to the different methods used. More
recently, Nakano (1994) sampled blue sharks across the
North Pacific and derived VBGF growth parameters that
Table 4
Size at age (cm) for the blue shark (Prionace glauca) calcu-
lated from von Bertalanffy equations based on tag-recap-
ture and vertebral methods.
Age (yr)
Vertebral method
Tag-recapture
GROTAG
method
combined
male
female
0
61.0
60.9
66.1
45
1
95.8
97.4
97.0
99
2
125.2
127.8
124.0
141
3
150.1
153.3
147.6
175
4
171.2
174.5
168.2
201
5
189.0
192.3
186.2
222
6
204.1
207.1
201.9
239
7
216.8
219.5
215.7
252
8
227.6
229.8
227.7
263
9
236.7
238.5
238.2
10
244.4
245.7
247.4
11
251.0
251.7
255.4
12
256.5
256.8
262.4
13
261.1
261.0
268.5
14
265.1
264.5
273.9
15
268.4
267.4
278.5
16
271.3
269.9
were similar to those of Tanaka et al. ( 1990), but estimated
growth rate to be slower than that of our present study. It
is difficult to ascertain whether interoceanic differences
in growth are real or are an artifact of method. Although
Tanaka et al. (1990) presented data to support the latter
within the North Pacific, the much larger maximum size
attained by this species in the North Atlantic (Strasburg,
1958; Tanaka, 1984) cannot be overlooked in relation to
interoceanic growth differences.
636
Fishery Bulletin 101(3)
Longevity estimates for the blue shark indicate that
they may Hve for 26 years when Taylor's ( 1958) method is
employed. On the other hand, Fabens" (1965) method for
>99% longevity produced an estimate of 20.7 years, which
may be more realistic. The maximum age determined from
vertebral band-pair counts was 16 and 15 years for males
and females, respectively. An analysis of maximum times at
liberty for tagged blue sharks supports the notion that this
species does not live as long as previously reported for the
North Atlantic. Of the 4967 blue sharks recaptured to date.
70 T
♦ Growth/yr T/R
Vertebral
60 -
♦
— ♦- Tag/recapture (Francis 1988)
♦
♦
50
^ ^
♦ ♦ ♦ ♦
■H- 40
*♦♦
.c
I
♦
2 30-
♦ <♦
O)
M
^"^^ *
i 20-
♦ ^^-...^^^
c
<
10 ■
♦ ^^--^ *
♦ ^""~<^^^
0 ■
50 100 150 200 250 300
Fork length (cm)
Figure 7
Comparison of the annual growth rates of the blue shark (Prionace
glauca) derived from multiple aging methods.
99% were at liberty for less than five years. The maximum
times at liberty are 9.1 and 8.5 years, despite the 39-year
history of the tagging program. The shark at liberty for 9.1
years was a male tagged at an estimated 122 cm FL; size at
recapture was not reported. According to our growth curve,
the shark was tagged at 1* years of age, which would cor-
respond to a maximum age of 10+ years at recapture. The
shark at liberty for 8.5 years, also a male, was estimated
to be 198 cm FL at tagging, which would correspond to 5*
years of age. Therefore, at recapture, this fish would be a
maximum age of 13.5 years, although its measured
FL at recapture actually corresponds to 11 years
on our growth curve. The largest long-term recap-
ture was a male, 244 cm FL at tagging and 266
cm FL at recapture 6 years later. This would cor-
respond to an estimated age of 10 years at tagging
and 16 years at recapture, which falls well within
the values of directly aged vertebrae (Fig. 5).
The occurrence of sexual differences in growth
is well documented in elasmobranchs; females
usually grow larger than males (Cortes, 2000).
Although the largest blue shark in our study was
a 312-cm-FL female, there is little evidence that
large females are highly abundant in the North At-
lantic. Maximum size male and female specimens
in our study, 284 cm FL and 312 cm FL, respective-
ly, represented the largest reliably measured blue
sharks from the North Atlantic, with the exception
of a 320-cm-FL specimen (sex unspecified) exam-
ined by Bigelow and Schroeder (1953). Indeed,
a thorough review of the literature reveals that
although 288-cm-FL and 279-cm-FL females were
reported by Gubanov and Grigoryev (1975) from
the Indian Ocean, males are consistently cited as
being very much larger than females in the world's
Table 5
Van Bertalanffy growth function parameters
and maximum age derived from vertebral bands in the blue shark (Prionace glauca) \
separated by location and sex.
Sex
Ocean
Region
n
Z.,
A'
T,
Max. age
Authors
Male
North Atlantic
All
287
282.3
0.18
-1.35
16
Current studv
East
112
309.0
0.12
-1.07
5
Silvaetal. (1996)
North Pacific
East
38
246.7
0.18
-1.11
9
Caillietetal. (1983)
West
43
308.1
0.10
-1.38
7
Tanakaet al.(1990)
All
148
319.5
0.13
-0.76
10
Nakano(1994)
Female
North Atlantic
All
119
286.8
0.16
-1.56
15
Current study
East
82
353.0
0.11
-1.04
6
Stevens (1975)
East
170
382.0
0.09
-1.19
5
Silvaetal. (1996)
North Pacific
East
88
202.6
0.25
-0.80
9
Caillietetal. (1983)
West
152
254.1
0.16
-1.01
8
Tanakaetal. (1990)
All
123
268.9
0.14
-0.85
10
Nakano(1994)
Combined
North Atlantic
All
411
285.4
0.17
-1.41
16
Current study
East
336
284,0
0.14
-1.08
5
Silvaetal. (1996)
East
159
314.4
0.12
-1.33
6
Henderson etal. (2001)
North Pacific
East
130
222.1
0.22
-0.80
9
Caillietetal. (1983)
Skomal and Natanson; Age and growth of Prionace glauca
637
oceans (Suda, 1953; Tucker and Newnham, 1957;
Aasen, 1966; McKenzie andTibbo, 1964; Dragonik
and Pelzarski. 1983; Stevens, 1984; Francis et al,,
2001 ). Although the largest blue shark reported
from the North Pacific was only 254 cm FL (Stras-
burg, 1958; Cailliet et al., 1983), individuals up to
331 cm FL have been reported from the South Pa-
cific and the largest sharks were all males (Francis
et al., 2001). The paucity of females exceeding 225
cm FL in the current study and the complete lack
of these specimens in the Stevens (1975), Silva et
al. (1996), and Henderson et al. (2001) samples
indicate that these fish are rare, inhabit unknown
or unfished areas of the Atlantic, or possibly avoid
fishing gear In our study, the VBGF parameters
(Table 5) show that females theoretically attain
larger sizes than males. However, the low number
of large females in this and previous studies may
indicate that natural mortality prevents them from
attaining these lengths. The occurrence of severe
lacerations on female blue sharks incurred during
courtship is well documented (Stevens, 1974; Pratt,
1979). Although highly speculative, the long-term
cumulative effects of such behavior may act as
a source of increased mortality in females of the
species, shortening their life-span and limiting the
number that reach the larger sizes.
Through an integrated approach incorporating
vertebral banding, OTC injection, and tagging
data, it has been shown that the blue shark grows
faster and lives a shorter life than previously
thought in the North Atlantic. We believe that
the validated vertebral interpretations generated
during this study for the first four years of growth,
combined with the vertebral counts and longevity
estimates from tag-recapture data, provide vigor-
ous estimates of age and growth for a large pelagic
carcharhinid, the blue shark.
Acknowledgments
We thank the many people who contributed to
the success of this study on research vessels, at
recreational fishing tournaments, and on board
private, chartered, and commercial fishing vessels.
This study would not have been possible without
the staff of the NMFS Apex Predators Program
(Narragansett, RI) including Nancy Kohler, Pat
Turner, and Ruth Briggs. We especially thank
retired NMFS researchers Jack Casey and H.
Wes Pratt for giving the senior author the tools
necessary to initiate and complete this work. We
are grateful to shark-aging pioneer Gregor Cailliet
for his moral support and relentless pursuit of this
publication. We are indebted to the thousands of
fishermen who voluntarily tagged and returned
sharks for the NMFS Cooperative Shark Tagging
Program. This study was partially funded with
support from the Sportfish Restoration Act.
300
A Tag/recapture-GROTAG
Tag/recapture-Gultand method
Vertebrae-current study
Aasen (1966)-LF modes
' Henderson el al (2001)
Silva etal. (1996)
0 2 4 6 8 10 12 14 16 18
B
Vertebrae-current study
/ Silva etal (1996)
12 14 16 18
250
c
^__^
^.^
Vertebrae-current study
OTC y^
y^ ■/■ Stevens (1975)
/^
/ y Silva etal (1996)
//
/
0 2 4 6 8 10 12 14 16
Age (years)
Figure 8
Von Bertalanfly growth curves generated from vertebral and recap-
ture data for (A) sexes combined, (B) male, and (C) female Prionace
glauca, as compared to OTC-injected recaptured blue sharks; included
for comparison are the von Bertalanffy growth curves of other North
Atlantic studies.
638
Fishery Bulletin 101 (3)
Literature cited
Aasen, O.
1966. Blahaien, Prionace^Zauca (Linnaeus, 1758). Fisken
ogHavet 1:1-15.
Bigelow, H.. B., and W. C. Schroeder.
1953. Fishes oftheGulfofMaine, 577 p. U.S. Dep. Int., Fish
and Wildl. Serv., Fish. Bull. 53.
Cailliet, G. M.
1990. Elasmobranch age determination and verification: an
updated review. In Elasmobranchs as living resources:
advances in the biology, ecology, systematics, and status of
the fisheries (H. L. Pratt Jr., S. H. Gruber, and T. Taniuchi,
eds.), p. 157-165. U.S. Dep. Commer., NOAA Tech. Rep.
90.
Cailliet, G. M., L. K. Martin, J. T. Harvey, D. Kusher, and
B. A. Welden.
1983. Preliminary studies on the age and growth of blue,
Prionace glauca, common thresher, Alopias vulpinus, and
shortfin mako, Isurus oxyrinchus, sharks from California
waters. In Proceedings of the international workshop on
age determination of oceanic pelagic fishes: tunas, billfishes.
and sharks (E. D. Prince and L. M. Pulos, eds.), p. 179-188.
U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8.
Campana, S. E.
2001. Accuracy, precision and quality control in age deter-
mination, including a review of the use and abuse of age
validation methods. J. Fish. Biol. 59:197-242.
Campana, S. E., M. C. Annand, and J. I. McMillan.
1995. Graphical and statistical methods for determining the
consistency of age determinations. Trans. Amer. Fish. See.
124:131-138.
Carey, F. G., and J. Scharold.
1990. Movements of blue sharks (Prionace glauca) in depth
and course. Mar Biol. 106:329-342.
Casey, J. G.
1982. Blue shark, Prionace glauca. Species synopsis. In
Ecology of the Middle Atlantic Bight fish and shellfish —
Monograph 15 — fish distribution (M.D. Grosslein and T.
Azarovitz, eds.), p. 45-^8. MESA New York Bight Atlas,
NY Sea Grant, Albany, NY.
1985. Trans-Atlantic migrations of the blue shark: a case
history of cooperative shark tagging. In World angling
resources and challenges; proceedings of the first world
angling conference (R. H. Stroud, ed.), p. 253-267. Int.
Game Fish Assoc, Ft. Lauderdale, FL.
Casey, J. G., H. L. Pratt Jn, and C. E. Stillwell.
1985. Age and growth of the sandbar shark (Carcharhinus
plumbeus) from the western North Atlantic. Can. J. Fish.
Aquat. Sci. 42(5):963-975.
Castro, J. L, C. M. Woodley, and R. L. Brudek.
1999. A preliminary evaluation of the status of shark
species. FAG Fisheries Technical Paper 380, 72 p. FAG,
Rome.
Compagno, L. J. V.
1984. FAG species catalogue. Sharks of the world. An anno-
tated and illustrated catalogue of shark species known to
date. Part 1. Hexanchiformes to Lamniformes. FAO Fish
Synop. 125, vol. 4, 250 p. FAO, Rome.
Connett, S.
1987. Blue sharks studied in the eastern Atlantic. In The
shark tagger: 1987 summary (J. Casey, H. L. Pratt Jr, N. E.
Kohler, and C. E. Stillwell, eds. ), p. 8-10. Newsletter of the
Cooperative Shark Tagging Program. U.S. Dep. Commer.,
NOAA, NMFS, Narragansett, RL
Cortes, E.
2000. Life history patterns and correlations in sharks. Rev.
Fish. Sci. 8(4):299-344.
2002. Catches and catch rates of pelagic sharks from the
northwestern Atlantic, Gulf of Mexico, and Caribbean. Col.
Vol. Sci. Pap. ICCAT 54(4): 1164-1 181.
Dragonik, B, and W. Pelzarski.
1983. The occurrence of the blue shark, Prionace glauca (L.),
in the North Atlantic. Rep. Sea Fish. Inst. 19:63-77.
Fabens. A. J.
1965. Properties and fitting of the von BertalanfTy growth
curve. Growth 29:265-289.
Ferreira, B. P., and C. M. Vooren.
1988a. Age, growth, and structure of the vertebrae in the
school shark Galeorhinus galeus (Linnaeus, 1758) from
southern Brazil. Fish. Bull. 89:19-31.
Francis, R. I. C. C.
1988a. Maximum likelihood estimation of growth and
growth variability from tagging data. NZ J. Mar Fresh-
water Res. 22:43-51.
1988b. Are growth parameters estimated from tagging and
age-length data comparable? Can. J. Fish. Aquat. Sci. 45:
936-942.
Francis, M. P., L. H. Griggs, and S. J. Baird.
2001. Pelagic shark by catch in the New Zealand tvma long-
line fishery. Mar. Freshwater Res. 52(2):165-178.
Gubanov, Ye. P., and V. N. Grigor'yev.
1975. Observations on the distribution and biology of the
blue shark Prionace glauca (Carcharhinidae) of the Indian
Ocean. J. Ichthyol. 15:37-43.
Gulland, J. A., and S. J. Holt.
1959. Estimation of growth parameters for data at unequal
time intervals. J. Cons. Int. Explor Mer 25:47-49.
Henderson, A. C, K. Flannery, and J. Dunne.
2001. Observations on the biology and ecology of the
blue shark in the North-east Atlantic. J. Fish. Biol. 58:
1347-1358.
Hoenig, J. M., and S. H. Gruber.
1990. Life history patterns in the elasmobranchs: implica-
tions for fisheries management. In Elasmobranchs as
living resources: advances in the biology, ecology, systemat-
ics, and status of the fisheries (H. L. Pratt Jr, S. H. Gruber,
and T Taniuchi, eds), p. 1-16. U.S. Dep. Commer, NOAA
Tech. Rep. 90.
ICCAT (International Commission for the Conservation of
Atlantic Tunas).
2002. ICCAT data preparatory meeting for Atlantic shark
stock assessment. Col. Vol. Sci. Pap. ICCAT, 54(4):
1064-1106.
Kohler, N. E.
1987. Aspects of the feeding ecology of the blue shark in the
western North Atlantic. Ph.D. diss., 163 p. Univ Rhode
Island, Kingston, RI.
Kohler, N. E., J. G. Casey, and P A. Turner
1987. Length-weight relationships for 13 species of sharks
from the western North Atlantic. Fish. Bull. 93:412-418.
Kohler, N. E., P. A. Turner, J. J. Hoey, L. J. Natanson, and
R. Briggs.
2002. Tag and recapture data for three pelagic shark species:
blue shark (Prionace glauca ), shortfin mako (Isurus oxyrin-
chus), and porbeagle (Lamna nasus) in the North Atlantic
Ocean. Col. Vol. Sci. Pap. ICCAT 54(4):1231-1260.
Lessa, R., F M. Santana, and R. Paglerani.
1999. Age, growth and stock structure of the oceanic whitetip
shark, Carcharhinus longnnanus. from the southwestern
equatorial Atlantic. Fish. Res. 42:21-30.
Skomal and Natanson: Age and growth of Phonace glauca
639
McKenzie, R. A., and S. N. Tibbo.
1964. A morphometric description of the blue shark i.Prio-
nace glauca ) from the Canadian Atlantic waters. J. Fish.
Res. Board Can. 21:865-866.
Nakano, H.
1994. Age. reproduction and migration of blue shark in the
North Pacific. Bull. Nat. Res. Inst. Far Seas Fish. 31:141-256.
Natanson, L. J.
1984. Aspects of age, growth, and reproduction of the Pacific
angel shark, Squatina californica, off Santa Barbara,
California. M.A. thesis, 71 p. San Jose State Univ., San
Jose, CA.
Natanson, L. J., J. G. Casey, and N. E. Kohler.
1995. Age and growth of the dusky shark, Carcharhinus
obscunis, in the western North Atlantic. Fish. Bull. 93:
116-126.
Natanson, L. J., and N. E. Kohler
1996. A preliminary estimate of age and growth of the dusky
shark Carcharhinus obscurus from the south-west Indian
Ocean, with comparisons to the western North Atlantic. S.
Afr. J. Mar. Sci. 17:217-224.
Natanson, L. J., J. J. Mello, and S. E. Campana.
2002. Validated age and growth of the porbeagle shark,
Lamna nasus, in the western North Atlantic. Fish. Bull.
100:266-278.
NMFS (National Marine Fisheries Service).
1999. Final fishery management plan for Atlantic tunas,
swordfish, and sharks., 1162 p. Highly Migratory Species
Management Division, Silver Spring, MD.
Parker, H. W., and F C. Stott.
1965. Age, size and vertebral calcification in the basking
shark, Cctorhinus maximus (Gunnerus). Zool. Meded.
40(34);305-319.
Pratt, H. L., Jr.
1979. Reproduction in the blue shark, Prionace glauca.
Fish. Bull. 77:445-470.
Pratt, H. L., Jr., and J. G. Casey
1983. Age and growth of the shortfin mako, Isurus oxyrin-
chus. using four methods. Can. J. Fish. Aquat. Sci. 40(11):
1944-1957.
1990. Shark reproductive strategies as limiting factors in
directed fisheries, with a review of Holden's method of esti-
mating growth parameters. In Elasmobranchs as living
resources: advances in the biology, ecology, systematics, and
status of the fisheries (H. L. Pratt Jr., S. H. Gruber, and T.
Taniuchi, eds), p. 97-110. NOAATech. Rep. 90.
Sciarrotta, T. C, and D. Nelson.
1977. Diel behavior of the blue shark, Prionace glauca, near
Santa Catalina Island, California. Fish. Bull 73:519-528.
Seki, T., T. Taniuchi, H. Nakano, and M. Shimizu.
1998. Age, growth and reproduction of the oceanic whitetip
shark from the Pacific Ocean. Fish. Sci. 64(1): 14-20.
Silva, A. A., H. M. Silva, and K. Erzini.
1996. Some results on the biology of the blue shark, Prio-
nace glauca. in the North Atlantic based on data from a re-
search cruise of the R/V Arquipelago in Azorean waters:
a summary paper, 9 p. Universidade dos Acores, Horta,
Acores, Portugal.
Simpfendorfer, C. A.
2000. Growth rates of juvenile dusky sharks, Carcharhinus
obscurus (Lesueur, 1818), from southwestern Australia esti-
mated from tag-recapture data. Fish. Bull. 98:811-822.
Simpfendorfer, C. A., R. B. McAuley, J. Chidlow, and P. Unsworth.
2002. Validated age and growth of the dusky shark, Carcha-
rhinus obscurus, from Western Australian waters. Mar.
Freshwater Res. 53:567-573.
Skomal, G. B.
1990. Age and growth of the blue shark, Prionace glauca, in
the North Atlantic. M.S. thesis, 82 p. Univ. Rhode Island,
Kingston, RI.
Stevens, J. D.
1973. Stomach contents of the blue shark (Prionace glauca
L.) of southwest England. J. Mar Biol. Assoc. U.K. 53:
357-361.
1974. The occurrence and significance of tooth cuts on the
blue shark (Prionace glauca L. ). J. Mar Biol. Assoc. U.K.
54:373-378.
1975. Vertebral rings as a means of age determination in
the blue shark (Prionace glauca L.). J. Mar. Biol. Assoc.
U.K. 55:657-665.
1976. First results of shark tagging in the northeast Atlan-
tic, 1972-1975. J. Mar. Biol. Assoc. U.K. 56: 929-937.
1984. Biological observations on sharks caught by sport-
fishermen off New South Wales. Aust. J. Mar. Freshwater
Res. 35:573-590.
1990. Further results from a tagging study of pelagic sharks
in the Northeast Atlantic. J. Mar Biol. Assoc. U.K. 70:
707-720.
Strasburg, D. W.
1958. Distribution, abundance, and habits of pelagic sharks
in the central Pacific Ocean. Fish. Bull. 58:335-361.
Suda, A.
1953. Ecological study of the blue shark (Prionace glauca
Linne'). South Sea Area Fish Res. Lab. Rep. 26:1-11.
Tanaka, S.
1984. Present status of fisheries biology, /n Elasmobranchs
as fishery resources (T. Taniuchi and M. Suyama, eds.), p.
46-59. Jpn. Soc. Sci. Fish., Fish. Ser. 49.
Tanaka, S., G. M. Cailliet, and K. G. Yudin.
1990. Differences in growth of the blue shark, Prionace
glauca: technique or population? In Elasmobranchs as
living resources: advances in the biology, ecology, systemat-
ics, and status of the fisheries (H. L. Pratt Jr., S. H. Gruber,
and T. Taniuchi, eds.), p. 177-187. U.S. Dep. Commer.,
NOAA Tech. Rep. 90.
Taylor, C. C.
1958. Cod growth and temperature. J. Cons. Int. Explor
Mer 23:366-370.
Tricas, T.
1978. Relationships of the blue shark, Prionace glauca, and
its prey species near Santa Catalina Island, California.
Fish. Bull. 77:175-182.
Tucker, D. W., and C. T. Newnham.
1957. The blue shark Prionace glauca breeds in British seas.
Ann. Mag. Nat. Hist., Series 12, 10:673-688.
von Bertalanffy, L.
1938. A quantitative theory of organic growth (inquiries on
growth laws II). Hum. Biol. 10:181-213.
Walter, J. P, and D. A. Ebert.
1991. Preliminary estimates of age of the bronze whaler
Carcharhinus brachyurus (Chondrichthyes: Carcharhini-
dae) from southern Africa, with a review of some life history
parameters. S. Afr. J. Mar. Sci 10:37-44.
Wintner,S. P, and G. Cliff.
1995. Age and growth determination of the blacktip shark,
Carcharhinus lintbatus, from the east coast of South Africa.
Fish. Bull. 94:135-144.
Wintner, S. P, and S. F J. Dudley
2000. Age and growth estimates for the tiger shark, Galeo-
cerdo cuvier, from the east coast of Africa. Mar Freshwa-
ter Res. 51:43-53.
640
Abstract — Little is known about the
ocean distributions of wild juvenile
coho salmon off the Oregon- Washington
coast. In this study we report tag recov-
eries and genetic mixed-stock estimates
of juvenile fish caught in coastal waters
near the Columbia River plume. To sup-
port the genetic estimates, we report
an allozyme-frequency baseline for 89
wild and hatchery-reared coho salmon
spawning populations, extending from
northern California to southern Brit-
ish Columbia. The products of 59 allo-
zyme-encoding loci were examined with
starch-gel electrophoresis. Of these, 56
loci were polymorphic, and 29 loci had
^0 9.5 levels of polymorphism. Average
heterozygosities within populations
ranged from 0.021 to 0.046 and aver-
aged 0.033. Multidimensional scaling of
chord genetic distances between sam-
ples resolved nine regional groups that
were sufficiently distinct for genetic
mixed-stock analysis. About 2.9% of the
total gene diversity was due to differ-
ences among populations within these
regions, and 2.6% was due to differences
among the nine regions. This allele-fre-
quency data base was used to estimate
the stock proportions of 730 juvenile
coho salmon in offshore samples col-
lected from central Oregon to northern
Washington in June and September-
October 1998-2000. Genetic mixed-
stock analysis, together with recoveries
of tagged or fin-clipped fish, indicates
that about one half of the juveniles
came from Columbia River hatcheries.
Only 22% of the ocean-caught juveniles
were wild fish, originating largely from
coastal Oregon and Washington rivers
(about 20%). Unlike previous studies
of tagged juveniles, both tag recoveries
and genetic estimates indicate the pres-
ence offish from British Columbia and
Puget Sound in southern waters. The
most salient feature of genetic mixed
stock estimates was the paucity of wild
juveniles from natural populations in
the Columbia River Basin. This result
reflects the large decrease in the abun-
dances of these populations in the last
few decades.
Genetic analysis of juvenile coho salmon
iOncorhynchus kisutch) off Oregon and
Washington reveals few Columbia River wild fish'
David J. Teel
Donald tA. Van Doornik
David R. Kuligowsl
a.
88 1 89
La Push
v/ ®^ . 83
84 82
63 ^^"^■'
+++++ 59-62 64 '^^-'^*
57-58 65 70-72
-46°
Location of
coastal pelagic
trawl surveys
66-68 69
+++++:■- . *-5 C
+++++ .53-56
+++++ 50-52 Washington
^''-^^ 38-39
+++++ 32-33
, , , , ^, 40-42 Columbia
**^^ 31 43-49
++++ 30
++++ 26-29
Cape Perpetua
23-25
^^^ 21-22
-42°
17-20
11-16 Oregon
\ V
• Rogue
5 —
£■ River
. ■/ ....... ^
6 ' Xallfornia
-38°
N
i
1
km
0 100 200
Figure 1
Locations of ocean sampling transect lines ( + ) and 89 coho salmon
populations in California, Oregon, Washington, and British Colum-
bia. Numbers correspond to population names in Table 1.
Genotypic frequencies of polymorphic loci for each base-
line sample were examined for departures from expected
Hardy-Weinberg proportions with a Fisher's exact test
(Guo and Thompson, 1992) by using GENEPOP version
3.1 (Raymond and Rousset, 1995). Hardy-Weinberg tests
were performed on isoloci (comigrating protein products of
duplicated loci) following Waples (1988).
We estimated allelic frequencies for each sample. Allelic
frequencies for isoloci were calculated as mean frequen-
642
Fishery Bulletin 101(3)
Table 1
Sample information and indices of genetic variability for coho salmon from the Pacific Northwest and Cahfomia. Map codes refer |
to Figure 1. Indices of genetic variability are '^tP^g^ = percentage of Pog^
loci and H = heterozygosity.
Source
Number of
Region and map code
Year sampled
fish
'-'''^95
H
California coast
1
Scott Creek
1994
21
12.5
0.039
2
Little River
1994
27
14.3
0.040
3
Warm Springs Hatchery
1994, 1994
160
16.1
0.041
4
Mad River Hatchery
1994
120
17.9
0.040
Klamath River to Cape Blanco
5
Iron Gate Hatchery
1994
120
9.0
0.021
6
Trinity Hatchery
1984, 1994
218
9.0
0.028
7
Rogue River (Illinois River, Greyback Creek)
1993
40
7.2
0.022
8
Cole Rivers Hatchery, stock no. 52 (Rogue River)
1993
100
9.0
0.030
9
North Fork Elk and Elk Rivers
1993
32
7.2
0.021
Oregon coast
10
Sixes River (Crystal and Edson Creeks)
1993
44
7.2
0.026
11
New River (Bether and Morton Creeks)
1993
62
10.7
0.034
12
Butte Falls Hatchery, stock no. 44 (Coquille River)
1993
100
9.0
0.036
13
Cole Rivers Hatchery, stock no. 37 (South Fork Coos River)
1993
129
10.7
0.034
14
Coos River (MilHcoma River and Marlow Creek)
1993, 1997
50
12.5
0.033
15
Butte Falls Hatchery, Eel River stock no. 63
1993
100
7.2
0.032
16
Ten Mile Lake
1992
56
7.2
0.030
17
Rock Creek Hatchery, stock no. 55 (Umpqua River)
1993
100
7.2
0.029
18
North Umpqua River (Williams Creek)
1993, 1997
67
7.2
0.025
19
Butte Falls Hatchery, stock no. 18 ( Umpqua River)
1993
100
7.2
0.027
20
Smith River (Halfway Creek)
1993
40
10.7
0.034
21
Fall Creek Hatchery, stock no. 113 (Talikenitch River)
1993
100
7.2
0.030
22
Siuslaw River
1996
51
9.0
0.029
23
Fall Creek Hatchery, stock no. 31 (Alsea River)
1993
100
14.3
0.040
24
Fall Creek Hatchery, stock no. 43 (Alsea River)
1993
95
9.0
0.037
25
Alsea River
1996
62
10.7
0.031
26
Beaver Creek
1993
62
9.0
0.035
27
Yaquina River
1996
54
12.5
0.043
28
Salmon River Hatchery, stock no. 33 (Siletz River)
1993
100
12.5
0.041
29
Siletz River (Forth of July, Sunshine, and Buck Creeks)
1993
50
10.7
0.033
30
Salmon River Hatchery, stock no. 36 (Salmon River)
1993
100
10.7
0.037
31
Trask River Hatchery, stock no. 34 (Trask River)
1992, 1993
220
16.1
0.039
32
Nehalem River Hatchery, stock no. 99 (Nehalem River)
1992
80
12.5
0.045
33
Nehalem River Hatchery, stock no. 32 (Nehalem River)
1993
100
14.3
0.044
Columbia River
34
Lewis and Clark River
1991, 1993
36
12.5
0.038
35
Big Creek Hatchery
1991
80
12.5
0.040
36
Grays River Hatchery
1987, 1991
200
7.2
0.033
37
Clatskanie River (Carcus Creek)
1991, 1992, 1996
113
10.7
0.033
38
Cowlitz Hatchery early-run
1991
80
9.0
0.027
39
Cowlitz Hatchery late-run
1991, 1992
180
7.2
0.031
40
Scappoose River (Siercks, Raymond, and Milton Creeks)
1991
44
14.3
0.041
41
Lewis River Hatchery early-run
1991
80
5.4
0.027
42
Lewis River Hatchery late-run
1991
80
12.5
0.032
43
North Fork Clackamas River early-run
1998«
48
16.1
0.036
44
North Fork Clackamas River late-run
1999»
45
14.3
0.028
45
Eagle Creek Hatchery
1991, 1992
180
7.2
0.037
continued
Teel et al.: Genetic analysis of juvenile Oncorhynchus kisutch
643
Table 1 (continued)
Source
Number of
Region and map code
Year samplet
fish
■^"^0.95
H
46
Sandy River Hatchery
1991, 1992
180
10.7
0.046
47
Sandy River
1991, 1992, 1996
124
10.7
0.043
48
Bonneville Hatchery
1991, 1992
180
10.7
0.043
49
Willard Hatchery
1991
80
7.2
0.032
South Washington coast
50
Naselle River Hatchery
1991
100
9.0
0.029
51
Nemah River Hatchery
1991
100
10.7
0.029
52
Willapa River Hatchery
1991
100
9.0
0.031
53
Chehalis River (Stillman Creek)
1995
71
9.0
0.026
54
Chehalis River (Satsop River, Bingham Creek)
1995
98
10.7
0.028
55
Bingham Creek Hatchery
1991,' 1992,'
1995
180
9.0
0.027
56
Chehalis River (Hope Creek)
1994, 1995, 1996
171
9.0
0.030
North Washington coast
57
Queets River
1995
99
9.0
0.028
58
Clearwater River
1995
100
7.2
0.029
59
Bogachiel River
1987
80
10.7
0.030
60
Sol Due Hatchery Summer Run
1994'
80
7.2
0.030
61
Sol Due River Summer Run
1995
120
10.7
0.030
62
Sol Due Hatchery Fall Run
1995'
80
9.0
0.032
63
Hoko River
1987
96
9.0
0.033
Puget Sound and Hood Canal
64
Dungeness Hatchery
1987
80
12.5
0.037
65
Quilcene Hatchery
1994'
100
9.0
0.025
66
North Fork Skokomish River
1994,' 1995'
126
7.2
0.030
67
Dewatto River
1994,' 1995,'
1996'
169
9.0
0.028
68
Minter Creek Hatchery
1992,' 1995'
80
9.0
0.035
69
Soos Creek Hatchery
1994,' 1995,
1996
680
9.0
0.034
70
Snoqualmie River (Harris Creek)
1987
120
7.2
0.034
71
Snoqualmie River (Grizzly Creek)
1994,' 1995,'
1996'
215
7.2
0.030
72
North Fork Skykomish River (Lewis Creek)
1995'
102
9.0
0.032
73
North Fork Stillaguamish River (Fortson Creek)
1987, 1989 '
200
9.0
0.031
74
North Fork Stillaguamish River (Mcgovem Creek)
1987
40
10.7
0.032
75
Upper Skagit River
1993
127
9.0
0.033
76
Skagit River (Carpenter Creek)
1993
139
9.0
0.032
77
Skagit River (West Fork Nookachamps Creek)
1987, 1993
220
9.0
0.035
78
Skagit River (Baker River)
1992'
303
10.7
0.036
79
Skagit River (Suiattle River, All Creek)
1987, 1993
200
10.7
0.032
80
Skagit River (Upper Sauk River)
1992, 1993
200
9.0
0.034
81
Skagit River (Upper Cascade River)
1992. 1993
224
9.0
0.031
82
Samish River (Ennis Creek)
1994, ' 1995,
' 1996
' 167
9.0
0.035
British Columbia
83
Chilliwack River Hatchery
1984
100
10.7
0.034
84
Cowichan River Hatchery
1984
80
9.0
0.036
85
Big Qualicum Hatchery
1989,' 1991
180
10.7
0.037
86
Robertson Creek Hatchery
1984
100
9.0
0.030
87
Capilano Hatchery
1989, ' 1991
200
12.5
0.038
88
Squamish River Hatchery
1988'
98
7.2
0.035
Upper Fraser River
89
Spius River Hatchery
1987
200
10.7
0.035
Meai
10.0
0.033
' Sample taken from adult fish. All other samples were from juvenile fish.
644
Fishery Bulletin 101 (3)
Table 2
Enzymes and study results for 59 loci in samples
of 89 coho salmon populations from the Pacific Northwest and California.
Number of
Range of
Enzyme or
Enzyme commission
Locus
populations
common allele
protein name
number
abbrev.
polymorphic
frequency
Aspartate aminotransferase
2.6.1.1
sAAT-1,2*
36
1.000-0.966
sAAT-3*
1
1.000-0.956
sAAT-4*
71
1.000-0.839
Adenosine deaminase
3.5.4.4
ADA-1*
34
1.000-0.924
ADA-2*
15
1.000-0.929
Aconitate hydratase
4.2.1.3
mAH-1*
2
1.000-0.992
mAH-2*
22
1.000-0.919
mAH-3*
3
1.000-0.944
sAH*
60
1.000-0.849
Adenylate kinase
2.7.4.3
AK*
4
1.000-0.993
Alanine aminotransferase
2.6.1.2
ALAT*
12
1.000-0.958
Creatine kinase
2.7.3.2
CK-Al*
8
1.000-0.971
CK-A2*
22
1.000-0.919
CK-Cl*
4
1.000-0.983
CK-C2*
9
1.000-0.972
CK-B*
1
1.000-0.999
Esterase
3.1.-.-
EST-1*
85
1.000-0.652
Fructose-bisphosphate aldolase
4.2.1.13
FBALD-3*
1
1.000-0.996
FBALD-4*
14
1.000-0.962
Formaldehyde dehydrogenase (glutathione)
1.2.1.1
FDHG*
33
1.000-0.954
Fumarate hydratase
4.2.1.2
FH*
43
1.000-0.835
b-N-Acetylgalactosaminidase
3.2.1.53
bGALA*
89
0.889-0.357
Glyceraldehyde-3-phosphate dehydrogenase
1.2.1.12
GAPDH-2*
64
1.000-0.713
GAPDH-3*
26
1.000-0.867
GAPDH-4*
9
1.000-0.975
GAPDH-5*
0
1.000-1.000
Glucose-6-phosphate isomerase
5.3.1.9
GPI-A*
17
1.000-0.906
GPI-Bl*
1
1.000-0.962
GPI-B2*
47
1.000-0.815
(ilutathione reductase
1.6.4.2
GR*
6
1.000-0.988
continued
cies over both loci and treated as a single tetrasomic locus.
Following the recommendations of Waples (1990), allelic
frequencies of samples taken in different years from the
same location were combined. In general, little temporal
allele-frequency variation was detected in coho salmon
populations sampled over years (Van Doornik et a!., 2002;
present study). Levels and patterns of genetic variation
within and between populations were estimated with 56
polymorphic loci (Table 2). Average expected heterozygos-
ity per locus (isoloci excluded) for each population was
calculated by using an unbiased estimator (Nei, 1978).
The proportion of P,, (,r, loci was computed for each popula-
tion, in which a locus was considered to be polymorphic
if the frequency of the most common allele was sO.95.
Chord distances (Cavalli-Sforza and Edwards, 1967) were
computed between all pairs of populations with BIOSYS
(Swofibrd and Selander, 1981), and relationships among
populations were depicted with multidimensional scaling
(MDS, NTSYS-PC, Exeter Software, NY). Allele-frequency
variation among baseline populations was partitioned
(Chakraborty et al., 19821 into two geographic levels: 1)
populations within regions; and 2) among regions (Table 1).
These regions were delimited by geography and by genetic
groupings in the MDS analyses.
We used the maximum likelihood procedures of Pella and
Milner (1987) and the Statistical Package for Analyzing
Mixtures (SPAM; Debevec et al., 2000) to estimate stock
contributions to simulated and actual mixtures of coho
salmon. Estimates were made by using 56 polymorphic loci
(Table 2) and 89 baseline populations, except for analysis
of marked (hatchery) fish where only hatchery populations
were used (Table 1). Allocations to individual baseline
populations were then summed to estimate contributions of
regional stock groups (Pella and Milner, 1987). Average mix-
Teel et al : Genetic analysis of juvenile Oncorhynchus kisutch
645
Table 2 (continued)
Number of
Range of
Enzyme or
Enzjone commission
Locus
populations
common allele
protein name
number
abbrev.
polymorphic
frequency
Isocitrate dehydrogenase
1.1.1.42
mIDHP-1*
4
1.000-0.964
mIDHP-2*
11
1.000-0.799
sIDHP-1*
11
1.000-0.948
sIDHP-2*
29
1.000-0.851
Lactate dehydrogenase
1.1.1.27
LDH-Al*
7
1.000-0.700
LDH-A2*
2
1.000-0.995
LDH-Bl*
18
1.000-0.942
LDH-B2*
20
1.000-0.956
LDH-C*
0
1.000-1.000
Malate dehydrogenase
1.1.1.37
sMDH-Al,2*
35
1.000-0.976
sMDH-Bl,2*
21
1.000-0.947
Mannose-6-phosphate isomerase
5.3.1.8
MPI*
41
1.000-0.897
o-Mannosidase
3.2.1.24
MAN*
5
1.000-0.981
Dipeptidase
3.4.-.-
PEPA*
63
1.000-0.895
Tripeptide amino peptidase
3.4.-.-
PEPB-1*
10
1.000-0.979
Peptidase-C
3.4.-.-
PEPC*
89
0.903-0.391
Proline dipeptidase
3.4.-.-
PEPD-2*
56
1.000-0.798
Leucyl-L-tyrosine peptidase
3.4.-.-
PEPLT*
19
1.000-0.953
Phosphogluconate dehydrogenase
1.1.1.44
PGDH*
7
1.000-0.967
Phosphoglycerate kinase
2.7.2.3
PGK-1*
14
1.000-0.930
PGK-2*
13
1.000-0.975
Phosphoglucomutase
5.4.2.2
PGM-1*
72
1.000-0.600
PGM-2*
32
1.000-0.958
Purine-nucleoside phosphorylase
2.4.2.1
PNP-1*
87
1.000-0.614
Pyruvate kinase
2.7.1.40
PK-2*
14
1.000-0.980
Triose-phosphate isomerase
5.3.1.1
TPI-1*
5
1.000-0.986
TPI-2*
0
1.000-1.000
TPI-3*
27
1.000-0.930
TPI-4*
2
1.000-0.994
ture estimates derived from 100 simulated mixtures were
used to evaluate the accuracy of estimated contributions to
each region with mixture sizes of 100, 300, and 500 fish.
We analyzed mixtures composed of fish entirely from each
region and also mixtures that excluded fish from regions
south and north of our marine sampling area. Precisions of
the stock composition estimates for the actual mixtures were
estimated by bootstrapping baseline and mixture genetic
data 100 times as described in Pella and Milner ( 1987).
Stock compositions were estimated for June and
September-October. We also combined samples over sur-
veys and made separate estimates from samples of marked
( fin-clipped and tagged hatchery fish ) and unmarked fish to
examine hatchery and wild stock compositions. However,
because not all hatchery fish are marked, unmarked fish
are a mixture of wild and hatchery fish. We therefore es-
timated the proportion of hatchery fish for a region in the
sample of unmarked fish (Pyf^) by
where P
MH
RvIRm
^U^^M
= the proportion of hatchery fish from a par-
ticular region in the sample of marked fish;
= the ratio of unmarked to marked releases in
a region; and
= the ratio of unmarked to marked fish in our
ocean samples.
The R,,/Rf^f for 1997 and 1998 brood years varied consider-
ably among regions; California coast 1.0, Klamath River to
Cape Blanco 0.01, Oregon coast 0.12, Columbia River 0.12,
southern Washington coast 0.03, northern Washington
coast 0.69, Puget Sound 0.43, southern British Columbia
0.09, and Upper Eraser River 0.80 (Lavoy'; PSMFC^). We
PuH =(Pmh(Ru'Rm^)/(S„/SJ,
(1)
Lavoy, L. 2001. Personal commun. Washington Department
of Fish and Wildlife, Olympia, WA. 98501.
■ PSMFC (Pacific States Marine Fisheries Commission).
2001. Regional Mark Information System (RMIS) coded-wire
tag on-line database. [Available from Pacific States Marine
Fisheries Commission, 45 SE 82°'' Dr., Suite 100, Gladstone,
OR 97027-2522.1
646
Fishery Bulletin 101(3)
then subtracted P^,^ for each region from the genetic esti-
mate of the region's contribution to the sample of unmarked
fish. The sum of the remaining values estimated the pro-
portion of wild fish in the sample of unmarked fish. When
Pijfj for a region was greater than the genetic estimate of
the region's contribution to the sample of unmarked fish,
the percentage of wild fish from that region was considered
to be zero.
We estimated regional proportions of hatchery and wild
coho salmon in the all-fish marine sample that included
both marked and unmarked coho salmon. Regional hatch-
ery contributions to the all-fish sample were made by sum-
ming each region's estimated contribution to the sampled
marked and unmarked fish, weighted by the proportion of
each of these sample types in the total sample. Regional
proportions of wild coho salmon in the all-fish sample
were made by multiplying a region's estimated proportion
of wild coho salmon in the unmarked sample by the propor-
tion of unmarked fish in the total sample.
Results
Baseline genetic data and population structure
Although coho salmon generally have low levels of genetic
variability in relation to other Pacific salmon, a sufficient
number of polymorphic loci were detected to distinguish
many populations and regional population groups. Of 59
loci screened in all 89 populations, 56 were polymorphic,
and 29 of these were at the Pq 95 level of polymorphism
in at least one population (Table 2). Allelic frequencies
are reported in an appendix that can be retrieved at
the Northwest Fisheries Science Center website [http:
//www.nwfsc. noaa.gov]. Twenty of the 56 polymorphic loci
had two alleles per locus, 24 had three alleles per locus,
nine had four alleles, two had five alleles, and one had six
alleles. Two loci (BGALA* and PEPC* ) varied in all popula-
tions studied. Three loci (GAPDH-5*, LDH-C*, and TPI-2*)
were monomorphic in all populations. Observed genotypic
proportions for polymorphic loci in 128 samples departed
significantly (P<0.05) from expected Hardy- Weinberg pro-
portions in 75 of 1476 tests (5.1% ). There were no consis-
tent trends by population or locus. Because the number of
significant tests is close to the number expected by chance
for this rejection level, we did not attach any biological
significance to these departures.
The percentages of Ppgr^ loci and average heterozygosi-
ties over 56 loci for each population appear in Table 1. The
percentage of P^^^ loci ranged from only 5.4% in Lewis
River hatchery early run (population 41) to 17.9% in the
Mad River hatchery (4). Average heterozygosities ranged
from 0.021 in Iron Gate hatchery (5) and Elk River (9) to
0.046 in Sandy River hatchery (46 ). Gene diversity analysis
of the 89 populations resulted in a total gene diversity (//•,.)
of 0.035 and an average sample diversity (H^) of 0.033.
Thus, 94.5% of the total genetic diversity was attributable
to within-sample variability and 5.5% was attributable to
variability among samples. About 2.9% of the total gene
diversity was due to variability among populations within
regions, and 2.6% was due to variability among the nine
regions.
Genetic relationships among populations of coho salmon
as revealed by two-dimensional MDS analysis showed that
genetic differences among populations were geographically
structured (Fig. 2). The first axis in the plot separated pop-
ulations in coastal Oregon and California from northern
populations. Several populations, including two from the
Rogue River in southern Oregon (numbers 7 and 8) and
Big Qualicum hatchery (85) on Vancouver Island, were po-
sitioned near the convergence of the southern and northern
population groups. The Iron Gate hatchery sample (5) from
the Klamath River, California, clustered with the northern
population group. Several genetically discrete groups ap-
peared on smaller geographical scales. However, samples
from Iron Gate hatchery (5), Yaquina River (27), Nehalem
hatchery (33), Willapa Bay area (50, 51, and 52), Dungeness
hatchery (64), McGovern Creek (74), upper Cascade River
(81), and Ennis Creek (82) did not cluster with nearby
populations. The single population in our study from the
upper Fraser River region — Spius hatchery (89) of the
Thompson River-was the most genetically distinct in the
MDS analysis (.r=-2.3,y=-0.9) and was positioned beyond
the scaling shown in Figure 2. The Little River (2) popula-
tion also fell outside the area of the plot {.v=5.1,y=2.0), but
was genetically most similar to other California coastal
populations (1, 3, and 4).
Genetic estimates of simulated stock mixtures
One demonstration of discreteness among regional groups
is the correct allocation in a mixed-stock analysis of simu-
lated samples from baseline populations to their stock of
origin. We used simulated sample sizes of 100, 300, and
500 taken from one region at a time; therefore the results
represent the accuracy of reallocation back to the region of
origin. Table 3 presents the average values of 100 bootstrap
resamplings of both the baseline and the mixture samples.
For simulated sample sizes of 100, reallocation accuracy
ranged from 81% (coastal northern Washington) to 98%
(upper Fraser River population) and averaged 88.7% over
the nine regions. Average accuracy increased to 92.9% with
an increase in the size of the simulated sample to 300. Only
marginal improvement (93.6%r accuracy) was achieved by
increasing the simulated sample size to 500.
We also used mixed-stock analysis of simulated samples
to examine the accuracy of composition estimates for Cali-
fornia, Puget Sound, and British Columbia regions when
fish from these areas were not present in mixtures. Average
values for sample sizes of 100 ranged from 0%^ (California
coast, upper Fraser River) to 4% (Oregon coast) and aver-
aged 1.8% over the five regions (Table 4). Increased sample
sizes of 300 and 500 resulted in small improvements in
average accuracy (1.4%^ and 1.0 %).
Stock compositions of ocean-caught coho salmon
Genotypes for 56 loci were scored for 730 juvenile coho
salmon captured in ocean trawls in 1998-2000 (Table 5).
About 65% of the 455 fish in June trawls were sampled
Teel et al.: Genetic analysis of juvenile Oncorhynchus kisutch
647
E
Q
British Columbia
upper Cascaae R _ ^ 84 ^^K,^© 64 Dongoness H
>.j O _?4
83 ^ai „K 87 o
Q27 YaquBiaR
California
coast
V
033
NehalemH
Puget Sound
north
Washington
coast
Oregon
coast
Klamath River
to Cape Blanco
south
Washington
Coast
■1 .0 -0.3 0 4 10 1 7
Dimension 1
Figure 2
Multidimensional scaling (MDS) of Cavalli-Sforza and Edwards ( 1967) chord distances based on 56
allozyme loci between samples of 89 populations of coho salmon extending from northern California
to southern British Columbia. Location numbers are given in Table 1 and Figure 1. Populations
within regions are identified with polygons where possible. Open circles indicate populations that
did not cluster closely with nearby populations. Populations 2 from the California coast and 89 from
the upper Fraser River fall beyond the scale of the plot.
Table 3
Mean estimated percentage contributions (± standard deviations) of 100 bootstrap resamplings
of mixtures composed of fish from
only one region. Population numbers are
explained in Table 1.
71=100
7^=300
n=500
Region of largest
Region (populations)
Estimate
Estimate
Estimate
misallocation
California coast (1-4)
95 ±4
97 ±3
97 ±3
Oregon coast
Klamath River to Cape Blanco (5-
-9)
94 ±5
96 ±3
96 ±3
Columbia River
Oregon coast (10-33)
86 ±7
91 ±4
92 ±3
Klamath River to Cape Blanco
Columbia River (34-49)
84 ±8
92 ±3
93 ±3
Oregon coast
South Washington coast (50-56)
88 ±8
95 ±3
95 ±3
North Washington coast
North Washington coast (57-63)
81 ±10
88 ±5
90 ±4
Puget Sound
Puget Sound (64-82)
85 ±7
90 ±5
90 ±4
British Columbia
British Columbia (83-88)
83 ±9
88 ±6
90 ±5
Puget Sound
Upper Fraser River (89)
98 ±2
99 ±1
99 ±1
Columbia River
in the two northern most transects along the Washington
coast, 24% in three transects closest to the Columbia River,
and 10% in the four most southern transects along the
Oregon coast. Samples from these three areas comprised
43%, 23%, and 33%, respectively, of the 275 fish caught
in September trawls. The numbers of offshore juveniles
caught in 1998 were too small to provide accurate mixed-
stock estimates; therefore the ocean samples collected in
1999 and 2000, and a sample pooled over 1998-2000, were
analyzed separately. In the 1998-2000 pooled sample,
Columbia River populations were estimated to be the
major contributing regional group in June (47%, SD=6%)
648
Fishery Bulletin 101(3)
Table 4
Actual percentage composition and mean
estimated percentage contributions
(± standard deviations) of 100 bootstrap
resamplings
of mixtures composed of 100, 300, and 500 fish
Popul
ation numbers are
explained in Table 1.
« = 100
n=300
n=500
Region (populations)
Actual
Estimate
Estimate
Estimate
California coast ( 1-4)
0
0±1
0±0
0±0
Klamath River to Cape Blanco (5-9)
0
3 ±4
2 ±2
1±2
Oregon coast (10-33)
20
21 ±8
20+5
20 ±4
Columbia River (34-49)
50
44 ±11
46 ±6
48 ±5
South Washington coast (50-56)
15
14 ±8
14 ±4
14 ±3
North Washington coast (57-63)
15
10 ±7
12 ±5
12 ±4
Puget Sound (64-82)
0
4 ±5
4 ±3
3 ±2
British Columbia (83-88)
0
2 ±3
1±2
1 ±1
Upper Fraser River (89)
0
0±1
0±1
0±0
and September (32%, SD=9%). The Oregon coastal region
contributed about 18% (SD=5%') to the June mixture and
21% (SD=7% ) to the September sample. The estimated con-
tribution of Puget Sound fish to the pooled ocean samples
was much greater in September ( 17%, SD=7%) than it was
in June(3%, SD=2%').
Genetic mixed-stock analysis of ocean-caught hatchery
fish with CWTs provided a direct comparison of genetic
estimates and a mixed-stock sample of known origins
(Brodziak et al. 1992). Only 41 fish had CWTs (Table 6).
No fish with CWTs appeared in the 1998 sample. Most of
the fish with CWTs in 1999 and 2000 originated from Co-
lumbia River (68%, n=28) and Oregon coastal (12%, n=5)
hatcheries. In the genetic analysis of the 41 fish, Columbia
River hatcheries were estimated to contribute about 22 fish
(53% , SD=2 1% ). Approximately 7 fish ( 16% , SD= 17% ) were
estimated to originate from Oregon coastal hatcheries.
Of the 730 juveniles sampled during the study, 501
(69% ) bore hatchery marks ( clipped adipose fins ). The per-
centage of unmarked fish in the September sample (35%)
was greater than that in June (29% ). Genetic mixed-stock
estimates for hatchery-marked fish alone indicated that
69% (SD=6%) originated from the Columbia River and
14% (SD=4%) from Oregon coastal hatcheries (Table 7).
The sample of unmarked fish, which contained a mixture
of wild and unmarked hatchery fish, was estimated to
have a much smaller proportion of Columbia River fish
(20%, SD=8% ) but a larger proportion of coastal Oregon
(36%, SD=9%) and northern Washington (25%, SD=7%)
fish (Table 7). About 30% of unmarked fish in the pooled
ocean sample originated from hatcheries (Eq. 1) and 70%
from wild populations. Estimated contributions from
hatchery and wild populations of all ocean juveniles
sampled (marked and unmarked) were 78% and 22%, re-
spectively. Coho salmon originating in the Columbia River
were estimated to comprise 54% of the total sample, but
only 1% consisted of wild fish. Oregon coastal rivers con-
tributed 21% to the total ocean sample, and nearly equal
proportions were contributed from hatcheries and wild
populations.
Discussion
Usefulness of coho salmon allozyme data
for mixed-stock analysis
Although the gene diversity analysis indicated that the
level of allele-frequency differentiation among populations
within regions was similar to that between regions, further
analyses showed that the magnitude of regional differen-
tiation in the baseline was sufficient to provide accurate
mixed-stock estimates. First, we found several genetically
discrete population groups of coho salmon over an area
extending from California to southern British Columbia.
Most of the samples in the MDS plot clustered with nearby
samples, and the north-south arrangement of neighboring
population groups indicated that isolation by distance is
an important component of genetic population structure
on this geographic scale. As with other species of Pacific
salmon, natal homing to spawning areas is an important
isolating mechanism between populations of coho salmon.
Second, the analysis of simulated stock mixtures also
demonstrated that regional differences were sufficient to
provide reliable estimates of coho salmon stock composi-
tions. Accurate estimates were obtained from simulated
sample sets composed of 100% contributions from each re-
gion (Table 3). Third, a more rigorous test of the adequacy
of the baseline was made by comparing genetic estimates
with direct determinations based on CWTs. These esti-
mates were reasonably accurate, especially for the largest
contributing regions (Table 6), given the small sample of
only 41 fish bearing CWTs. Both the simulation and CWT
mixture results are consistent with the findings of Wood et
al. ( 1987 ) that estimation accuracy decreases substantially
when mixture sample sizes are small and when genetic
separation among stocks is limited. Lastly, the analyses
of ocean-caught mixture samples themselves appeared to
provide reasonable composition estimates (Table 5). Ad-
ditionally, estimates for samples pooled over years tended
to be intermediate between the two annual estimates, as
would be expected from pooling.
Teel et al : Genetic analysis of juvenile Oncorhynchus kisutch
649
Table 5
Estimated percentage stock compositions (standard deviations), samp
e sizes (n
). and recoveries of coded wire
tags (CWT) for coho 1
salmon sampled in trawl surveys along the Oregon and Washington
coasts in
June and September 1998, 1999, and 2000
. Stock
compositions were not estimated for June [n
=43) and September (n--
= 18) 1998 because of small
sample sizes. None of the 1998 |
samples contained coded wire tags.
Region
June
September
Est.
CWT
Est.
CWT
1999
n
California coast
278
152
0±1
0
0±0
0
Mamath River to Cape Blanco
6 ±6
0
0±0
0
Oregon coast
25 ±7
5
25 ±8
0
Columbia River
46 ±9
8
20 ±14
4
South Washington coast
11 ±4
2
9 ±5
0
North Washington coast
10 ±5
2
18 ±15
0
Puget Sound
3 ±4
0
25 ±9
1
British Columbia
0±1
0
3 ±3
0
Upper Fraser River
0±0
0
0±0
0
2000
n
California coast
134
105
0±0
0
1±3
0
Klamath River to Cape Blanco
1±7
0
0±0
0
Oregon coast
11 ±7
0
17 ±8
0
Columbia River
40 ±11
11
48 ±16
5
South Washington coast
17 ±7
0
6 ±9
0
North Washington coast
21 ±11
0
10 ±16
0
Puget Sound
11 ±8
0
14 ±7
1
British Columbia
0±0
0
3 ±8
2
Upper Fraser River
0±0
0
0±0
0
1998, 1999, and 2000 combined
n
California coast
455
275
0±0
0
0±0
0
Klamath River to Cape Blanco
7 ±4
0
0±0
0
Oregon coast
18 ±5
5
21 ±7
0
Columbia River
47 ±6
19
32 ±9
9
South Washington coast
11 ±3
2
9 ±4
0
North Washington coast
13 ±4
2
19 ±11
0
Puget Sound
3 ±2
0
17 ±7
2
British Columbia
0±0
0
2 ±2
2
Upper Fraser River
0±0
0
0±0
0
Nonetheless, the usefulness of the allozyme baseline that
we compiled for coho salmon is limited by two factors. First,
few samples in the baseline are from California and British
Columbia populations. Although the baseline appears to be
adequate to analyze stock mixtures of juvenile coho salmon
off Oregon and Washington, mixed stock analyses of sam-
ples from other marine areas, particularly to the north, re-
quires the sampling of additional populations. Second, our
study demonstrated that estimates of stock compositions
are not sufficiently accurate to effectively identify stock
groups that are absent from mixtures or present in small
proportions (Tables 4 and 6). Estimation accuracy can be
improved by using additional gene markers. These markers
will likely be based on DNA variability because coho salmon
minisatellite (Miller et al., 1996; Beacham et al., 1996) and
microsatellite (Small et al., 1998a; 1998b; Beacham et al.,
2001) loci show much higher levels of polymorphism than
do allozyme loci. Recently, variation at eight microsatellite
DNA loci and one Mhc locus in coho salmon populations
in British Columbia and Washington was used to estimate
the stock compositions of fisheries off the west coast of
Vancouver Island (Shaklee et al., 1999; Beacham et al.,
2001). However, the use of highly polymorphic microsatel-
lite loci may not provide increased discrimination among
populations on large geographical scales because of allelic
convergence from multiple mutations (Nauta and Weiss-
650
Fishery Bulletin 101(3)
ing, 1996). Nonetheless, the extension of a DNA baseline to
include populations in Oregon and California, may resolve
fine-scale (geographic and temporal) differences between
coho salmon populations in southern coastal areas.
Stock compositions of ocean-caught juvenile
coho salmon
Studies using large purse seines conducted in 1981-85
revealed that juvenile coho salmon were the most abundant
of the Oncorhynchus species in the nearshore areas along
the Oregon and Washington coasts ( Pearcy and Fisher, 1988;
1990). Pearcy and Fisher (1988; 1990) captured hatchery-
Table 6
Actual composition
and estimated contributions
± stan-
dard deviations) of a mixture of 41-CWT fish.
Region
Actual
Genetic estimate
Number
7c
Number
%
California coast
0
0
1
3 ±4
Klamath River to
Cape Blanco
0
0
0
0±0
Oregon coast
5
12
7
16 ±17
Columbia River
28
68
22
53 ±21
South Washington coast
2
5
0
0±0
North Washington coast
3
7
5
11 ±9
Puget Sound
1
2
5
11 ±18
British Columbia
2
5
2
6 ±11
Upper Fraser River
0
0
0
0±0
tagged juvenile coho salmon and concluded they were not
highly migratory, often remaining close to their point of sea
entry for several months. Our genetic results corroborate
that finding. Genetic estimates indicate that about 89%
of ocean juveniles caught in June and 81% in September
originated from the Columbia River and adjacent coastal
rivers. Recoveries of hatchery-tagged fish in=41) also indi-
cate that juveniles remain near river mouths in their first
few months after ocean entry; only three of these CWT-
marked fish came from hatcheries in other regions.
However, our genetic results indicate that a change has
occurred in the distribution of Washington coastal and
Puget Sound juvenile coho salmon. In the 1980s, juvenile
coho salmon from Washington coastal hatcheries were
not recovered along the Washington and Oregon coasts
after mid summer, apparently having migrated northward
(Pearcy and Fisher, 1988). Pearcy and Fisher (1990) also
found that Puget Sound coho salmon did not migrate along
the Washington and Oregon coast until sometime between
their first and second summer at sea. However, our genetic
results showed that in 1998-2000 fish from Washington
coastal streams and hatcheries comprised substantial
proportions of the juveniles in nearshore areas along
the Washington and Oregon coast in both early and late
summer (24% and 28%). We also found that juvenile coho
salmon from Puget Sound are present in late summer. Our
finding that coho salmon from northern stocks move south
along the coast during their first summer was substanti-
ated by the catch of CWT-marked fish originating from
Puget Sound (n=2) and southern British Columbia {n=2).
Recent reductions in the number of coho salmon smolts
released from the region's hatcheries have not resulted in
a decrease in the proportion of hatchery juveniles along the
Oregon and Washington coasts. Annual releases of hatch-
Table 7
Estimated percentage stock
compositions and sampl
3 sizes for populations of marked (fish with clipped adipose fins) and unmarked
coho salmon sampled in trawl surveys along the Oregon and Washington
coasts in 1998
1999,
and 2000. Samples from June and
September were combined.
Separate estimates for the contributions of hatchery and wild stocks were made by
using estimates of |
hatchery marking rates for each region.
Marked fish
Unmarked fish
(hatchery fish)
(hatchery and Wild fish)
All fish
Genetic estimate
Genetic estimate
(n=501)
(n=229)
Hatchery
Wild
Hatchery
Wild
Total
Region
(%)
(%)
(%)
(%)
(%)
(%)
(%)
California coast
0±0
1±2
0
1
0
0
0
Klamath River to Cape Blanco 1 ±2
1±7
0
1
1
0
1
Oregon coast
14 ±4
36 ±9
4
32
n
10
21
Columbia River
69 ±6
20 ±8
18
2
53
1
54
South Washington coast
4 ±4
9 ±5
0
9
3
3
6
North Washington coast
1±7
25 ±7
2
23
1
7
8
Puget Sound
6 ±5
8 ±5
6
2
6
1
7
British Columbia
5 ±2
0±0
0
0
3
0
3
Upper Fraser River
0±0
0±0
0
0
0
0
0
Total
100
100
30
70
78
22
100
Teel et a\: Genetic analysis of juvenile Oncorhynchus ktsutch
651
ery smolts exceeded 64 million fish during the early 1980s
but have decreased to about 39 million in recent years, a
40% reduction (PSMFC^; NRC^). Nonetheless, the propor-
tion of hatchery coho salmon in nearshore marine waters
has remained high, averaging 74% in 1981-85 (Pearcy and
Fisher, 1990) and 78% in 1998-2000 (present study). This
result, therefore, leads to the conclusion that the number
of naturally produced juveniles in Oregon and Washington
coastal waters has also decreased proportionately during
this period. If so, wild populations of coho salmon may also
have experienced a decline in abundance on the order of
40%.
Steep declines in Columbia River wild populations are
particularly evident. At the beginning of the 20"^ century,
populations in the Columbia River are thought to have
been the largest producers of coho salmon in the region
(Chapman, 1986; Lichatowich, 1989) and likely contributed
a substantial proportion to the nearshore population of ju-
venile salmon. At present, Columbia River juveniles pre-
dominate along the coast. However, these fish are almost
entirely releases from hatchery facilities and Columbia
River wild coho salmon are conspicuously absent.
Acknowledgments
We are grateful to George Milner and Paul Aebersold who
developed much of the allozyme baseline for coho salmon.
Sewall Young, Laurie Weitkamp, Kathleen Neely, Bill
Waknitz, Kathryn Kostow, Orlay Johnson, Ken Currens,
Eric Beamer, Scott Chitwood, Doug Cramer, Marc Miller,
and Jennifer Nielsen provided baseline samples. We thank
Ed Casillas, Ric Brodeur, Bob Emmett, Cindy Bucher, Susan
Hinton, Cheryl Morgan, Paul Bentley, and Joe Fisher for
providing coho salmon samples and data from their coastal
salmon surveys. This study was supported in part by funds
from the Bonneville Power Administration and the U.S.
GLOBEC program as part of an initiative to understand
the effects of ocean dynamics on salmon populations.
Literature cited
Aebersold. P. B., G. A. Winans, D. J. Teel, G. B. Milner, and
F. M. Utter.
1987. Manual for starch gel electrophoresis: a method for the
detection of genetic variation. U.S. Dep. Conimer., NOAA
Tech. Report NMFS 61, 19 p.
Beacham, T. D., J. R. Candy, K. J. Supemault, T. Ming, B. Deale,
A. Schulze, D. Tuck, K. H. Kaukauna. J. R. Irvine, K. M. Miller,
and R. E. Wither.
2001. Evaluation and application of microsatellite and major
histocompatibility complex variation for stock identification
of coho salmon in British Columbia. Trans. Am. Fish. Soc.
130;1116-1149.
3 NRC (Natural Resource Consultants, Inc.). 1995. Database
of artificially propagated anadromous salmon (database).
[Available from Environmental and Technical Services Division,
NMFS, 525 N.E., Oregon Street, Portland, OR 97232.1
Beacham, T. D., K. M. Miller, and R. E. Withler
1996. Minisatellite DNA variation and stock identification
of coho salmon. J. Fish Biol. 49:411-429.
Beamish, R. J., D. McCaughran, J. R. King, R. M. Sweeting, and
G. A. McFarlane.
2000. Estimating the abundance of juvenile coho salmon in
the Strait of Georgia by means of surface trawls. North
Am. J. Fish. Manage. 20:369-375.
Brodziak, J., B. Bentley, D. Bartley, G.A.E. Gall, R. Gomulkiewicz,
and M. Mangel.
1992. Tests of genetic stock identification using coded wire
tagged fish. Can. J. Pish. Aquat. Sci. 49:1507-1517.
Cavalli-Sforza, L. L., and A. W. F Edwards.
1967. Phylogenetic analysis: models and estimation pro-
cedures. Evolution 21:550-570.
Chakraborty, R., M. Hagg, N. Ryman, and G. Stahl.
1982. Hierarchical gene diversity analysis and its applica-
tion to brown trout population data. Hereditas 97:17-21.
Chapman, D. W.
1986. Salmon and steelhead abundance in the Columbia
River in the nineteenth century. Trans. Am. Fish. Soc.
115:662-670.
Debevec, E. M., R. B. Gates, M. Masuda, J. Pella, J. Reynolds, and
L. W. Seeb.
2000. SPAM (version 3.2): statistics program for analyzing
mixtures. J. Hered. 91:509-511.
Emmett, R. L., and R. D. Brodeur.
2000. Recent changes in the pelagic nekton community off
Oregon and Washington in relation to some physical oceano-
graphic conditions. North Pacific Anadr. Fish Comm. Bull.
2:11-20.
Godfrey, H.
1965. Coho salmon in offshore waters. In Salmon of the
North Pacific Ocean. Part IX. Coho, chinook, and masu
salmon in offshore waters, p. 1-39. Int. North Pacific
Fish. Comm. Bull. 16.
Guo, S. W., and E. A. Thompson.
1992. Performing the exact test of Hardy- Weinberg propor-
tions for multiple alleles. Biometrics 48:361-372.
Guthrie, C. M., E. V. Fariey Jr., N. M. L. Weemes, and
E. C. Martinson.
2000. Genetic stock identification of sockeye salmon cap-
tured in the coastal waters of Unalaska island during
ApriL/May and August 1998. North Pacific Anadr. Fish
Comm. Bull. 2:309-315.
Hartt, A. C, and M. B. Dell.
1986. Early oceanic migrations and growth of juvenile
Pacific salmon and steelhead trout. Int. N. Pac. Fish.
Comm. Rep. 46, 105 p.
Johnson, O. W, T A. Flagg, D. J. Maynard, G. B. Milner, and
F W. Waknitz.
1991. Status review for lower Columbia River coho salmon.
U.S. Dep. Commerce, NOAA Tech. Memo. NMFS F/NWC-
202, 94 p.
Lawson, P. W., and R. M. Comstock.
2000. The proportional migration selective fishery model.
In Sustainable fisheries management: Pacific salmon (E. E.
Knudsen, C. R. Steward, D. D. MacDonald, J. E. WilUams,
and D. W Reiser, eds. ), p. 423-433. Lewis Pubhshers, Boca
Raton, FL.
Lichatowich, J. A.
1989. Habitat alteration and changes in abundance of coho
{Oncorhynchus kisutch) and chinook [Oncorhynchus tshaw-
ytscha) salmon in Oregon's coastal streams. In Proceed-
ings of the national workshop on effects of habitat alteration
on salmonid stocks, May 6-8, 1987, Nanaimo, B.C. (C. D.
652
Fishery Bulletin 101 (3)
Levings, L. B. Holtby, and M. A. Henderson, eds.), p. 92-99.
Can. Spec. Publ. Fish. Aquat. Sci. 105.
Miller, K. M., R. E. Withler, and T. D. Beacham.
1996. Stock identification of coho salmon (Oncorhynchus
kisutch) using minisatellite DNA variation. Can. J. Fish.
Aquat. Sci. 53:181-195.
Milner, G. B., D. J. Teel, F M. Utter, and G. A. Winans.
1985. A genetic method of stock identification in mixed
populations of Pacific salmon, Oncorhynchus spp. Mar.
Fish. Rev. 47:1-8.
Nauta, M. J., and F J. Weissing.
1996. Constraints on allele size at microsatellite loci: implica-
tions for genetic differentiation. Genetics. 143:1021-1032.
Nei, M.
1978. Estimation of average heterozygosity and genetic
distance from a small number of individuals. Genetics
89:583-590.
Orsi, J. A., M. V. Sturdevant, J. M. Murphy, D. G. Mortensen, and
B. L. Wing.
2000. Seasonal habitat use and early marine ecology of juve-
nile Pacific salmon in southeastern Alaska. North Pacific
Anadr Fish Coram. Bull. 2:111-122.
Pearcy W. G., and J. P Fisher.
1988. Migrations of coho salmon, Oncorhynchus kisutch,
during their first summer in the ocean. Fish. Bull. 86:
173-195.
1990. Distribution and abundance of juvenile salmonids off
Oregon and Washington, 1981-1985. U.S. Dep. Commerce,
NOAA Tech. Report NMFS 93, 83 p.
Pella, J., and G. B. Milner
1987. Use of genetic marks in stock composition analyses.
In Population genetics and fishery management (N. Ryman
and F Utter, eds.), p. 247-276. Washington Sea Grant,
Univ. Washington Press, Seattle, WA.
Raymond, M., and F Rousset.
1995. GENEPOP (versions 1.2 and 3.1): population genet-
ics software for exact tests and ecumenism. Heredity 86:
248-249.
Shaklee, J. B., F W. Allendorf, D. C. Morizot, and G. S. Whitt.
1990. Gene nomenclature for protein-coding loci in fish.
Trans. Am. Fish. Soc. 119:2-15.
Shaklee, J. B., T. D. Beacham, L. Seeb, and B. A. White.
1999. Managing fisheries using genetic data: case studies
from four species of Pacific Salmon. Fish. Res. 43:45-78.
Small, M. P, R. E. Withler, and T. D. Beacham.
1998a. Population structure and stock identification of Brit-
ish Columbia coho salmon iOncorhynchus k!sutch)based on
microsatellite DNA variation. Fish. Bull. 96:843-858.
Small, M. P, T. D. Beacham, R. E. Withler, and R. J. Nelson.
1998b. Discriminating coho salmon (Oncorhynchus kisutch )
populations within the Eraser River, British Columbia,
using microsatellite DNA markers. Mol. Ecol. 7:141-155.
Swofford, D. L., and R. B. Selander.
1981. BIOSYS-1: a FORTRAN program for the comprehen-
sive analysis of electrophoretic data in population genetics
and systematics. J. Hered. 72:281-283.
Utter, F, P. Aebersold, and G. Winans.
1987. Interpreting genetic variation detected by electro-
phoresis. In Population genetics and fishery management
(N. Ryman and F Utter, eds.), p. 21^5. Washington Sea
Grant, Univ. Washington Press, Seattle, WA.
Van Doomik, D. M., M. J. Ford, and D. J. Teel.
2002. Patterns of temporal genetic variation in coho sal-
mon: estimates of the effective proportion of 2-year-olds in
natural and hatchery populations. Trans. Am. Fish. Soc.
131:1007-1019.
Waples, R. S.
1988. Estimation of allele frequencies at isoloci. Genetics
118:371-384.
1990. Temporal changes of allele frequencies in Pacific
salmon: implications for mixed-stock fishery analysis. Can.
J. Fish. Aquat. Sci. 47:968-976.
Weitkamp, L. A., T C. Wainright, G. J. Bryant, G. B. Milner,
D. J. Teel, T G. Kope, and R. S. Waples.
1995. Status review of coho salmon from Washington, Ore-
gon, and California. U.S. Dep. Commer., NOAA Tech.
Memo. NMFS-NWFSC-24, 258 p.
Wood, C. C, S. McKinnell, T. J. Mulligan, and D. A. Foumier.
1987. Stock Identification with the maximum likelihood
mixture model: sensitivity analysis and application to com-
plex problems. Can. J. Fish. Aquat. Sci. 44:866-881.
653
Abstract— Recreational fisheries in
the waters off the northeast U.S. target
a variety of pelagic and demersal fish
species, and catch and effort data
sampled from recreational fisheries are
a critical component of the information
used in resource evaluation and man-
agement. Standardized indices of stock
abundance developed from recreational
fishei-y catch rates are routinely used in
stock assessments. The statistical prop-
erties of both simulated and empirical
recreational fishery catch-rate data
such as those collected by the National
Marine Fisheries Service (NMFS)
Marine Recreational Fishery Statis-
tics Survey (MRFSS) are examined,
and the potential effects of different
assumptions about the error structure
of the catch-rate frequency distribu-
tions in computing indices of stock
abundance are evaluated. Recreational
fishery catch distributions sampled by
the MRFSS are highly contagious and
overdispersed in relation to the normal
distribution and are generally best
characterized by the Poisson or nega-
tive binomial distributions. The model-
ing of both the simulated and empirical
MRFSS catch rates indicates that one
may draw erroneous conclusions about
stock trends by assuming the wrong
error distribution in procedures used to
developed standardized indices of stock
abundance. The results demonstrate
the importance of considering not only
the overall model fit and significance of
classification effects, but also the pos-
sible effects of model misspecification,
when determining the most appropriate
model construction.
The statistical properties of recreational catch rate
data for some fish stocks off the northeast U.S. coast
Mark Terceiro
Northeast Fisheries Science Center
National Marine Fishenes Service, NOAA
166 Water Street
Woods Hole, Massachusetts 02543
E-mail address: mtercer@whsunl wh whoiedu
Manuscript approved for publication
30 January 2003 by Scientific Editor.
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:653-672 (2003).
Major recreational fisheries in the wa-
ters off the northeast U.S. coast target
a wide variety of pelagic and demer-
sal fish species (NMFS, 1995, 1996).
Fishery data collected in the National
Marine Fishery Service ( NMFS ) Marine
Recreational Fishery Statistics Survey
(MRFSS) are the basis of fishery catch
and effort estimates for most of these
recreational fisheries and for indices
of population abundance used in stock
assessments (USDOC, 1992, 2001). For
some stocks, reliable fishery-indepen-
dent data such as research trawl survey
indices are not available, and therefore
the recreational fishery data are essen-
tial for tracking stock abundance. The
intercept (creel sampling) portion of the
MRFSS is an interview-type survey of
recreational fishing trips and is con-
ducted at public fishing sites such as
marinas, launching ramps, fishing
piers, and beaches. MRFSS catch esti-
mates are made by expanding intercept
survey sample catch rates in numbers,
calculated on a per trip basis, by the
estimated total number of recreational
fishing trips. The estimated total
number of fishing trips is calculated
from data collected in a MRFSS tele-
phone survey of households located in
coastal counties. The U.S. Department
of Commerce (USDOC, 1992, 2001) has
provided overviews of the MRFSS inter-
cept and telephone survey methods and
catch estimation procedures.
In many cases recreational and com-
mercial catch rates used as abundance
indices are standardized by using
general linear models that assume a
lognormal error distribution (Gulland,
1956; Robson, 1966; Gavaris, 1980;
Kimura, 1981). Commercial fishery
catch-rate data generally meet tests
of normality when log-transformed
(Gulland, 1956; O'Brien and Mayo,
1988). Because of the efficiency and
"integrating" property of commercial
fishing gear (including trawls, fixed
nets, and longlines), even catch rates
on a per tow or per set basis are usu-
ally lognormally distributed (Taylor,
1953). An important characteristic of
commercial data is that catch rates of
zero (tows or sets with no catch of the
target species) are rare.
With the assumption that there is an
underlying lognormal error distribu-
tion, general linear models have often
been used to standardize recreational
fishery catch rates and compute indi-
ces of abundance. This approach has
been used in the assessments of blue-
fin tuna (Brown and Browder, 1994),
summer flounder (TerceiroM, black
sea bass (NEFSC^), tautog (NEFSC^),
winter flounder (NEFSC^), and bluefish
' Terceiro, M. (ed.). 1993. Assessment
of summer flounder (Paralichthys denta-
tus), 1993: report of the stock assessment
workshop summer flounder working group.
Northeast Fisheries Science Center refer-
ence document 93-14, 72 p. Northeast
Fisheries Science Center, Woods Hole, MA
02543.
- NEFSC (Northeast Fisheries Science Cen-
ter). 1996a. Report of the 20th north-
east regional stock assessment workshop
(20th SAW): Stock Assessment Review
Committee (SARC) consensus summary of
assessments. Northeast Fisheries Science
Center reference document 95-18, 210 p.
Northeast Fisheries Science Center, Woods
Hole, MA 02542.
'NEFSC. 1996b. Report of the 21st north-
east regional stock assessment workshop
(21st SAW): Stock Assessment Review
Committee (SARC) consensus summary
of assessments. Northeast Fisheries Sci-
ence Center reference document 96-05d.
200 p. Northeast Fisheries Science Center,
Woods Hole, MA 02543.
654
Fishery Bulletin 101(3)
(NEFSC"*; Gibson and Lazar''). However, Banneret and
Austin (1983) noted that the sampling distribution of rec-
reational catch data is often highly skewed with a longer
right-hand tail than might be expected even from a log-
normal distribution. Furthermore, depending on the way
the catch rate is defined (i.e. catch per trip, day, or hour),
recreational fishery catch-rate distributions may contain a
high proportion of zero catches.
Hilbom ( 1985 ) presented a frequency distribution of num-
bers of salmon caught per trip in the British Columbia sport
fishery that appears to be best characterized by the nega-
tive binomial distribution, with a catch per hour frequency
best characterized by the Poisson distribution. Jones et al.
(1995) investigated the statistical properties of recreational
fishery sampling data collected in angler surveys in Virginia
and noted that the non-normality of recreational fishery
data may violate assumptions of lognormality in methods
used to develop indices of abundance, and especially the
validity of confidence intervals. Power and Moser (1999)
expressed similar concerns about sampled distributions of
fish and plankton collected by research trawl nets, noting
that the assumption of an underlying normal or lognormal
distribution for these types of data is commonplace, and
perhaps in error, and that distributions such as the Pois-
son or negative binomial may be more appropriate. Smith
(1990, 1996) recommended various nonparametric resam-
pling methods (e.g. bootstrap confidence intervals) for char-
acterizing the dispersion of highly skewed research trawl
survey catch distributions having a large proportion of zero
catches. Smith (1999) modeled angling success for salmon,
expressed as the catch after the first hour of angling, using
a negative binomial distribution model.
In addition to the Poisson and negative binomial, al-
ternatives to the lognormal error model for recreational
fishery catch rates also include the delta-lognormal and
delta-Poisson error models. These models are combinations
of the delta distribution (Pennington, 1983) and lognormal
or Poisson model approaches. The delta distribution has
been used in modeling fish and plankton abundance indices
from research trawl survey data, which are characterized
by highly skewed distributions with a relatively high pro-
portion of zero catches (Pennington, 1983). In the combined
delta-lognormal and delta-Poisson approaches, indices of
abundance are modeled as a product of binomially distrib-
uted probabilities of a positive catch and lognormal or Pois-
son distributed positive catch rates. The delta-lognormal
model has been used in modeling fish-spotter data (Lo et
al., 1992) and in the standardization of recreational fishery
catch rates forbluefin tuna (Brown and Porch, 1997; Turner
^ NEFSC. 1997. Report of the 23rd northeast regional stock
assessment workshop (23'''' SAW): Stock Assessment Review
Committee I SAR(^ ) consensus summary of assessments. North-
east Fisheries Science Center reference document 97-05. 191 p.
Northeast Fisheries Science Center, Woods Hole, MA 03543.
■'■' Gibson, M. R.. and N. Lazar 1998. Assessment and projection
of the Atlantic coast bluefish using a biomass dynamic model. A
report to the Atlantic States Marine Kisheri(>s Commission Blue-
fish Technical Committee and Mid-Atlantic Fishery Manage-
ment Council Scientific and Statistics Committee, 29 p. Rhode
Island Division of Fish and Wildlife, Jamestown, RI 02835
et al., 1997; Brown, 1999; Ortiz et al., 1999), both character-
ized by a highly contagious spatial distribution and a large
proportion of zeroes. Bluefin and yellowfin tuna catch rates
in the commercial and recreational fisheries have also been
standardized by using Poisson (Brown and Porch, 1997),
negative binomial (Turner et al., 1997), and delta-Poisson
error distributions (Brown, 2001; Brown and Turner, 2001)
to address these distributional characteristics.
In this study I first examine the statistical properties
of recreational fishery catch-rate data as sampled by the
MRFSS. Next, I examine the goodness of fit to different
statistical distributions of empirical MRFSS catch rates,
on both per trip and per hour bases. I then explore the
effects of five different assumptions about the error struc-
ture of the catch-rate frequency distributions (lognormal,
delta-lognormal, Poisson, delta-Poisson, and negative
binomial) in deriving standardized indices of abundance
with general linear models, using simulated recreational
fishery and empirical MRFSS catch per trip (zero catches
included) data.
Materials and methods
Overview of statistical methods
This work focuses on catch number per trip sampled in
the MRFSS as the index of abundance. The distributional
properties of MRFSS catch-per-hour rates are also exam-
ined, in order to explore whether the general conclusions
reached for catch-per-trip rates are likely to be similar to
catch-per-hour rates. Directed trips are defined as those
for which interviewed anglers indicated that they were
intending to catch a particular species as a primary or
secondary target, whether successful or not (zero catches
included). In analyses of trips for all species, all trips were
used regardless of target or success (zero catches included).
Catch rates were expressed as integer (natural) numbers
of fish per trip or per hour
A value of 1 was added to all observations when applying
a lognormal transformation to allow inclusion of the zero
catch rate observations (this constant was subtracted upon
retransformation to the original scale). Expected sample
values for the lognormal distribution were calculated by
using the normal distribution and log-transformed catch
rates (Sokol and Rohlf 1981). Previous work on MRFSS
catch-per-trip data has shown that the value of 1 is the
appropriate constant to be added (Terceiro'; NEFSC'^) be-
cause it tends to minimize the sum of the absolute value of
skew and kurtosis for these distributions (Berry, 1987). The
standard logarithmic transform bias correction was applied
to express results in the original arithmetic scale (Finney,
1951; Bradu and Mundlak, 1970). No constant was added
when data were analyzed under the assumption of bino-
mial, Poisson, or negative binomial error distributions.
The binomial distribution is a discrete frequency (prob-
ability) distribution of the number of times an event oc-
curs in a sample in which some proportion of the members
possess some variable attribute (Snedecor and Cochran,
1967). Each event is assumed independent of other prior
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
655
Table 1
Descriptive statistics for MRFSS (Marine Recreational Fishery Statistics Survey) 1981, 1988, and 1996 northeast U.S. coast catch
per trip, including zero catches. Catch is given in numbers offish. CV is the coefficient of variation (% ). D is the Kolmogorov test
statistic for normality. Test statistics significant at the 1% level (P<0.01 ) are shown by **, indicating rejection of the null hypothesis
that catch rates follow a normal distribution.
Species
No. of trips
Mean
Median
Variance
CV
Skew
D
1981
Bluefish
4615
3.80
0.00
155.09
328
27.78
0.380**
Summer Flounder
3135
1.88
0.00
14.69
204
4.71
0.312**
Atlantic cod
509
2.55
1.00
13.49
144
2.48
0.244**
Scup
269
8.44
2.00
275.10
196
4.08
0.305**
All species
20,280
3.45
0.00
355.94
547
65.48
0.427**
1988
Bluefish
7294
1.60
0.00
18.65
270
6.42
0.355**
Summer Flounder
4779
2.26
0.00
18.48
190
3.70
0.300**
Atlantic cod
1558
4.56
2.00
21.55
154
3.68
0.258**
Scup
960
9.28
3.00
312.65
190
4.48
0.300**
All species
48,423
2.29
0.00
43.66
289
8.94
0.365**
1996
Bluefish
5457
1.20
0.00
13.12
301
8.40
0.370**
Summer Flounder
7047
2.33
1.00
13.49
157
3.40
0.263**
Atlantic cod
1099
3.97
1.00
43.34
166
3.29
0.273**
Scup
643
13.83
4.00
524.60
165
3.44
0.273**
All species
81,057
2.57
0.00
47.32
268
10.45
0.354**
events in the same sample (Sokal and Rohlf, 1981). In the
present study, the binomial distribution was used only to
model the probabilities of a positive catch (as opposed to a
zero catch; thus the variable attribute of the observation is
either catch or no catch ) in the combined delta-lognormal
and delta-Poisson models.
The Poisson distribution is also a discrete frequency dis-
tribution of the number of times an event (such as catching
a fish during a trip ) occurs in a sample and is characterized
by a small mean value in relation to the observed maxi-
mum number of events within the sample (Sokal and Rohlf,
1981). For a Poisson distribution, the expected variance is
equal to its mean, and Poisson frequency distributions are
more highly skewed than normal or lognormal distribu-
tions (Bliss and Fisher, 1953).
The negative binomial is a discrete frequency distribu-
tion with a higher degree of dispersion than the Poisson
distribution, such that the variance is significantly larger
than the mean. A negative binomial distribution will con-
verge to a Poisson as the variance approaches the mean
(Bliss and Fisher, 1953). Although not as widely applied
as the Poisson in the analysis of count data, there is a
growing literature describing the properties of negative
binomial regression methods to be used when analyzing
"over-dispersed Poisson" frequency distributions (Manton
et al., 1981; Lawless, 1987). The dispersion parameter of
the negative binomial distribution, k, is a positive exponent
relating the mean and variance of the distribution such
that as the variance of a distribution exceeds the mean, the
value ofk decreases and the "over-dispersion" of the distri-
bution in relation to a Poisson distribution increases. The
most efficient estimate of the sample parameter, k', is esti-
mated by maximum likelihood (Bliss and Fisher, 1953).
Descriptive statistics and frequency distributions of
MRFSS catch per trip and catch per hour observations
were compiled by using the SAS FREQ and UNIVARIATE
procedures (SAS, 2000 ). Tests of normality were made with
the Kolmolgorov-Smirnov D-statistic for normality (test
significance expressed as probability < D; SAS, 2000). Eval-
uation of the most appropriate distributional fit to the data
was based on inspection of the frequency distribution plots,
the parametric chi-square (x^) and G-statistic goodness-
of-fit tests, and the nonparametric Kolmogorov-Smimov
(Z3-statistic) goodness-of-fit test for an intrinsic hypothesis
(because the expected distributions were calculated from
the observed sample moments; Sokol and Rohlf, 1981).
For the chi-square and G-tests, when intervals (classes)
of catch per trip with fewer than 3 expected instances
occurred, expected and observed frequencies for these in-
tervals were pooled with the adjacent intervals to obtain a
joint class with an expected frequency of occurrence of 3 or
more (Sokol and Rohlf, 1981). Because of the large sample
sizes involved (»100), the G-test correction suggested
by Williams (1976) proved to be very small in a few test
calculations and therefore was not routinely applied. Un-
realistic (for recreational fishery catch-rate data) negative
expected values computed for the lognormal distributions
were excluded, and the remaining positive distribution was
raised to the observed sample total, so that the expected
proportions at each interval summed to 1.0.
656
Fishery Bulletin 101(3)
Standardized annual indices of abundance derived from
the simulated recreational and empirical MRFSS data
were calculated by using maximum likelihood estimation
to fit generalized linear models with the SAS GENMOD
procedure (SAS, 2000). The SAS (2000) defaults for model
specification were generally followed. An identity link func-
tion was used under the lognormal distribution assump-
tion (catch rates were In-transformed prior to analysis). A
logistic link function was used under the binomial distri-
bution assumption applied for the probability of positive
catch component in the delta-lognormal and delta-Poisson
model approaches. A logarithmic link function was used
under the Poisson and negative binomial assumptions
(SAS, 2000). Type-3 general linear models were fitted in
all cases because the results of this type of analysis do not
depend on the order in which the terms of the model are
specified. The significance of the individual classification
effects (factors) in the models was judged by the chi-square
statistic (Searle, 1987; SAS, 2000).
The overall goodness of fit of the standardization mod-
els was evaluated by using the deviance and log-likelihood
statistics. The deviance is defined to be twice the difference
between the maximum achievable log likelihood and the
log likelihood at the maximum likelihood estimates of the
model parameters (McCullagh and Nelder, 1989). The devi-
ance has a limiting chi-square distribution, and so signifi-
cance is judged by comparison to critical values of the chi-
square distribution. The scale parameter (i.e. for normal
distributions) was held fixed at 1 for all models to facilitate
evaluation of goodness of fit and the degree of overdisper-
sion for models with different error distribution assump-
tions. Holding the scale parameter fixed has no effect on the
estimated intercept or model regression coefficients (e.g.
in the study, the year coefficients that serve as the annual
indices of abundance), but allows equivalent calculation
among models of a "dispersion estimate" (SAS, 2000). This
"dispersion estimate," measured after model fitting as the
deviance divided by the degrees of freedom (deviance/df),
is used to judge whether the data are overdispersed or un-
derdispersed with respect to the error distribution used in
model fitting and is therefore useful in evaluating whether
the correct error distribution assumption has been used in
the model (McCullagh and Nelder, 1989; SAS, 2000).
Descriptive statistics for MRFSS catch rates
The descriptive statistics (mean, median, variance, skew-
ness, and Kolmolgorov-Smirnov (D) normality test statis-
tic) and frequency distributions of MRFSS sample catch
rates for 1981, 1988, and 1996 were examined for four
species from U.S. Atlantic coast waters (Maine to the east
coast of Florida), and in aggregate for all species sampled
along the U.S. Atlantic coast. The following individual
species were considered: bluefish (Pomatomus saltatrix,
an example of a Atlantic coast predatory "gamefish");
summer flounder (Paralichthys dentatus, a Mid Atlantic
Bight demersal flatfish); Atlantic cod (Gadus morhua, a
New England demersal roundfish); and scup (Stenotomus
Table 2
Descriptive statistics for MRFSS (Marine Recreational Fis
hery Statistics Survey) 1981,
1988, and 1996 northeast U.S. coast catch 1
per trip, positive catches only. Catch is given
in numbers
offish. CV is the coefficient of variation (9<
. D is the Kolmogorov test
statistic for normality. Test
statistics significant at the I'X
level (P<0.01)
shown by **,
ndicating rejection of the null hypothesis |
that catch rates follow a normal distribution.
Species
No. of trips
Mean
Median
Variance
CV
Skew
D
1981
Bluefish
2288
7.66
4.00
283.32
220
22.02
0.340**
Summer Flounder
1380
4.26
2.67
23.21
113
3.87
0.222**
Atlantic cod
298
4.36
3.00
15.19
89
2.26
0.188**
Scup
165
13.76
7.33
375.88
141
3.39
0.262**
All species
9484
7.33
3.00
732.27
368
47.02
0.395**
1988
Bluefish
2445
4.75
2.33
40.35
133
4.40
0.254**
Summer Flounder
2326
4.64
3.00
26.92
112
2.94
0.209**
Atlantic cod
1065
6.67
4.00
58.02
114
3.46
0.219**
Scup
614
14.52
7.67
413.03
140
3.88
0.245**
All species
19,094
5.76
3.00
90.39
165
6.49
0.278**
1996
Bluefish
1666
3.93
2.00
32.26
144
5.61
0.258**
Summer Flounder
4196
3.91
2.66
16.46
104
3.13
0.203**
Atlantic cod
679
6.43
4.00
54.39
115
2.85
0.210**
Scup
438
20.31
12.50
638.90
124
3.05
0.220**
All species
39,094
5.30
2.67
83.43
172
8.35
0.286**
Tercelro: The statistical properties of recreational catch data off the northeastern U.S. coast
657
Table 3
Descriptive statistics for MRFSS (Marine Recreational Fishery Statistics Survey) 1981, 1988, and 1996 northeast U.S. coast catch
per hour, including zero catches. Catch is given in numbers offish. CV is the coefficient of variation (%). D is the Kolmogorov test
statistic for normality. Test statistics significant at the 1% level (P<0.01) shown by (**), indicating rejection of the null hypothesis
that catch rates follow a normal distribution.
Species
No. of trips
Mean
Median
Variance
CV
Skew
D
1981
Bluefish
4615
0.77
0.00
5.57
305
12.21
0.371**
Summer Flounder
3135
0.36
0.00
0.64
221
6.43
0.325**
Atlantic cod
509
0.38
0.17
0.43
172
4.04
0.280**
Scup
269
1.59
0.44
7.96
178
2.79
0.287**
All species
20,280
0.74
0.00
13.23
491
40.11
0.419**
1988
Bluefish
7294
0.39
0.00
1.26
287
7.26
0.363**
Summer Flounder
4779
0.45
0.00
0.82
200
6.11
0.309**
Atlantic cod
1558
0.96
0.50
1.99
147
3.32
0.249**
Scup
960
2.12
0.67
14.03
177
3.60
0.286**
All species
48,423
0.54
0.00
3.07
325
15.18
0.379**
1996
Bluefish
5457
0.34
0.00
1.23
322
7.73
0.378**
Summer Flounder
7047
0.52
0.22
0.729
164
4.17
0.271**
Atlantic cod
1099
0.86
0.28
1.99
165
3.12
0.272**
Scup
643
3.06
1.17
27.78
172
3.79
0.281**
All species
81,057
0.62
0.00
4.51
341
35.84
0.385**
Table 4
Descriptive statistics for MRFSS (Marine Recreational Fishery Statistics Survey) 1981, 1988, and 1996 northeast U.S. coast catch
per hour, positive catches only. Catch is given in numbers offish. CV is the coefficient of variation ('7f ). D is the Kolmogorov test
statistic for normality. Test statistics significant at the 1% level (P<0.01) shown by (**), indicating rejection of the null hypothesis
that catch rates follow a normal distribution.
Species
No. of trips
Mean
Median
Variance
CV
Skew
D
1981
Bluefish
2288
1.56
0.75
10.00
203
9.48
0.317**
Summer Flounder
1380
0.82
0.50
1.07
126
5.35
0.229**
Atlantic cod
298
0.65
0.45
0.56
115
3.64
0.220**
Scup
165
2.56
1.67
10.41
125
2.16
0.242**
All species
9484
1.58
0.67
26.96
328
29.00
0.381**
1988
Bluefish
2445
1.16
0.63
2.85
145
4.91
0.253**
Summer Flounder
2326
0.93
0.60
1.24
120
5.45
0.212**
Atlantic cod
1065
1.40
1.00
2.29
108
3.14
0.196**
Scup
614
3.31
1.71
17.98
128
3.09
0.223**
All species
19,094
1.37
0.67
6.67
188
11.08
0.301**
1996
Bluefish
1666
1.13
0.55
3.15
157
4.80
0.270**
Summer Flounder
4196
0.87
0.58
0.90
109
3.93
0.198**
Atlantic cod
679
1.39
0.80
2.49
114
2.68
0.205**
Scup
438
4.49
2.73
34.38
131
3.38
0.228**
All species
39,094
1.29
0.63
8.49
226
28.09
0.331**
658
Fishery Bulletin 101(3)
Table 5
Summary of goodness
of fit tests for 1996 MRFSS (Marine Recreational Fishery Statistics Si
jrvey) catch
per trip distributions, |
including zero catches
for bluefish, summer
flounder, Atlantic cod, scup, and all
species.
Expected number
Degrees of
X-'
G
D
Species
of intervals
freedom
statistic
statistic
X 0 01
statistic
^0 01
Bluefish
Mean = 1.20
Variance = 13.12
n = 5457
Lognormal
9
6
7314
5722
17
0.462
0.014
Poisson
7
5
5315
4251
15
0.394
0.014
Negative binomial
23
20
68
27
38
0.007
0.014
Summer flounder
Mean = 2.33
Variance = 13.49
n = 7047
Lognormal
11
8
8654
4714
20
0.281
0.012
Poisson
10
8
10,902
5772
20
0.307
0.012
Negative binomial
22
19
139
101
36
0.011
0.012
Atlantic cod
Mean = 3.97
Variance = 43.34
n = 1099
Lognormal
12
9
2138
1068
22
0.360
0.031
Poisson
12
10
8284
2212
23
0.425
0.031
Negative binomial
25
22
48
22
40
0.015
0.031
Scup
Mean = 13.83
Variance = 524.60
n =643
Lognormal
28
25
389,173
3850
44
0.541
0.041
Poisson
25
23
6.67e+07
6391
42
0.544
0.041
Negative bionomial
51
48
305
235
74
0.053
0.041
All species
Mean = 2.57
Variance = 47.32
n = 81,057
Lognormal
14
11
180,754
83,230
25
0.382
0.004
Poisson
13
11
306,000
129,928
25
0.440
0.004
Negative binomial
51
48
1577
1146
74
0.020
0.004
chrysops, a Mid-Atlantic demersal schooling roundfish,
likely to yield a relatively high catch per trip). These spe-
cies were selected as examples because they occur over a
broad range along the northeast U.S. coast, are among the
most frequently caught by recreational fishermen, and
their catch-rate distributions are representative of most
species caught by recreational fishermen in the northeast
US (USDOC, 1992). Four configurations of catch rate dis-
tributions were examined: 1 ) catch per trip distributions
including zero catches, 2) catch per trip distributions with
positive catches only, 3 ) catch per hour distributions includ-
ing zero catches, and 4) catch per hour distributions with
positive catches only.
Goodness-of-fit statistics for the lognormal, Poisson,
and negative binomial distributions were calculated for
the four individual species and for all species to help judge
which error structure best characterized the MRFSS
catch-rate data. A single year ( 1996) is presented because
of the similarity of the catch distributions across species
and time. Given the results of the Kolmogorov-Smirnov
D tests from the descriptive statistics work, which in-
dicated that none of the catch rates were normally dis-
tributed (see "Results" section), that error structure was
not examined further. As with the descriptive statistics
analysis, both catch-per-trip and catch-per-hour rates
were examined in the goodness-of-fit exercise, both for
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
659
Table 6
Summary of goodness
-of-fit tests for 1996 MRFSS (Marine Recreational Fishery Statistics Survey) catch
per trip distributions, |
positive catches only, for bluefish, summer flounder, Atlantic
cod, scup, and all
species.
Expected number
Degrees of
X^
G
D
Species
of intervals
freedom
statistic
statistic
y2
A 0-01
statistic
£'ooi
Bluefish
Mean = 3.93
Variance = 32.26
n = 1666
Lognormal
9
6
1803
1026
17
0.312
0.025
Poisson
11
9
2091
1211
22
0.312
0.025
Negative binomial
21
18
425
347
35
0.196
0.025
Summer flounder
Mean = 3.91
Variance = 16.46
n = 4196
Lognormal
10
7
6068
2863
18
0.270
0.016
Poisson
12
10
3821
2234
23
0.240
0.016
Negative binomial
20
17
699
587
33
0.143
0.016
Atlantic cod
Mean = 6.43
Variance = 54.39
n = 679
Lognormal
12
9
3376
962
22
0.379
0.040
Poisson
14
12
3419
925
26
0.365
0.040
Negative binomial
27
24
121
88
43
0.147
0.040
Scup
Mean = 20.31
Variance = 638.90
n = 438
Lognormal
30
27
3.74e+ll
6565
47
0.543
0.049
Poisson
32
30
8.09e+07
3477
51
0.475
0.049
Negative binomial
50
47
204
147
72
0.089
0.049
All species
Mean = 5.30
Variance = 83.43
n = 39,094
Lognormal
11
8
70,234
24,957
20
0.254
0.001
Poisson
16
14
169,662
59,516
29
0.391
0.001
Negative binomial
50
47
12,293
10,217
72
0.201
0.001
all directed trips including zero catches and for positive
catches only.
Simulated recreational fishery catch rates
To isolate the consequences of possible model misspecifica-
tion in deriving standardized indices of abundance, nega-
tive binomial distributions with characteristics like those
of MRFSS recreational catch-per-trip distributions were
simulated by using the SAS RANTBL function (SAS, 2000).
The simulated distributions were arranged to provide con-
tinuously decreasing, continuously increasing, and peaked
(increasing to a peak and then decreasing) trends in an
11-year time series of catch per trip. For the decreasing
trend, the simulation procedure began with year 1 set at
a mean catch per trip = 3.0, maximum catch per trip of 50
fish per trip, and variance = 81.0, which are characteristic
of the MRFSS catch-per-trip distributions for all species
(Table 1). For year 1, this combination of mean and vari-
ance provided a maximum likelihood estimate of the nega-
tive binomial dispersion parameter, k, of 0.23.
The vector of expected probabilities of catch per trip
for these initial moments, assuming a negative binomial
distribution, was then used to randomly generate 1000
660
Fishery Bulletin 101 (3)
MRFSS 1996 samples
5OO0
■ Bluefish
2500
1
0
l-._
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
llAM^
Summer
flounder
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
500 1
3k
■ Atlantic
C
cr
250
I cod
III.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Scup
ll«-i — f- -
-m- — — »
0
10 20
50
0000 -1
All
5000
species
0
!■■..
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Catch number per trip (Including zero catches)
Figure 1
Marine Recreational Fishery Statistics Survey (MRFSS)
1996 sample data for bluefish, summer flounder, Atlantic
cod, scup, and all species, Maine to the Florida east coast:
catch number per trip (including zero catches). The 15 and
50 fish intervals are "plus groups'" because they include
totals for larger intervals.
Il
MRFSS 1996 samples
Bluefish
1
lu.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Summer
flounder
0 12 3 4 5 6 7
ll
9 10 11 12 13 14 15
Atlantic
cod
1«1
U^
^A
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Scup
llll.li.i-l... I.-I.l ^
10
20
50
j
b
All
species
1^
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Catch number per trip (positive catches only)
Figure 2
Marine Recreational Fishery Statistics Survey (MRFSS)
1996 sample data for bluefish, summer flounder, Atlantic
cod, scup, and all species, Maine to the Florida east coast:
catch number per trip ( positive catches only). The 15 and 50
fish intervals are "plus groups" because they include totals
for larger intervals.
observations of catch per trip (including zeroes) for year
1 (n = 1000). The initial mean for year 2 was then set at 10
percent less than year 1 (i.e. 2.7) and the year 2 set of 1000
observations generated under the negative binomial as-
sumption. The dispersion parameter, k, was held constant
at the year 1 maximum likelihood estimate of 0.23, result-
ing in a decrease in variance, a relatively stable coefficient
of variation (CV), and less frequent occurrence of large
catch-per-trip values, as the mean decreased. These condi-
tions were felt to best reflect the true changes in angler
catch per trip as stock abundance declines. The exercise
was repeated for years 3 to 11, providing a time series of
decreasing simulated recreational fishery catch per trip.
The simulated annual distributions, scaled (normalized)
to the 1 1-year time series mean of 1.75, were re-ordered to
create the increasing and peaked time series.
Standardized indices of abundance were then calculated
from the simulated, trended series by using lognormal, Pois-
son, negative binomial, delta-lognormal, and delta-Poisson
models, with year serving as the single classification vari-
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
661
MRFSS 1996 samples
LL
Bluefish
0123456789
LiL
Summer
flounder
0123456789
Atlantic
cod
0123456789
ilL
All
species
0123456789
Catch number per hour (including zero catches)
Figure 3
Marine Recreational Fishery Statistics Survey (MRFSS)
1996 sample data for bluefish, summer flounder, Atlantic
cod, scup, and all species, Maine to the Florida east coast:
catch number per hour (including zero catches). The 9 fish
interval is a "plus group" because it includes totals for larger
intervals.
0 1
MRFSS 1996 samples
LL
Bluefish
^ F- : . ^
0123456789
Ll
Summer
flounder
0123456789
LL
Atlantic
cod
0123456789
Scup
5 6
Ll
8 9
All
species
0123456789
Catch number per hour (positive catches only)
Figure 4
Marine Recreational Fishery Statistics Survey (MRFSS)
1996 sample data for bluefish, summer flounder, Atlantic
cod, scup, and all species, Maine to the Florida east coast:
catch number per hour (positive catches only). The 9 flsh
interval is a "plus group" because it includes totals for
larger intervals.
able and index of abundance. Modeled in this way, the
negative binomial model is expected to provide year-effect
coefficients very close in absolute value to the unstandard-
ized, mean simulated catch per trip of the true underlying
negative binomial distribution because no other classifica-
tion effects are present to account for variance from the
unstandardized mean. The deviance of the year coefficients
provided by the models, assuming the other error distribu-
tions, then provides an indication of the degree of model
misspecification because virtually all the estimated vari-
ance in this particular exercise is due to model (process)
error, except for the small amount generated by the random
draw from the starting probability distributions.
MRFSS standardized indices of abundance, 1981-98
The potential effect of the assumed error structure on the
calculation of standardized indices of abundance was fur-
ther explored with empirical examples using the 1981-98
MRFSS time series of catch-per-trip rates (zero catches
662
Fishery Bulletin 101(3)
MRFSS 1996 samples
Bluefish
Observed
Lognormal
Poisson
Neg binomial
t -»m , , .
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Catch number per trip (Including zero catches)
Rgure 5
Observed and expected catch number per trip (including
zero catches) frequency distributions for bluefish, summer
flounder, Atlantic cod, scup, and all species, Maine to the
Florida east coast. The 15 and 50 fish intervals are "plus
groups" because they include totals for larger intervals.
0 1 2 3 4 5 6 7 e 9 10 11 12 13 14 15
Catch number per trip (positive catches only)
Figure 6
Observed and expected catch number per trip (positive
catches only) frequency distributions for bluefish, summer
flounder, Atlantic cod, scup, and all species, Maine to the
Florida east coast. The 15 and 50 fish intervals are "plus
groups" because they include totals for larger intervals.
included) for bluefish, summer flounder, Atlantic cod, scup,
and for all species. Annual indices of stock abundance were
developed from these MRFSS catch rate data following pro-
cedures in previous Atlantic coast bluefish and summer
flounder stock assessments (Terceiro'; NEFSC'; Gibson and
Lazar'). Standardized indices were calculated by applying
lognormal, Poisson, negative binomial, delta-lognormal,
and delta-Poisson models, using the main effects classifica-
tion variables determined in these stock assessments to be
statistically significant factors: year, fishing mode (shore,
private or rental boat, party or charter boat), state of land-
ing (Maine to Florida), fishing wave (two-month sampling
period, e.g. Jan-Feb), fishing area (>3 miles from shore,
s 3 miles from shore), and daysl2, the angler-reported days
of saltwater fishing during the previous 12 months ( a proxy
for angler avidity, experience, or skill, or a proxy for all
three characteristics). The retransformed, bias-corrected
(when necessary) year coefficients serve as the annual
indices of stock abundance. Calculation and evaluation of
the MRFSS standardized indices followed the general pro-
cedures described in the "Overview of statistical methods"
in the "Materials and methods" section.
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
663
Table 7
Summary of goodness-of-fit tests for 1996 MRFSS (Marine Recreational Fishery Statistics Survey) catch
per hour distributions,
including zero catches,
for bluefish, summer
flounder, Atlantic cod, scup, and all
species.
Expected number
Degrees of
x'
G
D
Species
of intervals
freedom
statistic
statistic
„2
X 001
statistic
^001
Bluefish
Mean = 0.35
Variance = 1.23
n = 5457
Lognormal
6
3
2806
3047
11
0.323
0.014
Poisson
5
3
514
41
11
0.028
0.014
Negative binomial
5
2
514
41
9
0.028
0.014
Summer flounder
Mean = 0.52
Variance = 0.72
n = 7047
Lognormal
7
4
2022
2430
13
0.245
0.012
Poisson
5
3
1408
1209
11
0.193
0.012
Negative binomial
5
2
1408
1209
9
0.193
0.012
Atlfuitie cod
Mean = 0.86
Variance = 2.00
n = 1099
Lognormal
7
4
289
300
13
0.243
0.031
Poisson
5
3
51
11
11
0.051
0.031
Negative binomial
5
2
51
11
9
0.051
0.031
Scup
Mean = 3.06
Variance = 27.78
n = 643
Lognormal
11
8
546
346
20
0.312
0.041
Poisson
10
8
1209
583
20
0.347
0.041
Negative binomial
19
16
54
39
32
0.032
0.041
All species
Mean = 0.62
Variance = 4.51
n = 81,057
Lognormal
13
10
144,556
126,529
23
0.575
0.004
Poisson
7
5
54,675
72,657
15
0.036
0.004
Negative binomial
7
4
54,675
72,657
13
0.036
0.004
Results
Descriptive statistics for MRFSS catch rates
Descriptive statistics of MRFSS catch rates for the four
catch rate configurations, four individual species, and for
all species are presented for the years 1981, 1988, and 1996
(Tables 1-4). These three years are characteristic of the
1981-2002 time series of MRFSS data. Given the similarity
among these years, frequency distributions are plotted only
for 1996 (Figs. 1-4). Catch rate means, both with and with-
out zero catches, are generally much higher than medisms,
variances are much larger than the means, skewness is
always much larger than zero, and there is a high propor-
tion of zero catch and one-fish catch-rate observations. In
all cases, the Kolmogorov-Smimov D test statistics were
significant at the 1% level. All of these factors indicate that
MRFSS catch-rate distributions are highly contagious and
overdispersed in relation to the normal distribution (Sokol
and Rohlf, 1981). Scup has highest frequency of high catch
rates (Figs. 1-4). The scup and Atlantic cod samples exhibit
modes at regular intervals of high catch-per trip rates (e.g.
10, 15, 20, 25, and 30 fish per trip) that may indicate some
degree of digit bias in the sampling.
664
Fishery Bulletin 101(3)
Table 8
Summary of goodness
of fit tests for 1996 MRFSS (Marine
Recreational Fishery Statistics Survey) catch
per hour distributions,
positive catches only, for bluefish, summer flounder, Atlantic cod
scup, and
all species.
Expected number
Degrees of
x''
G
D
Species
of intervals
freedom
statistic
statistic
X 001
statistic
^0.01
Bluefish
Mean = 1.13
Variance = 3.15
n = 1666
Lognormal
6
3
1508
1462
11
0.446
0.025
Poisson
5
3
590
590
11
0.269
0.025
Negative binomial
5
2
590
590
9
0.269
0.025
Summer flounder
Mean = 0.87
Variance = 0.90
n = 4196
Lognormal
6
3
3005
3146
11
0.416
0.016
Poisson
5
3
838
921
11
0.210
0.016
Negative binomial
5
2
838
921
9
0.210
0.016
Atlantic cod
Mean = 1.39
Variance = 2.49
n = 679
Lognormal
6
3
414
368
11
0.356
0.040
Poisson
5
3
145
113
11
0.213
0.040
Negative binomial
5
2
145
113
9
0.213
0.040
Scup
Mean = 4.49
Variance = 34.38
n = 438
Lognormal
10
7
601
292
18
0.270
0.049
Poisson
11
9
725
324
22
0.280
0.049
Negative binomial
17
14
127
99
29
0.166
0.049
All species
Mean = 1.29
Variance = 8.49
n = 39,094
Lognormal
7
4
38,171
33,641
13
0.434
0.001
Poisson
8
6
31,475
16,429
17
0.270
0.001
Negative binomial
8
5
31,475
16,429
15
0.270
0.001
For the catch-per-trip configurations, catch rates were best
characterized by the negative binomial di.stribution (Table.s
5-6, Figs. 5-6). Note that the calculated ehi-square, G-. and
D-test statistics were generally significant at the 1% level, so
that based on strict interpretation of these results, the null
hypothesis that the observed distributions come from one of
the theoretical distributions was rejected in all cases. How-
ever, the calculated test statistics for the negative binomial
distributions were at least an order of magnitude smaller
than those for the Poisson and lognormal distributions, sug-
gesting that an underlying negative binomial distribution
was much more likely. The distributions of the catch-per-
hour rates generally had a truncated range compared to
the catch-per-trip rate configurations (Figs. 1-4). For most
of the catch-per-hour distributions, the maximum likelihood
solution for the negative binomial k parameter occurred at
very large values (>1000). The expected frequencies for the
negative binomial distribution therefore converged to those
expected for a Poisson distribution, resulting in identical test
statistic values and indicating that the catch-per-hour rates
are best characterized by the Poisson distribution (Tables
7-8, Figs. 7-8).
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
665
MRFSS 1996 samples
Bluefish
Observed
Lognormal
Poisson
Neg. binomial
Summer
Observed
flounder
--a- Lognoimal
— •* — Poisson
■ Neg. binomial
Observed
— a — Lognormal
— ■• — Poisson
— •— Neg, binomial
12 3 4 5 6 7
Catch number per hour (including zero catches)
Figure 7
Observed and expected catch number per hour (including
zero catches) frequency distributions for bluefish, summer
flounder, Atlantic cod, scup, and all species, Maine to the
Florida east coast. The 9 fish intervals is a "plus group"
because it includes totals for larger intervals.
MRFSS 1996 samples
Bluefish
Observed
- ^ — Lognofmal
- -• — Potsson
— ■ — Meg binomial
Scup
Observed
- -A — Lognormal
- -• — Poisson
-•— Neg. binomial
— Obsen/ed
— Lognormal
— Poisson
Neg. binomial
0123456789
Catch number per hour (positive catches only)
Figure 8
Observed and expected catch number per hour (positive
catches only) frequency distributions for bluefish, summer
flounder, Atlantic cod, scup, and all species, Maine to the
Florida east coast. The 9 fish intervals is a "plus group"
because it includes totals for larger intervals.
Simulated recreational fishery catch rates
The eleven simulated distributions of catch per trip had
means ranging from 2.80 to 0.98 fish per trip, variances
ranging from 31.81 to 4.39, and CVs of about 200%. Simu-
lated variance decreased as the simulated mean decreased
because the negative binomial dispersion parameter, k, was
held constant at 0.23. The resulting unstandardized, simu-
lated index of abundance declined by 65'7( over the 11 year
series (Table 9).
All standardization model fits were highly significant
(P<0.001), as characterized by the chi-square statistics for
the year effect (Table 10). The three different time series
trends had no effect on the results, and therefore only the
results for the decreasing series are reported. The Poisson
and negative binomial models generated year coefificients
as standardized indices of abundance that were very simi-
lar to each other and, as expected, virtually identical to the
unstandardized annual means, indicating a 65% dechne
over the time series (Fig. 9). Interestingly, the diagnostic
666
Fishery Bulletin 101(3)
Table 9
Summary statistics for the simulated
recreational fishery catch per trip assuming a
negative binomial distribution, configured to
decline by ID'S
in successive time periods (years). For year 1
, starting
maximum catch per trip was 50 fish pe
r trip, mean was 3.0,
variance was 81.00, coefficient of variation (CV) of 300%, and the dispersion parameter of the negative binomial distribution, k. was |
0.23. In years
2-11, k was held constant at the year-1 value
of 0.23, a
llowing the variance to decrease as the
mean catch declined.
Annual simulated means were scaled to the 11 year time series mean
(1.75) for comparability with standardized indices calculated |
for decreasing
increasing, and peaked
time series trends.
Simulated
Simulated
Scaled
mean catch
maximum catch
Simulated
Simulated
simulated catch
Year
per trip
per trip
variance
CV (%)
per trip
1
2.80
47
31.81
201
1.60
2
2.49
39
24.75
200
1.42
3
2.27
37
20.97
202
1.30
4
2.05
34
17.23
203
1.17
5
1.85
31
14.24
204
1.06
6
1.67
29
11.84
206
0.95
7
1.49
26
9.65
209
0.85
8
1.32
21
7.36
206
0.75
9
1.21
21
6.47
211
0.69
10
1.09
19
5.38
213
0.62
11
0.98
17
4.39
215
0.56
£ 2.0
1.5
Simulated catch per trip
scales to series means
123456789 10 11
1.0
0.5
12 3 4 5 6 7
Year
9 10 11
Figure 9
Simulated recreational fishery indices of abundance mod-
eled under different error distribution assumptions.
statistics indicated a better determined year effect (more
precise year coefficients) for the Poisson than for the nega-
tive binomial. However, the dispersion estimate (deviance/
df) for the Poisson model was much greater than 1.0, indi-
cating that the input data were overdispersed with respect
to the Poisson distribution (Table 10). The latter was the
expected result, given that the variance of the annual
simulated data sets was much larger than the mean. The
results indicated that the negative binomial was a more
appropriate model, with a dispersion estimate closer to 1.0,
which was also the expected result given the true negative
binomial distribution of the simulated data (SAS, 2000).
The consequence of assuming a lognormal model for the
true underlying negative binomial distribution was a more
extreme smoothing of the true time series trends than with
the other model assumptions, with a decline of only 28%
over the time series (Fig. 9). The diagnostic statistics for the
lognormal model indicated a significant model fit, but the
dispersion estimate was much less than 1.0, indicating that
the input data were underdispersed with respect to the
lognormal distribution (Table 10). This finding is reflective
of the large number of 0 and 1 catch-per-trip observations,
and a lack of observations near the mean of the input prob-
ability distribution (SAS, 2000). In this simulation exercise,
therefore, the lognormal model dispersion estimate of much
less than 1.0 is indicative of model misspecification.
As noted in the "Materials and methods" section, the in-
dices of abundance from the delta models are calculated as
the product of the year-effect coefficients from the two com-
ponent models. The interaction of the year coefficients from
the binomial proportion positive catches and lognormal or
Poisson positive catches components of the delta models
Terceiro The statistical properties of recreational catch data off the northeastern U.S. coast
667
Table 10
Summary of model fits for simulated recreational fishery
catch per trip (including zero catches) with a decreasing
time series trend. Total model degrees of freedom were
10,989; for the positive catches component of the delta
models, degrees of freedom were 4,184. Year-model-effect
degrees of freedom were 10, and the year-model effect was
highly significant (P<0.0001) in all five models.
Criterion
Value
Dispersion
estimate
(value/df)
Lognormal model
Deviance
7330
0.6670
Log-likelihood
-13,773
Year chi-square
183
Poisson model
Deviance
51,719
4.7064
Log-likelihood
-7483
Year chi-square
2049
Negative binomial model
Deviance
10,699
0.9736
Log-hkelihood
Year chi-square
7524
239
Delta models: binomial proportion positive catch
Deviance 14,546 1.3237
Log-likelihood
Year chi-square
-7273
78
Delta-lognormal model: lognormal positive catches
Deviance 3474 0.8303
Log-likelihood
-5557
Year chi-square
119
Delta-Poisson model: Poisson positive catches
Deviance 15,822
3.7815
Log-likelihood
10,466
Year chi-square
936
Table 11
Summary of model fits for estimating indices of abundance
from empirical MRFSS (Marine Recreational Fishery
Statistics Survey) bluefish catch per trip (including zero
catches), 1981-98. Total model degrees of freedom (df) were
130,300; for the positive catches component of the delta
models, degrees of freedom were 48,447. All model fits and
classification effects were highly significant (P<0.001).
Dispersion
estimate
Criterion
Value
(value/df)
Lognormal model
Deviance
84,150
0.6458
Log-likelihood
-156,444
Year chi-square
1835
Poisson model
Deviance
675,791
5.1864
Log-likelihood
-19,680
Year chi-square
20,604
Negative binomial model
Deviance
99,393
0.7628
Log-likelihood
190,140
Year chi-square
2104
Delta models: binomial proportion positive catch
Deviance 157,674 1.2101
Log-likelihood -78,837
Year chi-square 854
Delta-lognormal model: lognormal positive catches
Deviance 39,963 0.8249
Log-likelihood -64,129
Yera chi-square 1240
Delta-Poisson model: Poisson positive catches
Deviance 249,112 5.1419
Log-likelihood 193,660
Year chi-square 10,501
provided some interesting results in this simulation exer-
cise. The binomial model component, common to both delta
models, provided a highly significant year effect and indi-
cated a 41% decline in abundance over the time series. The
dispersion estimate indicated some overdispersion of the
data with respect to the binomial distribution (Table 10).
The lognormal positive catches component of the delta-
lognormal model also provided a highly significant year
effect and indicated a 39% decline in abundance over
the time series, producing a smoothing effect similar to
that observed for the lognormal model of catch per trip
including zeroes. The dispersion estimate indicated some
underdispersion of the data with respect to the lognormal
distribution (Table 10). The product of the annual year coef-
ficients from the two delta-lognormal model components,
which individually indicated less decline than the unstan-
dardized indices, provided final indices of abundance that
declined 64% over the time series (due to the product of two
positive fractional values <1 providing a even smaller value
<1) — nearly identical to the unstandardized, Poisson, and
negative binomial series (Fig. 9).
The Poisson positive catches component of the delta-
Poisson model provided a highly significant year effect and
indicated a 51% decline in abundance over the time series.
The dispersion estimate was much greater than 1.0, indicat-
ing overdispersion of the data with respect to the Poisson
model (Table 10). The product of the annual year coefficients
from the two delta-Poisson model components provided in-
dices of abimdance that dechned 71% over the time series,
a slightly greater decrease than for the other models (Fig.
9). Note again that the delta-lognormal and delta-Poisson
models share the same binomial proportion positive catch
model components, and therefore annual year coefficients
for this component. The decrease estimated by the delta-
Poisson model was greater than that for the delta-lognor-
mal because the year coefficients from the Poisson positive
catch model were all smaller, and more closely matching the
unstandardized positive catch series, than the comparable
668
Fishery Bulletin 101(3)
Table 12
Summary of model fits for estimating indices of abundance
from empirical MRFSS (Marine Recreational Fishery Sta-
tistics Survey) summer flounder catch per trip (including
zero catches), 1981-98. Total model degrees of freedom
(df) were 102,162; for the positive catches component
of the delta models, degrees of freedom were 52,507. All
model fits and classification effects were highly significant
(P<0.001).
Dispersion
estimate
Criterion
Value
(value/df)
Lognormal model
Deviance
66,452
0.6505
Log-likelihood
-122,989
Year chi-square
2663
Poisson model
Deviance
444,657
4.3525
Log-likelihood
-14,827
Year chi-square
14,053
Negative binomial model
Deviance
96,698
0.9465
Log-likelihood
97,777
Year chi-square
2560
Delta models: binomisil proportion positive catch
Deviance 130,341 1.2758
Log-likelihood -65,171
Year chi-square 2498
Delta-lognormal model: lognormal positive catches
Deviance 36,780 0.7005
Log-likelihood -65,202
Year chi-square 1203
Delta-Poisson model: Poisson positive catches
Deviance 183,019 3.4856
Log-likelihood 115,991
Year chi-square 5675
Table 13
Summary of model fits for estimating indices of abundance
from empirical MRFSS (Marine Recreational Fishery
Statistics Survey) Atlantic cod catch per trip (including
zero catches), 1981-98. Total model degrees of freedom
(df) were 20,629; for the positive catches component of
the delta models, degrees of freedom were 13,160. All
model fits and classification effects were highly significant
(P<0.001).
Dispersion
estimate
Criterion
Value
(value/df)
Lognormal model
Deviance
19,425
0.9416
Log-likelihood
-28,697
Year chi-square
380
Poisson model
Deviance
142,834
6.9239
Log-likelihood
54,501
Year chi-square
4090
Negative binomial model
Deviance
21,824
1.0579
Log-likelihood
98,335
Year chi-square
323
Delta models: binomial proportion positive catch
Deviance 24,997 1.2117
Log-likelihood -78,837
Year chi-square 191
Delta-lognormal model: lognormal positive catches
Deviance 11,657 0.8858
Log-likelihood -17,920
Year chi-square 353
Delta-Poisson model; Poisson positive catches
Deviance 75,359 5.7264
Log-likehhood 88,239
Year chi-square 2805
lognormal positive catch year coefficients over the course
of the time series. For example, the year- 11 coefficient from
the binomial proportion positive catches model was 0.59;
the year-11 lognormal positive catches coefficient was 0.61,
providing a product for the year-11 index of 0.36. In con-
trast, the year-11 Poisson positive catches coefficient was
0.49, providing a product for the year-1 1 index of 0.29. When
these and the other annual coefficients were scaled to the
respective series means, the delta-Poisson model indicated
a slightly greater decline over the time series.
MRFSS standardized indices of abundance,
1981-98
All standardization models of the MRFSS catch per trip
(including zero catches), for the four individual species
and for all species, fitted well. In part because of the large
number of observations, the overall model fits and the indi-
vidual classification effects (year, mode, state, wave, and
days 12 ) were all highly significant. Only the year effect chi-
square statistics are tabulated because the year effect coef-
ficients serve as the annual indices of abundance (Tables
11-15). The year effect was generally the second or third
most important effect in the models, after mode and state.
The dispersion estimates (deviance/df) for the lognormal
models indicated the data were generally underdispersed
with respect to the lognormal; the dispersion estimates for
the Poisson models indicated overdispersion with respect
to that distribution. The dispersion estimates for the nega-
tive binomial models and binomial components of the delta
models were generally close to 1.0, indicating appropriate
model specification (Tables 11-15).
As in the simulated catch-rate exercise, the lognormal
standardized abundance indices generally show lower
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
669
Table 14
Summai'v of model fits for estimating indices of abundance
from empirical MRFSS (Marine Recreational Fishery Sta-
tistics Survey) scup catch per trip (including zero catches),
1981-98. Total model degrees of freedom (df) were 17,604;
for the positive catches component of the delta models,
degrees of freedom were 11,124. All model fits and classifi-
cation effects were highly significant (P<0.001).
Dispersion
estimate
Criterion
Value
(value/df)
Lognormal model
Deviance
32,270
1.8331
Log-likelihood
-30,346
Year chi-square
332
Poisson model
Deviance
375,924
21.3545
Log-likelihood
309,490
Year chi-square
12,094
Negative binomial model
Deviance
18,668
1.0604
Log-likelihood
466,529
Year chi-square
369
Delta models: binomial proportion positive catch
Deviance 22,027 1.2512
Log-likelihood -11,013
Year chi-square 174
Delta-lognormal model: lognormal positive catches
Deviance 14,340 1.2891
Log-likelihood -17,225
Year chi-square 350
Delta-Poisson model: Poisson positive catches
Deviance 212,250 19.0804
Log-likelihood 391,327
Year chi-square 8793
Table 15
Summary of model fits for estimating indices o
f abundance
from empirical MRFSS (Marine Recreational Fishery Sta-
tistics Survey catch per trip for all species (including zero
catches), 1981-98. Total model degrees of freedom (df) were
1,033,367; for the positive catches component
of the delta
models, degrees of freedom were 457,598. All model fits and
classification effects were highly significant (P<0.001).
Dispersion
estimate
Criterion Value
(value/df)
Lognormal model
Deviance 861,881
0.8341
Log-likelihood -1,372,576
Year chi-square 7246
Poisson model
Deviance 8,048,246
7.7884
Log-likelihood 277,042
Year chi-square 28,734
Negative binomial model
Deviance 870,357
0.8422
Log-likelihood 3,118,822
Year chi-square 2243
Delta models: binomial proportion positive catch
Deviance 1,351,532
1.3079
Log-likelihood -675,766
Year chi-square 11,867
Delta-lognormal model: lognormal positive catches
Deviance 466,644
1.0198
Log-likelihood 653,838
Year chi-square 655
Delta-Poisson model: Poisson positive catches
Deviance 3,773,909
8.2472
Log-likelihood 2,414,210
Year chi-square 10,785
rates of change in abundance than do the unstandardized,
Poisson, or negative binomial indices, with the CV of the
lognormal series about 25-50'7f of the CV of the unstan-
dardized indices (Figs. 10-14). In effect, the lognormal
standardization of MRFSS per trip catch rates had an
unintended (and undesirable) smoothing effect on the
independent annual indices abundance. The Poisson and
negative binomial models generally provided interpreta-
tions of the trend and annual changes in abundance very
similar to those of the unstandardized indices.
For bluefish, the delta-lognormal, and delta-Poisson
models provided time series of indices with about the
same variability and trend, but slightly different annual
changes, as those from the unstandardized, Poisson, and
negative binomial models. For summer flounder, Atlantic
cod, scup, and all species, the delta-lognormal and delta-
Poisson models provided time series of abundance indices
that were more variable, with slightly different trends and
annual changes, than the unstandardized, Poisson, and
negative binomial series. This last result is comparable to
that observed for the delta models used with the simulated
data and is therefore likely due in part to model misspecifi-
cation of the positive catch-per-trip component (recall that
catch per trip for these examples is best characterized
by the negative binomial distribution) and a comparable
interaction of the binomial, lognormal, and Poisson model
year coefficients.
Discussion
The frequency distributions of recreational fishery catch-
rate data as sampled by the MRFSS are highly skewed,
often with a significant proportion of zero catch observa-
tions. The present study indicates that MRFSS catch rates
generally are not normally or lognormally distributed
670
Fishery Bulletin 101 (3)
Bluefish catch per trip
indices scaled to series means
• Unstandardtzed
■ Lognormal
- Passon
- Neg BirwfTiial
82 84 86 88 90 92 94 96 98
86 88 90
Year
96 98
Rgure 10
Indices of bluefish abundance (catch number per trip
including zero catches) modeled under different error
distribution assumptions, 1981-98.
Atlantic cod catch per trip
indices scaled to series means
2.0
1.5
1.0
0.5
• Unstandardized
- Lognormal
-#~ Posson X
-•- Heq. Binomial ^V
flj
'
,
'
■
■
E
a
2.0
82 84 86
88 90
92
94
1
96
98
o
^^ Unstarwjardized
1
1.5
1.0
0.5
0.0
-V- Delta-Lognornial
-O- Delta-Poisson
^^'^
J
\l
V
/
82 84 86
88 90
Year
92
94
96
98
Figure 12
Indices of Atlantic cod abundance (catch number per
trip including zero catches) modeled under different
error distribution assumptions, 1981-98.
Summer flounder catch per trip
indices scaled to series means
2.0
15
J!-^ yA
ID
U
C
5
■»rA/'\^
H-
\..*-^
0.5
0.0
' V
-T- Lognormal
-•- Poisson
— •— Nog Biriomial
'
3
82 84 86 88 90
92 94 96 98
(0
O
X
2.0
''^ h /k
K /
10
0.5
0.0
/vy
^
'\^
-O- Della-Potsson
,
82 84 86 88 90
92 94 96 98
Year
Figure 11
Ind
ces of summer flounder abundance (catch number
per trip including zero catches) modeled under differ- |
ent
error distribution assumptions,
1981-98.
2.5 1
2.0
1.5
1.0
g 05
1 0.0
0}
c 2.0
1.5
1.0
0.5
0.0
Indices
includinj
distribut
Scup catch per trip
indices scaled to series means
^"^ Unstandardized
T Lognormal
-^- Poisson
-■- Neg- Binomial
\^^
82 84 86 88 90 92
94 96 98
gx2
•^^ UnstarKJardaed
-V- Delta-Logrxxmal
-O- Defta-Poisson
vA-^
^
82 84 86 88 90 92 94 96 9
Year
Figure 13
i( scup abundance (catch number per tri
; zero catches) modeled under different errc
ion assumptions, 1981-98.
3
P
r
Terceiro: The statistical properties of recreational catch data off the northeastern U.S. coast
671
All species catch per trip
indices scaled to series means
1.5
1.0
^HAv*"-**^^
0.5
■ Unslandafdlzed
-▼- Lognormal
-•- Poisson
-•- Neg. Binomial
u
c
■o
82 84 86 88 90 92 94 96 98
Index of abu
b by
h^^^^^A>^
05
0.0
b-^
^^ Unstandanllzed
-V- Della-Lognormal
-Q- Delta -Poisson
82 84 86 88 90 92 94 96 9
8
Year
Figure 14
Indices of all species abundance (catch number per trip
including zero catches) modeled under different error
distribution assumptions, 1981-98.
but usually best characterized by the Poisson or nega-
tive binomial distribution, depending on the manner in
which the catch rate is configured. This finding suggests
that standardization methods for MRFSS catch-rate data
where Poisson (in the case of per hour rates) or negative
binomial (for per trip rates) error structures are assumed
would usually be more appropriate than methods where
normal or lognormal error structures are assumed.
The modeling of both the simulated and empirical
MRFSS catch rates indicates that one may draw errone-
ous conclusions about stock trends by assuming the wrong
error distribution in procedures used to developed stan-
dardized indices of abundance. The results demonstrate
the importance of considering not only the overall model
fit and significance of classification effects, but also the pos-
sible effects of model misspecification, when determining
the most appropriate model construction. In particular, the
simulation exercise indicates that assuming a lognormal
model in the calculation of indices of abundance from
recreational fishery catch-per-trip data with a true under-
lying negative binomial distribution will provide indices
that will strongly underemphasize the true trends in the
indices, and therefore in stock abundance. This underesti-
mation applies equally to populations that may be declin-
ing or increasing faster than the lognormally standardized
indices might indicate.
The MRFSS catch-per-trip indices standardized with the
negative binomial model, which the descriptive statistics
and goodness-of-fit results suggest should be the appropri-
ate model, differ relatively little from the unstandardized
indices, indicating that the model effects accounted for a
low percentage of the variation in mean catch rate. The
classification categories recorded in the general MRFSS
sampling are broad, and even measures of angling avidity
such as "angler-reported days of saltwater fishing during
the previous 12 months" may not be adequate proxies for
the real factors (besides stock abundance) that account for
variation in recreational fishery mean catch rates. To make
standardization analysis of MRFSS catch rate data poten-
tially more useful, by accounting for a significantly larger
part of the unexplained variance and thus providing more
accurate indices of abundance, more information on the
characteristics of individual fishing trips may be needed.
Such information might include details on the type of
equipment used, the skills, experience, avidity, and identity
of the individual fishermen, and detailed temporal and spa-
tial information about fishing trips. In the future, collection
of detailed trip data for general recreational fisheries may
be best accomplished by the identification and sampling of
"test fleets" of known, individual fishermen.
Acknowledgments
I thank Vic Crecco of the Connecticut Department of Envi-
ronmental Protection, for raising questions about the best
way to calculate indices of abundance from recreational
fishery catch rate data during debates over the bluefish
assessments; Paul Rago of the Northeast Fisheries Science
Center, for numerous discussions about statistical distribu-
tions and tests; and two anonymous Fishery Bulletin ref-
erees, whose comments helped improve the quality of the
analyses and therefore the usefulness of the results.
Literature cited
Banneret, S. P., and C. B. Austin.
1983. Using frequency distributions of catch per unit effort
to measure fish-stock abundance. Trans. Am. Fish. Soc.
112:608-617.
Berry, D. A.
1987. Logarithmic transformations in ANOVA. Biometrics
43:439-456.
Bliss, C. I., and R. A. Fisher.
1953 . Fitting the negative binomial distribution to biological
data. Biometrics 9:177-200.
Bradu, D., and Y. Mundlak.
1970. Estimation in lognormal linear models. J. Am. Stat.
Assoc. 65:198-211.
Brown, C. A.
1999. Standardized catch rates for yellowfin tuna (Thun-
nus albacares) and bigeye tuna (Thunnus obesus) in the
Virginia-Massachusetts (US) rod and reel fishery. ICCAT
(International Commission for the Conservation of Tunas)
Col. Vol. Sci. Pap. Vol. 49(3):357-369.
2001. Standardized catch rates for yellowfin tuna (Thun-
nus albacares) in the Virginia-Massachusetts (U.S.) Rod
and reel fishery during 1986-1999. ICCAT (International
Commission for the Conservation of Tunas) Col. Vol. Sci.
Pap. Vol. 52:190-201.
672
Fishery Bulletin 101(3)
Brown, C. A., and J. A. Browder.
1994. Standardized catch rates of small bluefin tuna in the
Virginia-Rhode Island (U.S.) rod and reel fishery. ICCAT
(International Commission for the Conservation of Tunas)
Col. Vol. Sci. Pap. Vol. 32(2):248-254.
Brown, C. A., and C. E. Porch.
1997. A numerical evaluation of lognormal, delta-lognormal
and Poisson models for standardizing indices of abundance
from west Atlantic bluefin tuna catch per unit effort data (pre-
hrainary results ). ICCAT ( International Commission for the
Conservation of Tunas) Col. Vol. Sci. Pap. Vol. 46(2):233-236.
Brown, C. A., and S. C. Turner.
2001. Updated standardized catch rates of bluefin tuna,
Thunnus thynnus, from the rod and reel/handline fishery
off the northeast United States during 1980-1999. ICCAT
(International Commission for the Conservation of Tunas)
Col. Vol. Sci. Pap. Vol. 52:984-1006.
Finney, D. J.
1951. On the distribution of a variate whose logarithm is
normally distributed. Suppl. J. Stat. Soc. 7:155-161.
Gavaris, S.
1980. Use of a multiplicative model to estimate catch rate
and effort from commercial data. Can. J. Fish. Aquat. Sci.
37:2272-2275.
Gulland, J. A.
1956. On the fishing eifort in English demersal fisheries.
Fish. Investig. Ser. II Mar. Fish. G. B. Minist. Agric. Fish.
Food 20(5), 41 p.
Hilborn, R.
1985. Fleet dynamics and individual variation: why some
people catch more fish than others. Can. J. Fish. Aquat.
Sci. 42:2-13.
Jones, C. M., D. S. Robson, H. D. Lakkis, and J. Kressel.
1995. Properties of catch rates used in analysis of angler
surveys. Trans. Am. Fish. Soc. 124:911-928.
Kimura, D. K.
1981. Standardized measures of relative abundance based
on modeling log (c.p.u.e.), and their application to Pacific
ocean perch {Sebastes alutus). J. Cons. Int. Explon Mer
39:211-218.
Lawless, J. F.
1987. Negative binomial and mixed Poisson regression.
Can. J. Stat. 15(3):209-225.
Lo, N. C, L. D. Jacobson, and J. L. Squire.
1992. Indices of relative abundance from fish spotter data
based on delta-lognormal models. Can. J. Fish. Aquat. Sci.
49:2515-2526.
Manton, K. G., Woodbury, M. A., and E. Stallard.
1981. A variance components approach to categorical data
models with heterogenous cell populations: analysis of
spatial gradients in lung cancer mortality rates in North
Carolina counties. Biometrics 37:259-269.
McCullagh, P., and J. A. Nelder.
1989. Generalized linear models, 511 p. Chapman and
Hall, London.
NMFS (National Marine Fisheries Service).
1995. Status of the fishery resources off the northeastern
United States for 1994. U.S. Dep. Commer, NOAA Tech.
Memo. NMFS-NE-108, 140 p.
1996. Our living oceans. Report on the status of U.S. living
marine resources, 1995. U.S. Dep. Commer, NOAA Tech.
Memo. NMFS-F/SPO-19, 160 p.
O'Brien, L. S,, and R. K. Mayo.
1988. Sources of variation in catch per unit effort of yellow-
tail flounder, Limanda ferruginea (Storer), harvested off the
coast of New England. Fish. Bull. 86(1):91-108.
Ortiz, M., S. C. Turner, and C. A. Brown.
1999. Standardized catch rates for bluefin tuna, Thunnus
thynnus, from the rod and reel fishery off the northeast
United States from 1980-1997. ICCAT (International
Commission for the Conservation of Tunas) Col. Vol. Sci.
Pap. Vol. 49(2 ):254-269.
Pennington, M.
1983. Efficient estimators of abundance, for fish and plank-
ton surveys. Biometrics 39:281-286.
Power, J. H., and E. B. Moser.
1999. Linear model analysis of net catch data using the
negative binomial distribution. Can. J. Fish. Aquat. Sci.
56:191-200.
Robson, D. S.
1966. Estimation of the relative fishing power of individual
ships. Comm. N.W. Atl. Fish. Res. Bull. 3:5-14.
SAS Institute.
2000. SAS OnlineDoc, version 8. SAS Institute Inc., Cary
NC. http://www.sas.com/ts.
Searle, S. R.
1987. Linear models for unbalanced data, 536 p. John
Wiley and Sons, Inc., New York, NY.
Smith, B. D.
1999. A probabilistic analysis of decision-making about trip
duration by Strait of Georgia sport anglers. Can. J. Fish.
Aquat. Sci. 56:960-972.
Smith, S. J.
1990. Use of statistical models for the estimation of abun-
dance from groundfish trawl survey data. Can. J. Fish.
Aquat. Sci. 47:894-903.
1996. Analysis of data from bottom trawl surveys. NAFO
Scientific Council Studies 28:25-53.
Snedecor, G W, and W G Cochran.
1967. Statistical methods, 593 p. Iowa State Univ. Press,
Ames, lA.
Sokal, R. R., and F J. Rohlf
1981. Biometry, 859 p. WH. Freeman and Co., New York, NY.
Taylor, C. C.
1953. Nature of the variability in trawl catches. Fish. Bull.
54:145-166.
Turner, S. C, C. A. Brown, and H. Huang.
1997. Standardized catch rates of small bluefin tuna, Thun-
nus thynnus, from the U.S. rod and reel fishery off Vir-
ginia-Rhode Island in 1980-1995. ICCAT (International
Commission for the Conservation of Atlantic Tunas) Col.
Vol. Sci. Pap. Vol. 46(2):295-310.
USDOC (U.S. Department of Commerce).
1992. Marine recreational fishery statistics survey, At-
lantic and Gulf coasts, 1990-1991, 275 p. Current Fisher-
ies Statistics 9204.
2001. Marine recreational fishery statistics survey. U.S.
Dep. Commer, Washington, DC. http://www.st.nmfs.gov/
stl/recreational/index. (Accessed 30 January 2001.1
Williams, D. A.
1976. Improved likelihood ratio tests for complete contin-
gency tables. Biometrika. 63:33-37.
673
Abstract — Spatial variation in demo-
graphic parameters of the red throat
emperor (Lethrinus miniatus) was
examined among 12 coral reefs in
three geographic regions (Townsville,
Mackay, and Storm Cay) spanning
over 3° of latitude of the Great Bar-
rier Reef, Australia. Estimates of
demographic parameters were based
on age estimates from counts of annuli
in whole otoliths because there was no
significant difference in age estimates
between whole and sectioned otoliths.
There were significant regional differ-
ences in age structures, rates of somatic
and otolith growth, and total mortality.
The Townsville region was character-
ized by the greatest proportion of older
fish, the smallest maximum size, and
the lowest rates of otolith growth and
total mortality. In contrast the Mackay
region was characterized by the highest
proportion of younger fish, the largest
maximum size, and the highest rates
of otolith growth and total mortality.
Demographic parameters for the Storm
Cay region were intermediate between
the other two regions. Historic differ-
ences in fishing pressure and regional
differences in productivity are two
alternative hypotheses given to explain
the regional patterns in demographic
parameters. All demographic param-
eters were similar among the four reefs
within each region. Thus, subpopula-
tions with relatively homogeneous
demographic parameters occurred on
scales of reef clusters. Previous stud-
ies, by contrast, have found substantial
between-reef variation in demographic
parameters within regions. Thus spa-
tial variation in demographic param-
eters for L. miniatus may differ from
what is assumed typical for a coral-reef
fish metapopulation.
Scales of spatial variation in demography of a large
coral-reef fish— an exception to the typical model?
Ashley J. Williams
School of Marine Biology and Aquaculture
and
CRC Reef Research Centre
James Cook University
Townsville, Queensland. 4811, Australia
E-mail address, astiley williams(a'|cu edu au
Campbell R. Davies
Bruce D. Mapstone
CRC Reef Research Centre
James Cook University
Townsville, Queensland 4811, Australia
Garry R. Russ
School of Marine Biology and Aquaculture
James Cook University
Townsville, Queensland, 481 1, Australia
Manuscript approved for publication
17 December 2002 by Scientific Editor.
Manuscript received 4 April 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:673-683 (2003).
Estimates of demographic parameters,
such as growth and mortahty rates, are
fundamental to the understanding of a
species population dynamics and for
predicting responses of populations to
exploitation. Processes affecting popu-
lation dynamics operate at a number
of spatial and temporal scales (Levin,
1992) and can result in subpopulations
with distinct demographics. Differences
in demography between populations
may suggest geographic or reproduc-
tive isolation (or both) and as such
have been used in stock identification
for fisheries assessment and manage-
ment purposes (e.g. Begg et al., 1999).
Identifying the "unit stock" has been
the primary focus of studies of spatial
structure of harvested populations in
most fisheries. Knowledge of spatial
structure within a unit stock is impor-
tant for both fisheries management,
because potential yields may vary
spatially within a population (Caddy,
1975), and for conservation, in order to
maintain intraspecific diversity (Niel-
sen, 1998). Hence, it is important to
estimate demographic parameters over
a range of temporal and spatial scales
to determine the scale(s) at which the
parameters vary significantly (Caley et
al., 1996) and, therefore, to infer which
scales are of greatest importance for
assessment and management purposes
(Sale, 1998).
Most coral-reef fish exist as metapop-
ulations of sedentary adult populations
linked by pelagic larval dispersal (Sale,
1998). Consequently, adult populations
of reef fish are commonly spatially seg-
regated and may be exposed to different
environmental, biological, and ecologi-
cal processes, resulting in spatial dif-
ferences in demographic parameters
at a range of spatial scales. Relatively
few studies, however, have focussed on
spatial variation in demographic pa-
rameters of harvested species of coral-
reef fish. Those that have, have gener-
ally focussed on spatial scales within
individual reefs or among reefs within
a single region (e.g. Ferreira and Russ,
1995; Hart and Russ, 1996; Newman et
al., 1996). Comprehensive multiscale
approaches are rare (but see Adams et
al., 2000; Meekan et al., 2001).
The spatial structure of coral-reef
populations has generated considerable
interest in terms of the use of spatial
closures, or marine protected areas
(MPAs), as an effective tool for their
management (Roberts and Polunin,
674
Fishery Bulletin 101(3)
1991). However, the lack of information about the stock
structure of, and connectivity among, adult populations has
hindered MPA design (Walters and Bonfil, 1999). Conser-
vation management of the Great Barrier Reef (GBR) has
included the use of spatial closures of areas to activities,
including fishing, for more than 15 years. The majority of
spatial closures to line fishing are of individual coral reefs
or groups of reefs. This spatial management strategy is
underpinned by the assumption of the metapopulation
model of coral-reef fish described above. That is, closing
individual reefs to fishing will protect the adult popula-
tions on those reefs, and potentially provide a source of
larvae to areas open to fishing. Management of line fishing
on the GBR currently includes bag limits for recreational
fishermen and minimum-size restrictions that are uniform
for all fishermen and across the entire area of the fishery.
Such management regulations are based on the assump-
tion that the demography of target species does not vary
substantially over the species range and on the assump-
tion that that populations on the GBR represent a single,
homogeneous stock.
The red throat emperor (Lethrinus miniatus) (also
known as the trumpet emperor) is a relatively long-lived
(>20 years) (Loubens, 1980: Brown and Sumpton, 1998)
member of the Lethrinidae and has a restricted distribu-
tion in the western Pacific and eastern Indian Oceans (Car-
penter and Allen, 1989). On the GBR it is the second most
important demersal species in a multispecies line fishery,
contributing up to 1000 metric tons annually to the com-
bined commercial and recreational catch (Mapstone et al.^;
Higgs-). As with many tropical lethrinids, information on
the biology and ecology of L. miniatus is scarce. The limited
data available indicate that L. miniatus is usually associ-
ated with coral reefs, but that it is also commonly caught
in deeper water, in sand, and rubble areas between reefs
(Carpenter and Allen, 1989; Newman and Williams. 1996;
Williams and Russ'^). The habitat of juvenile L. miniatus
is unknown, but Williams and Russ^ have suggested that
juveniles may occupy the deeper rubble areas adjacent to
reefs. Like some other coral-reef fish, L. miniatus is thought
to form large aggregations associated with spawning
' Mapstone, B. D., J. P. McKinlay. and C. R. Davies. 1996. A
description of commercial reef line fishery logbook data held
by the Queensland Fisheries Management Authority. Report
to the Queensland Fisheries Management Authority from the
Cooperative Research Centre for the Ecologically Sustainable
Development of the Creat Barrier Reef, and the Department of
Tropical Environmental Studies and Geography. James Cook
University, Queensland. Australia, 480 p. lAvailable from
the Queensland Fisheries Service, G.P.O. Box 46, Brisbane.
Queensland, Australia 4001.1
- Higgs, J. 2001. Experimental recreational catch estimates
for (Queensland residents. Results from tbi' 1999 diary round.
KFISH technical report no. .'!. Queensland Fisheries Sei-vice, Aus-
tralia, 62 p. lAvailable from the Queensland Fisheries Service,
G.RO. Box 46, Bri.sbane, Queensland, Australia 4001.1
■' Williams, D. McB., and G. R. Russ. 1994. Review of data on
fishes of commercial and recreational fishing interest on the
Great Barrier Reef Report to the Great Barrier Reef Marine
Park Authority, 103 p. [Available from the Great Barrier Reef
Marine Park Authority, P.O. Box 1379, Townsville, Queensland,
Australia, 4810.1
(RusselH). These available data suggest that L. miniatus
adults have the capacity to move among individual reefs on
the GBR. This movement pattern contrasts with informa-
tion on movement patterns of other large coral-reef spe-
cies such as the coral trout (Plectropomus leopardus) (also
known as the leopard coral grouper, Heemstra and Randall,
1993) where adults show limited movement within a single
reef and very restricted movements between reefs (Davies,
1995). It also contrasts with movement patterns of the ma-
jority of coral-reef fish, where adults are known to have
very restricted home ranges and display little, if any, move-
ment between reefs (Lewis 1997; Sale, 1998). Therefore the
relevant spatial scale affecting demographic parameters of
L. miniatus may be larger than an individual reef and thus
is different from that for most "typical" coral-reef fish.
The central objective of this study was to determine how
the spatial patterns in demogi'aphy of large, more mobile
reef fish differ from smaller site-attached reef-fish species.
To achieve this we used validated age estimates to examine
spatial variation in demographic parameters of populations
of L. miniatus across two spatial scales most relevant to as-
sessing and managing the species on the GBR: 1 ) among
individual reefs within regions and, 2) among geographic
regions. Specifically, we estimated age structures, growth,
mortality, and otolith growth rates for among four reefs (all
closed to fishing) within each of three geographic regions
spanning over 500 km (over 3° of latitude) of the GBR.
Materials and methods
Collection methods
Samples of L. miniatus were collected from three geo-
graphic regions of the GBR as part of a large-scale manip-
ulative experiment to examine the effects of line fishing
on the GBR (Davies et al.^; Mapstone et al.'^). The three
regions cover most of the distribution of L. miniatus on
the GBR (Fig. 1), which is restricted to the southern 50%
of the GBR. Within each region L. miniatus were collected
from six individual reefs. Four of these reefs were zoned
"Marine National Park B" and were closed to all forms of
fishing (referred to as "closed reefs" in this article) whereas
the other two reefs were zoned "General Use B" and were
'' Russell, M. 2001. Spawningaggregationsof reef fishes on the
Great Barrier Reef: implications for management. Report from
the Great Barrier Reef Marine Park Authority. 37 p. [Available
from the Great Barrier Reef Marine Park Authority, P.O. Box
1379. Townsville, Queensland, Australia. 4810|.
= Davies, C. R., B. D. Mapstone, A. Ayling, D. C. Lou, A. Punt, G. R.
Russ, M. A. Samoilys, A. D. M. Smith. D. J. Welch, and D. McB.
Williams. 1998. Effects of line fishing experiment 1995-1997:
project structure and operations. -Supplementary to progress
report. CRC Reef Research Centre, Townsville, Australia, 28 p.
lAvailable tioni the CRC Reef Research Centre. PO. Box 772,
Townsville, Queensland, Australia 4810).
6 Mapstone, B. D.. C. R. Davies. I). C. Lou, A. E. Punt, G. R. Russ, D.
A. .1. Ryan. A. D. M. Smith, and D. McB. Williams. 1998. Effects
ofline fishing experiment 199.5-1997: progress report. CRC Reef
Research Centre, 86 p [Available from the CRC Reef Research
Centre, P.O. Box 772, Townsville, Queensland, Austraha 4810).
Williams et al.: Scales of spatial variation in demography of a large coral-reef fishi
675
140°00E
fv
150°OOE
io°oos-
\ "
imsl
imsvilleV,^ 20°oos-
Mackay\
Gladstones.
_L| Op Reef
Faraday Reef
QowReef
Yankee Reef
/ c
Tov
20-136^^ BaxReef
20-137 20-142^
21-130
/ 21-131
21-133 \
21-132
AUSTRALIA
Figure 1
Location of reefs sampled for L. miniatus from October 1995 to January 1996 in the 1)
Townsville, 2 ) Mackay, and 3 ) Storm Cay regions of the Great Barrier Reef, AustraUa.
Reefs maps indicate the relative position of the four reefs closed to fishing that were
sampled in regions 1, 2, and 3.
open to line and spear fishing (referred to as "open reefs").
Fishing had been prohibited from the closed reefs for at
least seven years prior to sampling. Each reef was sampled
for two days by the same four commercial line fishermen
using gear and sampling designs standardized across all
reefs (Davies et al.''). Fork length (FL) was measured to
the nearest millimetre immediately upon capture. Sagittal
otoliths were removed from frozen frames in the laboratory,
cleaned of any residual material, dried, and weighed to the
nearest 0.1 mg.
A total of 1015 L. miniatus were collected from the four
closed reefs in each region between October 1995 and Janu-
ary 1996. Sample sizes from the open reefs were small and
mortality and growth estimates from these reefs were
unreliable. Therefore, these samples were used only to in-
crease the sample size of older fish for a comparison of the
two methods for reading otoliths (whole and sectioned).
Comparison of otolith reading methods
The annual periodicity of opaque increment formation in L.
miniatus otoliths has been validated (Brown and Sumpton
1998). A subsample of 355 L. miniatus otoliths from both
open and closed reefs was used to assess whether readings
of whole otoliths provided age estimates similar to those
from sectioned otoliths, but at substantially lower cost (in
time). Otolith weight was used to select a broad range of
age classes for this assessment on the assumption that
otolith weight was a coarse indicator of age, thus avoiding
the need to preread otoliths to obtain a sample covering
all age classes. Each otolith in the subsample was read,
both whole and sectioned, on three separate occasions in
random order with no prior knowledge of collection date,
location, or fish size. For consistency, the right otolith was
chosen to estimate the age of all fish unless it was missing
or damaged, in which case the left one was used. Otoliths to
be read whole were placed in a small black dish of immer-
sion oil and examined under reflected light with a stereo
dissecting microscope. Counts of opaque increments were
made from the nucleus to the dorsoposterior edge on the
convex face of the otolith. For otoliths from older fish it was
necessary to rotate the otolith approximately 45° to clearly
observe increments on the otolith margin.
Otoliths to be sectioned were embedded in epoxy resin
and cut transversely, adjacent to the anterior side of the
nucleus with a Buehler Isomet low-speed saw. The poste-
rior portion of the otolith was retained and mounted on a
glass microscope slide with Crystalbond adhesive. A second
transverse cut adjacent to the posterior side of the nucleus
resulted in a thin section, incorporating the otolith nucleus,
remaining on the slide. Otolith sections were then ground
on 800- and 1200-grade sandpaper to remove saw marks
and a single drop of immersion oil was placed on sections to
fill surface irregularities. Otolith sections were examined
under a stereo dissecting microscope with reflected light
and a black background. Counts of opaque increments were
676
Fishery Bulletin 101(3)
made from the nucleus to the proximal surface, along the
dorsal margin of the sulcus acousticus.
The precision of age estimates from whole and sectioned
otoliths was calculated by using the index of average per-
cent error (Beamish and Fournier 1981). The estimates of
age from whole and sectioned otoliths were compared by
a paired ?-test. Difference in bias between the two reading
methods was observed by plotting the difference between
the two readings (sectioned age minus whole age) against
sectioned age, based on the assumption that sectioned age
provided the best estimate of true age (Beamish 1979). The
results from this comparison indicated no significant dif-
ference between whole and sectioned otolith readings and
there was no discernible difference in bias in the plot. As a
result, all remaining otoliths were read whole for greater
efficiency. Age estimates from whole otoliths were accepted
and used in subsequent analyses when counts from the
first two readings agreed. If the counts differed, otoliths
were read a third time. The otolith was excluded from sub-
sequent analyses if no two counts agreed, but included if
any two counts agreed.
Comparison of demographic parameters
The central objective of this study was to estimate the vari-
ation in demographic parameters of L. miniatus, specifi-
cally otolith and somatic growth rates, age structure, and
mortality, at different spatial scales. In the first instance,
parameters were compared among the four reefs within
each region to estimate the magnitude of variation at the
inter-reef scale. Data were then pooled from individual
reefs within each region to generate regional parameter
estimates, which were used to estimate the magnitude of
variation at the regional spatial scale.
The relationship between otolith weight and age (rep-
resenting otolith growth) was examined for each reef by
least-squares regression analysis, with otolith weight as
the dependent variable. The relationship was compared
among reefs within each region and among regions by us-
ing analysis of covariance (ANCOVA).
Reef-specific age-frequency distributions were construct-
ed for all reefs. Multidimensional contingency tables were
used to compare age frequencies among reefs within re-
gions and among regions. Age classes 4 years and younger
and age classes 10 years and older were pooled into 4 and
10* age classes, respectively, because of low frequencies in
the tails of the age distributions. As a result, the analyses
included a total of seven age classes.
Age-based catch curves (Ricker, 1975) were used to esti-
mate the instantaneous rate of total mortality (Z) at each
reef expres.sed on an annual basis. The number offish in
each age class was regressed against the corresponding
age, and the descending slope provided an estimate of Z.
Regressions were fitted from the first age class that was
fully selected by the sampling gear through to the oldest
age class that was preceded by no more than two consecu-
tive zero frequencies. As a result, the age range used to
estimate mortality varied slightly among reefs. Mortal-
ity rates were compared among reefs within regions and
among regions by using ANCOVA.
The von Bertalanffy growth function (VBGF) provided
the best fit to length-at-age data for most reefs according
to the parameter estimates of the Schnute (1981) growth
function. For consistency, and to enable spatial compari-
sons of growth, the VBGF was used to estimate growth
parameters for each reef and region:
L, = L^[1-^
-KH-lr,)]
where L, = the fork length at age t;
L^ = the mean asymptotic fork length;
K = the rate at which L^ is approached; and
^Q = the age at which fish have a theoretical length
of zero.
It was difficult to obtain a reliable estimate of initial growth
because the youngest fish collected was 2 years old. There
are also no published size-at-age data for larval or juve-
nile L. miniatus, or any other lethrinid. We constrained the
VBGF parameter t^ to zero to provide a better description
of the likely early growth of L. miniatus. This procedure
also allowed growth curves to be compared among reefs
within regions and among regions by using 959'f confidence
regions of the VBGF parameters L^ and K described by
Kimura(1980).
Results
Comparison of otolith reading methods
Age estimates from whole and sectioned otoliths did not
vary significantly over the range of ages between 2 and 21
years (^00.5 2 3.54-'^-'^^' ^=0.73). That is, for each age class
estimated from sectioned otoliths, the average difference
between whole and sectioned otolith readings did not
differ significantly from zero (Fig. 2). The index of average
percent error was very low for whole ( 1.6^* ) and sectioned
( 1.4%) otolith readings, indicating that otolith readings for
both methods were highly repeatable. This low index was
reflected in the agreement of at least two age estimates
for all whole otoliths, and hence no otoliths were excluded
from analyses.
Otolith growth
There was a significant positive linear relationship between
otolith weight and age for all reefs, with regression coef-
ficients ranging from 0.64 to 0.90. ANCOVA revealed that
the slope of this relationship was not significantly different
among reefs within each region (Townsville: F., .j^g=1.91,
P=0.13; Mackay: F.^.^^^=1.02, P 0.38; Storm Cay: Fl~i^=l.?>h.
P=0.20). Thus, otolith weight and age data were pooled for
each region to compare the region-specific relationships
between otolith weight and age (Fig. 3). The slopes of the
region-specific relationships differed significantly among
all regions (F.^y4g=28.9, P<0.001). The average growth in
otolith weight was greater in the Mackay region (26.91
mg/yr) than in the Storm Cay region (24.48 mg/yr), and
was least in the Townsville region (19.50 mg/yr).
Williams et al : Scales of spatial variation in demography of a large coral-reef fisfi
677
10
^ 0.5
o
5
I
6 years) in the Towns-
ville region than in the Mackay and Storm Cay regions
(Fig. 4). However the oldest fish were from the Storm Cay
region, where a small number offish persisted in the older
age-classes up to 19 years of age. The relative abundances
of age classes 4 and 5 were greater in the Mackay region
than in the Townsville and Storm Cay regions (Fig. 4).
Mortality
Estimates of annual total mortality rates (Z) for individ-
ual reefs were generally similar among reefs within each
600
500
400
300
200
100
A Townsville
« *
:
:.r'
m
500
fc
^
4(10
U)
%
300
r
■M)
n
L)
100
0
B Mackay
600
C storm Cay
«
500
400
• It^
♦
300
,llir
200
.#•
100
n
r*
0 5 10 15 20
Age (years)
Figure 3
Least-squares linear regression of
otolith weight on age for L. miniatus
from three regions of the Great Barrier
Reef: (A) Townsville, (B) Mackay, and
(C) Storm Cay.
region, with the exception of the Storm Cay region where
the estimated Z appeared much lower for reef 21-131 than
for other reefs in that region (Table 1). ANCOVA indicated
no significant difference in mortality among reefs in any
region (Townsville: F3 3o=0.80, P=0.50; Mackay: F3 2o=0.08,
P=0.97; Storm Cay: P.j ,5=1.14, P=0.37). Therefore mortal-
ity rates were estimated for each region from the pooled
age structures for all reefs within each region (Fig. 5). A
comparison among regions of the regression slopes from
the pooled age structures indicated significant differences
among regions {F2^g=T-ll, P=0.005). Tukey's multiple
comparison tests revealed that the estimated Z for the
Townsville (Z=0.42) and Mackay (Z=0.71) regions differed
significantly, whereas the estimate from the Storm Cay
region (Z=0.60l did not differ significantly from either
Townsville or Mackay.
678
Fishery Bulletin 101(3)
A Townsvllle
n = 347
B Mackay
10 12 14 16 18 20
n = 370
0) 40
2 4 6 8 10 12 14 16 18 20
C Storm Cay
n = 298
2 4 b b 10 12 14 16 18 20
Age (years)
Figure 4
Age structures for L. miniatus from
three regions of the Great Barrier
Reef: (A) Townsvllle, (B) Mackay, and
(C) Storm Cay. Age structures were
pooled from four closed reefs within
each region.
Somatic growth
Estimates of VBGF parameters varied considerably among
reefs within the Townsvllle and Mackay regions but in the
Storm Cay region, estimates of L^ and particularly K' were
very similar (Table 1). Examination of BS^f confidence
regions for VBCJF parameters for individual reefs (Fig. 6)
indicated considerable uncertainty in the estimates of both
K and L.^ and no clear differentiation among reefs within
regions. The similarity in VBGF parameters for individual
reefs within the Storm Cay region was particularly evident
from the 95% confidence regions. In both the Townsvllle
and Mackay regions, three reefs showed overlap in 95%
J= 2
TownsvNIe
Mackay
0 2 4 6 8 10 12 14 16 18
Age (years)
Figure 5
Age-based catch curves fori, miniatus from
three regions of the Great Barrier Reef • =
Townsvllle; ■, Mackay; ▲ = Storm Cay.
Catch curves were derived from pooled age
structures from four closed reefs within each
region. Open data points were not used in
the regressions.
confidence regions, whereas only a single reef in each
region appeared to have significantly different VBGF
parameters from the others (Fig. 6).
Given the lack of differentiation in growth among reefs,
the data from individual reefs were pooled for each region
to examine regional patterns in growth. VBGF parameters
varied significantly among regions (Table 1) with no over-
lap in the 95% confidence regions (Fig. 7). It appeared that
L. miniatus in the Mackay region attained a larger average
asymptotic size (L^=472.21 mm FL) than in the Storm Cay
region (L^=462.83 mm FL), where in turn these fish grew
larger than fish in the Townsvllle region (L^=453.36 mm
FL). It should be noted that the constrained fitting of the
VBGF itg set to zero) provided a conservative estimate of
regional variation in growth, and regional differences were
considerably larger when the VBGF parameter £
O 600
U.
500
400
300
200
100
0
C Storm Cay
^
«♦
500
400
300
200
100
0
490
480
470
460
450
B Mackay
D 95% confidence regions
Mackay
5 10 15 20
Age (years)
0 34 0 39 0 44 0 49 0 54
K (per year)
Figure 7
Von Bertalanffy growth function { VBGF ) fitted to size-at-age data for L. miniatus
from three regions of the Great Barrier Reef (A) Townsville, (B) Mackay, and
(C) Storm Cay. The VBGF parameter Jq is constrained to zero in all cases. Dashed
lines represent predicted growth between 0 and 2 years of age. (D) 95% confidence
regions for the VBGF parameters K and L^ for fit to pooled data from four reefs
within each of the three regions.
accuracy in relation to those estimated from sectioned oto-
liths. In contrast, Brown and Sumpton (1998) concluded
that whole otoliths from larger and presumably older L.
miniatus underestimated age by up to 40% with respect to
sectioned otoliths. The discrepancy between studies may
be due to differences in the techniques used to count incre-
ments in whole otoliths. It was noted early in the present
study that otoliths from older fish needed to be rotated to
reveal a number of increments close to the otolith margin.
By not using this technique Brown and Sumpton (1998)
may have underestimated ages from whole otoliths of
older fish. Readings from whole otoliths have been shown
to consistently underestimate the age of a number of reef
fish species (e.g. Ferreira and Russ, 1994; Newman et al.,
2000) resulting in biased estimates of mortality and sub-
sequent yield estimates (Newman et al., 2000). The results
from this study suggest that whole otoliths are adequate
for estimating the age of L. miniatus and that estimates of
demographic parameters presented in the present study
were not biased by underestimates of age.
The spatial patterns in the demography of L. miniatus de-
scribed in the present study are based on data collected from
a single survey in one year, thus leaving the temporal stabil-
ity of the patterns open to question. Continued monitoring
of the populations will be required to determine the stability
of the patterns, and focussed stock structure studies are
required to determine the most likely causal mechanism(s)
of the patterns. Notwithstanding the need for this work, the
significant regional differences in demographic parameters
found in the present study suggest different levels of produc-
tivity of L. miniatus populations in each region. Consequent-
ly, there is the potential for less productive populations to
be overfished, even where the fishing effort for the stock
as a whole is managed at sustainable levels (Caddy, 1975;
Sheperd and Brown, 1993). This argues for assessments and
management of L. miniatus stocks to explicitly consider the
regional structure in demography in order to meet both
sustainable use and conservation objectives for the Great
Barrier Reef World Heritage Area overall and on a regional
basis. Furthermore, this study highlights a more general
need for the use of multiscale sampling and analyses offish
populations to understand the relative importance of the
processes affecting demographic parameters, and the scales
at which these processes operate.
Acknowledgments
We acknowledge financial support from the Cooperative
Research Centre for the Great Barrier Reef World Heritage
Area, the Fisheries Research and Development Corpora-
tion, the Great Barrier Reef Marine Park Authority, and
the CRC Reef Research Augmentative Grant Scheme. The
VBGF ConfRegion program developed by J. Kritzer, CRC
Reef Research Centre, was used to estimate the VBGF 95%
confidence regions. We would like to thank Robin Stewart,
Mary Petersen, and the ELF field team for assistance in
the collection and processing of the fish from which this
682
Fishery Bulletin 101(3)
work arose. We would also like to thank three anonymous
reviewers for comments that improved the quality of
this manuscript. This manuscript is a contribution from
the CRC Reef Effects of Line Fishing Project, CRC Reef
Research Centre, Townsville, Australia.
Literature cited
Adams, S., B. D. Mapstone, G. R. Russ, and C. R. Davies.
2000. Geographic variation in the sex ratio, sex specific size,
and age structure of Plectropomus leopardus (Serranidae)
between reefs open and closed to fishing on the Great Bar-
rier Reef. Can. J. Fish. Aquat. Sci. 57:1448-1458.
Beamish, R. J.
1979. Differences in age of Pacific hake (Merluccius produc-
tus) using whole otoliths and sections of otoliths. J. Fish.
Res. Board Can. 36:141-151.
Beamish. R. J., and D. A. Fournier.
1981. A method for comparing the precision of a set of age
determinations. Can. J. Fish. Aquat. Sci. 38:982-983.
Begg, G. A., J. A. Hare, and D. D. Sheehan.
1999. The role of life history parameters as indicators of
stock structure. Fish. Res. (Amst.) 43:141-163.
Brown, I. W., and W. D. Sumpton.
1998. Age, growth and mortality of redthroat emperor Leth-
rinus miniatus (Pisces: Lethrinidae) from the southern
Great Barrier Reef Queensland, Australia. Bull. Mar.
Sci. 62:905-917.
Caddy, J. F
1975. Spatial model for an exploited shellfish population,
and its application to the Georges Bank scallop fishery. J.
Fish. Res. Board Can. 32:1305-1328.
Caley, M. J., M. H. Carr, M. A. Hixon, T P Hughes. G. P Jones,
and B. A. Menge.
1996. Recruitment and the local dynamics of open marine
population. Annu. Rev. Ecol. Syst. 27:477-500.
Carpenter, K. E., and G. R. Allen.
1989. FAO species catalogue. Emperor fishes and large-eyed
breamsof the world (family Lethrinidae). An annotated and
illustrated catalogue of lethrinid species known to date.
FAOFish. Synop. 125, 118p. FAO, Rome.
Conover, D. O.
1992. Seasonality and the scheduling of life history at differ-
ent latitudes. J. Fish Biol. 41(suppl. B):161-178.
Davies, C. R.
1995. Patterns of movement of three species of coral reef
fish on the Great Barrier Reef PhD diss., 203 p. Dep.
Marine Biology, James Cook Univ., Townsville, Queensland,
Australia.
DeMartini, E. E.
1993. Modelling the potential of fishery reserves for manag-
ing Pacific coral reef fishes. Fish. Bull. 91:414-427.
Doherty P. J.
1983. Tropical territorial damselfishes: is density limited by
aggregation or recruitment? Ecology 64:176-190.
Doherty, P. J., and A. Fowler
1994. Demographic consequences of variable recruitment to
coral reef fish populations: a congeneric comparison of two
damselfishes. Bull. Mar Sci. 54:297-313.
Doherty P. J., S. Planes, and P. Mather.
1995. Gene flow and larval duration in seven species offish
from the Great Barrier Reef Ecology 76:2373-2391.
Dudgeon, C. L., N. Gust, and D. Blair
2000. No apparent genetic basis to demographic differences
in scarid fishes across continental shelf of the Great Barrier
Reef Mar Biol. 137:1059-1066.
Ferreira, B. P., and G. R. Russ.
1994. Age validation and estimation of growth rate of the coral
trout, Plectropomus leopardus. (Lacepede, 1802) fi'om Lizard
Island, Northern Great Barrier Reef Fish. Bull. 92:46-57.
1995. Population structure of the coral trout, Plectropomus
leopardus (Lacepede 1802), on fished and unfished reefs off
Townsville, Central Great Barrier Reef Aust. Fish. Bull.
93:629-642.
Hart, A. M., and G. R. Russ.
1996. Response of herbivorous fishes to crown-of-thorns
starfish Acanthaster planci outbreaks. III. Age, growth,
mortality and maturity indices of Acanthurus nigrofuscus.
Mar Ecol. Prog. Sen 136:25-35.
Heemstra, P. C, and J. E. Randall.
1993. FAO species catalogue, vol. 16. Groupers of the world
(family Serranidae, subfamily Epinephelinae). An anno-
tated and illustrated catalogue of the grouper, rockcod, hind,
coral grouper, and lyretail species known to date. FAO
Fish Synop. 125, 292 p. FAO, Rome.
Jones, G. P.
1987. Competitive interactions among adults and juveniles
in a coral reef fish. Ecology 68:1534-1547.
Kimura, D. K.
1980. Likelihood methods for the von Bertalanffy growth
curve. Fish. Bull. 77:765-776.
Levin, S. A.
1992. The problem of pattern and scale in ecology. Ecology
73:1943-1967.
Lewis, A. R.
1997. Recruitment and post-recruit immigration affect the
local population size of coral reef fishes. Coral Reefs 16:
139-149.
Loubens, G.
1980. Biologie de quelques especes de poissons du lagon Neo-
Caledonien. 111. Croissance Cah. I'lndo-Pac. 2:101-153.
Lough, J.M.
1994. Climate variation and El Nino-Southern Oscillation
events on the Great Barrier Reef 1958 to 1987. Coral
Reefs 13:181-195.
Meekan, M. G., J. L. Ackemian, and G. M. Wellington.
2001. Demography and age structures of coral reef dam-
selfishes in the tropical eastern Pacific Ocean. Mar. Ecol.
Prog. Sen 212:223-232.
Newman, S. J., M. Cappo, and D. McB. Williams.
2000. Age, growth, mortality rates and corresponding yield
estimates using otoliths of the tropical red snappers, Lut-
janus ery'thropterus, L. malabancus and L. sebae, from the
central Great Barrier Reef Fish. Res. 48:1-14.
Newman, S. J., D. McB. Williams.
1996. Variation in reef associated assemblages of the Lut-
janidae and Lethrinidae at different distances offshore in
the central Great Barrier Reef Environ. Biol. Fishes 46:
123-138.
Newman, S. J., D. McB. Williams, and G. R. Russ.
1996. Variability in the population structure of Luljanus
adetii (Castelnau, 1873) and L. quinquelineatus (Bloch,
1790) among reefs in the central Great Barrier Reef Aust.
Fish. Bull. 94:313-329.
Nielson, J. L.
1998. Population genetics and the conservation and man-
agement of Atlantic salmon (.Salmo salar). Can. J. Fish.
Aquat. Sci. 55(suppl. 1):145-152.
Ricker.W. E.
1975. Computation and interpretation of biological statis-
Williams et al.: Scales of spatial variation in demography of a large coral-reef fish
683
tics offish populations. Bull. Fish. Res. Board Can. 191,
382 p.
Roberts, C. M., and N. V. C. Polunin.
1991. Are marine reserves effective in management of reef
fisheries? Rev. Fish Biol. Fish. 1:65-91.
Russ, G. R., A. C. Aleala, and A. S. Cabanban.
1992. Marine reserves and fisheries management on coral
reefs with preliminary modeling of the effects of yield per
reci-uit. Proc. of the 7"^ int. coral reef svmp. 2, p. 988-995.
Univ. Guam, Marine Lab., Mangilao, Guam.
Sale, P. F.
1998. Appropriate spatial scales for studies of reef-fish
ecology. Aust. J. Ecol. 23:202-208.
Schnute, J.
1981. A versatile growth model with statistically stable
parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140.
Sheperd, S. A., and L. D. Brown.
1993. What is an abalone stock: implications for the role of refii-
gia in conservation. Can. J. Fish. Aquat. Sci. 50:2001-2009.
Shulman, M. J., and E. Bermingham.
1995. Early life histories, ocean currents, and the population
genetics of Caribbean reef fishes. Evolution 49:897-910.
van Herwerden, L., J. Benzie, and C. R. Davies.
In press. Microsatellite variation and population genetic
structure of red throat emperor {Lethrinus miniatus) in
the Great Barrier Reef Australia. J. Fish Biol.
Walters, C. J., and R. Bonfil.
1999. Multispecies spatial assessment models for the Brit-
ish Colombia groundfish trawl fishery. Can. J. Fish. Aquat.
Sci. 56:601-628.
684
The occurrence of yellowfin tuna iThunnus albacares)
at Espiritu Santo Seamount in the Gulf of California
A. Peter Klimley
Salvador J. Jorgensen
Bodega Marine Laboratory
University of California, Davis
Westside Road
Bodega Bay, California 94923
Present address (for A. P. Klimley): Department of Wildlife, Fisfi, and Conservation Biology
University of California Davis
Davis, California 95616
E-mail address (for A. P. Klimley): apklimley@ucdavis.edu,
Arturo Muhlia-Melo
Centre de Investigaciones Biologicas del Baja Norte
Apartado Postal 128
La Paz, Mexico
later identifying them from these tags.
This method results in the removal of
individuals from the population and
yields a percentage of individuals
that have either left the area or have
been captured (Holland et al., 1999).
Detecting coded ultrasonic tags by an
automated monitor provides additional
information because marked individu-
als can be detected repeatedly over a
period of time. However, fewer tags can
be deployed because of their greater
cost. We used this method to reveal
synchronicity among visits of yellowfin
tuna, time of visits, and duration of
visits at the Espiritu Santo Seamount
in the Gulf of California.
Methods
Sallie C. Beavers
Bodega Marine Laboratory
University of California, Davis
Westside Road
Bodega Bay, California 94923
Pelagic fishes are not evenly dispersed
in the oceans, but aggregate at dis-
tinct locations in this vast and open
environment. Nomadic species such
as mackerels, tunas, and sharks form
assemblages at seamounts (Klimley
and Butler, 1988; Fontenau, 1991).
Fishermen have recognized this
behavior and have placed moorings
with surface buoys in deep waters to
provide artificial landmarks, around
which fish concentrate and are more
easily captured. These fish aggregating
devices (termed FADs) are common in
the tropical oceans (see review, Hol-
land, 1996). In a sense, it may only be
the larger size that separates a sea-
mount from a man-made FAD.
Fish may aggregate at seamounts for
very different reasons. The opportunity
to feed is greater because biomass at all
trophic levels, from primary producer
to apex consumer, is greater than in
the open ocean (Boehlert and (Jenin,
1987). The disturbance of flow by the
seamount creates eddies downstream
that retain nutrients critical to the
growth of phytoplankton, and this
enrichment supports a greater abun-
dance of consumers from zooplankton
to apex predators. The dipole nature of
seamount magnetic fields and the out-
ward radiating valleys and ridges of
magnetic minimums and maximums
might provide landmarks in oceanic
landscape that fish use as a reference to
guide migration (see discussion of mag-
netic "topotaxis" in Klimley, 1993). Yel-
lowfin (Thunnus albacares) and bigeye
(Thiinniis obesus) tunas do not reside
long at the Cross Seamount near Ha-
waii, an observation inconsistent with
the theory that tunas feed on prey that
remain aggregated at the site; rather
their rapid passage suggests that the
site is a landmark used to guide migra-
tions (Holland et al., 1999). Adult yel-
lowfin tuna also stay briefly (<5 min) at
FADs off Kaena Point, Oahu (Klimley
and Holloway 1999).
Describing the degree of residency of
pelagic fishes at different geographic
locations helps ascertain whether the
affinity to seamounts and FADs is com-
mon throughout the oceans. Holland
et al. (1999) determined the rates of
dispersion of tuna by attaching unique
tags to individuals, releasing them, and
We tagged 23 yellowfin tunas with
coded ultrasonic beacons during a
five-month period between 11 April
and 12 September 1998. They were
tagged <150 m from two monitoring
stations: Espiritu Santo North (ESN)
and South (ESS), separated by 500
m at the Espiritu Santo Seamount
(24°42'N; 110°18'W) in the south-
ern Gulf of California (Fig. 1). The
seamount rose to within 18 m of the
surface and extended 700 m along a
northwesterly-southwesterly axis.
Monitoring station ESN was situated
at the northwest end of the seamount
ridge at a depth of 47 m; station ESS
was at 37 m on the southwest end.
The monitors were deployed for 30
months, during which they recorded
when the tagged tuna swam within
the 150-m range of reception of the
monitors. Using SCUBA, we removed
the monitors from the moorings at
four-month intervals, downloaded
the records of tuna presence near the
seamounts to a laptop computer, and
replaced the monitors during the same
day We located a station by the rosette
of buoys, which floated at a depth of
<10 m and which was visible from
the surface, by towing a diver at the
surface near the GPS coordinates for
the mooring.
Manuscript approved for publication
30 January 2003 by Scientific Editor.
Manuscript received 4 ApriL 2003 at NMFS
Sciontifu- Publications OfTicc.
Fish. Bull. 101:684-692 (2003).
NOTE Klimley et al.: Occurrence of Thunnus albacares in the Gulf of California
685
115°
110°
Espiritu Santo Seamount (ES)
Figure 1
Bathymetric contour map of seamount Espiritu Santo (ES). The circles with
cross-hatching indicate the range of the tag-detecting monitor from the sea-
mount. The insert shows the geographic location of the seamount (ES) in the
Gulf of Cahfomia.
We determined the maximum range of signal-detection
of one monitor by lowering a transmitter to 10 m under
a small boat and lowering the monitor to a similar depth
under a larger support vessel. We recorded the separation
distance between the two boats using radar because the
small boat and transmitter drifted away from the support
vessel that was anchored in place at the highest point on
the seamount. The VROl monitor (Vemco Ltd., Shad Bay,
Nova Scotia, Canada) detected tags at a distance of 150 m
in seas with waves <0.5 m high (see circles. Fig. 1). Later
models (Vemco Ltd., VR02) used in the study have a pub-
lished reception range of aSOO m in calm seas (see http://
www. vemco.com). The range of tag detection by the moni-
tors decreases with rising sea state because of the increase
in wave-generated ambient noise.
The tuna were caught by rod and reel and lifted aboard 1-
30 minutes after being hooked. Smaller individuals (^15 kg)
were weighed with a scale with a hook that fit into the oper-
culum; intermediate sized fish (>15 and s25 kg) were
weighed in the net and the net's mass subtracted from the
cumulative value; and the masses of largest tuna (>25 kg)
were estimated on the basis of their length by using the re-
gression equation, >'=0.216ac + 2.981 given in Moore (1951).
The tags were inserted into the peritoneum of the tuna while
salt water was flushed over their gills by using the technique
described in Klimley and Holloway (1999). The tuna were
retained on board for tag implantation less than a minute.
The transmitters (Vemco Ltd., V16-6L) were cylindrical
and had a diameter of 16 mm, length of 106 mm, and net
mass in water of 16 g. They emitted individually coded
tone bursts of 70 kHz separated by 60-90 s intervals. The
amplitude of the pulses was 147 dB (re: 1 ^P) at a distance
of 1 m. The theoretical operating life of a transmitter was
476 days. Each tag was distinguished on the basis of a
unique pulse burst by an automated tag-detecting moni-
tor attached to the ESS and ESN detection stations. Water
686
Fishery Bulletin 101(3)
Table 1
Length and mass of the 23 yellowfin tuna {Thunnus albacares) tagged in the present study and the date and time of tagging. "N"
indicates tagging near northern monitor; "S" denotes tagging near southern monitor. An asterisk in front of a measurement indi-
cates that the value is derived from the mathematical relationship between mass and length given in Moore (1951); "TL" denotes
total length.
Tuna
no.
Date
Time
(h)
TL
(cm)
Mass
(kg)
Site
(N/S)
1
11 Apr
1998
13:04
80.0
7.3
S
2
11 Apr
1998
13:21
96.0
10.8
S
3
12 Apr
1998
08:46
91.0
10.3
8
4
12 Apr
1998
08:51
106.0
13.8
S
5
12 Apr
1998
09:54
104.0
15.5
S
6
17Jun
1998
09:54
91.5
17.0
S
7
24Jun
1998
10:38
86.5
11.3
S
8
26 Aug
1998
10:05
138.0
*51.7
N
9
26 Aug
1998
10:45
58.0
4.5
N
10
26 Aug
1998
11:43
66.0
5.5
N
11
26 Aug
1998
12:16
76.0
7.0
N
12
26 Aug
1998
10:14
155.0
*73.1
N
13
28 Aug
1998
10:50
71.0
7.2
N
14
28 Aug
1998
11:25
155.0
*73.1
N
15
10 Sep
1998
17:44
149.9
*66.2
S
16
10 Sep
1998
18:32
91.5
18.50
S
17
10 Sep
1998
18:44
*75.0
8.50
s
18
10 Sep
1998
19:07
111.8
*27.6
s
19
11 Sep
1998
17:25
114.5
20.5
N
20
11 Sep
1998
17:51
71.0
7.00
N
21
11 Sep
1998
18:25
106.5
20.5
N
22
12 Sep
1998
6:41
104.5
23.0
N
23
12 Sep
1998
7:30
141.0
*55.1
N
temperature was recorded every half hour at the seamount
by a Stoaway Tidbit temperature logger ( Onset Computers
Corp., Pocassett, MA) attached to the mooring line adjacent
to the tag-detecting monitor. We calculated a daily tem-
perature by averaging the half-hourly temperatures.
We used log-survivorship analysis (Fager and Young,
1978) to ascertain whether the tunas returned to the moni-
toring stations after favored time periods. A frequency his-
togram of the time intervals between randomly occurring
point events in a Poisson process is described by a negative
exponential distribution. A log-survivor plot of these inter-
vals generates a straight line with a slope proportional to
the probability of an event occurring at a given time after
the preceding event. This analysis is used to identify inter-
vals between events that occur more frequently than ex-
pected by chance because inflections in the resulting curve
are more easily distinguished from a .straight line than the
shape of the distribution on a frequency histogram with a
negative exponential distribution. An inflection in the log-
survivor curve indicates a change in the probability of an
event occurring at a given time after the last event — in our
case the time between successive arrivals of tunas within
the ranges of the two monitors.
Results
Twenty-three yellowfin tunas were tagged from 11 April
1998 to 12 September 1998 (Table 1). Individuals were
tagged during daylight hours from 6:41 to 19:07 hours.
The tunas ranged in length from 71.0 to 155.0 cm TL.
They ranged in mass from 7.25 to 73.1 kg. There appeared
to be two discrete size classes, small individuals varying
from 7.25 to 23.0 kg and large ones from 51.7 to 73.1 kg.
The masses of the larger individuals were determined
from their lengths by using a regression equation (Moore,
1951).
The yellowfin tunas stayed at seamount Espiritu Santo
over varying time periods (Fig. 2). Nine of the 23 tunas
left the seamount on the same day that they were tagged
(Fig. 2A). Two of the nine returned to the seamount once
for a single day. one within a week of tagging and the
other after two and one-half months. Six tunas stayed
intermediate periods of time after tagging, ranging from
two to six weeks. One of these tunas (no. 9) was eventually
caught at the seamount. Another tuna (no.lO) visited for
a single day after an absence of five weeks and returned
again after a similar period to stay for 15 months. Four
NOTE Klimley et a\. Occurrence of Thunnus albacares in the Gulf of California
687
Apr
May
Aug
Sep
Oct
Nov
Dec
'I '
Jan
Feb
Apr
28-
c
2000 ^ v^--^-'^~Jt..v^;^
24-
/ v'^ ;
20 J
16-
■ ' ' '1
^-, r i V — ...,.., \
23-
21-
19
IT-
IB-
IS-
11
9
7
5
3
1
—
Beacon not yet
recovered.
T
Tuna tagged
with beacon.
C
Tuna captured
with beacon.
F
Beacon drops
to surface of
seamount.
Apr
(itey
Aug
Sep
Oct
Jun Jul
IVIonth of year
Figure 2
Chronology of daily visits by 23 tagged yellowfin tuna to the seamount and temperature record over a 30-
month period beginning April 1998 and ending October 2000. Each visit, indicated by a solid diamond, is
based on the detection of a tag during a 24-h period by either the north (ESN) or south (ESS) monitoring
stations. The lines in the graphs show that the ultrasonic tag had yet to be recovered from a yellowfin tuna.
T = day of tagging, C = day of capture, and F = date of shedding of tag.
individuals (nos. 5, 19, 21, and 23) stayed for longer peri-
ods of time, ranging from six to 18 months. One of these
tunas (no. 51 was also caught by a fisherman. It is likely
that some tunas are nomadic and stay only a single day,
whereas others are resident, remaining at the seamount
throughout the year.
688
Fishery Bulletin 101(3)
It is unlikely that the tags on the two tunas (nos. 10
and 23), which stayed at the seamount longest, were shed
and lay on the bottom. The reasons supporting their being
attached to living tunas are as follows. First, the two tags
were not recorded with equal frequency during all times of
the day as might be expected of a tag lying at one location
within the range of the monitors. The tags were usually de-
tected for a few hours and then absent for a similar period.
This pattern of detection is consistent with the tunas mov-
ing within the range of the monitor and later outside its
range. Second, the two tags were jointly detected after long
periods of absence or ceased being detected simultaneously
after long periods of presence. This reception pattern is
consistent with the two tunas moving in and out of the de-
tection range of the monitors within the same school. Third,
one tuna (no. 23) was detected by the monitor on the south
side of the seamount, but not on the north side during one
day; the same tuna was detected by the northern monitor,
but not the southern monitor on the next day. This pattern
of detection was consistent with the tuna swimming over
the northern region of the seamount on the first day and
over the southern region on the second day.
The yellowfin tunas were present at the seamount at
all seasons of the year Five of the tunas tagged during
August and September 1998 (nos. 7, 8, 9, 16, and 17) emi-
grated during early fall as the water temperature began
to decrease (Fig. 2A). However, three individuals (nos. 10,
21, and 23) remained at the seamount from January 1999
to April 1999 when the temperature dropped to 18°C. Two
(nos. 10 and 23) remained present when the subsurface
water temperature descended to 16°C during the following
winter of 2000 (Fig. 2B).
The yellowfin tunas remained at the seamount at all
times of the day. This is evident from a 24-h record of the
arrivals of 10 tunas during a 15-d period from 16 to 30
September 1998 (Fig. 3). The tunas were present more
often during daytime, from 06:00 to 18:00 hours, during
the first 12 days. Notice the clustering of the different
symbols in Figure 3, each indicating a specific tuna, in
separate columns during the period from 06:00 to 18:00
hours. However, the amount of time spent at the sea-
mount became more evenly distributed between daytime
and nighttime by 28 September Note the even dispersion
of the symbols over the 24-h period during the last three
days of the 15-day period. There was little variation evi-
dent in the frequency of arrivals at different times of the
day when the arrivals were summed over the entire study
(Fig. 4). The percentage of arrivals during each hour of
the day (see crosshatched polygon) differed little from an
even distribution of arrivals (4.2''// /h) throughout the day
(see dashed circle).
We determined the frequency of various lengths of
stays at the north (Fig. 5A) and south sites (Fig. 5B) at
the Espiritu Santo Seamount. A stay for a particular tuna
was defined as the period of detections with no separation
intervals greater than 15 min. Let us say that tuna 1 was
detected at 08:00, 08:14, 08:28, and 09:00 hours. The dura-
tion of the stay of tuna 1 would be 28 min. The second detec-
tion followed the first by 14 min (<15 min), and the third
followed the second also by 14 min (also < 15 min). However,
the fourth detection followed the third by 32 min (>15 min)
and was therefore not pooled into a single duration. This
stay would then be placed in the 15:00-29:59 min time
class in Figure 5. Twenty-seven percent of the detections
at ESN and 33% of those at ESS were separated by greater
than 15 min and were thus considered single detections
and included in the 00:00-h class. Fifty-three percent of the
visits to ESN and 37% of the visits to ESS were between
00:01 and 14:59 min. Twenty-one percent of the visits to
ESN and 20% of the visits to ESS were between 15:00 and
59:59 min. The majority of visits were less than 1 hour in
duration and only a few exceeded an hour.
The intervals spent away from the seamount were simi-
larly short. Sixty percent of all absences at ESN were less
than 1 h (Fig. 6) as were 65% of the absences from ESS.
Ninety percent of the absences from both sites were less
than 5 hours. Only 0.1% of the visits exceeded 23 hours.
There appeared to be no favored period of absence as
indicated by the smooth slope of both curves in the log-
survivor plot. Only 72 periods of absence at ESN and 114
periods at ESS exceeded a day. Of these longer periods,
42% of the absences from ESN (Fig. 7A) and 46% of the
absences from ESS (Fig. 7B) were for two days. Only 7 %
of the absences from ESN and 4% of the absences from
ESS were between 10 and 19 days. Only 2% of periods of
absence from ESN exceeding a day were greater than 100
days (Fig. 7A).
Discussion
We found that yellowfin tuna remained at the seamount
for periods ranging from a few days to greater than a year.
Fifty percent of 458 yellowfin tuna tagged with dart tags
at the Cross Seamount off Hawaii were recaptured at that
seamount within 15 days of tag application (Holland et
al., 1999). This "half-life" of tuna residence was short, sug-
gesting that the seamount served as a landmark to guide
migration and not as a destination for feeding.
Thirty-eight yellowfin tuna were tagged with ultrasonic
beacons at two buoys off the western coast of Oahu and
monitored over a 13-month period by automated "listen-
ing" monitors (Klimley and Holloway, 1999). These moni-
tors (VR20) possessed a more sensitive receiver than our
monitors (VROl and VR02). The former had a maximum
range of 1.1 km. The maximum published range of our
monitors was 0.5 km. Twenty-seven of the tuna returned
to the buoys a mean of 4.2 visits per tuna. The mean dura-
tion of each visit was only 40.1 min and the mean period of
absence was 17.2 days. Seventy-three percent of the tuna
tagged on the same day returned together The tunas often
arrived at the same time of the day and returned only to the
buoy at which they were tagged. This allegiance of tunas
to one school, their predilection for returning to the site of
tagging, and the precise timing of their visits are consistent
with the theory that the species has migratory pathways
consisting of way-points that are visited with regularity.
That the tuna spent little time at the FAD suggests that
the buoys are not feeding destinations, but rather land-
marks used in migration.
NOTE Klimley et al.: Occurrence of Thunnus albacares in the Gulf of California
689
Tuna arrivals at Espiritu Santo North
24 00
22 00
02 00
- 02 00
-r? 00:00
T — f M' ' 1 **■ I ■ " r— r-» r**- r"'^* h oooo
:£ 15-Sep-98 16-Sep-98 17-Sef)-98 18-Sep-98 19-Sep-98 20-Sep-98 21-Sep-98 22-Sep-98 23-Sep-98 24-Sep-98
V)
E
r (Order of ten symbols from left to right, indicate tunas witti tags 5, 8, 9, 16, 17, 19, 20,21,22 and 23)
23-Sep-98 24-Sep-98
25-Sep-98
26-Sep-98 27-Sep-98 28-Sep-98 29-Sep-98
30-Sep-98
1-Oct-98
Figure 3
Twenty-four hour chronology of visits by 10 tagged tuna to the monitoring station ESN during 15 days from 16 to 30 September
1998. A unique symbol indicates the presence of a particular individual within the range of the monitor during a 15-min position.
Note the predominance of daytime visits during the first nine 24-h periods and then a progressive shift to an equal number of visits
during daytime and nighttime (see 28-30 Sept. 1998).
Tuna repeatedly moved in and out of the monitor range
over many days or left for the duration of the study. Sixty
percent of all absences at ESN and 65 % of the absences
from ESS were for less than 1 hour. If these tunas were to
swim at a sustained rate of 0.5 m/s (see Magnuson, 1978),
they would not move more than 900 m out the reception
range of the monitors (60 min x 60 s x 0.5 m/s/2).This close
attachment to the seamount contrasts with the behavior of
tuna at FADs offshore of Hawaii. Tunas visited the FADs
there rarely and spent little time within the range of the
monitor before departing for a period of several weeks
(Klimley and Holloway, 1999). The present study suggests
that the Espiritu Santo Seamount is a substantial feeding
ground that can support a year-round resident population
of yellowfin tunas. However, other tunas may stay only
briefly at the seamount, using it as a landmark, before
continuing on their nomadic migrations.
Seamoimts have dipole magnetic fields associated with
them because of the antiparallel polarity of magnetite within
volcanic magma extruded during periods when the earth's
polarity was reversed (Parker et al., 1987). Furthermore,
maxima (ridges) and minima (valleys) in the magnetic field
often lead outward from seamounts due to the extrusion of
magma. Klimley (1993) proposed that hammerhead sharks
use these for guidance during their nocturnal migrations into
the surrounding water to forage. This physical property of the
sea floor, originating fair below where the fishes swim, could
provide a fixed reference (or waypoint) for yellowfin during
their migrations. This species of tuna has been shown to
sense distinct patterns in a magnetic field (Walker, 1984).
690
Fishery Bulletin 101(3)
Espiritu
Santo
North
(n=7,654)
Espiritu
Santo
South 2100
(n=1 3,754) 2159
Time
Percent ^^'^^
Arrivals qIOO-
0159
B
0900-
0959
Figure 4
Percentage of visits by 23 yellowfin tuna at dif-
ferent times of day at monitoring stations ESN
(A) and ESS (B). Circle with dashed line indi-
cates an equal percentage of returns occurring
every hour of the day.
60,
A
50-
40-
Espiritu Santo North
30
n=7,654
20-
1
_ 10-
o
I
1.
0)
1 1 1 1 1 1 1 1 i I 1 T — 1 1 r — r -T 1 1
Perce n
B
40-
Espiritu Santo South
n=1 3,754
30-
20-
10-
1
ll.
R^CMTrin^csiTj-io^csi-^io^CNTj-in^P
'-'00000000000000000°
■^oooooooooooooooo
O000O00000C3O0OOO0
oinoin^u^omomoioomoino
o^co-^o^cots-o^co-^o^co-^o
OOOOO^^r-fNCNCSICNCncOCOCO'^
OOOOOOOOOOOOOOOOO
Period of presence (hh:mm;ss)
Figure 5
Percentage of visits of increasing duration recorded for 23 yellowfin tuna
at monitoring stations ESN (A) and ESS (B). A single detection of a
tagged fish would be placed in the 00:00-min time category.
Conclusions
Twenty-three yellowfin tuna were tagged with coded
ultrasonic beacons during a five-month period between
11 April and 12 September 1998. These tunas were cap-
tured, tagged, and released <150 m from two monitoring
stations: Espiritu Santo North (ESN) and Espiritu Santo
South (ESS), which were separated by 500 m, at the
Espiritu Santo Seamount in the southern Gulf of Califor-
nia (24°42'N: 110°18'W). The monitors were deployed for
a period of .'SO months, ranging from April 1998 to October
2000, during which they recorded tagged tunas swimming
within their 150 m range of reception. The tunas ranged
in length between 71.0 and 155.0 cm TL and in mass from
7.2 to 73.1 kg. The tunas stayed at the Espiritu Santo
Seamount for varying time periods. Nine of the 23 tunas
left the seamount on the same day that they were tagged.
Two of the nine returned to the seamount twice for a single
day, one within a week of tagging and another after 2.5
months. Five additional tunas stayed at the seamount for
intermediate periods, ranging from two to six weeks. Four
individuals stayed for longer periods of time, ranging from
6 tol8 months. Tunas were present at the seamount at all
times of the day. They moved in and out of the range of the
monitors, most often staying for periods <14:59 min. Fifty-
three percent of the visits to ESN and 37*7? of the visits
to ESS were of this duration. Smaller percentages of the
visits, 2K^ and 20'/r^ , lasted 15:00 to 59:59 min, respectively.
The majority of visits were <1 hour in duration and only
a few exceeded an hour. The intervals spent away from
the seamount were also brief Sixty percent of all absences
at ESN and 65';; of the absences from ESS were <1 hour.
Ninety percent of the visits to both sites were <5 hours.
Only 0.17c of the visits exceeded 23 hours. Tuna individuals
NOTE Kllmley et al : Occurrence of Thunnus albacares in the Gulf of California
691
100.o^
I
A
Iv^ -•- Espiritu Santo North (W=7841)
> 100-
B
c
^W,^ -•- Espiritu Santo Soulti(W=13. 854)
cn
ra
^^•^■^
\
0 2 4 6 8 10 12 14 16 18 20 22 24
Interval t between visits (hrs)
Figure 6
Log-survivor plots of percentages of intervals between successive
tuna arrivals greater than time ( over 24-h periods at monitoring
stations ESN (A) and ESS (B).
may use the site either as a landmark during their migra-
tory transit or as a feeding destination as suggested by the
short and long periods of time spent at the seamount.
Acknowledgments
We would like to thank those on the staffof Centro de Inves-
tigaciones Biologicas del Baja Norte of La Paz, Mexico, who
helped us tag yellowfin tuna at seamount Espiritu Santo.
This work was funded by the Biological Oceanography
Program of the National Science Foundation (grant: OCE-
9802058) and CONACYT of Mexico (grant: PN-9509-1995
and PN-1297-1998).
SOt
A
45^
40-
35-
Espiritu Santo North
30-
(t > 1 day, N=72)
25
20-
15-
10-
m
5-
Mil .1
1 1 1 I i"i i™i"i"i 1 1 i"; i™i i"i
o T-cvjco-^Lncoi^cpcpo
■^ oooooooocS'T
o
°- 50.
B
45-
40^
35-
30
Espiritu Santo South
(t> 1 day, /V=114)
25-
20
15-
10-
5
IbI_bbI| _■-
O^ 1 1 1 1 1 1 1 1 1 I I 1 1 1 1 i 1 :
'-cvjco'^tncor^cocno
OOOOOOOOOT
'-fytofi/itor^ooo) '^
Interval between visits (days)
Figure 7
Percentages of visits of greater than one day made
by tuna to two monitoring stations with single day
intervals ranging from 2-9 days and 10-day inter-
vals ranging from 10-19 to 90-99 days.
Literature cited
Boehlert, G. W., and A. Genin.
1987. A review of the effects of seamounts on biologi-
cal processes. In Seamounts, islands, and atolls (B. H.
Keating, P. Fryer, R. Batiza, and G. W. Boehlert, eds.),
p. 319-334. Geophys. Monogr. Sen 43.
Pagan, R. M., and D. Y. Young.
1978. Temporal patterns of behavior: durations, intervals,
latencies, and sequences. In Quantitative ethology (P. W.
Colgan, ed), p.78-114. John Wiley & Sons, New York, NY.
Fonteneau, A.
1991. Seamounts and tuna in the tropical Atlantic. Aquat.
Living Resour. 4:13-25.
Holland, K. N.
1996. Biological aspects of the association of tunas with
FADs. SPC Fish Aggregating Device Information Bull.
2:2-7.
Holland, K. N., P. Kleiber, S. M. Kajiura.
1999. Different residence times of yellowfin tuna, Thun-
nus albacares, and bigeye tuna, T. obsesus, found in mixed
aggregations over a seamount. Fish. Bull., 97:392-395.
Klimley A. P.
1993. Highly directional swimming by scalloped hammer-
head sharks, Sphyrna lewini, and subsurface irradiance,
temperature, bathymetry, and geomagnetic field. Mar.
Biol. 117:1-22.
1985. Schooling in the large predator, Sphyrna lewini, a
species with low risk of predation: a non-egalitarian state.
Ethology, 70:297-319.
Klimley A. P, and S. B. Butler
1988. Immigration and emigration of a pelagic fish assem-
blage to seamounts in the Gulf of California related to water
mass movements using satellite imagery. Mar. Ecol. Progr.
Ser 49:11-20.
Klimley, A. P., and C. Holloway.
1999. Homing synchronicity and schooling fidelity by yel-
lowfin tuna. Mar Biol. 133: 307-317.
Magnuson, J. J.
1978. Locomotion by scombrid fishes: hydrodynamics, mor-
phology, and behavior Fish Physiol. 239-313.
Parker, R. L., L. Shure, and J. A. Hildebrand.
1987. The application of inverse theory to seamount
magnetism. Rev. Geophys. 25:1-65.
692 Fishery Bulletin 101(3)
Moore, H. L. Walker, M. M.
1951. Estimation of age and growth of yellowfin tuna (A^eo- 1984. Learned magnetic field discrimination in yellow-
thunnus macropterus) in Hawaiian waters by size fre- fin tuna, Thunnus albacares. J. Comp. Physiol 155:673-
quencies. Fish. Bull. 52:131-149. 679.
693
Larvae of Dactylopsaron dimorphicum
(Perciformes: Percophidae) from oceanic islands
in the southeast Pacific
Mauricio F. Landaeta
Laboratorio de Oceanografia Pesquera y Ecologia Larval
Departamento de Oceanografia
Universidad de Concepcion
Casilla 160-C
Concepcion, Chile
Francisco J. Neira
Faculty of Fisfienes and Marine Environment
Australian Mantime College
PO Box 21
Beaconsfield, Tasmania 7270, Australia
Leonardo R. Castro
Laboratorio de Oceanografia Pesquera y Ecologia Larval
Departamento de Oceanografia, Universidad de Concepcion
Casilla 160-C
Concepcion, Cfiile
E-mail addres (for L R Castro, contact author): lecastro@udec.cl
Percophids are a family of small marine
benthic fishes common over soft bot-
toms from inshore to the outer slopes
in tropical to teinperate regions of the
Atlantic and in the Indo-West and
southeast Pacific (Reader and Neira,
1998; Okiyama, 2000). Five species
belonging to four genera have been
recorded around the Salas y Gomez
Ridge in the southeast Pacific, all of
which are endemic to the area except
for Chrionema chryseres, a species
which also occurs off the Hawaiian
Islands and Japan (Parin, 1985, 1990;
Parin et al., 1997). Of these five species,
larval stages have been described only
for Osopsaron karlik and Chrionema
pallidum (Belyanina 1989, 1990).
Dactylopsaron dimorphicum (Parin
and Belyanina, 1990) is a dwarf per-
cophid (29 mm maximum body length)
previously recorded only at the Cupole
(26°S; 86°W) and Baral (25°S; 96°W)
seamounts located to the west of
the Salas y Gomez Ridge and at the
junction of this and the Nazca Ridge,
respectively, at depths of 240-345 m
(Parin, 1990; Parin et al., 1997 ). Aduhs
of this monotypic genus differ from
other percophids in that the first dorsal
fin is positioned at the back of the head
and is in line with the mid-operculum,
8-10 digitiform processes are present
on the posterior upper opercular mar-
gin, and expanded lobes are present at
the distal end of the medial branchio-
stegal rays (Parin, 1990). This species
is sexually dimorphic, males have a
thicker and much longer first dorsal-fin
spine than females (Parin, 1990). There
is no information on their reproductive
biology and eggs are unknown (Watson
etal., 1984).
We describe the postflexion larvae
of D. dimorphicum using material col-
lected around Salas and Gomez and
Easter Islands in the southeast Pacific.
We also provide information on the spa-
tial distribution of this species around
both islands, and on how to distinguish
the larvae from those of teleosts with
similar larvae in the area. This note
constitutes the first record of D. dimor-
phicum off Easter Island, as well as the
first record of the larvae in nearshore
waters of both Pacific islands.
Methods
Field work
Larvae were obtained during an ocean-
ographic expedition (CIMAR-5) to
Easter Island (27°10'S; 109°20'W)
and Salas y Gomez Island (26°30'S;
105°20'W), approximately 3750 km
west of Chile, in November 1999. Sam-
ples were collected onboard the Chil-
ean navy research vessel AGOR Vidal
Gormaz by using a bongo sampler
equipped with two conical nets (0.6-m
diameter mouth openings, 3 m long,
350-nm mesh size). The mouth of each
net was fitted with an OSK flowmeter to
estimate volume of water filtered. Tows
were carried out for 10 min obliquely to
the surface from either the maximum
permissible depth in shallow (<200 m)
stations or from 400 m in deeper sta-
tions. Samples around Easter Island
were obtained at 10 stations located
approximately one nautical mile
(nmi) from the coast both during day
and night, and along four transects
(NW-SE and NE-SW) each containing
four stations located at 3, 7, 12, and 20
nmi offshore (Fig. 1). Samples around
Salas y Gomez Island were obtained
along four transects (N-S and E-W),
each containing four stations at 1, 3,
6, and 10 nmi offshore (Fig. 2). Addi-
tionally eight deep stations (>1500 m)
were also sampled between Easter and
Salas y Gomez islands. All samples
were fixed in 5% formalin and later pre-
served in 70% ethanol. Water volume
sampled per tow ranged between 112.6
and 517.7 m-'. Larval abundances were
standardized to 1000 m-^ and mapped
by using SURFER® (Golden Software,
Golden, CO). Statistical analyses were
performed using STATISTICA (Star-
Soft, Inc., Tulsa, OK).
Larval identification and
processing
Postflexion larvae were identified as
those of Dactylopsaron dimorphicum
by a combination of dorsal and anal-fin
meristics (D. IV [III-V] + 22 [20-22]
Manuscript approved for publication
15 January 2003 by Scientific Editor.
Manuscript received 4 April 2003 at NMFS
Scientific Publications Office.
Fish Bull. 101:693-697 (2003).
694
Fishery Bulletin 101(3)
and A. 24 [23-25]; Table 1), and by the presence of the
unique digitiform opercular processes (Parin, 1990; Oki-
yama, 2000). Identification was verified by using fin meris-
tics from cleared and stained specimens (Potthoff, 1984).
26.70
26 90
27.10
5
zr.x>
27 50
Dactylopsaron dimorphicum
A total of 55 postflexion larvae of D. dimorphicum
(8.2-15.3 mm standard length) were examined to describe
morphometries, meristics, and pigmentation. Three larvae
(9.1, 13.1 and 13.4 mm SL) were cleared and stained fol-
lowing the method of Potthoff ( 1984). Terminology
and morphometric measurements follow Neira et
al. (1998). Measurements were made to the near-
est 0.01 mm by using a dissecting microscope
fitted with an eyepiece micrometer Body length
(BL, Neira et al., 1998) in postflexion larvae cor-
responds to standard length (SL), i.e. tip of snout
to posterior margin of hypurals. Measurements of
body depth (BD), head length (HL), and preanal
length (PAL) were converted to a percentage (%) of
SL (Table 2). Eye diameter (ED) and snout length
(SnL) were converted to a percentage (%} of HL.
Pigment described refers solely to melanin. Illus-
trations were made with the aid of a camera lucida.
Larvae per 1000 m
o 0
1 -5
5-10
10-50
50-100
>100
109 80
109 60
109 40
Longitude (W)
109 20
109 00
Figure 1
Spatial distribution of postflexion larvae of Dactylopsaron dimorphi-
cum (numbers/1000 m^) around Easter Island in November 1999.
ID
■a
3
itl
Dactylopsaron dimorphicum
26 35
rtJ
o
o
26.45
Salasy Gbmez Island
0
o
0
•
0
• Laivae per 1000 m^
26 55-
o 0
• 1-5
o
# 5-10
% > 10
26.66
106.56
105.45
105.35
Longitude (W)
105.25
Figure 2
Spatial distribution of postflexion larvae of Dactylopsaron dimorphi-
cum (numbers/1000 m^) around Salas y G6mez Island in November
1999.
Results
Description of larvae
Postflexion larvae of Dactylopsaron dimorphicum
are elongate (BD 13.1-18.3%; Table 2), and have
a moderate to large head (HL 28.6-36.7%) and
an elongate snout (Fig. 3). Eyes are round and
pigmented by 8.2 mm SL. The mouth is large, pro-
trusible, and has a long ascending premaxillary
process giving a characteristic "duckbill" appear-
ance. Small villiform teeth are present along the
premaxilla and dentary. There are no head spines.
The digitiform processes on the upper opercular
margin are present in larvae >13.2 mm SL; the
lower, rearward-directed process that reaches the
end of the pectoral-fin base in adults was still
forming in the largest larva examined (15.3 mm
SL). The branchiostegal membranes are free from
the ithsmus. The short first dorsal fin is located at
the nape and lies in line with the mid-operculum;
pterygiophores of the five first-dorsal fin spines in
two of the cleared and stained larvae ( 13.1 and 13.4
mm SL) were located between the neural spines of
the second and third trunk vertebrae. The 9. 1-mm-
SL cleared and stained larva possessed only 15
of the 17-20 pectoral-fin rays, and first dorsal-fin
spines were developing. The elongate pelvic fins
are thoracic, i.e. inserted in front of the pectoral-
fin bases. Lateral line scales begin to form at >13
mm SL. Larvae are unpigmented, although a few
had a small melanophore at the base of the 17 or
18* dorsal-fin ray There are 31-35 myomeres. The
number of vertebrae in the cleared and stained
larvae is 34-35 (11-12 -h 22-24).
Larval distribution
Postflexion D. dimorphicum larvae were collected
within 6 nmi off both Easter and Salas and Gomez
NOTE Landaeta et a\ Larvae of Dactylopsaron dimorphicum in the southeast Pacific
695
Figure 3
Postfiexion larvae of Dactylopsaron dimorphicum collected around Easter Island in
November 1999. (A) 8.5 mm SL; (B) 14.8 mm SL. Illustrated by F. J. Neira.
Table 1
Meristic counts of percophid species recorded in submarine ridges of Salas y Gomez Island in the southeastern Pacific (from Parin
11985, 19901 and Okiyama [2000]).
Dorsal
Anal
Pectoral Pelvic Caudal (branched) Vertebrae
Chrionema chryseres
Chrionema pallidum
Dactylopsaron dimorphicum
Enigmapercis acutirostris
Osopsaron karlik
VI +16
26
23
1,5
15(11)
27-28
VI +14-15
18
20-22
1,5
15(11)
27-28
IIII-V] + 22 [20-22]
24 (23-251
18 [17-20]
1,5
14(8)
34-35
11+21
25
21
1,5
15 (8-9)
—
V- VI +19-20
22-23
19-20
1,5
14
32
This study.
Table 2
Standard length range (mm), and mean values (±1 SD) of
selected body proportions (given as a percentage of body
length) of postfiexion larvae o( Dactylopsaron dimorphi-
cum from Easter and Salas y Gomez islands in the south-
eastern Pacific Ocean («=55).
Standard length (mm)
8.2-15.3
Head length (%SL)
28.6-36.7 (32.3 ±3.3)
Eye diameter (%HL)
18.2-24.7 (22.0 ±3.1)
Snout length (%HL)
25.6-36.2 (31.0 ±5.51
Body depth (7fSL)
13.1-18.3 (15.3 ±2.2)
Preanal length C/fSL)
45.4-56.9 (49.5 ±4.0)
Islands (Figs 1 and 2). Around Easter Island, larvae were
caught only in nearshore stations ( <2 nmi ) over the narrow
shelf and were more abundant along the southern edge.
8 9 10 11 12 13 14
Body length (mm)
Figure 4
Combined body length (SL, mm) frequency distribution of
postfiexion larvae of Dactylopsaron dimorphicum around
Easter Island and Salas y Gomez Island in November 1999.
696
Fishery Bulletin 101(3)
The highest larval concentrations (>100 larvae/1000 m'')
occurred at the southeastern tip of the island and averaged
27 ±46 larvae/1000 m'' (Fig. 1). No significant differences
were found between day and night larval concentrations
(Kruskal-Wallis test=0.047; P>0.05). Around Salas and
Gomez Island, larvae were caught only at two stations
6 nmi west and south of the island, and in mean concentra-
tions <10 larvae lOOO/m-^ (Fig. 2). No larvae were caught in
any of the eight stations sampled between the two islands.
Body lengths of larvae caught in both islands ranged from
8 to 16 mm SL, and over 30% of the larvae were around 12
mm SL (Fig. 4).
Discussion
Postflexion larvae of D. dimorphicum are likely to be
confused with those of four other co-occurring percophid
species (see Table 1), and those of the creediid Crystallo-
dytes pauciradiatus that occur in the same region (Castro
and Landaeta, 2002) and have similar bodies with little or
no pigment. In the case of the percophids, the digitiform
opercular processes exclusive to D. dimorphicum , together
with dorsal and anal-fin meristics, should be sufficient
to distinguish between postflexion larvae of all species.
Larval C. pauciradiatus can be identified by using myo-
mere counts (56-58 vs. 31-35 in D. dimorphicum ) and their
small, early forming posterior preopercular spines (Reader
et al., 2000).
Our collection of D. dimorphicum larvae at Easter Island,
some 453 km to the southwest of Salas y Gomez Island
where it was first described (Parin, 1990), constitutes the
first record for Easter Island, thereby extending the known
range of this species over the South Pacific plate. Despite
numerous past fish surveys around Easter Island (i.e. Ran-
dall and Cea-Egaiia, 1984; Mujica, 1993), adults of this dwarf
percophid had not been reported there, a fact that could be
attributed to factors such as collection methods, depth of
surveys, and the very small size of these larvae. However, the
presence and abundance of larval D. dimorphicum reported
in this study, and the fact that they were among the five most
abundant larval taxa caught around Easter Island (Castro
and Landaeta, 2002), implies the existence of a well-estab-
lished breeding population. Biogeographically, this finding
also suggests that larval drift could play an important role in
the expansion of this and other fish species that have pelagic
larvae in this region of the southeast Pacific. In this context,
it is perhaps relevant that expansions offish ranges are not
uncommon in this region, even though both Easter and Sa-
las y Gomez Islands lie in different biogeographic provinces
(Parin et al., 1997). A good example is the pentacerotid Pen-
taceros decacanthus, which was regarded as endemic of the
Nazca and Salas y Gomez Ridges until it was recorded in
Easter Island (Parin and Kotlyar, 1988).
Acknowledgments
We would like to thank Paula Rosenborg and all the crew
from the AGOR Vi.dal Gormaz for their help with sampling.
We also thank Muneo Okiyama for his comments on iden-
tification of percophid larvae. This research was funded
by the Comite Oceanografico Nacional (CONA), Chile, and
forms part of a study on distribution patterns and larval
accumulation around oceanic islands headed by Leonardo
Castro.
Literature cited
Belyanina, T. P.
1989. Ichthyoplankton in the regions of the Nazca and Salas
y Gomez submarine ridges. J. Ichthyol. 29(5):84— 90.
1990. Larvae and fingerlings of little-known benthic and
benthopelagic fishes from the Nazca and Salas y Gomez
ridges. J. Ichthyol. 30(6):1-11.
Castro. L. R., and M. F. Landaeta.
2002. Patrones de distribucion y acumulacibn larval en tomo
a islas oceanicas: Isla de Pascua y Salas y Gomez. Cienc.
Tecnol. Mar. CONA 25(1):131-145.
Mujica, A.
1993. Zooplancton de las aguas circundantes a la Isla de
Pascua (27°08'S-109°26W). Cienc. Tecnol, Mar. CONA
16:55-61.
Neira, F. J., A. G. Miskiewicz, and T. Tmski.
1998. Larvae of temperate Australian fishes. Laboratory
guide for larval fish identification, 474 p. Univ. Western
Australia Press, Nedlands, Australia.
Okiyama, M.
2000. Percophidae (sandfishes, duckbills). //; The larvae of
Indo-Pacific coastal fishes: an identification guide to marine
fish larvae (J. M. Leis and B. M. Carson-Ewart, eds.), p.
554-560. Brill, Leiden, The Netheriands.
Parin, N. V.
1985. A new hemerocoetine fish, Osopsaron karlik ( Percophi-
dae, Trachinoidei) from the Nazca submarine ridge. Jpn.
J. Ichthyol. 3 1(4 1:358-361.
1990. Percophid fishes (Percophidae) from the Salasy Gomez
ridge (Southeast Pacific). J. Ichthyol. 30(l):68-79.
Parin, N. V., and A. N. Kotlyar
1988. A new boarfish, Pentaceros quinquespinis (Pentacero-
tidae), from the Southeast Pacific. Vopr Ikhtyol. 28(3):
355-360.
Parin, N. V., A. N. Mironov, and K. N. Nesis.
1997. Biology of the Nazca and Sala y Gomez submarine
ridges, an outpost of the Indo-West Pacific fauna in the
Eastern Pacific Ocean: composition and distribution of the
fauna, its communities and history. Adv. Mar Biol. 32:
147-242.
PotthofT, T
1984. Clearing and staining techniques. In Ontogeny and
systematics of fishes (H. G. Moser, W. J. Richards. D. M.
Cohen, M. P. Fahay A. W. Kendall, and S. L. Richardson,
eds.). p. 35-37. Am. Soc. Ichthyol. Herpetol. Special Pub-
lication 1.
Randall, J. E., and A. Cea-Egana.
1984. Native names of Easter Island fi.shes, with comments
on the origin of the Rapanui people. Occas. Pap. Bernice
P Bi.shop Mus. 25(12):1-16.
Reader, S. E., and F J. Neira.
1998. Percophidae: sandfishes, duckbills. In Larvae of
temperate Australian fishes. Laboratory guide for larval
fish identification IF J. Neira, A. G. Miskiewicz, and T.
Tmski. eds.). p. 358-361. Univ. Western Australia Press;
Nedlands, Australia.
NOTE Landaeta et a\ Larvae of Dactylopsaron dimorphicum in the southeast Pacific 697
Reader, S. E., J. M. Leis, and D. S. Rennis. Watson, W., A. C. Matarese, and E. G. Stevens.
2000. Creediidae (tommyfishes). In The larvae of Indo- 1984. Trachinoidea: development and relationships. In
Pacific coastal fishes. An identification guide to marine fish Ontogeny and systematics of fishes (H. G. Moser, W. J. Rich-
larvae (J. M. Leis and M. Carson-Ewart leds.). p. 575-578. ards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr, and S. L.
Brill, Leiden, The Netherlands. Richardson, eds.), p. 554-561. Am. Soc. Ichthyol. Herpetol.
Special Publication 1.
698
Assessment of sampling methods to estimate
horseshoe crab (JLimu/us polyphemus L.)
egg density in Delaware Bay
Penelope S. Pooler
David R. Smith
us. Geological Survey
Leetown Science Center
11700 Leetown Road
Kearneysville, West Virginia 25430
E-mail address (for D. R Smith, contact auttior): david_r_smitti@usgs gov
Robert E. Loveland
Department of Ecology and Evolution
Cook College
Rutgers University
New Brunswick, New Jersey 08901
Understanding the reliability of egg
density estimates at multiple scales
will help develop effective monitoring
programs.
We addressed all three questions
with respect to eggs found in both
shallow (0-5 cm) and deep (0-20 cm)
sediments. Horseshoe crabs are gener-
ally thought to lay most of their eggs at
a depth of 15-20 cm (Brockmann, 1990;
Botton et al., 1994). Through processes
of bioturbation and wave-generated
sediment activation, horseshoe crab
eggs are brought onto the beach and
made available to foraging shorebirds
(Botton et al., 1994; Kraeuter and Feg-
ley, 1994; Jackson et al., 2002).
Materials and methods
Mark L. Botton
Fordtiam University
113 West 60"' Street
New York, New York 10023
Stewart F. Michels
Delaware Division of Fisfi and Wildlife
RO. Box 330
Little Creek, Delaware 19961
Each spring horseshoe crabs {Limtilus
polyphemus L.) emerge from Delaware
Bay to spawn and deposit their eggs on
the foreshore of sandy beaches (Shuster
and Botton, 1985; Smith et al., 2002a).
From mid-May to early June, migra-
tory shorebirds stopover in Delaware
Bay and forage heavily on horseshoe
crab eggs that have been transported
up onto the beach (Botton et al., 1994;
Burger et al., 1997; Tsipoura and
Burger, 1999). Thus, estimating the
quantity of horseshoe crab eggs in
Delaware Bay beaches can be useful
for monitoring spawning activity and
assessing the amount of forage avail-
able to migratory shorebirds.
We evaluated procedures to estimate
horseshoe crab egg density by asking
three questions that address sampling
at a different spatial scale. 1) How
many samples of sediment are needed
for precise estimation of egg density
within a segment of beach? 2) Does egg
density within a segment of beach ad-
equately represent egg density across
a larger stretch of beach? 3) How many
beach segments should be sampled
to monitor bay-wide egg density? We
chose these three questions because
the objective of egg studies might focus
on any of these scales. We ask the first
question to determine the sampling
effort necessary to detect changes in
egg density over time within a specific
beach segment. The second question
allows us to examine the reliability of
using egg density in a beach segment to
infer egg density over a larger stretch
of beach. The third question deals with
the level of precision in estimates of
bay-wide egg density and how many
beaches must be sampled to detect
bay-wide declines in density over time.
During May and June 1999, we col-
lected sediment on 16 beaches in
Delaware Bay (Fig. 1), eight along the
eastern shore (New Jersey) and eight
along the western shore (Delaware), to
estimate egg density. Methods used to
collect sediment and extract horseshoe
crab eggs are summarized in the pres-
ent study, but are presented in detail
in Smith et al. (2002b). Beach sediment
was collected in cores (5 cm diameter)
within a 3-m wide strip along a 100-m
segment of beach. Each 3-m wide strip
was centered on the mid-beach eleva-
tion where a majority of horseshoe
crab nests occur (Botton et al., 1988).
The mid-beach elevation is halfway
between the spring high water level
and the beach break at the low tide ter-
race. Within each egg-sampling strip,
40 locations were selected randomly for
sediment collection. At each location, a
pair of core samples was taken: one to a
depth of 5 cm and the other to a depth
of 20 cm. We sampled eggs on 25-26
May and 14-15 June 1999, which fol-
lowed the heaviest spawning activity
in Delaware Bay that year (Smith et
al., 2002a). We mixed the entire core
contents thoroughly and then removed
80-mL aliquots. We ran the aliquots
Manuscript approved for publication
12 February 2003 by Scientific Editor.
Manuscript received 4 April 2003 at NMFS
Scientific Publications Office.
Fish Bull. 101:698-703 (2003).
NOTE Pooler et al.: Assessment of sampling methods to estimate egg density of Limulus polyphemus
699
Woodland\,^ "^Sea Breeze
NEW JERSEY
\ Raybins
,Fortescue
~\
Kitts Hummockr
Kimbles
/Reeds
North Bowers*
South Cape .
Shore Lab /
/Highs \<
Big Stone\^
North (
)
Slaughter^
DELAWARE powieK
Prime Hool^^
Cape May /
^
BroadkilK^
«i =
10 0 10 20 Kitometers
~1
s
Figure 1
Delaware Bay beaches ( • ) where horseshoe crab eggs were sampled
in May £ind June 1999.
through a 1-mm sieve to separate eggs and larvae from
ambient sediments and then counted eggs (dead or live)
and larvae in each aliquot. Depth of aerobic sand varied;
thus we measured core volume prior to extrapolating egg
counts to totals per core and then estimated the total den-
sity of eggs and larvae. The larvae comprised a small frac-
tion of total eggs and larvae, and for the purposes of this
paper we evaluated the sampling of eggs only.
Question 1 : How many sediment cores should be
sampled per beach segment?
We addressed this question in two steps. First, we deter-
mined the precision of egg-density estimates as a function
of egg density and sample size. Second, we translated the
precision of the estimates into statistical power to detect
change in egg density over time. For simplicity, variance
of the egg-density estimate was calculated from a random
sample from an infinite population. Coefficient of variation
(CV) was calculated as
CV = Vvar(.v)/H/y,
where var(jy) = variance of eggs among cores; and
y = egg density.
We modeled the relationship between egg density and
variance among cores (i.e. var[y]=/'[ y] to predict coefficient
of variation (CV) for different sample sizes and across
the observed range of egg densities (i.e. CV = ■]f\y\l n jy).
Using predicted CVs, we estimated the probability of de-
tecting a change in egg density over time. The probability
of detecting decline (i.e. statistical power) was calculated by
using a one-tailed ^test with a type-I error rate of 0.2 and a
constant rate of annual change for CVs = {0.1, 0.2, 0.3, 0.4)
with the software program TRENDS (Gterrodette, 1993).
Question 2: Is egg density within a beach segment
representative of egg densities along a larger
stretch of beach?
Smith et al. (2002b) modeled the relationship between
counts of spawning females and egg densities within beach
segments. Spawning females are counted annually as part of
a bay-wide survey of spawning activity (Smith et al., 2002a),
and in 1999, egg sampling was conducted on some of the
same beaches as the spawning survey (Smith et al., 2002b).
For eggs that were sampled in May 1999 on six New Jersey
beaches, the relationship was fairly strong, linear, and pre-
dictive (/•2=0.62; Smith et al., 2002b). Although we sampled
for eggs on only one 100-m segment of beach, we used the
above relationship to predict egg densities for all 100-m
segments along the beach where spawning females were
counted. We limited the predictions to the six New Jersey
beaches where we felt the relationship between spawning
females and egg densities was sufficiently strong (Smith
et al., 2002b). We compared egg density in the observed
100-m segment to the distribution of densities predicted in
all 100-m segments on the beach. If the observed density
was within the interquartile range of the distribution of
predicted densities, we concluded that the 100-m segment
was representative of the larger stretch of beach.
Question 3: How many beaches should be sampled?
Using the observed variation in egg density among the 16
beaches sampled in 1999, we predicted the CV for bay-wide
egg density estimates as a function of the number of beaches
sampled and under a stratified sampling design where the
two strata were New Jersey smd Delaware. We could not
evaluate CV across a range of bay-wide densities because
the 1999 results provided only one datum point, and we
expected variation among beaches to be a function of egg
700
Fishery Bulletin 101(3)
Table 1
Mean eggs per core and standard errors (SE) for horseshoe
crabs [Limul
IS polyphemus) at 16 beaches sampled in Delaware Bay
that were
sampled in May and June 1999. Cores were 5 cm
in diameter. At 40 random locations on
each beach, a pair of
sediment
cores were
sampled: one core at 5 cm
depth (shallow sediment) and another at 20 cm
depth (deep sediment).
No. of eggs per core on 25-26 May 1999
No. of eggs per core on 14-15 June 1999
Shallow
Deep
Shallow
Deep
State
Beach
sediment
SE
sediment
SE
sediment
SE
sediment
SE
Delaware
Broadkill
0.0
0.00
1.5
1.47
1.2
0.36
101.7
60.42
Prime Hook
0.2
0.08
81.9
76.17
7.4
2.06
223.9
112.14
Fowler
0.1
0.05
1.8
0.65
2.7
1.22
211.3
116.56
Slaughter
11.7
2.69
814.8
186.04
41.7
5.35
664.5
97.81
Big Stone
0.1
0.05
11.3
5.09
0.7
0.53
24.2
14.27
North Bowers
23.0
6.49
950.3
234.18
105.1
23.54
400.4
70.81
Kitts Hummock
26.4
8.23
325.1
78.63
15.2
5.49
124.8
43.94
Woodland
0.5
0.17
0.1
0.06
7.0
3.68
60.9
29.75
New Jersey North Cape May
0.3
0.25
0.0
0.00
0.5
0.33
0.7
0.37
South Cape Shore Lab.
25.5
0.86
1085.4
140.29
4.5
0.81
1399.2
144.03
Highs
2.1
0.71
1128.6
96.99
4.4
0.94
1456.8
173.80
Kimbles
9.7
4.80
1561.3
286.32
1.7
0.55
1008.0
105.63
Reeds
2.0
0.52
540.4
79.90
18.2
2.52
468.0
62.67
Raybins
3.5
1.91
65.8
43.88
0.1
0.06
6.7
4.57
Fortescue
2.0
0.43
645.9
108.71
20.6
3.85
465.7
193.64
Sea Breeze
27.5
7.95
347.3
94.70
0.2
0.09
3.1
2.01
density. However, we examined the probability of detecting
a percentage change in bay-wide egg density over time as a
function of the number of beaches sampled by using the 1999
bay-wide egg density as the initial value in the time series.
Results
When the objective is to monitor egg density within a seg-
ment of beach, a sample size of 40 sediment cores is suf-
ficient for detecting substantial changes in egg density in
the top 20 cm of sediment, but >40 cores would be needed to
monitor egg density in the top 5 cm of sediment. Distribu-
tions of egg densities were skewed right with median densi-
ties of 3 and 275 eggs per core for shallow and deep cores,
respectively (Table 1 ). A sample size of 40 cores resulted in
a CV of 0.26 for a median density of eggs 0-20 cm deep (Fig.
2B). In contrast, about 100 cores would need to be sampled
to bring the CV down to 0.3 when .sampling shallow sedi-
ment and when egg density was at the median (Fig. 2A). A
CV of 0.3 corresponds to a 75% chance of detecting a 509f
decline in egg density over 5 years (Fig. 3A) and an 80%
chance of detecting a 40% decline over 10 years (Fig. SB).
A sample size of 60 shallow cores would result in CV of 0.4
for median egg density (Fig. 2A), which would be sufficient
for monitoring over 10 years, but not over 5 years. A CV
of 0.4 would lead to a better than 85% chance of detecting
a 50% decline in density over 10 years (Fig. 3B). Precision
and power would improve when sampling higher densities
of eggs (Fig. 2).
At most beaches, observed egg densities within a 100-m
segment of beach were not representative of egg densities
throughout a larger beach. On only two of the six New Jer-
sey beaches examined (South Cape Shore Lab and Reeds)
did the observed egg density fall within the interquartile
range of beach-wide densities (Fig. 4). On three beaches
the observed egg density was greater than all predicted
densities, and on one beach observed egg density was less
than all predicted densities.
With egg density at the 1999 level and sampling at 16
beaches (i.e. eight beaches per state) distributed throughout
the bay, the CV for densities of eggs in 0-20 cm of sediment
was 0.26 in May and 0.29 in June (Fig. 5). For densities of
shallow eggs, the CV was 0.33 for egg densities in May and
0.43 in June. Variability in egg densities among beaches was
greater for sampling in June 1999 than in May 1999.
Discussion
Eggs in shallow sediment (0-5 cm) consistently yielded
lower densities and higher variability than eggs in deep
sediment (0-20 cm). A sample size of 40 sediment cores was
sufficient for estimating and monitoring density of eggs
0-20 cm deep within a 100-m beach segment. However, a
larger sample size (260 sediment cores) would be needed
for estimating and monitoring density of eggs 0-5 cm deep
within a segment of beach.
Because egg density in a 100-m segment of beach is not
necessarily representative of the larger surrounding beach.
NOTE Pooler et al : Assessment of sampling methods to estimate egg density of Limutus polyphemus
701
S 0.6
a 0.4
0) 0.2
0.0
500
1000
1500
Density of eggs per core
Figure 2
The relationship between density and coefficient of variation (CV) for (A) shallow
sediment ( 0-5 cm ) and ( B ) deep sediment ( 0-20 cm ). Curves in each figure depend
on sample size: circle is n = 20, triangle is a; = 40, square is n = 60, diamond is
n = 80, and x is n = 100. Vertical lines represent median egg densities that we
observed in 1999.
09
= 0,6
0.4
0 3 0 4 0 5
Rate of decline over five years
0.3 04 0.5
Rate of decline over ten years
Figure 3
Probability of detecting a decline (i.e. statistical power) for various magnitudes
of decline and for annual surveys over five (A) and 10 IB) years. Statistical power
was calculated for a one-tailed f-test with a type-I error rate of 0.2, and a constant
annual rate of change.
it is important to realize that if sampling is restricted to
a short segment of beach, then the scope of inference is
strictly limited to that segment. If a rehable estimate of
egg density along a beach is required, then it will be nec-
essary to take samples along the entire beach. Because of
the logistics of sampling sediment it would be difficult to
sample throughout a long stretch of beach in one stage of
sampling. However, a two-stage sampling design could be
considered in which beach segments are selected at the 1*'
stage and sediment cores within segments are selected at
the 2"'' stage.
Consistent with our findings on sampling within a beach,
bay-wide egg density can be more precisely estimated for
eggs 0-20 cm deep than for eggs 0-5 cm deep. A stratified
702
Fishery Bulletin 101(3)
0 15000 30000 45000 60000 0 15000 30000 45000 60000
Predicted density of egg cores
Figure 4
Density curves of predicted egg densities on 100-ni beach
segments at six New Jersey beaches. An arrow represents
the egg density in the observed 100-m segment. These
density curves were generated by dividing the area sur-
veyed for spawning females into 100-m segments and
using the observed relationship between egg densities
and spawning females to predict egg density for each seg-
ment. The beaches shown are (A) Fortescue. (B) Highs,
(C) Kimbles, (D) North Cape May, (E) Reeds, and (F) South
Cape Shore Laboraton,'.
random sample of eight beach segments per state would
result in CV sO.3 for estimates of egg densities 0-20 cm
deep. If this level of effort were maintained, it would be
sufficient to detect biologically significant declines in egg
density over a 5- or 10-year period. However, greater effort
would be required to monitor change in egg densities 0-5
cm deep. According to results from the May samples, to
estimate egg densities in shallow sediment with CV sO.3,
a stratified random sample of 10 segments per state would
be required.
Sampling eggs is a costly process; therefore sampling
efficiency and reducing sample size are important consider-
ations. Although sediment can be collected quickly, the pro-
cess of extracting and enumerating eggs from the sediment
can be time consuming. Quantifying the eggs in surface
sediments to assess shorebiixi forage biomass is likely to be
the main objective of many egg sampling studies because
horseshoe crab spawning activity can be assessed by other
methods, such as through counts of spawning horseshoe
crabs (Smith et al., 2002a). However, a primary finding
in the present study is that estimating eggs in 0-5 cm
5 7 10 12 15 17 20 22
Number of beaches sampled per state
Figure 5
Predicted coefficient of variation (CV) shown for the pos-
sible range of number of beaches sampled per state. This
figure is based on the observed levels of bay-wide density
during the two sampling periods in 1999. Curves are based
on egg densities found at different depths and time periods:
triangle is shallow sediment in June, circle is shallow sedi-
ment in May, diamond is deep sediment in June, and square
is deep sediment in May. Shallow sediment is 0 to 5 cm deep,
and deep sediment is 0 to 20 cm deep.
of sediment will be more costly than estimating eggs in
0-20 cm of sediment. In the future, alternatives in survey
design, such as stratification of the beach foreshore, should
be considered to reduce the amount of sediment that needs
to be collected for precise estimates of horseshoe crab egg
density.
Acknowledgments
This work was funded through the USGS/State Partner-
ship Project (no. 99HQAG0050). Additional funding was
provided through the New Jersey Endangered & Nongame
Species Program.
Literature cited
Botton, M. L., R. E. Loveland, and T. R. Jacobsen.
1988. Beach erosion and geochemical factors: influence on
spawning success of horseshoe crabs {Limulus polyphemus)
in Delaware Bay Mar Biol. 99:325-332.
1994. Site selection by migratory shorebirds in Delaware
Bay, and its relationship to beach characteristics and
abundance of horseshoe crab (Limnlus polyphcmus) eggs.
Auk 111:605-616.
Brockmann, H. J.
1990. Mating behavior of horseshoe crabs, Limulus
polyphemus. Behaviour 114:206-220.
Burger, J.. L. Niles, and K. E. Clark.
1997. Importance of beach, mudflats and marsh habitats to
migrant shorebirds on Delaware Bay. Biol. Conserv. 79:
283-292.
NOTE Pooler et a\: Assessment of sampling methods to estimate egg density of Limulus polyphemus
703
Gerrodette, T.
1993. TRENDS: software for a power analysis of linear
regression. Wildl. Soc. Bull. 21:515-516.
Jackson, N. L., K. F. Nordstrom, and D. R. Smith.
2002. Geomorphic-biotic interactions on beach foreshores in
estuaries. J. Coast. Res. 36:414-^24.
Kraeuter. J. N., and S. R. Fegley.
1994. Vertical disturbance of sediment by horseshoe crabs
(Limulus polyphemus) during their spawning season. Es-
tuaries 17:288-294.
Shuster, C. N., Jr., and M. L. Botton.
1985. A contribution to the population biology of horseshoe
crabs, Limulus polyphemus (L.), in Delaware Bay. Estu-
aries 4:363-372.
Smith, D. R., R S. Pooler, B. L. Swan, S. F. Michels, W. R. Hall,
R J. Himchak, and M. J. Millard.
2002a. Spatial and temporal distribution of horseshoe crab
{Limulus polyphemus) spawning in Delaware Bay: implica-
tions for monitoring. Estuaries 25:115-125.
Smith, D. R., P. S. Pooler, R. E. Loveland, M. L. Botton,
S. F. Michels, R. G. Weber, and D. B. Carter.
2002b. Horseshoe crab (Limulus polyphemus) reproductive
activity on Delaware Bay beaches: interactions with beach
characteristics. J. Coast. Res. 18:730-740.
Tsipoura, N., and J. Burger.
1999. Shorebird diet during spring migration stopover on
Delaware Bay. Condor 101:635-€44.
704
Larval abundance, distribution, and spawning habits
of spotted seatrout iCynoscion nebulosus)
in Florida Bay, Everglades National Park, Florida
Allyn B. Powell
Center for Coastal Fisheries and Habitat Research
National Ocean Service
National Oceanic and Atmospheric Administration
101 Pivers Island Road
Beaufort, North Carolina 28516
E-mail address; allyn. powelliaJnoaa.gov
The spotted seatrout (Cynoscion nebu-
losus ) is one of the most sought after
recreational fish in Florida Bay. and it
spends its entire life history within the
bay (Rutherford et al. ,1989b). The biol-
ogy of adult spotted seatrout in Florida
Bay is well known (Rutherford et al.,
1982, 1989b) as is the distribution and
abundance of juveniles within the bay.
The habitats and diets of juveniles are
well documented (Hettler, 1989; Ches-
ter and Thayer, 1990; Thayer et al.,
1999; Florida Department of Environ-
mental ProtectionM. Nevertheless, the
spatial and temporal spawning habits
of spotted seatrout and the distribu-
tion of larvae have only been partially
described (Powell et al., 1989: Ruther-
ford et al., 1989a i.
An excellent description of the eco-
logical history of Florida Bay is given
by Fourqurean and Robblee (1999).
Briefly, Florida Bay is subtropical and
is generally oligotrophic. The bay is a
network of shallow basins, mud banks,
and mangrove islands (keys). Tides are
influenced by the Gulf of Mexico and
Atlantic Ocean, but mud banks, which
are connected to basins by channels,
restrict circulation in the bay and at-
tenuate tidal energy very quickly. As a
result there is essentially no lunar tide
over most of the central and northeast-
ern portion of the bay.
This impediment to circulation could
have a negative effect on the recruit-
ment of early-stage planktonic larvae
into these portions of the bay. Within the
next few decades, plans to restore the
Everglades include increasing freshwa-
ter flows to Florida Bay. Prerestoration
information on larval distribution and
spawning patterns of spotted seatrout
is a high priority because increased
freshwater flows can have potential
positive and negative impacts. At low
salinities, the planktonic eggs of spot-
ted seatrout could sink to the bottom
and would not be viable (Holt and Holt,
2002; Alshuth and Gilmore^). On the
other hand, increased freshwater flows
can alleviate hypersaline conditions
that could result in an expansion of
the distribution of the early life stages
of spotted seatrout (Thayer et al., 1999;
Florida Department of Environmental
Protection'). The objective of the pres-
ent study is to document the distribu-
tion and abundance of spotted seatrout
larvae to determine their early life his-
tory habitats and spawning habits in
Florida Bay.
Methods and materials
To describe the distribution and
abundance of spotted seatrout larvae
in Florida Bay, I devised a series of
ichthyoplankton surveys between
1994 and 1999. The initial survey was
conducted during nine nonconsecutive
months between September 1994 and
August \i)9r^. A total of 14 fixed sta-
tions wore selected in basins of Florida
Bay (Fig. 1). In accordance with rec-
ommendations by the South Florida
Ecosystem Restoration Prediction and
Modeling (SFERPMi, Program Man-
agement Committee (PMC), Florida
Bay was divided into six zones for ease
of reporting results (Table 1, Fig. 1).
These zones are based on the benthic
moUuscan and benthic plant commu-
nities (Fourqurean and Robblee, 1999).
Paired bongo nets, 60 cm wide, were
fitted with 0.333-mm mesh and fished
from the port side of a 5.4-m boat. Nets
were towed during daylight, approxi-
mately 1 m below the surface for 5
minutes and volume estimates were
obtained from flowmeter readings.
In 1996, sampling was conducted
monthly from April through Septem-
ber at stations where recently hatched
spotted seatrout occurred during
1994-95 (stations 5, 6, 9-13). In 1996, 1
used a paired 60-cm bow-mounted push
nets with 0.333-mm mesh similar to
that described by Hettler and Chester
( 1990). Nets were fished approximately
1 m below the surface for 3 minutes.
The volume of water sampled with the
push net was slightly greater than that
sampled with the bongo nets (60 m'^ vs.
50 m-'). To test the efficiency of the two
gears, both were fished simultaneously
at 23 stations during 1996. A Kruskal-
Wallis nonparametric test was used to
evaluate differences (Sokal and Rohlf
1981 ). No significant differences in den-
sities offish larvae were found between
gear types (P=0.50).
During and after September 1997
sampling for spotted seatrout was
limited to four stations in four zones
(Table 1; stations 6. 15, 16, 17) where
paired bow-mounted push nets were
employed. Sampling occurred during
July and September 1997; March, May,
June, July, and September 1998; and
May, July, and November 1999.
' Florida Department of Environmental
Protection. 1995. Fisheries-indepen-
dent monitoring program, annual report.
Florida Department of Environmental
Protection, Florida Marine Research In-
stitute, 100 8'h Avenue SE. St. Petersburg,
FL 33701.
^ Alshuth, S., and R. G. Gilmore Jr 1994.
Salinity and temperature tolerance limits
for lai'vai spotted seatrout. Cynoscion neb-
(//<).s».s- C. (Pisces: Sciaenidac). Int. Coun.
Explor Sea, Coun. Meet. Pap., ICES-CM-
1994/L: 17. 19 p.
Manuscript accepted for publication
19 February 2003 by Scientific Editor
Manuscript received 4 April 2003 at NMFS
Scientific Publications Office.
Fish Bull. 101:704-711 (2003).
NOTE Powell: Larval abundance, distnbution, and spawning habits of Cynosaon nebulosus
705
— 25°20'
— 25°10'
— 25°00'
24°50'
8roo'
80°45'
80°30'
Figure 1
Location of stations in Florida Bay sampled in 1994-99. See Table 1 for station latitudes and longitudes.
Table 1
Florida Bay sampling stations including zone locations as defined by the South Florida Ecosystem Restoration Prediction and
Modeling Program, Program Management Committee. Stations 1-14 were sampled in 1994-95; stations 5 ,6, 9-13 in 1996; and
stations 6, 15-17 in 1997-99.
Station
Latitude
Longitude
(degrees and minutes)
(degrees and minutes)
Florida Bay zones
24 59.42
80 34.06
Atlantic transition
25 04.42
80 31.24
eastern
25 10.54
80 29.12
eastern
25 009.24
80 37.12
eastern
25 08.30
80 43.19
central
25 04.57
80 46.32
central
25 03.54
80 40.12
central
24 52.46
80 47.31
Atlantic transition
24 55.60
80 55.40
Gulf transition
24 58.48
80 59.48
Gulf transition
25 06.49
8105.16
Gulf transition
25 07.22
80 55.62
Gulf transition
24 59.98
80 55.46
western
24 59.06
80 46.54
central
25 10.80
80 37.80
northern
25 06.00
80 52.50
western
25 07.67
80 57.32
Gulf transition
Location
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Cowpens Cut
Butternut Key
Duck Key
between Eagle Key and Madeira Point
Big Key
Whipray Basin
between Calussa and Russel Keys
between Old Dan and Peterson Key banks
Sprigger Bank
between Oxfoot and Sprigger Banks
Cape Sable
Dave Foy Bank
between Blue and Ninemile Banks
between Rabbit and Gopher Keys
Little Madeira Bay entrance
Palm Key Basin
Bradley Key
706
Fishery Bulletin 101(3)
At stations where replicate tows were taken, densities
were averaged. Ichthyoplankton samples were preserved
in 95% ethanol. Temperature was measured at all stations
with a hand-held thermometer and salinity was measured
with a refractometer. Size at age of larvae was estimated
from the equation L,, standard length = -1.31 + 1.2162
(L,, age in daysl (Powell et al.^).
Because of the high coefficient of variation associ-
ated with ichthyoplankton samples (Cyr et al., 1992), my
sampling design was probably inadequate for multiway
statistical comparisons. Therefore, I used nonparametric
Kruskal-Wallis tests with o = 0.10 (Sokal and Rohlf, 1981)
and relied on patterns and trends to infer differences in
densities of spotted seatrout between stations and time
periods. In the period 1994-95 we tested densities among
nine months and 14 stations to determine trends in spatial
and temporal spawning habits. We also tested differences
between the period 1994-95 and the period 1996 to de-
termine interannual spatial and temporal differences or
similarities. Only those months (April through August)
that were sampled during the two periods were included.
During the period from 1997 through 1999 it was only
appropriate to determine spatial differences because we
sampled irregularly during this period.
A general description of the diverse habitats in relation
to stations in the present study is described by Holmquist
et al. ( 1989, decapod and stomatopod communities ); Thayer
and Chester (1989, fish distribution, seagrass distribution
and abundance, sediment depth, and organic content); Zie-
man et al. ( 1989, macrophyte distribution); and Fourqurean
and Robblee (1999, general description of the Florida Bay
ecosystem).
Results
In 1994-95, salinities were lowest and most variable at sta-
tions 1-5 and 7 in the eastern part of Florida Bay (Figs. 1
and 2 ). Hyperhaline conditions were never observed during
this period. Salinities in 1996, which were recorded monthly
from April through September at stations where recently
hatched spotted seatrout were collected in 1994-95, were
generally euhaline and, as in 1994-95, hyperhaline con-
ditions were never observed. From July 1997 through
November 1999 at four trout monitoring stations, mean
salinities were similar at stations 6, 16, and 17 but were
most variable at station 6. Station 15 had the lowest mean
salinity and the greatest variation. At this station salini-
ties ranged from 10.0 (March 1998) to 33.0 psu (May 1999).
Highest salinities for all four stations were observed in
May 1999. Hyperhaline conditions were observed only at
station 6 in June 1998 and May 1999 and at station 16 in
May 1999 (Fig. 2).
In general, spawning in Florida Bay occurred between
March and October and peaked in June, August, and Sep-
3 Powell, A. B., R. Cheshire, E. H. Laban, J. Colvocnresses, P.
O'Donnell.and M. Davidian. In review. Growth, mortality and
hatchdate distributions of larval and juvenile spotted seatrout,
Cynoscion nehulosus, in Florida Bay, 26 p.
15
10
40 -1
35
S 30
Q.
$• 25
c
« on -
15
10
~i — I — I — I — I — I — I — I — I — I — I — I — r
1 2 3 4 5 6 7 8 9 10 11 12 13 14
B
i ■ i i « i
~i I I I r I I
5 6 9 10 11 12 13
40
35
30 -
25 -
20
15
10
\ I
1 1 1 1
6 15 16 17
Station
Figure 2
Mean salinities SD for (A) 14 stations sampled
from September 1994 through August 1995,
(B) 7 stations sampled monthly from April
through September 1996, and (C) 4 stations
sampled from July 1997 through November
1999. For months that samples were taken,
refer to Tables 2 and 3.
tember ( Table 2 ). Densities of spotted seatrout were signifi-
cantly different among months in 1994-95 (P<0.01) and
1996 (P=0.01). Spotted seatrout larvae were absent during
December, February, and April in 1994—96 (Table 2), and
in November 1997-99 (Table 3). Most spawning, based on
larval collections, occurred between 26° and 34°C (Fig. 3).
The coldest temperature at which larval spotted seatrout
were collected was 20°C in March 1998 at station 6. Spotted
seatrout larvae were collected mainly at salinities between
25 and 40 psu (Fig. 3), although larvae were collected in
salinities as low as 12 psu at station 15.
NOTE Powell: Larval abundance, distribution, and spawning habits of Cynoscion nebulosus 707
Spatially, there were significant differences in densities
of spotted seatrout among 14 stations from 1994 through
1995 (P=0.01), and densities were highly variable (Table
2). In addition, a considerable number of zero catches oc-
curred at stations where spotted seatrout were collected
at least one time. This high variability indicated that the
sampling design was inadequate to properly evaluate the
spatial and temporal abundance of spotted seatrout larvae.
However, some patterns could be discerned. Generally, lar-
val spotted seatrout were absent or rarely collected in the
eastern (stations 2, 3, and 4), Atlantic transition (stations
1 and 14), northern transition (station 15) zones and in a
portion of the central zone (station 15) (Tables 2 and 3).
They were consistently collected at station 6 in the central
zone (Tables 2 and 3).
There were no significant differences (P=0.14) in densi-
ties of spotted seatrout among stations in 1996; stations
where trout occurred on more than one occasion in 1994-95
(Table 2). As in 1994-95, high mean densities of spotted
seatrout occurred at station 6 (Table 2).
Significant differences in spotted seatrout densities
at certain stations were observed between the periods
1994—95 and 1996. Differences were observed at stations
9 (P=0.02), station 10 (P=0. 02), and station 13 (P=0.02). At
these three stations in 1996 spotted seatrout were collected
only in one month (August). They were never collected at
station 5 in 1996 (Table 2).
Size of larvae was indicative of spawning locations.
Recently hatched spotted seatrout larvae ( 1.0-1.9 mm no-
tochord length; s5 d old) were collected mainly in central
(stations 5 and 6), Gulf transition (stations 9, 10, 11, 12, 17)
and western (stations 13 and 16) zones. Recently hatched
larvae were rare at station 4 (eastern zone), station 14
(central zone) and station 15 (northern transition zone).
They were absent at station 8 (Atlantic transition zone)
(Figs. 4 and 5).
Discussion
Evidence from previous studies (Powell et al., 1989; Ruth-
erford et al., 1989a) and the present study establishes the
spatial and temporal spawning habits of spotted seatrout
in Florida Bay. Length-frequency distributions of spotted
seatrout larvae collected in 1984-85 (Powell et al., 1989)
and data from the present study indicate that spotted seat-
rout spawn mainly in the Gulf transition, central, and west-
em zones of Florida Bay and that there is limited spawning
in the northern transition and eastern zones (Figs. 4 and
5). Spawning also occurs in the far northeastern portion of
Florida Bay in Little Blackwater and Blackwater Sounds
(Rutherford et al., 1989a). However, there is no evidence for
spawning in the Atlantic transition zone. The distribution
of planktonic larvae is not necessarily a good indicator of
postsettlement habitat requirements because abiotic fac-
tors related to transport could influence their distribu-
tion. However, in Florida Bay larvae are not distributed
homogeneously throughout the bay, and mudbanks impede
circulation (Fourqurean and Robblee, 1999). The adults are
generally nonmigratory and inhabit shallow seagrass-rich
environments (Chester and Thayer, 1990). Hence, the dis-
cu
u
Q
■
S
o
-rr
CO
CO
,-H
rA
CD
a
►J
S
M
"
00
to
•H
Oi
lO
o
cq
^~*
M
o
t^
M
o
o
O
CD
o
CO
00
X, -6
M
Z
Z
in
Z
a, OJ
< a
q
-a 6
Tf
c cs
+1
a m
02 C
CO
00
_Q CO
C to
11
d
+1
cfl
C/2
o
CO
o
o
o
o
o
CO
CO
CD
S
Z
Z
Z
o
a a.
■ -, CO
T3
2; -^
bo
o
M
M
CO
CO
m
CO
CO
CO
CO
c^
t g
3
<
•z
Z
z
Z
Z
(M
2 6
CM
. 0)
CO
V,
O
1-t
i>
o
in
°°
-t ^
^^
o
o
Tf
o
r-i
CO
o
,_;
cc
o
^
m aJ
3
,-*
r-»
,-»
ro -a
1-5
i-4 3
a^
•S|
in
CD
Oi
"*
2 c
05
1— 1
c
CO
lO
o
iri
CO
CO
CTJ
o
in
o
■H
o
S^
3
,— 1
00
eg
■-5
Cfi CO
^5
q
,-H
+1
Oi
,-H
00
o
00
ca cd
>>
a
o
o
o
o
rH
CX)
§
o
o
Tf
o
e4
o
O
o
O
o
C3
g "
Z
o "
-*
CM
^ci
OJ
CO
~£ -■"
I— 1
+1
^ S
Tf
135
o
t~:
o
CO
00
"^
o.
o
CO
Tf
o
d
,-i
CO
o
CD
Tf
,-4
S-2
2 a
?— 1
?— 1
M
Ol
,-H
a
|i
o
'-5
C
CO
rt
OJ
^ -S
-S
■^
in
CO
00
Ol
o
O)
CO
■*
§
Q CO
M
,-»
708
Fishery Bulletin 101(3)
Total fish larvae Spotted seatrout
10000
10000
1000
100
10
f 1
'•••*•• 1000
••• • : ' '•••.' <•: . 10
• • ••
1
o 18 20 22 24 26 28 30 32 34 36 18 20 22 24 26 28 30 32 34 36
^ Temperature ("C)
E
ID
C
J 10000
10000
S 1000
100
10
1
• :»«. ' •* 1000
• •••• ?ri^
1
• • : .V'
•••1 fcV
1
) 10 15 20 25 30 35 40 45 10 15 20 25 30 35 40 45
Salinity (psu)
Figure 3
Density of total fish larvae and spotted seatrout versus temperature and salinity.
"Total fish larvae" indicates the complete range of temperature and salinity where
all collections were made and is provided for visual comparison with spotted seatrout
collections.
Table 3
Densities (numbers/100 m-') of spotted seatrout collected at monitoring stations with a
mesh.
bow-mounted push net with 0.333-mm
.Station
1997
1998
1999
Jul Sep
Mar May
Jun
Jul Sep
May
Jul Nov
Mean ±SD
6
15
16
17
Mean ±SD
0 6.2
3.2 20.1
7.1 1.1
6.6 1.0
4.2 ±3.3 7.1 ±9.0
1.2 40.2
0 0
1,0 1.2
30.8 0
8.2+15.0 10.3 ±10.4
57.0
0
75.6
0
33.1 ±39.0
16.0 93.1
0 0
0 5.2
0 0
4.0 ±8.0 24.6 ±45.7
56.8
0
0
0
14.2 ±28.4
1.1 0
0 0
0 0
6.9 0
2.0 ±3.3 0
30.2 ±33.2
2.6+6.7
9.1 ±23.5
5.0 ±10.1
tribution of larvae presented in the present study is most
likely a good indication of adult spawning areas.
As indicated by the larval collections in this study,
spotted seatrout have a protracted spawning period from
March through October, which is similar to that observed
in Tampa Bay, Florida ( McMichael and Peters, 1989). To the
contrary, Stewart (1961) reported that spotted seatrout in
Florida Bay spawn throughout the year, and Rutherford
et al. (1989a) indicated that some spawning occurred as
early as February and continued into December. Powell et
al.,'* studying hatchdate distributions of juveniles, reported
peak spawning in early May, late June, and late August.
Seagrass meadows appear to be critical habitats for ju-
venile spotted seatrout (Chester and Thayer, 1990;Tolanet
al., 1997; Rooker et al.. 1998; Thayer et al., 19991. Rooker
et al. (1998) reported that juvenile spotted seatrout in a
Texas estuary prefer H. wrightii over T. testudinum. In
another Texas estuary, Tolan et al. (1997) reported that
NOTE Powell: Laa/al abundance, distribution, and spawning habits of Cynoscion nebulosus
709
juvenile spotted seatrout prefer H. wrightii over Syringo-
dium fiUforme. In Florida Bay juveniles are collected at
highest densities in western Florida Bay basins near the
Gulf of Mexico in habitats with deeper and more organic
sediments and with greater density and biomass of S. fili-
forme. In areas where spotted seatrout juveniles are rare
or absent, which generally reflects the distribution of their
larvae, organic matter and sediment depth are minimal,
water depth is generally deeper, and seagrass standing
crop, short shoot densities, and diversity are lower (Thayer
and Chester, 1989; Chester and Thayer 1990),
Spotted seatrout larvae were collected consistently at
relatively high densities in Whipray Basin (station 6;
central zone) and length-frequency distributions indicate
spawning most probably occurs in this area. The major-
ity of larvae collected in this area were 2.0-2.4 mm SL (5
to 6 d old), and it is possible that larvae could have been
transported into this area from western Florida Bay. Nev-
ertheless, Whipray Basin is a nursery area for juvenile
spotted seatrout (Florida Department of Environmental
Protection^. However, Whipray Basin has a relatively
sparse standing crop of the seagrass Thalassia testudinum
(12 g/m'-) compared to Palm Key (station 16; western
zone) which has a higher standing crop of T. testudinum
(28 g/m-) and Halodule wrightii (14 g/m^), and has been
demonstrated to be an important nursery area for spotted
seatrout juveniles (Florida Department of Environmental
Protection').
Spotted seatrout eggs have been collected in other
waters ranging from 15 to 50 psu (Holt and Holt, 2002).
Presumably, larval spotted seatrout eggs sink to the bot-
tom at salinities <15 psu and are not viable (Alshuth and
Gilmore^). Therefore, it is surprising that recently hatched
(<5-d old) larval spotted seatrout were collected, although
infrequently, at Little Madeira Bay (station 15; northern
transition zone) and only at 12 psu. Whether these low sa-
linities, which occurred in July and September 1997, were
a result of drastic changes in salinities caused by weather
events that occurred after hatching is unknown. Still, it is
highly unlikely that a significant number of viable eggs can
be produced at those low salinities.
The qualitative description of spotted seatrout spawn-
ing habits provides necessary baseline data in relation
to restoration activities, specifically freshwater inflow.
Restoration activities could have both a negative and posi-
tive effect because salinity can have significant effects on
spotted seatrout reproduction and early life history stage
processes ( Holt and Holt, 2002 ). For example, low salinities
(as discussed above) could be detrimental to egg viability;
whereas, alleviating hypersaline areas could expand the
spawning area, particularly in the central zone where hy-
persalinity conditions are persistent (Orlando et al., 1997;
Thayer et al., 1999). At high and low salinities, growth and
development rates of larval spotted seatrout have been
reported to be reduced because these processes are con-
strained by undeveloped osmoregulatory functions (Holt
and Banks, 1989). On the other hand, there is evidence that
spotted seatrout populations have adapted to reproduce in
extreme-salinity environments where spawning salinities
influence egg buoyancy and the salinity tolerance of early
Sta 4
^^
"1 — r-
* Oi <• C7» * O) O
— — c^ f>4 r^ t^ ^
I I I I I I ^
o tn o ■• o> ♦ a> o
— — c«j ri -n ^ V
I ) I I I I M
O lA O )0 o «o
— •-^ c« * rj m »o «
I I I I I I ^
O U^ O U^ o in
— — IM (N lO •^
Sla 1 1
10 ■
0 -
-n^
* Ot •♦ o> * Ol o
— ^ rx M K> to *
t I I I I I /<
O T ai -^ cjj at
1— T— CNCNCM T— 1— CNJtNtN
oinoin'^ oinoin'^
T-^c>jcs( T-^CNrsi
Size class (mm)
Figure 5
Length-frequency distributions of spotted seatrout from four stations where a bow-
mounted push net was used to sample fish from July 1997 through November 1999.
See Table 1 for station locations.
larval stages. This adaptation would allow spotted seatrout
to spawn over a wide range of salinities.
Future monitoring of spotted seatrout larval abundances
to evaluate restoration activities would probably require
numerous samples per station because of the high degree
of variability as shown by the present study. Therefore, it
would seem prudent to continue monitoring spatial spawn-
ing habits from larval collections, but to develop a juvenile
abundance index to monitor the success of restoration in
the Everglades, as well.
Acknowledgments
Sincere appreciation is extended to Al Crosby who was
field party chief and processed numerous samples. I am
also indebted to other Beaufort Laboratory staff members,
notably Robin Cheshire, Peter Crumley, Mike Greene,
Donald Hoss, Michael Johnson, and Michael Lacroix for
their able assistance in the field; Robin Cheshire, Curtis
Lewis, and Harvey Walsh for graphics; James Waters for
computer programming; and Dean Ahrenholz, Patti Mar-
raro Joe Smith and an anonymous reviewer for valuable
comments in their review of the manuscript. The staff of
the Polish Sorting and Identification Center processed
many of the ichthyoplankton samples. This study was sup-
ported through joint funding from the National Oceanic
and Atmospheric Administration Coastal Ocean Program
and National Marine Fisheries Service base funds to the
Beaufort Laboratory.
Literature cited
Chester, A. J., and G. W. Thayer.
1990. Distribution of spotted seatrout {Cynoscion nebu-
losus) and gray snapper {Lutjanus griseus) juveniles in
seagrass habitats of western Florida Bay. Bull. Mar. Sci.
46:345-357.
Cyr, H., J. A. Downing, S. Lalonde, S. B. Baines, and M. L. Price.
1992. Sampling lai-\'al fish: choice of sample number and
size. Trans. Am. Fish. Soc. 121:356-368.
Fourqurean, J. W., and M. B. Robblee.
1999. Florida Bay: a history of recent ecological changes.
Estuaries 22:345-357.
Hettler, W. F., Jr.
1989. Food habits of juveniles of spotted seatrout and
gray snapper in western Florida Bay. Bull. Mar. Sci. 44:
152-165.
NOTE Powell: Lan/al abundance, distribution, and spawning habits of Cynoscion nebulosus
711
Hettler, W. R, and A. J. Chester.
1990. Temporal distribution of ichthyoplankton near Beau-
fort Inlet, North Carolina. Mar. Eeol. Prog. Sen 68:157-
168.
Holmquist, J. G., G. V. N. Powell, and S. M. Sogard.
1989. Decapod and stomatopod communities of seagrass-
covered mud banks in Florida Bay: inter-and intra-
bank heterogeneity with special reference to isolated
subenvironments. Bull. Mar. Sci. 44:251-262.
Holt, G. J., and M. A. Banks.
1989. Salinity tolerance and development of osmoregulation
in larval sciaenids. In The early life history offish (J. H. S.
Blaxter, J. C. Gamble, and H. von Westernhagen, eds.); the
third ICES symposium, Bergen, Norway 35 October 1988,
p. 4-89. Rapp. P-V. Reun. 191.
Holt, G. J., and S. A. Holt.
2002. Effects of variable salinity on reproduction and early
life stages of spotted seatrout. It} Biology of the spotted
seatrout (S. Bortone, ed.), p. 135-145. CRC Press. Wash-
ington, DC.
McMiehael. R. H., Jr., and K. M. Peters.
1989, Early life history of spotted seatrout, Cynoscion nebu-
losus (Pisces: Sciaenidae), in Tampa Bay, Florida. Estuar-
ies 12:98-110.
Orlando, S. P, Jr,, M. B. Robblee, and C, J. Klein.
1997. Salinity characteristics of Florida Bay: a review of
the archived data set (1955-95), 89 p. Office of Ocean
Resources Conservation and Assessments, National Oceanic
and Atmospheric Administration, Silver Spring, Maryland,
Powell, A. B., D. E. Hoss, W. F. Hettler, D. S. Peters, and
S. Wagner.
1989. Abundance and distribution of ichthyoplankton in
Florida Bay and adjacent waters. Bull. Mar. Sci. 44:35-
48.
Rooker, J. R., S. A. Holt, M. A. Soto, and G. J. Holt.
1998. Postsettlement patterns of habitat use by sciaenid
fishes in subtropical seagrass meadows. Estuaries 21:
318-327.
Rutherford, E. S., E. S. Thue, and D. G. Buker.
1982. Population characteristics, food habits and spawning
activity of spotted seatrout, Cynoscion nebulosus, in Ever-
glades National Park. Rep. T-668, 48 p. South Florida
Research Center, U. S. National Park Service, Homestead,
Florida.
Rutherford, E. S., T. W Schmidt, and J. T. Tilmant.
1989a. Early life history of spotted seatrout (Cynoscion
nebulosus) and gray snapper iLutJanus griseus) in Florida
Bay, Everglades National Park, Florida. Bull. Mar. Sci.
44:49-64.
Rutherford, E. S., J. T. Tilmant, E. B. Thue, and T. W. Schmidt.
1989b. Fishery harvest and population dynamics of spotted
seatrout, Cynoscion nebulosus, in Florida Bay and adjacent
waters. Bull. Mar. Sci. 44:108-125.
Sokal, R. R., and F J. Rohlf
1981. Biometry, 2nd ed., 859 p. W. H. Freeman and Co.,
San Francisco, CA.
Stewart, K. W.
1961. Contributions to the biology of the spotted seatrout
(Cynoscion nebulosus) in the Everglades National Park, Flor-
ida. M.S. thesis, 103 p. Univ. Miami, Coral Gables, FL.
Tolan, J. M., S. A. Holt, and C. P. Onuf.
1997. Distributoin and community structure of ichthyo-
plankton in Laguna Madre seagrass meadows: potential im-
pact of seagrass species change. Estuaries 20:450^64.
Thayer, G. W., and A. J. Chester.
1989. Distribution and abundance of fishes among basin
and channel habitats in Florida Bay. Bull. Mar. Sci. 44:
200-219.
Thayer, G. W., A. B. Powell, and D. E. Hoss.
1999. Composition of larval, juvenile and small adult fishes
relative to changes in environmental conditions in Florida
Bay Estuaries 22:518-533.
Zieman, J. C, J. W Fourqurean, and R. L. Iverson.
1989. Distribution, abundance and productivity of sea-
grass and macroalgae in Florida Bay. Bull. Mar. Sci. 44:
292-311.
712
Effect of analytical conditions in wavelength
dispersive electron microprobe analysis on the
measurement of strontium-to-calcium (Sr/Ca) ratios
in otoliths of anadromous salmonids
Christian E. Zimmerman
Oregon State University
Department of Fisheries and Wildlife
Corvallis, Oregon 97331
Present address: US. Geological Survey
Alaska Science Center
101 1 E Tudor Road
Anchorage, Alaska 99503
E-mail address: czimmermanfSusgs gov
Roger L. Nielsen
College of Oceanic and Atmospheric Sciences
104 Ocean Administration
Oregon State University
Corvallis, Oregon 97331
The use of strontium-to-calcium (Sr/Ca)
ratios in otoliths is becoming a standard
method to describe life history type and
the chronology of migrations between
freshwater and seawater habitats in
teleosts (e.g. Kalish, 1990; Radtke et al.,
1990; Secor, 1992; Rieman et al., 1994;
Radtke, 1995; Limburg, 1995; Tzeng et
al. 1997; Volk et al., 2000; Zimmerman,
2000; Zimmerman and Reeves, 2000,
2002). This method provides critical
information concerning the relation-
ship and ecology of species exhibiting
phenotypic variation in migratory
behavior (Kalish, 1990; Secor, 1999).
Methods and procedures, however, vary
among laboratories because a standard
method or protocol for measurement of
Sr in otoliths does not exist. In this
note, we examine the variations in
analytical conditions in an effort to
increase precision of Sr/Ca measure-
ments. From these findings we argue
that precision can be maximized with
higher beam current (although there
is specimen damage) than previously
recommended by Gunn et al. ( 1992).
Wavelength dispersive electron mi-
croprobe analysis (WD-EM) has been
used by most researchers, although
other methods such as ()rot()n-induced
x-ray emission (PlXExBabaluk et al..
1997; Markowitz et al, 2000 ) have been
used. WD-EM remains a common and
relatively inexpensive method. The
conceptual approach among research-
ers using WD-EM is similar but the
methodological approach or analytical
(operating) conditions vary. In a com-
parison of laboratories using common
otoliths, Campana et al. (1997) found
among-laboratory variation in mean
Sr concentrations that could not be
described by otolith variability. Al-
though the laboratories were internally
consistent in applying their methods,
comparisons between laboratories dif-
fered. Campana et al. suggested that
the sensitivity of WD-EM to operating
conditions might have led to this varia-
tion between laboratories.
Development of analytical techniques
for measuring Sr/Ca ratios has been re-
viewed to validate techniques in specific
studies (Kalish, 1990; Secor, 1992; Toole
and Nielsen, 1992; Limburg, 1995).
Gunn et al. (1992) analyzed effects of
counting times, beam current, accel-
erating voltage, and beam diameter
on measures of Sr and other elements
and they warned that beam powers
required for WD-EM were sufficient
to cause specimen damage including
pitting and chemical change. As a re-
sult, Gunn et al. (1992) recommended
limiting beam power densities to
< 3pW/pm''. This recommendation has
been followed in most studies using Sr/
Ca ratios to reconstruct the chronology
of migrations between the freshwater
and marine environments (Table 1).
Toole and Nielsen ( 1992), however, con-
cluded that Sr/Ca precision could be in-
creased, with no loss of accuracy, by using
analytic conditions that lead to a beam
power density of just over 15pW/pm-
(5-pm beam diameter; accelerating
voltage=15 loA; beam current=25 kV).
The inherent beam damage was not
critical because of the similar behavior
of Sr and Ca during progressive beam
damage.
In published studies using WD-EM
to measure Sr/Ca ratios in otoliths, the
operating conditions, including beam
power densities, have varied greatly
(Table 1). Establishing a microprobe
protocol for measurement of Sr/Ca
ratios in otoliths involves a balancing
act of counting times, beam current,
and beam diameter. The selection of
optimum conditions is constrained
by financial resources, allocation of
time for use of instruments, and the
required resolution of Sr/Ca ratios
for any specific application. Each re-
searcher must weigh the benefits and
costs to best answer the question at
hand. Generally, these parameters are
manipulated to optimize precision and
accuracy of analyses in relation to vari-
ability within the otolith and implica-
tions of the results.
For Sr/Ca ratios to remain an ac-
cepted and accurate means of describ-
ing migration histories and other life
history events, continued analytic and
technical refinement and validation
are required. We examined the effects
of crystal choice, beam diameter, beam
current, and beam power densities on
Sr/Ca measurements (expressed as
atomic ratios) in salmonid otoliths: 1)
we measured Sr using both the TAP
and PET crystals in regions with high
Sr/Ca 00.003) and low Sr/Ca XO.OOl)
Manuscript approved for publication
12 February 2003 by Scientific Editor
Manu.script received 4 April 2003 at NMFS
Scientific Publications Office.
Fish Bull. 101:712-718 (2003).
NOTE Zimmerman and Nielsen: Measurement of strontium-to-calcium ratios in otoliths of anadromous salmonids
713
Table 1
Analytic conditions reported by researchers using
wavelength-dispersive (WD) electron spectroscopy to measure Sr/Ca ratios in
otoliths. Beam power density was
calculated for this study and minimum
limit of detection
is either directly
reported from the work
cited or from personal communications.
Beam power
Minimum limit
Beam
Accelerating
Beam current
density
of detection
Source
diameter Ip
m) voltage (kVl
(nA)
(pW/pm2)
(Sr ppm)
Brown and Severin (1999)
6
15
20
10.61
—
Campana et al. (1997) WD-1
10
25
5
1.59
175
Campana et al. (1997) WD-2
9
15
4
4.25
480
Kafemannetal. (2000)
5x8
15
10
3.75
490
Kahsh(1990)
10x10
15
10
1.5
—
Kawakami et al(1998)
1
15
50
954
—
Limburg(1995)
20
20
25
1.59
290
Radtke(1995)
5
15
10
7,63
—
Rieman et al. (1994)
5
15
50
38.14
—
Secor(1992)
5x5
25
20
20.00
580
Thresher etal. (1994)
14
15
25
2.44
311
Toole etal. (1993)
5
15
20
15.27
—
Volk etal. (2000)
10
15
15
2.86
237
Zimmerman and Reeves (2000)
7
15
50
19.50
43
levels; 2) we then compared the results of repeated Sr/Ca
measurements collected at the same spots using various
beam diameters, while holding accelerating voltage and
beam current constant to determine the effect of beam
damage on Sr/Ca measurements; and 3) we compared the
results of repeated Sr/Ca measurements collected at the
same spots using various beam currents, while holding ac-
celerating voltage and beam diameter constant. We argue
that increased precision of Sr measurements afforded by
higher beam current (and hence, higher beam power densi-
ties) is preferable for studies where only measurements of
Sr/Ca ratios are required.
Materials and methods
Otolith preparation
Sagittal otoliths from an adult sockeye salmon (Oncorhyn-
chus nerka) collected in the Deschutes River, Oregon, and
a juvenile chinook salmon (O. tshawytscha) collected in
the Umatilla River, Oregon, were used to represent high
(>0.003) Sr/Ca and low (<0.001) Sr/Ca ratios, respectively
(Zimmerman, unpubl. data). High Sr/Ca ratios character-
ized the saltwater growth region in the sockeye salmon
otolith and low Sr/Ca ratios characterized the freshwater
growth region of the chinook salmon otolith. Each oto-
hth was mounted sulcus side down with thermo-setting
plastic resin on a microscope cover slip attached at one
end with super-glue to a standard microscope slide. The
otolith was then ground with 1200-grit sandpaper in the
sagittal plane to the level of the nucleus. The mounting
medium was heated and the otolith turned sulcus side-up.
The otolith was then ground with 1200-grit and 2000-grit
sandpaper in the sagittal plane to the level of the primor-
dia and polished with 0.05-pm alumina paste. The cover
slip was then cut with a scribe and mounted with other
prepared otoliths (those used in other studies) on a petro-
graphic slide for microprobe analysis. The slide contain-
ing several otoliths was rinsed with deionized water, air
dried, and carbon coated (400 A). Elemental analysis was
conducted with a Cameca SX-50 wavelength dispersive
microprobe. Strontiantite (SrC03, USNM R10065) and
calcite (CaCOs, USNM 136321) were used as standards
for Sr and Ca, respectively. Standards were calibrated
with a 30-pm-diameter beam and 10-s counts resulting in
minimal effects of beam damage.
Effect of spectrometer (crystal) choice
To evaluate differences in diffracting crystals, we con-
ducted a series of tests where Sr was measured by using
both the PET and TAP crystals. A 15 kV, 50 nA beam was
used for these comparisons. With a 7-pm-diameter beam,
Sr was measured by using the TAP crystal (Sr La) and
Ca was measured by using the PET crystal (Ca Ka). Two
transects of 10 points each were sampled so that the points
on adjacent transects covered the same temporal location
on the otolith. Sr and Ca were analyzed simultaneously;
counting times for the Sr and Ca peaks were 40 s, and back-
ground counts were 40 s. A second set of transects covering
the same temporal locations in the otolith was sampled, but
Sr was measured with the PET crystal (Sr La). Because
our microprobe has only one PET crystal, simultaneous
measurement of elements was not possible. Transects were
conducted on both high and low Sr/Ca regions. Sr/Ca ratios
714
Fishery Bulletin 101 (3)
were calculated from normalized mole fractions of
Sr and Ca. Limit of detection (3a: Potts, 1987) was
calculated for all points in both high and low Sr/Ca
regions.
Effect of beam diameter and beam current
We conducted five repeat measurements at each of
five points within the high and low Sr/Ca regions
using a 1-. 7-, 15-, 20-. and 25-pm-diameter beam at
15 kV and 50 nA. This resulted in beam power densi-
ties of 961, 19.5, 4.2, 2.39, and 1.52pW/pm^ respec-
tively. We conducted five repeated measurements at
each of four locations within the high and low Sr/Ca
regions using beam currents of 5, 10, 20, and 30 nA
with a 10-m-diameter beam and accelerating voltage
of 15 kV, with resulting beam power densities of 1.0,
1.9, 3.8, and 5.7pW/pm-, respectively. The coefficient
of variation (CV) of Sr/Ca ratios was calculated as the
^^Sr/Ca ^ Mean Sr/Ca"' for each beam power density.
If beam damage affects precision and accuracy of
Sr/Ca ratios, subsequent measurements at the same
spot should be increasingly divergent from the first
and such divergence should be evident in high coeffi-
cients of variation. The limit of detection was used as
a measure of precision. Limit of detection (3a; Potts,
1987 ) for Sr was calculated for the first measurement
taken at each beam power density in each region.
Results
Effect of spectrometer (crystal) choice
Spectrometer (crystal) choice for the measurement of Sr
had an apparent systematic effect on Sr/Ca ratios at high
Sr/Ca levels but no effect at low Sr/Ca levels (Fig. lA). The
mean Sr/Ca level in the high Sr/Ca region was significantly
lower (-15%) when Sr was measured on the PET crystal
«=7.189; P<0.001; df=38). Crystal choice had an effect on
measurement of both Ca and Sr in the high Sr/Ca region.
Ca did not differ significantly between the high and low Sr/
Ca regions (P>0.05) but was approximately 2% lower when
Sr was measured with the PET crystal (Fig. IB). This dif-
ference was attributable to beam damage, which occurred
as Sr was measured. The mean Ca was 197.400 ppm when
Sr was measured on the PET crystal and 202,100 ppm
when Sr was measured on the TAP crystal.
In the high Sr/Ca region, mean Sr was significantly
lower when Sr was measured on the PET crystal (^=11. 58;
P<0.001;df=38)(Fig. lBi.Themean(±SD)was7270±4ppm
when Sr was measured with the PET crystal and 8870
±3 ppm when Sr was measured with the TAP crystal. This
difference is also reflected in the higher minimum limit of
determination for Sr for PET in both the high Sr/Ca (695
ppm) and low Sr/Ca regions (126 ppm). Using the TAP
crystal to measure Sr, we found that the minimum limit of
determination for Sr was 103 ppm and 65 ppm in the high
Sr/Ca and low Sr/Ca regions, respectively. To achieve simi-
lar counting statistics for Sr with the PET crystal, count
0 005 -
t-^-^-^^-^i^^^
0.0O4 ■
r T V
O
0 002 ■
^"'^^^^^
0 001 ■
0 000 ■
22 ■
20 -
18 -
^^-^^-^^^^-^!^
^=0=^=^
|- 0 20
■ 0.15
B
^^=i^fe^#-4
t=:n9r==m — E
o
1 "
CO i; -
O
10 -
es=crptTS
Sr atomic %
o O
8 ■
6 -
^—^ f=.=^=A^-A.-^
^ vv ^-W\
- 0 00
)23456?89 10
Transect pent
Figure 1
Transects of (A) Sr/Ca atomic ratio and (B) Ca (circles) and Sr
(triangles) as atomic percentage when Sr is measured on the PET
crystal (solid symbols) and on the TAP crystal (open symbols).
Error bars represent 95% confidence intervals.
times would need to be increased to 200 seconds on both
the peak and background.
Effect of beam diameter and beam current
In repeated measurements with different beam diameters
(same beam current) at the same locations, Sr/Ca ratios
did not vary greatly with the exception of measurements
made with the 1-pm beam (Fig. 2A). The CV of the Sr/Ca
ratios for the 1-m beam was high and led to significant
variation of Sr/Ca ratios in subsequent measurements at
the same point (Table 2). The Sr/Ca ratio was least vari-
able for the 7-pm beam in the high Sr/Ca region (Table 2).
Limit of detection (3a) for Sr ranged from 80 ppm to 172
ppm in the low Sr/Ca region and from 299 ppm to 315 ppm
in the high Sr/Ca region under the various beam diameters
(Fig. 3A).
Beam current had a significant effect on variation of
Sr/Ca and limit of detection for Sr. The greatest variation
in Sr/Ca ratios was observed at beam currents 5nA and 10
nA (Fig. 2B). The CV of Sr/Ca ratios was negatively related
to beam current (r=-0.93; P>0.05) (Table 3). The CV of Sr
was high in all treatments, ranging from 0.23 to 1.17 in the
high Sr/Ca region and from 0.11 to 0.46 in the low Sr/Ca
region. The CV of Ca was 0.04 for all beam configurations.
The limit of detection of Sr as measured at the first sample
NOTE Zimmerman and Nielsen: Measurement of strontium-to-calcium ratios in otoliths of anadromous salmonids
715
o
5
4
3
2
1
0
5 -
4
3 -
2
1
0
5
4 -
3
2
1
0
5 -
4
3
2
1
0 -
5
4
3
2
1
0
7 Lim
•-
20 Mm
B
Sample sequence at same location
Figure 2
Repeat measures of atomic Sr/Ca with (A) varying beam diameters, common beam
current (50 nA), and (B) varying beam currents, common beam diameter (10 jim).
Open circles represent low Sr/Ca regions and solid circles represent high Sr/Ca
regions.
for each beam-current configuration ranged from 99 ppm to
290 ppm in the low Sr/Ca region and from 312 ppm to 844
ppm in tiie high Sr/Ca region (Fig. 3B).
Discussion
In our experiment, for measuring Sr/Ca ratios in otoUths
with WD-EM analysis, the TAP crystal was the best choice
for the measurement of Sr because it provided the advan-
tage of higher count rates and higher resolution of Sn Use of
the PET crystal to measure Sr has not been reported in the
literature. In fact, crystal choice is frequently not reported,
making it difficult to know whether TAP or PET crystals
have been used to measure Sr. Personal communication
with several researchers confirmed that the TAP crystal
is commonly used to measure Sr and has been cited as the
crystal used (Kalish, 1990; Thresher et al., 1994). Given the
different results possible with the use of different crystals to
measure Sr, reporting the crystals used should be included
in papers reporting Sr/Ca ratios measured on WD-EMs.
Note that even though measurements of Sr with the TAP
crystal appeared to be the best method for otoliths, the TAP
crystal should not be used to measure Sr in materials con-
taining Si because of analytical interference from Si.
Variation of beam diameter has very little effect on
the limit of detection of Sr, even at extremely high beam
current densities. Rather, variation in beam current had
716
Fishery Bulletin 101(3)
1000
800
\
1. 600 -
a.
\
W
\.,^
400
^^--— ____^
200
v^_____^
o o
1 7 15 20 25 0 10 20 30 40 50
Beam diameter (nm) Beam current (nA)
Figure 3
Detection limit of Sr with (A) varying beam diameter and (B) varying beam current. Open
circles represent low Sr/Ca regions and solid circles represent high Sr/Ca regions.
Table 2
Coefficient of variation of Sr/Ca ratios and Sr (in parenthe-
ses) in high and low Sr/Ca regions of otoliths.
Beam diameter
(|im)
CV Sr/Ca and Sr
in low region
CV Sr/Ca and Sr
in high region
1
0.076
(0.02)
0.416
(0.41)
7
0.082
(0.07)
0.009
(0.04)
15
0.218
(0.23)
0.041
(0.05)
20
0.107
(0.03)
0.028
(0.02)
25
0.133
(0.13)
0.030
(0.04)
Table 3
Coefficient of variation of Sr/Ca ratios and Sr (in paren-
theses) in high and low Sr/Ca regions of otoliths under
varying beam current.
Beam current
(nA)
CV Sr/Ca and Sr
in low region
CV Sr/Ca and Sr
in high region
5
0.354
(0.94)
0.099
(0.11)
10
0.429
(1.11)
0.098
(0.41)
20
0.239
(1.12)
0.039
(0.45)
30
0.248
(0.23)
0.039
(0.46)
50
0.082
(1.17)
0.009
(0.11)
significant efTects on the limit of detection of Sr. As a re-
sult, higher beam currents (>20 nA) were appropriate for
measuring Sr/Ca ratios in spite of beam damage observed
at higher beam power densities. Beam diameters between
7 and 10m provide the best temporal resolution (i.e. cover-
ing fewer daily increments). The lower CV of Sr/Ca ratios
observed with the 7-m beam diameter was likely due to the
lower temporal variation afforded to smaller beam diam-
eters (compared to larger beam (liamctcrs) and lower error
related to specimen damage (com[)ared to the 1-m beam).
The lower CV of Sr/Ca ratios at the 7-pm beam diameter
suggested that in spite of beam damage, the Sr/Ca ratio
was not dramatically affected by beam damage. However,
the increase in CV for the smaller diameters suggested that
there are limits to usable beam densities.
Greater precision of Sr/Ca measurements is critical to
understanding life history of some species (Markowitz et al.,
2000) or in situations where differences in environmental
Sr/('a raios are less than those observed between ocean wa-
ter and freshwatcrs (Rieman et al., 1994; Volk et al., 2000).
Volk et al. (2000) found that timing of freshwater entry
and length of freshwater residence by summer steelhead
(O. mykiss) and spring chinook salmon had efTects on oto-
lith core or primordia Sr/Ca levels. Summer steelhead and
spring chinook enter freshwater and stay for up to several
months before spawning. Volk et al. (2000) suggested that
NOTE Zimmerman and Nielsen: Measurement of strontium-to-calcium ratios in otoliths of anadromous salmonids
717
significant egg development during this extended prespawn-
ing freshwater residence led to a dilution of the Sr signature
in these anadromous fish. Zimmerman and Reeves (2000,
2002) were able to distinguish between resident rainbow
trout and summer steelhead in the Deschutes River, Or-
egon, by comparing the Sr/Ca ratios in primordia and the
first summer of juvenile growth (freshwater growth region).
In essence the freshwater gi-owth region acts as a proxy for
the freshwater environment and significantly higher Sr/Ca
ratios in the primordia suggest an anadromous maternal
origin. The greater precision of Sr measures afforded by
higher beam currents may be important in distinguishing
differences in seasonal ecotypes, such as summer steelhead
and spring chinook salmon, or in distinuishing estuary
habitats from freshwater and ocean environments.
These results are applicable only to otolith calcium car-
bonate in the mineral form of aragonite. Like Brown and
Severin (1999), we have found that crystalline structure
affects the distribution of Sr. Vateritic regions should be
avoided when measuring Sr/Ca ratios in otoliths. In vater-
itic portions of otoliths from chinook salmon and steelhead,
Sr is often below our minimum detection limit of 43 ppm,
yet the concentration of Ca does not differ from that found
in aragonitic otolith regions (Zimmerman, unpubl. data).
Studies offish migration between marine and freshwater
environments are based on the general difference between
Sr in marine and freshwater environments. Sr concentra-
tions in seawater are generally an order of magnitude
greater than in freshwaters (Bagenal et al., 1973; Kalish,
19901. Sr is substituted for Ca in the calcium carbonate
matrix of the otolith at levels that correspond to those
in the environment (Kalish, 1989; Farrell and Campana,
1996). Given this relationship, it has become a convention
to report Sr as a fraction of Ca (Secor and Rooker, 2000).
However, Secor and Rooker (2000) pointed out that Ca is
relatively invariant in aragonitic otoliths and rarely var-
ies more than 5% within an individual fish. At 8074 points
sampled in the primordia, freshwater growth regions, and
saltwater gi-owth regions of several species of salmonids
the Sr/Ca ratio was entirely driven by differences in Sr
(Zimmerman, unpubl. data). At these 8074 points, Sr was
highly correlated with the Sr/Ca ratio {r^=99A5%) and Ca
was not correlated with the Sr/Ca ratio (r-<0.01%). Given
this relationship, increasing precision of Sr is desirable to
increase precision of the Sr/Ca ratio.
Our results suggest that tests of hypotheses related to
Sr/Ca ratios can be conducted at higher beam power densi-
ties than suggested by Gunn et al. ( 1992 ). High beam power
densities resulting from higher beam current and beam
diameter of 7 to 10-pm provide greater precision (spatial on
the otolith and temporal in the life of the fish) of Sr. This is
not true for studies of stock discrimination, such as those
described by Thresher ( 1999), that rely on absolute values
of multiple elements, including Sr.
Acknowledgments
Several people provided unpublished information concern-
ing analjftic conditions and detection limits. Gordon Reeves,
of the U.S. Forest Service Pacific Northwest Research Sta-
tion, kindly provided office and laboratory space to CEZ. We
thank Eric Volk, Ken Severin, and two anonymous review-
ers for comments that improved this manuscript.
Literature cited
Babaluk, J. A., N. M. Halden, J. D. Reist, A. H. Kristofferson,
J. L. Campbell, and W. J. Teesdale.
1997. Evidence for non-anadromous behaviour of Arctic
charr iSalvelinus alpinus) from Lake Hazen, EUesmere
Island, Northwest Territories, Canada, based on scan-
ning proton microprobe analysis of otolith strontium dis-
tribution. Arctic 50:224-233.
Bagenal, T. B., F. J. H. MacKereth, and J. Heron.
1973. The distinction between brown trout and sea-trout
by the strontium content of their scales. J. Fish Biol. 5:
555-557.
Brown, R., and K. P. Severin.
1999. Elemental distribution within polymorphic inconnu
iStenodus leucichthys) otoliths is affected by crystal
structure. Can. J. Fish. Aquat. Sci. 56:1898-1903.
Campana, S. E., S. R. Thorrold, C. M. Jones, D. Gunther,
M. Tubrett., H Longerich, S. Jackson, N. M. Halden, J. M.
Kalish, P. Piccoli, H. de Pontual, H. Troadec, J. Panfili,
D. H. Secor, K. P Severin, S. H. Sie, R. Thresher, W. J. Teesdale,
and J. L. Campbell.
1997. Comparison of accuracy, precision, and sensitivity in
elemental assays of fish otoliths using the electron micro-
probe, proton-induced x-ray emission, and laser ablation
inductively coupled plasma mass spectrometry. Can. J.
Fish. Aquat. Sci. 54:2068-2079.
Farrell, J., and S. E. Campana.
1996. Regulation of calcium and strontium deposition on the
otoliths of juvenile tilapia, Oreochromis niloticus. Comp.
Biochem. Physiol. 115A:103-109.
Gunn, J. S., I. R. Harrowfield, C. H. Proctor, and R. E. Thresher
1992. Electron probe microanalysis offish otoliths — evalua-
tion of techniques for studying age and stock discrimination.
J. Exp. Mar Biol. Ecol. 158:1-36.
Kafemann, R., S. Alderstein, and R. Neukamm.
2000. Variation in otolith strontium and calcium ratios as an
indicator of life-history strategies of freshwater fish species
within a brackish water system. Fish. Res. 46:313-325.
Kahsh, J. M.
1989. Otolith microchemistry: validation of the effects of
physiology, age and environment on otolith composition. J.
Exp. Mar Biol. Ecol. 132:151-178.
1990. Use of otolith microchemistry to distinguish the progeny
of sympatric anadromous and non-anadromous salmonids.
Fish. Bull. 88:657-666.
Kawakami, Y, N. Mochioka, K. Morishita, T. Tajima,
H. Nakagawa, H. Toh, and A. Nakazono.
1998. Factors influencing otolith strontium/calcium ratios in
Anquilla japonica elvers. Env. Biol. Fishes 52:299-303.
Limburg, K. E.
1995. Otolith strontium traces environmental history of
subyearling American shad Alosa sapidissima . Mar Ecol.
Prog. Ser 119:25-35.
Markowitz, A., D. Grambole, F. Herrmann, W. J. Trompetter,
T. Dioses, and R. W. Gauldie.
2000. Reliable micro-measurement of strontium is the key
to cracking the life-history code in the fish otolith. Nucl.
Instr and Meth. B. 168:109-116.
718
Fishery Bulletin 101(3)
Potts, P. J.
1987. A handbook of silicate rock analysis. Chapman and
Hall, New York, NY.
Radtke, R. L.
1995. Otolith microchemistry of charr — use in life history
studies. Nordic J. Freshwater Res. 71:392-395.
Radtke, R. L., D. W. Townsend, S. D. Folsom, and M. A. Morrison.
1990. Strontiumxalcium concentration ratios in otoliths of
herring larvae as indicators of environmental conditions.
Env. Biol. Fish. 27:51-61.
Rieman, B. E., Myers, D. L, and Nielsen, R. L.
1994. Use of otolith microchemistry to discriminate Oncor-
hynchus nerka of resident and anadromous origin. Can. J.
Fish. Aquat. Sci. 51:68-77.
Secor, D. H.
1992. Application of otolith microchemistry analysis to
investigate anadromy in Chesapeake Bay striped bass
Morone saxatilis. Fish. Bull. 90:798-806.
1999. Specifying divergent migration patterns in the concept
of stock: the contingent hypothesis. Fish. Res. 43:13-34.
Secor, D. H., and J. R. Rooker
2000. Is otolith strontium a useful scalar of life cycles in
estuarine fishes? Fish. Res. 46:359-371.
Thresher, R. E.
1999. Elemental composition of otoliths as a stock delineator
in fishes. Fish. Res. 43: 165-204.
Thresher, R. E., C. H. Proctor, J. S. Gunn, and I. R. Harrowfield.
1994. An evaluation of electron-probe microanalysis of
otoliths for stock delineation and identification of nursery
areas in a southern temperate groundfish, Nemadactylus
macropterus (Cheilodactylidae). Fish. Bull. 92:817-840.
Toole, C. L., D. F Markle, and P H. Harris.
1993. Relationships between otolith microstructure, micro-
chemistry, and early life history events in Dover sole,
Microstomus pacificus. Fish. Bull. 91:732-753.
Toole, C. L., and R. L. Nielsen.
1992. Effects of microprobe precision on hypotheses related
to otolith Sr:Ca ratios. Fish. Bull. 90:421-427.
Tzeng, W. N., K. P. Severin, and H. Wikstrom.
1997. Use of otolith microchemistry to investigate the
environmental history of European eel, Anquilla anquilla.
Mar. Ecol. Prog. Ser. 149:73-81.
Volk, E. C, A. Blakley, S. L. Schroder, and S. M. Kuehner.
2000. Otolith microchemistry reflects migratory characteris-
tics of Pacific salmonids: using otolith core chemistry to dis-
tinguish maternal associations with sea and freshwaters.
Fish. Res. 46:251-266.
Zimmerman, C. E.
2000. Ecological relation of sympatric steelhead and resi-
dent rainbow trout in the Deschutes River, Oregon. Ph.D.
diss., 116 p. Oregon State Univ., Corvallis, OR.
Zimmerman, C. E., and G. H. Reeves.
2000. Population structure of sympatric anadromous and
non-anadromous Oncorhynchus mykiss: evidence from
spawning surveys and otolith microchemistry. Can. J.
Fish. Aquat. Sci. 57:2152-2162.
2002. Identification of steelhead and resident rainbow trout
progeny in the Deschutes River, Oregon, revealed with oto-
lith microchemistry. Trans. Am. Fish. Soc. 131:986-993.
Fishery Bulletin 101(3)
719
Superintendent of Documents Publications Order Form
*5178
I I YEd, please send me the following publications:
Subscriptions to Fishery Bulletin
for $55.00 per year ($68.75 foreign)
The total cost of my order is $ .
. Prices include regular domestic
postage and handling and are subject to change.
(Company or Personal Name)
(Please tjrpe or print)
(Additional address/attention line)
(Street address)
(City, State, ZIP Code)
(Daytime phone including area code)
(Purchase Order No.)
Charge
your
order.
IT'S
EASY!
Please Choose Method of Payment:
I I Check Payable to the Superintendent of Documents
I I GPO Deposit Account
I I VISA or MasterCard Account
(Credit card expiration date)
-D
To fax
your orders
(202) 512-2250
(Authorizing Signature)
Mail To: Superintendent of Documents
P.O. Box 371954, Pittsburgh, PA 15250-7954
Thank you for
your order!
Fishery Bulletin
Guidelines for contributors
Content of papers
Articles
Articles are reports of 10 to 30 pages (double
spaced) that describe original research in one or
a combination of the following fields of marine
science: taxonomy, biology, genetics, mathematics
(including modeling), statistics, engineering, eco-
nomics, and ecology.
Notes
Notes are reports of 5 to 10 pages without an
abstract that describe methods and results not
supported by a large body of data. Although all
contributions are subject to peer review, responsi-
bility for the contents of articles and notes rests
upon the authors and not upon the editor or the
publisher It is therefore important that authors
consider the contents of their manuscripts care-
fully. Submission of an article is un-derstood to
imply that the article is original and is not being
considered for publication elsewhere. Manuscripts
must be written in English. Authors whose native
language is not EngUsh are strongly advised to
have their manuscripts checked for fluency by
English-speaking colleagues prior to submission.
Preparation of papers
Text
TMe page should include authors' full names and
mailing addresses (street address required) and
the senior author's telephone, fax number, e-mail
address, as well as a list of key words to describe the
contents of the manuscript. Abstract must be less
than one typed page (double spaced) and must not
contain any citations. It should state the main scope
of the research but emphasize the author's con-
clusions and relevEmt findings. Because abstracts
are circulated by abstracting agencies, it is impor-
tant that they represent the research clearly and
concisely. General text must be typed in double-
spaced format. A brief introduction should state the
broad significance of the paper; the remainder of
the paper should be divided into the following sec-
tions: Materials and methods. Results, Discussion
(or Conclusions), and Acknowledgments. Headings
within each section must be short, reflect a logical
sequence, and follow the rules of multiple subdi-
vision (i.e. there can be no subdivision without at
least two subheadings). The entire text should be
intelUgible to interdisciplinary readers; therefore,
all acronyms and abbreviations should be written
out and all lesser-known technical terms should be
defined the first time they are mentioned. The
scientific names of species must be written out the
first time they are mentioned; subsequent mention
of scientific names may be abbreviated. Follow Sci-
entific style and format: CBE manual for authors,
editors, and publishers (6th ed.) for editorial style
and the most current issue of the Aimerican Fish-
eries Society's common and scientific names of
fishes from the United States and Canada for
fish nomenclature. Dates should be written as fol-
lows: 11 November 1991. Measurements should be
expressed in metric units, e.g. metric tons (t). The
numeral one (1) should be typed as a one, not as a
lower-case el (1).
Footnotes
Use footnotes to add editorial comments regarding
claims made in the text and to document unpub-
lished works or works with local circulation. Foot-
notes should be numbered with Arabic numerals
and inserted in 10-point font at the bottom of the
first page on which they are cited. Footnotes should
be formatted in the same manner as citations.
If a manuscript is unpubhshed, in the process
of review, or if the information provided in the
footnote has been conveyed verbally, please state
this information as "unpubl. data," "manuscript
in review," and "personal commun.," respectively.
Authors are advised wherever possible to avoid ref-
erences to nonstandard literature (unpublished lit-
erature that is difficult to obtain, such as internal
reports, processed reports, administrative reports,
ICES council minutes, IWC minutes or working
papers, any "Research" or 'Svorking" documents,
laboratory reports, contract reports, and manu-
scripts in review). If these references are used,
please indicate whether they are available fi'om
NTIS (National Technical Information Service) or
from some other public depository. Footnote format:
author (last name, followed by first-name initials);
year; title of report or manuscript; type of report
and its administrative or serial number; name and
address of agency or institution where the report is
filed.
Literature dted
The Hterature cited section comprises works that
have been published and those accepted for pub-
lication (works in press) in peer-reviewed jour-
nals and books. Follow the name and year system
for citation format. In the text, write "Smith and
Jones (1977) reported" but if the citation takes
the form of parenthetical matter, write "(Smith
and Jones, 1977)." In the literature cited section,
list citations alphabetically by last name of senior
author: For example, Alston, 1952; Mannly, 1988;
Smith, 1932; Smith, 1947; Stalinsky and Jones,
1985. Abbreviations of journals should conform
to the abbreviations given in the Serial sources
for the BIOSIS previews database. Authors are
responsible for the accuracy and completeness of
all citations. Literature citation format: author
(last name, followed by first-name initials); year;
title of report or article; abbreviated title of the
journal in which the article was published, volume
number, page numbers. For books, please provide
publisher, city, and state.
Tables
Tables should not be excessive in size and must be
cited in numerical order in the text. Headings in
tables should be short but ample enough to allow
the table to be intelligible on its own. All unusual
symbols must be explained in the table legend.
Other incidental comments may be footnoted (use
italic arable numerals for footnote markers). Use
asterisks only to indicate probability in statistical
data. Place table legends on the same page as the
table data. We accept tables saved in most spread-
sheet software programs (e.g. Microsoft Excel).
Please note the following:
• Use a comma in numbers of five digits or more
(.e.g. 13,000 but 3000).
• Use zeros before all decimal points for values
less than one (e.g. 0.31).
Rgures
Figures include line illustrations, computer-gener-
ated line graphs, and photographs (or slides). They
must be cited in numerical order in the text. Line
illustrations are best submitted as original draw-
ings. Computer-generated line graphs should be
printed on laser-quality paper Photographs should
be submitted on glossy paper with good contrast.
All figures are to be labeled with senior author's
name and the number of the figure (e.g. Smith,
Fig. 4). Use Helvetica or Arial font to label ana-
tomical parts (line drawings) or variables (graphs)
within figures; use Times Roman bold font to label
the different sections of a figure (e.g. A, B, C).
Figure legends should explain all symbols and
abbreviations seen within the figure and should be
typed in double-spaced format on a separate page
at the end of the manuscript. We advise authors to
peruse a recent issue of Fishery Bulletin for stan-
dard formats. Please note the following:
• Capitalize the first letter of the first word of
axis labels.
• Do not use overly large font sizes to label axes
or parts within figures.
• Do not use boldface fonts within figures.
• Do not create outline rules around graphs.
• Do not use horizontal lines through graphs.
• Do not use large font sizes to label degrees of
longitude and latitude on maps.
• Indicate direction of degrees longitude and
latitude on maps (e.g. 170°E).
• Avoid placing labels on a vertical plane
(except on y axis).
•Avoid odd (nonstandard) patterns to mark
sections of bar graphs and pie charts.
Copyright law
Fishery Bulletin, a U.S. government pubhcation, is
not subject to copyright law. If an author wishes to
reproduce any part of Fishery Bulletin in his or her
work, he or she is obliged, however, to acknowledge
the source of the extracted literature.
Submission of papers
Send four printed copies (one original plus three
copies ) — clipped, not stapled — to the Scientific Edi-
tor, at the address shown below. Send photocopies
of figures with initial submission of manuscript.
Original figures will be requested later when the
manuscript has been accepted for publication.
Do not send your manuscript on diskette until
requested to do so.
Dr. Norm£m Bartoo
National Marine Fisheries Service, NCAA
8604 La Jolla Shores Drive
LaJoUa, CA 92037
Once the manuscript has been accepted for pub-
lication, you will be asked to submit a software
copy of your manuscript. The software copy should
be submitted in WordPerfect or Word format (in
Word, save as Rich Text Format). Please note that
we do not accept ASCII text files.
Reprints
Copies of published articles and notes are avail-
able free of charge to the senior author (50 copies)
and to his or her laboratory (50 copies). Additional
copies may be purchased in lots of 100 when the
author receives page proofs.
U.S. Department
of Commerce
Volume 101
Number 4
October 2003
Fishery
Bulletin
U.S. Department
of Commerce
Donald L Evans
Secretary
National Oceanic
and Atmospheric
Administration
Vice Admiral
Conrad C. Lautenbacher Jr.,
USN (ret.)
Under Secretary for
Oceans and Atmosphere
National Marine
Fisheries Service
William T. Hogarth
Assistant Administrator
for Fistienes
.^>^^'''°'%.
^^ATES 0» '^
The Fishery Bulletin (ISSN 0090-0656)
is published quarterly by the Scientific
Publications Office, National Marine Fish-
eries Service, NOAA, 7600 Sand Point Way
NE, BIN C 15700, Seattle, WA 981 15-0070.
Periodicals postage is paid at Seattle, WA,
and at additional mailing offices. POST-
MASTER: Send address changes for sub-
scriptions to Fishery Bulletin, Superin-
tendent of Documents, Attn.: Chief, Mail
List Branch. Mail Stop SSOM, Washing-
ton, DC 20402-9373.
Although the contents of this publica-
tion have not been copyrighted and may
be reprinted entirely, reference to source
is appreciated.
The Secretary of Commerce has deter-
mined that the publication of this peri-
odical is necessary according to law*for
the transaction of public business of this
Department. Use of funds for printing of
this periodical has been approved by the
Director of the Office of Management and
Budget.
For sale by the Superintendent of
Documents, U.S. Government Printing
Office, Washington, DC 20402. Subscrip-
tion price per year: $55.00 domestic and
$68.75 foreign. Cost per single issue:
$28.00 domestic and .$35.00 foreign. See
back for order form.
^
Scientific Editor
Dr. Norman Bartoo
Associate Editor
Sarah Shoffler
National Marine Fisheries Service, NOAA
8604 La Jolla Shores Drive
La Jolla, California 92037
Managing Editor
Sharyn Matriotti
National Marine Fisheries Service
Scientific Publications Office
7600 Sand Point Way NE, BIN C15700
Seattle, Washington 981 15-0070
Editorial Committee
Dr. Harlyn O. Halvorson
Dr. Ronald W. Hardy
Dr. Richard D. Methot
Dr Theodore W. Pietsch
Dr. Joseph E. Powers
Dr. Harald Rosenthal
Dr. Fredric M. Serchuk
Dr. George Watters
University of Massachusetts, Boston
University of Idaho, Hagerman
National Marine Fisheries Service
University of Washington, Seattle
National Marine Fisheries Service
Universitat Kiel, Germany
National Marine Fisheries Service
National Marine Fisheries Service
Fishery Bulletin web site: fishbull.noaa.gov
The Fishery Bulletin carries original research reports and technical notes on investigations in
fishery science, engineering, and economics. It began as the Bulletin of the United States Fish
Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery
Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through
volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing
through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system
began in 1963 with volume 63 in which papers are bound together in a single issue of the
bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a
periodical, issued quarterly. In this form, it is available by subscription from the Superintendent
of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in
limited numbers to libraries, research institutions. Slate and Federal agencies, and in exchange
for other scientific publications.
U.S. Department
of Commerce
Seattle, Washington
Volume 101
Number 4
October 2003
Fishery
Bulletih
'/'O'lne Biological Laboratory/
Wooos Hole Ocoanngraphic Institution
Library
NOV 2 1 2003
WQSUft I iijiNi tiit\ tjiUi
Contents
The conclusions and opinions expressed
in Ftshery Bulletin are solely those of the
authors and do not represent the official
position of the National Marine Fisher-
ies Service INOAA) or any other agency
or institution
The National Marine Fisheries Service
(NMFS) does not approve, recommend, or
endorse any proprietary product or pro-
prietary material mentioned in this pub-
lication No reference shall be made to
NMFS. or to this publication furnished by
NMFS, in any advertising or sales pro-
motion which would indicate or imply
that NMFS approves, recommends, or
endorses any proprietary product or pro-
prietary material mentioned herein, or
which has as its purpose an intent to
cause directly or indirectly the advertised
product to be used or purchased because
of this NMFS publication
Articles
721-731 Baum, Julia K., Jessica J. Meeuwig, and
Amanda C. J. Vincent
Bycatch of lined seahorses (Hippocampus erectus) in a
Gulf of Mexico shrimp trawl fishery
732-736 Bjorndal, Karen A., Alan B. Bolten, and Helen R. Mattins
Estimates of suiA/ival probabilities for oceanic-stage loggerhead
sea turtles (.Caretta caretta) in the North Atlantic
737-744 Chen, Yong, Margaret Hunter, Robert Vadas, and
Brian Beal
Developing a growth-transition matnx for the stock assessment of
the green sea urchin (Strongylocentrotus droebachiensis) off Maine
745-757 de Lestang, Simon, Norman G. Hall, and Ian C. Potter
Reproductive biology of the blue swimmer crab (Portunus pelagicus,
Decapoda: Portunidae) in five bodies of water on the west coast of
Australia
758-768 Dew, Jodi R., Jim Berkson, Eric M. Hallerman, and
Standish K. Allen Jr.
A model for assessing the likelihood of self-sustaining populations
resulting from commercial production of tnploid Suminoe oysters
(Crassostrea ahakensis) in Chesapeake Bay
769-777 Diaz-Jaimes, Pindaro, and Manuel Uribe-Alcocer
Allozyme and RAPD vanation in the eastern Pacific yellowfin tuna
(Thunnus albacares)
778-789 Govoni, John Jeffrey, Elisabeth H. Laban, and
Jonathan A. Hare
The early life history of swordfish iXiphias gladius) in the western
North Atlantic
790-799 Heales, Donald S., David T. Brewer, You-Gan Wang, and
Peter N. Jones
Does the size of subsamples take from multispecies trawl catches
affect estimates of catch composition and abundance?
Fishery Bulletin 101(4)
800—808 Laidig, Thomas E., Donald E. Pearson, and Lorraine L. Sinclair
Age and growth of blue rockfish (Sebastes mystinus) from central and northern California
809—821 Marin E., Baumar J., Antonio Quintero, Dany Bussiere, and Julian J. Dodson
Reproduction and recruitment of white mullet iMugil curema) to a tropical lagoon (fVlargarita Island, Venezuela) as
revealed by otolith microstructure
822—834 McDonough, Christopher J., William A. Roumillat, and Charles A. Wenner
Fecundity and spawning season of striped mullet (Mugil cephalus L.) in South Carolina estuaries
835-850 Nelson, Peter A.
Manne fish assemblages associated with fish aggregating devices (FADs): effects of fish removal, FAD size, fouling
communities, and prior recruits
851—859 Pajuelo, Jose G., Jose M. Lorenzo, and Muriel Gregoire
Age and growth of the bastard grunt iPomadasys inasus) Haemulidae) inhabiting the Canarian archipelago.
Northwest Africa
860-873 Punt, Andre E.
Evaluating the efficacy of managing West Coast groundfish resources through simulations
874-888 Purves, Martin G., David J. Agnew, Guillermo Moreno, Tim Daw, Cynthia Yau, and Graham Pilling
Distribution, demography, and discard mortality of crabs caught as bycatch in an experimental pot fishery for toothfish
(Dissostichus deginoides) in the South Atlantic
889-899 Stabenau, Erich K., and Kimberly R. N. Vietti
The physiological effects of multiple forced submergences in loggerhead sea turtles (Carena caretta)
900-909 Stephenson, Peter C, and Norm G, Hall
Quantitative determination of the timing of otolith ring formation from marginal increments in four marine teleost
species from northwestern Australia
Notes
910-914 Chan, Ricky W, K., Patricia I. Dixon, Julian G. Pepperell, and Dennis D. Reid
Application of DNA-based techniques for the identification of whaler sharks (Carcharhinus spp) caught in protective
beach meshing and by recreational fishenes off the coast of New South Wales
915-922 Ebert, Thomas A., and John R. Southon
Red sea urchins iStrongylocentrotus franascanus) can live over 100 years; confirmation with A-bomb '''carbon
923-932 Fulling, Gregory L., Keith D Mullin, and Carrie W. Hubard
Abundance and distribution of cetaceans in outer continental shelf waters of the U.S. Gulf of Mexico
933-938 Hata, David, and Jim Berkson
Abundance of horseshoe crabs iLimulus polyphemus) in the Delaware Bay area
939-948 Kerstetter, David W., Brian E. Luckhurst, Eric D. Prince, and John E. Graves
Use of pop-up satellite archival tags to demonstrate sun/ival of blue marlin (Makaira nigricans) released from pelagic
longline gear
949-950
2003 revieweis
951-960
2003 index
961
Subscription form
721
Abstract— Bycatch studies have largely
ignored population level efTects on fish
species of little commercial interest.
Here we analyze bycatch of the lined
seahorse ^Hippocampus erectus) in the
bait-shrimp trawl fishery in Hernando
Beach, Florida, providing the first
fisheries data for this species. Based
on catch per unit of effort (CPUE),
size, sex, and reproductive status of
trawled H. erectus, 1) approximately
72,000 seahorses were caught annu-
ally by this fleet, from a population of
unknown size, 2) trawling affected pop-
ulation cohorts differentially because of
temporal and spatial variation in CPUE
and population size, and 3) a greater
proportion of females than males was
removed in trawling. Our findings sug-
gest that trawling may affect seahorse
populations through direct mortality,
social disruption, and habitat damage.
However, the lack of specific abundance
or catchability estimates for H. erectus
means that the precise impact of trawl-
ing on this fish remains uncertain. This
paper focuses attention on the need
for research and monitoring of small
fishes that are caught incidentally in
nonselective gear.
Bycatch of lined seahorses (Hippocampus erectus)
in a Gulf of Mexico shrimp trawl fishery
Julia K. Baum
Jessica J. Meeuwig
Amanda C. J. Vincent
Proiect Seahorse
Department of Biology
McGill University,
1205 Dr Penfield Ave.
Montreal, Quebec, H3A 1B1, Canada
Present address (for J. K Baum); Department of Biology
Dalhousie University
1355 Oxford St.
Halifax, Nova Scotia, B3H 4J1, Canada
E-mail address (for J K Baum) baum@mathstat.dal.ca
Manuscript approved for publication
2 June 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull 101:721-731 (2003).
The incidental capture of marine or-
ganisms is now recognized as a seri-
ous problem in fisheries management
and marine conservation (Alverson et
al., 1994; lUCN, 1996; Alverson, 1997;
Jennings and Kaiser, 1998; FAO, 1999).
Shrimp trawl fisheries are the single
greatest source of bycatch, accounting
for 35% of the world's total bycatch
(Alverson et al., 1994). Bycatch research
has focused on marine megafauna, sea-
birds, and commercially important fish
species (see as examples Polacheck,
1989; Graham, 1995; Weimerskirch
et al., 1997; Julian and Beeson, 1998;
Pikitch et al., 1998; Galloway and Cole,
1999; Diamond et al., 2000). The conser-
vation impacts of bycatch for noncom-
mercial fishes and invertebrate species
remain largely unstudied (but see
Chan and Liew, 1986; Pettovello, 1999;
Milton, 2001). The few studies that
have evaluated incidental capture of
these species have focused on survival
rates of individuals (Hill and Wassen-
berg, 1990, 2000; Kaiser and Spencer,
1995; Probert et al., 1997; Mensink et
al., 2000) without addressing popula-
tion level effects of bycatch. However,
even species that comprise only a small
portion of the bycatch in a fishery may
experience significant impacts of inci-
dental harvest on their population size
and structure.
Seahorses are among those spe-
cies inferred to be greatly affected by
nonselective fishing gear, both because
intense trawling often covers seahorse
habitat and because their life history
traits likely render these fishes vul-
nerable to overexploitation (Vincent,
1996). Most studied seahorse species
are strictly monogamous (i.e. sexually
and socially), meaning that removal can
disrupt pairs and may reduce reproduc-
tive output (e.g. Vincent, 1995; Vincent
and Sadler, 1995; Kvarnemo et al.,
2000; Perante et al., 2002). Obligatory
parental care by males, combined with
relatively low fecundity, may reduce the
potential for population recovery from
overexploitation, although potentially
high survival of young may also offset
this apparent cost. In addition, sparse
distributions and low mobility suggest
that seahorses will be slow to recolonize
depleted areas (Perante et al., 2002;
Vincent et al.^).
Seahorses derived from bycatch ap-
pear to be contributing greatly to the
large and growing international trade
in these fishes (Vincent, 1996; Vincent
and Perry2). Consumer demand for
seahorses — both dried for traditional
medicine and curiosity trades, and,
less frequently, live for the aquarium
trade — is very high (Vincent, 1996).
Global demand for seahorses has sur-
passed supply and therefore trade has
increased and expanded geographically
' Vincent, A. C. J., K. L. Evans, and A. D.
Marsden. 2003. Home range behaviour
of the monogamous Australian seahorse.
Hippocampus whitei. Manuscript in
review.
722
Fishery Bulletin 101(4)
(Vincent, 1996; Vincent and Perry'^), placing populations
around the world under greater pressure. Much of this
market demand is met from retention of incidental land-
ings in shrimp trawls. Where no market has yet developed,
incidentally caught seahorses are discarded and the sur-
vival rate for these discarded seahorses is unknown.
Trade records and anecdotal evidence from other coun-
tries indicate that the United States has imported and
exported considerable numbers of both live and dried
seahorses in recent years (Vincent, 1996; Vincent and
Perry^). In Florida, the primary source of live seahorses in
the United States (Larkin and Degner, 2001), these fishes
ranked as the seventh most economically important orna-
mental fish group from 1990 to 1998, and seahorse landings
rose 184% during this period, whereas landings of each of
the more valuable fish groups declined (Adams et al., 2001).
Many seahorses in the United States are probably obtained
from bycatch of the shrimp trawl fisheries that operate in
known seahorse habitats along the Atlantic coast of the
United States.
Our objective was to document bycatch of the lined sea-
horse (Hippocampus erectus) in a live-bait shrimp trawl
fishery. This fish is often retained for the aquarium trade.
We quantify number, sex, size, and reproductive status of
trawl-caught seahorses and examine how these parameters
vary temporally and spatially. We also comment on poten-
tial conservation concerns resulting from this fishery.
Materials and methods
We focused our study on the lined seahorse (Hippocam-
pus erectus) because it was caught much more often than
its sympatric congenerics in the Gulf of Mexico — the
longsnout seahorse (H. reidi) and the dwarf seahorse (H.
zosterae). Hippocampus erectus is a large, deep-bodied
seahorse (adult height 5.5 to 18.5 cm: Lourie et al., 1999)
and has a geographic range that extends from southern
Canada to Argentina. Most species of seahorse have short
lifespans and low fecundity: Hippocampus erectus lives for
about four years (Lourie et al., 1999) and has broods of
about 100-1500 young (Teixeira and Musick, 2000). Hip-
pocam.pus erectus is found in shallow waters and offshore to
depths of over 70 m, primarily in mangroves and seagrass
beds (Vari, 1982). Like many other seahorse species, H.
erectus is listed as Vulnerable (A2cd) by the International
Union for Conservation of Nature and Natural Resources
(lUCN, 2002), based on suspected declines resulting from
habitat degradation and exploitation. However, as is the
case for many small fishes, there is little information
known about the biology of H. erectus, and no fishery data
exist for it.
Seahorse bycatch was assessed in the live-bait shrimp
trawl fishery operating from Hernando Beach, Florida
(Fig. 1). The fishery, using roller beam trawls, targets pink
shrimp iPenaeus duorarum) at night in seagrass beds and
relocates seasonally in Florida. Hernando Beach was cho-
'^ Vincent, A. C. J., and A. Perry (eds). 2003. Global trade in
seahorses. Manuscript in prep.
sen as our study site because it is a moderate-size fishing
port with 31 licensed trawlers and is active during the
summer sampling period. Boats were equipped with trawls
that had a slotted roller along the bottom of the frame, and
stainless steel finger bars attached vertically, 5 cm apart
along the length of the frame, to limit collection of benthic
substratum and other debris (Berkeley et al., 1985). Trawls
were towed from each side of the vessel in four configura-
tions: 1) one trawl per side, each measuring either 3.66 m,
4.27 m, or 4.88 m in length, or 2) two trawls per side, each
measuring 3.66 m in length. Net mesh sizes were 3.18-3.81
cm and the tail bag stretched mesh size was 2.54-3.18 cm.
Trawls usually lasted between 30 and 60 minutes, and fish-
ermen made multiple successive trawls in a night. Shrimp
were culled from the catch and held live onboard in aerated
holding tanks. Most bycatch was discarded overboard, al-
though some fishermen retained certain species, including
seahorses, for sale as aquarium fishes.
Bycatch sampling and data set description
Data on seahorse bycatch in the live-bait shrimp fishery
were collected on 95 fishing nights, from June to August
1998 and from June to July 1999, using three methods: 1)
we sampled bycatch onboard on 50 nights; 2) we recorded
seahorse bycatch data onshore on 14 nights from fishers
who retained seahorses to sell; and 3) we received data
from fishermen on their seahorse catches (including loca-
tion, time, number of trawls, and number of seahorses
per trawl) on a total of 31 nights. Onboard sampling was
semistratified in that we targeted our sampling to cover all
lunar and tidal phases and a variety of areas. However, we
were dependent on fishersmen's decisions about the time
and location of tows. We also collected anecdotal informa-
tion from 14 experienced fishermen about seahorse catches
over time.
During onboard sampling, we recorded the number of
seahorses caught, start and end time of trawl, depth and lo-
cation (Loran co-ordinates) of trawl, tidal and lunar phases,
and presence or absence of all bycatch species, including
biogenic habitat species. Hippocampus erectus found in
the catch were placed in a container of surface water
while measurements were made. They were then released,
except when fishermen chose to retain them for sale. We
measured, weighed, determined the sex, and recorded the
life history stage, reproductive status, and any injury for
each seahorse. Measurements were taken according to
Lourie et al. (1999) and included standard length, defined
as the length from the tip of the snout to the opercular
ridge and from the opercular ridge to straightened tail tip.
Seahorses that had lost tail rings were not included in the
length analysis. Unless precluded by logistic constraints,
wet weights were obtained onboard by using a 60-g Pesola
spring scale and onshore by using a 200-g Ohaus electronic
balance. Adult males were distinguished from females and
juveniles by the presence of a brood pouch. The standard
length of the smallest seahorse with a brood pouch (105.3
mm) was used as the division between adults andjuveniles
with the assumption that males and females matured at
the same size. Such an assumption may overestimate the
Baum et al.: Bycatch of Hippocampus erectus in a Gulf of Mexico shirimp trawl fisfiery
723
United Stales
of America
GULF OF
MEXICO
ATLANTIC
Florida OCEAN
OI& Ofto C2003 - Km
82^50'
82^40'
Figure 1
Map of the eastern Gulf of Mexico, showing the study site of Hernando Beach on the western
coast of Florida. The shore is shaded; additional solid lines indicate depth contours in feet.
Letter and number combinations (e.g. N3) represent fishing areas along the coast.
number of adult females because males that matured
after that size would have been included in our analysis
as females. We defined males as reproductively active if
they were pregnant or had recently released young (as
indicated by a loose pouch). Female reproductive state
was not included in this analysis because it is difficult to
determine reliably. We defined mortality to include sea-
horses already dead when the net was hauled and those
that died onboard.
We evaluated temporal and spatial patterns in catch per
unit of effort (CPUE), standard length, population struc-
ture, and reproductive status of the seahorse bycatch. Spe-
cifically, we tested 1) temporal effects of year, lunar and
tidal phase, and 2) spatial effects of area and depth. In
our estimates of seahorse CPUE, we used standard length
instead of biomass because female weight changes with
egg hydration and male weight increases greatly when
carrying embryos. Calculating CPUE per length (m) of
roller beam trawl controlled for variation in gear size. The
sampling unit was thus defined as the total or cumulative
standard length of seahorses caught during each tow (per
hour), per meter of trawl (per meter). Lunar phases were
defined as continuous variables by converting lunar day to
its angle, 6, based on a cycle of 29.5 days (=360°), with new
moon defined as 6 = 0°. These angles were then converted
to their cosine and sine functions for inclusion in linear
regression (deBruyn and Meeuwig, 2001). Tides were
semidiurnal in the Gulf of Mexico, ranging 1.3 m in tidal
level in the study area. High and low tides were defined as
each lasting two hours, with the remaining time classified
as ebb or flood accordingly. Spatial effects were analyzed
by dividing the total fishing ground into eight subareas ac-
724
Fishery Bulletin 101(4)
cording to their position with respect to the depth
contour and by identifying discrete geographical
clusters of trawls (Fig. 1). We then compared varia-
tion among fishing areas within years for those
areas with at least five observ'ations (areas II, 12,
Nl, N2, SI, S2 in 1998, and areas Nl, N3 and N4
in 1999). Interannual spatial comparisons were
not possible because there was little overlap in
sampled areas between the two years.
Statistical analysis
We based CPUE estimates on all trawls (?!=445).
Statistical analyses evaluating temporal and
spatial variation in seahorse bycatch included
only those trawls with nonzero seahorse observa-
tions (;i=205). The analysis required the data or
their residuals to be normal; this was achieved by
inverse hyperbolic sine transformations following
exclusion of zeros (Zar, 1996). It should be noted
that by excluding zeros we overestimated seahorse
CPUE and we lost information about areas where
seahorses were absent or rare. Our analyses
should be interpreted as applying to locations
where and times when seahorses were found in
sufficient numbers as to be caught.
We examined the data using ANOVA, AN-
COVA, linear regression and chi-square analyses
(Zar, 1996; SYSTAT, version 7.0, SPSS Inc., Chicago, ID.
All two-way and three-way ANOVAs included tests for
interactions. We used a general linear model because the
data were unbalanced. Interactions were removed from the
model if they were found to be nonsignificant. Models were
then rerun, followed by pairwise Tukey tests to indicate
where significant differences occurred (Zar, 1996; SYSTAT,
version 7). We report results for the final ANOVA only. The
Yates correction was applied to 2x2 chi-squares (Zar, 1996).
All significance levels were set to reject Hq at P<0.05 and
all means are reported with standard errors.
Results
Bait shrimp fishermen trawled from 11 to 24 km offshore,
between 1.8 and 6.4 m deep water (mean=3.76 ±0.87 m).
Fishermen typically left port between 17:00 and 19:00
and spent 5.8 ±0.23 h actively trawling per night (n=50
nights). Trawls lasted 40.2 ±11.4 min (n=445) on average,
and fishermen usually set 8 to 9 trawls per night. Distance
trawled could not be estimated because we were unable to
track trawler trajectories continuously and because they
changed direction during the tows. The benthic habitat was
composed primarily of seagrass iThalassia testudinum)hut
also included algae, coral, and sponge. Bycatch included at
least 118 species of fishes, invertebrates, and marine flora.
Catch per unit of effort
Hippocomptis erectus was the only seahorse species com-
monly caught in this fishery. Almost half of the trawls (46%)
6-
1998 • € O » • C
1999 O » • © O
I
4
O O •
3
Q.
U
4 i
y^^WY^^
150 160 170 180 190 200 210 220 230
Day of year
Figure 2
Variation in mean CPUE estimates by lunar phase transformed to
inverse hyperbolic sine. The curve represents the predicted values as
a function of the cosine ( 1998) and sine ( 1999) of day. The lunar phases
for both years are indicated. Filled circles and black line are for 1998;
open circles and gray line are for 1999.
caught H. erectus, and the number per trawl ranged from 0
to 16, whereas noH. reidi and only two of the much smaller
species, H. zosterae, were caught. In total, 916 H. erectus
were caught during the 95 documented fishing boat nights
of the two fishing seasons, resulting in an overall mean of
9.64 seahorses per fishing boat night. Mean CPUE for H.
erectus was 24.25 ±2.15 mm/{hxm), about one and a half
seahorses per hour per boat. If only trawls with seahorses
were included, CPUE was 52.52 ±3.80 mm/(hxm), or about
three seahorses per hour per boat (;i=205). Very high
CPUE was recorded on three nights: 16 July 1998 (mean
CPUE=122.0 ±22.5 mmAhxm), n=12 trawls), 28 June
1999 (mean CPUE=118.1 ±24. mm/(hxm), « = 12 trawls)
and 30 June 1999 (mean CPUE=154.9 ±36.6 mm/(hxm),
n=8 trawls). Bycatch is characterized by a high number of
low catches and infrequent large catches. Because the large
catches more likely reflect the spatial-temporal distribu-
tion characteristics of fish stocks rather than outliers of
the data (Ortiz et al. 2000), we analyzed the entire dataset
and then tested the robustness of our models by excluding
these three nights in order to assess their influence on the
CPUE patterns.
CPUE of nonzero trawls varied between years and with
lunar phase (Table 1), but not with tidal phase (P=0.15).
Trawls captured significantly more seahorses in 1999 than
in 1998 (P<0.0005). The effect of lunar phase varied be-
tween years: CPUE was highest on the lunar third quarter
in 1998, but only weakly significant and had slightly higher
CPUE on the full moon in 1999 (Fig. 2). The temporal varia-
tion in CPUE was largely driven by the three high CPUE
nights, but the effect persisted when these were excluded
(Table 1).
Baum et al Bycatch of Hippocampus erectus in a Gulf of Mexico sfnrimp trawl fishery
725
Table 1
General linear model of effects of year, lunar phase, and ai'ea on
Hernando Beach bait-shrimp trawl fishery.
the CPUE (n-
m/lh
xm) for
nonzero H. erectus bycatch trawls in the
CPUE by
traw
1
lexcluding
CPUE by trawl
CPUE of three highest nights)
Source n
F
P
n
F
P
Year 205
15.2
<0.0005
173
3.9
0.049
Lunar phase 205
173
cost 6)
19.1
<0.0005
4.1
0.045
sinte)
8.9
0.003
1998 areas III, 12, N1,N2, SI, S2) 116
7.4
<0.0005
105
2.3
0.037
1999 areas INI, N2,N3.N4) 87
8.3
<0.0005
68
0.3
0.85
In both years, there were significant differ-
ences in nonzero CPUE trawls among sites.
In 1998, CPUE was significantly higher in S2
than in 12, Nl, N2 and SI iFig. 3A, Table 1). In
1999. CPUE was significantly higher in area
Nl than in areas N3 and N4 IFig. 3B, Table
1). However, both of these spatial patterns,
like the lunar patterns, were driven primar-
ily by the high CPUE nights. Removing the
three outliers left a significant difference in
CPUE by area in 1998 only iTable 1). CPUE
did not vary with depth of the fishing ground
IP=0.67).
Size of seahorses
Mean standard length of adults (standard
length >105.3 mm) was 139.5 ±21.7 mm (n=465,
range 105.3-202 mm) and mean weight was
11.6 ±5.5 g (n=232, range 3-31 g). Hippocam-
pus erectus was sexually dimorphic: males had
a brood pouch and were significantly longer
than females (n=465, P<0.0005, Table 2), and
had a greater weight to standard length ratio,
although this latter difference was relatively
weak (P=0.04). Juveniles (standard length
<105.3 mm) had a mean standard length of
83.3 ±16.7 mm ln=65, range 41.4-105 mm)
and a mean weight of 2.4 ±1.0 g (?!=38, range
0.9-4.0 g).
In adults, standard length varied by year
{n=425, F=l.\, P=0.008) and by lunar phase
(7!=425, cos(e):P=5.4, P=0.02, sin(0):P=7.7,
P=0.006). Mean standard length was greater
in 1998 than in 1999 and highest on the new
moon. In 1998 significantly larger adult sea-
horses were caught in areas 12 and S2 than
in II and SI (n=229, P=13.7, P<0.0005; Fig.
4). There was no effect of area in 1999 (n=2\2,
F=1.9, P=0.14). Standard length was not re-
lated to depth or tide.
1 1() -
A 1
IIHI -
go -
II B
.so -
70 -
hO ■
-Ml -
<• .\Q
Catcfi per unit of effort (mm/(hym))
1
11
B
1
12 Nl N2 SI S2
1 B
J\0 -
(ill -
40 -
^ >0.10) or tidal phase (P>0.90). However, significantly
more juveniles were caught during the first quarter than
during other lunar phases (Table 3). Fishing area had a
significant effect on size class, with the highest numbers
of juveniles caught in N2 and N3 (Table 3).
The sex ratio (males as fraction of total) of 0.42 differed
significantly from a 1:1 ratio (;t:;-^= 19.56, df=l. P<0.001).
The sex ratio did not vary temporally or spatially (Table
3), but it did vary as a function of size class. There were
proportionally more males in the larger size class (>150
mm, 0.61) than in the smaller size class (<150 mm, 0.30)
(X,^=14.95, df=l,P<0.005).
About 25% of the male seahorses captured in 1998 were
considered to be reproductively active, whereas fewer
than 1% were reproductively active in 1999. Indeed, male
reproductive activity was higher in 1998 than 1999 even
after controUing for smaller male size in 1999 (Table 3). The
proportion of reproductively active males did not vary with
lunar or tidal phase but did vary significantly with area
Baum et al : Bycatch of Hippocampus erectus in a Gulf of Mexico sfirimp trawl fisfiery
727
Table 2
Descriptive statistics (sample size [n], mean, standard deviation [SD], minimum [mini and maximum [max]) for standard length
and weight of female (F) and male (M) seahorses. P values and sample sizes (n) indicate results off-tests evaluating sexual dimor-
phism in SL and weight.
Sex
n
mean
SD
min
max
Standard length (mm)
P<0.0005,n=465
Weight (g)
P<0.002, n=232
M
F
M
F
201
264
105
140
146.1
134.4
12.5
11 0
21.9
20.1
5.0
5.8
105.3
105.3
6.0
30
200.5
202.0
27.0
31.0
Table 3
Contingency tables on effects of year, lunar phase, and area on population structure of seahorses in the bycatch of the Hernando
Beach bait shrimp trawl fishery. Yates corrections were applied to 2x2 contingency tables.
Source
Juveniles: Adults
Sex ratio
df
df
X~
Reproductive state
df
P
Year
Lunar phase
Area
2.46
11.00
8.84
>0.10
<0.025
<0.05
0.15
3.69
13.3
>0.50
>0.25
>0.05
27.89
7.80
21.64
<0.001
>0.05
<0.001
M
(Table 3); almost half of the reproductively ac-
tive males were found in one area (S2),and 83%
of them were caught on one of the three nights
with very high CPUE (16 July 1998).
Mortality and injury
Fewer than 1% of seahorses died during tows or
sorting, but 4,7% (;i=28 of 588 seahorses) lost
tail rings. The mean loss was 22 of the usual 36
tail rings (Lourie et al., 1999), or 61% of the tail
(range=8-31 rings lost). Approximately 14% of
the losses (n=4 of 588 seahorses) appeared to
be the result of new wounds, probably caused
by the focal trawl. There was no effect of year
(P=0.25), sex (P=0.75), or reproductive status
(P=0.75) on incidence of seahorse injury. Postre-
lease mortality is unknown.
Discussion
Direct effects of the trawl fishery on seahorse
mortality
We estimate that this fleet catches almost 72,000 sea-
horses incidentally per annum, based on the mean CPUE
and given that 31 boats fished approximately 240 nights
per year. Most seahorses were returned to the wild in the
Hernando Beach fishery, but this may not be the case in
other live-bait shrimp trawl fleets in Florida (Vincent, pers.
_a
jiMi
i
wkk
41) Ml Sll M)ll i:0 1411 IM) IXd :iKi
Standard length (mm)
Figure 5
Length-frequency histogram for seahorse bycatch in 1998 and 1999.
obs.). We could not determine the potential fishing-induced
mortality for the Hernandez Beach H. erectus population,
even when all trawled seahorses were retained, because
seahorse catchability in roller beam trawl gear is unknown
and no studies have estimated H. erectus density in the
wild. Focal behavioral studies on congenerics similar in
728
Fishery Bulletin 101(4)
size to H. erectus have found varied densities: H. comes, an
exploited species associated with coral reefs in the tropics,
had localized densities of 0.019/m- in a marine protected
area, and much less elsewhere (Perante et al., 2002), and
an unexploited species, H. whitei, had localized densities
of 0.088-0. 215/m^ in a study area, and no seahorses were
found over large adjacent areas (Vincent et al.'). Our study
also suggests very patchy distributions of//, erectus (549f
of the trawls had no seahorses at all and the number of
seahorses per trawl set ranged from 0 to 16).
Although variation in CPUE may reflect differential
catchability by habitat, we suggest that in areas where
seahorses were caught, temporal rather than spatial ef-
fects drove CPUE. It is difficult to make conclusions about
variation in CPUE because data were unbalanced, in that
the areas trawled differed between years and among lu-
nar phases. However, analysis of variance on a subset of
data for three sites (Nl, N2, N3) on three lunar phases
(1^' quarter, full moon, 3'''' quarter) for which we had data
in both years («=149 trawls), indicated that there was a
strong effect of year, a weaker effect of lunar phase, and no
effect of site. These results suggest that CPUE was mainly
affected by temporal variation. Lunar patterns in CPUE
as a result of fish behavior and ecology are common (e.g.
Parrish, 1999). This would be consistent with observations
for other species of seahorses; H. comes in the Philippines
(Vincent et al.-*) and H. spinosissimus and H. trimaculatus
in Vietnam (Meeuwig et al.**) exhibited patterns in CPUE
with respect to lunar phase, although these species were
also distributed in patches in space.
Data from this study suggest that the H. erectus popu-
lation was spatially structured. In 1999, the mean size of
incidentally caught adult seahorses decreased, reflecting
the absence of the largest size class of males and an in-
crease in smaller females that year (Fig. 5). We attribute
this difference to spatial structuring: the shallower areas
(12, S2) where the largest male and female seahorses were
caught in 1998 were not fished during the 1999 sampling
season. Most of the seahorse bycatch were adult H. erectus;
the dearth of juvenile H. erectus (and dwarf seahorses, H.
zosterae) in the trawls reflects low catchability or retention
due to mesh size. Similar proportions of juvenile seahorses
were caught over the two sampling seasons. The ratio of
juveniles to adults appears to be temporally influenced
I proportionally more juveniles were caught on new moons),
but this variation probably also reflects spatial structuring
because these trawls occurred primarily in deeper offshore
areas (N2, N3) that were fished almost exclusively during
this lunar phase. Perhaps//, erectus undergoes ontogenetic
movement, between juvenile and adult life history stages,
and adults maintain site fidelity. Spatial size structuring
probably also occurs in other seahorse species, for the en-
tire population and for adults alone (//. comes, Meeuwig''';
H. guttulatus, Curtis''). A better understanding of the
spatial structuring of populations could allow for spatial
control of fishing effort to minimize bycatch.
We found a consistent, female-biased sex ratio in the
catch across the two years of our study, with only 42%
males. This bias may reflect the sex ratio of the H. erectus
population: a similar sex ratio (40% males) was found in a
population of//, erectus in Chesapeake Bay, Virginia (Teix-
eira and Musick, 2000). Female-biased sex ratios have also
been found in H. zosterae (33% males) when sampled by
pushnet (Strawn, 1958), and in H. abdominalis studied
underwater in Australia (Martin-Smith^). Many other
wild populations of seahorses studied underwater, however,
have documented equal numbers of males and females (//.
breviceps: H. comes: Moreau and Vincent*; Perante et al.,
1998; H. reidi: Dauwe, 1993; H. whitei: Vincent and Sadler,
1995). Sexual dimorphism in H. erectus was too slight to
explain different catchability of the two sexes and would,
in any case, have favored the capture of males. The dispro-
portionate catch of females could have arisen from spatial
segregation by sex; the greater catches of reproductively
active males in shallower areas suggests that males may
spend most of their time inshore of the trawled area. We
also cannot discount the possibility that some seahorses
classified as females may have been immature males, and
the sex ratio in the population could in fact be 1:1.
The proportion of reproductively active seahorses in the
bycatch was lower than expected, particularly in 1999. Our
study occurred during summer, within the breeding season
for the congeneric and sympatric //. zosterae in Florida (Feb-
ruary to October; Strawn, 1958), and for H. erectus in Ches-
apeake Bay (May to October; Teixeira and Musick, 2000;
Vincent, personal obs.). Males of all studied seahorse species
were reproductively active almost continuously throughout
the breeding season (Dauwe, 1993; Nijhoff, 1993; Vincent
and Sadler, 1995; Perante et al., 2002), often remating the
same day that they release their young (Vincent and Sadler,
1995). In our study, trawling may have occurred outside the
primary breeding areas for male H. erectus, but catches of re-
productively nonactive adult males during the breeding sea-
son also suggest that repeated trawling may have disrupted
breeding in the population. A further indication of possible
spatial structuring in the population (by reproductive sta-
tus and size) is that almost half of the reproductively active
males caught in 1998 were found in S2, the shallowest area;
this area was not sampled in 1999 when few reproductively
active males were found. Such spatial structuring offers the
possibility of trawling outside the breeding area.
^ Vincent, A. C. J., J. J, Meeuwig, M. G. Pajaro, and N. C. I'eranto.
Seahorse catches in the central Philippines: characteristics and
conservation implications. Manuscript in prep.
" Meeuwig, J. J., D. H. Hoang, T S. Ky, S.D. Job, and A. C. J. Vincent.
Bycatch landings of seahorses in central Vietnam. Manuscript
in prep.
•'' Meeuwig, J. J. Life history parameters of the exploited seahorse
Hippocnmpus comes: a length based analysis. Manuscript in
prep.
^ Curtis, J. 2002. Unpubl. data. Project Seahorse. Fisheries
Center, The University of British Columbia, 2204 Main Mall,
Vancouver, BC, V6T 1Z4, Canada.
" Martin-Smith, K. 2002. L'npubl. data. Project Seahorse,
Fisheries Center. The University of British Columbia, 2204
Main Mall, Vancouver, BC, V6T 1Z4, Canada.
8 Moreau, M-A., and A. C. J. Vincent. 2000. Unpubl. data.
Project Seahorse, Fisheries Center, The University of British
Columbia, 2204 Main Mall, Vancouver, BC,V6T 1Z4, Canada.
Baum et a\ : Bycatch of Hippocampus erectus in a Gulf of Mexico sfirimp trawl fisfiery
729
Indirect effects of the trawl fishery on mortality
Direct immediate mortality from trawling and culling was
rather low, probably in part because trawl sets were of very
short duration in this live-bait fishery More importantly,
most H. erectus caught in the Hernando Beach fishery were
returned to sea, rather than retained as in some other Flor-
ida trawl fisheries (Vincent, pers. obs.). Indirect impacts of
the fishery may, however, be considerable.
Seahorses caught in trawls may experience high postre-
lease mortality. A study in the live-bait shrimp trawl fishery
in Tampa Bay Florida (Meyer et al, 1999), found that only
one of four of the congeneric dwarf seahorse (H. zosterae)
caught as bycatch were alive in the holding tank of seawater
36 hours after collection (Meyer^). Such tows lasted only 5
minutes (Meyer et al., 1999); therefore trawl-induced mor-
tality could be greater in the Hernando Beach fishery (with
trawls of 30-50 minutes), although H. erectus are larger and
perhaps more robust than H. zosterae. Like other discarded
bycatch, seahorses may also be subject to intense predation
upon release. Predation on discarded fish has been observed
on prawn trawlers in Australia (Hill and Wassenberg, 1990)
and within the bait shrimp fishery of Tampa Bay ( Meyer et
al., 1999). Captains of bait-shrimp boats concurred that this
is commonplace in the Hernando Beach fishery, and we fre-
quently observed bottlenosed dolphins (Tursiops truncatus)
and schools of ladyfish (Elops saurus) swimming alongside
the boats, feeding on discarded bycatch.
Trawling may significantly disrupt seahorse populations,
particularly if they are spatially structured, as the present
study suggests. The disproportionate removal of females
could reduce mating opportunities, especially if//, erectus
is monogamous, as are most studied seahorse species (e.g.
Vincent, 1995; Vincent and Sadler, 1995; Kvarnemo et al.,
2000; Perante et al., 2002). Trawling, on account of re-
peated intrusion onto breeding grounds, could also disrupt
courtship and negatively affect reproduction, In heavily
exploited areas of the fishery where fishermen repeatedly
trawl productive areas, seahorses may face cumulative
stress. For example, tail injuries are likely a serious wound
for seahorses, given that their tails are essential to grasp
holdfasts and may play a key role in mating competition, as
they do with Hippocampus fuscus (Vincent, 1994).
Benthic habitat degradation is another potential indirect
effect of live-bait shrimp trawling on seahorses. Bottom-fish-
ing gear can reduce habitat complexity by removing emer-
gent epifauna, smoothing sedimentary bedforms and by re-
moving structure-forming species such as corals and sponges
(Hutchings, 1990; Auster et al., 1996; Auster and Langton,
1999; Thrush and Dayton, 2002), Roller beam trawls also
affect habitat complexity by redistributing macroalgae and
seagrass (Meyer et al., 1999). We estimated that seagrasses
comprised between 50% and 80% of the volume of the catch
for each trawling operation. Although roller beam trawls are
assumed to have low impact on seagrass habitat (Tabb and
Kenny, 1969; Meyer et al., 1999), the effects of long-term re-
petitive trawling have not been tested, and it is possible that
species composition and abundance, including that ofH. erec-
tus, have been adversely affected (Watling and Norse, 1998).
Summary
Despite the relatively low direct mortality of seahorse per
boat, the live-bait trawl fishery has the potential to affect
seahorse populations, both directly and indirectly. The key
question is whether the level of exploitation, and subse-
quent impacts, represents a conservation concern. Our
evidence is inconclusive. Perhaps only the skewed sex ratio
and low proportion of reproductively active males suggest
a potential problem. However, fishermen have consistently
reported that seahorse catch per boat has declined greatly
over the past two decades in this area. Effects of trawling
are also almost certainly greater in food shrimp trawl fish-
eries, which trawl with much larger gear for longer peri-
ods, and obtain substantially more bycatch, with higher
mortality. Our analysis should thus be seen as a first step
in identifying areas for which more information is needed,
specifically estimating abundance and fishing mortality,
and understanding spatial structuring in H. erectus.
This paper focuses attention on the need for research on
and monitoring of small fishes that may be affected by non-
selective fishing gear Management responses to minimize
bycatch have focused primarily on seabirds, sea turtles,
and commercially important finfishes, but trawl fisheries
may also have significant impacts on the many small ma-
rine organisms obtained as bycatch, even if they comprise
only a small proportion of the bycatch. Bycatch excluder
devices are unlikely to be effective in reducing incidental
catches of these species. Temporal variation in CPUE and
spatial population structuring, as observed in our present
study for H. erectus, suggest that time-area closures may be
a pragmatic solution for reducing incidental catch.
Acknowledgments
This paper is a contribution from Project Seahorse. We
thank Jana Schulz for her assistance with fieldwork, Daniel
and Patricia Mohr for their support during fieldwork, James
Boxall for preparing the map, A. DeBruyn, L. Crowder M.
Kaiser, and an anonymous reviewer for providing helpful
comments on an earlier draft of this manuscript. This study
would not have been possible without the cooperation and
support of many of the shrimp boat captains and crew in
Hernando Beach. This research was funded through an
NSERC summer undergraduate award to JKB, support
from the Community Fund (UK) and Guylian Chocolates
Belgium for JJM, and an NSERC operating grant to ACJV.
Literature cited
Meyer, D. 1999. Personal commun. NOAA Center for
Coastal Fisheries and Habitat Research, Beaufort Laboratory,
101 Fivers Island Road, Beaufort, NC 28516.
Adams, C, S. Larkin, and D. Lee.
2001. Volume and value of marine ornamentals collected in
Florida, 1990-98. Aquar Sci. Conserv. 3(l-3):25-36.
730
Fishery Bulletin 101(4)
Alverson, D. L.
1997. Global assessment of fisheries bycatch and discards:
a summary overview. In Global trends: fisheries manage-
ment {E. K. Pikitch, D. D. Huppert, and M. P. Sissenwine,
eds.), 115-125 p. Am. Fish. Soc. Symp., Vol. 20.
Alverson, D. L., M. H. Freeberg, S. A. Murawski, and J. G. Pope.
1994. A global assessment of fisheries bycatch and discards,
233 p. FAOFish. Biol. Tech. Pap. 339. FAO, Rome.
Auster, P. J., and R. W. Langton.
1999. The effects of fishing on fish habitat. In Fish habitat:
essential fish habitat (EFH) and rehabilitation (L. Benaka,
ed. ), p. 150-187. Am. Fish. Soc. Rep. 22.
Auster, P. J., R. J. Malatesta, R. W. Langton, L. Watling,
P. C. Valentine, C. L. S. Donaldson, E. W. Langton,
A. N. Shepard, and I. G. Babb.
1996. The impacts of mobile fishing gear on seafloor habi-
tats in the Gulf of Maine (Northwest Atlantic): implica-
tions for conservation of fish populations. Rev. Fish. Sci.
4:185-202.
Berkeley, S. A., D. W. Pybas, and W. L. Campos.
1985. Bait shrimp fishery of Biscayne Bay, 16 p. Fla. Sea
Grant Ext. Prog. Tech. Pap. 40. Florida Sea Grant, Gaines-
ville, FL.
Chan, E. H., and H. C. Liew.
1986. Characteristics of an exploited tropical shallow-water
demersal fish community in Malaysia. In Proceedings of
the first Asian fisheries forum (J. L. Maclean, L. B. Dizon,
and L. V. Hosillows, eds.), p. 349-352. Asian Fisheries
Society, Manila, Philippines.
Dauwe, B.
1993. Ecology of the seahorse Hippocampus reidi on the
Bonaire coral reef (N.A.): habitat, reproduction and com-
munity interaction. M.S. thesis, 65 p. Rijksuniversiteit,
Groningen, Netherlands.
deBruyn, A. M. H., and J. J. Meeuwig.
2001. Detecting lunar cycles in marine ecology: periodic
regression versus categorical ANOVA. Mar. Ecol. Prog.
Ser. 214:307-310.
Diamond, S. L., L. G. Cowell, and L. B. Crowder
2000. Population effects of shrimp trawl bycatch on Atlantic
croaker Can. J. Fish. Aquat. Sci. 57(10):2010-2021.
FAO (Food and Agriculture Organization of the United Nations).
1999. The state of world fisheries and aquaculture 1998,
112 p. FAO, Rome.
Galloway, B. J., and J. G. Cole.
1999. Reduction of juvenile red snapper bycatch in the Gulf
of Mexico shrimp trawl fishery. N. Am. J. Fish. Manag.
19(2):342-355.
Graham, G. L.
1995. Finfish bycatch from the Southeastern shrimp fishery.
In Solving bycatch: considerations for today and tomorrow,
p. 115-119. Univ. Alaska Sea Grant College Program
Report 96-03, Fairbanks, AK.
Hill, B. J., and T. J. Wassenberg.
1990. Fate of discards from prawn trawlers in Torres Strait.
Aust. J. Mar Freshw. Res. 41:53-64.
2000. The probable fate of discards from prawn trawlers fish-
ing near coral reefs. A study in the northern Great Barrier
Reef Australia. Fish. Res. 48(3):277-286.
Hutchings, P.
1990. Review of the effects of trawling on macrobenthic
epifaunal communities. Aust. J. Mar Freshw. Res. 41:
111-120.
lUCN (International Union for Conservation of Nature and
Natural Resources).
1996 World Conservation Congress, 1"' session. Resolutions
and recommendations. 1.16 Fisheries by-catch; Montreal,
Canada, 14-23 October 1996. lUCN, Gland, Switzerland.
2002. 2002 lUCN red list of threatened species. lUCN,
Gland, Switzerland and Cambridge, UK. http://www.redlist.
org/ [Accessed 21 January 2003].
Jennings, S., and M. J. Kaiser
1998. The effects of fishing on marine ecosystems. Adv.
Mar Biol. 34:201-352.
Julian, F, and M. Beeson.
1998. Estimates of marine mammal, turtle and seabird
mortality for two California gillnet fisheries: 1990-1995.
Fish. Bull. 96:271-284.
Kaiser, M. J., and B. E. Spencer
1995. Survival of by-catch from a beam trawl. Mar Ecol.
Prog. Ser. 126:31-38.
Kvarnemo, C, G. I. Moore, A. G. Jones, W. S. Nelson, and
J. C. Avise.
2000. Monogamous pair bonds and mate switching in the
Western Australian seahorse Hippocampus subelongatus.
J. Evol. Biol. 13:882-888.
Larkin, S. L., and R. L. Degner.
2001. The U.S. wholesale market for marine ornamentals.
Aquar Sci. Conserv. 3(l):13-24.
Lourie, S. A., A. C. J. Vincent, and H. J Hall.
1999. Seahorses: an identification guide to the world's
species and their conservation, 214 p. Project Seahorse,
London, UK.
Mensink, B. P., C. V. Fischer, G. C. Cadee, M. Fonds,
C. C. Ten Hallers-Tjabbes, and J. P. Boon.
2000. Shell damage and mortality in the common whelk
Buccinum undatum caused by beam trawl fishery. J. Sea
Res. 43(1 ):53-64.
Meyer, D. L., M. S. Fonseca, P. L. Murphey, R. H. McMichael, Jr.
M. M. Byerly, M. W. LaCroix, P E. Witfield, and G. W. Thayer
1999. Effects of live-bait shrimp trawling on seagrass beds and
fish bycatch in Tampa Bay Florida. Fish. Bull. 97:193-199.
Milton, D.
2001. Assessing the susceptibility to fishing of populations of
rare trawl bycatch: sea snakes caught by Australia's north-
em prawn fishery. Biol. Conserv. 101:281-290.
Nijhoff, M.
1993. Reproductive cycle of the seahorse Hippocampus reidi
on the Bonaire coral reef M.S. thesis, 49 p. Rijksuniver-
siteit, Groningen, Netherlands.
Ortiz, M., C. M. Legault, and N. M. Ehrhardt.
2000. An alternative method for estimating bycatch from the
U.S. shrimp trawl fishery in the Gulf of Mexico, 1972-1995.
Fish. Bull. 98:583-599.
Parrish, J. K.
1999. Using behavior and ecology to exploit schooling fishes.
Environ. Biol. Fishes 55:157-181.
Perante, N. C, M. G. Pajaro, J. J. Meeuwig, and A. C. J. Vincent.
2002. Biology of Hippocampus comes in the central
Philippines. J. Fish Biol. 60(4):821-837.
Perante, N. C, A. C. J. Vincent, and M. G. Pajaro.
1998. Demographics of Hippocampus comes seahorses in
Bohol, central Philippines. In Proceedings of the third
international conference on the Marine Biology of the South
China Sea; Hong Kong 1996, p. 439-148. Hong Kong Univ
Press, Hong Kong.
Pettovello, A. D.
1999. By-catch in the Patagonian red shrimp {Pleoticus
muelleri) fishery. Mar. Freshw. Res. 50:123-127.
Pikitch, E. K., J. R. Wallace, E. A. Babcock, D. L. Erickson,
M. Saclens, and G. Oddsson.
1998. Pacific halibut bycatch in the Washington, Oregon, and
Baum et a\ Bycatch of Hippocampus erectus in a Gulf of Mexico shrimp trawl fishery
731
California groundfish and shrimp trawl fisheries. N. Am.
J. Fish. Manag. 18:569-586.
Polacheck.T.
1989. Harbour porpoises and the gillnet fishery: Incidental
takes spur population studies. Oceanus 32:63-70.
Probert, P. K., D. G. McKnight, and S. L. Grove.
1997. Benthic invertebrate bycatch from a deep-water trawl
fishery, Chatham Rise, New Zealand. Aquatic conserva-
tion: marine and freshwater ecosystems 7(l):27-40.
Strawn, K.
1958. Life history of the pygmy seahorse, Hippocampus zos-
ierae Jordan and Gilbert, at Cedar Key, Florida. Copeia
1958:16-22.
Tabb, D. C, and N. Kenny
1969. A brief history of Florida's live bait shrimp fishery
with description of fishing gear and methods. FAO Fish.
Rep. 57:1119-1134.
Teixeira, R. L., and J. A. Musick.
2000. Reproduction and food habits of the lined seahorse.
Hippocampus erectus (Teleeostei: Syngnathidae) of Chesa-
peake Bay, Virginia. Rev. Bras. Biol. 61(l):79-90.
Thrush, S., and P K. Dayton.
2002. Disturbance to marine benthic habitats by trawling
and dredging: implications for marine benthic biodiversity.
Ann. Rev Ecol. Syst. 33:449-473.
Vari, R. P.
1982. Fishes of the western North Atlantic, subfamily Hippo-
campinae. The seahorses, p. 173-189. Sears Foundation for
Marine Research Memoir, vol. 1, Yale Univ., New Haven, CT.
Vincent, A. C. J.
1994. Operational sex ratios in seahorses. Behaviour 128:
153-167.
1995. A role for daily greetings in maintaining seahorse pair
bonds. Anim. Behav. 49:258-260.
1996. The international trade in seahorses, vii + 163 p.
TRAFFIC International, Cambridge, UK.
Vincent, A. C. J., and L. M. Sadler
1995. Faithful pair bonds in wild seahorses, Hippocampus
whitei. Anim. Behav. 50:1557-1569.
Watling, L., and E. A. Norse.
1998. Disturbance of the seabed by mobile fishing gear: a
comparison to forest clear cutting. Conserv. Biol. 12(6):
1180-1197.
Weimerskirch, H., N. Brothers, and P. Jouventin.
1997. Population dynamics of wandering albatross Diomedea
exulans and Amsterdam albatross D. amsterdamensis in the
Indian Ocean and their relationships with long-line fisher-
ies; conservation implications. Biol. Conserv. 79:257-270.
Zar, J. H.
1996. Biostatistical analysis, 3"^ ed., 662 p. Prentice Hall,
Upper Saddle River, NJ.
732
Abstract— Estimates of instantaneous
mortality rates (Z) and annual appar-
ent survival probabilities (.0) were
generated from catch-curve analyses
for oceanic-stage juvenile loggerheads
(Caretta caretta) in the waters of the
Azores. Two age distributions were
analyzed: the "total sample" of 1600 log-
gerheads primarily captured by sight-
ing and dipnetting from a variety of
vessels in the Azores between 1984 and
1995 and the "tuna sample" of 733 log-
gerheads (a subset of the total sample)
captured by sighting and dipnetting
from vessels in the commercial tuna
fleet in the Azores between 1990 and
1992. Because loggerhead sea turtles
begin to emigrate from oceanic to neritic
habitats at age 7, the best estimates of
instantaneous mortality rate (0.094)
and annual survival probability (0.911)
not confounded with permanent emi-
gration were generated for age classes
2 through 6. These estimates must be
interpreted with caution because of the
assumptions upon which catch-curve
analyses are based. However, these are
the first directly derived estimates of
mortality and survival probabilities for
oceanic-stage sea turtles. Estimation
of survival probabilities was identified
as "an immediate and critical require-
ment" in 2000 by the Turtle Expert
Working Group of the U.S. National
Marine Fisheries Service.
Estimates of survival probabilities for
oceanic-stage loggerhead sea turtles
{Caretta caretta) in the North Atlantic
Karen A. Bjorndal
Alan B. Bolten
Archie Carr Center for Sea Turtle Research and Department of Zoology
University of Florida
P.O Box 118525
Gainesville, Florida 32611
E-mail address (for K, A. B|orndal): kab@zoology ufl edu
Helen R. Martins
Deparlamento de Oceanografia e Pescas
Universidade dos Azores
PT-9901-862 Horta
Acores, Portugal
Manuscript approved for publication
21 April 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:732-736(2003).
A major gap in our understanding of
sea turtle demography is the level of
mortality — both natural and human-
induced — experienced by wild popula-
tions. Lack of directly derived estimates
of mortality (or survival) probabilities
for the juvenile oceanic-stage in sea
turtle populations is a critical source of
uncertainty in development of popula-
tion models and evaluation of man-
agement plans. In current population
models, survival estimates for juvenile,
oceanic-stage sea turtles are fitted
parameters, not directly derived esti-
mates, of survival (Chaloupka, in press;
Heppell et al., in press). The Turtle
Expert Working Group ( 2000) identified
the estimation of survival probabilities
as "an immediate and critical require-
ment." Population models indicate that
survival probability of juvenile oceanic-
stage loggerhead sea turtles (Caretta
caretta) has a substantial effect on
overall population growth (Chaloupka,
in press; Heppell et al., in press).
Catch-curve analyses have been used
for many years to estimate survival
probabilities for species harvested in
commercial fisheries and, less fre-
quently, for other species (Seber, 1982).
Estimates of survival probabilities
have been generated from catch-curve
analyses for subadult neritic-stage
populations of loggerhead sea turtles
(Frazer, 1987; Epperly et al., 2001) and
Kemp's ridley sea turtles (Lepidochelys
kempii) (Turtle Expert Working Group,
2000) based on stranding data. Catch-
curve analyses confound mortality and
permanent emigration, and thus gen-
erate estimates of apparent survival
probability (CP),
0 = S (1 - emigration ),
where S = true survival probability;
and
emigration = the probability of perma-
nent emigration.
We estimated survival probabilities
(both and S) for juvenile oceanic-
stage loggerhead sea turtles in the wa-
ters around the Azores, a developmental
habitat for the population of loggerhead
sea turtles that nest on beaches in the
southeastern United States (Bolten
et al., 1998). We applied catch-curve
analyses to two age distributions of
loggerhead sea turtles.
Methods
Two size distributions were compiled
for this study The first ("total sample")
comprised 1600 oceanic-stage logger-
head sea turtles that were captured
from 1984 through 1995 in the waters
of the Azores. Except for a few of the
smallest of these sea turtles found as
stranded carcasses, they were collected
in dipnets after being sighted at the
surface of the ocean from the decks of a
B|orndal et al.: Estimates of survival for Caretta caretta In the North Atlantic
733
variety of vessels. Loggerhead sea turtles captured
on longline hooks were excluded from this sample
to meet the requirement of equal probability of cap-
ture across the age range ( see "Results" section ). The
turtles were measured, tagged, and released soon
after capture. The juvenile loggerhead sea turtles
ranged in size from 8.5 to 82.0 cm curved carapace
length (mean=33.1 cm, SD=11.6) measured from
the anterior point at midline to the posterior notch
at midhne between the supracaudals (Bolten, 1999).
For 248 turtles, straight-line carapace measure-
ments were converted to curved carapace length,
as described in Bjorndal et al. (2000).
The second age distribution ("tuna sample") was
a subset of the total sample and comprised 733 log-
gerhead sea turtles captured by crews of commercial
tuna vessels in the Azores between 1990 and 1992.
We analyzed the tuna sample in addition to the to-
tal sample because the tuna sample was collected
over a shorter interval ( 1990-92) than was the total
sample (1984-95). In addition, sizes of vessels from
which turtles were captured were more consistent
for the tuna sample. This collaborative project with
the tuna fleet is described in Bolten et al. (1993);
sea turtles are not bycatch in the tuna fishery. Sea
turtles were sighted at the surface while the crews
were scanning for indications of tuna feeding ac-
tivity. The turtles were then captured in dipnets,
tagged and measured by a crew member, and re-
leased at sea. The juvenile loggerhead sea turtles
ranged from 11.0 to 82.0 cm curved carapace length
( mean=33.5, SD= 1 1.2 ). No conversion from straight
to curved measurements was required.
The size distributions were converted to age dis-
tributions by using a size-at-age function developed
for this population based on a skeletochronological
study (Table 1; Bjorndal et al., 2003). Catch curves
were generated for each age distribution by plot-
ting the natural log of A^^ against x, where N^ is
the number of turtles of age x. The catch curves
were truncated by excluding age classes with
fewer than five individuals, as recommended by Seber
(1982). The age at which the population fully recruited
to the capture method (threshold age) was identified as
the age with the highest ln(A'',). Linear regression analy-
ses of the values on the right-hand or declining slope of
the distribution were used to generate estimates of total
instantaneous mortality rate (Z), which is expressed on
an annual basis and is the absolute value of the slope of
the regression line. Annual apparent survival probability
(46 cm — the size at which they
begin to leave oceanic habitats and recruit to
neritic habitats (Bjorndal et al., 2000). Thus, the
sharper decline beginning with age 7 reflects the
migration of turtles out of the sampling area.
Estimates of total instantaneous mortality rate
(Z) and annual apparent survival probability
(4>) were generated for three age ranges for each
sample: from threshold age to final age, from
threshold age to pre-emigration age (6 years),
and from pre-emigration age to final age. For the
total sample, the age ranges were 2 to 12 years, 2
to 6 years, and 6 to 12 years, respectively; for the
tuna sample, the ages were 4 to 11, 4 to 6, and 6
to 11 years (Fig. 2, Table 2).
Discussion
Estimates of mortality and survival generated
from catch curves should be interpreted with
caution for several reasons. First, the analysis
assumes a stable age distribution, which we
cannot confirm for North Atlantic loggerhead sea
turtles. Second, the analysis assumes that mor-
tality rates were consistent over the years of the
study. The similarity of the mortality and survival
estimates between the total sample ( 1984-95) and
the tuna sample ( 1990-92) suggests that the esti-
mates for the total sample are not greatly affected
by heterogeneity among years. Third, converting
size distributions to age distributions based on a
size-at-age function introduces some level of error.
We believe the error from our size-at-age function
is small, as discussed in Bjorndal et al. (in press).
Fourth, the analysis assumes no size or age effect on mor-
tality rates. The catch curves for both the total sample and
tuna sample reveal a size or age effect with a pivot point at
age 6. This size or age effect reflects the beginning of emi-
gration out of our study area. Loggerhead sea turtles begin
to leave oceanic habitats around the Azores and recruit to
neritic habitats at 7 years of age, at -46 cm curved carapace
length (Bjorndal et al., 2000, 2003). This change in slope
demonstrates the fifth difficulty in interpreting catch-
curve estimates: permanent emigration and mortality are
confounded in the estimates. That is, declines in numbers
with age, whether they are due to emigration or mortality,
are included in the estimate of mortality. The confounding
of emigration and mortality can introduce a major error in
estimates of mortality in populations — such as sea turtle
populations in oceanic and neritic habitats — that undergo
developmental migrations.
— I 1 1 1 1 1 1 1 1 1 1 r-
01 23456789 10 11 12
- 6n
5
4
3 -
2 -
1 -
0 ^
B
• •---^--_
0 1
5 6 7
Age (years)
10
— r—
11
Figure 2
Catch curves for (A) total sample and (B) tuna sample. Dashed lines
are linear regressions for entire sample, threshold age to age 6, and
age 6 to final age (see Table 2 for regression statistics).
Because little permanent emigration apparently occurs
before the age of 7, the survival estimates for the ages
prior to 7 years are our best estimates of true survival (S).
As can be seen in Table 2, the estimates of total instanta-
neous mortality and annual survival are similar for the
two samples. We believe that the estimate of S (0.911 and
0.894, respectively, for the total sample and tuna sample)
would apply to the entire life stage over the size range from
20 to 65 cm CCL for most sources of mortality other than
fisheries biased to large sizes, such as longline fisheries.
For predation, as sea turtles increase in size, they outgrow
the prey size of some fish predators, but they also grow into
the prey size of the largest predators, such as killer whales
and humans (although the latter source is now very low in
the Azores as a result of legislation and education [senior
author, personal obs.]). Death from ingestion of or entangle-
ment in marine debris would probably not vary substan-
tially over this size range. However, mortality from inciden-
Bjorndal et al.: Estimates of survival for Caretta caretta in the North Atlantic
735
Table 1
Size ranges of age classes and age distributions for tota
of turtles in each age class x, and YOY is young of year.
sample
and tuna
sam
pie. CCL is curved carapace length
N^
is the
number
Size range
(cm CCL) Age
Total
sample
Tuna sample
K
ln(N, 1
iV,
\n{N^)
< 15,0 YOY
9
2.197
—
—
15.0-20.5 1
101
4.615
49
3.892
20.5-26.1 2
287
5.659
118
4.771
26.1-317 3
248
5.513
111
4.710
31.7-36.9 4
248
5.513
120
4.787
36.9-41.7 5
224
5.412
108
4.682
41.7-46.5 6
189
5.242
96
4.564
46.5-49.9 7
127
4.844
58
4.060
49.9-52.3 8
56
4.025
27
3.296
52.3-55.0 9
37
3.611
17
2.833
55.0-58.2 10
33
3.497
16
2.773
58.2-61.6 11
19
2.944
7
1.946
61.6-65.0 12
13
2.565
—
tal capture in longline fisheries in the Azores does increase
with size, with the 2 to 6 year age classes experiencing very
little mortahty (Bolten,in press). Thus, if our estimate ofS
(calculated for the age classes between threshold age and
age 6) were applied to the entire oceanic stage, the effect
of mortality in longline fisheries, or other fisheries biased
to large size classes, would be underestimated.
The estimate of 0.911 for annual survival probabilities
for oceanic-stage loggerhead sea turtles in the waters of
the Azores indicates high survival in this lifestage without
mortality from longline fisheries. Species characterized by
long life and late sexual maturity, such as loggerhead sea
turtles, require very high survival throughout immature
stages to maintain populations (Congdon et al., 1993;
Grouse, 1999). This high probability of survival is also
consistent with the theory that lower predation in oceanic
habitats compared to neritic habitats is the selective pres-
sure that maintains oceanic juvenile stages in most species
of sea turtles (Bolten, 2003).
In two updated matrix models for North Atlantic log-
gerhead sea turtles (that differed in stage lengths). Hep-
pell et al. (in press) derived fitted estimates of 0.745 and
0.875 for annual survival probabilities of the oceanic stage,
which they defined as spanning 5 to 45 cm carapace length.
Chaloupka (in press) derived an estimate of annual sur-
vival probability for oceanic-stage loggerhead sea turtles
in Australia of 0.67 sampled from a logistic probability
density function that ranged from 0.60 to 0.76 and had
a mode at 0.67. The tuned estimate of 0.67 was derived
from a stochastic simulation model that incorporated em-
pirically based survival probability estimates for all age
classes in the model except the oceanic phase (Chaloupka
and Limpus, 2002; Chaloupka, in press). The estimate of
0.67 was generated for a size range from posthatchlings
that have left the waters directly adjacent to the nesting
Table 2
Estimates of instantaneous mortality rates (Z) and annual
apparent survival probabilities {0, estimated as e"^) for
oceanic-stage loggerheads in the waters of the Azores
generated from catch-curve analyses, r^ (coefficient of
determination) and P values are from linear regression
analyses (see Fig. 2).
Age range (years
Z
0
r2
P
Total sample
2 to 12
0.333
0.720
0.935
< 0.001
2 to 6
0.094
0.911
0.923
= 0.009
6 to 12
0.441
0.643
0.974
< 0.001
Tuna sample
4 to 11
0.421
0.656
0.954
< 0.001
4 to 6
0.112
0.894
0.999
= 0.021
6 to 11
0.498
0.608
0.966
< 0.001
beach to subadults that begin to leave the oceanic habitats
at a size of 69 cm curved carapace length (Chaloupka and
Limpus, 2002).
The fitted estimates for annual survival from the Heppell
et al. (in press) models and the Chaloupka (in press) model
are lower than the estimates in our study, but the size ranges
are different. In the Heppell et al. (in press) models and the
Chaloupka (in press) model, the oceanic stage includes the
posthatchling phase during which loggerhead sea turtles
migrate from nesting beaches to their oceanic habitats. We
could not include this early posthatchling phase in our esti-
mates of survival of oceanic-stage loggerhead sea turtles in
the waters of the Azores because many turtles in this phase
736
Fishery Bulletin 101(4)
have not reached the Azores and they are younger than
our threshold ages. We beheve that mortality in this early
transitional stage when loggerhead sea turtles first cross
the Atlantic may be high. In addition to high rates of pre-
dation. winds and currents can overwhelm the swimming
and orientation abilities of the posthatchling sea turtles,
transporting the turtles to habitats, such as waters off the
British Isles, that cannot sustain them (Carr, 1986; Hays
and Marsh, 1997). Generating directly derived estimates
of survival probabilities of loggerhead sea turtles younger
than 2 years of age should be a high priority.
Acknowledgments
This study would not have been possible without the sup-
port of our colleagues in the Azores: "Equipa Tartaruga"
at the Department of Oceanography and Fisheries (DOP),
University of the Azores; the captains and crews of the
commercial tuna fleet based in Horta and Pico; and J. and
G. Franck of the MY Shanghai. We thank M. Chaloupka
for encouragement to pursue catch-curve analysis. We
thank M. Chaloupka and J. Seminoff for comments on
earlier drafts of the manuscript and P. Eliazar for technical
assistance. This project was funded by the U.S. National
Marine Fisheries Service and the Disney Wildlife Con-
servation Fund. All work was conducted in compliance
with the Institutional Animal Care and Use Committee,
University of Florida.
Literature cited
Bjorndal, K. A., A. B. Bolten, T. Dellinger, C. Delgado, and
H. R. Martins.
2003. Compensatory growth in oceanic loggerhead sea
turtles: response to a stochastic environment. Ecology
84:1237-1249.
Bjorndal, K. A., A. B. Bolten, and H. R. Martins.
2000. Somatic growth model of juvenile loggerhead sea
turtles Caretta caretta: duration of pelagic stage. Mar.
Ecol. Prog. Ser 202:265-272.
Bolten, A. B.
1999. Techniques for measuring sea turtles. In Research
and management techniques for the conservation of sea
turtles (K. L. Eckert, K. A. Bjorndal, F. A. Abreu-Grobois,
and M. Donnelly, eds.), p. 110-114. lUCN/SSC Marine
Turtle Specialist Group Publication 4, Washington, DC.
2003. Variation in sea turtle life history patterns: neritic
versus oceanic developmental stages. In Biolog>' of sea
turtles, vol. 2 (P. L. Lutz, J. A. Musick, and J. Wyneken,
eds.), p. 243-257. CRC Press, Boca Raton, PL.
In press. Active swimmers — passive drifters: the oceanic
juvenile stage of loggerheads in the Atlantic system. In
Loggerhead sea turtles (A. B. Bolten and B. E. Witherington,
eds.). Smithsonian Institution Press, Washington, DC.
Bolten, A. B., K. A. Bjorndal, H. R. Martins, T Dellinger,
M. J. Biscoito, S. E Encalada. and B. W. Bowen.
1998. Transatlantic developmental migrations of logger-
head sea turtles demonstrated by mtDNA sequence analysis.
Ecol. Appl. 8:1-7.
Bolten, A. B., H. R. Martins, K. A. Bjorndal, and J. Gordon.
1993. Size distribution of pelagic-stage loggerhead sea
turtles (.Caretta caretta) in the waters around the Azores
and Madeira. Arquipelago llA:49-54.
Carr, A.
1986. Rips, FADS, and little loggerheads. BioScience 36:
92-100.
Chaloupka, M.
In press. Simulation modeling of population viability for log-
gerhead sea turtles exposed to competing mortality risks
in the western south Pacific region. In Loggerhead sea
turtles (A. B. Bolten and B. E Witherington, eds.). Smith-
sonian Institution Press, Washington, DC.
Chaloupka, M. Y., and C. J. Limpus.
2002. Survival probability estimates for the endangered log-
gerhead sea turtle resident in southern Great Barrier Reef
waters. Mar Biol. 140:267-277.
Congdon, J. D., A. E. Dunham, and R. C. van Loben Sels.
1993. Delayed sexual maturity and demographics of
Blanding's turtle (Emydoidea blandingii): implications
for conservation and management of long-lived organisms.
Conserv. Biol. 7:826-833.
Grouse, D. T
1999. The consequences of delayed maturity in a human-
dominated world. Am. Fish. Soc. Symp. 23:195-202.
Epperly S. P., M. L. Snover, J. Braun-McNeill, W. N. Witzell,
C. A. Brown, L. A. Csuzdi, W. G. Teas, L. B. Crowder, and
R. A. Myers.
2001. Stock assessment of loggerhead sea turtles of the
western North Atlantic. In Stock assessments of logger-
head and leatherback sea turtles and an assessment of the
impact of the pelagic longline fishery on the loggerhead and
leatherback sea turtles of the western North Atlantic, p.
3-66. NCAA Tech. Memo. NMFS-SEFSC-455.
Frazer, N. B.
1987. Preliminary estimates of survivorship for wild juve-
nile loggerhead sea turtles (Caretta caretta). J. Herpetol.
21:232-235.
Hays, G. C, and R. Marsh.
1997. Estimating the age of juvenile loggerhead sea turtles
in the North Atlantic. Can. J. Zool. 75:40^6.
Heppell, S. S., L. B. Crowder, D. T Grouse, S. P Epperly and
N. B. Frazer
In press. Population models for Atlantic loggerheads: past,
present and future. In Loggerhead sea turtles (A. B.
Bolten and B. E. Witherington, eds.). Smithsonian Insti-
tution Press, Washington, DC.
Isaac, V. J.
1990. The accuracy of some length-based methods for fish
population studies, 81 p. International Center for Living
Aquatic Resources Management, Manila, Philippines.
Seber, G. A. F
1982. The estimation of animal abundance and related
parameters, 654 p. Macmillan Publishing Co., New York,
NY.
Turtle Expert Working Group.
2000. Assessment update for the Kemp's ridley and logger-
head sea turtle populations in the western North Atlantic.
U.S. Dep. Commer, NOAATech. Memo. NMFS-SEFSC-444,
115 p.
737
Abstract— The green sea urchin iStron-
gyloccntrotus droebachiensis) is impor-
tant to the economy of Maine. It is the
state's fourth largest fishery by value.
The fishery has experienced a con-
tinuous decline in landings since 1992
because of decreasing stock abundance.
Because determining the age of sea
urchins is often difficult, a formal stock
assessment demands the development
of a size-structured population dynamic
model. One of the most important com-
ponents in a size-structured model is a
growth-transition matrix. We developed
an approach for estimating the growth-
transition matrix using von Bertalanffy
growth parameters estimated in previ-
ous studies of the green sea urchin off
Maine. This approach explicitly consid-
ers size-specific variations associated
with yearly growth increments for
these urchins. The proposed growth-
transition matrix can be updated read-
ily with new information on growth,
which is important because changes in
stock abundance and the ecosystem will
likely result in changes in sea urchin
key life history parameters including
growth. This growth-transition matrix
can be readily incorporated into the
size-structured stock assessment model
that has been developed for assessing
the green sea urchin stock off Maine.
Developing a growth-transition matrix for
the stock assessment of the green sea urchin
(Strongylocentrotus droebachiensis) off Maine
Yong Chen
School of Marine Sciences
218 Libby Hall
University of Maine
Orono, Maine 04469
E-mail address ychen@maine.edu
Margaret Hunter
Maine Department of Marine Resources
P.O. Box 8
West Boothbay Harbor, Maine 04575
Robert Vadas
Department of Biological Sciences
University of Maine
Orono, Maine 04469
Brian Beal
Division of Environmental and Biological Sciences
University of Maine
Machias, Maine 04654
Manuscript approved for publication
17 April 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:737-744 (2003).
The green sea urchin (Strongylocentro-
tus droebachiensis) fishery is the state's
fourth largest fishery by value, worth
$20.3 million to harvesters in 1999.
The fishery is managed by limited
entry, a limited number of opportunity
dates that are established each year
by recommendation of the sea urchin
zone council (SUZC), and minimum
and maximum size limits. The fishery
is further regulated seasonally by two
zones that correspond to variation in
spawning time along the coast (Vadas
etal., 1997).
The Maine sea urchin fishery began
in the late 1980s and reached its peak
in landings in 1992. It has since experi-
enced a continuous decline in landings,
mainly resulting from large decreases
in sea urchin stock abundance (Fig. 1).
Although the large decrease in abun-
dance is evident in many studies (Ste-
neck and Vadas'; Harris^) and apparent
to the sea urchin fishing industry, the
catch-per-unit-of-effort (CPUE) data
derived from the fishery have shown
no significant decreases over the last
10 years (Fig. 1). We need to perform
a formal stock assessment to better
understand the population dynamics of
the sea urchin stock and to develop an
optimal management strategy.
A population dynamics model for the
sea urchin stock should provide reliable
estimates of model parameters with
suitable statistical methods (Hilborn
and Walters, 1992; Chen and Paloheimo,
1998; Walters, 1998). A size-structured
population dynamics model is needed
for the sea urchin fishery because sea
urchins are difficult to age and growth
varies widely among individuals (Quinn
and Deriso, 1999).
One of the key components of a size-
structured population dynamics model
is a growth-transition matrix, which
describes the probability of an organ-
1 Steneck, R., and R. L. Vadas. 2002. Per-
sonal commun. School of Marine Sciences,
University of Maine, Orono, ME 04469.
2 Harris, L. 2002. Personal commun.
Department of Zoology, University of New
Hampshire, Durham, NH 03824.
738
Fishery Bulletin 101(4)
ism growing from one size class to another size class in a
given unit of time (Sullivan et al., 1990; Sullivan, 1992).
In practice, two approaches can be used to incorporate a
growth-transition matrix into a stock assessment: one is
to incorporate the growth-transition matrix and simul-
taneously estimate matrix parameters with parameters
that describe other biological processes in the fishery
(Sullivan et al., 1990), and the other approach is to esti-
mate the growth-transition matrix independent of other
stock assessment models (Chen et al., 2000). The former
considers covariance among different processes by esti-
mating all parameters simultaneously, but makes the
analysis more complicated. The latter approach reduces
the complexity of modeling but does not consider the
covariance of growth and other biological processes. Be-
cause size-structured models are often complicated and
have many parameters to be estimated, the estimation
of a growth-transition matrix outside the main modeling
process may be preferable (Chen et al., 2000). In either
case, the quality of the growth-transition matrix can
greatly influence the quality of the stock assessment. It
is thus essential to develop a growth-transition matrix
for the Maine sea urchin stock that can capture the
variations in growth increments among individuals.
The information required in estimating a growth-
transition matrix includes the mean growth increment
in a given unit of time and its associated variation for
sea urchins of different sizes. Because growth rates of
sea urchins vary with size, growth increments also vary
with size, and this variation in growth with size is rarely
constant. This has been implicit in the statements of
model assumptions in many papers (e.g. Sullivan et al.,
1990; Sullivan, 1992, Quinn and Deriso, 1999). However,
because the variance for growth increments is difficult
to estimate, it is often assumed to be constant for organ-
isms of different sizes (Quinn and Deriso, 1999). Such
an assumption of constant variation in growth incre-
ment is rather unrealistic and may introduce biases in
estimating a growth-transition matrix. Thus, for the
Maine sea urchin we need to develop an approach that
can explicitly consider nonconstant variances for gi"owth
increments of sea urchins of different sizes.
Growth of the sea urchin along the Maine coast has
not been studied extensively and the data are limited.
The data we used for this study were from Vadas et al.
(2002) who collected size-at-age data on sea urchins in
two habitats (barren and kelp) from three areas along
the coast of Maine.
Methods and materials
Previous studies have indicated that many environmental
variables might influence the growth of the sea urchin
(Meidel and Scheibling, 1998; Russell, 1998). Sea urchins
in favorable habitats, feeding on preferred seaweeds, grow
faster than those feeding on less favorable algae and
mussels, and sea urchins on barren grounds grow slower
Even in the same habitat, different rates of growth were
identified (Vadas, 1977). Previous studies divided the
LU
Q.
o
1985
1990
1995
2000
2005
80
60
40
20
0 I : . ^ - - - ' 1
1993 1994 1995 1996 1997 1998 1999 2000 2001
180
160
140
120
1993 1994 1995 1996
1997
Year
1998 1999 2000 2001
Figure 1
Observed catch measured in metric tons, effort measured in
diver-hours, and catch per unit of effort measured in pounds
per diver-hour for the sea urchin fishery in management zone 2
in Maine. Zone 1 has a similar temporal pattern.
coast of Maine into three regions, northeast, center, and
southwest (Vadas et al., 1997). For each region, sea urchin
samples were randomly taken from two habitats, barren
and kelp. Size-at-age data were collected in 1997-98 for
sea urchins in each habitat and area (Vadas and BeaP).
Detailed descriptions about the derivation of size and age
'' Vadas, R. L., and B. F. Beal. 1999. Temporal and special vari-
ability in the relationships between adult size, maturity and
fecundity in green sea urchins: the potential use of a roe-yield
standard as a conservation tool. Report to the Maine Depart-
ment of Marine Resources, Augusta, Maine 04333.
Chen et al : Developing a growth-transition matrix for stock assessments of Strongylocentrotus droebachiensis
739
information, justification for dividing the Maine coast, and
selection of the habitats can be found in Vadas et al. ( 1997)
and Vadas et al. (2002).
Vadas et al. (1997) modeled the size-at-age data using
the von Bertalanffy growth function (VBGF) described as
L, = L„{\-e
-All-I,,)
(1)
where L, = size at age t;
L^ = defined as the mean asymptotic length that
the sea urchin may attain;
K = the Brody growth parameter; and
tg = the hypothetical age of size 0 (Ricker, 1975).
For each area and habitat, a VBGF was used to fit the size-
at-age data. Three parameters in the VBGF (i.e. L„, K, and
tf,) and their standard errors were estimated by using the
nonlinear least squares method. These estimates were pre-
sented in Vadas and BeaP and Vadas et al. (2002), and were
made available to the authors of the present study (Table
1). Clearly there were large differences in the estimates of
L^ and K and their associated variations among different
areas and habitats (Table 1).
The LJs estimated for different areas and habitats
ranged from 63.1 (northeast region with barren habitat)
to 95.2 mm (southeast region with kelp habitat) (Table 1)
and tended to be smaller than some large individuals ob-
served in the fishery (about 100 mm, Vadas, 1977; Hunter,
unpubl. data). This likely resulted from relatively small
sample sizes that covered relatively small areas, in a
relatively short period, compared with the fishery catch,
which targeted larger-size individuals. The exclusion of
individuals in the fishery catch that were larger than the
L.,'s estimated in Vadas and Beal'^ and Vadas et al. (2002)
from the calculation of the growth-transition matrix may
underestimate the variability in sea urchin growth, thus
introducing errors in stock assessment. Based on the data
collected in the Maine sea urchin fishery (Hunter, unpubl.
data) and previous studies (Vadas, 1977), 100 mm was
considered a reasonable value for the average asymptotic
size (LJ for sea urchins on the coast of Maine. However,
more extensive sampling needs to be done in the future to
verify this estimate.
We might be able to derive an estimate of L^ for the
Maine sea urchin stock based on the examination of the
data collected from the fishery and other studies (Ricker,
1975; Moreau, 1987; Chen et al., 1992). An estimate of K
for the whole Maine urchin stock is, however, more difficult
because K is an abstract rate describing how fast organ-
isms approach the L_.^ and there are no observations or
background information with which to compare estimates
(Ricker, 1975; Moreau, 1987). We thus need to develop an
approach to estimate K for the Maine sea urchin stock
which corresponds to the value we assumed for the L^.
Many studies have indicated that estimates of K and L„
tend to be highly and negatively correlated (e.g. Moreau,
1987; Chen and Harvey, 1994). Thus, a fish population or
species with a large L^^ tends to have a low ii" value, and vice
versa (Gallucci and Quinn, 1979; Chen et al., 1992). This
suggests a strong relationship between L^ and /f estimates
Table 1
The average asymptotic
size (L_) and Broady growth coef-
ficient (K) estimated for different areas and habitats along |
the coast of Maine in the study
done by
'/adas et al
.(1997,
2002). Coefficient of variation (CV) was calculated by using |
Equation 2.
Coefficient of
Area
Habitat
Parameter
variation (CV)
^„
A'
CV(L„)
CV(AD
Northeast
Barren
63.1
0.1404
0.242
1.209
Northeast
Kelp
88.. 5
0.1263
0.224
0.543
Center
Barren
67.0
0.2315
0.084
0.354
Center
Kelp
6.3.4
0.3268
0.065
0.248
Southeast
Barren
80.1
0.1776
0.099
0.397
Southeast
Kelp
95.2
0.1181
0.128
0.338
(Pauly, 1980; Stergiou, 1993). Such a relationship may be
used to estimate K for a given L„ or to estimate L„ for a
given K. In this study we developed and used the follow-
ing empirical approach to derive K for a given value of L^
and its associated uncertainties in the development of a
growth-transition matrix: 1) conduct a regression analysis
for K and L„ estimated for different areas and habitats
along the coast of Maine (Table 1); 2) calculate coefficients
of variation (CV) for each K and L„ (Table 1) as
„ standard error for K
CV(K) = and
CV(L,) =
K
standard error for L^
(2)
and conduct a regression analysis of CV(/0 and CV(L„)
estimates of different areas and habitats (data in Table 1);
3) use 100 mm to approximate L^„ and use this L„ to esti-
mate K from the regression analysis between K and Lj,
and 4) calculate the average CV for LJs of different areas
and habitats and then use the average CV(L„) to estimate
CW(K) from the CW{K)-CV(LJ regression equation.
Because K and L„ were estimated for different areas
and habitats and had different precisions, outliers might
arise in the regression analyses. To avoid possible bias in-
troduced by outliers, we used a reweighted least squares
(RLS) method for the regression analyses (Chen et al.,
1994). This method involves conducting a robust least
median of squares (LMS) analysis to identify outliers
(Rousseeuw and Leroy, 1987) and justifying the identified
outliers by using background information, followed by a
weighted LS analysis where justified outliers are weighted
by 0 and other data have a weight of 1 (Chen et al., 1994).
In the two regression analyses (i.e. steps 1 and 2), L^ and
CViLJ were used as the independent variables and K and
CW(K) were used as the dependent variables. The reason
for this choice (instead of the other way around) is thatL„
is often estimated more reliably and with much smaller
740
Fishery Bulletin 101(4)
variations (Chen at al., 1992; also see Table 1), whereas
K is often less reliably estimated (Moreau, 1987). One of
the basic assumptions for a regression analysis is that the
independent variable is error free. In practice, this assump-
tion is often relaxed when the independent variable has a
much smaller error than the dependent variable (McArdle,
1988). The violation of the normal distribution assump-
tion for the errors in the regression analyses may bias the
test for the significance of the regression model and its
parameters using common parametric tests {F- or /-tests),
but does not necessarily result in biases in the regression
analysis (Sen and Srivastava, 1990)."
Given K and L_, the growth increment during a unit of
time (i.e. year) can be calculated as
Ai. =(L„-L„)(l-e"').
(3)
where K and L^ are the true values without errors; n
indexes size class; and L,, is the middle point of the n"' size
class. With Equation 3, we can develop two approaches to
estimate the growth-transition matrix. One approach is a
Monte Carlo simulation. We can randomly sample H sets
ofK and L^ values from their joint distributions ( thus con-
sider their covariance) and then use them in Equation 3
to calculate H sets of AL for each size group. We can then
derive the probability distribution for AL from these H sets
of AL values for each size group. The Monte Carlo simula-
tion approach is straightforward but requires extensive
calculations, in particular when there are a large number
of size groups. It is also inconvenient to update the growth-
transition matrix when there are new growth data or large
changes in growth due to changes in the environment. The
second approach is analytic and not so straightforward,
but it is easy to update with new information and is less
computationally intensive. It is likely that the growth-
transition matrix for the Maine sea urchin fishery will
need to be updated because of possible changes in growth
caused by changes in the sea urchin population size and
its ecosystem. Thus we used the second approach, which is
described as follows.
Assuming the uncertainties associated with the VBGF
parameters L^ and K are AL„ and AK respectively, where,
AZ._ e N{0.ai_ ) and AK e N{0.al ), we have
L = L+AL ami K = K + AK.
(4)
where Z,„ and K^ are the estimated parameters. Replacing
the true values of L„ and K in Equation 3 with Equation
4 and using the approximation e-^ = 1 + AX for small AX,
we have
AL„=(L„-L„)(\ -€-'') +
[aL_(1 -e *)-(!„- Z.JAA:^"*' -AL„AA:e''l = aZ„ -I- f„,
(5)
where
AL„=(L„-L„){\-e-'') (6)
e„=ALJ\-e-^)-{L^-L„)AKe-^ -AL„AKe~^- (7)
Thus, the expected (mean) value of AL,, is AL^ and vari-
ance of AL^ can be estimated from Equation 7 as
Var(AZ,„) = CT^d -e '^ ) +(L„ - L„)-(7j.e"
2Cov( L„. Ar)( 1 - e"'' )(ZL - i,„ )?'*".
(8)
Items with the order of three and above for AL„ and AK
are omitted in deriving Equation 8 from Equation 7. From
Equation 8, it is clear that the variance of the growth incre-
ment varies among different size classes.
From AL,, estimated in Equation 6, an expected average
yearly growth increment was calculated for each size class.
The variability for the average yearly growth increment
was assumed to follow a normal distribution with a mean
of AL„and variance ofVar (AL,,) estimated from Equation 8.
This distribution was used to determine the vector of prob-
abilities of growing from size class k to other size classes.
If d,^„, and d„ are the lower and upper ends of size class d,
the probability of a sea urchin growing from size class n to
size class d can be computed as
P..^., = I /(
.v|AL„,Var(AL„)(/.v,
(9)
where x is a random variable having a density probability
distribution defined by /(.vj AL„,Var(AL„ )) with its expected
value of AL,;, and variance of Var(AL,^) (Quinn and Deriso,
1999). In the present study we assumed that the .v variable
was a normal density distribution function with a mean of
AL^, defined by Equation 6 and with a variance of Var ( AL,^ )
defined by Equation 8. The probability of a sea urchin grow-
ing from one size to another was estimated for all size classes
to form the matrix. Negative growth increments were not
permitted. The largest size class acts as a plus group; there-
fore sea urchins in this group have a probability of 1 of
remaining in the group. The model contains 61 size classes,
each with 1-mm interval width, ranging from 40 mm in
size (midpoint value for size class from 39.5-40.5 mm)
to 100 mm.
Because no negative growth was allowed, the summation
of the probabilities of a sea urchin of size class k growing
into all other size classes was smaller than 1 (because the
normal distribution is symmetric). This problem was avoided
by standardization which involved dividing the probability
of an urchin in a given size class n growing into each size
class by the summation of the probabilities of growing from a
given size n to all the size classes. All calculations were done
in MS-Excel©( Microsoft Office 2000, Microsoft Corporation,
Redmond, WA). A worksheet for estimating a growth-transi-
tion matrix as described above is available upon request.
Results
The LMS analysis suggested that the logarithmic K and
L^ data for the barren habitat in the Southwest area was
an outlier in the K and L,, regression analysis (Fig. 2). The
estimated K and L^ values for the barren habitat in the
Southwest had CVs over 120% and 24%, respectively, much
Chen et a\: Developing a growth-transition matrix for stock assessments of Strongylocentrotus droebachiensis
741
Ln(/..
0.0 -
4
-0.5
-1.0
-1.5
-2.0
-2.5 J
4.2
4.3
4.4
4.5
4.6
♦ Observed
■ Outlier
— Predicted
Figure 2
The regression analysis of logarithmic K and /-„ for different
locations and habitats of Maine sea urchins.
1,4
1.2
♦ Observed
-m- Outlier
■
1,0
be 0,8 -
— Predicted
£ 0,6 -
S 0.4 -
^^^^ ^
0.2 -
♦
0.0 -
1 1 1
0
0 0.1 0.2 0.3
CV for /...,
Figure 3
The regi
location.
■ession analysis of CVs for K and L^ for different
5 and habitats of Maine sea urchins.
larger than the estimates for other locations and habitats
(Table 1). This was the only site where the K estimate was
not significantly different from 0 (thus the VBGF was not
significant). We thus concluded that this data point was
an outlier because of the poor fit of the VBGF, and subse-
quently it was given a zero weight in the RLS analysis. The
RLS regression equation for K and L^ was estimated by
LniK) = 8.653 - 2.3777 LniLJ,
P=0"0038, adj. r2=0.94.
(10)
The standard deviations for the intercept and slope were
1.2605 and 0.28923, respectively. The P value for Equation
10 indicates that the regression model is significant. The
adj. r^ is the coefficient of determination adjusted for the
sample size, suggesting 94% of the variance in \n{K) could
be explain by the model.
The LMS analysis of the CVs of parameters K and L_
also suggested that the barren habitat in the southwest
area was an outlier because it had an exceptionally large
CV for K (Fig. 3). We thus concluded that this data point
was an outlier and should be given a weight of zero in the
RLS analysis. The RLS regression equation for the CVs of
parameters K and L„ was estimated by
F
7
h
6
r
Q>
F
5
ID
4
C
sz
3
^
o
2
CJ)
Ti
(IJ
1
CJ
n
II
UJ
50 100
Midpoint of size class (mm)
150
Figure 4
The expected annual growth increment for Maine sea
urchins of different size classes.
CVm = 0.189 + 1.5602 CV (LJ,
P=0.034, adj.
r2 = 0.76. (11)
The standard deviations for the intercept and slope were
0.0561 and 0.42319, respectively. The P value suggested the
regression model was significant (P<0.05). The value of r^
suggests 76% of the variance in CV(K) could be explained
by the model.
The average CV for LJs of different areas and habitats
was 15%. The L^ was assumed to have a value of 100 mm
in this study as discussed previously. This gave the L^ a
standard error estimate of 15.0 mm, making its 95% con-
fidence intervals 70 mm to 130 mm. The iiT value was esti-
mated to be 0.1006 using Equation 10 and L^ of 100 mm.
Using Equation 11 and the CV for L,^ the CV for K was
estimated to be 42.3%, which yielded the value of 0.0426
for the standard error for K.
The annual expected growth increment decreased quick-
ly with sea urchin size (Fig. 4). The largest expected annual
increment was 6 mm for the smallest size class (39.5-40.5
mm) included in the study. The variance for annual growth
increments calculated by using Equation 8 was large for
small sea urchins. It decreased initially with size, reaching
the smallest value at the 59 mm size class (58.5-59.5 mm),
followed by a progressive increase with size (Fig. 5). The
expected annual growth increment for the largest size class
included in this study had the highest variance, which was
over eight times as high as the smallest variance (Fig. 5).
The probability distribution of annual growth increment
varied among size classes (Fig. 6), reflecting the differences
in variances associated with different size classes. The last
size class was a plus class, with the probability of staying
m the same size class being 1. Figure 6 clearly indicated
that no negative growth was allowed.
Discussion
Great variation in growth was observed in the Maine sea
urchin stock (Vadas et al., 2002). Such a pattern of variation
was reflected in estimating the VBGF parameters for dif-
742
Fishery Bulletin 101(4)
60 80
Midpoint of size class (mm)
Figure 5
The variances of growth increment estimated for dif-
ferent sea urchin size classes by using Equation 8.
ferent areas and habitats (Table 1). Large standard errors
were estimated for the VBGF parameters for sea urchins of
the same area and habitat, and large differences occurred
in the estimated VBGF parameters between different areas
and habitats (Table 1). The approach developed in the pres-
ent study considered observations made in both the fishery
and scientific studies and provided a systematic way to
incorporate the large variation in growth into the estima-
tion of a growth-transition matrix, and subsequently into
the sea urchin stock assessment.
It should be noted that the algorithm developed for esti-
matmg the variance of growth increments is approximate,
and violations of the assumptions used in deriving the
algorithm may introduce errors in estimating a growth-
transition matrix. For example, large errors in estimat-
ing K and L„ will introduce errors in Equation 5, which
was derived by assuming small errors for the two growth
parameters. Nonnormal distribution of AL with its mean
defined by Equation 6 and variance defined by Equation 8
will also result in errors in developing a growth-transition
matrix. Other factors that may influence the quality of the
estimated growth transition matrix include errors in esti-
mating CVs for K, L^ estimated from Equations 10 and 11,
and omitting high order items in deriving Equation 8.
Unlike most studies in which the variance for the annual
growth increment was assumed to be the same for all size
classes (Quinn and Deriso, 1999), our study explicitly sug-
gested that the variance for the annual growth increment
changed with size (Fig. 4). The differences in the variance
were large between size classes, and changed nonlinearly
with size. If a constant variance were used for all size
classes, the variance in growth increment would be se-
verely underestimated for large and small fish. This could
introduce large biases in a stock assessment.
Size-dependent variation might better describe the
variation in annual growth increment. Fish in small size
classes tend to grow fast, but their growth tends to be more
susceptible to environmental variation than adult growth,
often resulting in large variation among individuals (Sum-
merfelt and Hall, 1987). Fish in large size classes (older
fish) have to divert some energy to reproduction but tend
to have considerable variation in energy allocation strate-
gies among individuals. Differences among adults in the
ability to grow can also be considerable because of genetics,
specific growth patterns during juvenile stages, and differ-
ences in energy allocation between growth and matura-
tion during younger ages (Nikolskii, 1969). This difference
may cause large variations in growth for large and old fish
(Summerfelt and Hall, 1987; Chen et al., 1988). Compared
with old and young ages, growth rates for medium-size and
medium-age fish may be less varied (Nikolskii, 1969). This
pattern can be reflected realistically in the estimated varia-
tion by using the approach derived in our study.
Although the choice of L^ was a bit arbitrary in our study,
it reflects observations from both the fishery and scientific
studies. The largest sea urchins observed in the different
scientific studies tend to be smaller than 100 mm, as in-
dicated by the estimated L^ values for different areas and
habitats (Fig. 1 ). The inability to observe larger sea urchins
in scientific studies may result from relatively small sam-
ple sizes, the focus of research (small areas), and the large
growth variations even in small spatial scales. The data
collected from the fishery were more extensive and covered
more areas. This, together with the tendency for taking
large individuals in the fishery, may suggest that large
individuals are more likely to appear in the fishery, rather
than in scientific studies. Thus, it may be reasonable to set
the expected value of L^ at 100 mm. Also, this higher value
corresponds more closely to the upper growth estimates for
green sea urchins from the northeast Pacific ( Vadas, 1977).
The CV was assumed to be 15% for L_, resulting in the 95%
confidence interval of L^ ranging from 70 mm to 130 mm.
This range was believed to be a reasonable estimate for the
maximum attainable length for green sea urchins on the
coast of Maine (Vadas, 1977).
The approach developed in our study can be readily used
to incorporate the VBGF parameters estimated from dif-
ferent studies. This can be accomplished by rerunning the
regression analyses between K and L^ and between CVs
for K and L^. As more information about the growth of
sea urchins on the coast of Maine becomes available, the
growth transition matrix can be easily updated to reflect
the variation identified in newer studies. The flexibility and
ability to easily update and incorporate new information
makes this approach desirable to the Maine sea urchin
fishery, which is currently undergoing large changes in its
population size and has only limited growth data.
The value of 100 mm chosen for L^ was rather arbitrary.
However, because we considered the negative correlation
between K and L_ in deriving the growth transition ma-
trix, a small error in the L^ estimate would not change the
growth-transition matrix greatly. In the future, however,
we can conduct a systematic sampling of the stock across
its geographical range and derive some forms of weighted
average size with a composite variance that captures the
range of sizes exhibited by the species. Such an approach
would provide us with a better estimate of L„.
The growth-transition matrix developed in our study
summarizes the growth patterns of sea urchins along the
coast of Maine. It can be updated whenever new growth
data become available. It can be readily incorporated into
Chen et al : Developing a growth-transition matrix for stock assessments of Strongy/ocentrotus droebachiensis 743
55
o
06
0.5
0.4
03
0.2
0.1
0
72
73
74
/ \ M ■' \ 1 \ \' ''
75
- - 76
77
/ ; 1 U \ \ \ '
78
,^XO'.K^VV■■.
- - - -79
-
78
83
88
0.5
0.2
63
80
/ \ '^
81
82
83
84
85
/ ' A' \-' ■•/'■' -.'l \ '•
86
1 i\iynA\\
87
.'.''// /)(y-/.V'A^v:^.
79
1 n
0.9
0.8 1
96
07
0.6
0.5
97
98
99
100
0.4 J
0.3 -
0.2
0.1 -
n -
99 92 94
Midpoint of size class (mm)
Figure 6
Probabilities of sea urchins growing from one size class to others. Each probability distribution was labeled with
the midpoint value of the current size class of the sea urchin.
744
Fishery Bulletin 101(4)
a size-structured stock assessment model to evaluate the
status of sea urchin stock and to evaluate alternative man-
agement strategies for the Maine sea urchin fishery (Chen
and Hunter, 2003).
Acknowledgments
We would like to thank the Maine Department of Marine
Resources, the Northeast Consortium, and the Maine Sea
Urchin Zone Council for supporting this study. Construc-
tive and detailed comments from two anonymous reviewers
and the scientific editor greatly improved an early version
of the manuscript, for which we are grateful.
Literature cited
Chen, S., S. Watanabe, and K. Takagi.
1988. Growth analysis on fish population in the senescence
with special reference to an estimation of age at end of
reproductive span and life span. Bull. Jpn, Soc. Sci. Fish.
54:1567-1572.
Chen, Y., P. Breen, and N. Andrew.
2000. Impacts of outliers and mis-specification of priors on
Bayesian fisheries stock assessment. Can. J. Fish. Aquat.
Sci. 57:2293-2305.
Chen, Y., and H. H. Harvey
1994. Maturation of white sucker, Catostomus commer-
soni, populations in Ontario. Can, J. Fish. Aquat. Sci. 51:
2066-2076.
Chen, Y., and M. Hunter
2003. Assessing the green sea urchin (Strongylocentrotus
drobachiensis) stock m Maine, USA. Fish. Res. (Amst.)
60:527-537
Chen, Y, D. A. Jackson, and H. H. Harvey.
1992. A comparison for von Bertalanffy and polynomial
functions in modeling fish growth data. Can. J. Fish.
Aquat. Sci. 49:1228-1235.
Chen, Y, D. A. Jackson, and J. E. Paloheimo.
1994. Robust regression approach to analyzing fisheries
data. Can. J. Fish. Aquat. Sci. 51:1420-1429.
Chen, Y., and J. E. Paloheimo.
1998. Can a more realistic model error structure improve
parameter estimation in modelling the dynamics of fish
populations? Fish. Res. (Amst.) 38: 9-19.
Gallucci, V. F, and T. J. Quinn II.
1979. Reparameterizing, fitting, and testing a simple growth
model. Trans. Am. Fish. Soc. 108:14-25.
Hilborn, R., and C. Walters.
1992. Quantitative fisheries stock assessment: choice, dy-
namics, and uncertainty, 570 p. Chapman and Hall, New
York, NY.
McArdle, B. H.
1988. The structural relationship: regression in biology.
Can. J. Zool. 66:2329-2339.
Meidel, S. K., and R. E. Scheibling.
1998. Size and age structure of the sea urchin Strongylocen-
trotus droebachiensis in different habitats. In Echinoderms
(R. Mooi, M. Telford, eds.), p 737-742. Proceedings of the
9"^ international echinoderm conference; San Francisco, 5-9
August 1996. A. A. Balkema, Rotterdam, Netherlands.
Moreau, J.
1987. Mathematical and biological expression of growth in
fishes: recent trends and further developments. In The age
and growth offish (R. C. Summerfelt and 0. E. Hall (eds.), p
81-113. Iowa State Univ. Press, Ames, lA.
Nikolskii, G. V.
1969. Theory of fish population dynamics, 323 p. Oliver &
Boyd, Edinburgh, UK.
Pauly, D.
1980. On the interrelationships between natural mortality,
growth parameters and mean environmental temperature
in 175 fish stocks. J. Cons. Int. Explor Mer 39:175-192.
Quinn, T. J., II, and R. B. Deriso.
1999. Quantitative fish dynamics, 542 p. Oxford Univ. Press,
New York, NY.
Ricker,W. E.
1975. Computation and interpretation of biological statistics
offish populations, 382 p. Bull. Fish. Res. Board Can., vol.
191.
Rousseeuw, P. J., and A. M. Leroy
1987. Robust regression and outlier detection, 352 p. John
Wiley & Sons, New York, NY.
Russell, M. P
1998. Resource allocation plasticity in sea urchin: rapid
diet induced, phenotypic changes in the green sea urchin,
Strongylocentrotus droebachiensis (MiiUer). J. Exp. Mar
Biol. Ecol. 220:1-14.
Sen, A. K., and M. 8. Srivastava.
1990. Regression analysis: theory, methods and applications,
350 p. Springer- Verlag, New York, NY.
Stergiou, K. I.
1993. Nutrient-dependent variation in growth and longev-
ity of the red bandfish, Cepola macrophthalma (L.), in the
Aegean Sea. J. Fish Biol. 42:633-644.
Sullivan, P J.
1992. A Kalman filter approach to catch-at-length analysis.
Biometrics 48:237-257.
Sullivan, P J., H. L. Lai, and V. R Gallucci.
1990. A catch-at-length analysis that incorporates a stochas-
tic model of growth. Can. J. Fish. Aquat. Sci. 47:184-198.
Summerfelt, R. C, and G. E. Hall.
1987. The age and growth of fish, 530 p. Iowa State Univ.
Press, Ames, lA.
Vadas, R. L.
1977. Preferential feeding: an optimization strategy in sea
urchins. Ecol. Monogr. 47:337-371.
Vadas, R. L., B. Heal, S. Dudgeon, and W. Wright.
1997. Reproductive biology of green sea urchins along the
coast of Maine: final report, 59 p. Maine Cooperative
Extension Service and Maine Sea Grant Program, Orono,
ME.
Vadas, R. L., B. Smith, B. Beal, and T. Dowling.
2002. Sympatric growth morphs and size bimodality in
the green sea urchin (Strongylocentrotus droebachiensis).
Ecol. Monogr. 72:113-132.
Walters, C. J.
1998. Evaluation of quota management policies for develop-
ing fisheries. Can. J. Fish. Aquat. Sci. 55:2691-2705.
745
Abstract— Par tun us pelagicus was
collected at regular intervals from two
marine embayments and two estuaries
on the lower west coast of Australia
and from a large embayment located
approximately 800 km farther north.
The samples were used to obtain data
on the reproductive biology of this
species in three very different envi-
ronments. Unlike females, the males
show a loosening of the attachment
of the abdominal flap to the cephalo-
thorax at a prepubertal rather than a
pubertal molt. Males become gonadally
mature (spermatophores and seminal
fluid present in the medial region of
the vas deferentia) at a very similar
carapace width (CW) to that at which
they achieve morphometric maturity,
as reflected by a change in the relative
size of the largest cheliped. Logistic
curves, derived from the prevalence
of mature male P. pelagicus, gener-
ally had wider confidence limits with
morphometric than with gonadal data.
This presumably reflects the fact that
the morphometric (allometric) method
of classifying a male P. pelagicus as
mature employs probabilities and is
thus indirect, whereas gonadal struc-
ture allows a mature male to be read-
ily identified. However, the very close
correspondence between the CWj^'s
derived for P. pelagicus by the two
methods implies that either method
can be used for management purposes.
Portunus pelagicus attained maturity
at a significantly greater size in the
large embayment than in the four
more southern bodies of water, where
water temperatures were lower and the
densities of crabs and fishing pressure
were greater. As a result of the emigra-
tion of mature female P. pelagicus from
estuaries, the CWj^'s derived by using
the prevalence of mature females in
estuaries represent overestimates for
those populations as a whole. Estimates
of the number of egg batches produced
in a spawning season ranged from one
in small crabs to three in large crabs.
These data, together with the batch
fecundities of different size crabs, indi-
cate that the estimated number of eggs
produced by P. pelagicus during the
spawning season ranges from about
78,000 in small crabs (CW=80 mm)
to about 1,000,000 in large crabs
(CW=180 mm).
Manuscript approved for publication
19 June 2003 by Scientific Editor.
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:745-757 (2003).
Reproductive biology of the blue swimmer crab
(Portunus pelagicus, Decapoda: Portunidae) in
five bodies of water on the west coast of Australia
Simon de Lestang
Norman G. Hall
Ian C. Potter
Centre for Fish and Fisheries Research
Division of Science and Engineering
Murdoch University
South Street
Murdoch, Western Australia 6150
E-mail (for S de Lestang): simond@murdoch edu.au
Portunid crabs, such as Portunus pelagi-
cus, Scylla serrata, and Callinectes sapi-
dus, form the basis of important
commercial and recreational fisheries.
The blue swimmer crab (P. pelagicus)
is found in sheltered nearshore marine
waters and estuaries throughout the
Indo-West Pacific (Stephenson, 1962;
Kailola et al., 1993). In Australia, the
commercial catches of this portunid
have increased greatly during the last
20 years, and annual catches in 1998
reached 4377 metric tons (t) (Anony-
mous, 2000). The commercial fishery
for P. pelagicus in Western Australia
is the largest in Australia; the catch in
the 1999-2000 financial year weighed
673 t and fetched a wholesale price of
approximately $A3 million fCAES^).
Large numbers of portunids fre-
quently enter estuaries as juveniles
and remain there for an extended
period (Hill, 1975; Potter et al., 1983;
Perkins-Visser et al., 1996; Potter and
de Lestang, 2000). Although female
portunids sometimes become ovigerous
in estuaries, such individuals emigrate
into coastal marine waters, where they
release their eggs (Van Engel, 1958;
Metcalf et al., 1995; Potter and de Les-
tang, 2000). In contrast, the individuals
of those assemblages of portunids that
occupy marine embayments often do
not leave these marine environments
to spawn and, in cases where there is
a salinity gradient, they spawn in the
high salinity regions of those systems
(e.g. Campbell, 1984; Sumpton et al.,
1994; Prager, 1996; Potter and de Les-
tang, 2000).
The most common method for deter-
mining the size at which male crabs at-
tain maturity is to estimate the size at
which the pattern of growth of one of
its appendages changes from that which
characterizes juvenile crabs to that
which characterizes adult crabs (e.g.
Hartnoll, 1974; Somerton, 1980; Reeby
et al., 1990; Muino et al., 1999). Howev-
er, this indirect approach is not precise
and requires careful measurements of
a considerable number of individuals
covering a wide size range. Despite the
fact that macroscopic characters can be
used to distinguish sequential stages in
the development of the vas deferentia
of portunids (Ryan, 1967a; Meagher,
1971), few studies have attempted to
use such staging to determine the body
size at which the gonads of male crabs
attain maturity (e.g. Reeby et al., 1990).
Sumpton et al. (1994) considered that,
as in female P. pelagicus, a marked loos-
ening of the attachment of the abdomi-
nal flap to the cephalothorax signaled
the attainment of maturity in male
P. pelagicus. However, this criterion
has yet to be shown to be valid for the
males of this species. Although varia-
tions in the size at which crustaceans
reach maturity among bodies of water
and geographical regions may reflect,
1 CAES (Department of Fisheries, Catch
and Effort Statistics). 2002. Unpubl.
data. Western Australian Department
of Fisheries, Catch and Effort Sta-
tistics. Fisheries Western Australia, WA
Marine Research Laboratories, West Coast
Drive, Waterman, 6020, Perth, Australia.
746
Fishery Bulletin 101(4)
in part, differences in such features as genetic
composition and density, there is a strong overall
tendency for the size at maturity of this crustacean
to be inversely related to water temperature (Pillai
and Nair, 1971; Jones and Simons, 1983; Polovina,
1989; Dugan et al., 1991; Miliou, 1996; Somerton
and Donaldson, 1996; Fisher, 1999).
Estimates of the fecundity of crabs have typically
been based on the number of eggs in a single batch
of eggs (e.g. Potter et al., 1983; Melville-Smith,
1987; Ingles and Braum, 1989). However, such an
approach does not take into account the fact that
female crabs often produce more than one batch of
eggs during a spawning season (Van Engel, 1958;
Pillai and Nair, 1971; Campbell, 1984).
The aims of this study were as follows. 1)
Compare the results of three methods directed
at determining whether male P. pelagicus have
attained maturity and elucidate whether each
method produces reliable results. 2) Compare
aspects of the reproductive biology oi P. pelagicus
in two estuaries and two marine embayments in
temperate Australia with those of this species in
a large marine embayment in a much warmer
and more northern subtropical environment.
Particular emphasis will be placed on comparing
the size at maturity of both sexes and the periods
during which ovigerous females are present, and
on proposing reasons for the significance of any
differences between the assemblages in these five
bodies of water 3) Use the data collected for one of
the marine embayments to determine the age and
time of year at which P. pelagicus becomes mature
and develop a method for deriving the annual fe-
cundity that takes into account the fact that the
larger individuals of this species are believed to
produce more than one batch of eggs in a spawn-
ing season.
■ 25 S
Indian Ocean
30 S
Cockburn Sound
Peel-Harvey Estuary ,
Leschenault Estuary 1 . >-
Koombana Bay
110E
120 E
Figure 1
Map showing locations of the five bodies of water in which Portunus pelagi-
cus was sampled on the west coast of Australia. The map of Australia (insert
top right) shows the location (box) of the five bodies of water, and the map
of the lower west of Australia (insert bottom left) shows the location of the
four bodies of water sampled in this region.
Materials and methods
Sampling regimes
Up to 100 Portunus pelagicus were collected monthly for
two years from the Leschenault Estuary (May 1997-April
1999), Koombana Bay, and Cockburn Sound (June 1998-
May 2000), for three years from the Peel-Harvey Estuary
(May 1995-April 1998), and bimonthly for two years from
Shark Bay (July 1998-May 2000). The first four bodies of
water are located on the lower west coast of Australia,
approximately 800 km to the south of Shark Bay (Fig. 1).
The nearshore, shallow waters in each of these bodies of
water (water depth <1.5 m) were sampled for P. pelagicus
by using a 21.5-m seine net with a bunt of 3-mm mesh;
whereas offshore deeper waters were sampled by employ-
ing a small otter trawl net with a codend of 25-mm mesh
and crab traps consisting of either 12- or 76-mm mesh (see
Potter and de Lestang, 2000 for further details of the nets
and traps). The mean water depths at the deeper offshore
sites of the above five bodies of water were 3, 9, 19, 3, and
10 m, respectively. The water temperature at the bottom
of the water column at each site was recorded on each
sampling occasion.
Measurements and changes at puberty
The carapace width (CW) of each crab, i.e. the distance
between the tips of the two lateral spines of the carapace,
was measured to the nearest 1 mm. The length and height
of the propodus of the largest cheliped, the length of the
merus of the second walking leg, and the length of the pri-
mary pleopod of each male crab in Cockburn Sound and
Shark Bay were measured to the nearest 0.1 mm. Because
the relationship between the length of the dorsal propo-
dus of the largest cheliped and the width of the carapace
showed the greatest change over the size range of male
crabs, that structure was chosen for allometric analysis to
determine the size at which males become morphometri-
cally mature.
De Lestang et al.: Reproductive biology of Portunus pe/agicus
747
Sex of small crabs, i.e. with a CW < about 30 mm, was
determined with a dissecting microscope to ascertain
whether their pleopods bore setae and thus the crabs were
females. At CWs > about 30 mm, the female crabs could
readily be distinguished from male crabs by their posses-
sion of a far wider abdominal flap (Van Engel, 1958; Warner
1977). During the pubertal molt of female portunids, the
abdominal flap changes from a triangular to oval shape
and from being tightly to loosely fixed to the cephalothorax
(Ryan, 1967b; Fielder and Eales, 1972; Ingles and Braum,
1980; Fisher, 1999).
The size and time of occurrence of all ovigerous females
were recorded. The ovary of each crab was assigned to one
of four stages by using macroscopic characters similar
to those described for the development of the ovaries of
P. pelagicus and other portunids (Ryan, 1967b; Meagher,
1971; Krol et al., 1992; Kumar et al.^). The assignment of
these stages was augmented by examining the characteris-
tics of a subset of 200 of these ovaries in 6-/.im histological
sections that had been stained with Mallory's trichrome.
For 5-10 ovaries of each macroscopic stage, the diameters
of 30 randomly selected oocytes that had been sectioned
through the nucleus were measured to the nearest 5 pm.
Two measurements (the longest diameter and shortest di-
ameter) for each oocyte were then averaged to provide an
estimate of each oocyte diameter.
Male crabs were designated as either morphometrically
immature or mature by using differences in the regression
equations for the relationships between the natural loga-
rithms of the length of the dorsal propodus of their largest
cheliped and carapace width in what were clearly either
juvenile (small and gonadally immature) or adult crabs
(large and gonadally mature). For full description of the
method see Somerton (1980).
On the basis of their macroscopic appearance, the vas
deferentia of each male crab were assigned to one of three
stages by using criteria derived from the description of go-
nadal development for P. pelagicus by Meagher ( 1971) and
for P. sanguinolentus by Ryan (1967a). Aquarium studies
by Meagher (1971) showed that male crabs with gonads at
stages I and II did not copulate and are thus considered im-
mature, whereas those with gonads at stage III copulated
successfully with females and thus have mature gonads.
Ovaries and vas deferentia from a wide size range of
at least 20 females and 20 males, respectively, from each
sampling occasion in each of the five bodies of water were
weighed to the nearest 0.01 g. The mean gonad weight
at a constant carapace width for each sex in each month
in each water body was determined by using analysis of
covariance (ANCOVA) of the natural logarithm of the go-
nad weight as the dependent variable, month as a fixed
factor, and the natural logarithm of the carapace width
as a covariate. The common constant carapace width of
'^ Kumar, M., Y. Xiao, H. Williams, G. Ferguson, G. Hooper, and S.
Venema. 1999. Blue crab fishery. South Australian Fisher-
ies Assessment Series. 99/02, 64 p. South Australian Research
Development Institute, Grenfell Centre Level 14. 2.5 Grenfell
Street Adelaide 5000, Australia.
crabs in all bodies of water was a default value calculated
by the ANCOVA.
Size frequency and reproductive data for the corre-
sponding months in the different years in each water body
were pooled for describing intra-annual trends in these
variables.
Size at maturity
The percentages of female crabs of different carapace
widths which, in each water body, had undergone a pubertal
molt, were subjected to logistic regression to determine the
size at which 50% of the female crabs would have become
mature sensu Hartnoll (1974). Data for each assemblage
were randomly resampled and analyzed to create 1000
sets of bootstrap estimates of the parameters of the logis-
tic regression and estimates of the probability of maturity
within the range of recorded carapace widths. The 95% con-
fidence intervals of the CWgg's were derived by using this
resampling technique, which produced slightly more con-
servative estimates than those obtained from the Hessian
matrix of the logistic regression and thus reflected better
the uncertainty of the parameter that was associated with
the data. The 95% confidence intervals of the probability
of maturity at each specified carapace width were taken
as the 2.5 and 97.5 percentiles of the corresponding pre-
dicted values resulting from this resampling analysis. The
point estimate of each parameter and of each probability
of maturity at the specified carapace width were taken as
the medians of the bootstrap estimates.
The percentages of mature male crabs at different cara-
pace widths in each of the five bodies of water, with matu-
rity being assigned by using firstly morphometric and then
gonadal criteria (see earlier), were subjected to logistic re-
gressions to determine the CWgg's for these variables. The
percentages of male crabs in Cockburn Sound and Shark
Bay, which possessed an abdomen that was loosely fixed
to the cephalothorax, were likewise subjected to logistic
regression analysis. The logistic regressions relating ma-
turity and carapace width for both the females and males
in the different assemblages were compared by using a
likelihood ratio test, as described by Cerrato (1990) and
emplo3ring a Bonferroni correction.
Fecundity
The total wet weight of eggs in each batch of eggs of 40
early-stage ovigerous females, i.e. with yellow eggs, from
Cockburn Sound and which covered a wide size range,
was weighed to the nearest 0.001 g. The number of eggs
in each of four replicate subsamples from each batch
were recorded, after which each of those subsamples was
weighed to the nearest 0.001 g. These data were then used
to estimate the total number of eggs in each batch of eggs
of each female. The relationship between batch fecundity
(BF) and carapace width (CW) was described by using the
equation \nBFj=m\nCWj+b.
The number of batches of eggs produced by a full size
range of mature females during the spawning period
was estimated by determining the spawning period (SP),
748
Fishery Bulletin 101(4)
defined as the time (days) when > 5% of all ma-
ture females were ovigerous, and the proportions of
ovigerous females among all mature females in se-
quential 10-mm CW intervals during the spawning
period. The proportion of ovigerous females (O^) in
the^th size class during this period also represents
the average time a mature female in this size class
is ovigerous during that period and takes into ac-
count the fact that an ovigerous female spawns at
least once during a spawning period and that the
brood period (BP) of an ovigerous female is about
18 days at 20°C (Meagher. 1971). Thus, the mean
number of batches (NB ) produced by the mature
female crabs in thejth size class during a spawn-
ing period (average water temperature 20.4°C ) can
be estimated with the equation NB=O^SP/BP.
The relationship between number of broods
and carapace width was described empirically
by fitting a modified logistic curve, NB=l+Nb^^J
1 1 -i-exp[-ln(l 9 )(CW-a)/(6-a)), ranging upwards
from a minimum of one batch to a maximum of
l+NB^^^ batches, where a and b are parameters.
The total fecundity of crabs at different carapace
widths was calculated as the product of batch
fecundity, BF, and the number of broods, NB, by
using the relationships between BF and CW and
NB and CW, as described above.
Results
o
-•— Shark Bay
-o — Cockburn Sound
-i- Peel-Harvey Estuary
-♦— Leschenault Estuary
30
28
26
24
a 22
k 20
E
(D
r 18
0)
14
12
10
JFIylAMJJASOND
Month
Figure 2
Mean monthly water temperatures for sampling sites in Shark Bay,
Cockburn Sound, Peel-Harvey Estuary, and Leschenault Estuary
(water temperatures in Koombana Bay were essentially the same as
those in Leschenault Estuary). Mean monthly water temperatures in
each water body were derived from data pooled for at least two years.
The black rectangles on the .v axis refer to summer and winter months
and open rectangles to autumn and spring months.
Water temperature
Mean monthly water temperatures at the bottom of the
water column in the Leschenault Estuary, Peel-Harvey
Estuary, Cockburn Sound, and Shark Bay followed the
same trends, with values rising to a maximum in mid to
late summer and declining to a minimum in mid-winter
(Fig. 2). Water temperatures in Koombana Bay were
essentially the same as those in Leschenault Estuary.
Although the mean monthly water temperatures in the
Leschenault and Peel-Harvey estuaries and Koombana
Bay in corresponding months were similar, they were
lower in these bodies of water than in Cockburn Sound in
eight of the twelve months of the year (Fig. 2). However,
the mean water temperatures in each month in Shark Bay
were greater than those in the corresponding months in
each of the above four more southern bodies of water. Thus,
for example, although the maximum mean monthly water
temperature was 28°C in Shark Bay, it never reached
25°C in any of the other bodies of water (Fig. 2). Likewise,
the minimum monthly water temperature was greater in
Shark Bay (19°C) than in either Cockburn Sound ( 16°C) or
the Leschenault and Peel-Harvey estuaries (12-13°C).
Macroscopic and histological gonad staging
Macroscopic examination of the gonads of a large number
of females and males oi P. pelagicus, covering a wide size
range, and, in the case of females, an histological examina-
tion of the ovaries of a subset of these crabs, showed that the
ovaries and vas deferentia could be classified into four and
three developmental stages, respectively (Tables 1 and 2).
Size at sexual maturity
The minimum carapace widths of female crabs that had
undergone their pubertal molt ranged from 61 mm in
both the Peel-Harvey Estuary and Shark Bay to 84 mm
in the Leschenault Estuary. Although the CWg^'s derived
for females at maturity in Cockburn Sound (86.4 mm)
and Koombana Bay (86.9 mm) were not significantly dif-
ferent (P>0.05), both of these values were significantly
less (P<0.05) than the 92.0 mm for females in Shark Bay
(Fig. 3). The high CWg^'s for female crabs in the Peel-
Harvey (97.5 mm) and Leschenault estuaries (98.0 mm)
were not representative of females in their populations as
a whole (see "Discussion" section).
The relationships between the dorsal length of the larg-
est cheliped and carapace width of male P. pelagicus in
each of the five bodies of water were described better by
using two log-log lines (Fig. 4A) rather than a single log-log
line. The CWjjg's of male crabs at morphometric maturity in
the four bodies of water on the lower west coast, esti-
mated with data obtained from an allometric approach
and employing the above log-log regressions, ranged only
from 86.2 mm in the Peel-Harvey Estuary and Cockburn
De Lestang et al : Reproductive biology of Portunus pelagicus
749
Table 1
Morphological characteristics of macroscopic stages in the development of the ovaries of Portunus pelagicus and the types of oocytes
found in each of those stages. Mean diameters of oocytes at different stages in development are shown in parentheses.
Maturity stage
Macroscopic appearance of ovary
Types of oocytes
I Immature Relatively small, flattened and off white to ivory in color
Anterior region is small, and does not displace the hepa-
topancreas. The central "H" shaped region, located in the
gastric region, is loosely joined to the dorsal surface of the
spermathecae. The posterior section, located in the cardiac and
intestinal regions, forms two parallel lobes.
II Early Conspicuously larger than stage-! ovaries, pale yellow, oval
development in cross section and slightly nodulated. The anterior region
marginally displaces the hepatopancreas and the central
region envelops the dorsal surface of the spermathecae, and
the two lobes of the posterior region are starting to become
convoluted.
III Late Large, yellow, and nodulated. Anterior region displaces the
development hepatopancreas, and the central and posterior regions occupy
almost all of the space in the gastric, posterior and intestinal
cavities. Most of the spermathecae are enveloped by ovarian
tissue.
IV Fully mature Very large, deep yellow to orange, and highly nodulated.
Hepatopancreas is now completely displaced from its former
position by the enlargement of the anterior region of the ovary.
The gastric, posterior, and intestinal cavities are completely
filled with the enlarged central and posterior sections of the
ovary. The spermathecae are totally enveloped by the ovary.
Loosely packed oocytes, comprising oogo-
nia (5 ^m) and, to a lesser extent, chromatin
nucleolar oocytes (10 ^m) and perinucleolar
oocytes (30 ^m). These three types of oocytes
are found in each of the next three ovarian
stages.
Yolk-vesicle oocytes (90 ^m) are present for
the first time.
Early yolk-granule oocytes (130 fjm) sur-
round small areas of early stage oocytes,
and some late yolk vesicle oocytes are
present.
Advanced oocytes all at the late yolk-
granule stage (250 /jm).
Table 2
Morphological characteristics of stages in the development of the vas deferens of Portunus pelagicus and the location of spermato-
phores in those stages.
Maturity stage
External appearance of vas deferentia
Histological appearance of the vas deferentia
I Immature I Vas deferentia not detectable macroscopically.
NA
II Immature II Anterior vas deferentia (AVD) becoming enlarged, middle Spermatophores present in AVD. MVD and PVD
and posterior vas deferentia (MVD and PVD, respectively) contain no spermatophores.
straight and opaque.
III Mature I AVD and MVD enlarged and white and PVD enlarged and Spermatophores present in AVD and MVD PVD
convoluted but still opaque. contains no spermatophores.
Sound to 87.2 mm in the Leschenault Estuary (Fig. 4B).
The CWjq's for each of these bodies of water, which were
not significantly different (P>0.05) from each other, were
significantly less at P < 0.05 or 0.001 than the 96.0 mm
determined for male crabs in Shark Bay.
The CWjq's for males at gonadal maturity in each water
body, derived from the prevalence of males with mature go-
nads, i.e. stage III (Fig. 4C), differed by only 0.3 to 2.2 mm
from those derived for males in each corresponding water
body by using the prevalence of morphometrically mature
males (Fig. 4B). The CWg^'s derived for male crabs from
gonadal data in the four southern bodies of water, which
ranged only from 86.5 to 88.4 mm, did not differ signifi-
cantly (P>0.05 ). However, on the basis of gonadal data, each
of these CWjg's differed significantly at P<0.05, 0.01, or
0.001 from the 97.0 mm estimated for male crabs in Shark
Bay (Fig. 4C). These trends were parallel to those derived
from morphometric data (Fig. 4B).
The logistic curves derived from gonadal data in each of
the four southern bodies of water were significantly differ-
750
Fishery Bulletin 101(4)
Female size at maturity
180
100
80
60
40
20
0
100
80
60
40
20
0
Koombana
Bay
Ij
III
86 9 mm
—-^■- \ (358)
20
60
100
140
180
Cockburn
Sound
180
60 100 140 180
Carapace width (mm)
60 100 140 180
Carapace width (mm)
Figure 3
Logistic regressions and their 95% confidence limits fitted to percentage contribu-
tions of those adult females which, at each size, had undergone their pubertal molt in
each of the five bodies of water sampled in Western Australia. Arrows and measure-
ments denote CW^q's and the numbers in parentheses refer to the number of crabs
used to create the regressions.
ent (P>0.05) and had steeper slopes than those determined
by using morphometric data (Fig. 4). The confidence limits
for the logistic curves constructed from gonadal data were
also usually tighter than those constructed from morpho-
metric data.
The CWr,Q for male crabs with a loose abdominal flap
in Cockburn Sound, i.e. 72.1 mm, differed significantly
(/'<0.05) from that in Shark Bay, i.e. 76.2 mm (data not
shown). However, all of the male crabs in Cockburn Sound
with carapace widths of 70 to 75 mm and loosely attached
abdominal flaps contained gonads at stage I or II and were
thus immature.
Trends exhibited by gonad weights and proportions of
ovigerous females
The mean monthly gonad weight of mature female crabs
with a standard carapace width ( 104 mm), as determined
by ANCOVA (see "Material and methods" section), rose to
a sharp peak of about 5 g in October in Koombana Bay and
in September in Cockburn Sound (Fig. 5). In contrast, the
mean monthly gonad weights of mature female crabs in
the Leschenault and Peel-Harvey estuaries remained at
<1.5 g and did not tend to peak sharply at any time of the
year. The mean monthly gonad weights of mature female
De Lestang et al.: Reproductive biology of Portunus pelagicus
751
T3
Q.
0)
Leschenault Estuary
Koombana Bay
Peel-Harvey Estuary
./
/
Cockburn Sound
Shark Bay
100
50
0
100
50
0
100
50
= 0
100
50
0
100
50
2 3 4 5 6
Log carapace width (mm)
96 0 mm
(602)
100
50
0
100
50
0
100
50
= 0
100
50
0
100
50
88.0 mm
J/^ (76)
J
86 5 mm
(67)
97,0 mm
f (813)
0 40 80 120 160
Carapace width (mm)
0 40 80 120 160
Carapace width (mm)
Figure 4
Maturity data for male Portunus pelagicus in the five bodies of water sampled in Western
Australia. (A) Relationship between the natural logof length of the dorsal propodus of the
largest cheliped and the natural log of carapace width. Logistic regressions and their 95%
confidence limits were fitted to percentage contributions of those adult males, which, at
each size, were (B) morphometrically mature and (C) possessed mature gonads. Arrows
and measurements denote CW^^'s and the numbers in parentheses refer to the number
of crabs used to create the regressions.
crabs in Shark Bay did not peak sharply at any time and
were > 1 g in all but two of the ten months in which this
embayment was sampled (Fig. 5).
The mean monthly gonad weights of male crabs with a
standard carapace width of 118.4 mm, as determined by
ANCOVA, varied little and never exceeded 1 g in any of
the five bodies of water (data not shown). However, they did
reach their maxima at a similar time of the year, i.e. late
summer (February) or early autumn (March), in the four
bodies of water on the lower west coast of Australia.
The monthly percentage contributions made by oviger-
ous female crabs from all mature female crabs in Koom-
752
Fishery Bulletin 101(4)
10
8
6
4
2
0
10
8
6
4
2
0
10
g, g Peel-Harvey
i 6 ■ ^s'^^^y
2
0
10
8
6
4
2
0
10
8
6
4
2
0
Leschenault
Estuary
Koombana
Bay
Cockburn
Sound
Shark Bay
iir~i-i — m
V \,
O -1'
%
'^ \ '^^'^ ^<5 \ '
Figure 5
Mean monthly gonad weights +95'/f confidence intervals for female Por-
tunus pelagicus and the percentage of ovigerous females among all adult
female P. pelagicus in the five bodies of water sampled. The black rectangles
on the -V axis refer to summer and winter months and open rectangles to
autumn and spring months.
bana Bay and Cockburn Sound peaked at 32-36% in No-
vember and December in Koombana Bay and at 35-45%
in October to December in Cockburn Sound. Although
the corresponding percentage contributions of ovigerous
female crabs were far lower in the Leschenault and Peel-
Harvey estuaries than in the above two embayments, they
still reached their maxima at the same time of the year
(Fig. 5). Few or no ovigerous female crabs were caught in
either of those estuaries or embayments between March
and August. Ovigerous females were found in Shark Bay
in each of the ten months in which that embayment was
sampled and, unlike the situation in the two more southern
marine embayments, their monthly contributions to the
overall number of adult female crabs did not vary markedly
throughout the year (Fig. 5).
Trends exhibited by oocyte development
The maximum diameter of the oocytes increased progres-
sively from 95 pm in stage-I gonads to 315 pm in stage-IV
gonads (data not shown). The modal oocyte diameter
of the distinct and largest cohort of oocytes in stage IV
(240-259 ^m) was only slightly less than that of the fertil-
ized yellow external eggs found under the abdominal flap
of ovigerous females (300-319 /jm). The most advanced
oocytes in ovaries at stages III and IV were at the early
and late yolk-granule stage, respectively. The presence of
two distinct size cohorts of oocytes in the ovaries of large
females with grey eggs under their abdomen (i.e. eggs that
had been fertilized for several days) is consistent with the
view that large female P. pelagicus are multiple spawners.
De Lestang et al : Reproductive biology of Portunus pelagicus
753
20
15
0 '■
30
15
0
30 t
Female crabs
r-Th-T-r-n
10
0
30 r
15
0 ^
30
0'
30
15
0
30 r
Male crabs
40 80 120 160 0
Carapace width (mm)
40
80
120
160
Figure 6
Frequency histogi-ams for the carapace widths of juvenile and adult females and males
of Portunus pelagicus in Cockburn Sound, n = number of crabs measured. Designation of
males as mature was based on gonadal criteria described in the text.
Age and time of sexual maturation of Portunus pelagicus
The carapace-width frequency data for P. pelagicus in
inshore and offshore waters in Cockburn Sound demon-
strated that, in this marine embayment, males are repre-
sented by two main size cohorts in January and February
(Fig. 6). The first size cohort represents the 0+ age class
that resulted from the spawning period that commenced in
the previous August-September, whereas the second cohort
corresponds to l-i- crabs, which start to decline markedly in
numbers after February and are rarely represented after
June (Fig. 6). Although similar trends are exhibited by the
data for females, the numbers of 1+ individuals of this sex
remained higher for a longer period, i.e. until May. The
above trends are entirely consistent with those reported
in detailed studies of the age composition and growth of
P. pelagicus in the Peel-Harvey Estuary (Potter et al., 1983)
and Leschenault Estuary (Potter and de Lestang, 2000).
None or very few of the female and male 0+ crabs caught
in Cockburn Sound in January, February, and March were
mature. However, some of the larger 0-t- crabs had become
mature by May, i.e. when they would mostly have been
754
Fishery Bulletin 101(4)
O 350
o
° 300
• •
X
- 250
r^^^^
"O 200
1 '50
■g 100
S 50
4
03
■5 3
m
^ • ^
o 2
y^
Number
0
5- 1000
0
0
'- 800
X
^^^^--"^^
Total fecundity
0000
^ ^--^^
80 90 100 110 120 130 140 150 160
Carapace width (mm)
Figure 7
Batch fecundity, mean number of batches, and total fecundity of Por-
tunus petagicus in Cockburn Sound.
between four and eight months old (Fig. 6). The prevalence
of mature crabs subsequently increased, with the result
that the vast majority of crabs in the following January,
i.e when they had just entered their second year of life,
were mature (Fig. 6). Thus, all crabs have typically become
mature when they are just over one year old.
Fecundity
In Cockburn Sound, the number of eggs recorded for
a single batch of eggs under the abdomen of a female,
ranged from 68,450 in a crab with a CW of 84 mm to
324,440 in a crab with a CW of 154 mm (Fig. 7). The rela-
tionship between batch fecundity (BF) and carapace
width (CW) is described by the following equation: lnfiF=
1.82081nCH'+3.2862.
The estimated mean number of egg batches, produced by
female crabs in the different size classes over the spawn-
ing period, ranged from about one in crabs of 100-109 mm
CW to about three in crabs of 150-159 mm CW (Fig. 7). A
range of one to three batches per instar corresponds to that
recorded by Campbell ( 1984) for P. pelagicus in aquaria ex-
periments. The empirical relationship between the number
of batches (NB) and carapace width (CW) is described by
A^Bj=l+2/{l+exp[-ln(19)(CW,-113.7)/13.8]|.
A combination of the equations for the relationships
between batch fecundity and CW and the number of egg
batches and CW was then used to determine the relation-
ship shown between total fecundity (TF) and carapace
width (CW) and which is shown in Figure 7.
Discussion
Designation of maturity in male crabs
Aquaria studies by Meagher (1971) demonstrated that
male crabs with gonads at stage 111, i.e. with spermato-
phores and seminal fluid in the medial vas deferentia, can
copulate successfully with females. Because this parallels
the situation recorded by Comeau and Conan (1992) for the
snow crab iChionoecetes opilio), we likewise regard such
gonads as mature. Our study also showed that, because
male P. pelagicus still possess immature gonads (stage II)
when their abdominal flap becomes loosely attached to the
cephalothorax, the latter change occurs at a prepubertal
De Lestang et a\ : Reproductive biology of Portunus pe/agicus
755
molt and thus, unlike the supposition of Sumpton et al.
(1994), does not coincide with the attainment of maturity.
The situation in males thus contrasts with that in female
P. pelaglcus, in which the abdominal flap becomes loose as
an outcome of the pubertal molt (Fielder and Eales, 1972;
Campbell, 1984; Potter and de Lestang, 2000; Smith^).
The very close similarity between the corresponding
CWjq's derived for male P. pelagicus in each of the five
bodies of water by using morphometric and gonadal data
demonstrates that morphological and gonadal maturity
are attained by this species at essentially the same cara-
pace width. However, the question of whether a male crab
of about the size of maturity has become morphometrically
mature depends on determining whether the relative
length of one of its appendages is closer to the regression
line which relates the length of that appendage to the cara-
pace width in either juvenile or adult crabs. Because the
overall relationship between cheUped length and carapace
width of P. pelagicus does not undergo a marked shift at
around the attainment of maturity, the use of the allome-
tric method never enabled us to determine with absolute
certainty whether, in the region of size overlap, a male
was morphometrically immature or mature. The lack of
precision, when determining maturity with morphometric
data, could account for the slopes of the logistic curves for
the prevalence of "mature" individuals of P. pelagicus de-
rived from these morphometric data in the four southern
bodies of water being shallower than those obtained from
gonadal data.
From the above, it follows that there would be an advan-
tage in determining the CW^q's for male P. pelagicus at ma-
turity by using data on gonadal state obtained by the simple
and direct procedure of examining the vas deferentia, rather
than relying on data obtained by an allometric method that
is indirect and relies on a careful measurement of the ap-
pendage lengths and carapace dimensions of a considerable
number of individuals. However, the remarkable similari-
ties between the CWg^'s derived by using morphometric and
gonadal data show that, if it is desirable to avoid damaging
the crabs, the data obtained from allometric analysis does
yield a close approximation of this important measure for
P. pelagicus. Thus, the CW^q derived from either gonadal
or morphometric data for male P. pelagicus can be used for
developing management plans for this species.
The very close correspondence between the size at which
gonadal and morphometric maturity are attained by the
males of P. pelagicus contrasts with the situation recorded
by Comeau and Conan (1992) and Sainte-Marie et al.
( 1997) for the males of the snow crab Chionoecetes opilio.
In this latter species, the males attain gonadal maturity at
a smaller body size than that at which morphometric ma-
turity is attained following the terminal molt. The males
of C. opilio with large cheliped and large body size are at a
competitive advantage over smaller males when courting
(Comeau and Conan, 1992; Sainte-Marie et al., 1997). Be-
•^ Smith, H. 1982. Blue crabs in South Australia — their status,
potential and biology. Report 6, p. 33-51. South Australian
Fisheries Industry Council Grenfell Centre Level 14, 25 Grenfell
Street Adelaide 5000, Adelaide, Australia.
cause the aquaria studies of Campbell (1984) have shown
that the large males of P pelagicus also have a similar com-
petitive advantage during courting, a male of this species
with mature gonads and differentiated chelipeds may not
be able to compete for females successfully if larger males
are present.
Influence of migration on estimates of CW50 for female crabs
The mean monthly gonad weights recorded for post-
pubertal individuals were less for females in estuaries
than in marine embayments, strongly indicating that
females often tend to emigrate from the estuaries to their
spawning grounds before their gonads are fully developed
(Van Engel, 1958; Potter and de Lestang, 2000). Such an
emigration from estuaries by mature female P. pelagicus
reduces the proportion of mature individuals within each
carapace width interval, thereby increasing the proportion
of immature females in these class intervals. This shifts the
logistic curve to the right and consequently increases the
CW5Q, which accounts for the significantly greater CWgg's
derived for females in estuaries than in marine environ-
ments on the lower west coast of Australia. For this reason,
subsequent comparisons of the CW^q's for female crabs in
the different bodies of water will focus on those derived for
assemblages in the three marine embayments.
In contrast to the CWgg's for females, the CWgg's for
males at maturity in the two estuaries and the two marine
embayments on the lower west coast of Australia were not
significantly different. This presumably reflects the fact
that, unlike mature females, the large males of P. pelagicus
tend to remain in estuaries during the spawning period
(Potter and de Lestang, 2000).
Influence of temperature on reproductive biology
The CWjq's derived for males at "maturity" in each of the
five bodies of water never differed by more than 2.2 mm,
irrespective of whether gonadal or morphometric data
were used. However, the maximum CWgg's determined for
males in the two estuaries and two embayments by using
gonadal and morphometric data, i.e. 88.4 mm for Cockburn
Sound and 87.2 mm for the Leschenault Estuary, respec-
tively, were 8.6 and 8.8 mm less than the corresponding
CWgg's determined for males in Shark Bay Furthermore,
the CWgg's for female P. pelagicus in Koombana Bay and
Cockburn Sound were 5.6 and 5.1 mm, respectively, less
than that of females in Shark Bay.
The greater CWgg's for P. pelagicus in Shark Bay than in
the other four bodies of water, which are located approxi-
mately 800 km farther south, runs counter to the general-
ization that the CWgg's for decapods tend to be inversely
related to water temperature (e.g. Campbell and Robinson,
1983; Jones and Simons, 1983; Dugan et al., 1991). Howev-
er, the opposite situation has sometimes been recorded and,
in those cases, has been attributed to differences among
populations of one or more of the following: density, preda-
tion pressure, and food availability (Hines, 1989; Polovina,
1989; Pollock, 1995; McGarvey et al., 1999). It thus appears
relevant that the mean density of P. pelagicus was far lower
756
Fishery Bulletin 101(4)
in the sampling sites in Shark Bay, 0.6 crabs/100 m'-, than
those in Cockburn Sound and Koombana Bay, 2.80 and 2.94
crabs/100 m-, respectively. The mean density in Shark Bay
was also far lower than those recorded in the Leschenault
and Peel-Harvey estuaries between the middle of spring
and middle of autumn, when P. pelagicus colonizes estu-
aries (Potter et al., 1983; Potter and de Lestang, 2000).
Furthermore, commercial or recreational fishing pressure
(or both), which leads to a reduction in CWgy's at maturity
in the spiny lobster (Polovina, 1989), is far greater for
P. pelagicus in the southern bodies of water than in Shark
Bay (Bellchambers''). Recent work with microsatellite
DNA has also shown that the assemblages of P. pelagicus
in Shark Bay are genetically distinct from those in more
southern bodies of water, such as Cockburn Sound and the
Peel-Harvey Estuary (Chaplin et al.^).
The marked differences between the CWjq's at maturity
for P. pelagicus in Shark Bay and bodies of water farther
south emphasize the need for managers to take into ac-
count this type of variation when determining a minimum
legal carapace width (MLCW) for capture. However, the
current MLCW for P. pelagicus in Western Australia,
127 mm, is well above even the CW^^ for this species at
maturity in Shark Bay.
The prevalence of ovigerous females did not peak sharply
at any time of the year in Shark Bay, whereas ovigerous
females were found predominantly during spring and
summer in Cockburn Sound and Koombana Bay. Moreover,
the mean monthly gonad weights of a female P. pelagicus
of standard carapace width lay within a relatively narrow
range of 0.9 to 1.8 g in Shark Bay, whereas they rose to a
sharp peak of about 5 g in spring and fell below 1 g in some
months in Cockburn Sound and Koombana Bay. The trends
exhibited by the reproductive variables of female P. pelagi-
cus thus provided strong evidence that reproductive activity
extends over much or all of the year in Shark Bay, whereas
it occurs predominantly in spring and summer in the two
southern embayments. The more protracted spawning
period in Shark Bay presumably reflects the presence of
higher water temperatures throughout the year and in par-
ticular during winter and early spring. Such a conclusion
is consistent with the results of other studies, which have
shown that water temperature influences ovulation and egg
development in P. pelagicus and other decapods (Rahaman,
1980; Campbell, 1984; Pollock, 1995; Kumar et al.^).
Fecundity
The vast majority of previous estimates of the fecundity
of crustaceans have been based on the number of eggs
borne by females at a particular time which, in the case of
multiple spawners, does not take into account the fact that
■' Bellchambers, L. 2002. Personal conimun. Fisheries West-
ern Australia, WA Marine Research Laboratories, West Coast
Drive, Waterman, 6020, Perth, Australia.
5 Chaplin, J., E.S.Yap.E.Sezmis.andl. C.Potter. 2001. Genetic
(microsatellite) determination of the stock structure of the blue
swimmer crab in Australia. FRDC project 98/1 18. 84 p. Mur-
doch University, South Street, Murdoch, 6150, Perth, Australia.
larger crabs can produce two or more batches of eggs within
a spawning period. The few previous attempts to obtain the
total fecundity of crustaceans have involved tracking the
number of batches of eggs borne by particular individuals
at different times (e.g. Chubb et al.^). The advantage of
the approach developed during the current study is that
it uses a combination of batch fecundity and an estimate
of the number of batches produced during the spawning
period by female P. pelagicus of different carapace widths
to determine the relationship between the total fecundity
and body size of this species in a given population. Because
the older crabs have a far longer intermolt period between
copulation and egg extrusion than younger crabs, i.e. eight
versus four months, they have a far greater amount of time
to accumulate the energy reserves required to produce
eggs. This difference accounts for the greater number of
egg batches produced by larger than small crabs.
Acknowledgments
Thanks are expressed to many colleagues and friends,
and particularly R. Melville-Smith, D. Fairclough, M.
Pember, T Linke, M. Travers, and W. White, who assisted
with sampling. Our thanks are also expressed to the three
anonymous referees for their constructive criticisms. Fund-
ing was provided by the Australian Fisheries Research and
Development Corporation and Murdoch University.
Literature cited
Anonymous.
2000. FAO yearbook. Fisheries statistics. Capture produc-
tion, 86/1, 713 p. FAO, Rome, Italy
Campbell, G. R.
1984. A comparative study of adult sexual behaviour and
larval ecology of three commercially important portu-
nid crabs from the Moreton Bay region of Queensland,
Australia. Ph.D. diss. 253 p. Univ. Queensland, Bris-
bane, Queensland, Australia.
Campbell, A., and D. G. Robinson.
1983. Reproductive potential of three American lobster
(Homarus americanus) stocks in the Canadian Maritimes.
Can. J. Fish. Aquat. Sci. 40:1958-1967.
Cerrato, R. M.
1990. Interpretable statistical tests for growth comparisons
using parameters in the von BertalaniTy equation. Can. J.
Fish. Aquat. Sci. 47:1416-1426.
Comeau, M., and G. Y. Conan.
1992. Morphometry and gonad maturity of male snow
crab, Chionoecetes opilio. Can. J. Fish. Aquat. Sci. 49:
2460-2468.
Dugan, J. E., A. M. Wenner, and D. M. Hubbard.
199 1 . Geographic variation in the reproductive biology of the
sand crab Emerita analoga (Stimpson) on the California
coast. J. Exp. Mar Biol. Ecol. 150:63-81.
'^ Chubb, C, C. Dibdcn, and K. Ellard. 1984. Studies on the
breeding stock of the western rock lobster, Paniilirus cygnus. in
relation to stock and recruitment. FIRTA project 85/87, 37 p.
Fisheries Western Australia, WA Marine Research Laboratories,
West Coast Drive, Waterman, 6020, Perth, Australia.
De Lestang et al : Reproductive biology of Portunus pelaglcus
757
Fielder, D. R., and A. J. Bales.
1972. Observations on courtship, mating and sexual matu-
rity in Portunus pelagicus (L., 1766). J. Nat. Hist. 6:
273-277.
Fisher, M. R.
1999. Effect of temperature and salinity on size at maturity
of female blue crabs. Trans. Am. Fish. Soc. 128:499-506.
Hartnoll, R. G.
1974. Variation in growth pattern between some secondary
sexual characters in crabs (Decapoda, Brachyura). Crus-
taceana 27:131-136.
Hill, B. J.
1975. Abundance, breeding and growth of the crab Scylla
serrata in two South African estuaries. Mar. Biol. 32:
119-126.
Hines, A. H.
1989. Geographic variation in size at maturity in brachy-
uran crabs. Bull. Mar. Sci. 45:356-368.
Ingles, J., and E. Braum.
1989. Reproduction and larval ecology of the blue swim-
ming crab Portunus pelagicus in Ragay Gulf, Philip-
pines. Int. Rev Hydrobiol. 74:471-490.
Jones, M. B., and M. J. Simons.
1983. Latitudinal variation in reproductive characteristics
of a mud crab, Helice crassa (Grapsidae). Bull. Mar Sci.
33:656-670.
Kailola, P. J., M. J. Williams, P C. Stewart, R. E. Riechelt,
A. McNee, and C. Grieve (eds.).
1993. Australian fisheries resources, 422 p. Bureau of
Resource Sciences, Canberra, Australia.
Krol, R. M., W. E. Hawkins, and R. M. Overstreet.
1992. Reproductive components. In Microscopic anatomy
of invertebrates (F. W. Harrison and A. G. Humes, eds.),
p. 295-343. Wiley-Liss, New York, NY.
McGarvey, R., G. J. Ferguson, and J. H. Prescott.
1999. Spatial variation in mean growth rates at size of
southern rock lobster, Jasi/s edwardsii, in South Australian
waters. Mar Freshw. Res. 50:333-342.
Meagher, T. D.
1971. Ecology of the crab Portunus pelagicus (Crustacea:
Portunidae) in south Western Australia. Ph.D. diss, 227 p.
Univ Western Australia, Perth, Australia.
Melville-Smith, R.
1987. The reproductive biology of Geryon maritae (Decapoda,
Brachyura) off south west Africa/Namibia. Crustaceana
53:259-275.
Metcalf K. S., J. van Montfrans, R. N. Lipcius, and R. J. Orth.
1995. Settlement indices for blue crab megalopae in the
York River, Virginia: temporal relationships and statistical
efficiency Bull. Mar. Sci. 57:781-792.
Miliou, H.
1996. The effect of temperature, salinity and diet on final
size of female Tisbe holothuriae (Copepoda, Harpacti-
coida). Crustaceana 69:742-754.
Muifio, R., L. Fernandez, E. Gonzalez-Gurriaran, J. Freire, and
J. A. Vilar.
1999. Size at maturity of Liocarcinus depurator (Brachy-
ura: Portunidae): a reproductive and morphometric study.
J. Mar Biol. Assoc. U.K. 79:295-303.
Perkins-Visser, E., T. G. Wolcott, and D. L. Wolcott.
1996. Nursery role of seagrass beds: enhanced growth of
juvenile blue crabs iCallinectes sapidus Rathburn). J.
Exp. Mar Biol. Ecol. 198:155-171.
PiUai, K. K., and N. B. Nair
1971. The annual reproductive cycles of Uca annulipes,
Portunus pelagicus and Metapenaeus affinis (Decapoda:
Crustacea) from the south-west coast of India. Mar Biol.
11:152-166.
Pollock, D. E.
1995. Changes in maturation ages and sizes in crustacean
and fish populations. S. Afr J. Mar. Sci. 15:99-103.
Polovina, J. J.
1989. Density dependence in spiny lobster, Panulirus mar-
ginatus, in the northwestern Hawaiian Islands. Can. J.
Fish. Aquat. Sci. 46:660-665.
Potter, I. C, P. J. Chrystal, and N. R. Loneragan.
1983. The biology of the blue manna crab Portunus pelagi-
cus in an Australian estuary. Mar. Biol. 78:75-85.
Potter, I. C, and S. de Lestang.
2000. Blue swimmer crab Portunus pelagicus in Leschen-
ault Estuary and Koombana Bay, south-western Aus-
tralia. J. R. Soc. West. Aust. 83:221-236.
Prager, M. H.
1996. A simple model of the blue crab, Callinectes sapidus,
spawning migration in Chesapeake Bay. Bull. Mar. Sci.
58:421-428.
Rahaman, A. A.
1980. Ecological observations on spawning of a few inver-
tebrates of the Madras coast. J. Madurai Kamaraj Univ.
9:71-77.
Reeby, J., P. N. Prasad, and M. S. Kusuma.
1990. Size at sexual maturity in the male crabs of Portu-
nus sanguinolentus and P. pelagicus. Fish. Technol. 27:
115-119.
Ryan, E. P
1967a. Structure and function of the reproductive system
of the crab Portunus sanguinolentus (Herbst) (Brachyura:
Portunidae). I. The male system. Mar. Biol. Assoc. India
Symp. Ser. 2:506-521.
1967b. Structure and function of the reproductive system
of the crab Portunus sanguinolentus (Herbst) (Brachyura:
Portunidae). II. The female system. Mar. Biol. Assoc. India
Symp. Ser 2:522-544.
Sainte-Marie, B., J. M. Sevigny, and Y Gauthier.
1997. Laboratory behaviour of adolescent and adult males
of the snow crab (Chionoecetes opilio) (Brachyura: Majidae)
mated noncompetitively and competitively with primipa-
rous females. Can. J. Fish. Aquat. Sci. 54:239-248.
Somerton, D. A.
1980. A computer technique for estimating the size of
sexual maturity in crabs. Can. J. Fish. Aquat. Sci. 37:
1488-1494.
Somerton, D. A., and W Donaldson.
1996. Contribution to the biology of the grooved and triangle
tanner crabs, Chionoecetes tanneri and C. angulatus, in the
eastern Bering Sea. Fish. Bull. 94:348-357.
Stephenson, W
1962. The evolution and ecology of portunid crabs, with espe-
cial reference to Australian species. In The evolution of
living organisms (G. W. Leeper, ed.), p. 34-67. Melbourne
Univ. Press, Melbourne, Australia.
Sumpton, W. D., M. A. Potter, and G. S. Smith.
1994. Reproduction and growth of the commercial sand
crab, Portunus pelagicus (L.) in Moreton Bay, Queensland.
AsianFish. Sci. 7:103-113.
VanEngel.W.A.
1958. The blue crab and its fishery in Chesapeake Bay Part
1. Reproduction, early development, growth and migration.
Commer. Fish. Rev. 20:6-17.
Warner, G. F.
1977. The biology of crabs, 202 p. Paul Elek (Scientific
Books), London.
758
Abstract— Culture of a non-native
species, such as the Suminoe oyster
iCrassostrea ariakensis). could offset
the harvest of the declining native
eastern oyster (Crassostrea virginica)
fishery in Chesapeake Bay. Because of
possible ecological impacts from intro-
ducing a fertile non-native species,
introduction of sterile triploid oysters
has been proposed. However, recent
data show that a small percentage
of triploid individuals progressively
revert toward diploidy, introducing the
possibility that Suminoe oysters might
establish self-sustaining populations.
To assess the risk of Suminoe oyster
populations becoming established in
Chesapeake Bay, a demographic popu-
lation model was developed. Parameters
modeled were salinity, stocking density,
reversion rate, reproductive potential,
natural and harvest-induced mortal-
ity, growth rates, and effects of vari-
ous management strategies, including
har\'est strategies. The probability of a
Suminoe oyster population becoming
self-sustaining decreased in the model
when oysters are grown at low salinity
sites, certainty of harvest is high, mini-
mum shell length-at-harvest is small,
and stocking density is low. From the
results of the model, we suggest adopt-
ing the proposed management strate-
gies shown by the model to decrease
the probability of a Suminoe oyster
population becoming self-sustaining.
Policy makers and fishery managers
can use the model to predict potential
outcomes of policy decisions, supporting
the ability to make science-based policy
decisions about the proposed introduc-
tion of triploid Suminoe oysters into
the Chesapeake Bay.
A model for assessing the likelihood of
self-sustaining populations resulting from
commercial production of triploid Suminoe oysters
iCrassostrea ariakensis) in Chesapeake Bay
Jodi R. Dew
Jim Berkson
Eric M. Hallerman
Department of Fisheries and Wildlife Sciences
106 Cheatham Hall
Virginia Polytechnic Institute and State University
Blacksburg, Virginia 24061-0321
E-mail address (for J. Berkson, conlaci auttior). |berkson(gvt.edu
Standish K. Allen Jr.
School of Manne Science
Virginia Institute of Manne Sciences
Gloucester Point, Virginia 23062
Manuscript approved for publication
16 June 2003 by Scientific Editor.
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull. lOhT.'JS-TeS (2003).
The native eastern oyster (Crassostrea
virginica) population in Chesapeake
Bay has declineti because of habitat
degradation, over-harvest, and dis-
ease- and parasite-mediated mortahty.
Efforts to restore the eastern oyster pop-
ulation in Maryland and Virginia have
been hindered by persistent diseases
and habitat degradation (Mann et al.,
1991; Gottlieb and Schweighofer, 1996).
Recent restoration efforts have included
intensified reef building programs. In
addition to restoring the native oyster,
discussions about introducing non-
native disease- and parasite-resistant
oyster species into the Chesapeake
Bay have gone forward since the early
1990s (Mann et al., 1991; Lipton et al.,
1992; Gottlieb and Schweighofer, 1996;
Hallerman etal., 2002).
In 1997, in-water testing of non-na-
tive oyster species (sterile triploids)
began in Virginia, first with the Pa-
cific oyster (Crassostrea gigas), then
with the Suminoe oyster (Crassostrea
ariakensis) (Calvo et al., 1999; Calvo
et al., 2001). Field studies with Pacific
oysters showed poor performance under
Chesapeake Bay conditions (Calvo et
al., 1999). However, field studies with
Suminoe oysters demonstrated disease
resistance and rapid growth, and indi-
viduals reached minimum harvest shell
length of about 77 mm in approximately
one year (Calvo et al., 2001). These re-
sults, and subsequent small-scale trials
by industry, evoked strong interest in
the commercial culture of Suminoe oys-
ters to supplement the eastern oyster
fishery.
Ideally, aquaculture with 100% tri"
ploid oysters would pose no risk of es-
tablishment of a self-sustaining oyster
population (Guo and Allen, 1994a).
However, a number of factors make the
use of triploids imperfect. For example,
recent data have shown that a small
percentage of triploid oysters progres-
sively revert toward diploidy with age
(Calvo et al., 2001; Zhou. 2002). Rever-
sion of triploids leads to mosaicism in
which individuals comprise both dip-
loid and triploid cells. Mosaics them-
selves are innocuous unless the re-es-
tablishment of diploid cells leads to re-
covered reproductive capability, which
could in turn lead to the establishment
of a self-sustaining Suminoe oyster
population. We define this hazard,
"reproductively effective reversion," as
the process of yielding mosaics with re-
covered reproductive capability. Repro-
ductively effective reversion introduces
the possibility that triploid Suminoe
Dew et al: Model for assessing populations of Crassostrea anakensis in Cfiespeake Bay
759
oysters planted for aquaculture could become
a self-sustaining population of diploid Suminoe
oysters and introduce numerous unknown eco-
logical consequences.
Another hazard associated with deployment
of triploid Suminoe oysters is the possibility
that nontriploids might be stocked inadver-
tently because of failure to detect them in a
mixed batch of triploid and diploid individu-
als. Although technology to produce "100%"
triploids is now available, as practiced on
Pacific oyster (Guo and Allen, 1994b; Guo et
al., 1996), the reliability of the approach for
producing "100%" triploids in Suminoe oyster
is yet undetermined. Diploids may enter the
population from several sources: chromosomal
nondisjunction in tetraploid males producing
haploid gametes, low level hermaphrodism
in diploid females yielding self-fertilized
embryos, and cross-contamination between
diploid and triploid cultures (cf Guo and Al-
len, 1997). Typically, flow cytometry has been
used to determine the presence or absence of
diploid cells (Allen, 1983). Flow cytometry has
the sensitivity to detect one diploid among a
thousand triploid oysters (Allen and Bushek,
1992); thus, the detection threshold is 0.001 with current
technology. Should the (nonzero) frequency of diploids be
greater than zero but less than one in a thousand, then
the batch would be certified 100% triploid. This failure to
detect diploid individuals in a mixed batch poses a hazard
for stocking other fertile diploid oysters in that batch into
culture systems.
Before substantial commercial introduction of triploid
Suminoe oysters into the Chesapeake Bay, any environ-
mental hazards of reproduction associated with a range
of management scenarios should be assessed. Hazards
are defined as undesirable outcomes from an activity
(Hallerman and Kapuscinski, 1995). Stocking triploid
Suminoe oysters produces two hazards in this model: the
inadvertent stocking of diploids and the reproductively
effective reversion of triploids. These two hazards may
lead to the establishment of a self-sustained Suminoe
oyster population and the probability of this occurring
is defined as a risk. Risk assessment is the process of 1)
identifying hazards posed by management actions, such
as deployment of triploid Suminoe oysters, 2) quantifying
the associated risks of hazards being realized (Hallerman
and Kapuscinski, 1995), such as the population becoming
self-sustaining, and 3) evaluating the consequences of
the hazards. Quantitative models often are used to as-
sess risk (Lackey, 1994). Building upon data collected on
growth, mortality, and reproductively effective reversion
for Suminoe oysters, we have developed a quantitative
model to estimate the risk associated with large-scale
deployment of triploid Suminoe oysters under a range of
management scenarios. The model predicts the likelihood
of out-planted triploid Suminoe oysters giving rise to a
self-sustaining population at a given site in the Chesa-
peake Bay given user-specified stocking, reproductively
Slocking, age-class zero |
Surviving
stocked
age-class zero
oysters
Starting
population
size
Next
year
Grow to
mean
shell
length
Surviving juvenile
oysters
^ Mortality ^
Reproduction
Natural
mortality
Reproductively
effective
reversion and
detection
threshold
Surviving adult oysters
Final
population
Figure 1
Flow chart depiction of the annual time step in the model for estimating
likelihood of estabhshing self-sustaining reproduction in triploid Suminoe
oysters (C. ariakensis).
effective reversion, reproduction, growth, and mortality
rates (both natural and harvest), as well as user-specified
management options.
Methods
Overview of model
A quantitative population model of the Suminoe oyster was
developed to evaluate the consequences of hazards asso-
ciated with introducing triploid Suminoe oysters under
a range of environmental conditions and management
strategies. The model includes set demographic parameters
(length-fecundity, oyster density-fertilization efficiency,
and salinity-fecundity relationships) and user-specified
variables (reproduction, growth, and natural and harvest
mortality rates). It includes options for varying stock-
ing rates, harvest rates, and other management actions.
Because little is known about Suminoe oyster reproduc-
tion, we assumed that Suminoe oysters would behave like
the congeneric eastern oysters in Chesapeake Bay; hence,
an eastern oyster fecundity model (Mann and Evans, 1998)
was used to estimate fecundity of Suminoe oysters. The
model assumes that the Suminoe oyster population is
closed, i.e. that natural immigration and emigration do not
occur The model is age-structured, and a yearly time step is
used. The state variable tracked through time is population
size. Intrinsic population growth rate is exponential and
without density dependence. The final output of the model
is the predicted population size of Suminoe oyster assum-
ing specified demographic parameters and environmental
and management variables. The model was programmed in
Visual Basic (Microsoft Corp., Redmond, WA).
760
Fishery Bulletin 101(4)
Modeling approach
In each annual time step for age classes one through six,
growth occurs to the mean shell length of the age class,
then natural mortality and harvest are imposed, and then
reproduction occurs (Fig. 1). Because Suminoe oysters
grow quickly in autumn (Cahn, 1950), the annual time
step begins in September Harvest occurs from October to
April. Natural mortality occurs at the greatest rates during
the summer months. Because an annual time step is being
used, the model is designed so that natural mortality and
harvest are imposed simultaneously. Reproduction occurs
during the summer months. The model simulates repro-
duction for fertile individuals in all mature age classes.
The final population size for a particular age class after
natural mortality and harvest becomes the starting value
for population size for the next age class in the next time
step. All individuals stocked each year are age-class zero
individuals. The starting population size for age-class one
in the next time step is equal to the sum of all individuals
less than one-year old produced by all age classes, plus the
number of individuals stocked.
Model variables, parameters, and equations
The initial conditions for the model are determined by the
user's choice of specific values for several variables (Tables 1
and 2). The key abiotic variable driving population growth
is salinity, because fecundity is highly dependent upon
salinity (Mann and Evans, 1998). Biotic variables of the
model include mean shell length for each age class, mor-
tality (natural and harvest) for each age class, disease
prevalence, total mortality of oysters less than one year
old, oyster population density, sex ratio for each age class,
and reproductively effective reversion rate for each age
class (Table 1). Other variable inputs are stocking rates,
harvest regulations, and management strategies.
Stochasticity is programed into the model to incorporate
both the uncertainty involved in estimating variable values
and environmental variation. Some variables are regarded
as stochastic variables because they vary around some
mean value from year to year, whereas other variables
(such as salinity and sex ratio of the population for each age
class) are deterministic in the model because they fluctuate
over a longer period of time in the absence of a catastrophe
(Kennedy et al., 1996). Stochasticity affects shell length,
natural mortality, and reproductively effective reversion
rates at each age, and the degree of variance is set by the
user as a constant for each year At each time step, a mean
shell length, mortality rate, and reproductively effective
reversion rate for each age class is randomly drawn from a
log normal distribution around a mean with an associated
variance.
We assume that the mean shell length of each age class
at the current time step does not affect the mean length of
the age class at a subsequent time step, because of large,
highly variable growth rates per year (Calvo et al., 2001).
Default mean shell length for each age-class values were
obtained from Cahn ( 1950). The user may, of course, specify
other mean shell lengths. Growth affects the potential for
Table 1
Definitions of model parameters and variables.
Symbol
Parameter and variable definition
A
Area (square meters)
C
Certainty in obtaining the desired harvest
rate (00.05) revealed little popu-
lation structure among samples. Mantel
tests demonstrated that the genetic
relationships among samples did not
correspond to an isolation-by-distance
model for either class of marker. Four
of eight comparisons of coastal and
offshore samples revealed differences
of allele frequencies at the Gpi-F* locus
(P<0.05), although none of these differ-
ences was significant after correction
for multiple testing (P>0.001). Results
are consistent with the hypothesis that
the CYRA yellowfin tuna samples com-
prise a single genetic stock, although
gene flow appears to be greater among
coastal samples than between coastal
and offshore samples.
Allozyme and RAPD variation in the eastern
Pacific yellowfin tuna Whunnus albacares)
Pindaro Dfaz-Jaimes
Instituto de Ciencias del Mar y Limnologia
Unlversidad Nacional Autonoma de Mexico
Circuito exterior de Ciudad Universitaria
Apdo. Postal 70-305
Mexico, D F 04510
E-mail address; pindaro@maricmyl-unam.mx
Manuel Uribe-Alcocer
Instituto de Ciencias del Mar y Limnologia
Universidad Nacional Autonoma de Mexico
Circuito exterior de Ciudad Universitaria
Apdo, Postal 70 305
Mexico, D.F 04510
Manuscript approved for publication
19 June 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:769-777.
Yellowfin tuna (Thunnus albacares) is a
cosmopolitan species inhabiting tropical
and subtropical waters in the Atlantic,
Pacific, and Indian oceans. This species
has accounted for more than a third of
the world's tuna production since 1970.
The eastern Pacific has contributed
from 21% to 26% of the global catch
from 1993 through 1997, represent-
ing 273,329 metric tons (t) in 1990 to
264,426 t in 1998 (lATTC, 1999).
Yellowfin tuna is a large pelagic fish
with a common size of 150 cm (Collette
and Nauen, 1983). Spawning occurs
throughout the year in the tropical
oceans, preferably near islands and
coasts (Leis et al., 1991). Growth is
rapid and individuals reach maturity
by the end of the second year (Suzuki
et al., 1978). Schooling of individuals
of similar size is observed near surface
waters and is often associated with
floating objects (Wild, 1994).
Yellowfin tuna is currently consid-
ered to comprise a single species (Gibbs
and CoUete, 1967), although significant
morphometric and meristic differences,
limited fish movements, and differences
among catch data, have been reported
for the different regions of the Pacific
Ocean (Godsil and Greenhood, 1951;
Schaefer, 1955; Joseph et al., 1964;
Suzuki et al., 1978; Schaefer, 1991).
Population structure in yellowfin tuna
has been addressed in the Pacific Ocean
by using several independent methods.
Morphometric and meristic based stud-
ies have shown significant differences
(Godsil and Greenhood, 1951; Schaefer,
1955; Kurogane and Hiyama, 1957),
and at least three stocks or discrete
units (western, central, and eastern Pa-
cific) have been proposed. More recent
studies using morphometric multivari-
ate analysis suggest the presence of dif-
ferent stocks between north and south
regions in the eastern Pacific (Schaefer,
1991), as well as across the Pacific
Ocean (Schaefer, 1991). Additionally,
differences in larval distribution, catch
rates, and size composition data of yel-
lowfin tuna caught along the equatorial
Pacific by longline and purse-seine have
been used by Suzuki et al. ( 1978) to dis-
tinguish between western, central, and
eastern Pacific groups.
Tagging experiments have shown
limited movement of yellowfin tuna
between western and eastern Pacific
waters (Joseph et al., 1964; Fink and
Bayliff, 1970). In the eastern Pacific,
the presence of two groups has been
suggested: a northern group off Baja
California coast and the Revillagigedo
Islands and a southern group from the
Maria Islands through Chile. Some
mixing occurs between them (Fink
and Bayliff, 1970). There seem to be
marked movements between north
and south groups along the coast with
limited westward movements (Joseph
et al., 1964).
770
Fishery Bulletin 101(4)
Some studies of population structure using genetic anal-
yses have not revealed the presence of discrete stocks along
the Pacific Ocean. Barret and Tsuyuki (1967) used transfer-
rin analysis and did not find differences in allele frequen-
cies between samples from Hawaii and eastern Pacific
samples, although heterogeneity was detected within the
eastern Pacific samples (lATTC, 1975). Allozyme variation
studies in the esterase locus (Fujino, 1970) did not show
enough evidence of genetic differentiation between east-
ern Pacific and Hawaii samples. Furthermore, Scoles and
Graves (1993) used restriction fragments length polymor-
phisms (RFLP) and analysis of mitochondrial (mt) DNA to
examine five samples collected across the Pacific Ocean and
one from the Atlantic Ocean. Although they detected 34
haplotypes and considerable genetic variation, no evidence
of genetic differentiation among samples was found.
However, more recent genetic studies have provided lim-
ited evidence of genetic heterogeneity. Ward et al. (1994)
analyzed four polymorphic allozyme loci and 18 mtDNA
haplotypes in yellowfin tuna from the Pacific Ocean. Al-
though no unique haplotypes were found in the analyzed
populations through RFLP analysis, the eastern Pacific
samples were found to be different from the central and
western Pacific samples in frequency differences at a sin-
gle locus GPI-F*, suggesting that the signal of population
structure exhibited is due to selective factors contributing
to the divergence. Eastern Pacific samples (n=41) were
collected in the northeast Pacific off California and at an
unspecified site off Mexico (n=40). Comparisons of GPI-F*
allele frequencies from eastern Pacific also included two
samples previously analyzed by Sharp (1978) from Roca
Partida (Central America) and Ecuador. Their results
showed population homogeneity at the GPI-F* locus for
this region.
To date, the methods and logistics used to study diver-
gence in the Pacific yellowfin tuna have been focused on
a global rather than a local scale, and sampling has been
focused on the wide areas of the west and central Pacific.
Local structure in the eastern Pacific yellowfin tuna has
not been addressed through a more intense sampling
strategy to examine genetic homogeneity in this region.
Because tagging studies have shown restricted longitudi-
nal movements by yellowfin tuna, population structure and
isolation by distance hypotheses can be tested. To evaluate
the stock structure of yellowfin tuna in eastern Pacific, we
employed analyses of allozymes and of randomly amplified
polymorphic DNA (RAPDs).
RAPDs have proven to be useful genetic markers because
of their high levels of polymorphisms (Williams et al.,
1990; Welsh et. al., 1991). They have been used to estimate
population structure in fishes, including the cod (Kenji,
1998), red mullet (Mamuris et al., 1998), and striped bass
(Bielawski and Pumo, 1997). The use of RAPDs, considered
as neutral markers, and the simultaneous use of allozyme
analyses with intense sampling in a more local area, might
provide evidence about the relationship between gene flow
and spatial distribution of the eastern Pacific yellowfin
tuna, as well as evidence of the presence of local selective
factors responsible for the divergence suggested by Ward
etal.(1994).
Materials and methods
Sampling
A total of 327 tissue samples from specimens of ten loca-
tions were obtained from commercial tuna boats fishing
in the tropical eastern Pacific from 1994 to 1996 (Fig. 1).
Muscle tissue samples were dissected from specimens at
the time of landing and were transported in liquid nitrogen
or on dry ice to the Laboratorio de Genetica de Organismos
Acuaticos of the Instituto de Ciencias del Mar y Limnologia
in Mexico City. Samples were maintained at -70°C until
processing.
Allozyme analysis
For allozyme analysis, 1 cm-* (about one gram) of tissue
sample was ground with a manual homogenizer in 1.5
mL of extraction buffer (O.OIM Tris-O.OOIM EDTA, pH
6.8, and 1% NADP) and centrifuged at 2500 g at 4°C.
Electrophoretic runs were performed in 12% (w/v) starch
gels ( Sigma Chemicals, St. Louis, MO ). Four buffer systems
were used to analyze nineteen enzymes that resolved 28
loci, eight of which showed polymorphism: Aat-S* (aspar-
tate aminotransferase), Glud (glutamate dehydrogenase),
Gpi-F* and Gpi-S* (glucose phosphate isomerase). La
(leucil-L-alanine), Lgg (L-leucil-glycil-glycine), Pap-F*
(L-leucil-L-proline) and 6-Pgd (phosphogluconate dehy-
drogenase). Enzymes AK (adenilate kinase), CK (cre-
atinine kinase), GAPDH (glyceraldehyde-3-phosphate
dehydrogenase), LDH (lactate dehydrogenase), MDH
(mMalate dehydrogenase), ME (malic enzyme) and SOD
(superoxide dismutase), displayed twenty more loci that
were presumably monomorphic. Buffer systems for enzyme
analysis were 1) amino-citrate: 0.04M citric acid, 15mL/L
of N-3-aminopropyl-morpholine, pH 6.5 (AAT, GPI, and
LA); 2) 0.008M Tris, 0.003 M citric acid, pH 6.7 (GLUD
and 6-PGD); 3) 0.025 M Tris, 0.192 glycine, pH 8.5 (GPI
and LGG); 4) 0.0.076 M Tris, 0.005 M citric acid, pH 8.7
(PAP). Enzyme assays were performed following Harris
and Hopkinson (1976). Enzymes showing polymorphism
were analyzed for all samples and subjected to population
genetic analysis.
RAPD analysis
For RAPD analyses, genomic DNA was extracted from
muscle tissue by using standard phenol-chloroform proto-
cols (Sambrooketal., 1989), resuspendedinTE buffer (lOmM
Tris-O.lmM EDTA pH 8.0), and quantified with a Hoefer
DyNA quant 200 fluorometer. DNA was amplified with
primer F-10 (Operon® Alameda, CA; 5'-GGAAGCTTGG-3').
Amplifying reactions were performed in a final volume of
22 //L consisting of 0.7 to 1 ng///L of DNA in amplification
buffer, 10 mM Tris-HCl, 50 niM KCl, 1.5 mM MgCl.^, 33
ng of primer, 10 mM dNTPs, and 1 U of Taq polymerase.
Amplification of genomic DNA was performed in a Perkin
Elmer®, Foster City, CA (mod. 480), thermal cycler. The pro-
gram was set for 1 cycle of 1 min. at 36°C, followed by 44
cycles of 1 min. at 36°C; 1 min. at 94°C; 2 min. at 72°C, and
DIaz-Jaimes and Uribe-Alcocer; Allozyme and RAPD variation in Thunnus albacares
771
126°
119°
Date (Nl
OCY
Jun94 (42)
WRE
Jun94 (49)
NCL
Aug96(6)
SCL
Feb9S(14)
CUE
Feb9f) (50)
MCH
Feb96(2())
COL
Feb96(30)
NAY
Apr94(25)
CSL
May94 (40)
GC
May94(51)
28°
21°
7° -
OCY
"O
WRE
€
126°
19°
Revillagigedo ,"
Islands
Gpi- F alleles
700
Figure 1
Thunnus albacares. Location of sampling sites, dates, and sample sizes, and Gpi-F* gene frequencies of eastern Pacific
yellowfin tuna. The sites are as follows: OCY = out of CYRA area; WRE = west of the Revillagigedo Islands; NCL =
North Clipperton islands; SCL = South Clipperton Islands; MCH = Michoacan; QUE = Guerrero; COL = Colima; CSL =
Cape San Lucas; NAY = Nayarit; GC = Gulf of California. Symbols are • coastal, A intermediate, ■ offshore.
a final cycle of 15 min. at 72°C. Optimal DNA concentra-
tions for amplification were determined by testing several
dilutions, one of which was taken as the standard for every
subsequent amplification.
Amplified fragments were resolved by electrophoresis in
1.5% agarose gels (Sigma Chemicals) for 3 to 4 h. at 90 mA
(100 V). A lOObp DNA Ladder (GibcoBRL, Gaithersburg,
MD, 15628-019) was used as size standard. After electro-
phoresis, gels were stained with ethidium bromide and
photographed in a UV light transilluminator.
Data analysis
Allelic frequencies, test of conformity of genotype distribu-
tions with Hardy-Weinberg, and heterozygous deficit were
determined by using Genepop version 3.3 (Raymond and
RoussetM. Homogeneity of allozyme and RAPD allele fre-
quencies was evaluated by using the exact probability test
(Raymond and Rousset, 1995) consisting of a contingency
analysis for every polymorphic locus and an estimation
of their probability values by the combined probability
of Fisher (Sokal and Rohlf, 1995) as implemented in the
TFPGA program (Miller^). Pairwise comparisons were
conducted to determine allele frequency differences among
samples in order to define sources of variation. Based on
the longitudinal differentiation pattern observed by Ward
et al. (1994) and the morphological latitudinal differences
within eastern Pacific samples reported by Schaefer (1991)
1 Raymond, M. L., and F. Rousset. 1995a. GENEPOP (ver-
sion 1.2): population genetics software for exact tests and
ecumenicist. J. Heredity 86:248-249.
2 Miller, M. P. 1997. Tools for genetic populations analyses
(TFPGA) 1.3: a windows program for the analysis of allozyme
and molecular population genetic data, 29 p. Computer
software distributed by the author at http://bioweb.usu.edu/
mpmbio.
772
Fishery Bulletin 101(4)
at north-south of the 15-20°N range, spatial homogene-
ity was tested at the following levels: overall samples (O),
among longitudinal regions (L; coastal, intermediate, and
offshore localities), and latitudinal regions (N; north-south
of the 15-20° range). The Gulf of California sample was
excluded from this analysis because the large variation
found in this sample (reflected in the significant differ-
ences shown in allele frequency homogeneity from pairwise
comparisons) would not allow an accurate assessment of
whether longitudinal or latitudinal differentiation exists or
not. Significance levels were adjusted for multiple testing
through the Bonferroni sequential method (Rice, 1989).
Population subdivision was estimated by using the
Weir and Cockerham (1984) method through the TFPGA
program. Standard error and confidence intervals were
obtained through jackknife and bootstrapping procedures,
respectively, with Fgj, Pro 1.0 (Weir, 1990). Estimates of pop-
ulation subdivision were partitioned into the following lev-
els: O, over all the samples; L, longitudinal regions (coastal,
intermediate and offshore); and N, latitudinal regions.
We used the 6 statistic to estimate gene values between
sample pairs (Slatkin, 1993) that are defined as Mg. An
"isolation by distance" model was evaluated from the cor-
relation between the distance between localities measured
as geographic separation in nautical miles (nmi), and the
M(j values by means of the Mantel test (Hellberg, 1994) in
both allozymes and RAPDs.
The patterns of the amplification products resulting from
the RAPD analysis were subjected to the same analyses
as allozymes with the procedures described in Lynch and
Milligan (1994). RAPD fragments were interpreted under
the following assumptions: 1) fragments were considered
to behave as dominant genes (Williams et al., 1990); 2)
every polymorphic fragment was considered derived from
a two-allele locus; 3) the equilibrium of Hardy- Weinberg
was assumed for all genotypes, and 4) each fragment was
considered to be an independent locus.
Only those fragments clearly defined and having consis-
tent intensity were recorded. Because of this, the Michoacan
sample, with poor consistency in the banding patterns, was
excluded from the RAPD analysis. The allele frequency of
every fragment was calculated on the basis of the inferred
homozygous recessive genotypes. Because of the dominant
nature of the alleles, and in order to correct the bias originat-
ed by calculating the recessive allelic frequencies, we chose
the estimation based on the Taylor expansion (Kendall and
Stuart, 1977, cited m Le Corre et al., 1997) as implemented
in TFPGA program. This reduction on the bias is based on
the equation resulting from the second order expansion
of Taylor (see details in Lynch and Milligan, 1994).
Results
Allozymes
eight polymorphic loci detected are shown in Table 1. After
adjusting levels of significance by the Bonferroni procedure,
significant deviation of genotypic frequencies from those
expected under Hardy-Weinberg was found in the loci Lgg
and Pap-F* for the Gulf of California sample and in the Aat-
S* locus for two localities — west of Revillagigedo Islands,
and the Gulf of California (P<0.0006, Table 1). Deviations
displayed for both locations corresponded to a heterozygous
deficit (P<0.0006, after Bonferroni correction).
Comparison of allozyme allele frequencies among all
collections (overall) by the exact probability test revealed
significant heterogeneity at loci Glud, La, and Lgg, after
Bonferroni adjustment (P<0.006). Pairwise comparisons
among samples to test allele homogeneity gave signifi-
cant differences for nine of 45 comparisons after correc-
tion for multiple tests (P<0.001), seven of which involved
comparisons with the Gulf of California (GC) sample. The
remaining significant differences were between Guerrero-
Nayarit, and Cape San Lucas-Nayarit comparisons (Table
2), resulting from significant heterogeneity at Glud, La,
and Lgg loci.
In general, allozyme analysis displayed low levels of dif-
ferentiation. The 9 value over all loci was different from
zero (P<0.05) and showed that 4.8% of the variance was
attributable to differences among samples (Table 3).
Individual loci showed d values ranging between 0.0037
and 0.27. Highly significant values at loci La (9=0.13
±0.089) and Lgg (0=0.27 ±0.253) evidently resulted from
their weak polymorphism in some samples (Table 3).
Allele frequency homogeneity was tested among coastal,
intermediate, and offshore regions (Table 4). Significant
heterogeneity was found by exact test between coastal
and offshore comparisons (P=0.0043) but was found to be
not significant between coastal and intermediate regions
(P=0.0632). Subdivision as measured by 0 among coastal,
intermediate, and offshore localities was not different from
zero. However, for the Gpi-F* locus the population subdivi-
sion among regions (0.0058) was twice as large as that noted
among samples (0=0.003), but neither value was significant
(Table 3). No latitudinal differentiation by the exact test or
population subdivision estimated by the 9 index was found
between north and south regions (data not shown).
The gene flow values (Mg) were high (mean 24.8 mi-
grants per generation). A lack of correlation between gene
flow estimations and geographic distance by means of the
Mantel test was observed (r'-=-0.144; P=0.22), resulting in
a rejection of the isolation by distance model.
For the Gpi-F* locus, paired tests of significance (data
not shown) showed discrepancies in the Gpi-F* 175 allele
frequencies among localities from the coast with those
located at the CYRA limits (Fig. 1, Table 5). Four of eight
comparisons of coastal and offshore samples revealed
differences of allele frequencies at this locus (P<0.05),
although none of these differences was significant after
correction for multiple testing (P>0.001).
Of the 28 analyzed allozyme loci, eight (28.5%) showed
polymorphism under the 0.95 criterion. The observed het-
erozygosities per sample over all allozyme loci ranged from
0.027 to 0.083 (mean 0.052). Allozyme frequencies for the
RAPDs
The primer OPF-10 produced 11 amplified fragments, with
sizes from 200 to 600 bp (base pairs). Four of the fragments
Diaz-Jaimes and Unbe-Alcocer: Allozyme and RAPD variation in Thunnus albacares
773
Table 1
Allele frequencies for allozymes and RAPDs, samples size (n) and agreement to the Hardy-Weinberg equilibrium (HW) for every
loci and sample of Thunnus albacares. Significance values for HW tests were adjusted for multiple comparisons with an initial a
level of 0.0006 ((0.05/(8 loci x 10 samples)]. P = probability of significance for allele frequency heterogeneity per locus, * = significant
at a = 0.006. — = no data. OCY = out of CYRA area; WRE = west of the Revillagigedo Islands; NCL = North Clipperton islands;
SCL = South Clipperton Islands; MCH = Michoacan; GUE = Guerrero; COL = Colima; CSL = Cape San Lucas; NAY = Nayarit; GC =
Gulf of California.
Collection
Locus
Aat-S*
Glud
Gpi-F*
Gpi-S*
La
Lgg
Pap-F*
6-Pgd
FIO-1
FlO-2
FlO-3
FlO-4
Allele OCY
WRE
NCL
SCL
GUE
MCH
COL CSL
NAY
-90
0.171
-100
0.829
n
41
HW
yes
100
0.700
85
0.300
n
40
HW
ves
135
0.071
100
0.262
75
0.667
n
42
HW
yes
-60
0.402
-100
0.598
n
41
HW
yes
120
0.0
100
1.0
n
42
HW
yes
115
0.0
100
1.0
n
42
HW
yes
110
0.0
100
1.0
n
42
HW
—
100
0.902
90
0.098
n
41
HW
yes
a
b
n
HW
a
b
n
HW
a
b
n
HW
a
b
n
HW
0.406
0.594
9
0.406
0.594
9
0.406
0.594
9
0.115
0.885
9
0.266
0.734
32
no
0.634
0.366
41
yes
0.134
0.280
0.586
41
yes
0.536
0.464
42
yes
0.024
0.976
42
0.024
0.976
42
0.0
1.0
42
0.903
0.097
31
yes
0.282
0.718
8
0.603
0.397
0.475
0.525
0.282
0.718
8
0.250
0.750
6
yes
0.667
0.333
6
yes
0.250
0.250
0.500
6
yes
0.417
0.583
6
yes
0.0
1.0
6
0.0
1.0
6
0.0
1.0
6
0.667
0.333
6
yes
0.234
0.766
7
0.441
0.559
7
0.329
0.671
7
0.234
0.766
7
0.231
0.769
13
yes
0.857
0.143
14
yes
0.091
0.500
0.409
11
yes
0.571
0.429
14
yes
0.0
1.0
14
0.0
1.0
14
0.0
1.0
14
0.821
0.179
14
yes
0.322
0.678
9
0.322
0.678
9
0.625
0.375
9
0.178
0.822
9
0.154
0.846
39
yes
0.663
0.337
49
yes
0.190
0.230
0.580
50
yes
0.360
0.640
50
yes
0.0
1.0
50
0.0
1.0
50
0.02
0.98
50
yes
0.939
0.061
49
yes
0.275
0.725
20
yes
0.775
0.225
20
yes
0.079
0.316
0.605
19
no
0.425
0.575
20
yes
0.0
1.0
20
0.0
1.0
20
0.05
0.95
20
yes
1.0
0.0
20
0.083
0.917
6
0.278
0.722
6
0.397
0.603
6
0.083
0.917
6
0.077
0.923
26
yes
0.914
0.086
29
yes
0.185
0.370
0.444
27
yes
0.414
0.586
29
yes
0.0
1.0
24
0.0
1.0
24
0.0
1.0
29
0.966
0.034
29
yes
0.505
0.495
9
0.092
0.908
38
yes
0,788
0.212
40
yes
0.125
0.363
0.512
40
yes
0.449
0.551
39
yes
0.0
1.0
40
0.0
1.0
24
0.0
1.0
40
1.0
0,0
40
0.355
0.645
10
0.322 0.355
0.678 0.645
9 10
0.406 0.529
0.594 0.471
9 10
0.246 0.219
0.754 0.781
9 10
0.0
1.0
25
0.680
0.320
25
yes
0.160
0.260
0.580
25
yes
0.479
0.521
24
yes
0.1
0.9
20
yes
0.0
1.0
24
0.0
1.0
25
0.840
0.160
25
yes
0.282
0.718
8
1.0
0.0
0.371
0.629
0,138
0.862
8
GC
0.234
0.766
47
no
0.395
0.605
43
yes
0.200
0.314
0.486
35
yes
0.338
0.663
40
yes
0.2
0.8
50
yes
0.349
0.651
43
no
0.083
0.917
48
no
0.938
0.062
48
yes
0.246
0.754
9
0.322
0.678
9
0.406
0.594
9
0.246
0.754
9
0011
0.526
0.015
0.055
0.007
0.572
0.369
0.912
0.929
774
Fishery Bulletin 101(4)
Table 2
Painvise-sample comparisons of allele frequency homogeneity for Thunnus albacares. Probability values in allozymes are above
the diagonal { — ) and RAPDs are below the diagonal ( — ). * = significant values after Bonferroni correction for multiple tests (initial
a was 0.05). na = data not available. OCY = out of CYRA area; WRE = west of the Revillagigedo Islands; NCL = North Clipperton
islands; SCL = South Clipperton Islands; MCH = Michoacan; GUE = Guerrero; COL = Colima; CSL = Cape San Lucas; NAY =
Nayarit; GC = Gulf of California.
Sample
OCY
WRE
NCL
SCL
GUE
MCH
COL
CSL
NAY
GC
OCY
0.5807
0.7391
0.3414
0.4386
0.8254
0.0295
0.1525
0.0194
*<0.001
WRE
0.9316
—
0.9482
0.6047
0.4330
0.1180
0.0012
0.0084
0.0192
*<0.001
NCL
0.9797
0.9770
—
0.9232
0.3494
0.7307
0.1567
0.0859
0.4779
0.0174
SCL
0.9798
0.8088
0.8915
—
0.3989
0.0717
0.3414
0.1093
0.0287
*<0.001
GUE
0.5935
0.4732
0.9095
0.7757
—
0.4137
0.1723
0.5932
*<0.001
*<0.001
MCH
na
na
na
na
na
—
0.0500
0.1532
0.0096
*<0.001
COL
0.9385
0.7864
0.9155
0.9794
0.4569
na
—
0.8237
0.0082
*<0.001
CSL
0.9475
0.9703
0.9824
1.000
0.6576
na
0.9849
—
0.0012
*<0.001
NAY
0.9656
0.9984
0.8418
0.4965
0.4101
na
0.4206
0.5707
—
*<0.001
GC
0.9458
0.9637
0.9987
0.9977
0.9743
na
0.9715
0.9874
0.6518
—
Table 3
Estimates of population subdivision 9 (Weir and Cocker-
ham, 1984) for allozymes of Thunnus albacares partitioned
into longitudinal regions 0^ (regions) (i.e. coastal-interme-
diate-offshore) and samples 9^ (overall), n.v. = negative
values. P = probability of significance of subdivision esti-
mations. Significance of single-locus values was corrected
with an initial level of 0.006 (0.05/8 loci). Means and stan-
dard error were obtained by the jackknife method. Con-
fidence intervals obtained by 1000 resamplings through
bootstrapping are also shown. * = signifcant values after
Bonferroni correction.
Locus
L (regions)
"O(overall)
Aat-2*
Glud
Gpil*
Gpi-2*
La
Lgg
Pap-1*
6-Pgd
Mean
CI 95%
0.0152
0.0009
0.0058
0.0014
n.v.
0.0073
n.v.
0.044
0.0067 ±0.0042
0.0003-0.0199
0.072
0.227
0.086
0.780
0.012
0.004
0.068
>0.006*
0.029
0.086
0.003
0.0037
0.13
0.27
0.024
0.02
0.048 ±0.022
0.019-0.101
0.007
>0.006*
0.210
0.007
>0.006*
>0.006
0.001
>0.006*
were polymorphic for all samples (Table 1). No significant
heterogeneity of RAPD allele frequencies was found for any
locus between any paired sample comparison, among all
collections, nor among latitudinal or longitudinal regions
(P=0.4806).
The mean lvalue for all fragments and samples (over-
all), as well as regional estimations derived from liAPDs
(0.0302), were not significantly different from zero and
displayed some negative values. Klstimations of gene
flow between sample pairs (Mg) from RAPU data aver-
Table 4
Pairwise-regions comparison of allele frequencies for
Thunnus albacares. Probabilities of nonheterogeneity for
allozymes (based on exact tests) are above the diagonal
( — ) and RAPDs are below the diagonal ( — ). * = significant
after corrected for multiple tests (Rice, 1989).
Region
Coastal
Intermediate
Offshore
Coastal
Intermediate
Offshore
0.9998
0.9039
0,0632
0.9126
0.0043*
0.5384
aged 29.2 migrants per generation. The evaluation of the
relationships between geographic distances and the gene
flow estimations in pairwise collections (Afg), through the
Mantel test, showed a nonsignificant correlation (r-=0.413,
P=0.984).
Discussion
The test of conformance to the Hardy-Weinberg frequen-
cies showed significant differences in Lgg and Pap-F* loci
only in the Gulf of California sample, where polymorphism
at those loci was also consistently found. Similar results
were obtained, with smaller differences in locus Aat-S*
from the west of the Revillagigedo Islands and the Gulf of
California samples. Considering the fact that our samples
were provided by the commercial fleet, they could have
included representatives of different schools with differ-
ences in genotypic distributions originated by differences
in age classes or sexual ratios (or for both) among schools
because recruitment of individuals into new schools has
been reported to be mainly by aggregating individuals of
Diaz-Jaimes and Uribe-Alcocer: Allozyme and RAPD variation in Thunnus albacares
775
Table S
Comparison
31,(1994).—
of Gpi-F* allele frequencies for Thunnus
= data absent.
albacares among
data from the present study and those reported
in Ward et
Locus
Allele
Western/Central'
Eastern'
Eastern (present
data)
Offshore
Intermediate
Coastal
Pooled
Gpi-F*
135
0.026
0,100
0.103
0.147
0.163
0.145
100
0.640
0.269
0.301
0.412
0.301
0.307
75
0.332
0.631
0.596
0.441
0.566
0.548
40
0.002
—
—
—
—
—
n
346
178
83
17
196
296
' Allele frequ
-ncies for Gpi-F* reported in Ward et al, ( 1994)
similar sizes (Collette an(i Nauen, 1983). Because recruit-
ment to the original tuna schools has been reported as
well (Kimley and Holloway, 1999), random processes could
also induce differences in genotypic frequencies that favor
aggregation of some genotypes, while segregating some
others, causing a kind of Wahlund effect that is reflected by
a heterozygous deficit as shown by the homozygous excess
for loci and locations having HW deviations, especially as
shown in the Gulf of California sample.
The estimations of population structure based on allo-
zymes showed a small but significant value different from
zero (6=0.048; P<0,01). The Gulf of California sample con-
tributed to the significant subdivision value as shown when
that collection was excluded from the regional subdivision
analysis, as well as to significant heterogeneity of its allele
frequencies when paired comparisons were made.
The small value of 0 for overall estimations on RAPD
data is probably due to the small sample size. The nega-
tive values of 6 from overall and regional estimations re-
sulted from subtracting the large value of the correction
derived from the variation expected of the sample sizes
from the small value of variation due to fluctuations in al-
lele frequencies. The fact that RAPD data are considered
dominant could reduce information about the true allele
distributions by subestimation of null allele frequencies
notwithstanding the correction applied to recessive geno-
types, which is dependent on the sample sizes (Lynch and
Milligan, 1994). Other assumptions for RAPD data limit
the value of this marker, especially when estimations are
derived from a small number of loci and sample sizes. Ad-
ditional constraints are related to the limited number of
alleles (two) to estimate dominant markers, which tend
to subestimate the polymorphism and thus reduce the
significance of relatively small discrepancies in allele
distributions.
No differentiation between coastal and offshore samples
was found in our study because of the slight, nonsignifi-
cant differences in the estimation of the subdivision by
regions. Although the overall estimation was not different
from zero, the allele homogeneity analysis showed allele-
frequency heterogeneity between coastal and offshore
samples, and nonheterogeneity between coastal and inter-
mediate samples.
These results are consistent with the migration reports
through tagging studies; evidence exists for the presence
of two main yellowfin tuna groups in the eastern Pacific
that mix to some extent (Fink and Bayliff, 1970) and that
migrate longshore from around the 20°N to the mouth of
the Gulf of California and to the zone between the Revilla-
gigedo and the Clipperton islands, and back again (Joseph
et al, 1964; Fink and Bayliff, 1970), although longitudi-
nal movements are restricted to the limits of yellowfin
regulatory area (CYRA). Similarly, important northward
movements along the coasts to the mouth of the Gulf of
California, and subsequently to the western coasts of Baja
California, have been reported. Although the estimation of
6 for allozymes showed a significant value, it was notably
influenced by the heterogeneity found between the Gulf of
California sample and all other samples.
Discarding the variation displayed by loci La, Lgg,
and Pap-F*, originating mainly from Gulf of California
sample, the estimation of subdivision was still marginally
significant after Bonferroni correction, which should be
considered as evidence that the Gulf of California sample
may represent a partially isolated population with differ-
ent allele frequencies. Oceanographic conditions inside the
Gulf are somewhat different from those of the Pacific Ocean
where there are warmer waters at the end of the year, es-
pecially during yellowfin tuna spawning seasons. There is
also high productivity characterized by the presence of sig-
nificant biomass abundance of sardine or anchovy schools
(Cisneros-Mata et al., 1995), which represents opportuni-
ties to establish the feeding and consequently the spawn-
ing grounds for eastern Pacific yellowfin tuna. Likewise,
there is a trend of migratory movements through the Gulf
of California by different groups of yellowfin tunas (Fink
and Bayliff, 1970). These movements promote stock mix-
ing and help to explain the wide polymorphism displayed
in this sample, in contrast to the weak variation found in
other samples from the coast and offshore regions. Further
genetic research, including sequential temporal sampling
of young fishes in order to ensure the presence of individu-
776
Fishery Bulletin 101(4)
als that originated in discrete spawning grounds, should be
undertaken to prove the presence of an independent unit
inside the Gulf of California, which, if confirmed, might
necessitate new stock management strategies.
Allele frequencies for Gpi-F* locus found in the present
study, apparently, correspond to those reported by Ward et
al. (1994 and 1997). These authors reported a higher pro-
portion of the allele Gpi-F*75 (0.571) in the eastern Pacific
region and a gradual decrement of the same allele toward
the central (0.423) and western Pacific regions (0.330),
where allele Gpi-F*100 (0.650) had the higher proportion.
In the present study, the highest frequencies for the allele
Gpi-F*75 corresponded to the region of the eastern Pacific,
situated in the limits of the yellowfin tuna regulatory area
(offshore region), and there was a slight decrease in fre-
quencies towards the coastal area (Table 5). Furthermore,
allele frequencies for the Gpi-F*75 allele from the coastal
locality, Colima (0.444), and the intermediate locality
southeast, Clipperton Islands (0.409), have coincidences
with those reported by Ward et al. ( 1994) for the collection
Hawaii 92 (0.423) in the central Pacific region.
The similarities in the Gpi-F* allelic frequencies be-
tween eastern (Colima and Clipperton) and central Pacific
samples (Hawaii 92) might possibly be attributed to the
extended migrations of yellowfin tuna in the eastern Pa-
cific brought about by the strong influence of warm waters
on tuna movements because of the increased depth of the
thermocline layer in that area, which was reflected by a
decrease in catches (Joseph and Miller, 1988; Wild, 1994)
and which possibly led to the mixing of the eastern and
central Pacific stocks.
The low number of RAPD loci analyzed and the uncer-
tainty of fulfilling some assumptions, such as the genetic
identity of each band needed for qualitative and quantita-
tive interpretation of data in terms of allelic frequencies,
do not allow us to consider our estimations of subdivision
reliable with the RAPD method. Additionally, the lack of
reliability of estimations associated with high sampling
variances by using randomly collected fishery samples
highlights the need to design more efficient spatial and
temporal sampling strategies in local and wide areas, as
well as the need for alternative hypervariable markers to
assess the divergence patterns observed in highly migra-
tory species.
Acknowledgments
We are grateful to Ernesto Escobar from Pescados Industri-
alizados S. A. PINSA for allowing the sampling, and Robert
Olson from lATTC for providing the Gulf of California
samples. We thank Monica Dominguez-Lopez, Yolanda
Hornelas-Orozco, Evangelina Castillo, and Alma Hernan-
dez-Perez, for collection and processing of samples, Luis
Eguiarte and Valeria Souza for the facilities provided in
their laboratory and three anonymous reviewers for their
valuable comments. This manuscript benefited from the
critical reading of John Graves, Jan McDowell, and Bar-
bara Rutan. Funding for this project was provided by the
Programa de Apoyo a Estudiantes de Posgrado (PADEP)
and by the project IN20598 of the Programa de Apoyo
a Proyectos de Investigacion e Innovacion Tecnologica.
Direccion General de Asuntos del Personal Academico,
Universidad Nacional Autonoma de Mexico.
Literature cited
Barret, I., and H. Tsuyuki,
1967. Serum transferrin polymorphism in some scombroid
fishes. Copeia 1967:551-.557.
Bielawski, J. P. and D. E. Pumo.
1997. Randomly amplified polymorphic DNA ( RAPD ) analy-
sis of Atlantic coast striped bass. Heredity 78:32^0.
Cisneros-Mata, M. A., Nevarez-Martinez, M. O., and
Hammann, M. G.
1995. The rise and fall of the Pacific sardine, Sardinops
sagax caeruteus Girard, in the Gulf of California, Mexico.
CALCOFI Rep. 36:136-143.
Collette, B. B., and C. E. Nauen.
1983. Scombrids of the world: an annotated and illustrated
catalogue of tunas, mackerels, bonitos and related species
known to date. In FAO species catalogue, vol. 2, 137 p.
FAO Fish. Synop. 1325. FAO, Rome.
Fmk, B. D., and W. H. Bayliff.
1970. Migrations of yellowfin and skipjack tuna in the east-
ern Pacific Ocean as determined by tagging experiments,
1952-1964. Inter-Am. Trop. Tuna Comm. Bull. 15:1-227.
Fujino, K.
1970. Immunological and biochemical genetics of tunas.
Trans. Am. Fish. Soc. 99:152-178.
Gibbs, R. H., and B. B. Collette.
1967. Comparative anatomy and systematics of the tunas,
genus Thunnus. Fish. Bull. 86:835-838.
Godsil, H. C, and E. C. Greenhood.
1951. A comparison of the populations of yellowfin tuna
(Neothunnus macropterus) from the eastern and central
Pacific, 33 p. Calif Dep. Fish Game Fish. Bull. 82.
Harris, H., and D. A. Hopkinson.
1976. Handbook of enzyme electrophoresis in human genet-
ics, 273 p. Am. Elsevier Publishing Co., New York, NY.
Hellberg, M. E.
1994. Relationships between inferred levels of gene flow and
geographic distance in a philopatric coral, Ballanophyllia
elegans. Evolution 48:1829-1854.
lATTC (Inter- American Tropical Tuna Commission).
1975. Annual report of the Inter- American Tropical Tuna
Commission, 169 p. lATTC. LaJolla, CA.
1999. Annual report of the Inter-American Tropical Tuna
Commission, 357 p. lATTC, LaJolla, CA.
Joseph, J., F G. Alverson, B. D. Fink, and E. B. Davidoff.
1964. A review of the population structure of yellowfin tuna,
Thunnus albacares. in the eastern Pacific Ocean. Inter-
Am. Trop. Tuna Comm. Bull. 9(2):53-112.
Joseph, J., and F. R. Miller
1988. El Nino and the surface fishery for tunas in the east-
ern Pacific. Proceedings of the Tuna Fisheries Research
Conference, Japan Fish. Far Seas Fish. Res. Lab. Maguro
Gyogyo Kyogikai Gijiroku, Suisancho-Enyo Suisan Kenky-
usho: 199-207
Kcnji, S,
1998. Genetic variation and local differentiation in the
Pacific cod Gadua macrocephalus around Japan revealed
by mtDNA and RAPD markers. Fish. Sci. 64(5):673-679.
Diaz-Jaimes and Unbe-Alcocer: Allozyme and RAPD variation in Thunnus a/bacares
777
Kimley, A. P., and C. F. Holloway.
1999. School fidelity and homing synchronicity of yellowfin
tuna, Thunnus albacares. Mar. Biol. 133:307-317.
Kurogane, K., and Y. Hiyama.
1957. Morphometric comparison of the yellowfin tuna taken
from the Equatorial Pacific. Jpn. Soc. Sci. Fish. Bull. 23:
388-293.
Leis, J. M., T. Trnski, M. Harmelin-Vivien, J.-P. Renon, V Dufour,
M. K. El Moundi, and R. Galzin.
1991. High concentrations of tuna lai-vae (Pisces: Scombri-
dae) in near-reef waters of french Polynesia (Society and
Taumotu islands). Bull. Mar. Sci. 48(1):150-158.
Le Corre, S., S. Dumolin-Lapegue. and A. Kremer
1997. Genetic variation at allozyme and RAPD loci in sessile
oak Quercus petraea (Matt.) Liebl.: the role of history and
geography Mol. Ecol. 6:519-529.
Lynch, M., and B. G. Milligan.
1994. Analysis of population genetic structure with RAPD
markers. Mol. Ecol. 3:91-99.
Mamuris, Z., A. P. Apostodilis, A. J. Theodoru, and
C. Triantaphyllidis.
1998. Application of random amplified polymorphic DNA
(RAPD) markers to evaluate intraspecific genetic varia-
tion in red mullet {Mullus barbatus). Mar Biol. 132:
171-178.
Raymond, M. L., and F. Rousset.
1995. An exact test for population differentiation. Evolu-
tion 49:1280-1283.
Rice, W. R.
1989. Analyzing tables of statistical tests. Evolution 43:
223-225.
Sambrook, J., E. F. Fritsch, and T. Maniatis.
1989. Molecular cloning: a laboratory manual, 2nd ed., 999 p.
Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Scoles, D. R., and J. E. Graves.
1993. Genetic analysis of the population structure of yellow-
fin tuna, Thunnus albacares, from the Pacific Ocean. Fish.
Bull. 91:690-698.
Schaefer, K. M.
1991. Geographical variation in morphometric characters
and gill-raker counts of yellowfin tuna Thunnus albacares,
from the Pacific Ocean. Fish. Bull. 9:690-698.
Schaefer, M. B.
1955. Morphometric comparison of yellowfin tuna from
southeast Polynesia, Central America, and Hawaii. Inter-
Am. Trop. Tuna Comm. Bull. 1:89-136.
Sharp, G. D.
1978. Behavioral and physiological properties of tunas and
their effects on vulnerability to fishing gear In The physi-
ological ecology of tunas (G. D. Sharp and A. E. Dizon, eds),
p. 397-449. Academic Press, New York, NY.
Slatkin, M.
1993. Isolation by distance in equilibrium and non-equilib-
rium populations. Evolution 47:264-279.
Sokal, R., and F. J. Rohlf
1995. Biometry, 3rd ed., 776 p. W. H. Freeman and Co.,
New York, NY.
Suzuki, Z., P. K. Tomlinson, and M. Honwa.
1978. Population structure of Pacific yellowfin tuna. Inter-
Am. Trop. Tuna Comm. Bull. 17:273-441.
Ward, R. D., N. G. Elliott, P. M. Grewe, and A. Smolenski.
1994. Allozyme and mitochondrial DNA variation in yellow-
fin tuna {Thunnus albacares) from the Pacific Ocean. Mar
Biol. 118:531-539.
Ward, R. D., N. G. Elliott, B. H. Innes, A. J. Smolenski, and
P. M. Grewe.
1997. Global population structure of yellowfin tuna, Thun-
nus albacares, inferred from allozyme and mitochondrial
DNA variation. Fish. Bull. 95:566-575.
Weir, B. S.
1990. Genetic data analysis, 377 p. Sinauer, Sunderland,
MA.
Weir, B. S., and C. C. Cockerham.
1984. Estimating F statistics for the analysis of population
structure. Evolution 38:1358-1370.
Welsh, J., C. Petersen, and M. McCelland.
1991. Polymorphisms generated by arbitrary primed PCR in
the mouse: application to strain identification and genetic
mapping. Nucl. Acids Res. 19:303-306.
Wild, A.
1994. A review of the biology and fisheries for yellowfin
tuna, T. albacares, in the eastern Pacific Ocean. FAO Fish.
Tech. Pap. 336(2):52-107.
Williams, J. G, A. R. Kubelik, J. Livak, J. A. Rafalaski, and
S. V. Tingey
1990. DNA polymorphisms amplified by arbitrary prim-
ers are useful as genetic markers. Nucl. Acids Res. 18:
6531-6535.
778
Abstract— Lengths and ages of sword-
fish (Xiphias gtadius) estimated from
increments on otoliths of larvae col-
lected in the Caribbean Sea, Florida
Straits, and off the southeastern United
States, indicated two growth phases.
Larvae complete yolk and oil globule
absorption 5 to 6 days after hatching
(DAH). Larvae <13 mm preserved
standard length (PSD grow slowly
(-0.3 mm/d); larvae from 13 to 115 mm
PSL grow rapidly (-6 mm/d). The accel-
eration in growth rate at 13 days fol-
lows an abrupt (within 3 days) change
in diet, and in jaw and alimentary canal
structure. The diet of swordfish larvae
is limited. Larvae <8 mm PSL from
the Caribbean, Gulf of Mexico, and
off the southeastern United States eat
exclusively copepods, primarily of one
genus, Corycaeus. Larvae 9 to 11 mm
eat copepods and chaetognaths; larvae
>11 mm eat exclusively neustonic fish
larvae. This diet indicates that young
larvae <11 mm occupy the near-surface
pelagia, whereas, older and longer
larvae are neustonic. Spawning dates
for larvae collected in various regions of
the western North Atlantic, along with
the abundance and spatial distribution
of the youngest larvae, indicate that
spawning peaks in three seasons and
in five regions. Swordfish spawn in the
Caribbean Sea, or possibly to the east,
in winter, and in the western Gulf of
Mexico in spring. Elsewhere swordfish
spawn year-round, but spawning peaks
in the spring in the north-central Gulf
of Mexico, in the summer off southern
Florida, and in the spring and early
summer off the southeastern United
States. The western Gulf Stream fron-
tal zone is the focus of spawning off the
southeastern coast of the United States,
whereas spawning in the Gulf of Mexico
seems to be focused in the vicinity of
the Gulf Loop Current. Larvae may use
the Gulf of Mexico and the outer con-
tinental shelf off the east coast of the
United States as nursery areas. Some
larvae may be transported northward,
but trans-Atlantic transport of larvae
is unlikely.
The early life history of swordfish (Xiphias gladius)
in the western North Atlantic
John Jeffrey Govoni
Elisabeth H. Laban
Jonathan A. Hare
Center for Coastal Fisheries and Habitat Research
National Oceanic and Atmospheric Administration
101 Rivers Island Road
Beauloa North Carolina 28516-9722
E-mail address (for J, J, Govoni): Jeff Govoni(5>noaa gov
Manuscript approved for publication
17 April 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Offfice.
Fish. Bull. 101:778-789 (2003).
Swordfish (Xiphias gladius) live in
warm waters of the world's oceans, as
well as in large enclosed basins such
as the Caribbean and Mediterranean
seas, and the Gulf of Mexico (Berkeley,
1983). Swordfish are highly migratory
throughout their global range. The
worldwide population structure, as
currently understood, has at least three
breeding units: Mediterranean, north-
western Atlantic to the tropical South
Atlantic, and Indo-Pacific (Kotoulas et
al., 1995; Chow and Takeyama, 2000;
Reeb et al., 2000). For the purpose of
fishery management, the International
Commission for the Conservation of
Atlantic Tunas (ICCAT) recognizes
only North Atlantic and South Atlan-
tic stocks. Possible genetic exchange
between eastern and western North
Atlantic populations is incompletely
documented.
Swordfish reportedly spawn year-
round in the western North Atlantic in
different seasons and regions. Spawn-
ing season and location has been
inferred from the abundance of small
larvae (Gorbunova, 1969; Richards
and Potthoff, 1980; Potthoff and Kel-
ley, 1982; Grail et al, 1983; Govoni et
al., 2000), gonad maturation (LaMonte,
1944; Beckett, 1975), or oocyte cytol-
ogy (Taylor and Murphy, 1992; Arocha,
1997; Arocha, 2002). The observation of
live females with running eggs, hooked
on long-lines, and followed to the fishing
vessel by several smaller males (Lee^;
Berkeley''^) corroborates spawning in
some seasons and locations. Although
gonad condition and oocyte status can
indicate spawning season, the resolu-
tion of spawning location can be am-
biguous with these methods because
mature gonads and hydrated oocytes
can be found in several seasons within
the range of these highly migratory
fishes. The determination of age and
the distribution of young larvae, along
with realistic estimates of water veloc-
ity and trajectory, help to resolve this
ambiguity.
Beyond spawning, the early life his-
tory of swordfish in any ocean is incom-
pletely described (Palko et al., 1981).
Larvae undergo a stark change in
physical appearance between ~8 and 13
mm preserved standard length (PSL),
from a typical scombroid larval form to
a juvenile istiophorid one (Collette et
al., 1984). At this juncture in develop-
ment, larvae develop characteristic
preorbital, supraorbital, posttemporal,
and preopercular spines; enlarged and
spinous dorsal, ventral, and lateral
scale anlagen; and a continuous long
dorsal fin that extends along most of
the dorsal aspect. Swordfish retain
these larval characters until they are
at least 188 mm PSL (Arata, 1954; Pot-
thoff and Kelley, 1982), a size at which
most fishes are considered juveniles.
By using the ages and lengths of larvae
hatched in the laboratory and reared
through yolk and oil globule absorp-
tion (Sanzo, 1910; Yasuda et al., 1978),
along with length frequencies of larvae
caught in the western North Atlantic,
' Lee, D. J. 1995. Personal commun.
Southeast Fisheries Science Center,
NMFS, 75 Virginia Beach Drive, Miami,
FL 33149.
'^ Berkeley, S. A. 1998. Personal commun.
Long Marine Laboratory, Univ. California,
Santa Cruz, 100 Shaffer Road, Santa Cruz,
CA 95060.
Govoni et al,: Early life history o\ Xiphias gladius in the western North Atlantic
779
Arata (1954) inferred the age and growth of larvae from 6
to 192 mm preserved total length (PTT). Aside from this
effort, the age of larval swordfish has been undetermined
and growth has not been described. Diets of lai-vae have
been reported (Gorbunova, 1969), but the apparent transi-
tion in diet has neither been detailed nor reconciled with
changes in physical features and growth. Similarly, the
vertical distribution of larvae has not been reconciled with
their diet or growth. Most larvae are collected near the
surface of the ocean, typically in neuston nets, but some
larvae are collected in nets that sample below the surface
(Govoni et al., 2000).
In the present study, we resolve and summarize the early
life history of swordfish in the western North Atlantic. We
report estimated age, describe growth, relate growth to
feeding, morphological features, and vertical distribution,
and infer spawning time and location and the sources and
fates of swordfish larvae. This study supplements that of
Govoni et al. (2000) by providing the dimension of time,
i.e. age of larvae, to the spatial distribution and possible
transport of these larvae.
Methods
Collections of larvae
Ichthyoplankton collections from cruises in 1989 (in the
northeastern Caribbean about the Lesser Antilles), 1991
and 1997 (off the southeastern United States), and 2000
(in the Straits of Florida and off the southeastern United
States), produced 63 larvae that were preserved in 95%
ethanol (for examination of otolith microstructure). Sam-
ples were collected either from the neuston (i.e. taken with
a 1.0x0.5 m neuston net) or from depth intervals (i.e. taken
with a 1-m MOCNESS [multiple opening and closing net
and environmental sampling system] [Wiebe et al., 1976]).
Larvae were measured for preserved standard length
(PSL), the conventional length measure for larval fishes
(Kendall et al, 1984), and lower-jaw-fork-length (LJFL),
the measure in common use for juvenile and adult sword-
fish (Megalofonou et al., 1995).
Otolith excision and examination
Of the 63 larvae collected, sagittae were found and success-
fully excised from 37 larvae, lapilli from 32, and asterisci
from six. Otoliths were mounted on glass slides and dried
before examination. Broken sagittae and lapilli, and some
large sagittae, were embedded in plastic, sectioned with a
saw, and polished (Secor et al., 1991).
Otolith growth increments were counted along the lon-
gest axis of each sagitta and lapillus by using a compound
light-transmission microscope; increments in asterisci
were fewer and less defined and were not counted. Three
blind counts were made by the same observer. Although
increments were consistently visible on both sagittae
and lapilli, counts from individual larvae were greater on
sagittae (Student's Mest; P<0.001). Standardized counts
(the standard deviate of each repeated count) on the right
and left sagittae were not significantly different (nested
ANOVA; P<0.05). Increment counts from either the left or
right sagitta, decided by coin toss, were used for age and
growth rate determination. The mean of three replicate
counts was rounded to a whole number.
Increment counts from sagittae were used to estimate
larval age. Increments were assumed to form daily (Cam-
pana, 2001). The core increment was assumed to form at
hatching (Jones, 1986). The first increment outside of the
core increment was counted as one. Age from hatching
(AFH) was the number of increments counted from the
core increment on sagittae. This definition differs from
that of Prince et al. (1991) who counted the core increment
as increment one for the istiophorid blue marlin (Makaira
nigricans) and for a single larval swordfish that was 8.5
PSL. The radius of each otolith was measured by image
analysis.
Growth model
The best empirical fit among a suite of regressions of
estimated age (AFH) and length (PSL) — linear, polyno-
mial, and piece-wise, and moving — was chosen to describe
somatic growth (Forbes and Lopez, 1989; Hare and Cowen,
1995; Rogers et al., 2001). Criteria for best fit were the
following: the interpretation of fit from graphical display;
regression coefficients (r^), and dispersion or convergence
of regression residuals. The model of best fit was also
applied to the estimated age and lower jaw fork length
(LJFL) to allow comparison with published accounts of
juvenile swordfish growth.
Diet
Sixty-eight specimens from the present collections and
from Govoni et al. (2000), all with undamaged alimentary
canals, were examined for gut contents. These specimens
were taken in the northeastern Caribbean (3 specimens),
Gulf of Mexico (5), and off the southeastern coast of the
United States (60). Food was identified to the lowest taxon
possible following Govoni et al. (1983).
Physical features of larvae
Histological sections of three larvae, 21.5, 30.0, and
52.0 mm PSL, were cut as a preliminary aid to the loca-
tion of otoliths within the cranium and to determine the
histological constitution of the larval alimentary canal.
Time and location of spawning
Spawning dates were estimated from the ages of larvae
(estimated from the growth model (AFH)), plus 3 days
(the incubation period at 25°C for swordfish eggs given
by Yasuda et al. [1978]). Spawning dates are thus days
from fertilization (DFF). Spawning location was inferred
by applying DFF to larval swordfish lengths reported in
the present study, as well as lengths given in Govoni et al.
(2000), taking into consideration the time that eggs and
larvae were at large and adrift (DFF) and the location where
780
Fishery Bulletin 101(4)
Figure 1
Otoliths of larval swordfish (Xiphiasgladius): (A) histological section through the braincase of a 21.0-mm larva; (B) left
sagitta from a 5.4-inni-PSL larva; (C) histological section of a right sagitta of a 33.3-mm-PSL larva; and (D) transverse
section of a right sagitta of a 47.5-mm-PSL larva ( C=core increment; l=first check; 2=second check; M=macula within
the otolith vestibule; 0=otolith; x's (B-D) mark intervals of measurement).
larvae were collected, and back-calculating the geographic
origin of eggs with mean axial trajectories and velocities
of water currents for the Yucatan, Gulf Loop, Florida Cur-
rents, and the Gulf Stream (-1.5 m/s (Maul and Vukovich,
1993; Olson et al., 1994; Boicourt et al., 1998]) and the
Caribbean Sea (-0.2 m/s [Mooers and Maul, 1998]).
Results
Otolith structure and increment counts
Sagittae and lapilli were round, extremely small, and
lacked rostra or sulci in larvae <5 mm PSL(Fig. 1, A and Bi.
A rostrum developed on sagittae at -5.5 mm PSL (Fig. IC).
Lapilli did not develop rostra, and remained symmetrical
with growth.
Two checks, distinct zones of irregular increment spacing
and opacity, were evident on most sagittae (Fig. 1, B-D).
The first check was evident at the third increment on all
sagittae examined. The second check was found on sagit-
tae from larvae >3.8 PSL but varied from the seventh to
tenth increment.
Growth model
A piece-wise regression (Table 1; Equations 5-7) with two
linear segments provided the best fit with biologically real-
istic parameters. An assigned intercept of 3.2 mm PSL was
used for the first segment; this value was obtained by adjust-
ing the length at hatching with the scale given by Yasuda
et al. ( 1978) and by accounting for shrinkage due to pres-
ervation. Growth rate for the first segment was 0.3 mm/d
and 5.9 mm/d PSL for the second segment (Fig. 2A). The
intersection of the two linear segments was at an estimated
age of 13.3 d AFH., 3 to 6 d after the observed second check.
The PSL of larvae at the intersection was 11.0 mm.
Growth rate in LJFL, also modeled with a piecewise
regression, was 0.2 mm/d for the first segment (the upper
and lower jaws of larvae <11 mm PSL are of equal length.
Govoni et al : Early life history o\ Xiphias gladius in the western North Atlantic
781
Table 1
Summary of models evaluated for describing growth of larval swordfish. Models 1-6 were preserved standard length (PSD as a
function of estimated age from hatching (AFH); model 7 was lower jaw fork length (LJFL) as a function of AFH. There were three
model types; linear regression (models 1 and 2), 2"'' order polynomial (models 3 and 4), and piecewise regression (models 5-7).
Y-intercepts (model parameter a) were estimated by the regression in Equations 1, 3, and 5 but were fixed at an observed length
at hatching of 3.2 mm from Yasuda et al. (1978) in Equations 2, 4, 6, and 7 ("ns" denotes that a was not significantly different from
an assigned a of 3.2 at a=0.05; * =y-intercept significantly different from 3.2 at a=0.05; na = not applicable; b and c are slopes; and
d is the inflection point).
Model
1. PSL = a + 6AFH
2. PSL = 3.2 +6AFH
3. PSL = a +6AFH+ fAFH2
4 . PSL = 3.2+6AFH + c AFH^
5. PSL = a + 6AFH + c(AFH - d) x (AFH d)
6. PSL = 3.2 + 6AFH + c(AFH-d)x(AFH>d)
7. LJFL = 3.2 + 6AFH+C (AFH- d)x (AFH >d)
Model parameters
Model fit
a
b
c
d
r2
Residuals
-39.11*
4.15
na
na
0.81
not normal
3.20 na
1.47
na
na
0.81
not normal
11.31ns
-2.77
0.22
na
0.92
normal
3.20 na
-1.75
0.19
na
0.91
normal
-2.22*
0.76
5.10
13.53
0.89
normal
3.20 na
0.26
5.60
13.29
0.89
normal
3.20 na
0.16
3.28
11.24
0.77
ns
Table 2
Diet composition of 68 larval swordfish (Xiph
ias gladius)
in the western North Atlantic.
% Frequency
% total
of occurrence
number
(among larvae
(among all
Diet item
with food)
food items)
Copepodites and
adult copepods
(unidentifiable)
5
2
calanoids
Eucalanus
2
1
cyclopoids
Corycaeus spp.
59
74
Oithona spp.
4
2
Chaetognaths
2
1
Larval and juvenile
fishes 33
16
Invertebrate eggs
2
2
Chyme
5
2
hence PSL=LJFL) and 3.4 mni/d for the second segment
(Fig. 2B).
The fit of the piecewise regression for PSL and LJFL was
unchanged by inclusion or exclusion of the estimated age
of the largest larva.
Diet
The diet of larvae is limited and transitional. Larvae
<8.3 mm PSL ate copepods exclusively, primarily a single
cyclopoid genus, Corycaeus spp., but also another cyclopoid,
Oithona spp., and the calanoid Eucalanus (Table 2). Larvae
9.0 to 11.0 mm PSL ate copepods (Fig. 3A) and chaeto-
120
A © /
e"
/
E
/
£
/
E" 80
/
m
/
•g
/
CO
/
CO
CO
• f •
1 40
/•
a>
CO
/?*
CD
D.
^T*
i-i-i^^ia
— — en^co*'',*' 1 1 1
0 7 14 21 28 35
120
B
e"
E.
,
£ 80
• /
O)
c
cu
/
■=£
/
£
/
5
/
a
• /
S 40
S
A^
_i
•Z7 •
J^^**
©rartBi*,*
0
(
) 7 14 21 28 35
Estimated age from hatching (days)
Figure 2
Growth of larval swordfish from the western
North Atlantic: (A) estimated age and PSL; (B)
estimated age and LJFL.
782
Fishery Bulletin 101(4)
B
■-«»Si,~,
<5 5 to 10 10 to IS
Standard lengtti (mm)
>1S
Figure 3
Diet composition of larval swordfish from the western North Atlantic: (A) a 3.9-mm-PSL larva and cyclopoid copepods
(genus Corycaeus) that were removed from the gut; (B) a 12.7-mm larva with a larval fish in its gut; and (C) the frequency
of occurrence of copepods and larval fishes in the guts of larval swordfish.
gnaths. Larvae > 1 1 .0 mm PSL ate almost exclusively larval
and juvenile fishes (Fig. 3, B and C). Remnant jaws and
heavy pigmentation of many of the fishes eaten, indicated
that most were neustonic. One exocoetid was identified by
intact pectoral fins and counts of vertebrae.
Jaw and alimentary canal structure
The structure of the alimentary canal and jaws changed
concomitantly. The alimentary canal began to change from
three segments (foregut, midgut and hindgut), typical of
larval fishes (Govoni et al., 1986a), to four segments (esoph-
agus, stomach, anterior intestine, and posterior intestine)
between 9.0 and 12.0 mm PSL. Jaws change during this
period from the beak-like jaws to the elongate rostral bill
of the istiophorids (Fig. 3, A and B). Gastric glands were
evident in the fundic region of the stomach (ventricili-gas-
tric cecum), close to the junction with the esophagus in the
30.0-mm-PSL larva (Fig. 4, A and B). The pyloric region of
the stomach (pars pylorica) was evident in the 21.5-mm-
PSL larva and the 30.0-mm larva.
Time and location of spawning
Back-calculated spawning dates demonstrated year-round
spawning and peaks in three seasons and five regions
(Fig. 5, A-C). Larvae collected in the eastern Caribbean
were spawned in the winter (northern hemisphere) only.
Larvae collected in the western Gulf of Mexico were
spawned in spring. In the north-central Gulf of Mexico,
larvae were spawned in all seasons, but spawning peaked
in spring. Off south Florida, larvae were spawned in
all seasons, and spawning peaked peaked in spring. Lar-
vae collected off the southeastern United States were
spawned throughout the year whereas larvae collected in
the north-central Gulf of Mexico and in southern Florida
waters were spawned mostly in spring and early summer.
Modes of the number of larvae collected and their esti-
mated DFF advanced slightly in day of the year from the
north-central Gulf of Mexico to off the southeastern United
States. Larvae <10 DFF were collected both in the north-
central Gulf of Mexico and off the southeastern United
States (Fig. 5D), but not off South Florida.
Govoni et al.: Early life history of Xiph/as gladius in the western North Atlantic
783
Figure 4
Photomicrographs of a 30.0-mm-PSL swordfish larva: (A) fundic stomach; (B) gastric glands in stomach
mucosa (Gg=gastric glands; M=mucosa; L=stomach lumen with larval fish remains; E=eye of eaten larval
fish).
Discussion
The first check on sagittae apparently corresponds with
the completion of yolk and oil globule absorption and the
beginning of feeding. Sanzo (1910) and Yasuda et al. (1978)
reported hatching 3 days after fertilization when larvae
were ~4 mm live total length (LTD, or 3.8 mm live stan-
dard length (scaled from their drawings), and complete yolk
and oil globule absorption 8 d after fertilization (or 5 DAH)
when larvae were -5 mm LTL, or 4.3 mm live standard
length. Larvae from the present material had completed
yolk and oil absorption between 3.8 and 3.9 mm PSL and
had the first check 3 increments after the core increment.
Temperature and feeding influence the growth rate of fish
larvae and their otoliths, but larvae are typically collected
in water 25 ±1''C (Arata, 1954; Taning, 1955; Tibbo and
Lauzier, 1969; Markle, 1975; Govoni et al., 2000), as were
the larvae collected for age determination. This tempera-
ture is common to the Gulf Stream and its progenitor cur-
rents (Schmitz et al., 1993; Hitchcock et al., 1994), and is
similar to the temperature used to rear larvae (Sanzo, 1910;
Yasuda et al., 1978). The difference in length at complete
yolk and oil absorption between Sanzo (1910) and Yasuda
et al. (1978) and the present collections probably owes to
shrinkage of larvae with death and preservation (e.g. Thei-
lacker, 1980).
The second check follows concomitant changes in diet
and morphological features that take place between 8 and
13 mm PSL or from 7 to 11 DFH. An acceleration in so-
matic growth follows the second check within a day or so.
Young swordfish larvae eat copepods; older larvae other
larval fishes. The most striking morphological change of
larval swordfish is in the jaws. Swordfish larvae <13 mm
SL have beak-like jaws that are typical of the larval scom-
broid fishes (Collette et al., 1984), particularly those of
the wahoo (Acanthocybium solandrl) and scaleless tuna
(Gymnosarda unicolor); older larvae develop bill-like jaws
with elongate rostral cartilages anterior of the premaxillar-
ies and equally elongate mandibles (McGowan, 1988). The
constitution of the alimentary canal changes as well. The
development of a functional stomach with gastric glands
in larval swordfish, which typically arises during the meta-
morphosis of fishes (Govoni et al., 1986a), is evident in the
larvae of other scombroid fishes where it is accompanied
similarly by a change in diet from zooplankton to fish (Kaji
et al., 1999; Shoji et al., 1999). A switch from zooplankton to
fish is common among istiophorid larvae, but it is neither
as exclusive nor abrupt (Voss, 1953; Lipskaya and Gorbu-
nova, 1977) as it is with swordfish. Accelerated growth af-
ter such a dietary shift is also a common trait of scombroid
larvae (Shoji etal, 1999).
Swordfish larvae grow rapidly, faster than other larval
fishes with reportedly rapid growth. Growth rates reported
in the present study are for larvae that have survived pre-
dation and possibly variable feeding success; these rates
do not necessarily represent average larval growth of the
overall population. Growth rates of larvae >11 mm PSL
(13 DAH), 5.6 mm/d, are nonetheless faster than the lar-
val growth rates of other fast-growing larvae that survive
in the sea, e.g. sablefish (Anoplopoma fimbria) (Boehlert
and Yoklavich, 1985), and the oceanic-pelagic common dol-
phinfish (Coryphaena hippurus) reared in the laboratory
at high food densities without predation (Hassler and Hog-
arth, 1977; Kraul, 1991). The growth rate of larval sword-
fish <13 mm LJFL, 3 mm/d, is slower than the maximum
(16 mm LJFL/d) and sustained (10 mm LJFL/d) growth
784
Fishery Bulletin 101(4)
#h.n ^JT^
200
Estimated Spawning date
300
400
78
Longitude (°W)
40
30-
20
10
0
16
Southeast US
iifn iMilniiFl] [1 0
12-
South Florida
0
60r
50-
40
30
20
10
0
25r
20
15
10
5
0
12
10
8
6
4
k
Northcentral Gulf of Mexico
iJIL.^ r
Z
c
3
Western Gulf of Mexico
0
Eastern Caribbean
0 100 200 300 400
Estimated
spawning date
Figure 5
Estimated spawning dates (day of the year) for swordfish larvae collected from the Caribbean Sea, the Gulf of
Mexico, and off the southeast coast of the United States (hatched bars and circles depict winter in the north-
ern hemisphere, filled bars and triangles spring-summer, and cross-hatched bars and squares summer-fall):
(A) number of larvae and estimated spawning dates; (B) geographical distribution of larval collections;
(C) number of larvae and estimated spawning date by region; (D) number of larvae by region and season.
rate oflarval blue marlin over the first 100 days or <10G0
mm LJFL (Prince et al., 1991).
Larval and juvenile swordfish from the western North
Atlantic and Mediterranean exhibit four growth phases.
Growth is linear for larvae <11 mm PSL, for larvae 11 to
115 PSL, and for juveniles 510 to 740 mm LJFL (Megalo-
fonou et al., 1995). Growth becomes allometric for larger
juveniles (Ehrhardt, 1992). Larvae <11 mm grow -0.1 mm
LJFL/d. After the acceleration of growth, larvae in the
western North Atlantic grow at -3 mm LJFL/d, whereas
young juveniles in the Mediterranean grow at 23 mm
LJFL/d (Megalofonou et al., 1995). Growth slows in older
juveniles <250 mm LJFL to -2.5 mm LJFL/d (Ehrhardt,
1992). Adult growth (Berkeley and Houdc, 1983; Tserpes
and Tsimenides, 1989) may constitute a fifth phase, as rec-
ognized by Yabe et al. (1959) for Pacific swordfish.
The limited diet oflarval swordfish is unusual; few larval
fishes prey almost exclusively upon either copepods or lar-
val fishes. Swordfish larvae 12.0 mm total length (TL) eat
zooplankton, and larger larvae >12.0 mm TL eat other fish
larvae (Gorbunova and Lipskaya, 1975), including conspe-
cifics (Arata, 1954). Larval fishes as a whole are selective
feeders; Corycaeus is selected by larval percoids in the Gulf
of Mexico (Govoni et al., 1986b). Young larval istiophorids
from the Florida Current eat primarily cyclopoid copepods
of the genera Corycaeus, Farranula, and Oithona (Post et
al, 1997), before they become piscivorous (Gorbunova and
Lipskaya, 1975). Closely related genera of fishes exhibit
Govoni et al.: Early life history o\ Xiphias gladius in the western North Atlantic
785
20
D
2()r
15
10
5
20|-
15
10
Winter spawning
ai
tru
20
40
no larvae
20
40
60
60
40
20-
15^
loi-
60 20|
Spring/summer spawning
ou
Summer/fall spawning
40
-
20
-
>i^
/_'
•.■.1-.-.I
20r
10k
5^
20 40
no larvae
no larvae
no larvae
Estimated age from spawning (days)
Figure 5 (continued)
c
CO
en
o
Pz
60
O 0)
IT)
S 3
0) m
cr u>
different diets, even when occupying the same space (e.g.
Govoni et al., 1983; 1986b). Larvae of the related istiophorids
have limited diets, but these are not as exclusive or abruptly
changing as that of swordfish. Diets of larvae examined in
our study showed no evidence of cannibalism.
The diet of larval swordfish helps to resolve their vertical
distribution. Most larvae have been collected at the surface
in neuston or dip nets (Taning, 1955; Yabe et al., 1959; Gor-
bunova, 1969; Nishikawa and Ueyanagi, 1975), although
some have been collected in plankton nets that fished
principally below the surface (Grail et al., 1983; Govoni et
al., 2000). The diet of swordfish larvae indicates that larvae
<11 mm PSL may live in the near surface water, whereas
larvae >11 mm are neustonic. Corycaeus is a common ne-
ritic copepod of the Caribbean, the Florida Current, and
the continental shelf off the southeast coast of the United
States; Corycaeus is not neustonic (Owre and Foyo, 1967;
1972; Paffenhofer, 1983; 1985). That Corycaeus is eaten
almost exclusively by young swordfish larvae implies that
these larvae occupy the near-surface pelagia. Istiophorid
larvae undertake dietary shifts (Voss, 1953; Gorbunova
and Lipskaya, 1975; Lipskaya and Gorbunova, 1977) and
changes in vertical distribution (Bartlett and Haedrich,
1968; Leis et al., 1987) that are similar to those of sword-
fish larvae, but conflicting evidence exists for vertical dis-
tribution of larval istiophorids. Gorbunova and Lipskaya
(1975) implied that istiophorid larvae accumulate in sur-
face waters during the day and disperse below the surface
786
Fishery Bulletin 101(4)
at night, but Bartlett and Haedrich (1968) indicated the
reverse. Large swordfish larvae are caught during the day
and night in the neuston. The restricted diets of both large
and small larvae implies little vertical movement.
The overall modal increase in larval age from the Gulf of
Mexico to the north indicates that spawning takes place in
the north-central Gulf of Mexico and off the southeast coast
of the United States and that there is possible northward
transport. Estimated ages, along with velocities and trajec-
tories of currents, indicate that larvae could be transported
from considerable distances, but only if they remain within
the axes of major currents. The smallest larva collected off
the Carolinas was 3.9 mm PSL and had an estimated age
of 7 d AFH (this specimen was previously reported as be-
ing approximately 4 days old in Govoni et al. [2000]). With
3 days incubation at 25°C added to this estimated age, a
swordfish egg and larva would be planktonic for 10 days.
With a mean axial trajectory and velocity of the Florida
Current and Gulf Stream of 1.5 m/s (Olson et al., 1994), a
larva 3.9 mm PSL could be transported from as far away
as 910 km, which could place the origin of this larva in the
Straits of Florida, if its northward progress had not been
checked by eddies so typical of the Gulf Loop Current (Maul
and Vukovich, 1993) and Gulf Stream, particularly after
the Gulf Stream exits the Straits of Florida (e.g. Lee et al.,
1991; Govoni and Hare, 2001). Off the southeast coast of
the United States, swordfish larvae aggregate in the west-
ern Gulf Stream frontal zone (Govoni et al., 2000), where
northward current velocities are considerably slower than
1.5 m/s (Marmorino et al., 1999). Because larvae reside
primarily in the western Gulf Stream frontal zone where
northward velocities are slower and where the front itself
is so frequently distorted by meanders and eddies, it is
unlikely, but not impossible, that a larva as young as 7 d
APH collected off the Carolinas of the United States could
have been transported from the Straits of Florida.
The largest and oldest larva examined, one collected off
South Carolina (Govoni and Hare, 2001), was 115 mm PSL
and had an estimated age of 30 d AFH; with 3 days incuba-
tion this fish could have been at large for 33 days and would
have traveled 4290 km, given the mean axial trajectories
and velocities of the Caribbean Sea (-0.2 m/s) and the
Yucatan, Gulf Loop, and Florida Currents (-1.5 m/s). This
calculation might place the spawning origin of this larva
in the eastern Caribbean Sea or south of the Sargasso Sea
if a direct, unchecked passage is assumed.
Inference of the seasonality and geography of spawn-
ing is limited and biased by the unsystematic temporal
and spatial distribution of the present collections of larval
swordfish and by uncertainties about the rate and trajec-
tory of transport of eggs and larvae. Yet, taken as a whole,
spawning dates, back-calculated from larvae collected in
various regions of the western North Atlantic, and the
abundance and spatial distribution of the youngest larvae
indicate a spawning distribution with modes in three
seasons and five regions. The western Gulf Stream frontal
zone is the focus of spawning off the southeastern coast of
the United States. Spawning in the Gulf of Mexico seems
to be focused in the vicinity of the northern most arc of the
Gulf Loop Current.
Estimated spawning dates and the spatial distribution
of young larvae offer an alternative to gonad condition and
oocyte status as a means of resolving spawning season and
location. Spawning season and location resolved in the
present study corroborate the scenario recently proposed
by Arocha (1997). Rather than the single breeding unit cur-
rently recognized for the western North Atlantic by ICCAT,
Chow and Takeyama (2000) and Arata ( 1997) proposed two
spawning groups: one south of the Sargasso Sea and east
of the Lesser Antilles, and the other in the Windward Pas-
sage of the Antilles, the Yucatan Channel, and the Straits
of Florida up to 35°N latitude. Accordingly, spawning begins
in December south of the Sargasso Sea. Larvae from this
spawning group transit into the Caribbean, are retained
there by its anticyclonic circulation, and use the southeast-
em Caribbean as a nursery area. Arocha (1997) implied that
the second group spawns progressively later in the year and
that larvae are transported with the Gulf Loop and Florida
Currents, and the Gulf Stream. Arocha (1997) speculated
that larvae and juveniles use the Gulf of Mexico and waters
inshore of the Gulf Stream as a nursery area. Spawning
dates and abundances of young larvae corroborate Arocha's
(1997) proposed scenario for the seasonality and location
of spawning and confirm spawning off the southeastern
United States in the late spring and summer in the north-
ern hermisphere. Spawning dates and abundance of young
larvae also indicate the Gulf of Mexico as a nursery area.
Further, large numbers of juveniles discarded from the long-
line fishery prosecuted in the vicinity of the Charleston Gyre
(Cramer, 2001) and the collection of larvae there (Govoni
and Hare, 2001) indicate that the waters off the southeast-
ern coast of the United States serve as a nursery area.
Swordfish larvae are collected elsewhere in the western
North Atlantic, although they are rarely caught north of
Cape Hatteras, North Carolina (Tibbo and Lauzier, 1969).
The trajectory of the Gulf Stream north of Cape Hatteras
is convoluted and its velocity is slower, ~1 m/s (e.g. Bow-
ers and Rossby, 1989; Flierl and Davis, 1993; Hare et al.,
2002); the transit period of plankton from Cape Hatteras
to the Azores is 120-300 days (Scheltema, 1971). Swordfish
spawned in the western North Atlantic would be juvenile
fish by the time they reached the eastern North Atlantic.
North of Cape Hatteras, the Gulf Stream sheds eddies into
the Sargasso Sea (McGuillicuddy et al., 1998); thus, the
general location for juvenile swordfish that are spawned
and not retained in the western North Atlantic may well
be the central Atlantic and Sargasso Sea.
Swordfish are multiple spawners (Arocha, 2002) and
adults may move and spawn among regions of the western
North Atlantic. Movement of spawning adults, along with
transport of larvae, may result in the genetically well-mixed
population of the western North Atlantic ( Alvarado Bremer
et al., 1995a). There is apparently no genetic exchange be-
tween northwestern Atlantic and Mediterranean reproduc-
tive populations (Alvarado Bremmer et al., 1995b; Chow
and Takeyama, 2000). Transoceanic migration of adult fish
is possible, but cross-Atlantic transfer of swordfish larvae
is not likely. Swordfish lars'ae, collected principally in water
>25°C (Govoni et al., 2000), probably perish as Gulf Stream
water cools when it traverses the northern western North
Govoni et al.: Early life history of Xiphias gladius in the western North Atlantic
787
Atlantic (Scheltema, 1971; Cowen et al., 2000; Gaylord and
Gaines, 2000). Larvae that are not retained by eddies of the
Gulf Loop Current and the Gulf Stream may be diverted to
the central Atlantic (McGuillicuddy et al., 1998) and prob-
ably do not transit to the eastern North Atlantic.
Acknowledgments
We thank L. R. Settle [NOAA - NOS, Center for Fisheries
and Habitat Research (CCFHR)] for the collection of some
of the swordfish specimens reported in our study. M. M.
Leiby (Florida Department of Natural Resources) loaned
SEAMAP collections. F. Arocha (Universidad de Oriente,
Venezuela), S. A. Berkeley (Oregon State University), and D.
L. Lee and J. L. Cramer (NOAA- Fisheries, Southeast Fish-
eries Science Center) provided invaluable counsel on sword-
fish biology. D. W. Ahrenholz and M. H. Prager (CCFHR,
National Marine Fisheries Service, NOAA), W. J. Richards
(Fisheries, Southeast Fisheries Science Center, NOAA), and
F Arocha provided reviews of the manuscript.
Literature cited
Alvarado Bremer, J. R., A. J. Baker, and J. Mejuto.
1995a. Mitochondrial DNA control region sequences indi-
cate extensive mixing of swordfish (Xiphias gladius) in the
Atlantic Ocean. Can. J. Fish. Aquat. Sci. 52:1720-1732.
Alvarado Bremer, J. R., J. Mejuto, T. W. Greig, and B. Ely.
1995b. Global population structure of the swordfish (Xiphias
gladius) as revealed by analysis of the mitochondrial DNA
control region. J. Exp. Mar Biol, Ecol. 197:295-310.
Arata, G. F.
1954. A contribution to the life history of the swordfish,
Xiphias gladius Linnaeus, from the south Atlantic coast of
the United States and the Gulf of Mexico. Bull. Mar Sci.
Gulf Caribb. 4:183-243.
Arocha, F.
1997. The reproductive dynamics of swordfish Xiphias gla-
dius L. and management implications in the northwestern
Atlantic. Ph.D. diss., 350 p. Univ Miami, Miami, FL.
2002. Oocyte development and maturity classification of
swordfish from the north-western Atlantic. J. Fish Biol.
60:13-27.
Bartlett, M. R., and R. L. Haedrich.
1968. Neuston nets and South Atlantic larval blue marlin
{Makaira nigricans). Copeia 1968:469-474.
Beckett, J. S.
1975. Biology o^ sworAfish, Xiphias gladius L., in the north-
west Atlantic ocean. In Proceedings of the international
billfish symposium Kailua-Kona, Hawaii, 9-12 August
1972. part 1: Report of the symposium (R. S. Shomura and
F Williams, eds.), p. 103-106. NOAA Tech. Rep. NMFS-
SSRF-675.
Berkeley, S. A.
1983. Atlantic swordfish stock structure data and sugges-
tions for its interpretation. Int. Comm. Conserv. Atl.Tuna,
Collect. Vol. Sci. Pap. 18:839-845.
Berkeley, S. A., and E. D. Houde.
1983. Age determination of broadbill swordfish, Xiphias
gladius, from the Straits of Florida, using anal fin spine
sections. In Proceedings of the international workshop on
age determination of oceanic pelagic fishes: tunas, billfishes.
and sharks (E. D. Prince and L. M. Pulos, eds.), p. 137-143.
NOAA Tech. Rep. NMFS-SSRF-8.
Boehlert, G. W., and M. M. Yoklavich.
1985. Larval and juvenile growth of sablefish, Anoplopoma
fimbria, as determined from otolith increments. Fish. Bull.
83:475-481.
Boicourt, W. C, W. J. Wiseman Jr, A. Valle-Levinson, and
L. P. Atkinson.
1998. Continental shelf of the southeastern United States
and the Gulf of Mexico: in the shadow of the western bound-
ary current. In The sea (A. R. Robinson and K. H. Brink,
eds.), vol. 11, p. 135-183. John Wiley & Sons, New York,
NY.
Bowers, A. S., and T Rossby.
1989. Evidence of cross-frontal exchange processes in the
Gulf Stream based on isopycnal RAFOS float data. J. Phys.
Oceanogr. 19:1177-1190.
Campana, S. E.
2001. Accuracy, precision and quality control in age deter-
mination, including a review of the use and abuse of age
validation methods. J. Fish Biol. 59:197-242.
Chow, S., and H. Takeyama.
2000. Nuclear and mitochondrial DNA analyses reveal four
genetically separated breeding units of the swordfish. J.
Fish Biol. 56:1087-1098.
Collette, B. B., T. Potthoff W. J. Richards, S. Ueyanagi,
J. L. Russo, and Y. Nishikawa.
1984. Scombroidei: development and relationships. In
Ontogeny and systematics of fishes (H. G. Moser, W. J.
Richards, D. M. Cohen, M. P Fahay, A. W. Kendall, and S. L.
Richardson, eds), p. 591-620. Am. Soc. Ichthyol. Herpetol.
Spec. Publ. 1.
Cowen, R. K., K. M. M. Lwiza, S. Sponaugle, C. B. Paris, and
D. B. Olson.
2000. Connectivity of marine populations: open or closed?
Science 287:857-859.
Cramer, J.
2001. Geographic distribution of longline effort and sword-
fish discard rates in the Straits of Florida and oceanic
waters of the continental shelf slope, and Blake Plateau
off Georgia and the Carolinas from 1991 to 1995. In Island
in the stream: the ecology and fisheries of the Charleston
Bump (G. R. Sedberry, ed.), p. 97-103. Am. Fish. Soc.
Symp. 25.
Ehrhardt, N. M.
1992. Age and growth of swordfish, Xiphias gladius, in the
northwestern Atlantic. Bull. Mar Sci. 50:292-301.
Flierl, G. R., and C. S. Davis.
1993. Biological effects of Gulf Stream meandering. J. Mar
Res. 51:529-560.
Forbes, T. L., and G. R. Lopez.
1989. Determination of critical periods in ontogenetic
trajectories. Funct. Ecol. 3:525-632.
Gaylord, B., and S. D. Gaines.
2000. Temperature or transport? Range limits in marine
species mediated solely by flow. Am. Nat. 155:769-789.
Gorbunova, N. N.
1969. Breeding grounds and food of the larvae of the sword-
fish Xiphias gladius Linneus (Pisces, Xiphiidae). Probl.
Ichthyol. 9:375-387.
Gorbunova, N. N., and N. Ya. Lipskaya.
1975. Feeding of larvae of the blue marlin, Makaira nigri-
cans (Pisces, Istiophoridae). J. Ichthyol. 15:95-101.
Govoni, J. J., G. W. Boehlert, and Y. Watanabe.
1986a. The physiology of digestion in fish larvae. Environ.
Biol. Fishes 16:59-77.
788
Fishery Bulletin 101(4)
Govoni, J. J., and J. A. Hare.
2001. The Charleston Gyre as spawning and larval nursery
habitat for fishes. In Island in the stream: the ecology and
fisheries of the Charleston Bump (G. R. Sedberry. ed.), p.
123-136. Am. Fish. Soc. Symp. 25.
Govoni. J. J., D. E. Hoss, and A. J Chester
1983. Comparative feeding of three species of larval fishes in
the northern Gulf of Mexico: Breuoortia patronus, Leiosto-
mus xanthurus, and Micropogonias undulati/s. Mar. Ecol.
Prog. Ser 13:189-199.
Govoni, J. J., P. B. Ortner, F. Al-Yamani, and L. Hill.
1986b. Selective feeding of spot, Leiostomus xanturus,
and Atlantic croaker, Micropogonias undulatus, larvae
in the northern Gulf of Mexico. Mar. Ecol. Prog. Ser 28:
175-183.
Govoni, J. J., B. W. Stender, and 0. Pashuk.
2000. Distribution of larval swordfish, Xtphias gla-
dius, and probable spawning off the southeastern United
States. Fish. Bull. 98:64-74.
Grail, C, D. R de Sylva, and E. D. Houde.
1983. Distribution, relative abundance, and seasonality of
swordfish larvae. Trans. Am. Fish. Soc. 112:235-246.
Hare, J. A., J. H. Churchill, R. K. Cowen, T J. Berger,
R C. CorniUon, P Dragos, S. M. Glenn, J. J. Govoni,
and T N. Lee.
2002. Routes and rates of larval fish transport from the
southeast to the northeast United States continental shelf
Limnol. Oceangr. 47: 1774-1789.
Hare, J. A., and R. K. Cowen.
1995. Effects of age, growth rate, and ontogeny on the
otolith size-fish size relationship in bluefish, Pomatomus
sattatrix, and the implications for back-calculation of size
in fish early life history stages. Can. J. Fish. Aquat. Sci.
52:1909-1922.
Hassler, W. W., and W. T Hogarth.
1977. The growth and culture of dolphin, Coryphacna hip-
purus, in North Carolina. Aquaculture 12:115-122.
Hitchcock, G. L., T. Rossby, J. L. Lillibridge, E. J. Lessard,
E. R. Levine, D. N. Connors, K. Y. Borsheim, and M. Mork.
1994. Signatures of stirring and mixing near the Gulf
Stream front. J. Mar Res. 52:797-836.
Jones, C.
1986. Determining age of larval fish with the otolith incre-
ment technique. Fish. Bull. 84:91-104.
Kaji, T, M. Tanaka, M. Oka, H. Takeuchi, S. Ohsumi, K. Teruya,
and J. Hirokawa.
1999. Growth and morphological development of laboratory-
reared yellowfin tuna Thunnus alhacares larvae and early
juveniles, with special emphasis on the digestive system.
Fish. Sci. 65:700-707.
Kendall, A. W., Jr, E. H. Ahlstrom, and H. G. Moser.
1984. Early life stages of fishes and their characters. In
Ontogeny and systematics of fishes (H. G. Moser, W. J.
Richards, D. M. Cohen, M. R Fahay, A. W. Kendall, and S. L.
Richardson, eds.), p. 11-22. Am. Soc. Ichthyol. Herpetol.,
Spec.Publ. 1.
Kotoulas, G., A. Magoulas, N. Tsimenides, and E. Zouros.
1995. Marked mitochondril DNA differences between
Mediterranean and Atlantic populations of the swordfish,
Xiphiasgladius. Mol. Ecol. 4:47.3-481.
Kraul, S.
1991. Larviculture of the mahimahi, Coryphaena hippurus,
in Hawaii, USA. J. World Aquacult. Soc. 24:410-421.
LaMonte, F.
1944. Note on breeding grounds of blue marlin and sword-
fish o(T Cuba. Copeia 4:258.
Lee, T. N., J. A. Voder, and L. P. Atkinson.
1991. Gulf Stream frontal eddy influence on productivity of
the southeast U.S. continental shelf J. Geophys. Res. 96:
22,191-22,205.
Leis, J. M., B. Goldman, and S. Ueyanagi.
1987. Distribution and abundance of billfish larvae (Pisces:
Istiophoridae) in the Great Barrier Reef lagoon and Coral
Sea near Lizard Island, Australia. Fish Bull. 85:757-766.
Lipskaya, N. Ya., and N. N. Gorbunova.
1977. Feeding of sailfish larvae. Oceanology 17:340-344.
Markle, G. E.
1975. Distribution of larval swordfish in the Northwest
Atlantic Ocean. In Proceedings of the international billfish
symposium Kailua-Kona, Hawaii, 9-12 August 1972, part
1: Report of the symposium (R. S. Shomura and F Williams,
eds. ), p. 252-260" NOAA Tech. Rept. NMFS-SSRF-675.
Marmorino, G. O., D. R. Lyzenga, and J. A. C. Kaiser
1999. Comparison of airborne synthetic aperture radar
imagery with in situ surface-slope measurements across
Gulf Stream slicks and a convergent front. J. Geophys.
Res. 104:1405-1422.
Maul, G. A., and F M. Vukovich.
1993. The relationship between variations in the Gulf
of Mexico Loop Current and Straits of Florida volume
transport. J. Phys. Oceanogr. 23:785-796.
McGowan, C.
1988. Differential development of the rostrum and mandible
of the swordfish iXiphias gladius) during ontogeny and its
possible functional significance. Can. J. Zool. 66:496-503.
McGuillicuddy, D. J., A. R. Robinson, D. A. Siegel, H. J. Jannasch,
R. Johnson, T. D. Dickey, J. McNiel, A. F. Michaels, and
A. H. Knap.
1998. Influence of mesoscale eddies on new production in the
Sargasso Sea. Nature 395:263-266.
Megalofonou, P., J. M. Dean, G. DeMetrio, C. Wilson, and
S. Berkeley.
1995. Age and growth of juvenile swordfish, Xiphias gladius
Linnaeus, from the Meditteranean Sea. J. Exp. Mar Biol.
Ecol. 188:79-88.
Mooers, C. N. K., and G. A. Maul.
1998. Intra-Americas sea circulation: coastal segment (3,W).
In The sea (A. R. Robinson and K. H. Brink, eds.), vol. 11, p.
183-208. John Wiley & Sons, New York, NY.
Nishikawa, Y., and S. Ueyanagi.
1975. The distribution of the larvae of swordfish, Xiphias
gladius, in the Indian and Pacific oceans. In Proceedings of
the international billfish symposium Kailua-Kona, Hawaii,
9-12 August 1972, part 1: Report of the symposium (R. S.
Shomura and F Williams, eds.), p. 261-264. NOAA Tech.
Rep. NMFS-SSRF-675.
Olson, D. B., G. L. Hitchcock, A. J. Mariano, C. J. Ashjian,
G. Peng, R. W. Nero, and G. R Podesta.
1994. Life on the edge: marine life and fronts. Ocean-
ography 7:52-60.
Owre, H. B., and M. Foyo.
1967. Copepods of the Florida Current. Fauna Caribaea 1,
Crustacea, part 1: Copepoda, 137 p. Inst. Mar Sci., Univ.
of Miami, Miami, FL.
1972. Studies on Caribbean zooplankton. Description of the
program and results of the first cruise. Bull. Mar Sci. 22:
483-521.
Paffenhofer, G.-A.
1983. Vertical zooplankton distribution on the northeast-
ern Florida shelf and its relation to temperature and food
abundance. J. Plankton Res. 5:15-33.
Govoni et a\ Early life history oi Xiphias gladius in the western North Atlantic
789
1985. The abundance and distribution of zooplankton on the
southeastern shelf of the United States. In Oceanography
of the southeastern U.S. continental shelf (L. P. Atkinson,
D. W. Menzel, andK.A. Bush, eds.), p. 104-117. Am. Geo-
physical Union, Washington, DC.
Paiko, B. J., G. L. Beardsley and W. J. Richards.
1981. Synopsis of the biology of the swordfish, Xip/iia.s g/a-
dius Linnaeus, 21 p. NOAA Tech. Rep. NMFS Circ. 441.
Post, J. T, J. E. Serafy, J. S. Ault, T. R. Capo, and D. P de Sylva.
1997. Field and laboratory observations on larval Atlantic
sailfish ilstiophorus platypterus) and swordfish {Xiphias
gladius). Bull. Mar Sci. 60:1026-1034.
Potthoff, T, and S. Kelley
1982. Development and structure of the vertebral column,
fins and fin supports, branchiostegal rays and squamation
in the swordfish A7p/uas g/arfii/s. Fish. Bull. 80:161-186.
Prince, E. D., D. W. Lee, J. R. Zweifel, and E, B. Brothers.
1991. Estimating age and growth of young Atlantic blue
marlin Makaira nigricans from otolith microstructure.
Fish. Bull. 89:441-459.
Reeb, C. A., L. Arcangeli, and B. A. Block.
2000. Structure and migration corridors in Pacific popu-
lations of the swordfish Xiphius gladius, as inferred
through analyses of mitochondrial DNA. Mar. Biol. 136:
1123-1131.
Richards, W. J., and T. Potthoff
1980. Larval distributions of scombrids (other than bluefin
tuna) and swordfish in the Gulf of Mexico in the spring of
1977 and 1978. Int. Comm. Conserv. Atl. Tuna, Coll. Vol.
Sci. Pap. 9:680-694.
Rogers, J. S., J. A. Hare, and D. G. Lindquist.
2001. Otolith record of age, growth, and ontogeny in larval
and pelagic juvenile Stephanotepis hispidus (Pisces:
Monacanthidae). Mar. Biol. 138:945-953.
Sanzo, L.
1910. Uovoe larva di Pesce-spada (A'lp/iiasg/adii/s L.). Riv.
Mens. Pesca Idrobiol. 12:206-209.
Scheltema, R. S.
1971. The dispersal of the larvae of shoal-water benthic
invertebrate species over long distances by ocean currents.
In Fourth European marine biology symposium (D. J. Crisp,
ed.), p. 7-28. Cambridge Univ. Press, Cambridge, UK.
Schmitz, W. J., J. R. Luyten, and R. W. Schmitt.
1993. On the Florida Current T/S envelope. Bull. Man Sci.
53:1048-1065.
Secor, D. H., J. M. Dean, and E. H. Laban.
1991. Manual for otolith removal and preparation for micro-
structural examination, 85 p. Electric Power Research
Institute and the Belle W. Baruch Institute for Marine
Biology and Coastal Research, Univ. South Carolina,
Columbia, SC.
Shoji, J., T. Maehara, and M. Tanaka.
1999. Short-term occurrence and rapid growth of Spanish
mackerel larvae in the central waters of the Seto Inland
Sea, Japan. Fish. Sci. 65:68-72.
Taning, A.
1955. On the breeding areas of the swordfish iXiphias).
Deep-Sea Res. 3(suppl.):438-450.
Taylor, R. G., and M. D. Murphy
1992. Reproductive biology of the s-w or Af\s\\ Xiphias gladius
in the Straits of Florida and adjacent waters. Fish. Bull.
90:809-816.
Theilacker, G. H.
1980. Changes in body measurements of larval northern
anchovy, Engraulis mordax, and other fishes due to han-
dling and preservation. Fish. Bull 78:685-692.
Tibbo, S. N., and L. M. Lauzier.
1969. Larval swordfish (Xiphias gladius) from three locali-
ties in the western Atlantic. J. Fish. Res. Board Can. 26:
3248-3251.
Tserpes, G., and N. Tsimenides.
1989. Age determination and growth of swordfish Xiphias
gladius L., 1958 in the Aegean Sea. Fish Res. 8: 159-168.
Voss, G. L.
1953. A contribution to the life history and biology of the
sailfish, Istiophorus americanus Cuv. and Val., in Florida
waters. Bull. Mar. Sci. Gulf Caribb. 3:206-240.
Wiebe, P H., K. H. Burt, S. H. Boyd, and A. W. Morton.
1976. A multiple opening/closing net and environmental
sensing system for sampling zooplankton. J. Mar. Sci. 34:
313-326.
Yabe, H., S. Ueyanagi, S. Kikawa, and H. Watanabe.
1959. Study on the life-history of the sword-fish, Xiphias
gladius Linnaeus. Rep. Nankai Reg. Fish. Res. Lab. 10:
107-150.
Yasuda, F., H. Kohno, A. Yatsu, H. Ida, P. Arena, F L. Greci,
and Y. Taki.
1978. Embryonic and early larval stages of the swordfish,
Xiphias gladius, from the Mediterranean. J. Tokyo Univ.
Fish. 65:91-97.
790
Abstract— Although subsampHng is
a common method for describing the
composition of large and diverse trawl
catches, the accuracy of these tech-
niques is often unknown. We deter-
mined the sampling errors generated
from estimating the percentage of
the total number of species recorded
in catches, as well as the abundance
of each species, at each increase in
the proportion of the sorted catch. We
completely partitioned twenty prawn
trawl catches from tropical northern
Australia into subsamples of about
10 kg each. All subsamples were then
sorted, and species numbers recorded.
Catch weights ranged from 71 to 445 kg,
and the number offish species in trawls
ranged from 60 to 138, and invertebrate
species from 18 to 63. Almost 70% of the
species recorded in catches were "rare"
in subsamples (less than one individual
per 10 kg subsample or less than one in
every 389 individuals).
A matrix was used to show the in-
crease in the total number of species
that were recorded in each catch as the
percentage of the sorted catch increased.
Simulation modelling showed that sort-
ing small subsamples (about 10% of
catch weights) identified about 50% of
the total number of species caught in a
trawl. Larger subsamples (509c of catch
weight on average) identified about 80%
of the total species caught in a trawl.
The accuracy of estimating the abun-
dance of each species also increased
with increasing subsample size. For
the "rare" species, sampling error was
around 80% after sorting 10% of catch
weight and was just less than 50% after
40% of catch weight had been sorted.
For the "abundant" species (five or
more individuals per 10 kg subsample
or five or more in every 389 individu-
als), sampling error was around 25%
after sorting 10% of catch weight, but
was reduced to around 10% after 40%
of catch weight had been sorted.
Does the size of subsamples taken from
multispecies trawl catches affect estimates
of catch composition and abundance?
Donald S. Heales
David T. Brewer
CSIRO Marine Research
233 Middle St
Cleveland, Queensland 4163, Australia
E mail address (for D S Heales) don.heales@marine csiro.au
You-Gan Wang
Dept of Biostatistics
Harvard University
Boston, Massachusetts 02115
Peter N. Jones
CSIRO Mathematical and Information Sciences
233 Middle St.,
Cleveland, Queensland 4163, Australia
Manuscript approved for publication
9 March 2003 by Scientific Editor.
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:790-799 (2003).
Concerns are held worldwide regarding
the sustainability of bycatch species
taken in trawls, particularly prawn
trawls. Under the voluntary FAO Code
for Responsible Fisheries, managers are
required to "take measures to conserve
target species, associated or dependent
species and nontarget species and their
environment" (FAO, 1995). An essen-
tial part of this process is the accurate
monitoring of population sizes and
structures.
With large trawl catches, subsam-
pHng is often the only cost-effective
or feasible way to describe the bycatch
composition. How well these subsam-
ples represent the total catch depends
on how diverse the catch is, how well
the catch is mixed before the subsam-
ples are taken, and what proportion of
the catch is taken as a subsample.
There is a large literature on sub-
sampling theory for terrestial insect
studies (Van Ark and Meiswinkel,
1992), aquatic macroinvcrtcbrate stud-
ies (Vinson, 1996; Walsh, 1997), and
marine ecological studies (Andrew and
Mapstone, 1987). However, in most of
these studies, samples of very small
animals collected in the field can be
resuspended in fluid and mixed evenly
in the laboratory before the subsam-
ples are taken. In fisheries, in direct
contrast, large catches are extremely
difficult to manipulate and redistrib-
ute evenly before subsampling. A few
fisheries studies have examined the
impact of subsampling on estimates of
the abundance and different size ranges
of one or a few dominant species. For ex-
ample, in the Crangon trawl fisheries in
Belgian waters, sampling strategy had
only a minor effect on the reliability of
estimates of size selectivity for the tar-
geted shrimp (Polet and Redant, 1999).
In UK waters, subsampling trawled fish
(both target species and discards) from
either the sorting conveyor or the pound
made no difference to catch composition
estimates {Tamsett et al., 1999).
However, in tropical trawl fisher-
ies, over one hundred species can be
recorded in a single catch. Under ESD
(ecological sustainable development)
guidelines, all these species (both tar-
get and bycatch) are equally important
but there has been very little research
on subsampling techniques applicable
to such diverse catches. A recent study
in Australia's Northern Prawn Fishery
(NPF) examined the accuracy of sub-
sampling from large, diverse catches
of fish and invertebrates (Heales et
al., 2000). For most of the "abundant"
Heales et al : Effect of size of subsamples on estimates of catchi composition and abundance
791
species, their position on trawler sorting
trays from which the bycatch subsamples
were collected, had little effect on the ac-
curacy in representing the catches.
Although the accuracy of subsampling
is a general problem for all multispecies
fisheries, few other studies have been pub-
lished on the topic. Reliable techniques for
subsampling are needed, however, espe-
cially with the demands that bycatch spe-
cies, as well as the target species, should
be ecologically sustainable. We describe
here research done in Australia's NPF but
the results and methods are applicable to
many sampling problems in fisheries.
The NPF is a large tropical trawl
fishery that extends from Cape York in
Queensland to Cape Londonderry in
Western Australia. In addition to catch-
ing penaeid prawns (e.g. 8531 metric tons
(t) in 1997-98, Taylor and Die, 1999), its
bycatch component is estimated at over
38,000 t a year (Pender et al., 1992) or
more than 80% of the total catch in the
tiger prawn fishery (Brewer et al., 1998).
The neighboring Torres Straits Prawn
Fishery (TSPF) also has an estimated an-
nual bycatch of 4800 t (Williams, 1985).
The NPF has a management require-
ment to assess the impact of trawling on
nontarget species. The bycatch of both
these prawn trawl fisheries (NPF and
TSPF) is very diverse. Ramm et al. (1990) recorded 115
fish taxa in their study of NPF waters, and Brewer et al.
( 1998) recorded over 250 species from one area of the NPF.
At least 390 fish species, 234 invertebrate taxa, and 43 elas-
mobranch species have been recorded in a current bycatch
project in the NPF (Stobutzki et al., 2000).
Despite the lack of knowledge on the ability of sub-
samples to accurately represent such diverse catches,
many studies of trawl communities in Northern Australia
have used subsampling techniques to estimate catch rates.
In two bycatch studies of the NPF, the smaller catches
were entirely sorted and the larger catches subsampled
(Poiner and Harris, 1986; Harris and Poiner, 1991). Trawl
catches in another bycatch study were spread evenly over
the sorting tray and a visually estimated fraction of the
catch was subsampled (Ramm et al., 1990). Other work-
ers simply subsampled without confirming the accuracy
of their subsampling techniques (Blaber et al., 1990 and
1994; Martin et al, 1995; Brewer et al, 1998; Wassenberg
etal., 1998).
A lack of knowledge of trawl impacts on nontarget spe-
cies has led to the present CSIRO study that describes the
bycatch from the NPF and provides a framework for any
future bycatch monitoring program. As part of that study,
we made the first assessment of the accuracy of subsam-
pling over a range of subsample sizes as a tool for esti-
mating the total catch composition of a large multispecies
fishery.
129"
134
139
Stu(jy area
\'L J Australia
129"
134
139"
144"
Figure 1
Map of the Torres Straits Fishery (area 1 — broken line) and Northern Prawn
Fishery (unbroken line), Australia, showing the managed area of the fisheries
and the regions where samples were collected. Regional codes are 1 = Torres
Straits, 2 = Weipa, 3 = east of Mornington Island, 4 = north of Mornington
Island, 5 = west of Mornington Island, 6 = east of Vanderlin Islands, 7 = south
of Groote Eylandt, 8 = north of Groote Eylandt, 9 = Melville Island.
Materials and methods
Data were collected from a series of 14 trawl samples taken
during two research cruises of the RV Southern Surveyor.
These samples were collected from one region of the Torres
Straits Prawn Fishery (TSPF), and from eight of the major
tiger prawn (Penaeus esculentus and P. semisulcatus) fish-
ing regions of the Northern Prawn Fishery (NPF), (namely
Weipa, east of Mornington Island, north of Mornington
Island, west of Mornington Island, north of Vanderlin
Islands, south of Groote Eylandt, north of Groote Eylandt,
and Melville Island) (Fig. 1). All trawls were undertaken
in either late summer 1997 (February-March, the end of
the wet season) or in mid-spring 1997 (September-October,
the dry season). We used a single 14-fathom Florida-Flyer
prawn trawl net so that the data would be comparable with
data from either of the two nets used by the twin-rigged
commercial NPF vessels in the tiger prawn fishery. All
trawls were done at night, again to be comparable with
the fishery Duration of trawls ranged from 1 to 3 hours
(Table 1), and depths ranged from 23 to 42.3 m.
A further six trawl catches were sampled by a scientific
observer on board commercial NPF vessels fishing for ti-
ger prawns north of Mornington Island in late May 1997,
and north of Groote Eylandt in late September 1997. Each
trawl sample consisted of the entire catch from one of the
two 14-fathom Florida-Flyer prawn trawl nets used by
these vessels. All trawls were done at night. Their duration
792
Fishery Bulletin 101(4)
Table 1
Summary of catch data for 20 entirely sorted trawls froi
n the Northern Prawn Fishery and Torres
Straits Prawn Fishery.
'Bycatch
individuals" refers to the total
number of bycatch animals (fish and invertebrates)
in each trawl.
"Fish species" refers to the total |
number offish species recorded in the bycatch of each trawl, (n) is
the total number
Duration of
Start
Catch
Bycatch
Fish
Invertebrate
All
trawl
time of
weight
Subsamples
individuals
species
species
species
Region
(h)
trawls
(kg)
in)
in)
In)
(;i)
Ul)
Torres Straits
1.2
0400
94
9
4151
74
21
95
Weipa
2.7
0220
166
16
4911
92
36
128
East of Mornington Island
1.0
0415
274
24
7313
90
18
108
East of Mornington Island
2.9
0250
87
8
2019
77
45
122
North of Mornington Island
3.3
1845
182
16
9762
101
36
137
North of Mornington Island
3.0
2215
194
17
11015
94
42
136
North of Mornington Island
3.2
1840
445
36
13,826
114
52
166
North of Mornington Island
2.0
0350
71
7
1771
60
30
90
North of Mornington Island
2.2
0330
100
10
5067
77
34
111
West of Mornington Island
1.6
0415
174
16
5976
100
21
121
West of Mornington Island
1.7
0400
269
26
6967
87
24
111
North of Vanderlin Islands
2.7
0310
156
15
7182
71
19
90
South of Groote Eylandt
2.0
0345
315
27
23,751
105
25
130
South of Groote Eylandt
2.7
0250
165
17
3856
64
25
89
North of Groote Eylandt
2.5
2230
147
12
5558
68
28
96
North of Groote Eylandt
3.0
0215
85
9
2792
60
34
94
North of Groote Eylandt
3.5
1830
158
13
5664
89
43
132
North of Groote Eylandt
3.5
2215
169
15
8289
96
63
159
North of Groote Eylandt
3.5
2215
189
16
5748
96
43
139
Melville Island
3.0
0130
170
14
4635
60
30
90
Total
3610
323
140,253
ranged from 3 to 3.5 hours (Table 1), and depths ranged
from 29 to 41 m.
Sample collection
On the research vessel, catches were spilled from the
codend onto the flat deck (equivalent to the sorting tray
on commercial vessels). The entire catch of each trawl was
progressively partitioned by shovelling the catch into con-
secutively numbered boxes (subsample replicates), each of
about 10 kg (according to the methods described by Heales
et al., 2000). Partitioning of the catch began at the outer
edges and continued in a clockwise direction, and sub-
samples were taken from each of four major compass direc-
tions: north, east, south, and west until the entire catch had
been collected in successively numbered boxes. The direc-
tion of the ship's bow was always designated as north.
Samples on the research vessel were sorted immediately.
Fish and invertebrates were identified to the lowest taxo-
nomic level possible (mostly species). Where this was not
possible, the data were grouped to genus and in a few cases
to family. Total numbers and weights were recorded for
each species in each subsample and entered directly into a
relational database.
On the commercial vessels, the catches were spilled onto
the sorting tray and the commercial-size prawns were
removed. The bycatch would then normally move down
a trash chute and spill overboard. However, to sample a
catch, the trash chute was diverted so that all the bycatch
was collected in consecutively numbered boxes (subsample
replicates) each of approximately 10 kg.
All samples collected from commercial vessels were
frozen on board and transported to the laboratory for sub-
sequent sorting, identification, and data entry according to
the methods described above.
Although most bycatch species were identified to spe-
cies level, some could only be identified to genus and in a
few cases to family. In order to be consistent in terms used
throughout this study, we use the term species (plural
form) even when referring to multispecies groups.
The methods used to collect subsamples on both the re-
search and commercial vessels differed only in the position
from which subsamples were taken (see earlier "Material
and methods" section). However, a previously published
study (Heales et al., 2000) showed that the majority of
the "abundant" bycatch species were evenly distributed
throughout the catch. Consequently, for all analyses we
combined the 14 catches collected from the research vessel
with the six catches from commercial vessels.
Data analysis
Abundance groupings Within each catch, there was a large
range of species (both fish and invertebrates) and they
Heales et al : Effect of size of subsamples on estimates of catcfi composition and abundance.
793
occurred at many different levels of relative abundance. To
obtain an overview of how rarely, or how frequently, differ-
ent species occurred in catches, we reduced each occurrence
of a species in a catch to an index of relative abundance. We
concentrated solely on determining the accuracy of taking
different size subsamples in representing the range of rela-
tive abundances (from very low to very high) of the species
in these catches.
The relative abundance indices were based on the av-
erage number of individuals of a given species that were
recorded in a standardized 10-kg subsaniple taken from
that catch. To generate this index, we used the following
equation;
n = 10 X (TotNum I Weight).
(1)
where n = the mean number of individuals of a given spe-
cies per 10-kg subsample;
TotNum = the total number of individuals of that species
in the whole catch; and
Weight = the total weight of the catch in kg.
We derived a separate index of abundance for each species
in every catch where it was recorded. Thus, a species that
occurred in all 20 catches would have 20 different abun-
dance indices in the analysis.
To highlight the differences in distribution between
the two extremes of the "rare" species and the "abundant"
species when estimating catch composition, we grouped
the indices of abundance into 11 categories, ranging from
less than one individual per 10-kg subsample, up to 10 or
more individuals per subsample. Species with abundance
indices of less than one individual per 10-kg subsample
were classed as "rare"; those with one to less than five
individuals were classed as "common"; and those with five
or more individuals per 10-kg subsample were classed as
"abundant."
For example, the common ponyfish (Leiognathus moreto-
niensi), was classed as "abundant" in 11 of the 20 catches, as
"common" in eight catches, and "rare" in one catch. Because
a species could have different abundance indices in each
catch, individual species are not referred to by name in the
results. Instead, we refer to the occurrence of each species in
a catch, as one case of some relative abundance index that
was recorded in that catch (i.e. one "case" of species by trawl
abundance). The relativefrequency of all the cases (i.e. abun-
dance indices) in each of the three abundance categories
(throughout the combined 20 catches) was then calculated.
To calculate the average number of bycatch individuals
per 10-kg subsample (over all the catches), the following
equation was used:
X = 10 X (Total number /Total weight)
(2)
where
X = the mean number of bycatch individuals
(per 10-kg subsample);
Total number = the total number of all bycatch individu-
als (summed over all 20 catches); and
Total weight = the total weight (kg) of all bycatch indi-
viduals (summed over all 20 catches).
We then examined the average occurrence ratios within
10-kg subsamples for the "rare," "common ," and "abundant"
species.
Catch composition To examine the relationship between
the number of recorded species and the weight of sorted
catch, the subsamples were first analyzed in the order
that they were collected. The cumulative number of spe-
cies (both fish and invertebrates) was plotted against
the cumulative weight of sorted catch, for each of the 20
catches. Each catch was also summarized in terms of the
percentage of species recorded for each 10% increment of
weight of sorted catch.
The order (position on the sorting tray) where the sub-
samples were collected on both the research and commercial
vessels was just one of the many possible ways that a catch
can be divided into 10-kg subsamples. To determine the level
of accuracy in recording the number of species in a catch,
we examined 200 combinations of subsample selection (with
no replacement), by randomly reordering the subsamples
using Monte Carlo simulations for each catch. We also cal-
culated the cumulative number and percentage of species
recorded, as well as the cumulative weight and percentage
of the sorted catch, for each catch. The proportion of species
recorded was fitted as a power function of the proportion
of the weight of sorted catch, as described by the following
asymptotic equation (Snedecor and Cochran, 1980):
y=p'' + e.
(3)
where y = the proportion of species recorded;
p = the proportion of the weight of catch sorted;
k = the mean exponential parameter; and
e = the random normal error term, with unequal
variance.
The variance of £ is assumed to be p (1-p) a^ to ensure
that the variance of y is fixed at zero when p = 0 and 1.
This formulation has the property that, when none of the
catch has been sorted, no species will have been recorded.
It also ensures thaty = 1 whenp =1, i.e. when all the catch
has been sorted, all of the species have been recorded. The
estimate of ct- was obtained by fitting the following model
according to the SAS procedure NLIN (version 7, SAS Inst.,
Cary NC):
y
* = p' I ^{pO-p)) + £'
(4)
where y* = y H(p(l-p)) and £* = E/^(p(l-p)); and
£* now has homogeneous variance structure.
Different k values were estimated for each catch to re-
flect the variation in the relationship. The mean ki value for
a given catch (! = l-20) was obtained from 200 analyses for
that catch. The predicted y values i.e. y^ (atp=0.1, 0.2 etc.
to 1.0) were obtained by averaging p*' values across_the 20
catches (note that this is different from p'- where k is the
mean k value for the 20 catches). We defined the y^ values
as the predicted expected proportion of species recorded
after p proportion of catches had been sorted.
794
Fishery Bulletin 101(4)
The corresponding 95% confidence interval for the pre-
dicted mean values (vp) was evaluated by using the width
1.96 a„ where 10 to <50, and >50 per subsample)
had mean sampling error rates below 25%.
For the "rare" species ( ^
^^ ^"^
Cumulative weight of catch sorted (l0.05) between
the southern CPFV and northern speared
samples (Fig. 6). The models representing
males were virtually identical, with param-
eters for the CPFV model of t^ = -0.94 years,
k = 0. 195/year, and L„ = 331.66 mm FL, and
for the speared model of t^ = -0.99 years,
k = 0.194/year, and L^ = 323.14 mm FL. There
was no significant difference between females;
<„ = -1.94 years, k = 0.107/year, and L„ = 430.74
mm FL for the CPFV samples and t^ = -1.14
years, k = 0.166/year and L„ = 393.34 mm FL
for the speared. Females from CPFV were
larger at ages after 15 years than those that
had been speared.
Discussion
100 200 300 400 500
Julian date of capture (day)
600
700
Figure 2
Growth of blue rockfish (Sebastes mystinus) collected nearshore during
their first 1.5 years. Solid line is the fitted linear growth model l/i=99;
r-=0.91). Vertical line represents 1 May, one year after the parturition
date of 1 January (i.e. 365 +120=485 days).
We estimated the age (using the break-and-
burn technique) of blue rockfish to be greater
than that reported in earlier studies. Aging the
scales of blue rockfish. Miller and Geibel ( 1973 )
reported maximum ages of 24 and 17 years for
females and males, respectively, whereas the
oldest of either sex reported by MacGregor
(1983) was only 13 years. Based on modal pro-
gression of length distributions, the estimate of
the oldest individuals of either sex calculated
by Karpov et al. (1995) was 17 years. In a study
of blue rockfish off Newport, Oregon, McClure
(1982) examined otolith surfaces and deter-
mined that the oldest female was 16 years,
and the oldest male was only 12 years. In aging
males to 44 years and females to 41 years, our
study more than doubled the recorded maxi-
mum ages for blue rockfish, demonstrating the
value of the break-and-burn section method for
accurate age determination.
Age data were validated by using an edge analysis and
the first translucent zone was validated as corresponding
to the first annual growth increment. Campana (2001)
pointed out that there are problems in using edge analysis
as a validation tool. Specifically the extension of younger,
validated ages to older, nonvalidated ages. In our study,
we validated ages up to 23 years for females and up to 25
30
2.5
2.0
g 1 5
05
0,0
Diameter = 0.02(FL) - 0.02
60
Fork length (mm)
120
Figure 3
Change in otolith diameter with fish length for young-of-the-year and
one-year-old blue rockfish (Sebastes mystinus). Solid line is the fitted
linear model ( n = 198; r2=0.95 ). Dashed line represents an estimated total
otolith diameter of 2.19 mm for a fish of 108.5 mm FL (i.e. fish length at
time of translucent zone completion).
years for males; ages of older fish could not be positively
validated. Therefore, caution must be taken when using
the older ages.
The growth rates of blue rockfish in our study were
similar to those estimated by others in California, but
slower than conspecifics off Oregon (Fig. 7). MacGregor
(1983) examined blue rockfish from southern California
and determined the combined male and female growth
804
Fishery Bulletin 101(4)
4.5
1.5
*,(« S» - » . - . ^i^^^ y» 4t - - , -
§^ -S**^
100
200 300
Fork length (mm)
400
500
Figure 4
Diameter of the first, second, and third translucent zones in otoliths from blue
rockfish iSebastes mystinus) at various lengths. Dashed lines represent the
mean diameters.
500
450
400
F
350
300
0)
250
o
200
150
100
20 25 30
Age (yr)
Figure S
Von Bertalanffy growth models for all male (solid line) and female (dashed line)
blue rockfish {Sehastes mystinus). (;i=348 for males and /i = 1348 for females.)
rate and calculated k (instantaneous growth rate) to
range from 0.13-0.16/years, which was comparable to k in
our study (0.2/year for males and 0.15/years for females;
mean ^=0.17/year). Karpov et al. (1995) calculated k for
the combined male and female growth rate from modal
progressions studies to be 0.12/year. This also was similar
but less than the k from our present study. On the other
hand, McClure ( 1982) estimated a much faster growth rate
for blue rockfish off Oregon, with a k for males of 0.23/years
and for females of 0.31/year. Although the Oregon fish were
larger at age (Fig. 7), maximum sizes from Oregon and
California were similar; the largest specimen from Oregon
was 460 mm FL (McClure, 1982) and the largest individual
from our study was 444 mm FL.
The difference in growth between studies may be attrib-
uted to a temporal difference in the collection offish. Two
thermal regime shifts have occurred in the Pacific Ocean
over the past 25 years; one in 1977 and the other in 1989
(Hare and Mantua, 2000). The samples from our study
came from two different thermal regimes, but the growth
Laidig et al.: Age and growth of Sebastes mystinus
805
curves were not statistically different (Table 1). Therefore,
these regimes did not appear to effect the growth of adult
fish. Out of the four other surveys mentioned above, three
came from one of these two regimes, and the fourth, Miller
and Geibel (1973), came from an earlier regime. If there
were any effects from the three different thermal regimes,
it would seem clear that these differences would show up
between samples from such varied regimes. However, the
only study with different measures of growth was that from
Oregon (McClure, 1982), with samples that were collected
during the same regime as two of the other studies (Mac-
Gregor, 1983; and the present study). Therefore, thermal
regime alone does not seem to have a major impact on the
growth of blue rockfish, although further analysis is needed
to confirm this point.
These differences in growth parameters between fish
from California and Oregon may be attributed to differ-
ences in aging methods. Wilson and Boehlert (1990) found
that estimates of growth based on aging of otolith surfaces
were higher for Sebastes pinniger, but were similar to
growth rates estimated from otolith sections for S. diplo-
proa. The ages of S. alutus determined from otolith surfaces
had poor correlation with ages from otolith cross-sections
for fish older than 17 years, but there was close agreement
for younger fish (Stanley and Melteff 1987). Reading ages
from the surface of an otolith may underestimate the age
of a rockfish (Munk, 2001) and thus result in greater size-
at-age and growth rate estimates. However, aging methods
may not be the only factor influencing the growth discrep-
ancies. Miller and Geibel (1973) and MacGregor (1983)
both used scales to age blue rockfish (which also can under-
estimate the age offish [Beamish and MacFarlane, 1987]),
and, yet, their growth models more closely approximated
the model produced by our study.
Faster growth estimated for blue rockfish off Oregon
may reflect a latitudinal difference in growth. Fraidenburg
(1980) examined length and age composition of Sebastes
flavidus and reported evidence of a north-to-south cline
of decreasing size-at-age. Pearson and Hightower (1991)
studied S. entomelas and noted smaller k values and larger
average maximum lengths with increasing latitude. Boe-
hlert and Kappenman ( 1980) reported faster growth in the
north for S. diploproa and no difference in growth with
latitude for S. pinniger. They postulated that because the
fish live demersally on the continental shelf latitudinal
variation in environmental factors may be insufficient to
explain the difference in growth rates and that differential
exploitation by the fishery may be a possible influence on
growth. Blue rockfish live at relatively shallow depths
where environmental and biological factors may have a
greater influence on their populations.
Although blue rockfish display a possible latitudinal
trend in growth rate between California and Oregon,
within California no latitudinal trend in growth rates was
observed. Specimens from both the southern CPFV sample
and the northern speared sample areas had translucent
zone completion by 1 May, which was consistent with the
findings of Miller and Geibel (1973) using ages from scales.
Individual fish in our study also had similar maximum ag-
es and maximum fish lengths in the north and south areas.
4bU
Males
400
350
300
/"""^
bpearea
250
200
/
150
^-
100
50
10
20
30
40
50
1° '*^° Females
400
350
300
250
200
150
100
50
10
20 30
Age (yr)
40
50
Figure 6
Von Bertalanffy growth models for male and female
blue rockfi.sh ^Sebastes mystinus) from the CPFV
(thick Hne) and speared samples (narrow line).
Individual data points are plotted for blue rockfish
from CPFV (open circle) and speared samples (open
triangles).
No latitudinal trend in growth rates was observed over
the 280 km between the centers of the two sampling ar-
eas. Although growth rates varied throughout their study
area from Half Moon Bay in the north to Morro Bay in
the south. Miller and Geibel (1973) likewise observed no
latitudinal trend in growth for blue rockfish.
Blue rockfish have average maximum ages and growth
rates when compared to other rockfish species. Maximum
ages for rockfishes {Sebastes spp.) range from 12 years for
the relatively small calico rockfish to 205 years for rough-
eye rockfish, one of the largest species (Cailliet et al., 2002;
Love et al., 2002). According to Love and Johnson (1998),
of the 38 species most accurately aged, most lived to more
than 40 years. Love et al. (1990) found growth rates for
three species that share the blue rockfish habitat (black,
/fe=0.12-0.21/year;yellowtail,^=0.16-0.20/year; and widow
rockfish, /j=G.14-0.22/year) to be similar to that for blue
rockfish (k=0.n years). Mean k values for rockfish varied
806
Fishery Bulletin 101(4)
500
400
^ 300
200
100
Females (OR, McClure, 1982)
Males (OR, McClure, 1982)
■Miller and Geibel (1973)
-MacGregor (1983)
9
Age (yr)
12
15
Figure 7
Growth models from five published studies of blue rockfish (Sebastes mystinus).
from 0.04/year for female silvergrey rockfish to 0.62/year
for the shorter-lived dwarf Puget Sound rockfish, with the
average range of ^ values occurring from 0.1 to 0.3/years
(Love et al,, 1990; Beckman et al., 1998). This considerable
longevity and relatively slow growth rate have significant
effects on the ability of many rockfish stocks to withstand
exploitation.
The age and growth relationships described in this
study indicate that both recruitment of blue rockfish to
the fishery and their maturity occur at younger ages than
previously reported. Blue rockfish enter the fishery at a
size of approximately 200 mm (Miller et al., 1967; Miller
and Geibel, 1973). This length equates to ages of 2-4 years
as determined in our study compared to 3-5 years as es-
timated by Miller et al. (1967). The new estimates for age
at which 50% of individuals are mature (using fish lengths
from Miller and Geibel, 1973) are even more striking: our
estimated age at 50% maturity is 5-6 years for males and
5 years for females, whereas estimates from Miller and
Geibel ( 1973) and Echeverria ( 1986) were 7 years for males
and 7-8 years for females. Similarly, the youngest mature
males and females in these early studies were 4-5 years,
whereas we estimated the age to be 3 years.
The changes observed in our study in age-at-length,
maximum age, recruitment age, and age at 50% matu-
rity have important implications for stock assessments.
Accurate information on age composition, weight-at-age,
age specific availability to the fishery, and maturity-at-age
is crucial to the proper functioning of the stock synthe-
sis model (Methot, 1990), which is used for Pacific coast
groundfish. If incorrect age data are used, it could lead to
erroneous estimates of population size, and subsequently
to either overfishing or an unnecessary reduction in allow-
able catch.
Acknowledgments
We would like to thank all the port samplers who collected
the CPFV catch data. We also thank the crew and scientific
personnel aboard the RV David Starr Jordan for collecting
samples. We thank James Chess, Edmund Hobson, Dan
Howard, and Kelly Silberberg for braving the cold waters
of the Pacific Ocean to collect the nearshore specimens.
We also thank Churchill Grimes and Mary Yoklavich for
their many constructive comments and the reviewers of
this manuscript.
Literature cited
Beamish, R. J., and D. A. Fournier.
1981. A method of comparing the precision of a set of age
determinations. Can. J. Fish. Aquat. Sci. 38:982-983.
Beamish, R. J., and G. A. MacFarlane.
1987, Current trends in age determination methodology. In
Age and growth of fish (R. C. Summerfelt and G. E. Hall,
eds), p. 15-42. Iowa State Univ, Press, Ames, lA,
Beckmann, A. T., D, R. Gunderson, B. S. Miller, R. M. Buckley,
and B. Goetz.
1998. Reproductive biology, growth, and natural mortality
of Puget Sound rockfish, Sebastes emphaeus (Starks, 1911).
Fish, Bull. 96:352-356.
Laidig et al : Age and growth of Sebastes mystinus
807
Boehlert, G. W., and R. F. Kappenman.
1980. Variation of growth with latitude in two species of
rockfish (Sebastes pinniger and S. diploproa) from the
northeast Pacific Ocean. Mar Ecol. Prog. Ser 3:1-10.
Brown, I. W., and W. D. Sumpton.
1998. Age, growth and mortality of redthroat emperor Leth-
rinus mintatus (Pisces: Lethrinidae) from the southern
Great Barrier Reef, Queensland, Australia. Bull. Mar
Sci. 62:905-917.
Cailliet, G. M., A. H. Andrews, E. J. Burton, D. L. Watters,
D. E. Kline, and L. A. Ferry-Graham.
2002. Age determination and validation studies of marine
fishes: do deep-dwellers live longer? Exp. Gerontology 36:
739-764.
Campana, S. E.
2001. Accuracy, precision and quality control in age deter-
mination, including a review of the use and abuse of age
validation methods. J. Fish. Biol. 59:197-242.
Chilton, D. E., and R. J. Beamish.
1982. Age determination methods for fishes studied by the
groundfish program at the Pacific Biological Station, 102 p.
Can. Spec. Publ. Fish. Aquat. Sci. 60.
Crabtree, R. E., and L. H. Bullock.
1998. Age, growth, and reproduction of black grouper, Mycte-
roperca bonaci, in Florida waters. Fish. Bull. 96:735-753.
Draper, N., and H. Smith.
1981. Applied regression analysis, 2d ed., 709 p. John Wiley
and Sons, New York, NY.
Echeverria, T.
1986. Sexual dimorphism in four species of rockfish genus
Sebastes (Scorpaenidae). Environ. Biol. Fishes 15:
181-190.
Echeverria, T., and W H. Lenarz.
1984. Conversions between total, fork, and standard lengths
in 35 species of Sebastes from California. Fish. Bull. 82:
249-251.
Fraidenburg, M. E.
1980. Yellowtail rockfish, Sebastes flavidus. length and age
composition ofTCalifornia, Oregon, and Washington in 1977.
Mar Fish. Rev 42:54-56.
Hare, S. R., and N. J. Mantua.
2000. Empirical evidence for North Pacific regime shifts in
1977 and 1989. Prog. Oceanogr. 47:103-145
Karpov, K. A., D. P Albin, and W. H. Van Buskirk.
1995. The marine recreational fishery in northern and cen-
tral California, 192 p. Calif Dep. Fish Game Fish Bull.
176.
Kimura, D. K., R. R. Mandapat, and S. L. Oxford.
1979. Method, validity, and variability in the age determi-
nation of the yellowtail rockfish (Sebastes flavidus) using
otoliths. J, Fish. Res. Board Can. 36:377-383.
Lea, R. N., R. D. McAllister, and D. A. VenTresca.
1999. Biological aspects of nearshore rockfishes of the genus
Sebastes from central Cahfomia, 109 p. Calif Dep. Fish
Game Fish Bull. 177.
Love, M. S., and K. Johnson.
1998. Aspects of the life histories of grass rockfish, Sebastes
rastrelliger . and brown rockfish. S. auriculatus, from south-
ern California. Fish. Bull. 87:100-109.
Love, M. S., P. Morris, M. McCrae, and R. Collins.
1990. Life history aspects of 19 rockfish species (Scorpae-
nidae: Sebastes) from the Southern California Bight, 38 p.
NOAATech. Rep. NMFS 87
Love, M. S., M. Yoklavich, and L. Thorsteinson.
2002. The rockfishes of the northeast Pacific, 405 p. Univ
California Press, Berkeley, CA.
MacGregor, J. S.
1983. Growth of the blue rockfish (Sebastes mystinus).
CALCOFI (Calif Coop. Ocean. Fish. Investig.) Rep. XXIV:
216-225.
Mayo, R. K., V. M. Gifford, and A. Jearid Jr
1981. Age validation of redfish, Sebastes marinus (L.), from
the Gulf of Maine-Georges Bank region. J. Northwest Atl.
Fish. Sci. 2:13-19.
McClure, R. E.
1982. Neritic reef fishes off central Oregon: aspects of
life histories and recreational fishery. M.S. thesis, 94 p.
Oregon State Univ., Corvallis, OR.
Methot, R. D.
1990. Synthesis model: an adaptable framework for analy-
sis of diverse stock assessment data. Bull. Int. North Pac.
Fish. Coram. 50:259-277.
Miller, D. J., and J. J. Geibel.
1973. Summary of blue rockfish and lingcod life histories;
a reef ecology study; and giant kelp, Macrocystis pyrifera,
experiments in Monterey Bay, California, 137 p. Calif
Dep. Fish Game Fish Bull. 158.
Miller, D. J., M. W. Odemar, and D. W. Gotshall.
1967. Life history and catch analysis of the blue rockfish
(.Sebastodes mystinus) off central California, 1961-1965.
Calif Dep. Fish Game Mar Res. Operations Ref 67-14:
1-130.
Munk, K. M.
2001. Maximum ages of groundfish in waters off Alaska
and British Columbia and considerations of age deter-
mination. Alaska Fish. Res. Bull. 8:12-21.
Pearson, D. E.
1996. Timing of hyaline-zone formation as related to sex,
location, and year of capture in otoliths of the widow rock-
fish, Sebastes entomelas. Fish. Bull. 94:190-197.
Pearson, D. E., and J. E. Hightower
1991. Spatial and temporal variability in growth of widow
rockfish (Sebastes entomelas), 47 p. NOAA Tech. Memo.,
NOAA-TM-NMFS-SWFSC 167.
Pearson, D. E., J. E. Hightower, and J. T. H. Chan.
1991. Age, growth, and potential yield for shortbelly rockfish,
Sebastes jordani. Fish. Bull. 89:403-409.
PFMC (Pacific Fishery Management Council).
2001. Status of the Pacific coast groundfish fishery through
2001 and recommended acceptable biological catches for
2002. Pacific Fishery Management Council, Portland,
OR.
Ratkowsky, D. A.
1983. Nonlinear regression modeling, 276 p. Marcel
Dekker, New York, NY.
Rogers, J. B., M. Wilkins, D. Kamikawa, F. Wallace, T. Builder,
M. Zimmerman, M. Kander, and B. Culver
1996. Status of the remaining rockfish in the Sebastes
complex in 1996 and recommendations for management
in 1997. Appendix E: Status of the Pacific coast groundfish
fishery through 1996 and recommended biological catches
for 1997: stock assessment and fishery evaluation, 59 p.
Pacific Fishery Management Council, Portland, OR.
Schnute, J.
1981. A versatile growth model with statistically stable
parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140.
Six, L. D., and H. F. Horton.
1977. Analysis of age determination methods for yellowtail
rockfish, canary rockfish, and black rockfish off Oregon.
Fish. Bull. 75:405-414.
808
Fishery Bulletin 101(4)
Stanley, R. D., and B. R. Melteff.
1987. A comparison of age estimates derived from the sur-
face and cross-section method of otoUth reading for Pacific
ocean perch (Sebastes alutus). Lowell Wakefield fisheries
symposium: proceedings of the international rockfish sym-
posium, Anchorage, Alaska, USA, Oct. 20-22, 1986 Alaska
Sea Grant Rep. 87-2, p. 187-196. Univ. Alaska, Anchorage.
AK.
Wales. J. H.
1952. Life history of the blue rockfish, Sebastodes mystinus.
Calif Fish Game 38(4):485-498.
Wilson, C. D., and G. W. Boehlert.
1990. The effects of different otolith ageing techniques on
estimates of growth and mortality for the splitnose rockfish,
Sebastes diploproa, and canary rockfish, S. Pinniger. Calif
Fish Game 76:146-160.
809
Abstract— The reproductive activity
and recruitment of white mullet (Mugil
curema) was determined by observa-
tions of gonad development and coastal
juvenile abundance from March 1992
to July 1993. Adults were collected
from commercial catches at three sites
in northeastern Venezuelan waters.
Spawning time was determined from
the observation of macroscopic gonadal
stages. Coastal recruitment was deter-
mined from fish samples collected
biweekly by seining in La Restinga
Lagoon. Margarita Island, Venezuela.
The examination of daily growth rings
on the otoliths of coastal recruits was
used to determine their birth date
and estimate the period of successful
spawning. Fish with mature gonads
were present throughout the year but
were less frequent between September
and January when spawning individu-
als migrated offshore. In both years,
juvenile recruitment to the lagoon
was highest between March and June
when high densities of 25-35 mm juve-
niles were observed. Back-calculated
hatching-date frequency distributions
revealed maximum levels of successful
spawning in December-January that
were significantly correlated with peri-
ods of enhanced upwelling. The rela-
tion between the timing of successful
spawning and the intensity of coastal
recruitment in white mullet was likely
due to variations in food availability for
first-feeding larvae as well as to varia-
tions in the duration of the transport
of larvae shoreward as a result of vary-
ing current conditions associated with
upwelling.
Reproduction and recruitment of
white mullet iMugil curema) to a tropical lagoon
(Margarita Island, Venezuela) as revealed by
otolith microstructure*
Baumar J. Marin E.
Antonio Quintero
Institute Oceanografico de Venezuela
Universidad de Oriente
Cumana 6101
Edado Sucre, Venezuela
E-mail address (for B. J Mann E.): bmann@sucre. udo.edu. ve
Dany Bussiere
Julian J. Dodson
Departement de biologle
Unlverslte Laval,
Ste-Foy
Quebec, Canada GIK 7P4
Manuscript approved for publication
10 June 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:809-821
White mullet (Mugil curema) is a wide-
spread coastal pelagic fish occurring
from Massachussetts to southern Brazil.
Considered to be catadromous, the juve-
nile fish recruit to lagoons and estuaries
following a period of offshore spawning
(Blaber, 1987; Ibahez-Aguirre, 1993;
Ditty and Shaw, 1996). White mullet
is an important economic resource
supporting many small communities
through both fishing and aquaculture
(AIvarez-Lajonchere, 1982; Gomez and
Cervigon, 1987). Small schools of mullet
are captured with gill and "atarraya"
nets near the coast and in neritic waters
and between 300 and 400 metric tons
are sold annually on Margarita Island,
Venezuela.
Reproductive periodicity in white
mullet varies over its geographic distri-
bution. Several authors have reported
protracted or continuous reproduction
in tropical waters and generally two
spawning peaks per year (Jacot, 1920;
Anderson, 1957; Angell, 1973; Moore,
1974; AIvarez-Lajonchere, 1976, 1980;
Yaiiez-Arancibia, 1976; Rodriguez and
Nascimento, 1980; Garcia and Bus-
tamente, 1981; Franco, 1986; Ibanez-
Aguirre, 1993). Figure 1 summarizes
previous work describing the spawning
periods of M. curema based on gonad
development and estimated according
to the arrival of juveniles in the coastal
zones. The spawning period is quite
variable. Angell (1973) suggested that
schooling occurs in coastal areas just
prior to the offshore spawning migra-
tion and that the departure of individu-
als for the spawning grounds causes a
reduction of the gonadosomatic index in
the nearshore populations. Moore ( 1974)
also reported that during the spawning
period fully ripe fish are rare in coastal
collections. Despite these studies, little
is known of the factors influencing re-
productive patterns of the white mullet.
Ibanez-Aguirre (1993) suggested that
the timing of reproduction in M. curema
in Tamiahua Lagoon, Mexico, is an ad-
aptation to avoid competition with juve-
niles of the conspecific Mugil cephalus.
In areas of favorable thermal regimes,
M. curema may penetrate a wider range
of salinities and competitively exclude
M. cephalus (Moore, 1974).
The periodicity of white mullet re-
production may be related to environ-
mental variability that signals periods
of optimal early growth and survival.
Stability of the water column and suit-
' Contribution of Quebec-Ocean, Pavilion
Alexandre-Vachon, Local 2078, Universite
Laval, Quebec, Qc. GIK 7P4.
810
Fishery Bulletin 101(4)
Location ^ FMAMJJ ASOND
I I I I I I I I I I I I
Cariaco Gulf, Venezuela
(Franco, 1986) 1.2
La Restinga Lagoon, Venezuela
(Angell, 1973) 1,2
Patanemo Lagoon, Venezuela
(Blanco, 1985) 2
Tunas de Zaza. Cuba
(Alvarez-Lajonchere, 1980)2
Cuba
(Garcia and Bustamante, 1981) 3
Guerrero Stale Lagoons, Mexico
(Yanez-Arancibia, 1976) 3
Gulf of fdlexico
(Ibanez-Aguirre, 1993) 3
Southern Texas, USA
(Moore, 1974) 3
South Florida, USA
(Anderson, 1957) 3
North Carolina, USA
(Jacol, 1920) 3
Figure 1
Spawning periods for white mullet (Mugil curema) in tropical waters, based on the
histological observations of gonads (1), macroscopic observations of gonads (2), and
estimated from the arrival of juveniles in coastal areas (3). Thin horizontal lines indi-
cate periods of continuous but minor spawning between periods of peak spawning.
able food in coastal lagoons, river deltas, and estuarine
mangrove areas have been identified as important factors
influencing the recruitment of juvenile Mugilidae (Yafiez-
Arancibia, 1976; Blaber and Blaber, 1980; Blaber, 1987;
Vieira, 1991), Based on macroscopic gonad observations of
schools of white mullet captured offshore, Etchevers (1974)
proposed that the spawning of white mullet recruiting
along the southern coast of Margarita Island, Venezuela,
occurs between La Tortuga Island and Margarita Island
in the vicinity of the 1000-m deep Cariaco trench (Fig. 2).
Seasonal environmental variability in this area is mainly
generated by the alternation between upwelling during the
dry season and freshwater discharge during the wet season
(Gomez, 1983;Muller-Kargeret al., 1989). The rainy season
strongly influences the eastern Caribbean as freshwater
plumes from the Amazon and Orinoco Rivers lower salini-
ties throughout the region. Both upwelling and freshwater
runoff produce intense peaks in coastal primary production
(Gines 1972; Ferraz-Reyes et al,, 1987; Miiller-Karger et
al,, 1989), which could influence spawning periodicity and
recruitment success. The purpose of this study was to docu-
ment the periodicity of reproduction and recruitment of A/,
curema along the southern coast of Margarita Island and
to examine their relationship with environmental signals,
particularly those associated with upwelling.
Methodological advances in counting daily growth
increments in otoliths of marine fishes (Pannella, 1971;
Campana and Nielsen, 1985) have greatly aided studies
of the age, growth, and recruitment of larval and juvenile
fishes (Wilson and Larkin, 1980; McBride and Conover,
1991; Jenkins and May, 1994; Sirois and Dodson, 2000).
For the striped mullet (M, cephalus), a close relative of the
white mullet, Radtke (1984) showed that the first incre-
ment is formed one day after hatching and that additional
increments are formed daily thereafter. Daily growth rings
have also been demonstrated in laboratory studies for M.
so-iiiy by Li et al, (1993). In the present study, we examined
the microstructure of the otoliths of juveniles recruiting
to a coastal lagoon in order to back-calculate the date of
hatching and hence the time of successful spawning. We
first validated that otolith growth increments of juvenile
M. curema were formed daily.
Material and methods
Reproductive periodicity was documented from samples
of adult fish taken monthly from commercial catches in
three fishery zones in Venezuela: 1) the Chacopata zone,
located between Chacopata lagoon and Coche and Cubagua
Islands; 2) the Cariaco Gulf zone; and 3) the Margarita zone
located along the southern coast of Margarita Island and
the northern coast of Cubagua Island (Fig, 2), Measure-
ments of water temperature, salinity, and rainfall were
Marin et al,: Reproduction and recruitment of Mugil curema
811
n
Sea
La Tonuga
>
— ■
/
—
150 m
\
\
500
s. lOOOm
lOOOm \
Cariaco Trench ^
^ (western part)
— 1 —
65°
Santa Fe Gulf
Figure 2
Map of the northeastern coast of Venezuela showing locations mentioned in the text.
collected periodically throughout the entire study period
at the La Salle meteorological station, next to La Restinga
Lagoon on Margarita Island.
Total (TL) and standard (SL) lengths of adult mullet
were measured to the nearest 0.5 cm and total and gut-
ted body mass were recorded to the nearest 0.1 g. Sexual
maturity was determined by observation of the gonads and
gonadal stages were classified as follows:
Stage I Ovaries transparent and inconspicuous, whit-
ish-yellow in color and rounded with a small
diameter. Testes longer than ovaries and rib-
bonlike in form.
Stage II Ovaries rounder and wider than in stage I, and
yellow in color. Testes thinner, and wider than
stage I, but still with thin edges and a ribbon-
Uke form; white in color.
Stage III Ovaries large, pale yellow, smooth in appear-
ance, turgid, and round. Ovocytes easily distin-
guished macroscopically (as granular). Testes
milky-white in color (bright), turgid, and wider
in appearance and having thicker edges than in
stage II.
Stage rV Spawned (spent) ovaries purple and wrinkled in
appearance. Testes whitish, or transparent with
white patches, and wrinkled in appearance.
Recruitment periodicity was documented from samples
of juveniles seined at semimonthly intervals at the mouth
of La Restinga Lagoon (Fig. 2). The 2-cm mesh beach seine
measured 1.5 m deep and 50 m long. Juvenile white mullet
were distinguished from other juvenile mullets according to
the descriptions of Alvarez-Lajonchere et al. ( 1976). White
mullet juveniles were characterized by a scaly gray appear-
ance as opposed to the shiny metallic gray appearance of a
sympatric mullet species (Mugil incilis). For white mullet,
recruitment is defined as the appearance of juveniles in
coastal areas (Vieira, 1991). We calculated catch per unit
of effort (CPUE) as the number of juveniles per seine haul.
For all samples, standard length of fish was measured to
the nearest 1 mm. Otolith analysis was restricted to one
sampling period per month. After examining size-fre-
quency distributions of juveniles captured in the lagoon,
the otoliths of approximately 20 individuals representing
all cohorts collected on a given sampling date were ana-
lyzed. The otoliths (sagittae) were removed with needles,
rinsed in water, and then attached to strips of masking
tape. The otolith was then sanded to obtain a transversal
section (Fig. 3) with a thickness of approximately 20 ^m
by using the technique described by Secor et al. ( 1992) and
a metallurgic jig adapted from Neilson and Geen (1986).
Readings of the number of increments were made along the
curvilinear surface running from the nucleus to the edge of
the otolith (Fig. 3). Because daily growth increments were
812
Fishery Bulletin 101(4)
Regular
rings
(141)
Central
rings
(14)
Nucleus
B
Central
Regular
nngs
1 1 null I I I I
1 1 mill I I I I
1 1 iiiiii I I I I
1 1 mill I I I I
I I I I I I I I I I I I I I
I I I I I I I I I I I I I I
II I I I I I I I I I I I I
II I I I I I I I I I I I I
Nucleus «r
Cut plane
Figure 3
(A) Photomicrograph in optical microscopy of the posterior radius of a polished otolith
from a 180-day-old juvenile white mullet iMugil curema). Different sequences of rings are
indicated which result from a horizontal plane not aligned with the changing growth plane
of the otolith. Numbers of rings in each sequence are shown in parentheses. (Bi Schematic
presentation of the changing growth plane of the otolith, the cut plane, and the resultant
increment distortion (black rectangles represent the distance between adjacent rings).
less consistent in the anterior field, otolith counts were
always made along the posterior radius of the sagittae. We
also measured the area of the nucleus, which represents
the prehatch zone. All measurements were made under a
microscope which was connected to an image analyzer and
computer.
To evaluate the error in counting growth increments
on otoliths, one reader made 7 independent counts of the
number of growth increments on otoliths obtained from
11 juveniles representing the size-range of sampled fish.
The number of growth increments ranged from 63 to 289
(mean=147) and the mean coefficient of variation was
8.71% (SD=2.13%) (Fig.4). We therefore considered an
error of approximately 10% for the counts of growth incre-
ments. In applying the technique to the subsamples of the
different cohorts sampled in the lagoon, at least two counts
were made for each otolith. All counts were made by the
same person.
To evaluate if otolith growth increments were formed
daily, we read the otoliths of juveniles sampled on succes-
sive sampling dates and compared the average increase
in the number of otolith increments to the number of days
between samplings (Struhsaker and Uchiyama, 1976; Jor-
dan, 1993; Jenkins and May, 1994). Birthdate was obtained
by subtracting the number of daily growth rings on otoliths
from the date of capture. We used the hatching mark as
defined for M. cephalus (Radtke, 1984) and M. iso-iuy {Li
et al., 1993) to locate the hatching mark on the otoliths of
the white mullet.
Knowing that white mullet embryos hatch from 24 to
40 hours after fertilization (Anderson, 1957; Houde et al.,
1976), we back-calculated hatching dates of recruits to es-
timate when successful spawning occurred. We examined
the relationship between the spawning dates of recruits
and an index of the intensity of upwelling. In calculat-
ing this index, we determined wind stress based on data
fnmi Fundacion La Salle, Margarita Island, and Cumana
Airport meteorological stations. The upwelling index (UI)
was based on Bowden's (.1983) theoretical calculations as
follows:
Mann et al ■ Reproduction and recruitment oi Mugil curema
813
Figure 4
Number of growth increments (counted on seven independent occasions) on otoliths
from 1 1 juvenile white mullet of varying size. Means, standard errors, and standard
deviations are illustrated.
Ul=-
100
where f = Coriolis parameter;
Tjj. = surface wind stress; and
p„, = average density of the water (1025 kg/m'^).
The term /'was calculated as
/■=2ojsin(/;),
where co = angular velocity of rotation of the earth
(7.29xl0-5s [seconds]): and
/j = latitudinal position at the place i.
The term t^^ represents surface wind stress measured in
the .r-axis perpendicular to the coast (Bowden, 1983), often
considered in terms of the empirical equation
Tj^ = ^ X p_j X W^,
where k = empirical drag coefficient (1.11 to 3.25, as a
function of wind velocity; Bowden 1983);
p^ = mean density of the air ( 1.25 kg/m^); and
W = wind velocity.
pling period during each of the 18 months of the study by
applying the frequency distribution of birth dates of aged
juveniles to the total catch for that date. A total of 398
juveniles were aged by otolith analysis. If x% of aged fish
captured on a given date were hatched on Julian Day y,
this percentage was applied to the total catch of juveniles
for that sampling date. Secondly, all fish hatched on a given
day were summed across the 18 monthly sampling dates.
This frequency distribution was then correlated with the
distribution of UI estimates over the same period of time
as that of the birth dates.
Before proceeding with correlation, trends in birth date
and UI data series were described by using a smooth-
ing spline. The spline fit uses a set of smoothly spliced
3'''' degree polynomial segments (Simonoff, 1996; JMP®
software, version 3.2.1, SAS Institute, Gary, North Caro-
lina). Predicted values were correlated with the raw data
points in order to optimize the value of lambda used to
fit the smoothing spline. Increasing the value of lambda
increases the degree of smoothing but weakens the correla-
tion between predicted and raw data. Pearson correlation
and cross-correlation functions were used to describe the
temporal relationship between upwelling and the date of
hatching offish recruited to the coastal lagoon.
The drag coefficient, k, changes as a function of wind
velocity and gives values equivalent to those of Bakun et
al.(1974).
The relationship between upwelling and the birth dates
of successful recruits (captured in La Restinga Lagoon)
was determined in two steps. First, we calculated the
birth dates of juveniles captured during one monthly sam-
Results
The white mullet surveyed in the commercial catches
measured from 4 to 36 cm SL. The largest fish were from
the Chacopata zone where the most abundant sizes classes
were those from 22 to 30 cm. The most abundant sizes in
the Cariaco Gulf and Margarita Island zones were 20 to
814
Fishery Bulletin 101(4)
26 cm and 18 to 26 cm, respectively. The mullet from the
Margarita zone were mainly juveniles and small adults
(Fig. 5).
Maturity and reproductive periodicity
An examination of gonad maturity revealed that 90% of
the male and female mullet in the Margarita zone were
immature or at developing stages (I and II) and only 10%
in developed and spawning stages (III and IV). In contrast,
53.8% of females and 42.8% of males of the population of
generally larger mullet sampled at Chacopata were in
developed and spawning stages. Finally in the Cariaco Gulf
zone, where the fish were of intermediate size, compared
to fish at Margarita and Chacopata, 41.2% of females and
32.7% of males were in developed and spawning stages.
Throughout our study, sexually mature (stage-Ill) fish
were present in the samples from the Chacopata fishery
(Fig. 6), and their abundance showed a marked seasonal
pattern. Mature and spent fish were least abundant
(<25% of the population) between September and January
and most abundant from April to August 1992 and May to
June 1993. In contrast, in the Margarita fishery, immature
(stage-I) fish dominated the samples and mature fish only
occurred sporadically (Fig. 6). Finally, in the Cariaco Gulf
zone, immature and maturing fish (stage-I, and stage-ID
generally dominated the population, except in July when
mature fish became abundant.
Otolith microstructure
The otolith of M. curema had a round nucleus with a mean
radius of 9.26 ^m (95% confidence interval (CI)=0.54, n=8),
and the dark area in the center had a mean diameter of
4.97 jjm (CI=0.82, n=8). The otolith was round during the
larval period and became ovoid when mullet reached 10-12
mm (SL). In the early juvenile stage (18-20 mm SL), the
otolith was strongly elongated and the anterior end was
arrow-like. In juveniles (>20 mm SL), the otolith was always
thin, concave, and umbrella shaped in form (Fig. 3).
Validation of otolith increment lines
The otolith increments counted for the first strong cohort
present in the lagoon during March and April 1992 dem-
onstrated that the average number of increments added
during the 14-day interval between sampling collections
was close to 14 days (Table 1). This indicated that otolith
increments were formed daily, as observed in other species
of mullet.
Chacopata
_LL
Margarita
160
140
120
100
80
60
40
20
0
160
140
120
- 100
I 80
C
I 60
40
20
0
180
160
140
120
100
80
60
40
20
0
12 14 12 16 18 20 22 24 26 28 30 32 34 36 38
Standard length (cm)
Figure 5
Length-frequency distributions of white mullet surveyed
in the commercial catches from the three fishery zones
of northeast Venezuela.
Cariaco Gulf
Juvenile recruitment
The catch-per-unit-of-effort measurements for juveniles
captured in La Restinga Lagoon demonstrated a seasonal
pattern and high recruitment from March to early July
1992 and from late March to May 1993, and low recruit-
ment during the remaining months (Fig. 7). The recruit-
ment peak in 1992 was more than twice that in 1993. The
periods of strong recruitment were associated with the
rainy season in northeastern Venezuela (Fig. 7), as previ-
ously reported by Okuda et al. (1978), Gomez (1983), and
Ferraz-Reyes(1989).
The discontinuous length-frequency distributions of
juveniles sampled biweekly suggested the presence of four
cohorts in the lagoon during the study period (Fig. 8). Two
cohorts were present on 5 March 1992, the first sampling
date. The cohort of smaller juveniles, referred to as the
first cohort, had a mean length of 29.8 mm (range of 18 to
Mann et a\ Reproduction and recruitment of Mugil curema
815
100
g 80
o. 60
B 40
I 20
100
80
60
40
20
0
Chacopata
Cariaco Gulf
Margarita
MAMJ JA sond;j FMAMJ J
1992 ■ 1993
Figure 6
Percent composition of four maturity stages (I to IV, see text
for definitions) of Mwg;/ curema (males and females combined)
captured from Margarita, the Cariaco Gulf, and the Chacopata
fisher)' zones of northeast Venezuela.
36 mm) and varied in age from 50 to 70 days (Fig. 9). The
cohort oflarger juveniles, referred to as the second cohort,
had a mean length of 105 mm (range of 90 to 130 mm), and
otolith analysis indicated an age of 160 to 240 days (Fig. 8).
This cohort was present until May 1992 and was largely
absent thereafter (Fig. 7). A third cohort was first observed
in mid October 1992. It measured from 50 to 100 mm in
length, overlapping the length distribution of cohort 1. As
such, no clear distinction could be made between these two
cohorts on these dates. The third cohort became more dis-
tinct in December 1992 and January 1993 as individuals
from the first cohort left the lagoon. These individuals were
aged from 95 to 200 days old in December. This cohort was
present until approximately April 1993 and disappeared
thereafter (Fig. 9). Finally, a fourth cohort first appeared in
March 1993 and consisted offish of similar size and age as
the first cohort observed in March of the previous year.
The back-calculated hatching dates showed that the sec-
ond, older cohort observed in the lagoon in March 1992 was
composed of mullet that had hatched between August and
October 1991 (Fig. 10). The younger cohort in the March
1992 sample originated from continuous hatching from
Table 1
Validation of daily increment formation in the otoliths of
juvenile white mullet (Mugil curema) sampled at 14-day
intervals in March and April 1992. /i= number of otoliths in
sample, mean age (standard deviation) is given in days on
date of capture, and "difference" is the difference in mean
age between successive sampling dates.
Sampling dates
5 March
1992
19 March
1992
2 April 2
1992
15 20 19
56.46(4.81) 70.85(9.88) 84.3(11.24)
14.39 13.45
Mean age (SD)
Difference
late December 1991 to late March 1992. The back-calculated
hatching dates indicated that larval production of successful
recruits was almost absent during April and May 1992 but
small increases were observed in June and July 1992. The
third cohort, which first appeared in September 1992, was
816
Fishery Bulletin 101(4)
jLil
50 ST
MAMJJASOMDJFMAMJJ
1992 1993
Figure 7
Histogram of monthly abundance of juvenile white mullet as indicated by catch
per unit of effort (CPUE) from semimonthly sampling in La Restinga lagoon from
March 1992 to July 1993. The continuous black line presented in the first annual
cycle represents the mean monthly rainfall averaged over the period 1985-93.
Table 2
Correlation analyses of hatching
weakens the r^ values between
contrast, increasing lambda va'
BD = birth date, UI = upwelling
date and upwelling index data series. Increasing lambda values
abserved and predicted values of birth date and upwelling index
ues strengthen the correlation (Pearson's correlation) between
index
used to fit the
over time (col
the two data
smoothing spline
umns 2 and 3). In
series (column 4).
Data series
Hatching dates
(r2, P)
Upwel!
(r
ing index
2,P)
BD versus UI
(r,P)
Raw data
0.28,
<0.05
Smoothing spline
A=l
0.82
<0.000
0.86,
<0.000
0.36,
<0.000
Smoothing spline
A= 10
0.73,
<0.000
0.74,
<0.000
0.41,
<0.000
Smoothing spline
A= 100
0.67,
<0.000
0.63,
<0.000
0.45,
<0.000
Smoothing spline
A = 1000
0.64,
<0,000
0.54,
<0.000
0.52,
<0.000
Smoothing spline
A = 10,000
0.56,
<0.000
0.45,
<0.000
0.57,
<0.000
Smoothing spline
A= 100,000
0.48,
<0.000
0.40,
<0.000
0.64,
<0.000
composed of individuals that had hatched between June
and August 1992. Finally, individuals in the fourth cohort,
which first appeared in the lagoon in March 1993, were fish
that had hatched in January and February 1993.
The hatching dates of recruits coincided with periods
of increasing upwelling, particularly during January and
February of 1992 and 1993 (Fig. 10). The use of increa.sing
levels of lambda to fit the smoothing spline increasingly
weakened the correlations between predicted and observed
values of birth dates and UI index and strengthened the
correlations between birth dates and the upwelling index
(Table 2). Choosing a lambda value of 10 resulted in r'^
values greater than 0.70 (/'<0.000) between the raw and
predicted data series and in an rvalue of 0.41 (P<0.000)
between birth date and the upwelling index. Cross-corre-
lation analysis between these two series revealed that the
strongest correlation (r=0.52, P<0.000) occurred when the
upwelling index lagged behind birth dates by 8 days and
by 46 days. These lag periods reflect the coincidence of the
major peak of upwelling with the two peaks of birth dates
that are separated by approximately 35 days. Given the
estimated 10-day error associated with aging otoliths, the
8-day lag cannot be interpreted.
Discussion
Our sampling of white mullet in the coastal waters of
northeastern Venezuela revealed the presence of mature
fish throughout the year, but abundance was lowest
Mann et al,: Reproduction and recruitment of Mugil curema
817
between August and January. Mullet from the
Margarita zone were small (4 to 36 cm in SL)
and mostly immature (>80%). Because size at
maturity of white mullet is 24 cm (Marin and
Dodson, unpubl. data), most of the adults in
the lagoon were probably in their first spawn-
ing cycle. Similarly, mullet from the Cariaco
Gulf zone also appeared to be young adults in
their first spawning cycle. In contrast, mullet
from the Chacopata zone were larger and
generally in more advanced stages of gonadal
development. This finding suggests that the
Chacopata mullet were part of a prespawning
aggregation, and the location of the aggrega-
tion agrees with the more offshore location of
the Chacopata fishery.
Because white mullet spawn offshore (Ja-
cot, 1920; Anderson, 1957; Ditty and Shaw,
1996), the small proportion of mature fish
in the coastal fisheries from July to April is
likely explained by the migration of adults to
the offshore spawning grounds. If this is so,
reproduction is not indicated by an increase in
the frequency of fish in advanced stages, but
rather is associated with the disappearance of
maturing and mature fish from coastal areas.
The disappearance of fish in advanced stages
from coastal areas as the spawning season ap-
proaches was also reported by Angell (1973)
and Moore (1974). The analysis of birth dates
of juveniles sampled in the La Restinga La-
goon indicates that successful spawnings are
concentrated in the periods of increased up-
welling and also coincide with the end of the
rainy season. The spawning season may or may
not be concentrated at these times but larvae
that hatch during upwelling events are most
likely to successfully recruit to the lagoon.
Although reproduction in tropical fishes
is often protracted, peaks in successful
spawning may nevertheless be initiated by
environmental cues (Redding and Patino,
1993). The white mullet possibly uses tem-
perature or other signals associated with
upwelling to synchronize its spawning with
upwelling events. The variations in the tim-
ing of recruitment of white mullet in differ-
ent geographical regions may be the result
of variation in the timing of favorable condi-
tions that enhance survival. Such conditions
may include increased primary production
(Ferraz-Reyes, 1983; Miiller-Karger et al.,
1989) or hydrographic mechanisms likely to
facilitate transport of larvae to coastal nurs-
ery areas (Blaber, 1987), so that survival is
increased. Populations likely have adequate
time to adapt to environmental conditions in
particular areas because local hydrographic
patterns develop over geological time scales
(Bakun, 1986; Sinclair, 1988; Heath, 1992).
Uj
a.
o
jlllu.
__1M.
miiL
noc
n = 416
iDnDnc
JlDnnnnnn n n n
_dQ
jmd
jaa
n = 103
jm
nn_
nnnr,n
nnn^nnnnnOnn r.
n = 88
,nnnn)
nflnn
Innr
_la_
DnJl
n = S7
-a □ □. Dnn nHn- ..
JUMbIjuu^
wn-.-n sx^M
■■■■■■111.
■■■■■Illll-
■ ■■ III
Uihi.
D n
JUlL
II 1
n = 4
ill
n = 104
Uu
n = 415
JlIlL
n = 252
Jim
iiliiiii.
ilii ..III
lUlllliIh^
^■lIUImlBllkMlL
JUI.Ulill.
-■Illl
n = 66
Mar-05
Mar-19
Apr-02
Apr-14
Apr-30
May-14
May-28
Jun-11
Jun-25
Jljl-08
Jul-23
Aug-05
Aug-20
Sep-03
Sep-17
Oct-02
OcM5
Ocl-29
Nov-10
Nov-26
Dec- 10
Dec-28
Jan-08
Jan-21
Feb-04
Mar-04
Mar-18
Apr-01
Apr-15
Apr-29
May-13
May-25
Jun-10
Jun-25
Jul-08
Jul-22
1992
1993
10 30 50 70 90 110 130 150 170 190 210
Standard length (mm)
Figure 8
Semimonthly size distributions of juvenile white mullet from March 1992
to July 1993 in La Restinga Lagoon. The abundance for each size class is
presented as the Log CPUE + 1. From top to bottom, open bars represent
cohort 1, black bars represent cohort 2, gray bars represent cohort 3, and
black bars represent cohort 4. Cohorts were identified by discontinuities
in size distributions of juveniles, n = number offish sampled.
818
Fishery Bulletin 101(4)
Strong offshore transport of surface waters
occurs during upwelling events, so that a rapid
metamorphosis to the demersal stage may be
critical for the coastal recruitment of white mullet.
This rapid metamorphosis is suggested for several
offshore spawning fishes with pelagic larvae that
subsequently recruit to estuaries (Creutzberg et
al., 1978; Heath, 1992) or that remain near the
bottom during ebb flow, once close to the coast,
thereby reducing offshore transport (Bartsch and
Knust, 1994). White mullet undergo metamorpho-
sis to the demersal stage 14 days after hatching
(Houde et al., 1976) at which time they would be
entrained in the inshore transported water that
occurs at depths greater than 50 m in the coastal
zone of northeastern Venezuela (Quintero, unpubl.
data). Several studies suggest that increased mor-
tality is caused by increased predation associated
with the change to bottom habitat (Johannes,
1978; Bakun, 1986). Heath (1992) suggested that
mortality from predation is particularly high dur-
ing migration to nursery areas. Given the time
to metamorphosis (14 days) and the age of white
mullet when they enter the lagoon (50 to 70 days
for the first cohort), metamorphosis to the demer-
sal stage most probably occurs at least one month
before entry into the lagoon (Anderson, 1957;
Caldwell and Anderson, 1959; Yanez-Arancibia,
1976;Vieira, 1991).
During the demersal period at sea, white mullet
may be exposed to considerable mortality due to
benthic predators. Variation in the abundance of
recruitment pulses into La Restinga Lagoon may
reflect the interplay between spawning time and
the mortality during transport to the coastal area.
At some point between metamorphosis and lagoon
entry, juvenile mullet also develop active swim-
ming behavior to facilitate passive transport. We
observed intensive recruitment of small mullet
into the lagoon between March and June by indi-
viduals that had hatched the previous December
to February. The timing of their hatching means
that their return to the lagoon was likely facilitat-
ed by prevailing currents. In contrast, recruitment
of mullet to the lagoon over the remainder of the
year was weak and sporadic, and fish were much
larger and older At its first appearance in the
lagoon, the third cohort was twice the age of the
first cohort. These fish were not produced during a
period when currents would likely have facilitated
lar\'al transport to the lagoon (little upwelling)
and their lower densities may partially reflect
increased mortality during the more prolonged
return to the lagoon. We propose that spawning
during periods of weak upwelling causes a delay
in transport to coastal nursery areas and conse-
quently decreased survival.
Periods of hatching leading to successful recruitment,
from late December to March, coincided with moderate
peaks in the upwelling index. This successful recruitment
JL
I ■ ii I
n nnlllLnn
. I. .
_oDoolUnOlk_
Mn
.ndnnn Jl I
—O a II nn n nHI n
I ru-n .,~,n^
ntlnnnn UnO n W
nnrlln,
jJlu
. on n rinnii n„
I .1.1
Jllll_
_llllll„
juL
n = 32 Mar-05
n = 27 Apr-02
n = 22 Apr-30
n = 20 May-28
n = 23 Jun-25
n = 22 Jul-23
n= 16 Aug-20
n = 20 Sep-17
n = 20 Oct-15
n=19 Nov-10
n = 25 Dec-10
n = 27 Jan-08
n=14 Feb-04
n = 6 Mar-18
n= 19 Apr-15
n = 27 May-13
n= 17 Jun-10
n= 16 Jul-08
40 80 120 160 200 240 280 320 360 400
Age (days)
Figure 9
Monthly age distributions of juvenik' white mullet from March 1992
to July 199.3. Age wa.s calculated by counting daily growth ring.s start-
ing with the hatch mark. From top to bottom, open bars represent
cohort 1, black bars represent cohort 2, gray bars represent cohort 3
and black bars represent cohort 4. n = number of otoliths examined.
may be the result of moderate levels of wind speed (<6 m/s-),
that promote moderate upwelling and yield optimal trophic
conditions for fish larvae (Cury and Roy, 1989). Coastal
upwellings in northeast Venezuelan waters are caused by
Mann et a\ ; Reproduction and recruitment oi Mugil curema
819
Hatching date
ooooooooooooooooooo
Date
Figure 10
(Upper panel) Frequency distribution of hatching dates for juvenile white mullet recruits caught in 1992
and 1993 in La Restinga lagoon. Points represent raw data. Continuous black line represents the trend
line generated after application of a smoothing spline (lambda=10). (Lower panel) Upwelling index (UI)
for northeastern Venezuela calculated from August 1991 to March 1993. See text for description of cal-
culation. Points represent raw data. Continuous black line represents the trend line generated following
application of a smoothing spline (lambda=10).
moderate levels of wind stress and this differs from the
strong upwellings observed in such places as Peru and Sen-
egal. The relation between the timing of successful spawn-
ing and the intensity of coastal recruitment in white mullet
is likely due to variations in the duration of the transport
of larvae and juveniles to the shore as a result of varying
current conditions as well as variations in food availability
for first-feeding larvae.
Acknowledgments
This work is part of a Ph.D. thesis submitted to Laval Uni-
versity by the senior author who was financially supported
by the Fundayacucho Program of Scholarships. We thank
the technical and analytical assistance of the zooplankton
staff of lOV-UDO (Institute Oceanografico de Venezuela,
Universidad de Oriente), Domingo Figueroa and Rafael
Briceno, and Caroline Berger for her work in the field
survey, figure preparation and the processing of otoliths
with Julie Paquet. Martin LLano, of the Metereological
Station of Fundacion La Salle, kindly supplied the enviro-
mental data. We thank Luis Trocolis, Jose Luis Fuentes, and
Alfredo Gomez from de Edimar-UDO, Nueva Esparta, for
laboratory support, Jose Bechara, Casimiro Quinones, and
Jean Paul Boulianne for advice and statistical support and
Idelfonso Liheros and Jesus Marcano for support and com-
plementary information. This work was funded by grants
from NSERC (Natural Sciences and Engineering Research
Council of Canada) and FCAR (Fonds pour la Formation
des Chercheurs et L'Aide a la Recherche, Quebec, Canada)
to Julian J. Dodson and GIROQ (Groupe Interuniversitaire
de Recherches en Oceanographie du Quebec) and by the
Consejo de Investigacion-Universidad de Oriente, proyecto
LISA-92 (CI-5-019-00554/92).
Literature cited
Alvarez-Lajonchere, L.
1976. Contribucion al estudio del ciclo de vida de Mugil
curema Valenciennes (Pisces: Mugilidae). Cien. Ser. 8
Investig. Mar (Havana) 28:3-130.
1980. Composicidn per especies y distribucion de las post-
larvas y juveniles de Lisas (Pisces, Mugilidae) en tunas de
Zaza, Cuba. Rev. Investig. Mar. Cuba l(2-3):28-60.
1982. The fecundity of mullet (Pisces, Mugilidae) from
Cuban waters. J. Fish. Biol. 21:607-613.
Anderson, W. W.
1957. Early development, spawning, growth and occurrence
of the white mullet (Mugil curema) along the south Atlantic
820
Fishery Bulletin 101(4)
Coast of the United States. Fish. Bull.Wildl. Serv. 57(120):
415-425.
Angell, C.
1973. Algunos aspectos de la biologia de la lisa, Mugil
curema Valenciennes, en aguas hipersalinas del nororiente
de Venezuela. Contrib. Fund. La Salle. Cienc. Nat. 51:
223-238.
Bakun, A.
1986. Local retention of planktonic early life stages in tropi-
cal reef bank demersal systems: the role of vertically-struc-
tured hydrodynamic processes. In lOC/FAO workshop
on recruitment in tropical coastal demersal communities,
Ciudad del Carmen, Campeche, Mexico. Workshop Rep.
44(suppl), p. 16-32.
Bakun, A., D. R. McLain, and F. V. Mayo.
1974. The mean annual cycle of coastal upwelling off west-
ern North America as observed from surface measurements.
Fish. Bull. 72:843-844.
Bartsch, J., and R. Knust.
1994. Simulating the dispersion of vertically migrating
sprat larvae [Spratlus sprattus) in the German Basin with
a circulation and transport model system. Fish. Oceanogr.
3:92-105.
Blaber, S. J. M.
1987. Factor influencing recruitment and survival of mugilids
in estuaries and coastal waters of Southeastern Africa. In
Common strategies of anadromous and catadromous fishes
(M. Dadswell, R. Klauda, C. Saunders. R. Rulifson, and J.
Cooper, eds.), p. 507-518. Am. Fish. Soc.Symp. 1?
Blaber, S. J. M., and T. G. Blaber
1980. Factors affecting the distribution of juvenile estua-
rine and inshore fish. J. Fish. Biol. 17:143-162.
Bowden, K. F.
1983. Physicaloceanography of coastal waters, 302 p. Ellis
Horwood Ltd., Chichester, UK.
Caldwell, D. K., and W. W. Anderson.
1959. Offshore occurrence of larval white mullet, Mugil
curema, in the western Gulf of Mexico. Copeia 1959:252.
Campana, S. E., and J. D. Nielsen.
1985. Microstructure of fish otoliths. Can. J. Fish. Aquat.
Sci. 42:1014-1032.
Creutzberg, F, A. T. G. W. Eltink, and G. J. van Noort.
1978. The migration of plaice larvae Pleuronectes platessa
in the western Wadden sea. In Physiology and behaviour
of marine organisms (D. S. Lunsky and A. J. Berry, eds.),
p. 243-251. Pergamon Press, New York, NY.
Cury, P, and C. Roy.
1989. Optimal environmental window and pelagic fish
recruitment success in upwelling areas. Can. J. Fish.
Aquat. Sci. 46:670-680,
Ditty, J. G., and R. F Shaw.
1996. Spatial and temporal distribution of larval striped
mullet {Mugil cephalus) and white mullet (A/, curema.
family: Mugilidae) in the northern Gulf of Mexico, with
notes on mountain muWet, Agonostomus monticola. Bull.
Mar. Sci. 59:271-288.
Etchevers, S. L.
1974. Fecundidad de la lisa {Mugil curema Valenciennes) en
el Oriente de Venezuela. Bol. CientiT Tdc. Ser. Recur. Mar.
CIC UDO (Centre de Ingenieria y Computacidn, Universi-
dadde Oriente) 1(1), 19 p.
Ferraz-Reyes, E.
1983. Estudio del fitoplancton de la Cuenca Tuy-Cariaco,
Venezuela. Bol. Inst. Oceanogr. Venez.Univ.Oriente. 22(1/2):
111-124.
1989. Influencia de los factores fisicos en la distribucion ver-
tical de la biomasa fitoplanctonica en el Golfo de Cariaco
(Venezuela). Bol. Inst. Oceanogr. Venez. Univ. Oriente.
28(l/2):47-56.
Ferraz-Reyes, E., E. Mandelli, and G. Reyes-Vasquez.
1987. Fitoplancton de la Laguna Grande del Obispo, Ven-
ezuela. Bol. Inst. Oceanogr. Venez. Univ. Oriente. 26(1/2):
111-124.
Franco, L.
1986. Biologia y reproduccion de la lisa, Mugil curema
Valenciennes (Pisces: Mugilidae) en el Golfo de Cariaco,
Venezuela. M.S. thesis, 103 p. Instituto Oceanografico
de Venezuela, Univ. Oriente, Cumana, Venezuela.
Garcia, A., and G. Bustamante.
1981. Resultados preliminares del desove inducido de lisa
{Mugil curema Valenciennes) en Cuba. Acad. Cienc. Cuba
Inf Cient.-Tec. 158:7-26.
Gines, H.
1972. Cartas pesqueras de Venezuela. Caracas, 328 p.
Fundacion La Salle de Ciencias Naturales. M.E.L.S.A.,
Madrid, Venezuela.
Gomez, A.
1983. Pigmentos clorofilicos, produccion primaria y abun-
dancia plantonica en el canal de entrada de la Laguna de
la Restinga, Venezuela. Bol. Inst. Oceanogr. Univ. Oriente
22:43-64.
Gomez, A., and F Cervigon.
1987. Perspectivas del cultivo de peces marines en el Caribe
Sur y noreste de Suramerica. Rev. Latinoam. Acuicult. 34:
40-50.
Heath, M. R.
1992. Field investigations of the early life stages of marine
fish. Adv. Mar. Biol. 28:1-173.
Houde, E. D., S. A. Berkley, J. J. Klinovsky, and R. C. Schekter.
1976. Culture of larvae of the white mullet, Mugil curema
Valenciennes. Aquaculture 8:36.5-370.
Ibaiiez-Aguirre, A. L.
1993. Coexistence de Mugil cephalus and M. curema in a
coastal lagoon in the Gulf of Mexico. Mar. Biol. 42:959-961.
Jacot, A. R
1920. Age, growth and scale characters of the mullets, Mugil
cephalus and Mugil curema. Trans. Am. Microsc. Soc. 39:
199-229.
Jenkins, G. P., and H. M. A. May.
1994. Variation in settlement and lai-val duration of King
George whiting, Silloginodes punctata (Sillaginidae),
in Swan Bay, Victoria, Australia. Bull. Mar. Sci. 54(1):
281-296.
Johannes, R. E.
1978. Reproductive strategies of coastal marine fishes in the
tropics. Environ. Biol. Fishes 3:65-84.
Jordan, A. R.
1993. Age, growth and back-calculated birthdate distribution
of lan'al jack mackerel, TYachurus declives (Pisces: Caran-
gidae) from eastern Tasmanian coastal waters. Aust. J.
Mar Freshw. Res. 45:19-33.
Li, C, S. Xueshen, Y Feng, Y Chunwu, and H. Ruidong.
1993. Daily growth increments in otoliths of mullet larva,
Mugil so-iuy Basilewsky, and determination from field-col-
lected ones. Oceanol. Limnol. Sin. 24(4):345-349.
McBride, R. S., and D. O. Conover
1991. Recruitment of young-of-the-year blucfish Pomatus
saltatrix to the New York Bight: variation in abundance
and growth of spring- and summer-spawned cohorts. Mar
Ecol. Prog. Ser. 78:205-216.
Moore, R. H.
1974. General ecology, distribution and relative abundance
Mann et a\ : Reproduction and recruitment of Mugil curema
821
of Mugil cephalus and Mugil curema on the south Texas
coast. Contr. Mar. Sci. 18:241-255.
Miiller-Karger, F. E., C. R. McClain, T. R. Fisher, W. E. Esaias,
and R. Varela.
1989. Pigment distribution in the Caribbean Sea: observa-
tions from space. Prog. Oceanogr. 23:23-64.
Neilson, J. D., and G. H. Geen.
1986. First-year growth rate of Sixes R. chinook salmon
as infered from otohths: effects on mortality and age at
maturity. Trans. Am. Fish. Soc. 115:28-33.
Okuda, T., J. Benitez, and G. Cedeno.
1978. Caracteristicas hidrograficas del Golfo de Cariaco,
Venezuela. Bol. Inst. Oceanogr. Venez. Univ. Oriente 17:
69-87.
Panella, G.
1971. Fish otoliths: daily growths layers and periodic
patterns. Science. 173:1124-1127.
Radtke, R. L.
1984. Formation and structural composition of larval striped
mullet otoliths. Trans. Am. Fish. Soc. 113:186-191.
Redding, M., and R, Patifio.
1993. Reproductive physiology. In Fish physiology (E.
Evans, ed. ), p. 503-534. ORG Press, Boca Raton, FL.
Rodriguez, C. L., and I. V. do Nacimiento.
1980. Estudio microscbpico dos ovarios de Mugil curema
Valenciennes, Brasil. In I simposio Brasileiro de aqua-
cultura, Recife, 1978, p. 213-219. Academia Brasileira de
Ciencias, Rio de Janeiro, Brazil.
Secor, D. H., J. M. Dean, and E. H. Laban.
1992. Otolith removal and preparation for microstructural
examination. In Otolith microstructure examination and
analysis (D. K. Stevenson and S. E. Campana, eds. ), p. 9-57.
Can. Spec. Publ. Fish. Aquat. Sci. 117.
Simonoff J. S.
1996. Smoothing methods in statistics. Springer-Verlag,
New York, NY.
Sinclair, M.
1988. Marine populations: an essay on population regulation
and speciation, p. 252. Univ. Washington Press, Seattle,
WA.
Sirois, P., and J. J. Dodson.
2000. Critical periods and growth-dependant survival of
larvae of an estuarine fish, the rainbow smelt Osnierus
mordax. Mar Ecol. Prog. Ser 203:233-245.
Struhsaker, P., and J. H. Uchiyama.
1976. Age and growth of the nehu, Stolephorus purpureus
(Engraulidae), from the Hawaiian islands as indicated by
daily growth increments in sagittae. Fish. Bull. 74:9-17.
Vieira, J. P.
1991. Juvenile mullets (Pisces: Mugilidae) in the estuary
Lagoa dos Patos, RS, Brazil. Copeia 1991:409-418.
Wilson, K. H., and R A. Larkin.
1980. Daily growth rings in the otoliths of juvenile sockeye
salmon Oncorhynchus nerka. Can. J. Fish. Aquat. Sci. 37:
1495-1498.
Yanez-Arancibia, L. A.
1976. Observaciones sobre Mugil curema Valenciennes, en
areas naturales de crianza, crianza, maduracibn, creci-
miento, madurez y relaciones ecologicas. Ann. Cent. Cienc.
Mar Limnol Univ Nac. Auton. Mexico 2:211-243.
822
Abstract— Fecundity in striped mullet
(Mugil cephalus) from South Carolina
correlated highly with length and
weight, but not with age. Oocyte counts
ranged from 4.47 x 10'^ to 2.52 x 10^
in 1998 for fish ranging in size from
331 mm to 600 mm total length, 2.13 x
105 to 3.89 X 10« in 1999 for fish ranging
in size from 332 mm to 588 mm total
length, and 3.89 x 10= to 3.01 x lO'' in
2000 for fish ranging in size from 325 mm
to 592 mm total length. The striped
mullet in this study had a high degree
of variability in the size-at-age relation-
ship; this variability was indicative of
varied growth rates and compounded
the errors in estimating fecundity at
age. The stronger relationship of fecun-
dity to fish size allowed a much better
predictive model for potential fecundity
in striped mullet. By comparing fecun-
dity with other measures of reproduc-
tive activity, such as the gonadosomatic
index, histological examination, and
the measurement of mean oocyte diam-
eters, we determined that none of these
methods by themselves were adequate
to determine the extent of reproductive
development. Histological examinations
and oocyte diameter measurements
revealed that fecundity counts could be
made once developing oocytes reached
0.400 nm or larger. Striped mullet
are isochronal spawners; therefore
fecundity estimates for this species are
easier to determine because oocytes
develop at approximately the same rate
upon reaching 400 idm. This uniform
development made oocytes that were
to be spawned easier to count. When
fecundity counts were used in conjunc-
tion with histological examination,
oocyte diameter measurements, and
gonadosomatic index, a more complete
measure of reproductive potential and
the timing of the spawning season was
possible. In addition, it was determined
that striped mullet that recruit into
South Carolina estuaries spawn from
October through April.
Fecundity and spawning season of striped mullet
{Mugil cephalus L.) in South Carolina estuaries*
Christopher J. McDonough
William A. Roumillat
Charles A. Wenner
Marine Resources Research Institute
South Carolina Department of Natural Resources
217 Fort Johnson Road
Charleston, South Carolina 29422-2559
E-mail address (for C J McDonougfi): mcdonoughc@mrd.dnrstate.sc,us
Manuscript approved for publication
19 June 2003 by Scientific Editor
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:822-834
The striped mullet iMugil cephalus L.)
is distributed circumglobally in tropi-
cal and semitropical waters between
latitudes 42°N and 42°S (Thompson,
1963; Rossi et al., 1998). Even though
considered a marine species, striped
mullet are euryhaline and can be found
year round throughout the full range of
estuarine salinities in the southeastern
United States (Jacot, 1920; Anderson,
1958). Striped mullet are a commer-
cially important fish throughout the
world sustaining both fisheries and
aquaculture industries. In the south-
eastern United States (North Carolina
and Florida) there are significant com-
mercial fisheries for striped mullet,
whereas in South Carolina and Georgia
landings are more limited (NMFS').
The primary fishery in most of these
states is for "roe" mullet during the
fall spawning migration. Throughout
the rest of the year mullet are fished
commercially for bait, if they are fished
at all (Anderson, 1958). Striped mullet
have a significant economic impact in
the areas where they are more heavily
fished by the commercial fisheries and
the landings of this species from 1994
to 1998 yielded a landings (wholesale)
value of 38.2 million dollars. Striped
mullet are also one of the most impor-
tant forage fishes that are found in the
estuaries of the southeastern United
States and represent a significant
food source for upper-level piscivores
(Wenner et al.^).
The biological features of striped
mullet has been well documented
(Jacot, 1920; Anderson, 1958; Thomson,
1963, 1966; Chubb et al., 1981), but
much less information is available on
the biological aspects of reproduction
in the wild (Anderson, 1958; Stenger,
1959; Greeley et al., 1987; Render et al,
1995). There is a large body of work con-
cerning striped mullet reproduction in
aquaculture but many of these studies
have used artificial manipulation of the
reproductive cycle. Although the matu-
ration process of oocytes may have been
the same as that in wild striped mullet,
the environment and conditions under
which maturation occurs were artifi-
cial (Shehadeh et al., 1973a; Kuo et al.,
1974; Pien and Liao, 1975, Kelly, 1990;
Tamaru et al., 1994; Kuo, 1995). In ad-
dtion, despite the demonstrated ability
to initiate reproduction (both in and
out of season) for striped mullet, the
majority of aquaculture studies have
had to rely on wild fish for broodstock
(Kuo et al., 1974; Pien and Liao, 1975;
Kuo, 1995).
* Contribution 522 of the Marine Resources
Institute, South Carolina Dept. of Natural
Resources, Charleston, SC 29422-2559.
' National Marine Fisheries Service. 2001.
Personal commun. Statistics and Eco-
nomic Division, 1315 East-West Highway,
Silver Spring, Md. 20910. http://www.
St. nmfs.gov/stl/index. html.
2 Wenner, C. A., W. A. Roumillat, J. E. Moran,
M. B. Maddox, L. B. Daniel, and J. W.
Smith. 1990. Investigations on the life
history and population dynamics of ma-
rine recreational fishes in South Carolina,
part 1. Completion report, project F-37,
p. 3-13 and project F-31, p. 6-35. South
Carolina Marine Resources Research
Institute, P.O. Box 12559, Charleston, SC
29422
McDonough et al : Fecundity and spawning season of Mugil cepha/us
823
In the southeastern United States the spawning season
lasts from two to five months depending on the coastal
area involved (Jacot, 1920; Broadhead, 1956; Anderson,
1958; Arnold and Thompson, 1958; Stenger, 1959; Dindo
and MacGregor, 1981; Greeley et al., 1987; Render et al.,
1995; Hettler et al., 1997). Striped mullet are considered
isochronal spawning fishes (Greeley et al., 1987; Render et
al., 1995), i.e. they have synchronous gamete development
and individuals spawn all their reproductive material at
once or in batches over a very short time period (days, as
opposed to weeks). There have been limited observations of
offshore spawning activity (Arnold and Thompson, 1958),
and few examples of eggs and larvae collected offshore (An-
derson, 1958; Finucane et al., 1978; Collins and Stender,
1989). Collins and Stender (1989) concluded that striped
mullet spawn in and around the edge of the continental
shelf off the coasts of North Carolina, South Carolina,
Georgia, and the east coast of Florida (an area often re-
ferred to as the South Atlantic Bight) and have a protracted
spawning season from October to April. This spawning
season contrasts with that estimated by most other stud-
ies (Jacot, 1920; Broadhead, 1956; Anderson, 1958; Arnold
and Thompson, 1958; Stenger, 1959; Dindo and MacGregor,
1981; Greeley et al., 1987; Render et al., 1995; Hettler et al,
1997 ). These studies based their estimates on the reproduc-
tive condition of migrating adults and the subsequent re-
cruitment of juvenile fish back into coastal estuaries — not
on actual data on the offshore presence of striped mullet
larvae. Female mullet have been shown to mature at two
to three years of age at a size range from 230 mm to 350
mm standard length (Thomson, 1951, 1963; Greeley et al.,
1987). Determination of spawning activity in mullet has
been estimated by using gonadosomatic indices (Dindo and
MacGregor, 1981; Render et al., 1995), examination of oo-
cyte size and maturity stage (Stenger, 1959; Greeley et al.,
1987 ), and by the presence of enlarged, developing ovaries
in migrating fish (Jacot, 1920; Anderson, 1958).
The purpose of the present study was to develop size- or
age-related estimates of fecundity in striped mullet from
South Carolina estuaries. These fecundity estimates were
determined in order to develop models for estimating
potential fecundity from catch data, such as length and
weight data. In addition, other indicators of reproductive
activity such as gonadosomatic indices and oocyte size were
examined to provide information on the duration of the
spawning season. Potential fecundity estimates can give a
barometer of reproductive potential based on morphologi-
cal information from catch curve data and size-frequency
distributions.
Materials and methods
Striped mullet were collected monthly from January 1998
through December 2000 by using a randomly stratified
sampling regime within three different estuarine systems
along the South Carolina coast: (Ashepoo-Combahee-
Edisto (ACE) Basin, Charleston Harbor, and the Cape
Remain estuary, Fig. D.The female striped mullet used for
fecundity estimates were collected from October through
February each year because these were the only months
when fecund females were present. Fecund female striped
mullet were defined as specimens in the tertiary stage of
vitellogenesis with oocyte diameters greater than 400 ^m
and that had gonadosomatic indices greater than 5.0. The
vitellogenic stage for each specimen used in the fecundity
estimates was confirmed histologically. Fish were captured
during daylight ebbing tides with water levels ranging
from 0.3 to 2.0 meters, and the majority (67.8%) were
caught during late ebb. The fish were primarily caught
with a 184-m trammel net with an outside stretch mesh of
350 mm and an inside stretch mesh of 63.5 mm, although
a few of the 1999 samples (5) and the 2000 samples (3)
were obtained by using a cast net 1.8 m in diameter and
equipped with 10-mm mesh. Eight fecund females were
also captured by using an electroshock boat in the fresh-
water and low-salinity areas of the Cooper River (one of the
three rivers that make up the Charleston Harbor Estuary)
in October 2000. The areas included in the present study
have been sampled on a monthly basis since 1991 with
trammel nets by the Inshore Fisheries Group of the South
Carolina Department of Natural Resources as part of a
gamefish monitoring program. During this period repro-
ductively developing striped mullet of both sexes were
generally observed from October through February and
were presumably heading offshore to spawn (Jacot, 1920;
Anderson, 1958; Arnold and Thompson, 1958). Male striped
mullet that were reproductively developed were easy to
discern because they were usually leaking milt and were
not analyzed further. All other fish were brought back to
the laboratory and eviscerated in order to determine sex
and to collect ovaries for reproductive analysis. All of the
samples were kept on ice and were generally examined
within twenty-four hours of capture.
Standard morphometric measurements included total
length (TL) in mm, fork length in mm (FL), and standard
length in mm (SL), total weight (TW) in grams, ovary
weight (OW) in grams, sex, and maturity. Body weight
(BW) was calculated as total weight minus ovary weight
{BW=TW-OW). Saggital otoliths were removed for aging.
A small section of the posterior end of the ovaries, at the
junction of the two lobes, was also removed for histo-
logical examination. The whole ovaries were fixed in 10%
seawater-buffered formalin and the histological sample
was fixed in 10% neutral-buffered formalin. Histological
samples were processed by using standard procedures for
paraffin embedding and sectioning (Humason, 1967). The
sections were dried on slides and stained by using standard
haematoxylin and eosin-Y staining techniques (Humason,
1967). Examination of the histological sections for maturity
stage was done by using a compound microscope at lOOx
magnification. Each histological section was evaluated by
two separate readers to determine agreement on maturity
stage. If there was a discrepancy in maturity staging for
any specimen, the discrepancy was either resolved by the
two readers or that specimen was not used in the analysis.
There were no discrepancies between readers on any of
the reproductively developing females. Maturity was as-
sessed according to a modified version of the schedule used
by Wenner et al. (1986) adapted to work with isochronal
824
Fishery Bulletin 101(4)
80° 30'
80° 00
80° 30'
80° 00'
79° 30'
Figure 1
The South Carolina coast with the locations of estuaries and major river basins where striped mullet were collected
from 1998 to 2000.
spawning fish, as well as previous models of reproductive
development (Stenger, 1959; Wallace and Selman, 1981)
(Table 1). This evaluation method was based on identifica-
tion of morphological characteristics evident in histological
sections. Each specimen was evaluated by two (in some
cases three) readers and discrepancies between readers
were either resolved or the specimen was excluded from
the analysis.
A gonadosomatic index (GSI) was calculated for each
specimen following the method of Render et al. (1995)
where GSI was expressed as gonad weight (GW) divided
by body weight (BW) such that
GSI = (GWIBW) X 100.
The GSI values of the fecundity specimens were compared
among the three sampling years, as well as with GSI values
for female striped mullet collected during the rest of the
year that were mature but not reproductively active.
Fecundity determinations were made from a total of 129
advanced-stage developing ovaries; 50 from 1998, 37 from
1999, and 42 from 2000. All of the ovaries were determined
by histological examinations and criteria outlined in Table
1 to be actively vitellogenic, having tertiary yolk-stage
oocytes. All oocytes counted for fecundity were 400 /jm
or larger and in the tertiary yolk stage (Pien and Liao,
1975) (Fig. 2). This 400-/jm threshold has also been used
in other studies of oocyte development in striped mullet for
determining the point at which the oocytes that would be
spawned during that year were identifiable (Shehadeh et
al., 1973a). Unlike other species (particularly batch spawn-
ers), where fecundity counts should ideally be conducted
by using hydrated oocytes only, the striped mullet oocytes
used for fecundity counts were still presumably several
weeks from hydration and spawning. However, because
mullet are synchronous spawners, it is relatively easy to
distinguish the developing oocytes from the undeveloped
ones because of the drastic difference in size between the
two, as well as by the uniformity in size of the developing
oocytes once they reach 400 pm.
Fecundity was estimated by using a modified gravimet-
ric method. The fixed whole ovary was patted dry and re-
weighed. The ovarian lobes were divided into four discrete
regions along each lobe's longitudinal axis and three sub-
samples (chosen at random) were taken between the two
lobes and prescn'ed in 50% isopropyl until oocyte counts
could be conducted. The subsamples ranged in weight from
0.025 g to 0.033 g. The subsamples from each specimen
McDonough et al.: Fecundity and spawning season of Mugil cephalus
825
Table 1
Histological criteria used to determine reproductive stage in female striped mullet {Mugil cephalus) once sexual differentiation has
occurred (Wenner et al.: see footnote 2 in the general text).
Reproductive stage
Description
1. Immature
2. Developing
3. Running ripe
4. Atretic or Spent
5. Inactive or Resting
Inactive ovary with previtellogenic oocytes and no evidence of atresia. Oocytes are < 80 ^m, lamellae
still contain somatic and connective tissue bundles. Ovary wall is very thin (one or two cell layers).
Developing ovary have enlarged oocytes generally greater than 120 /jm in size. Cortical alveoli are
present and actual vitellogenesis occurs after oocytes reach 180 /jm in size and continue to increase
in size. Abundant yolk globules with oocytes reaching a size of >600 pm.
Completion of yolk coalescence and hydration in most oocytes.
More than 30% of developed oocytes undergoing the atretic process.
Previtellogenic oocytes only but traces of atresia possible. In comparison to immature females, most
oocytes are >80 ^m, lamellae have some muscle and connective tissue bundles. Lamellae are larger,
have moore oocytes, and are elongated. A thicker ovarian wall with blood vessels, muscle, and nerve
tissue.
were then teased apart. After separation, the oocytes were
spread out on a Bogorov tray and counts of ooc5ftes, greater
than 400 |jm, were made by using a dissecting microscope at
12x magnification. Each subsample was counted twice and
counts were averaged. A third count was performed if the
first two counts differed by more than 10%. Oocyte density
was calculated by dividing the mean number of oocytes by
the mean weight of all three subsamples for each specimen.
The oocyte density was then used to calculate the total
oocyte number for each ovary, or individual fecundity, by
multiplying mean oocyte density by whole ovary weight.
In order to determine mean oocyte diameter for each
specimen, 20-30 oocytes were removed from each counted
subsample and grouped together in a petri dish. Each oo-
cyte was then measured along the longest axis by using
Optimas''''^ Image Analysis software (version 6, Media
Cybernetics, Bothell,WA). Mean oocyte diameter was
calculated as the average of all measurements for each
826
Fishery Bulletin 101(4)
ou -
^m 1998
1 1 1999
25 -
^m 2000
20 -
15 -
10 -
1
5 -
1
IJ
•1
I
1 I
300-350 351-400 401-450 451-500 501-550 551-600
Size class (mm)
Figure 3
Size-frequency distribution of female striped mullet used in fecundity determinations from
South Carolina estuaries from 1998 to 2000. n (number offish in sample) = 50 fish for 1998,
ri = 37 for 1999. and n = 42 for 2000.
subsample. The overall mean oocyte diameter for each
specimen resulted from the calculated average of the
means of the three subsamples. Measurements were not
made on fresh oocytes but shrinkage was estimated from
the amount of whole ovary shrinkage because fresh ovary
weight and preserved (in 10% formalin) ovary weight were
known. The estimated unpreserved oocyte diameter was
determined by multiplying the preserved oocyte diameter
by 1 and adding the percentage of ovary shrinkage. The dif-
ference in the preserved oocyte diameter and the estimated
fresh oocyte diameter was then compared by using a paired
t-test to determine if there was a significant difference be-
tween preserved and unpreserved oocyte diameters.
Age was determined by using the left sagittal otolith,
which was embedded in epoxy resin. A 0.5-mm transverse
section encompassing the otolith core was cut with an
Isomet low-speed saw with diamond wafering blades. The
thin section of the otolith was embedded in epoxy and ob-
served with a dissection microscope at the magnification
appropriate for the otolith's size. Age was recorded as the
number of rings (annular bands) present. The otoliths were
initially aged by one reader. A second reader then evalu-
ated a subsample of specimens from 1998 and 2000 and all
the otoliths from 1999. Ages were validated by the percent-
age of agreement between the two age determinations, an
analysis of variance (ANOVA) between the two groups of
ages, and a paired Mest comparing the means and vari-
ances of the two groups (Campana et al., 1995).
Results
Fecund female striped mullet (again, defined as those
females with ovaries containing oocytes >4G0 ^m in the
tertiary stage of vitellogenesis) were collected from late
October through February; most of the specimens were
caught in November and December for all three years.
Size-frequency distributions did not vary over the three
years of the study (Fig. 3). Fewer fish (n=37) were taken in
1999 versus 1998 (n=50) and 2000 (/i=42). The most abun-
dant size class for each year was that from 401 to 450 mm.
In the overall size-frequency distributions, fecund females
made up greater than 44% of those fish larger than 400 mm
in 1998 and comprised all of the specimens over 500 mm.
In 1999 the fecund females made up 12.5"% offish in the
401-450 mm size range, 39% of the fish in the 451-500 mm
size range, and 100% of the fish in the size classes over 500
mm. In 2000, fecund females made up 17.7% of the 401^50
mm size class, 35.7% in the 451-500 mm size range and,
like 1998 and 1999, all of the specimens over 500 mm.
The majority of females used in our fecundity study
were 3 or 4 years old (Fig. 4), accounting for 80.0% of the
specimens in 1998 and 73.3% in 1999. However, the age
distribution in 2000 showed that the frequency of 2-, 3-, and
4-year-olds was the same and that these three ages made
up 82.0% of the fecund fish sampled that year. Three-year-
old fish made up the largest single group in 1998 and 1999.
The age determined from the otoliths was validated as part
McDonough et aL: Fecundity and spawning season of Mugil cephalus
827
25 -
■■1 1998
1 1 1999
^H 2000
20 -
15 -
!
10 -
i
5 -
n
r
1
0 -
1
1 1 1 1 1 .n. 1
— g — ,
1 2
5 6
Age (yr)
10
Figure 4
Age-frequency distribution of fecund striped mullet in South Carolina estuaries from 1998 to
2000. n (number offish in sample) = 50 fish for 1998, n = 37 for 1999, and /; = 42 for 2000.
of another study where size and age structure of striped
mullet in South Carohna was examined (Wenner and Mc-
Donough, unpubl. data^). A comparison of multiple read-
ings of the same group of otoliths assessed aging precision.
One year (1999) was chosen at random and all of the oto-
liths («=1234) from that year were aged by a second reader.
The ages of the two independent determinations were then
compared by using a one-way ANOVA and a ^test. The
variance statistic was 2.78 for the original ages and 2.81
for the second age reads, which were not significantly dif-
ferent (P=0.001) and both had almost identical normalized
residuals. Overall, there was an 83.4% agreement on ages
between readers. The results from the ANOVA (F= 1555.0,
df=10, P=0.000) and the t-tesi «=2.898, df=1233, signifi-
cance |2-tailed] =0.004) both confirmed that there was no
significant difference between the separate age determina-
tions. Therefore, the age recorded by the first reader for all
specimens was used in the analysis.
The length-weight relationship for fecund female striped
mullet was compared by using a linear regression of (natu-
ral) log-transformed body weight against total length to
see if there were any differences between years. The
regression coefficients from each year were compared by
Wenner, C. A., and C. J. McDonough. 2001. Cooperative
research on the biology and assessment of nearshore and estua-
rine fishes along the southeast coast of the U.S. Part IV: Striped
mullet, Mugil cephalus. Final rep, Grant no. NA77FF0550,
82 p. Marine Resources Research Institute, South Carolina
Dept. of Natural Resources, P.O. Box 12559 Charieston, S.C.
29422-2559.
using a test of significance between more than two slopes
(Zar, 1984). The weight measurement used was total body
weight minus gonad weight (TW-OW=BW) because ovary
weight had a considerable influence on total body weight
in the fecund specimens (GSI values of 7.7 to 27.7). There
was no significant difference in the total length to Ln
body weight regressions between different years (F=9.22,
P=0.001, df=129). Because there was little difference in the
regressions between years and in order to increase sample
size, data from all three years were combined to obtain the
overall total-length to Ln-body-weight relationship of Ln
BW = -11.1 + 2.92 (TL) (Fig. 5). In contrast, the length and
body-weight-at-age relationships were highly variable; a
wide range of sizes occurred in the 2-, 3-, and 4-year age
classes (Fig. 6). The high degree of variability was also ex-
acerbated by the smaller number offish age 5 or older.
The gonadosomatic index (GSI) for fecund mullet ranged
from 7.7 to 27.5 in 1998, 9.3 to 27.7 in 1999, and 9.5 to 26.6
in 2000. In contrast, the GSI for mature females that were
not undergoing any reproductive development ranged from
almost zero to 4 for all three years of the study. The rela-
tionship of GSI to size (TL or BW) was not very strong in
any year. However, GSI was positively correlated (P=0.01)
with oocyte diameter and negatively correlated with oo-
cyte density (Table 2) because of the inverse relationship
of oocyte density and oocyte diameter. The correlation coef-
ficient for GSI and age were very close to zero and slightly
negative (Table 2).
Mean GSI by month for males and females (Fig. 7) in-
creased from October through April, peaking in November-
828
Fishery Bulletin 101(4)
December. The duration of the reproductive season, as evi-
denced by advanced reproductive condition determined by
the GSI, was also confirmed from histological assessments
of maturity stages for all gonads collected during this time
period (not just those used for the fecundity study) which
indicated that reproductively developing males were present
5
7 0 -
65
60
5.5 -
5.0
Ln BW=-]] A +29
r'=949. F= 9690.4
/I = 1 29
250
300
350
400
450
500
550
600
Total length (mm)
Figure 5
Regression analysis of log-transformed (Ln) body weight on total length
for fecund striped mullet in South Carolina estuaries from 1998 to 2000.
n ( number of fish in sample) = 129.
August through February, whereas reproductively develop-
ing females were present August through April (Fig. 8).
There was no significant difference in oocyte density
among subsamples taken from different areas of the ovary
lobe. This result was obtained by using an ANOVA of oocyte
densities between the four divided areas of the ovary lobes
where subsamples were taken (F=0.421,
df=3). This analysis allowed us to accept
the assumption that oocytes were equally
distributed throughout the ovary lobes,
which provided validation for the random
sampling of oocytes from different areas of
the lobe in order to determine individual
fecundity.
The regression of individual fecundity
with total length (TL) was not a linear
relationship, whereas the regression of
fecundity on body weight (BW) was linear.
Therefore, the comparisons of individual
fecundity to total length (TL) and body
weight (BW) were made by using both the
raw data and the data with natural log
transformations. The range of specimen
total lengths was 291 to 600 mm in 1998,
332 to 588 mm in 1999, and 325 to 592 mm
in 2000 and for body weight 242 to 2149 g
in 1998, 335 to 2008 g in 1999, and 284 to
2144 g in 2000. Mean fecundity, compared
between years with a two sample Mest,
was significantly different between 1999
and 2000 {t=0.019, df=78, P=0.985) but
was not significantly different between
1998 and 1999 (<=0.974, df=86, P=0.336)
650
Table 2
Pearson correlation coefficients, with significance values, for the morphological
Carolina estuaries from 1998 to 2000. n (number offish in sample) = 129. TL
diameter, ODN = oocyte density, GSI = gonadosomatic index, FEC = fecundity.
variables and fecundity of striped mullet in South
= total length, BW = body weight, ODM = oocyte
Age
TL
BW
ODM
ODN
GSI FEC
Age
Pearson correlation
significance (2-tailed)
1.000
TL
Pearson correlation
significance (2-tailed)
0.117
0.186
1.000
BW
Pearson correlation
significance (2-tailed)
0.164
0.062
0.951**
0.000
1.000
ODM
Pearson correlation
significance (2-tailed)
0.004
0.964
0.101
0.254
0.029
0.741
1.000
ODN
Pearson correlation
significance (2-tailed)
0.575
-0.050
0.927
^0.008
0.282
0.095
0.000
-0.628**
1.000
GSI
Pearson correlation
significance (2-tailed)
-0.029
0.746
0.128
0.147
-0.006
0.947
0.543**
0.000
-0.645**
0.000
1.000
FEC
Pearson correlation
significance (2-tailed)
0.113
0.200
0.892**
0.000
0.888**
0.000
0.059
0.,'504
0.142
0.105
0.260** 1.000
0.003
" Correlation is significant at the 0.01 leve
1 (2-tailed).
McDonough et al : Fecundity and spawning season of Mugil cephalus
829
and between 1998 and 2000 «=1.368, df=92,
P=0.179). However, given that the mean fecun-
dity was 1.18 miUion oocytes in 1998, 1.16 mil-
hon oocytes in 1999, and 1.09 milhon oocytes in
2000, the difference in mean fecundity between
1999 and 2000 was probably not biologically
significant. It was determined that data could
be pooled across years for several reasons. The
coefficients of determination for each year in-
dicated that there was a similarly strong rela-
tionship of fecundity to total length and body
weight in all three years and the coefficients of
variation for each year (0.408 for 1998, 0.594
for 1999, and 0.457 for 2000 at P=0.001) were
not significantly different. By pooling the data
from all three years we were able to determine
two models of potential fecundity based on total
length (TL) and body weight (BW) (Fig. 9):
Ln Fecundity = -6.86 + 3.42(Ln Total Length)
[r2=0.803,F=527.2, df=129]
Ln Fecundity = 6.95 + 1.05(Ln Body Weight)
[r2=0.804, F=530.6, df=129].
The r^ values for untransformed data were
very close to the values obtained with trans-
formed data (r2=0.795, i^=502.9 for fecundity
on total length [TL] and r2=0.787, F=479.4 for
fecundity on body weight [BW]). The high r^
values, as well as the high correlation coeffi-
cients between fecundity and total length and
body weight (Table 2) indicated that potential
fecundity was size dependent.
Unlike fecundity, oocyte density did not
change with size (TL or BW) in 1998 and 1999
and increased with size in 2000. The increase
in density in 2000 was due to a group of fish
captured in freshwater in October having
relatively low GSIs and high densities of
oocytes that also happened to be some of the
largest fish captured that year. Oocyte density
was negatively correlated with GSI (Table 2), and thus
indicated that increasing GSI resulted in lower oocyte
densities.
There was not a high degree of variability in oocyte di-
ameter over the entire size range for the three years of the
study. Oocyte diameter did appear to increase with age in
2000 and remained stable for 1998 and 1999. However, the
increase in oocj^te diameter in 2000 was not statistically
significant.
In a comparison of mean oocyte diameter in each size
class (total length) by month of capture, the data for all
three years were pooled in order to obtain adequate repre-
sentation in each month. Oocyte size ranged from 463 to
682 /jm and the mean size was 596 ^m. The largest mean
oocyte diameters were found in specimens captured in Jan-
uary and February. Specimens were captured during the
months of November and December for all size classes and
there was an increase in oocyte diameter with each pro-
o
1998
700
A
1999
O
2000
o
A
A
o
600 -
k
o
500 -
O
o
i
i
o
e
O
o
400 '
i
o
2
#
o
o
A
300 -
o
700
A
3000
1500 -
600
o
o
o
o
B
5 6
Age (yr)
Figure 6
Relationship of total length (A) and body weight (B) to age for fecund
striped mullet in South Carolina estuaries from 1998 to 2000. n (number
offish in sample) = 129.
gressive month. In particular, females in the 400-500 mm
size range (which represented the largest number of
specimens) were examined and there was a progression of
increasing oocyte diameter with month of capture through
the reproductive season.
The increase in oocyte size, as the reproductive season
progressed, was more apparent when mean oocyte diam-
eter by month for each year separately was examined.
Specimens were collected from October through February
in 1998, November through January in 1999, and October
through December in 2000. Equal effort was made during
all of these months of each year to capture specimens, but
they were not always available for capture. There was an
increase in mean oocyte size per month as the spawning
season progressed in all three years (Fig. 10). Even though
the largest oocyte size measured was 682 /jm, this measure-
ment was that of a preserved oocyte. If we factor in a mean
shrinkage of 4%, maximum oocyte size becomes 709 ^m.
830
Fishery Bulletin 101(4)
The paired i-test showed no significant difference between
the preserved oocytes and the predicted size of fresh oo-
cytes (t=-26.2, df=128, P=0.000).
Discussion
Fecundity in striped mullet from South Carolina correlated
highly with length and weight, but not with age. Oocyte
counts ranged from 4.47 x IQS to 2.52 x 10'' in 1998 for fish
ranging in total length from 331 mm to 600 mm, 2.13 x 10^
to 3.89 X 10® in 1999 for fish ranging in total length from
332 mm to 588 mm, and 3.89 x 10^ to 3.01 x 10® in 2000 for
fish ranging in total length from 325 mm to 592 mm. These
fecundity levels correspond with general fecundity levels
(2.0 x 10'' to 14.0 X 10®) found in striped mullet in northeast
Florida (Greeley et al., 1987), the Gulf Coast (Render et
al., 1995) as well as studies in Europe and Asia (review by
Alvarez-Lajonchere, 1982). One marked difference in fecun-
dity between the present study and some in the literature
was the difference in oocyte density. Render et al. (1995)
found densities ranging from 798 to 2616 oocj^es/g ovary
weight, whereas densities in the present study ranged from
1710 to 14,817 oocytes/g ovary weight. However, although
fecundity increased with both total length and body weight
in 1998 and 1999, densities did not. The lower oocyte densi-
ties in the larger fish were most likely indicative of larger
oocytes. This feature is common in both synchronous and
asynchronous spawning fishes (Greeley et al., 1987; Render
et al., 1995; Fox and Crivelli, 1998; DeMartini and Lau,
1999). Because total length and body weight were more
highly correlated with increased fecundity, the larger speci-
mens would have made a greater individual reproductive
o
O
Males
Females
M.^Y JUN JUL AUG SEP
OCT NOV
Month
DEC JAN FEB MAR APR
Figure 7
Mean gonadosomatic index value by month for male and female striped
mullet from South Carolina estuaries from 1998 to 2000. n (number of fish
in sample) = 455.
contribution during any given spawning season (Korhola
et al., 1996; Kaunda-Arara and Ntiba, 1997; DeMartini
and Lau, 1999). Making estimates of potential fecundity
from age alone was difficult because of the variability in
the size-at-age relationship; however, fecundity estimates
from total length or body weight appeared more reliable
and reflected values closer to the fecundity levels observed
in the present study.
Previous studies have reported that female mullet be-
come reproductively mature at three years of age (Thom-
son, 1951; 1963; Stenger, 1959; Chubb et al., 1981; Render
et al., 1995). Greely et al. (1987) suggested that fecundity
specimens collected in northeastern Florida were as young
as two years at maturity. But, Greely et al. (1987) did not
age the striped mullet used in their study and inferred age
from a size-at-age growth schedule (Thomson, 1966). In
the present study, two-year-olds made up only a small per-
centage of the fecund fish in 1998 and 1999. However, the
results were different for the 2000 specimens in that there
was an almost equal distribution in the age frequency of
two-, three-, and four-year- olds. However, three- and four-
year-olds made up the greatest percentage of females with
advanced ovaries. Maturing fish were those undergoing ac-
tive vitellogenic development and were generally captured
in and around inlets or estuaries. Our study would suggest
that female striped mullet reach 50% maturity at age 2 and
100% maturity at age 3. Three- and four-year-olds made up
the majority of reproductively advanced fish in all years,
whereas less abundant older fish made less of a contribu-
tion towards total reproductive effort.
There are several possible explanations accounting for the
wide age distribution in maturity stages; the most likely
is that fecundity is size related despite the highly variable
growth rates and the widely ranging size at
age in adult striped mullet. Size at maturity
has been found to range widely from 230 mm
standard length (Thomson, 1963; Greeley et
al., 1987; Tamaru et al, 1994) up to 410 mm
standard length (Thomson, 1963; 1966; Chubb
et al., 1981) for two- and three-year-old fish.
The lower end of this size range agrees quite
readily with the lower size range (291 mm
TL=239 mm SL) found in our study. In a con-
current study of maturity schedules related to
size and age in South Carolina striped mullet
(McDonough, unpubl. data), male striped mul-
let were found to mature at two years of age
and as small as 250 mm total length ( 190 mm
standard length). Other species of mullet have
been shown to mature over a wide range of siz-
es. The Pacific mullet (Mugil so-iuy) becomes
mature upon reaching approximately 430 mm
total length (Okumus and Bascinar, 1997).
Monthly GSI levels clearly showed that
the time period of reproductive activity is
from October through April. Female striped
mullet in all reproductive developmental
stages were observed during the course of
our sampling, with the exception of stage-3
(hydrated oocytes) females and females with
McDonough et al.: Fecundity and spawning season o\ Mugil cephalus
831
recently spawned ovaries (characterized by the presence
of postovulatory foHicles). Atretic ovaries were observed
from December through May. There were no postovu-
latory follicles observed, indicating that any atretic
ovaries were not from recently spawned fish. The fish
with atretic ovaries were characteristically emaciated
for their size (TL and BW) and were most common from
January through March. The presence of females with
atretic ovaries starting in December is strong evidence
that spawning occurred in November, if not earlier,
and females with atretic ovaries caught as late as May
demonstrate that spawning may still occur as late as
April. Additional evidence for the October through April
spawning period has also been shown in backcalculated
birth dates for juvenile striped mullet by daily growth
increments (McDonough and Wenner, 2003). This
evidence supports the concept of offshore spawning
in striped mullet and a yet undetermined time period
required for moving from the estuaries to the spawning
areas and for returning again to the estuaries. Other
authors have come to the same conclusion from similar
evidence in estuaries throughout the southeast (Jacot,
1920; Broadhead, 1956; Anderson, 1958; Stenger, 1959;
Shireman, 1975; Dindo and MacGregor, 1981; Greeley
et al., 1987; Render et al., 1995; Hettler et al., 1997).
All of the fecundity specimens were caught from Oc-
tober through February when the mean monthly GSI
was highest. Pien and Liao (1975) found that mullet
oocytes reached a hydrated size of 900 to 1000 ^m. The
size of oocytes used for fecundity counts in the present
study ranged from 463 to 682 ^m. The maximum size of
oocytes in the tertiary stage of vitellogenesis from our
study was 600 ^m or greater. This result agrees with
those of previous studies where the maximum size of
oocytes prior to either hydration or atresia (if spawning
did not occur) ranged from 600 to 700 ^m (Shehadeh et
al., 1973b; Kuo et al., 1974). There was no evidence of
prespawning atresia in any of the specimens used for
fecundity estimates.
The appropriateness of using a GSI alone to determine
the level of reproductive development has been questioned,
particularly for serial or asynchronous spawning fishes (De
Vlaming et al., 1980; Hunter and Macewicz, 1985). Striped
mullet can have a wide range of GSI values that range from
practically zero to over thirty (Render et al., 1995). The GSI
range for females in our study ranged from almost zero to
27.7. Because of the high variability in GSI with size, it
does not appear appropriate to use GSI alone in order to
assess reproductive development in striped mullet. When
used in conjunction with histological analysis and mean
oocyte diameter of tertiary-stage oocytes, GSI does provide
excellent supporting evidence of reproductive schedules
and spawning season duration. GSI is probably more ap-
propriately used for isochronal spawning fishes than for
serial spawning fishes because of the uniform development
of oocytes in the former. However, it is still difficult to meet
all the basic assumptions of the GSI index as given by De
Vlaming et al. ( 1980) because of the high variability of GSI
with size. Another technique that has been used in aquacul-
ture situations to assess maturity and sex involves the use
g. 0
o
6
c^
>^
g 200
0)
3
CT
Developing males
n = 340
n
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
B
Developing females
n = 277
I
n
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 8
Frequency distribution of reproductively developing male (A)
and female (B) striped mullet by month in South Carolina
estuaries from 1998 to 2000.
of a cannula to remove oocytes from the ovaries of live fish
which are then evaluated (Shehadeh, et al, 1973a; Kuo et
al., 1974). This technique, although useful for determining
sex and the extent or stage of reproductive development,
would be inappropriate for estimating potential fecundity.
Historically, reproductively developing mullet have been
found in southeast United States waters from November
through December (Jacot, 1920; Anderson, 1958; Stenger,
1959). During our study, reproductively developing mullet
were caught in the Charleston Harbor Estuary from Oc-
tober through February. Gonad development of these fish
was discernible through gross morphological observation,
and histological sections showed that vitellogenesis was
well underway. Other studies have shown that previtel-
logenic oocytes were usually less than 160 jum and that
the onset of vitellogenesis began when the oocytes reached
a size of 180 pm (Dindo and MacGregor, 1981; Greeley et
al., 1987; Tamaru et al., 1994; Render et al., 1995). Devel-
oping individuals caught during our study that were not
used for fecundity counts had less-developed ovaries, GSI
values less than 7, and mean oocyte diameters less than
832
Fishery Bulletin 101(4)
16
15
14
13
12 -
^ 11
1 5
O
5 16
15
14
Prediclion Inlerval (2 SD)
Confidence Intenal (95%)
Ln Fecundity = -6.86 + 3.42 (Ln TL)
r^ = 0.W3, F = 527.2. df= 129
50
5.75
6.00
Ln Total length
6-25
6.50
13
12
Prediction Inlerval (2 SD)
Confidence Interval (95%)
^ -' . ^
-
^ ,<>^-^^*
-
^ • . » y^
^^^'
-
•
Ln Fecurulitx = 6.95 + 1.05 (Ln BW)
B
1 1
r = 0.804. F = 530.6. rf/-= 129
1 1 ! 1
5.0
5.5
6.0
6.5
Ln Body weight
70
7.5
80
Figure 9
Regression of the natural log (Ln) transformation of individual fecundity on
total length (A) and body weight (B) for striped mullet combined data, 1998 to
2000, from South Carolina estuaries, n (number offish in sample) = 129.
350 jjm. The vitellogenic activity in these ovaries was still
in the primary and secondary stages. These specimens
could not be used for fecundity counts because not all
of the oocytes destined for that year's spawning batch
had developed enough to be separable from the smaller
oocytes that would not develop. Once the developing
oocytes reached a size of 400 ^m or larger, they became
more uniform in size and in appearance and it was obvi-
ous which oocytes would constitute that year's spawn. At
that point, fecundity could be determined more accurately
because all the oocytes to be counted were significantly
and equally larger. This point in particular is important
in that it makes fecundity estimates from nonhydrated oo-
cytes more accurate for isochronal spawning fishes, such
as striped mullet. Fecundity estimates made in fishes that
are batch-spawners should only be made from hydrated
oocytes because of the presence of multiple developmental
stages (Hunter and Macewicz, 1985). The presence of dif-
ferent vitellogenic stages in the ovary of a repeat-spawn-
ing fish makes it necessary to determine individual batch
fecundity and spawning frequency before any estimate
of annual fecundity can be made. In isochronal spawn-
ing fishes, such as striped mullet, this process is made
simpler by the fact that oocytes mature at a similar rate
(Greeley et al., 1987). During early vitellogenesis (180 ^m
to 350 fjm), there is a higher degree of variability in the
rate of development and a range of developmental stages
would be present from the presence of precortical alveoli
through secondary and tertiary vitellogenesis (Render et
al., 1995). Estimating numbers of oocytes (uniform in an
individual but varying in size because of season) during
this stage would naturally make oocyte density and oocyte-
size relationships inconsistent. This could possibly be cor-
rected by using some timing factor such as month during
the spawning season. The present study did demonstrate
an inverse relationship between oocyte density and oocyte
diameter When month of capture was taken into consider-
ation, oocyte density decreased with increasing oocyte size
as the spawning season progressed.
In conclusion there were several biological aspects of
striped mullet reproduction demonstrated in this study.
Fecundity levels in striped mullet increased with total
McDonough et al,; Fecundity and spawning season oi Mugil cephalus
833
0 1998
-^ 1999
___ 700 -
O 2000
^
0)
1 650 -
o
E.
11
I <> °
S 600 -
<> ()
o <>
■o
B "" "
>.
o
o
O
f
!
L
500 -
OCT
NOV DEC JAN FEB
Month
Figure 10
Mean oocyte diameter per
month during the spawning season for striped
mullet in South Carolina
estuaries 1998 to 2000. Error bars represent
standard deviation.
length (TL) and body weight (BW=TW-OW). Oocyte den-
sity remained relatively stable with size in the fecund fish
and this allowed reasonable estimates of potential fecun-
dity based on total length and body weight. Age-specific
fecundity was highly variable and there appeared to be no
consistent relationship. The reproductive season for striped
mullet in South Carolina extends from October through
April as determined by mean monthly GSI levels and
histological confirmation of reproductive state. Potential
spawning periods would also occur within this period as
evidenced by the elevated GSI levels and the presence of
atretic (or possible postspawning) ovaries from December
through May. The gonadosomatic index itself is more useful
to evaluate reproductive potential when used in conjunc-
tion with other techniques, such as histological analysis
and oocyte diameter. The models of potential fecundity as
they relate to size (total length and body weight) could be
useful when applied to catch statistics of length and weight
in populations with known size- and age- frequency distri-
butions. This application would allow reasonable estimates
of potential fecundity for these populations.
Acknowledgments
This study would not have been possible without the assis-
tance of everyone in the Inshore Fisheries group at the
Marine Resources Research Institute of the South Carolina
Department of Natural Resources, which includes Myra
Brouwer, John Archambault, Hayne Von Kolnitz, Will
Hegler, Erin Levesque, Alice Palmer, Chad Johnson, Richie
Evitt, Larry Goss, and Travis Waits. We especially thank
Chad Altman of the South Carolina Department of Health
and Environmental Control for the freshwater specimens.
We also thank Myra Brouwer for assistance with map fig-
ures and the anonymous referees for careful review and
suggestions for this manuscript. This research was made
possible by National Marine Fisheries Service MARFIN,
grant no. NA77FF0550.
Literature citations
Anderson, W. W.
1958. Larval development, growth, and spawning of striped
mullet (Mugil cephalus) along the south Atlantic coast of
the United States. Fish. Bull. 58:501-519.
Alvarez-LaJonchere, L.
1982. The fecundity of mullet (Pisces, Mugilidae) from
Cuban waters, J. Fish Biol. 21:607-613,
Arnold, E. L,, and J. R. Thompson,
1958. Offshore spawning of the striped mullet, Mugil cepha-
lus, in the Gulf of Mexico. Copeia 1958: 130-132.
Broadhead, G. C.
1956. Growth of the black mullet, Mugil cephalus, in west
and northwest Florida. Mar. Lab. Tech, Ser 25:1-29.
Campana, S. E., M. C. Annand, and J. L McMillan,
1995, Graphical and statistical methods for determining
the consistency of age determinations. Trans. Am. Fish.
Soc. 124:131-138.
Chubb, C. F., I. C. Potter, C. J. Grant, R. C. J. Lenanton, and
J. Wallace.
1981. Age, structure, growth rates, and movements of sea mul-
let, Mugil cephalus L., and yellow eye mullet, Aldrichetta
forsteri (Valenciennes), in the Swan-Avon river system. West-
ern Austrailia. Aust. J. Mar. Freshw. Res. 32:605-628.
Collins, M. R., and B. W. Stender.
1989. Larval striped mullet (Mugil cephalus) and white
mullet (Mugil curema) off the southeastern United States.
Bull. Mar. Sci. 45(3):580-589.
834
Fishery Bulletin 101(4)
DeMartini, E. E., and B. B. Lau.
1999. Morphometric criteria for estimating sexual maturity
in two snappers, Etelis carbunculus and Pristipomoides
seiboldii. Fish. Bull. 97:449-458.
De Vlaming, V. L., H. S. Wiley, G. Delahunty, and R. Wallace.
1980. Goldfish (Carassius auratus) vitellogenin: induction,
isolation, properties and relationship to yolk proteins.
Comp. Biochem. Physiol. B 67B:613-623.
Dindo, J. J., and R. MacGregor
1981. Annual cycle of serum gonadal steroids and serum
lipids in striped mullet. Trans. Am. Fish. Soc. 110:403-
409.
Finucane, J. H., L. A. ColHns, and L. E. Barger
1978. Spawning of the striped mullet, Mugil cephalus, in
the northwestern Gulf of Mexico. Northeast Gulf Sci. 2:
148-150.
Fox, M. G., and A. J. Crivelli.
1998. Body size and reproductive allocation in a multiple
spawning centrarchid. Can. J. Fish. Aquat. Sci. 55(3):
737-748.
Greeley, M. S., D. R. Calder, and R. A. Wallace.
1987. Oocyte growth and development in the striped mullet,
Mugil cephalus, during sesonal ovarian recrudescence: rela-
tionship to fecundity and size at maturity. Fish. Bull. 85:
187-200.
Hettler, W. F., D. S. Peters, D. R. Colby, and E. H. Laban.
1997. Daily variability in abundance of larval fishes inside
Beaufort Inlet. Fish. Bull. 95:477-493.
Humason, G. L.
1967. Animal tissue techniques, 426 p. W.H. Freeman and
Co., San Francisco, CA.
Hunter, J. R., and B. J. Macewicz.
1985. Measurement of spawning frequency in multiple
spawning fishes. In An egg production method for esti-
mating spawning biomass of pelagic fish: application to the
northern anchovy, Engraulis mordax. NOAA Tech. Rep.
NMFS 36:79-94.
Jacot, A. P.
1920. Age, growth, and scale characters of the mullets, Mugil
cephalus and Mugil curema. Trans. Am. Fish. Soc. 39(3):
199-229.
Kaunda-Arara, B., and M. J. Ntiba.
1997. The reproductive biology of Lutjanus fulviflamma
(Forsskal, 1775) (Pisces: Lutjanidae) in Kenyan inshore
marine waters. Hydrobiologia 353(1/3):153-160.
Kelly, C. D.
1990. Effects of photoperiod and temperature on ovarian
maturation in the striped mullet, Mugil cephalus. Pac.
Sci. 44(2):187-188
Korhola, A., P. Solemdal, P. Bratland, and M. Fonn.
1996. Variation in annual egg production in individual cap-
tive Atlantic cod (Gadhus morhua). Can. J. Fish. Aquat.
Sci. 53(3 ):610-620.
Kuo, C. M.
1995. Manipulation of ovarian development and spawn-
ing in grey mullet, Mugil cephalus L. Israel J. Aquacult.
Bamidgeh 47(2):43-58.
Kuo, C. M., C. E. Nash, and Z. H. Shehadeh.
1974. A procedural guide to induce spawning in grey mullet
(Mugil cephalus L.). Aquaculture 3(1974): 1-14.
McDonough, C. J., and C. A. Wenner
2003. Growth, recruitment, and abundance of juvenile
Mugil cephalus in South Carolina estuaries. Fish. Bull.
101:343-357.
Okumus, I., and N. Bascinar.
1997. Population structure, growth, and reproduction of
introduced mullet, Mugil so-iuy, in the Black Sea. Fish.
Res. 33:131-137.
Pien, P. C, and J. C. Liao.
1975. Preliminary report of histological studies on the grey
mullet gonad related to hormone treatment. Aquaculture
5:31-39.
Render, J. H., B. A. Thompson, and R. L. Allen.
1995. Reproductive development of striped mullet in Loui-
siana estuarine waters with notes on the applicability of
reproductive assessment methods for isochronal species.
Trans. Am. Fish. Soc. 124( 1 ):26-36.
Rossi, A. R., M. Capula, D. Crosetti, D. E. Campton, and L. Sola.
1998. Genetic divergence and phylogenetic inferences in five
species of Mugilidae (Pisces: Perciformes). Mar Biol. 131:
213-218.
Shehadeh, Z. H., C. M. Kuo, and K. K. Milisen.
1973a. Validation of an in vivo method for monitoring ovar-
ian development in the grey mullet (Mugil cephalus L.). J.
Fish Biol. 1973(5):489-496.
Shehadeh, Z. H., W. D. Madden, and T P Dohl.
1973b. The effect of exogenous hormone treatment on
spermiation and vitellogenesis in the grey mullet, Mugil
cephalus L. J. Fish Biol. 1973(5):479-487.
Shireman, J. V.
1975. Gonadal development of striped mullet (Mugil cepha-
lus) in freshwater Prog. Fish Cult. 37(4):205-208.
Stenger. A. H.
1959. A study of the structure and development of certain
reproductive tissues of Mugil cephalus Linnaeus. Zoolog-
ica 44(2):53-70.
Tamaru, C. S., C. S. Lee, C. D. Kelley, G. Miyamoto, and
A. Moriwake.
1994. Oocyte gi'owth in the Striped mullet, Mugil cepha-
lus L., at different salinities. J. World Aquacult. Soc. 25:
109-115.
Thomson, J. M.
1951. Growth and habits of the sea mullet, Mugil dobula
Gunther, in Western Australia. Aust. J. Mar Freshw. Res.
2:193-225.
1963. Mullet life history strategies. Aust. J. Sci. 25:414-416.
The grey mullets. Oceanogr. Mar. Biol. Annu. Rev. 4:
301-335.
Wallace, R. A., and K. Selman.
1981. Cellular and dynamic aspects of oocyte growth in
teleosts. Am. Zool. 21:325-343.
Wenner, C. A., B. A. Roumillat, and C. W. Waltz.
1986. Contributions to the life history of black seabass, Cen-
tropristis striata , off the southeastern United States. Fish.
Bull. 84:723-741.
Zar, J. H.
1984. Biostatistical analysis, 2"'' ed., p. 292-305. Prentice
Hall Inc., Englewood Cliffs, NJ.
835
Abstract— Fishes are widely known
to aggregate around floating objects,
including flotsam and fish aggregating
devices (FADs). The numbers and diver-
sity of juvenile fishes that associated
with floating objects in the nearshore
waters of the eastern tropical Pacific
were recording by using FADs as an
experimental tool. The effects of fish
removal. FAD size, and the presence
or absence of a fouling community at
the FAD over a period of days, and the
presence of prior recruits over a period
of hours were evaluated by using a
series of experiments. The removal of
FAD-associated fish assemblages had a
significant effect on the number of the
dominant species (Abudefduf troschelii)
in the following day's assemblage com-
pared to FADs where the previous day's
assemblage was undisturbed; there was
no experimental effect on combined spe-
cies totals. Fishes do, however, discrimi-
nate among floating objects, forming
larger, more species-rich assemblages
around large FADs compared to
small ones. Fishes also formed larger
assemblages around FADs possessing
a fouling biota versus FADs without a
fouling biota, although this effect was
also closely tied to temporal factors.
FADs enriched with fish accumulated
additional recruits more quickly than
FADs that were not enriched with fish
and therefore the presence of prior
recruits had a strong, positive effect on
subsequent recruitment. These results
suggest that fish recruitment to float-
ing objects is deliberate rather than
haphazard or accidental and they sup-
port the hypothesis that flotsam plays
a role in the interrelationship between
environment and some juvenile fishes.
These results are relevant to the use
of FADs for fisheries, but emphasize
that further research is necessary for
applied interests.
Marine fish assemblages associated with
fish aggregating devices (FADs):
effects of fish removal, FAD size,
fouling communities, and prior recruits
Peter A. Nelson
Department of Biological Sciences
Northern Arizona University
Flagstaff, Arizona 86011-5640
Present address: Center for Marine and Biodiversity & Conservation
Scripps Institution of Oceanography
University of California, San Diego
La Jolla, California 92093-0202
E-mail address, pnelson@ucsd.edu
Manuscript approved for publication
13 June 2003 by Scientific Editor.
Manuscript received 26 June 2003
at NMFS Scientific Publications Office.
Fish. Bull. 101:835-850
Fishes associate with floating objects
in nearly all oceans of the world (Good-
ing and Magnuson, 1967; Hunter and
Mitchell, 1967; Klima and Wickham,
1971; Crawford and Jorgenson, 1993;
Kingsford, 1993; Druce and Kingsford,
1995; Massuti et al., 1998; Hampton
and Bailey, 1999; Parin and Fedoryako,
1999). Fishes also gather around fish
aggregating devices (FADs), floating
objects deployed to concentrate target
species or bait fishes and improve the
catch for artisanal, sport, or commercial
fisheries. The physical attributes of a
floating object, such as a FAD, may affect
the ability of potential fish recruits to
locate the floating object or may affect
the adaptive advantages of associating
with that object (or both) — a topic that
has been addressed in numerous prior
studies (e.g. Hunter and Mitchell, 1968;
"Wickham et al., 1973; Wickham and
Russell, 1974; Fedoryako, 1989; Roun-
tree, 1989; Safran, 1990; Safran and
Omori, 1990; Friedlander et al., 1994;
Hall et al., 1999b). However, the present
study is apparently the first to address
empirically the effects of disturbance,
fouling communities, and prior recruits
by examining both the number and
the diversity of fishes that aggregate
around FADs. In addition, this study
addresses the effect of FAD size, a factor
well represented in prior studies but
frequently confounded by temporal or
design issues. FADs are widely used to
enhance sport and commercial fisheries,
but are expensive to build, deploy, and
maintain; therefore better information
about the effects of FAD size and foul-
ing could aid design efforts. Given the
bycatch associated with the FAD fishery
for tuna in the eastern tropical Pacific
(Hall et al., 2000), for example, we need
a better understanding of how fishes
use FADs in order to manage fisheries
for FAD-associated species (Lennert-
Cody and Hall, 2000). Finally, careful
study of how differing characteristics
of floating objects affect fish recruit-
ment may provide important clues
regarding the adaptive significance of
fish associations with flotsam and drift
algae — a phenomenon widely noted but
poorly understood.
Prior research has suggested that
rates of immigration and fish removal
from FADs similar to those seen in the
present study were high from one day to
the next (Nelson, 1999), and "Wickham
and Russell (1974) reported that mid-
water FADs, which were fished daily,
produced a larger cumulative catch
than mid-water FADs, which were un-
disturbed during the same period and
then fished once at the end of the con-
clusion of the study. I tested the hypoth-
esis that, over time, the size and diver-
sity of FAD-associated fish assemblages
are reduced by the repeated removal of
these fishes compared with undisturbed
assemblages. Effective management
or use of FADs deployed for fisheries
purposes and an understanding of the
ecological relationship between flotsam
and fishes associated with flotsam will
depend in part on patterns of immigra-
tion and loss to fish assemblages.
836
Fishery Bulletin 101(4)
There have been numerous attempts to
equate flotsam structure (size, complexity,
orientation, etc.) with the number of associ-
ated fishes (e.g. Hunter and Mitchell, 1968;
Dooley, 1972; Wickham et al., 1973; Wickham
and Russell, 1974; Rountree, 1989; Druce and
Kingsford, 1995), but the results have been
equivocal, except when the analysis was re-
stricted to a single species (e.g. Histrio histrio,
Dooley, 1972;Decapterus punctatus, Rountree,
1989). Huge aggregations have been associ-
ated with very small objects — lATTC (Inter-
American Tropical Tuna Commission ) records
include a report of 55 metric tons of mostly
yellowfin tuna {Thiinnus albacares) fished
from beneath a 1-m length of floating poly-
propylene rope (Hall et al., 1999b); therefore,
despite the intuitive appeal, there is no clear
reason to expect that size of FAD per se is an
important factor in determining the size of
associated assemblages. Thus, object size re-
mains an unresolved problem in understand-
ing flotsam-associated communities. If there
are optimal FAD sizes, these may be species
specific, and economical FAD design depends
upon controlled experiments in the field.
Fouling organisms (sessile invertebrates
and algae that colonize flotsam) are believed
to have a strong, positive effect on the sub-
sequent recruitment and retention of fishes
by commercial fishermen (Gaertmer and
Medina-Gaertner, 1999; Hall et al,, 1999a;
Hallier and Parajua, 1999; Suzuki, 1999).
However, prior to the results presented here,
there appear to have been no controlled tests
of the hypothesis that the presence of fouling
organisms enhances fish recruitment to a
floating object. I also compared the numbers and diversity
of fishes associated with FADs that are equipped with ar-
tificial (lead weight) fish versus FADs without these artifi-
cial fish. The latter experiment was intended to determine
the importance of prior recruits to subsequent patterns of
recruitment. To test a similar hypothesis over the short
term (hours versus days) and using living fish instead
of painted models, I also compared recruitment to FADs
enriched with real fish (juvenile Abiidefduf troschelii) to
unenriched FADs.
I tested the hypothesis that each of these factors would
affect the number of fishes associated with FADs (combined
and individual species), as well as the species diversity of
FAD-associated fish assemblages. Both the size of these
FAD-associated fish assemblages and their species diver-
sity provide insight on recruitment processes and the use of
floating objects by fishes. Although the association of fishes
with floating objects has been well documented, very little
is known regarding the behavioral and ecological processes
behind these assemblages. The results reported in the pres-
ent study provide new information on the role of flotsam
and FAD characteristics in determining the number and
diversity of these assemblages, and some clues towards
Figure 1
Location of the study site at Achotines, Panama, Central America.
understanding why and how fishes aggregate beneath
floating objects.
Materials and methods
Study site and FAD construction
All research was conducted between July and October 1997
on the Pacific coast of Panama, Central America, from the
Inter-American Tropical Tuna Commission laboratory at
Achotines, near the tip of the Azuero Peninsula (Fig. 1).
Experimental FADs were constructed of three tuna purse-
seine buoys lashed together and anchored to the substrate
with a 25-kg cast concrete block unless otherwise noted
(Fig. 2). The length of the anchor lines allowed the FADs
to rest at the surface at all tidal heights. Each buoy was
roughly 25 cm in cross sectional diameter, and approxi-
mately 35 cm in length. The FADs were detachable from
their moorings by detaching a large (2 m diameter) loop on
the anchor line that held a 2-kg line weight (Fig. 2). This
design allowed me to change FADs for another treatment.
The FAD arrays were deployed nearshore (within 1.5 km;
Nelson: Fad characteristics and associated fish assemblages
837
Fig. 3) and in shallow water (14-25 m). The FAD
treatments were not assigned randomly to FAD
positions; instead, I assigned treatments uni-
formly across the FAD array because the total
number of FADs was relatively low (8-10) and,
given the small number of experimental units,
in order to reduce the possibility that the results
be confounded by positional effects.
The anchored FADs were spaced approximately
100 m apart. The maximum horizontal underwa-
ter visibility measured was 27 m, and typically
averaged much less. Assuming that vision was
the principal means by which fishes located
these objects, it is therefore highly unlikely that
fishes treated the FAD array as a single "object"
or moved from FAD to FAD within the array. I
believe that it was unlikely that any fish trans-
ferred from one FAD to another for the following
reasons. 1) The horizontal underwater visibility
was always much less than the distance between
FADs. 2) Observations suggest that short-term,
daytime fidelity was high; once a fish associated
with a floating object, it was unlikely to leave that
object during the day (Nelson, 1999). 3) Crossing
an open stretch of water for another floating ob-
ject (presumably within detection range) entails
a potential risk for a fish. Moser et al. (1998) did
note that the larger juveniles and adult fishes
associated with floating Sargassum showed little
apparent fidelity to this habitat and would move
between their research boats, floating observation
equipment, and the Sargassum habitat. However,
the fishes that were observed to move between
these floating objects were juvenile carangids be-
tween 10 and 20 cm in length (Moser et al., 1998),
whereas the fishes in the present study were gen-
erally much smaller and presumably less vagile.
Longshore currents ran roughly west to east
through the experimental area, and rarely in the
reverse (Fig. 3). I recorded an estimate of cur-
rent direction using an underwater compass and
the angle of the FAD anchor lines. This estimate
represented the sum of the forces due to windage
on the FAD buoys and currents.
Censusing FADS
FAD-associated fish assemblages were censused
by direct visual observation by divers using mask
and snorkel. Most other studies offish assemblages
associated with floating material have employed
nets or quantitative fishing methods for sampling
purposes (e.g. Kojima, 1960; Dooley, 1972; ICings-
ford, 1992, 1995), but Hunter and Mitchell (1968)
compared data from net captures, automated
photography and direct visual observations and
found that the visual observations agreed well
with the other methods, and provided behavioral
information not available with the other methods.
In my study, a FAD was approached by swimming
Fish Aggregation Device
(FAD)
2 kg line weight
cast concrete with 6.25 mm
polypropylene loop
FAD line & anchor line
9 mm blue polypropylen
to bottom & 25 kg
concrete anchor
Figure 2
Design of detachable fish aggregating devices (FADs) used in the
present study.
— '-, ^Laboratorio A^fiotines
long-shore Currents
North
\to Los Frailles
i9-12km
rocky coast
sand beach
experimental FAD array
(general location)
t
drifting FAD experinnents
conducted in this area
Figure 3
Coastline and general location of experiments and observations. The
laboratory iLaboratorio Achotines) is located at approximately T'S'N lat,
SOnCW long.
838
Fishery Bulletin 101(4)
slowly and quietly at the surface from a distance of at least
12 m. All fishes associated (defined below) with the floating
object were counted and identified; therefore the statistical
unit in all of the experiments described below was a single
FAD-associated assemblage of fishes at a given date and
time. Horizontal underwater visibility, measured with a
Secchi disk, was always sufficient to allow the identification
of species and to count individual fishes from a minimum
distance of 2 meters.
Any fish observed within 2 m of a FAD was considered to
be "FAD-associated." Very few fishes were observed outside
of this range, and with rare exceptions, fishes responded
to the approach of an observer first by swimming towards
the observer and then by moving closer to the FAD, rather
than away from it. Different species of fishes used the
space around and below the FAD differently, as both Good-
ing and Magnuson (1967) and Hunter and Mitchell (1967)
described, but the juvenile fishes that predominated in the
present study were unambiguous regarding their relation-
ship to the FADs. After fishes resumed their prior positions
in relation to a FAD, continued observations of these fishes
revealed that there was no inclination to abandon that FAD.
The appearance of potential predators invariably resulted
in a tightening of the spatial distribution around the FAD.
When the experiment required the capture of FAD-as-
sociated fishes, I used a smaller (1.1x1.3 m) version of the
diver-operated liftnet described by McCleneghan and Houk
(1978). Captured fishes were preserved for further studies,
held in grow-out facilities in the laboratory to verify spe-
cies identification, or released 1.5 km down-current over
rocky reef habitat to ensure that they had effectively been
removed from the FAD array.
Diversity calculations
Measurements of species diversity provided a means of
monitoring treatment effects on the composition of FAD-
associated assemblages. I measured species diversity using
species richness (S, the raw number of species observed),
and the Brillouin index (HB). S is simple and widely
used, but increases with sample size, and, where sample
sizes are unequal, HB provides a less biased measure of
diversity (Magurran, 1988). In addition, HB was chosen
over one of the more commonly used information theory
indices (e.g. the Shannon-Wiener index) because 1) FAD-
associated assemblages are not a random sample of poten-
tial recruits (different species vary in their attraction to
floating objects) and 2) each of these assemblages was
completely censused — not sampled (Magurran, 1988, and
references therein).
HB is calculated as
///? =
In/V!
-Z'"".^
where N = the total number of fishes of all species
observed; and
;i, = the number of individuals within the ith spe-
cies (Magurran, 1988).
Statistical analyses
1 used a two-way repeated measures ANOVA (a=0.05) to
test for treatment differences, differences among observa-
tion dates and evidence of treatment-by-sample interaction
for assemblage sizes (no. of fish(es)), species richness (S)
and species diversity (HB). Because individual species dif-
fered in their relative abundance and had different ecologi-
cal requirements, there was the potential for the dominant
species to bias comparisons of experimental treatments
where assemblage size (a combinination of all species) was
used. For all experiments except for the recruit-enriched
experiment, I repeated the statistical analyses twice: once
using the number of sergeant major damselfish (Abudefduf
troschelii) only and again using all fishes combined but
with A. troschelii removed. For the artificial fish experi-
ment where rainbow runner (Elagatis bipinnulata) were
particularly abundant, I ran separate analyses for A. tros-
chelii alone, £. bipinnulata alone, and for all species minus
the numbers of A. troschelii. When no fishes were present
at a FAD, HB was undefined; the "missing" data were re-
placed according to the procedures of Zar ( 1996) and the
degrees of freedom were reduced accordingly.
Fish-removal experiments
Observations on similar FADs during a previous field
season at the same location suggested that the turn-
over rate of fish associated with anchored FADs is high,
especially when the initial assemblage is large, but that
some recognizable individuals did persist from day-to-day
(Nelson, 1999). If immigration and emigration rates were
as high as suspected, FADs cleared offish on a daily basis
should not differ significantly from undisturbed FADs in
their mean assemblage size or in the average number of
species associated with these FADs. Wickham and Russell
(1974) compared the catches of bait fishes associated with
FADs subject to daily purse-seine sets versus those allowed
to "soak" undisturbed for three days prior to a single
purse-seine set and concluded that sufficient emigration
and immigration occurred on a daily basis to remove any
appreciable effect of daily removals. I sought to address
similar questions, but by using a different system (juvenile
reef fishes versus bait fishes).
To test these hypotheses, 1 deployed eight identical FADs
on 30 June 1997 in two lines of four FADs each, oriented
roughly parallel to shore (Figs. 2 and 3). Fishes associ-
ated with all eight FADs were counted and identified on
a daily basis, beginning 3 July 1997. Alternate FADs were
cleared of all fish, following the daily counts; the remaining
FADs were left undisturbed in such a way as to distribute
the treatments evenly among the FAD array. After three
consecutive days of these observations (series 1), the treat-
ments were reversed, and previously undisturbed FADs
were cleared, and those that had been cleared regularly
were left undisturbed (series 2). The treatments were re-
versed in an attempt to control for possible positional effects
of the FADs. However, the two series were necessarily run
consecutively, not concurrently; therefore treatment posi-
tion was confounded by sample date. I used a 2 (cleared vs.
Nelson: Fad characteristics and associated fish assemblages
839
undisturbed) by 2 (first series vs. second series) by 6
(sample date) model and I used a repeated measures
ANOVA (repeated on sample date) on the follow-
ing dependent variables: assemblage size (total no.
of fishes), species richness (S), and HB. I repeated
analyses of assemblage-size effects looking at the
number of A. troschelii only, and the total number of
fishes minus the number of A. troschelii.
FAD size
To determine the effect of FAD size on the associ-
ated assemblage size and diversity, I compared
FAD-associated fish assemblages between triple-
size FADs and single FADs. An existing anchored
array of eight FADs (two lines parallel to the coast
of four FADs each. Fig. 3) was cleared of fishes on 24
July 1997. As the fish were removed from the FADs,
each FAD was replaced with a fresh (i.e. clean and
unfouled) single or triple-size FAD, placed at alter-
nating positions. The single FADs were constructed
as described above and in Figure 2; the triple-size
FADs were identical to the single FADs, except
that they consisted of nine, rather than three,
purse-seine buoys lashed together and had the
effect of nearly tripling the wetted surface area
(although inner buoys are less exposed than outer
ones) and the volume of the FAD, and of increasing
the maximum linear dimension of the FAD by a factor of
two. Treatments were not reversed for this or subsequent
experiments because sample date appeared to be the major
factor determining assemblage size for any species, based
on the previous experiment. Note that in each of these
experiments, except for the recruit-enrichment experi-
ment that used drifting FADs, treatments were assigned
uniformly throughout the FAD arrays so that onshore,
offshore, or longshore biases in recruitment due to oceano-
graphic processes would not confound the results. Fishes
at all FADs were counted and identified on three dates (26,
28, and 30 July 1997), each observation separated from the
next by 48 hours. No fish were collected, with the exception
of one balistid, taken from the array on 26 July because
it was the first of that species to be observed associated
with a FAD. Data were analyzed for experimental effects
on total assemblage size, species diversity (S and HB), the
number of A. troschelii, and total number of fishes minus
the number of A. troschelii.
Presence of absence of a fouling community
To determine whether the presence of a fouling community
on a floating object affected the associated fish assemblage,
I compared FAD-associated assemblage sizes and species
richness between fouled and unfouled (control) FADs. Con-
trol FADs were scrubbed of all fouling organisms. Fouled
FADs had been deployed for a minimum of 14 days (range:
14-22 days) in the study area, and had accumulated fouling
that completely covered the wetted surface of the FAD with
gooseneck barnacles (Lepas sp.), hydroids, and bryozoans.
Grapsid crabs and polychaete worms (Amphimone vagans)
Fish Aggregation Device
(FAD)
steel hoop to
distribute model \
aggregation
1 oz lead
1 oz lead
weight,
weight,
vertical
model fish
orientation
(painted)
to line weight and anchor
Figure 4
Treatments for the artificial fish experiment involved suspending
artificial fish, lead weights of an equivalent mass and volume, or
nothing (control) from a steel hoop lashed beneath the FAD.
were also intermittent associates of fouled FADs. Control
and fouled FADs were deployed on 8 September 1997 in an
alternating array of eight buoys, with four FADs per treat-
ment (layout and spatial distribution of treatments follow
that of the FAD size experiment). All fishes were cleared
from FAD positions prior to deploying the FADs, and data
collection commenced 24 hours later Data were collected
on four consecutive days (9, 10, 11, and 12 September 1997)
and analyzed for experimental effects on total assemblage
size, species diversity (S and HB), the number of A. tros-
chelii, the number of E. bipinnulata, and total number of
fishes minus the number of A. troschelii.
Artificial fish experiment
I tested the hypothesis that potential recruits would dis-
tinguish between FADs with an "assemblage" of artificial
fish suspended beneath them, FADs with an "assemblage"
of suspended material equal to the artificial fish in size
but not resembling fish in appearance, and control FADs
without anything suspended beneath them (Fig. 4). I con-
structed artificial fish from 31.25-g lead fishing weights.
These weights were flattened, tear-drop-shaped objects,
painted a dull yellow with black bars to resemble juvenile
Abudefduf troschelii and suspended, by using monofila-
ment (20 lb. test) and a steel hoop, beneath the "artificial
fish FADs" (Fig. 4). I suspended oblong 31.25-g lead fish-
ing weights beneath "weighted FADs," and the control
FADs had only a steel hoop beneath each (Fig. 4). These
FADs were deployed in an anchored array, and the vari-
ous treatments were distributed in an alternating pattern
throughout the array. FAD positions were cleared of fishes,
840
Fishery Bulletin 101(4)
Table 1
Fish species and life history stages observed at all experimental FADs (combined data)
and abundance. 1 = coastal pelagic species; 2 = substrate-associated species; and 3 = po
A = adult.
with relative importance by frequency
ssible flotsam specialists. J = juvenile;
No,
Species
Family
Stage
Frequency (%) Abundance {%)
1
Abudefduf troschelii^
Pomacentridae
J
34.1 814
2
Etagatis bipinnulata^
Carangidae
J
16.9 8.1
3
Polydactylus approximans'^
Polynemidae
J
6.4 1.8
4
Mugil sp. ■^
Mugilidae
J
6.4 1.3
5
Lutjanus argentiventris^
Lutjanidae
J
6.4 1.1
6
Epinephelus panamensis'^
Serranidae
J
3.4 0.5
7
Hoplopagrus guntheri-
Lutjanidae
J
3.2 0.5
8
Canthidermis maculatus^
Balistidae
J
2.8 0.4
9
Gnathanodon speciosus-
Carangidae
J
1.5 0.2
10
Atectis ciliaris^
Carangidae
J
11
Caranx caninus^
Carangidae
J
12
Caranx caballus^
Carangidae
J
14.7 3.9
13
Caranx vinctus^
Carangidae
J
(nos. 10-14)
14
Seriola peruana^
Carangidae
J
15
Tylosaurus acus pacificus^
Belonidae
A
16
T. crocoditus fodiator^
Belonidae
A
17
Fistularia commersonii^
Fistulariidae
J
18
Syngnathus auliscus-
Syngnathidae
J
19
Lobotes pacificus^
Lobotidae
J and A
20
Mulloidichthys dentatus^
Mullidae
J
4.1 0.7
21
Sectator ocyurus^
Kyphosidae
A
(nos. 15-26)
22
Parapsettus panamensis^
Ephippidae
J
23
Hypsoblenmus breviceps^
Blenniidae
?
24
goby (unidentified) 2
Gobiidae
7
25
Aluterus scriptus'^-^
Balistidae
J and A
26
Batistes polytepis'^
Balistidae
J
26 species
16 families
100 100
and treatment FADs were deployed on 24 September 1997.
FADs were monitored daily as described above, from 25
September through 3 October 1997 (sampling days=9).
Data were analyzed for experimental effects on total
assemblage size, species diversity (S and HB), number of
A. troschelii. number of £. hipinnulata, and total number
of fishes minus the number of A. troschelii.
Recaiit-enriched vs. nonenriched FADs
I tested the hypothesis that the presence of prior recruits
(juvenile sergeant major damsclhsh, Abudefduf troschelii)
would have a positive effect on subsequent recruitment to
a FAD. I used A. troschelii because these were the most
important species associated with FADs by frequency and
abundance (Table 1). It is possible that the selection of a
particular species as the prior recruit might affect the sub-
sequent recruitment of the same or different species (via
intra- or interspecific competition for example), but I had
no basis for predicting the direction of such effects.
Given the strong day-to-day changes in assemblage sizes,
this test required frequent, short-interval observations of
the experimental FADs. I used drifting, rather than an-
chored, FADs to provide a more realistic (and conservative)
test of the effect. (Drifting objects should result in fewer
chance encounters by potential fish recruits carried by cur-
rents through a fixed FAD array, but anchored FADs are
much easier to track for longer experiments.) I deployed
four drifting FADs (constructed from 3 buoys — the "single"
size) in the stippled area indicated in Figure 3. Two of
these FADs were enriched with nine A. troschelii per FAD,
previously collected from anchored FADs and released in
close proximity to drifting FADs immediately after deploy-
ment. The two control FADs received no sergeant majors
to start. Both groups were checked immediately following
deployment to verify that the fish had associated with the
experimental FADs and to check against quick recruitment
to the control FADs. To minimize the potential transfer of
fish with the boat, I accelerated sharply when leaving a
FAD enriched with sergeant majors and when checking
Nelson: Fad characteristics and associated fish assemblages
841
the FADs, entered the water from the boat a minimum of
10 m from each FAD.
The FADs were deployed from an inflatable boat at 50-m
intervals in a roughly linear array, and checked at hourly
intervals. The FADs did not maintain their initial spatial
arrangement, but I did not move any FAD once the drift
began unless FAD-to-FAD distance had been reduced to
less than 10 m. In this instance, I moved one or more FADs
to a minimum FAD-to-FAD distance of 50 m after checking
for any FAD-associated fishes. In none of these instances
were any FAD-associated fishes observed. I monitored the
drift for four hours; deteriorating weather and fading light,
however, did not permit additional observations.
I used linear regression to test the hypothesis that the
number of FAD-associated fishes changed over time for
the enriched FADs and for the nonenriched FADs. I used
a ^-test to compare the slopes of the two regression models
and to test the hypothesis that the treatments accumulated
fish at different rates.
Results
Twenty-six species of fishes from 16 families were recorded,
including species associated with reef, soft bottom, and
coastal pelagic habitats as adults (Table 1). Only juvenile
specimens were observed clearly associated with FADs,
with the exception of Ahiterus scriptus and Lobotes paci-
ficus, of which both juvenile and adult forms were observed
in close, continuous proximity to the FADs. Two needlefish
species (Tylosaurus acus pacificus and T. crocodilus fodia-
tor) appeared occasionally in close proximity to the FADs,
but they were not clearly associated with the FADs. An
adult Lobotes pacificus (tripletail) was observed once and
a single adult Al uterus scriptus (scrawled filefish) were
observed on three separate instances. Horizontal under-
water visibility averaged 13.4 m (±1.7 SE) for all sampling
days combined.
Juvenile sergeant major damselfish (Abudefduf trosche-
lii) were the dominant species by frequency of occurrence
and numerical abundance (Table 1) for all experiments.
The damselfish was followed in rank overall by juvenile
rainbow runner (Elagatis bipinnulata), although this spe-
cies was observed with the FADs only during the fouling
and model fish experiments. Juvenile threadfin {Polydac-
tylus approximans), mullet (Mugil sp.), and yellow snapper
(Lutjanus argentiuentris) were equally frequent but dif-
fered slightly in abundance (P approximans>Mugil sp.>L.
argentiventns; Table 1). The latter pattern was consistent
across all experiments. Specimens from a suite of juvenile
carangids (excluding E. bipinnulata) were also observed
frequently.
Fish-removal experiments
Sample date, series, and treatment combined to have a
significant effect on A. troschelii abundance (three way
interaction, P=0.03), but there was no clear pattern; the
remaining species (combined species less numbers of A.
troschelii) were influenced by sample date (date by series
30
25
20
IS
10
LU
U)
A 0
A Abudefduf troschelii only
□ observed
□ removed -j-
1
^
8-1
0)
1 r
3-Jul 4-Jiil 5-Jul 6-Jul 7-Jul 8-]ul 9-Jul 10-|ul
B combined spp. - A. trosclielii
T
MA
i
I
3-Jul 4-Jiil 5-Jul 6-Jul 7-Jiil 8-JliI 9-Jul 10-]ul
Figure 5
Repeated fish-removal effects (disturbed vs. undisturbed
[observed only]) on aggregation size (mean no. of fishes) for
Abudefduf troschelii alone (A) and for all species combined
less A. troschelii (B). See Tables 3-5 for sample sizes and
ANOVA results for assemblage size and diversity measures.
interaction, P<0.01) but not by treatment (P=0.73, Table 2,
Fig. 5). Measures of diversity varied between series (series:
S, P<0.01; HB, P=0. 01) but were unaffected by treatment.
Thus, fish removal or fish disturbance may contribute to
assemblage sizes for individual species (e.g. A. troschelii),
but, in the present study, the total number of combined
species was unaffected.
FAD size
Abudefduf troschelii was strongly affected by a combina-
tion of treatment and sample date (date by treatment
interaction, P=0.03, Table 3, Fig. 6). Results from the
remaining species combined were comparable with larger
total numbers at the larger FADs, although not statisti-
cally significant (treatment, P=0.07). Although both mea-
sures of diversity (S and HB) suggested that the treatment
may have had a positive effect on diversity (S, treatment,
P=0.02), species richness was positively correlated with
sample size. HB, a diversity measure comparatively unaf-
842
Fishery Bulletin 101(4)
Table 2
Repeated measures ANOVA results (cleared FADs vs. undisturbed FADs, n.
=^2=12) in
fish-removal experiment.
Dependent variable
Factor(s)
F
df
P
1-/J
Number of fishes
treatment
0.12
1, 12
0.73
0.06
(all species combined)
series
35.3
1,12
<0.01
>0.99
treatment x series
0.64
1, 12
0.44
0.11
date
5.53
2,24
0.01
0.81
date X series
48.5
2,24
<0.01
>0.99
date X treatment
3.22
2,24
0.06
0.55
3-way interaction
0.06
2,24
0.94
0.06
Number offish
treatment
16.7
1,12
<0.01
0.97
(A. troschelii only)
series
5.92
1,12
0.03
0.61
treatment x series
0.27
1, 12
0.61
0.08
date
3.10
2,24
0.06
0.53
date X series
7.71
2,24
<0.01
0.93
date X treatment
2.95
2,24
0.07
0.51
3-way interaction
4.24
2,24
0.03
0.69
Number of fishes
treatment
0.05
1, 12
0.82
0.06
(all spp. minus A. troschelii)
series
13.3
1, 12
<0.01
0.93
treatment x series
0.01
1,12
0.94
0.05
date
2.93
2,24
0.07
0.51
date x series
4.06
2,24
0.03
0.66
date X treatment
0.07
2,24
0.93
0.06
3-way interaction
0.31
2,24
0.74
0.09
Species richness (S)
treatment
0.63
1,12
0.44
0.11
series
11.8
1,12
<0.01
0,90
treatment x series
0.63
1,12
0.44
0.11
date
0.43
2,24
0.66
0.11
date X series
0.63
2,24
0.54
0.14
date X treatment
0.69
2,24
0.51
0.15
3-way interaction
2.08
2,24
0.37
0.37
Species diversity (UB)'
treatment
1.05
1, 10
0.33
0.15
series
8.79
1, 10
0.01
0.79
treatment x series
0.54
1,10
0.48
0.10
date
1.43
2,22
0.26
0.27
date X series
2.80
2,22
0.08
0.49
date X treatment
0.12
2,22
0.12
0.07
3-way interaction
0.61
2,22
0.55
0.14
' Missing data were replaced accordi
to interactions between treatment
ng to the directions in Zar (1996), and the degrees of freedom
series, and dates.
were reduced accordingly. "3-way interaction" refers
fected by sample size (Magurran, 1988), was marginally
nonsignificant (HB, treatment, P=0.07, Table 3).
cies diversity (HB), though not richness (S), was signifi-
cantly affected by sample date (P=0.02).
Presence or absence of a fouling community
Treatment and sample date combined to have a significant
efTect on the number of A. troschelii (date by treatment
interaction, P<0.01) — an effect contributing to the similar
significant interaction effect for all species combined (Fig.
7, Table 4). Although the mean numbers of fish(es) were
consistently higher at fouled FADs i'or E. bipinnulata alone
and for all species minus A. troschelii, the only significant
main efTects were due to sample date (Table 4, Fig. 7). Spe-
Artificial fish experiment
Experimental treatments (FADs with model fish, with lead
weights or with nothing, Fig. 4) had no effect on any mea-
sured parameter — combined species, A. troschelii alone, E.
bipinnulata alone, combined species less A. troschelii, spe-
cies richness and diversity (Fig. 8, Table 5). All measures
were significantly affected (P<0.01) by sample date except
for£. bipinnulata alone (date, P=0. 48). Although individual
FADs varied in the number of associated E. bipinnulata.
Nelson: Fad characteristics and associated fish assemblages
843
Table 3
FAD-size
effects and repeated measures ANOVA results (single FADs vs
triple-size FADs
fll=n2=12).
Dependent variable
Factor(s)
F
df
P
1-/3
Number of of fishes
treatment
24.1
1,6
<0.01
0.99
(all species combined)
date
10.7
2,12
<0.01
0.97
date X treatment
2.56
2,12
0.12
0.41
Number offish
treatment
10.9
1,6
0.02
0.79
(A. troschelii only)
date
18.2
2,12
<0.01
>0.99
date X treatment
4.55
2,12
0.03
0.66
Number of fishes
treatment
4.84
1,6
0.07
0.45
(all spp-A. troschelii)
date
0.42
2,12
0.67
0.10
date X treatment
0.29
2, 12
0.75
0.09
Species richness (S)
treatment
11.3
1,6
0.02
0.81
date
0.38
2,12
0.69
0.10
date X treatment
0.38
2,12
0.69
0.10
Species diversity (HB)
treatment
5.00
1,6
0.07
0.46
date
3.39
2,12
0.07
0.52
date X treatment
0.40
2,12
0.68
0.10
these numbers were strikingly constant across sample date
and, to a lesser extent, across treatments (Fig. 8).
Recruit-enriched vs. nonenriched FADs
Enriched FADs showed significantly higher rates of
recruitment than nonenriched FADs: the regression line
for the enriched FADs had a significant slope (f , j g|=20.76,
P<0.01), but the regression line for the nonenriched FADs
did not (F,i g)=2.29, P=0.17; Fig. 6). All additional fish were
juvenile sergeant major damselfish, Abudefduf troschelii .
These slopes are significantly different (f=3.05, 2 tailed
test, v=6, P=0.02; Fig. 9); enriched FADs accumulated fish
at a significantly higher rate (2.5 fish per hour) than did
nonenriched FADs that accumulated fish at a rate of 0.1
fish per hour Horizontal underwater visibility was 15 m at
the beginning of the experiment.
Discussion
FAD size, the presence of a fouling community, and the
presence of prior recruits all had positive effects on the size
of FAD-associated assemblages, although the latter factor
was assessed over a period of hours, whereas the former
were assessed over days. The repeated removal of an exist-
ing assemblage also had significant effects due at least
partially to treatment, but in all of these analyses sample
date appeared to play the largest role in determining the
numbers of fish(es) at these FADs. The presence of artifi-
cial fish or comparable-size weights did not significantly
affect assemblage sizes. There was little support for the
hypothesis that any of these factors might affect the spe-
cies diversity of these assemblages; only species richness
was significantly increased along with an increase in FAD
size and this result may be an effect of assemblage size
rather than object characteristics. Where treatment effects
did significantly affect the numbers of fishes, their effects
50-
45-
40-
35-
30-
25
20
15-
10-
5
0
10-
9-
8-
7
6
5-
4-
3-
2-
1-
0-t
A Abudefduf troschelii only
D single
D triple
26-Jul
28-Jul
30-Jul
B comb ned spp. - A. troschelii
m
26-Jul
28-Jul
30-Jul
Figure 6
FAD-size effects (single FADs vs. triple-size FADs) on
aggregation size (mean no. of fishes) for Abudefduf trosche-
lii alone (A) and for all species combined less A. troschelii
(B). See Tables 3-5 for sample sizes and ANOVA results for
assemblage size and diversity measures.
844
Fishery Bulletin 101(4)
A Abudi'fduf irosdiclii only
.w-
D Clean
T
2S-
n Fouled
20-
X.
Ky?.
T
m.
]b-
%
ID-
■/>.
S-
T
'■*
T
T
0-
1
1 '
^
6
5
4-
3-
9-Sep
10-Sep
11 -Sep
12-Sep
B combined spp. - A. troschclii
p.,,
0-
m
W
6-
5-
4-
3-
2 -
1-
9-Sep 10-Sep 11 -Sep
C Lla%at'\s bipi)nuilnta only
12-Sep
mi
9-Sep
10-Scp
11 -Sop
12-Sep
Figure 7
Effects of fouling community (fouled vs. unfouled [control])
on aggregation size (mean no. of fishes) ior Ahudefduf tros-
chelii alone (A), for all species combined less A. troschelii
(B), and for Elagatis bipinnulata alone (C). See Tables 3-5
for sample sizes and ANOVA results for assemblage size
and diversity measures.
on Abudefduf troschelii were generally the strongest. It is
not clear whether this is a species-specific effect or if these
results are due to the fact that A. troschelii was the most
numerically important species.
The absence of a significant treatment main effect in the
fish-removal experiments suggests that recruitment and
loss from these anchored FADs is sufficiently rapid so that
I20-|
A Abudefduf troschelii only
1011-
n Control
■-
so-
□ VVi'ighls
J
60-
■ Models
r
40-
_ -p
■r T
20-
0 J
fi.Jf1
\\
\i\
1
C^D-D-Q-Ci-C-t; r, T,
^ ^ ^ rj Z} zj f^ r~^^ r^
0.99
(A. troschelii only)
sate
19.7
3,18
<0.01
>0.99
date X treatment
25.2
3,18
<0.01
>0.99
Number offish
treatment
1.06
1,6
0.34
0.14
(£, bipinnulata only)
date
4.89
3,18
0.01
0.84
date X treatment
2.15
3,18
0.13
0.45
Number of fishes
treatment
3.01
1,6
0.13
0.30
(all spp. -minus A. troschelii)
date
3.88
3,18
0.03
0.73
date X treatment
2.27
3,18
0.12
0.47
Species richness (S)
treatment
2.49
1,6
0.17
0.26
date
3.00
3,18
0.06
0.60
date X treatment
0.60
3,18
0.62
0.15
Species diversity (HB)
treatment
1.74
1,6
0.23
0.19
date
4.35
3,18
0.02
0.79
date X treatment
1.91
3,18
0.16
0.40
Table 5
Repeated measures ANOVA
results (control [no weights
and no artificial fish] vs.
weights and vs.
artificial fish, n
=^2=13=27) in
the artificial fish experiment.
Dependent variable
Factor(s)
F
df
P
1-/3
Number of fishes
treatment
0.13
2,6
0.88
0.06
(all species combined)
date
3.52
8,48
<0.01
0.97
date X treatment
0.37
8,48
0.98
0.20
Number offish
treatment
0.12
2,6
0.89
0.06
(A. troschelii only)
date
3.34
8,48
<0.01
0.95
date X treatment
0.41
8,48
0.97
0.22
Number of fishes
treatment
0.29
2,6
0.76
0.08
(all spp. minus A. troschelii)
date
4.82
8,48
<0.01
>0.99
date X treatment
0.73
8,48
0.75
0.41
Number offish
treatment
0.05
2,6
0.95
0.06
(E. bipinnulata only)
date
0.95
8,48
0.48
0.38
date X treatment
0.36
8,48
0.94
0.20
Species richness (S)
treatment
0.41
2,6
0.68
0.10
date
9.87
8,48
<0.01
>0.99
date X treatment
0.90
8,48
0.58
0.50
Species diversity (HB)'
treatment
0.13
2,5
0.88
0.06
date
4.44
8,47
<0.01
0.99
date X treatment
1.00
8,47
0.45
0.56
' Missing data were replaced according to the directions in Zar (1996), and the degrees of freedom were reduced accordingly.
results are consistent with those obtained by Wickham and
Russell (1974). A similar result would occur if these FADs
had a predictable carrying capacity and recruitment was suf-
ficiently rapid that removal of the assemblage was followed
by its replacement before the next observation. However,
assemblage sizes within treatments varied widely from one
day's observations to the next; therefore recruitment, rather
than carrying capacity, seems to determine assemblage size.
846
Fishery Bulletin 101(4)
f 1?-
E 10- -
Abudcfduf twfchelii
enriched i-ADs
Y =- 2.2X + s.y
non-enriched FADs
Y = 0.1X-0,1
0-$
00:00
-*-
-^
■V
01:00 02:00 03:00 04:00
Elapsed time (tih.mm)
Figure 9
Changes in the number of fish associated
with enriched (solid circles) and nonen-
riched (open circles) drifting FADs over time.
Enriched FADs showed significantly higher
rates of recruitment than did nonenriched
FADs; the slopes of the regression lines are
significantly different ((=3.05, 2-tailed test,
v=6,P=0.02).
There appears to be insufficient time or stability for such
factors as competition or predation to influence the size or
diversity of these FAD-associated fish assemblages. De-
spite these results, individual fish do remain with a specific
FAD for days: On at least five separate occasions associated
with some of the other experiments described in this paper,
individuals recognizable by scars and bite marks were
sighted repeatedly as many as six days after the initial
observation (Nelson, unpubl. data). Although the FADs and
the associated fishes described in the present study are not
directly comparable to FADs and fishes targeted in fisheries-
scale operations, these experiments are among the first con-
trolled efforts at understanding the effects of disturbance or
fishing for FAD-associated assemblages.
The average assemblage size for all experiments and
treatments varied considerably, often significantly, over
time (Tables 2-5). Significant interaction effects between
sample date and FAD treatments may be indicative of day-
to-day recruitment fluctuations, dependent upon recruit-
ment variability. A significant interaction may result when
these effects are large and are in evidence regardless of the
experimental treatment (i.e. occur in concert across treat-
ments). Significant sample date effects (and series effects
in the fish-removal experiment, Table 2) are likely a result
of temporal fluctuations in the numbers of fishes available
to recruit to the FADs.
Note that the two series in the fish-removal experiment
differed not only in which FADs were given a particular
treatment (positional effects), but also in time — the two
scries were necessarily run consecutively, not concurrently.
I believe, however, it to be unlikely that positional effects
influenced any of the results reported in the present study:
treatments were assigned to FADs within the arrays in
such a way as to ensure that inshore, offshore, or longshore
positions were equally weighted among treatments. Signifi-
cant series main effects, independent of additional factors,
were found only for species richness and diversity (HB) — a
result I attribute to changes in the availability of potential
recruit species. Temporal patterns of juvenile reef fish re-
cruitment are often variable and may be affected by such
factors as spawning periodicity (Love et al., 1990), variable
predation (Nelson, 2001), or changing physical oceano-
graphic processes (Doherty, 1991; Levin, 1994; Kingsford
and Finn, 1997). Rountree ( 1989), also, found that the mean
numbers of the most abundant species observed around a
FAD array off South Carolina varied widely during FAD
deployment, albeit over a much longer time period (nearly
200 days). Thus, differences in assemblage size and diver-
sity over time are not unexpected.
FAD size had significant, positive effects on assemblage
size and species richness. Although tripling the FAD size
resulted in a nearly threefold increase in the number of as-
sociated fishes (combined species), the response may not be
linear. (Note, however, that Rountree (1989) demonstrated
that the number of Decapterus punctatus associated with
midwater FADs exhibited a significant, positive linear re-
sponse to FAD size.) Further research will be necessary to
resolve the effect of FAD size on numbers of aggregating
fishes. Also of interest is the significant increase in species
richness attributable to increased FAD size. Bortone et al.
(1977) suggested that species diversity may be a function
of "clump size" for Sargassum-associated fish assemblag-
es, and Moser et al. (1998) found greater numbers offish
species under large ( 10-20 m diameter) mats of floating
Sargassum than they did under smaller clumps (<1 m
diameter) or in open water. However, the changes in spe-
cies richness from this experiment could well be an effect
of assemblage size; treatment effects on species diversity
measured using the Brillouin index (HB) were marginally
nonsignificant (Table 3, P=0.07). Significant sample date
differences in treatment and evenness are due to large
fluctuations in the abundance of the dominant species,
Abudefduf troschelii, ranging at the triple-size FADs from
1 to 55 individuals over the course of 11 days.
Fishes were five times more numerous on average at
fouled FADs than they were at comparable FADs lacking
fouling organisms, but measures of diversity showed no
significant treatment effect (Table 4). There was a signifi-
cant interaction between treatment and sample date for
the present experiment (Table 4) that may have been due
to fluctuations in assemblage sizes among sample dates
across both FAD treatments. The species composition of
these assemblages was similar to that of other experi-
ments, except that Elagatis bipinnulata were regularly
observed: Abudefduf troschelii were the dominant species
by abundance, followed by E. bipinnulata, and Mugil sp.
All were small, young-of-the-year fishes (the largest E. bi-
pinnulata individuals reached approximately 80 mm SL)
and seemed not to be feeding on the larger invertebrates
forming much of the colonizing community. During casual
observations of FAD-associated fishes, I observed fish feed-
ing on plankton carried past the FADs, but no physical
contact with the FAD or fouling organisms. Ibrahim et al.
Nelson: Fad characteristics and associated fish assemblages
847
(1996) reported that none of the gut contents from FAD-
associated fishes included sessile organisms found on their
FADs (fish size ranges included specimens 8-14, 15-99, and
>100 mm SL — the first two size categories are comparable
to the fishes in the present study). Larger, piscivorous fishes
do feed at least occassionally on smaller fishes associated
with floating objects (Gooding and Magnuson, 1967), but
published gut content studies are conflicting. Some sug-
gest that piscivorous species that associate with flotsam
rely on other sources of food (e.g. Gooding and Magnuson,
1967; Hunter and Mitchell, 1967; Brock, 1985), while others
suggest that flotsam- (or algae-) associated fishes form an
important food resource for these larger piscivorous fishes
(Dooley, 1972; Manooch et al., 1984; Coston-Clements et al,
1991). Morgan et al. (1985) noted the occurrence of at least
two members of the Sargosswrn -associated invertebrate
fauna among the stomach contents of several species of
pelagic fishes. From the perspective of flotsam- or FAD-as-
sociated fishes, opportunistic predation by piscivores that
do not associate with FADs may be more important than
predation by other members of the assemblage. Additional
gut content data from juvenile and nonpiscivorous fishes
are sorely lacking. I address possible explanations for the
results of the present study below.
I recorded no significant treatment effect attributable
to differences between FADs with artificial fish, FADs de-
ployed with artificial-fish-sized weights, or control FADs.
I attribute significant sample date effects to day-to-day
changes in constituent individuals and the fluctuating
availability of potential recruits. Numbers of £. bipin-
nulata were strikingly constant across treatments and
sample dates in this experiment (Table 5, Fig. 8) and in
the fouling experiment (Fig. 7) and seemed to indicate an
apparently unusual characteristic of this species — individ-
uals remaining associated with a given FAD for multiple
days. Although the experiment was intended to distinguish
between FADs with prior recruits versus FADs without
prior recruits, the lack of a significant treatment effect does
not negate the possibility that potential recruits would
distinguish between occupied and unoccupied FADs. The
painted artificial fish and lead weights clearly lacked many
attributes of living fish. However, comparable numbers of
recruits found at all treatments suggest that a change in
the structural complexity of the FADs did not affect assem-
blage size or diversity. Although the addition of four small
lead weights (artificial fish were painted and oriented
differently but were still lead weights) did not appear to
increase appreciably the visible surface area of those FADs,
the subsequent experiment with live fishes instead of ar-
tificial fish had a dramatic effect on recruitment; therefore
sizeable changes in the physical size of a FAD may be nec-
essary to yield a response in fish recruitment. The potential
roles of structural complexity and orientation of FADs will
be informative areas for future research. Past investiga-
tions in these areas (e.g. Hunter and Mitchell, 1968; Klima
and Wickham, 1971; Wickham et al., 1973) have provided a
useful beginning, but more work is needed.
Although sample sizes were small, the presence of prior
Abudefduf troschelii "recruits" (enriched FADs) had a sig-
nificant effect on patterns of subsequent recruitment; this
effect contrasted sharply with FADs lacking fish at the
start of this experiment (nonenriched FADs). For this spe-
cies, these results point to a social aspect to these aggrega-
tions, and sociality may also be involved in the recruitment
of other species, particularly the schooling fishes Caranx
spp., Polydactylus approximans, and Mugil spp., as sug-
gested for some scombrids (e.g. Dagorn and Freon, 1999).
The addition of fishes below a FAD may increase recruit-
ment rates by rendering the object more visible, although
the artificial fish experiment indicated that simply adding
fish-size objects beneath a FAD does not affect recruitment.
Comparisons between these two experiments are tenuous,
however, because the artificial fish experiment employed
anchored FADs observed over a period of days, whereas
the enriched FAD experiment used drifting FADs observed
over a course of hours.
Why do FAD size, the presence of a fouling community,
and the actual presence of prior recruits at a FAD each
have the effect of increasing the size and, possibly, the di-
versity of FAD-associated assemblages of juvenile fishes?
The simplest explanation is that these factors contribute
to the target strength of the object, increasing the visual,
olfactory, or auditory stimulus (or some combination) of
the floating object. Larger objects should be easier to find,
especially if potential recruits rely on vision to explore
their environment. Kellison and Sedberry (1998) found
that the fishes associated with mid-water floating struc-
tures that were tethered to an artificial reef decreased
in abundance over time (193 days), and suggested that
the loss in buoyancy associated with the development of
a fouling community may have reduced the effective size
of these floating objects, accounting for fewer associated
fishes (see also Hunter and Mitchell, 1968; Rountree, 1989).
To account for the positive effects of a fouling community
observed in the present study, it seems reasonable to sup-
pose that fouling organisms may be detected by olfactory
means; Sweatman ( 1988) has shown that some larval fishes
use olfactory cues for settlement on reefs. Further experi-
ments, for example experiments controlling for FAD size,
odor cues, and visibility of the FADs, are needed to deter-
mine why some of these factors exhibit these effects.
Future research on the role of flotsam as shelter from
predators and as a conveyance to suitable habitat could
yield evolutionary explanations for the attraction to float-
ing objects. For these small fishes, such objects likely repre-
sent a shelter from predators (Mitchell and Hunter, 1970).
Some species do respond to the approach of an observer by
positioning themselves so that the FAD is between them
and the observer. Particularly during daylight and crepus-
cular hours when visually-oriented predators are most
active, flotsam may offer refuge in a habitat where there
is little alternative refuge. During the day, when onshore
winds drive drifting objects towards shallow water, flotsam
and drift algae, unlike anchored FADs, may also offer a
comparatively safe conveyance to more suitable habitat.
Thus, there may be adaptive advantages for juvenile reef
fishes in associating with floating objects.
Although the juvenile fishes associated with the FADs
used in the present study are not of interest to any fishery,
the patterns observed from them may be relevant to FADs
848
Fishery Bulletin 101(4)
deployed commercially to aggregate fish species at various
life history stages. FAD size is clearly relevant to those in-
terested in studying potential improvements to FAD design.
Carefully controlled studies on the importance of surface
area versus volume and the orientation of FAD structures
are needed. The role of a fouling community, too, deserves
further investigations. Although a fouling community may
weigh down streamers (trailing pieces of buoyant material
intended to increase the subsurface of area of a FAD), such
a community may also improve recruitment and possibly
retention of recruits around a FAD. Finally, the importance
of the initial recruits to a floating object should be studied
further. Enriching a FAD may increase the speed at which
additional fishes are recruited. Improved artificial fish may
prove more effective than the items used in the present re-
search. FADs are an important tool in a number of artisanal
(small-scale fishery based on traditional methods), sport,
and commercial fisheries, especially in tropical waters
where FAD fisheries particularly target tunas (Scomb-
ridae), jacks (Carangidae), and Coryphaena spp. (Galea,
1961; Klima and Wickham, 1971; Beets, 1989; Hilborn
and Medley 1989; Friedlander et al., 1994; Higashi, 1994;
Hall et al., 1999b). Due largely to the potential for fisheries
enhancement, considerable research has been focused on
the importance of floating-object characteristics and the
numbers of fishes attracted to such objects; however, the
results have been difficult to interpret and are often con-
flicting (Rountree, 1989; Kingsford, 1993; Druce and Kings-
ford, 1995). Because log sets in tuna purse-seine fisheries
(where fishermen target fish associated with drifting logs or
FADs) are associated with high levels of bycatch (Hall, 1998;
Lennert-Cody and Hall, 2000), the behavior and ecology of
flotsam-associated species is in urgent need of study so that
a means of reducing bycatch may be devised.
This study made use of FADs floating at the surface;
studies by other researchers have employed similar tools
or they have used FADs tethered in mid-water. No one has
examined the effects of FAD position in relation to the sur-
face, and the implicit assumption appears to be that there is
no biologically significant difference. This assumption has
not been tested, although comparisons between data from
floating structures, whether at the surface, mid-water, or
tethered close to the bottom, are common in the literature.
I have made comparisons between my data from surface
FADs and results from mid-water FADs (e.g. Wickham and
Russell, 1974; Rountree, 1990); such comparisons may be
misleading and should be interpreted with caution.
The results from the present study indicate that turn-
over rates at nearshore anchored FADs are high and that
undisturbed FAD assemblages may show little difference
in these rates from disturbed FADs. Fishes recruiting to
these FADs discriminate among potential floating objects,
forming larger, more species-rich assemblages around tri-
ple-size FADs than around single FADs. FADs possessing
a fouling biota also attract larger (though no more diverse)
assemblages than do clean FADs. The latter effect was
complicated by temporal fluctuations that overlay these
treatment effects, resulting in day-to-day changes in the
total numbers of fishes in both treatments (Table 4, Fig. 7).
Further, the presence of prior recruits in the enrichment
experiment had a strong effect on subsequent recruitment.
Thus, the association of juvenile fishes with floating objects
is not a haphazard process, and floating-object character-
istics play potentially important roles in fish recruitment
to these objects. These results suggest that associating
with flotsam may be adaptive, rather than an accidental
behavior and support Kingsford's hypothesis (Kingsford,
1993 ) that floating material is an important environmental
component in the relationship between environment and
some juvenile fishes.
Acknowledgments
For help in the field, I am grateful to I. Nelson and D.
Mansue. W. L. Montgomery, S. Shuster, and to two anony-
mous reviewers who provided helpful criticism. Funding
was provided by the American Museum of Natural His-
tory (Lerner-Gray Fund), American Society of Ichthyolo-
gists and Herpetologists (Raney Award), Animal Behavior
Society, International Women's Fishing Association (Max
Coan Memorial Scholarship), Seaspace/Houston Underwa-
ter Club, Smithsonian Tropical Research Institute (STRI),
and Sigma Xi. D. R. Robertson of STRI and D. Margulise,
R. Olson, and V. Scholey of the Inter-American Tropical
Tuna Commission provided invaluable advice and logisti-
cal support.
Literature cited
Beets, J.
1989. Experimental evaluation of fish recruitment to com-
binations offish aggregating devices and benthic artificial
reefs. Bull. Mar. Sci. 44:973-983.
Bortone, S., P. A. Hastings, and S. B. Collard.
1977. The pelagic-Sargassum ichthyofauna of the eastern
GulfoflMexico. Northeast Gulf Sci. 1:60-67.
Brock, R. E.
1985. Preliminary study of the feeding habits of pelagic
fish around Hawaiian fish aggregation devices or can fish
aggregation devices enhance local fisheries productivity?
Bull. Mar Sci. 37:40-49.
Coston-Clements, L., L. R. Settle, D. E. Hoss, and F. A. Cross.
1991. Utilization of the Sar^assum habitat by marine inver-
tebrates and vertebrates — a review. NOAA Tech. Memo.
NMFS-SEFSC-296:32.
Crawford, R. E., and J. K. Jorgenson.
1993. Schooling behaviour of arctic cod, Boreogadiis saida,
in relation to drifting pack ice. Environ. Biol. Fish. 36:
345-357.
Dagorn, L., and P. Freon.
1999. Tropical tuna associated with floating objects: a simu-
lation study of the meeting point hypothesis. Can. J. Fish.
Aquat. Sci. 56:984-993.
Doherty, P. J.
1991. Spatial and temporal patterns in recruitment. InThe
ecology of fishes on coral reefs (P. F. Sale, ed.), p. 261-294.
Academic Press, San Diego, CA.
Dooley, J.
1972. Fishes associated with the pelagic Sargassum com-
plex, with a discussion of the Sargassum community. Con-
Irib Mar Sci. 16:1-32.
Nelson: Fad characteristics and associated fish assemblages
849
Druce, B. E., and M. J. Kingsford.
1995. An experimental investigation on the fishes associ-
ated with drifting objects in coastal waters of temperate
Australia. Bull. Mar. Sci. 57;378-392.
Fedoryako, B. I.
1989. A comparative characteristic of oceanic fish assem-
blages associated with floating debris. J. Ichthyol. 29:
128-137.
Friedlander, A., J. Beets, and W. Tobias.
1994. Effects offish aggregating device design and location
on the fishing success in the U.S. Virgin Islands. Bull. Mar.
Sci. 55:592-601.
Gaertmer, D., and M. Medina-Gaertner.
1999. An overview of the relationship between tunas and
floating objects in the south of Caribbean Sea. lATTC
(Inter-American Tropical Tuna Commission), Spec. Rep.
11:66-86.
Galea, J. A.
1961. The "Kannizzati" fishery. Gen. Fish. Counc. Mediterr.
Sess. Rep. 6:85-91.
Gooding, R. M., and J. J. Magnuson.
1967. Ecological significance of a drifting object to pelagic
fishes. Pac. Sci. 21:486-497.
Hall, M. A.
1998. An ecological view of the tuna-dolphin problem:
impacts and trade-offs. Rev. Fish Biol. Fish. 8:1-34.
Hall, M. A., D. L. Alverson, and K. I. Metuzals.
2000. By-catch: problems and solutions. Mar. Pollut. Bull.
41:204-219.
Hall, M. A., M. Garcia, C. Lennert-Cody, R Arenas, and F. Miller.
1999a. The association of tunas with floating objects and dol-
phins in the eastern Pacific Ocean: a review of the current
purse-seine fishery. lATTC, Spec. Rep. 11:87-194.
Hall, M. A., C. Lennert-Cody, M. Garcia, and P. Arenas.
1999b. Characteristics of floating objects and their attrac-
tiveness for tunas. lATTC, Spec. Rep. 11:396-446.
Hallier, J. P., and J. I. Parajua.
1999. Review of tuna fisheries on floating objects in the
Indian Ocean. lATTC, Spec. Rep. 11:195-221.
Hampton, J., and K. Bailey.
1999. Fishing for tunas associated with floating objects:
review of the western Pacific fishery. lATTC, Spec. Rep.
11:222-284.
Higashi, G. R.
1994. Ten years of fish aggregating device (FAD) design
development in Hawaii. Bull. Mar Sci. 55:651-666.
Hilborn, R., and P. Medley.
1989. Tuna purse-seine fishing with fish-aggregating
devices (FAD): models of tuna FAD interactions. Can. J.
Fish. Aquat. Sci. 46:28-32.
Hunter, J. R., and C. T Mitchell.
1967. Association of fishes with flotsam in the offshore
waters of Central America. Fish. Bull. 66:13-29.
1968. Field experiments on the attraction of pelagic fish to
floating objects. J. Cons. Int. Explor Mer 31:427-434.
Ibrahim, S., M. A. Ambak, L. Shamsudin, and M. Z. Samsudin.
1996. Importance offish aggregating devices (FADs) as sub-
strates for food organisms offish. Fish. Res. 27:265-273.
Kellison, G. T, and G. R. Sedberry
1998. The effects of artificial reef vertical profile and hole
diameter on fishes off South Carolina. Bull. Mar. Sci. 62:
763-780.
Kingsford, M. J.
1992. Drift algae and small fish in coastal waters of north-
eastern New Zealand. Mar. Ecol. Progr. Ser. 80:41-55.
1993. Biotic and abiotic structure in the pelagic environ-
ment: importance to small fishes. Bull. Mar. Sci. 53:393-
415.
1995. Drift algae: a contribution to near-shore habitat com-
plexity in the pelagic environment and an attractant for
fish. Mar Ecol. Progr Ser. 116:297-301.
Kingsford, M. J., and M. Finn.
1997. The influence of phase of the moon and physical pro-
cesses on the input of presettlement fishes to coral reefs. J.
Fish Biol. 51:176-205.
Klima, E. F, and D. A. Wickham.
1971. Attraction of coastal pelagic fishes with artificial
structures. Trans. Am. Fish. Soc. 1:86-99.
Kojima, S.
1960. Fishing for dolphins in the western part of the Japan
Sea. V. Species of fishes attracted to bamboo rafts. Bull.
Jpn. Soc. Sci. Fish. 26:379-382.
Lennert-Cody, C. E., and M. A. Hall.
2000. The development of the purse seine fishery on drift-
ing fish aggregating devices in the eastern Pacific Ocean:
1992-1998. In Peche thoniere et dispositifs de concentra-
tion de poissons: colloque DCP, Martinique, Octobre 1999,
p. 78-107. Institut Frangais de Recherche pour I'Exploita-
tion de la Mer, Issy-les-Moulineaux, France.
Levin, P S.
1994. Fine-scale temporal variation in recruitment of a tem-
perate demersal fish: the importance of settlement versus
post-settlement loss. Oecologia 97:124-133.
Love, M. S., M. H. Carr, and L. J. Haldorson.
1990. The ecology of substrate-associated juveniles of the
genus Sebastes. Environ. Biol. Fish. 30:225-243.
Magurran, A. E.
1988. Ecological diversity and its measurement, 179 p.
Princeton Univ. Press, Princeton, NJ.
Manooch, C. S., Ill, D. L. Mason, and R. S. Nelson.
1984. Food and gastrointestinal parasites of dolphin Cory-
phaena hippurus collected along the southeastern and Gulf
coasts of the United States. Bull. Jpn. Soc. Sci. Fish. 50:
1511-1525.
Massuti, E., S. Deudero, P. Sanchez, and B. Morales-Nin.
1998. Diet and feeding of dolphin (Coryphaena hippurus)
in western Mediterranean waters. Bull. Mar. Sci. 63:
329-341.
McCleneghan, K., and J. L. Houk.
1978. A diver-operated net for catching large numbers of
juvenile marine fishes. Calif. Fish Game 64:305-307.
Mitchell, C. T, and J. R. Hunter.
1970. Fishes associated with drifting kelp, Macrocystis pyr-
ifera, off the coast of southern California and northern Baja
California. Calif Fish Game 56:288-297.
Morgan, S. G., C. S. Manooch III, D. L. Mason, and J. W. Goy
1985. Pelagic fish predation on Cerataspis, a rare larval
genus of oceanic penaeoids. Bull. Mar. Sci. 36:249-259.
Moser, M. L., P. J. Auster, and J. B. Bichy.
1998. Effects of mat morphology on large Sargassum-assoc'i-
ated fishes: observations from a remotely operated vehicle
(ROV) and free-floating video camcorders. Environ. Biol.
Fish. 51:391-398.
Nelson, P. A.
1999. The ecology and behavior of flotsam-associated ma-
rine fish aggregations. Ph.D. diss., 137 p. Northern Ari-
zona Univ., Flagstaff, AZ.
2001. Behavioral ecology of young-of-the-year kelp rock-
fish, Sebastes atrovirens Jordan and Gilbert (Pisces: Scor-
paenidae). J. Exp. Mar. Biol. Ecol. 256:33-50.
850
Fishery Bulletin 101(4)
Parin, N. V., and B. I. Fedoryako.
1999. Pelagic fish communities around floating objects in the
open ocean. lATTC, Spec. Rep. 11:447-458.
Rountree, R. A.
1989. Association of fishes with fish aggregation devices:
effects of structure size on fish abundance. Bull. Mar. Sci.
44:960-972.
1990. Community structure of fishes attracted to shallow
water fish aggregation devices off South Carolina, U.S.A.
Environ. Biol. Fish. 29:241-262.
Safran, P.
1990. Drifting seaweed and associated ichthyofauna: fioat-
ing nursery in Tohoku waters. Mer 28:225-239.
Safran, P., and M. Omori.
1990. Some ecological observations on fishes associated
with drifting seaweed off Tohoku coast, Japan. Mar. Biol.
105:395-402.
Suzuki, Z.
1999. Distribution of floating logs in the Pacific and purse
seine sets on tunas associated with logs by Japanese boats
in the tropical western and central Pacific. lATTC, Spec.
Rep. 11:459-479.
Sweatman, H. P. A.
1988. Field evidence that settling coral reef fish larvae
detect resident fishes using disolved chemical cues. J.
Exp. Mar Biol. Ecol. 124:163-174.
Wickham, D. A., and G. M. Russell.
1974. An evaluation of mid-water artificial structures for
attracting coastal pelagic fishes. Fish. Bull. 72:181-191.
Wickham, D. A., J. W. Watson Jr., and L. H. Ogren.
1973. The efficacy of midwater artificial structures for attract-
ing pelagic sport fish. Trans. Am. Fish. Soc. 3:563-572.
Zar, J. H.
1996. Biostatistical analysis, 3rd ed., 662 p. Prentice Hall,
Upper Saddle River, NJ.
851
Abstract— The bastard grunt 'Pomada-
sys incisus) is one of the most abundant
coastal demersal fishes inhabiting
the Canary Islands. Age and growth
were studied from samples collected
between October 2000 and September
2001. Growth analysis revealed that
this species is a fast growing and
moderately short-lived species (ages
up to seven years recorded). Length-
at-age was described by the von Ber-
talanffy growth model (L_=309.58 mm;
A=0.220/year; ((,=-1.865 year), the
Schnute growth model (>'j= 126.66 mm;
.V2=293.50 mm; a=-0.426; 6= 5.963),
and the seasonalized von Bertalanffy
growth model (L„=309.93 mm; k=0:218/
year; <„= -1.896 year; C=0.555; (^=0.652).
Individuals grow quickly in their first
year, attaining approximately 60% of
their maximum length; after the first
year, their growth rate drops rapidly
as energy is probably diverted to repro-
duction. The parameters of the von
Bertalanffy weight growth curve were
W, =788.22 mm; *=0.1567/year; <„=
-1.984 year Fish total length and otolith
radius were closely correlated, r'^=Q.912.
A power relationship was estimated
between the total length and the oto-
lith radms (a=49.93; v'=0.851). A year's
growth was represented by an opaque
and hyaline (translucent) zone — an
annulus. Backcalculated lengths were
similar to those predicted by the growth
models. Growth parameters estimated
from the backcalculated sizes at age
were L_=315.23 mm; ft=0.217/year; and
(q= -1.73 year
Age and growth of the bastard grunt
(Pomadasys incisus: Haemulidae) inhabiting
the Canarian archipelago. Northwest Africa
Jose G. Pajuelo
Jose M. Lorenzo
Departament of Biology
University of Las Palmas de Gran Canaria
Campus Universitano de Tafira
35017 Las Palmas de Gran Canaria, Spam
E-mail address (for J G Pajuelo) |paguelo@dbio.ulpgc.es
Muriel Gregoire
Faculty of Biology
University of Liege
Liege, Belgium
Present address: Departament of Biology
University of Las Palmas de Gran Canaria
Campus Universitano de Tafira
35017 Las Palmas de Gran Canaria, Spain
Manuscript approved for publication
19 June 2003 by Scientific Editor
Manuscript received 26 June 2003
at NMFS Scientific Publications Office.
Fish. Bull. 101:851-859 (2003).
The family Haemulidae consists of
16 genera (126 species), including
the genus Pomadasys, This genus is
represented by 37 species distributed
around the world. Members of this
family are commonly referred to as
grunt (Bauchot and Huraeu, 1990).
Only the bastard grunt (Pomadasys
incisus (Bowdich, 1825)) is found off the
Canary Islands. The bastard grunt is a
coastal demersal fish species inhabiting
marine and brackish waters along the
eastern central Atlantic coasts from the
Strait of Gibraltar to Angola, and also
in the Canaries, Azores, and Cape Verde
Islands (Bauchot and Hureau, 1990). In
the Canary Islands, where the bastard
grunt is one of the most abundant spe-
cies, it has been observed in high densi-
ties in schools along coastal waters.
Information on the biology of the
bastard grunt is not available any-
where in the world. Despite its wide-
spread occurrence, the bastard grunt
has no commercial value for its low
quality meat and it is discarded in the
Canarian archipelago. The need for a
biologically based discard management
strategy and the paucity of data avail-
able on the biology of this species have
prompted an investigation into aspects
of its life history. We report aspects of
age and growth, which are important
parameters in models for managing
the population of the bastard grunt off
the Canary Islands.
Materials and methods
A total of 878 individuals of P. incisus
were collected at weekly intervals from
discarded commercial catches taken
between October 2000 and September
2001 off Gran Canaria (Canary Islands,
central-east Atlantic, 27°57'24"N-
15°35'23"W).
Each fish was measured to the nearest
mm for total length (L,) and weighed to
the near 0.1 g for total body weight (Wp.
Sex was assessed visually and sagittal
otoliths were removed, cleaned, and
stored dry for later age determination.
Age estimation was made by identify-
ing and counting annuli following Wil-
liams and Bedford (1974). An annulus
was defined as a hyaline zone formed
annually in the winter season when
there is low growth and an opaque zone
formed annually in the summer season
when there is increased growth. The
whole otoliths were placed in a black-
ened bottom watch glass containing
water and examined under a compound
microscope (lOx) with reflected light.
Counts of the growth bands were made
852
Fishery Bulletin 101(4)
Figure 1
Saggital otoliths of P. incisus collected off the Canary Islands: (A) from a two-year-old (182 mm); (B) from a
three-year-old (202 mm); and (C) from a five-year-old (265 mm) fish. Ri is the radius of the i"' band (distance
from the centre of the otolith to the outer margin of the annulus), Re is the radius of the otolith at capture, and
MI is the marginal increment measured.
by two readers without knowledge of the size and sex of the
specimens, or previous counts of the other reader. Counts
were made for otoliths of each individual on two separate
occasions, and only coincident readings were accepted. The
same approach was used to determine the final number
of bands in each specimen, and a consensus was reached
between readers on the final counts. Reproducibility of the
resultant age estimates was evaluated with the coefficient
of variation (CV) (Chang, 1982) and the index of average
percent error (lAPE) (Beamish and Fournier, 1981).
To validate that rings were formed annually in the
bastard grunt, we analyzed the monthly mean marginal
increment (Hyndes and Potter, 1997). Marginal increment,
estimated as the distance between the outer edge of the
outermost annual ring and the periphery of each otolith,
was measured (to the nearest 0.01 mm) with an ocular
micrometer (Fig. 1). Measurements were always made
along the longest axis of the otolith. The pattern expected
in the marginal increment would be a minimal value at the
start of the growth period, increasing with time until the
measurement fell to a minimum again at the formation of
the next period of growth (Pilling et al., 2000). The size of
the marginal increment varies both with the age of the fish
and the time of sampling during the year. Because older
fish grow slower than younger fish, a smaller marginal
increment is expected. For this reason, to assess the pos-
sibility of false annulus formation among either younger
or older bastard grunt, quantitative marginal increment
analyses should be standardized for age. Therefore, we
used age class to standardize our analyses. Owing to the
wide range of ages encountered, however, there were insuf-
ficient samples to fully accomplish this standardization.
It was necessary to combine the ages in two or more age
groups representing fast, moderate, and slow-growing in-
dividuals (Pilling et al., 2000). Mean marginal increments
were plotted against month of capture, and the minimum
was used to indicate the month of annulus formation.
Once the periodicity and timing of ring formation were
verified, the age of each fish was determined from the num-
ber of annuli, the assumed birthdate, and the sampling date.
It was assumed that annulus formation began 1 January,
corresponding to the peak of spawning in the species
(Gregoire'). The difference between the date of capture and
the birthdate was used to estimate a fractional age (Gordoa
and Moll, 1997). This fraction was added to the number of
annuli read in the otoliths to avoid any potential bias
in growth estimates due to differences in sampling date
(Gordoa and Moli, 1997).
Length-at-age was described by the three-parameter
specialized von Bertalanffy growth model (Ricker, 1973):
L, =LJl-e
-kH-l„)
),
the four-parameter Schnute growth model (Schnute,
1981):
Gregoire, M. 2001. Edad y crecimiento del roncador
Pomadasys incisus (Bowdich, 1825) de Gran Canaria (Islas
Canarias). Unpubl.data. ULPGC Socrates/Erasmus Research
Report, 34 p. Departament of Biology, University of Las Palmas
de Gran Canaria. 35017 Las Palmas de Gran Canaria. Spain.
Pajuelo et al.: Age and growth of Pomadasys incisus
853
Z., =
.y +(.v;'
and the seasonalized von Bertalanffy growth model
(Pitcher and Macdonald, 1973)
L, = L
\-e
-Al/-/„ )-{ — sin(2;r(/-/, ))
where Lc = the length-at-capture; and
V = a constant derived from the relationship of
total length to otolith radius (Francis, 1990).
The von Bertalanffy growth curve was fitted to the backcal-
culated length-at-age by means of Marquardt's algorithm
for nonlinear least squares parameter estimation (Gay-
anilo et al., 1996).
where A, = the smallest age in the sample;
A., = the largest age in the sample;
Tg = the age at zero length;
L, = the length-at-age;
L„ = the predicted asymptotic length;
yj = the estimated mean length of Aj-year-old
fish;
^2 = the estimated mean length of A2-year-old
fish;
C = the amplitude of the fluctuation in seasonal
growth;
t^ = the point of the minumum growth {t^=WP+
0.5); and
k = the Brody growth constant (Schnute, 1981;
Sparre and Venema, 1995).
The models were fitted to data with the Marquardt's algo-
rithm for nonlinear least squares parameter estimation
(Gayanilo et al., 1996). A nonparametric, one-sample
test was applied to test for residual randomness and a
Bartlett's test was used to test for their homoscedasticity.
Von Bertalanffy growth model parameters were also esti-
mated for observed W, as a function of age by substituting
weights in place of lengths in the growth equation and
incorporating b derived by the weight-length regression
(Sparre and Venema, 1995):
W, = WJ]-e
-l(i-r„l,*
where W, = the weight at age; and
W^= the predicted asymptotic weight.
Backcalculated size of each fish at the time of forma-
tion of each annulus was determined by a backcalculation
formula consistent with the body proportional hypothesis
(Campana, 1990; Francis, 1990). The measurements were
the following: the radius of the ('*' band (Ri, 0.01 mm, dis-
tance from the center of the otolith to the outer margin of
the annulus) and the radius of the otolith at capture (Re,
0.01 mm, distance from the center of the otolith to the pe-
riphery). These measurements were always made along the
longest axis of the otolith ( Fig. 1 ). The relationship between
the radius of the otolith at capture iRc) and the total length
was estimated as a power function (nonlinear relationship).
It is estimated by fitting the data by a regression of log(L,)
on log(i?c) consistent with the body proportional hypothesis
(BPH). The length of an individual when the i"' band was
laid down {Li, mm) was calculated as
Li = (Ri/Rc)'Lc,
Results
Of the 878 fish examined, 377 (42.9%) were males and 412
(46.9%) females. The remaining 89 (10.1%) individuals
were immature and could not be identified macroscopically.
Fish varied in size from 103 to 304 mm L,, and weighed
between 8.7 and 137.1 g W,. Males varied from 143 to 298
mm L,, and total mass was from 9.3 to 114.7 g W^. Female
total length varied between 134 and 304 mm and total
mass was from 13.2 to 137.1 g. Immature fish varied from
103 to 186 mm L, and from 8.7 to 29.6 g W,. No significant
differences were found between males and females in
mean size (Student's i-test, <=1.03<-f'o 05 2 1560^'
therefore, the specialized von Bertalanffy growth model
1 -
0.8"
--
0.6"
n
r~
0.4"
R
-p 1
rn T I S
0.2'
i r
1
p-|
-L
±
? A-
^
E 0
E 08"
c
£ 0.6-
o
—
o
1 0,4-
s
n:'
--
-L
-r-
-[- -r-
s
I
s^
0.8-
0.6-
0.4
T -r 15
T -r-
0.2-
^r*-
1
1
1
1 1 1
1 1
10 II 12 1 2 3 4 5 6 7 8 9
I 2000 I 2001 I
Month
Figure 2
Mean monthly marginal increment from otoliths with one and
two, three, and more than three annuli, representing fast, mod-
erate, and slow-growing individuals of P. incisus off the Canary
Islands. Standard errors are identified by the bars.
was chosen because it has fewer parameters making it sta-
tistically more robust, its parameters are commonly used in
mortality estimates and per recruit modelling, and because
it allows for comparison between growth studies conducted
on other species (Booth, 1997). No significant differences
were found between mean lengths-at-age of males and fe-
males with a Student's t-test {t=0.52 ;gQggjQ = 1.65). Backcalculatcd size at time of
annulus formation was used to provide length-at-age
data unbiased by differences in sampling date and
to estimate the von Bertalanffy equation (Table 3).
Backcalculated lengths were similar to those pre-
dicted by the growth models. Growth parameters
estimated from the backcalculated sizes at age were
L„=315.23 mm; A=0.217/year; and tg=-1.733 year. The
data was pooled as a single growth model because no
significant differences in the growth parameters were
found between males and females (HoteUing's T'-'-test,
7^=48.3>ro2Q.o5,3,784=7-89).
Discussion
Age estimation in fishes is complicated by the phe-
nomenon of "stacking" of growth zones towards the
otolith margin, particularly in older fish (Buxton and
Clarke, 1991). In many cases age determination is
difficult because whole otoliths are so thick that light
does not pass through (Buxton and Clarke 1991); how-
ever, in the bastard grunt off the Canary Islands the
translucency of the otoliths allows aging with relative
ease. The values of the lAPE and the CV suggested
that the precision levels obtained are according to the
reference point values indicated by Campana (2001).
The oldest age estimate obtained in the present study
was seven years and the phenomenon of stacking was
not evident.
The otoliths of the bastard grunt have a ring pat-
tern common to teleost fishes. Marginal increment
analysis demonstrated that one annulus, consisting of
one opaque zone and one hyaline zone, is formed an-
nually. These rings are believed to be deposited during
periods of fast and slow growth, respectively (Williams
and Bedford, 1974). Seasonal growth cycles might be
related to physiological changes produced by the influ-
ence of temperature, feeding regime, and reproductive
cycle (Morales-Nin and Ralston, 1990). The seasonal-
ized von Bertalanffy growth model reveals the reduc-
Pajuelo et al.: Age and growth of Pomodasys incisus
855
Table 1
Sample size per age group
off the Canary Islands, n
intervals.
(age-length key) and percentage (in parentheses) within each age group for all fish of P
s the number fish by age class and SD is the standar deviation. Total length classes are
ncisus collected
given in 10-mm
Size (mm)
Age groups (year)
0
I
II
III
IV
V
VI
VII
100
2(40.0)
3(60.0)
110
1(8.1)
10(90.9)
120
1(5.6)
17(94.4)
130
19(100.0)
140
15(71.4)
6(28.6)
150
4(17.4)
19(82.6)
160
33(100.0)
170
47(95.9)
2(4.1)
180
69(70.4)
29(29.6)
190
52 (46.6)
61(53.4)
200
25 (20.5)
96(78.6)
1 (0.9)
210
14(12.8)
89(81.6)
6(5.5)
220
5 (5.9)
57(67.0)
23(27.1)
230
14(27.0)
38(73.0)
240
2(10.0)
17(85.0)
1(5)
250
1 (6.6)
9(60.0)
4 (26.8)
1(6.6)
260
2(20.0)
6(60.0)
2(20.0)
270
1(25.0)
1(25.0)
2 (50.0)
280
1(33.3)
1(33.3)
1(33.3)
290
1 (100)
300
1(100.0)
Mean
107
125
176
212
236
261
275
295
n
4
68
270
351
97
13
7
2
SD
5.77
12.64
17.28
18.06
12.36
11.43
11.40
14.14
Table 2
Parameters estimates, standard errors, and 95% cofidence intervals for the
specialized von Bertalanffy, Sehnute, and seasonalized
von Bertalanffy growth models for all P. incisus collected off the Canary Islands. All models were
pooled without the
age-0 class.
Parameter
95% confidence intervals
Estimate
Standard error
Lower
Upper
Specialized von Bertalanffy growth model (r^=0.9l)
L_ (mm)
309.58
8.06
294.07
325.08
k (/year)
0.220
0.031
0.157
0.283
;,i(year)
-1.865
0.055
-2.007
-1.723
Sehnute growth model (r^=0.87)
V, (mm)
126.66
1.51
124.54
128.78
v., (mm)
293.50
9.50
286.68
300.32
A
-0.426
0.077
-0.578
-0.274
B
5.963
0.664
4.659
7.267
Seasonalized von Bertalanffy growth model (r'-=0.91)
L_ (mm)
309.93
7.68
295.21
324.65
if (/year)
0.218
0.032
0.153
0.282
;„(year)
-1.896
0.049
-2.067
-1.725
C
0.555
0.212
0.138
0.971
/,
0.652
0.061
0.532
0.773
856
Fishery Bulletin 101(4)
350 -
300
^
jm^^^^^jm •
• ^-^^^ ^J©"*"^ *
251) -
?
^ 200
c
OJ
— 150
TO
...s^^"^'
|5
^Btt*"* • ^'^^ Bertalanffy curve
ion
^ ZjSf H Seasonalized von Bertalanffy curve
-•-Schnule curve
-♦Observed data
50 -
0 . : 1 , , T
0 12 3 4 5 6 7 8
Age (year)
Figure 3
Von Bertalanffy specialized and seasonalized growth curves and the Schnute
growth curve for all individuals of P incisus collected off the Canary Islands.
4(H)
350 -
..
300
,• vr\ '
3 250
g 2110
m
,o
*- 150
l(K) -
* *** JjaUKBlwir*^ " H>788(l^-"«"*""')='"
V^MfWrnnZ^^^^*^
50 -
• • mW)^^^^ •' ''="*» 1
m-'^^iVt ' *
^"SUS^
" - 1 1 1 I
0 2 4 6 8
Age (year)
Figure 4
Von Bertalanffy growth curve derived from observed weight at age for all P. incisiis
collected off the Canary Islands.
tion of somatic growth and the formation of the hyahne
zone during the winter months. The high correlation found
between L, and otolith radius indicates that otoliths are a
useful structure for estimating the age and for indicating
the past growth history of bastard grunt.
The coefficients of determination for each fitted curve
show that the three models explain more than 88% of the
growth pattern. The use of the von Bertalanffy model to
describe growth has been criticized for several reasons
(Booth, 1997). These include the use of parameters that
Age and growth of Pomadasys incisus
857
320
280 -
« ^'^
• • ^^X*
240
•liilK^*^^**
mm)
iliililW^'^
£
.,.^^^''
al leng
O 1 "*()
F
J^'V* * * L,=49.93 R"*^'
**tl* • n=812
80 -
• • r=0.912
40 -
0 1 1 1 1 1 '
12 3 4 5 6 7 8
Otolith radius (mm)
Figure 5
The relation of otolith length to total length for P. incisus collected off the Canary
Islands.
Table 3
Backcalculated 1
ength-at-age for all P.
incisus collected off the Canary Islands.
Annu
us number
Age
Number
(year)
offish
I 11
Ill
rv
V
VI
VII
1
68
119
2
270
123 179
3
351
127 172
219
4
97
132 168
212
235
5
13
129 178
205
245
26
6
7
122 170
221
233
257
275
7
2
128 174
214
239
260
279
293
Mean
126 173
214
238
259
277
293
Number of backcalculated
lengths at age
808 740
470
119
22
9
2
Annual
growth
ncrement (mm)
126 47
41
24
21
18
16
Annual
growth
ncrement (%)
43.0 16.1
13.9
8.2
7.2
6.1
5.4
have little biological meaning (Schnute, 1981) and the
absence of parameters that take into account seasonal
changes in growth rate (Pauly, 1980; Moreau, 1987). Nev-
ertheless, the von Bertalanffy growth model has been used
extensively to describe the growth of grunts. The growth
model provides a simple description of growth which can
be compared between species and species groups (Booth,
1997). The special and the seasonal forms of the von Ber-
talanffy growth model were chosen for the present study
because they contain fewer parameters than the Schnute
growth model.
Backcalculated lengths-at-age are in close agreement
with the length estimated from otoliths readings. The
results obtained with the backcalculation method are very
satisfactory because they show the consistency in the in-
terpretation of the sequence of growth increments of the
bastard grunt off the Canary Islands and reduce the effect
of size-selective sampling bias on the length estimates for
youngest fish in the sample (Campana, 2001).
The growth parameters obtained are reasonable because
the theoretical maximal length value is higher than the
size of the largest fish sampled and the growth coefficient
858
Fishery Bulletin 101(4)
1 -
0.8 -
/ ©^^ \
„~
P an>\reus~\^
sfficlent (year
o
a-
1
v.y
\^*«\
1 0.4^
o
0.2 -
P sirtalus
P
\/-\ \ P opfnitluris
' ~---___^ jubenile ^— N. ^\ \
incistis ^^^/^^ \s, y^ ^
P argmleus ^--.^^^^^ /\# \
P fiaculalus p-SigTOmAieoo/mii
0 -
\.
1
1 1 ' ■ ^^
0 200
400 600 800 1000 1200
Asymptotic length (mm)
Figure 6
The relation of asymptotic
length to the growth coefficient for different species of the
Pomadasys genus.
value indicates a relatively fast attainment of maximum
size, characteristic of the moderately short life cycle for this
species. However, the estimations of ^g tend to be negative
and different from zero for values affected by the small
sample size of smaller fish. These estimations suggest that
the von Bertalanffy growth model does not accurately de-
scribe growth in the early stages. Pomadasys incisus grows
quickly in its first year, attaining approximately 60% of its
maximum length. After the first year, the annual growth
rate drops rapidly. This change in growth rate is attribut-
able to the utilization of available energy for reproduction
instead of somatic growth; in the study area the maturation
process begins in the second year of life (GregoireM.
Two different patterns of growth rate in relation to as-
ymptotic length are observed for Pomadasys species (Fig.
6). The pattern off! incisus is similar to that observed
for P. striatus, P. jubeline, P. kaakan, P. maculates, and P.
commersonnii — species characterized by a high or moder-
ate asymptotic length and low or moderate growth coef-
ficient (Latif and Shenouda, 1972; Wallace and Schleyer,
1979; Edwards at al., 1985; Iqbal, 1989; Al-Husaini et al.,
2001; Pauly''^). However, it differs substantially from that
observed for P. argyreus, a species with a very low asymp-
totic length (<151 mm) and a very high growth coefficient
(0.62-0.83/year), and for P. opercularis and P. argenteus,
species characterized by a high asymptotic length and a
high growth coefficient (550-741 mm; 0.28-0.52/years)
(Deshmukh, 1973; Nzioka, 1982; Brothers and Mathews,
1987; Majid and Imad, 1991; Ingles and Pauly^).
Results of models used in fisheries management, e.g.
analytical yield-per-recruit models (Beverton and Holt,
1957), are sensitive to uncertainty in the estimates of
input parameters such as the von Bertalanffy growth
parameters. Several estimations of growth in Pomadasys
species have been derived through length-based methods,
which for slow growing species are uncertain. The growth
parameters from this study are the first otoliths-based es-
timates of growth for P. incisus. Similar estimates obtained
from different growth models and methods suggest that the
current estimation could be considered a good estimation
of the growth pattern for the species and adequate for use
as an input parameter in models for the management of
the species.
Acknowledgments
The authors are grateful to the three anonymous review-
ers for their constructive, critical, and useful comments on
the manuscript.
^ Pauly, D. 1978 A preliminary compilation of fish lengtli
growth parameters. Ber Inst. Meereskunden, Universitat an
der Kiel 55, 200 p. Institut fur Meereskunde, Dusternbrooker
Weg 20, 24105 Kiel, Germany.
'' Ingles, J., and D. Pauly. 1984. An atlas of the growth, mortal-
ity and recruitment of Philippine fishes, ICLARM Technical
Repport 13, 127 p. World Fish Center (ICLARM) Jalan Batu
Maung, Batu Maung, 1 1960 Bayan Lepas, Penang, Malaysia.
Paiuelo et al,: Age and growth of Pomadasys Incisus
859
Literature cited
Al-Husaini, M., S. Al-Ayoub, and J. Dashti.
200 1 . Age validation of nagroor, Pomadasys kaakan ( Cuvier,
1830) (family: Haemulidae) in Kuwaiti waters. Fish. Res.
53:71-81.
Bauchot, M. L., and J. C. Hureau.
1990. Haemulidae. //; Check-list of the fishes of the eastern
tropical Atlantic (J. C. Quero, J. C. Hureau, C. Karrer, A.
Post, and L. Saldanha, eds. ), p. 786-787. UNESCO, Paris,
France.
Beamish, R. J., and D. A. Fournier.
1981. A method for comparing the precision of a set of age
determinations. Can. J. Fish Aquat. Sci. 38:982-983.
Beverton, R. J. H., and S. J. Holt.
1957. On the dynamics of exploited fish populations. Fish-
eries Investigations Series II, XIX, 533 p. Her Majesty's
Stationery Office, London, UK.
Booth, A. J.
1997. On the life history of the lesser gurnard (Scorpaeni-
formes: Triglidae) inhabiting the Agulhas Bank, South
Africa. J. Fish. Biol. 51:1155-1173.
Brothers, E. B., and C. P. Mathews.
1987. Application of otolith microstructural studies to age
determinations of some commercially valuable fish of the
Arabia Gulf Kuwait Bull. Mar Sci. 9:127-157.
Buxton, C. D., and J. R. Clarke.
1991. The biology of the white musselcracker Sparadon
durbanensis (Pisces: Sparidae) on the Eastern Cape Coast,
South Africa. S. Afr J. Mar Sci. 10:285-296.
Campana, S. E.
1990. How reliable are growth back-calculations based on
otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227.
2001. Accuracy, precision and quality control in age deter-
mination, including a review of the use and abuse of age
validation methods. J. Fish. Biol. 59:197-242.
Chang, W. B.
1982. A statistical method for evaluating the reproduc-
ibility of age determinations. Can. J. Fish. Aquat. Sci. 39:
1208-1210.
Deshmukh,V. M.
1973. Fishery and biology of Pomadasys hasta (Bloch).
Indian J. Fish. 20:497-522.
Edwards, R. R. C, A. Bakhader, and S. Shaher
1985. Growth, mortality, age composition and fishery yields
offish from the Gulf of Aden. J. Fish. Biol. 27:13-21.
Francis, R. I. C. C.
1990. Back-calculation of fish length: a critical review. J.
Fish Biol. 36:883-902.
Gayanilo, F. C, R Sparre, and D. Pauly.
1996. FAO-ICLARM stock assessment tools (FiSAT) user's
guide, 126 p. FAO (Food Agric. Organ. U. N.) Comp. Inf
Ser (Fish.)8.
Gordoa, A., and B. Moll.
1997. Age and growth of the sparids Diplodus vulgaris, D.
sargus and D. annularis in adult populations and the dif-
ferences in their juvenile growth patterns in the north-west
Mediterranean Sea. Fish. Res. 33:123-129.
Hyndes, G. A, and I. C. Potter
1997. Age, growth and reproduction ofSillago schomhurgkii
in south-western Australin, nearshore waters and compari-
sions of life history styles of a suite ofSillago species. En-
viron. Biol. Fish. 49 435-447.
Iqbal, M.
1989. A note on the population dynamics of Pomadasys
kaakan (Haemulidae) from Pakistan. Fishbyte 7:4-5.
Latif, A., and S. Shenouda.
1972. Biological studies on Rhonciscus striatus (family
Pomadasidae) from the Gulf of Suez. Bull. Inst. Ocean.
Fish. 2:103-134.
Majid, A., and A. Imad
1991. Growth of Pomadasys kaakan (Haemulidae) off the
coast of Pakistan. Fishbyte 9:30-33.
Morales-Nin, B., and S. Ralston.
1990. Age and growth of Lutjanus kasmira (Forskal) [sic]
in Hawaiian waters. J. Fish Biol. 36:191-203.
Moreau, J.
1987. Mathematical and biological expression of growth in
fishes: recent trends and further developments. In Age
and growth offish (R. C. Summerfelt and G. E. Hall, eds.),
p. 81-113. Iowa State Univ. Press, Ames, lA.
Nzioka, R. M.
1982. Biology of the small spotted grunt Pomadasys opercu-
laris (Playfair 1866) (Pisces: Pomadasyidae) around Malindi
in Kenya. Kenya J. Sci. Tech. 3:69-81.
Pilling, G. M., R. S. Millner, M. W. Easey, C. C. Mees,
S. Rathachasen, and R. Azemia.
2000. Validation of annual growth increments in the otoliths
of the lethnnid Lethrinus mahsena and the lutjanid Apnon
virescens from sites in the tropical Indian Ocean, with notes
on the nature of growth increments in Pristipomoides
filamentosus. Fish. Bull. 98:600-611.
Pauly, D.
1980. On the interrelationships between natural mortality,
growth parameters, and mean enviromental temperature in
175 fish stocks. J. Cons. Int. Explor Mer 39:175-195.
Pitcher, T J., and R D. M. Macdonald.
1973. Two models for seasonal growth in fishes. J. Appl.
Ecol. 10:597-606.
Ricker, W. E.
1973. Linear regressions in fishery research. J. Fish. Res.
Board Can. 30:409-434.
Schnute, J.
1981. A versalite growth model with statistically stable
parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140.
Sparre, P., and S. C. Venema.
1995. Introduccion a la evaluacion de recursos pesqueros
tropicales, parte 1, manual, 420 p. FAO (Food Agric.
Organ. U. N.) Doc. Tec. Pesca 306(1).
Wallace, J. H., and M. H. Schleyer
1979. Age determination in two important species of South
Africa angling fishes, the knob {Argyrosomus hololepidotus
Lacep) and the spotted grunter (Pomadasys commersonnii
Lacep). Trans. R. Soc. S. Afr. 44:15-26.
Williams, T., and B. C. Bedford.
1974. The use of otoliths for age determination. In The
ageing of fish (T. B. Bagenal, ed.), p. 114-123. Unwin
Brothers, Old Woking, Surrey, UK.
860
Abstract— Management of West Coast
groundfish resources by the Pacific
Fishery Management Council involves
Federal government and academic
scientists conducting stock assess-
ments, generally using the stock syn-
thesis framework, applying the 40-10
rule to determine harvest guidelines
for resources that are not overfished
and conducting rebuilding analyses
to determine harvest guidelines for
resources that have been designated
as overfished. However, this manage-
ment system has not been evaluated
in terms of its ability to satisfy the
National Standard 1 goals of the Sus-
tainable Fisheries Act. A Monte Carlo
simulation framework is therefore
outlined that can be used to make such
evaluations. Based on simulations tai-
lored to a situation similar to that of
managing the widow rockfish {Sebastes
entomelas) resource, it is shown that
catches during recovery and thereafter
are likely to be highly variable (up to
±30'?f from one year to the next). Such
variability is far greater than has been
presented to the decision makers to
date. Reductions in interannual vari-
ability in catches through additional
data collection are, however, unlikely.
Rather, improved performance will
probably arise from better methods for
predicting future recruitment. Rebuild-
ing analyses include quantities such as
the year to which the desired probabil-
ity of recovery applies. The estimates
of such quantities are, however, very
poorly determined.
Evaluating the efficacy of managing West Coast
groundfish resources through simulations
Andre E. Punt
School of Aquatic and Fishery Sciences
University of Washington
1122 NE Boat Street
Seattle, Washington 98195-5020
E-mail address aepunt@u. Washington edu
Manuscript approved for publication
24 April 2003 by Scientific Editor
Manuscript received 26 June 2003
at NMFS Scientific Publications Office.
Fish. Bull. 101:860-873 (2003).
National Standard 1 of the Sustainable
Fisheries Act (SFA) of 1996 states that
"Conservation and management mea-
sures shall prevent overfishing while
achieving, on a continuing basis, the
optimum yield from each fishery for
the United States industry." The need
to satisfy this National Standard has
led inter alia to the requirement for the
eight Regional Fishery Management
Councils to develop control rules that
are used to assess whether overfishing
is occurring^ or a stock is in an over-
fished state (e.g. Restrepo and Powers,
1999). In addition, the SFA specifies
that a rebuilding plan has to be devel-
oped for any fish stocks that are des-
ignated as overfished. This plan needs
to include the time period by which the
stock will be rebuilt to fi^gY (the aver-
age biomass associated with maximum
sustainable yield, MSY), and the strat-
egy by which the stock is to be rebuilt.
The Pacific Fishery Management
Council (PFMC) has adopted the "40-
10" rule to manage groundfish stocks
that are not designated as being over-
fished. This rule determines the harvest
guideline for each groundfish stock by
computing the catch corresponding to
an i^MSY proxy (F^q,,/ for flatfish, FgQ.^
for rockfish in the Sebastes complex, and
Fjjr; for other species ) and reducing it if
the spawning output is estimated to be
less than 40% of the estimated Bq. This
reduction in catch is linear with spawn-
ing output, being 0 at 0.4Sq and lOO'J
at O.IBq. For stocks that are designated
as being in an overfished state (defined
for West Coast groundfish as being
that the spawning output is less than
0.25Bq) a rebuilding plan is developed. ■*
The main features of the technical as-
pects of a rebuilding plan (referred to as
a rebuilding analysis) identified by the
Scientific and Statistical Committee of
the PFMC are outlined in Appendix 1.
In brief, the rebuilding analysis used by
the PFMC involves projecting the best
estimates of the current age-structure
of the overfished population forward
under a range of alternative fishing
mortality rates and selecting the fish-
ing mortality rate that has a Council-
selected probability that the population
recovers to the proxy for Bygv of 0.4Sg
within a time frame consistent with the
specifications of the SFA.
Detailed stock assessments are avail-
able for only a small subset of the 81
species included in the PFMC Ground-
fish Management Plan. Of these species,
nine (bocaccio [Sebastes paucispinis],
canary rockfish [Sebastes pinniger[,
cowcod [Sebastes levis], darkblotched
rockfish [Sebastes crameri], lingcod
[Ophiodon elongates], Pacific ocean
perch [Sebastes alutus]. Pacific whiting
[Merluccius productus], widow rockfish
[Sebastes entomelas], and yelloweye
rockfish [Sebastes ruberhmiis]) have
been designated overfished and rebuild-
ing plans have been or are being devel-
oped for them. The direct consequences
' In the present study, and consistent with
usage by the Pacific Fishery Management
Council, "overfishing" means that the level
of fishing mortality exceeds that associ-
ated with MSY and "being in an overfished
state" means that the current spawning
output is less than 25'7c of the pre-exploi-
tation equilibrium spawning output. Sq
(spawning output is the product of egg
production-at-age and numbers-at-age).
2 F^,^ is the fishing mortality rate at which
the spawning output-per-recruit is reduced
to x% of its unfished level.
3 One implication of this is that the 40-10
rule is not actually used if the stock is
assessed to be below 0.25B„.
Punt: Managing West Coast groundfish resources through simulations
861
for industry of the implementation of a rebuilding plan
can be substantial (e.g. a reduction in the catch of canary
rockfish from 883 metric tons (t) in 1999 to only 90 t in
2001), although there are also indirect consequences in the
form of reductions in the harvest of nonoverfished species
to prevent overharvesting of overfished species through
technical interactions.
The performance of the method commonly used for as-
sessments of West Coast species has been evaluated to
some extent (e.g. Sampson and Yin, 1998; lanelli, 2002).
However, the performance of this assessment method in
combination with the rules used to determine harvest
guidelines has not been evaluated.
Management procedures'* are combinations of stock as-
sessment methods and catch control laws that have been
evaluated by means of Monte Carlo simulation to assess
the extent to which they are able to satisfy the manage-
ment objectives for a fishery. Evaluation of management
procedures by means of Monte Carlo simulation has been
argued to be essential because "if a management procedure
is unable to perform adequately in the ideal world repre-
sented on a computer, what basis is there to assume that
it will perform adequately in the real world?" (Sainsbury^).
One caveat to this argument is that it is only possible to
evaluate a management procedure if it is fully specified and
if it will be followed for several years in reality.
Management procedures have been adopted by the In-
ternational Whaling Commission for managing commer-
cial and aboriginal whaling (e.g. IWC, 1992, 2001) and by
southern African nations for managing a variety of pelagic
and demersal resources (Butterworth and Bergh, 1993; Co-
chrane et al., 1998; Geromont et al., 1999). Management
procedures are under consideration in Australia (Punt et
al., 2001) and New Zealand (Starr et al., 1997). If it can
be assumed that the same rules will be applied to modify
rebuilding plans each time new information on abundance
and year-class strength becomes available, it is possible to
consider the combination of the assessment method, the
default 40-10 rule, and rebuilding plans as a "manage-
ment procedure" and evaluate it by means of Monte Carlo
simulation. This study therefore involves determining from
past practice the "management procedure" being applied
by the PFMC. However, this "management procedure" has
not been formally adopted in any way and the approach to
managing West Coast groundfish could change in time.
This paper first outlines a simulation framework (a
management procedure evaluation, MPE, framework)
within which the expected performance of the approach
used by the PFMC to determine harvest guidelines can
be evaluated. It then evaluates variants of this approach
for scenarios similar to that of managing the fishery for
widow rockfish.
'' Also referred to as "harvest strategies" ( Punt et al ., 200 1 ), "man-
agement decision rules" (Starr et al., 1997 ), "fisheries control sys-
tems" ( Hilborn, 1979 ), and "operational management procedures"
(Barnes, 1999).
^ Sainsbury, K. G. 2001. Personal commun. CSIRO Marine
Research, Castray Esplanade, Hobart, TAS 7000, Australia.
Materials and methods
The steps in evaluating management procedures are as
follows:
1 Identification of the management objectives and rep-
resentation of these by using a set of quantitative
performance statistics.
2 Identification of a range of alternative management
procedures.
3 Development and parameterization of a set of alterna-
tive structural models (called operating models) of the
system.
4 Simulation of the future use of each management
procedure to manage the system (as represented by
each operating model). For each year of the projection
period, the simulations involve the following steps;
a Generation of the data available for assessment
purposes.
b Application of a method of stock assessment to
the generated data to determine key assessment-
related quantities (e.g. current age-structure,
spawning output in relation to target and limit
levels, historical trends in recruitment) and any
inputs to the catch control law.
c Application of the catch control law element of the
management procedure to determine a harvest
guideline.
d Determination of the biological implications of
this harvest guideline by setting the catch for
the "true" population represented in the operat-
ing model based on it. The step can potentially
include "implementation uncertainty" (Rosenberg
and Brault, 1993).
The harvest guideline is not updated every year in the
simulations described in this article, but rather every
third year (co-incident with the results from each new
survey) and thus reflects the intended frequency with
which assessments for West Coast groundfish species are
conducted. Each simulation trial (i.e. each combination of
an operating model variant and candidate management
procedure) involves 100 simulations of an 80-year manage-
ment period. The four steps listed above are discussed in
detail below.
Note that for the application considered in this paper
then, there are three "models": 1) the operating model that
represents "reality" for the simulations, 2) an assessment
model (a stock synthesis-like approach), and 3) a model to
calculate the harvest guidelines. The data available to the
last two models are generated from the first model.
The operating model
The operating model has been taken to be virtually iden-
tical to that on which the population assessments and
rebuilding analysis calculations are based (Appendix 1),
with two exceptions: 1) the approach used to generate
recruitment and 2) the allowance for variability over time
in commercial selectivity. Commercial selectivity is given
862
Fishery Bulletin 101(4)
o _
8
O -1
B
CM
\
Selectivity
3 •"
CO
>
3
O
O
o ■
in
\ AA
elative
1.0
^ \
/\^
o
^/-'"^^ '^
S.
s
J
J 0
o
'
C
) 5 10 15 20
0 10 20 30
40
Age (yr)
Year
Figure 1
Biological parameters (A) and catch history (B) for
widow rockfish iSebastes entomelas).
by the following double-logistic equation:
5,, „ = S[ ^, / max 5' ,,.
I 1
(1)
5,' =-
\ + e
■6f(u-iiy^+y, I
l + e
-Slti^-a)
where S^^ = the selectivity on fish of age a during
yeary;
a^Q, oIq, 5j, ^2 = the parameters of the double-logistic
equation;
7 = the deviation from the average selectiv-
ity pattern in yeary:
Yy = Ps rv-l + ^v
el ~ N(0;a;),
pg = the interannual correlation in the de-
viation from average selectivity; and
Gg = a measure of the standard deviation of
the interannual deviations from aver-
age selectivity.
Recruitment is assumed to be governed by a Beverton-Holt
stock-recruitment relationship:
' 4/i + (5/i-l)(B, /B„-l)
E^~N{0-ol). (2)
where Rq = the "virgin recruitment" (the number of zero-
year-olds at the pre-exploitation equilibrium
level);
By - the spawning output at the start of year y;
h = the"steepness"of the stock-recruitment rela-
tionship (the fraction of virgin recruitment
expected at 0.2B„); and
a^ - the standard deviation of the logarithms of
the random fluctuations in recruitment about
its expected value.
The biological parameters of the operating model are set
to those for widow rockfish (Fig. lA), and the catches for
Table 1
The baseline parameters of the operating
values used in the tests of sensitivity. N/A
model and the
= not available.
Parameter
Baseline value
Sensitivity values
Ps
0.707
N/A
^s
0.4
N/A
h
0.4
0.25; 0.7
cffl
0.6
0.4; 1
M
Spawning output
in year 41
0.15/yr
0.2S„
N/A
0.1Bo;0.4Bo
the 40 years prior to the year in which the management
procedure is first applied (referred to as "projection year 1")
are set to the actual catches for widow rockfish (Fig. IB).
The baseline values for the parameters /;, (T^, pg, and Og
(Table 1) are educated guesses. The baseline choice for
steepness, /;, is lower than the posterior mean for this
quantity (0.65) obtained by Dorn (2002) because, increas-
ingly. West Coast rockfish are being found to be less pro-
ductive than initially anticipated (e.g. lanelli, 2002). The
value assumed for the extent of variation in recruitment,
a^, although based on the collection of estimates of this
parameter by Beddington and Cooke (1983), is neverthe-
less also largely an educated guess. Sensitivity to the
values for both h and a^j is explored.
The biomass at the start of year 1 is assumed equal to
Bq, which is defined as the mean of the distribution for
the unfished biomass which would arise given variability
in recruitment about its expected value. However, this
specification has little impact on the results. For example,
the alternative that is defined to be the median of the
distribution for the unfished biomass would only change
B„ by about 5%.
The value for Bg for each simulation is selected so that
the spawning output at start of year 41 (projection year 1)
equals a prespecified fraction of Bq (baseline fraction
Punt: Managing West Coast groundfish resources through simulations
863
Table 2
The parameters on which the generation of future data is based
n'' is the sample size for the
multinom
al distribution.
Data source First year collected
Frequency
Precision
Catch rates 14
Every year
a'= 0.4
Fishery age-composition 21
Every year
^"=200
Survey indices 13
Every third year
CT''=0.5
Survey age-composition 13
Every third year
n'=200
0.2 — i.e. just below the level that defines an overfished
stock). Sensitivity to alternative values for the ratio of the
spawning output at the start of year 41 to Bq is explored
(Table 1).
Generating future data
The data available for assessment purposes are survey
indices of relative abundance, age-composition data from
surveys, catch-rate-based indices of relative abundance,
and age-composition data from the commercial catches.
Table 2 lists the baseline specifications regarding the fre-
quency at which the various data sources are collected and
the parameters that determine the sampling variability
associated with each data source.
The survey and catch-rate indices are generated by using
the equations
B':">" = B'/'-"'''''\
/= By' -'"■'■ '\
ef ~/v(0;(CT'r); (3a)
f;:~/v(0;(cr')-); (3b)
where £*■"''''' = the survey index for year j;
B^ = the survey selected-biomass during yeary:
fi: = £H-„5;,/V„,„.-''"";
(4a)
w^ = the mass of an animal of age a;
Sj, = the selectivity of the survey gear on animals of age
a (assumed to be governed by a logistic function
and to be independent of time);
^\a ~ ^^^ number of animals of age a at the start of year
.v;
Z,,^ = the total mortality on animals of age a during year
y\
cTj = the standard deviation of the random fluctuations
in survey catchability;
"max - ^^^^ oldest age considered in the operating model;
/^, = the catch-rate index for year y;
B^ = is the exploitable biomass during yeary;
fi:=X»'.%^^..,(" -<-"'■■');
(4b)
CT'' = the standard deviation of the random fluctuations in
fishery catchability.
Note that Equations 3a and 3b assume that the survey and
fishery catchability coefficients are unity. This assumption
can be made without loss of generality because the stock
assessment method is not provided with this information
and instead estimates these catchability coefficients. Note
also that the key difference between the survey index
and the catch-rate index is that selectivity for the latter
changes over time (see Eq. 1), whereas selectivity for the
former is time-invariant.
The age-composition data are generated by selecting
a sample multinomially from the age-composition of the
survey catch and of the fishery catch (see Eqs. 5a and 5b
for the relative survey and fishery catches-at-age):
Sln^.
•7^«v.a(l-^ )•
(5a)
(5b)
F^ = the fully selected fishing mortality during year y;
and
The PFMC management procedure
The "PFMC management procedure" (see Fig. 2 for an
overview) involves first conducting a stock assessment
by fitting an age-structured population dynamics model
to the available data by maximizing a likelihood func-
tion. This approach mimics the common use of the stock
synthesis framework (Methot, 2000) when conducting
assessments of West Coast groundfish resources. The
likelihood function is determined by assuming that the
age-composition data are multinomially distributed (in
the simulations with effective sample sizes given by the
actual effective sample sizes) and by assuming that the
survey and catch-rate series are log-normally distributed
about the appropriate model quantities. The estimable
parameters of the model are the annual recruitments,
the annual fishing mortalities, the catchability coef-
ficients, and the parameters that determine selectivity
(the survey and fishery selectivity are [correctly] assumed
to be governed by logistic and double-logistic equations).
The values for the remaining parameters (weight-at-age,
fecundity-at-age, and natural mortality) are assumed to
be known without error. The key outputs from the assess-
ment are time-series of recruitments and spawning out-
864
Fishery Bulletin 101(4)
i
Conduct a Stock Assessment
i
Select an approach for
generating future recruitment
(recruits or recruits / spawner)
No Currently under a Yes
Rebuilding Plan''
Is stock assessed to be
currently overfished''
Apply the 40-10 rule
Yes
♦ Declare stock overfished
Has stock recovered to
above the t.arget level''
Apply the rebuilding software
(Do not update the target rebuild
year unless the stock has just
been declared overfished)
^ Apply the constraints on ^ i
the extent of inter-annual variation
in harvest guidelines
Constrain the harvest guideline
to fall within the maximum and minimum limits
Figure 2
Flowchart of the Pacific Fishery Management Council management procedure.
puts, and the age structure at the start of the last year of
the assessment.
An estimate of the pre-exploitation equilibrium spawn-
ing output (i.e. Bq) is obtained by multiplying the average
recruitment for the first ten years of the assessment period
by the spa-wning output-per-recruit in the absence of fish-
ing. This approach to estimating Sg has been used for sever-
al rebuilding analyses for West Coast groundfish species. If
the estimate of the current spawning output exceeds 0.4Sg
or if it exceeds 0.25 fig and the resource is not currently
under rebuilding (i.e. has not yet been declared to be in an
overfished state), a raw harvest guideline is computed us-
ing the 40-10 rule. On the other hand, if the estimate of the
current spawning output is less than 0.25 fig or the stock is
currently under a rebuilding plan and the spawning output
has not yet recovered to 0.4 fig, the raw harvest guideline
is based on the application of the rebuilding analysis (see
Appendix 1 for further details).
It is necessary to know the maximum possible rebuilding
period, T^^^, when using a rebuilding analysis to calculate
a harvest guideline. If the stock is declared overfished in
the present year, T^^^ is computed as described in Appen-
dix 1. On the other hand, if the stock is currently under a
rebuilding plan, T^^^ is taken to be the value computed
when the stock was first declared overfished. Therefore,
the implementations of the rebuilding plans considered
in this paper are based on the assumption that the 7",,,,^,^
and the probability of recovery by T^^g^ are set when the
first rebuilding analysis is conducted and not changed
thereafter. The probability of recovery by T„^^^ is taken
to be 0.6 in this paper because this is the probability on
which management of widow rockfish is currently based.
This probability ranges between 0.55 and 0.92 among the
seven overfished groundfish resources for which it has been
selected by the PFMC.
Calculation of a harvest guideline using the 40-10 rule
and application of the rebuilding analysis requires the
ability to generate future recruitment. For the purposes of
the present study (and consistent with current practice),
future recruitment is either generated from the estimates
of recruitment from the assessment or by multiplying the
spawning output by a generated value for the recruits-per-
spawning output ratio. The pool of recruitment to recruits-
per-spawning output is taken to be those for the last 23
years of the assessment period less those for the last three
years. The last three years are excluded because of their
known poor precision. The approach used to generate re-
cruitment therefore leads to the set of recruitments used
to conduct projections changing with time. Allowing the set
of recruitments to change with time is needed to avoid an
inconsistency between the recruitments used for projections
and the recruitments on which the estimate of Bg is based.
Allowance is made for the raw harvest guideline to be
constrained so as not to change by more than a prespecified
percentage from that for the previous year and not to fall
outside of specified limits, although this option is not part
of the baseline simulations.
One aspect of the actual management process that is
ignored in the simulation of the PFMC management pro-
cedure is the time-lag between the collection of data and
their use in assessments (for example, catch-at-age infor-
mation from surveys conducted in one year would usually
not be available for use in the assessments conducted
in the following year) and that between assessments
Punt: Managing West Coast groundfish resources through simulations
865
Table 3
The performance statistics used in the present study. For consistency with the definition of recovery used by the Pacific Fishery
Management Council, "recovery by year x" is defined as the spawning output being larger than OABq at or before year x. Some of
the statistics are based on the "actual" (i.e. operating model) spawning output and others are based on the "assessed" (i.e. assess-
ment model) spawning output.
Abbreviation
rec
°decl
5%D/50%D
AAV
Description
The fraction of the simulations in which the stock is assessed to be overfished at the start of the first
projection year that actually recover by the maximum possible recovery year determined from the
rebuilding analysis conducted in projection year 1.
The median year in which the actual spawning output first reaches QABg.
The proportion of simulations in which the spawning output is assessed to be below 0.25Bq (i.e. overfished)
at the start of projection year 1.
The lower 5th and median of the distribution of the actual spawning output in projection years 20 and 60
expressed in relation to the actual pre-exploitation spawning output, Bq.
Average annual absolute change in catch evaluated after 20 and 60 years, i.e.
where C^ is the catch during yeary.
Average annual catch over projection years 1-20 and 1-60.
The fraction of simulations in which actual spawning output reached 0.4Bo sometime between projections
years 1 and 20 and between projection years 1 and 60 (but may have dropped below 0.4Sq again).
being conducted and their being used for management
purposes.
The performance statistics
A variety of performance statistics are considered (Table 3).
These consider both the performance of the management
procedure in terms of the behavior of the rules used for
management (statistics F^^^, Y^^^, and P^g^^) and of satis-
fying the goals established by the SFA in relation to the
status of the population and the fishery (statistics 5%D,
50%D, C, AAV, and P^J. The choice of years 20 and 60
in the definitions of the latter five statistics is meant to
capture "short"-term and medium-term considerations. For
instance, recovery should have occurred by year 60 in most
cases and the population should be well above 0.25Bq after
20 years. The catch and catch variability statistics for the
first 20 years provide an indication of the likely impacts of
recovery on the industry.
The need to examine aspects of the behavior of the man-
agement rules can be understood from Figure 3, which
shows results for four simulations for the combination of
a PFMC management procedure and an operating model
variant. The solid lines are the "true" time-trajectories of
spawning output (expressed in relation to the pre-exploi-
tation level) and the dotted lines reflect the estimates of
this ratio each time an assessment is conducted (every
third year for the analyses shown in Fig. 3). The up ar-
rows indicate when the assessment first indicates that the
population is overfished (based on the model estimates of
spawning output) — note that a population may be identi-
fied to be overfished more than once during a given simula-
tion. The down arrows indicate the years in which recovery
is predicted by the rebuilding analysis software (with the
estimates from the assessment) to occur with 60% probabil-
ity. The solid bar parallel to the j:-axis indicates the years in
which management is based on the rebuilding plan (rather
than the 40-10 rule). The bar will stretch from the up ar-
row to the down arrow unless the population is assessed
to have recovered to 0.4Bq (when management reverts to
being based on the 40-10 rule).
There are several possible impacts of the difference
between the perceived and true state of the system. For
example, the population can erroneously be assessed not
to be overfished in the first projection year (e.g. simulation
1 in Fig. 3). The statistic Pj^^, is designed to capture the
frequency of this possibility. Even if the population is as-
sessed to be overfished, there is no guarantee that it will
recover with the expected probability and in the "correct"
year. For example, for simulation 1, the stock assessment
indicates that recovery occurs in year 71 (the solid bar con-
sequently stops in year 71) even though the true population
size is less than 30% of fig at that time. The statistic F^^^
attempts to capture whether the rebuilding analysis per-
forms as expected given that the population is assessed to
be overfished at the start of the first projection year.
There are other aspects to evaluating the behavior of the
management rules in relation to the perceived and true
state of the system (e.g. the difference between the true
and estimated biomasses and recruitments). Although it
is straightforward to evaluate these aspects (e.g. Patter-
son and Kirkwood, 1995; Punt et al., 2002), they are not
considered in detail in this paper to reduce the volume of
results presented.
866
Fishery Bulletin 101(4)
Simulation 1
60 80 100
Simulation 3
Simulation 2
60 80 100 120
Simulation 4
Year
Figure 3
Time-trajectories of the "true" and the assessment model-estimated ratio of the spawning output to Sq
(depletion) for four simulations. The up arrows indicate the years in which the stock was declared to be in
need of rebuilding and the down arrows show the values of T^^^. The horizontal bars indicate the years
during which the stock is under a rebuilding plan. Year 41 is the first "projection year," i.e. the first year
in which the management procedure is used to determine the catches (the catches for the years prior to
year 41 are set equal to the historical catches — see Fig. lA)
Results and discussion
Detailed results for a single operating model variant and
management procedure
Figures 4 and 5 and Table 4 summarize aspects of a
simulation trial in which the operating model has its
baseline parameterization (Tables 1 and 2) and in which
the management procedure used to set harvest guidelines
is the PFMC management procedure with no constraints
on interannual variation in harvest guidelines other than
an upper limit of 10,000 t. The lack of any constraints on
changes in harvest guidelines has been imposed because
the PFMC has not adopted any such constraints. The har-
vest guideline is updated every third year.
Figure 4 shows the time-trajectories of catch, spawning
output in relation to the pre-exploitation equilibrium level
("true" and estimated), and the perceived fishing mortality
on which the harvest guideline is based for three of the 100
simulations that constitute a simulation trial. The horizon-
tal bars on the x-axis again reflect the year during which
the stock is managed by using the results from the rebuild-
ing analysis rather than the 40-10 rule. The most notable
feature of Figure 4 is the high variability in annual catches.
This variability arises for several reasons: 1) the additional
information on population biomass obtained each time a
survey occurs changes the perceived status of the resource
and hence how far the spawning output is from the target
level of OABq-, 2) an extension of the assessment period
changes the set of recruitments on which generation of
future recruitment is based; and 3) a change from being
under a rebuilding plan to being managed by means of the
40-10 rule can lead to marked changes in catch. The lat-
ter is evident by the change in fishing mortality and catch
when the spawning output is estimated to reach 0.4Sg (i.e.
the end of the horizontal bar). A marked impact due to
the addition of data for a single 3-year period may appear
surprising. However, effects of this nature have already
been observed for West Coast species (see, for example, the
2002 update to the sablefish [Anoplopoma fimbria] stock
assessment [Schirripa and Methot^l).
'^ Schirnpa. M. J., and R. Methot. 2002. Status of the sablefish
resource off the continental U.S. Pacific Coast in 2001. In Stock
assessment and fishery evaluation: appendix to the status of the
Pacific Coast groundfish fishery through 2001 and acceptable bio-
logical catches for 2002, x + 122 p. Pacific Fishery Management
Council, 7700 NE Ambassador Place, Portland, OR 97220.
Punt; Managing West Coast groundfish resources through simulations
867
r
o
E
Simulation 1
Simulation 2
Simulation 3
Year
Year
Figure 4
Time-trajectories of catch (upper panels), spawning output in relation to the pre-exploitation level (solid line
is "true"; dotted line is estimated) (center panels), and perceived fishing mortality (used to set the harvest
guideline [solid line]; dotted line=F^,gY proxy) (lower panels) for three individual simulations. The results in
this figure pertain to the baseline operating model and baseline Pacific Fishery Management Council manage-
ment procedure.
o
o
o 1
ID
^ .
*"
Depletion
0.4 08
Catch (t)
15000
1h
\ .^,<'rrrrrT7mTrrr: § -
o
6
y
|^^^^^_^-,.wv„-— -^
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Year Year
Figure 5
Piecewise medians (solid lines) and 90% intervals (dotted lines) for spawning output in relation to the pre-exploi-
tation equilibrium level (left panel) and catch (right panel). The results in this figure pertain to the baseline
operating model and baseline Pacific Fishery Management Council management procedure.
The extent of variability in catch in Figure 4 differs
markedly from the way advice on expected catches during
the rebuilding period is presented to the decision makers
(e.g. Fig. 6). One way to improve the presentation of in-
formation on expected catches would be to include some
individual catch trajectories from those on which the
rebuilding analysis is based. However, even these would
severely underestimate the actual extent of uncertainty
868
Fishery Bulletin 101(4)
Table 4
Performance statistics (see Table 3 for definitions) for six alternative management procedure variants. All of the calculations
table relate to the baseline operating model. PFMC = Pacific Fishery Management Council. N/A = not applicable.
m this
Management procedure F^^^.
Yre.
^decl
Results after 20 years
Results after 60 years
5%D
509-fD
AAV
C
^rec
5%D
50%D
AAV
C
P
rec
Baseline 0.22
72
0.82
0.22
0.33
0.33
1759
0.32
0.23
0.36
0.25
2847
0.80
With constraints 0.27
61
0.82
0.24
0.40
0.38
591
0.54
0.24
0.41
0.17
2440
0.89
No 10 years and estimated f ^jsy 0.42
68
0.82
0.24
0.34
0.30
1652
0.27
0.25
0.41
0.24
2649
0.84
Preferred 0.59
62
0.82
0.25
0.39
0.31
950
0.49
0,28
0.54
0.21
1961
0.96
PFMC (baseline) N/A
95
N/A
0.19
0.29
0.23
2273
0.07
0.24
0,33
0.20
2851
0.55
PFMC (preferred) N/A
64
N/A
0.23
0.36
0.30
1239
0.45
0.30
0.48
0.20
2265
0.93
O
° 1
Catch (t)
1000
_^— :^^^
o
2000 2005 2010 2015 2020 2025 2030
Year
Figure 6
Time-trajectories of catch (median and 95% intervals) for
the annual catch for widow rockfish based on a rebuilding
analysis conducted in 2002.
because they are conditioned on knowing the age-structure
of the population at the start of the projection period and
are based on fixed levels of fishing mortality during the
rebuilding period.
The impact of estimation uncertainty is also evident
in Figure 4. The following are three examples of this: 1)
management based on the rebuilding plan only starts in
year 53 in simulation 1 because, prior to this year, the stock
assessment indicates (erroneously) that the stock is above
rather than below 0.25Bq; 2) the resource is predicted to
have recovered to 0.4Bg in year 71 in simulation 1 (and
hence management is based on the 40-10 rule thereaf-
ter)— however, the spawning output is really only slightly
larger than O.Sfiy at this time; and 3) in simulation 3 the
assessment model indicates that the spawning output has
recovered to above 0.4Bg in year 65 when, in fact, it recov-
ered to 0.4Bq three years earlier.
The results of all 100 simulations are summarized by
the time-trajectories in Figure 5. The trajectories of catch
in Figure 5 are notably less variable that the individual
trajectories in Figure 4 because, for instance, the 5"^,
median, and 95'^ intervals for the catch in year 80 are
obtained by sorting all 100 year-80 catches and taking the
appropriate percentiles. Unlike the individual trajectories,
the median trajectories of catch and spawning output show
quite smooth changes over time. This result highlights the
importance of the AAV statistic that captures interannual
variation in catches within individual simulations.
Overall, there is a high probability (0.82) that the as-
sessment model identifies that the spawning output is less
than 0.25Bq at the start of the projection period (Table 4).
However, the probability that recovery occurs at or before
the Tjjjax y6ar predicted from the rebuilding analysis con-
ducted in projection year 1 is rather low (0.22) and 50%
of simulations exceed 0.4Sq only in year 72 (i.e. after 30
years). The probability of being below the overfished level
of 0.25Bq still exceeds 5% after 60 years of management
with this management procedure although there is an 80%
probability that the spawning output recovers to 0.4B„
sometime during the first 60 years of management with
the management procedure.
It should be noted that the impact of recruitment vari-
ability and assessment errors following recovery to OABq
can be consequential. For example, the probability of hav-
ing reached 0.45q after 60 years of management by using
the management procedure exceeds 0.8 but the median
value of the ratio of the spawning output in year 60 to Bq
is nevertheless still less than 0.4 (Table 4, Fig. 5). One rea-
son for the spawning output not stabilizing at 0.4 Bg is a dis-
crepancy between the fishing mortality rate that stabilizes
the population at Bq (deterministically) and Fgg,. For the
baseline steepness of 0.4, the fishing mortality required to
stabilize the spawning output at 0.4 Bq actually corresponds
to a lower fishing mortality than ^50% (closer to J^63%^-
Sensitivity to alternative management procedures
Table 4 includes results for a range of variants of the
baseline management procedure designed to improve its
performance. The following are areas where improved
performance is desirable: 1) the extent of interannual
variability in catches; 2) the similarity between the year
Punt: Managing West Coast groundfish resources through simulations
869
in which the rebuilding analysis indicates recovery will
occur and the year at or before which it actually occurs;
and 3) the probability of being below the overfished level
after 20 and 60 years.
The first variant of the baseline management procedure
("with constraints" in Table 4) involves imposing maximum
and minimum catch limits of 30 and 8000 t and constrain-
ing changes in harvest guideline not to exceed 25% from
one year to the next, except in the first year when reduc-
tions of up to 99% are allowed. This variant leads to much
lower interannual variation in catches when a 60-year pe-
riod is considered ( 17% compared with 25% ) but the AAV
is actually higher for the first 20 years. This variant also
leads to higher probabilities of recovery. However, there is
still a large discrepancy between the actual year of recov-
ery to 0.4Sg and the year that underlies the management
procedure (the value of F^^^ in Table 4 is only 0.27 for the
"with constraints" variant).
The second variant considered ("no 10 year and esti-
mated ^msy" see Table 4) drops the requirement that T^^^
be defined as 10 years if the resource can be recovered in 10
years and instead always sets T^^^ to T^^^^ plus one mean
generation. It also allows the Fj^gy proxy used when apply-
ing the 40-10 rule to differ from the default value of Fgg^^
by setting it to F (Jakobsen, 1993) ii^ F is lower than
^MSY- Estimating (rather than fixing) F^^gy is consistent
with the recommendation of Brodziak (2002). The major
performance difference between this variant and the base-
line management procedure is the increased value of F^.^^.
The "preferred" variant in Table 4 combines the features
of the "with constraints" and "no 10 years and estimated
^msy" variants. Compared with the baseline management
procedure, it leads to a markedly increased value for F^^^
(remarkably close, in fact, to the desired value of 0.6),
slightly lower catch variability, a less than 5% chance of
being overfished after 20 years, and higher probabilities of
being recovered to 0.4Bg after 20 and 60 years of manage-
ment. The major disadvantage of this variant is the lower
catches and that it leaves the spawning output well above
40% of Bg after 60 years (see row "preferred" in Table 4).
Prior to the adoption of Amendment 11 of its Groundfish
Management Plan, the PFMC set harvest guidelines using
only the 40-10 rule.^ Table 4 therefore also lists results for
management procedures based on the 40-10 rule. When
the 40-10 rule is applied without any constraints ("PFMC
(baseline)" in Table 4), the probability of recovery and the
values for the "50%'D" statistic are lower (particularly
the former) than for the "preferred" variant. In contrast,
application of the 40-10 rule with constraints ("PFMC
(preferred)" in Table 4) leads, arguably, to no more than a
slight difference in cateh (the 40-10 rule achieves higher
catches) and probability of recovery (the "preferred" vari-
ant achieves a higher probability of recovery ). The remain-
ing analyses of this paper focus on the "preferred" variant.
Future consideration of management procedures for West
Coast groundfish resources should consider a management
procedure that is based simply on the 40-10 rule and has
no associated rebuilding analysis component, at least for
' Albeit with different target fishing mortality levels.
comparative purposes. At present, however, such a man-
agement procedure would be inconsistent with the SFA
because it would not specify the time to recover to the proxy
for 5^,gY 'even if the results of this paper suggest that there
is considerable uncertainty associated with the estimation
of this particular quantity).
Sensitivity to alternative operating model specifications
The values assumed for h and a^ in the baseline operating
model are somewhat arbitrary. Table 5 therefore examines
the sensitivity of the results for the "preferred" manage-
ment procedure to varying the values assumed for these
parameters, as well as that of the size of spawning output
at the start of the first projection year to fig.
The results are, as expected, sensitive to all three of the
factors considered. Increasing a^ from 0.4 through 0.6 to 1
leads to lower and more variable catches, a slightly higher
probability of recovery in the first 20 years and a markedly
higher value of 50%D after 60 years (0.74 for <7;j=l com-
pared to 0.46 for cr^=0.4). The ability to detect an overfished
stock declines slightly as the extent of variation in recruit-
ment increases. The management procedure behaves as
expected as steepness is increased from 0.25 through 0.4 to
0.7; the probability of recovery is markedly higher for high
values of steepness even though the management proce-
dure does identify cases with low steepness, and accordingly
sets very low harvest guidelines in such cases. However,
it is perhaps noteworthy that the probability of correctly
identifying that the resource is overfished is lowest for the
least productive scenario. The catches for the scenario in
which the spawning output is 10% of Bg at the start of the
first projection year are much lower than for the baseline
scenario, particularly over the first 20 years. However, these
lower catches are necessary to achieve recovery (the median
value of the statistic 50%I> after 60 years is 0.52 and there
is a 0.93 probability of the spawning output having recov-
ered to 0.4Bg after 60 years for this scenario).
The behavior of the management procedure can be evalu-
ated in terms of whether it eventually allows the stock to
recover to 0.4Bg and whether it keeps the stock away from
the overfished level of 0.25Bq. The "preferred" management
procedure can be argued to satisfy this criterion, except
possibly for the scenario with the lowest steepness but,
even in this case, the probability of recovery is 0.6 after
60 years.
The value for the F^^^ statistic varies markedly depend-
ing on steepness and the ratio of the spawning output at
the start of the first projection year to Bg. Although the
"preferred" management procedure performs well for the
baseline scenario in terms of recovering the resource by the
predicted value for T^^^, this good performance is clearly a
fortunate anomaly. However, it does help to highlight that
predictions of the year-to-recovery from rebuilding analy-
ses should be interpreted with considerable caution.
Sensitivity to data quality
The data-related specifications for the baseline trial
(Table 2) could be considered to be data-rich. It is therefore
870
Fishery Bulletin 101(4)
Table 5
Performance statistics (see Table 3 for definitions) for 10 variants of the base)
ine operating
model.
All of the calculations in this |
table relate to the preferred management proce
dure. N/A = not
applicable.
Operating model scenario
f'.ec
^ec
■fdecl
Results after 2C
years
Results
after 6C
years
5%D
50%D
AAV
C
^rec
5<7rD
50%D
AAV
C
P,e,
Baseline
0.59
62
0.82
0.25
0.39
0.31
950
0.49
0.28
0.54
0.21
1961
0.96
Structural changes
a^ = 0.4
0.59
63
0.86
0.24
0.38
0.26
1242
0.44
0.25
0.46
0.18
2379
0.87
a^=l
0.59
61
0.72
0.23
0.41
0.43
417
0.54
0.32
0.74
0.32
592
0.96
/i=0.25
0.15
94
0.76
0.20
0.28
0.76
86
0.02
0.23
0.38
0.50
126
0.60
/!=0.7
0.84
53
0.87
0.31
0.46
0.16
3427
0.93
0.40
0.61
0.14
3951
1.00
Initial spawning out = 0.1 B„
0.42
72
1.00
0.19
0.29
0.43
417
0.05
0.27
0.52
0.23
1375
0.93
Initial spawning out = 0.4 Bg
N/A
N/A
N/A
0.31
0.50
0.21
2881
0.92
0.30
0.66
0.19
2849
0.97
Data-related changes
Deterministic data
0.68
61
0.84
0.29
0.38
0.30
957
0.51
0.31
0.55
0.20
2050
0.98
n'=5Q
0.68
60
0.82
0.26
0.39
0.32
785
0.56
0.29
0.55
0.22
1938
0.97
a<-=l
0.56
62
0.79
0.20
0.39
0.31
987
0.48
0.31
0.57
0.22
1962
0.97
5-yr update frequency
0.55
62
0.80
0.21
0.38
0.27
1160
0.49
0.27
0.53
0.19
1980
0.95
important to assess the sensitivity of the results to the
quahty of the data. The row "deterministic data" in Table
5 provides results for a trial in which the survey biomass
index, the catch-rate index, and the age-composition data
are known without error. The results from this trial provide
an upper bound on the impact of improved data quality
on the assessment results.^ Somewhat surprisingly, the
results for this trial are not notably better than for the
baseline trial — the most notable difference between the
baseline trial and the "deterministic data" trial being the
higher values for the "5%D" statistics for the latter trial.
The lack of major improvement in performance arises
because, even with perfect information on spawning
output and recruitment, it is still not possible to estimate
Bq exactly by multiplying average recruitment for the first
10 years of the assessment period by spawning output-per-
recruit in the absence of fishing (hence the value of 0.84
for Pjed'- Furthermore, the rebuilding analyses are still
based on generating future recruitment by using spawn-
ing output and recruitment data for only 20 years, which is
clearly a major source of variability in the predictions from
the rebuilding analysis.
Decreasing the catch-at-age sample size from 200 to 50
has relatively little impact on the values for the perfor-
mance statistics (the AAV statistic is marginally higher
and the average catch, particularly for the 20-year pro-
jection horizon, is lower). Decreasing the precision of the
catch-rate data has a rather larger impact. This is most
evident in the value for the "5%D" statistic which is 0.2
rather than 0.25, as is the case for the baseline trial. The
" The assessment still ignores interannual changes in selectivity;
therefore the assessment results will not be exactly the same as
the true values.
"5-yr update frequency" scenario in Table 5 examines the
implications conducting assessments every fifth rather
than every third year. The results are not markedly sensi-
tive to the interassessment period although the lower val-
ues for the "5%D" statistics are perhaps noteworthy.
General remarks
The framework developed in this paper provides an objec-
tive basis for contrasting different management procedures
and evaluating their sensitivity to uncertainty. Given such
a framework, it becomes possible to compare variants of
one class of management procedure (e.g. Table 4) and to
compare variants among different classes of management
procedure.
The management procedure options presented in this
paper are but a small subset of those possible. In particular,
it should be possible to improve performance by modifying
the approach used to generate future recruitment when
conducting rebuilding analyses to make use of some form
of stock-recruitment relationship. One reason for expected
improved performance is that it may then be feasible to
estimate the fishing mortality rate corresponding to 0.4Bg
rather than having to set it to the default value of F^j,^, or
basing it on F . Other possible management procedure
options include 1) not increasing the rebuilding fishing
mortality rate selected when the rebuilding analysis was
first conducted if a stock is recovering faster than initially
anticipated; 2) not decreasing the rebuilding fishing mor-
tality rate as long of the probability of recovery by T^^^^ is
at least 0.5; and 3) smoothing the discontinuity that arises
when a stock changes status from being under a rebuild-
ing plan to being managed with the 40-10 rule when the
Punt: Managing West Coast groundfish resources through simulations
871
stock has recovered to 0.4fig. In terms of the last option,
one of the issues considered an early rebuilding analysis for
widow rockfish involved fishing mortality increasing to its
target level as the stock approaches 0.4BQ(MacCall^).
The values for the F^^^ statistic highlight that the predic-
tions of the time to recovery (even in a probabilistic sense)
from rebuilding analyses are highly uncertain. The uncer-
tainty of this estimate of the time to recovery is due to the
uncertainty about current stock size and that associated
with making long-term predictions based on a short time-
series of spawning output and recruitment data.
Although the performance of the management proce-
dures is less than ideal, the results are almost certainly
optimistic because the operating model is extremely simple
and considers no major structural uncertainties (except for
variability in selectivity over time). In contrast, Punt et al.
(2002) found that including spatial structure in an oper-
ating model and assessing the stock by using a spatially
aggregated assessment approach led to assessments that
were markedly in error. However, the simulations con-
ducted by Punt et al. (2002) were developed for a far more
data-poor situation than that for West Coast groundfish,
although there is also clearly spatial structure in the West
Coast groundfish fishery. Another source of uncertainty not
considered in this paper but that may be of critical impor-
tance to the management of West Coast groundfish species
is the impact of environmental regime shifts, which have
been argued to impact long-term trends in recruitment (e.g.
Francis et al., 1998).
An important aspect of this study is the ability to focus on
the relationship between the overall performance of a man-
agement procedure and the performance of its constituent
parts. For example, the results for the "deterministic data"
scenario in Table 4 show that given the approach used to
conduct the future projections, even perfect information
from surveys and very large age-composition samples are
unlikely to lead to marked improvements over the current
situation if that situation is adequately modeled by the
baseline operating model. Identification of the key sources
of uncertainty could be used to focus future management-
related research activities.
The computational requirements of the calculations out-
lined above are substantial. In particular, the need to apply
a fairly complicated method of stock assessment once every
three years means that rapid evaluation of management
procedures is (currently) computationally not feasible.
It is possible, in principle, to simplify the management
procedure considerably by assuming that the results from
a stock assessment can be mimicked by generating a bio-
mass estimate based on the "true" biomass but with some
random error (e.g. Hilborn et al., 2002). However, although
such an approach may be satisfactory for some manage-
ment procedures (e.g. those that set the harvest guideline
equal to some fraction of the current biomass), this is not
the case for PFMC-type management procedures that de-
pend on the (assessed) age-structure of the population.
9 MacCall, A. D. 2002. Personal commun. NMFS Santa Cruz
Laboratory, 110 Shaffer Rd, Santa Cruz, CA 95060.
It needs to be recognized that any simulation study is by
design case-specific. However, the conclusions of this study
may be relevant to a fairly broad set of West Coast rock-
fish species owing to their similar biology and exploitation
history — the two factors most likely to impact the relative
performance of different management procedures.
Acknowledgments
Discussions with Alec MacCall, John DeVore, and Richard
Methot are gratefully acknowledged as are the comments
on an earlier version of this paper by Pamela Mace and
two anonymous reviewers. This work was funded through
NMFS grant NA07FE0473.
Literature cited
Barnes, W. R.
1999. Viewpoint: an industry view of the application of
operational management procedures to setting total allow-
able catches for the South African pelagic fishery. ICES.
J. Mar. Sci. 56:1067-1069.
Beddington, J. R., and J. G. Cooke.
1983. The potential yield of fish stocks. FAO Fish. Tech.
Pap. 242:1-47.
Brodiak, J.
2002. In search of optimal harvest rates of west coast
groundfish. N. Am. Fish. Manage. 22:258-271
Butterworth, D. S., and M. O. Bergh.
1993. The development of a management procedure for the
South African anchovy resource. In Risk evaluation and
biological reference points for fisheries management (S. J.
Smith, J. J. Hunt, and D. Rivard, eds. ), p. 83-89. Can. Spec.
Publ. J. Fish. Aquat. Sci. 120.
Cochrane, K. L., D. S. Butterworth, J. A. A. De Oliveira, and
B. A. Roel.
1998. Management procedures in a fishery based on highly
variable stocks and with confiicting objectives: experiences
in the South African pelagic fishery. Rev Fish. Biol. Fish.
8:177-214.
Dorn, M. W.
2002. Advice on west coast rockfish harvest rates from
Bayesian meta-analysis of stock-recruit relationships. N.
Am. Fish. Manage. 22:280-300.
Francis, R. I. C. C.
1992. Use of risk analysis to assess fishery management
strategies: a case study using orange roughy (Hoplostethus
atlanticus) on the Chatham Rise, New Zealand. Can. J.
Fish. Aquat. Sci. 49:922-930.
Francis, R. S., A. Hare, A. Hollowed, and W. Wooster.
1998. Effects of interdecadal variability on the oceanic eco-
systems of the NE Pacific. Fish. Oceanogr. 7:1-21.
Geromont, H. F, J. A. A. De Oliveira, S. J. Johnston, and
C. L. Cunningham.
1999. Development and application of management proce-
dures for fisheries in southern Africa. ICES J. Mar. Sci.
56:953-966.
Hilborn, R.
1979. Comparison of fisheries control systems that uti-
lize catch and effort data. J. Fish. Res. Board Can. 36:
1477-1489.
872
Fishery Bulletin 101(4)
Hilborn, R.. A. Parma, and M. Maunder.
2002. Exploitation rate reference points for west coast rock-
fish: are they robust and are there better alternatives? N.
Am. Fish. Manage. 22:365-375.
lanelli.J. N.
2002. Simulation analyses testing the robustness of pro-
ductivity determinations from West-Coast Pacific ocean
perch stock assessment data. N. Am. Fish. Manage. 22:
301-310.
IWC (International Whaling Commission).
1992. Report of the Fourth Comprehensive Assessment
Workshop on Management Procedures. Rep. Int. Whal.
Comm. 42:305-321.
2001. Report of the standing working group on the develop-
ment of an Aboriginal subsistence whaling management
procedure. J. Cetacean Res. Manag. 4(suppl.): 148-177.
Jakobsen, T.
1993. The behaviour of F^^^, F^^^ and F(„g|, in response to
variation in parameters used for their estimation. /;! Risk
evaluation and biological reference points for fisheries man-
agement (S. J. Smith, J. J. Hunt, and D. Rivard, eds.), p.
1 19-125. Can. Spec. Publ. J. Fish. Aquat. Sci. 120.
Methot, R. D.
2000. Technical description of the stock synthesis assess-
ment program, 56 p. NOAA Tech. Memo., NMFS-NWFSC-
43.
Patterson, K. R., and G. P. Kirkwood.
1995. Comparative performance of ADAPT and Laurec-
Shepherd methods for estimating fish population param-
eters and in stock management. ICES J. Mar. Sci. 52;
183-196.
Punt, A. E., A. D. M. Smith, and G. Cui.
2001. Review of progress in the introduction of management
strategy evaluation (MSE) approaches in Australia's South
East Fishery Mar Freshw. Res. 52:719-726.
2002. Evaluation of management tools for Australia's South
East Fishery. 2. How well do commonly-used stock assess-
ment methods perform? Mar. Freshw. Res. 53:631-644.
Restrepo, V. R., and J. E. Powers.
1999. Precautionary control rules in US fisheries manage-
ment: specification and performance. ICES. J. Mar. Sci.
56:846-852.
Rosenberg, A. A., and S. Brault.
1993. Choosing a management strategy for stock rebuilding
when control is uncertain. In Risk evaluation and biologi-
cal reference points for fisheries management (S. J. Smith,
J. J. Hunt, and D. Rivard, eds.), 243-249. Can. Spec. Publ.
J. Fish. Aquat. Sci. 120.
Sampson, D. B., and Y. Yin.
1998. A Monte Carlo evaluation of the stock synthesis
assessment program. In Fishery stock assessment models
(R Funk, T J. Quinn II, J. Heifetz, J. N. lanelli, J. E. Powers,
J. F. Schweigert, P. J. Sullivan, and C. I. Zhang, eds.), p.
315-338. Alaska Sea Grant, Fairbanks, AK.
Starr, P. J., P. A. Breen, R. H. Hilborn, and T H. Kendrick.
1997. Evaluation of a management decision rule for a New
Zealand rock lobster substoek. Mar Freshw. Res. 48:
1093-1101.
Appendix 1 : An overview of the technical aspects of
the PFMC's rebuilding analysis
The key steps of the PFMC's rebuilding analysis are 1)
to select the maximum allowable rebuilding time iT^^^),
R,
if a = a„,„
W,,_
,„-,<-""*'"-"
if a„,„ PFMC (Pacific Fishery Management Council). 2001. SSC
terms of reference for groundfish rebuilding analysis, 9 p.
Pacific Fishery Management Council, 7700 NE Ambassador
Place, Portland, OR 97220.
Punt: Managing West Coast groundfish resources through simulations
873
It should also be noted that no account is taken of uncer-
tainty regarding the current age structure, natural mortal-
ity, selectivity, etc., although the projections do account for
uncertainty about future recruitment
year in which the spawning output first reaches 0.48^. T^^^
is the median of the distribution for this year constructed
by conducting projections for many different (random)
realizations of future recruitment.
Selecting the maximum allowable rebuilding period
The maximum allowable rebuilding time, T^^^, is defined
as the maximum of 10 years and the sum of the mean
generation time and the minimum possible rebuilding
time. This specification implements the requirement of
the SFA to "take into account the status and biology of any
overfished stocks offish, [and] the needs of fishing commu-
nities." The minimum possible rebuilding period for a given
future projection is computed by projecting the population
forward with zero fishing mortality and by identifying the
Calculating the target fishing mortality rate
The target fishing mortality rate and hence the harvest
guideline are determined by projecting the population
forwards many times ( 100 times for the purposes of this
paper), each time with a different sequence of future
recruitment and for a variety of alternative Fs and then
identifying the level of F that corresponds to the spawning
output having reached 0.4B(, by T^^^ with the prespecified
probability p.
874
Abstract— Between March 2000 and
April 2001 two commercial fishing ves-
sels fished for toothfish (Dissostichus
eleginoides) off South Georgia using
pots. A significant number of lithodid
crabs (three species of Paralomis
spp.) were caught as bycatch. Paralo-
mis spinosissima occurred in shallow
water, generally shallower than 700 m.
Paralomis anamerae, not previously
reported from this area and therefore
representing a considerable southerly
extension in the reported geographic
range of this species, had an interme-
diate depth distribution from 400 to
800 m. Paralomis formosa was present
in shallow waters but reached much
higher catch levels (and, presumably,
densities) between 800 and 1400 m.
Differences were also noted in depth
distribution of the sexes and size of
crabs. Depth, soak time, and area were
found to significantly influence crab
catch rates. Few crabs (3% of P. spi-
nosissima and T7c of P. formosa) were
males above the legal size limit and
could therefore be retained. All other
crabs were discarded. Most crabs ( >99%
of P. formosa, >97% off! spinosissima,
and >90% of P. anamerae) were lively
on arrival on deck and at subsequent
discard. Mortality rates estimated
from re-immersion experiments indi-
cated that on the vessel where pots
were emptied directly onto the factory
conveyor belt 78-89% of crabs would
survive discarding, whereas on the
vessel where crabs were emptied down
a vertical chute prior to being sorted,
survivorship was 38-58%. Of the
three, P. anamerae was the most vul-
nerable to handling onboard and sub-
sequent discarding. Paralomis spino-
sissima seemed more vulnerable than
P. formosa.
Distribution, demography, and discard
mortality of crabs caught as bycatch
in an experimental pot fishery
for toothfish {Dissostichus eleginoides)
in the South Atlantic
Martin G. Purves
Marine and Coastal Management
P.O. Box X2, Roggebaai 8012
Cape Town, South Africa
Present address: Resources Assessment Group
47 Pnnce's Gate
London SW7 7QA, United Kingdom
E-mail address; m purvesigimpenal acuk
David J. Agnew
Renewable Resources Assessment Group
Imperial College
Royal School of Mines
Prince Consort Road
London, SW7 2BP, United Kingdom
Guillermo Moreno
Tim Daw
MRAG Ltd
47 Pnnces Gate
London, SW7 2QA, United Kingdom
Cynthia Yau
Zoology Department
University of Aberdeen
Aberdeen, AB24 2TZ, United Kingdom
Graham Pilling
Centre for Environment, Fisheries and Aquaculture Science (CEFAS)
Pakefleld Road
Lowestoft, NR33 OHT, United Kingdom
Manuscript approved for publication
17 April 2003 by Scientific Editor
Manuscript received 26 June 2003
at NMFS Scientific Publications Office.
Fish Bull. 101:874-888 (2003).
The Commission for the Conservation
of Ajitarctic Marine Living Resources
(CCAMLR) and its Scientific Committee
were pioneers in the development of the
"ecosystem approach" for the manage-
ment of fisheries. Using this approach
the Commission is bound to consider
the impact of any fishery on both the
target, dependent, and related species.
Currently, the most important fishery in
CCAMLR waters is the longline fishery
for Patagonian toothfish (Dissostichus
eleginoides) and fishing grounds near
South Georgia Island and Shag Rocks in
CCAMLR subarea 48.3 (South Atlantic
sector) are among the most important.
Mitigation measures, including require-
ments for setting at night, in the winter,
and with specialized gear, have been
introduced to reduce incidental mortal-
ity of birds being hooked by longlines.
However, these measures impose severe
Purves et al.: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic
875
operational restrictions on the fishery, and low levels of
bird mortality still occur (CCAMLRM.
Pot fishing for toothfish has recently been tried around
South Georgia (Agnew et al., 2001), and although pots do
not catch birds, they do take lithodid crabs as bycatch.
Crabs are largely a "nuisance" catch when fishing for
toothfish, but this bycatch is clearly of concern in rela-
tion to crab populations, and must be considered within
the CCAMLR's ecosystem approach. A small amount of
exploratory crab fishing has already taken place around
South Georgia Island and Shag Rocks. Only 798 metric
tons (t) of crabs have been taken in directed crab fisheries
since 1992; by the FV Pro Surveyor (July-August 1992;
299 t; CCAMLR2), the FV American Champion (September
1995-January 1996; 497 t; CCAMLRM), and the FW Argos
Helena (August 1999; 2 t) (CCAMLR''). A pot fishery for
toothfish is likely to take place in deep water where cur-
rent longline fishing concentrates (around the 1000 m con-
tour; Agnew et al., 1999) rather than in shallower waters
(<400 m) where crab fishing has taken place (Otto and
Macintosh, 1996; Watters, 1997 ). Toothfish pot fishing may
therefore impact different crab population components
from those impacted by the crab fishery.
We investigated the likely effects of toothfish pot fishing
on crabs caught as bycatch on board two commercial fish-
ing vessels. These vessels conducted three separate trials
fishing for toothfish around South Georgia between March
2000 and April 2001. This paper reports on the species of
crab taken during these trials, as well as the distribution of
crabs and their biological characteristics. In common with
many other crab fisheries (Hoggarth, 1991; Schmidt and
Pengilly, 1993) retention size limits are set for the fishery
around South Georgia. Only males greater than 102 mm
carapace width for Paralomis spinosissima and 90 mm for
P. formosa may be retained. All undersize and female crabs
from both the toothfish pot fishery and the crab fishery
must be discarded. We also report results of experiments
on survival of such discards.
Methods
During the first cruise (March to May 2000), two observ-
ers were deployed on board the FV Argos Georgia (cruise
Gl). Detailed information was collected on the number
of toothfish and the numbers and species composition of
' CCAMLR (Commission for the Conservation of Antarctic Marine
Living Resources). 1999. Report of the Working Group for
Fish Stock Assessment, 110 p. Annex 5 in the report of the
eighteenth meeting of the Scientific Committee. CCAMLR, RO.
Box 213, North Hobart, Tasmania 7002, Australia.
- CCAMLR. 1992. Report of the working group for fish stock
assessment, 164 p. Annex 5 in the Report of the eleventh meet-
ing of the Scientific Committee. CCAMLR, RO. Box 213, North
Hobart, Tasmania 7002, Australia.
^ CCAMLR. 1997. Report of the working group for fish stock
assessment, 169 p. Annex 5 in the Report of the sixteenth meet-
ing of the Scientific Committee. CCAMLR, P.O. Box 213, North
Hobart, Tasmania 7002, Australia.
■• CCAMLR. 2000. CCAMLR statistical bulletin, 153 p.
CCAMLR, RO. Box 2 13, North Hobart, Tasmania 7002, Australia.
I ■ P. spinosissima
n P formosa
aP anamerae
Depth range (m)
Figure 1
Catch in numbers per pot of the three dominant crab
species by depth range as found during fishing opera-
tions of the Argos Georgia from May to April 2000. Pots
were set between 211m and 1651 m depth.
crabs caught in the pots. Information was also collected on
fishing results, such as catch rates, fish bycatch, and the
commercial viability of this fishing method (Agnew et al.,
2001), and the diet of toothfish (Pilling et al., 2001). In the
second and third toothfish pot cruises single observers were
deployed on the Argos Georgia (cruise G2) and another
vessel, the FV Argos Helena (cruise H), which fished simul-
taneously from January to April 2001 . Fishing gear and the
configuration of gear was similar for all three cruises. The
semiconical pots of approximately 80 cm height were con-
structed of steel frames and covered with 80-mm polysteel
(Movline) mesh. A collapsible funnel entrance was situated
on the side of the pot, orientated horizontally, and tapering
to the pot's interior. A drawstring held the bottom mesh
together in the middle. This configuration allowed the
pots to be emptied easily when hauled aboard and to be
stacked on top of each other when not in use. A panel was
sewn into the pots by using biodegradable sisal twine to
ensure that crabs could eventually escape from lost pots
and to prevent "ghost fishing." However, catch handling
methods were different on the two vessels; the pots were
emptied directly onto the factory conveyor belt on the Argos
Georgia and emptied down a chute to the factory level on
the Argos Helena.
During the first cruise (March to May 2000) depth of fish-
ing, determined as water depth by onboard echo sounders,
was related to the species of crab caught. Paralomis spino-
sissima were generally caught in relatively shallow waters,
whereas P. formosa tended to be caught in much greater
numbers in deeper waters (Fig. 1). Paralomis anamerae
876
Fishery Bulletin 101(4)
Figure 2
Map of areas used in the study. SR=Shag Rocks. NG=North South Georgia, EG=East South Georgia,
SG=South South Georgia, and the 200-m, 1000-m and 3000-m bathymetric contours are shown. The
total number of hauls in each of these areas were as follow: SR=145, NG=50, EG=30, and SG=15.
Inset shows the position of the study area in relation to the South Atlantic.
(Macpherson, 1988) were caught at intermediate depths.
Other influences on crab catch rates were investigated by
using two cruises in 2001, which covered the study area
more evenly (most hauls in cruise Gl were concentrated
in a small area around southeast Shag Rocks; see Fig. 2).
For each species two generalized linear models were con-
structed with Splus statistical software (version 2000,
Mathsoft Engineering & Education, Inc., Cambridge, UK):
a binomial link model for the probability of encountering a
crab of that species (pe) and a Gaussian link model for the
natural logarithm of CPUE (catch in numbers per pot) for
all nonzero catches. Both models were of the form A=pjX
depth+p.2xarea-(-p3Xvessel+p4Xsoak time, where for the
binomial models, A was set to 1 if crabs had been caught
and 0 if they had not, and for the Gaussian models, A was
set to ln(numbers per pot) for all sets catching crabs. Area
and vessel parameters were factors. Depth and soak time
were modeled as linear continuous variables, except in the
case of the Gaussian model for P. forniosa, where a third-
order polynomial best described the relationship between
CPUE and depth. Predictions from the two models were
combined to predict crab catch rates per pot,
(} = ;7(c)xexp[H±1.96x5£],
where/'/ = the predicted In(CPUE) from the Gaussian
model; and
SE = the standard error from the Gaussian model.
Biological data were collected from all crabs in randomly
selected pots. Carapace widths, carapace lengths, chela
height, and chela length were measured to the nearest
millimeter below by using calipers. Weights were measured
to the nearest 5 g below with spring-balanced scales. Sex,
maturity stage, condition of the carapace, and an index of
vitality (Table 1) were recorded for a subsample of crabs,
selected as a random portion across species from the
contents of selected pots. Identification of the uncommon
species, Paralomis anamerae, Neolithodes diomedeae, and
Lithodes murrayi, was confirmed by using dried specimens
in London (at Imperial College and the British Museum of
Natural History).
Male size at maturity was determined from the allometric
relationship between carapace length (L) and right (domi-
nant) chela size (height, CH, or length, CD. The slope of
the L-CH or L-CL relationship is assumed to change when
crabs reach maturity. Around South Georgia, Otto and
Macintosh (1996) used L and CH to determine the size at
maturity of male P. spinosissima, and around the Falkland
Islands Hoggarth (1991) used L and CL. For both P. spino-
sissima and P. fonnosa we found that the intersection of
the two lines corresponding to the onset of maturity was
not easy to identify from the relationship between CL and
L. We therefore used CH in our relationships. Following the
methods ofSomerton and Otto (1986), two linear regression
lines were fitted to natural logarithms of L and CH or CL.
These lines represent juvenile and adult phases in the L-CH
or CL relationship and intersected at a point taken as the L
at which males became mature. The regression lines of best
fit were determined by minimizing the combined residual
sums of squares, and standard errors were estimated by
using 500 bootstraps (sampling with replacement).
Females were classified into two categories, "eggs absent"
and "eggs present." The vast majority of "eggs absent" fe-
males were small immature animals, but some large ani-
mals were also encountered in this category. Consequently,
for estimating size at maturity, females in the "eggs pres-
ent" category were defined as "mature" and those in the
"eggs absent" category were classified as "immature" up to
the size at which the proportion of females with eggs (i.e.
mature) reached 90%, after which they were classified as
"mature without eggs." Female size at maturity was deter-
Purves et al.: Distribution, demography, and discard mortality of crabs caugtit as bycatch in the South Atlantic
877
Relative
Table 1
index used for assessing vitality in Paraloniis spp.
Vitality
index
Description
Characteristics
1
2
3
4
Lively
Lively but limp
Dead
Dead and eaten
Limbs supported and held out. Limbs resist manipulation,
objects. Can hang on smooth end of forceps by 1 claw (weakest
Legs hang when picked up. Claws weak and can be opened — not
own weight on forceps. Mouthparts move, indicating life
seawater.
No signs of life. Mouthparts do not move when submerged.
Only shell or carapace remaining.
Crabs actively pinch
on large crabs).
capable of supporting
when submerged in
mined by plotting the proportion of mature females against
size (carapace width) and determining the point of 50%
maturity (Sragg). Logistic curves of the form {proportion
mature = 1/(1 + expi-ricarapace width-Sm^Q))}] were used
to estimate 50% maturity (Sm^g) and its standard error by
using the nonhnear fitting function in Splus.
Three different experiments were conducted between
March 2000 and April 2001 to assess crab discard survival
rates. During cruise Gl a number of alive and active crabs
from one haul were tagged through one of the lateral plates
of the abdomen with Hallprint plastic T-bar tags and main-
tained in running seawater before they were placed in pots
prior to the next setting. Once these pots were rehauled, the
vitality of the tagged crabs was assessed by using the four
point relative scale shown in Table 1. A control group of crabs
were similarly tagged and kept onboard in running seawater
to monitor any effect that tagging might have had on sur-
vival. Survival experiments conducted on cruises G2 and H
differed in that crabs were selected at random and included
individuals that were already "limp" prior to re-immersion.
Crabs were tagged with thin strips of masking tape around
their legs prior to re-immersion. To ensure that the same
crabs were assessed for vitality after rehauling, pots were
marked and sealed off to prevent any new captures.
Estimates of the total survival that can be expected
after discarding were made in the following manner. By
observing crabs on arrival on deck we determined the pro-
portion of animals that arrived on deck as lively {lively^),
or limp (limp J, or dead. Using the data from the survival
experiments we set p(liuely,lively) as the probability that
a lively animal that is discarded will recover to a lively
condition (this was estimated by calculating the proportion
of experimentally re-immersed lively animals that were
recovered as lively). We defined p(limp,liuely) similarly.
The proportion of discarded animals that were lively and
would continue to be lively is lively ^xpdiuely, lively) and the
proportion of discarded animals that were limp but would
recover to a lively condition is limp ^xpdimpjively). The
overall survivorship, S, is then
S = lively ^xpilivelyMvely) + limp^xpilimp, lively).
In our experiments, some of the damage may have
occurred on rehauling the pots after re-immersion, a
situation that would not normally occur once crabs have
been discarded. p(lively,lively) can be corrected for this by
adding to it the proportion of animals that were not lively
when first hauled (i.e. 1-livelyJ, termed the rehaul correc-
tion. For instance, suppose that 1% of the crabs were not
lively on the first hauling, and in the experiment 4% of
the re-immersed lively crabs were not lively on recovery.
The rehaul correction would indicate that 1% of the re-im-
mersed crabs would have been damaged anyway simply by
the hauling process, and that therefore the correct damage
rate attributed simply to the initial capture and discarding
would be 3%.
Results
Crab catch
The majority of the bycatch comprised two species of litho-
did crabs (Anomura: Lithodidae), Paralomis spinosissima
and P. formosa. Both species have been previously reported
in catches around Shag Rocks and South Georgia (Otto and
Macintosh, 1996) and have formed a large proportion of the
total catch of crab (Table 2). Crab species formed 69.5% of
the total weight of all species caught, including toothfish,
and 98.2% of the total numbers of individuals caught.
Three other species of crab were also identified during
the pot trials. The most abundant of these was Paralo-
mis anamerae (Macpherson, 1988), which like the other
Paralomis species, was subject to detailed biological sam-
pHng. 12,370 individuals of this species (1 721 kg) were
caught. All individuals were discarded because they were
smaller than the minimum size limit for the smaller of
the two regulated species, P. formosa. Two other species,
Neolithodes diomedeae and Lithodes murrayi, were also
caught in small numbers.
Distribution
Crab distribution was investigated for the areas as defined
in Figure 2. There were too few data from South of South
Georgia; therefore the analysis was restricted to data from
Shag Rocks, North South Georgia, and East South Georgia.
A few pot strings had been left for several days because
878
Fishery Bulletin 101(4)
Table 2
Species proportion (no. of
crabs and percentage)
and discard rates for crabs caught during
the pot trials around South Georgia |
during the period March 2000 to April 2001.
Number
Number
% total crab catch
% total crab catch
Discard
kept
discarded
by number
by weight
rate
Paralomis formosa
22,803
300,660
70.1%
63.1%
93%
Paralomis spinosissima
3576
121,580
27.1%
35.4%
97%
Paralomis anamerae
0
12,370
2.7%
1.4%
100%
Neolithodes diomedeae
0
<0.1%
<0.1%
Lithodes murrayi
0
<0.1%
<0.1%
Table 3
The results of fitting generalized linear models on the probability of encountering crabs, and the catch rate of crabs (numbers per
pot) for nonzero catches. ANOVA results; the significance of adding each parameter to the general linear model is presented. For
the binomial model of the probability of encounter, a chi-squared significance test was used; for the Gaussian model of ln(CPUEl
an F test was used. Coefficients p,..£)j are given with standard error in parentheses. For those models including "area" the standard
case was for East South Georgia (EG), and the parameters for North South Georgia and Shag Rocks (NG, SR) are given. Analyses
were performed with Splus statistical software. Final models were constructed by using only the significant parameters (italicized
in this table), n = number of crabs in sample. Poly (1), Poly (2), and Poly (3) are the coefficients of each of the three orders in a 3rd
order polynomial.
Probability of encountering
Parameter
P. formosa
ANOVA results
n
101
Depth (m)
<0.001
Area
>0.05
Vessel
>0.05
Soak time (h)
>0,05
Coefficients P[ p^
n
101
Intercept
-2.2358(1.4373)
Depth (m)
0.0098(0.0035)
Area
Soak time (h)
Ln(CPUE) for nonzero
catches of P. formosa
Probability of encountering
P. spinosissima
Ln(CPUE) for nonzero
catches of P. spinosissima
91
<0.001
>0.05
>0.05
<0.05
91
1.7229(0.4301)
Poly(l)
Poly(2):
Poly(3):
4.6387 (5.3605)
-8.4647 (6.2257)
-1.3614 (4.4808)
101
<0.001
>0.05
>0.05
>0.05
101
8.5065(1.5822)
0.0106 (0.0022)
0.0447 (0.0221)
82
<0.001
<0.01
>0.05
<0.01
82
3.217(0.4877)
-0.0055 (0.0008)
NG: 1.1304(0.3241)
SR: 1.2771(0.2875)
0.0638(0.0191)
of bad weather, and to eliminate these the maximum
soak time was Hmited to 39 hours. The mid-depth of the
set (average of the depth at the start and end of the set)
was used to indicate setting depth, and the analysis was
restricted to sets whose depth range was less than 200 m.
The data from 45 sets (19% of total) were omitted from
analyses because they did not meet these criteria for area,
depth, or soak time.
For the binomial encounter models, the only significant
factor was depth (Table 3). F'or the Gaussian catch per pot
model, depth, and soak time were significant for both spe-
cies and area was significant for P. spinosissima. However,
the depth effects were opposite for the two species, so that
P. spinosissima decreased in abundance with depth and P.
formosa increased in abundance with depth, at least up to
about 1000 m depth (Fig. 3). There were insufficient data
from the cruises in 2001 to establish the effect of depths
shallower than 300 m and deeper than 1100 m, although
the limited sampling at depths greater than 1200 m in year
2000 (Fig. 1) suggests that catch rates of P formosa would
continue to decline at these depths, as suggested by the
generalized linear models in Figure 3.
The sex ratio of P. formosa was skewed towards males
in shallow water (<800 m) and females in deep water
(Fig. 4). The mean size of P. formosa of both sexes also
decreased significantly with an increase in depth (Fig.
5). Although catch rates in numbers were usually much
higher in deeper water, smaller crabs of no commercial
Purves et al.: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic
879
value often dominated catches. Only about 38% of
P. spinosissima sampled between 200 and 400 m
were males, whereas this proportion increased to
about 76% in depths of 600 to 800 m (Fig. 4). The
mean carapace width of males remained relatively
constant between 200 to 800 m, but female P. spi-
nosissima decreased in size with an increase in
depth (Fig. 5).
Size frequencies
Males achieved larger sizes than females in all
three species (Fig. 6). Only 5.7% of the sampled
P. spinosissima individuals were of legal size
(carapace width greater than 102 mm), of which
only 6% were females. A difference was also noted
in the percentage of legal-size P. spinosissima for
the different areas: 10% at South Georgia and
only 3.8% at Shag Rocks. For P. formosa only
11.6% (/! = 1012) were larger than the minimum
legal size of 90 mm. Of these legal crabs only 6%-
(n=63) were females, indicating that few females
would be processed if carapace width was the only
criterion used to select crabs that could be taken
legally. No obvious difference was noted in the
percentage of legal-size crabs caught in the two
main fishing areas.
Although a legal-size limit is not specified for P.
anamerae (sizes ranged from 39 to 96 mm), only
two crabs (0.2% of sample) were larger than the
legal limit for P. formosa (90 mm). A peak in the
length distribution of this species occurred be-
tween 55 and 57 mm, and few crabs were larger
than 77 mm.
Differences in the size-frequency distribution
of the sexes was more pronounced for P. formosa
than for the other two species; females peaked at
65-72 mm and males peaked at between 85 and
90 mm carapace width (Fig. 6). For P. spinosissima
female size distribution peaked between 78 and 82
mm and males peaked at 87 to 92 mm carapace
width (Fig. 6). Males and females of P. anamerae
had a relatively even size distribution up to cara-
pace widths of 65 mm. Most of the larger crabs
were males; only 6.8% of individuals larger than
65 mm were females. Maximum widths recorded
were 121 mm (P. spinosissima), 120 mm {P. for-
mosa), and 91 mm (P. anamerae).
Size at maturity
There was no significant difference between
female size at maturity at Shag Rocks and South
Georgia for either species (^-tests on Sm^g esti-
mated by fitting logistic models to the proportion
mature at size, P>0.05). The close allometric
relationship found between carapace width and
length (Table 4) made it possible to alter between these two
types of measurements (Table 5). Combining the two areas,
female size at maturity was 55.1 mm carapace length
200 300 400 500 600 700 800
Depth (m)
— P formosa - -*- - P spinosissima. EG
900 1000 1100 1200
.fi. ■ . P spinosissima. SR
Figure 3
General-linear-model-predicted CPUE (numbers per pot) for the
two species (Paralomis formosa and P. anamerae) at different depths
standardized for a soak time of 18 h (the average used in the study).
For P. spinosissima predictions at East South Georgia and Shag Rocks
are presented separately (EG, SR, respectively).
n = 1988
80
70 ■
60 -
50
40
30
20
10
% IVIales of P formosa
% IVIales of P spinosissima
200-400 400-600 600-800
800-
1,000
1 ,000-
1,200
1,200-
1,400
Depth range (m)
Figure 4
The percentage of males of Paralomis formosa and P. spinosissima
found at different depth ranges during random sampling of the crab
bycatch; all data from cruises Gl (Argos Georgia cruise 1), G2 (Argos
Georgia cruise 2), and H (Argos Helena cruise) were used.
(57.1 mm carapace width) for P. formosa and 61.2 mm
carapace length (67.7 mm carapace width) for P. spinosis-
sima (Fig. 7). These sizes are very similar to the 61.7 mm
880
Fishery Bulletin 101(4)
90 1
~ 80
70
65
60
P formosa M
PT?^~-^-.=^_^-- n=1987
P formosa F
^^^~-^^....^^^
P spinosissima M
,\
N
V
P spinosissima F
n= 1504
n= 1749
200-400 400-600 600-800 800-1.000
Depth range (m)
1,000-1.200
1.200-1,400
Figure 5
Mean carapace width of the males (M) and females (F) of P formosa and P. spinosissima
found at different depth distributions.
Table 4
Parameters of the
regression
carapace width -
-- carapace length x slope
-1- intercept (all ages and sexes combined), n
= number of
crabs in sample.
Species
Intercept
Slope
Correlation coefficient
n
P. spinosissima
6.457
0.994
0.977
367
P. formosa
1.976
1.032
0.969
351
P. anamerae
5.700
0.917
0.979
28
The size at maturity (Sm^^) for ma
are shown in parentheses.
Table 5
es and females of both Paralomis species. The standard errors (SE) for male carapace lengths
P. spinosissima P. formosa
Carapace length
(mm) Carapace width (mm) Carapace length (mm) Carapace width (mm)
Males 67.3(0.10)
Females 61.2
73.4 64.0(0.16) 68.0
67.7 55.1 57.1
carapace length reported by Otto and Macintosh (1996) for
P. spinosissima from Shag Rocks and South Georgia.
Unfortunately, owing to limited sampling of male crabs
around South Georgia, estimates of male sexual maturity
were only available for Shag Rocks. Too few samples off!
anamerae were available for determination of the onset
of female or male maturity in either area. Male size at
maturity (Sm^g) was determined at a carapace length of
67.3 mm (SD=2.3 mm derived from bootstrap resampling
with replacement) for P. spinosissima and at 64.0 mm (SD=
3.6 mm) for P. formosa (Fig. 8).
Crab survival rate
Approximately 5000 crabs were examined for carapace
damage on the cruises in 2001. The results (Table 6) indi-
cated that the level of visible damage to these crabs prior
to discarding was very low (2% of crabs of all species). The
vitality of these crabs was also assessed (according to the
index in Table 1). Most crabs were lively on arrival on deck
and prior to discard (Table 6). Differences were, however,
noted in both carapace condition and the vitality of crabs
between the two fishing vessels. Pairwise x^ comparisons
Purves et a\: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic
881
P spinosissima
P. anamerae
50
60
70 80 90
Carapace width (mm)
Figure 6
Carapace-width frequency distribution for the three Paralomis species (mm). (A) P.
spinosissima, (B) P. formosa, (C) P. anamerae. The minimum legal size (102 mm for P.
spinosissima and 90 mm for P. formosa) is indicated with a vertical line and arrow.
between the vitality indices of P. formosa and P. spinosis-
sima clearly showed that both species displayed signifi-
cantly lower vitality on the Argos Helena than on the Argos
Georgia (Table 7).
The processing environment of the two vessels may ex-
plain these differences. On the Argos Helena crabs were
likely to sustain more damage as pots were emptied down
a vertical chute before entering the processing area below
deck. On the Argos Georgia pots were emptied on a hori-
zontal conveyer belt leading to the factory. Interestingly,
there was no significant difference between the vitality
displayed by P. anamerae between the two vessels, al-
though this may be a result of the smaller sample size for
this species. For the Argos Georgia, the vitality of P. ana-
merae was significantly lower than for either P. formosa
or P. spinosissima.
The results of the three survival experiments are shown
in Table 8. Experiment 1 re-immersed only lively crabs
and included a control set of animals retained on deck
in a large tank for the same length of time as the re-im-
882
Fishery Bulletin 101(4)
1
0.9
0.8
0.7
0 6
0.5
0.4 -
0.3
0.2
0.1
0
1
0.9
0.8
0.7 ■
0.6
0.5
0.4
0.3
0.2
0.1
0
10 20 30 40 50 60 70 80 90 100 110 120
B
p. formosa
0 10 20 30 40 50 60 70 80 90 100 110
Carapace width (mm)
Figure 7
Size at female maturity for (A) P. spinosissima and (B) P. formosa. The size
at maturity U^q) carapace width was 67.7 mm and 57.1 mm for the two
species, respectively.
Table 6
The number and percentages of crabs of the different species
assessed for
carapace
condition and
vitality prior to being discarded
during 2001 iG2=Argos Georgia
cruise no
2, H=A
rgos Helena
).
Species and cruise
Carapace condition
Vitality ind
ex
Damaged
Undamaged
Total
Lively
Limp
Dead
Total
No.
%
No.
%
No.
%
No.
%
No.
%
Paralomis formosa
G2
48
1.9
2431
98.1
2479
2294
98.9
26
1.1
0
0.0
2320
H
23
2.7
814
97.3
837
745
97.4
19
2.5
1
0.1
765
Total
71
2%
3245
98%
3316
3039
98%
45
1%
1
0%
3085
Paralomis spinosissima
G2
17
1.5
1143
98.5
1160
1122
98.4
17
1.5
1
0.1
1140
H
18
2.0
874
98.0
892
767
94.8
39
4.8
3
0.4
809
Total
35
2%
2017
98%
2052
1889
97%
56
3%
4
0.2%
1949
Paralomis anamerae
G2
0
0.0
45
100.0
45
31
86.1
5
13.9
0
0
36
H
2
2.7
72
97.3
74
62
92.5
5
7.5
0
0
67
Total
2
2%
117
98%
119
93
90%.
10
10%
0
0%
103
Purves et al,: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic
883
Table 7
Results of
pairwise x'^ comparisons of the
vitality of crabs
on first being caught by different cruises (from Table 6). G2 = Argos
Georgia cruise no. 2 and H = Argos
Helena,
F = Rformosa.S
= P. spinosissima,A = P. anamerae. x'^ values are given, together with
significance, df = 2 for all except i
,alicized
results, when df
was 1 beca
use of the absence of anv dead crabs in the comparisons.
NS = not significant.
G2-F
H-F
G2-S H-S G2-A
G2-F
H-F
10.49, P<0.01
G2-S
2.9, NS
—
H-S
—
6.99, P<. 05
20.7,P<0.001
G2-A
35.2. P<0.001
—
29.2,P<0.001 —
H-A
—
5.5, NS
— 1.14, NS 0.49. NS
P. spinoxissima
mersion process. Out of the
35 lively control animals a
similar proportion died dur-
ing the experiment as in the
re-immersed group (8%), but a
lower proportion of the control
group were lively following the
experiment (63%). The lower
proportion of lively animals
in the control group may have
been a result of interruptions
in the supply of oxygenated
water and the continual dis-
turbance on deck due to the
ship's motion. Consequently,
controls were not performed in
experiments 2 and 3.
Experiments 2 and 3 took
a random sample of lively
and limp crabs and subjected them to re-immersion. The
proportions of lively crabs at the beginning of the re-im-
mersion experiments were lower than the proportions
estimated for the population as a whole in Table 6, with
the exception off! spinosissima in experiment 2. However,
sample sizes were much smaller in the experiments and the
crabs were subject to greater handling times than those as-
sessed in Table 6; therefore the results presented in Table 6
are more likely to be representative of the condition of
discarded crabs than results presented in Table 8.
In experiment 1 on cruise Gl, no limp animals were
subjected to re-immersion. In comparing the results of
immersing lively animals, there was no significant differ-
ence between the results for experiment 1 and 2 (cruises
Gl and G2 on the Argos Georgia) for either P. formosa or
P. spinosissima. Combining the results of the experiments
for Gl and G2, there was a significantly lower vitality
after re-immersion for both species on the Argos Helena
compared with the Argos Georgia (Table 7). On the Argos
Georgia there was no significant difference in vitality af-
ter re-immersion between P. formosa and P. spinosissima,
whereas on the Argos Helena vitality for P. spinosissima
was significantly lower than vitality for P. formosa. These
2 It
x; JO
TO
ii
c
_i
25
2
Ln {carapace length)
Figure 8
Chela allometry in male (A) P. spinosissima and (B) P. formosa sampled at Shag Rocks in
2001, The intersections of fitted curves were used to determine size at maturity.
results are similar to the initial assessment of vitality
prior to re-immersion (Table 6), where significantly fewer
P. spinosissima were lively in comparison with P. formosa
on the Argos Helena , but there was no difTerence between
the two species on the Argos Georgia.
A single re-immersion experiment performed on 15 P.
anamerae crabs (11 were "lively" and 4 were "limp"), on
the Argos Helena in April 2001, resulted in a mortality
rate of 73%. Only 27% of the crabs that were "lively" before
re-immersion were still "lively" after the rehaul. Although
more data need to be collected on the survival rate of this
species, this high mortality rate, together with the higher
incidence of individuals off! anamerae found to be "limp"
during vitality assessments (Table 8; 10% compared to 3%
of P. spinosissima and 1% of P. formosa), seems to indicate
that this bycatch species might be particularly vulnerable
to onboard handling and discarding.
Crabs that were physically damaged (i.e. had missing
legs or cracked carapaces) before being subjected to re-im-
mersion were less hkely to survive than undamaged crabs.
Of the 19 damaged P. spinosissima (13 of these were "lively"
and 6 were "limp"), 58% did not survive re-immersion and
only 32% of these were still "lively." The effect of damage
884
Fishery Bulletin 101(4)
Table 8
Results of survival-rate experiments. For
each experiment the number of re-immersions is given
together with the total
number of
crabs in lively and
imp condition that were re-immersed and their condition on
rehauling the c
•ab pots
after re
immersion.
Number of
Number of
re-immersions
crabs
Initial condition
Condition
after re-
mmersion
P. formosa
Experiment 1
Lively
Limp
Dead
(cruise Gl)
2
30
Lively
20
6
4
0
Limp
0
0
0
Experiment 2
Lively
Limp
Dead
(cruise G2)
6
98 (93%)
Lively
81
11
6
7
Limp
4
1
2
Experiment 3
Lively
Limp
Dead
(cruise H)
3
49 (91%)
Lively
27
21
1
5
Limp
0
2
3
P. spinosissima
Experiment 1
Lively
Limp
Dead
(cruise Gl)
2
42
Lively
35
3
4
0
Limp
0
0
0
Experiment 2
Lively
Limp
Dead
(cruise G2)
6
60(100%)
Lively
OLimp
55
0
5
0
0
0
Experiment 3
Lively
Limp
Dead
(cruise H)
10
167 (88%)
Lively
71
67
29
23
Limp
1
6
16
Table 9
Estimation of total survival rate for P. formosa and P. spinosissima based on results from re-immersion experiments. Because there
was no significant difference between the responses of lively animals between experiments 1 and 2, these data were pooled to give a
single estimate for the Ar^^os Georgia. On the Argos Georgia no Hnip P. spinosissima were encountered during the re-immersion exper-
iments; therefore p(?imp,/iiie/y) was set equal to p(limp,liuely) from experiment 3. Three calculations are presented: the first according
to the text description without the rehaul correction, the second including this rehaul, and the third where the original proportions of
lively anima\s pdiuely J were used from the experimental data in Table 8 rather than the data from the larger sample in Table 6.
pdiuely J from
Table 6, without
rehaul correction
pdivelyj from
Table 6, with
rehaul correction
pdivelyj from
Table 8, without
rehaul correction
pdivelyj from
Table 8, with
rehaul correction
P. formosa
Argos Georgia
78.7%
79.8%
77.5%
83.7%
Argos Helena
53.7%
56.2%
50.0%
58.4%
P. spinosissima
Argos Georgia
86.9%
88.5%
88.2%
88.2%
Argos Helena
40.5%
45.4%
37.9%
48.5%
was not so pronounced for P. formosa, although of the 12
crabs that were damaged prior to re-immersion (10 were
"lively" and 2 were "limp"), only 58% were "lively" upon
recovery. The mortality rate of 17% for damaged specimens
was more than double the 8% overall mortality found dur-
ing re-immersion experiments for this species. Most of the
dead crabs examined after re-immersion had been attacked
by isopods and amphipods and only the shell remained. It
is possible that these organisms were in fact responsible
for killing the crabs, particularly where damage to the shell
allowed access to the softer tissues of the crab.
Calculations of survival rate are given in Table 9 both
with and without the re-haul correction. As discussed
above, the more accurate estimate of lively ^^ is probably
from Table 6 because of the additional handling stress
associated with the experiment. However, Table 9 also
presents results obtained from data in Table 8 to estimate
this probability. The results suggest that the survival rate
Purves et al.: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic
885
of discarded crabs would be high on the Argos Georgia, be-
tween 77% and 88% for both P. formosa and P. spinosissima
(77-84% for the former, 87-88% for the latter). On the Ar-
gos Helena discard survival rate was much lower, 50-58%
for P. formosa and 38-49% for P. spinosissima.
Discussion
Crab species
Three previously unreported or rarely reported crab spe-
cies, Paralomis anamerae, NeoUthodes diomedeae (Bene-
dict, 1894), and Lithodes murrayi (Henderson, 1888), were
found in our study. The most abundant of these was P. ana-
merae, found at mid-range depths. The only other record of
P. anamerae is the original description of the species based
on four specimens obtained from the continental shelf of
Argentina at depths of 132-135 m (Macpherson, 1988).
The specimens obtained from South Georgia, at depths of
530-1210 m, therefore represent a considerable southerly
extension in the reported geographic distribution of this
species, as well as a notable increase in its bathymetric
range. Lopez Abellan and Balguerias (1994) reported both
P. spinosissima and P. formosa from a 1986-87 trawl survey
on the shelf, but no other species of crabs.
A certain amount of confusion surrounds the identifica-
tion of Paralomis species around South Georgia. Paralomis
aculeata is found in the CCAMLR database, but almost cer-
tainly because of its inclusion in the FAO species identifica-
tion guide for the Southern Ocean CArnaud 1985, in Fischer
and Hureau, 1985), attributed to Henderson (1888). This
species is not mentioned in Macpherson (1988), even as
a junior synonym. Conversely, none of the Paralomis spe-
cies identified in the present paper appear in Fischer and
Hureau (1985). It is not clear, therefore, which species of
Paralomis CCAMLR scientific observers have been identi-
fying as P. aculeata. Twenty-two specimens ofN. diomedeae
were collected at depths ranging from 420 to 1294 m. This
species has previously been recorded from South Georgia
(Macpherson, 1988).
Sixteen L. murrayi specimens were found on southeast
Shag Rocks and west South Georgia, approximately 60 nmi
apart, at depths of between 450 m and 605 m, Lithodes.
murrayi is mainly reported from the southern Indian
Ocean around Prince Edward, Crozet, and Possession
Islands, as well as Macquarie Islands, Mozambique Chan-
nel, and southern New Zealand at depths of 35-200 m
(Hale, 1941; Yaldwyn and Dawson, 1970; Arnauld, 1976;
Arnauld and Do-Chi, 1977; Kensley, 1977). However, it has
been reported in small numbers in CCAMLR statistical
catch records from 1993-94, 1997-98 and 1998-99 and
by CCAMLR observers (CCAMLRi). We have confirmed
the identification and the extension of the range of this
species to South Georgia. Klages et al. (1995) reported on
the distribution of L. murrayi off Peter Island, close to the
Antarctic continent between 180 and 260 m depth, and
a circum-Antarctic distribution has been claimed for the
species (Macpherson, 1988). The present study therefore
represents the greatest depth recorded for it.
Distribution
Catch rates of P. spinosissima encountered in our study
were lower than experienced in the September 1995-Janu-
ary 1996 crab fishery by the FV American Champion. Wat-
ters (1997) reported that average catch rates of legal-size
male P. spinosissima were between 14.2 and 28.4 males
per pot in the Shag Rocks and northwest South Georgia
areas (53.5-54°S, 37-40°W). Although the toothfish pots
operated by the Argos Georgia from March and May 2000
produced catch rates between 0.5 and 4 crabs/pot, the pro-
portion of legal-size males was very low (3%), resulting in
legal (retained) crab catch rates of less than 1/pot (note
that the retention rates given in Table 2 are lower than
those calculated from the length-frequency sampling. Fig.
6). Retention rates for the 1992 FV Pro Surveyor cruise
(July-August 1992) were 36% (Otto and Macintosh^). The
retention rates on the Argos Helena's experimental crab
fishery in 1999 were much lower than this (8% and 14%
for P. spinosissima and P. formosa respectively) (Purves^).
The low retention rate is most likely to be a consequence
of the pot design used on vessels G and H, where the col-
lapsible funnel entrances might have restricted the catch
to smaller size crabs.
A further feature of the American Champion crab fishery
was the restriction of fising effort to depths of less than 500
m. The present trials were targeted at toothfish rather than
crabs and were conducted according to an experimental
plan that distributed fishing effort over time, area, and
the full depth range of longlines used in the main toothfish
fishery. Accordingly, fishing occurred over a much wider
depth range than was used by the previous crab fisheries.
Our very high catch rates of P formosa in deep water were
therefore not reported by Watters (1997). However, even
our catch rates did not result in high numbers of retained
legal-size crabs on the Argos Georgia because the propor-
tion legally permitted was only 10.5%. Interestingly, even
in shallow water (400-800 m) P. formosa appeared to be
more common than P. spinosissima. Only in waters less
than 400 m deep did P. spinosissima become the dominant
species. This finding confirms the results of Watters (1997)
who found that P. spinosissima catch rates declined at
depths deeper than 300 m. Catch rates of P. spinosissima
were low even in these depths (5 crabs/pot). Only 9 of the
total of 110 sets of the Argos Georgia were conducted in
depths shallower than 400 m because the main target of
the fishery was toothfish.
The differences found in our study in the distribution
by depth of the different sexes and sizes of crabs might
■^ Otto, R. S., and R. A. Macintosh. 1992. A preliminary report
on research conducted during experimental crab fishing in
the Antarctic during 1992 (CCAMLR Area 48). Document
WG-FSA-92/29, 20 p. CCAMLR, RO. Box 213, North Hobart,
Tasmania, Australia.
* Purves. M. G. 1999. Report of the South African designated
CCAMLR observer on board the British registered longliner
"Argos Helena" in Statistical Subarea 48.3, 31 August to 23
September 1999, 13 p. CCAMLR, RO. Box 213, North Hobart,
Tasmania, Australia.
886
Fishery Bulletin 101(4)
indicate that recruitment of P. formosa takes place in
deeper water. This conclusion is based on a decrease in
size with increasing depths, the higher proportion of fe-
males encountered in deeper water, and increasing crab
densities at depth. For P spinosissima this trend was not so
pronounced, although crabs of this species were also gener-
ally of a smaller size in deeper water However, contrary to
the case with P. formosa, females were more prevalent in
shallower water. Very few females of P. spinosissima were
encountered in deep water These rather unusual findings
might suggest ecological partitioning of the benthic habi-
tat, and warrant further investigation.
Another unexpected result of our work was the discovery
of a third species of Paralomis, P. anamerae, at intermedi-
ate depths. This species was apparently not present in the
American Champion or Pro Surveyor catches, presumably
again because of the depth restriction in these cruises.
Maturity
For P. spinosissima our estimate of 67.3 mm carapace
length at 50% male maturity is similar to the 66.4 mm
carapace length found by Otto and Macintosh (1996) at
Shag Rocks for this species. Unfortunately, as can be seen
from Figure 8, relatively few small P. formosa males were
encountered and size at maturity for this species (64.0 mm
carapace length) is likely to have been poorly estimated in
our analysis. However, if it is assumed that male P. formosa
mature at the same size in relation to female P. formosa,
as in the case of P. spinosissima, the female maturity data
presented in Table 5 would suggest that male P. formosa
would mature at 60.4 mm carapace length (64.3 mm cara-
pace width) rather than the 64.0 mm shown in Figure 8.
Watters and Hobday (1998) have also examined size at
maturity for P. spinosissima and P. formosa, although their
samples were taken from South Georgia rather than Shag
Rocks. Using a method based on finding the maximum
of the second derivative of smoothing spline fits to chela
height and carapace width data, they found that size at
morphometric maturity for P. spinosissima was 73 mm
carapace length. This size is similar to that which Otto
and Macintosh (1996) obtained for P. spinosissima at South
Georgia using the same technique as we did. Watters and
Hobday's (1998) results for P. formosa are, however, for a
higher size at maturity (80 mm carapace length) than that
for P. spinosissima, which would seem to be at variance
with our results and the apparent relative sizes of the two
species (see Fig. 6 and CCAMLR^).
The minimum size limits for crabs at South Georgia were
set by CCAMLR in 1992 but, in common with many crab
stocks (Schmidt and Pengilly, 1993), these measures were
not accompanied by rigorous analysis of the effectiveness
of these measures in meeting management objectives. For
P. spinosissima. Otto and Macintosh's^ male maturity re-
sults for P. spinosissima were used, and allowing males
at least one opportunity to breed and an assumed growth
per moult of 15%, minimum size limits were calculated as
94 mm and 84 mm carapace width at South Georgia and
Shag Rocks, respectively (CCAMLR^). These results are
very similar to our own, but the CCAMLR limit of 102 mm
width was based on the then-existing processing require-
ments rather than on these calculations. Our results sug-
gest, allowing at least one opportunity to breed, that the
limit should be 83 mm for P. spinosissima. For P. formosa,
taking our more conservative figure of 64.0 mm carapace
length at 50% maturity, the catch size limit should be set
at 78 mm carapace width (the less conservative figure, 60.4
mm carapace length, would suggest a size limit of 74 mm
carapace width).
Hoggarth (1991) reviewed minimum size limits for a
number of stocks of lithodid crabs and found that mini-
mum legal sizes were about 70% of the maximum size for
males, which would suggest 85 mm and 84 mm carapace
width for P. spinosissima and P. formosa, respectively. It
should, however, also be taken into account that these
estimates were probably biased because of the greater
sampling effort made at Shag Rocks. Note that the length-
frequency distribution for P. formosa in Figure 6B appears
to indicate a lower maximum size for males of this species
than for males of P. spinosissima. However, the largest
P formosa actually encountered was 120 mm carapace
width. Furthermore, Figure 6B seems to be truncated at
the larger sizes, suggesting perhaps that a proportion of
the large adult population was not encountered during
fishing.
Discard mortality
Our results demonstrate that, although a high proportion of
crabs is likely to survive the physical strain of being hauled
to the surface from potentially great depths, some under-
size individuals and nontarget females can be expected
to die following discarding. The most significant factor
affecting discard survivorship was handling on board the
vessel. On the Argos Georgia, where crabs were unloaded
from pots and sorted on a conveyor belt, survivorships were
high, up to 88%, and P. spinosissima survived better than
P. formosa. By contrast, on the Argos Helena, where crabs
went down a chute prior to processing, survival rate was
between 38% and 58% and P formosa survived consider-
ably better than P spinosissima. In general, P. anamerae
was the most vulnerable species, followed by P spinosis-
sima, and the least vulnerable — P formosa.
Studies of the discard mortality of lithodid crabs in
North Pacific fisheries have produced a variety of results.
Stevens ( 1990) found that crabs discarded from commercial
sole trawls suffered high mortalities (47.3%), but Byersdor-
fer and Watson'' and Zhou and Shirley ( 1995) both reported
relatively low mortalities (<2%) resulting from handling
when fishing with pots. Our results support these previous
studies and extend them to the Antarctic, clearly indicating
that where handling on a pot vessel is reduced, mortali-
ties are relatively low (<15% mortality). When crabs are
' Byersdorfer, S., and L. J. Watson. 1992. A summary of bio-
logical data collected during the 1991 Bristol Bay red king crab
tagging study. Technical Fishery Report 92-14, 30 p. Alaska
Department of Fish and Game, Division of Commercial Fisheries,
P.O. Box 25526, Juneau, AK 99802-5526.
Purves et al.: Distribution, demography, and discard mortality of crabs caught as bycatch in the South Atlantic 887
handled on a pot vessel, as they would be on a trawl
vessel, mortalities are higher.
Other factors, which could not be tested in the
re-immersion experiment, may also affect the rate
of crab survival. We re-immersed crabs in pots,
whereas normally they would be simply dropped
into the sea and would be subject to predation from
birds and fish before they reached the bottom. The
effect of discarding crabs away from their original
habitat is unknown, but our results demonstrate
a clear depth separation between the two species;
therefore one would expect at least an energetic
cost if crabs have to relocate. The crabs subjected
to re-immersion experiments were sampled imme-
diately before being discarded. They might suffer
further damage through the actual discard process;
for instance Stevens (1990) speculated that, while
traveling through offal chutes, they could become
entangled in machinery or suffer further damage
upon impact with the surface of the water Ide-
ally crabs should have been sampled after being
through the full discarding procedure, but this was
not practical. Finally, eggs often became dislodged
during handling and this loss possibly impacted
reproductive success.
Zhou and Shirley (1995) presented results that indi-
cate that there are no long-term effects of handling on
crab survival, feeding rate, or crab condition; therefore
we might reasonably expect that the survival rates seen
in our experiments would also be the relevant long-term
survival rates. However, even with relatively low discard
mortality, the impacts of repeated catching and discarding
of individuals will have a cumulative effect on crab popula-
tions. Both retained and discarded bycatch should there-
fore continue to be reported and be incorporated into crab
population models. The presence of such a large discarded
bycatch might provide the opportunity for the retention
(and removal from the population) of parasitized crabs, as
suggested by Basson (1994).
Crabs are an inconvenience in a fishery targeting tooth-
fish. In situ observations made during the AUDOS ex-
periments on the UK's January 2000 survey confirm that
toothfish seek to avoid direct contact with crabs (Yau et al.,
2002), although crabs do form a component of their food
(Pilling et al., 2001). An inverse relationship was found
in the present study between toothfish numbers in pots
and crab numbers in pots, suggesting toothfish avoid pots
with large crab populations (Fig. 9). Therefore, conducting
the toothfish pot fishery in an area of low crab abundance
is sensible, and our data do suggest that, at intermedi-
ate depths, the crab catch should be low and composed
primarily of large P. formosa. Avoidance of areas of high
crab bycatch will also reduce the mortality associated with
discarding female and undersize male crabs. These discard
levels are very high (>93%) — considerably higher than
those in the Bering Sea (85%: Stevens, 1996). Such high
discard levels could be reduced further by developing new
pot designs to limit crab catches to larger, legal-size ani-
mals— for instance designs with excluder panels (Stevens,
1996) — or perhaps by reducing the minimum size limit.
100
A
80
\
60
\
40 -
\
20 -
n
■.
■iW^
14
12 ro
u
10 o
D Frequency of
number/pot
toothfish
■ Number/pot
P spinosissima
0-1
1-2
2-3
CPUE (number/pot) of toothfish
Figure 9
The frequency distribution of the number of toothfish caught per
pot compared with the CPUE's of Paralomis spinosissima and
P. formosa (numbers/pot) and of toothfish {Dissostichus eleginoides)
catches (numbers/pot), as found during the first cruise of the Argos
Georgia during March to May 2000.
Acknowledgments
We wish to acknowledge the assistance of Argos Ltd. and
the excellent cooperation of the captains of the two ves-
sels. Captain Joaquin Abraldes Gonzalez and Captain Jose
Andres Sampedro and their crew. Permission to publish
these data was kindly granted by Argos Ltd. The Govern-
ment of South Georgia and the South Sandwich Islands
funded the research by MRAG Ltd. on South Georgia
fisheries.
Literature citations
Agnew, D. J., T. M. Daw, M. G. Purves, and G. M. Pilling.
2001. Fishingfor toothfish using pots: results of trials under-
taken around South Georgia, March-May 2000. CCAMLR
(Commission for the Conservation of Antartic Marine
Living Resources) Sci, 8:93-105.
Agnew, D. J., L. Heaps, C. Jones, A. Watson, K. Berkieta, and
J. Pearce.
1999. Depth distribution and spawning pattern of Dissosti-
chus eleginoides at South Georgia. CCAMLR Sci. 6:19-36.
Arnauld, P. M.
1976. Peches experimentales de Lithodes murrayi Hender-
son, 1888 (Crustacea, Anomura) aux iles Crozet (SW de
I'Oceanlndien). Thetys 3:167-172.
Arnauld, R M., and T. Do-Chi.
1977. Donnees biologiques et biometriques sur les lithodes
Lithodes murrayi (Crustacea: Decapoda: Anomura) des lies
Crozet (SW de I'Ocean Indien). Mar Biol. 39:147-159.
Basson, M.
1994. A preliminary investigation of the possible effects of
rhizocephalan parasitism on the management of the crab
fishery around South Georgia. CCAMLR Sci. 1:175-192.
Benedict, J. E.
1894. Descriptions of new genera and species of crabs of
the family Lithodidae, with notes on the young oi Lithodes
888
Fishery Bulletin 101(4)
camtschaticus and Lithodes brevipes. Proc. U.S. Natl. Mus.
17:479-488.
Fischer, W., and J. C. Hureau (eds.)
1985. FAO species identification sheets for fishery purposes,
Southern Ocean, vol. l.,p. 232. FAO (Food and Agriculture
Organization of the United Nations), Rome, Italy.
Hale, H. M.
1941. Decapod Crustacea. Br. Aust. N.Z. Antarct. Res.
Exped. 1929-1931. 48:257-285.
Henderson, J. R.
1888. Report on the Anomura collected by H.M.S. Challenger
during the years 1873-76. Rep. Sci. Res. Voy H.M.S. Chal-
lenger, Zool. 27(I-vii): 1-221, plates 1-21.
Hoggarth, D. D.
1991. An ecological and economic assessment of the Falk-
land Islands inshore crab, Paralomis granulosa. Ph.D.
diss., 312 p. Imperial College, Univ. London, London. UK.
Kensley, B.
1977. The South African Museum's Meiring Naude cruises.
Part 2. Crustacea, Decapoda, Anomura and Brachyura.
Ann. S. Afr. Mus. 72:161-188.
Klages, M., J. Gutt, A. Starmans, and T. Bruns.
1995. Stone crabs close to the Antarctic continent: Lithodes
murrayi Henderson, 1888 (Crustacea; Decapoda; Ano-
mura) off Peter I Island (68°51-S, 90°51-W). Polar Biol.
15:73-75.
Lopez-Abellan, L, J., and E. Balguerias.
1994. On the presence of Paralomis spinosissima and
Paralomis formosa in catches taken during the Spanish
survey Antartida 8611. CCAMLR Sci. 1:165-173.
Macpherson, E.
1988. Revision of the family Lithodidae Samouelle, 1819
(Crustacea, Decapoda, Anomura) in the Atlantic Ocean.
Mongr. Zool. Mar 2:9-153.
Otto, R. S., and R. A. Macintosh.
1996. Observations on the biology of the Lithodid crab
Paralomis spinosissima from the Southern Ocean near
South Georgia. In Proceedings of the international sympo-
sium on biology, management and economics of crabs from
high latitude waters. Anchorage, Alaska, October 1995, p.
627-647. Alaska Sea Grant, Fairbanks, AK.
Pilling, G. M., M. G. Purves, T. M. Daw, D. J. Agnew, and
J. C. Xavier.
2001. The stomach contents of Patagonian toothfish
around South Georgia (South Atlantic). J. Fish Biol. 59:
1370-1384.
Stevens, B. G.
1990. Survival of king and tanner crabs captured by com-
mercial sole trawls. Fish. Bull. 88:731-744.
1996. Crab bycatch in pot fisheries: causes and solutions. In
Proceedings of the solving bycatch workshop, September
25-27, 1995, Seattle, Washington (TWray,ed.), p. 151-158.
Alaska Sea Grant, Fairbanks, AK.
Schmidt, D. C, and D. Pengilly
1993. Review of harvest strategies used in the management of
lithodid crab in Alaska. In Proceedings of the international
symposium on management strategies for exploited fish pop-
ulations: October 21-24, 1992, Anchorage, Alaska (G. Kruse,
D. M. Eggers, R. J. Marasco, C. Pautzke, and T. J. Quinn, eds.),
p. 385-407. Alaska Sea Grant, Fairbanks, AK.
Somerton, D. A., and R. S. Otto.
1986. Distribution and reproductive biology of the golden
king crab, Lithodes aequispina, in the eastern Bering Sea.
Fish. Bull. 84:571-584.
Watters, G.
1997. Preliminary analyses of data collected during experi-
mental phases of the 1994/95 and 1995/96 antarctic crab
fishing seasons. CCAMLR Sci. 4:141-159.
Watters, G., and A. J. Hobday.
1998. A new method for estimating the morphometric size at
maturity of crabs. Can. J. Fish. Aquat. Sci. 55:704-714.
Yaldwyn, J. C, and E. W. Dawson.
1970. The stone crab Lithodes murrayi Henderson: the
first New Zealand record. Rec. Dom. Mus. (Wellingt.) 6:
275-284.
Yau, C, M. A. Collins, P M. Bagley, I. Everson, and I. G. Priede.
2002. Scavenging by megabenthos and demersal fish on the
South Georgia slope. Antarct. Sci. 14:16-24.
Zhou, S., and T C.Shirley
1995. Effects of handling on feeding, activity and survival of
red king crabs Paralithodes camtschaticus (Tilesius, 1815).
J. Shellfish Res. 14:173-177.
889
Abstract— Sea turtles are subjected to
involuntary submergence and potential
mortality due to incidental capture by
the commercial shrimp fishing indus-
try. Despite implementation of turtle
excluder devices (TEDs) to reduce at-
sea mortality, dead stranded turtles
continue to be found in near-record
numbers along the coasts of the west-
ern Atlantic Ocean and northern Gulf
of Mexico. Although this mortality may
be due to an increase in the number of
turtles available to strand, one alterna-
tive explanation is that sea turtles are
repetitively submerged (as one fishing
vessel follows the path of another) in
legal TEDs. In the present study, labo-
ratory and field investigations were
undertaken to examine the physiologi-
cal effects of multiple submergence of
loggerhead sea turtles iCaretta caretta ).
Turtles in the laboratory study were
confined during the submersion epi-
sodes, whereas under field conditions,
turtles were released directly into
TED-equipped commercial fishing
nets. Under laboratory and field condi-
tions, pre- and postsubmergence blood
samples were collected from turtles
submerged three times at 7.5 min per
episode with an in-water rest interval
of 10, 42, or 180 min between submer-
gences. Analyses of pre- and postsub-
mergence blood samples revealed that
the initial submergence produced a
severe and pronounced metabolic and
respiratory acidosis in all turtles. Suc-
cessive submergences produced sig-
nificant changes in blood pH, Pco„ and
lactate, although the magnitude of the
acid-base imbalance was substantially
reduced as the number of submergences
increased. In addition, increasing the
interval between successive submer-
gences permitted greater recovery of
blood homeostasis. No turtles died
during these studies. Taken together,
these data suggest that repetitive sub-
mergence of sea turtles in TEDs would
not significantly affect their survival
potential provided that the animal has
an adequate rest interval at the surface
between successive submergences.
The physiological effects of multiple forced
submergences in loggerhead sea turtles
(.Caretta caretta)
Erich K. Stabenau
Kimberly R. N. Vietti
Department of Biology
Bradley University
1501 W Bradley Ave.
Peoria, Illinois 61625
E-mail address (for E, K. Stabenau): eks@bradleyedu
Manuscript approved for publication
25 March 2003 by Scientific Editor.
Manuscript received 26 June 2003
at NMFS Scientific Publications Office.
Fish. Bull. 101:889-899 (2003).
The five sea turtle species inhabiting
the waters of the U.S. Gulf of Mexico
and Atlantic Ocean are considered to
be threatened or endangered. One con-
tributing factor to sea turtle mortality
is incidental capture in the nets of com-
mercial shrimping vessels. The National
Research Council's Committee on Sea
Turtle Conservation (1990) suggested
that as many as 5500 to 55,000 log-
gerhead (Caretta caretta) and Kemp's
ridley (Lepidochelys kempi) sea turtles
were killed annually during shrimp-
ing-related activities. More recently,
two independent studies statistically
confirmed the relationship between
shrimping activity and the appearance
of stranded sea turtles in the U.S. Gulf
of Mexico and the Atlantic Ocean (Cail-
louet et al., 1991; Crowder et al., 1995).
Because of the impact of trawl-related
mortality on sea turtle populations, the
U.S. government passed regulations
in 1987 requiring that commercial
shrimping vessels pull nets equipped
with certified turtle excluder devices
(TEDs). TEDs are designed to exclude
any turtle that may enter into shrimp-
ing nets, while not affecting the catch of
the target species. Crowder et al. (1995)
reported that the sea turtle population
off the coast of South Carolina contin-
ued to decline when TED regulations
were implemented; however, the rate
of decline decreased significantly after
full-time TED use.
In spite of the TED regulations,
near-record numbers of dead stranded
sea turtles have been found on U.S. Gulf
of Mexico and Atlantic Ocean beaches
(Shaver-Miller'). Although there may
be other man-related or natural causes
for this continued sea turtle mortality,
there are two plausible reasons for the
increased mortality during shrimping
activities. First, commercial shrimp
fishermen generally do not carry le-
gally certified TEDs in their trawl
nets and the TEDs that are used are
often installed incorrectly or purposely
sewn shut. Second, the shrimp fisher-
men may pull legal TEDs; however,
the turtles are repetitively submerged
as they are caught in the TEDs of ves-
sels that follow each other These suc-
cessive submergences may exacerbate
the physiological effects experienced
by sea turtles during a forced submer-
sion, and thus, may limit their survival
potential.
Sea turtles spend approximately
99% of their time under the surface
of the water. During the brief period
at the surface, the turtle will exhale
and inhale a solitary breath and then
dive under the surface (Jackson, 1985).
In fact, multiple breaths by sea turtles
are generally seen only after prolonged
dives. Minimal information is available
on the physiological effects of forced
submergences of sea turtles. It has been
suggested that voluntary dives by sea
turtles are aerobic in nature (Wood et
al., 1984), whereby oxygen availability
minimizes the metabolic production of
lactic acid. The turtles may accumulate
carbon dioxide, resulting in a respira-
Shaver-Miller, D. 2002. Personal com-
mun. Texas coordinator, Sea Turtle
Stranding and Salvage Network, USGS,
Corpus Christi, Texas 78406.
890
Fishery Bulletin 101(4)
tory acidosis that is ameliorated by hyperventilation at the
surface. Therefore, voluntary diving in the absence of any
other external stressor does not limit sea turtle survival
potential.
In contrast, forced submergence of Kemp's ridley and log-
gerhead sea turtles produces significant blood respiratory
and metabolic derangements. Stabenau et al. ( 1991) report-
ed that forced submergence of Kemp's ridley sea turtles for
less than 7.5 min in shrimp nets equipped with TEDs re-
sulted in significant increases in blood lactic acid and PcOj,
and decreases in blood pH. Moreover, several hours were
required for these turtles to fully recover blood homeostasis
(National Marine Fisheries Service, unpubl. data^). How-
ever, the study by Stabenau et al. ( 1991) did not address the
physiological effects of multiple forced submergences of sea
turtles. It is plausible that repeated submergence induces
progressive, significant blood acid-base disturbances, and
limits sea turtle survival potential. Therefore, the present
study examined the physiological effects of multiple forced
submergences on loggerhead sea turtles.
This investigation was divided into two phases. First,
a laboratory component was conducted to examine the
feasibility of a multiple submergence study. This phase of
the research permitted characterization of the magnitude
of the acid-base disturbance under controlled conditions.
Second, a field investigation was conducted to expose tur-
tles to TED-equipped commercial fishing nets. Data from
these studies may offer greater insight into potential sea
turtle mortality caused by multiple capture in commercial
shrimping nets carrying legal TEDs.
Materials and methods
Laboratory study
Thirty-nine headstarted 2-year-old loggerhead sea turtles
reared in captivity at the National Marine Fisheries Ser-
vice (NMFS) Galveston Laboratory were used in this phase
of the study. Each turtle was randomly placed into experi-
mental (submerged, 37.0 ±0.2 cm, 6.51 ±0.06 kg, n=21) or
control (nonsubmerged, 36.9 ±0.2 cm, 6.45 ±0.10 kg, n=18)
treatments. All turtles were of comparable size and weight
and therefore any alterations in blood parameters between
experimental and control turtles represented treatment
effects rather than size effects. It should be noted that the
turtles used in our study were representative of the aver-
age size of dead stranded turtles and those animals used
in annual TED certification tests.
The study was initiated by collecting presubmergence
blood samples from the experimental turtles immediately
prior to their individual confinement in a weighted canvas
bag. Each turtle was then submerged for 7.5 min in sea-
water filled tanks. Postsubmergence blood samples were
collected within 30 s of bringing the turtle out of the water
to minimize blood acid-base changes. Following an in-water
2 National Marine Fisheries Service. 1994. Unpubl. data.
(Available from E. K. Stabenau, Bradley University, 1501 W.
Bradley Ave., Peoria, IL 61625]
rest interval of 10 (treatment 1), 42 (treatment 2), or 180
(treatment 3) min, a presubmergence blood sample was
collected and the turtle was submerged a second time. A
postsubmergence blood sample was then collected imme-
diately upon surfacing. The turtle was then submerged a
third time, following the same rest interval between the
first and second submergence episodes, and pre- and post-
submergence blood samples were collected as described
above. The seventh serial blood sample was collected 180
min after the final submergence in all turtles. Blood sam-
ples were also collected from control turtles over the same
time intervals to ensure that repetitive handling and blood
sampling did not alter blood homeostasis. All blood samples
were collected into heparinized vacutainers from the dorsal
cervical sinus as described by Owens and Ruiz (1980). No
more than 4-6% of blood volume was collected during the
serial sampling to minimize potential physiological effects
associated with blood volume depletion.
Field study
Thirty-six headstarted 2-year-old loggerhead sea turtles
reared in captivity from the NMFS Galveston Laboratory
were used in this phase of the study. The turtles were trans-
ported from Galveston, TX, to Panama City, FL, where they
were placed into large pens in St. Andrews Bay. The sub-
mergence study was initiated after a minimum of 21 days
of natural conditioning in the in-water pens. Each turtle
was randomly placed into experimental (submerged, 35.9
±0.2 cm, 6.77 ±0.09 kg, n=24) or control (nonsubmerged,
35.4 ±0.3 cm, 6.46 ±0.12 kg, n=12) treatments. As in the
laboratory study, all experimental and control turtles were
of comparable size and weight.
The study was initiated by collecting presubmergence
blood samples from the experimental turtles immediately
prior to their individual confinement in a weighted mesh
bag. Each turtle was then submerged using the standard
protocol for TED certification tests. Briefly, the mesh bag
containing a turtle was placed onto a line connecting the
trawl vessel to the headrope on the shrimp net. Divers then
released the turtle (without handling the animal) into the
mouth of the trawl. Often, turtles were observed vigorously
swimming in the trawl until being overcome by the net.
Although the shrimp net was equipped with a TED, divers
held the escape door closed for 5 min. The turtle was then
permitted to leave the trawl and surface. Thus, the total
submergence time was approximately 7.5 min, including
the time for the weighted mesh bag containing the turtle
to reach the headrope for release into the trawl, the 5 min
within the trawl, and the time for the turtle to surface.
Turtles were immediately captured at the surface and
returned to the trawl vessel for postsubmergence blood
sampling. Typically, postsubmergence blood samples were
collected within 1-2 min of the turtle surfacing. Following
a rest interval of 10 (treatment 4), 42 (treatment 5), or 180
(treatment 6) min in water-filled containers on the trawl
vessel, a presubmergence blood sample was collected and
the turtle was submerged a second time. A postsubmer-
gence blood sample was then collected immediately upon
surfacing. The turtle was then submerged a third time,
Stabenau and Vietti: The physiological effects of multiple forced submergences of Caretta caretta
891
following the same rest interval between the first and
second submergence episodes, and pre- and postsubmer-
gence blood samples were collected as described above. A
seventh serial blood sample was collected 180 min after
the final submergence in all turtles. Blood samples were
also collected from nonsubmerged control turtles over the
same time intervals to ensure that repetitive handling and
blood sampling did not alter blood homeostasis. The blood
sampling technique and volume collected was identical to
that described for the laboratory component of the study.
Blood and plasma analyses
In the laboratory study, blood PcOj and pH were analyzed
immediately following collection by using a clinical blood
gas analyzer with electrodes thermostatted at 37°C. Both
variables were corrected to turtle cloacal temperature
using requisite correction factors for sea turtle blood and
plasma (Stabenau and Heming, 1994). In the field study,
blood gases (Po., and PcOr,) and pH were analyzed on the
trawl vessel immediately following collection using a
blood gas analyzer with electrodes thermostatted to turtle
body temperature (27-28. 5°C). The remaining analyses
were comparable for both the laboratory and field compo-
nents of the submergence study. Packed red cell volume
(hematocrit) was determined by following centrifugation
of heparinized microcapillary tubes. Two hundred micro-
liters of whole blood were then added to 10% trichloro-
acetic acid for lactate analysis. The deproteinized samples
were centrifuged, and the supernatant removed and stored
at -70°C. Lactate was determined spectrophotometrically
by using standard enzymatic techniques (Sigma, kit 826-
B, Saint Louis, MO). The remaining whole blood was then
centrifuged, the plasma removed and stored at -70°C.
Plasma Na+ and K* were measured with flame photometry
( Jenway, model PFP7, Essex, England), and plasma CL was
determined with electrometric titration (Haake-Bucher,
model 4425000, Saddle Brook, NJ). Plasma glucose was
measured spectrophotometrically (Sigma, kit 16-20), and
plasma osmolality was determined with a vapor pressure
osmometer (Wescor, model 5500, Logan, UT). For the labo-
ratory study, plasma norepinephrine was analyzed with
HPLC (BAS, model LC-300, West Lafayette, IN).
All data are expressed as means ±SE. Where appropri-
ate, the data was analyzed with one-way ANOVA. Post-hoc
comparisons between means were analyzed with Tukey's
multiple comparison test. A fiduciary level of P<0.05 was
regarded as significant.
Results
Blood pH, PcOj, and lactate
The initial submergence of loggerhead sea turtles under
laboratory and field conditions produced a dramatic and
severe acidosis in all experimental turtles. Blood pH fell
an average of 0.54 ±0.03 (range 0.49 to 0.59 pH units)
and 0.63 ±0.06 (range 0.53 to 0.73 pH units) m labora-
tory turtles and field turtles, respectively, following ini-
tial submergence (Figs. lA and 2A). The blood acidosis
was derived from respiratory and metabolic components
as evident from a positive proton-lactate deficit (Buffer
capacityx4pH-/l[Iactatel), and from significant increases
in blood PcOj and lactate (Figs. 1 and 2). The initial sub-
mergence also produced significant decreases in blood Pog
and increases in plasma norepinephine (P<0.05, n=24 for
Po., and n = ll for norepinephrine). In contrast, minimal
changes in blood pH, PcOg, and lactate were observed fol-
lowing collection of the first two blood samples in nonsub-
merged control turtles (Figs. 1 and 2).
Recovery of the respiratory and metabolic derangements
in submerged turtles was dependent on the interval be-
tween successive submergences. A 10-min in-water rest
interval between the first and second submergence (treat-
ment-! and -4 turtles) permitted partial recovery of blood
pH (Figs. lA and 2A) and Pco,, (Figs. IB and 2B), but blood
pH remained significantly different from presubmergence
values. Washout of additional lactate was also detected in
these animals, whereby blood lactate concentration in-
creased higher than the postsubmergence value (Figs. IC
and 2C ). Turtles with a 42-min surface interval (treatment-
2 and -5 turtles) between the first and second submergence
had partial to complete recovery of blood pH (Figs. lA and
2A), complete recovery of blood Pco.^ (Figs. IB and 2B), and
slight recovery of blood lactate (Figs. IC and 2C). Only
the blood lactate remained significantly different from
the initial presubmergence value after the 42-min rest in-
terval. Turtles with a 180-min in-water recovery interval
(treatment-3 and -6 turtles) showed complete recovery of
blood pH and PcOr,, although the lactate concentration was
slightly higher than baseline levels (Figs. 1 and 2). Blood
Pog and plasma norepinephrine recovered completely re-
gardless of the surface interval (P>0.05, n=24 and n = ll
for Poj and norepinephrine, respectively). Nonsubmerged
control turtles in the laboratory and the field exhibited few
significant changes in blood pH, Pcoo, or lactate, whether
the interval between the second and third serial blood
sample was 10, 42, or 180 min (Figs. 1 and 2).
The second 7.5-min submergence produced a drop in
blood pH and an increase in Pco,^ (Figs. 1 and 2) in all of the
experimental animals, and significant differences occurred
in treatment 2-6 turtles. It is noteworthy, however, that
the severity of the acid-base imbalance was not as drastic
as the acidosis measured following the first submergence.
The mean pH difference (4pH) between the second pre- and
postsubmergence ranged from 0.11 and 0.16 in treatment-1
and -4 turtles (animals with a 10-min interval between
submergences), respectively, to 0.50 in treatment-3 turtles
and 0.66 in treatment-6 turtles (animals with a 180-min
interval between submergences). The acidosis in treat-
ment-1 and -4 turtles resulted, in part, from the continual
elevation in blood lactate. In contrast, the longer surface
interval between the two submergence episodes resulted
in enhanced recovery of acid-base variables. Therefore,
the turtles with a surface interval of 42 or 180 min had in-
creased production of CO2 and lactate in relation to turtles
with a brief surface interval (Figs. 1 and 2). Comparable
changes in blood Poj and norepinephrine were measured
following the second submergence (P<0.05, n=24 and n=9
892
Fishery Bulletin 101(4)
Blood collection
Figure 1
Blood pH (A), PcOj (B), and lactate (C) measured prior to and after three succes-
sive forced submergence episodes in loggerhead sea turtles in the laboratory. Blood
collection 1, 3, and 5 are presubmergence samples, whereas blood collection 2, 4,
and 6 are postsubmergence samples. Blood collection 7 was taken 180 min after
the final submergence. The surface interval between the submergences was 10
min (T), 42 min (•), or 180 min (■). Data from control sea turtles (A) are shown
for comparison.
for Po.2 and norepinephrine, respectively). Collection of the
fourth sample from nonsubmerged control turtles revealed
no significant changes in blood pH, PCO2, or lactate when
compared to the third sample (Figs. 1 and 2).
The remaining serial blood samples revealed comparable
patterns in the blood pH, Pco.,, Po,, lactate, and norepi-
nephrine. Turtles given a longer rest interval at the surface
(after the second submergence^ had enhanced recovery of
Stabenau and Vietti The physiological effects of multiple forced submergences of Caretto caretta
893
Blood collection
Figure 2
Blood pH (A), PcOj (B), and lactate (C) measured prior to and afler three succes-
sive forced submergence episodes in loggerhead sea turtles in TED-equipped nets.
Blood collection 1, 3, and 5 are presubmergence samples, whereas blood collection
2, 4, and 6 are postsubmergence samples. Blood collection 7 was taken 180 min
after the final submergence. The surface interval between the submergences was
10 min (T), 42 min (•), or 180 min (■). Data from control turtles (A) are shown
for comparison.
blood acid-base variables, whereas a brief surface interval
permitted minimal recovery of blood homeostasis (Figs. 1
and 2). Submersion of experimental turtles a third time
resulted in similar changes in blood pH, PCO2, and lactate
to that measured following the second submersion, and
the length of the at-surface rest interval affected the mag-
nitude of recovery of blood acid-base status. The seventh
serial sample collected 180 min after the final postsubmer-
894
Fishery Bulletin 101(4)
Table 1
Mean (±SE) plasma Na*, K*, and plasma osmotic pressure (OP) prior to an
of sea turtles with a 10-min, 42-min, or 180-min rest interval. Serial blood
methods" section. Significant differences between samples 1 and 2, 3 and 4,
significant differences of samples from the initial blood sample (serial sampl
d following laboratory multiple forced submergences
sampling regime is described in the "Materials and
and 5 and 6 are indicated by an asterisk (*), whereas
e 1) are denoted by a pound sign (#).
Treatment
10 min
42 min
180 min
Na+
(mM)
(mM) (
OP
mosm/kg)
Na*
(mM)
(mM)
OP
(mosm/kg)
Na*
(mM)
(mM)
OP
(mosm/kg)
Control
153 ±3
4.0 ±0.2
319 ±3
152 ±4
3.9 ±0.3
305 ±2
158 ±4
4.3 ±0.4
322 ±4
Serial sample
1
144 ±5
4.5 ±0.3
319 ±6
158 ±6
3.9 ±0.2
314 ±11
162 ±2
4.1 ±0.1
296 ±3
2
159 ±6
5.9 ±0.6
341 ±4
163 ±3
6.1 ±0.6"
364 ±10**
187 ±2*«
6.9±0.3*«
342 ±5*"
3
145 ±3
4.9 ±0.2
330 ±5
156 ±2
4.1 ±0.3
336 ±8
160 ±4
4.4 ±0.1
308 ±4
4
166 ±7«
6.2 ±0.3"
351 ±14"
160 ±6
5.5 ±0.2"
342 ±15
179 ±4
6.7±0.5*»
339 ±4*»
5
158 ±6
5.1 ±0.1
335 ±11
147 ±6
4.1 ±0.3
334 ±12
158 ±6
3.9 ±0.2
305 ±5
6
154 ±5
6.1 ±0.3«
340 ±12
157 ±9
4.8 ±0.3
345±12«
181 ±2
5.7 ±0.5*
323 ±8"
7
139 ±3
4.8 ±0.4
323 ±8
149 ±6
4.4 ±0.3
331 ±7
158 ±10
4.4 ±0.5
305 ±3
gence sample revealed that blood pH, Pcog, and lactate
recovered completely for all experimental turtles (Figs. 1
and 2). Minimal changes in blood pH, PcOj and lactate were
detected in laboratory and field control turtles during col-
lection of the 5-7 serial blood samples (Figs. 1 and 2).
Ions, glucose, and osmotic pressure
Postsubmergence blood samples from laboratory turtles
revealed elevations in plasma Na*, K+, and osmotic pres-
sure when compared to the corresponding presubmergence
values (Table 1). Significant increases in the plasma Na*,
K+, and osmotic pressure were observed more frequently
in turtles with a longer in-water rest interval between suc-
cessive submergences (Table 1). In contrast, the plasma ion
concentrations and osmotic pressure of control turtles did
not substantially change (P>0.05, n=9) during serial blood
sample collection. In addition, no significant differences
in plasma glucose and CI" (P>0.05, n = 10) were measured
in any of the experimental turtles. Although most of the
postsubmergence changes in the blood parameters in
experimental turtles were not significant (Table 1), and
minimal alterations in blood chemistry were observed
in control turtles, the results suggested that there was a
relationship between blood acid-base status and plasma
osmolality and ion concentration. Therefore, correlation
analyses were used to determine the interdependence of
these variables.
Figure 3 shows the results of the correlation analyses,
where pH is plotted versus ion concentration (i.e. Na*,
K*, and CI" concentration), osmolality, or hematocrit.
Nonsubmerged control turtles had a significant correla-
tion between blood pH and plasma chloride, and pH and
hematocrit (Fig. 3). As pH declined, there were slight, yet
significant, increases in the [CI"] and hematocrit. How-
ever, no correlation was detected between pH and plasma
[Na*], [K+], or osmolality in these animals. In contrast,
a significant correlation was detected between blood pH
and plasma [Na*], [K+], [CI"], osmolality, and hematocrit
in experimentally submerged turtles (Fig. 3). In each case,
a decrease in blood pH led to an increase in the correlated
variable. These data are consistent with significant water
movement into and out of the red blood cells during and
after forced submersion.
Brief forced submergence of loggerhead turtles in trawl-
equipped fishing nets had a profound effect on the plasma
ionic status (Table 2). Plasma [K*] increased significantly
immediately following submergence in all experimen-
tal turtles. Significant increases were also observed in
the plasma [Na*] and osmotic pressure, although these
changes did not occur in turtles from all of the experimen-
tal treatments (Table 2). Turtles partially to completely
recovered from the ionic imbalances, although subsequent
submergences caused significant increases in plasma K"^
and nonsignificant increases in plasma Na* and osmolality
in most experimental turtles (Table 2). Ionic homeostasis
in forcibly submerged turtles was achieved within 180 min
of the final submergence, whereby plasma ion concentra-
tions were comparable to the initial presubmergence val-
ues (Table 2). The plasma ion concentrations and osmotic
pressure in nonsubmerged control turtles were unaffected
by serial blood sampling. Thus, ionic changes in experimen-
tal turtles resulted from the forced submergence and not
from handling and repetitive blood sampling.
Discussion
Acid-Base status
Multiple submergences of 2-year-old loggerhead sea tur-
tles under laboratory and field conditions produced sig-
Stabenau and Vietti: The physiological effects of multiple forced submergences of Caretta carelta
895
7Sn
Control Treatment
7 6-
7 4-
•'i'iH.; = '
7 2-
7 0-
66-
r^<00001
P=0 9717
7
76-
74-
7 2
7 0-
68
Submergence Treatment
r =0 1965
P=0 0001
[Na*]. mM
r' =00165
P=0,3147
7 8'
7 6-
74
72-
70-
68
66
7 6-
74-
72-
7 0
6 8-] r''=0 0640
P=0,0454
78
76
74
72
70-
6 8-1 r'=0 6301
P<0 0001
SV.
[K*], mM
7 8-|
7 6-
74-
\
v^^V.
72-
70-
. a • .\
fi«-
r'
=0 1946
•• ...X
P=
=0,0002
.
66-
, ,
[Cr], mM
r=0.0434
P=0 1014
■
:c^^^
■
"
B
. r^=0 4724
•'..
V
P<0 0001
■
Osmolality, mosmoles/kg
I
Q.
•^ft*^.
r=0 0975
P=0 0123
7.6
7.4'
7.2
7.0
68
66
r=0 2968
P<0 0001
Hematocrit (%)
Figure 3
Relationship between blood pH and plasma |Na*|, (K*], [C!"], osmolality, and hematocrit in
control (left column) and submerged (right column) loggerhead sea turtles. The lines are best
fits to the data. Significance of the correlated variables is noted on each figure.
896
Fishery Bulletin 101(4)
Table 2
Mean l±SE) plasma Na*, K*, and plasma osmotic pressure (OP) prior to and following multiple forced submergences of sea turtles
in TED-equipped nets with a 10-min, 42-min, or 180-min rest interval. Serial blood sampling regime is described in the "Materials
and methods" section. Significant differences between samples 1 and 2, 3 and 4, and 5 and 6 are indicated by an asterisk (*),
whereas significant differences of samples from the initial blood sample (serial sample 1) are denoted by a pound sign (#).
Treatment
10 min
42 min
180 min
Na*
(mM)
K*
(mM)
OP
(mosm/kg)
Na*
(mM)
K*
(mM)
OP
(mosm/kg)
Na*
(mM)
K*
(mM)
OP
(mosm/kg)
Control
150+3
3.0 ±0.2
313 ±8
139 ±6
3.4 ±0.2
321 ±7
151 ±1
3.1 ±0.1
310+5
Serial sample
1
153 ±2
3,3 ±0.3
318 ±4
160 ±4
3.1 ±0.2
331 ±12
164 ±2
4.5 ±0.7
325 ±9
2
171 ±8
5.5 ±0.3*»
345 ±4*»
186 ±8*
5.0 ±0.4*»
368 ±10
188 ±4
7.0 ±0.6*"
355 ±3*»
3
156+6
4.3 ±0.0"
332 ±4
163 ±3
2.8 ±0.1
338 ±11
163 ±10
3.6 ±0.3
314 ±3
4
171 ±8
5.3±0.1*»
349 ±1"
181 ±3
4.9 ±0.3*'*
361 ±13
176+10
6.2+0.3*»
352 ±9*
5
166 ±4
4.3 ±0.1"
334 ±2
160 ±8
2.9 ±0.2
332 ±9
173 ±10
4.0 ±0.2
323 ±3
6
166 ±12
5.1±0.1«
335 ±11
185 ±4*
4.5 ±0.4*«
343 ±14
175 ±18
5.3 ±0.0
333 ±11
7
157 ±4
3.7 ±0.1
325 ±2
161 ±6
2.6 ±0.2
326 ±9
159 ±11
3.6 ±0.6
320 ±4
nificant blood metabolic and respiratory disturbances.
The most dramatic changes in blood pH, Pco.^, and lactate
occurred following the first of the three forced submer-
gences in all of the experimental turtles (Table 3). Under
laboratory conditions, the turtles exhibited an average pH
change of 0.54 U following the first submergence, whereas
initial submergence of 2-year-old loggerhead sea turtles
in TED-equipped commercial fishing nets induced a pH
decrease of 0.63 U. The initial acid-base disturbances mea-
sured in our study were comparable in magnitude to those
measured in Kemp's ridley and loggerhead sea turtles in
standard TED certification trials (Table 3).
The second and third submergences of 2-year-old log-
gerheads sea turtles did not result in similar changes in
blood pH, PcOj, and lactate, as was measured following the
initial submergence (Table 3). To our knowledge, no infor-
mation is available in the literature on the physiological
effects of multiple submergences in sea turtles for compari-
son. Obviously, the interval between the submergence epi-
sodes directly influenced the magnitude of the blood acid-
base imbalance during successive submergences. A longer
time interval at the surface led to enhanced recovery of
blood pH, Pcog, and lactate. Lutz and Dunbar-Cooper ( 1987)
reported that loggerhead sea turtles captured during trawl-
ing at Cape Canaveral, Florida, exhibited a 16.8% decline
in lactate 180 min following submergence. Those authors
proposed that the rate of lactate decline was dependent on
the magnitude of the lactate concentration, so that 10 mM
of lactate would decline at a rate of 1.25 niM lactate/h.
However, in the present study, the rate of lactate decline
was considerably higher than that suggested by Lutz and
Dunbar-Cooper (1987). Lactate declined 70.0% and 79.6%
within 180 min of the submergence episodes in treatment
3 turtles, whereas no decline was measured in treatment
1 turtles (10 min interval) between submergences. In fact,
it was apparent that lactate continued to washout into the
bloodstream during the 10-min recovery phases in these
turtles (Fig. 1, Table 3). Thus, turtles with a brief period
between the submergence episodes would have a limited
ability to release the CO2 retained during submersion or
to break down lactic acid produced during the course of
the forced dive. Lactate declined 15.2%- and 18.7% during
the 42-min interval between submergences in treatment-2
turtles. Blood lactate declined 80.9%, 76.0%, and 82.5% in
treatment-1, -2, and -3 turtles, respectively, during the final
180-min recovery period. Thus, the overall rate of lactate
decline in the final 180 minutes of the laboratory study was
2.6 ±0.2 mM/h. Finally, the elevated lactate concentration
in sea turtles during the 180-min postsubmergence recov-
ery time interval suggests that the samples were collected
too soon to permit complete recovery of blood lactate.
Comparable rates of lactate clearance measured in the
laboratory submergence study were detected following
forced submergences of loggerhead sea turtles in TED-
equipped fishing nets. Substantial retention of CO, and
additional washout of lactate occurred during the 10-min
postsubmergence recovery interval in treatment-4 turtles.
Treatment-5 turtles exhibited a 6% drop in the blood
lactate concentration during the first 42-min postsubmer-
gence recovery interval and a 17.5% decrease in the blood
lactate during the second recovery interval. Thus, the 42-
min postsubmersion recovery interval permitted recovery
of blood gases, but was inadequate to clear the blood lactate
(Fig. 2, Table 3). Lactate declined 80.4% and 83.8%, respec-
tively, during the first two 180-min postsubmergence re-
covery intervals in treatment-6 turtles. As was the case for
laboratory submerged sea turtles, a longer surface interval
ultimately resulted in an increased ability to recover from
the submersion episodes. In fact, lactate declined 82.7%,
82.8%, and 87.9%, respectively, in treatment-4, -5, and -6
turtles 180 minutes after the final submersion episode
(Fig. 2, Table 3).
Stabenau and Vietti: The physiological effects of multiple forced submergences of Caretta caretta
897
Table 3
Effects of forced submergence on blood pH, Pco.„ and lactate in
Kemp's ridley (LK) and
loggerhead (CC) sea turtles. Data are |
expressed as the mean difference (J) between post- and presubme
rgence values. Data from
this study
are provided from the three
submergence episodes of treatment 1-3 turtles under the laboratory protocol and treatment 4-6 turtles in the field protocol, ND = |
not determined.
Turtle size Submergence
dPco^
iilactate
Species (kg) duration (min)
4pH (
mm Hg)
(mMl
Reference
LK 5-16.5 <7.3
0.37
12.8
8.5
Stabenau etal. (1991)
5-6 <7.3
0.31
24.5
15.1
TED certification tests'
CC 5-6 4.3
0.33
ND
13.4
TED certification tests'
5-6 12.5
0.52
ND
17.2
CC 6.5-7.0 7.5 treatment 1
1)0.49
61.1
7.6
Laboratory study
2)0.11
16.3
-0.1
3)0.10
15.3
1.1
treatment 2
1)0.57
70.8
9.3
2)0.20
20.9
2.3
3)0.23
21.1
1.1
treatment 3
1)0.59
98.7
9.6
2)0.50
68.6
7.2
3)0.46
67.3
5.9
treatment 4
1)0.63
45.8
10.2
Field study
2)0.16
24.5
1.9
3)0.11
9.3
0.9
treatment 5
1)0.53
36.3
9.1
2) 0.38
19.9
3.5
3) 0.28
17.5
3.0
treatment 6
1)0.73
54.2
11.2
2)0.66
31.3
9.2
3)0.65
27.5
9.3
' Data were collected by one of the authors (EKS) during standard TED certification tests
in 1993-94. S
amples were
collected from the cervical sinus
of Kemp's ridley and loggerhead sea turtles prior to and following forced submergences
in a commercial shrimp net equipped with a TED, Turtles |
m these studies were permitted to exit the TED-equipped net.
It must be noted that any discussion on lactate produc-
tion and recovery following submersion is applicable to
environmental conditions comparable to those reported
in this study. For example, lactate formation and recovery
rates of lactate build-up would be significantly influenced
by water temperature. Longer recovery rates may take
place in cold water, whereas warmer waters may lead to
additional lactate production thereby influencing the rate
of lactate elimination. In addition, the blood lactate con-
centrations measured in this study may underestimate the
true lactate burden. Lactate has been shown to partition
into other tissues, including the shell, following submer-
sion of freshwater turtles (Jackson et al., 1999). Finally,
sea turtle size could potentially alter lactate production
and elimination. Results from submersion experiments
conducted in our laboratory indicate that smaller animals
exhibit a significant acidosis and lactate build-up in com-
parison to larger sea turtles. Whether less acidosis and
lactate build-up is due to additional lactate buffering by
the larger sea turtles warrants further investigation.
Ions, osmolality, and hematocrit
There are three primary mechanisms for recovery of blood
pH following an acid-base disturbance: cellular buffer-
ing, and respiratory and renal compensation. Cellular
responses occur immediately following the disturbance,
whereas respiratory and renal adjustments occur within
minutes to hours, respectively. Previously, Stabenau et al.
(1991) reported that Kemp's ridley sea turtles exhibited
a significant increase in plasma [K+] following trawl sub-
mergences. However, those authors reported that trawl
stress had no effect on plasma [Cli, [Na+], or hematocrit.
In the present study, a cellular response to the severe
acid-base disturbance caused by the multiple forced
submergences was suggested by alterations in plasma
ion concentrations, osmolality, and hematocrit during the
blood acidosis. As shown in Figure 3, decreases in blood
pH were correlated with increases in [K*], [Na*l, [CI"],
osmolality, and hematocrit.
Hematocrit (percent packed red blood cells) changes may
result from washout of additional red blood cells into the
bloodstream, from areas such as the spleen, in order to pro-
vide more red blood cells during the hypoxic phases of the
forced submergence. This explanation, however, is unlikely
given that substantial fluctuations in hematocrit were ob-
served during the course of the submergence experiments
and that a normal hematocrit was measured in the final se-
rial blood sample. A more plausible explanation is that there
was an osmotically obliged influx of water into the red blood
898
Fishery Bulletin 101(4)
cells, swelling the cells, and leading to increases in hemato-
crit, and in plasma ion concentration and osmotic pressure.
Red cell volume is regulated in animals through transport
of intracellular and extracellular solutes. Although there is
minimal information available in the literature concerning
regulatory volume transport in reptiles, the mechanisms of
regulatory volume increase (RVI) and regulatory volume
decrease (RVD) are known in other lower vertebrates. For
example, Cala (1983) reported that in Amphiuma (am-
phiuma [common name]) red cells, the mechanism of RVD
is K+„^j/H+jj^ counter-transport coupled with Chg^j/HCOg'jj,
exchange (where the subscripts in and out represent
transport into and out of the cell, respectively), whereas
RVI is accomplished by Na^j^/H+^j^^j transport coupled with
Cl-,^/HC03-„,jt exchange (Cala, 1983). Other studies have
suggested that red cell RVD occurs because of electroneu-
tral KCl cotransport out of the cell and RVI occurs because
of electroneutral NaK2Cl or NaCl cotransport into the cell
(Haussinger and Lang, 1991). It is impossible to determine
which of these mechanisms, if any, were involved in regu-
lating red cell volume in sea turtles during and following
forced submergence. These transporters, however, have
been shown to be sensitive to cellular hypoxia (i.e. low Pog)
and low blood pH (Cossins and Gibson, 1997) — conditions
present in the experimental turtles following submergence.
In addition, hypoxic and acidotic conditions were absent
in nonsubmerged control turtles which did not experience
substantial shifts in plasma ion concentrations, osmotic
pressure, or hematocrit.
Effects of handling
Significant changes in blood pH, Pco.2, and lactate were
occasionally detected in nonsubmerged control turtles.
However, it is impossible to determine if these changes
resulted from repetitive handling during blood sampling or
from increased activity while free-swimming in a large cir-
cular tank following blood collection. Nevertheless, control
turtle blood lactate concentration was substantially less
than the lactate measured following forced submergence in
experimental turtles (Figs. 1 and 2). In addition, the blood
pH remained fairly constant in the control turtles during
collection of the seven serial samples.
Laboratory versus field experimentation
It should be noted that conducting the study under labo-
ratory and field conditions provided unique benefits for
analyzing the physiological effects of submersion. For
example, the laboratory conditions permitted collection of
blood samples immediately upon termination of the sub-
mersion period, whereas in the field, sea turtles had to be
transported back to the trawl vessel for postsubmersion
blood sampling. Turtles forcibly submerged under labora-
tory or field conditions hyperventilated upon surfacing.
Stabenau et al. (1991) reported a 9- to 10-fold increase
in the breathing frequency of trawled Kemp's ridley sea
turtles. Comparable breathing rates were observed in the
present study after submersion and, thus, it is plausible
that the blood PcOg measured in turtles under field condi-
tions underestimated the actual buildup in blood COg (see
Table 3 for a comparison of the blood Pco, under labora-
tory and field conditions). In contrast, the field experiment
permitted examining the physiological stress of semiwild
turtles in TED-equipped commercial fishing nets following
a minimum of 21 days of in-water conditioning. The greater
acidosis measured in forcibly submerged turtles resulted
from increased swimming activity during the forced sub-
mergence. This is confirmed by a postsubmergence increase
in blood lactate of 10.1 niM under trawling conditions
versus 8.8 mM following laboratory submergence. Under
laboratory and field conditions, the behavior of the turtles
following submergence was monitored up to their release.
It is unclear, however, if the acid-base and ionic imbal-
ance caused by forced submersions would alter long-term
normal physiology and behavior. It is plausible that repeti-
tive alteration of blood pH by the magnitude measured in
the present study may have pathological consequences. For
example, no information is available on whether turtles
resume normal diving and feeding behavior following pro-
longed or multiple forced submersions, or whether turtles
become more susceptible to repeated submersions in TED-
equipped nets.
Use of turtles reared in captivity
Two-year-old loggerhead sea turtles reared in captivity
were used for all of the submergence experiments. It was
assumed that these animals were adequate surrogates
for wild sea turtles. In fact, similar-size animals from the
NMFS Galveston Laboratory are used in annual TED
certification trials. Nevertheless, there may be differences
in the physiology of captive and wild turtles subjected to
forced submergences. For example, it is possible that wild
sea turtles would be exposed to forced submergences fol-
lowing lengthy, voluntary dives. No information is available
in the literature on the acid-base and ionic status of wild
sea turtles following prolonged voluntary dives or forced
multiple submergences. If dives are anaerobic, then sub-
jecting wild sea turtles to multiple forced submergences
may adversely affect survival potential.
Conclusions
The data suggest that forced submergences of 2-year-old
loggerhead sea turtles reared in captivity produce signifi-
cant blood metabolic and respiratory acidosis. Repetitive
submergences did not augment the acidosis, rather subse-
quent submergences resulted in less severe acid-base dis-
turbances. Under trawl conditions, the turtle must recover
from any physiological acid-base disturbance when it is
freed from a TED-equipped net. Recovery is accomplished,
in part, by the turtle immediately surfacing and hyper-
ventilating (Jackson, 1985; Stabenau et al., 1991). This
behavior was observed following each submergence epi-
sode. Turtles would then resume normal voluntary diving
behavior, presumably after partial-to-eomplete recovery
from the acid-base disturbance. These data suggest that
repetitive submergences of sea turtles in TED-equipped
Stabenau and Vietti: The physiological effects of multiple forced submergences of Caretta caretta 899
nets would not significantly affect their survival potential,
provided that the turtles have a recovery interval between
successive submergences. However, it should be noted that
the latter statement is based on comparable-size turtles
that may be submerged in shrimp nets equipped with
legally certified and installed turtle excluder devices. Poor
installation or lack of use of legal TEDs would result in aug-
menting the acid-base imbalance in the turtles. Increasing
the magnitude of the blood acid-base and ionic disturbance
during each submersion would increase the length of time
necessary to achieve partial or complete recovery.
Acknowledgments
Grateful appreciation is expressed to personnel from the
National Marine Fisheries Service Galveston and Pasca-
goula Laboratories for their assistance in turtle husbandry
and in conducting the submersion protocol under labora-
tory and very difficult field conditions. These studies were
conducted under appropriate threatened and endangered
species permits issued by the U.S. Fish and Wildlife Ser-
vice, Texas Park and Wildlife Department, and the Florida
Department of Natural Resources.
Literature cited
Caillouet, C. W., M. J. Duronslet, A. M. Landry, D. B. Revera,
D. J. Shaver, K. M. Stanley, R. W. Heinly, and E. K. Stabenau.
199L Sea turtle strandings and shrimp fishing effort in
the Northwestern Gulf of Mexico, 1986-89. Fish. Bull.
89:712-718.
Cala, R M.
1983. Volume regulation by red blood cells: mechanisms of
ion transport. Mol. Physiol. 4:33-52.
Cossins, A., and J. Gibson.
1997. Volume-sensitive transport systems and volume
homeostasis in vertebrate red blood cells. J. Exp. Biol.
200:343-352.
Crowder, L. B., S. R. Hopkins-Murphy, and J, A. Royle.
1995. Effects of turtle excluder devices (TEDs) on loggerhead
sea turtle strandings with implications for conservation.
Copeia 1995:773-779.
Haussinger, D., and F. Lang.
1991. The mutual interaction between cell volume and cell
function: a new principle of metabolic regulation. Biochem.
Cell Biol. 69:1-4.
Jackson, D. C.
1985. Respiration and respiratory control in the green turtle,
Chelonia mydas. Copeia 1985:664-671.
Jackson, D. C, Z. Goldberger, S. Visuri, and R. N. Armstrong.
1999. Ionic exchanges of turtle shell in vitro and their rel-
evance to shell function in the anoxic turtle. J. Exp. Biol.
202(part5):513-.520.
Lutz, R L., and A. Dunbar-Cooper
1987. Variations in blood chemistry of the loggerhead sea
turtle Caretta caretta. Fish. Bull. 85:37-44.
National Research Council.
1990. Decline of sea turtles: causes and prevention, 259 p.
National Academy Press, Washington, DC.
Owens D. W., and G. J. Ruiz.
1980. New methods of obtaining blood and cerebrospinal
fluid from marine turtles. Herpetologica 36:17-20.
Stabenau, E. K., and T A. Heming.
1994. The in vitro respiratory and acid-base properties of
blood and tissue from the Kemp's ridley sea turtle, Lepido-
chelys kempi. Can. J. Zool. 72:1403-1408.
Stabenau, E. K., T. A. Heming, and J. A. Mitchell.
1991. Respiratory, acid-base and ionic status of Kemp's
ridley sea turtles (Lepidochelys kempi) subjected to
trawling. Comp. Biochem. Physiol. 99A:107-111.
Wood, S. C, R. N. Gatz, and M. L. Glass.
1984. Oxygen transport in the green sea turtle. J. Comp.
Physiol. B. 154:275-280.
900
Abstract— The sectioned otoliths of
four fish species from a tropical demer-
sal trawl fishery in Western Australia
revealed a series of alternating trans-
lucent and opaque zones in reflected
light. The translucent zones, referred
to as growth rings, were counted to
determine fish ages. The width of the
opaque zone on the periphery of the
otolith section as a proportion of the
width of the previous opaque zone
(index of completion) was used to de-
termine the periodicity of growth-ring
formation.
This article describes a method for
modeling changes in the index of ring
completion over time, from which a
parameter for the most probable time
of growth-ring formation (with confi-
dence intervals) can be determined.
The parameter estimate for the timing
of new growth-ring formation for Leth-
rinus sp. 3 was from mid July to mid
September, for Lutjanus vitta from
early July to the end of August, for
Nemipterus furcosus from mid July to
late September, and for Lutjanus sebae
from mid July to mid November. The
confidence intervals for the timing of
formation of growth rings was variable
between species, being smallest for L.
vitta, and variable between fish of the
same species with different numbers of
gi'owth rings.
The stock assessments of these
commercially important species relies
on aging information for all the age
classes used in the assessment. This
study demonstrated that growth rings
on sectioned otoliths were laid down
annually, irrespective of the number
of growth rings, and also demonstrated
that the timing of ring formation for
these tropical species can be deter-
mined quantitatively (with confidence
intervals).
Quantitative determination of the timing of
otolith ring formation from marginal increments
in four marine teleost species from northwestern
Australia
Peter C. Stephenson
Western Australian Marine Research Laboratories
West Coast Drive (off Elvire St)
Waterman, Western Australia, 6020, Australia
E-mail address: pstephensongifish.wa.govau
Norm G. Hall
School of Biological and Environmental Sciences
Murdoch University
Murdoch, Western Australia, 6150, Australia
Manuscript approved for publication
3 June 2003 by Scientific Editor.
Manuscript received 26 June 2003 at
NMFS Scientific Publications Office.
Fish. Bull 101:900-909 (2003).
The Pilbara fish trawl fishery, operat-
ing on the North West Shelf of Western
Australia, has (developed rapi(ily in
the last ten years and is now the most
valuable commercial scalefish fishery in
Western Australia. Catch from this fish-
ery was valued at $7 million (wholesale
value) in 2001. In this multispecies fish-
ery, Lutjanus vitta (Quoy and Gaimard,
1824) (brownstripe red snapper),
Nemipterus furcosus (Valenciennes,
1830) (fork-tailed threadfin bream, also
known as rosy threadfin bream), Leth-
rinus sp. 3 (Carpenter and Niem, 2001)
(lesser spangled emperor, known locally
as blue-spot emperor) made up 8%, 10%,
and 20% respectively of the total sca-
lefish trawl catch in 2000. The highly
prized species, Lutjanus sebae (Cuvier,
1828) (red emperor), although compris-
ing only 4% of the catch, is important
because of its high market value.
In 1993 a research project was com-
menced to determine the fishing effort
required for optimal level of catches in
the Pilbara trawl fishery (Stephenson
and DunkM. The project relied on vali-
dated age composition data for L. vitta,
L. sp. 3, N. furcosus, and L. sebae. The
growth rings on otoliths have been
shown to be formed annually for only
one to three growth rings for N. furcosus
(Sainsbury and Whitelaw, 1984), and for
two to three growth rings for L. sebae
(McPherson and Squire, 1992). After
pooling of all age classes, Davis and
West (1992) showed that growth rings
of L. vitta were formed annually.
Determining age composition in-
volves counting growth rings on hard
parts of fish (otoliths, scales, spines,
bones) and determining the timing of
growth-ring formation. Sagittal otoliths
are commonly used for aging teleost
fishes and recent studies (Hyndes et al.,
1992; Milton et al., 1995; Newman et
al., 1996) have indicated that for some
species sectioned otoliths give more
reliable age estimates than whole, or
broken-and-burnt otoliths. The peri-
odicity of ring formation is commonly
determined by the mark-recapture
method in which fish are injected with
chemical markers and the number of
rings created between injection and
recapture are compared (Ferreira and
Russ, 1992; Francis et al., 1992; New-
man et al., 1996).
An alternative to mark-recapture is
marginal increment analysis in which
the distance from the growth ring to the
edge of the otolith, for a sample offish,
is tracked over time (Campana, 2001)
and a sharp drop in this marginal
increment, once a year, is taken as an
indication of annual ring formation.
The analysis is often performed on
Stephenson, P. C, and I. Dunk. 1996. Re-
lating fishing mortality to fish trawl
effort on the North West Slope of Western
Australia. Final report of project 93/25 to
the Fisheries Research and Development
Corporation, 1995, 44 p. Western Austra-
lia Marine Research Laboratories, PO Box
20, North Beach, Western Australia 6092,
Australia.
Stephenson and Hall; Timing of otolith ring formation in marine teleosts from northwestern Australia
901
pooled age classes (Barger, 1985; Manickchand-Heileman
and Kenny, 1990; Murphy and Taylor, 1990; Ross et al.,
1995; Pearson, 1996; Morales-Nin and Moranta, 1997; Van
der Walt and Beekley, 1997) or on a restricted number of
age classes (Sainsbury and Whitelaw, 1984; McPherson and
Squire, 1992).
Analysis with pooled data has limited value because
there may be different patterns of gi'owth-ring formation
at different life stages (e.g. at sexual maturity) and pooled
data may have interage differences masked by dominant
age groups (Beamish and McFarlane, 1983; Hyndes et al. ,
1992). Studies in which data were pooled only for young
and old fish, due to low fish numbers, reduced these prob-
lems and improved the credibility of the results (Hyndes et
al., 1992; Fletcher and Blight, 1996; Hasp et al., 2002).
Accounts of statistical analysis of the marginal incre-
ment data are rare. Davis and West ( 1992) used AN OVA to
show that there were differences in the marginal increment
of urohyal bones of L. vitta with time of year. As this sea-
sonal pattern was the same for age classes 1 to 6, Davis and
West ( 1992) pooled the data and used a graphical represen-
tation to show the time of formation of the annual rings.
This article describes a method for modeling changes
in the index of completion of an otolith growth increment
over time. This method enables quantitative determination
of the most probable time of growth-ring formation (with
confidence intervals) and is illustrated for the species L.
vitta, L. sp. 3, N. furcosus, and L. sebae, from the Pilbara
fish trawl fishery.
Materials and methods
Between October and November 1993 and between October
and November 1994, samples of 30 fish of each species were
randomly selected each month from fishery-independent
trawl surveys. For the other months between January 1994
and March 1995, samples of 30 fish of each species were
randomly selected each month from commercial catches.
The samples came from an area between 115°30'E longi-
tude and 120°E longitude; between the 50 meter and 100
meter depth isobaths.
The sagittal otoliths were extracted from each sampled
fish and the right otolith was embedded in epoxy resin and
then sectioned transversely through the otolith core to a
thickness of 0.4 mm. A Gemmaster high speed saw with a
100 mm by 0.1 mm diamond tipped saw blade was used for
sectioning. The otolith sections were set on 76 mm by 50
mm glass slides with casting resin and covered with cover
slips. The sections were viewed with a dissecting microscope
with an attached color video camera connected to a per-
sonal computer and a color monitor. Transmitted light
revealed alternating wide opaque and narrow translucent
zones. The translucent zones, referred to in the present
study as growth rings, were counted to determine fish ages.
The distance from the outer extremity of the last wide,
dark band to the otolith edge, u',, is referred to as the mar-
ginal increment and the distance between the outer edges
of the second to last and the last dark band is denoted by
«',_[. The distances w^ and w^^ on the portion of the otolith
ventral to the sulcus towards the proximal margin were
measured on the computer screen. The index of comple-
tion, c,, was determined by using the formula of Tanaka
etal. (1981)
(1)
and written to a file by using a computer program written
in the programming language "HiSoft Basic" (version 2.0.
MichTron, Auburn Hills, MI).
The index of completion, c,, we expect to increase over
time, and then decrease abruptly when a new growth ring
is formed. The timing of formation of a new ring would occur
at the same time for a fish species with the same number of
rings, but there would be considerable variability in timing
and detection between species and individuals (Fig. 1).
The increase in the index of completion over time, t, is
modeled as a strictly increasing function /■(<, a, b, d) with the
following parameters: maximum value, a, rate of increase,
6, and horizontal translation, d.
For our study, data were collected over a period of 18
months (October 1993 to March 1995) and the relation
between the index of completion and time was expressed
as two functions, denoted Fj and Fg
Fj : f-j = fit, a. b, c/,) and F., : i .-. = f\t, a, b, d.J,
(2)
where c ^ and t ,, = the estimates of the index of comple-
tion;
t = the time in months from t = 0 il Octo-
ber 1993) to < = 18 (31 March 1995);
and
d^, d.2 = the translation parameters for func-
tions Fj and F.2 respectively.
Ifthe point (c,, t) is associated with function Fj, the value
of the normal probability density function of the observed
deviation from Fj, evaluated at observation, i, is given by
A,. = — p=exp
and similarly the value of the normal probability density
function of the observed deviation from F.^, evaluated at
observation, i, is given by
'^., =•
a\!lK
^exp
(c-t\)-
where a- is the variance of the residuals when Fj is fitted
to the data and where it is assumed to be equal to the vari-
ance of the residuals when the function F^ is fitted.
To ensure the tractability of the subsequent analysis,
we assume that the probability, P,, of a point with index of
completion c,, at time t, being represented by Fj, is given
by the logistic function
P.=
1
1 +exp
ln(i9)^'^-')
(S-R)_
(3)
902
Fishery Bulletin 101(4)
Q.
E
Month
Figure 1
Values of index of completion at time t, (c,. t) forL. vitta with five growth rings during
the monthly time intervals from 1 October 1993 (t=Q) to 31 March 1995 (t=\S). Also
illustrated are the functions Fj and Fj fitted to (C,, t) and the function P, representing
the probability that (C,, t) is most hkely represented by function Fj. The likelihood
functions, Aj and A.^ are shown for the values of index of completion when t = 10.4.
where R, S, and R-iS-R) are the values off corresponding
to the 50'*^ , 95*, and S"' percentiles of the logistic function.
The probability that the index of completion is associated
with Fj, rather than Fj, is calculated as 1-P,.
Figure 1 illustrates typical values of the index of comple-
tion at time t and the functions Fj and Fj representing
these points before and after new growth-ring detection.
The most likely time at which a new growth ring is detected
is given by the value oft when P,=0.5. The likelihood func-
tions Aj and Ag are illustrated for t = 10.4 months. When
X, j is high, (Cj, t) is likely to lie closest to Fj and when A^ ,
is high, (c,, t) is likely to lie closest to F2.
As a point (c,, t) will be associated with either Fj or Fg
(but not both), it follows that the likelihood function K is
given by
the probability
of the observed
deviation
from Fj, X
given that
the point is
associated
with F,
That is,
probability
that the
point is
represented
hyF,
the probability
of the observed
deviation
from Fji
given that
the point is
associated
withF,.
X
probability
that the
point is
represented
byF,.
K,, = KA + ^^-^-P,^-
The overall log-likelihood associated with all the observed
points (c,, t), , for j = 1 to n in a particular age class is given
by
n n
^ln(^,,) = ^ln[A,,/^ + A,,(l-P,)].
The parameters of the functions Fj and Fj (i.e. a, b, d^,
and 1^2' as well as a, R, and S) were estimated separately
for each value of the number of rings by maximizing the
log-likelihood.
The value oft = R corresponds to the month where a val-
ue of the index of completion is equally likely to be on either
Fj or F2; that is, the point where the drop in the index of
completion occurs. The value of S and R-iS-R) correspond
to the 95'*^ and the 5'^ percentiles for the time at which a
new growth ring is likely to be detected, indicating reliabil-
ity of the estimate of the time of ring formation t = R.
Results
The plots of index of completion versus time reflected this
growth pattern in the four species we studied, with the
growth rate decreasing as the time of new growth-ring
detection approached.
The temporal pattern of growth of the otolith suggests
the index of completion could be modeled with a logistic
function
F{t.a.b,d) = -
1 -i-exp
lnl9
(cl't)
with the maximum value o = 1, phase shift d, and rate of
increase b.
A characteristic of otolith growth is that the distance
between growth rings decreases each year, thus, the rate
of increase in the marginal increments will be greater for
fish with few rings and less for fish with many rings. On
the other hand, the index of completion, being the ratio of
Stephenson and Hall: Timing of otolith ring formation in marine teleosts from northwestern Australia
903
? 08
0.6
0.4-
02
• ^
4 rings
• •
X • 1
•
•
•
•v.— — — T"
•
• 1 .
u-
1
• \ •
! 1 .
. • 1
• •
1 \ * *
A \
0123456789 101112131415161718 0123456789 1011121314 15161718
ONDJ FMAMJJASONDJFM ONDJ FMAMJJASONDJFM
Month
Figure 2
Index of completion at time<,(c„ t), for L. sp. 3 for growth-ring categories 2,3,4,5,6, 7,8, and 9-11 sampled
between the months 1 October 1993 {t=0} to 31 March 1995 it=\8). The sohd circles (•) represent those (c,,
t) most likely represented by c^ and lying closest to the function Fj and the solid triangles (A) are those
most likely represented by c,, and lying closest to the function Fj. The logistic function, P,, indicates the
probability that points (c,, t) are most likely represented by function F,.
the marginal increment to the width of the previous band,
would be expected to be constant at the same time of year
for a particular species, regardless of the ring count.
Thus, during the maximization of the objective function
n
the rate of increase parameter b is assumed constant for
all ring counts for a species, but the other parameters (dj ,
d^, R, and S) are estimated for each number of rings for
each species.
Figure 2 shows the pattern of changes in the index
of completion for L. sp. 3 for otoliths with two to eleven
growth rings from 1 October 1993 («=0) to 31 March 1995
(<=18). The data were pooled for the older age classes (nine
or more growth rings) because of the small numbers of old
fish in the samples. The points indicating the values of in-
dex of completion, c^, are represented with different symbols
according to whether they are most likely to be represented
by Cj (lying closest to the function Fj) or most likely to be
represented by c^ dj^ng closest to the lower line F^). The
function P, represents the probability of points being repre-
sented by line Fj. The time where P,=0.5 is the most likely
time of detection of the formation of a new growth ring.
In the other three species, L. vitta, N. furcosus, and L.
sebae, the points representing index of completion, (c,, t)
and functions, F^, Fj, and P, are illustrated in Figures 3, 4,
904
Fishery Bulletin 101(4)
3 rings
^
0.8-
•
•
•
Ar
-♦-
•
0.6-
• •
1
^i :
1
•
•
A
J»r
•
* A *^
r
0.4-
o
%
(B
(1^-
^^
—
■% '
^
CL
\
A
t
nn-
1-0
5 rings
•
• *
•
1
0.8-
•
■
•
V
•
0.6-
0,4-
•
• •
•
•
•
• •
A
•
A
P
0 2-
nn-
•
• .
A
A
^
d:
i*-i r-
(
0123456789 101 1 12 13 14 15 1617 1£
ONDJ FMAMJ J ASONDJFM
0 12 3 4 5 6 7
ONDJ FMAM
Month
Figure 3
Index of completion at time t , (c,, t), for L. vitta for growth-ring categories 2 ,3, 4, 5, 6, 7, 8, and 9-12 sampled
between the months 1 October 1993 (;=0) to 31 March 1995 (?=18). The solid circles (•) represent those (c,,
t) most likely represented by t-j and lying closest to the function Fj and the solid triangles (A) are those
most likely represented by c, and lying closest to the function F.,. The logistic function, P, , indicates the
probability that points (c,, I) are most likely represented by function F,.
and 5. Because of the smaller numbers offish, the data was
pooled for L. vitta with 9-12 growth rings, for N. furcosus
with 7-9 growth rings, and for L. sebae with 13-14 and
15-19 growth rings.
The estimates of the parameters b, d^, d, (the rate of in-
crease and phase shift of Fj and F^), R, the time of formation
of a new growth ring, with the range ±(S-R), and the stan-
dard deviation (cj) for the four species are listed in Table 1.
The estimated time of formation of a new ring varied
between age classes and occurred for L. sp. 3 from mid
July to mid September, for L. vitta from early July to early
September, for N. furcosus from mid July to late September,
and for L. sebae from mid July to mid November.
The confidence intervals are small for L. sp. 3 and L. vitta
and generally larger for A^. furcosus and L. sebae where read-
ing increments was more difficult. The confidence interval
was smaller for L. sebae with 12-19 growth rings because
the clarity of the rings generally improved for older fish.
The difference between the phase shift for functions f j
and F^, that is d^-di, was between 11 to 13 months. This
difference indicated an annual cycle of growth-ring forma-
tion. The standard deviation, a, of the values of index of
completion was lowest for forL. vitta (0.1-0.14) and highest
fori, sebae (0.11-0.21) and L. sp. 3 (0.11-0.24).
Discussion
With different starting values for the parameters, the
method we described found that the time of detection of
growth-ring formation for the four species was consistent.
Although the timing varied considerably for different num-
bers of growth rings, the estimate was generally similar
for each species.
Other marginal increment studies on these species pro-
duced estimates of the times for growth-ring formation
Stephenson and Hall: Timing of otolith ring formation in marine teleosts from northwestern Australia
905
0123456789 1011121314 15161718 0123456789 1011121314 15161718
ONDJ FMAMJ J ASONDJFM ONDJ FMAMJ J ASONDJFM
Month
Figure 4
Index of completion at time t, (c,, t), for N. ftircostis for growth-ring categories 2 ,3, 4, 5, 6, and 7-9 sampled
between the months 1 October 1993 (^=0) to 31 March 1995 (/=18). The solid circles (•) represent those (c,,
t), most likely represented by c"j and lying closest to the function F, and the solid triangles (A) are those
most likely represented by c"., and lying closest to the function Fj. The logistic function, P,, indicates the
probability that (c,, t) are most likely represented by function Fj.
that are consistent with those determined in the present
study (Table 2). Davis and West ( 1992) found that the time
of formation of a new translucent growth-ring on urohyal
bones for L. vitta was October, later than the timing found
in our study. Sainsbury and Whitelaw (1984) found the
marginal increment values on whole otoHths from N. fur-
casus had low values in July 1979 and in May 1980 (earlier
than observed in the present study for sectioned otoliths)
but the sampling reported by Sainsbury and Whitelaw
(1984) was very sparse: four sampling times in 1979 and
two in 1980. McPherson and Squire (1992) reported that
the mean monthly marginal increment of the first two age
classes for L. sebae appeared to have a minimum between
July and September that is consistent with the present
study.
In our study, the growth zones on L. vitta were generally
clearly defined; the opaque zone was easily distinguished
from the translucent zone, and there were few discontinui-
ties (areas of dissimilar structure or optical density within
the growth zone). This clear definition was especially no-
ticeable for the outer growth zones of the otolith for older
fish which often had very clear dark zones. This finding is
consistent with the small confidence intervals for this spe-
cies, especially in the fish with a greater number of growth
rings in their otoliths. The growth rings on L. sp. 3 and N.
furcosus had poor contrast and had many discontinuities
which made the analysis difficult. For young L. sebae, there
were many discontinuities within the growth zones which
made locating the translucent zone difficult. For fish with
two to four growth rings, low values of the index of comple-
tion occurred when t = 7 and also when t = 12 (Fig. 5). For
the older L. sebae (where the number of rings is greater
than or equal to 12) the wide zone was very dark and by in-
creasing the magnification, the marginal increment could
be measured relatively easily. The narrow confidence limits
for the timing of growth-ring formation for L. sebae are
consistent with this explanation but the small number of
data points results in less reliable measures in the timing
of new ring formation.
The time of formation of the new growth ring was
slightly earlier for L. vitta than for L. sp.3, N. furcosus, or
L. sebae. The calculation of an earlier growth-ring forma-
tion may be attributed to the more clearly defined trans-
lucent zone in L. vitta which may be detectable earlier in
906
Fishery Bulletin 101(4)
0,8-
5 rings
•
: 1
•
•
(
•
•
U.b-
• •
. r^^T
«
.
'>'"^*
O.4.
• • ^^
-r"*^
-— '
i
0.?-
'
A *
A
nn.
A
1— i. .
10 rings
•
N.«-i-^
■
• •
•
-^
A
•
>^f^
\
* ,
^- *
•
•
^v
— -"^
• *
A
\i
0.8-
8 rings
t
•
•
•
■*.
t
.1
• •
• •
0-6-
0.4-
02-
I
I.
•
•
•
i
«
V
nn.
*
0123456789 101112131415161718 0123456789 1011121314 15161718
ONDJFMAMJJASONDJFM ONDJFMAMJJASONDJFM
Month
Figure 5
Index of completion at time t, (c,, t). for L. sebae for growth-ring categories 2-3, 3, 4, 5, 6, 7, 9, 10, 11, 12,
13-14, and 15-19 sampled between the months 1 October 1993 «=0) to 31 March 1995 (t=18). The solid
circles (•) represent those (c,, t) most likely represented by i", and lying closest to the function Fj and the
solid triangles (A) are those most likely represented by c'j and lying closest to the function F^. The logistic
function, P, , indicates the probability that (c,, t) are most likely represented by function F,.
Stephenson and Hall: Timing of otolith ring formation in marine teleosts from northwestern Australia
907
Table 1
Parameter estimates for the fit of two functions to the index of completion data. The parameters are the
phase shifts rfj and d.,, the time of new growth-rmg detection t = R with the confidence interval (5"' and 9£
standard deviation, a, of the observed values of index of completion for points fitted to function Fj and F^.
rate of increase, b, the
"> percentiles), and the
rings
;i
h
rf;
d..
R
Month
a
Lutjanus sp. 3
2
36
20.3
3.6
15.3
9.5 ±0.1
Jul
0.14
3
134
20.3
2.0
15.4
9.9 ±2.0
Jul
0.13
4
146
20.3
1.7
16.2
10.2 ±2.1
Aug
0.13
5
72
20.3
2.3
14.8
10.5 ±0.1
Aug
0.24
6
38
20.3
1.7
14.9
11.5 ±0.2
Sep
0 11
7
22
20.3
0.6
13.7
10.7 ±0.1
Aug
0.12
8
22
20.3
1.1
14.0
10.2 ±0.1
Aug
0 12
9-11
29
20.3
0.9
14.2
10.9 ±1.3
Aug
0.15
Lutjanus vitta
2
31
15.8
3.0
14.3
10.9 ±0.1
Aug
0.13
3
62
15.8
1.5
18.3
10.5 ±0.2
Aug
0.14
4
96
15.8
3.4
15.9
10.5 ±0.2
Aug
0 10
5
127
15.8
2,7
16.3
10.2 ±2.0
Aug
0.12
6
66
15.8
2.5
16.5
10.0 ±1.3
Aug
0.12
7
56
15.8
2.0
15.4
10.3 ±1.7
Aug
0.11
8
24
15.8
3.3
15.8
9.0 ±0.2
Jul
0.10
9-12
39
15.8
3.1
16.2
11.1 ±0.2
Sep
0.10
Nemipterus furcosus
2
85
11.9
2.0
13.1
10.6 ±2.7
Aug
0.18
3
95
11.9
3.9
14.2
9.5 ±2.2
Jul
0.17
4
116
11.9
4.0
15.0
9.5 ±4.3
Jul
0.14
5
82
11.9
3.3
16.4
11.8 ±2.6
Sep
Oil
6
53
11.9
4.0
16.4
11.8 ±4.2
Sep
0.10
7-9
39
11.9
4.5
15.8
11.4 ±3.0
Sep
0.10
Lutjanus sebae
2-3
34
17.4
1.9
14.4
7.8 ±2.5
May
0.17
4
62
17.4
5.0
15.9
11.9 ±1.2
Sep
0.21
5
43
17.4
5.1
13.4
11.6 ±6.7
Sep
0.14
6
60
17.4
3.3
17.0
9.6 ±3.7
Jul
0.13
7
63
17.4
2.6
17.0
10.6 ±4.9
Aug
0.16
8
60
17.4
2.8
15.8
13.5 ±7.6
Nov
0.15
9
52
17.4
3.2
15.9
11.0 ±3.3
Sep
0.19
10
32
17.4
3.3
15.9
10.6 ±3.4
Aug
0.15
11
34
17.4
3.1
14.9
11.6 ±2.6
Sep
0.17
12
22
17.4
2.1
13.9
9.6 ±0.1
Jul
0.12
13-14
23
17.4
4.7
14.6
13.3 ±0.4
Nov
0.21
15-19
21
17.4
4.0
17.6
11.5 ±0.2
Sep
0.11
Table 2
The timing of growth-ring formation for the four
species in the present
study and for comparative
studies
Lutjanus sp.3
L. vitta
Nemipterus furcosus
L. sebae
Present study July-September
July-September
July-September
July-November
Davis and West (1992)
October
Sainsbury and Whitelaw ( 1984)
May-July
McPherson and Squire (1992)
July-September
908
Fishery Bulletin 101(4)
the year than it is in the other three species. Similarly, the
apparent earlier timing of new ring creation in our study,
compared to the findings of Davis and West (1992) may have
been due to the fact that the translucent zone can probably
be detected closer to the time of formation in sectioned
otoliths than in urohyals (Reshetnikov and Claro, 1976).
In summary, modeling the change in index of completion
over time enabled estimates to be made of the time of the
formation of a new growth ring (with confidence intervals)
for four tropical species. Although the index of completion
was modeled with a logistic function in the present study,
alternative functions (e.g. sine or linear), gave very similar
results. The technique is a useful addition to marginal in-
crement analysis because it can be used in place of previous
subjective methods to determine quantitatively the timing
of new ring formation.
Acknowledgments
This project was financed by the Fisheries Research and
Development Corporation (FRDC) (project 94/25) and the
Department of Fisheries, Western Australia where the
main author was employed for the duration of the project.
The authors thank Mike Moran (Department of Fisher-
ies, Western Australia) for obtaining FRDC funding and
providing critical advice and encouragement. The senior
author also thanks Robert Black (University of Western
Australia) for his valued suggestions and encouragement.
Stephen Newman (Department of Fisheries, Western
Australia) provided advice on reading and interpretation
of otolith bands, Iain Dunk (Department of Fisheries,
Western Australia) collected samples, sectioned otoliths,
and acted as the second reader for fish aging and mar-
ginal increment analysis. Tony Paust, Ken Bryers, Justin
Chidlow, Daryn Payne (Department of Fisheries, Western
Australia) assisted in sample collection. I also acknowledge
the assistance of M. G. Kailis, Kraus Fishing Company, and
Westmore Seafoods whose vessels were used for sample
collection, and two anonymous referees for their construc-
tive comments.
Literature cited
Beamish, R. J., and G. A. McFarlane.
1983. The forgotten requirement for age validation in fisher-
ies biology. Trans. Am. Fish. Soc. 112:735-743.
Barger, L. E.
1985. Age and growth of Atlantic Croakers in the Northern
Gulf of Mexico based on sectioned otoliths. Trans. Am.
Fish. Soc. 114:847-850.
Campana, S. E.
2001. Accuracy, precision and quality control in age deter-
mination, including a review of the use and abuse of age
validation methods. J. Fish Biol. 59:197-242.
Carpenter, K. E., and V. H. Niem.
2001. FAG species identification guide for fishery purposes.
The living marine resources of the Western Central Pacific,
vol. 5. Bony fishes, part 3 (Menidae to Pomacentridae), p.
2791-3380. FAG (Food and Agriculture Organization of
the United Nations), Rome, Italy.
Davis, T. L. G., and G. J. West.
1992. Growth and mortality of Lutjanus vittus (Quoy and
Giamard) from the North West Shelf of Australia. Fish.
Bull. 90:395-404.
Ferreira, B. P., and G. R. Russ.
1992. Age, growth and mortality of the inshore coral trout
Plectropornus maculatus (Pisces: Serranidae) from the Cen-
tral Great Barrier Reef Australia. Aust. J. Mar Freshw.
Res. 43:1301-1312.
Fletcher, W. J., and S. J. Bhght.
1996. Validity of using translucent zones of otoliths to age
the pilchard Sardinops sagax neopilchardus from Albany,
Western Australia. Mar Freshw. Res. 47:617-624.
Francis, R. I. C. C, L. J. Paul, and K. P Mulligan.
1992. Ageing of adult snapper (Pagrus Auratus), from otolith
annual ring counts: validation by tagging and oxytetacy-
cline injection. Aust. J. Mar Freshw. Res. 43:1069-1089.
Hesp, S. A., I. A. Potter, and N. G. Hall.
2002. Age and size composition, growth rate, reproductive
biology, and habitats of the West Australian dhufish, Glau-
cosoma hebraicum, and their relevance to management of
this species. Fish. Bull. 100:214-227.
Hyndes, G. A., N. R. Loneragan, and I. C. Potter
1992. Influence of sectioned otoliths on marginal increment
trends and age and growth estimates for the flathead Platy-
cephalus speculator. Fish. Bull. 90:276-284.
Manickchand-Heileman, S. C, and J. S. Kenny.
1990. Reproduction, age, and growth of the whitemouth
croaker Micropogonias furnieri (Desmarest 1823) in Trini-
dad waters. Fish. Bull. 88:523-529.
McPherson, G. R., and L. Squire.
1992. Age and growth of three dominant Lutjanus species
of the Great Barrier Reef inter-reef fishery. Asian Fish.
Sci. 5:25-36.
Milton, D. A., S. A. Short, M. F O'Neill, and S. J. Blaber
1995. Ageing of three species of tropical snapper (Lutjani-
dae) from the Gulf of Carpentaria, Australia, using radiom-
etry and otolith counts. Fish. Bull. 93:103-115.
Morales-Nin, B., and J. Moranta.
1997. Life history and fishery of the common dentex
(Dentex dentex) in Mallorca (Balearic Islands, western
Mediterranean). Fish. Res. 30:67-76.
Murphy, M. D., and R. G. Taylor
1990. Reproduction, growth and mortality of red drum
Sciaenops ocellatus in Florida waters. Fish. Bull. 88:
531-542.
Newman, S. J., D. M. Williams, and G. R. Russ.
1996. Age validation, growth and mortality rates of the tropi-
cal snappers (Pisces: Lutjanidae ) Lu(/a/?!zs adetii (Castelnau,
1973) andL. quinquelineatiis (Bloch, 1790) from the Central
Great Barrier Reef Aust. J. Mar Freshw. Res. 47:575-584.
Pearson, D. E.
1996. Timing of hyline-zone formation as related to sex, loca-
tion, and year of capture in otoliths of the widow rockfish,
Sebastes entomelas. Fish. Bull. 94:190-197.
Reshetnikov, Y. S., and R. M. Claro.
1976. Cycles of biological processes in tropical fishes with
reference to Lutjanus synagris. J. Ichthyol. 16:711-733.
Ross, L. J.,T. M. Stevens, and D. S. Vaughan.
1995. Age, growth, mortality, and reproductive biology of
red drums in North Carolina waters. Trans. Am. Fish.
Soc. 124:37-54.
Sainsbury, K. J., and A. W. Whitelaw.
1984. Biology of Peron's threadfin bream Nemipterus pero-
nii (Valenciennes), from the North West Shelf of Australia.
Aust. J. Mar Freshw. Res. 35:167-185.
Stephenson and Hall: Timing of otolith ring formation In marine teleosts from northwestern Australia 909
Tanaka, K.. Y. Mugiya, and J. Yamada. Van der Walt, B.A., and L. E. Beckley.
1981. Effects of photoperiod and feeding on daily growth 1997. Age and growth of Sarpa salpa (Pisces: Sparidae) off
patterns in otoliths of juvenile Ttlapia nilotica. Fish. the coast of South Africa. Fish. Res. 31:241-248.
Bull. 79:459-465.
910
Application of DNA-based techniques for the
identification of whaler sharks (Carcharhinus spp.)
caught in protective beach meshing and
by recreational fisheries off the coast of
New South Wales
Ricky W. K. Chan
School of Biological, Earth and Environmental Sciences
The University of New South Wales, UNSW
Sydney, New South Wales 2052, Australia
Present address: Educational Testing Centre
The University of New South Wales
ULD 3 East Parcel Centre
Rosebery, New South Wales 2018, Australia
E-mail address sharkman@etc.unsw,edu au
Patricia I. Dixon
Centre for Manne and Coastal Studies
The University of New South Wales, UNSW
Sydney, New South Wales 2052, Australia
Julian G. Pepperell
Pepperell Research and Consulting
PO Box 1475
Noosaville DC, Queensland 4566, Australia
vessels may experience difficulties in
identification if distinguishing parts of
a shark are discarded prior to confir-
mation of species (Stevens and Wayte^).
Similarly, observers of protective beach
meshing may find species identifica-
tion difficult on severely decomposed
sharks. Without proper identification,
the exact number of individual species
inhabiting NSW waters and the num-
ber of each species being landed cannot
be determined (Chan, 2001).
The rise of molecular biological tech-
niques in marine forensic science has
facilitated the development of accu-
rate taxonomic identification of shark
species by sampling biological tissue
(Martin, 1991; Lavery 1992; Heist and
Gold, 1999). DNA techniques require
only muscle tissue, allowing biopsy
tissue to be taken from specimens that
can be released, rather than having
to sacrifice the shark to obtain liver
and heart tissue for allozyme analysis
(Godfrey, 1997). Methods of taxonomic
identification include PCR-based
restriction fragment length poly-
Dennis D. Reid
New South Wales Fishenes
PO Box 21
Cronulla, New South Wales 2230, Australia
The International Union for the Conser-
vation of Nature's (lUCN) development
of the Shark Specialist Group is indica-
tive of the increasing environmental
awareness of sharks' crucial ecological
role as apex predators and that they
are being threatened by human activi-
ties. Although the conservation status
of certain carcharhinid species (Car-
charhinus limbatus, C. obscurus, and
C. plumbeus) are presently considered
at low risk or near threatened accord-
ing to the lUCN's threatened species
categories,' species from the genus
Carcharhinus are known to inhabit
the waters of New South Wales (NSW),
Australia (Stevens, 1984; Last and
Stevens, 1994); however their conser-
vation status has not been determined.
Known as whaler or "requiem" sharks,
they are also commonly caught off
the coast of New South Wales in com-
mercial fisheries (Stevens and Wayte^;
Tanner and Liggins^), recreational fish-
eries (Pepperell, 1992; Gartside et al.,
1999; Steffe et al.*) and by protective
beach meshing (Reid and Krogh, 1992;
Dudley 1997).
Because of morphological similari-
ties between a number of shark spe-
cies in the genus Carcharhinus (Last
and Stevens, 1994; Naylor and Marcus,
1994), taxonomic identification to spe-
cies level has been difficult or inac-
curate (or both)(Stevens and Wayte^).
Historical catches of certain species of
sharks in NSW commercial fisheries,
recreational fisheries, and protective
beach meshing have been recorded to
genus level only (Pepperell, 1992; Reid
and Krogh, 1992; Tanner and Liggins'^).
Formally trained Australian Fisheries
Management Authority (AFMA) ob-
servers aboard commercial longlining
1 Musick, J., and S. Fowler. 2000. Car-
charhinus limbatus, C. obscurus and C.
plumbeus In lUCN 2002. 2002 lUCN
red list of threatened species, http:
//www. iucn.org/redlist/2000index. html.
[Accessed 1 October 2002.1
2 Stevens,J. D..andS. E.Wayte. 1998. A
review of Australia's pelagic sharks
resources. Fisheries Research and Devel-
opment Corporation project 98/107, 64
p. CSIRO Marine Research, GPO Box
1538, Hobart, Tasmania 7001 Australia.
3 Tanner, M., and G. W. Liggins. 2000.
New South Wales commercial fisheries
statistics 1993/94 to 1996/98, 82 p. New
South Wales Fisheries, PO Box 21,
Cronulla, NSW 2230 Australia.
■* Steffe, A. S., J. J. Murphy, D. J. Chapman,
B. E. Tarlinton, G. N. G. Gordon, and A.
Grinberg. 1996. An assessment of the
impact of offshore recreational fishing in
New South Wales waters on the manage-
ment of commercial fisheries. Fisheries
Research and Development Corporation
Project 94/053, 139 p. New South Wales
Fisheries, PO Box 21, Cronulla, NSW 2230
Australia.
Manuscript approved for publication
24 June 2003 by Scientific Editor
Manuscript received 13 July at NMFS
Scientific Publications Office.
Fish. Bull. 101:910-914 (2003).
NOTE Chan et al.: Identification of Carcharhinus spp, by DNA-based techiniques
911
morphism (PCR-RFLP; Martin, 1991), DNA sequencing
techniques (Heist and Gold, 1999), isoelectric focusing of
muscle proteins (Renon et al., 2001; Smith and Benson,
2001) and direct multiplex PCR amplification (Shivji et
al., 2002). These techniques use the differences in the se-
quences of nucleotide bases within a DNA strand among
species. DNA techniques have high sensitivity, are easily
reproduced, and allow the development of a unique "DNA
fingerprint" for each species (Martin, 1991; Innes et al.,
1998; Pepperell and Grewe, 1999). It was the aim of this
project to initiate a shark DNA database for species iden-
tification of pelagic sharks (beginning with species from
the genus Carcharhinus) in New South Wales by using
PCR-RFLP techniques.
Materials and methods
Sharks of the genus Carcharhinus landed by recreational
fisheries and caught in NSW beach meshing were identified
to species level by using morphometric taxonomic guides
and dentition identification (Cliff and Wilson, 1994; Last
and Stevens, 1994; Naylor and Marcus, 1994). Positively
identified sharks were retained as voucher specimens (see
"Acknowledgments" section). A tissue biopsy (5-10 g) from
the dorsal region on either side of all voucher specimens
and unidentified sharks was taken and stored in 75% etha-
nol prior to DNA extraction. Mitochondrial DNA (mtDNA)
of specimens from six species of the genus Carcharhinus
(C. brachyurus, C. brevipinna, C. falciformis, C. leucas, C.
limbatus, and C ohscurus; see Table 1 for sample sizes) was
extracted by using a Fastprep DNA Extraction kit (BIOlOl,
Integrated Sciences, Sydney, New South Wales) following
the manufacturer's instructions. Approximately 200-300
mg of tissue was placed in a sterilized 1.5-mL eppendorf
tube and after the addition of 1 mL Fastprep lysis buffer,
incubated at 56°C for three hours prior to the extraction
stage (Chan, 2001). Following the extraction procedure of
the Fastprep protocol, quality and quantity of DNA was
measured by using a GeneQuant DNA/RNA calculator
( Amersham Biosciences, Sydney, New South Wales).
The polymerase chain reaction (PCR) was used to am-
plify the 1146 nucleotide base-pair (bp) cytochrome b (cyt
h) region of the mtDNA (Martin and Palumbi, 1993; Kita-
mura et al., 1996). For each 50 pL PCR reaction, 100-200
ng of template mtDNA was used, 1.5 niM MgCU, IX PCR
buffer, 2 mM of dNTP, 5 mM of each external primer
(5'-TGACTTGAARAACCAYCGTTG-3' and 3'-CTCCAG-
TCTTCGRCTTACAAG-5') and two units of DyNAzyme
EXT DNA polymerase (Finnzymes, GeneWorks, Adelaide,
South Australia) were added to a sterilized 200 pL PCR
tube. The PCR was undertaken in a MJR MiniCycler (MJR,
GeneWorks) with a heated bonnet on a cycle of 94°C for
three minutes, followed by 35 cycles of 94°C denaturing
for 45 seconds, 48°C annealing for 30 seconds, 72°C exten-
sion for 90 seconds, and a final 10-minute extension time
at 72°C (Chan, 2001).
To determine if the cyt 6 region was successfully ampli-
fied, 10 pL of PCR product was added to 2 pL loading dye
(25% bromophenol blue, 40% sucrose in distilled water) and
loaded into wells of a 1.5% w/v agarose gel submerged in
0.5X TBE (Tris-borate-EDTA, pH 8) buffer, with 1 pg of a
100-bp DNA ladder (Sigma-Aldrich, Sydney, New South
Wales) added to 5 pL distilled water added to each side-
end well. The gel was subject to electrophoresis at 125 V
for 45-60 minutes and then stained in ethidium bromide
for 10 minutes, de-stained in fresh distilled water for 20
minutes prior to illumination under ultraviolet (UV) light
to determine the success and yield of the amplification.
To obtain species-specific profiles, restriction fragment
length polymorphism (RFLP) was used on the entire 1146
bp cyt b fragment (Martin, 1991; Chan, 2001). The suc-
cessful amplified reaction products had the primers, taq
polymerase, and buffer chemicals removed by using a
BRESAspin PCR purification kit (GeneWorks). For each
RFLP reaction, 30 pL of purified PCR-amplified cyt b
mtDNA (300-1000 ng; 30 pL of distilled water was used
for control reactions), 5 pL of lOX buffer, one unit of a re-
striction enzyme (Alu l,Hae III, Ps? I, Taq I,andXAo I) and
distilled water up to 50 pL volume was added to a steril-
ized 200 pL PCR reaction tube and incubated at 37°C for
2 hours in the MJR minicycler (with heated bonnet), with
the exception of Taq I (incubated at 65°C for two hours in
a water bath). After the allotted digestion time, 10 pL of
loading dye was added to each tube prior to loading into a
1.5% w/v agarose gel submerged in 0.5XTBE buffer In both
end wells, 5 pL of distilled water -i- 1 pg of a 100 bp DNA
ladder was added. The gel was subject to electrophoresis
at 125 V for 60-90 minutes, stained in ethidium bromide
for 10 minutes, and destained in fresh distilled water for
20 minutes prior to illumination under UV light. Enzyme-
digested DNA fragments >100 bp were then "scored" to the
nearest 25 bp based upon migration of the DNA fragment
(the smaller the fragment, the faster the migration) and
recorded for each enzyme and sample (Martin, 1991) by
using the 100-bp DNA ladder as a standard measuring
guide for size estimation.
Results and discussion
PCR-RFLP profiles were successfully developed for six spe-
cies of the genus Carcharhinus; distinct and discrete pat-
terns were observed for each species with five restriction
enzymes (Table 1, and Chan, 2001). The only intraspecific
polymorphism observed was for two specimens of C. bre-
vipinna with the Xho I restriction enzyme. Increasing the
sample size of all species may identify more intraspecific
polymorphisms. Because of the relatively small sample
sizes, no statistical analyses were undertaken. Other
restriction enzymes were tested (Chan, 2001), and with
the possible inclusion of other species from the genus Car-
charhinus into this database in the future, these restriction
enzymes may be required in order to discern the additional
species. Because some of the fragment sizes were rounded
to the nearest 25 bp, the sum of the fragments for a restric-
tion enzyme of a species may be more than 1146 bp, the
size of the C3rt b uncleaved region for sharks (Martin and
Palumbi, 1993). Fragments <100 bp were not recorded
because the DNA ladder had a lower limit of 100 bp.
912
Fishery Bulletin 101(4)
Table 1
Summary of PCR-RFLP banding patterns for the cytochrome b (cyt 6)
region
n Carch
arhinus
spp.
Fragment
sizes are given in
number of base pairs
(bp) and have
been rounded to the nearest 25 bp.
Where the enzyme appeared
not to have cleaved the cyt h |
region, it
was scored '
1146
" n = denotes sample size.
/ = denotes fragment size
present
Carcharhinus species
C.
brevipinna
C brevipinna
Fragment
C.l
imhatus
C. brachvurus
C. leucas C. obscurus
C. falciformis
Haplotype 1
Haplotype 2
Enzyme
size
(n=9)
(;i=12)
(n=3) (
/!=29)
(I
1 = 12)
(;!=6)
(n=2)
Alu I
1000
700
600
500
450
350
300
200
/
/
•
•
/
/
/
/
/
•
•
Hae III
1100
975
750
225
/
/
/
•
/
/
/
/
175
/
/
/
/
/
Pst\
1146
975
175
/
/
•
/
/
/
/
/
Taql
1146
1100
850
650
325
300
/
/
/
/
/
/
/
/
/
Xho I
1146
850
325
/
/
/
/
•
/
/
These techniques can be used to complement morphomet-
ric identification (Chff and Wilson, 1994; Last and Stevens,
1994; Naylor and Marcus, 1994) or can be used to identify
"cryptic" species when morphological identification cannot
be done. Other "cryptic" species caught in beach meshing
and by recreational fisheries can be added to the DNA da-
tabase, such as hammerhead sharks (Sphyrna spp.) which
are commonly caught and are recorded in catch records to
genus level only (Pepperell, 1992; Reid and Krogh, 1992;
Chan, 2001). Although this project positively identified
six species from the genus Carcharhinus, other species of
this genus are known to inhabit the NSW coastline (Ste-
vens, 1984; Last and Stevens, 1994). During the warmer
months, when the northern currents extend farther south
to the Sydney region, transient tropical Carcharhinus spp.
may appear off the coast. In the northern regions of NSW,
there have been recorded catches of the blacktip reef shark
(C melanopterus) by shore-based anglers (Gartside et al.,
1999). Although transient tropical whaler sharks may not
have permanent stocks in NSW waters, it is important to
discern them from resident Carcharhinus spp. prior to
any species-specific stock assessment. Given the number
of shark species and difference in life histories (Last and
Stevens, 1994; Smith et al., 1998), identification to species
level is crucial.
The use of genetic techniques allows, for the first time,
accurate identification of species of whaler sharks that
were landed by recreational fisheries and caught in pro-
tective beach meshing in NSW and that have been his-
torically recorded to genus level. Continual sampling and
formal identification are required for comparison of catches
between species oi Carcharhinus. Genetic techniques have
the potential to be used for all other shark species and
fisheries within the Australian Fishing Zone (AFZ). The
use of genetic techniques has been employed in the field
of law enforcement to prevent the selling of protected fish
species at local fish markets where the majority of the
carcass is not retained (Ward et al., 1999). This use could
NOTE Chan et a\. Identification of Carcharhinus spp by DNA-based techniques
913
be extended to ensure that protected shark species such
as the grey nurse shark (Carcharias taurus), white shark
(Carcharodon carcharias), and the smalltooth sand tiger
shark (Odontaspis ferox) are not sold.
This project is the first time that Carcharhinus spp. have
been formally identified to species level in the 60-year his-
tory of NSW protective beach meshing and only the second
time in NSW recreational fisheries after Stevens (1984).
The depositing of voucher specimens and all DNA biopsies
at the Australian Museum ensures that these valuable and
irreplaceable biological samples can be used in future re-
search. It is evident that DNA techniques can be used to
taxonomically identify "cryptic" specimens, especially Car-
charhinus spp., and Sphyrna spp. to species level that were
once recorded to genus level only in many fisheries based
in NSW (Pepperell, 1992; Reid and Krogh, 1992; Chan,
2001; Tanner and Liggins^). It is important that sharks
that are caught be recorded to the lowest taxonomic level
for management and conservation strategies. Long-term
routine sampling and recording to species level will provide
useful data on which conservation management strategies
can be developed as part of the Australian national plan of
action for the conservation and management of sharks.
Acknowledgments
The authors would like to acknowledge the assistance of
the NSW Game Fishing Association and all NSW recre-
ational gamefishing clubs, their officials and their anglers
who cooperated with the research, NSW protective beach
meshing contractors and observers, NSW Fisheries staff,
and the numerous volunteers (Joanne Bennett, Tanya
Compton, Rikke Dano, Paul Godfrey, Gary Henry, Andrew
Hodges, Alex Irwin, Jeff Murphy, Julie Needham, Milena
Rantala, and Clint Wilson) who helped collect samples. We
thank Ed Heist and Andrew Martin for their comments on
the manuscript and specially thank Bill Sherwin (UNSW)
and Marie Roseline Yardin for their assistance in this
project. This project was funded by NSW Fisheries and
the National Heritage Trust Coast and Clean Seas' Marine
Species Protection Program (CCS Project no. 9856). Voucher
shark specimens were retained at the Australian Museum,
Sydney, NSW, Australia (Collection Manager, Fish Section)
and NSW Fisheries, NSW, Australia (Dennis Reid).
Literature cited
Chan, R.W. K.
2001. Biological studies on sharks caught off New South
Wales. Ph.D. diss., 314 p. School of Biological Science,
Univ. New South Wales, Sydney, Australia.
Cliff, G., and R. B. Wilson.
1994. Natal Sharks Board's field guide to sharks and other
marine animals, 57 p. Group Editors, Durban, South
Africa.
Dudley, S. R J.
1997. A comparison of the shark control programs of New
South Wales and Queensland (Australia) and KwaZulu-
Natal (South Africa). Ocean Coast. Manag. 34:1-27.
Gartside, D. F., B. Harrison, and B. L. Ryan.
1999. An evaluation of the use of fishing club records in the
management of marine recreational fisheries. Fish. Res.
41:47-61.
Godfrey, P.
1997. Identification of sharks caught in NSW waters using
allozyme electrophoresis. B.S. (Hons.) thesis, 54 p. Cen-
tre for Marine and Coastal Studies, Univ. New South Wales,
Sydney, Australia.
Heist, E. J., and J. R, Gold.
1999. Genetic identification of sharks in the U.S. Atlantic
large coastal shark fishery. Fish, Bull. 97:53-61.
Innes, B. H., P. M. Grewe, and R. D. Ward.
1998. PCR-based genetic identification of marlin and other
billfish. Mar. Freshw. Res. 49:383-388.
Kitamura, T., A. Takemura, S. Watabe, T. Taniuchi, and
M. Shimizu.
1996. Mitochondrial DNA analysis for the cytochrome 6 gene
and D-loop region from the bull shark Carcharhinus leucas.
Fish. Sci. 62:22-27.
Last, P. R., and J. D. Stevens.
1994. Sharks and rays of Australia, 513 p. CSIRO (Com-
monwealth Scientific and Industrial Research Organisa-
tion) Publishing, Melbourne, Australia.
Lavery, S,
1992. Electrophoretic analysis of phylogenetic relationships
among Australian carcharhinid sharks. In Sharks: biology
and fisheries (J. G. Pepperell, ed.), p. 97-108. Aust. J. Mar
Freshw. Res. 43.
Martin, A. P.
1991. Application of mitochondrial DNA sequence analysis
to the problem of species identification of sharks. In Con-
servation biology of elasmobranchs (S. Branstetter, ed.), p.
53-59. NOAA Tech. Rep. NMFS 1 15.
Martin, A. R, and S. R. Palumbi.
1993. Protein evolution in different cellular environments:
cjftochrome b in sharks and mammals. Mol. Biol. Evol. 10:
873-891.
Naylor, G. J. P., and L. F. Marcus.
1994. Identifying isolated shark teeth of the genus Carcha-
rhinus to species: relevance for tracking phyletic change
through the fossil record. Am. Mus. Novit. 94:1-53.
Pepperell, J. G.
1992. Trends in the distribution, species composition and
size of sharks caught by gamefish anglers off South-eastern
Australia, 1961-90. In Sharks: biology and fisheries (J. G.
Pepperell, ed.), p. 213-225. Aust. J. Mar Freshw. Res. 43.
Pepperell, J. G., and P. M. Grewe.
1999. A field guide to Indo-Pacific billfishes, 16 p. CSIRO
Publishing, Melbourne, Australia.
Reid, D. D., and M. Krogh.
1992. Assessment of catches from protective shark mesh-
ing off New South Wales beaches between 1950 and 1990.
In Sharks: biology and fisheries (J. G. Pepperell, ed.), p.
283-296. Aust. J. Mar Freshw. Res. 43.
Renon, P., M. M. Colombo, F. Colombo, R. Malandra, and
PA. Biondi.
2001. Computer-assisted evaluation of isoelectric focusing
patterns in electrophoretic gels: identification of smooth-
hounds (Mustelus mustelus, Mustelus asterias) and com-
parison with lower value shark species. Electrophoresis
22:1534-1538.
Shivji, M., S. Clarke, M. Pank, L. Natanson, N. Kohler, and
M. Stanhope.
2002. Genetic identification of pelagic shark body parts for
914
Fishery Bulletin 101(4)
conservation and trade monitoring. Conserv. Biol. 16:
1036-1047.
Smith, P. J., and P. G. Benson.
2001. Biochemical identification of shark fins and fillets
from the coastal fisheries in New Zealand. Fish. Bull. 99;
351-355.
Smith, S. E., D. W. Au, and C. Show.
1998. Intrinsic rebound potentials of 26 species of Pacific
sharks. Mar. Freshw. Res. 49:663-678.
Stevens, J. D.
1984. Biological observations on sharks caught by sport
fishermen off New South Wales. Aust. J. Mar Freshw.
Res. 35:573-590.
Ward, R. D., R. K. Daley, J. Andrew, and G. K. Yearsley
1999. Protein fingerprinting. In Australian seafood
handbook: an identification guide to domestic species (G.
K. Yearsley P. R. Last, and R. D. Ward, eds.), chap. 9, p.
358-392. CSIRO Marine Research, Hobart, Australia.
915
Red sea urchins (Strongylocentrotus franciscanus)
can live over 100 years: confirmation with
A-bomb carbon
Thomas A. Ebert
Department of Zoology
Oregon State University
Corvallis, Oregon 97331-2914
E mail address eberttfn'sciences oregonstate edu
John R. Southon
Center for Accelerator Mass Spectrometry
Lawrence Livermore National Laboratory
Livermore, California 94551-9900
Red sea urchins (Strongylocentrotus
franciscanus) along the west coast of
North America, like most large sea
urchins in temperate waters world-
wide, are the focus of a commercially
important fishery. In a review of bio-
logical data for purposes of fishery
management, the life span of red sea
urchins was suggested to be 7-10
years (Sloan, 1986) and they have
been included with much shorter-lived
species for illustrating complex popula-
tion dynamics (Hastings and Higgins,
1994). Recent work with tetracycline
and calcein tagging (Ebert, 1998; Ebert
et al., 1999), however, has shown that
individuals continue to grow through-
out life, although at a very slow rate,
and large individuals are estimated to
be in excess of 100 years old. A poten-
tial problem with the studies using
tetracycline and calcein is that one-
year time intervals were used between
tagging and recapture and therefore it
is possible that occasionally there may
have been very good years for growth
that were missed. If occasional growth
spurts occurred, largest sizes would
have been attained in much less than
100 years. The potential problem of
missed good years for growth could be
resolved with a marker that captures
a longer period of time. The accuracy
of age estimates has consequences
for resource management where size
limits may need adjustment in order to
protect older individuals (Hilborn and
Walters, 1992; Congdon et al., 1994;
Ebert, 1998). There is also the need to
understand the evolution of life histo-
ries of species where long life tends to
be an indicator of uncertainty in indi-
vidual reproductive success (Murphy,
1968; Roff 1992; Stearns, 1992).
Enhanced radiocarbon in the oceans
due to atmospheric testing of nuclear
weapons that began in the 1950s (Ny-
dal and Lovseth, 1983; Broecker et
al., 1985, Duffy et al., 1995) provides
a permanent marker in carbonate-
based skeletal elements that are not
reworked by resorption and deposition
during growth and hence has a long
time period between mark and recov-
ery. The enhanced radiocarbon marker
has been used in various studies to
validate the periodic (usually annual)
nature of growth zones in fish (Kalish,
1993, 1995; Campana, 1997; Campana
et al., 2002) and invertebrates (Tureki-
an et al., 1982; Witbaard et al.. 1994;
Peck and Brey, 1996) where validation
by chemical tags such as tetracycline
has been impractical. Red sea urchins
lack interpretable growth zones (Breen
and Adkins, 1976) and therefore there
is no natural feature to serve as a cross
check for studies using chemical tags.
In the present study we present a test
and confirmation of age in red sea
urchins estimated from tetracycline
tagging using an enhanced ^*C signal
in the ocean from nuclear weapons
testing.
Materials and methods
Red sea urchins were tagged with tet-
racycline from 1989 to 1992 in northern
California, Oregon, and Washington
and collected after time intervals of
approximately one year (details pre-
sented in Ebert et al., 1999). It is not
possible to determine whether a live
sea urchin has a tetracycline mark
and therefore large collections had
to be made. Skeletal elements were
cleaned with sodium hypochlorite
bleach to remove all organic material
not bound in the calcite of the skeleton,
and then skeletal ossicles were exam-
ined by using UV illumination to detect
the tetracycline marks, which fluoresce
yellow. Growth increments were mea-
sured in jaws of Aristotle's lantern of
1582 tagged-recovered red sea urchins
and used to estimate growth parame-
ters. Jaw ossicles, the demipyramids of
Aristotle's lantern, are internal skeletal
elements that grow around all surfaces
but not equally in all directions so that
a change in jaw length, AJ , is mostly
at the end closest to the esophagus
and there is little growth closest to the
mouth, the labial end, where the teeth
extend from the jaw.
The Tanaka function (Eq. 1) was used
to describe growth (Tanaka, 1982, 1988)
because it can model data that show an
initial lag, an exponential phase with a
maximum, and can include continuing
growth throughout life. This function
is described in greater detail else-
where (Tanaka, 1982, 1988; Ebert et
al., 1999). The usual formulation of the
Tanaka model is zisize as a function of
size at time t and At is assumed to be
fixed for all individuals in the sample,
usually at zit = 1 year (Tanaka, 1982,
1988; Ebert, 1998; Ebert et al., 1999)
and not included explicitly in the equa-
tion. In the present study we estimated
the amount of jaw that would have to
be removed to represent the time span
from the time of collection in the 1990s
with relatively high i*C levels to the
time before atmospheric testing of
atomic bombs (relatively low '■'O and
Manuscript approved for publication
10 July 200.3 by Scientific Editor.
Manuscript received 13 July 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101(4):915-922 (2003).
916
Fishery Bulletin 101(4)
therefore the Tanaka model was modified from previous
uses to make AJ a function ofJf^^f the size on the date of
recapture rather the date of marking, which is the usual
way of estimating growth parameters. Also, At was explic-
itly included as a variable (Eq. 2),
A/ = y„
V7
In
lG + 2^G-+fa
where
and
G = EIA-falE-fAt
E = exp{4fU„^-d})
(1)
(2)
(3)
The three parameters of the Tanaka function, a, d, and f,
have biological meaning: "a" is related to maximum growth
rate, which is approximately l^/a; "d" shifts the size at
which growth is maximum; and "/" is a measure of the rate
of change of the growth rate. A graphical presentation of
how changes in these parameters change the growth curve
is given in Ebert et al. (1999).
Explicit use of At and making AJ a function of Jf+^t
required a modification of the usual presentation of the
Tanaka function. In Ebert et al. (1999) Equation 2 was
written as
G = E/4-fa/E+f (4)
with no At and with "+ /". Equation 3 was written as
E = exp[4f{J,-d)). (5)
with Jf, rather than J^^^^. Tetracycline tagging for a period
of one year, At = 1, provides the Tanaka parameter esti-
mates and these parameters were used to estimate a Ajaw
size that would cover the time from the date of collection to
a time. At, before A-bomb testing; At is time run backwards
from the date of collection, which is the reason for the sign
change from Equation 4 to Equation 2.
The samples of red sea urchins that were selected for
radiocarbon analysis were part of the tagging study at
Halftide Rocks off San Juan Island, Washington (Ebert et
al., 1999). Individuals were tagged with tetracycline on 26
October 1991 and collected again on 21 October 1992. The
recaptured tagged individuals {n=365) are part of the 1582
tagged sea urchins from northern California, Oregon, and
Washington that were used to estimate Tanaka param-
eters. For ''*C analysis, specimens were selected from the
Halftide Rocks collection that did not show fluorescence in
the skeleton and therefore probably had not been handled
in 1991. The use of untagged individuals for radiocarbon
analyses avoids any possible contamination from handling
and tagging in 1991.
Cleaned jaws for ''*C analysis were cemented to alumi-
num blocks with a two-part epoxy cement and aligned so
that the esophageal margin was approximately parallel
with the block base. The block was held on the stage of a
small milling machine and the stage tilted so that the jaw
was as parallel as possible with the milling bit. Approxi-
mately 0.5 mm of the jaw surface was removed and sides
were milled to remove recently deposited calcite and to ex-
pose the underlying older skeleton. The jaw was measured
and successive samples were milled from the esophageal
edge to a depth of 0.5 mm, which produced samples larger
than 1 mg of carbonate in most cases. Samples were placed
in individual reaction chambers, evacuated, acidified with
orthophosphoric acid, and heated. The evolved CO2 was
converted to graphite by reduction with an excess of hydro-
gen in individual reactors with iron powder as a catalyst
(Vogel et al., 1987). Analysis of "C in the graphite targets
was done at the Center for Accelerator Mass Spectrometry,
Lawrence Livermore National Laboratory, and reported as
A^'^C7cc (Stuiver and Polach, 1977), which includes a correc-
tion for a S^^C of -3 based on stable isotope analyses. Mean
precision (1 standard deviation) of radiocarbon measure-
ments was 4.2%o (range: 3.0-7.9).
Results
Of the 1582 tag recoveries from all sites, 739 jaws showed a
growth increment, AJ, of <0.02 cm and of these only 13 had
a labial measurement >0, which is at the end of the jaw at
the mouth opening. The smallest nonzero measurements
were 0.001 cm and therefore growth less that this was
recorded as 0; 54 sea urchins in the sample had clear tetra-
cycline marks but 0 measurable growth. For large jaws, the
measured labial component was too small to be measured
and therefore all of the calculated AJ since the late 1950s
was milled from the esophageal end of the jaw only.
Tetracycline tagging indicated that annual jaw growth
(Fig. 1) was very slow for large sea urchins and many in-
dividuals showed annual increments of less than 0.01 cm.
The resulting growth curve of jaw length as a function of
age (Fig. 2A) showed that at least some large individuals
would be expected to have ages in excess of 100 years. If
this age estimate is correct, a drop in ^^C should be found
in successive small slices removed from large jaws, which
would first show current ^^C levels and then drop to pre-
bomb levels. Because the Halftide Rocks samples were col-
lected in 1992 we used At = 35 years, which would go back
to 1957. Using Equations 1-3, growth parameters given in
Fig. 1, and At = 35 years, we estimated the increment to
be between 1 and 2 mm for jaws between 2.5 and 2.6 cm
(Fig. 2B).
Successive milled samples from the esophageal ends
of large jaws (Fig. 3, A-D) showed a precipitous drop in
radiocarbon to prebomb levels over 1-2 millimeters, in
agreement with predictions. Variations across replicates
and samples probably are the result of differences in the
width of milled samples and an inability to remove all re-
cently deposited calcite or to follow the exact growing edge
of the jaw with the milling machine. Smaller jaws (Fig.
3, E-G) were not expected to show a prebomb signature,
and indeed they did not. They do, however, indicate the
^*C level to be expected in recent skeletal material and
emphasize the rapid change in radiocarbon shown in large
jaws. Changes in ^•*C in successive milled samples in jaws
NOTE Ebert and Southon: Confirmation of longevity for Strongy/ocentrotus franascanus with '""carbon
917
0.06 ■
0 04'
?
=
0 03-
t
B
0 02-
i^
□o »
" °
o
fejj^ ft •¥
» % ■(.
001 ■
^
^m
^\
0 nil ■
^?vS
Jf^^f^trr
—
2.5
3.0
Figure 1
Jaw growth increments, AJ, for tagged red sea urchins iStrongylocen-
trotus franciscanus) from northern California, Oregon, and Washing-
ton; fitted Hne is the Tanaka function (Tanaka, 1982, 1988; Ebert et
al., 1999) v/ithf= 10.95650 ±0.35064 SElstandard error), d = 0.04937
±0.01664, and a = 8.63029 ±0. 16659; M is approximately 1 year for all
samples; (A) entire data set; n = 1582; (B) restricted scale to show just
jaws larger than 2.0 cm and degree of scatter; n = 336; conversion to
1 981 7
body diameter, D, from jaw length, J, is D =4.8951J ; therefore
jaw lengths of 2.0 and 2.7 cm would have expected test diameters of
11.9 and 17.9 cm, respectively.
E-G are similar to changes shown in coral samples from
the Galapagos (Guilderson and Schrag, 1998) and may in-
dicate that '■^C levels in surface waters in regions of strong
upwelling were still rising when sea urchin were collected
in 1992. The conclusion is that '''C analysis supports the
age estimates based on tetracycline tagging and use of the
Tanaka function: large red sea urchins are old and may
have ages of 100 years or more
Discussion
The largest reported red sea urchins, with body diam-
eters over 19 cm, are from British Columbia, Canada,
(Bureau, 1996) and with estimated jaw lengths of about
2.8 cm would be expected to be around 200 years old (Fig.
2A), Age estimates of lOO-i- years far exceed estimates of
life span for other sea urchins (Table 1) based on growth
lines in ossicles. Natural growth lines, however, tend to
underestimate ages of old individuals because very small
increments will have alternations of dark and light areas
that are difficult or impossible to resolve and hence counts
underestimate age (Ebert, 1988). For example, the maxi-
mum age estimate for Strongylocentrotus droebachiensis,
the commercial species of the U.S. east coast, is 25 years
by counts of growth lines (Robinson and Maclntyre, 1997)
but at least twice this if tagging and size structure (Russell
et al., 1998) are used. Similarly, tagging and size structure
of Evechinus chloroticus (Lamare and Mladenov, 2000)
have indicated survival rates similar to S. franciscanus
but the maximum number of growth lines reported was
only 10 (Dix, 1972). Survival rate, however, is not a fixed
parameter for a species and there is local variation, as well
as geographic patterns, evident in the survival rate for S.
franciscanus (Ebert et al. 1999).
Estimates of annual survival rates based on growth pa-
rameters and mean size for red sea urchins from southern
California to Alaska (Ebert et al., 1999) indicate that very
old individuals would not be expected in southern Califor-
nia where few individuals attain ages of 50 years. At more
northern locations, the probability of long life increases
(Fig. 4) and ages of lOO-i- are expected, particularly in
Washington and Alaska. The mechanism causing the lati-
tudinal pattern are unclear. Latitudinal differences in sur-
vival may be due to increased disease outbreaks associated
with higher temperatures in the south (Ebert et al. 1999) or
the presence of more predator species in the south (Tegner,
2001). Physiological senescence related to temperature is
unlikely because there is no pattern to growth differences
associated with latitude (Ebert et al., 1999) and no evidence
for physiological decline in relative gonad size in the south
(Tegner and Levin, 1983) or north (Kramer and Nordin,
1975). The largest individuals continue to develop gonad
masses in accord with the same allometric relationships
as smaller individuals. It is reasonable to conclude that
senescence does not occur in red sea urchins.
918
Fishery Bulletin 101(4)
28
24
? 2°
S 16-
^ 12
S
en
-> 0,8 i,
0.4
00
20 40
60 80 100
Age (yr)
120 140 160
2.4 2.5
Jaw length (cm)
2.7
Figure 2
(A) Growth curve for jaws of red sea urchins from northern California, Oregon, and Washington usingTanaka parameters
and an initial jaw size of 0.44 cm, the approximate jaw size for a 1-year-old red sea urchin; (B) change in jaw size, Ajaw,
with Atime = 35 years starting with a final jaw length; Ajaw is the estimate of how much of the jaw would have to be
removed to expose prebomb calcite for sea urchins collected in 1992.
Table 1
Maximum age estimates for sea
urchins based on growth zones
in skeletal ossicles.
Species
Years
Reference
Lytechinus variegatus
4
Beddingfield and McCUntock (2000)
Strongylocentrotus nudus
6
Kawamura (1966)
Psammechinus miliaris
7
Jensen (1969)
Paracentrotus lividus
8
Crapp and Willis (1975)
Sphaerechinus granulans
9
Lumingas and Guillou (1994)
Evechmus chloroticus
10
Dix(1972)
Psammechinus miliaris
10
Gage (1991)
Strongylocentrotus intermedlus
10
Agatsuma(2001)
Echinus acutus var. norvegicus
11
Gage etal. (1986)
Loxechinus albus
11
Gebauer and Moreno (1995)
Echinus esculentus
12
Nichols etal. (1985)
Allocentrotus fragilis
15
Sumich and McCauley (1973)
Echinus elegans
21
Gage etal. (1986)
Strongylocentrotus droebachiensis
24
Robinson and Maclntyre (1997)
Echinus affinis
28
Gage and Tyler (1985)
Sterechinus neumayeri
40
Brey etal. (1995)
Sterechinus antarcticus
75
Brey (1991)
Red sea urchins larvae spend at least two months in the
plankton (Strathmann, 1978) during which time they can
be carried far along the coast or out to sea. There is year-
to-year variation in settlement and recruitment and years
of zero success and greater variation at northern sites (Ber-
nard and Miller, 1973, Low, 1975,Tegner and Dayton, 1981,
Duggins, 1983, Pearse and Hines, 1987, Sloan et al., 1987,
Ebert et al., 1994). An important point, however, is that
these authors reported some recruitment at study sites
and so extreme longevity would at first seem unnecessary
for species survival. The important issue for evolution of
life histories, however, is not whether some individuals
recruit to the population but how successful an individual
is each year in leaving offspring. The long life of adult red
sea urchins emphasizes the difficulties individuals have in
successfully having offspring that settle in suitable habitat
and survive to reproductive age. Many annual reproductive
episodes appear to be required to succeed and therefore red
sea urchins are classic bet hedgers that use resources to
promote annual survival of adults as well as to reproduce
(Stearns, 1992),
Attributes of a long life span have consequences for
resource management. The implications for management
of the red sea urchin resource have been explored by us-
NOTE Ebert and Southon: Confirmation of longevity for Stmngylocentrotus franciscanus with '''carbon
919
-130
130
0.0 0-5 1.0 1.5 2.0 2,5 0.0 0.5 1.0 1.5 2.0 2.5
Jaw length (cm)
Figure 3
Change in J C %c with successive milled samples from the esophageal end of red sea
urchin jaws collected at Halftide Rocks, Washington (48°28.8'N, 1220°59.8'W), 21 Oct.
1992. Step width in lines is the width of a milled sample. Jaws of three sea urchins are
drawn but a total of 7 were sampled; jaw lengths are A = 2.58 cm, B = 2.54, C = 2.53,
D = 2.51, E = 2.29, F = 2.01, G = 1.80 cm; primes represent replicate jaws from the same
sea urchin; jaw C is labeled just at the final milled sample (initial sample is just below
the initial sample for jaw Al; bands labeled 1 through 4 on jaw G show how samples
were milled from the jaw after the surface layer had been removed by approximately
0.5 mm. Shaded areas of the jaw, which were not sampled, indicate an area below the
top surface where a shelf exists for articulation with the epiphysis, another ossicle of
Aristotle's lantern.
ing elasticity analysis (de Kroon et al., 1986) of a matrix
model and have shown that small changes in survival of
individuals larger than 9 cm would have a greater effect
on population maintenance than survival of smaller sea
urchins (Ebert, 1998). The conclusion from matrix analysis,
which is supported by the '''C test of growth and age of
Strongylocentrotus franciscanus we present in our study, is
that the preservation of large individuals must be included
in long-term management plans for this species as well
as for other long-lived sea urchins in developing fisher-
ies such as that for Evechinus chloroticus (Barker, 2001).
Finally, our work strongly suggests that life spans of other
exploited sea urchm species should be explored in greater
detail in developing management plans because preserva-
tion of large and old individuals may be very important for
the long-term viability of these fisheries.
Acknowledgments
Tagging and processing sea urchins were done in collabora-
tion with S. Schroeter and J. Dixon, with support from the
Pacific States Fishery Commission, a self-imposed landing
tax of sea urchin fishermen administered by the Calif
Dept. Fish & Game, Oregon Sea Urchin Community Com-
mission, and Ore. State Univ. Sea Grant; field work was
facilitated by resource managers in California, Oregon, and
Washington (P. Kalvass, N. Richmond, A. Bradbury). Stable
isotope analysis was done at the University of California
at Davis and radiocarbon analysis was funded by a Center
for Accelerator Mass Spectrometry minigrant at Lawrence
Livermore National Laboratory and was carried out under
the auspices of the US DOE. The manuscript benefited
from a critical reading by G. Fox.
920
Fishery Bulletin 101(4)
1 1
,01 1
Q. .001 1
.0001
.00001
50
80
Age (yr)
100
Figure 4
Probabilities of attaining ages of 50. 80, and 100 years for red sea urchins
(Strongylocentrotus franciscanus) at sites from Alaska to northern
California; estimates were based on Tanaka growth parameters and size-
frequency distributions (Ebert et al., 1999); numbers of sites are shown
at the tops of bars and there are two samples at each site; error bars are
2SEs.
Literature cited
Agatsuma, Y.
2001. Eco\ogy o{ Strongylocentrotus intermedius. In Ed-
ible sea urchins: biology and ecology (J. M, Lawrence, ed.),
p. 333-346. Developments in aquaculture and fisheries
science, no. 32. Elsevier Science B. V., Amsterdam, Neth-
erlands.
Barker, M. F.
2001. The ecology of Evechinus chloroticus. In Edible
sea urchins: biology and ecology (J. M. Lawrence, ed.), p.
245-260. Developments in aquaculture and fisheries
science, no. 32. Elsevier Science B. V., Amsterdam, Neth-
erlands.
Beddingfield, S. 0., and J. B. McClintock.
2000. Demographic characteristics of Lytechinus variega-
tiis (Echinoidea: Echinodermata) from three habitats in a
north Florida bay Gulf of Mexico. Mar. Ecol. 21:17-40.
Bernard, E R., and D. C. Miller.
1973. Preliminary investigation on the red sea urchin
resources of British Columbia [Strongylocentrotus fran-
ciscanus (Agassiz)). Fish. Res. Board Can. Tech. Rep. 400:
1-37.
Breen, P. A., and B. E. Adkins.
1976. Growth rings and age in the red sea urchin, Strongy-
locentrotus franciscanus. Fish. Res. Board Can. Manuscr.
Rep. Ser 1413.
Brey T
1991. Population dynamics o{ Sterechinus antarcticus
(Echinodermata: Echinoidea) on the Weddell Sea shelf and
slope, Antarct. Antarctic Sci. 3:251-256.
Brey.T., J. Pearse, L. Basch, J. McClintock, and M. Slattery
1995. Growth and production of Sterechinus neumayeri
(Echinoidea: Echinodermata) in McMurdo Sound, Antarct.
Mar Biol. 124:279-292.
Broecker, W. S.. T.-H. Peng, G. Ostlund, and M. Stuiver
1985. The distribution of bomb radiocarbon in the ocean. J.
Geophys. Res. 90:6953-6970.
Bureau, D.
1996. Relationship between feeding, reproductive condition,
jaw size and density in the red sea urchin, Strongylocentro-
tus franciscanus. M.S. thesis, 90 p. Simon Eraser Univ.,
Burnaby, Canada.
Campana, S. E.
1997. Use of radiocarbon from nuclear fallout as a dated
marker in the otoliths of haddock Melanogrammus
aeglifinus. Mar Ecol. Prog. Ser 150:49-56.
Campana, S. E., L. J. Natanson, and S. MyklevoU.
2002. Bomb dating and age determination of large pelagic
sharks. Can. J. Fish. Aquat. Sci. 59:450-455.
Congdon, J. D., A. E. Dunham, and R. C. van Loben Sels.
1994. Demographics of common snapping turtles iChelydra
serpentiana): implications for conservation and manage-
ment of long-lived organisms. Am. Zool. 34:397-408.
Crapp, G. B., and M. E. Willis.
1975. Age determination in the sea urchin Paracentrotus
lividus (Lamarck), with notes on the reproductive cycle. J.
Exp. Mar Biol. Ecol. 20:157-178.
de Kroon, H., A. Plaisier, J. van Groenendael, and H. Caswell.
1986. Elasticity: the relative contribution of demographic
parameters to population growth rate. Ecology 67:
1427-1431.
Dix, T G.
1972. Biology of Evechinus chloroticus (Echinoidea: Echi-
nometridae) from different localities. 4. Age, growth, and
size. N. Z. J. Mar Freshw. Res. 6:48-68.
Duffy, P. B., D. E. Eliason, A. J. Bourgeois, and C. C. Covey.
1995. Simulation of bomb radiocarbon in two global ocean
general circulation models. J. Geophys. Res. 100:22545-
22563.
NOTE Ebert and Southon: Confirmation of longevity for Strongylocentrotus franclscanus witfi '''carbon
921
Duggins, D. O.
1983. Starfish predation and the creation of mosaic patterns
in a kelp-dominated community. Ecology 64:1610-1619.
Ebert, T. A.
1988. Calibration of natural growth lines in ossicles of two
sea urchins, Strongylocentrotus purpuratus and Echinome-
tra mathaei, using tetracycline. In Echinoderm biology:
proceedings of the sixth international echinoderm confer-
ence (R. D. Burke, P. V. Mladenov, P. Lambert, and R. L.
Parsley, eds.), p. 435-443. A. A. Balkema, Rotterdam,
Netherlands.
1998. An analysis of the importance of Allee effects in
management of the red sea urchin Strongylocentrotus
franciscanus. In Echinoderms: San Francisco. Proceed-
ings, 9th international echinoderm conference ( R. Mooi and
M. Telford, eds.), p. 619-627. A. A. Balkema, Rotterdam,
Netherlands.
Ebert, T. A., J. D. Dixon, S. C. Schroeter, P. E. Kalvass,
N. T. Richmond, W. A. Bradbury, and D. A. Woodby
1999. Growth and mortality of red sea urchins Strongylo-
centrotus franciscanus across a latitudinal gradient. Mar
Ecol. Prog. Ser 190:189-209.
Ebert, T. A., S. C. Schroeter, J. D. Dixon, and P. E. Kalvass.
1994. Settlement patterns of red and purple sea urchins
{Strongylocentrotus franciscanus and S. purpuratus) in
Cahfornia, USA. Mar Ecol. Prog. Ser 111:41-52.
Gage, J. D.
1991. Skeletal growth zones as age-markers in the sea
urchin Psammechinus miliaris. Mar Biol. 110:217-228.
Gage, J. D., and PA. Tyler
1985. Growth and recruitment of the deep-sea urchin Echi-
nus affinis. Mar Biol. 90:41-53.
Gage, J. D., P A. Tyler, and D. Nichols.
1986. Reproduction and growth of Echinus acutus var nor-
vegicus Duben & Koren and E. elegans Duben cfe Koren on
the continental slope off Scotland. J. Exp. Mar Biol. Ecol.
101:61-83.
Gebauer, P., and C. A. Moreno.
1995. Experimental validation of the growth rings of i-o.r-
echmus albus (Molina, 1782) in southern Chile (Echinoder-
mata: Echinoidea). Fish. Res. 21:423-435.
Guilderson, T P. and D. P Schrag.
1998. Abrupt shift in subsurface temperatures in the tropi-
cal Pacific associated with changes in El Nino. Science
281:240-243.
Hastings, A., and K. Higgins.
1994. Persistence of transients in spatially structured eco-
logical models. Science 263:1133-1136.
Hilborn, R., and C. J. Walters.
1992. Quantitative fisheries stock assessment: choice,
dynamics, and uncertainty, 570 p. Chapman & Hall. New
York, NY.
Jensen, M.
1969. Breeding and growth of Psammechinus miliaris
(Gmehn). Ophelia 7:65-78.
Kalish,J. M.
1993. Pre- and post-bomb radiocarbon in fish otoliths.
Earth Planet. Sci. Lett. 114:549-5.54.
1995. Radiocarbon and fish biology. In Recent develop-
ments in fish otolith research (D. H. Secor, J. M. Dean, and
S. E. Campana, eds.), p. 637-653. Univ. South Carolina
Press, Columbia, SC.
Kawamura, K.
1966. On the age determining character and growth of a
sea urchin, Strongylocentrotus nudus. Sci. Rep. Hokkaido
Fish. Exp. Station 6:56-61.
Kramer, D. E., and D. M. A. Nordin.
1975. Physical data from a study of size, weight and gonad
quality for the red sea urchin (Strongylocentrotus francis-
canus (Agassiz)) over a one-year period. Fish. Res. Board
Can. Manuscr Rep. Ser 1372:68.
Lamare, M. D., and P. V. Mladenov.
2000. Modelling somatic growth in the sea urchin Evechins
chloroticus (Echinoidea: Echinometridae). J. Exp. Mar
Biol. Ecol. 243:17-43.
Low, C. J.
1975. The effect of grouping of Strongylocentrotus francis-
canus, the giant red sea urchin, on its population biology.
Ph.D. diss., 157 p. Univ. British Columbia, Vancouver,
Canada.
Lumingas, L. J. L., and M. Guillou.
1994. Growth zones and back-calculation for the sea
urchin, Sphaerechinus granulans, from the Bay of Brest,
France. J. Mar Biol. Assoc. U K. 74:671-686.
Murphy, G. L
1968. Pattern in life history and the environment. Am.
Nat. 102:391-411.
Nichols, D., A. A. T. Sime, and G. M. Bishop.
1985. Growth in populations of the sea-urchin Echinus
esculentus L. (Echinodermata: Echinoidea) from the Eng-
lish Channel and Firth of Clyde. J. Exp. Mar. Biol. Ecol.
86:219-228.
Nydal, R., and K. Lovseth.
1983. Tracing bomb "C in the atmosphere 1962-1980. J.
Geophys. Res. 88:3621-3621.
Pearse, J. S., and A. H. Hines.
1987. Long-term population dynamics of sea urchins in a
central California kelp forest: rare recruitment and rapid
decline. Mar Ecol. Prog. Ser 39:275-283.
Peck, L.S., and TBrey
1996. Bomb signals in old Antarctic brachiopods. Nature
380:207-208.
Robinson, S. M. C, and A. D. Maclntyre.
1997. Aging and growth of the green sea urchin Bull.
Aquacul. Assoc. Can. 91:56-60.
Roff, D. A.
1992. The evolution of life histories, 535 p. Chapman &
Hall, New York, NY.
Russell, M. P, T A. Ebert, and P S. Petraitis.
1998. Field estimates of growth and mortality of the green
sea urchin, Strongylocentrotus droebachiensis. Ophelia
48:137-153.
Sloan, N. A.
1986. Red sea urchin. Underwater world. DFO/2322 UW/
53, 4 p. Fisheries and Oceans Canada, Ottawa, Canada.
Sloan, N. A., C. P. Lauridsen, and R. M. Harbo.
1987. Recruitment characteristics of the commercially har-
vested red sea urchin Strongylocentrotus franciscanus in
southern British Columbia, Canada. Fish. Res. 5:55-69.
Stearns, S. C.
1992. The evolution of life histories, 249 p. Oxford Univ
Press, New York, NY.
Strathmann, R.
1978. Length of pelagic period in echinoderms with feeding
larvae from the northwest Pacific. J. Exp. Mar Biol. Ecol.
34:23-27.
Stuiver, M., and H. A. Polach
1977. Discussion: reporting of "C data. Radiocarbon 19:
355-363.
Sumich, J. L., and J. E. McCauley
1973. Growth of a sea urchin, Allocentrotus fragilis, off the
Oregon Coast. Pac. Sci. 27:156-167.
922
Fishery Bulletin 101(4)
Tanaka, M.
1982. A new growth curve which expresses infinite increase.
Publ. Amakusa Mar Biol. Lab. 6:167-177.
1988. Eco-physiological meaning of parameters of ALOG
growth curve. Publ. Amakusa Mar Biol. Lab. 9:103-106.
Tegner, M. J.
2001. The ecology of Strongylocentrotus franciscanus and
Strongylocentrotus purpuratus. In Edible sea urchins: biol-
ogy and ecology (J. M. Lawrence, ed.), p. 307-332. Devel-
opments in aquaculture and fisheries science, no. 32.
Elsevier Science B. V., Amsterdam, Netherlands.
Tegner, M. J., and P. K. Dayton.
1981. Population structure, recruitment and mortality
of two sea urchins (Strongylocentrotus franciscanus and
S. purpuratus) in a kelp forest. Mar Ecol. Prog. Ser 5:
255-268.
Tegner, M. J., and L. A. Levin.
1983. Spiny lobsters and sea urchins: analysis of a predator-
prey interaction. J. Exp. Mar Biol. Ecol, 73:125-150.
Turekian, K. K., J. K. Cochran,. Y. Nozaki, L Thompson, and
D. S. Jones.
1982. Determination of shell deposition rates of Arctica
islandica from the New York Bight using natural '--'Ra
and 228'pi5 gfjjj bomb produced '■*€. Limnol. Oceanogr. 27:
737-741.
Vogel, J. S., J. R. Southon, and D. E. Nelson.
1987. Catalyst and binder effects in the use of filamentous
graphite for AMS. Nucl. Instru. Methods Phys. Res. Sec.
B 29:50-56.
Witbaard, R., M. L Jenness, K. van der Borg, and G. Ganssen.
1994. Verification of annual growth increments in Arctica
islandica L. from the North Sea by means of oxygen and
carbon isotopes. Neth. J. Sea. Res. 33:91-101.
923
Abundance and distribution of cetaceans
in outer continental shelf waters of the
U.S. Gulf of Mexico
Gregory L. Fulling
Keith D. Mullin
Carrie W. Hubard
Southeast Fisheries Science Center
National Marine Fisheries Service, NOAA
3209 Frederic Street
Pascagoula, Mississippi 39567
E-mail address (for G. L Fulling) Greg Fulling(a)noaa, gov
The U.S. Marine Mammal Protection
Act (MMPA) requires that stocks of
marine mammal species in U.S. waters
be maintained at or above their opti-
mum sustainable population (OSP)
level, defined as the number of animals
that results in the maximum net pro-
ductivity. To meet this requirement for
each stock, the U.S. National Marine
Fisheries Service (NMFS) estimates
annual human-caused mortality and
potential biological removal (PBR), the
maximum number of animals that may
be removed from a stock due to human
activities (e.g. fisheries bycatch) while
allowing the stock to reach or maintain
its OSP PBR is calculated by follow-
ing specific criteria and using the
estimated abundance of the stock, its
maximum net productivity rate (theo-
retical or estimated), and a recovery
factor (Barlow et al., 1995; Wade and
Angliss, 1997). The NMFS is required
to prepare an annual stock assessment
report (SAR) for each stock to update
abundance, stock structure, maximum
net productivity, human-caused mor-
tality, PBR, and status (e.g. Waring et
al.,2001).
Cetaceans in the U.S. Gulf of Mexico
(U.S. GOM) occur in two species assem-
blages that overlap in upper continen-
tal slope waters (-200-1000 m). The
oceanic waters (>200 m) are routinely
inhabited by 20 species that, in most
cases, inhabit deep warm-temperate to
tropical waters throughout the world.
Bottlenose dolphins iTursiops trim-
catus) and Atlantic spotted dolphins
[Stenella frontalis) are the only two
species commonly found in continen-
tal shelf waters (<200 m) (Mulhn and
Hansen, 1999).
In the U.S. GOM the distribution of T.
fruncatiis ranges from inshore waters
to deep waters of the continental slope
(Blaylock and Hoggard, 1994; Hansen
et al.'; Mullin and Hoggard'^). In the
U.S. GOM, the NMFS divides T. trun-
catus into 38 stocks: 33 inshore stocks
(bays, sounds, and estuaries); 3 coastal
stocks (western, northern, and eastern)
from shore to 9 km seaward of the 18-m
(10-fm) isobath; 1 outer continental
shelf (OCS) stock from the coastal
boundary to 9 km seaward of the 183-m
(100-fm) isobath; and 1 continental
shelf edge and slope stock from the
OCS boundary out to the U.S. Exclu-
sive Economic Zone (FEZ) (Waring et
al., 2001). The abundance estimate for
the OCS T. truncatus stock is 50,247
dolphins (CV=0.18) and is based on
aerial surveys conducted during fall
which covered all the U.S. GOM shelf
waters over 3 years in sections, west,
central, and east, in 1992, 1993, and
1994, respectively (Blaylock and Hog-
gard, 1994; Waring et al., 2001).
One U.S. GOM S. frontalis stock is
recognized, and the abundance, 3213
dolphins (CV=0.44), is estimated from
ship surveys of shelf edge and oceanic
waters >100 m deep conducted from
1991-94 (Hansen et al.M. Abundance
estimates for S. frontalis for the U.S.
GOM OCS were not made from the
1992-94 aerial surveys although S.
frontalis groups were sighted (War-
ing et al., 2001). The majority of S.
frontalis are thought to inhabit the
shelf-edge region. However, data from
opportunistic sightings (e.g. Mills and
Rademacher, 1996) and a summer 1994
ship survey of the eastern GOM (Hof-
stetter, 2002) have indicated that they
are common throughout eastern GOM
shelf waters >10 m deep, and in oceanic
waters <500 m.
The NMFS Southeast Fisheries Sci-
ence Center (SEFSC) conducts annual
spring and fall ichthyoplankton sur-
veys in the U.S. GOM. The spring sur-
vey targets the entire oceanic portion
of the U.S. GOM, and the fall survey
focuses on shelf waters from the U.S.-
Mexico border to southern Florida.
Since 1991, abundance estimates of
oceanic cetacean species in the U.S.
GOM have been based primarily on
data collected during annual spring
surveys (Hansen et al.'; Mullin and
Hoggard'^; Mullin and Fulling^). Be-
cause of the lack of current assessment
information on and the uncertainty of
abundance estimates for T. truncatus
and S. frontalis in OCS waters, ceta-
cean surveys were conducted during
the fall ichthyoplankton surveys from
1998 to 2001. From these surveys, we
report the abundance and distribution
of cetaceans in OCS waters (20-200 m
deep) ofthe U.S. GOM.
1 Hansen, L. J., K. D. Mullin, and C. L.
Roden. 1995. Unpublished report. Es-
timates of cetacean abundance in the
northern Gulf of Mexico from vessel sur-
veys, 20 p. Southeast Fisheries Science
Center, 3209 Frederic St., Pascagoula, MS
.39.567.
2 Mullin, K.D., and W. Hoggard. 2000. Vi-
sual surveys of cetaceans and sea turtles
from aircraft and ships. In Cetaceans,
sea turtles and seabirds in the northern
Gulf of Mexico: Distribution, abundance
and habitat associations. Volume II: Tech-
nical report (R.W. Davis, W. E. Evans, and
D. Wursig, eds.), p. 111-172. OCS Study
MMS 96-0027. Minerals Management
Service, Gulf of Mexico OCS Region, New
Orleans, LA. 70123.
3 Mullin, K.D., and G.L. Fulling. 2003. Un-
published report. Abundance of ceta-
ceans in the oceanic northern Gulf of
Mexico, 1996-2001, 35 p. Southeast
Fisheries Science Center, 3209 Frederic
Street, Pascagoula, MS 39567.
Manuscript approved for publication
10 July 2003 by Scientific Editor.
Manuscript received 20 July 2003 at
NMFS Scientific Publications Office.
Fish. Bull. 101:923-932 (2003).
924
Fishery Bulletin 101(4)
i 28-
Figure 1
Survey effort in Beaufort sea state of <3 (dark lines), east (1342 km) and west (2202 km) of Mobile Bay,
Alabama (bgld vertical line), in the northern U.S. Gulf of Mexico outer continental shelf (20-200 m) during
fall 1998-2001. The 20- and 200-m isobaths (thin lines) are shown.
Methods
Study area
The study area (245,800 km^) included continental shelf
waters of the U.S. GOM between the U.S-Mexico border and
Key West, Florida, between the 20- and 200-m isobaths ( Fig.
1). However, survey effort did not extend south of 26.0°N in
the southeastern GOM and therefore abundance estimates
were extrapolated for this region. The shelf is wide (up to
200 km) off the Florida peninsula and off northern Texas
and Louisiana, and narrower off the Florida Panhandle
near DeSoto Canyon, the Mississippi River Delta, and
southern Texas. The continental slope is a steep escarp-
ment from 1000 to 2000 m in the eastern GOM.
Survey design and data collection
Surveys were conducted from the 68-m NOAA Ship Gordon
Gunter (1998, 1999, and 2001) and the 52-m NOAA Ship
Oregon II (2000). The four surveys ranged from 28 to 32
days between 28 August and 2 October and were divided
into two legs of 12 to 19 days. Standard ship-based, line-
transect survey methods for cetaceans, similar to those
used in the Pacific Ocean and U.S. GOM, were used (e.g.
Barlow, 1995; Hansen et al.'*). Surveys were conducted
24 hours a day along a predetermined trackline between
plankton stations uniformly spaced 30 nmi apart. The
trackline uniformly covered the shelf waters roughly
10-200 m deep in 1998-2001 (Fig. 1).
■• Hansen, L. J., K. D. MuUin, T. A. Jefferson, and G. P. Scott.
1996. Visual surveys aboard ships and aircraft. In Distribu-
tion and abundance of marine mammals in the north-central
and western Gulf of Mexico: Final report. Volume II: Techni-
cal report (R.W. Davis and G.S. Fargion, eds.), p. 55-132. OCS
Study MMS 96-0027. Minerals Management Service, Gulf of
Mexico OCS Region, New Orleans, LA. 70123.
Data were collected by two teams of three observer — one
team positioned on the flying bridge 14.5 m above the wa-
terline (Gunter) and the other team positioned 9.2 m above
the waterline (Oregon II) during daylight hours while the
vessels moved between plankton stations, weather permit-
ting (i.e. no rain, Beaufort sea state <6). Each team had at
least two members experienced in ship-based, line-transect
methods and in identification of tropical cetaceans. The
left- and right-side observers searched to the horizon in the
arc from 10° right and left of the ship's bow to the left and
right beams (90°), respectively, using 25x binoculars. The
third observer searched, using unaided eye or 7x hand-held
binoculars, and recorded data. Observers changed position
every 30-40 minutes, and the two teams alternated 2-h
watches throughout daylight hours. Survey speed was usu-
ally 18 km/h (-10 knots) but varied with sea conditions.
Data were recorded on a computer interfaced with a glob-
al positioning system (GPS) by an in-house BASIC data
acquisition program (Southeast Fisheries Science Center,
NMFS, Pascagoula, MS). For each cetacean sighting, time,
position, bearing and reticle (a measure of radial distance)
of the sighting, species, group-size, behavior, bottom depth,
sea surface temperature, and associated animals (e.g. sea-
birds, fish) were recorded. The bearing and radial distance
for groups sighted without 25x binoculars and close to the
ship were estimated. Survey effort data were automatically
recorded every 2 minutes and included the ship's position
and direction, effort status, observer positions, and envi-
ronmental conditions that could affect the observers' abil-
ity to sight animals (e.g. Beaufort sea state, sun position).
Typically, if a sighting was within a 5.5-km strip on either
side of the ship, the ship was diverted from the trackline
to approach the group to allow the observers to identify
species and estimate group-size by consensus.
Cetaceans were identified to the lowest taxonomic
level possible. Observers' ability to make identifications
depended on weather and animal behavior. Differences
between T. truncatus and S. frontalis could not always be
NOTE Fulling et al : Abundance and distribution of cetaceans in the U.S. Gulf of Mexico
925
distinguished at long distances and were therefore some-
times recorded as "T. truncatus + S. frontalis."
Analytical techniques
Survey effort that was parallel to the bathymetry gradi-
ents, occurred in waters outside the OCS study area, or
occurred in a Beaufort sea state >4 was excluded from
analyses (Fig. 1). Survey effort used in analyses is sum-
marized in Table 1. Survey effort was not uniformly dis-
tributed throughout the study area due to poor survey
conditions, particularly in the eastern GOM, during two of
the four years. Because S. frontalis sightings were clearly
more numerous in the east, we delineated the study area
into west (106,186 km^) and east (139,614 km^) regions at
88°1.5.0'W (ca. Mobile Bay, Alabama) and estimated abun-
dances separately for each region. A combination of line-
transect and strip-transect methods were used to make
abundance estimates. Line-transect methods were used for
sightings detected with 25( binoculars, which constituted
the majority of sightings (129/140). Strip-transect methods
were used for the 11 sightings that were made without
the 25x binoculars (naked-eye sightings) and that were
observed by the primary team.
Line-transect estimates
For each species or species group (i) [i.e. T. truncatus, S.
frontalis, rough-toothed dolphins iSteno bredanensis) and
T. truncatus+S. frontalis] detected by 25x binoculars, and
for each region (J) (east and west), abundance estimates
were made with line-transect methods (Af^, j) by using the
software program DISTANCE (Colorado Coop. Fish and
Wildlife Research Unit, Colorado State Univ., Fort Collins,
CO) (Laake et al., 1993; Buckland et al., 2001 ) and by mcor-
porating data into the following equation:
A',,
2L-g{0)
(1)
where A^ = area of region j;
n^ = number of group sightings of species; in region
7;
^Li I ~ niean group size of species i in region j;
f (0) = sighting probability density function at per-
pendicular distance zero for species i;
L = total length of transect line in region j; and
g(0) = probability of seeing a group on the transect
line.
The parameter g(0) was not estimated; g(0) = 1 was used
for each abundance estimate. Abundances were negatively
biased because observers usually miss some groups at the
surface on the transect line, and some groups were under
the surface while in the observation area, therefore g(0) <1
(see "Discussion" section). The log-normal 95% confidence
interval was computed (Buckland et al., 2001) for each
abundance estimate because it was a product of estimates
and tended to have a skewed distribution. The variance of
^Li ^^^ estimated by using
Table 1
Total survey
efTort (km) during
1998-
-2001
in
waters
20-200
m
and
under Beaufort sea
state conditions <3.
Year
West
East
Total
1998
174
67
241
1999
477
120
597
2000
281
0
281
2001
448
629
1077
Total
1380
816
2196
var(A',,^) = A'-,^
var(/i,, ) var(S,,
^l.
) van
- +
[/(O)]
f,(or
(2)
The sampling unit was the length of the transect completed
on-effort each day with Beaufort sea state <3 in a region.
The formula used to estimate each component of the vari-
ance is given in Buckland et al. (2001). Varin^ , ^) was
length-weighted and based on the variation in the number
of on-effort group sightings between sampling units that
ranged up to 191 km/d.
Estimation of f(0)
The perpendicular distance (y ) was estimated by using bear-
ing and reticle measurements. The reticle readings were
converted to radial sighting distances {R) by the method of
Lerczak and Hobbs ( 1998;y=R sin(6), where 6=angle between
the sighting and the transect line). Because of the difference
in observer height (5.3 m) between the Oregon II and Gunter,
each ship could potentially yield a different sighting function,
g(x). However, only seven sightings were made in sea states
<3 from the Oregon II during the one year it was used; there-
fore data from both ships were pooled. Estimates of/'/O) were
made by using a hazard-rate, uniform, or half-normal model
with exact perpendicular sighting distances and no adjust-
ments. Model selection was determined by using Akaike's
Information Criterion (AIC; Buckland et al., 2001).
The number of S. bredanensis groups and the number of
T. truncatus+S. frontalis groups sighted was insufficient
to estimate /ID) for each. Because the S. bredanensis group
and T. truncatus+S. frontalis group had similar sighting
characteristics (e.g. body size, group-size, surface behavior),
we pooled them with sightings of T. truncatus to estimate
f'iO). Total number of sightings for both T. truncatus and
S. frontalis was sufficient to estimate fiO) for each without
pooling with other species. Truncation for T. truncatus, S.
bredanensis, and T. truncatus + S. frontalis was 3300 m,
and was 5000 m for S. frontalis. Each estimate of /"/O) was
based on pooled sightings from the east and west regions.
Estimation of mean group-size
Group-sizes tend to be related toy, because in many cases
larger groups are easier to see than small groups with
926
Fishery Bulletin 101(4)
increasing^'. In general, the arithmetic mean of group-size
may be an overestimate of the true mean group-size and
could lead to positively-biased abundance estimates. There-
fore, a regression of group-size by j' was used to estimate an
"expected mean group-size" (program DISTANCE) and it
was used if the regression was significant (P<0. 15). VarCS^^)
was the analytical variance for mean group-sizes based on
arithmetic means or was estimated as in Buckland et al.
(2001:74) for expected mean group-sizes.
Strip-transect estimates
One requirement for unbiased line-transect estimates of
abundance is that the cetacean group should not move in
response to the ship before it is sighted (Buckland et al.,
2001). If cetaceans are not sighted before they respond
to the ship, in cases of attraction to the ship, /TO) and
abundance will be overestimated. During previous U.S.
GOM surveys, groups of T. truncatus or S. frontalis were
consistently attracted to ride the bow waves as the ship
approached (Wiirsig et al., 1998). Therefore, the abundance
and variance of groups sighted by naked eye (Ng) were
estimated by
N.,
2-L-w,
and
var(/V„J = N=.^
var(/i^,J ^ ^ar(-^5,.;)
(3)
(4)
where w^ = l//",(0) which was treated as a constant, i.e. strip
width, w^ , was equal to the line-transect effec-
tive strip half-width [l//](0)] with vardu^) = 0.
For each region, species total abundance (Nj.^ ) was the
line-transect and strip-transect estimates added, iV^.^ =
Nj^, j + Ns,,j. Total U.S. GOM OCS abundance for each
species was Nj,^ = 27V j., ^. The coefficient of variation (CV)
for each abundance was estimated as CViN) = [vartAOl'^^W
and the CV for each summed abundance as
CV(/V,„„)
■(I
CV'{N)N
r/1
N.
(5)
Results
Abundance estimates were based on 2196 km of effort and
140 sightings (Figs. 1 and 2). For east and west regions,
there was 816 km of effort and 73 sightings, and 1380 km of
effort and 67 sightings, respectively (Tables 1 and 2). Only
three cetacean species were encountered. Groups of J! trun-
catus (30 east region, 45 west region) and S. frontalis (34
east, 12 west) were the most frequently encountered (Fig.
2, Table 2) and S. bredanensis groups ( 1 east, 2 west) were
also sighted. Tiirsiops truncatus and S. frontalis were esti-
mated to have flO) of 0.6238/km (CV=0.12) and 0.4101/km
(CV=0.11), and an effective strip half-width of 1603 and
2438 m, respectively (Figs. 3 and 4). Steno bredanensis and
T. truncatus+S. fron talis abundances were based on an /! 0 ) =
0.6059/km(CV=0.11) and an effective strip half-width of
1650 m.
Mean group-sizes (from 25x binocular sightings) of T.
truncatus for east (9.8, 0.25) and west (10.0, 0.18) regions
were similar (Table 2), and had an overall range of 1-68
animals. The mean group size of S. frontalis was larger
in the east (24.3, 0.19) than the west (15.6, 0.21) with an
overall range of 1-267 animals. Group-sizes of S. bredanen-
sis were 8, 11, and 20 animals. The east mean group-size
for both T. truncatus and S. frontalis is the size-biased
or expected mean group-size because the expected mean
was significantly smaller that the arithmetic mean, 10.9
(P=0.14) and 31.9 (P=0.08), respectively
The most abundant species (number of individuals; CV)
found in U.S. GOM OCS waters was S. frontalis (30,772;
0.27); the vast majority (91%) occurring in the east
(27,997; 0.29). The density of S. frontalis was about eight
times greater in the east compared to the west (20.1 and
2.6 dolphins/100 km^, respectively). The abundance of T.
truncatus was 25,320 (0.26); there was greater abundance
in the east (15,198; 0.34) than m the west (10,122; 0.29)
but with similar densities (10.9 and 9.5 dolphins/100 km^,
respectively). The total OCS abundance of S. bredanensis
was 1238 (0.65), and that of 7: truncatus+S. frontalis, 1868
(0.37).
Discussion
Both T. truncatus and S. frontalis occur in northern GOM
waters outside the OCS (i.e. waters <20 m or >200 m).
About 23,000 T. truncatus inhabit inshore and coastal
waters (<20 m) (Waring et al., 2001) and nearly 3000 occur
in oceanic waters (Mullin and Fulling-). Both the "coastal"
and "offshore" ecotypes of T! truncatus (Hersh and Duffield,
1990) occur in the northern GOM (LeDuc and Curry 1998).
How these ecotypes are distributed in the northern GOM
and western North Atlantic is being investigated from skin
biopsy samples collected, in part, during the 1998-2001
OCS surveys. Using mitochondrial DNA, obtained from
biopsy samples collected during a U.S. Atlantic ship survey,
Torres et al. (2003) reported no offshore form was sampled
within 6 km of shore and no coastal from was sampled
beyond 39 km from shore or in waters >34 m deep. Forty-
seven percent (35/75) of the GOM OCS T. truncatus groups
were in waters >34 m deep.
Ship surveys of northern GOM waters indicate that very
few S. frontalis (<500 animals) occur in oceanic waters, and
those that do are usually found close to the shelf edge in
waters <500 m deep (Davis et al, 1998; Mullin and Full-
ing2). The smaller "offshore" or "Gulfstream" S. frontalis
that occurs in parts of the oceanic Atlantic (Perrin, 2002)
has not been recorded for the northern GOM. During the
1998-2001 surveys, S. frontalis was sighted in waters <20
m deep. However, because sampling was not perpendicular
to bathymetry, abundance estimates were not calculated.
This species is not known to occur in U.S. GOM inshore
waters (Mullin and Hansen, 1999).
NOTE Fulling et al.: Abundance and distribution of cetaceans in the US Gulf of Mexico
927
3 28-
30-
28-
26-
98 96
Figure 2
Locations of all on-effort sightings of Turswps truncatus (top), Stenelta frontalis
(center), Steno bredanensis, and T. truncatus+S. frontalis (bottom) in the northern U.S.
Gulf of Mexico outer continental shelf (20-200 m) during fall 1998-2001. Numbers of
sightings shown are prior to truncation. The 20- and 200-m isobaths (thin lines) are
shown (AL=Alabama, FL=Florida, LA=Louisiana, MS=Mississippi, TX=Texas).
Abundance
The abundance estimates for cetaceans reported in the
present study are the first ship-based estimates for the U.S.
GOM OCS. Abimdance estimates for T. truncatus on the OCS
(25,320; 0.26) are half the estimate in the pre-2002 SARs
(50,247; 0.18) (e.g. Waring et al, 2001), that were based on
aerial surveys conducted during fall 1992-94 (Blaylock et
al, 1994; Waring et al, 2001). The abundance estimate for
S. frontalis for the entire U.S. GOM in SARs prior to 2002
(3,213; 0.44) was based on data from ship surveys of OCS and
oceanic waters >100 m deep (Waring et al., 2001; Hansen et
al.'). Our current abundance estimate of S. frontalis (30,772;
0.27) for the OCS is almost an order of magnitude larger
During the 1991-94 aerial surveys, there were 13 sight-
ings of S. frontalis groups and 10 sightings that were iden-
tified as T. truncatus+S. frontalis in OCS waters (SEFSC,
NMFS, Pascagoula, MS, unpubl. data). Using these sight-
ings and 139 T. truncatus sightings to estimate /(O), we esti-
mated the abundance of S. frontalis from the aerial survey
data to be 14,866 ( 0.37) for the U.S. GOM OCS [west, 3,526
(0.86); east, 11,340(0.40)].
There are several potential reasons for the differences in
abundances of the two species from ship and aerial surveys.
The U.S. GOM OCS east of 85.5°W makes up about 44%
of the U.S. GOM OCS. Aerial survey abundance estimates
in this area were based on a small number of transect
lines grouped in two places and most of the area was not
928
Fishery Bulletin 101(4)
Table 2
Group-size, density and abundance estimates of cetaceans in the northern U.S.
Gulf of Mexico
outer continental shelf (waters
20-200 m deep) during fall 1998-
-2001 (n =
number of
group sightings, 5=
:mean
group-size, Z)=animals/100 km-
Niabundance esti-
mate, CV=coefficient of variation
. R=reticl
e sightings
and K=naked eye sightings
.
Species and stratum
;;
S
CV(S)
D
N
cv^N)
95"r CI
T\irsiops truncatus
East-R
27
9.8
0.25
10.1
14,132
0.40
6426-31,082
East-K
3
6.7
0.70
0.8
1066
0.85
139-8182
East total
30
10.9
15,198
0.38
7442-31,035
West-R
41
10.0
0.18
9.2
9786
0.30
5502-17,406
West-K
4
3.5
0.34
0.3
336
0.67
94-1201
West total
45
9.5
10,122
0.29
5790-17,696
OCS total
75
10.3
25,320
0.26
15,457-41.478
Stenella frontalis
East-R
32
24.3
0.19
19.5
27.226
0.30
15,09.3-49,113
East-K
2
11.0
0.09
0.6
771
0.55
252-2358
East total
34
20.1
27,997
0.29
15,978-49,057
West-R
11
15.6
0.21
2.6
2712
0.42
1192-6169
West-K
1
4
—
0.6
63
1.17
9-433
West total
12
2.6
2775
0.41
1279-6023
OCS total
46
12.5
30.772
0.27
18,418-51,412
Steno bredanensis
East-R
0
—
—
—
0
—
—
East-K
1
11
—
0.4
586
0.85
118-2902
East total
1
0.4
586
West-R
2
14
0.43
0.6
652
0.98
115-3715
West-K
0
—
—
—
0
—
—
West total
2
0.6
652
OCS total
3
0.5
1238
0.65
384-3990
T. truncatiis+S. frontalis
East-R
8
2.4
0.22
0.7
983
0,57
324-2983
East-K
0
—
—
—
0
—
—
East total
8
0.7
983
West-R
8
4.8
0.28
0.8
885
0.47
355-2207
West-K
0
—
—
—
0
—
—
West total
8
0.8
885
OCS Total
16
0.8
1868
0.37
920-3793
surveyed (see Fig.l in Baumgartner [1997]). Complete
coverage would have certainly led to more S. frontalis
sightings and it is possible the lines that were surveyed
were in areas with more T. triuicatus. Blaylock and Hog-
gard (1994) estimated from aerial surveys that about 31%
of the T. truncatus in OCS waters west of Mobile Bay were
in rather a small area from the Mississippi River Delta
west to about 90.5°W. Our ship survey effort in this area
was small and resulted in only one sighting of T! truncatus
(Fig. 2). Therefore, our ship-based estimates may have un-
derestimated the abundance ofT. truncatus in the western
OCS. Aerial abundances were based on survey lines that
extended from 9.3 km past the 18 m (10 fm) curve to 9.3
km past 183 m ( 100 fm) curve; therefore the area surveyed
was somewhat different than our 20-200 m OCS study
area for ship surveys. Aerial survey effort in waters >200 m
may have resulted in more sightings ofT. truncatus than
S. frontalis because the deeper waters are not the common
habitat of S. frontalis (Mullin and Fulling'^) and sightings
in waters <20 m would have also been biased toward T.
truncatus.
Stenella frontalis and T. truncatus are similar in length
and shape. Stenella frontalis are born without spots and
become progressively more spotted with age, but young ani-
mals look very similar to T. truncatus (see Herzing, 1997).
Therefore, depending on the composition of the group, from
a distance S. frontalis are not always easily distinguished
from T. truncatus: therefore it is possible that some groups
were misidentified as T. truncatus during aerial sui-veys,
leading to bias in the relative abundance of each species.
The annual PBR for the OCS stock of T. truncatus was
432 dolphins, and for the U.S. GOM stock of S. frontalis,
NOTE Fulling et a\ Abundance and distribution of cetaceans in the US Gulf of Mexico
929
1 27713
1 14942
1 0217
Tui slops truiicatus
B 089399
CD
n
2 0766277
Q.
c
9 0 638564
o
OJ
■S 0 510851
Q
\
.
0 383138
0 255426
0 127713
\.
0
500
1000 1500 2000 2500 3000 3500
Perpendicular distance in meters
Figure 3
Plot of the detection
funct
ion of pooled sightings ofTursiops truncatus in the northern U.S.
Gulf of Mexico.
Stenella frontalis
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Perpendicular distance in meters
Figure 4
Plot of the detection function of pooled sightings of Stenella frontalis in the northern U.S.
Gulf of Mexico.
23 dolphins (Waring et al., 2001). Using the abundances,
we estimated that the annual PBR would be 204 dolphins
for T. truncatus and 246 dolphins for S. frontalis (Table 2).
Although these changes in both PBRs are large, the annual
fishery-related mortality and serious injury for each spe-
cies is estimated to be <3 dolphins in the U.S. GOM OCS
(Waring etal., 2001).
they were achieved after four years of effort. In cases where
there is human-caused mortality in a cetacean stock, abun-
dance estimates with a CV < 0.50 are generally required
to avoid incorrectly classifying a cetacean stock as "stra-
tegic" under the U.S. MMPA (i.e. annual human-caused
mortality>annual PBR) less than 10% of the time (Wade
and DeMaster, 1999).
Precision
The precision of the abundance estimates for T. truncatus
(CV=0.26) and S. frontalis (CV=0.27) was good, although
Bias
The surveys were designed to meet the assumptions of line-
transect theory (Buckland et al., 2001). However, the abun-
930
Fishery Bulletin 101(4)
dance estimates were negatively biased because the central
assumption that all cetacean groups on the transect line are
detected (i.e.^(G)=l), certainly was not met, and data were
not collected to correct estimates for perception and avail-
abihty bias (Marsh and Sinclair, 1989). Barlow (1995) esti-
mated perception bias in a ship survey in the Pacific Ocean,
and although the group-sizes were not estimated at close
range, the majority of groups missed by the primary team
were apparently small groups. From this. Barlow (1995)
estimated g(0) to range from 0.73 and 0.79 for small groups
of delphinids (<21 animals). Delphinids have relatively short
dive-cycles but diving synchrony among members of a group
can affect availability bias; if dives are asynchronous, the
probability that at least one animal will be at the surface
increases with group-size. Because availability bias varies
by species due to differences in individual dive cycles, group
diving behavior, and group-sizes, we were not able to address
this potential bias based on Barlow's ( 1995) results.
The use of the effective strip half-width [l//",(0)] from
the 25x binocular sightings for the strip width for the
strip-transect estimates (Table 2) was assumed to be con-
servative and somewhat negatively biased. The distance
from which animals will come to the ship to ride the bow
is unknown and variable, depending on factors such as
the animals' previous behavior, number of bowriding op-
portunities, and the type of ship. If the strip width was too
narrow, the strip-transect estimates of abundance would
be positively biased.
Our abundance estimates were for the entire U.S. OCS,
but the surveys did not extend south of 26.0°N in the
eastern Gulf Sightings from a 1994 survey of the eastern
Gulf (Hoffstetter, 2002) indicated that the distribution of
T. truncatus and S. frontalis does not change dramatically
between 26.0°N and Key West; therefore we believe this
potential bias is minimal.
Because our estimates are from four combined years, an-
other source of bias would occur if there were annual shifts
in cetacean distribution, that is, if the majority of animals
of any species occurred in a different part of the OCS in
one year during fall compared to others years. However,
there was no indication that this variation in distribution
occurred and therefore potential bias is probably minimal.
Potential bias due to the seasonality of the survey is also
possible but cannot currently be addressed.
Additionally, survey effort from the 2001 cruise was the
most complete effort of all years and may have carried more
weight than all the other cruises. However, the 2001 survey
provided adequate eastern GOM coverage. Variable survey
effort in the fall is common because tropical weather can
create rough sea conditions. Additionally, fall surveys al-
ways began in the west and terminated in the east. Because
the same cruise track was always followed, we rarely had
the opportunity to survey those areas not surveyed previ-
ously during nighttime transit, and thus may have created
both a spatial and temporal bias.
Distribution
The observed distributions of both T. truncatus and
S. frontalis were not surprising given previous descriptions
of their distributions. The greater number of S. frontalis in
the U.S. GOM off Florida compared to the western GOM
was suggested by Schmidley and Melcher (1974), and the
distribution of sightings reported by Mills and Rademacher
(1996) supported this finding. The density of S. fi-ontalis
was much greater in the eastern GOM OCS than the west-
ern GOM OCS but the density of T! truncatus was similar
in the two regions (Table 2).
The West Florida Shelf and Texas-Louisiana Shelf are
very different marine environments, but how habitat dif-
ferences specifically affect cetacean density patterns is not
clear. The oceanography of the U.S. GOM continental shelf
is complex, variable both spatially and temporally, and dif-
ficult to characterize briefly. Nevertheless, there are some
clear distinctions between eastern and western OCS. First,
there are 3415 active oil and gas platforms in the U.S. GOM
OCS (0-200 m); the vast majority of these platforms (with
their attendant boat and helicopter traffic) occur in waters
west of Mobile Bay (MMS^). Also, -95% of the U.S. GOM
fisheries landings by weight occur west of Mobile Bay (10
years of NMFS^ data). Additionally, sediment- and nutri-
ent-laden fresh water from the Mississippi River and its
distributary, the Atchafalaya River, usually moves west and
predominately affects the Texas-Louisiana and Mississippi-
Alabama shelves. The bottom of the Texas-Louisiana Shelf
is primarily clay-slit mud and sand, and that of the West
Florida Shelf is a mosaic of sand, gravel, shell, and coral
(Rabalais et al., 1999). Primary production associated with
the Mississippi River outfiow is the highest measured in the
GOM (Lohrenz et al., 1999). However, productivity on the
West Florida Shelf can be enhanced by a variety processes
(e.g. Gilbes et al., 1996). The deep eastern GOM is subject
to the quasi-annual incursion of the Loop Current, which
can extend to the Mississippi-Alabama Shelf (Wiseman
and Sturges, 1999). This incursion can lead to upwelling
episodes along the Loop Current front that may increase
productivity along the shelf edge and on the West Florida
Shelf (Paluszkiewicz et al., 1983; Gilbes et al., 1996).
Baumgartner et al. (2001) reported greater sighting rates
of cetaceans in the eastern GOM shelf-edge and oceanic wa-
ters and suggested that greater feeding opportunities may
occur because of the influence of the Loop Current. Griffin
and Griffin (2003), whose study included coastal waters
(<20 m), reported that S. frontalis on the West Florida Shelf
was found in deeper, more saline, and less turbid water than
those where T. truncatus was found.
Demersal fish (e.g. sciaenids) are abundant and diverse
on the western GOM OCS, but less abundant on the east-
ern OCS (Darnell et al"; Darnell et al.^). The known prey of
^ Mineral Management Service, Gulf of Mexico Region website:
http://www.gomr.mms.gove/hompg/fastfactsAVaterDepthAVater
Depth.html. [Accessed on 7/8/2003.]
^ National Marine Fisheries Service web site: http://www.
st.nmfs.gov/stl/commercial/. [Accessed on 8 July 2003.]
■ Darnell, R. M., R. E. Defenbaugh, and D. Moore. 1983. North-
western Gulf shelf bio-atlas; a study of the distribution of
demersal fishes and penaeid shrimp of the soft bottoms of the con-
tinental shelf from the Rio Grande to the Mississippi River Delta.
Open File Report No. 82-04, 438 p. Minerals Management Ser-
vice, Gulf of Mexico OCS Region, New Orleans, LA 70123.
NOTE Fulling et al : Abundance and distribution of cetaceans in the US, Gulf of Mexico
931
T. trunccitus from the GOM consist primarily of demersals,
at least close to shore, but they also prey on pelagic spe-
cies (Barros and Odell, 1990). The prey of S. frontalis are
not well characterized but descriptions include epipelagic
and mesopelagic fish and squid, and benthic invertebrates
(Perrin, 2002). Richard and Barbeau ( 1994) observed "spot-
ted dolphins" feeding on flyingfish (Exocoetidae) in waters
28-35 m deep on the West Florida Shelf This is not uncom-
mon because S. frontalis have been routinely observed feed-
ing on flyingfish at night during haulback of longline gear
during NMFS fisheries assessment surveys (Grace^). Fertl
and Wiirsig ( 1995) describe S. frontalis feeding on a school
of small clupeid fish at the surface south of the Florida
Panhandle. A S. frontalis satellite-tracked for 24 days
off Texas stayed in waters 12-63 m deep (mean, 32.6 m)
and 58.1% of its dives were <10 m (Davis et al., 1996).
These shallow dives observed by Davis et al. may indicate
feeding on epipelagic species.
The occurrence of S. bredanensis in continental shelf
waters of the U.S. GOM is interesting because this spe-
cies is usually described as inhabiting oceanic waters (e.g.
Jefferson, 2002). In the northern GOM, the estimated
density of S. bredanensis was larger in OCS waters during
fall (0.50 dolphins/100 km^; Table 2) than that estimated
for oceanic waters during spring (0.32 dolphins/100 km'-)
(MuUin and Fulling-). In fact, if there is no OCS-oceanic
shift in distribution between spring and fall, there may be
similar numbers of S. bredanensis in northern GOM shelf
waters (1238; 0.65) as in oceanic waters (1231; 0.45). One
of the groups sighted in OCS waters was near the shelf-
edge 1 183 m) but the other two sightings were at depths of
31 m and 33 m off Texas (Fig. 2). The use of shelf waters
in the U.S. GOM by this species may not be atypical; two
sightings of S. bredanensis were made on the West Florida
Shelf in waters <55 m deep during August 1994 (Hofstetter,
2002). Pitman and Stinchomb (2002) provide evidence that
S. bredanensis may be specialized predators of dolphinfish
[Coryphaena hippurus) in the Pacific Ocean. Dolphinfishes
have a circumtropical distribution but occur in oceanic and
shelf waters in the northern GOM commonly associated
with Sargassiim and other drifting materials (Hoese and
Moore, 1998). Steno bredanensis in the northern GOM are
commonly found near flotsam, as they are in the Pacific — a
place where dolphinfish tend to aggregate.
The abundance estimates presented in this study are
the first ship-based estimates of T. tursiops and S. fron-
talis from Gulf of Mexico OCS waters. Although probably
negatively biased, these estimates provide reliable data for
the management of these species. Our results suggest that
the diverse U.S. GOM environments provide an excellent
natural experiment and opportunity to further understand
» Darnell, R. M., J. A. Kleypas, and R. E. Defenbaugh. 1987.
Eastern Gulf shelf bio-atlas; a study of the distribution of
demersal fishes and penaeid shrimp of the soft bottoms of the
continental shelf from the Mississippi River Delta to the Florida
Keys. OCS Study 86-004 1 , 548 p. Mmerals Management Ser-
vice, Gulf of Mexico OCS Region, New Orleans, LA 70123.
^ Grace, M. A. 2003. Personal commun. NOAA, 3209 Frederic
Street, Pascagoula, MS 39567.
the ecology of these sympatric cetacean species in OCS pe-
lagic waters.
Acknowledgments
Many people made significant contributions to the success
of the surveys including the officers and crews of NOAA
ships Gordon Gunter and Oregon II and the field party
chiefs. The marine mammal observers were H. Adams, N.
Baertlein, C. Brown, J. Brusher, C, Burks, C. Gates, J. Con-
tillo, L. Csuzdi, A. Debose, A. Beier-Engelhaupt, K. Maze-
Foley, J. Henne, W. Hoggard, K. Hough, J. Litz, T. Martinez,
M. Newcomer, C. Palmer, K. Rademacher, C. Roden, C.
Sinclair, S. Stienessen, J. Tobias, CWH, KDM, and GLF.
The comments of two anonymous reviewers improved the
manuscript significantly. This work was conducted under
Marine Mammal Research permit 779-1339 issued to the
SEFSC and supported by Interagency Agreement 15958
between the NMFS, SEFSC and the Minerals Management
Service, Gulf of Mexico Region.
Literature cited
Barlow, J.
1995. The abundance of cetaceans in California waters. Part
I: Ship surveys in summer and fall of 1991. Fish. Bull. 93:
1-14.
Barlow, J., S. L. Swartz, T. C. Eagle, and R R. Wade.
1995. U.S. marine mammal stock assessments: guidelines
for preparation, background, and a summary of the 1995
assessments. NOAA Tech. Memo. NMFS-OPR-6:73.
Barros, N. B., and D. K. Odell.
1990. Food habits of bottlenose dolphins in the southeastern
United States. In The bottlenose dolphin (S. Leatherwood
and R. R. Reeves, eds.), p. 309-328. Academic Press, San
Diego, CA.
Baumgartner, M. F.
1997. The distribution of Risso's dolphin (Grampus griseus)
with respect to the physiography of the northern Gulf of
Mexico. Mar. Mamm. Sci. 13:614-638.
Baumgartner, M. F., K. D. Mullin, L. N. May, and T, D. Leming.
2001. Cetacean habitats in the northern Gulf of Mexico.
Fish. Bull. 99:219-239.
Blaylock, R. A., and W. Hoggard.
1994. Preliminary estimates of bottlenose dolphin abundance
in the southern U.S. Atlantic and Gulf of Mexico continental
shelf waters, 10 p. NOAA Tech. Memo. NMFS-SEFSC-356.
Buckland, S. T, D. R. Anderson, K. P. Burnham, J. L. Laake,
D. L. Borchers, and L. Thomas.
2001. Introduction to distance sampling: estimating abun-
dance of biological populations, 432 p. Oxford University
Press, New York, NY.
Davis, R. W., G. S. Fargion, N. May, T D. Leming,
M. Baumgartner, W. E. Evans, L. J. Hansen, and K. D. Mullin.
1998. Physical habitat of cetaceans along the continental
slope in the north-central and western Gulf of Mexico. Mar
Mamm. Sci. 14:490-507.
Davis, R. W., G. A. J. Worthy, B. Wiirsig, S. K. Lynn, and
F I. Townsend.
1996. Diving behavior and at-sea movements of an Atlantic
spotted dolphin in the Gulf of Mexico. Mar Mamm. Sci.
12:569-581.
932
Fishery Bulletin 101(4)
Fertl, D.,andB. Wursig.
1995. Coordinated feeding by Atlantic spotted dolphins
iStenella frontalis) in the Gulf of Mexico. Aquat. Mamm.
21{l):3-5.
Gilbes, F., C. Tomas, J. J. Walsh, and F. E. Miiller-Karger.
1996. An episodic chlorophyll plume on the West Florida
Shelf Cont. Shelf Res. 16:1201-1224.
Griffin. R. B., and N. J. Griffin.
2003. Distribution, habitat partitioning and abundance of
Atlantic spotted dolphins, bottlenose dolphins, and logger-
head sea turtles on the eastern Gulf of Mexico continental
shelf GulfMex.Sci. 21:23-34.
Hersh, S. L., and D. A. Duffield.
1990. Distinction between northwest Atlantic offshore and
coastal bottlenose dolphins based on hemoglobin profile and
morphometry. In The bottlenose dolphin (S. Leatherwood
and R. R. Reeves, eds.), p. 129-142. Academic Press, San
Diego, CA.
Herzing, D. L.
1997. The life history of free-ranging Atlantic spotted dol-
phins (Stenella frontalis): age classes, color phases and
female reproduction. Mar. Mamm. Sci. 13:576-595.
Hoese, H. D., and R. H. Moore.
1998. Fishes of the Gulf of Mexico. Texas, Louisiana, and
adjacent waters, 422 p. Texas A&M Univ. Press, College
Station, TX.
Hoffstetter, T. C.
2002. Distribution and abundance of marine mammals in
the northeastern Gulf of Mexico relative to sea surface tem-
perature and depth. M.Sc. thesis, 74 p. Univ. Southern
Mississippi, Hattiesburg, MS.
Jefferson, T. A.
2002. Rough-toothed dolphin, Steno bredanensis. In Ency-
clopedia of marine mammals (W. F. Perrin, B. Wiirsig, and
J. G. M. Thewissen, eds.), p. 1055-1059. Academic Press,
San Diego, CA.
LeDuc, R, G., and B. E. Curry.
1998. Mitochondrial DNA sequence analysis indicates need
for revision of the genus Tursiops. Rep. Int. Whal. Comm.
47:393.
Lerczak, J. A., and R. C. Hobbs.
1998. Calculating sighting distances from angular readings
during shipboard, aerial, and shore-based marine mammal
surveys. Mar Mamm. Sci. 14:590-599.
Lohrenz, S. E., D. A. Wiesenburg, R. A. Arnone, and X. Chen.
1999. What controls primary production in the Gulf of
Mexico? In The Gulf of Mexico large marine ecosystem (H.
Kumpf, K. Steidinger, and K. Sherman, eds.), p. 151-170.
Blackwell Science, Maiden, MA.
Marsh, H., and D. F Sinclair
1989. Correcting for visibility bias in strip transect aerial
surveys of aquatic fauna. J. Wildl. Manag. 53:1017-1024.
Mills, L. R., and K. R. Rademacher.
1996. Atlantic spotted dolphins (Stenella frontalis) in the
Gulf of Mexico. GulfMex.Sci. 14:114-120.
Mullin. K. D., and L. J. Hansen.
1999. Marine mammals of the northern Gulf of Mexico. In
The Gulf of Mexico large marine ecosystem (H. Kumpf K.
Steidinger, and K. Sherman, eds.), p. 269-277. Blackwell
Science, Maiden, MA.
Paluszkiewicz, T, L. P. Atkinson, E. S. Posmentier, and
C. R. McClain.
1983. Observations of a Loop Current frontal eddy intru-
sion onto the West Florida Shelf J. Geophys. Res. 88:
9639-9651.
Pitman. R. L., and C. Stinchcomb.
2002. Rough-toothed dolphins (Steno bredanensis) as
predators of mahimahi (Coryphaena hippurus). Pac. Sci.
56: 447-450.
Perrin, W. R
2002. Stenella frontalis. Mamm. Species 702:1-6.
Rabalais, N. N., R. S. Carney, and E. G. Escobar-Briones.
1999. Overview of continental shelf benthic communities of
the Gulf of Mexico. In The Gulf of Mexico large marine
ecosystem (H. Kumpf, K. Steidinger, and K. Sherman, eds.),
p. 171-195. Blackwell Science, Maiden, MA.
Richard, K. R., and M. A. Barbeau.
1994. Observations on spotted dolphins feeding nocturnally
on flying fish. Mar. Mamm. Sci. 10:473-477.
Schmidley, D. J., and B. A. Melcher
1974. Annotated checklist and key to the cetaceans of Texas
waters. Southwest. Nat. 18:453-464.
Torres, L. G., P E. Rosel, C. D'Agrosa, and A. J. Read.
2003. Improving management of overlapping bottlenose
dolphin ecotypes through spatial analysis and genetics.
Mar. Mamm. Sci. 19:502-514.
Wade, P R., and R. P Angliss.
1997. Guidelines for assessing marine mammal stocks:
report of the GAMMS workshop April 3-5, 1996, Seattle,
WA, 93 p. NOAA Tech. Memo. NMFS-OPR-12.
Wade, P R., and D. P DeMaster.
1999. Determining the optimum interval for abundance
surveys. In Marine mammal survey and assessment
methods (G. W. Garner, S. C. Amstrup, J. L. Laake, B. F. J.
Manly, L. L. McDonald, and D. G. Robertson, eds. ), p 53-66.
A. A. Balkema. Rotterdam, Netherlands.
Waring, G. T, J. M. Quintal, and S. L. Swartz (eds.l.
2001. U.S. Atlantic and Gulf of Mexico marine mammal
stock assessments— 2001, 310 p. NOAA Tech. Memo.
NMFS-NE-168.
Wiseman, W. J., Jr., and W. Sturges.
1999. Physical oceanography of the Gulf of Mexico: processes
that regulate its biology. In The Gulf of Mexico large
marine ecosystem (H. Kumpf, K. Steidinger, and K. Sher-
man (eds. ), p 77-92. Blackwell Science, Maiden, MA.
Wiirsig, B., S. K. Lynn, T A. Jefferson, and K. D. Mullin.
1998. Behavior of cetaceans in the northern Gulf of Mexico
relative to survey ships and aircraft. Aquat. Mamm. 24:
41-50.
933
Abundance of horseshoe crabs
(Limulus polyphemus)
in the Delaware Bay area
David Hata
Jim Berkson
Department of Fisheries and Wildlife Sciences
Virginia Polytectinic Institute and State University
Blacksburg, Virginia 24061-0321
E-mail address (for J Berkson, contact author) iberksoniffivtedu
III recent years, increasing commercial
landings of horseshoe crabs {Limulus
polyphemus) along the Atlantic coast of
the United States have raised concerns
that the present resource is in decline
and insufficient to support the needs
of its user groups. These concerns have
led the Atlantic States Marine Fisher-
ies Commission ( ASMFC ) to implement
a fishery management plan to regulate
the harvest (ASMFCM. In order to
properly manage any species, specific
management goals and objectives
must be established, and these goals
depend on the resource users involved
(Quinn and Deriso, 1999). Horseshoe
crabs present a distinct resource
management challenge because they
are important to a diverse set of users
(Berkson and Shuster, 1999).
Horseshoe crabs lay their eggs on
sandy beaches in spring and summer,
and migrating shorebirds rely heav-
ily on the eggs to supply the energy
required to complete their migration
(Rudloe, 1980; Shuster and Botton,
1985; Castro and Myers, 1993; Botton
et al., 1994; Myers, 1996; Thompson,
1998; Tsipoura and Burger, 1999).
Biomedical companies catch horseshoe
crabs for their blood, from which they
produce Limulus Amebocyte Lysate
(LAL) (Novitsky 1984; ASMFCi). LAL
is used to detect contamination of in-
jectable drugs and implantable devices
by Gram-negative bacteria and is the
most sensitive means available for
detecting endotoxins (Novitsky, 1984).
Finally, horseshoe crabs are harvested
commercially for bait in the American
eel (Anguilla rostrata), catfish (Ictci-
lurus spp.), and whelk (Busycon spp.)
fisheries (ASMFC).
The goal of the ASMFC fishery man-
agement plan is to ensure a sustainable
population level that will support the
continued use by these diverse ecologi-
cal, biomedical, and fishing interests
(ASMFC). Proper management of the
resource requires information on the
status and dynamics of the horseshoe
crab population (Berkson and Shus-
ter, 1999). However, the status of the
population is poorly understood, and
there is currently no reliable informa-
tion on which to base any management
scheme. Available fishery-independent
surveys were not designed for horse-
shoe crabs, and are of little or no value
in assessing their status (ASMFC'^).
Towards this end, the states of New
Jersey, Delaware, and Maryland in
conjunction with the ASMFC and the
National Fish and Wildlife Foundation,
funded a pilot benthic trawl survey for
the fall of 2001. Data collected during
this pilot trawl survey were used to es-
timate the horseshoe crab population
size in the Delaware Bay area.
Methods
This study was conducted in the vicin-
ity of Delaware Bay, which is the center
of abundance for horseshoe crabs on
the Atlantic coast (Shuster, 1982). The
study area extended from north of Cape
May, New Jersey, to south of Ocean City,
Maryland (39°10'N to 38°10'N), and
from shore out to 22,2 km (Fig. 1). The
area was divided into four strata based
on distance from shore and topography,
both of which influence crab distribu-
tion. Distance from shore was con-
sidered important because horseshoe
crab abundance decreases with depth
(Botton and Ropes, 1987a). Therefore,
the area was split into an inshore zone
from 0 to 5.6 km (0 to 3 nautical miles
[nmi] ) from shore and an offshore zone
from 5.6 to 22.2 km (3 to 12 nmi) from
shore. Topography was also consid-
ered important because commercial
fishermen stated that crabs are more
abundant in troughs (Burke^; Eutsler'*;
Munson^). For this study, troughs were
defined as at least 2.4 m deep, no more
than 1.8 km wide, and more than 1.8
km long. These dimensions are common
for troughs identified as important by
the fishermen. The inshore and off-
shore zones were both further divided
into trough and nontrough areas. The
resulting strata were inshore trough,
inshore nontrough, offshore trough,
and offshore nontrough.
The study area was divided into grids
of one-minute latitude by one-minute
longitude. A grid was considered in-
shore if the majority of its area was in
water and inshore of the 5.6-km divid-
ing line. A grid was considered offshore
if the majority of its area was offshore
of the 5.6-km dividing line and inshore
of the 22.2-km boundary. A grid was
also considered a trough if the long ax-
is of a trough passed through the grid.
• ASMFC (Atlantic States Marine Fisheries
Commission). 1998. Interstate fishery
management plan for horseshoe crab.
Fishery management report no. 32, 58 p.
Atlantic States Marine Fisheries Com-
mission. 1444 Eye Street, NW, Sixth Floor,
Washington, DC 20005.
2 ASMFC. 1999. Horsehoe crab stock
assessment report for peer review. Stock
assessment report No. 98-01 (supplement),
47 p. Atlantic States Marine Fisheries
Commission, 1444 Eye Street, NW, Sixth
Floor, Washington, DC 20005.
■* Burke, C. 2001. Personal commun. 25
Cove Drive, North Cape May, NJ 08204.
■* Eutsler, J. 2001. Personal commun.
11933 Gray's Corner Road, Berlin, MD
21811.
^ Munson, R. 2001. Personal commun.
Box 358, Newport, NJ 08345.
Manuscript approved for publication
6 March 2003 by Scientific Editor
Manuscript received 22 July 2003 at
NMFS Scientific Publications Office.
Fish Bull. 101:933-938 (2003).
934
Fishery Bulletin 101(4)
3830'
Figure 1
Study area and sampling locations. Symbols indicate type
and location of strata. Day and night tows were made at each
location.
A grid was considered nontrough if no trough long axis
passed through it. Each grid was therefore assigned to one
of the four strata. Twelve grids were randomly selected in
each stratum, for a total of 48 unique sampling locations.
The fishermen also stated that time of day influenced
horseshoe crab catchability (Burke^; Eutsler"*; Munson^).
Therefore, grids were sampled both in daylight and at
night. The second tow in a grid (day or night) was made
near the location of the first to reduce location variability,
but slightly offset to avoid possible influence of the first
tow on the catch of the second. The second tow was also
made more than 24 hours after the first to avoid interac-
tions, but no more than four days later, to avoid introducing
other unknown variability Abundance estimates from the
daytime and nighttime samples were calculated separately
for comparison.
Our study was conducted in the fall, between 10 Sep-
tember and 16 October 2001. The stock assessment model
adopted by the ASMFC requires abundance information on
newly mature crabs, and identification requires that crabs
have undergone a terminal molt. Crabs reportedly molt in
the late summer and fall in the Delaware Bay area (Burke^;
Eutsler"*; Munson-'').
Sampling was conducted from a chartered 16.8-meter
commercial fishing vessel. For capturing horseshoe crabs,
commercial fishermen typically use a flounder trawl
equipped with a Texas sweep (Burke'^; Eutsler'*; Munson^;
Michels''). This modified sweep consists of a chain line in-
^ Michels, S. 2001. Personal commun. Delaware Department
of Natural Resources and Environmental Control, Division of
Fish and Wildlife, 89 Kings Hwy, P.O. Box 1401, Dover, DE 19901.
NOTE Hata and Berkson: Abundance of Limulus polyphemus in the Delaware Bay area 935
stead of rope, which runs from wing to wing of the
net (Fig. 2). The net ropeUne is attached behind the
sweep chain. In addition, usually three rows of weight
chain are attached behind the sweep chain. The chain
sweep is considered more effective in digging crabs
out of the bottom than the typical ground gear of
most research trawls. We used a standard two-seam
flounder trawl with an 18.3-m headrope and 24.4-m
footrope. The net consisted of 14-cm stretched mesh
polypropylene throughout and was equipped with
chafing gear on the bag. The net was attached to the
trawl doors by 91-m ground cables wrapped in rubber
cookies. Tow duration was usually 15 minutes (bottom
time), except for one tow in the Delaware Bay ship-
ping channel, which was reduced to 7.5 minutes. We
assumed that density was not affected by tow dura-
tion (e.g. gear saturation was not a factor).
All horseshoe crabs were culled from the catch,
and either all or a subsample were examined. For
subsamples of a large catch, 50 crabs greater than
150 mm prosomal width were examined, as well as
all small, soft, and shedding crabs. Horseshoe crabs
that were not examined were counted separately by
sex. Examined crabs were measured for prosomal
width and identified to sex and maturity. Maturity
classifications were as follows: immature; primipa-
rous (mature horseshoe crabs that had not spawned
yet); and multiparous (crabs that had spawned at
least once (Table 1| ). When catches were subsampled,
characteristics of examined crabs were extrapolated
to all crabs in that tow. Abundance was estimated for
each demographic group as well as for the total.
Tow distances were determined for most tows
from beginning and ending positions and recorded
by using Loran C. These are minima because they
do not consider any deviations from a straight path.
Distances were not recorded for three tows; therefore they
were estimated as the mean distance of all other tows. Net
width was estimated as half of the mean of the headrope
and footline lengths (Fridman, 1986). The tow distance and
net width were used to calculate the swept area to deter-
mine the density of horseshoe crabs. We assumed that the
ground cables and trawl doors were not effective in catch-
ing crabs; therefore all fishing was done only by the net.
No information is available on the efficiency of the ground
cables or doors for horseshoe crabs, but we do not believe
horseshoe crabs are mobile enough, nor swim fast enough,
to be effectively herded by them.
The mean density (crabs/km^) and variance in each
stratum were calculated by assuming a zl-distribution
(Aitchison and Brown, 1957; Pennington, 1983), and these
estimates were combined by using formulas for a stratified
random design (Cochran, 1977). The /i-distribution model
is applicable to skewed data that consist of a portion of zero
catches when the frequency of nonzero catches follows a
lognormal distribution (Pennington 1983; Pennington
1996). With such skewed data, the estimator of the mean as
defined for the 4-distribution model is more efficient than
the sample mean estimator derived from the normal distri-
bution (Smith, 1988). Areas by stratum and total area were
\
to ground cable
1
•■"■'^^
OTrai^^^v
WvVAW. ^^''k.
i;,;::'^m,^.
■ ! ■ X '! ■ X' C<)OvM>i»«******"**!a***^'*'*"**^^»^*"****^**'****^^^
280 -
^V^ **
26,0 -
240 -
A
220 -
on n
0 10 20 30 40 50 60 70 80 90 100 110 120
340 -
Tag 5D-2
320 -
^T**^***^ ^ It /***^ K »r*«***nr*n m
30 0 -
Y^ 7 V A V L / \r^
280 -
i\ ^ /V \r V
26,0 -
i^ Y i Y
240 -
B
22 0 -
on n
0 10 20 30 40 50 60 70 80 90 100 110 120
34 0 -
Tag 5D-4
32 0 -
5 30 0 -
2 280 -
Q. 26 0 -
a> 24 0 -
H
22 0 -
20 0 ~
c
0 10 20 30 40 50 60 70 80 90 100 110 120
34 0 -
Tag 5D-5
32 0 ^
A^*VVVV********N 7"********^ /***-******,»♦***
30 0 -
280 -
260 -
240 -
D
22 0 -
20 0 -
0 10 20 30 40 50 60 70 80 90 100 110 120
340 -|
Tag 5D-7
320 -
V******^K f**^ AK/^f f***^ ti