\r, 51 53 55 5? 59 6l 63 65 67 69 71 7? 75 77 79 8l 83
median of CL. class (mm)
Figure 4. Area graph showing size frequencies of male Guinea chick
lobsters {P. guttatus) by physiographic province (data from Transects
1 and 2). Legend key: P-reef 3-6f, patch reefs, 3 to 6 fathoms (6 to 12
m); R-flat 2-5f, reef flat, 2 to 5 fathoms (4 to 10 m); F-reef 7-9f, fore
reef, 7 to 9 fathoms (14 to 18 m); M-terr lOf, main terrace (i.e., the
reef-front terrace), 10 fathoms (19.4 m); Edge 13-20f, edge, 13 to 20
fathoms (25 to 39 m).
male) was lowest for the reef flat, just inside the reef crest, indi-
cating that females were localized to the atoll rim inside the reef
crest during the autumn and early winter, whereas during Transect
2 (May-September), it was lowest on the edge of the island shelf,
and on the fore-reef slope and the reef-crest area adjacent, indi-
cating a seasonal localization of females to the rim reefs and outer
reefs during the breeding season (Table 2).
The St. Catherine's study area (patch reefs) showed the lowest
sex ratio of trapped male:female animals (April-May), followed
by the North Reefs (reef flats and patch reefs) (September-
December) and the East Ledge Flat (reef flats) (February-March).
These areas are all within the 3 fathom contour and suggest a
localization of females to the area inside the reef crest (three
fathom line) in the autumn, winter, and spring (Table 3A and
B). A similar sex ratio to that at East Ledge Flat was found
off Soldier's Point, Cooper's Island, around patch reefs just inside
the breaker line or reef crest (October-January). On Transect
1 (October-January), this was the only location where females
were found (Table 2). Trapping in the high-energy environment
(frequently with powerful "surge" on the sea floor) just seaward
of the reef crest-boiler reefs off Cooper's Island (October-
Figure 3. The Transect across the northeastern reef system, through Kitchen Shoals, eastern study areas, and section through the reef system.
(A) The Transect 2 trap line across the northeastern reef system and the Kitchen Shoals and Sea Venture Shoals study areas. The positions of
single traps on the 7 mile long transect line are marked by a cross. Soundings are in fathoms with feet as a subscript (1 fathom = 1.94 m).
Sampling was on average twice weekly, from May through October. An average of seven traps per trip were hauled at Kitchen Shoals. The
number of stations on the transect was reduced to 13 by the end of the experiment, owing to loss of traps and insufficient replacement. The trap
rates of the physiographic provinces traversed by the transect were nevertheless obtained, including that of the reef-crest area at Kitchen Shoals.
A capture-recapture experiment was also carried out at the Kitchen Shoals study area, simultaneous to the trapping along the transect line, (b)
The reef profile. Sectional diagram showing the major physiographic provinces of the Bermuda reef system. The profile shown is the section from
west to east along the line of the transect (Fig. 3a). Key: L, lagoon; SL, sea level; PRZ, patch reef zone in the lagoon, landward of the atoll rim,
of 6 to 12 metres deep (3 to 6 fathoms); RCA, reef crest area comprising the reef flats (rf) and the algal cup reefs or "boiler reefs" (br) that break
the surface, forming the atoll rim; the reef flats are located at ca. 6 to 8 m deep (3 to 4 fathoms); FRS, fore-reef slope, sloping reef surface seaward
of the reef crest or "breaker line," ranging from ca. 12 to 18 m in depth (6 to 9 fathoms); RFT, reef-front terrace or "main terrace" at 19 m (10
fathoms); E, edge of the island shelf, or "platform edge," i.e., the terraces and slopes seaward of the reef-front terrace out to the dropoff; in this
province, P. guttatus was only found at 22 to 45 m deep (1 1 to 22 fathoms); DO, The "drop off," i.e., the beginning of the steep, precipitous slope
to the ocean floor, at ca. 68 to 78 m deep (35 to 40 fathoms).
398
Evans and Lockwood
Northeast 8-U fa
northeast 7-9 fa
Worth Reefs 3-9 la
Ito 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85
median of C.L. class(mm)
Figure 5. Line graph showing size frequency distributions for male
Guinea chick lobsters {P. guttatus from the northern reefs by depth.
The fore reef associated with Northeast Tower (Northeast 7-9fa) has a
pronounced peak (modal class 54 to 56 mm CD of small lobsters of 50
to 58 mm CL. The same peak is much less apparent in the reef-front
terrace, i.e., main terrace (Northeast 8-12fa) size frequency distribu-
tion and even less so in that for the reef-crest area and patch reef zone
(North Reefs 3-9fa). It should be noted, however, that the other two
categories both include the 7 to 9 fathom range. The smallest individ-
uals were on the fore-reef slope at 7 to 9 fathoms (14 to 18 ml in depth;
the medium-sized males were on the reef-front terrace at 8 to 12
fathoms (16 to 23 m) in depth; and larger males comprised the "North
Reefs" sample, taken from the North Reefs itself at 3 fathoms (6 m) in
depth, and the pinnacle patch and table reefs inside of them, at 3 to 9
fathoms (6 to 18 m) in depth.
December) resulted in a high sex ratio of trapped Guinea chicks
(Table 3A).
The median sex ratio (male:female) of Guinea chicks caught in
the study areas seaward of the reef crest (28:1) was greater than the
median sex ratio from the study areas landward of the reef crest
(11:1) (see statistical analysis, Table 3C). This inferred a general
localization of females to the reef tract.
Change in Size Composition at Kitchen Shoals
The results of a study of the change in size composition of male
Guinea chick lobsters over the summer at the northeastern reef
crest are presented in Figure 1 la to c. Figure 1 la and b show the
change in size composition for the Kitchen Tripod shoal, and
Figure 12c presents the change in the size composition of animals
caught under overhanging coral-algal cliffs at the reef crest itself,
close to Kitchen Tripod. The three figures show the size compo-
sition changing through the summer (May to September), to one of
increasingly smaller size, then changing toward its original com-
position in October.
Relative Abundance of Spawning Females
Along the line of Transect 2 (through Kitchen Shoals), al-
though the trap rates were low, ovigerous P. guttatus occurred in
a spatial and temporal sequence near the shelf edge in May, at the
reef-front terrace in June, at the fore-reef slope in early July, and
at the reef-crest area from mid-July to the latter part of August
(Table 4), suggesting an inward breeding migration of gravid fe-
males.
The relative abundance (CPUE) of spawning females (i.e.,
females bearing eggs and/or eroded spermatophores) was greatest
at the reef crest (along Transect 2 for the months May to Septem-
ber inclusive). CPUE of breeding females was found to peak in a
spatial and temporal sequence from deeper waters near the eastern
platform edge at 26 to 35 m deep ( 13 to 18 fathoms) in May to the
reef crest in August (Fig. lid). In September, the reef-crest area
was the only location of Transect 2 where breeding females were
captured.
Seasonality of Breeding Females
Traps were maintained either on the rim reefs or outside of
them for the entire period from August 1986 to March 1987 and
were located in the patch reef zone in March-April (St. Cather-
ine's study area). They were located right across the reef system
from early May 1987 to late October 1987.
Breeding females with eggs and/or eroded spermatophores
were only found in the months of May to September and only on
the outermost reefs, comprised by the rim reefs and the fore-reef
P-reef 3-6J
[ID
R-flat 2-5f
>>
V
c
CD
h
; \
19 51 53 55 57
median
1 61 63 6! 6
of C.L.
F-reef 7-9f
M-terrace lOf
Edee 13-20f
1 73 75 77 79 81
class(mm)
W M 51 53 55 57 59 bl £.3 kb (.7 69 71
median of C.L. class(mm)
Figure 6. Size frequency distribution of female P. guttatus. (a) Area
graph showing the size frequencies of females by physiographic Prov-
ince (Data from Transects 1 and 2). (b) Line graph showing size fre-
quency distribution of females taken from the northern and eastern
reef systems (North Rock-Cooper's Island) in 1986.
Ecology and Behavior of Bermudan Lobster Panuurus guttatus
399
5t 53 55 57 59 61 tl iS 47 fcQ 71 7> 75
MEDIAN OF CL CLASS (m*l
Figure 7. Size frequency distributions of male and female Guinea
chick lobsters IP. guttatus) from the reef crest (breaker ledge) of St.
David's Head. 21 August to 3 October. The curve for males is a solid
line, and for females, it is a dashed line.
slope, the reef-front terraces, and the inner shelf edge (the latter
ranges from 26 to 39 m in depth ( 13 to 20 fathoms). They were not
found in the patch reef zone or in waters >43 m in depth (22
fathoms). There were always traps hauled from 49 to 55 m in
depth (25 to 28 fathoms) during the Transect 2 experiment.
Population Density, Catchability, and Sex Ratio of Trappable Guinea
Chick Lobsters
The results of a study of trap rates of males and females at
Kitchen Shoals over the period 7 May-21 October 1987 are pre-
sented in Figure 12. Catch per trap haul was greatest around the
time of the new moon. Sutcliffe (1956) found a similar relation-
ship for P. argus at Bermuda.
There was a very strong and wide model peak of trap rates (for
male P. guttatus) during the period 16 July to 17 August at
Kitchen Shoals (Fig. 12).
Results of capture-recapture experiments on the abundance,
catchability. and sex ratio of trappable P. guttatus in the study
areas are presented in Tables 5 to 13, starting with relative abun-
dance (CPUE) (Table 5A to C). Correction factors, which were
needed to correct the raw population density estimates, are given
in Table 6. Raw density estimates for St. David's Head, Cooper's
Island, and Kitchen Shoals made by the Hayne's Index Capture-
Recapture Method (Hayne 1949) (Tables 7 to 9) were collated,
corrected, and averaged (Table 10A to D); these results indicate a
mean population density of 29 (±7.6) ha~' (95% confidence
limit).
Separate male and female population density estimates for
Kitchen Shoals, real sex ratio, and the difference in catchability
between the sexes are presented in Table 1 1 . The estimate for
females was derived by subtraction of the estimate for males from
the estimate for males and females together. The low number of
recaptures of females made it impossible to estimate the female
population number separately.
Results of capture-recapture studies for Kitchen Shoals using
the Bailey Triple Catch Method (Chalmers and Parker 1989) are
w-
?c_
— ^-i
*20-
c
1)
(U
u
I
5-
0-
— 1 — ; — 1 — 1 — 1 — t ~'r
— 1 r
-1 — r
-T —
1
\
- I ■
ED
res.
&3 45 47
19 51 ':■' 55 r'7 I ■
median of C.L.
'7 69 71 "3 75
:lass(mm)
81
19 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81
median of C.L. class(mm)
male
ED
53 55 57 59 61 63 65 67 69 71 73 75 7 79 81
median of C.L, class (mm)
Figure 8. Size frequency distributions of male and female (fern.) P.
guttatus from the northeastern reef system (North Rock-Cooper's Is-
land), (a) The breaker ledge just seaward of the breaker line (reef
crest) of Cooper's Island, October to December, (b) The fore reef
associated with the north reefs, around northeast breaker, July and Au-
gust, (c) Combined data from the study areas and Transects 1 and 2.
400
Evans and Lockwood
Inside (r-flat)
Bkr-line (f-reef
October November December January Oct-Jan
Month
Figure 9. Showing monthly Guinea chick trap rates on Transect 1, off
Soldier's Point, Cooper's Island, October-January, and variation
with depth. There were two traps inside the breaker line (reef crest),
one "at the breaker line" (bkr-line), i.e., just seaward of the boiler
reefs, but at 5 fathoms (10 m), and one at each of 5, 10, 15, and 20
fathoms (10, 19, 29, and 39 m). Total effort in trap nights, October
through January, was 109 (Inside), 96 (Bkr-line), 104 (5 fathoms), 104
(10 fathoms), 99 (15 fathoms), and 95 (20 fathoms). Surge conditions
are a feature of wave energy around the reef crest of the Bermuda atoll
system, November through May. For the period October to January,
CPUE was greatest "inside" the boiler reefs, i.e., on the reef flat, or
atoll rim, and second greatest on the 5 fathom terrace. CPUE was
third greatest at the breaker line (i.e., on the breaker ledge just sea-
ward of the coral-algal cliffs forming the boiler reefs). Monthly CPUE
"Inside" was relatively high in October and in December, whereas at
the 5 fathom station, it was relatively high in November and January.
CPUE at the very high-energy breaker line station declined continu-
ously from October to January. There is evidence of a migration away
from the breaker line from October through December, chiefly to the
reef-front terraces but also to the area inside of the breaker line.
presented in Table 12. Population density of P. guttatus in the
physiographic provinces is presented as 95% confidence bands in
Table 13.
Statistical tests of the four Hayne's Index Method population
estimates for the St. David's Head, Cooper's Island, and Kitchen
Shoals reef-crest areas showed that these results were conclusive
findings (see the test tables in Appendix: Table 14A to D). Sta-
tistical tests of population estimates from similar capture-recapture
studies at East Ledge Flat and Sea Venture Shoals showed them to
be inconclusive owing to low recapture rates.
The mean population density of fishable Guinea chick lobsters
in the reef-crest areas in the summer and autumn periods of 1986/
87 was 29 (±7.6) Guinea chicks ha~ ' (95% confidence limits).
Similarly, the mean catchability coefficient (q) was 1.3 x 10~3
(±0.84 x 10~3) trap night' ', and the mean effective area fished
(E) (Miller 1986) was 13 x 10~3 (±6.7 x 10~3) ha. trap
night"1) (Table 10D).
The average real sex ratio of fishable Guinea chicks at the
northeastern reef crest in the period 23 May to 21 October 1987
was estimated at 7:1 (male:female) (Table 1 1 A and B), by the
Hayne's Index capture-recapture population estimates for this pe-
riod for males and females (1) and males only (2). subtracting (2)
from ( 1 ) to obtain a female population estimate. The sex ratio of
trapped animals was 13:1 (male:female). The factor for converting
the sex ratios of trapped animals to real sex ratios was estimated at
7/13 = 0.54 (Table 11B). The catchability of females at Kitchen
Shoals was lower than that of males both for catchability coeffi-
cient (q) and for the effective area fished (E) (Table 1 1C): female
catchability q was 2.6 x 10 ~4 trap night" ', and for males, it was
4.8 x 10~4 trap night'1.
Analyses by Bailey's Triple Catch (Table 12) yielded informa-
tion on population change over summer and autumn at the north-
eastern reef crest. Although the confidence intervals overlap, the
results suggest that the highest population density of fishable
Guinea chick lobsters over this period probably occurred at the
height of the breeding season, i.e. from 1 July to 30 September
(Table 12). There was a large, broad, modal peak in trap rates at
the northeastern reef crest during mid-July to mid-August (Fig.
12). This supports the suggestion that density was greatest for the
months July through September.
DISCUSSION
Transects Across the Reef System
The results of the Transect 1 experiment (Fig. 9) suggest that,
prima facie, there was a migration offshore (i.e., an outward mi-
gration) during the period from 3 October to 5 December. There
are three possible explanations:
( 1 ) It may have been undertaken to avoid the surge conditions
V
CD
3»
sz
a "•:
ed "
u
■p
O
a
w
Cm
U
Fatch-reefe
on
Reef-flat
10 fa terrace
Edse(13-Ufa)
May
Jun Jul
Aug
Month (19871
Sep May-Sep
Figure 10. Showing monthly Guinea chick trap rates on Transect 2,
east-west across the Northeastern reef system, through Kitchen
Shoals, May to September. The reef flat bars refer to the catch and
effort data from pot no. 8 in the Kitchen Shoals study area, located
close to Kitchen Tower under overhanging cliffs of coral at the reef
crest. The total effort for each physiographic province during May
through September (in traps hauled) was as follows: patch reefs, 121;
reef flat (single pot: pot 8, at Kitchen Tower), 29; fore reef, 30; 10
fathom terrace, 74; and edge, 31. During the summer months. May to
September inclusive, CPUE was greatest at the reef flat (i.e., the atoll
rim) and second greatest at the 10 fathom (19 m) terrace. CPUE on the
reef flat increased from May to July, peaked in July, and then leveled
off in August and September, whereas on the 10 fathom terrace, and
near the edge of the island shelf, CPUE peaked in June and August.
Ecology and Behavior of Bermudan Lobster Panulirus guttatus
401
TABLE 2.
Sex ratio of trapped animals (S.R.) of the Guinea chick lobster P.
guttatus on Transects 1 and 2 (S.R. = ratio of males caught to
females caught).
Transect 1 (Autumn)
Transect 2 (Sumt
Male Female
ner)
Male
Female
Province"
Caught
Caught
S.R.
Caught
Caught
S.R.
PRZ
b
b
b
12
0
12:0'
RCA
31
3
10.1
63d
4d
16:1
FRS
34
0
34:0
10
1
10:1
RFT
12
0
12:0
89
2
45:1
E
9
0
9:0
9
1
9:1
a Key; see also Figure 4b. PRZ. patch reef zone; RCA. reef crest area;
FRS. fore-reef slope. RFT. reef-front terrace (or main terrace); E. edge of
platform (inner shelf edge).
b No patch reef zone on Transect 1 .
L Absence of females in the patch reef zone provides indirect evidence of
an outward breeding migration from the patch reef zone to the rim and
outer reefs.
d From Pot 8 at Kitchen Tower.
at the breaker line (reef crest). These conditions became fully
established in early November.
(2) Alternatively, it may have resulted from the migration of
newly recruited animals at the breaker ledge (just seaward of the
reef crest) into feeding areas on the reef-front terraces and slopes.
(3) Most probably, it was a return migration to the reef-front
terrace and outer terraces after a summer breeding migration to the
eastern reef-crest areas.
During the period October to January, the greatest abundance
of fishable Guinea chick lobsters off the steeper southeastern shore
of the island occurred in the somewhat quieter waters of the atoll
rim, inside the breaker line, or reef crest. The second greatest
abundance occurred on the reef- front terrace (at 10 m or 5 fath-
oms). Monthly CPUE at the high-energy breaker line station de-
clined continuously from October to January, possibly as a result
of animals moving away from the worst areas of surge or a return
migration to home foraging grounds on either side of the reef crest.
The results on relative abundance for the Transect 2 experiment
support those from the Transect 1 experiment in that, for the
period from May to September inclusive. CPUE was greatest on
the reef flat adjacent to the reef crest and second greatest on the
reef-front terrace (at 20 m or 10 fathoms) (Fig. 10). Thus, the
results of both transect experiments indicate that the greatest abun-
dance of Guinea chick lobsters was on the atoll rim and on the
reef- front terrace.
Sex and Depth Found
Consideration of the sex ratio of trapped animals at various
locations and time sheds further light on the field behavior of
Guinea chick lobsters. During the Transect 1 experiment, off the
eastern end of the island (October to January), the sex ratio of
trapped animals was greatest inside the reef crest and females were
only caught at the two stations inside the reef crest.
Whereas, on the Transect 2 trap line during the period from
May to September, the sex ratio of trapped animals was greatest at
the inner shelf edge and at the fore-reef slope (and the reef crest
adjacent to it) (Table 2). Females may thus prefer to be in an area
of high reef slope during the breeding season. The cliff-like, lab-
yrinthine structure of such areas of slope may offer protection to
the relatively small females at the critical time of larval release,
which is generally followed by molting in palinurid lobsters.
Larval release in spiny lobsters (P. argus) occurs at the extreme
edge of the island shelf (Sutcliffe 1953a). Palinurids (spiny lob-
sters) "inject" their larvae into the ocean circulation systems by
this method and so facilitate genetic mixing and recruitment after
lengthy larval drift around the ocean gyre systems (Pollock 1992).
Females may migrate away from areas of powerful surge con-
ditions seaward of the reef crest (on the breaker ledges) and on
fore-reef slopes, once these conditions begin in late October/early
November, and move to quieter areas such as the reef flats inside
the reef crest and the patch reef zone further inside. This is sup-
ported by results presented in Tables 2 and 3. Perhaps the larger
females are particularly susceptible to the wave action on the outer
reefs, and it is chiefly they that migrate to the more comfortable
locations inside the reef crest.
There were high (male:female) sex ratios at the Cooper's Island
breaker ledge during the stormy autumn/early winter period from
October to January (41:1) and at the exposed fore reef ( 1 4 to 1 8 m
or 7 to 9 fathoms) between North Rock and Northeast Tower in
July and August (35; 1) (Table 3A). There are two comparable
"reference points" for the latter sample: the progressively lower
energy environments of the inner terraces at 16 to 18 m deep (8 to
9 fathoms) and the main terrace (16 to 24 m or 8 to 12 fathoms
deep) between North Rock and Northeast Tower, each in the pe-
riod from July to August (25:1 and 18:1 male:female, respec-
tively). Higher energy environments may thus be avoided by the
females in the winter, because females are smaller than males, or
females may suffer a higher mortality rate in these environments.
This is given support by the observation that the sex ratio (male:fe-
male) of Guinea chick lobsters caught at the breaker ledge off St.
David's Head increased from autumn through winter to spring,
indicating a female flight from this high-energy environment over
the winter period (Table 3A).
At the reef flats and patch reefs of the North Reefs, the sex ratio
(male:female) of trapped animals was the second lowest observed
in the present study: 223 males to 31 females caught (7:1), in the
period September to December (Table 3B). Sex ratio was lowest
for the patch reefs in the St. Catherine's area: six males to two
females caught (3:1) during April-May (Table 3B). The areas
landward of the reef crest have the lowest Guinea chick sex ratios
(Table 3A to C).
The results in fact infer that there is both a habitat and a sea-
sonal localization of female Guinea chick lobsters on the platform.
A general, habitat localization of females was tentatively put for-
ward by Sutcliffe (1953), who was puzzled not only by the low
proportion of females caught, by also by the fact that most females
had "spent" their eggs and were found at locations well inside of
the reef crest.
In the present study, ovigerous females were only found on the
rim reefs and outer reefs, during the breeding season (Tables 2 and
4; Figs. 10 and 1 Id); the median sex ratio of samples from the
study areas seaward of the reef crest (male:female) was higher than
that for samples from areas landward (Table 3C). Sutcliffe's
(1953) hypothesis that there is a localization of female Guinea
chicks on the platform is supported by these results, but the hy-
pothesis does not explain the low proportion of females that were
caught in both studies.
In summary, the results of the transect studies (Tables 2 and 4;
Figs. 10 and 1 Id) indicate that, in summer, trappable females are
402
Evans and Lockwood
TABLE 3.
P. guttalus sex ratio (S.R.I from the study areas (ratio of males caught to females caught).
Sample Location
Depth & Period
Males
Females
Caught
Caught
S.R.
132
9
15:1
64
4
16:1
22
1
22:1
A. Locations seaward of the reef crest
St. David's Head, Breaker Ledge. (5-6 fathoms)
Aug-Oct
Dec-Jan
April-May
Cooper's Island. Breaker Ledge (5-6 fathoms)
Oct-Jan
Eastern Edge
Aug-Jan (10-28 fathoms)
May-Sep (10-20 fathoms)
Fore-Reef Slope. North Rock-Northeast Tower (7-9 fathoms)
Jul-Auga
Inner Reef Terraces, North Rock-Northeast Tower-Kitchen Tower (8-9 fathoms)
July
Main Terrace, North Rock Northeast Tower (8-12 fathoms)
Jul-Auga
Fore-Reef Slope off Kitchen Shoals (7-9 fathoms)
May-Sep
Reef-Front Terrace off Kitchen Shoals (10 fathoms)
May-Sep
Median S.R. for samples seaward of reef crest
B. Locations landward of the reef crest
*The North Reefs, inside the reef crest (3-9 fathoms)
Sept-Deca
*Off Cooper's Island, inside the boiler reefs (3-5 fathoms)
Oct-Jan
East Ledge Flat, inside the reef crest (3-5 fathoms)
Feb-Mar
St. Catherine's patch reefs, well inside the reef crest (3-6 fathoms)
Apr-May
Kitchen Shoals, on the reef crest (3-5 fathoms)
23 May-21 Oct)
Sea Venture Shoals, inside reef crest (3-5 fathoms)
Jun-Aug
Median S.R. for samples landward of reef-crest
123
67
9
414
76
229
10
89
223
31
49
6
291
19
41:1
3
22:1
1
9:1
12
35:1
3
25:1
13
18:1
1
10:1
2
45:1
28:1
31
7:1
3
10:1
5
10:1
2
3:1
22
13:1
1
19:1
11:1
From landings by Guinea chick fisherman Lynwood Outerbridge.
in breeding condition and are chiefly located on the rim reefs
(reef-crest areas) and the reefs seaward, including the outer ter-
races near the edge of the platform. They probably seasonally
migrate there in connection with the release of their larvae from
areas having the maximum encounter with the ocean current gyre.
Results from the study areas suggest that, in the autumn and win-
ter, trappable females are located in the reef flats and patch reefs
of the reef tract (i.e., the outer reefs), extending along the rim of
the plateau, inside of the reef crest (Table 3A to C), where they
may mate in the spring.
Finally, Sutcliffe (1953) found only larger females of 58 mm
CL or greater in a year-round transect across the Bermuda Plateau
from rim to rim. These were "well inside of the reef crest." This
agrees with the relatively high sex ratios found in the patch reefs
of the North Reefs (Table 3B) and correlates with the size at first
physical maturity of female Guinea chick lobsters (56.4 to 62.3
mm CL, 95% confidence limits) (Evans and Lockwood 1995).
Size and Depth
The smallest sizes of male Guinea chick lobsters caught in the
present study were found on the fore reef slope ( 14 to 18 m or 7
to 9 fathoms) and reef- front terrace ( 16 to 24 m or 8 to 12 fathoms)
off the North Reefs (Fig. 5).
The smallest females were on the fore-reef slope (Figs. 6a and
8b) and the reef- front terrace (or main terrace) and inner platform
edge (Fig. 6a). The smallest female caught was a 42.5 mm CL.
dark-colored individual captured on 3 September 1986, with a
small mesh trap just seaward of the reef crest associated with St.
David's Head (on the breaker ledge, see Fig. 2, close under ver-
tical coral-algal cliffs, representing the fore-reef slope).
The largest males caught in the present study were on the
reef-front terrace and reef tract (rim reefs and adjacent patch reefs)
between North Rock and Northeast Breaker (the North Reefs; Fig.
5). The largest females were found on the reef flat or atoll rim
(Fig. 6).
The results suggest ( 1 ) that the fore-reef slope from reef crest
to main terrace is an area of recruitment into the fishery and (2)
that larger animals are located at, or inside, the reef crest. Subadult
P. argus lobsters, in contrast with (1). migrate offshore from the
inshore waters and recruit into the fishery at the edge of the platform
at 28 to 32 fathoms (Evans 1988. 1989); but in agreement with (2)
larger P. argus animals were found in the reef tract (Evans 1988).
Ecology and Behavior of Bermudan Lobster Panulirvs guttatus
403
TABLE 3C.
Mann-Whitney U-test on difference in the medians for the areas
seaward of the reef crest and the areas landward of it.
Sample A: sex ratios from seaward of reef crest, excluding
landings during the breeding season: 9, 10, 15, 16, 22, 22, 25, 41,
45
Sample B: sex ratios from landward of the reef crest (none were
from landings in the breeding season). 3, 7, 10. 10, 13, 19
No. of Sample A Measurements Smaller
Sample B Than Sample B Measurement.
3
7
in
10
13
19
Total
0
0
1.5
1.5
2
4
9
Therefore. UA = 9.
Repeat in reverse manner
Sample A
No. of Sample B Measurements Smaller
Than Sample A Measurement.
9
10
15
16
22
22
25
41
45
Total
3
5
5
6
6
6
6
6
45
Therefore. UB = 45.
The test statistic is therefore 9 (the lower of the two values of UA and UB).
The critical value of U at the 59c level (for samples of 9 and 6) is 10
(Chalmers and Parker 1989. p. 100, Table III (2)). The null hypothesis that
there is no difference in (he population medians can therefore be rejected
at the 5<7r level.
Three factors suggest that further measures should be taken
toward the conservation of the Bermudan female gene pool and the
Bermudan breeding stock: ( 1 ) size at first physical maturity of
female P. guttatus at Bermuda is 56.4 to 62.3 mm CL (95%
confidence limits) (Evans and Lockwood 1995); (2) the smallest
females caught in traps during the present study were 48 to 50 mm
CL (Fig. 6b); and (3) low proportions of female Guinea chicks are
caught in the northeastern and eastern reef systems (Figs. 4 and 6
to 8). Three possible measures are: ( 1) a total ban on the taking of
females; (2) a minimum size limit of 59 mm CL. based upon
Evans and Lockwood (1995); and (3) a prohibited trap zone in the
Guinea chick nursery area.
The latter may be the fore-reef slope and the inner reef-front
terraces (the terraces at 12 to 18 m or 6 to 9 fathoms in depth),
which are associated with the northern rim reefs, especially the
North Reefs (which extend North Rock to the Northeast Breaker)
(Figs. 5 and 8b).
It is unlikely that the main Guinea chick lobster nursery areas
are located in the coastal and inshore areas enclosed by the estab-
lished prohibited trapping zone (Fig. 1). Research would need to
be undertaken to establish that the fore reef adjacent to the reef
crest of the North Reefs is a nursery area.
Change in Size Composition at kitchen Shoals
One very likely contribution to the change in male size com-
position over the spring and summer period from May to Septem-
ber at the Kitchen Shoals study area, including the reef crest at
Kitchen Shoals Tripod, is recruitment into the fishery. The results
suggest that the fore-reef slope, breaker ledge, and inner reef- front
terrace at 12 to 18 m (6 to 9 fathoms) seaward of the reef crest are
areas of recruitment into the fishery. The change in size compo-
sition observed at Kitchen Shoals may therefore be the spring and
summer recruitment of animals of size 53 to 61 mm CL into the
fishery. Throughout this period, but especially in July and August,
many animals on the rim reefs may become subject to commercial
trap fishing for the first time, by entering traps and becoming
trappable.
A study of the trap rates of male lobsters at the Kitchen Shoals
study area (Fig. 12) found that there was a surge of trap rates from
mid-July to mid-August and that this resulted in a broad, high,
modal peak of trap rate from 27 July to 5 August. These results
infer a fundamental change in the population size at the northeast-
ern reef crest mid summer, at the height of the breeding season.
The results of the capture-recapture experiments by Bailey's Triple
Catch Method at Kitchen Shoals also support this inference (Table
12).
A second highly probable factor in the change of size compo-
sition is in-migration of smaller males from the reef-front terraces
and slopes. This would have progressively diluted the population
of larger animals on the reef flat with smaller individuals. How-
ever, there was no evidence of a relatively long migration of males
from the trap rates observed on Transect 2. Such a migration
should have been apparent unless it was balanced by recruitment
on the fore reefs and inner terrace, as suggested, or by a counter-
migration of animals from the patch reef zone migrating to the rim
reefs, reef-front terraces, and slopes. Sutcliffe (1952, 1953a) re-
corded such an outward breeding migration for P. argus.
There is in fact a persuasive case for the in-migration of males
to the reef flat from the reef-front terraces and slopes: there is
evidence of a breeding migration of egg-bearing females from
these localities to the reef crest and flats of the atoll rim during
May-August (Fig. lid; Table 4). Smaller males would have ac-
companied these breeding females, as occurs in migrations of
egg-bearing P. argus females and smaller males from the waters of
the lagoon and the inshore waters to the outer reefs (Sutcliffe
1953a). These migrations probably entrain new recruits to the
fishery.
In conclusion, the observations perhaps reflect both annual
recruitment into the fishery at the reef crest, fore reef, and inner
reef-front terrace, and a breeding migration of smaller males (ac-
companying egg-bearing females) from the broad reef-front ter-
race to the flats and breaker line (reef crest). This migration may
entrain new fishery recruits.
Relative Abundance of Spawning Females
The results of the study of trap and catch rates of breeding
females on Transect 2 (Fig. lid; Table 4) indicate that there was
an inward migration of breeding females (bearing eggs) from the
reef-front terraces and slopes to the reef-crest area during the pe-
riod May-August. Some mating of P. guttatus lobsters at Ber-
404
Evans and Lockwood
jo-
Jun
CD
Jul
m
15
Aug
>>
0
\=\
ti
Sep
d)
□
D1
Oct
3 m depth (1.5 fathoms) on from the shores and jetties of
Government Cut. Miami, during the summer period from June to
October and found that, in June. 81% of the females were oviger-
ous. This declined sharply thereafter to 24% in July. 18% in Au-
gust, 11% in September, and 1% in October. This may be com-
pared with the Transect 2 experiment of the present study, in
which all females captured in the period of 13 July-20 August
were ovigerous (Table 4).
Caillouet et al. ( 1971 ) found that the sex ratio was about unity:
of 894 P. guttatus lobsters caught. 55% were males and 457c were
females; males and females in Miami were almost equally repre-
sented in June. August, and October, but males outnumbered fe-
males about 2: 1 in July and September. Water temperature peaked
in July, and salinity was about the same throughout the period; it
is therefore unlikely that these factors reduced the mobility of
females over the surfaces of the boulders at night. It is more likely
that the females were relatively less mobile than males, remaining
within the cracks and interstices of the boulders because of post
spawning premolt condition.
Population Density and Real Sex Ratio
Linear regression analyses of the Hayne's Index capture-
recapture data, i.e., of the "proportion marked in catch" and
"number of animals marked previously," showed a significant
positive correlation between these variables in each case for the St.
David's Head. Cooper's Island, and Kitchen Shoals experiments
(Table 14A to D, Appendix). These population estimates (Tables
7 to 9 and 1 1A) are therefore conclusive findings.
The estimate of population density at St. David's Head breaker
ledge is based on an estimate of the area of breaker ledge fished
( 12 hectares), and the area fished is not accurately known because
of the linear nature of the group of pots close to the seaward edge
of the boiler reefs. Research to estimate the radius of attraction of
baited arrowhead traps would enable a refinement of this estimate.
Results from electromagnetic tracking work at Seven Mile Beach
TABLE 6.
Calculation of correction factors for capture-recapture population density estimates — summary of correction factor estimates: A. St. David's
Head: raw density estimate x 1.6; B. Cooper's Island: raw density estimate x 1.2: C. Kitchen Shoals (Hayne's Index): (1) Males +
females: raw density estimate x 1.1, (2) Males only: raw density estimate x 1.1; D. Kitchen Shoals (Bailey Triple Catch): (1) 1 Jun-31
Aug: raw density estimate x 1.0, (2) 1 Jul-30 Sep: raw density estimate x 1.0, (3) 1 Aug-21 Oct: raw density estimate x 1.0.
A. St. David's Head tS.D.H.)
"Trap happiness" may be estimated and taken into account by application of the correction factor applied by Morgan (1974) for P. cygnus: raw
density estimate x 1.6. It should be noted that the response to recapture in the baited Bermudan arrowhead traps is not precisely known.
Marking was by punching a coded sequence of holes in the tail fan; no tags were applied; therefore, there was no need to take loss of tags into
account or to take into account mortality or immigration arising from tagging and/or displacement. Estimate of S.D.H. correction factor: raw
density estimate X 1.6.
B. Cooper's Island (C.I); results of Cooper's Island experiment to assess the combined effect of tagging and displacement.
Number tagged along the boiler reefs off Soldier's Point, C.I. (having both tail fan mark and Floy tag) = 38. Number marked with holes in the
tail fan only = 35.
Number of marked animals killed by in-trap predation up to 5 November 1986 (midterm of study) with tail fan mark only = 1.
Ratio of only tail fan marked to tagged + tail fan mark animals at large is therefore 34:37 = 1:1.088.
Number of marked animals recaptured with only tail fan mark was 29. but 5 of these had also been tagged, losing the tag from other than
emigration, mortality, or trap shyness; these causes were as follows: 1 from molting; 3 from contact with mesh; and 1 from in-trap predation. The
"true" number of marked animals we caught with solely tail fan mark is 24.
Number of animals recaptured with tag + tail fan mark = 9. but there were 5 recaught known to have had a tag besides these 9. The "true"
number of marked animals recaught with a tag + tail fan mark is 14.
Therefore, because 24 animals with tail fan mark only were recaptured, we should have recaptured about 24 x 1.088 tagged animals = 26
tagged animals.
The apparent reduction in the availability of tags for recapture is given by (26-14) x 100/26% or ca. 46% (probably caused by mortality and
emigration).
About half (34/71 or 48%) of the animals that were marked were not tagged, so the estimate of the combined loss of tags from the population
from tag- or displacement-induced emigration should be adjusted accordingly (by 48%). The adjusted factor for tag- and displacement-induced
mortality and emigration is given by: 0.46 - (0.48 x 0.46) = 0.24; the raw density estimate should therefore be reduced by a factor of 1-0.24 =
0.76, in order to take this into account.
Estimate of tag loss from moiling: Tag loss from molting need not be taken into account because tagged animals were double marked by tail fan
mark and tag.
Estimate to adjust for "trap happiness" . "Trap happiness" may be taking into account by multiplying the raw density estimate by 1.6 (as in the
Morgan [1974] experiment). It should be noted that the response to recapture in the baited Bermudan arrowhead traps is not precisely known.
The combined correction factor for tag- and displacement-induced mortality and emigration and "trap happiness' is therefore 0.76 x 1.6 = 1.2.
408
Evans and Lockwood
TABLE 6C.
Kitchen Shoals (Hayne's Index: Studies 1 and 2): results of Kitchen
Shoals experiment to assess the combined effect of tagging
and displacement.
Tail Fan Mark (TFM)
(single 0.25
inch diameter
hole punched in telson)
Spaghetti Tag
No. marked No.
f TFM
1987
No. marked
No. recaught
recaught
7 May
4
0
0
0
12 May
0
1
0
0
15 May
0
1
0
0
21 May
2
0
0
0
23 May
2
0
0
0
26 May
0
0
1
0
29 May
0
0
0
0
1 June
0
0
0
0
4 June
3
0
2
0
9 June
0
2
2
0
1 1 June
0
0
5
0
16 June
0
0
4
0
19 June
0
1
8
0
21 June
5
0
2
0
25 June
2
0
2
3
29 June
3
0
8
0
2 July
4
1
5
0
10 July
1
0
3
2
13 July
0
0
5
1
16 July
0
1
6
7
21 July
0
0
13
1
24 July
0
2
15
2
27 July
0
2
16
6
5 August
0
0
0
4
Totals
26
11
97
26
Mortality and emigration arising from the combined effect of tagging and
displacement would tend to an overestimation of the population estimate,
and therefore, a correction factor should be applied to the population
estimates from this consideration alone. However, in the course of the
census (32 May-21 Oct), 20 out of 202 animals marked were solely tail fan
marked = 9.9%
For 97 animals tagged, we should have observed on average 97 x ( 1 1/26)
recaptures of tagged animals = 41 recaptures. Therefore, the apparent
reduction in the availability of tags for recapture is given by:
(41-26) x 100/41 = 36.6%, or about 37%. (Animals were kept in a
bucket of fresh water before and after handling; this procedure was not
carried out at Cooper's Island).
Final correction factor for this "effect" is given by: 1 - (0.366 - (0.366
x 0.099)) = 0.67.
Estimate of adjustment factor to take into account tag loss from molting:
Tag loss from molting was not a problem because tagged animals were
double marked with a tail fan mark and a tag.
Estimate to adjust for "trap happiness": "Trap happiness" may be taken
into account by multiplying the raw density estimate by 1.6 (as in the
Morgan [1974] experiment). It should be noted that the response to recap-
ture in the baited Bermudan arrowhead traps is not precisely known.
The combined correction factor for tag- and displacement-induced mortal-
ity and emigration is therefore 0.67 x 1.6 = 1.1.
in Australia on foraging patterns of juvenile P. cygnus rock lob-
sters indicate that a juvenile may be attracted to a baited pot as far
away as 80 m from its home reef (Jernakoff and Phillips 1986).
The breaker ledge at the St. David's Head study area is 10 to 175
m in width and is about 100 m wide on average. It is therefore
Table 6D.
Capture-recapture studies by Bailey's Triple Catch Method at
Kitchen Shoals (Studies 1-3): Adjustment of the raw population
density estimates may be mde by application of the following
combined correction factor, composed of three components.
Component effects that must be taken into account
1 "Trap happiness": multiply by 1.6 (estimated "trap happiness"
effect, Morgan [1974]).
2. Combined effect of tag- and displacement-induced mortality and
emigration: multiply by 0.67 (correction factor estimated in Table
6C above).
3. Tag loss from molting: multiply by 0.90 (from estimate of 10%
per molt tag loss from molting4 (Davis 1978)). It was necessary
to take this effect into account because the secondary mark did
not permit determination of the time of last capture if a tag had
been lost.
Combined correction factor for Bailey's Triple Catch Studies at Kitchen
Shoals: combined correction factor taking "trap happiness," tag- and dis-
placement-induced mortality and emigration, and tag loss from molting
into account is given by: raw density estimate x 1 .6 x 0.67 x 0.90, which
is the same as raw density estimate X 1.0.
The combined correction factor for "trap happiness," tag- and displace-
ment-induced mortality and emigration, and tag loss from molting is 1.0.
a Other evidence indicated that only one molt occurred (Evans, 1988,
1989).
unlikely that the entire area of the ledge in the study area was
fished. However, the coverage was in fact made greater by the
random resetting of pots at their stations.
The combined effect of mortality from tagging and release and
of emigration induced by tagging and displacement (the latter was
also found to occur in western Australian rock lobsters; Chittle-
TABLE 7.
Population estimate at the Breaker Ledge off St. David's Head,
autumn 1986.
Sampling
w
X
y
wx2
wxy
21 Aug
->
0
0
0
0
24 Aug
7
3
0
63
0
27 Aug
15
10
1/15
1.500
10
30 Aug
13
24
1/13
7,488
24
3 Sept
14
36
5/14
18,144
180
8 Sept
10
45
5/10
20,250
225
1 1 Sept
14
50
7/14
35.000
350
16 Sept
14
57
2/14
45,486
114
20 Sept
14
65
6/14
59.150
390
23 Sept
12
72
7/12
62,208
504
26 Sept
9
77
2/9
53,361
154
30 Sept
12
82
5/12
80,688
410
3 Oct
10
89
5/10
79,210
445
Totals
462,548
2,796
For each recapture phase: w, total number caught; x, total number previ-
ously handled up to, but not including, that phase (i.e., the total marked up
to that time); y, the proportion of the catch previously handled (i.e., the
proportion of the catch that is marked).
Raw population estimate, St. David's Head breaker ledge: p = Swx:/
Swxy (Hayne's Index multiple census).
p = 462,548/2,796 = 165 fishable Guinea chick lobsters. Area of breaker
ledge measured by tracing and use of small square graph paper = 12
hectares. Raw estimate of population density = 165/12 = 14 ha-1
Ecology and Behavior of Bermudan Lobster Panulirus currATUs
409
TABLE 8.
Population estimate at the Breaker Ledge off Cooper's Island, late
autumn/early winter 1986.
Sampling w x y wx wxy
6 Oct
9 Oct
13 Oct
18 Oct
21 Oct
24 Oct
27 Oct
31 Oct
5 Nov
8 Nov
17 Nov
25 Nov
5 Dec
Totals
15
7
10
6
8
20
2
5
24
3
9
14
12
0
12
17
23
29
32
46
46
49
65
66
66
74
0
0
2/10
1/6
3/8
5/20
1/2
1/5
6/24
2/3
7/9
6/14
4/12
0
1.008
2,890
3,174
6,728
20.480
4,232
10,580
57,624
12,675
39.204
60,984
65,712
285.291
0
0
34
23
87
160
46
46
294
130
462
396
296
1,974
For each recapture phase: w, total number caught; x. total number previ-
ously handled up to, but not including, that phase (i.e. , the total marked up
to that timel; y, the proportion of the catch previously handled (i.e., the
proportion of the catch that is marked).
Raw population estimate. Cooper's Island breaker ledge: p - £wx2/2wxy
(Hayne's Index multiple census), p = 285.291/1,974 = 145 fishable
Guinea chick lobsters. Area of breaker ledge measured by tracing and use
of small square graph paper = 5.7 hectares. Raw estimate of population
density = 145/5.7 = 25 ha"1
borough 1974) was estimated to be about 46% at the Cooper's
Island Breaker Ledge, in the late autumn/early winter of 1986
(Table 6B), and to be about 37% at Kitchen Shoals in the summer
of 1987 (Table 6C). The loss rate of tags from molting was as-
sumed to be the same as that evaluated by the Davis (1978) study
(10% per molt).
There was only one molt of tagged lobsters during the study
period June-October (Evans 1988). previous capture-recapture
work in Australia from trapping P. cygnus rock lobsters in baited
lobster pots (Morgan 1974) resulted in an estimate of the effect of
previous capture on the probability of recapture. Morgan (1974)
found that estimates of density had to be adjusted to correct for a
"trap-happiness" effect, by multiplying the estimate by 1.6. The
rock lobsters are only slightly larger in size than P. guttatus lob-
sters, and so. the same adjustment factor was used in the present
study with some modification. (The correction factors that were
calculated and applied to raw population density estimates in the
present study were presented in Table 6.)
The estimate of the real sex ratio of fishable Guinea chicks
(7:1, male:female) at the northeastern reef crest for the summer
and autumn was less than that observed from trapping (13:1), but
much more than unity. This does not reflect a higher mortality for
trappable females than for trappable males, because the annual
total mortality coefficients were estimated at 0.81 for males and
0.77 for females (Evans 1989, p. 122). Prima facie, it points
further toward a localization of females on the island shelf, al-
though the transect, area, and capture-recapture studies did not
find concentrations elsewhere that were great enough to explain
the disparity. Furthermore, an inshore juvenile lobster sampling
program by artificial lobster shelters resulted in a nil catch of Guinea
chick lobsters, yet attracted 64 early benthic juvenile P. argus lob-
sters into residence (Evans 1988 and 1989, Evans et al. 1994).
A more plausible explanation is that the sex ratio of the early
benthic juvenile stage is near unity and that the mortality rate of
juvenile female Guinea chick lobsters is significantly higher than
that of juvenile males. This would be likely, because the female of
the species is smaller: trappable females have a slower growth rate
than males. The annual growth coefficient K was estimated to be
0.155 for females, as opposed to 0.22 for males (Evans 1988,
1989).
The mean retention size of the 4. 1 m hexagonal mesh traps is
50 mm CL (Munro 1974), and therefore, a proportion of the pop-
ulation of females is precluded from sampling by traps; this may
also contribute to the disparity in the sex ratio of Guinea chicks
TABLE 9.
Population estimate at Kitchen Shoals (Kitchen Tripod Shoal) by
Hayne's Index (male and female combined) for the summer and
autumn of 1987.
Sampling
wx
wxy
23 May
26 May
1 Jun
4 Jun
9 Jun
11 Jun
16 Jun
19 Jun
21 Jun
25 Jun
29 Jun
2 Jul
10 Jul
13 Jul
16 Jul
21 Jul
24 Jul
27 Jul
5 August
17 August
20 August
27 August
2 Sept
10 Sept
14 Sept
17 Sept
21 Sept
24 Sept
5 Oct
12 Oct
16 Oct
21 Oct
Totals
2
1
0
7
4
5
4
10
8
7
12
9
6
6
14
16
19
25
19
7
4
6
3
7
5
8
25
16
16
15
2
25
313a
13
17
25
32
36
47
55
59
64
70
83
100
118
133
137
139
143
144
147
151
157
176
187
202
202
202
0
0
0
0
2/4
0
0
1/10
0
3/7
0
1/9
2/6
1/6
8/14
1/16
2/19
6/25
4/19
3/7
2/4
2/6
0
4/7
0
2/8
6/25
5/16
1/16
10/15
1/2
5/25
0
4
0
63
256
320
676
2,890
5,000
7,168
15,552
19.881
18,150
20,886
57,344
78,400
130,891
250,000
264,556
123,823
75,076
115,926
61,347
145,152
108,045
182,408
616,225
495,616
559,504
612,060
81,608
1,020,100
5,068,927
0
0
0
0
16
0
0
17
0
96
0
47
110
59
512
70
166
600
472
399
274
278
0
576
0
302
942
880
187
2,020
202
1,010
9.235
For each recapture phase: w. total number caught; x. total number previ-
ously handled up to. but not including, that phase (i.e., the total marked up
to that time); y, the proportion of the catch previously handled (i.e., the
proportion of the catch that is marked).
Raw population estimate. Kitchen Shoals, is given by: p = 2wx2/2wxy
(Hayne's Index multiple census), p = 5068927/9235 = 549 fishable
Guinea chick lobsters. Shoal area measured by small square method = 17
ha. Raw estimate of population density = 549/17 = 32 ha"'
" Catch was composed of 291 males and 22 females. In constructing the
above table allowances have been made for animals dead from in-trap
predation.
410
Evans and Lockwood
TABLE 10.
Calculation and collation of CPUE data, population density
estimates and catchability estimates for trappable Guinea chick
lobsters in the reef-crest areas of Bermuda, in the summer and
autumn periods of 1986-87.
A. SDH. (21 August-October 1986)
Raw population estimate = 165 Guinea chicks
Corrected population estimate = 165 x 1.6 = 264 Guinea chicks
Raw density estimate = 14 Guinea chicks ha' '
Corrected density estimate = 14 x 1.6 = 22 Guinea chicks ha ''
CPUE = 137 Guinea chicks/322 trap nights = 0.425 Guinea chicks
trap night ' '
q = 1.6 X 10~3 trap night'1
E = 19 x 10'3ha. trap night '
B C.I. (3 October-5 December 1986)
Raw population estimate = 145 Guinea chicks
Corrected population estimate = 145 x 1.2 = 174 Guinea chicks
Raw density estimate = 25 Guinea chicks ha '
Corrected density estimate = 25 x 1.2 = 30 Guinea chicks ha' '
CPUE = 137 Guinea chicks/427 trap nights = 0.321 Guinea chicks
trap night- '
q = 1.8 x 10 ~3 trap night' '
E = 11 x 10'3 ha trap night'1
C. K.S male and female combined (23 May-21 October 1987)
Raw population estimate = 549 Guinea chicks
Corrected population estimate = 549 x 1 . 1 = 604 Guinea chicks
Raw density estimate = 32 Guinea chicks ha '
Corrected density estimate = 32 x 1.1 = 35 Guinea chicks ha' '
CPUE = 313 Guinea chicks/1.146 trap nights = 0.273 Guinea
chicks trap night' '
q = 0.45 x 10'' trap night
E = 7.8 x 10' ' ha trap night' '
S.D.H., St. David's Head; C.I., Cooper's Island; K.S . Kitchen Shoals.
Catch per unit of effort (CPUE) = q.P. where q is the coefficient of
catchability and P is the population number (Nicholson and Bailey 1935.
Ricker 1975). CPUE = ED, where E = effective area fished (Miller
1986) and D = population density.
TABLE 11.
Male and female population densities, real sex ratio, and the
difference in catchability between the sexes at the Northeastern reef
crest summer/autumn.
TABLE 10D.
Summary of results on the density and catchability of Guinea chicks
at the reef crest.
Density*
Catchability
Location
q"
Ec
SDH.
22
1.6 x
10 '3
19 x 10 3
C.I.
30
1.8 x
10"3
11 x 10"3
K.S.
35
0.45 ■
10 '3
7.8 x 10'3
Mean
29
1.3 x
10"'
13 x 10"3
SD (SD)
6.6
0.73 x
10'3-
5.8 x 10'3
95% Confidence
limits
±7.6
±0.84 x
10"3
±6.7 x 10 "3
* Guinea chicks ha '.
b Trap night ' ' .
c ha. trap night '.
d 95% confidence limits (2 x standard error) calculated from: standard
error = SD/(N)~"2, where SD is the standard deviation and N is the
number in the sample (Mathers 1981).
Sampling
w
X
y
wx2
wxy
23 May
1
0
0
0
0
26 May
1
1
0
1
0
1 Jun
0
2
0
0
0
4 Jun
5
2
0
20
0
9 Jun
3
7
1/3
147
7
11 Jun
5
7
0
245
0
16 Jun
4
12
0
576
0
19 Jun
10
16
1/10
2,560
16
21 Jun
8
24
0
4,608
0
25 Jun
7
31
3/7
6,727
93
29 Jun
12
35
0
14,700
0
2 Jul
9
46
1/9
19,044
46
10 Jul
6
54
2/6
17,496
108
13 Jul
3
58
1/3
10.092
58
16 Jul
14
60
8/14
50,400
480
21 Jul
16
66
1/16
69,696
66
24 Jul
19
79
2/19
118.579
158
27 Jul
25
96
6/25
230,400
575
5 Aug
16
114
4/16
207.936
456
17 Aug
7
126
3/7
111.132
378
20 Aug
3
130
2/3
50.700
260
27 Aug
5
131
2/5
85.805
262
2 Sept
3
134
0
53.868
0
10 Sept
6
135
4/6
109.350
540
14 Sept
4
137
0
75.076
0
17 Sept
8
141
2/8
159,048
282
21 Sept
22
147
6/22
475,398
882
24 Sept
15
163
5/15
398.535
815
5 Oct
14
173
1/14
419.006
173
12 Oct
15
186
10/15
518,940
1.860
16 Oct
2
186
1/2
69,192
186
21 Oct
23
186
4/23
795.708
744
Totals
291
4,074,985
8.446
For each recapture phase: w, total number caught; x, total number previ-
ously handled up to. but not including, that phase (i.e.. the total marked up
to that time); y, the proportion of the catch previously handled (proportion
of catch with a mark).
Raw population estimate. Kitchen Shoals, is given by: p = Zwx2/Swxy
(Hayne's Index multiple census), p = 4074985/8446 = 482 male fishable
Guinea chicks. Shoal area measured by small square method = 17 ha.
Raw estimate of population density = 482/17 = 28 ha' '. In constructing
the above table, allowances have been made for animals dead from in-trap
predation.
caught. This was also suggested by Munro (1974) to explain a
similar sex ratio of trapped animals from the population of P.
guttatus lobsters at Jamaica (77 males to 37 females).
The population density of trappable Guinea chicks at the north-
eastern reef crest in the period May-October was estimated to be
greatest in the middle of this period. July through September (Ta-
ble 12). This correlates with the modal peak in trap rates that
occurred at the reef crest at Kitchen Shoals from mid-July to
mid-August (Fig. 12) (which resulted from the very sharp rise in
trap rates there during the last 2 weeks of July). It also correlates
with the above interpretation of the change in size composition at
Kitchen Shoals, viz. recruitment coupled with an inward breeding
Ecology and Behavior of Bfrmudan Lobster Panulirvs guttatus
TABLE MB.
Real sex ratio at Kitchen Shoals May-October.
The raw population estimate for females is 549^182 = 67 female
fishable Guinea chicks, by subtraction from the male + female
population estimate (Table 9).
The real sex ratio is therefore 7:1.
The sex ratio of trapped animals at Kitchen Shoals over the same
period was 13:1.
A conversion factor may therefore be computed for transforming the
sex ratio of trapped animals to real sex ratio: conversion factor =
7/13 = 0.54.
migration of egg-bearing females and smaller males, to the reef
crest (in the period May to August), entraining new fishery re-
cruits.
Catchabilily
The catchability of spiny lobsters and rock lobsters varies with
a whole series of environmental factors and conditions, including
TABLE 11C.
Difference in catchability between the sexes at Kitchen
Shoals, Mav-October.
1. Males
Raw population estimate = 482 Guinea chicks
Correct population estimate 482 x 1.1 = 530 chicks
Raw density estimate = 28 Guinea chicks ha" '
Corrected density estimate = 28 x 1.1 = 31 chicks ha '
C.P.U.E. = 291/1146 = 0.254 Guinea chicks trap night-1 or
CPUE. = 291/237 = 1.23 Guinea chicks trap haul '
q = 4.8 x 10"4 trap night '
E = 8.2 x 10"3 ha. trap night " '
q' = 2.3 x 10" ' trap haul"1
E' = 0.040 ha. trap haul '
2. Females
Raw population estimate = 67 Guinea chicks
Corrected population estimate 67 x 1.1 = 74 chicks
Raw density estimate = 67/17 = 3.9 Guinea chicks ha" '
Corrected density estimate = 3.9 x 1.1 = 4.3 chicks ha" '
C.P.U.E. = 22/1146 = 0.0192 Guinea chicks trap night"1 or
C.P.U.E. = 22/237 = 0.0928 chicks trap haul" '
q = 2.6 x 10"4 trap night
E = 4.5 x 10"' ha. trap night"1
q' = 1.3 x 10 "' trap haul"1
E' = 0.022 ha. trap haul '
3. Transformation of estimates
The conversion factor to transform catchability q from units of trap
night ' to trap haul"1 is given by:
Multiply by 4.8 for males
Multiply by 5.0 for females
Multiply by 4.9 for males + females
The conversion factor to transform effective area fished E from
units of ha. trap night" ' ha. trap haul ' is given by:
Multiply by 4.9 for males
Multiply by 4.9 for females
Multiply by 4.9 for males + females
For the period 23 May-21 October 1987, total catch = 291 males and 22
females: total effort = 1.146 trap nights, or 237 trap hauls.
Equations relating to estimates of catchability (q) and effective area fished
(E). CPUE, catch per unit of (fishing) effort: q. catchability coefficient: P.
population number: D. population density. CPUE, q.P (Nicholson and
Bailey 1935); CPUE, E.D (Miller 1986).
TABLE 12.
Capture-recapture population estimates at the Northeastern reef
crest (at Kitchen Shoals) for different periods during the summer
and autumn of 1987, by Bailey's Triple Catch Method.
V The Technique: Bailey's Triple Catch Method (Chalmers and
Parker. 1989) required a data set as follows:
( 1 ) The number of individuals that were marked in the first sample
and recaptured in the second, R, 2.
(2) The number of individuals that were marked in the second
sample and recaptured in the third, R2 3.
(3) The number of individuals that were marked in the first sample
and recaptured in the third, R, ,.
(4| The total number of animals caught in the second sample, S2.
The procedure carried out was as follows:
1. The "population of marked individuals" (M) was calculated
from the equation:
M = ([S2 x R1>3]/R2,3) + R,.;.
2. Then, the value of M was used to estimate the total population
N from the equation:
N = (M x S:)/R|,:.
3. The 95% confidence limit of the population estimate were
calculated from the equation associated with the Method
(Chalmers and Parker 1989):
the lunar and reproductive cycles. The chief of these are water
temperature, water salinity, and percentage of lobsters in a premolt
condition (Morgan 1974). Temperature and salinity are positively
correlated with catchability. and the percentage in premolt condi-
tion is negatively correlated (Morgan 1974).
The catchability of Guinea chick lobsters was higher at St.
David's Head and Cooper's Island in the autumn and early winter
than at Kitchen Shoals in the summer and autumn (Table 10D).
This may reflect the influence of the reproductive cycle and its
associated molting cycles, because the period of study at Kitchen
Shoals included the height of the breeding season, whereas the
other two study periods were after the height of season had passed.
TABLE 12B.
Spiny lobsters (P. guttatus) in the exploited phase at Kitchen Shoals
study area: Table for capture-recapture estimates of population
number by Bailey's Triple Catch Method.
Time of Capture
1
2
3
4
5
Time of last capture
June
4
13
1
1
1
Julv
—
8
7
10
4
August
—
—
3
3
1
September
—
—
—
3
9
October
—
—
—
—
2
Total marked
4
21
11
17
17
Total unmarked
53
74
25
47
41
Total caught
57
95
36
64
58
Total released
49
92
36
62
16
412
Evans and Lockwood
TABLE 12C.
Estimate for period 1, 1 June-31 August.
Rl.2 = 13
R2.3 = 1
R1.3 = 1
Si = 95, but 92 were released.
92 x 1
M = — = — + 13 = 26
26 x 95
N = — - — = 190
95% confidence limit of the population estimate (from the equation in
Part A of Table 12) = 190 ± 171. or 190 ± 90.0%.
TABLE 12D.
Estimate for period 2, 1 July-30 September.
R1.2
=
17
R2.3
=
3
R1.3
=
10
s2
=
36
M
=
36 x
3
10
— +
7 =
127
N
_
127 x
36
= 653
95% confidence limit of the population estimate (from the equation in
Part A of Table 12) = 653 ± 351 . or 653 ± 53.8%.
TABLE 12E.
Estimate for period 3, 1 August-21 October.
R,.: = 3
R2.3 = 9
R1.3 = 1
S; = 64. but 62 were released.
62 x 1
M = — — + 3 = 10
10 x 64
N = = 213
95% confidence limit of the population estimate (from the equation in
Part A of Table 12) = 213 ± 251. or 213 ± 118%.
TABLE 12F.
Summary of the population density estimates at Kitchen Shoals
using Bailey's Triple Catch Method.
Period 1 (1 June-31 August):
N =
= 190
Guinea chicks ± 90.0%.
Period 2 (1 July-30 Sept.):
N =
= 653
Guinea chicks ± 53.8%.
Period 3 (1 August-21 Oct.):
N =
= 213
Guinea chicks ± 118%.
The correction factor to take into account ( 1 ) predilection to reenter traps,
(2) the combined effect of tag/displacement-induced mortality and emigra-
tion, and (3) tag loss from molting was estimated to be x 1 .0, which when
applied, left these estimates unchanged (see Table 6D for estimation of
components 1 to 3). N, population estimate with 95% confidence limits,
calculated from the formula associated with the estimation equation (Table
12A).
At Bermuda, seawater temperatures remain relatively high during
the months May to November inclusive (22 to 28°C) but are some-
what lower in the months December to April ( 17 to 22°C) (Morris
etal. 1977).
Catchability can also be affected by intertrap separation if there
is overlap between the areas of attraction of each pot. Intertrap
separation was on average 159 m at St. David's Head, 123 m at
Cooper's Island, and 116 m ('/la of a nautical mile) at Kitchen
Shoals (nine traps in a cross-shaped pattern). It is not known if this
influenced the results, because the area of attraction is not accu-
rately known. However, it seems likely that only the traps set at
St. David's Head study area were free from interference by over-
lap, because Western rock lobsters can be attracted 80 m to a
baited lobster pot.
The lower catchability of females compared with males (both
in terms of q and E) at the reef crest in the period May to October
(including the height of the breeding season) indicates that females
are less catchable than males in such circumstances. Two possible
explanations are (1) that females are smaller and less powerful
than males and feed rather more within the labyrinthine reef struc-
ture than do males, and/or (2) that breeding females are less catch-
able than at other times.
The population density estimate for determination of the catch-
ability coefficient is preferably assessed by an independent method
(Miller 1986). Morgan (1974) however derived catchability coef-
ficient values from simultaneous observations on the population
density, catch, and effort, and from these values, he derived the
linear correlations with temperature, salinity, and percentage in
premolt, which are referred to above.
In the present study, the trap rate of the special central station
(pot 8) of the Transect 2 trapline (in a favored location at the reef
crest surrounded on three sides by overhanging labyrinthine coral
cliffs) was used for the calibration of the population density esti-
mate ( i . e . . to determine E for the purpose of the conversion of trap
rates: Table 13A to C). This station was a deep, sand-floored
TABLE 13.
Calculation of population density in the different physiographic
provinces: Table 13A. Density and catchability in the midsummer
period at the northeast reefcrest (Kitchen Shoals.)
From Table 12D, the estimate of population density for period 2 (1
July-30 September) was 653 ± 351 Guinea chicks (95% confidence
limits of Bailey's Triple Catch Method).
The upper and lower limits of this estimate are 302 and 1,004 Guinea
chicks, respectively. The study area was 17 ha. in area. Thus, the
upper and lower limits of the density estimate (D) for this period are
302/17 and 1.004/17 or 18 and 59 Guinea chicks ha~ ' respectively.
The catch per unit of fishing effort (CPUE) at the reef-crest station on
Transect 2 (pot 8) for this period was 43/13 = 3.3 Guinea chicks
trap hauP ' (see Table 13B below).
An estimate of the effective area fished (E) for this period may be
derived from the equation CPUE = E.D. by substitution of the
upper and lower limit of D:
Lower limit of E = 3.3/59 = 0.056 ha. trap haul" '
Upper limit of E = 3.3/18 = 0.183 ha. trap haul '
These may be used to estimate the 95% confidence upper and lower
limits of population density at the other physiographic provinces,
from the catch and effort data in Table 13B below. These estimates
are presented in Table 13C below.
TABLE 13B.
Spiny lobster P. guttalus in Bermuda: catch and effort on Transect 2.
Patch Reef Zone,
Reef Hat,
Fore Reef,
Main Terrace,
Edge (13-18 fathoms),
Pots 1-
-7
Pot 8
Pot 9
Pots 10-12
Pot 13
1987
c
f
CPLE
c
f
CPUE
c
f
CPUE
c
f
CPUE
c
f
CPUE
May
2
45
0.044
9
7
1.286
0
5
0
7
16
0.437
2
6
0.333
June
1
46
0.022
15
9
1.667
2
8
0.25
23
17
1.353
4
7
0.571
July
7
27
0.259
24
6
4
3
7
0.429
19
17
1.118
2
6
0.333
August
1
1
1
11
4
2.75
6
6
1
38
17
2.235
2
6
0.333
September
1
2
0.5
8
3
2.667
0
4
0
4
7
0.571
0
6
0
f, no. of traps hauled; c, no. caught
TABLE 13C.
Population density for the period 1 July-30 September, 1987, in the
different physiographic provinces as lower and upper limits of a
95% confidence band (densities in Guinea chicks ha ').
PRZ
RCA
FRS
RFT
1.6-5.4
18-59a
2.9-9.5
8.2-27
1.2-3.9
PRZ. patch reef zone; RCA. reef-crest area; FRS, fore-reef slope; RFT.
reef-front terrace; E, Edge.
* Density of fishable Guinea chicks estimated by the Bailey Triple Catch
Method, for Kitchen Shoals area. Example of calculation method (density
estimate for the reef- front terrace):
CPUE (Table 13B) = (19 + 38 + 4)/(17+ 17 + 4)
= 61/41
= 1 .5 Guinea chicks trap haul " '
Lower limit (LL) of estimate DLL =CPUE EUL
= 1.5/0.183 = 8.2 ha~ '
Upper limit (UL) of estimate DUL = CPUE/ELL
= 1.5/0.056 = 27 ha"'.
TABLE 14A.
Statistical test on the Hayne's Index population estimate for St.
David's Head.
Tabulate the proportion marked in catch (Y data) and the number
of animals marked previously (X data):
X
0
3
10
24
36
45
50
57
65
72
77
82
89
Y
0
0
0.067
0.077
0.357
0.500
0.500
0.143
0.429
0.583
0.222
0.417
0.500
Linear regression of above data by programmable calculator for
Y = A 4- BX:
A = 0.04455
B = 0.00527
r = 0.7511
The tabulated value of r in Murdoch and Barnes (2nd ed. 1970) is 0.6835
(p = 0.005) for a significant positive correlation (and there are 1 1 degrees
of freedom). The calculated value of r is greater than this, so there is a
significant positive correlation of the "proportion of marked animals in
catch" with "number of animals marked previously" at the 0.5% level of
significance (p = 0.005, single-tailed test for positive correlation). Fur-
thermore, the tabulated value for r at the 5% level is 0.4762. The result is
therefore a conclusive finding.
depression situated close to Kitchen Tower in the center of the
shoal on which it stands (the study area). The area's catch rates
were not used for the calibration, but the surrounding study area
was the site of the capture-recapture experiment. In this way, a
degree of independence was attained in the determination of the
density.
Robust mean estimates of effective area fished (E) may be used
to convert similarly obtained catch rates to density at other times
and places (Miller 1986).
Examples of the potential and actual application of conversion
are cited below within the context of the present study and the
potential for future research.
The estimate of E for the reef crest off Cooper's Island (7.8 x
10 ~3 ha trap night" ': Table 10D) may be used to transform the
trap rates of stations along Transect 1 to population density.
The estimate of E (as a 95% confidence band), which was
determined from the Bailey Triple Catch Method for midsummer
at Kitchen Shoals, was actually used to transform the simultaneous
station trap rates of Transect 2 to population density estimates in
TABLE 14B.
Statistical test on the Hayne's Index population estimate for St.
Cooper's Island.
Tabulate the proportion marked in catch (Y data) and the number
of animals marked previously (X data):
X
0
12
17
23
29
32
46
46
49
65
66
66
74
Y
0
0
0.200
0.167
0.375
0.250
0.500
0.200
0.250
0.667
0.778
0.429
0.333
Linear regression of above data by programmable calculator for
Y = A + BX:
A = 0.012418
B = 0.007595
r = 0.7728
The tabulated value of r in Murdoch and Bames (1970) is 0.6835 {p =
0.005) for a significant positive correlation (and there are 11 degrees of
freedom). The calculated value of r is greater than this, so there is a
significant positive correlation of the "proportion of marked animals in
catch" with ' 'number of animals marked previously ' ' at the 0. 5% level of
significance (p = 0.005, single tail test for positive correlation). Further-
more, the tabulated value of r at the 5% level is 0.4762. The result is
therefore a conclusive finding.
414
Evans and Lockwood
TABLE 14C.
Statistical test on the Hayne's Index population estimate for Kitchen
Shoals males and females combined.
TABLE 14D.
Statistical test on the Hayne's Index population estimate for Kitchen
Shoals males.
Tabulate the proportion marked in catch (Y data) and the number
of animals marked previously (X data):
X Y
0 0
2 0
3 0
3 0
8 0.500
8 0
13 0
17 0.100
25 0
32 0.429
36 0
47 0.111
55 0.333
59 0.167
64 0.571
70 0.0625
83 0.105
100 0.240
118 0.211
133 0.429
137 0.500
139 0.333
143 0
144 0.571
147 0
151 0.250
157 0.240
176 0.313
187 0.0625
202 0.667
202 0.500
202 0.200
Linear regression of above data by programmable calculator
for Y = A + BX:
A = 0.09234
B = 1.38 x 10 -}
r = 0.4541
There are 30 degrees of freedom, and the tabulated value of r in Murdoch
and Barnes (1970) is 0.4487 at the 0.5% level (single-tailed test for pos-
itive correlation). The calculated value of r is greater than this, and there-
fore, there is a positive correlation of the "proportion marked in catch"
with "number of animals marked previously" at the 0.5% level of signif-
icance (p = 0.005). Furthermore, the tabulated value for r at the 5% level
is 0.2960. The result is therefore a conclusive finding.
Tabulate the proportion marked in catch (Y data) and the number
of animals marked previously (X data):
X Y
0 0
1 0
2 0
2 0
7 0.333
7 0
12 0
16 0.100
24 0
31 0.429
35 0
46 0.111
54 0.333
58 0.333
60 0.571
66 0.0625
79 0.105
96 0.240
1 14 0.250
126 0.429
130 0.666
131 0.400
134 0
135 0.666
137 0
141 0.250
147 0.273
163 0.333
173 0.0714
186 0.666
186 0.500
186 0.174
Linear regression calculator for Y
A = 0.08548
B = 1.70 x 10~5
r = 0.4917
A + BX:
There are 30 degrees of freedom, and the tabulated value of r in Murdoch
and Barnes (1970) for the 0.5% level of significance is 0.4487. The cal-
culated value of r is greater than this, and therefore, there is a positive
correlation of the "proportion marked in catch" with "number of animals
marked previously" at the 0.5% level of significance (p = 0.005). single-
tailed test for positive correlation). Furthermore, the tabulated value for r
at the 5% level is 0.2690. The result is therefore a conclusive finding.
the present study (Table 13A to C). (These are two examples of
application for different place, same time.)
The mean estimate of E derived for the reef-crest areas (13 x
10~3 ha trap night-1) could be used to estimate density from
Guinea chick trap rates at any time interval (in the months June
through November) and for other areas of the Bermudan reef sys-
tem.
Such studies of catchability in terms of effective area fished (E)
thus have practical value as a tool for the assessment of spiny
lobster resources of tropical islands.
ACKNOWLEDGMENTS
We thank Mr. Arthur Evans for field work assistance, for tech-
nical assistance in trap construction and maintenance, and for the
line drawings. We also thank fisherman Roger Hollis for the use of
his fishing vessel Jocelyn for the transects and work in the study
areas. We also thank fisherman Lynwood Outerbridge for allow-
ing us to measure the Guinea chicks he landed from the northern
reef system.
Ecology and Behavior of Bf.rmudan Lobster Panulirus cuttatus
415
Grateful thanks are also given to Jack Ward, Dr. Brian Luck-
hurst, and Dr. James Burnett-Herkes of the Bermuda Department
of Agriculture, Parks and Fisheries, for logistical support with
traps and ropes and for suggesting transect studies.
We also thank Dr John Tarbit of the U.K. Overseas Devel-
opment Administration for the sponsorship and funding of this
project through the O.D.A. Natural Resources and Environ-
ment Division. Dr. Anthony Knap, of Bermuda Biological
Station, also assisted the project through an internship award
for laboratory fees, and we thank him for this and other logis-
tical support. We also thank Dr. Colin Bannister of the U.K.
Ministry of Agriculture, Fisheries and Food for encourage-
ment and help in regard to the fisheries aspects of the early anal-
yses. Finally, we thank the Sigma Xi Society for Scientific Re-
search for a grant-in-aid toward the cost of the 'Floy' tagging
equipment.
APPENDICES:
STATISTICAL ANALYSES OF THE POPULATION ESTIMATES
MADE BY THE HAYNE's (1949) METHOD.
The Hayne's Index population estimates are based upon
Hayne's (1949) equation for population number P:
P = 2wx2/2wxy
Schumacher and Eschmeyer (1943) used this method to
compute an estimate of fish population number from records of
netting, marking, and releasing.
The formula is the inverted form of the usual expression for slope
of a regression line passing through the origin (Hayne 1949).
The linear regression analyses carried out in Tables 14A to
D below are "free-standing," i.e., the regression lines are not
"forced through the origin."
LITERATURE CITED
Caillouet, C. W.. G. L. Beardsley & N. Chitty. 1971. Notes on size, sex
ratio and spawning of the spiny lobster Panulirus guttatus (Latreille),
near Miami Beach, Florida. Bull. Mar. Sci. 21:944-951.
Chalmers. N. & P. Parker. 1989. The O.U. Project Guide: fieldwork and
statistics for ecological projects. In: J. Crothers, F.R.C The Field
Studies Council, No. 9 in a Series of Occasional Publications. 2nd ed.
108 pp.
Chittleborough. R. G. 1974. Home range, homing and dominance in ju-
venile Western rock lobsters. Aust. J. Mar. Freshwater Res. 25:227-
234.
Davis, G. E. 1978. Field evaluation of a tag for juvenile spiny lobsters
Panulirus argus. Trans. Am. Fish. Soc. 107:100-103.
Evans, C. R. 1988. Final Report of the Research Project Population Dy-
namics of Spiny Lobsters Panulirus argus and P. guttatus (Latreille) at
Bermuda.' Res. Rep. Dep. Oceanogr. Univers. Solon. July 1988. 2
vols., 241 pp
Evans, C. R. 1989. Population dynamics and ecology of spiny lobsters
Panulirus argus and P. guttatus at Bermuda Ph.D. Thesis, Depart-
ment of Oceanography, University of Southampton. March 1989, 2
vols.. 474 pp.
Evans, C. R. & A. P. M. Lockwood. 1995. Field studies of the repro-
ductive biology of the spiny lobsters Panulirus argus (Latreille) and P
guttatus (Latreille) at Bermuda. Journal of Shellfish Res. (in press).
Evans. C. R.. A. J. Evans & A. P.M. Lockwood. 1994. Scaled-down
shelters (casitas) with predator barriers having potential for spiny lob-
sters Panulirus argus nursery habitat enhancement. Paper accepted for
Langosta '94: Mexico-Cuba International Workshop on the Manage-
ment of Artificial Shelters in the Fisheries for Spiny Lobsters and
Population Dynamics, Havana. Cuba. 17-21 October 1994. 30 pp.
Hayne. D. W. 1949. Two methods for estimating population from trap-
ping records. J. Mammal. 30:399-411.
Jernakoff, P. & B. F. Phillips. 1986. Electromagnetic tracking of juvenile
rock lobsters. Aust. Fisheries 1986(May):32-34.
Mathers, P. 1981. Environmental Biology Course Lectures, B.Sc. Com-
bined Sciences Degree Course. Dorset Institute of Higher Education.
Bournemouth.
Miller, R. J. 1986. Traps as a survey tool for animal density. Paper pre-
sented at the 39th Annual Sessions of the Gulf and Caribbean Fisheries
Institute, Hamilton. Bermuda. November 1986. 16 pp.
Morgan, G. R. 1974. Aspects of the population dynamics of the Western
rock lobster, Panulirus cygnus (George). II Seasonal changes in the
catchability coefficient. Aust. J . Mar. Freshwater Res. 24:249-259.
Morns. B . J Bames. F. Brown & J. Markham (eds.) 1977. The Bermuda
Marine Environment: A Report of the Bermuda Inshore Waters Inves-
tigations 1976-1977. Bermuda Biological Station Special Publication,
(15); 120 pp.
Murdoch, J. & J. A. Barnes. 1970. Statistical Tables for Science, Engi-
neering. Management and Business Studies. 2nd ed. Macrmllan, Lon-
don, 46 pp.
Munro. J. L. 1974. The biology, ecology, exploitation and management of
Caribbean reef fishes. Scientific Report of the O.D.A. /University of
West Indies Fisheries Ecology Research Project, 1962-1973. Part VI.
The biology, ecology, and bionomics of Caribbean reef fishes-
crustaceans (spiny lobsters and crabs). Univ. West. Indies Zool. Dep.
Res. Rep. 3:1-57.
Nicholson, A. J. & V. A. Bailey. 1935. The balance of animal popula-
tions. Proc. Zool. Soc. ton. 1935:551-598.
Pollock, D. E. 1992. Palaeoceanography and speciation in the spiny lob-
ster genus Panulirus in the Indo-Pacific. Bull. Mar. Sci. 51:135-146.
Ricker, W. E. 1975. Computation and interpretation of biological statistics
of fish populations. Fish. Res. Bd. Can. Bull. 191, 302 pp., Ottawa.
Schumacher. F. X. & R. W. Eschmeyer. 1943. The estimation of fish
population in lakes or ponds. J . Tenn. Acad. Sci. 18:228-249.
Sutcliffe, W. H. Jr. 1952. Some observations on the breeding and migra-
tion of the Bermuda spiny lobster Panulirus argus. Proc. Gulf. Carihh.
Fish. Inst. 4:64-69.
Sutcliffe, W. H. Jr. 1953. Notes on the biology of a spiny lobster Panu-
lirus guttatus in Bermuda. Ecology 34:794-796.
Sutcliffe, W. H. Jr. 1953a. Further observations on the breeding and mi-
gration of the Bermuda spiny lobster Panulirus argus. J. Mar. Res.
12:173-183.
Sutcliffe. W. H. Jr. 1956. Effect of light intensity on the activity of the
Bermudan spiny lobster Panulirus argus. Ecology 37:200-201.
Journal of Shellfish Research, Vol. 13, No. 2, 417-423, 1994.
SPAWNING CYCLE OF THE RED CLAM MEGAPITARIA AURANTIACA (SOWERBY, 1831)
(VENERIDAE) AT ISLA ESPIRITU SANTO, BAJA CALIFORNIA SUR, MEXICO
FEDERICO GARCIA-DOMlNGUEZ,*
SILVIA ALEJANDRA GARClA-GASCA AND
JOSE LUIS CASTRO-ORTIZ
Centro Interdisciplinario de Ciencias Marinas
Institute Politecnico National
Apdo. Postal 592. La Pa:. B.C.S. 23000. Mexico
ABSTRACT Adult red clams. Megapitaria aurantiaca, were collected at Isla Espiritu Santo, B.C.S. , Mexico, from May 1991 to
May 1992. Gonadal development was analyzed using histological techniques. Development phases were categorized into five stages:
indifferent, developing, ripe, partially spawned, and spent. Spawning occurred all year except in September, with peaks in December
and March. A clear relation between temperature and spawning was not observed. Spawning occurs all during the year. Nevertheless,
through the year, two distinct periods are observed: one, from May to November, with the mean of the partially spawned population
at 8.8, coincides with warmer months of the year; and the second, from December to April, with a mean of 38.9%, coincides with
cooler months.
KEY WORDS: Spawning cycle, bivalves, Megapitaria. histology
INTRODUCTION
The red clam Megapitaria aurantiaca (Sowerby, 1831 ) is com-
mercially fished along the west coast of Baja California Sur (Hol-
guin 1976). It lives in the sublittoral zone to 25 m depth in medium
sand, coarse sand, and coarse sand-pebble areas (Baqueiro 1979).
The increasing demand for this species, its high cost, and the
decrease of natural stocks make it a prime candidate for commer-
cial culture (Baqueiro 1989).
Studies of reproduction using histological techniques have been
made on Veneridae clams like Cyprina islandica (Loosanoff
1953), Venerupis japonica (Holland and Chew 1974), Ameghino-
mya antiqua (Verdinelli and Schuldt 1976), Venus striatula
(Ansell 1961), Megapitaria squalida and Dosinia ponderosa
(Baqueiro and Stuardo 1977), Venus antiqua (Lozada and Bustos
1984). Callista chione (Valli et al. 1984), Mercenaria mercenaria
(Eversole et al. 1980, Manzi et al. 1985). Chione fluctifraga (Mar-
tinez-Cordova 1988). C. undatella (Baqueiro and Masso 1988),
Mercenaria spp. (Hesselman et al. 1989). and Chione californi-
ensis (Garcfa-Domi'nguez et al. 1993). For M. aurantiaca. only
one study of reproduction, also including some ecological and
biological aspects, was made in Bahia Zihuatanejo, Guerrero,
Mexico (Baqueiro and Stuardo 1977).
This work was made to study the spawning cycle of a wild
adult population of M. aurantiaca at Isla Espiritu Santo. B.C.S..
Mexico, in relation to temperature and also to condition index
(standard weight), because reproductive events have an influence
on the weight of an organism. It was expected that the weight
would increase during gametogenesis and maturity and decrease
during spawning.
MATERIALS AND METHODS
From May 1991 to May 1992, 25 specimens per month of an
adult population of clams located at Isla Espiritu Santo, B.C.S.
(Fig. 1 ) were collected by scuba dive at 3 m depth. A total number
of 290 organisms were captured. At the time of the collection of
biological samples, surface water temperature was recorded.
Before dissection, shell length was measured with a 0.001 mm
resolution caliper. Shell and tissue were blotted dry, and total and
wet weight without the shell was measured with a balance that read
to the nearest 0.1 g. Mantle, adductor muscles, gills, labial palps,
and siphons were removed, keeping only the visceral mass (gonad,
liver, and gastrointestinal tract) and the foot. These tissues were
fixed in buffered 10% formalin and processed using histological
techniques (Humason 1979). Paraffin sections 7 to 9 u,m thick
were stained with hematoxylin and eosin.
Categories of Gonadal Condition
The reproductive process (either spermatogenesis or oogenesis)
of M. aurantiaca was divided into five stages (indifferent, devel-
oping, ripe, partially spawned, and spent) based solely on mor-
phological observations. Categories comparable to those already
in use for other species have also been used in this study where
appropriate (Brousseau 1981, for Petricola pholadiformis; Brous-
seau 1982, for Geukensia demissa; Manzi et al. 1985, for M.
mercenaria: Brousseau 1987, for Mya arenaria: Malachowski
1988, for Hinnites giganteus: Hesselman et al. 1989, for Merce-
naria spp.; Ponurovsky and Yakovlev 1992, for Tapes phillippi-
narium: Jaramillo et al. 1993, for Chlamys amandi: and Garcia-
Dominguez et al. 1993, for C. californiensis) .
Developmental Stages of the Male
Indifferent Stage (Fig. 2a)
This was characterized by an absence of gametes; however,
residual spermatozoa were occasionally observed in some speci-
mens. Vesicular connective tissue between follicles occupies al-
most all of the space. Gonoduct secondary ducts were open and
empty. These ducts have columnar, ciliated epithelium, with basal
nuclei and acidophilic granules in the cytoplasm.
Developing Stage (Fig. 2b)
The spermatogenic cells begin to proliferate around the follicle
walls. Inside the follicle, a variable quantity of germinal cells and
ripe gametes were observed. Spermatozoa are stored as a dense
mass in the lumen of the follicle waiting to spawn. The connective
417
418
GARCfA-DOMlNGUEZ ET AL.
— 24- 3D' N
BAHIA DE
I A PAZ
110" 15' W
Figure 1. Study area. Isla Espiritu Santo, B.C.S., Mexico.
tissue between follicles decreases when follicles increase because
of the accumulation of sperm.
Ripe Stage (Fig. 2c)
The follicles were distended and filled with dense, radiating
bands of spermatozoa, the tails of which project into central lu-
men. Almost all connective tissue between follicles has been sub-
stituted by follicles full of spermatozoa.
Partially Spawned Stage (Fig. 2d)
This is the reproductive stage, and spermatozoa are expelled
into environment. The follicles were partially empty. There is a
marked decrease in the number of spermatozoa filling the lumen.
This is the only stage where some secondary ducts are full of
spermatozoa.
Spent Stage (Fig. 2e)
The follicles collapsed or had decreased in size and are invaded
by amebocytes, which phagocytosed a small amount of unspent
spermatozoa. There is no evidence of active spermatogenesis tak-
ing place.
Developmental Stages of the Female
Indifferent Stage (Fig. 3a)
This stage is characterized by absence of gametes. The follicles
are partially compressed in shape or size and are empty except for
occasional residual free oocytes. Vesicular connective tissue oc-
cupies the space between follicles.
Developing Stage (Fig. 3b)
This stage is a continuous process, involving a proliferation
and growth of the oocytes. As the number of mature ova (free in
the lumen) increased, the amount of connective tissue decreased.
The developing, oocytes, which begins as hemispherical stalked
cells attached to the wall of the follicle, become enlarged spherical
cells, 45.9 |jun (standard deviation (SD], 5.5) in diameter, as
maturity approaches.
Ripe Stage (Fig. 3c)
The ripe ovary is characterized by the presence of distended
follicles filled with ripe oocytes, some of which are attached to the
follicular wall by slender stalks. Little or no connective tissue was
present.
Partially Spawned Stage (Fig. 3d)
Some follicles contain oocytes, whereas others are empty.
There is a noticeable reduction in the number of free large oocytes
present in the lumen, 52.5 u,m (SD, 4.7) in diameter. Little con-
nective tissue was present. Follicular walls are broken.
Spent Stage (Fig. 3e)
The follicles were empty except for a few, large, unspent
oocytes free in the lumen that are phagocytosed by amebocytes.
Follicular walls are reestablished.
A condition index, the standard weight (SW) (Searcy 1984),
was used to relate reproductive events with weight variation. Con-
dition is interpreted as the fatness of the organism. SW was esti-
mated using the equation:
SW = a lb
where 1 is a constant length of 95 mm and a and b are parameters
estimated with a nonlinear regression analysis using the length and
weight data of each monthly sample.
RESULTS
Reproductive Cycle
The annual reproductive cycle of the red clam M. aurantiaca
from Isla Espiritu Santo is summarized in Figure 4. Ripe clams
were found every month, except April 1992. The ripe phase pre-
dominated in August 1991 and May 1992 (60 and 55% of the
clams). Reproductive individuals (partially spawned) were ob-
served all years except in September 1991. The maximum was in
December 1991 and March 1992. when spawning individuals were
50 and 68% of the population and the temperature was 22°C. Spent
red clams were observed in May. July, August. November, and
December 1991, and January, February. March, and April 1992.
Clams in the active phase were encountered throughout the study
except February and March 1992. The largest number of active
clams occurred in April 1991 (61% of the individuals). Indifferent
clams were observed all year except October 1991. Gonadic ac-
tivity was synchronic for both sexes. The sex ratio was 44.8%
males and 55.2% females.
Standard Weight
The SW varied between 4.5 g in June 1991 and 13.9 g in April
1992 (Fig. 5). The mean length of the clams was 95.0 mm (SD,
12.01).
Temperature
The minimum temperature of 20°C was recorded in January
and February, and the maximum of 30°C was in September
(Fig. 5).
DISCUSSION
The spawning cycle of M. aurantiaca shows the same phenom-
ena observed for other venerid clams: growth and development of
Spawning Cycle of the Red Clam
419
*a#r
" *' &&
Figure 2. Photomicrographs of gonadal stages of the male red clam M. aurantiaca. (a) Indifferent male ( x 100); (b) developing male i > iooi: (c)
ripe male (X100); (d) partially spawned male (SC, secondary ducts full of spermatozoa) (X100); (e) spent male (X400).
420
GARCfA-DOMfNGUEZ ET AL.
rf""* ^
-
4' •
.V
* . »
...a'
■"*■ *^
/ *
K
'
1
.
1
•
» » •*
T^fi*
.
•».
"
» 1
1*
to
v'
4
1
■Jw9\Jr *•
3b ^ '^*J
■f
v^
Figure 3. Photomicrographs of gonadal stages of the female red clam M. aurantiaca. (a) Indifferent female (xlOO); (b) developing female
(xlOO); (c) ripe female (xlOO); (d) partially spawned female (XlOO); (e) spent female (XlOO).
Spawning Cycle of the Red Clam
421
100
>-
o
z
LU
Z)
o
LU
LT
LL
LU
>
h-
LU
DC
SPENT
PART SPAWN
RIPE
DEVELOPING
INDIFFERENT
S O N D J
MAY 1991 -MAY 1992
M
M
Figure 4. Relative frequency of M. aurantiaca population with gonads in each gonadal stage during 1991 to 1992. Observations of males and
females are combined.
gametes, ripening, spawning, residual gametes absorbed, and rest
or inactivity (Baqueiro and Stuardo 1977, Manzi et al. 1985; Hes-
selman et al. 1989).
The result of gonadal examination showed no clearly defined
seasonal reproductive cycle for M. aurantiaca at Isla Espiritu
Santo, Mexico. Spawning occurs all during the year as a contin-
uous phenomenon. Nevertheless, through the year, two distinct
periods are observed. The first one coincides with the warmer
months of the year, from May to November 1991, where the
maximum partially spawned population is 28% in August, the
minimum is zero in September, and the monthly mean is 8.8%.
The second period coincides with cooler months, from December
1991 to April 1992, and has two major peaks: December 1991
(22°C), when partially spawned individuals were 68% of the pop-
F M A
19 9 2
Figure 5. Relation of some gonadal stages (developing, ripe, and par-
tially spawned) with water temperature and SW. Observations of
males and females are combined.
ulation. and March 1992 (22°C) with 40% of the population par-
tially spawned. There was a minimum of 16% in January 1992,
and the monthly mean was 38.9%.
Baqueiro and Stuardo (1977) studied a population of M. au-
rantiaca at Bahia Zihuatanejo, Mexico, and found that reproduc-
tion extends all year, independent of temperature and with two
maximum periods: October (28.8°C), when spawning individuals
were 62% of the population, and May (27.8°C), with 40% of the
population spawning.
This does not indicate important differences in the reproductive
pattern of this species because, per Sastry (1970). the reproductive
cycle of a species living in different climatic zones can vary be-
cause a species' reproduction is in response to the environment.
Differences in the gonadal cycle of different populations of M.
mercenaria have been suggested as a result of different phenotypic
responses to the variation of environmental factors (Porter 1964,
Hesselman et al. 1989). In populations of other marine bivalves,
regional and local differences of gonadic cycles have been sug-
gested (Ropes and Stickney 1965, Holland and Chew 1974;
Thompson et al. 1980, Eversole et al. 1980, Garcia-Domi'nguez et
al. 1993).
In some bivalves, like Placopecten magellanicus, reproductive
output varies not only between populations from different sites,
but also between consecutive years in a given population. This
suggested that gamete production is strongly influenced by envi-
ronmental factors (MacDonald and Thompson 1985), such as tem-
perature and food availability set in a seasonal context, that con-
dition both the reproductive effort and the timing of reproductive
events (Bayne and Newell 1983).
In this work, only temperature was considered and we did not
observe a very clear relation with spawning. Similarly, Baqueiro
and Stuardo (1977) did not observe a clear relation between
spawning and temperature. This phenomenon has been observed
in other venerid clams like M. squalida and D. ponderosa (Baque-
iro and Stuardo 1977). In V. japonica (Holland and Chew 1974),
V. antiqua (Lozada and Bustos 1984), and M. mercenaria (Manzi
422
GARCfA-DOMfNGUEZ ET AL.
et al. 1985), a clear relation between temperature and gonadic
activity has been established.
In the first period of the reproductive cycle of M. aurantiaca,
we observed that changes in gonadic ripeness are reflected in SW
variations. From May to August 1991. a gradual increase of the
SW is seen until a maximum in August 1991, coinciding with a
maximum of ripe and spawning clams. Subsequently, the SW and
the frequency of ripe and spawning organisms decreased. Searcy
(1984) found that, in some months, variations of SW of Tivela
stultorum, a Veneridae clam, can be associated with spawning
because, when SW decreases, the frequency of spawning organ-
isms increases.
In the second period of the cycle of M. aurantiaca, correspond-
ing to cooler months, a clear relation between SW and ripe stage,
gametogenesis, or spawning is not observed. This indicates that
SW is not a good index to quantify reproductive activity. Condi-
tion indices, such as SW, are important because of the possibility
of using them to make general inferences about the reproductive
cycle of a species (Hickman and Illingworth 1980), but they can be
modified by other factors like rain (Searcy 1984) and nutritional
state (Crosby and Gale 1990), neither of which were considered in
this study.
ACKNOWLEDGMENTS
Our gratitude to the Direccion de Estudios de Posgrado e In-
vestigacion del Institute Politecnico Nacional (IPN). who gave us
the funds for this work, to M. Sc. Arturo Tripp Quezada for his
help in collecting samples, and Dr. Ellis Glazier for his editorial
help on the English manuscript. The following IPN fellowships
were awarded: Comision de Operacion y Fomento de Actividades
Academicas to F. Garcfa-Dominguez and J. L. Castro-Ortiz and
Programa Institucional de Formacion Investigadores to S. A. Gar-
cia-Gasca.
REFERENCES
Ansell. A. D. 1961. Reproduction, growth and mortality of Venus stria-
tula (Da Costa) in Karnes Bay, Millport. J. Mar. Biol. Assoc. U.K.
41:191-215.
Bayne, B. L. & R. C. Newell. 1983. Physiological energetics of marine
molluscs, pp. 491-498. In: A. S. M. Saleuddin and K. M. Wilbur
(eds.). The Mollusca. vol. 4. Academic Press, New York.
Baqueiro. E. 1979. Sobre la distribucion de Megapitaria aurantiaca (Sow-
erby), M. squalida (Sowerby) y Dosinia ponderosa (Gray) en relacion
a la granulometria del sedimento (Bivalvia: Veneridae): Nota cienti-
fica. An. Centro Cienc. del Mar y Limnol. Univ. Nat. Auton. Mexico
6:25-32.
Baqueiro, E. 1989. Clam culture in Mexico: past, present and future, pp.
383-394. In: J. J. Manzi and M. Castagna (eds.). Clam Manculture in
North America. Elsevier Scientific Publishers, Amsterdam.
Baqueiro, E. & J. A. Masso. 1988. Vanaciones poblacionales y repro-
duccidn de dos poblaciones de Chione undatella (Sowerby, 1 835). bajo
diferentes regimenes de pesca en la Bahia de La Paz. B.C.S., Mexico.
Cienc. Pesq. Inst. Nal. Pesca. Mexico 6:51-67.
Baqueiro. E. & J. Stuardo. 1977. Observaciones sobre la biologia, eco-
logia y explotacion de Megapitaria aurantiaca (Sow., 1835), M.
squalida (Sow., 1835) y Dosinia ponderosa (Gray, 1838) (Bivalvia:
Veneridae) de la Bahia de Zihuatanejoe Isla Ixtapa, Gro., Mexico. An.
Centro Cienc. del Mar y Limnol. Univ. Nal. Auton. Mexico 4:161-
208.
Brousseau. D. J. 1981. Spawning cycle in a population of Petricola pho-
ladiformis (Pelecypoda: Petncolidae) from Milford. Connecticut. The
Veliger 24:56-61.
Brousseau, D. J. 1982. Gametogenesis and spawning in a population of
Geukensia demissa (Pelecypoda: Mytilidae) from Westport. Connecti-
cut. The Veliger 24:247-251.
Brousseau, D. J. 1987. A comparative study of the reproductive cycle of
the soft-shell clam, Mya arenaria in Long Island Sound. J. Shellfish
Res. 6:7-15.
Crosby, M. P. & L. D. Gale. 1990. A review and evaluation of bivalve
condition index methodologies with a suggested standard method. J
Shellfish Res. 9:233-237.
Eversole, A. G., W. K. Michener & P. J. Eldndge. 1980. Reproductive
cycle of Mercenaria mercenaria in a South Carolina estuary. Proc.
Natl. Shellfish. Assoc. 70:22-29.
Garcfa-Dominguez. F., G. Garcia-Melgar & P. Gonzalez-Ramirez. 1993.
Reproductive cycle of the clam Chione californiensis (Broderip, 1835),
in Bahia Magdalena, Baja California Sur. Mexico. Ciencias Marinas
19:15-28.
Hesselman, L. D., B. J Barber & N. J. Blake. 1989. The reproductive
cycle of adult hard clams. Mercenaria spp. in the Indian River Lagoon.
Florida. J. Shellfish Res. 8:43-19.
Hickman. R. W. & J. Illingworth. 1980. Condition cycle of the green-
lipped mussel Perna canaliculatus in New Zealand. Mar. Biol. 60:27-
38.
Holgufn, Q. O. 1976. Catalogo de especies mannas de importancia com-
ercial en Baja California Sur. Subs. Pesca. Mexico. Inst. Nal. de
Pesca. 177 p.
Holland. A. D. & K. K. Chew. 1974. Reproductive cycle of the manila
clam (Venerupis japonica), from Hood Canal. Washington 1963.
Proc. Natl. Shellfish. Assoc. 64:53-58.
Humason, G. L. 1979. Animal Tissue Techniques. W. H. Freeman and
Co.. San Francisco, 661 pp.
laramillo. R , J. Winter, J. Valencia & A. Rivera. 1993. Gametogenic
cycle of the chiloe Scallop (Chlamys amandi). J Shellfish Res. 12:59-
64.
Loosanoff, V. L. 1953. Reproductive cycle in Cyprina islandica. Biol.
Bull. (Woods Hole) 104:146-155.
Lozada, E. & H. Bustos. 1984. Madurez sexual y fecundidad de Venus
antiqua antiqua King y Brodenp 1835 en la Bahia Ancud (Mollus-
ca:Bivalvia:Veneridae). Rev. Biol. Mar. Valparaiso 20:91-1 12.
MacDonald, B. A. & R. J. Thompson. 1985. Influence of temperature
and food availability on the ecological energetics of the giant scallop
Placopecten magellanicus. II. Reproductive output and total produc-
tion. Mar. Ecol. Prog. Ser. 25:295-303.
Malachowski, M. 1988. The reproductive cycle of the rock scallop Hin-
nites giganteus (Grey) in Humboldt Bay, California. J. Shellfish Res.
7:341-348.
Manzi, J. J., M Y. Bobo & V G. Burrell. 1985. Gametogenesis in a
population of the hard clam, Mercenaria mercenaria (Linnaeus), in
North Santee Bay, South Carolina. The Veliger 28:186-194.
Martinez-Cordova. L. R. 1988. Bioecologia de la almeja negra Chione
fluctifraga (Sowerby, 1853). Rev. Biol. Trop. 36:213-219.
Ponurovsky, S. K. & Y. M. Yakovlev. 1992. The reproductive biology of
the Japanese httleneck. Tapes phillippinarium (A. Adams and Reeve,
1850) (Bivalvia: Veneridae). J. Shellfish Res. 11:265-267.
Porter, H. J. 1964. Seasonal gonadal changes of adults clams Mercenaria
mercenaria (L.), in North Carolina. Proc. Natl. Shellfish. Assoc. 55:
35-52.
Ropes, J. W. & A. P. Stickney. 1965. Reproductive cycle of Mya are-
naria in New England. Biol. Bull. (Woods Hole) 128:315-327.
Sastry, A. N. 1970. Reproductive physiological variation in latitudinal
Spawning Cycle of the Red Ceam
423
separated populations of the bay scallop, Aequipecten aequisulcatus
Lamarck. Biol. Bull. (Woods Hole) 138:56-65.
Searcy, B. R. 1984. Un estudio sobre la condicion de la almeja pismo
Tivelu stultorum con datos de longitud y peso de la carne de capturas
comerciales. Ciencias Marinas 9:19-30.
Thompson. I.. D. S. Jones &J. W Ropes 1980. Advanced age for sexual
maturity in the ocean quahog Arctica islandica (Mollusca: Bivalvia).
Mar. Biol. 57:35-39.
Valli, G, E. Vidoli & C. Marussi. 1984. Osservazioni preliminari sulla
riproduzione e sulla biometna di Callista chione (L.) (Mollusca: Bi-
valvia) del Golfo di Trieste. Nova Thalassia 6:97-103.
Verdinelli, M. A. & M. Schuldt. 1976. Consideraciones preliminares so-
bre aspectos de la dinamica poblacional y reproduccidn de la almeja
rayada [Ameghinomya antique! King-Chiomdae) en punta Loma. Golfo
Nuevo — Chubut. Rev. Museo de La Plata (Nueva Serie) Seccion Zool.
22:183-202.
Journal of Shellfish Research. Vol. 13, No. 2, 425-431, 1994.
QUANTIFYING SEASONAL VARIATION IN SOMATIC TISSUE: SURFCLAM SPISULA
SOLIDISSIMA (DILLWYN, 1817)— A CASE STUDY1
JOSEPH G. LOESCH AND DAVID A. EVANS
The College of William and Mary
Virginia Institute of Marine Science
School of Marine Science
Gloucester Point, Virginia 23062
ABSTRACT Condition indexes are commonly derived from bivalve species. Usable meat yields (UMY. in 1/bu) from 181 daily
landings of Atlantic surfclams. Spisula solidissima (Dellwyn. 1817), at a Virginia processing plant in 1974 and 160 landings in 1975
were used as an index in our analysis. The data were fitted to a basic sinusoidal model and a two-compartment sinusoidal model to
demonstrate the utility of these models for quantifying cyclic events. The basic model, x = xa + A cos2m + B sin2ir/. is linear in
its independent variables and fitted by multiple regression, with x = UMY, t = time in years, where x0. A, and B are constants
determined by the regression procedure U0 = mean UMY). Its alternate form is x = .v„ + r cos2rr(r - („), with x, x0, and t as before.
r = amplitude of the sinusoidal variation, and i0 = time when the maximal UMY occurs; r and tu are related to A and B as r =
Va- + B2. and t0 = ( l/2n|tan~ ' {BIA). The sinusoidal fit to the 1974 data was highly significant (/> < 0.0005); therefore, the null
hypothesis that the data are not a function of time was rejected. The annual mean yield. a„. was 5.93 1/bu, ;„ was 0.45 (i.e., the
maximal UMY occurred about mid-June), and the amplitude r was 0.730; thus, the difference between the lowest and highest yields.
2r. was almost 1.5 1/bu. Similar estimates were determined from the 1975 data and the combined data. The fit was recalculated for
both data sets after excluding apparent outliers. As expected, the root-mean-square residual (/?MSrf>) decreased, whereas the coef-
ficient of determination (R2) increased with the removal of the apparent outliers, but the fitted parameters were inconsequentially
affected. A fit of the data to a two-component sinusoidal model, x = x0 + A , cos2ir/ + B, sin2Tr/ + A2 cos4ir; + B2 sin4Tr/, modeled
an annual variation with an asymmetric rise and fall. As a demonstration, the data were also fitted to a parabolic model, x = a0 +
att + a2r. Although this model produced fits comparably as close as the sinusoidal models, the coefficients are not interpretable in
a simple manner, as is the case with the sinusoidal fits, and it does not allow asymmetric behavior.
KEY WORDS: Spisula solidissima. condition index, usable meat yields, seasonal variation, maximum, minimum, sinusoidal,
parabolic
INTRODUCTION Source of Data
Condition indexes are commonly derived for bivalve species.
Various index models have been used; in general, the condition
indexes reflect a relationship between soft tissue weight and the
size of the cavity formed by the two valves. The indexes are used
primarily to estimate seasonal meat quality or the effects of disease
and pollution on meat quality. It has been suggested that a con-
dition index for oysters be used to monitor pollution. Lawrence
and Scott (1982) and Crosley and Gale (1990) reviewed and eval-
uated bivalve condition index methodologies; in each study, the
authors recommended that a standardized index be used, although
their models were somewhat different. The presentations and lit-
erature cited by those authors and references in the index of papers
published in the Journal of Shellfish Resource (Castagna et al.
1992, Mann et al. 1993) provide an ample introduction to bivalve
condition indexes.
Herein, we present methodologies for estimating seasonal in-
dexes, the maximal and minimal annual values, associated confi-
dence intervals, and tests of significance, regardless of the condi-
tion index used.
METHODS
Condition Index
To demonstrate the model, we use a condition index defined as
usable meat yields (UMY) in liters per bushel of the Atlantic
surfclam. Spisula solidissima (Dillwyn. 1817).
The UMYs were determined from daily landings of surfclams
at the C&D Seafood Co. in Oyster, Virginia — 181 landings total-
ing 167,564 bushels in 1974 and 160 landings totaling 270,170
bushels in 1975. In both years, the surfclams were harvested in an
area approximately between 8.5 to 17.5 nautical miles offshore of
Cape Henry and south to the North Carolina state line.
Sinusoidal Model
Loesch (1977) reported the relationship between mean monthly
water temperature and mean monthly usable meat yield per bushel
(mean UMY) for surfclams. The data in terms of daily UMYs are
resurrected herein to assess parameters not previously considered
in order to demonstrate the utility of sinusoidal functions for quan-
tifying cyclic events exhibited in the life history of many marine
species.
The basic sinusoidal model used was
where
x = xq + A cos2-rrf + B sin 2-rrf
x = UMY in 1/bu
'Contribution 1896 of The College of William and Mary, Virginia Institute
of Marine Science. School of Marine Science, Gloucester Point, Virginia.
USA.
1 = time of the year (in years)
and the model parameters determined by regression procedure are
xq (annual mean UMY in 1/bu), and A and B.
The sample data were fitted to the model by regressing x on cos2-n7
and sin2-n7. Although cos2tt/ and sin2trr both depend on t. they
are linearly independent of each other and therefore can be used
425
426
Loesch and Evans
as independent variables in a multiple linear regression pro-
cedure.
The model is alternatively expressed
x = xq + r cos2ttU - fn)
where ,v0 is the mean UMY. r is the amplitude of the sinusoidal
variation, and tn is time when the maximal UMY occurs; r and /0
are related to A and B as follows:
r = (A2 + B2)"2
Meat yield (UBu)
and
f0 = (l/2ir) tan \BIA). [see footnote 2]
Two-Component Sinusoidal Model
A feature of the basic sinusoidal model is that the rise and fall
on either side of the maximum (or minimum) are symmetrical.
This could be regarded as an unrealistic constraint to put upon the
model. The problem is addressed by including additional terms to
account for the additional feature. The appropriate extension of the
sinusoidal model is to include two additional terms that constitute
an additional sinusoidal component with a period of 6 months,
i.e., one-half of the period of the basic sinusoid:
x = x0 + Al cos2it? + fi, sin2iTf + A2 cos4tt7 + B2 sin4-n7
The function is still linear in the parameters, and the fit can again
be performed using a standard regression procedure. As with the
one-component model, an alternative expression is:
x = x0 + r, cos2ttU - /J,") + r2 cos4ir(Z - t',2'),
where r2 is the amplitude of the second component. The interpre-
tation of t(0" and 4"' in terms of time of maximum is. however,
now more complex.
Alternative Quadratic (Parabolic) Models
For method comparison purposes, in addition to fitting the data
to a sinusoid, we consider the quadratic function:
x = an + o,t + a-,r
2There are two angles in the range 0-2tt radiants whose tangent is BIA. The
appropriate one lies in the quadrant where its cosine has the same sign as
A and its sine has the same sign as B
Meal yield (UBu)
40
JFMAMJJASOND
1974
Figure 1. Observed clam meat yield data and sinusoidal fit for IM74.
50
45
Figure 2. Observed clam meat yield data and sinusoidal fit for 1975.
as an alternative model. This function describes a parabola, con-
taining a single maximum (when a2 < 0) or minimum (when a2 >
0). The position of the maximum (or minimum) is given in terms
of the model parameters by the following expressions:
"i "I
'max — -, • -*max — ^0 ,
2«: 4«2
In order to treat the feature of asymmetry, a term in i3 can be added
to the quadratic model to give a cubic model:
a No of observations
40 -
nl
D
00 1 0
residual (UBu) (1974)
O-
No of observations
n
a
"-2 0 -10 00 10 20
residual (UBu) (1975)
Figure 3. Distribution of residuals from the fit to data from (a) 1974
and (b) 1975.
Si asonai Variation in Somatic I ism i
427
TABLE 1.
Results from fitting a sinusoidal model x = x„ + A cos2irf + B sinl-nl to the clam meat yield data. The model is alternatively expressed as x
= x„ + r cns2n({ - („).
Year/Cut
N
*o
.4
B
r
= 2tttu. Figure 4 shows the two points corresponding to
the two years. The cross arms represent the standard errors in the
estimation of A and B. They are all approximately equal to 0.053.
As is discussed later, these errors are uncorrelated so that the
standard error in r for each year is also approximately 0.053. The
difference between the two amplitudes is 0.007; the standard error
in this quantity is approximately \fl x 0.053 = 0.075. This gives
a r-statistic of 0.09. There is, therefore, no evidence of a differ-
ence in amplitude between the years.
The treatment of the phase angles is different. Because the
error in r is much smaller than r itself, one may say that, approx-
imately:
8 0. i.e.. a concave-up
parabola with a single minimum. On the other hand, fitting the
sinusoidal function to the July 1974 to June 1975 data gives values
for the model parameters r and t„ (Table 4) that are very similar to
those from the two fits for the data from January through Decem-
ber in 1974 and 1975. These two fitted functions and the data are
shown in Figure 1 1 .
DISCUSSION
It is seen that the sinusoidal model is superior to the quadratic:
the parameters are interpretable in terms of meaningful quantities
such as annual mean value, the amplitude of the annual variation,
and the phasing of the sinusoid, which relates to the time of oc-
currence of the maximum and minimum. The values of the pa-
rameters are relatively insensitive to where the year's data begin,
whereas the quadratic fits give totally different descriptions. The
values of the sinusoidal parameters for successive years can be
compared in a meaningful way by considering the changes in the
overall mean, the amplitude, and the phasing.
A further advantage of the sinusoidal characterization of the
data is that the year-to-year comparison of amplitude and phase
can be made quite simply with a graphical presentation. The pa-
rameters in the quadratic fit do not lend themselves to a similar
simple geometric interpretation. Calculation of the errors in func-
tions of the parameters (such as differences) is more complicated
because the errors of determination in the parameters are corre-
Seasonal Variation in Somatic Tissue
431
Meat yield (L'Bu)
M A
1975
Figure 11. Quadratic and sinusoid fits for the July 1974 to June 1975
sequence.
The heavy curve is the sinusoid.
lated. It is a feature of the sinusoidal functions that, if the obser-
vations are evenly spaced over a complete period (or multiple of
periods), the determinations of the parameters are perfectly uncor-
rected. For a relatively large number of points that are distributed
approximately uniformly over a complete period, as is the case
here, the correlation is negligible. The addition to the model of a
second sinusoidal component with a 6-month period allows the
characterization of an asymmetric rise and fall. The coefficients of
the basic sinusoid are little affected by the addition of these extra
terms to the model. This is because the determinations of the
parameters are uncorrected with each other as noted above.
The quadratic fit has no intrinsic merit and is merely an arbi-
trary parametrization of the observations. The model is suggested
because the data for a complete year from January to December
show the presence of an apparent maximum and the quadratic
function can reproduce this feature. However, 12 months of data
from July through June exhibit the opposite appearance, with the
presence of a minimum, which gives rise to a totally different fit.
LITERATURE CITED
Castagna, M., N. Lewis. C. M. McFadden & M. Gibbons. 1992. Index of
papers published in the Journal of Shellfish Research. J . Shellfish Res.
11:561-581.
Crosley. M. P. & L. D. Gale. 1990. A review and evaluation of bivalve
index methodologies with a suggested standard method. J Shellfish
Res. 9:233-237.
Lawrence, D. R. & G. I. Scott. 1982. The determination and use of con-
dition index of oysters. Estuaries 5:23-27.
Loesch. J. G. 1977. Usable meat yields in the Virginia surf clam fishery.
Fish Bull. 75:640-642.
Mann, R. E.. M. Lynch, M. Castagna, B. M. Campos, & N. Lewis. 1993.
Index of papers published in the Proceedings of the National Shellfish-
enes Association. J. Shellfish Res. 12:158-182.
Journal of Shellfish Research, Vol. 13, No. 2.433-441, 1994.
AGE, GROWTH RATE, AND SIZE OF THE SOUTHERN SURFCLAM, SPISULA SOLIDISSIMA
SIMILIS (SAY, 1822)
RANDAL L. WALKER1 AND PETER B. HEFFERNAN1 2
'Shellfish Research Laboratory
Marine Extension Service
University of Georgia
20 Ocean Science Circle
Savannah, Georgia 3141 1-101 1
'Marine Institute
80 Harcourt Street
Dublin 2, Ireland
ABSTRACT The age, growth rale, and size of the southern surfclam, Spisula solidissima similis, were determined by shell-
sectioning techniques for clams collected from beach drift off of Wassaw Island, Georgia (Atlantic coast), and Cape San Bias, St
Joseph Bay, Florida (Gulf of Mexico coast). The shell-sectioning results for the Georgia population were validated by analysis of
monthly size-frequency data for a field population collected from St. Catherines Sound, Georgia. The southern surfclam deposited a
single age band within the shell matrix during the summer months at both sites. A distinct alternating pattern of translucent to opaque
to translucent zones in the outer shell was evident for clams from both sites. The translucent zone is formed from May to October,
whereas the opaque zone is formed from November to April. The annual band occurs within the translucent zone. According to the
von Bertalanffy growth regressions, maximum size estimates of 76 and 122 mm for Georgia and Florida surfclam populations,
respectively, are predicted. In Georgia, surfclams obtained a maximum shell length of 74 mm and were aged to a maximum of 4 years,
compared with 106 mm in shell length and 5.5 years for clams from Florida. In Georgia, the majority of surfclams (92%) collected
alive or from beach drift lived to a mean age of 1.5 years. Clams from Florida tended to survive to a mean age of 3.5 years. Clam
cohorts collected from St. Catherines Sound grew to a mean shell length of 48 mm in 1990 and 47 mm in 1991 in 1.5 years before
dying. Southern surfclams from Georgia were found to differ in age. growth rate, and size from a population from the Gulf coast of
Florida, whereas both southern groups contrasted greatly with the Atlantic surfclam. Spisula solidissima, which has been shown to
grow to 226 mm and has a lifespan of 37 years.
KEY WORDS: Spisula solidissima similis, surfclam, growth, size, longevity
INTRODUCTION
Because of the commercial importance of the Atlantic surf-
clam, Spisula solidissima (Dillwyn, 1817), much is known about
its life history. Although the clam ranges from Nova Scotia to
South Carolina (Abbott 1974). the fishery is centered between
Virginia and Massachusetts. The clam obtains an average shell
length of 100 to 130 mm (Abbott 1974), but has been reported to
reach 226 mm in shell length (Ropes and Ward 1977) and lives to
37 years of age (Sephton and Bryan 1990). See reviews of the life
history (Fay et al. 1983, Ropes 1980) of the Atlantic surfclam and
its fishery (Ropes 1980 and 1982; Ropes et al. 1969, Yancey and
Welch 1968) for further information.
Unlike the Atlantic surfclam, little other than taxonomic range
and size is known about the life history of the southern surfclam.
Spisula solidissima similis (Say 1822). This subspecies ranges
from Cape Cod, Massachusetts, to both coasts of Florida and to
Texas (Abbott 1974) and grows to 127 mm (Andrews 1977). Be-
cause of the dearth of life history information pertaining to this
subspecies, we investigated its aquacultural potential by studying
its relative growth rate, size, and longevity patterns from the
coastal waters of Georgia and the Gulf coast of Florida.
METHODOLOGY
Articulated shells from beach drift were collected from two
sites (Fig. 1): the beaches at the northeastern end of Wassaw
Island, Georgia (Atlantic coast), and Cape San Bias, St. Joseph
Bay, Florida (Gulf of Mexico coast). Articulated shells only were
collected by hand from beach drift at low tide. Each shell was
numbered and labeled as to collection site and date. Shells were
collected from Wassaw Island on March 25, April 6, July 23,
September 21 and 23, October 25, November 2 and 27, December
28, 1988, and January 6, 1989. Because of the distance factor.
Cape San Bias, Florida, shells were collected only on March 9,
June 15, and December 6, 1988.
Clams were returned to the laboratory where they were mea-
sured for shell length (i.e.. anterior-posterior measurement), and
aged by shell-sectioning techniques (see Rhoads and Lutz 1980,
Rhoads and Pannella 1970). Growth curves for each clam were
constructed by measuring each summer (translucent) and winter
ring increment (opaque zone) per clam (Fig. 2). Population growth
curves were constructed by averaging the shell length measure per
ring per clam for each population. Mean shell length data per ring
were fitted to a von Bertalanffy growth equation for each popula-
tion using an IBM PC computer with the Fishery Science Appli-
cations System developed by the University of Rhode Island (Saila
et al. 1988). The resulting initial parameters of the von Bertalanffy
growth model generated from mean shell length data were then
used as starting parameters for a SAS program (SAS Institute Inc. ,
1989) designed to determine the parameters of the growth model
using all of the growth data.
To validate the results of the shell-sectioning technique for the
Georgia sample, monthly samples of live clams were collected via
trawling with a 12.2 m otter-trawl net or a 1 m planning dredge
pulled by the 10.6 m R/V SEA DAWG in St. Catherines Sound,
Georgia. Clams occurred in sand ridges at the mouth of the Sound
in approximately 8 m depth. An otter trawl and planning dredge
were used as collection devices, because they would knock off the
tops of the ridges, momentarily suspending the surfclams within
the water column, while the net or bag was being pulled through
433
434
Walker and Heffernan
Wassaw Island
Si. Catherines
Sound
-30°
32°
28°
-26°
84° 82°
Figure 1. The sampling sites for S. s. similis shells collected from
beach drift from Wassaw Island, Georgia, and Cape San Bias, St.
Joseph Bay, Florida, and live animals collected from St. Catherines
Sound, Georgia.
Cape San Bias, Florida
March
June
n=97
n = 38
(1 = 142
I
Wassaw Island, Georgia
March/April
July
September OctJJov
November
n = 49
n = 41
n = 24 n = 13
n = 21
St. Catherines Sound, Georgia
January February March April May
n z 60 n = 60 n = 60 n = 50 n = 30
Hill
St. Catherines Sound, Georgia
August September October November December
I I I
Terminal Shell Band Phase
3 Translucent ^B Opaque
Figure 3. The phase (translucent or opaque) of the terminal ring with
time of samples of S. s. similis collected from Wassaw Island, Georgia,
Cape San Bias, St. Joseph Bay, Florida, and St. Catherines Sound,
Georgia. Note that a lag time may occur between time of dominance
between one phase and other for live animals and those shells from
beach assemblages. See text for Discussion.
the impact area. Other sampling devices (i.e., benthic dredges, box bands are laid down within the translucent zone. The first year's
corers, oyster dredges. Van Veen grab samplers) proved unsuccessful annual band is generally absent from within the shell matrix. Fur-
at sampling for surfclams. Trawling continued until at least 30 ther year bands are quite distinct within the shell matrix,
clams were obtained. Clams were measured for shell length. A single annual band is laid down during the summer months
RESULTS
It became apparent during the course of this study that the
southern surfclam can be aged, even without sectioning, by hold-
ing the shell up to a strong light source. A distinct alternating
pattern of translucent, then opaque, then translucent zones is ob-
served in the outer shell (Fig. 2). Each of these phases represent
0.5 years' growth (Fig. 3). Within the shell matrix, the annual
First Translucent Zone
First Opaque Zone
Second Translucent Zone
Second Opaque Zone
-Third Translucent Zone
0 12 4
Years
Figure 2. A graphic illustration of the translucent and opaque zona- Figure 4. Growth curves of S. s. similis from Georgia and Florida
tion observed in the shells of S. s. similis. populations.
LU
S3 100^
+i
E
E 80
_c
•^"^^T JB
H 60-
03
_l
Jj 40-
C/5
-B- Wassaw Island, GA
« 20-
-•- Cape San Bias. FL
0-
' 1 '
i ' i
Life Cycle of Spisula solidissima similis
435
0 60-
'0.40-
0 20-
■000-
= 0 60^
» 0 JO —
n = 32
n = 96
Cape San Bias. St Joseph Bay. Flordia
n = 131 n = 93 n = 19
Wassaw Island, Georgia
n = 56 n = 3 n = 1
n = 2
n = 0
0.20-^
0 00 J
0.5 1
1.5 2 2.5 3 35 4
Terminal Shell Band Phase
4 5 5 years
i Translucent ^H Opaque
Figure 5. The percentage of S. s. similis that died in each 0.5 year
increment.
for the southern surfclam. March and April surfclam samples from
Wassaw Island showed 90% of the clams were in the opaque phase
at the terminarnng (Fig. 3). By July. 95% of clam terminal rings
were in the translucent phase. With fall, the dominance shifts back
to the opaque phase. November (76%). December, and January
(both 100%) samples showed the majority of terminal rings were
in the opaque phase (Fig. 3). The same pattern was observed for
clams from Florida (Fig. 3). For live clams collected from St.
Catherines Sound, Georgia, the terminal ring is in the opaque
phase January through April, before changing from a dominance
of opaque phases in May to a dominance of translucent phases in
June. Dominance of shells in the opaque phase occurs again by
October (Fig. 3).
The growth curves of the two surfclam populations are given in
Figure 4. Up to age 2. the surfclams from Georgia show growth
similar to those from Florida. By age 2. most Georgia clams have
died (see Fig. 5), and among those surviving, growth rate has
slowed; however, growth continues in the Florida population.
Shell length data per ring (translucent and opaque per year) were
fitted to a von Bertalanffy equation, and the results are given in
Table 1.
Figure 5 shows the stage of the terminal ring at time of death
for both populations of surfclams. Maximum age for the Georgia
surfclam was 4 years, with the majority (92%) dying before 2
years of age. In the Florida surfclam population, maximum age
was 5.5 years with only 19.5% of the clams dying before 2 years.
Only 3% of the Georgia clams lived longer than 3 years compared
with 41% in the Florida population.
The shell length size-frequency data for surfclams from the two
populations are given in Figure 6. Surfclams from Florida ranged
in shell length from 21.1 to 105.5 mm. with a mean of 60.0 ±
1.21 (standard error |SE]) mm. Clams from Georgia ranged in size
from 24.4 to 73.9 mm. with a mean of 43.9 ± 0.79 (SE) mm. The
von Bertalanffy growth equation yielded maximum size estimates
(Lmax) of 76 and 122 mm for the Georgia and Florida clam pop-
ulations, respectively (Table 1 ). Surfclams obtained a greater size
in the Florida population because of their longer life span.
The monthly size-frequency data for the live clams collected
from St. Catherines Sound are given in Figure 7. Beginning in
January 1990, a single cohort at a mean shell length size of 33.5
mm occurred. This cohort grew to a mean size of 42.3 mm by
April, when the occurrence of the new 0+ cohort emerged at a
mean size of 20.2 mm. By September, the 0+ cohort had a mean
size of 26.7 mm and the remaining 1 + cohort had a mean size of
47.9 mm. By September, the majority of the 1+ cohort, then
approximately 1.5 years old, had died. No 1 + cohort clams were
collected after September. The 0+ cohort grew little from August
8 (x = 27.3 mm) to December 5 (,v = 30.7 mm), after which,
rapid growth occurred from December U = 30.7 mm) to February
(.v = 41 mm). From February to April (,v = 43.5 mm), little
growth occurred. At that time, clams were approximately 1 year
old. The mean shell length of the cohort increased, reaching a
maximum of 46.6 mm by July before dying off. By June, the new
0+ cohort emerged and obtained a mean shell length of 16.3 mm.
By July, they reached 22.3 mm.
For the cohort that was monitored from emergence in April
1990 to death in July 1991 , growth rate was 3.6 mm/month for the
first year and 1 .0 mm/month for the next 3 months, assuming that
clams were spawned in March to April of 1990 (Kanti et al. 1993).
The growth of the cohort over its lifespan (1.25 years) is 3.1
mm/month.
DISCUSSION
The southern surfclam occurs in a unique habitat within the
sounds of coastal Georgia. Clams inhabit the sand ridges in areas
at the mouth of sounds where a major creek or river, draining from
the marsh system of a barrier island, enters into the sound. Be-
cause of the relatively high current speed present in these areas,
the bottom tends to be sandy, forming a series of ridges along the
bottom. Clams generally occur at the top of these ridges. Further-
more, with help from the Florida Department of Natural Re-
sources, attempts at acquiring reproductively active surfclams
from the west coast of Florida revealed that the only clams located
(<10 mm in shell length) by scuba occurred in the top of sand
ridges in Tampa Bay, at a site previously described with abundant
clams (Godcharles 1971).
TABLE 1.
The results of a von Bertalanffy growth function for the Georgia and Florida 5. s. similis populations. The von Bertalanffy growth equation
is L, = L„
(I
").
Population
K ± SE
t„ ± SE
Death
assemblage
For t = 0.5
year
increments
Georgia
Florida
75.77 ± 12.12
121.50 ± 11.85
0.74 ± 0.26
0.46 ± 0.08
-0.04 ± 0.10
0.156 ± 0.04
0.9719
0.9563
o
03
03
09
09
cc
Cape San Bias, Florida
n = 277
Range 21 .1 to 1 05.5 mm Mean 60.0 ± 1 .21 (SE) mm
Wassaw Island, Georgia
n = 159
Range 24.4 to 73.9 mm Mean 43.9 ± 0.79(SE) mm
n i i i i i i
St. Catherines Sound, Georgia
n = 915
Range 9.9 to 53.8 mm Mean 33.2 ± 0.30 (SE) mm
n — i — i — i — r
22.5 32.5 42.5 52.5 62.5 72.5
Class Mid-point Size (mm)
~ i i i i i
82.5 92.5 102.5
Figure 6. The size distribution of S. s. similis collected from Wassaw Island, Georgia, Cape San Bias, Florida, and St. Catherines Sound,
Georgia.
Life Cycle of Spisula solidissima similis
437
The southern surfclam is easily aged, and shell-sectioning tech-
niques are not necessary to achieve good estimates of age and size
tor this species from either Florida or Georgia coastal waters. The
alternating pattern of translucent to opaque to translucent zones
within the shell precludes the necessity of using expensive and
time-consuming shell-sectioning techniques to determine age.
However, for exact size measurements, the shell-sectioning tech-
nique may be required, because growth slows and the translucent
zone becomes less obvious with increased age (>4 years). In the
case of the Georgia clams from the sounds, shell sectioning is not
required, because the first visible annual band is not laid down in
the shell matrix until year 2, by which time the cohort has died out.
This method of aging bivalves according to the alternating patterns
of transparency in the shell has been used for Spisula sachalinensis
in Japan and Korea ( Sasaki 1981, Kang and Kim 1983, respectively),
for Meretrix meretrix in China (Zhang and Fuxue 1988), and for
the mussel Mxtilus galloprovincialis in Japan (Hosomi 1983).
Although this method is reliable for determining the age and
size of the southern surfclam, caution is needed when determining
the exact timing of the transitional phases between translucent to
opaque zones for specimens collected from death assemblages.
For shells collected from the beach drift at Wassaw Island in
October 1989. the majority (92%) were still in the translucent
phases. By contrast, live specimens collected in October 1989
were all in the opaque phase. It is to be expected that a lag time
occurred between the death of the clam and the time that the shell
is washed ashore.
Southern surfclams are smaller and shorter lived than their
northern counterpart. The results of both the shell-sectioning anal-
ysis of beach-death assemblages and size-frequency data for live
cohorts for the Georgia surfclam population show that clams gen-
erally live less than 2 years. No live clams older than 1.5 years
were collected, although a few clams older than 2 years of age of
unknown origin were found in the beach drift samples from Geor-
gia (Fig. 5).
For the surfclam population within the sounds of coastal Geor-
gia, the 1 + year cohort dies during the August to September
period. By August, ambient water temperatures may reach as high
as 31°C (Winker et al. 1985). Sustained water temperatures of
30°C occur from June well into September in most years. Georgia
also possesses roughly 30% of the total area of salt marshes along
the Atlantic coast, so coastal waters experience heavy organic
loading (Odum and de la Cruz 1967). The heavy organic loading,
high ambient water temperatures, and consequent low dissolved
oxygen levels (Winker et al. 1985) that occur during summer may
be responsible for the dieoff of 1 + year class. Physiological test-
ing on the southern surfclam is required to test this hypothesis.
The Atlantic surfclam requires an environment with ample dis-
solved oxygen. Thurberg and Goodlet (1979) found that oxygen
levels of 1.4 mg 1~ ' at 10°C were lethal to 50% of adult clams
exposed for up to 10 weeks. Major surfclam mortalities were
recorded off of New Jersey in 1976, in which an estimated 62% of
the standing stocks of surfclams were killed (Ropes 1980). Ap-
proximately 259c of the ocean quahog, Arctica islandica, popula-
tion occurring within this area of anoxic conditions also died.
Anoxic water combined with hydrographic water conditions, or-
ganic loading of coastal waters, and a massive bloom and dieoff of
a dinoflagellate were responsible (Steimle and Sinderman 1978).
Low oxygen levels adversely affected surfclam burrowing by low-
ering the rate of activity (Savage 1976). Furthermore, water tem-
perature, which is inversely proportional to dissolved oxygen lev-
els, may play a role. Northern surfclams tolerate water tempera-
tures from 14 to 30°C (Loosanoff and Davis 1963), with juvenile
clams surviving higher temperatures better than adults (Saila and
Pratt 1973). Experimental field grow-out studies, using S. solidis-
sima stocks from Connecticut and Delaware cultured in coastal
Georgia, showed excellent growth and survival of clams from
October to May. but total mortality if clams are not harvested in
May (Goldberg and Walker 1990, Walker and Heffernan 1990).
Ambient water temperatures may reach 28 to 30°C (lethal levels)
by the end of May in Georgia.
From the monthly size-frequency data for the St. Catherines
Sound population (Fig. 7). an accurate record of growth for the
southern surfclam can be constructed. In January 1990, a single
cohort exists. This cohort grew little from January (.v = 33.5 mm)
to February (x = 33.7 mm) but grew rapidly from February to
May, followed by a slowing phase from May (x = 43.6 mm)
through August (x = 44.3 mm). By August, the 1 + cohort was
dying off and was completely gone by October 1991. This cyclic
form of growth was closely matched by its gametogenic cycle
(Kanti et al. 1993). In early winter, surfclams have entered the late
active stage of their maturation process (Kanti et al. 1993). Thus,
a large portion of their energy budget probably was devoted to
gametogenesis. During March and April, clams have reached their
final maturation phase and food is again available for growth. By
April, the clams have begun spawning and food presumably is
being devoted primarily to the build-up of somatic reserves with
shell growth, consequently, slowed.
The 0+ cohort was detected in samples during May (x = 20.9
mm) and grew rapidly until August ix = 27.3 mm), when growth
rate slowed down. Growth remained slow until November before
it accelerated again. This growth slowdown during summer pre-
sumably is the result of the physiological stress of high water
temperatures, low dissolved oxygen levels, and high organic load-
ing (Steimle and Sinderman 1978), which apparently kill the more
susceptible adults (Saila and Pratt 1973). The 0+ cohorts growth
slowed during winter, when sexual maturation probably required
the majority of the energy available.
Growth rates of Georgia and Florida surfclam populations were
similar during 0+ year groups. Overall, the Florida population
continues to grow at a relatively good rate during the second year
before the rate begins to decline. For the Georgia population, the
growth rate decreases significantly after the first year.
Jones et al. (1978), studying an inshore-and-offshore popula-
tion of surfclams off of New Jersey , found a similar growth pattern
between the two sites for the first 3 years, after which, the growth
in the inshore site was markedly reduced. In addition, clams from
the inshore site obtained a lower maximum size, grew slower, and
had a shorter lifespan (Chang et al. 1976, Jones et al. 1978). No
explanation was presented to explain the differences observed.
Cerrato and Keith (1992) also found that a surfclam population
from an estuary in New York was dominated by two age classes
and had an apparent life span of 10 years, whereas, in an inshore
population, clams ranged in age from 2 to 22 years with no age
class dominance. Adults from the estuarine environment grew
slower and obtained a smaller maximum size than did clams from
the inshore population. They hypothesized that physiological
stresses due to reduced salinity and more extreme temperatures
found under estuarine conditions were probably responsible for the
observed differences in population parameters.
A substantial body of literature shows that, with decreases in
latitude, greater growth rates, earlier age at maturity, and shorter
Count
January 1990 February 1990 March 1990 April 1990 May 1990
0 2 4 6 8101214 16 1820 0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12
— December 1990
January 1991
February 1991
March 1991
April 1991
09 02 4 6 8 10 12 02468 10 12 14 0246 8 10 12 14 16 18 20 0 2 4 6 8 101214 16 18 0 2 4 6 8 10 12 14161
10-
20-
■ ' ■
N = 30
x = 36-8 1 8.3 mm
30-
40-
50-
May 1991 July 1991
024 68 10 12 0246
60-1-
30
40
50"
N = 25
x = 163 ±08mm
N = 6
x = 46.7±2 3 mm
20
f
10 12
N = 8
x = 22.3 t 20mm
30 -
40 I
50 ■
60 "
N = 6
x = 467i 23mm
Figure 7. The size-frequency data showing the growth of two cohorts of S. s. similis over time from a population in St. Catherines Sound.
Georgia.
Life Cycle of Spisula solidissima similis
439
TABLE 2.
Annual growth increments of the northern surfclams versus the southern surfclams for the first 3 years of life from natural populations
along the Atlantic coast of the United States. (Based on Table 4 of Ropes 1980).
Shell Length at:
1 year
2 years
3 years
Location
(in mm)
Source
S. solidissima
P.E.I. Canada"
15.9
34.9
53.2
Kerswill 1944
Canada
15.0
31.0
49.0
Caddy and Billard 1976
Cardigan Bay
12.0
46.2
72.8
Sephton and Bryan 1990
Northumberland Strait
32.2
57.9
77.6
Sephton and Bryan 1990
Gulf of St. Lawrence
23.5
46.0
64.2
Sephton and Bryan 1990
Monomoy Point. MA
25.4
40.6
81.3
Belding 1910
Long Island, NY
34.0
60.0
79.0
Ropes 1980
Pt. Pleasant. NJ
Jones et al. 1978
Inshore
38.0
62.0
79.0
Offshore'
29.0
55.0
81.0
Barnegat Bay, NJ
34.0
56.0
73.0
Change! al. 1976
Offshore, NJ
47.2
79.2
100.8
Loesch and Ropes 1977
Central, NJ
45.7
78.7
99.1
Ropes 1980
Ocean City, MD
39.0
57.0
94.0
Chang etal. 1976
Chincoteague Bay, VA
42.2
68.6
90.5
Ropes et al. 1969
S. s. similis
Wassaw Island. GA
44.3
57.7
70.8
This study
St. Joseph Bay, FL"
37.1
78.8
92.3
This study
St. Catherines Sound. GA
43.5
This study
P.E.I.. Prince Edward Island.
1 Gulf of Mexico coast.
lifespans occur for intraspecific comparisons of marine bivalve
species (Newell 1964). The growth rate of the southern surfclam is
comparable to that of the Atlantic surfclam in its southern distri-
butional range (Table 2). An increase in growth rate for the At-
lantic surfclam with a decrease in latitude does appear evident
(Table 2 and Ropes 1980). On the basis of the mean size achieved
at 1 year, the southern surfclam achieves size comparable to that
of the northern species at its lower distributional limit for both the
Georgia and the Florida populations. At 2 years, only the Florida
population is comparable, because the majority of the Georgia
clams have died.
In terms of achieving sexual maturity at an earlier age, the
southern surfclam in Georgia behaves more like a semelparous
species rather than an iteroparous species, like the northern spe-
cies. The 0+ cohort begins sexual development within its first
year and spawns at approximately age 1 at a size of 40 mm (Kanti
et al. 1993). After spawning, clams survived for another 5 to 6
months before the cohort ( 1 + year) died off. The gametogenic
cycle of the Florida population was not studied, but it is likely that
it would be similar in terms of age and size at sexual maturity to
that found in the Georgia population. For the Atlantic surfclam,
Sephton and Bryan (1990) stated that sexual maturity occurred at
80 mm in shell length at an age of 4 years for a population of clams
from Prince Edward Island, Canada, whereas. Ropes et al. ( 1969)
observed that sexual maturity occurred at 45 mm and 1 year of age
for clams from an inshore population in Chincoteague Inlet. Vir-
ginia. Belding (1910) in Massachusetts found that sexual maturity
could occur in 1 year olds at a size of 39 mm, but that the majority
of clams matured at 2 years of age and at 67 mm in shell length.
The Atlantic surfclam in offshore populations commonly at-
tains ages of 25 and occasionally 30 years (Jones et al. 1978).
Southern surfclams in Georgia from inshore populations rarely
survive 2 years. By contrast. Florida clams reach a mean age of
3.5 years, with some individuals surviving to 5 years. Thus, there
is a definite decrease in lifespan with a decrease in latitude when
comparing the southern and the Atlantic surfclams.
In Georgia, the southern surfclam resembles the life history
traits of Spisula subtruncata and Notospisula trigonella. It is in-
teresting to note that N. trigonella, S. subtruncata, and the Geor-
gia surfclam are all basically annual species and occur within the
estuary, whereas, commercial Spisula species (S. polynyma, S.
sachalinensis, and S. solidissima) are long-lived species that occur
primarily offshore. For the three estuarine species, a single re-
cruitment period occurs: N. trigonella recruits from August to
November in Australia (Jones et al. 1988), 5. s. similis recruits
from April to June in Georgia (Kanti et al. 1993), and S. subtrun-
cata recruits over the summer in Italy (Ambrogi and Ambrogi
1987). These estuarine species also achieve the smallest maximum
size of the Spisula species; N. trigonella reaches 21 mm (Jones et
al. 1988), S. s. similis in inshore areas of Georgia reaches 57 mm
(this study), and S. subtruncata grows to 22 mm (Ambrogi and
Ambrogi 1987). The stressful environmental parameters common
to estuarine conditions may well regulate the lifespan of these
Spisula species.
The feasibility of establishing a fishery for this animal is poor,
but some potential may exist at least for the development of an
aquacultural product in Georgia. The southern surfclam occurs in
a rather unique habitat in Georgia. These habitats are generally
small in overall area and occur in only half of the sounds of coastal
Georgia. Although no density estimates have been determined for
surfclams in these habitats (because of the inefficiency of most
bottom-sampling devices within these sand ridges), it is our opin-
440
Walker and Heffernan
ion that overall densities are low and insufficient to support even
a small natural fishery. The distribution pattern of surfclams in
offshore areas is unknown, but past natural resource surveys by the
R/V SILVER BAY during 1959 and 1960 failed to report any
findings of surfclam beds in offshore areas of Georgia (Cummins
1966); however, some surfclams were found at one site off of
North Carolina (Cummins et al. 1962). The sampling devices used
in these surveys were geared toward harvesting larger sized clams
(R. Cummins, Jr., 1992, personal communication), so small-sized
southern surfclams may have been missed during those surveys.
There is some potential for the development of an aquaculture
product for the southern surfclam. Clams attained a mean size of
40 to 50 mm in shell length, which is about the ideal size for a raw
or steamer product. In previous studies evaluating the aquaculture
potential of growing the Atlantic surfclam over the winter in Geor-
gia, we had targeted 50 mm as an ideal size for a steamer product
(Goldberg and Walker 1990). In those studies, 10 to 15 mm seed
planted in October grew to mean sizes of 45 to 60 mm by May,
depending on location, initial planting size, tidal-planting height,
cage-mesh size, and substrate type (Goldberg and Walker 1990,
Walker and Heffernan 1990). To develop southern surfclam aqua-
culture. hatchery and nursery-rearing protocols will need to be
developed. Increasing the growth rate of the southern surfclam by
selective breeding programs or by triploidy induction may yield
larger clams for production purposes than are currently found in
natural inshore populations in coastal Georgia. These avenues of
research are being investigated.
ACKNOWLEDGMENTS
We thank Ms. F. Hodges and Mr. D. Head for their help in
sectioning and preparing the shells for analysis. Mr. G. Paulk is
thanked for his aid in collecting shells. Special thanks to Mrs. D.
Thompson for typing the manuscript, Ms. R. Rivers for editing the
manuscript, and Ms. A. Boyette and Ms. S. Macintosh for the
graphics herein. We also thank Capt. J. Whitted of the R/V SEA
DAWG. We also thank Mr. B. Arnold, Dr. D. Marelli, and the
crew of the R/V ALLMOND of the Florida Department of Natural
Resources for their effort in attempting to locate live clams from
the Tampa Bay area. Mr. R. Goldberg of the National Marine
Fisheries Service is thanked for stimulating our interest in the
southern surfclam. The work was supported by the Georgia Sea
Grant Program under grant number NA84AA-D-00072.
REFERENCES
Abbott, R. T. 1974. American Seashells. 2nd ed. Van Nostrand Reinhold,
New York, 663 pp.
Ambrogi, R. & A. D. Ambrogi. 1987. Temporal variations of the sec-
ondary production in the marine bivalve Spisula sublruncala off the Po
River Delta (Italy). Est. Coast. Shelf Sci. 25:369-379.
Andrews. J. 1977. Shells and Shores of Texas. University of Texas Press.
Austin, 365 pp.
Belding, D. L. 1910. The growth and habits of the sea clam {Mactra
solidissima). Rep. Comm. Fish. Game Mass. 1909 Publ. Doc. 25:26-
41.
Caddy, J. F. & A. R. Billard. 1976. A first estimate of production from
unexploited population of the bar clam, Spisula solidissima. Can. Fish.
Mar. Sen\ Tech. Rept. 648, 13 p.
Cerrato, R. M. & D. L. Keith. 1992. Age structure, growth, and mor-
phometric variations in the Atlantic surf clam, Spisula solidissima,
from estuarine and inshore waters. Mar. Biol. 114:581-593.
Chang, S., J. W. Ropes & A. S. Merrill. 1976. An elevation of the surf
clam population in the Middle Atlantic Bight. Int. Counc. Explor. Sea.
Special meeting on population assessments of shellfish stocks. Paper
No. 33, 43 p. U.S. Natl. Mar. Fish. Serv., Northeast Fisheries Center,
Sandy Hook, N.J. Lab. Ref. No. 76-1.
Cummins. R.. Jr. 1966. Hard-clam explorations off southeastern United
States. Comm. Fish. Rev. 28:33-42.
Cummins. R., Jr., J B. Rivers & P J. Struhsaker. 1962 Exploratory
fishing off the coast of North Carolina. September 1959— July I960.
Comm. Fish. Rev. 24:1-9.
Fay. C. W., R. J. Neves & G. B. Pardue. 1983. Species profiles: life
histories and environmental requirements of coastal fishes and inver-
tebrates (mid- Atlantic) . . . surfclams. U.S. Fish and Wildlife Service
82/11, U.S. Army Corps of Engineers, TR EL-82-4.
Godcharles, M. F. 1971 . A study of the effects of a commercial hydraulic
dredge on benthic communities in estuarine areas. Fla. Dept. Nat.
Resources. Mar. Research Lab., St. Petersburg, FL. Tech. Ser. 64,
51 p.
Goldberg, R. & R. L. Walker. 1990. Cage culture of yearling surf clam,
Spisula solidissima in coastal Georgia. J. Shellfish Res. 9:187-193.
Hosomi, A. 1983. Ecological studies on the mussel Mytilus galloprovin-
cialis (Lamarck) II A simple age estimation method by light penetra-
tion. Venus 42:269-288.
Jones, D. S., I. Thompson & W. Ambrose. 1978. Age and growth rate
determinations for the Atlantic surf clam, based on internal growth
lines in shell cross section. Mar. Biol. 47:63-70.
Jones. A. R . A. Murray & G. A. Skilleter. 1988. Aspects of the life
history and population biology of Noiospisula trigonella (Bivalvia:
Mactridae) from the Hawkesbury Estuary, southeastern Australia. The
Veliger 30:267-277.
Rang, Y. J. & C. K. Kim. 1983. Studies on the structure and production
processes of biotic communities in the coastal shallow waters of Korea.
III. Age and growth of Spisula sachalinensis from the eastern waters of
Korea. Bull. Korean Fish. Soc. 16:82-87.
Kanti, A., P. B. Heffernan & R. L. Walker. 1993. Gametogenic cycle of
the southern surf clam, Spisula solidissima similis (Say, 1822). from
St. Catherines Sound. Georgia. J. Shellfish Res. 12:255-261.
Kerswill, C. J. 1944. The growth rate of barclams. Fish. Res. Bd. Canada
Prog. Repl. All. Coast Sin. 35:18-20.
Loesch, J. G. & J. W. Ropes. 1977. Assessment of surf clam stocks in
nearshore waters along the Delmarva peninsula and in the fishery south
of Cape Henry. Proc. Natl. Shellfish. Assoc. 67:29-34.
Loosanoff, V. L. & H. C. Davis. 1963. Rearing of bivalve mollusks. pp.
130-136. In: F. S. Russell (ed.). Advances in Marine Biology, vol. 1.
Academic Press. New York.
Newell. N. D. 1964. Physiology aspects of the ecology on intertidal mol-
luscs, pp. 59-81. In: R. M. Wilbur and C. M. Yonge (eds.). Physi-
ology of Mollusca. vol. 1. Academic Press, London.
Odum, E. P. & A. A. de la Cruz. 1967. Particulate organic detritus in a
Georgia salt-marsh estuarine ecosystem, pp. 383-388. In: G. H. Lauff
(ed.). Estuaries, vol. 83. AAAS Publication, Washington.
Rhoads. D. C. & G. Pannella. 1970. The use of molluscan shell growth
patterns in ecology and paleoecology. Lethaia 3:143-161.
Rhoads. D. C. & R. A. Lutz. 1980. Skeletal Growth of Aquatic Organ-
isms: Biological Records of Environmental Change Plenum Press.
New York. 750 pp.
Ropes, J. W. 1980. Biological and fisheries data on the Atlantic surfclam,
Spisula solidissima (Dillwyn). Woods Hole Lab. Tech. Ser. Repl. No.
24. 91 p.
Ropes. J. W. 1982. The Atlantic Coast surf clam fishery, 1965-1974.
Mar. Fish. Rev. 44:1-14.
Ropes, J. W & G. E. Ward. Jr. 1977. The Atlantic coast surf clam
fishery— 1974. U.S. Natl. Mar. Fish. Ser. Mar. Fish. Rev. 39:18-23.
Ropes, J. W..J. L. Chamberlin& A. S.Merrill. 1969. Surf clam fishery.
Life Cycle of Spisula soudissima simius
441
In: F. E Firth (ed.). The Encyclopedia of Marine Resources. Van
Nostrand, Reinhold, Co., New York. pp. 1 14-125.
Saila, S. B. & S. D. Pratt. 1973. Mid-Atlantic Bight fisheries, pp. 6-1-
6-125. In: S. B. Saila (coordinator). Coastal and Offshore Environ-
mental Inventory, Cape Hatteras to Nantucket Shoals. Marine Exper-
imental Station. University of Rhode Island, Kingston. Marine Publi-
cation Series No. 2.
Saila. S. B , C. W. Recksiek & M. H. Prager. 1988. Basic Fishery Sci-
ence Programs: A Compendium of Microcomputer Programs and Man-
ual of Operations. Development in Aquaculture and Fisheries Science,
vol. 18. Elsevier, New York, 233 pp.
SAS Institute. Inc. 1989. SAS Users Guide: Statistics. Version 6. SAS
Institute Inc.. Cary. NC. 846 pp.
Sasaki. K. 1981. Growth of the Sakhalin surf clam. Spisula sachalenesis
(Schrenck) in Sendai Bay. TohukuJ. Agricull. Res. 32:168-180.
Savage, N. B. 1976. Burrowing activity in Mercenaria mercenaria and
Spisula solidissima as a function of temperature and dissolved oxygen.
Mar. Behav. Physiol. 3:221-234.
Sephton. T. W. & C. F. Bryan. 1990. Age and growth determinations for
the Atlantic surf clam, Spisula solidissima (Dillwyn. 1817). in Prince
Edward Island, Canada. J. Shellfish Res. 9:177-186.
Steimle. F. W. &C. J. Sinderman. 1978. Review of oxygen depletion and
associated mass mortalities of shellfish in the Mid-Atlantic Bight in
1976. U.S. Nail. Mar. Fish. Sent. Mar. Fish. Rev. 40:17-25.
Thurberg. F. P. & R. D. Goodlet. 1979. Low dissolved oxygen concen-
trations and surf clams: a laboratory study, pp. 277-280. In: R. L.
Swanson and C. J. Sinderman (eds.). Oxygen Depletion and Associ-
ated Benthic Mortalities in New York Bight. 1976. Paper No. 1 1, Part
2. U.S. Department of Commerce, NOAA Professional. Washington.
DC
Walker, R. L. & P. B. Heffernan. 1990. Intertidal growth and survival of
northern quahogs, Mercenaria mercenaria (Linnaeus, 1758) and At-
lantic surf clams, Spisula solidissima (Dillwyn. 1817), in Georgia. J.
World Aquacul. Soc. 21:307-313.
Winker. C. D , L. C. Jaffe & J. D. Howard. 1985. Georgia Estuarine
Data 1961-1977. University of Georgia. Georgia Marine Science Cen-
ter. Tech. Rept. Ser. No. 85-7, vol. I, Athens pp. 420.
Yancey. R M & W R Welch. 1968. The Atlantic coast surf clam with
a partial bibliography. U.S. Dept. of Interior, Fish and Wildl. Serv.,
Circular No. 288. pp 13.
Zhang. J. & L. Fuxue. 1988. Determination of age of the Asian hard clam.
Merelrix merelrix (Linnaeus). J. Fish. China 12:251-258.
Journal of Shellfish Research, Vol. 13, No. 2,443-449, 1994.
EXPOSURE OF THE SANDY-BEACH BIVALVE DONAX SERRA RODING TO A HEATED AND
CHLORINATED EFFLUENT 1. EFFECTS OF TEMPERATURE ON BURROWING
AND SURVIVAL
J. M. E. STENTON-DOZEY AND A. C. BROWN
Zoology Department.
University of Cape Town,
Rondebosch 7700,
South Africa
ABSTRACT The potential impact of the thermal plume from a nuclear power station on burrowing and survival of the bivalve Donax
serra. which inhabits a nearby sandy beach, was investigated. Median lethal time and median lethal temperature were used to define
the size-related upper thermal tolerance of D. serra. Small individuals would best tolerate heated effluent from the power station.
Temperatures above 32°C were lethal to all sizes of D. serra. After extended exposure to temperatures between 24 and 29°C, 50% of
the animals no longer remained buried. Because these temperatures occur in the thermal plume, such displacement from the sand can
result in exposure to predation as well as possible stranding on the beach.
KEY WORDS: Bivalve, burrowing, Donax serra. effluent, sandy beach, survival, temperature
INTRODUCTION
The burrowing bivalve Donax serra Roding inhabits the inter-
tidal zone of high-energy sandy beaches along the southern Afri-
can coastline, where it often comprises 80 to 90% of the total
macrofaunal biomass (McLachlan 1977a and b. Bally 1981 , Cook
and Birkett 1984). It is the largest member of its genus, attaining
a shell length of 80 mm and a dry tissue mass of 5 to 6 g. South
Africa's first nuclear power station was constructed in the imme-
diate vicinity of an expansive sandy beach densely populated by
D. serra. The power station uses sea water as a coolant, the heated
effluent being discharged directly into the surf zone at tempera-
tures approximately 10°C warmer than ambient. Inshore thermal
plumes are thereby created that have the potential to range in
temperature from 18 to 27°C. As part of a larger ecological impact
assessment (Cook and Birkett 1984 and 1986), a study of the
possible effects of the plume on the physiological fitness of D.
serra was undertaken.
The extensive research on the thermal tolerances of marine
organisms has focused strongly on the survival of fish larvae and
juveniles near power plants using sea water as a coolant (reviewed
by Schubel et al. 1978) and, from a different perspective, the ways
in which intertidal animals survive desiccation and increases in
temperature during aerial exposure (reviewed by Newell 1979).
The thermal tolerance of an organism is a product of an interaction
between exposure temperature and the duration of that exposure
and it is imperative to recognize both of these factors in any
experiments designed to determine upper lethal limits. Earlier data
were obtained by slowly heating the water in which animals were
contained at an arbitrary rate of 1°C in 5 to 10 minutes (Huntsman
and Sparks 1924. Henderson 1929. Broekhuysen 1940, Evans
1948, Southward 1958, McLachlan and Erasmus 1974). This
method resulted in the estimation of upper lethal limits higher than
conditions experienced in nature (Newell 1979).
The importance of the interaction between the duration of ex-
posure and temperature was recognized by researchers working on
fish (see Schubel et al. 1978) and the methods they established are
now widely used to measure temperature tolerances of marine
organisms in terms of both natural and human-induced changes. In
this approach, the time taken to reach 50% mortality is deter-
mined, or alternatively, the temperature at which 50% mortality
occurs can be used as the criterion of upper temperature tolerances
(Kennedy and Mihursky 1971). In this study, both median lethal
time and median lethal temperature are used to define survival
limits and burrowing responses in D. serra. This is the first in a
series of three papers, the subsequent two considering the effects
of chlorine on burrowing and survival and the effects of temper-
ature and chlorine on the heart rate of D. serra.
MATERIALS AND METHODS
Collection and Maintenance
Bivalves were collected from Ouskip beach 1.5 km south of the
power station, an area unaffected by the thermal plume. Animals
were kept in flowing sea water in 25 1 tanks fitted with air-lift
pumps to facilitate the circulation of water at 15°C through a bed
of sand and gravel and fed mixtures of natural detritus and the alga
Tetraselmis suecica. Experiments were designed to investigate the
effect of acute and stepwise changes in temperature on survival
(percent mortality) and burrowing.
Acute Temperature Exposure Experiments
These experiments simulated acute temperature changes as
may arise when the prevailing wind direction shifts to an onshore
northwesterly, resulting in the concentration of the thermal plume
and a rapid increase in temperature in the surf zone (Rattey and
Potgieter 1987). Experimental temperature changes used in the
laboratory exceeded natural swift variations in ambient tempera-
tures to which D. serra may be acclimatized.
The temperature at which 50% of test animals died after acute
exposure for 4 days was defined as the median lethal temperature
(LT50) (see Kennedy and Mihursky 1971). Median lethal time,
that is. the time to 50% mortality at a particular temperature (New-
ell 1979) was used to further define mortality in terms of thermal
resistance lines.
D. serra were grouped into three size classes on the basis of
shell width (in millimeters), measured as the greatest distance
between the dorsal and ventral shell extremities. Juveniles were
divided into two groups, namely <7 mm and between 7 and 35
443
444
Stenton-Dozey and Brown
mm, whereas breeding individuals were taken as >35 mm. This
grouping was chosen because it corresponds closely to the three
major growth cohorts in the population (Cook and Birkett 19861
and the size-related intertidal distribution of D. serra. The large
animals occur subtidally. whereas the smallest individuals are
most abundant at midtide levels.
Forty bivalves from each size group were placed in 20 1 tanks
at temperatures between 20 and 38°C at 2°C intervals for exposure
times of 1. 3. 6, 12, 24. 48, 72, and 96 hours. The tanks were
fitted with air-lift pumps that circulated sea water through a bed of
sand, thus ensuring an even temperature throughout. Percent mor-
tality was assessed together with the upper temperature limit at
which 50% of the bivalves retained an ability to burrow (BTS0)
(see Ansell et al. 1980a). Throughout the experiment, controls
were maintained at 15°C and any incidental deaths in this group
were used to correct data. During each observation, the numbers
of dead individuals, individuals that were buried, and the number
that were not buried but in a state of stress were noted. Stress was
evident from shell gaping and flaccid siphons and foot. The ab-
sence of reaction to mechanical stimulation in the foot and siphons
and the mantle edge and adductor muscles were taken as criteria
for death in the same manner as used by Ansell et al. (1980a) and
Ansell and McLachlan (1980). The absence of valve closure on
tactile stimulation of the cruciform muscle also proved a useful
criterion. If death was uncertain, possible recovery was monitored
by placing the individual in fresh, circulating sea water at 15°C for
2 days.
Four individuals from each exposure temperature were trans-
ferred daily to 15°C in order to monitor recovery. The criterion for
recovery was the ability of 50% of the test individuals to burrow
and ventilate after 4 days at 15°C.
Stepwise Temperature Exposure Experiments
These experiments were designed to establish whether stepwise
exposure to the temperatures above would improve survival rate.
Bivalves, approximately 120 within each size group, were pro-
gressively exposed to temperatures from 14 to 36°C at 2°C incre-
ments per day over 16 days. Sea water, at the appropriate exper-
imental temperature, was replaced every second day. Recovery
was determined by transferring an individual to 15°C after 24
hours of exposure to a test temperature after stepwise introduction
to that temperature. The criteria for recovery were the same as
used before.
In all experiments, the LT50 was determined graphically. Num-
bers dead, expressed as percentages of the original number, were
plotted for each temperature. Time interval and LT50 were deter-
mined from the resulting plots. The temperature at which 50% of
the bivalves originally present had burrowed, BT5„, was deter-
mined in a similar manner. The same method has been used by a
number of researchers (Bodoy and Masse 1977 and 1978, Ansell
et al. 1980a and b, Ansell and McLachlan 1980). Probit analysis
(Finney 1964) enables a statistical measure of 50% lethal limits,
and this was applied to some of the data. This method provided
similar or identical results, as found by Lent (1968), and because
it is a long and tedious procedure, the graphical method described
was used throughout.
Median lethal time was also determined graphically by plotting
percent mortality for each exposure temperature against exposure
time on a logarithmic scale. In turn, median lethal times were
plotted as a function of exposure temperature on a semilogarithmic
scale to obtain thermal resistance lines for each size group; these
lines define the zone of resistance of an organism (Newell 1979).
RESULTS
Acute Temperature Exposure Experiments
The rapid decline in LT50 values within the first 24 hours (Fig.
1 ) indicates the extreme sensitivity of the bivalves to temperatures
>31°C, where a 2°C increase sometimes meant the difference
between 0 and 100% mortality. Between 24 and 96 hours, toler-
ance limits were size dependent, a stable LT50 of 30°C being
reached after 48 hours for the small ones, after 72 hours at 28°C
for individuals >7 but <35 mm. and at 27°C for large D. serra
(Fig. 1). Thus, the smallest showed marginally better survival of
acute temperature exposure, resulting in a steady LT50 value 3°C
higher than that for the largest animals after 4 days.
It is evident from thermal resistance lines (Fig. 2) that the
survival time for all D. serra shows an exponential decline as
exposure temperature increases. Size-related differences in tem-
perature tolerance were only marginal because there were no sig-
nificant differences between the slopes and elevations of the three
resistance lines in Fig. 2 (slopes: F = 0.42, df = 18, p > 0.25;
elevations: F = 1.89, df = 20. 0.25 > p > 0.10). Data were
therefore combined to produce a single zone of tolerance where the
resistance line (Y) = 42.02 - 7.72(logX). With extrapolation.
this line defines the upper lethal temperature as 42°C and the
incipient temperature as 25°C (highest temperature to which an
organism can be continuously exposed for an indefinite period
without increasing mortality).
BT50 values indicated that the smallest individuals burrowed
immediately on transfer to temperatures near the lethal limits (36
to 38°C). whereas large ones displayed initial thermal shock by
lying on the surface for about 1 hour before burrowing (Fig. 3).
Longer exposure ( > 1 hour) to these high temperatures resulted in
the reemergence of D. serra <35 mm, followed by thermal pa-
ralysis and death. BT50 values increased for large D. serra as a
consequence of many individuals on the sand surface burying
themselves after initial thermal shock.
Between 12 and 48 hours, BT50 values stabilized for all sizes,
50% of the larger animals remaining buried at temperatures 2 to
3°C higher than individuals <35 mm. Over the last 2 days of the
experiment, values remained steady for the smallest bivalves but
declined to 26°C for the larger ones as they gradually emerged
from the sand to lie on the surface with their shells gaping. At this
O
o
• < 7 mm
A > 7 mm < 35 mm
■ > 35 mm
m — r
0 36 12
24
48
72
96
duration of exposure (h)
Figure 1. LT5n after acute exposure of D. serra to temperatures be-
tween 20 and 38°C for periods of 0, 3, 6, 12, 24. 48, 72, and 96 hours.
Three size groups were used with shell widths <7 mm, >7 but <35
mm, and ~>35 mm.
Exposure of D. serra
445
44 -i
■r- "2
40
B 38 -
36 -
34
32 -
E
a
ra
w
0)
a
E
0)
c
ro
•a
o
E
30-
28-
26
< 7 mm
> 7 mm < 35 mm
>35 mm
common regression
ZONE OF UPPER
TEMPERATURE TOLERANCE
I I I
10
-i — r
50
TH
100
median lethal time (h)
Figure 2. Median lethal times plotted against upper lethal tempera-
tures; data were regressed onto semilogarithmic axes to produce ther-
mal resistance lines, the common regression Isolid line) designating the
zone of upper temperature tolerance (shaded area).
point. BT5(, and LT50 values were nearly equal for each respective
size group, emphasizing the close association between emergence
from the sand and death due to thermal stress (Figs. I and 3).
In Figure 4, recovery of test individuals is defined as the ability
of 50% of animals to burrow and ventilate after being returned to
15°C for 4 days from temperatures between 20 and 38°C. On the
basis of these criteria, all sizes recovered after 4 days' exposure to
temperatures between 20 and 28°C and after 1 day at 30°C. Only
large individuals recovered after 2 days at 30°C, whereas more
than 50% of medium and small sizes lay on the sand surface with
their shells gaping and eventually died. No further recovery was
observed for any size beyond 2 days' exposure to 30°C. and at
32°C and above, there was no recovery, even after only 1 day of
exposure.
Stepwise Temperature Exposure Experiments
For any one particular temperature, the percent mortality was
plotted for the number of days exposed to that temperature after
the temperature was reached by stepwise daily increments of 2°C
from 14°C. All bivalves <7 mm held at 36°C for 1 day (after
percentage recovery
1 > 50% recovery
l< 50% recovery
O
o
0)
i—
*-»
03
J—
d)
Q.
E
0)
»—
V)
o
a.
x
o
E
E.
in
CO
A
to
CO
V
r-.
A
V
m
co
A
co
V
A
V
in
CO
A
in
CO
V
r-
A
V
in
co
A
m
CO
V
1^
A
V
34
32
30
28
26
24
1
2
3
4
duration of exposure (days)
Figure 4. Recovery rate of three size groups of D. serra over 4 days at
15°C after acute exposure for 4 days to temperatures ranging from 20
to 38°C.
stepwise introduction) died, whereas for individuals >7 mm, this
occurred at 32°C, resulting in corresponding LT50 values of 33°C
for the small size and 30°C for the larger ones (Fig. 5). The decline
in LT50 values from these values over 4 days was similar to that for
acute exposure for the same duration (Fig. 1), values being only
slightly higher by 1 to 2°C; thus, stepwise exposure to increasing
temperatures did little to enhance survival.
As exposure increased to 12 days, LT50 values remained fairly
constant between 27 and 30°C for all sizes. For small D. serra,
longer exposure resulted in LTS0 declining rapidly to 20°C and
then stabilizing from 14 to 16 days of exposure. LT50 values for
larger bivalves followed the same trend, but stability was reached
at higher LT50 values — 24°C for individuals >7 but <35 mm and
O
o
38 -i
rr
i
• < 7 mm
A > 7 mm < 35 mm
■ > 35 mm
T
03 6 12 24
I
48
72
I
96
duration of exposure (h)
Figure 3. BT5„ after acute exposure of D. serra to temperatures be-
tween 20 and 38°C for periods of 0, 3, 6, 12, 24, 48, 72, and 96 hours.
Three size groups were used with shell widths <7 mm, >7 but <35
mm, and >35 mm.
O
o
36 -l
32 -
28 -
24 -
20
• < 7 mm
A > 7 mm <35 mm
■ > 35 mm
1 I r
8
10
12
T — I-
14
16
duration of exposure (days)
Figure 5. LT50 values for D. serra gradually exposed to temperatures
increased daily by 2°C from 14 to 38°C over 16 days.
446
Stenton-Dozey and Brown
26°C for those >35 mm. Unlike acute exposure, stepwise expo-
sure up to 16 days resulted in large bivalves displaying the better
thermal tolerance.
BT50 values followed LT50 values for the first 8 days of ex-
posure (Fig. 6), when individuals died as soon as they emerged
from the sand. Once again, a slight increase in BT50 for large sizes
during the first 2 days of exposure represents an initial delay in
burrowing, even after gradual introduction to increasing temper-
atures. Beyond 8 days, BT50 values lagged behind LT50 values as
more individuals emerged from the sand before dying. By the end
of the experiment BT50 = LT50 for small animals (20°C), but for
larger animals. BTMI = 23°C, whereas LT50 = 26°C. This dif-
ference demonstrates that large animals surface at a temperature
below the 50% lethal limit and lie gaping for some time before
dying.
More than 50% of all sizes recovered after 16 days' exposure
to temperatures less than and equal to 26°C. after 13 days at 28°C.
and after 1 day exposure to 30°C (Fig. 7). After 14 and 15 days at
28°C, only large individuals recovered, and after 16 days, more
than 50% of all sizes showed no signs of recovery. At 30°C, after
2 days' exposure, medium and small sizes recovered, but after 3
and 4 days, only small ones fulfilled the recovery criteria; beyond
4 days, no recovery was noted. At 32°C, large D. serra recovered
after 1 day, but beyond this time and at higher temperatures, all
sizes lay on the surface until death.
o
2
ra
0)
a
E
o
2
in
O
a.
percentage recovery
shell widths
(mm)
■ > 50% recovery
< 50% recovery
in n n
CO V V
a r*-
A
*\
34
32
30
28
26
24
22
12 3 4 5 6 7 8
9 10 11 12 13 14 15 16
duration of exposure (days)
Figure 7. Recovery rate for three size groups of D. serra after 4 days
at 15°C after stepwise exposure at 2°C per day to temperatures rang-
ing from 14 to 38°C.
DISCUSSION
Intertidal distribution of D. serra at Ouskip is distinctly size
related, with the smallest animals at midtide, middle sizes between
mid- and low tide, and the largest in the surf zone. This implies
different temperature regimens, the smallest individuals being ex-
posed to the highest and most fluctuating temperatures. The tem-
peratures to which marine bivalves are exposed in their natural
environments are a function of mierohabitat plus latitudinal posi-
tion (Henderson 1929. Dickie 1958, Kennedy and Mihursky 1971,
Ansell et al. 1980a and b, Ansell and McLachlan 1980). These
factors interact with characteristics such as body shape and size,
growth, and reproductive condition to influence the upper temper-
ature tolerances of a species (Bayne 1976). Thus, size-related
differences in D. serra (Figs. 1 and 5) may be ecologically im-
portant, even though such differences, after acute exposure at
least, proved nonsignificant according to median lethal times (Fig. 2).
During low- water spring tides, temperatures in the sand at
depths of 10 to 15 cm at midtide, where small D. serra are found.
36 -I
O 32"
o
28 -
24 -
20 -
m
• < 7 mm
A > 7 mm < 35 mm
■ > 35 mm
1^
10
1 — I — I 1 1
12
14
16
duration of exposure (h)
Figure 6. BT;o values for D. serra gradually exposed to temperatures
increased daily by 2°C from 14 to 38°C over 16 days.
are usually warmer than in the surf zone by 1 to 3°C. Physiological
adjustment to such temperature changes is reflected in LT50 values
for small D. serra being 2 to 3°C higher than those for the larger
subtidal individuals after 4 to 1 1 days' exposure to near-lethal
temperatures (Figs. 1 and 5). The smaller size would therefore best
tolerate heat loading from the power station, largely by virtue of
intermittent exposure to the thermal plume at high tide. It has been
shown in Mytilus edulis that exposure to a high temperature in 6
hour cycles, rather than continuously, improved survival time by
83% (Pearce 1969).
The greater temperature tolerance of small individuals is fur-
ther demonstrated by the fact that they burrowed more rapidly than
adults at near-lethal temperatures. However, even though the
speed of burrowing declined with size, animals of large size still
retained the ability to burrow. This is of ecological significance,
because burrowing in nature, sometimes to a depth of 30 cm.
enables individuals to escape predation, dislodgment, and unfa-
vorable temperatures. Although it is unlikely that the D. serra
population near the power station would ever be exposed to lethal
temperatures (ca. >32°C). sublethal effects, such as reemergence
from the sand, could be of major importance. The ability to main-
tain position is especially important in the immediate vicinity of
the outfall, where the scouring effect of the effluent water, which
can reach a velocity of 80 m' r', is most concentrated.
After about 12 days of stepwise temperature exposure in the
laboratory, BT50 values were between 20 and 24°C, depending on
size (Fig. 6). These temperatures are found in the thermal plume
at the outfall, and long-term exposure is possible during extended
periods of onshore northwesterly winds, which trap the plume in
the surf zone (Rattey and Potgieter 1987). If such conditions pre-
vailed, individuals, especially the subtidal adults, would lose their
position in the sand to be swept ashore or preyed on by sandsharks,
crabs, and seabirds.
D. serra has an extensive distribution in southern Africa, from
Walvis Bay on the west coast to just north of East London on the
Exposure of D. serra
447
TABLE 1.
Comparison of LT50 values meaned over 48 and 72 hours of
exposure when LT50 values had stabilized for small (juvenile) and
large (adult) Donax species from the southern temperate (South
Africa). European warm-temperate (Mediterranean), and
Mediterranean-boreal (North Atlantic) (after Ansell and
McLachlan 1980).
Mean 48-72 Hour LT,„ (°C)
Species
Size
Locality
Small
1 arge
S. Africa
(south coast)
D.
sordidus
30.6
D.
serra
27.1
29.0
S. Africa
(west coast)
D.
serra
29.2
28.5
Mediterranean
D
trunculus
32.9
32.6
D.
semislriatus
29.9
N. Atlantic
D
villains
28.2
24.8
east. Intertidal distribution in Algoa Bay. on the south coast, is
also size related but is the reverse of the west coast pattern, with
juveniles in the swash zone at low tide and the adults in the
intertidal. A number of reasons, including differences in interspe-
cific competition, food supply, and temperature regimens (Donn
1986), have been proposed for this reversal. On the west coast, sea
temperatures range from 8 to 14°C in summer and from 1 1 to 17°C
in winter (Walker et al. 1984). and food supply is mainly detritus
and nearshore phytoplankton. which bloom in response to up-
welling (Stenton-Dozey 1989). In Algoa Bay, the summer maxi-
mum is 26°C with an annual mean maximum of 2 1 to 22°C ( Ansell
and McLachlan 1980), and in winter, temperatures drop to 15 to
17°C (Hanekom 1975). D. serra and another smaller species,
Donax sordidus. which is absent in the west, feed mostly on
diatoms sustained within the surf zone in a semienclosed ecosys-
tem (McLachlan and Bate 1984).
LT50 data presented by Ansell and McLachlan (1980) for D.
serra in Algoa Bay are directly comparable with data presented
here. After 48 and 72 hours of acute exposure, large south coast
individuals showed a higher tolerance than small ones (Table 1).
The burial response also differed in that, on exposure to high
temperatures, south coast adults burrowed immediately with a
BT50 after 1 hour of 34.5°C, compared with 29.5°C for adults
from the west coast, which also showed a marked delay in bur-
rowing (Fig. 4). Small individuals from both areas burrowed im-
mediately, but after 1 hour, those near the power station displayed
a BT50 of 37°C, compared with only 31°C for a similar size from
Algoa Bay. These differences suggest that upper temperature tol-
erances within the two populations are most strongly influenced by
differences between their respective microhabitats.
A characteristic stability of LT5„ values after 48 to 72 hours'
exposure allows direct comparison of Donax species from South
Africa with those from European waters (Table 1). Such a com-
parison provides an insight into the interaction between upper
thermal tolerances and latitudinal and bathymetric distributions.
Ansell et al. (1980b) attribute differences between the European
species to differences in zonational position on the shore and zoo-
geographic locality where higher LT50 values correspond to
greater tidal exposure and a more southerly distribution. Donax
trunculus, followed by Donax semis triatus, has the greater thermal
tolerances, reflecting the distribution of the former species in shal-
lower water and the more southern range of both compared with
Donax vittatus from the North Atlantic. The two South African
TABLE 2.
LT50 values after 24 hours exposure to near-lethal temperatures for small (juvenile) and large (adult) burrowing bivalves acclimated
between 15 and 20V from intertidal and suhtidal habitats in the northern and southern hemisphere.
24 Hour LT5(
Size
(°Cl
Species
Small
Large
Reference
Northern hemisphere
Subtidal
Placopeeten magillanicus
T.fabula (European N. Atlantic)
T.fabula (Mediterranean)
T tenuis (Mediterranean)
D. semislriatus
D trunculus
Intertidal
M. arenaria
G. gemma
M. ballhica
Modiolus demissus
T. tenuis (European N. Atlantic)
D. vittatus
Southern hemisphere
Intertidal
D. sordidus
D. serra
D. serra
31.6
32.0
29.0
32.0
22.5
Dickie (1958)
27.0
Ansell et al. (1980a)
29.0
Ansell et al. (1980a)
33.5
Ansell et al. (1980a)
30.0
Ansell et al. (1980b)
33.5
Ansell et al. (1980b)
30.5
Kennedy and Mihursky (1971)
37.2
Kennedy and Mihursky (1971)
31.5
Kennedy and Mihursky (1971)
38.4
Waugh and Garside (1971)
31.5
Ansell et al. (1980a)
29.0
Ansell et al. (1980b)
33.0
Ansell and McLachlan (1980)
31.0
Ansell and McLachlan (1980)
31.0
This study
448
Stenton-Dozey and Brown
species compare most closely with D. semistriatus and D. vittatus;
the size-related difference in D. vittatus does not reflect different
microhabitats but is believed to be an influence of age. In South
Africa, D. sordidus has the highest thermal tolerance, which is
indicative of its truly intertidal habitat where it migrates with the
tides. However, there is little difference in tolerance between the
west and south coast D. serra at the same tidal levels and this
probably reflects their similar latitudinal positions between 33 and
34°S.
Data for burrowing bivalves of other genera allow for a broader
comparison of LT50 values, but only after acute exposure to near-
lethal temperatures for 24 hours. Table 2 compares bivalves from
different shoreline distributions and latitudes in both northern and
southern hemispheres. The high LT5U values among the intertidal
group, which compare well with values for South African Donax,
support the general maxim that molluscs experiencing tidal expo-
sure have the greater thermal tolerances (Henderson 1929. South-
ward 1958, Kennedy and Mihursky 1971). Indeed, likeD. serra,
the greater tolerance of small-sized Mya arenaria and Macoma
balthica is directly related to a difference in microhabitat, where
juveniles are more exposed to the warming of mud flats on reced-
ing tide.
The thermal response of species compared within one study do
reflect zoogeographic influences. Between M. arenaria, M. bal-
thica, and Gemma gemma, the last species, followed by M. bal-
thica, have the greatest tolerances, coinciding with their wider
distribution range on the east coast of the United States. European
Tellina fabula and Tellina tenuis from the Mediterranean have
higher LT50 values than do populations in the North Atlantic, even
though T. tenuis is subtidal in its southern distribution.
Gradual exposure to increasing temperatures slightly increased
the upper temperature tolerances of D. serra (Fig. 5). This sug-
gests an ability to adjust upper limits and can be compared with
geographically separated Tellina species (Table 2), as well as to
seasonally acclimated Mya. Gemma, Macoma, Tellina, and Eu-
ropean Donax species, in which shifts in upper tolerances corre-
sponded to seasonal changes in temperature. Thus, it should be
borne in mind that LT50 values determined in the laboratory do not
represent a finite limit but rather an average upper tolerance,
which can assist in understanding not only an individual species
but also the thermal load that its biotic environment can tolerate.
Furthermore, it should be recognized that many factors, including
salinity, p02, the thermal limits of protein stability (Stenton-
Dozey 1989), and the thermal history of an organism (Newell
1979) interact to influence upper temperature tolerances, and each
combination of factors can be specific to a species and in some
cases to individuals within a species.
In conclusion, the upper temperature tolerances estimated for
D. serra in terms of median mortality indicate that the population
near the outfall of the power station would experience no adverse
effects from the thermal plume. However, Schubel et al. (1978)
have criticized the use of only 50% lethal limits and suggest that
a family of mortality curves be used ranging from 10 to 90% in
intervals of 10%. Such an approach would cover the eventuality of
unmeasured sublethal effects and provide a broader estimate of
thermal tolerances. In this study, estimates of burrowing success
provided a median sublethal measure that showed that D. serra
surfaced when exposed to temperatures that were not necessarily
lethal. Such a response must be considered when assessing effects
of a warmed effluent, because the loss of anchorage may mean
eventual death.
ACKNOWLEDGMENT
Financial support was provided by the South African Founda-
tion for Research Development.
REFERENCES
Ansell, A. D. & A. McLachlan. 1980. Upper temperature tolerances of
three molluscs from South African sandy beaches. J. Exp. Mar. Biol.
Ecol. 48:243-251.
Ansell, A. D., P. R. O. Barnett. A. Bodoy & H. Masse. 1980a. Upper
temperature tolerances of some European molluscs. I. Tellina fabula
and T. tenuis. Mar. Biol. 58:33-39.
Ansell, A. D., P. R O. Bamett. A. Bodoy & H. Masse. 1980b. Upper
temperature tolerances of some European molluscs. II. Donax vittatus,
D. semistriatus and D. trunculus. Mar. Biol. 58:41-46.
Bally, R. 1981. The ecology of three sandy beaches on the west coast of
South Africa. Ph.D. thesis. University of Cape Town, South Africa,
404 pp.
Bayne. B. L. 1976. Marine Mussels. Their Ecology and Physiology. Cam-
bridge University Press, London.
Bodoy, A. & H. Masse. 1977. Etude sur la resistance a la temperature de
quelques mollusques marine des cotes de Provence. Bull. Ecol. 8:91-
101.
Bodoy, A. & H. Masse. 1978. Etude experimentale de 1'influence de la
temperature sur la survive de mollusques bivalves marins endoges.
Haliotis 7:131-132.
Broekhuysen. C.J. 1940. A preliminary investigation of the importance of
desiccation, temperature and salinity as factors controlling the vertical
distribution of certain marine gastropods in False Bay, South Africa.
Trans. Roy. Soc. S. Afr. 28:255-292.
Cook, P. A. & D. A. Birkett. 1984. Koeberg Nuclear Power Station En-
vironmental Monitoring Programme: Baseline Ecological Survey. Uni-
versity of Cape Town, South Africa 1981-1983. 180 pp.
Cook, P. A. & D. A. Birkett. 1986. Koeberg Nuclear Power Station En-
vironmental Monitoring Programme: Baseline Ecological Survey. Uni-
versity of Cape Town. South Africa 1983-1985, 200 pp.
Dickie, L. M. 1958. Effects of high temperature on survival of the great
giant scallop. J. Fish. Res. Bd. Canada 15:1189-121 1 .
Donn, T. E. (ed.). 1986. Biology of the Genus Donax in Southern Africa.
Institute for Coastal Research U.P.E. Number 5, University of Port
Elizabeth, South Africa 51 pp.
Evans, R. C. 1948. The lethal temperatures of some common British
littoral molluscs. J. Anim. Ecol. 17:165-173.
Finney. D. J. 1964. Statistical Methods in Biological Assay. 2nd ed.
Hafner, New York.
Hanekom. N. 1975. A study of Donax serra in the eastern Cape. M.Sc.
thesis. University of Port Elizabeth, South Africa, 70 pp.
Henderson, J. T. 1929. Lethal temperatures of lamellibranchiata. Contr.
Can. Biol. Fish. 4:399-411.
Huntsman, A. G. & M. I. Sparks. 1924. Limiting factors for marine an-
imals. 111. Relative resistance to high temperature. Contr. Can. Biol.
Fish. 2:97-114.
Kennedy, V. S. &J. A. Mihursky. 1971. Upper temperature tolerances of
some estuarine bivalves. Chesapeake Sci. 12:193-204.
Lent, C. M. 1968. Air-gaping by the ribbed mussel. Modiolus demissus
(Dillwyn): Effects and adaptive significance. Biol. Bull. 123:60-73.
Exposure of D. serra
449
McLachlan. A. 1977a. Studies on the psammolittoral fauna of Algoa Bay.
South Africa. II. The distribution, composition and biomass of the
meiofauna and macrofauna. Zool. Afr. 12:33-60.
McLachlan, A. 1977b. Composition, distribution, abundance and biomass
of the macrofauna and meiofauna of four sandy beaches. Zool. Afr.
12:279-306.
McLachlan, A. & T. Erasmus. 1974. Temperature tolerances and osmo-
regulation in some estuarine bivalves. Zool. Afr. 9:1-13.
McLachlan, A. & G. Bate. 1984. Carbon budget for a high energy surf
zone. Vie Milieu 34:67-77.
Newell, R C. 1979. Biology of Intertidal Animals. Marine Ecological
Surveys, Faversham, Kent. England.
Pearce. J. B. 1969. Thermal addition and the benthos: Cape Cod Canal.
Chesapeake Sci. 10:227-233.
Rattey, D. & F Potgieter. 1987. Koeberg Nuclear Power Station: Warm
water plume report. 7 pp. + appendix. ESCOM, South Africa.
Schubel. J. R., C. C. Cputant & P. M. J. Woodhead. 1978. Thermal
effects of entrainment. pp. 20-93. In: J. R. Schubel and B. C. Marcy,
Jr. (eds.) Power Plant Entrainment: A Biological Assessment. Aca-
demic Press, New York.
Southward, A. J. 1958. Note on the temperature tolerance of some inter-
tidal animals in relation to environmental temperatures and geograph-
ical distribution. J. Mar. Biol. Assoc. U.K. 37:49-66.
Stenton-Dozey. J. M. E. 1989. The physiology and energetics of the
sandy-beach bivalve Donax serra (Roding) with special reference to
temperature and chlorine tolerance. Ph.D. Thesis, University of Cape
Town, 318 pp
Walker, N. D., J. Taunton-Clark & J. Pugh. 1984. Sea temperatures off
the South African west coast as indicators of Benguela warm events. S.
Afr. J. Sci. 80:72-77.
Waugh. D. L. & E. T. Garside. 1971. Upper lethal temperatures in rela-
tion to osmotic stress in the ribbed mussel. Modiolus demissus. J. Fish.
Res. Bd. Canada 28:527-532.
Journal of Shellfish Research. Vol. 13, No. 2.451 454, 1994.
EXPOSURE OF THE SANDY-BEACH BIVALVE DONAX SERRA RODING TO A HEATED AND
CHLORINATED EFFLUENT II. EFFECTS OF CHLORINE ON BURROWING AND SURVIVAL
J. M. E. STENTON-DOZEY AND A. C. BROWN
Zoology Department,
University of Cape Town,
Rondebosch 7700,
South Africa
ABSTRACT The effect of chlorine as free residuals in the range of 0.1 to 1.2 ppm on the survival and hurrowing ability of the
sandy-beach bivalve Donax serra was investigated. Median lethal times were calculated from daily observations over 2 weeks of the
number of dead and buried D. serra. Recovery in nonchlorinated sea water was monitored for 12 days after the transference of some
individuals every 24 hours. The addition ot chlorine resulted in immediate valve closure, a position maintained for 6 hours at
concentrations <0.3 ppm and up to 8 days at >0.6 ppm. No median lethal times were measurable at <0.6 ppm. but above this
concentration, the median lethal times approximated 10 days. After 14 days' exposure to between 0.6 and 1.2 ppm, 90 to 100% of
the bivalves died. Full recovery in fresh sea water occurred after 6 days' exposure to all chlorine concentrations. Longer exposure to
>0.6 ppm resulted in 50% mortality during the recovery test, whereas below this concentration. >50% recovered but burrowing was
delayed. Chemoreceptors on the siphons and mantle edge of D. serra probably enable the rapid detection of chlorine and hence
immediate valve closure. Possible modes of action of chlorine on physiological fitness are discussed. Long-term (> 14 days) survival
and retention of burrowing ability were invariable at chlorine concentrations <0.6 ppm
KEY WORDS: Bivalve, burrow, chlorine, power station effluent, sandy beach, survival
INTRODUCTION
A nuclear power station on the west coast of South Africa, 35
km north of Cape Town, discharges a heated and chlorinated ef-
fluent directly into the surf zone of a sandy beach densely inhab-
ited by the burrowing bivalve Donax serra Roding. Thermal ef-
fects from the plume were considered in a previous article in terms
of survival (% mortality) and burrowing ability (Stenton-Dozey
and Brown 1994). Temperatures >32°C were lethal to all sizes of
D. serra, whereas sublethal effects were noted after extended ex-
posure (4 to 11 days) to between 24 and 29°C, when 50% of the
animals could no longer burrow. The loss of this ability in the surf
zone near the outfall would mean displacement from the sand and
eventual death by predation or stranding on the beach.
At the inlet of the power station, chlorine iNaOCl) is added to
prevent fouling of the condenser cooling system's pipes. In water,
chlorine reacts immediately to form a mixture of hypochlorous
acid (HOC1) plus hypochlorite ion (OC1-), depending on pH,
temperature, and dissolved solids (Sugam and Helz 1977). This
mixture, referred to as free available chlorine (FAC), has strong
oxidizing properties and, in sea water, reacts rapidly with bromide
to produce hypobromous acid and hypobromite (Sugam and Helz
1977, Morgan and Carpenter 1978), which in turn combine readily
with organics in sea water (Goldman et al. 1978). Thus, the water
discharged at the outlet contains not only FAC, but also varying
amounts of bromine halites and halogens. Because chlorine chem-
istry in sea water is complex and because of the restrictive nature
of analytical techniques (Johnson 1978). most researchers elect to
measure chlorine concentrations as FAC or combined chlorine,
with the fate of the derivatives remaining unknown.
We selected to measure FAC: at the power station outfall, these
concentrations do not normally exceed 0.5 ppm but can reach 2.0
ppm if shock dosing is undertaken. Once the discharged water
reaches the surf zone, however, FAC is rapidly dispersed (Rattey
and Potgieter 1987), so that spatially, the possible impact on the
population of D. serra would be more limited than that of the
thermal plume (Stenton-Dozey and Brown 1994). The aim of this
article is to investigate the effect of FAC at concentrations present
at the discharge point on the survival and burrowing ability of D.
serra.
MATERIALS AND METHODS
Collection and Maintenance
Bivalves were collected 1 .5 km south of the power station, an
area unaffected by the discharge plume. Animals were kept in
flowing sea water in 25 1 tanks that were fitted with air-lift pumps
to facilitate the circulation of water at 15°C over a sand bed.
Because water in the aquarium was replaced regularly, natural
detritus was available as food but this was supplemented by peri-
odic additions of cultured algae (Tetraselmis suecica).
Effect of Chlorine on Survival
Stock solutions of 2 g of calcium hypochlorite per 1 of double-
distilled water were made up daily in drip bags fitted with adjust-
able wheels whereby 24 hour dosages could be controlled by set-
ting the number of drops per minute. The addition of chlorine in
distilled water did not reduce salinity levels, which were main-
tained between 34 and 35%<- . Because dosing with chlorine in this
manner did not allow precise control over concentration, it proved
more practical to work within chlorine ranges between 0. 1 and 1.2
ppm.
The concentration of FAC in sea water was determined by the
colorimetric, stabilized, neutral orthotolodine method (Johnson
and Overby 1969). This procedure minimizes stoichiometrical in-
terference from chloramines as well as iron, nitrite, alkalinity, and
acidity by using sulfosuccinate as a stabilizer so that a neutral
phosphate buffer could be applied at pH 7.0.
Because chlorine toxicity is affected by salinity. pH, and am-
monium-nitrogen (Burton 1977), these parameters were measured
every day. Salinity was adjusted to 34%o by dilution with double-
distilled water; pH was maintained at 7, and by adding fresh sea
water, ammonium-nitrogen was kept below 0.4 mg l_1, the
451
452
Stenton-Dozey and Brown
threshold concentration for the production of toxic, slow-decaying
chloramines (Inman and Johnson 1978).
Five batches of 30 adult bivalves each were placed in separate
20 1 tanks with slow-flowing, aerated sea water at 15°C, which
was circulated through a bed of sand via air-lift pumps. Four
batches were exposed to chlorine concentrations ranging from 0. 1
to 0.3, 0.3 to 0.6, 0.6 to 0.9, and 0.9 to 1.2 ppm for 2 weeks,
whereas the fifth batch was a control without chlorine. Median
lethal times were calculated from daily observation of dead and
buried D. serra, the numbers being corrected by deaths in con-
trols. Recovery was followed by transferring one individual from
each chlorine level to a tank with fresh, flowing, nonchlorinated
sea water every 24 hours and noting the time taken to reburrow or
die over 12 days.
RESULTS
Effect of chlorine in the range from 0. 1 to 1 .2 ppm on percent
mortality and burial response in adult D. serra at 15°C is illus-
trated in Figure 1. On dosing at all concentrations, D. serra im-
mediately retracted the siphons and foot and closed the valves
while buried in the sand. At the lowest chlorine concentration, this
position was maintained intermittently for approximately 3 to 6
hours; thereafter, the valves and siphons remained open. After 14
days' exposure, no adverse effects were apparent; mortality was
zero, and all individuals remained completely buried with their
siphons fully or partially open on the surface.
After the initial dosing between 0.3 and 0.6 ppm, animals
remained retracted for 1 to 2 days and then periodically opened
and closed their valves and siphons to ventilate over the next 12
days; a low percent mortality (<10%) was preceded by the emer-
gence of approximately 30% of individuals. No median lethal
times were measurable below 0.6 ppm.
At concentrations above 0.6 ppm, animals remained withdrawn
with their siphons and valves tightly closed for 7 to 8 days, during
which time the tissues were effectively isolated and protected from
the external medium. Thereafter, shell gape increased and the
siphons became limp; eventually, all emerged from the sand to lie
on the surface with their shells wide open. The median lethal time
for the range from 0.6 to 0.9 ppm was 10.5 days, and between 0.9
and 1 .2 ppm, it was 10 days. By the end of the 14 day experiment,
90% of those exposed to 0.6 to 0.9 ppm were dead, as were 100%
at 0.9 to 1.2 ppm.
On transference to fresh, nonchlorinated sea water, there was
full recovery after 6 days' exposure to all concentrations of chlo-
rine, although some delay in reburrowing was observed in the
range from 0.6 to 1.2 ppm (Fig. 2). However, beyond the sixth
exposure day, >50%< of individuals subjected to >0.6 ppm never
reburied in fresh sea water. Time on the surface before dying
ranged from 5 days, after 7 days' exposure, to less than 1 day by
the end of the experiment. On the other hand, at concentrations of
<0.6 ppm, >50% of transferred animals recovered fully, although
reburial was slow after 12 days' exposure.
DISCUSSION
The immediate withdrawal of the siphons and foot, as well as
valve closure, by D. serra on dosing with chlorine is a common
escape response among burrowing bivalves suddenly exposed to a
chemical pollutant, as well as to drastic changes in salinity (Block
1977, Akberali and Black 1980, Trueman and Akberali 1981,
Akberali and Davenport 1982, Trueman 1983). At low FAC con-
centrations (<0.3 ppm). D. serra reemerged and resumed pump-
Chlorine
concentration
(ppm)
column L = 0.1 -0.6
1 completely buried
■
R3 dead
partially buried
or on surface
3
CO
TJ
re
Ol
•D
CD
o>
re
01
a
14
duration of exposure (days)
Figure 1. Effect of chlorine in the range from 0.1 to 1.2 ppm on the
mortality rate and burial response of D. serra at I5°C.
duration of exposure (days)
Figure 2. Recovery of adult D. serra at 15°C with no chlorine after
exposure to chlorine in the lower range (column L) of 0.1 to 0.6 ppm
and the higher range (column H) of 0.6 to 1.2 ppm. Total recovery was
recognized by an ability to burrow (speckled area), partial recovery by
an inability to burrow (black area), and no recovery by ultimate death
(striped area). Recovery criteria were set by the relevant behavior of
more than 50% of the bivalves.
Effects of Chlorine
453
ing within 3 to 6 hours of chlorine addition. At levels approaching
lethal limits (>0.6 ppm), the valves remained closed for up to 8
days, although full recovery in nonchlorinated sea water only oc-
curred after 6 days' exposure. During this time, the mantle mar-
gins often protruded, but because they were always held tightly
together, it seems likely they afforded the same protection as
sealed valves.
The estuarine bivalve Scrobicularia plana can remain closed
for a similar period, 5 to 7 days, in the presence of copper and low
salinities (Trueman and Akberali 1981, Trueman 1983), and
Mytilus edulis responded to salinity decline by closing for 4 days
(Davenport, 1981 ). The duration of valve closure must be a critical
period because eventually depletion of energy resources (Stenton-
Dozey 1989). build-up of an oxygen debt (van Wijk et al. 1989).
or the accumulation of metabolites would force the animals to
open their valves and interact with the environment (Akberali and
Black 1980). Valve closure is thus effective in isolating tissues
from unfavorable conditions, provided that these conditions are of
a transient or recurrent short-term nature. Such avoidance behavior
is not restricted to bivalves but also occurs in other valved organ-
isms, for example, barnacles and cirripede cyprids. It is for this
reason that power stations usually chlorinate continuously, be-
cause intermittent addition is ineffective in removing such ani-
mals.
At moderate FAC levels, 0.3 to 0.6 ppm, D. serra periodically
opened and closed valves and siphons to draw in water, and this
action is probably a way of sampling external conditions. Hodgson
and Fielden (1984) have shown the existence of ciliated sensory
receptors on the inner and outer side of both siphons, as well as on
the lobes, of the tentacles and mantle edge of D. serra, and there
are strong indications that these are chemoreceptors. The cruci-
form muscle complex could be another site of mechano- and
chemoreception, as found in other burrowing bivalves (Odiete
1978. Pichon et al. 1980). Siphonal receptors have also been ob-
served in M. edulis (Davenport 1981). on the mantle tentacles of
the giant scallop Placopecten magellanicus (Moir 1977). and on
Lima luans (Owen and McCrae 1979). The sensitivity of isolated
siphonal preparations to low levels of chemical stimuli (Akberali
etal. 1981. Hodgson 1982) is further evidence of the efficiency of
chemoreception in bivalves.
The percentage of D. serra dying on exposure to FAC was not
gradual but rather displayed a stepwise response, with the thresh-
old between chronic and acute toxicity at around 0.6 ppm. Above
0.6 ppm, median lethal time approximated 10 days, but below this
concentration, percent mortality never reached 50% over a period
of 14 days. The American oyster Crassostrea virginica shows
even greater tolerance than D. serra with <10% mortality when
exposed to chlorine in the range from 0.35 to 0.85 ppm for 15 days
(Scott and Middaugh 1978). Such resilience to chlorination was
not evident in a model suggested by Mattice and Zittel ( 1976) that
predicted the threshold between acute and chronic toxicity for a
heterogeneous array of organisms to be only 0.02 ppm. However,
this model could be biased toward a low threshold, because a high
proportion of fish that are extremely sensitive to chlorine (Morgan
and Carpenter 1978. Jolly et al. 1978, Hocutt et al. 1980) were
included.
The effect on marine organisms of the variety of halogenated
organics that are rapidly formed when chlorine is added to sea
water is, as yet, not well understood. There is some evidence that
they may be more toxic than chlorine itself (Waugh 1964, Morgan
and Carpenter 1978, Hileman 1982, Helz and Kosak-Channing
1984). However, irrespective of the effective pollutant being chlo-
rine or its derivatives, it is not known which aspect of the organ-
ism's physiology is most impaired. Interference with ehemosenory
ability in C. virginica has been documented (Hillman 1980), with
secondary consequences such as lack of detection of food and
predators. Research on juvenile fish and zooplankton (Capuzzo
1977, Capuzzo et al. 1977a and b) suggests that the mode of action
of chlorine appears to be some form of metabolic inhibition, al-
though the actual mechanisms remained unknown. Those authors
also found that respiration rates in fish declined by 50% at standard
threshold concentrations, possibly indicating physical damage to
the gills. Furthermore, decreases in the respiration rates of zoo-
plankton larvae occurred at levels of residual chlorine that were
virtually undetectable. Such results negate the value of the median
lethal chlorine concentration and highlight the necessity to set
threshold dosage below concentrations that have sublethal effects.
This aspect is addressed in the final article in this series in terms
of effects on the heart rate of D. serra.
CONCLUSIONS
Laboratory studies have shown that D. serra is able to protect
itself against a wide range of chlorine levels as long as exposure
does not exceed 6 days. During this period, the animal remains
buried and effectively isolated from the environment by closing
the valves but still retaining an ability to detect and respond to
external changes via chemoreceptors on the mantle margin. Nev-
ertheless, on a wave-swept beach, such a response could result in
dislodgment from the sand if the animal is not buried deeply, with
lethal consequences. However, this danger is far less at chlorine
concentrations of <0.6 ppm, because D. serra is able to resume
burrowing and pumping within 2 days after the initial closure
response.
ACKNOWLEDGMENT
Financial support was provided by the South African Founda-
tion for Research Development.
REFERENCES
Akberali, H. B. & J. E. Black. 1980. Behavioural responses of the bivalve
Scrobicularia plana (da Costa) subjected to short-term copper (Cu II)
concentrations. Mar. Emir. Res. 4:97-107.
Akberali, H. B. & J. Davenport. 1982. The detection of salinity changes
by the marine bivalve molluscs Scrobicularia plana (da Costa) and
Mytilus edulis L. J. Exp. Mar. Biol. Ecol. 58:59-72.
Akberali, H B , T. M. Wong & E. R Trueman. 1981. Behavioural and
siphonal tissue responses of Scrobicularia plana to zinc. Mar. Emir.
Res. 5:251-264.
Block, R. M. 1977. Physiological responses of estuarine organisms to
chlorine. Chesapeake Sci. 18:158-160.
Burton, D. T. 1977. General test conditions and procedures for chlorine
toxicity tests with estuarine and marine macro-invertebrates and fish.
Chesapeake Sci. 18:130-136.
Capuzzo, J. M. 1977. The effects of chlorine and chloramines on growth
and respiration rates of larval lobsters (Homarus americanus). Water
Res. 11:1021-1024.
Capuzzo, J. M., J. A. Davidson, S. A. Lawrence & M. Libm. 1977a. The
454
Stenton-Dozey and Brown
differential effects of free and combined chlorine on juvenile marine
fish. Esluar. Coastal Mar. Sci. 5:733-741 .
Capuzzo, J. M., J. C. Goldman, J. A. Davidson & S. A. Lawrence.
1977b. Chlorinated cooling waters in the marine environment: devel-
opment of effluent guidelines. Mar. Poll. Bull. 8:161-163.
Davenport. J. 1981. The opening response of mussels (Mytilus edulis)
exposed to rising sea water concentrations. J . Mar. Biol. Assoc. U.K.
61:667-678.
Goldman, J. C, J. M. Capuzzo & G. T. F. Wong. 1978. Biological and
chemical effects of chlorination at coastal power plants, pp. 291-305.
In: R L. Jolly, H. Gorchev and D. H. Hamilton, Jr. (eds.). Water
Chlorination: Environmental Impact and Health Effects, vol. 2. Ann
Arbor Science Publishers, Inc. Ann Arbor.
Helz, G. R. & L. Kosak-Channing. 1984. Dechlorination of waste water
and cooling water. Environ. Sci. Technol. 18:48A-55A.
Hileman, B. 1982. The chlorination question. Environ. Sci. Technol. 16:
15A-18A.
Hillman, R. E. 1980. Behaviour of bivalve molluscs, pp. 309-325. In:
C. H. Hocutt, J. R. Stauffer. Jr., J. E. Edinger. L. W. Hall, Jr &
R. P. Morgan II (eds.). Power Plants: Effects on Fish and Shellfish
Behaviour. Academic Press, New York.
Hocutt, C. H., J. R. Stauffer, Jr., J. E. Edinger, L. W Hall, Jr. & R. P.
Morgan II (eds.). 1980. Power Plants: Effects on Fish and Shellfish
Behaviour. Academic Press, New York, 346 pp.
Hodgson, A. N. 1982. Some behavioural and electrical responses of Scro-
bicularia plana (Bivalvia: Tellinacea) to siphonal wounding. J. Moll.
Stud. 48:87-94.
Hodgson, A. N. & L. J. Fielden. 1984. The structure and distribution of
peripheral ciliated receptors in the bivalve molluscs Dona* serra and
Donax sordidus. J. Moll. Stud. 50:104-1 12.
Inman, G. W. Jr. & J. D. Johnson. 1978. The effect of ammonia concen-
tration on the chemistry of chlorinated sea water, pp. 235-252. In:
R L. Jolly, H. Gorchev and D. H. Hamilton, Jr. (eds). Water Chlo-
rination: Environmental Impact and Health Effects, vol. 2. Ann Arbor
Science Publishers, Inc. Ann Arbor.
Jolly. R. L., H. Gorchev & D. H. Hamilton, Jr. (eds). 1978. Water
Chlorination; Environmental Impact and Health Effects, vol. 2. Ann
Arbor Science Publishers. Inc. Ann Arbor.
Johnson, J. D. 1978. Measurement and persistence of chlorine residuals in
natural waters, pp. 37-63. In: R. L. Jolly (eds). Water Chlorination:
Environmental Impact and Health Effects, vol. 1. Ann Arbor Science
Publishers. Inc. Ann Arbor
Johnson, J. D. & R. Overby. 1969. Stabilised neutral orthotolodine
method for chlorine. Anal. Chem. 41:1744-1750.
Mattice.J. S.&H. E. Zittel. 1976. Site-specific evaluation of power plant
chlorination. J. Wat. Pollut. Control. Fed. 48:2284-2308.
Moir, A. J. G. 1977. Ultrastructural studies on the ciliated receptors of the
long tentacles of the giant scallop. Placopecten magellanicus (Gme-
lin). Cell Tissue Res. 184:367-380.
Morgan. R. P. & E. J. Carpenter. 1978. Biocides. pp. 95-133. In: J. R
Schubel and B. C. Marcy, Jr. (eds.). Power Plant Entrainment: A
Biological Assessment. Academic Press, New York.
Odiete, WO. 1978. The cruciform muscle and its associated sense organs
in Scrobicularia plana (da Costa). J. Moll. Stud. 40:180-189.
Owen. G. & J. M. McCrae. 1979. Sensory cell/gland complexes associ-
ated with the pallial tentacles of the bivalve Lima Hans (Gmelin), with
a note on specialised cilia on the pallial curtains. Phil. Trans. Roy. Soc.
Lond. 287:6-62.
Pichon. Y.. M. Moueza & L. Frenkiel. 1980. Mechanoreceptor properties
of the sense organ of the cruciform muscle in a Tellinacea lamelli-
branch. Donax trunculus L.: an electrophysiological approach. Mar.
Biol. Lett. 1:273-284.
Rattey. D. & F. Potgieter. 1987. Koeberg Nuclear Power Station: Warm
Water Plume Report, ESCOM, South Africa. 7 pp. + appendix.
Roberts, M H . R J Diaz, M. E Bender & R. Hugget. 1975. Acute
toxicity of chlorine to selected estuanne species. J . Fish. Res. Board
Can. 32:2525-2528.
Scott, G. I. & D. P. Middaugh. 1978. Seasonal chronic toxicity of chlo-
rination to the American oyster, Crassostrea virginica. pp. 311-328.
In: R. L. Jolly. H. Gorchev & D. H. Hamilton. Jr. (eds.). Water
Chlorination: Environmental Impact and Health Effects, vol. 2, Ann
Arbor Science Publishers, Inc. Ann Arbor.
Stenton-Dozey, J. M. E. 1989. The physiology and energetics of the
sandy-beach bivalve Donax serra Roding with special reference to
chlorine and temperature tolerance. Ph.D. Thesis, University of Cape
Town. 318 pp.
Stenton-Dozey, J. M. E. & A. C. Brown. 1994. Exposure of the sandy-
beach bivalve Donax serra Rddmg to a heated and chlorinated effluent.
I. Effects of temperature on survival and burrowing. J. Shellfish Res.
13(2 ):443-M9.
Sugam, R. & G. R. Helz. 1977. Speciation of chlorine produced oxidants
in manne waters: theoretical aspects. Chesapeake Sci. 18:116-118.
Trueman, E. R. 1983. Behavioural and physiological responses of a bur-
rowing bivalve to stress, pp. 669-673. In: A. McLachlan & T. Eras-
mus (eds.). Sandy Beaches as Ecosystems. Junk Publications. The
Hague.
Trueman. E. R. & H. B. Akberali. 1981. Responses of an estuarine bi-
valve Scrobicularia plana (tellinacea) to stress. Matacologia 21:15-
21
Van Wijk. K... J. M. E. Stenton-Dozey & A. C. Brown. 1989. Oxycon-
formation in the burrowing bivalve Donax serra Roding. J. Moll. Stud.
55:133-134.
Waugh. G. D. 1964. Observations on the effects of chlorine on the larvae
of oysters (Oslrea edulis [L.]) and barnacles {Elminius modestus [Dar-
win]). Ann. Appl. Biol. 54:423-440.
Journal of Shellfish Research, Vol. 13, No. 2. 455-459. 1994.
EXPOSURE OF THE SANDY-BEACH BIVALVE DONAX SERRA RODING TO A HEATED AND
CHLORINATED EFFLUENT III. EFFECTS OF TEMPERATURE AND CHLORINE ON
HEART RATE
J. M. E. STENTON-DOZEY AND A. C. BROWN
Zoology Department.
University of Cape Town,
Rondebosch 7700,
South Africa
ABSTRACT The singular and synergistic effects of temperature and free available chlorine in the range of 0. 1 to 1 .2 ppmon the heart
rate of the sandy-beach bivalve Donax serra were investigated. The aim was to identify the possible sublethal impact of a heated and
chlorinated effluent on a dense bivalve population near the outfall. The heart rate of adults, in the presence or absence of chlorine, was
recorded at 15, 20, 25, and 30°C after daily increments of 5°C from 15°C. One individual from each exposure temperature was
transferred daily to fresh sea water at 15°C to monitor recovery. Heart rate reached a maximum of 44 beats min~ ' at 25°C from a basal
rate of 15 beats min~ ' at 15°C. The heart rate of those exposed to 30°C did not return to basal frequency in fresh sea water at 15°C.
On dosing with chlorine, beat frequency immediately dropped to half the basal rate as the valves closed and ventilation stopped. Below
0.6 ppm and <25CC. the basal rate returned within 24 hours, and full recovery occurred in fresh sea water at 15°C. Results indicated
that plume conditions would not be lethal to D. serra but that sublethal effects relating to burrowing activity and heart rate can be
expected.
KEY WORDS: Bivalve, chlorine, effluent, heart rate, temperature, sandy beach
INTRODUCTION
This article is the final in a series of three concerned with the
effects of a heated and chlorinated effluent from a nuclear power
station on the physiological fitness of Donax serra Roding in
South Africa. Dense populations of this intertidal sandy-beach
bivalve are directly exposed to a thermal plume 10°C above the
ambient seasonal range of 8 to 17°C and chlorine concentrations
between 0. 1 and 0.4 ppm. Initial laboratory studies on the effects
of temperature and chlorine on survival and burrowing (Stenton-
Dozey and Brown 1994a and b) showed that the discharge plume
was unlikely to have a lethal effect. However, D. serra did display
distinct postural changes, such as increased valve adduction and
siphonal closure, which indicated possible sublethal effects. This
necessitated the selection of a physiological parameter to quantify
these behavioral responses.
A common physiological measure of the response of bivalves
to raised temperatures, as well as to the presence of chemical
pollutants, is change in heart activity (Coleman 1974. Lowe 1974.
Earll 1975, Trueman and Akberali 1981. Trueman 1983). Heart
rate is easy to monitor using impedance techniques (Trueman et al.
1973). which in addition, provide evidence of gross changes in
behavior, such as valve movements, burrowing, and periods of
inactivity or heightened activity. These responses in turn corre-
spond closely to changes in the physical and chemical environment
(Earll 1975). In this article, heart rate is used as an indicator of the
sublethal effects on D. serra of raised temperatures both in the
presence and in the absence of chlorine.
MATERIALS AND METHODS
Effect of Temperature on Heart Rate
The electronic recording technique developed by Trueman
(1967) was used to monitor heart rate. Two small holes were
drilled in each valve over the heart. Fine silver-wire electrodes
were inserted through these holes to lie on opposite sides of the
pericardium. The electrodes were sealed in place with dental wax
and connected in series to an impedance pneumograph (Alternate
Current coupling) and oscillograph by a lightly screened cable.
Changes in impedance between electrodes resulting from pulsatile
changes in heart volume or from movements of the animal within
its shell were recorded on the oscillograph.
Adult bivalves with implanted electrodes, plus others without
electrodes, which served to demonstrate any detrimental effect of
the implant on postural behavior, were kept in 20 I tanks and
allowed to equilibrate overnight at 15°C. Because starvation can
markedly depress heart rate in bivalves (Bayne 1976). all animals
were fed cultured algae during the equilibration period. Heart rate
was initially monitored by transferring individuals directly from
15°C to either 20, 25, or 30°C for 1 to 2 days. However, this
involved long periods of reequilibration before recording could
commence, so as an alternative, the temperature was raised, with-
out disturbing the animals, by 5°C every 24 hours from 15 to 30°C.
Both methods of increasing temperature resulted in similar heart
activity.
Preliminary experiments involving continuous recordings indi-
cated that no unpredictable short-term variations in heart activity
occurred. The heart rate of each individual was. therefore, mon-
itored for 15 minutes in every hour for periods up to 18 hours.
The experiment was repeated six times at each temperature, and
five individuals were used per run, four with implants and one as
a control.
One individual from each exposure temperature was transferred
daily to fresh sea water at 15°C to monitor recovery over 24 hours.
An equilibration period of 6 hours was allowed before the heart
rate of these animals was measured.
Combined Effect of Temperature and Chlorine on Heart Rate
Adult bivalves were simultaneously exposed to the temperature
regimen described above and free available chlorine (FAC) con-
centrations ranging from 0. 1 to 0.3. 0.3 to 0.6. 0.6 to 0.9, and 0.9
to 1.2 ppm. Chlorine was administered as calcium hypochlorite
455
456
Stenton-Dozey and Brown
and measured as FAC, as described in Stenton-Dozey and Brown
(1994b). Experiments were repeated three times for each chlorine
range and temperature, using five individuals per run in which one
served as a control without implanted electrodes. Some individuals
were removed daily after 24 hours' exposure to each treatment,
placed in nonchlorinated sea water at 15°C, and allowed to equil-
ibrate for 6 hours before the recovery heart rate was monitored for
24 hours.
RESULTS
Effect of Temperature
While buried and ventilating at 15°C, heart activity steadied at
13 ± 2 beats min" '; experiments commenced once this rate was
maintained for 24 hours (Fig. 1A). When the temperature was
exposure temperatures
(5°C increments 24 hrs '1 )
15°C 20°C 25°C 30°C
50
40 -
30
B 20 H
n
r
a
10 -
Hi ftt
*♦
H
fr
irregular
"•• beats or
death
12 3 4
consecutive 24 hr exposure periods for 4 days
raised to 20°C. heat frequency increased immediately, the maxi-
mum rate of 26 beats min" ' being reached after 5 hours. After 24
hours at 20°C and at the beginning of exposure to 25°C, heart
activity was less stable, fluctuating between means of 18 and 33
beats min' ', with standard deviations much higher than for data
at 15 or 20°C. Just before the temperature was raised from 25 to
30°C, heart rate had risen to 44 beats min"1, but at 30°C, it
became irregular, followed most often by death after 12 hours.
Because controls showed stress but did not die at 30°C, death must
to some degree be attributed to implanted electrodes. At maximum
heart rate, Ql0 values equaled 4.0 between 15 and 20°C and 2.9
from 20 to 25°C.
Bivalves exposed to 20 and 25°C recovered completely when
returned to 15°C, as shown by return to a normal heart rate of ± 13
beats min"' (Fig. IB). Those from 30°C displayed erratic heart
rates on transference to 15°C and. after monitoring for 24 hours,
either retained an irregular beat while lying on the surface gaping
or were dead. No controls died during recovery, but some indi-
viduals did emerge to lie on the sand surface, thereby indicating
stress related to temperature alone rather than to electrode im-
plants.
On increasing the temperature to 20°C, a rapid increase in beat
frequency was apparent from 15 to 22 beats min" ' in 5 minutes
(Fig. 2B); adductions were less frequent, more protracted, and
often followed by cessation of beats for 30 seconds when the
valves closed (Fig. 2C). At 25°C, regular beating often gave way
to extensive contraction and retardation or suppression (Fig. 2D).
Suppression often alternated with a reduction in beat amplitude,
followed by rapid reextension and commencement of ventilation
(Fig. 2E). Throughout exposure to 25°C, the shell gaped and the
siphons were, most often, fully extended.
Shell gape was maximal at 30°C, and both siphons remained
fully extended but limp. Beat frequency, although rapid and often
steady, was of low and irregular amplitude, with infrequent ad-
ductions, followed by an extended suppression of heart activity
(Fig. 2F). At death, the shell gaped, the siphons collapsed, and
beats became more indistinct, fading away or ceasing after a pe-
riod of regular, strong beats (Fig. 2G).
Synergistic Effect of Chlorine and Temperature
previous exposure temperature
_ 30
c
I
« 20
g 10-
t
3
15°C
20°C
1
25°C
\
30°C
irregular
■Hi'" beats or
death
24 hrs
24 hrs
24 hrs
24 hrs
separate 24 hr recovery periods at 15°C
Figure 1. (A) Effect of temperature on the heart rate of adult D. serra
when raised 5°C per day from 15 to 30°C. Vertical bars represent one
standard deviation around the mean. (B) Recovery heart rates at !?"(
after exposure to 20, 25, and 30°C for 24 hours.
On dosing between 0. 1 and 0.3 ppm at 15°C, the valves closed,
resulting in a fall in heart rate from 13 to around 8 beats min" '
(Fig. 3). Over the next 24 hours, heart rate increased to 10 beats
min" ' as D. serra reemerged. A rise to 20°C and then 25°C over
the next 48 hours resulted in beats increasing to 1 1 and 14 beats
min" '. respectively. This frequency was nearly four times lower
than that reached at 25°C without chlorine (see Fig. 1 ). At 30°C,
13 to 14 beats min " ' were maintained for 6 hours, and then, heart
activity became indistinct as animals began dying. In the range
from 0.3 to 0.6 ppm at 15°C, heart rate dropped to 6 and only rose
to 10 at 20°C when the valves began gaping. At 25 and 30°C, the
heart responded as for the dosage of <0.3 ppm (Fig. 3).
On dosing with chlorine between 0.6 and 0.9 ppm at 15°C,
heartbeat dropped immediately to 5 beats min " ' , only increasing
to 7 over the next 2 days with an increase from 20 to 25°C (Fig.
3). After 18 hours at 25°C. the heartbeat showed no regular pattern
and bivalves began dying at 30°C. A similar response occurred in
the range from 0.9 to 1.2 ppm, when heart rate never rose above
6 beats min " ' , becoming indistinct after 3 hours at 25°C; indi-
viduals began dying before the temperature was raised to 30°C.
Effects on Heart Rate
457
Regular beats at 15 C interspersed with
adductions (a) associated with burrowing
Increase in beat frequency as temperature
was increased trom 15 to 20 °C (arrowed)
At 20 °C fewer adductions (a) olten followed
immediately by cessation (c) of heart beat
At 25 C regular beating often gave way to periods of
extensive contraction (ec).... and retardation (r) or suppression (s)
At 25 °C periods of suppression alternated
with
periods of reduced beatinq
E
ill
llklWw jfclWrrJ.
1
|p 1 IfW
\v V • M\
min
Amplitude of beats at 30 C was greatly
reduced and often irregular
At 30 C beats became more and more indistinct and as death approached, either
faded away or suddenly ceased after beating strongly
Figure 2. Traces of heart activity showing the effect of temperature (IS to 30°C) on the heart rate of adult D. serra.
Q10 was negative above 0.6 ppm, demonstrating the predom-
inance of the chlorine effect, so that heart rate was no longer
positively temperature dependent. Furthermore, in this chlorine
range, heart activity was suppressed to a steady, slow beat, with
small standard deviations from means, irrespective of temperature,
whereas at concentrations below 0.6 ppm, deviations were higher
and more variable as temperature and chlorine interacted to raise
and suppress beat frequency (Fig. 3).
On transference to nonchlorinated sea water at 15°C, all indi-
viduals exposed to FAC <0.6 ppm and 25°C or less recovered
fully, as indicated by a return to the normal basal rate of 13 beats
min-1 (Fig. 3). However, individuals from 30°C either died or
continued to display irregular heart activity. In the range from 0.6
to 1.2 ppm, a return to normal beat frequency was only observed
in individuals from 15°C (Fig. 3). Those from 20°C displayed a
suppressed heart rate (10 beats min-1) during the recovery test,
whereas those exposed to 25°C retained a disrupted beat pattern.
Traces of heart activity showed that at FAC concentrations
<0.6 ppm, heartbeats became greatly protracted (Fig. 4A), but
once the temperature was raised to 20°C, the normal pattern re-
turned (Fig. 4B). At 25°C, beats were once again protracted but
weaker, with periods of total suppression (Fig. 4C), and at 30°C,
they were even weaker and hence lower in amplitude and indistinct
(Fig. 4D). Above 0.6 ppm, beat frequency became irregular on
dosing at 15°C and this pattern persisted at 20°C (Fig. 4E). At
25°C, regular protracted beats disintegrated into an indiscreet pat-
tern as animals neared death (Fig. 4F).
DISCUSSION
With an increase in temperature, the heart rate of D. serra
increased immediately and did not acclimate over the range from
15 to 30°C. two responses that make this physiological parameter
a sensitive and hence convincing measure of sublethal effects.
B
EXPOSURE RECOVERY AT 15°C
TO CHLORINE NO CHLORINE
(at 5°C increments 24hr') previous exposure temperature
15°C 20°C 25°C 30°C
fHfrf^fH
111 ii
TTrTT
t*H^#-
o ^
_M+
N
#
*♦—*)!
24
^-H
24
24
24
consecutive 24 hr exposure
periods for 4 days
24 hr recovery periods
Figure 3. (A) Combined effect of temperature and chlorine on the
heart rate of adult D. serra. Temperature was increased by 5°C daily
for 4 days from 15 to 30°C within the FAC range from 0.1 and 1.2
ppm. (B) Recovery heart rates at 15°C without chlorine after exposure
to 15, 20, 25, and 30°C with chlorine.
458
Stenton-Dozey and Brown
Nonacclimation is common among bivalves, for example, Isog-
numon alatus (Trueman and Lowe 1971), My a arenaria (Lowe
and Trueman 1972), Mytilus edulis and Mytilus californianus
(Pickens 1965, Widdows 1973), Crassostrea gigas (Lowe 1974)
and Perna perna (Bayne 1976). Those species, like D. serra,
displayed Q10 values >2 between 10 and 25/27°C, beyond which,
beat frequencies generally became erratic, thereby reflecting the
critical temperature at which homeostatic mechanisms begin to
break down. For D. serra, this coincided with failure to return to
a normal heartbeat at 15°C after exposure to 30°C.
Also common among bivalves is an immediate increase in heart
rate on increasing temperature. It has been suggested that ther-
moreceptors, possibly in the mantle tissue, play an important role
in this response (Trueman and Lowe 1971, Lowe 1974). Because
there is an abundance of sensory cells on the siphons and mantle
edge of D. serra, even though most appear adapted to detect
chemical or mechanical changes in the external medium (Hodgson
and Fieldin 1984), it seems reasonable to assume that at least some
of these cells function as thermoreceptors with a possible neural
connection to the heart.
The addition of chlorine between 0. 1 and 1 .2 ppm to sea water
at 15°C led to immediate bradycardia in D. serra; beat frequency
drastically declined as the siphons and foot withdrew and the
valves closed tightly. This is a common stress response among
bivalves to a pollutant, a salinity drop, or aerial exposure, and in
all of these instances, valve closure is accompanied by a decline in
heart rate (Trueman 1967, Coleman and Trueman 1971, Earll
1975, Akberali and Black 1980, Trueman and Akberali 1981).
Valve closure leads to a drastic drop in p02 and an increase in
pCO, levels of the mantle cavity water in M. edulis (Bayne 1971 )
and Scrobicularia plana (Akberali and Trueman 1979), and it is
this drop in oxygen tension, rather than any mechanical effect of
closed valves, that is believed to attenuate beat frequency (Bayne
1976).
D. serra became acclimatized to continuous exposure to low
FAC concentrations at 15°C, as demonstrated by a return to nor-
mal heart rate, paralleled by increasing shell gape and pedal and
siphonal extension, during exposure to <0.3 ppm. This is sup-
ported by the low mortality rate of D. serra exposed to 0.1 to 0.3
ppm at 15°C for 2 months in experiments after biochemical
changes in the body tissues (Stenton-Dozey 1989). Above 0.3
ppm, a suppression of heart activity and valve closure persisted for
the 24 hours of the experiment. Because D. serra is able to main-
tain valve closure for 7 to 8 days when continuously exposed to
this FAC range (Stenton-Dozey 1989), the resultant oxygen debt
(Bayne 1976) would prevent heart activity from returning to nor-
mal and anaerobic respiration would probably sustain basal me-
tabolism (De Zwaan 1977). The accumulated endproducts of
o
15 C, 0.1-0.6 ppm: Beats protracted directly
after dosing with chlorine
25 C, 0.1 -0.6 ppm: Beats protracted and
weaker with periods of total suppression
20 °C
0.9 -1.2 ppm: Beats irregular directly
after dosing with chlorine
20 C, 0.1 - 0.6 ppm: As temperature increases
to 20 °C, normal beat pattern returns
30 C, 0.1 -0.6 ppm: Beats weaker and more
irregular
AAA^w^nA/v^^AAAv^^AATv^
mm
X
3_
25 C, 0.6 - 0 9 ppm: Regular, protracted beats
becoming indistinct and intermittent near dealh
Figure 4. Traces of heart activity showing the combined effect of temperature (15 to 39°C) and FAC (0.1 to 1.2 ppm) on the heart rate of adult
D. serra.
Effects on Heart Rate
459
anaerobiosis could be buffered by calcium mobilized from the
calcareous shell, as occurs in S. plana (Akberali 1980, Akberali et
al. 1977). Once the valves open after 8 days and if the chlorine
dose is maintained indefinitely >0.3 ppm. eventual death would
be inevitable.
A combination of high temperatures and chlorine resulted in an
increase in the number of valve adductions and thereby an increase
in ventilation, so that heart activity was raised above levels ob-
served in solely chlorine treatments. Lacking the protection af-
forded by continuous closure of the valves, internal organs were
intermittently exposed to FAC. Below 0.3 ppm at temperatures
<25°C, D. serra would survive prolonged exposure (8 days or
more), but at higher concentrations (especially >0.6 ppm), pas-
sive shell gaping, as a result of weakened and/or paralyzed adduc-
tor muscles (Trueman and Lowe 1971), would lead to eventual
death.
Recovery after short-term exposure (±24 hours) followed by
transference to 15°C without FAC was rapid for individuals from
the range 0.1 to 1.2 ppm at 15°C and <0.6 ppm at 20 and 25°C.
In the first group, continual valve closure at high concentrations
assured protection, whereas in the second, intermittent valve open-
ing and closing prevented permanent internal damage during the
24 hour experiment During recovery in fresh sea water, valve
adductions were accompanied by an overshoot in heart rate, in-
dicative of oxygen deprivation. This is consistent with the fact that
D. serra is an oxyconformer. incurring an oxygen debt with a
decline in oxygen tensions (Van Wijk et al. 1989). Other bivalves
have shown a similar recovery response not only after exposure to
a pollutant (Akberali and Black 1980, Trueman and Akberali
1981) but also on reimmersion after aerial exposure (Trueman
1967, Coleman and Trueman 1971).
CONCLUSION
A discharge plume with FAC concentrations >0.6 ppm in
combination with temperatures >25°C for longer than 24 hours
will have a sublethal effect on D. serra from which recovery is
unlikely. In most instances, however, chlorination within the
plume is unlikely to exceed 0.3 ppm (Rattey and Potgieter 1987),
and at this concentration, D. serra was able to tolerate tempera-
tures up to 25°C for at least 24 hours. However, longer exposure
results in a slow burrowing response (Stenton-Dozey and Brown
1994a and b), so that even though heart rate may normalize at
25°C and <0.6 ppm. the danger of dislodgment from the sand near
the outfall, with possible lethal consequences, still persists. Con-
servative dosing with FAC levels at <0.3 ppm and temperatures
±20°C would result in invariable tolerance by D. serra.
ACKNOWLEDGMENT
We thank Prof. E. R. Trueman for his advice. Financial sup-
port was provided by the South African Foundation for Research
Development.
REFERENCES
Akberali. H. B. 1980. 45Calcium uptake and dissolution in the shell of
Scrobicularia plana (da Costa). J. Exp. Mar. Biol. Ecol. 43:1-9.
Akberali. H. B. & E. R. Trueman. 1979. pO, and pC02 changes in the
mantle cavity of Scrobicularia plana (Bivalvia) under normal and
stress conditions. Esiuar. Coast. Shelf Sci. 9:499-507.
Akberali, H. B.&J. E. Black. 1980. Behavioural responses of the bivalve
Scrobicularia plana (da Costa) subjected to short-term copper (Cu II)
concentrations. Mar. Emir. Res. 4:97-107.
Akberali, H B , K. R. M. Marriott & E. R. Trueman. 1977. Calcium
utilisation during anaerobiosis induced by osmotic shock in a bivalve
mollusc. Nature 256:852-853.
Bayne, B. L. 1971. Ventilation, the heart rate and oxygen uptake by
Mytilus edulis L. in declining oxygen tensions. Comp. Biochem. Phys-
iol. 40A:1065-1085.
Bayne, B. L. 1976. Marine Mussels. Their Ecology and Physiology. Cam-
bridge University Press. London.
Coleman, N. 1974. The heart rate and activity of bivalve molluscs in their
natural habitats. Oceanogr. Mar. Biol. Annu. Rev. 12:301-313.
Coleman. N. & E. R. Trueman. 1971. The effect of aerial exposure on the
activity of the mussels Mytilus edulis L. and Modiolus modiolus (L.).
J Exp. Mar. Biol. Ecol. 7:295-304.
De Zwaan. A. 1977. Anaerobic energy metabolism in bivalve molluscs.
Oceanogr. Mar. Biol. Annu. Rev. 15:103-187.
Earll. R. 1975. Temporal variation in the heart activity of Scrobicularia
plana (Da Costa) in constant and tidal conditions. J. Exp. Mar. Biol.
Ecol. 19:257-274.
Hodgson, A. N. & L. J. Fielden. 1984. The structure and responses of
Scrobicularia plana (Bivalvia: Tellanacea) to siphonal wounding. J.
Moll. Stud. 48:87-94.
Lowe, G. A. 1974. Effect of temperature change on the heart rate of
Crassostrea gigas and Mya arenaria (Bivalvia). Proc. Malac. Soc.
Lond. 41:29-36.
Lowe, G. A. & E. R. Trueman. 1972. The heart and water flow rates of
Mya arenaria (Bivalvia: Mollusca) at different metabolic levels.
Comp. Biochem. Physiol. 14A:487^f94.
Pickens, P. E. 1965 Heart rate of mussels as a function of latitude, in-
tertidal height and acclimation temperature. Physiol. Zool. 38:390-
405.
Rattey, D. & F. Potgieter. 1987. Koeberg Nuclear Power Station: Warm
Water Plume Report. ESCOM, South Africa. 7 pp. + appendix.
Stenton-Dozey, J. M. E. 1989. The physiology and energetics of the
sandy-beach bivalve Donax serra Roding with special reference to
temperature and chlorine tolerance. Ph.D. Thesis, University of Cape
Town, 318 pp.
Stenton-Dozey, J. M. E. & A. C. Brown. 1994a. Exposure of the sandy-
beach bivalve Donax serra Roding to a heated and chlorinated effluent.
I. Effects of temperature on burrowing and survival. J. Shellfish Res.
13:443^49.
Stenton-Dozey, J. M. E. & A. C. Brown. 1994b. Exposure of the sandy-
beach bivalve Donax serra Roding to a heated and chlorinated effluent.
II. Effects of chlorine on burrowing and survival. J. Shellfish Res.
13:451-454.
Trueman, E. R. 1967. Activity and heart rate of bivalve molluscs in their
natural environment. Nature 214:832-833.
Trueman, E. R. 1983. Behavioural and physiological responses of a bur-
rowing bivalve to stress, pp. 669-673. In: A. McLachlan & T. Eras-
mus (eds.). Sandy Beaches as Ecosystems. Junk Publications. The
Hague.
Trueman. E. R. & G. A. Lowe. 1971. The effect of temperature and
littoral exposure on the heart rate of a bivalve mollusc. Isognomum
alalus in tropical conditions. Comp. Biochem. Physiol. 38A:555-564.
Trueman. E. R. & H. B. Akberali. 1981. Responses of an estuarine bivalve
Scrobicularia plana (Tellinacea) to stress. Malacologia 21:15-21.
Trueman. E. R., J. G. Blatchford. H. D. Jones & G. A. Lowe. 1973.
Recordings of the heart rate and activity of molluscs in the natural
habitat. Malacologia 14:377-383.
Van Wijk, K., J. M. E. Stenton-Dozey & A. C. Brown. 1989. Oxycon-
formation in the burrowing bivalve Donax serra Roding. J. Moll. Stud.
55:133-134.
Widdows, J. 1973. Effect of temperature and food on the heart beat,
ventilation rate and oxygen uptake of Mytilus edulis. Mar. Biol. 20:
269-276.
Journal of Shellfish Research, Vol. 13, No. 2,461-465. 1994.
THE IN VITRO LIFE CYCLE OF A PERKINSUS SPECIES (APICOMPLEXA, PERKINSIDAE)
ISOLATED FROM MACOMA BALTHICA (LINNEAUS, 1758)
S. J. KLEINSCHUSTER,1 F. O. PERKINS,2 M. J. DYKSTRA3 AND
S. L. SWINK'
Haskin Shellfish Research Laboratory
Rutgers University
RD#1, Box B-8
Port N orris, New Jersey 08349
School of Marine Science
Virginia Institute of Marine Science
College of William and Mary
PO Box 1346
Gloucester Point, Virginia 23062
Department of Microbiology, Pathology & Parasitology
North Carolina State University
4700 Hillsborough St.
Raleigh, North Carolina 27606
ABSTRACT Using standard sterile techniques and a single medium previously described (Kleinschuster and Swink 1993), the in
vitro culture of a Perkinsus species isolated from Macoma halthka was possible. Zoosporulation. the release of zoospores, and the
reestablishment of secondary cultures from the zoospores completed an in vitro life cycle.
KEY WORDS: Perkinsus species. Macoma balthica, cell culture
INTRODUCTION
Species of the apicomplexan protozoan Perkinsus have been
reported to cause major mortalities in bivalve populations and
cross-transmission of Perkinsus species infections between certain
bivalve species is possible (Goggin et al. 1989). Although a simple
in vitro culture technique for the propagation of Perkinsus marinus
Levine 1978 has been established (Kleinschuster and Swink
1993). techniques for the routine induction of zoosporulation and
release of zoospores in quantity have not been available in recent
years either in vivo or in vitro. Consequently, the role of the
zoospore in the infection/invasion process has not been fully in-
vestigated.
We report herein simple methodology for the in vitro culture of
a Perkinsus species isolated from Macoma balthica including veg-
etative propagation of the isolate, zoosporulation, and release of
zoospores followed by the reestablishment of secondary cultures
from the zoospores, thereby completing an in vitro life cycle.
MATERIALS AND METHODS
Infected clams were obtained from King's Creek, ( 14 ppt salin-
ity), a tributary of the York River in Virginia, and maintained in
aquaria at the Virginia Institute of Marine Science. Clams used as
a source of Perkinsus species were transported to the Haskin Shell-
fish Research Lab and maintained in recirculating sea water at 12
to 15°C. Cells of the parasite were obtained from hemolymph
aspirated from the blood sinuses of the adductor muscles. Imme-
diately after aspiration, hemolymph and cellular components were
transferred to 25 cm2 T-flasks containing 1 to 2 ml of sterile sea
water and appropriate antimicrobics (streptomycin, 0.2 mg ml" ',
penicillin, 200 U ml-1, and amphotericin B 0.25 u-g ml~') and
held at room temperature for 3 hours. After this treatment, the
isolation medium was replaced by a medium used for the in vitro
culture of P. marinus (Kleinschuster and Swink 1993). and the
cultures were kept at 28°C under ambient air. The medium was
exchanged (50%) every 3 days. Phase contrast microscopy was
used to obtain photomicrographs.
Routine roller bottle technique and subculturing were used to
upscale the production of cultured cells. Aliquots of these cultures
were centrifuged at low speed to pellet the organisms, which were
then washed two times with sterile sea water and resuspended in
T-flasks with sterile sea water to induce zoospore formation.
For ultrastructural analysis, prezoosporangia. zoosporangia,
and zoospores in sterile sea water were sent to N.C. State Uni-
versity in T-flasks. Contents of the flasks were decanted into a
centrifuge tube, and after the bottom of the flask was scraped with
a rubber policeman, the cells were pelleted at 1000 x g for 5
minutes. The medium was gently pipetted from the tube and re-
placed with 3% glutaraldehyde in 0. 1 M sodium cacodylate buffer,
pH 7.8 (Azevedo 1989). After 1 hour of primary fixation at room
temperature, the cells were rinsed in fresh buffer several times.
The cells were pelleted after each rinse. After the final rinse, the
cells were pelleted, the medium was removed, and the cells were
resuspended in molten (approximately 50°C) 4% water agar and
quickly centrifuged at 1000 x g for 30 seconds. Once the agar had
solidified, the agar containing the cells was sliced into 1 mm thin
slices with a razor blade and placed into 1% osmium tetroxide in
the same buffer at room temperature. After 1 hour, the samples
were rinsed three times in distilled water and subsequently dehy-
drated in a graded ethanol series, passed through 100% acetone,
and infiltrated with Spun resin (Dykstra 1993). Ultrathin sections
were obtained, stained with methanolic uranyl acetate and lead
citrate, and evaluated with a transmission electron microscope
(TEM).
RESULTS
Low-density in vitro cultures of the Perkinsus isolate are rep-
resented in Figures 1 and 2. In nutrient-rich medium, vegetative
461
462
Kleinschuster et al.
Figures 1 and 2. Photomicrographs of low-density in vitro cultures of a Perkinsus species isolated from M. balthica. Note schizonts typical of
Perkinsus species (arrowheads) in nutrient-enriched medium and the discharge tubes of zoosporangia present in sea water cultures (arrows).
Scale bar, 0.05 mm (Fig. I) and 0.1 mm (Fig. 2).
reproduction was vigorous and typically consisted of schizonts in
various stages of merogeny; meronts and merozoites were evident
(Fig. 1). Generally, vegetative stages of this isolate were larger
and more ovoid than P. marinus under similar conditions; how-
ever, merogeny appeared to be similar. A typical low-density sea
water culture is represented by Figure 2. Various presporulation
stages were evident, including prezoosporangia and zoosporangia
with discharge tubes.
Several developmental stages of induced zoosporulation are
seen in Figures 3 to 7. Typical two-, four-, and eight-cell stages
of developments are seen in Figures 3 to 5. respectively. Dis-
charge tubes were routinely evident at the two-cell stage (Fig. 3),
Figures 3 to 6. Photomicrographs of a low-density sea water culture of a Perkinsus species isolated from M. balthica. Represented are typical
two-, four-, and eight-cell stages of zoosporulation (arrows). Figure 6 exhibits a sporangium with two discharges tubes (arrows). Blurred
background images are motile zoospores. Scale bar. 0.1 mm.
In Vitro Culture of a Perkinsus Species from Macoma balthica
463
464
Kleinschuster et al.
and multiple discharge tubes were not uncommon (Fig. 6). Suc-
cessive karyokineses and cytokineses resulted in the formation and
release of motile zoospores (Fig. 7). The blurred and dotted back-
grounds of Figures 6 and 7 represent discharged and motile zoo-
spores.
A representative TEM photomicrograph of a mature prezoo-
sporangium with typical thickened cell wall, discharge tube, and
multiple prezoospores is seen in Figure 8.
A TEM photomicrograph of a zoospore displaying apicom-
plexan structures, including rhoptries, a conoid, and subpellicular
microtubular and microneme-like organelles is seen in Figure 9.
The kinetosome substructure was the same as in P. marinus (Per-
Figurc 7. Photomicrographs of a low-density sea water culture of Perkinsus species isolated from At. balthica. Notice discharge of zoospores from
/.oosporangium (arrow). Blurred background images are motile zoospores. Scale bar, 0.1 mm.
Figure 8. Zoosporangium of a Perkinsus species containing zoospores (arrows). Note the thick zoosporangial wall and the discharge tube and
plug of wall material derived from the inner layer of the zoosporangial wall (arrowhead). Original magnification, x2,240.
Figure 9. Zoospore of a Perkinsus species showing apicomplexan structure: Conoid (C), rhoptries (R), subpellicular microtubules (arrowheads),
microneme-like organelles (mn), flagella (F), and mitochondrion (m). Original magnification, X22.800.
In Vitro Culture of a Perkinsus Species from Macoma balthica
465
kins 1988). Whole mounts of zoospores negatively stained in
0.5% aqueous uranyl acetate exhibited filamentous mastigonemes
and spur-like structures identical to those of P. marinus (Perkins
1991).
In general, this Perkinsus species exhibited more rigorous
growth and reproduction than P. marinus under similar culture
conditions and had a very short doubling time (approximately 20
hours in log phase). Additionally, sea water-induced zoospores
were readily returned to the vegetative-propagation state by sub-
stitution of the sea water with nutrient-enriched medium and cul-
tured as described above.
DISCUSSION
Isolation and in vitro propagation of a vigorous Perkinsus spe-
cies will facilitate studies of the basic biology of this parasite. The
ability to induce zoospores and their subsequent release directly
from vegetative cultures without the use of the Ray thioglycolate
technique (1952) together with the reestablishment of vegetative
cultures from zoospores may provide an impetus toward our un-
derstanding of this organism's parasitic profile through experimen-
tal manipulation. Although species identification was not an ob-
jective of this study, in consideration of the morphologial charac-
teristics of this isolate as seen in the host, as well as the possibility
of cross-species infection, it is suggested that this organism may
be Perkinsus atlanticus, which has been described from Ruditapes
decussatres, a Portuguese clam (Azevedo 1989). Because of the
ease of culturing this parasite through an in vitro life cycle, it may
be appropriate to develop this system as a model for the study of
Perkinsus species/host interactions.
ACKNOWLEDGMENTS
This work is identified as Hatch Project No. 32100 and is
identified as paper No. D-32100-6-94 by the New Jersey Agricul-
tural Experiment Station and paper No. 94-25 by the Institute of
Marine and Coastal Sciences, Rutgers University; contribution
No. 1901 of the School of Marine Science and Virginia Institute of
Marine Science, College of William and Mary.
REFERENCES
Azevedo, C. 1989. Fine Structure of Perkinsus attenticus n. sp. (Apicom-
plexa, Perkinsea) parasite of the clam Ruditapes decussatus from Por-
tugal. J. Parasitol. 75:627-635.
Dykstra, M. J. 1993. A Manual of Applied Techniques for Biological
Electron Microscopy. Plenum Press, New York.
Goggin, C L.. K. B. Sewell & R. T. G. Lester. 1989. Cross-infection
experiments with Australian Perkinsus species. Dis. Aquatic Organ.
7:55-59.
Kleinschuster. S. J. & S. L. Swink. 1993. A simple method for the in
vitro culture of Perkinsus marinus. The Nautilus 107:76-78.
Perkins, F. O. 1988. Structure of protistan parasites found in bivalve mol-
luscs, pp. 93-111. In: W. S. Fisher (ed.). Disease processes in marine
bivalve molluscs. American Fisheries Society Special Publication 18,
Bethesda, MD.
Perkins, F. O. 1991. Sporozoa. pp. 261-331. In: F. W. Harrison and
J O. Corliss (eds.). Microscopic Anatomy of Invertebrates, vol. I.
Protozoa. Wiley-Liss. New York.
Ray, S. M. 1952. A culture technique for the diagnoses of infection with
Dermocystidium marinum. Mackin. Owen and Collier in oysters. Sci-
ence 116:360-361.
Journal of Shellfish Research. Vol. 1?, No. 2, 467-472, 1994.
A COMPARATIVE STUDY OF THE GAMETOGENIC CYCLES OF THE CLAMS TAPES
PHILWP1NARUM (A. ADAMS & REEVE 1850) AND TAPES DECUSSATUS (LINNAEUS) ON THE
SOUTH COAST OF IRELAND
QIUSHI XIE AND GAVIN M. BURNELL
Department of Zoology
University1 College Cork
Cork, Ireland
ABSTRACT A comparative study was carried out on the gametogenic cycles of the introduced Manila clam Tapes phitippinarum and
the native carpet-shell clam Tapes decussatus during the period from September 1990 to September 1991 . Examination of histological
preparations showed that obvious gonadal development began in March for the Manila clam, and in April for the carpet-shell clam.
Histological evidence showed that, in the Manila clam, ripe specimens appeared in May with major spawning starting in September
Spawning of this exotic species was not previously recorded in Ireland. In the carpet-shell clam, first ripe specimens were observed
in June and major spawning commenced in August. Both species had a unimodal gametogenic cycle in southern Irish waters.
Measurement of oocyte size was facilitated by application of an image analysis system, and the resultant data were used to determine
the various stages of gonadal development. Mean diameter of free oocytes was 43.2 u.m for the Manila clam and 42.6 u.m for the
carpet-shell clam.
KEY WORDS: gametogenic cycle, oocyte size, image analysis. Tapes philippinarum. Tapes decussatus
INTRODUCTION
Manila clams. Tapes philippinarum, are indigenous to the
Indo-Pacific region. They were introduced into Ireland in 1982
because of faster growth rate and the ability to tolerate a wider
range of environmental conditions than the native carpet-shell
clam Tapes decussatus. Despite their commercial importance,
there are no published records on the gametogenic cycles of these
two species in Irish waters. It is well documented that gametogenic
cycles in marine invertebrates are influenced by exogenous as well
as endogenous factors (Giese 1959, Sastry 1975, Sastry 1979).
The most important exogenous factor, temperature, is closely as-
sociated with geographic location. Studies carried out in Japan,
North America, and Europe have shown geographic variations in
the gametogenic cycles of these two species (Ohba 1959, Holland
and Chew 1974, Breber 1980, Beninger and Lucas 1984, Sa-
rasquete et al. 1990, Shafee and Daoudi 1991). It is therefore
likely that the gametogenic cycles of the Manila clam and the
carpet-shell clam in Ireland are somewhat different from those in
other areas. It was observed that the carpet-shell clam occurs only
sporadically along the Irish coast (Partridge 1977a, Partridge
1977b). This may be attributed to biological as well as hydro-
graphic factors; for example, abnormality in gametogenesis. asyn-
chronous spawning, poor fecundity, and disruption of metamor-
phosis may all contribute to the patchy distribution of the carpet-
shell clam in Irish water. The main objective of this study was to
compare gametogenic cycles of these two species in Ireland and
correlate spawning time with environmental conditions in Irish
waters.
MATERIALS AND METHODS
Approximately 9 month old juveniles of both species were
purchased from a commercial hatchery in May 1990. They were
held in trays ( 1 mm mesh) at a density of 5 ,000 m ~ 2 over a muddy
bed in a sheltered inlet at the Atlantic Shellfish Ltd. in Cork
Harbour, in the south of Ireland (51°50'N, 8°17'W). Average air
exposure was approximately 15%. Density was reduced to 2,500
m~2, and mesh size was increased to 6 mm in July 1990 to
facilitate water flow and reduce siltation.
Water temperature and salinity were measured during high tide
about every 14 days between May 1990 and March 1992. From
September 1990 to September 1991 , 12 to 20 clams of each spe-
cies were used for histological examination of gonadal develop-
ment and gametogenesis. The visceral mass was fixed in Helly's
fixative for 12 to 24 hours. Tissue was then rinsed in running tap
water to remove excessive fixative before being stored in 70%
ethyl alcohol. Whole visceral mass or, for large specimens, the
portion between the labial palps and the foot, was dehydrated in
ethyl alcohol and then embedded in paraffin wax. Cross sections
of 7 u.m were then made in the middle region of the visceral mass.
These thin sections were stained with Harris's hemotoxylin and
eosin and were examined under a light microscope to determine
the stages of the gametogenic cycle.
Changes in cytological characteristics were used to categorize
specimens into one of six arbitrary gonadal stages. The five stages
used by Holland and Chew (1974) for the Manila clam were
adopted, but the criteria were modified, and one extra stage, the
inactive stage, was added for sexually undifferentiated specimens
(Mann 1979a). The criteria used were as follows.
Inactive
Gonad was predominantly composed of connective tissue; sex
was not distinguishable.
Female
Early Active
Gonad proliferation had started, as seen by increasing numbers
of discernible oocytes; oocytes were still small; no free oocytes
were present in the lumen; mean oocyte diameter was <20 u.m; no
oocytes were >30 u,m; follicle boundaries were not easy to dis-
tinguish.
Late Active
Connective tissue was decreasing; free oocytes were present in
the lumen but accounted for less than half of the total oocytes in
467
468
XlE AND BURNELL
the follicles; many young oocytes of various sizes were attached to
the follicle wall; the mean oocyte diameter was between 20 and 35
u.m; more than half of the oocytes had a diameter greater than 20
u,m; most of the follicle walls were thick; individual follicles were
relatively small but easily discernible.
Ripe
Half or more than half of the oocytes were free in the lumen
and had a polygonal configuration; half or more than half of the
oocytes had a diameter of 3=35 |xm and the mean oocyte diameter
was 3=35 |xm; follicle size increased; follicle wall was thin.
Partially Spent
Number of free oocytes per follicle had declined; empty folli-
cles appeared; some oocytes had undergone cytolysis.
Spent
Half or more than half of the follicles were empty; follicles
became shrunk, fused, or scattered; follicle wall was broken; few
residual free oocytes remained; in some specimens, only oogonia
and small oocytes remained among the connective tissue and
phagocytes.
Male
Early Active
Gonad proliferation started; spermatogonia and spermatocytes
were present in the follicles; in more developed specimens, sper-
matids were also found; no spermatozoa were present.
Late Active
Spermatogonia, spermatocytes, spermatids, and spermatozoa
coexisted in the follicles; in less developed specimens, there was
no dominant cell type, but in more developed specimens, sperma-
tids and spermatozoa were the major cell types; spermatozoa
formed centric or elongate bands in the follicles; spermatozoa
mass had a radius of less than half of that of the follicle.
Ripe
Gonad was mainly composed of mature spermatozoa, which
formed centric or elongate bands or "plugs'" in the follicles; the
spermatozoa mass had a radius of more than half of that of the
follicle; in very ripe specimens, the spermatozoa bands were close
to the follicle wall.
oocytes per specimen for May to September 1991 samples as the
variance of oocyte size increased. Video images of gonads were
digitized and displayed on a color monitor. The boundaries of
oocytes were defined by hand, and the areas were then calculated
automatically by the computer. Calibration was made with a stage
micrometer. For each field, all of the oocytes were sectioned
through the nucleus, or for small oocytes, those with nucleolus
clearly visible were measured. Several sequential fields were re-
quired to achieve the predefined number of oocytes. The "spent"
specimens had few measurable oocytes and therefore were ex-
cluded from measurement (Grant and Tyler 1983). The area of
each oocyte was then converted to the diameter of a circle with
equal area. Measuring oocyte size by the image analysis system is
considered to be more accurate and less tedious than the traditional
method, which used an eyepiece graticule to measure the long axis
only (Barber and Blake 1983) or both long and short axes (Brown
1984). These traditional methods are likely to overestimate oocyte
size, particularly the near-ripe oocytes with slender stalk. The
mean oocyte size and the oocyte size distribution were used in
some of the staging criteria to minimize the subjectiveness of a
qualitative description.
The size of oocytes lying free in the lumen was measured by
the same image analysis system but with a different technique:
oocytes were highlighted and only those sectioned through the
nucleus were selected for measurement. For each species, 50 free
oocytes from each of the 10 randomly selected ripe specimens (2
from June, 4 from July, and 4 from August) were measured.
RESULTS
Data for surface and depth samples were averaged, and the
seasonal variations in temperature and salinity are illustrated in
Figure 1 . Figure 1 shows a cyclical change in temperature with a
minimum temperature of 5.6°C in December 1990 and a maximum
temperature of 20.2°C in July 1990. Temperature rose to above
8CC, a reported lower temperature limit for gonad activation in the
Manila clam (Mann 1979b. Bourne 1982), at the end of February.
Each year, there were over 4 months when the temperature was
above 14°C. the reported lower temperature limit for spawning in
the Manila clam (Mann 1979b). Salinity fluctuated between 24.6
to 34.9%c. Salinity was generally lower during winter and spring
than during the summer months.
No hermaphrodites were found in this study in either species.
Of the 248 Manila clam specimens examined, 34 were sexually
undifferentiated, 109 were female, and 105 were male (Table 1).
Of the 250 carpet-shell clam specimens examined. 99 were sexu-
ally undifferentiated, 75 were female, and 76 were male in the
Partially Spent
Mature spermatozoa started to release; an empty space ap-
peared in the center in over 20% of the follicles.
Spent
Follicles were shrunk, fused, scattered, and disorganized; sper-
matozoa mass occupied approximately 20% or less of the follicle
space; in completely spent specimens, only residual spermatozoa
could be found in some follicles; connective tissue and phagocytes
became increasingly prominent.
For the March and April 1991 samples, 50 oocytes per speci-
men were measured by the use of Arclmage 5 image analysis
software on an ACORN computer. This was increased to 100
o 15
M J J
1990
A S O N
D J F M
1991
D J
F M A
992
36
34
32
-30
-28
-26
24
AMJJASON
Month
Figure 1. Seasonal variation in sea temperature and salinity at the
sampling site (51°51'N, 8°71'W).
Gametogenic Cycles of Clams
469
TABLE 1.
Seasonal changes in frequency of various gonadal stages in T. philippinarum in the period from September 1990 to September 1991.
Male"
Female"
Date
Examined
EA
LA
R
PS
SP
Total
EA
LA
R
PS
SP
Total
Inactive
20 Sept. 90
12
0
2
2
0
0
4
0
0
6
0
1
7
1
19 Oct. 90
18
0
0
1
7
1
9
0
0
0
1
8
9
0
4 Nov. 90
20
0
0
0
1
4
5
0
0
1
4
10
15
0
4 Dec. 90
20
0
0
0
0
3
3
0
0
0
0
10
10
7
31 Jan. 91
20
0
0
0
0
4
4
0
0
0
0
6
6
10
14 Feb. 91
20
0
0
0
0
3
3
0
0
0
0
8
8
9
19 Mar. 91
20
9
1
0
0
0
10
5
0
0
0
0
5
5
15 Apr. 91
19
4
6
0
0
0
10
4
4
0
0
0
8
1
14 May 91
20
0
10
0
0
0
10
1
2
6
0
0
9
1
26 June 91
20
0
7
6
0
0
13
0
1
6
0
0
7
0
15 July 91
20
0
0
12
0
0
12
0
0
8
0
0
8
0
13 Aug. 91
19
0
0
8
0
0
8
0
0
11
0
0
11
0
10 Sept. 91
20
0
0
11
3
0
14
0
0
2
2
2
6
0
■ EA, early active; LA, late active; R. ripe; PS, partially spent; SP, spent.
same period (Table 2). The sex ratio was not significantly different
from a 1:1 ratio for either species (p > 0.05, chi-square test, Zar
1984). The seasonal changes in the frequency of various gonadal
stages for the two species are listed in Table 1 for the Manila clam
and in Table 2 for the carpet-shell clam.
Manila Clams
Gametogenesis began in March. This was seen by the increas-
ing numbers of discernible oogonia and small oocytes in females
and the occurrence of spermatocytes or even spermatids in males.
Active gamete proliferation took place in April and May. In April.
10 of 19 samples were in the late active stage. The first ripe
specimens appeared in May. and all specimens were in the ripe
stage in July and August (Table 1). Although minor release of
spermatozoa from more developed follicles might occur as early as
May. mature oocytes appeared to be retained in the follicles
whereas young oocytes grew to full size. This was reflected by the
declining percentage of oocytes of <35 p.m between May and
August (Fig. 2a). From March to August, mean oocyte size stead-
ily increased (Fig. 3) as a result of oocyte growth and the associ-
ated decline in the number of young oocytes. During vitellogen-
esis, both follicle size and number of mature oocytes per follicle
appeared to increase significantly. In August, very few young
oocytes were present. As the pressure in the follicles increased, the
mature oocytes acquired a polygonal configuration. Major spawn-
ing did not commence until September, when partially spent and
spent specimens were found. Development of a new batch of
oocytes was observed in some specimens in this month; however,
these young oocytes failed to grow because water temperature was
decreasing. Mean oocyte size decreased in September as a result of
spawning and redevelopment (Fig. 3). Most of the specimens were
either in the partially spent stage or in the spent stage in October.
Although occasional ripe specimens could still be found in No-
vember, most of the clams were in the spent stage by this time.
Beginning in December, the gonads entered a resting period dur-
ing which there was no sign of gonadal activity. This resting
period ended in March when gametogenesis started once again.
TABLE 2.
Seasonal changes in frequency of various gonadal stages in T. decussatus in the period from September 1990 to September 1991.
No. of Clams
Examined
Male"
Female"
Date
EA
LA
R
PS
SP
Total
EA
LA
R
PS
SP
Total
Inactive
20 Sept. 90
14
0
0
0
1
5
6
0
0
0
1
6
7
1
19 Oct. 90
19
0
0
0
0
5
5
0
0
0
0
3
3
11
4 Nov. 90
19
0
0
0
0
4
4
0
0
0
0
2
2
13
4 Dec. 90
20
0
0
0
0
0
0
0
0
0
0
1
1
19
31 Jan. 91
20
0
0
0
0
0
0
0
0
0
0
0
0
20
14 Feb. 91
20
0
0
0
0
0
0
0
0
0
0
0
0
20
19 Mar. 91
20
0
0
0
0
0
0
12
0
0
0
0
12
8
15 Apr. 91
20
7
0
0
0
0
7
6
0
0
0
0
6
7
14 May 91
20
5
3
0
0
0
8
3
9
0
0
0
12
0
26 June 91
19
0
7
3
0
0
10
0
7
2
0
0
9
0
15 July 91
20
0
4
10
0
0
14
0
1
5
0
0
6
0
13 Aug. 91
19
0
0
10
1
0
11
0
0
4
3
1
8
0
10 Sept. 91
20
0
0
1
7
3
11
0
0
0
8
1
9
20
EA, early active; LA, late active; R, ripe; PS, partially spent; SP, spent.
470
XlE AND BURNELL
philippinarum
□ < 15 urn
□ 15-25 urn
H 25-35 u.m
■ > 35 u.m
Sept /9 1
T. decussatus
0 < 1 5 u.m
□ 15-25u.m
B 25-35 (im
I > 35 u.m
Sept / 9 1
Figure 2. Changes in the relative frequency distribution ( % ) of oocyte
size in the period from March 1991 to September 1991 for T. philip-
pinarum (a) and T. decussatus (b). N = 250 to 1,200.
Carpet-Shell Clams
Although female gonads showed early signs of gonadal activity
in March, male gonads were still in the resting stage (Table 2).
Obvious gonadal development was not observed for either sex
until April. The first late active specimens were observed in May,
and the first ripe clams were seen in June. The highest frequency
of ripe stages was observed in July and August. Major spawning
activity began in August when 1 of 1 1 of males was partially
spent, 3 of 8 of females were partially spent, and 1 of 8 of females
was spent. Spawning appeared to be completed by the end of
September or early October. There was a higher percentage of
oocytes <35 p.m in the follicles each month between May and
August in the carpet-shell clam than in the Manila clam (Fig. 2a
and b|. Follicle size and number of mature oocytes per follicle also
appeared to be smaller than those of the Manila clam. Mean oocyte
size steadily increased from March to July and remained more or
less unchanged in August and September (Fig. 3). Gonadal rede-
velopment may also have occurred in August and September; how-
ever, the persistent presence of small oocytes made it difficult to
state with confidence whether it truly happened. After spawning
was completed, residual gametes were absorbed and the gonad of
the carpet-shell clam entered a resting period beginning in Sep-
tember (Table 2).
The size of free oocytes from morphologically ripe females
ranged between 20.2 and 58.7 u.m for the Manila clam and be-
tween 23.0 and 56.3 u,m for the carpet-shell clam. The mean
diameter of 500 free oocytes was 43.2 ± 6.2 u.m (standard devi-
ation [SD]) for the Manila clam and 42.6 ± 6.1 u.m (SD) for the
carpet-shell clam. There was no significant difference in mean
oocyte diameter between the two species when all of the data were
pooled (N = 500) (p > 0.05, /-test, Zar 1984). However, a
statistically significant difference existed among individuals of the
same species sampled in the same month (p < 0.05, analysis of
variance, Zar 1984) in July and August for the Manila clam and in
August for the carpet-shell clam. Most of the free oocytes (78.6%
for the Manila clam and 77.8% for the carpet-shell clam) had a
diameter between 35 and 50 |j,m. Oocytes that were larger than 50
u.m only accounted for 13.0% of the total number in the Manila
clam and 9.6% in the carpet-shell clam (Fig. 4).
DISCUSSION
It is well established that temperature is one of the most im-
portant environmental factors that influence the gametogenic cy-
cles of molluscs (Giese 1959, Mann 1979a and b, Sastry 1975).
The effect of temperature on the gametogenic cycles of T. philip-
pinarum and T. decussatus was also evident when these results
were compared with those obtained from other areas.
The Manila clam has been reported to have one spawning pe-
riod in Washington state (Holland and Chew 1974). two spawning
periods in Japan (Ohba 1959), and three spawning periods in
southwest Spain (Sarasquete et al. 1990). From these records and
others (Mann 1979b), it appeared that, for the Manila clam, the
lower temperature limit for gonadal activity is approximately 8°C,
12°C for gamete ripening, and 14°C for spawning. Low temper-
ature affects gametogenesis as well as spawning (Mann 1979b,
Bourne 1982). In the south of Ireland, the temperature require-
bU -
„ i
t
40 -
J^-^""
30 -
20 -
- T decussatus
10 -
0 ■
March
April
May
July
August September/91
June
Month
Figure 3. Changes in mean oocyte size in the period from March 1991
to September 1991 for T. philippinarum and T. decussatus, with ver-
tical bars representing SD. N = 250 to 1 ,200.
32 5 37 5 42 5 47 5
Oocyte diameter (u.m)
Figure 4. The relative frequency distribution ( % ) of free oocyte size
(N = 500) of T. philippinarum and T. decussatus.
Gametogenic Cycles of Clams
471
merits for various stages of the gametogenic cycle were generally
met (Fig. 1). Histological evidence obtained in this study showed
that most Manila clams completed their gametogenic cycle. How-
ever, the relatively low temperature encountered by this popula-
tion in the south of Ireland seemed to delay gonadal development
and spawning when compared with populations in warmer cli-
mates (Ohba 1959, Sarasquete et al. 1990). This time lag could be
explained by a time-temperature effect (Mann 1979b). Neverthe-
less, these results were generally consistent with findings from
areas with similar temperature ranges (Beninger and Lucas 1984,
Coleman 1989). Although there is little doubt that the temperature
in Irish waters is sufficient for adult Manila clams to reach the ripe
stage, questions have been raised as to whether the ripe gametes
were actually released by spawning or resorbed (Coleman 1989).
This study showed that, in September 1991, 4 of 6 females were
either partially spent or spent and 3 of 1 1 males were partially
spent. Temperature at that time was =18°C, well above the lower
temperature limit (14°C) for spawning. This would therefore ap-
pear to be the first recorded case of the Manila clam spawning in
Irish waters. There were approximately 2 more weeks before the
water temperature decreased below the 14°C limit. During this
period, further spawning could have occurred. Although only 1 of
12 clams was in the spent stage on 20 September 1990, most
of specimens were in the partially spent or spent stages on 19
October (Table 1). Between 20 September 1990 and 19 October
1990, water temperature was above 14°C for approximately 7
days, during which spawning could have taken place. Resorption
of these ripe gametes during this period was not very likely be-
cause no massive cytolysis was observed. One of the 20 specimens
was in the ripe stage in November, the rest were either in the
partially spent or spent stage. Mature gametes remaining in the
follicles in this month were believed to be resorbed because the
temperature was too low to stimulate spawning. The delay in
spawning, which appeared to occur in the apparently ripe clams
sampled in the summer of 1991. suggested a degree of synchro-
nization in spawning activity among individuals in the population.
Since its introduction to Ireland in 1982. natural recruitment of
the Manila clam has not been recorded. Late spawning is believed
to be one of the major reasons for no recruitment because the water
temperature in the months after spawning would not be conducive
for larval development.
These results of the gametogenic cycle of the carpet-shell clam
showed that this species had only one spawning period. Gamcto-
genesis began in late spring (March/April), and the first ripe clams
appeared in June. Spawning commenced in August and ceased in
September. These findings compared well with previous observa-
tions of this species from temperate regions (Breber 1980.
Beninger and Lucas 1984). In warmer years, gonadal development
and spawning began earlier in the year, as observed by Partridge
( 1977b) in populations from the west of Ireland. The timing of the
major events in the reproductive cycle was also clearly associated
with latitude, with clams from more southerly waters of the north-
cm hemisphere tending to reach the ripe stage and spawn earlier in
the year. For example, spawning started in early May in the At-
lantic coast of Morocco and there were two major spawning pe-
riods instead of one (Shafee and Daoudi 1991).
Although there is natural recruitment, the carpet-shell clams
are believed to be less successful in Irish waters than in southern
European waters (Partridge 1977a and b). Examination of the his-
tological preparations alone did not provide conclusive evidence to
suggest that gametogenesis in the carpet-shell clam was abnormal.
However, visual assessment of the gonads of the two species
showed that follicle size, number of free oocytes per follicle, and
overall density of free oocytes appeared to be markedly smaller in
the carpet-shell clam than in the Manila clam. Whether this is an
interspecific difference or a sign of poor fecundity is not clear. The
late spawning of both species in Ireland, as observed in this study.
is believed to reduce the chance of larval survival, which is opti-
mal at 23 to 26°C for T. philippinarum (Helm and Pellizzato 1990)
and at 25°C for T. decussatus (Partridge 1977b).
The wide range of free oocyte size is accounted for by the
presence of some small oocytes and also by the fact that some
specimens had significantly different mean free oocyte size. There
is no direct evidence that the small oocytes (<35 u.m) were fully
mature. An artificial fertilization is required to test if these small
oocytes are in fact physiologically ripe.
ACKNOWLEDGMENTS
We express our appreciation to the Atlantic Shellfish Ltd. for
their help throughout the sampling period. We are also very grate-
ful to Dr. Eamonn Twomey for his guidance with image analysis
and critical comments for the manuscript.
LITERATURE CITED
Barber, B. J. & N. J. Blake. 1983. Growth and reproduction of the bay
scallop, Argopecten irradians (Lamarck) at its southern distributional
limit. J. Exp. Mar. Biol. Ecol. 66:247-256.
Beninger, P. G. & A. Lucas. 1984. Seasonal variations in condition,
reproductive activity, and gross biochemical composition of two spe-
cies of adult clam reared in a common habitat: Tapes decussatus L.
(Jeffreys) and Tapes philippinarum (Adams & Reeve). J. Exp. Mar.
Biol. Ecol. 79:19-37.
Bourne, N. 1982. Distribution, reproduction, and growth of Manila clam.
Tapes philippinarum (Adams and Reeves), in British Columbia. J.
Shellfish Res. 2:47-54.
Breber, P. 1980. Annual gonadal cycle in the carpet shell clam Venerupis
decussata in Venice, Italy. Proc. Nail. Shellfish Assoc. 70:31-35.
Brown, R. A. 1984. Geographical variations in the reproduction of the
horse mussel. Modiolus modiolus (Mollusca: Bivalvia). J. Mar. Biol.
Assoc. U.K. 64:751-770.
Coleman, E. M. 1989. An investigation into the performance of the Ma-
nila clam Tapes semidecussalus (Reeve) in Drumcliff Bay Co. Sligo.
MSc Thesis, National University of Ireland. 71 pp.
Giese, A. C. 1959. Comparative physiology: annual reproductive cycles
of marine invertebrates. Annu. Rev. Physiol. 21:547-576.
Grant, A. & P. A. Tyler. 1983. The analysis of data in studies of inver-
tebrate reproduction. I. Introduction and statistical analysis of gonad
indices and maturity indices. Int. J . Invert. Reprod. 6:259-269.
Helm. M. M. & M. Pellizzato. 1990. Hatchery breeding and rearing of the
Tapes philippinarum species. In: Giuseppe Alessandra (ed.). Tapes
philippinarum Biologia e Sperimentazione. Ente Sviluppo Agricolo
Veneto (ESAV) Press.
Holland. D. A & K. K Chew. 1974. Reproductive cycle of the Manila
clam (Venerupis japonica). from Hood Canal. Washington. Proc. Nat.
Shellfish Ass. 64:53-58.
Mann. R. 1979a. Some biochemical and physiological aspects of growth
and gametogenesis in Crassostrea gigas and Ostrea edulis grown at
sustained elevated temperatures. J. Mar. Biol. Assoc. U.K. 59:95-1 10.
Mann, R. 1979b. The effect of temperature on growth, physiology, and
gametogenesis in the Manila clam Tapes philippinarum (Adams &
Reeve, 1850). J. Exp. Mar. Biol. Ecol. 38:121-133.
Ohba, S. 1959. Ecological studies in the natural population of a clam.
472
XlE AND BURNELL
Tapes japonica. with special reference to seasonal variations in the size
and structure of the population and to individual growth. Biol. J.
Okayama Univ. 5:13—47.
Partridge, J. K. 1977a. Littoral and benthic investigations on the west
coast of Ireland. IV. (Section A: Faunistic and ecological studies),
annotated bibliographies of the genus Tapes (Bivalvia: Veneridael: Part
l-Tapes decussatus (L. ). Part U-Tapes semidecussatus Reeve. Pro. R.
Ir.Acad.. Seel. B 77:1-63.
Partridge, J. K. 1977b. Studies on Tapes decussatus in Ireland. PhD the-
sis. National University of Ireland. 414 pp.
Sarasquete, M. C. S. Gimeno & M. L. Gonzalez de Canales. 1990. Cy-
cle reproducteur de la palourde Ruditapes philippinarum (Adams &
Reeve, 1850) de la cote sud ouest atlantique (Espagne). Rev. Int.
Oceanogr. Med. 97-98:90-99.
Sastry, A. N. 1975. Physiology and ecology of reproduction in marine
invertebrates, pp. 279-299. In: F. J. Vernberg (ed). Physiological
Ecology of Esturine Organisms. University of South Carolina Press,
Columbia, South Carolina.
Sastry, A. N. 1979. Pelecypoda (excluding Ostreidae). pp. 113-292. In:
A. C Giese and J. S. Pearse (eds.). Reproduction of Marine Inverte-
brates. Academic Press, New York.
Shafee, M. S. & M. Daoudi. 1991. Gametogenesis and spawning in the
carpet-shell clam, Ruditapes decussatus (L.) (Mollusca: Bivalvia),
from the Atlantic coast of Morocco., Aquacul. Fish. Manage. 22:203-
216.
Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall International, New
Jersey, 718 pp.
Journal of Shellfish Research, Vol. 13, No. 2. 473-478. 1994.
IN SITU GROWTH RATES OF THE OCEAN QUAHOG, ARCTICA ISLANDICA
(LINNAEUS, 1767), IN THE MIDDLE ATLANTIC BIGHT
MICHAEL J. KENNISH,1 RICHARD A. LUTZ,1
JOSEPH A. DOBARRO,2 AND LOWELL W. FRITZ3
Institute of Marine and Coastal Sciences
Rutgers University
New Brunswick, New Jersey 08903
Bureau of Shcllfisheries,
Department of Environmental Protection
Port N orris. New Jersey 08349
National Marine Fisheries Service
Alaska Fisheries Science Center
7600 Sand Point Way, NE
Seattle, Washington 98115
ABSTRACT Shell morphometric measurements of laboratory-spawned and Gulf of Maine ocean quahogs [Arctica islandica) trans-
planted at a site of commercially important shellfish beds off Cape May, New Jersey, indicate slow adult growth rates. Clams greater
than 50 mm in length grew very slowly or not at all for extensive periods of time after transplantation. The mean increase in the shell
length of adult clams from the Gulf of Maine amounted to 0.35, 0.39, -0.02, 0.40. and -0.10 mm when measured 213. 307, 368,
520. and 606 days after transplantation, respectively. Younger, laboratory-spawned clams exhibited greater growth rates than adults,
with a mean increase in shell length of 1 .87. I 34, 5.28, and 0.54 mm being recorded 75, 120. 288. and 374 days after transplantation,
respectively. These findings support those of other investigations, which have shown that the growth of A. islandica is among the
slowest of all continental-shelf bivalves.
KEY WORDS: ocean quahogs. Arctica islandica, transplantation experiments, growth rates
INTRODUCTION
The ocean quahog Arctica ( =Cyprina) islandica Linne. a sea
clam of great commercial value, is currently harvested in large
quantities off the East Coast of the United States. In 1991, U.S.
commerical landings of A. islandica totaled 22,275 mt of meats,
with most landings derived from the Exclusive Economic Zone off
of New Jersey and the Delmarva Peninsula (Mid-Atlantic Fishery
Management Council 1992). Although the Delmarva Peninsula
had the greatest landings and fishing effort between 1982 and
1989, New Jersey is now the principal area for commercial ex-
ploitation of ocean quahogs. Concentrated fishing activity off New
Jersey reduced the catch per unit effort (CPUE) by 29% between
1986 and 1992. and it now appears to be depleting the stock
(Weinberg 1993). According to Weinberg (1993), the ocean qua-
hog resource off New Jersey may be exhausted within 10 years at
the current removal rates.
Because of the increasing importance of the ocean quahog to
the sea clam industry during the 1970s, the Mid-Atlantic Regional
Fisheries Council initiated a management plan in 1977 that in-
cluded research on the biology and ecology of the species (Ropes
1979). As a result, detailed studies have been conducted on the
distribution, abundance, growth, and reproduction of A. islandica
during the past 17 years (Fogarty 1981, Jones 1981, Mann 1982.
Murawski et al. 1982 and 1989, Ropes 1984, Ropes et al. 1984a.
Rowell et al. 1990, Fritz 1991, Kraus et al. 1991). Despite these
studies, considerable uncertainty still exists with regard to the size
of the newly recruited resources in the Mid-Atlantic Bight area and
whether these resources can continue to adequately support the
fishery in the near and distant future. Early management of the
stock was based on the concept of a very limited productivity
potential of the resource, ascribable to slow growth and poor re-
cruitment rates. Hence, harvest rates greater than a few percent of
the extant stock were predicted to rapidly deplete the accumulated
stock (Murawski et al. 1989). Presently, annual landings represent
about 2 to 6% of the total harvestable stocks.
Proper management of sea clam stocks is clearly dependent on
accurate age and growth rate data as well as reliable recruitment
estimates of natural populations. Although there has been a grad-
ually accumulating data base on the growth rates and age structure
of ocean quahogs along the East Coast of the United States, there
is a dearth of information on recruitment of the clams in this
region. Several studies have shown that growth rates are extremely
slow and longevity very long in the species (Jones 1980a and
1980b. Thompson et al. 1980a and 1980b, Forster 1981 , Ropes et
al. 1981 and 1984b, Ropes and Pyoas 1982, Murawski et al. 1982,
Turekian et al. 1982, Ropes and Murawski 1983, Murawski and
Serchuk 1989). Age estimates as high as 157 and 225 years have
been reported for clams dredged off central New Jersey and south-
ern Massachusetts, respectively, on the basis of analyses of "an-
nual" growth patterns within the organism's shell (Ropes and
Murawski 1983). Murawski and Serchuk (1983), using data on
shell growth, have projected that 17% of the New Jersey resource
and 16% of the Delmarva resource is in excess of 100 years old.
In addition, the median age of these exploited stocks may be more
than 70 years. These sea clams may take 20 years or more to reach
commercial size, with depleted stocks requiring 50 to 100 years to
replenish themselves (Weinberg 1993). These data, if correct,
have serious implications for the prudent management of this valu-
able resource.
National Oceanic and Atmospheric Administration (NOAA)
surveys of ocean quahogs along the East Coast of the United States
473
474
Kennish et al.
since 1965 have not detected any new recruitment in the fishery,
although only two surveys ( 1989 and 1992) have been conducted
since 1986 (Mid-Atlantic Fishery Management Council 1992). Es-
timates of the resource distribution are as follows: southern Vir-
ginia-North Carolina = 1%. Delmarva Peninsula = 8%, New
Jersey = 21%, Long Island = 28%, southern New England =
28%, Georges Bank = 22%. Although the proportion of the re-
source off southern New England (28%) is substantial, it does not
appear to be economically harvestable because of problems with
bottom topography. In addition, the resource on the Georges Bank
(22%) cannot be harvested because the area is closed as the result
of paralytic shellfish poisoning. Hence, at least 50% of the ocean
quahog resource is currently unharvestable. Current management
strategies for the ocean quahog resource are predicated on the
stock as essentially a mine, with limited harvest rates relative to
standing stock, but no known frequency of recruitment. That no
new recruitment has been observed anywhere in the Mid-Atlantic
Bight raises concern regarding the long-term management plan for
the resource (Mid-Atlantic Fishery Management Council 1992).
At present, there is little interannual variability in population size
or structure of ocean quahogs in the Mid-Atlantic Bight, owing to
the absence of recruitment, long generation time of the species,
slow adult growth rates, and low mortality rates (Weinberg 1993).
Data on recruitment of ocean quahogs in exploited populations
of the Mid-Atlantic Bight are critically important to decision mak-
ers in the sea clam industry and government fisheries programs
who must provide effective long-term management of this valu-
able resource. The paucity of recruitment data presently available
on this species places a significant constraint on effective man-
agement of the fishery. Recruitment data are especially needed in
the deeper coastal waters (i.e., 40 to 65 m), where the greatest
densities of the clam are found. Plans are underway to conduct
initial recruitment studies of ocean quahogs along the continental
shelf off of New Jersey.
MATERIALS AND METHODS
Between June 1987 and June 1992, investigations of in situ
growth rates of A. islandica were conducted at a continental-shelf
site located 65 km off of Cape May, New Jersey (39°00'45"N.
74°04'32"W) at a depth of 45 m (Fig. 1 ). This site was selected for
several reasons: ( 1 ) the area is representative of the most actively
fished locations off the coasts of New Jersey and the Delmarva
Peninsula, (2) the presence of dense populations of clams suggests
a most suitable habitat for survival of transplanted clams, and (3)
the wreck of a Norwegian vessel, the Varanger, provides an ideal
location for transplantation experiments. These investigations
served as a precursor to recruitment studies being planned at this
location. The principal objective of the study was to monitor the
growth of ocean quahogs (some of known age) transplanted in
experimental predator-exclusion cages at the coastal site. To ob-
tain growth rates on the clams, SCUBA divers deployed and re-
covered the animals on regularly scheduled cruises to the site,
enabling the acquisition of a time series of size measurements. The
experimental cages were initially used to mollify the effect of sea
star predation, which was a major problem early in the study.
Investigators followed several procedural steps when trans-
planting the clams:
I Numbered specimens were first placed within 0.64 cm
mesh, polyethylene bags (one to five individuals per bag;
bag dimensions measured 20.32 x 20.32 x 7.62 cm).
2. The bags were subsequently sealed onboard the ship and
transported to the bottom by SCUBA divers, where the bags
were placed within 1.27 cm mesh, polyvinylchloride
(PVC)-coated galvanized wire cages (one to four bags per
cage; cage dimensions measured 60.96 x 60.96 x 17.78 cm).
3. The bags were then nestled within the wreckage of the
forward, starboard section of the stern of the military vessel
(MV) Varanger, a vessel that was torpedoed and sunk in
1942 during World War II.
The use of polyethylene bags decreased the handling time of
individual clams, thereby easing deployment operations. The
cage-within-a-cage system not only effectively eliminated sea star
predation but also facilitated rapid recovery and measurement of
the clams. Placement of the cages directly in the wreck of the
Varanger was intended to mitigate dredge/weather interference,
thus increasing the probability of achieving the goals of the proj-
ect.
Early in the research program, 96 sea clams were transplanted
in four cages at the study site, including 67 specimens (shell length
= 48.5 to 63.3 mm) transferred from a Gulf of Maine population
in November 1989 and 29 known-age, laboratory-spawned sea
clams (shell length = 9.2 to 19.9 mm) placed at the site in July
1990. These latter 29 clams were spawned in laboratories of the
Virginia Institute of Marine Science at Wachapreague. Virginia,
under the direction of Michael Castagna. Sequential numbers were
etched into the shells of the clams transplanted from the Gulf of
Maine, allowing the growth rate of each individual to be carefully
tracked. Because of the small size of the laboratory-spawned
clams at the time of transplantation and concerns with potentially
adverse effects of the etching process, the shells of these clams
were not numbered.
During the 5-year study period, 16 cruises were successfully
completed, with the last cruise to the site being June 1, 1992. At
that time, it was discovered that nearly all experimental animals
had been lost, probably when two major coastal storms C'nor'east-
ers") passed through the area, one in October 1991 and the other
in January 1992. On the final trip to the study site, SCUBA divers
found that the clam cages had been extensively damaged (i.e.,
they were broken apart and/or buried deep in sediment), and only
two live clams were recovered.
RESULTS
Laboratory-Spawned Clams
Laboratory-spawned clams (N = 29) were transplanted to the
study site on July 1, 1990. Standard shell morphometric measure-
ments (i.e., shell length, height, and width) of these clams were
made on September 16 and November 16, 1990, and April 19 and
July 12, 1991. These dates represent in situ terms of 75 , 120. 288,
and 374 days, respectively, for the known-age specimens. As
noted previously, the clams were also recovered on June 1, 1992,
but extensive damage to the cages and poor survivorship of the
transplants resulted in a limited number of growth measurements.
The total length of time that the laboratory-spawned clams were
transplanted at the experimental site is estimated to be between
485 and 577 days.
Table 1 lists the mean incremental growth rate of the labora-
tory-spawned clams during the period of transplantation. On the
basis of length measurements, the clams exhibited steady, albeit
slow growth over the year. Growth was generally variable between
sampling periods, with the greatest increase in the axial measure-
ments occurring during the colder months between November
Growth Rates of the Ocean Quahog
475
39*
oo M»mrr mwh
00 NAUTCAL M.6S
Figure 1. Map of research area showing site of in situ growth studies of A. islandica off Cape May, New Jersey.
1990 (120 days) and April 1991 (288 days). Mean increases in
length, height, and width during this period amounted to 4.55,
3.53, and 2.73 mm, respectively (Fig. 2). Growth was relatively
uniform during the other sampling periods, except in the final
year, when the least amount of growth was recorded for each axial
measurement.
Difficulties of measuring precisely with calipers along the same
plane of growth (i.e., length, height, and width planes) each time.
together with operator error with the calipers, resulted in spurious
growth measurements on some clams. In most cases, when dis-
crepancies were noted during the recording process, the clams
were remeasured for verification. In cases when the discrepancies
were not uncovered during the recording process, an effort was
made to resolve the spurious measurement during the subsequent
cruise.
Until the last cruise to the study site on June 1, 1992. mortality
476
Kennish et al.
TABLE 1.
In situ mean incremental growth rate (in millimeters) of
laboratory-spawned transplants.
Days
•p u
75
120
288
374
485/577
Length
Height
Width
0
0
0
1.87
1.91
1.09
1.34
1.03
0.98
5.28
4.24
2.81
0.54
1.37
0.46
0.16
0.16
0.36
' Tn. time of initial planting.
TABLE 2.
In situ mean incremental growth rate (in millimeters) of Gulf of
Maine transplants.
Days
T a
*0
213
307
368
520
606
717/809
Length
0
0.35
0.39
-0.02
0.40
-0.10
0.51
Height
0
0.40
0.38
-1.45
-1.07
0.80
0.94
Width
0
0.44
0.41
0.16
0.36
0.30
0.80
1 T0, time of initial planting.
of the known-age clams was low. No individuals died between
July 1 and November 1990. On both April 19, 1991. and July 12.
1991, four dead clams were found. On June 1, 1992. SCUBA
divers recovered the cages, but they were severely damaged. The
polyethylene bags containing the clams were also missing. Two of
the bags were located more than 15 m from the project site, and all
but two of the clams in these bags were dead.
Gulf of Maine Transplants
In November 1989, 67 clams were transplanted from the Gulf
of Maine to the study site after being measured and numbered. The
length of these clams ranged from 48.5 to 63.3 mm. Size mea-
surements on the clams were registered on November 13, 1989.
June 14, 1990. September 16. 1990, April 17. 1991, July 12,
1991, and June 1, 1992. Not all of the clams were measured on
September 16, 1990, because only three of the cages were recov-
ered. SCUBA divers were unable to locate one cage because of
poor visibility and because the cage had been displaced from the
study site. It was recovered on the subsequent cruise date and
relocated to the site of the Varanger. All of the cages containing
the Gulf of Maine transplants incurred extensive damage before
their recovery on June 1 , 1992, and mortality of the experimental
population was substantial. The total length of time that the Gulf
of Maine clams were transplanted at the experimental site is esti-
mated to be between 717 and 809 days.
The mean incremental growth rate of the Gulf of Maine trans-
plants during the period of transplantation is presented in Table 2.
Negative growth in shell height was recorded in November 1990
(368 days) and April 1991 (520 days), amounting to - 1.45 and
— 1.07 mm, respectively. This negative growth may have been
due to inaccuracies in measuring procedures or actual shell disso-
lution. Overall, the population surviving through July 1991 (606
days) had a small positive mean growth increment along each axis
of the shell. The total increase in axial measurements for the entire
period of transplantation was as follows: length. 0.51 mm; height,
0.94 mm; and width, 0.80 mm. The greatest increase in growth
occurred during the final year of transplantation.
The growth rate of the Gulf of Maine transplants was signifi-
cantly lower than that of the smaller laboratory-reared stock. On
the first two cruises after transplantation, June and September
1990, the mean increase in the shell length of the Gulf of Maine
transplants was 0.35 and 0.39 mm, respectively. These measure-
ments represent in situ time periods of 213 and 307 days.
Mortality of the Gulf of Maine transplants was low before the
major coastal storms of November 1991 and January 1992. During
LENGTH
HEIGHT
WIOTH
TIME Pays)
Figure 2. In situ mean cumulative growth rate (length, height, width) of laboratory-spawned transplants.
Growth Rates of the Ocean Quahog
477
the 2-year period before these storms, only 29 of the original 67
transplants had died. Seventeen clams died before the redundant
caging technique was implemented, presumably from sea star pre-
dation. Twelve clams died for unexplained reasons after being
placed in the redundant cages. Several of the clams had broken or
damaged shells, possibly the result of being tossed about in the
cages by shear stresses along the bottom, associated with storm
surges.
DISCUSSION
On the basis of the transplantation experiments described
above, a general pattern of growth is beginning to emerge for A.
islandica in the New York Bight off of Cape May. New Jersey.
The in siru data suggest that, although growth may be variable, it
is relatively rapid in small, juvenile clams. Smaller clams (<25
mm in length) may grow at rates of 10 to 12 mm per year. Larger
clams (>50 mm in length) appear to have substantially reduced
growth rates and may not grow at all for long periods of time.
These findings support the results of other investigators (e.g.,
Jones 1980b, Thompson et al. 1980a, Murawski et al. 1980 and
1982. Ropes et al. 1984b), who reported extremely slow (in situ)
growth of sea clam populations (at least in clams >20 mm).
Murawski et al. (1982) conducted one of the most comprehen-
sive studies of ocean quahog growth to date by planting about
42,000 marked clams at a deep-water (53 m) site located 48 km
SSE of Shinnecock Inlet, Long Island, New York (40°25.1'N,
72°23.7'W), in 1978 and recovering individuals annually thereaf-
ter. By examining annuli formed in the shells of these recovered
specimens. Murawski et al. (1982) recorded annual increases in
shell length of the experimental animals amounting to 6.37c at age
10, 0.5% at age 50, and 0.2% at an estimated age of 100 years.
These investigators also developed a growth rate relationship for
the marked clams (59 to 104 mm in shell length [SL]) recaptured
1 year after their release (SLt+ , = 2.081 1 + 0.9802 SL,) and an
age/growth relationship [SL = 75.68-81.31 (0.9056)'] for
younger, smaller specimens sampled from a natural population of
unmarked clams in the vicinity of the experimental site. These
results reveal that ocean quahogs are among the slowest growing
bivalves inhabiting continental-shelf waters, except perhaps dur-
ing the first 20 years of their life. They are also one of the longest
lived bivalve species, with a potential lifespan of about 225 years
(Ropes and Murawski 1983).
Growth rates of the ocean quahog vary along its geographic
range, and considerable variability may occur in the size/age re-
lationships and longevity of the species. Ropes and Pyoas (1982).
comparing the annuli in the shells of ocean quahogs from sites off
Long Island, New York, off Sable Island, Canada, and from
Georges Bank, found significant variation in growth rates of the
clams. Specimens from Georges Bank exhibited substantially
greater growth rates than those from off Long Island, New York,
and off Sable Island, Canada. However, as noted by Murawski et
al. (1980). data are not comprehensive enough to state conclu-
sively that a latitudinal cline in ocean quahog growth exists.
It is unclear what effect the artificial environment created by
the cages had on the growth or mortality of the experimental
clams. On all dives to the seafloor, SCUBA divers observed sed-
iment accumulation in the cages that buried the clams. Hence, the
conditions for the clams in the cages may not have been substan-
tially different from those along the neighboring sea floor. The
artificial environment of the cages could have contributed to the
observed mortality, but it has not been possible to determine how
or why this is the case.
ACKNOWLEDGMENTS
This is New Jersey Agricultural Experiment Station Publication
No. D-32402-2-94 and Contribution No. 94-14 of the Institute of
Marine and Coastal Sciences, Rutgers University, supported by
New Jersey State funds, the Fisheries and Aquaculture Technol-
ogy Extension Center, and New Jersey Sea Grant (Grant No.
NA85AA-D-SG084).
LITERATURE CITED
Fogarty, M. J. 1981. Distribution and relative abundance of the ocean
quahog Arctica islandica in Rhode Island Sound and off Martha's
Vineyard, Massachusetts. J. Shellfish Res. 1:33-39.
Forster. G. R. 1981. A note on the growth of Arctica islandica. J. Mar.
Biol. Assoc. U.K. 61:817.
Fritz, L. W. 1991. Seasonal condition change, morphometries, growth.
and sex ratio of the ocean quahog, Arctica islandica (Linneaus. 1767)
off New Jersey. U.S.A. J. Shellfish Res. 10:79-88.
Jones, D. S. 1980a. Annual cycle of shell growth and reproduction in the
bivalves Spisula solidissima and Arctica islandica. Ph.D. Thesis.
Princeton, NJ
Jones, D. S. 1980b. Annual cycle of shell growth increment formation in
two continental shelf bivalves and its paleoecological significance.
Paleobiology 6:331-340.
Jones, D. S. 1981 . Reproductive cycles of the Atlantic surf clam Spisula
solidissima. and the ocean quahog Arctica islandica off New Jersey. J.
Shellfish Res. 1:23-32.
Kraus, M. G., B. F. Beal & L. McMartin. 1991. Growth and survivorship
of ocean quahogs, Arctica islandica (Linnaeus) in an intertidal mudflat
in eastern Maine. J. Shellfish Res. 10:290.
Mann. R. 1982. The seasonal cycle of gonadal development in Arctica
islandica from the southern New England shelf. Fish. Bull. 80:315-
326.
Mid-Atlantic Fishery Management Council. 1992. 1993 optimum yield.
domestic annual harvest, domestic annual processing, joint venture
processing, and total allowable level of foreign fishing recommenda-
tions for surf clams and ocean quahog FMP. Technical Report, Mid-
Atlantic Fishery Management Council, Dover, DE, 67 pp.
Murawski. S. A. & F. M. Serchuk. 1983. An Assessment of Ocean Qua-
hog. Arctica islandica. Resource and Fishery in FCZ Waters off the
Northeastern USA — Autumn 1983. Woods Hole Laboratory Reference
Document No. 83-25, National Marine Fisheries Center, Woods Hole,
MA.
Murawski. S. A. & F. M. Serchuk. 1989. Mechanized shellfish harvest-
ing and its management: the offshore clam fishery of the eastern United
States, pp. 479-507. In: J. F. Caddy (ed.). Marine Invertebrate Fish-
eries: Their Assessment and Management. Wiley Interscience, New
York.
Murawski, S. A. J. W. Ropes & F. M. Serchuk. 1980. Growth studies of
the ocean quahog, Arctica islandica. ICES/CM. 1980/K:38, Shellfish
Comm., 24 p.
Murawski. S. A., J W Ropes & F M Serchuk. 1982. Growth of the
ocean quahog, Arctica islandica. in the Middle Atlantic Bight. Fish.
Bull. 80:21-34.
Murawski, S. A., F. M. Serchuk, J. S. Idoine & J. W Ropes. 1989.
Population and fishery dynamics of ocean quahog in the Middle At-
lantic Bight, 1976-1990. J. Shellfish Res. 8:464.
Ropes, J. W. 1979. Biology and distribution of surf clams (Spisula so-
478
Kennish et al.
lidissima) and ocean quahogs (Arnica islandica) off the northeast coast
of the United States. In: Proceedings of the Northeast Clam Industries:
Management for the Future. University of Massachusetts and Massa-
chusetts Institute of Technology. Sea Grant Program SP-1 12:47-66.
Ropes. J. W. 1984 Procedures for preparing acetate peels and evidence
validating the annual periodicity of growth lines formed in the shells of
ocean quahogs. Arnica islandica. Mar. Fish. Rev 46:27-35.
Ropes, J. W. & S. A. Murawski. 1983. Maximum shell length and lon-
gevity in ocean quahogs. Arnica islandica Linne. ICES/CM. 1983/
K:32. Shellfish Comm., 8 p.
Ropes. J. W. & D. Pyoas. 1982. Preliminary age and growth observations
of ocean quahogs. Arnica islandica Linne, from Georges Bank. ICES/
CM. 1982/K:15, Shellfish Committee, 6 p.
Ropes, J. W., F. M. Serchuk & S. A. Murawski. 1981. Studies of ocean
quahogs off Shinnecock Inlet, Long Island. Coastal Oceanogr. Climal.
News 3:31-32.
Ropes, J. W.,S. A. Murawski & F. M. Serchuk. 1984a. Size, age. sexual
maturity, and sex ratio in ocean quahogs. Arnica islandica L., off
Long Island, New York. Fish. Bull. 82:253-267.
Ropes, J. W., D. S. Jones, S. A. Murawski, F. M. Serchuk & A. Jerald.
Jr. 1984b. Documentation of annual growth lines in ocean quahogs.
Arnica islandica Linne. Fish. Bull. 82:1-19.
Rowell, T. W , D R Chaisson & J. T. McLane. 1990. Size and age of
sexual maturity and annual gametogenesis cycle in the ocean quahog.
Arnica islandica (Linnaeus, 1767), from coastal waters in Nova
Scotia, Canada. J. Shellfish Res. 9:195-203.
Thompson, [., D. S. Jones & D. Dreibelbis. 1980a. Annual internal
growth banding and life history of the ocean quahog Arnica islandica
(Mollusca: Bivalvia). Mar. Biol. 57:25-34.
Thompson, I.. D. S. Jones & J. W. Ropes. 1980b. Advanced age for
sexual maturity in the ocean quahog. Arnica islandica (Mollusca: Bi-
valvia). Mar Biol. 57:35-39.
Turekian, K. K., J. K. Cochran, Y. Nozaki, I. Thompson, and D. S.
Jones. 1982. Determination of shell deposition rates of Arnica islan-
dica from the New York Bight using natural 228Ra and "8Th and
bomb-produced l4C. Limnol. Oceanogr. 27:737-741.
Weinberg. J. R. 1993. Ocean quahog populations from the Middle Atlan-
tic to the Gulf of Maine in 1992. NOAA/NMFS, Northeast Fisheries
Science Center Reference Document No. 93-02, Woods Hole, MA, 18
pp.
Journal of Shellfish Research, Vol. 13, No. 2, 474-486, 1994.
TEMPORAL AND SPATIAL EFFECTS OF TIDAL EXPOSURE ON THE GAMETOGENIC
CYCLE OF THE NORTHERN QUAHOG, MERCENARIA MERCENARIA (LINNAEUS, 1758), IN
COASTAL GEORGIA
RANDAL L. WALKER1 AND PETER B. HEFFERNAN2
Shellfish Research Laboratory
Marine Extension Service
20 Ocean Science Circle
Savannah, Georgia 31411-1011
Marine Institute
80 Harcourt St.
Dublin 2, Ireland
ABSTRACT Experimental reproductive studies of the gametogenic cycle of the northern quahog. Mercenaria mercenana (Linnaeus.
1758). were performed on clams planted at various tidal heights within their natural tidal distribution. Quahogs from a homogenous
natural population were planted at the subtidal, mean low-water mark, oyster zone at the 2 hours above mean low-water mark, and
within the salt marsh for a 12 month acclimation period before sampling. Clams sampled starting 12 months after being moved to the
marsh were significantly (p < 0.0001) smaller in size than were clams from the other three tidal heights. There were no significant
(p = 0.5458) differences in ages of transplanted quahogs at the four tidal heights. Qualitative results show a similar timing of oogenic
cycles at all four sites. Females from the marsh site had consistently lower gonadal index values, delayed peak gonadal index values,
and a delay of I month in spawning. The pattern for males from the marsh is dissimilar to that for males found at other sites, with
no peak values occurring and spawning occurring a month later. For quantitative results, quahogs from the marsh site had significantly
lower values (% male gonadal area. % gonadal area occupied by spermatozoa, % female gonadal area, % gonadal area occupied by
ova, mean egg number per field of view, and mean egg diameter) than quahogs from the other sites. Although reproductive parameters
for the three lower tidal sites were not significantly different, there was a consistent decrease in all reproductive parameters measured
with increases in tidal exposure. Cage fouling appeared to alter the gametogenic pattern for clams planted subtidally, as observed in
decreases in percent gonadal area and percent area occupied by gametes, followed a month later by decreases in mean egg size and
numbers of eggs present. Subtidal clams had a more intense spawn compared with clams from the other sites. Temporal alterations
in the reproductive pattern of quahogs occurred according to tidal exposure.
KEY WORDS: Clam, gametogenesis, intertidal, Mercenaria, oocyte, reproduction, spawning
INTRODUCTION
The northern quahog, Mercenaria mercenaria (Linnaeus.
1758). primarily inhabits intertidal areas within coastal Georgia
(Walker and Tenore 1984, Walker 1987), but it occurs principally
in subtidal areas throughout the remainder of its natural range. In
the salt-marsh ecosystem of coastal Georgia, quahogs inhabit both
subtidal and intertidal areas (creek bottoms and banks, river banks,
tidal flats), up into the lower edge of the salt marsh itself. There
were no significant differences in survival of quahogs planted from
the spring low-water mark up to 3 hours above mean-low water;
however, growth decreased with increases in aerial exposure time
(Walker and Heffernan 1990). Because gametogenesis is directly
related to size, it is assumed that reproductive effort may also be
adversely affected by increases in intertidal exposure time. In-
creased intertidal exposure can result in the reduction in the
amount of gametes produced by bivalve populations (Seed 1969.
Harvey and Vincent 1989), alternation in spawning patterns
(Hunter 1949, Boyden 1971, Griffiths 1981, Borrero 1987), or a
combination of both effects (Seed and Brown 1977).
The northern quahog represents an important commercial spe-
cies for the eastern U.S. coastal fisherman. The gametogenic cycle
of the northern quahog has been studied throughout its natural
range (reviewed by Eversole 1989). However, spatial effects on
reproduction studies within a single locale are lacking for this
species with two exceptions. Keck et al. ( 1975) determined that no
differences in spawning patterns occurred between two quahog
beds located within the Delaware Bay: however, differences in the
maturation process were detected. Yet, these beds occurred in
typical subtidal habitats where large differences in gametogenesis
are less likely to occur. Eversole et al. (1980), using standard
histological analysis techniques, found no differences in reproduc-
tive parameters between quahogs grown at different densities or
between intertidal and subtidal areas in South Carolina. However,
gonadal-somatic indices (GSI) analysis revealed that clams grown
at lower densities and at the subtidal site had higher GSI values
than did those grown at high densities or at the intertidal site
(Eversole et al. 1984). The intertidal zone represents a harsh
environment for quahogs, with the degree of stress increasing with
increases in tidal exposure time. This study uses quantitative his-
tological techniques to determine spatial (increased tidal exposure)
effects on quahog gametogenesis in coastal Georgia.
MATERIALS AND METHODS
In August 1990, approximately 1.200 quahogs ranging in size
from 30 to 65 mm in shell length (i.e., longest possible measure-
ment: anterior- posterior) were collected from House Creek, Little
Tybee Island, Wassaw Sound. Georgia. Quahogs were divided
into four equal groups of 300 clams. A set of quahogs was placed
within a 12.7 mm mesh, vinyl-coated wire cage, 1 x 1 x 0.5 m
that was placed just below the spring low-water mark, at the mean-
low water, at 2 hours above mean-low water (the oyster zone), and
just within the salt marsh adjacent to the Shellfish Research Lab-
oratory on Skidaway Island, Georgia. Quahogs were allowed to
grow at these sites until October 1991. before sampling began, to
479
480
Walker and Heffernan
allow clams to fully acclimate to their respective intertidal heights.
By allowing a 12 month delay in sampling, the quahogs would
have presumably gone through at least two natural reproductive
episodes (Heffernan et al. 1989) before the start of this experi-
ment.
Beginning October 1991 , monthly samples of 10 to 20 quahogs
were collected per tidal site. Each quahog was measured for shell
length with Vernier calipers, aged according to shell band analysis
(Walker 1987), and dissected along a standard axis from umbo-
to-foot with a gonadal sample removed for histological analysis
(Howard and Smith 1983). The visceral mass section containing
the gonadal sample was preserved in Davidson's solution under
refrigeration for 48 hours. After 48 hours, the samples were
washed with 50% ethanol and replaced with 70% ethanol for stor-
age until samples were processed. Tissue samples were dehydrated
in an alcohol series, cleared in toluene, and embedded in para-
plasm Sections were cut 7 to 10 u.m in thickness with a rotary
microtome. Sections were stained with Harris hematoxylin and
counterstained with eosin (Howard and Smith 1983).
Qualitative Reproductive Analysis
Prepared histological slides were examined with a Zeiss Axio-
vert 10 microscope (x20), sexed. and assigned to a development
stage as described in Heffernan et al. (1989) where sexually un-
differentiated = 0; male and female active = 2; male and female
ripe = 3; and male and female ripe and spawning = 1 . Male and
female monthly gonadal index (Gl) values were computed for each
of the four intertidal sites. The monthly GI for both sexes was
determined by multiplying the number of specimens ascribed to
each stage, summing all those values, and dividing this figure by
the total number of clams analyzed.
Quantitative Reproductive Analysis
Quantitative analysis of gonadal preparations was performed
with Color Image Analyzed Densitometry Microscopy housed at
Skidaway Institute of Oceanography, Savannah, Georgia (Kanti et
al. 1993). Photomicroscope images (10 x 1.25 optivar) were
viewed and captured by a Hitachi Model DK-7000 SU-3 Chip
CCD color camera. The images were then viewed on a Trinitron
color-video monitor. The image analyzer is capable of performing
detailed area measurements and statistical analysis on features de-
tected within the blue thresholds (operator controlled). Two mi-
croscopic fields of view (xlO) per specimen were analyzed to
ensure detection of variations in gametogenic development within
specimen.
An operator-controlled marker was used to edit nongonadal
tissue (e.g., intestines and digestive diverticula) in the evaluation
of percent gonadal area per field. Females were analyzed for per-
cent gonadal area, percent gonadal area occupied by oocytes,
oocyte number per field of view, and mean oocyte diameter. Egg
number was manually counted from the Trinitron screen, and
oocyte diameter of nucleolated oocytes was measured directly on
the screen. Males were analyzed for percent gonadal area and
percent gonadal area occupied by spermatozoa.
Mean individual values for each data category were analyzed.
Mean monthly values were then computed and used in the quan-
titative assessment of gametogenesis. Sex ratios were tested
against a 1:1 ratio with a Chi-square statistic (Elliott 1977).
Statistical analysis of data was performed by analysis of vari-
ance (ANOVA) and Tukey's multiple range tests (MRT) with SAS
for a personal computer (SAS Institute, Inc. 1989). All percentage
data were arcsine transformed before statistical analysis. Mean
quahog reproductive values for all qualitative and quantitative pa-
rameters analyzed were compared statistically (/-test) on a monthly
basis during periods of suspected spawning activity.
Water temperature and salinity data were recorded daily at
0800 hours from September 1991 to October 1992 from the dock
of the Marine Extension Service, Skidaway River, directly adja-
cent to the grow-out site.
RESULTS
The results of the ANOVA reveal that no significant differ-
ences (p = 0.5458) occurred in the age (range, 2 to 1 1 years; x =
4.6 years) of quahogs collected from the four tidal heights; how-
ever, significant (p < 0.0001) differences occurred in size (range,
39 to 81 mm). Tukey's MRT shows that quahogs from the marsh
(x = 55.3 mm) were smaller than clams from the other three sites
(x = 60.4 mm for quahogs at the oyster site and 63 mm for clams
at the two lower sites).
A total of 406 quahogs were sexed during this experiment. The
ratio of male to females was 1.00:0.91 and did not significantly
differ from unity (Chi-square = 0.9852, p > 0.05). Sex ratios
were equal for quahogs collected from each tidal level (Table 1).
Qualitative Results
A significantly higher (p = 0.0052) percentage of quahogs
from the marsh site (21.2%) were sexually undifferentiated com-
pared with clams (Table 1) from the oyster (6.5%), mean low-
water (7.2%), and subtidal sites (4.5%). Sexually undifferentiated
clams occurred from October to February. November to January,
and April to June for the oyster zone, mean low-water, and sub-
tidal sites, respectively, but occurred from October to June in the
marsh site (Fig. 1).
GI for male and female quahogs planted at the various tidal
heights are given in Fig. 2. The results show similar timing of
maturation for both males and females from the three lower tidal
TABLE 1.
Sex ratios and number of quahogs, M. mercenaria (Linnaeus, 1758), sexually undifferentiated at each tidal level
Percent
Chi-Square
Tidal Level
No. of Males
No. of Females
No.
Undifferentiated
Undifferentiated
Value
Marsh
33
34
18
21.2
0.0149
Oyster
65
51
8
6.5
1 .6897
Mean low water
55
61
9
7.2
0.3103
Subtidal
60
47
5
4.5
1.5790
Total
213
193
40
9.0
0.9852
Tidal Exposure and Gametogenesis in Quahogs
481
Subtidal
Ripe
J 1 Active
Ripe and spawning
Undifferentiated
Nov
1991
Jan Feb Mar Apr May Jun
1992
Figure 1. Qualitative data illustrating the sex and gonadal developmental stages of quahogs, M. mercenaria, sampled from subtidal (A), mean
low water (B), 2 hours above mean low water (C), and the marsh zone (D) from the Skidaway River, Georgia, from October 1992 to June 1993.
The height of each the shaded areas represents the percent frequency of clams in each developmental stage.
482
Walker and Heffernan
-»- Subtidal Male
A.
-•- MLWMale
3.00-
-*- Oyster Male
2 50-
— •— Marsh Male
2.00-
■ ^^* » /r~~ * *~AA\
1.50-
1.00-
0.50-
nnn-
B.
300
2 50
2 00
1 50
1.00
0 50-
0 00
-■ — Subtidal Female
-• — MLW Female
-* — Oyster Female
-• — Marsh Female
Oct
Dec Jan Feb Mar
Apr
1992
May
Jun
Nov
1991
Figure 2. GI ± 1 standard error (SE) for male (Al and female (B)
quahogs, M. mercenaria, planted at four tidal heights in the coastal
waters of Georgia. MLW, mean low water.
heights, whereas a markedly different pattern exists for the clams
from the marsh-grass tidal height. A decrease in the GI for males
from the three lower tidal sites occurred from November to De-
cember, possibly indicating the end of fall spawning in clams (Fig.
2A). An increase in GI values followed from December to April,
when values peaked. A significant and rapid drop in GI values
occurred between April and June for clams from the oyster and
mean low- water sites, whereas male GI values from the subtidal
area significantly decreased from April to May with no change
from May to June. The pattern exhibited by the males from the
high intertidal marsh site increased in GI value from 1 .0 in Octo-
ber to 2.0 in November and stayed at this value until April, when
the value decreased to 1 .0 by June. No significant change in GI (a
= 0.05, f-test) values occurred between April (GI = 2.0) and
May (GI = 1.8), but decreased significantly to 1 .0 by June. Thus,
males from the marsh site never achieved the maximum GI values
exhibited by clams from the subtidal (GI = 2.6) or other two sites
(GI = 3.0). A low percentage of males (20%) and females (20%)
from the marsh site were ripe by May. The decrease in GI values
between May and June indicates spawning, and 55% of the males
and 11% of the females were in the ripe and spawning stage in
June. Approximately 92% of the clams from the three lower sites
were in the ripe stage in May, whereas 88% were in the ripe and
spawning stage in April (Fig. 1).
The pattern in GI values for females was similar for quahogs
from all three lower tidal sites. The pattern for the females from
the marsh site follows a trend similar to that of females from the
other sites, but at reduced levels and lower peak values; also,
clams spawned a month later than clams at the other sites (Fig.
2B). The GI values for clams at the three lower intertidal sites
gradually increased from low values in October (range, 1.0 sub-
tidal to 1.6 mean-low water) to peak values in April (3.0 for
subtidal and mean-low water to 2.86 for the oyster site), before
decreasing significantly to 1.0 by May, indicating spawning. GI
for the clams from the marsh site were significantly (p < 0.0001)
lower than values for the other sites in April (GI = 2.0) and
significantly higher in May (GI = 2.33). Thus, females from the
marsh site achieved a peak in GI value a month later, and appar-
ently spawned a month later, than quahogs from the lower tidal
sites.
Quantitative Results
Spermatogenesis
The results of ANOVA tests for the four tidal heights show that
significant differences (p < 0.0001) occurred for both the percent
gonadal area and percent gonadal area occupied by spermatozoa.
Quahogs from the marsh had lower levels (p < 0.05, Tukey's
MRT) of gonadal area and gonadal area occupied by spermatozoa.
Data analysis by month showed an alternation of separation of
means of reproductive parameters for the three lower tidal heights.
The most consistent pattern over time occurred in the quahogs
from the mean low-water level, with the most dramatic change in
pattern for the subtidal and oyster-level quahogs occurring in De-
cember 1991 (Fig. 3). Subtidal clams showed a significant de-
crease in percent gonadal area and percent gonadal area occupied
« 0
O-
Oct Nov
Dec Jan Feb Mar Apr May
1991 1992
Figure 3, Percent male gonadal area (A) and percent area occupied by
spermatozoa (B) for quahogs, M. mercenaria, planted at four tidal
heights in the coastal waters of Georgia. MLW, mean low water.
Tidal Exposure and Gametogenesis in Quahogs
483
by spermatozoa in December, whereas oyster-level clams had a
significant increase in both parameters. The decrease in parame-
ters in December for the subtidal males may be explained by
adverse effects of cage fouling on gametogenesis (see Discussion
below). A similar decrease occurs for female parameters (Fig. 4A
and B). The rapid increase in parameters for the oyster-level males
and subsequent decrease between December and January is not
reflected in the female parameters (Fig. 4A and B).
Oogenesis
The pattern of oogenesis is best seen in the development ex-
hibited by quahogs from the mean low-water and oyster sites. A
gradual increase in the percent gonadal area (Fig. 4A), percent
gonadal area occupied by oocytes (Fig. 4B). and mean number of
eggs per field (Fig. 4C) occurs from October through April. A
rapid decline in these parameters from April through July probably
indicates spawning. ANOVA of the various parameters reveals
A.
£ 50 -
15
S4°-
/
f 30-
CO
1 20-
LL.
§10-
a>
°- n
i
B.
\j. 22
D. E
Nov Dec Jan Feb Mar Apr
1991 1992
Figure 4. Percent female gonadal area ± 1 SE (A), percent gonadal
area occupied by oocytes ± 1 SE (B), mean oocyte number per field of
view ± 1 SE (C), and mean nucleolated oocyte diameter ± 1 SE (D) for
quahogs, M. mercenaria, planted at four intertidal heights in coastal
Georgia. MLW, mean low water.
significant differences (p < 0.0001) among treatments for each
reproductive parameter.
Tukey's MRT revealed that, in all cases, data from the marsh
site are significantly lower than values from the other three sites.
Data from the subtidal site differ from the mean low-water and
oyster sites where in October and November, mean percent go-
nadal area (p = 0.0012 and p = 0.0044. respectively) and go-
nadal area occupied by oocytes (p < 0.0001 and p = 0.0010,
respectively) were significantly higher in values. A significant
decrease in reproductive values for percent gonadal area and per-
cent gonadal area occupied by oocytes in subtidal clams occurred
between November and December, before a rapid increase
achieved peak values in April. Furthermore, a more rapid decrease
in percent female gonadal area and percent gonadal area occupied
by oocytes occurred for clams from the subtidal site (Fig. 4A and
B) between April and May compared with the more gradual de-
crease in parameters for clams from the mean low-water and oyster
sites (Fig. 4). In April, mean percent area occupied by oocytes of
subtidal clams was significantly higher (p = 0.0001 ) in value than
for clams from the intertidal sites. By May, subtidal clams had
significantly lower (p = 0.0001) area occupied by oocytes than
did clams from the intertidal sites. Thus, it appears that the sub-
tidal quahogs had a more intense spawning between April and May
than did clams at the other sites. A gradual increase in the mean
number of eggs per field of view occurred from December through
March-April for quahogs from the mean low-water and oyster-
level sites, whereas mean egg number increased from November
to December at the subtidal site, but decreased from December to
January. A rapid increase in mean number of eggs per field of
view occurred in clams from the subtidal site from January (x =
2.6 eggs per field) to February (x = 14 eggs per field). This
increase in egg numbers was correlated with an increase in the
mean oocyte diameters found during this time period (Fig. 4D).
Finally, mean number of egg patterns (Fig. 4C) showed a temporal
shift in peak occurrence. Maximum egg diameter values occurred
as early as February in the clams from the subtidal cage compared
with peak values in March for mean-low water, March-April for
oyster site, and May for marsh cage clams.
Mean ambient water and salinity values taken at 0800 hours at
the Marine Extension Service dock on the Skidaway River are
given in Figure 5. Ambient river temperatures ranged from a mean
high of 28°C in September 1991 to a mean low of 10°C in mid-
January before increasing to 29°C by July 1992. River salinity
gradually increased from 17 ppt in September 1991 to 30 ppt in
January 1992, before decreasing to 23.5 ppt in April, and then
rapidly increasing to 30 ppt by July 1992.
DISCUSSION
The results of this study showed that spatial differences in
gametogenesis occurred within a Georgia population of quahogs.
Intertidal exposure did affect gametogenesis where animals at the
upper limits of its intertidal distribution exhibited lower values of
percent male gonadal area, percent gonadal area occupied by sper-
matozoa, percent female gonadal area, percent gonadal area oc-
cupied by oocytes, mean egg numbers per field of view, and mean
egg diameters. Furthermore, cage fouling appears to adversely
alter the pattern of gametogenesis for clams planted in subtidal
areas. Quahogs were found to be reproductively active from Oc-
tober to June, with a major spawning period occurring from April
to June for the lower three tidal sites and from May to June for
484
Walker and Heffernan
1 — I — I — I — I — I — I — r
J F M A M
i — i — i — i — i — i — i — r
J J A S
1991
1992
Figure 5. Biweekly means of ambient water temperature (°C) and salinity (ppt) for the site where northern quahogs, M. mercenaria, were
planted from the subtidal zone to the marsh zone adjacent to the Marine Extension Service's dock located on Skidaway Island, Georgia.
Temperature and salinity readings were taken at 0800 hours each morning.
clams from the marsh site. These findings are in general agreement
with two earlier reproductive studies showing major spawning
season in the spring for quahogs in Georgia (Pline 1984, Heffernan
et al. 1989); however, spawning times differed. Various spawning
times between these Georgia reproductive studies may result from
a number of factors, e.g., sampling in differing years, from dis-
similar habitats, and under varying saline or water temperature
regimens. In this study, quahogs from the Skidaway River
spawned from April through June, whereas quahogs from a Was-
saw Island site spawned from March to May in 1982, and chow-
ders spawned sooner and for a longer period than littlenecks (Pline
1984). Quahogs from House Creek site spawned from March to
May in 1982, from April to May in 1985, and from April to June
in 1986 (Heffernan et al. 1989). These annual variations in spawn-
ing patterns are probably tied to environmental parameters (loca-
tion, habitat, water temperature, and salinity). Quahogs normally
spawn when water temperatures increase from 20°C. The Skid-
away River temperatures increased from a mean of 21 to 27°C
from April to June (Fig. 5).
Temporal and spatial variability in spawning patterns between
oyster beds has been documented within the Chesapeake Bay
(Kennedy and Krantz 1982) as well as within the James River of
the Chesapeake Bay (Cox and Mann 1992). A number of factors
can contribute to the variability of spawning patterns, including
differences in water temperatures, salinity, disease pressure, size
of oyster, and time since last spawning (Cox and Mann 1992), as
well as geographical variables (Brown 1984, Newell et al. 1982,
Eversole 1989) or depth of occurrence (MacDonald and Thompson
1986). This study shows that the degree of tidal exposure can
effect the reproductive pattern of quahogs.
Although, overall, there were no significant differences in the
various reproductive parameters for quahogs grown at the oyster-
level, mean low-water, or subtidal sites, there is a consistent de-
crease in the various quantitative values according to tidal expo-
sure. Reproductive variables become statistically lower when
clams occurred within the salt marsh — the upper limit of the in-
tertidal distribution of this species in Georgia. Harvey and Vincent
(1989) found, for a population of Macoma balthica, that clams
from the upper limit of its intertidal limit were reproductively
senescent compared with individuals from the lower intertidal
area. In contrast, only a small percentage (21%) of Mercenaria
species from the upper intertidal distributional limit in this study
were reproductively undifferentiated, with the majority showing
some reproductive potential. At peak spawning, all clams showed
some differentiation (Fig. 1); however, whether any of the clams
from the marsh site produced viable gametes is unknown. Only
40% of the marsh-site clams achieved the ripe stage, all during the
month of May, whereas 92% of the clams from the other lower
sites were in the ripe stage in April, and 88% were ripe and
spawning by May (Fig. 1).
Differences in the maturation of other bivalves according to
tidal exposure have been documented. Borrero (1987) found that
the onset of gametogenesis and spawning occurred earlier in 10
populations of mussels, Geukensia demissa, from low intertidal
habitats than in 10 populations from the higher intertidal zone. For
a population of Hiatella species from the Arctic, subtidal clams
Tidal Exposure and Gametogenesis in Quahogs
485
were reproductively active year around, whereas clams from the
littoral zone did not produce ova from June through October
(Hunter 1949). In a Modiolus modiolus population, mussels from
the intertidal zone spawned in fall to winter, whereas mussels from
the subtidal zone were reproductively active year around (Seed and
Brown 1977). For cockles, Cerastoderma edule, higher percent-
ages of mature cockles occur in clams collected from the extreme
low-water, spring tidal level than in clams collected from mean
low- water mark (Boy den 1971).
For Mercenaria species, Eversole et al. (1980), using qualita-
tive histological analysis, found no significant differences in re-
productive parameters for clams planted subtidally and intertidally
in South Carolina. Qualitative GSI analysis revealed that clams
from the subtidal zone had higher values than clams from the
intertidal zone (Eversole et al. 1984). Landers (1954) suggested
that intertidal quahogs spawned earlier than subtidal ones. The
quantitative data from this experiment suggest that subtidal qua-
hogs spawned at the same time as clams from the mean low-water
and oyster-level sites, but subtidal clams appear to have had a
more intense initial spawn from April to May (Figs. 2 through 4),
followed by a resting phase from May to June, with another rapid
decline in reproductive parameters in June to July. In clams from
the oyster and mean low-water sites, a gradual decline in all pa-
rameters occurred from April to July.
A decline in reproductive parameters for quahogs according to
increases in tidal exposure is expected. In intertidal growth studies
(Walker and Heffernan 1990). growth rate of quahogs decreased
according to increase in intertidal exposure time. No significant
differences in survival occurred in that experiment for quahogs
planted from the spring low-water mark up to the 3 hours above
the mean low-water mark. In this experiment, quahogs from the
marsh site were significantly (as determined by Tukey's MRT)
smaller than animals from the lower intertidal sites. This size
difference is not associated with any difference in age of clams
sampled. Clams were not significantly different in age.
Greater initial mortalities occurred for quahogs at the marsh
site, resulting in only 10 clams being sampled per month. The
heavy mortality experienced within the marsh site may be due to
substrate differences, as well as increased physiological stress,
compared with earlier intertidal quahog growth and survival stud-
ies (Walker and Heffernan 1990). In this study, quahogs were
replanted in a muddy substrate occurring at the base of live Spar-
tina alternaflora, whereas in the previous study, clams were
planted in a substrate of predominately sand material. Sand sub-
strates are the preferred substrate of quahogs for greater growth
and survival (Pratt 1953). Because quahogs roughly allocated
equal amounts of energy to gamete production as to flesh produc-
tion (Ansell and Lander 1967, Hibbert 1977), the observed de-
creases in growth rate associated with increases in intertidal ex-
posure are interpreted as representing decreased total energy avail-
ability, with similar negative effects on gametogenic development
expected.
Fouling by sea squirts, Molgula species, on the subtidal cage
presumably had a negative impact on the gonadal development of
the quahogs. Decreases in percent gonadal area and percent area
occupied by oocytes, followed a month later by decreases in mean
egg diameters and mean numbers of eggs, occurred for clams in
the fouled subtidal cage. No major epibenthic fouling (excepting
some minor oyster spat settlement) occurred on intertidal cages
during the course of this experiment, but in November 1991, a
heavy set of Molgula species was observed on the subtidal cage.
By December 1991, Molgula species completely covered the top
and sides of the subtidal cage, with consequent negative effects on
flow rate and food availability presumed for the caged quahogs.
Approximately 90% of the cage was cleaned in December, and on
the January sample date, the remaining corner had become cleared
of Molgula species. The negative effect of this fouling can best be
seen in females, where a sharp contrast to the patterns in the two
intermediate intertidal height cages occurred. There was a signif-
icant decline in both percent gonadal area and percent gonadal area
occupied by oocytes for clams from the subtidal cage from No-
vember to December as compared with increases in these param-
eters for quahogs from the intermediate tidal cages. These param-
eters showed a recovery, presumably associated with cage clean-
ing and increased tood supply, by February, with oocyte area
showing a staggered resurgence. Similarly, but delayed by 1
month, there were significant decreases in mean egg numbers
observed in the subtidal quahogs in December-January, whereas
the mean egg-size pattern was shown to be 1 month out of syn-
chrony with the two intermediate tidal levels. Became gameto-
genesis is coupled with growth and feeding, we believe that this
pattern observed for the subtidal quahogs is a direct result of the
cage fouling that restricted growth and feeding during the Novem-
ber-to-December time period.
One can only speculate as to the results if cage fouling on the
subtidal cage had not occurred. In October and November, the
mean percent female gonadal area and percent gonad occupied by
eggs at the subtidal level were significantly higher than values at
the mean low-water mark and oyster site. Given the higher game-
togenic output at the subtidal level before the impact of cage
fouling, it is reasonable to speculate that peak values may have
been attained earlier and/or at higher levels in the absence of cage
fouling.
Although there is a consistent decrease in the various repro-
ductive parameters according to tidal placement, the results of this
study show that no significant differences in reproductive param-
eters were detected between quahogs planted from the subtidal to
the 2 hours above mean low-water mark. This zone is where most
quahogs occur naturally in the coastal waters of Georgia. A sig-
nificant reduction in gamete production does occur for clams
grown at the upper limit (i.e.. in the salt marsh) of its tidal dis-
tribution, and it is unknown if the gametes produced there would
produce viable offspring. Thus, reproductive senescence does not
occur for all individuals from the upper limits of its distribution, as
was observed in M. balthica populations (Harvey and Vincent
1989). In light of the results of this study, resource managers can
assume that spawning occurs during single events (fall and spring)
for most of the Georgia quahog population, with the major repro-
ductive contribution coming from clams within the lower intertidal
and subtidal populations.
ACKNOWLEDGMENTS
We thank Ms. P. Adams and Ms. M. Sweeney for preparing
the histological slides. Mr. D. Hurley is thanked for collecting
some of the field samples. Ms. D. Thompson prepared the manu-
script, and Ms. S. Mcintosh and A. Boyette prepared the graphs.
Special thanks are given in memory of the late Dr. Walter Isaac,
Department of Psychology, University of Georgia, for allowing us
to "borrow" many of the histological supplies that made this work
possible. This work was funded by the Georgia Sea Grant Program
under project number NA84AA-D-00072.
486
Walker and Heffernan
Literature Cited
Ansell, A. D. & K. F. Lander. 1967. Studies on the hard-shell clam,
Venus mercenaria, in British waters. III. Further observations on the
seasonal biochemical cycle and on spawning. J. Appl. Ecol. 4:425-
435.
Borrero, F. J. 1987. Tidal height and gametogenesis reproductive varia-
tion among populations of Geukensia demissa. Biol. Bull. 173:160-
168.
Boyden, C. R 1971. A comparative study of the reproductive cycles of
the cockle Cerastoderma edule and C. glaucum. J Mar. Biol. Assoc.
U.K. 51:605-622.
Brown, R. A. 1984. Geographical variations in the reproduction of the
horse mussel. Modiolus modiolus (Mollusca: Bivalvia). J. Mar. Biol.
Assoc. U.K. 64:751-770.
Cox, C. & R. Mann. 1992. Temporal and spatial changes in fecundity of
eastern oysters, Crassostrea virginica (Gmelin. 17911 in the James
River, Virginia. J. Shellfish Res. 11:49-54.
Elliott, J. M. 1977. Some Methods for the Statistical Analysis of Samples
of Benthic Invertebrates. Freshwater Biological Association Scientific
Publication No. 25, Titus Wilson & Son LTD., Kendal. 160 pp.
Eversole, A. G. 1989. Gametogenesis and spawning in North American
populations: Implications for culture, pp. 75-1 10. In: J. J. Manzi and
M. A. Castagna (eds.). Clam Mariculture in North America, vol. 19.
Development in Aquaculture and Fisheries Science. Elsevier. New
York.
Eversole, A. G., W. K. Michener, & P. J. Eldridge. 1980. Reproductive
cycle of Mercenaria mercenaria in a South Carolina estuary. Proc.
Nail. Shellfish. Assoc. 70:22-30.
Eversole, A. G., W. K. Michener & P. J. Eldridge. 1984. Gonadal con-
dition of Mercenaria mercenaria (Linne) in a South Carolina estuary.
J. Shellfish Res. 4:88 (abstract).
Griffiths, R. J. 1981. Production and energy flow in relation to age and
shore level in the bivalve Choromytilus meridonalis (Kr). Est. Coastal
Shelf Sci. 73:477^193.
Harvey, M. & B. Vincent. 1989. Spatial and temporal variations of the
reproductive cycle and energy allocation of the bivalve Macoma balth-
ica (L.) on a tidal flat. J. Exp. Mar. Biol. Ecol. 129:199-217.
Heffernan, P. B., R. L. Walker & J. L. Carr. 1989. Gametogenic cycles
of three bivalves in Wassaw Sound, Georgia: I. Mercenaria merce-
naria (Linnaeus, 1758). J. Shellfish Res. 8:51-60.
Hibbert. C.J. 1977. Energy relations of the bivalve Mercenaria merce-
naria on an intertidal mudflat. Mar. Biol. 44:77-84.
Howard, D. W. & C. S. Smith. 1983. Histological Techniques for Marine
Bivalve Mollusks. U.S. Dept. Commerce, NOAA Technical Memo-
randum NMFS-F/NEC-25. 97 pp.
Hunter. W. R. 1949. The structure and behaviour of Hiatella gallicana
(Lamarck) and H. arctica (L), with special reference to the boring
habitat. Proc. Royal Soc. Edinb. 63:271-289.
Kami, A., P. B. Heffernan & R. L Walker. 1993. Gametogenic cycle of
the southern surf clam, Spisula solidissima similis (Say), from St.
Catherines Sound, Georgia. J. Shellfish Res. 12:255-261.
Keck. R. T., D. Mauer & H. Lind. 1975. A comparative study of the hard
clam gonad development cycle. Biol. Bull. 148:243-258.
Kennedy, V. S. & L. B. Krantz. 1982. Comparative gametogenic and
spawning patterns of the oyster Crassostrea virginica (Gmelin) in cen-
tral Chesapeake Bay. J. Shellfish Res. 2:133-140.
Landers, W. S. 1954. Seasonal abundance of clam larvae in Rhode Island
waters 1950-1952. U.S. Fish Wildl. Spec. Rep. Fish. 117:1-29.
MacDonald, B. A. & R. J. Thompson. 1986. Influence of temperature
and food availability on the ecological energetics of the giant scallop
Placopecten magellanicus. Mar. Biol. 93:37—48.
Newell, R. I. E., T. J. Hilbish, R K Koehn & C. J. Newell. 1982.
Temporal variation in the reproductive cycle of Mytilus edulis L. (Bi-
valvia. Mytilidae) from localities on the eastern coast of United States.
Biol. Bull. 162:299-310.
Pratt, D. M. 1953. Abundance and growth of Venus mercenaria and Cal-
locardia morrhuana in relation to the character of the bottom sedi-
ments. J. Mar. Res. 12:60-74.
Pline, M. 1984. Reproductive cycle and low salinity stress in adult Mer-
cenaria mercenaria L. of Wassaw Sound, Georgia. Master Thesis,
Georgia Institute of Technology, 74 pp.
SAS Institute Inc. 1989. SAS User's Guide: Statistics. Version 6. SAS
Institute Inc., Cary, NC, 846 pp.
Seed, R. 1969. The ecology of Mytilus edulis L. (Lamellibranchiata) on
exposed rocky shores I. Breeding and settlement. Oecologia 3:277-
316.
Seed, R. & R. A. Brown. 1977. A comparison of reproductive cycles of
Modiolus modiolus (L.), Cerastoderma edule (L.) and Mytilus edulis
L. in Strangfold Lough. Northern Ireland. Oecologia 30:173-188.
Walker, R. L. 1987. Hard Clam, Mercenaria mercenaria (L), Popula-
tions of Coastal Georgia. University of Georgia, Georgia Marine Sci-
ence Technical Report Series, No. 87-1. Athens. 73 pp.
Walker, R. L. & P. B. Heffernan. 1990. Intertidal growth and survival of
northern quahogs Mercenaria mercenaria (Linnaeus 1758) and Atlan-
tic surf clams Spisula solidissima (Dillwyn 1817) in Georgia. J. World
Aqua. Soc. 21:301-313.
Walker, R. L. & K. R. Tenore. 1984. The distribution and production of
the hard clam, Mercenaria mercenaria. in Wassaw Sound, Georgia.
Estuaries 7:357-363.
Journal of Shellfish Research, Vol. 13, No. 2, 487^491, 1994.
THE EFFECT OF AQUACULTURE ON THE GENETICS OF NATURAL POPULATIONS OF
THE HARD CLAM, MERCENARIA MERCENARIA (L)
KATHARINE LEE METZNER-ROOP1
Graduate Program in Marine Biology
University of Charleston
Charleston, South Carolina 29412
ABSTRACT From 1980 to 1984. Mercenaria mercenaria seed clams from Aquaculture Research Corporation (ARC) in Massachu-
setts were imported by Trident Seafarms Company, located on the Folly River near Charleston, South Carolina. Although broadly
similar genetically to naturally occurring M. mercenaria populations, ARC stocks show substantially elevated frequencies of two
ordinarily rare glucose phosphate isomerase (GPI) alleles and have been marked with notata, a shell coloration gene. The purpose of
this investigation was to estimate the extent to which the ARC genome has become established in natural populations of M. mercenaria
remaining 3 years after Trident Seafarms discontinued aquaculture operations in 1989. I examined seven samples of over 300
individual clams each: large and small individuals from Trident Seafarms itself and from two of its former grow-out sites, plus small
clams only from a fourth site where no known aquaculture had occurred. I obtained one significant result — apparent linkage dis-
equilibrium between notata and a GPI marker at one grow-out site. Over 12 tests, such findings may be attributable to type I statistical
error. So. although some follow-up study may be indicated, generally, the effect of aquaculture on the genetics of natural populations
of hard clams in South Carolina seems to have been negligible.
KEY WORDS: clams, Mercenaria. electrophoresis, isozymes, notata. introductions, aquaculture
INTRODUCTION
Important questions have recently arisen regarding the fre-
quency and consequences of genetic interaction between intro-
duced aquaculture stocks and native populations. One possible
outcome would be the selectively neutral spread of a foreign ge-
nome through the host population until it is diluted beyond detec-
tion (Dillon 1988). A second outcome would be that introduced
genomes prove deleterious and disappear more rapidly than ex-
pected from simple diffusion (Helle 1981, Krueger et al. 1981.
Verspoor 1988. Hindar et al. 1991). A third possibility would be
that introduced genomes prove selectively advantageous and in-
crease in frequency, to the benefit or detriment of other species,
including mankind (Helle 1981. Krueger et al. 1981, Strand and
Lavan 1990. Hindar et al. 1991). An analysis of the Trident Sea-
farms operation, which from 1980 to 1989 raised a New England
stock of Mercenaria mercenaria near Charleston, South Carolina,
affords an opportunity to examine the genetic consequences of one
aquaculture operation in some detail.
The establishment and operation of Trident Seafarms has been
reviewed by Brown et al. ( 1983), Manzi et al. ( 1984), Stevens et
al. (1984), and Manzi et al. (1986). In September 1980, M. mer-
cenaria seed clams were first imported from Aquaculture Research
Corporation (ARC) in Massachusetts to the Trident Seafarms fa-
cility on Folly Island. 15 km south of peninsular Charleston, South
Carolina. Here, clams were grown to about 8 to 10 mm in nursery
systems of various design. Older clams were also held in vinyl-
coated wire trays in the adjacent Folly River marshes. From 1980
to 1981. experimental field units were placed at Little Oak Island,
also near the Folly River, about 4 km east of the central facility.
In 1982, field units were planted in Bass Creek at Kiawah Island,
8 km west. The clams at Bass Creek were removed in 1985.
Trident Seafarms maintained larger clams in the marshes of the
Contribution No. 1 18 from the Gnce Marine Biological Laboratory
'Present Address: % Dr. R. T. Dillon, Department of Biology, College of
Charleston. Charleston. South Carolina 29424.
Folly River near its central facility until going out of business in
1988. The nursery holding ARC seed clams was also discontinued
at that time. Hurricane Hugo caused some damage to experimental
clam pens remaining in place in the marshes around Trident Sea-
farms in September 1989, after which all remaining cages were
recovered by state personnel (N. H. Hadley, personal communi-
cation).
With repeated planting of ARC stocks in local waters over a 9
year period, it seems possible that these clams may have been able
to spawn and hybridize with the local wild population of M. mer-
cenaria. Dispersal in the veliger larval stage may have spread the
ARC genome substantial distances from original nursery and
grow-out sites (Dillon and Manzi 1989a, McCay 1990).
Although allozyme studies have not generally detected genetic
divergence among natural populations of M. mercenaria inhabit-
ing most of the Atlantic coast of North America. ARC stock may
be characterized by the loss of some rare alleles and by allele
frequencies significantly different from the wild at several enzyme
loci (Dillon and Manzi 1987). The difference is most striking at
the glucose phosphate isomerase (GPI) locus. Dillon and Manzi
(1987 and 1989b) reported that a sample of 213 M. mercenaria
wild collected in the vicinity of Charleston showed a frequency of
0.038 for an allele they designated "GPI 70" and did not show a
"GPI 60" allele. Their estimates for the ARC broodstock (N =
1 10) were 0. 195 for GPI 70 and 0.027 for GPI 60. Adamkewicz
et al. (1984) have verified that the inheritance of GPI isozyme
phenotype is Mendelian in Mercenaria species.
A striking shell color polymorphism, notata, may also be used
as a marker for ARC contribution to natural stocks. Chanley
(1961) reported that this trait is controlled by a single allele in-
herited in Mendelian fashion. Notata homozygotes are dark "red"
colored with white longitudinal bands, heterozygotes have thin
zig-zag stripes, and clams homozygous for the normal allele are
uncolored. Notata coloration has been selectively bred into ARC
stocks as a marketing tool. In their survey of heterozygosity and
growth in hatchery stocks of Mercenaria species, Dillon and
Manzi (1988) sampled 248 pure ARC clams aged 1 year. An
examination of 239 of these individuals showed 16 homozygotes
487
488
Metzner-Roop
and 108 heterozygotes, for a gene frequency of 0.293. The gene is
quite rare in the wild, however. Eleven South Carolina populations
showed from 0.71 to 2.17% notata heterozygotes, with a mean
gene frequency of 0.006 (Eldridge et al. 1976, Humphrey and
Walker 1982).
The ARC stocks of M. mercenaria have been selectively bred
for improved growth under aquaculture conditions. Released to the
wild, such clams might be able to reach reproductive maturity
faster than their neighbors and release propagules earlier, giving
ARC offspring an advantage over other larvae in terms of space
and food acquisition. Therefore, placing ARC genes into the nat-
ural gene pool might be considered advantageous to humans, as it
has been in "stock enhancement programs" for other species in
the past (Helle 1981, Strand and Lavan 1990, Gaffney and Allen
1992). On the other hand, the movement of the New England M .
mercenaria genome may model other accidental introductions that
have, in the past, proved catastrophic (Groves and Burdon 1986,
Mooney and Drake 1986. Strayer 1991).
In this investigation, I examine the evidence that aquaculture
operations have affected the genetics of natural Mercenaria pop-
ulations in the Charleston area. Data on GPI and notata frequen-
cies at sites where ARC clams were formerly cultured by Trident
Seafarms will constitute a baseline against which future studies
can assess any long-term genetic consequences of aquaculture.
MATERIALS AND METHODS
The four samples locations are shown in Figure 1 . Site 1 is the
site of the former Trident Seafarms, where ARC clams (and sev-
eral other aquaculture stocks) were held from 1980 to 1989. From
March 21 to June 2, 1992, I sampled in the drainage ditch where
water flowed from the nursery system and holding tanks into the
Folly River. Site 2 is located in Bass Creek at Kiawah Island,
where clams where held in grow-out units from 1982 to 1985. I
sampled clams from this site between December 4, 1992, and
March 1 1, 1993. At Little Oak Island, site 3, clams were planted
from 1980 to 1981, and I sampled here from December 8, 1992,
to March 23, 1993. Some data from this site were from McMillan
(1993). Site 4 was a mud flat on Cole Island. 2.75 km from site 1 ,
4 km from site 2. and 6 km from site 3. I sampled here from July
23 to November 18, 1992. Although no aquaculture operations are
known to have taken place at site 4, it seemed possible that intro-
Figurc 1, Study area, showing collection sites 1 to 4 and site of 1986 control collection (C).
Aquaculture Effects on Mercenaria
489
duced genomes might, in one generation, spread here from site 1
or 2.
At sites 1 to 3, I collected about 600 clams, half "large" and
half "small." The large sample included individuals with maxi-
mum shell dimension greater than 50 mm. and the small sample
included individuals 50 mm or less. At site 4, I sampled over 300
clams, all small. My sample size of 300 was chosen in order to be
95% certain of obtaining a statistically significant contingency
chi-square (a = 0.05), given the control frequency of GP1 70
estimated by Dillon and Manzi (1989b) and a 10% contribution by
the ARC genome. Each small individual collected was scored as to
the presence of notata markings. Because shells tend to wear and
darken with age, I did not score large individuals for coloration.
Methods for horizontal starch gel electrophoresis were gener-
ally those of Dillon (1985 and 1992). I ground frozen cuttings of
siphon tissue and resolved isozymes in the supernatant by electro-
phoresis in an N-(3-aminopropyl) morpholine (pH 6) buffer sys-
tem. Gels were stained for GPI (E.C. 5.3.1.9) by an agar overlay
technique.
The 213 M. mercenaria serving as GPI controls for this study
were collected by Dillon and Manzi (1989b) near the origin of
Cole Creek on November 12, 1986 (Fig. 1). This site was located
5 km distant from the nearest known aquaculture site and seems to
show no significant divergence from wild Virginia or Massachu-
setts Mercenaria populations at the GPI locus. Because their av-
erage standard shell length was 75.4 mm, many of these individ-
uals may have been born before the establishment of Trident Sea-
farms. Humphrey and Walker's (1982) estimate of 0.006 for the
natural frequency of the notata gene in South Carolina, based on
N = 1,539 individuals pooled from 11 sites, was used as the
notata control.
My first statistical tests were for temporal heterogeneity in GPI
allele frequency. I combined alleles into two categories: the mark-
ers GPI 60 + GPI 70 and all others. Then at each of the sites 1 to
3, 1 compared gene frequencies in the large clams with those in the
small clams using Fisher's exact tests (two tailed). Where no sig-
nificant heterogeneity was detected, samples of large and small
individuals were combined. In my second set of tests. I compared
GPI allele frequencies obtained at all four of the sites to the control
frequencies of Dillon and Manzi (1989b). I used Fisher's exact
tests on the one-tailed hypothesis that present frequencies of GPI
60 + GPI 70 are greater. My third set of tests were also one tailed,
assuming that the present frequencies of the notata allele at each of
the four sites are greater than the control frequencies derived by
Humphrey and Walker (1982). Here, I used chi-square contin-
gency tests, with Yate's correction. My final tests were for dis-
equilibrium between notata and GPI marker alleles. All individ-
uals were categorized by the presence or absence of GPI 70 or GPI
60 (together) and by the presence or absence of notata coloration.
Then, the significance of the composite linkage disequilibrium
(AAB) was tested using chi-square (program LD86.FOR of Weir
1990).
RESULTS
Figure 2 shows the size distributions (standard length of shell)
of the Mercenaria species sampled for this study. Note that these
are not random samples of the population at each site. Because
smaller clams are more difficult to find, a 300 individual sample of
clams smaller than 50 mm taken from the wild will be strongly
cm
O
O
.2
a
250
150
50
250
150
50
2. Bass Creek at
Kiawah Island
3 250
u
x>
a
3
150
50
250
150
50
4. Cole Island
20 40 60 80 100 120
Standard Length (mm)
Figure 2. Size distribution (standard shell lengths) of Mercenaria spe-
cies sampled at four sites in the vicinity of Charleston, SC. Individuals
are graphed by 10 mm increments; the size given is the maximum for
each category.
biased to the 40 to 50 mm size class. The GPI gene frequencies
observed at the four sites are shown in Table 1 . Combining alleles
into marker (GPI 60 + GPI 70) and nonmarker classes, Fisher's
exact tests uncovered no significant differences between large and
small clams at any site. Two-tailed values of p were 0.76, 0.36,
and 0.85 for sites 1, 2, and 3, respectively.
There is no evidence of ARC contribution to the stocks of
Mercenaria species currently inhabiting the four sites, judging by
490
Metzner-Roop
TABLE 1.
Gene frequencies and sample sizes (N) at the GPI locus in large (L)
and small (S) M. mercenaria from the four sites shown in Figure I.
Site 1
Site 2
Site 3
Site 4
Alleles
L
S
L
S
L
S
S
110
0.020
0.024
0.003
0.005
0.000
0.003
0.022
105
0.006
0.006
0.018
0.022
0.021
0.028
0.002
100
0.910
0.900
0.915
0.893
0.866
0.859
0.920
90
0.021
0.023
0.028
0.030
0.063
0.055
0.027
80
0.009
0.017
0.023
0.032
0.029
0.031
0.019
70
0.029
0.020
0.007
0.017
0.017
0.024
0.006
60
0.005
0.010
0.005
0.002
0.004
0.000
0.003
N
328
351
301
300
350
309
314
allele frequencies at the GPI locus. Counter to expectation. Table
1 shows that the combined frequency of marker alleles was lower
at all of the sites than the control 0.037 observed by Dillon and
Manzi (1989b). The incidences of notata coloration in my four
samples of small clams are given in Table 2. None of these is
significantly greater than the control frequency (one-tailed p =
0.08, 0.34, 0.21 , and 0. 1 1 . respectively).
However, two of the small individuals collected at site 3 (Little
Oak Island) carried both notata coloration and the marker allele
GPI 70. The composite disequilibrium coefficient (AAB) was
0.0028, highly significant (chi-square = 10.74, one-tailed p =
0.0005). Such disequilibrium between notata and GPI markers is
consistent with the expectation from an ARC contribution to the
Mercenaria population inhabiting site 3.
DISCUSSION
At least some escape of foreign gametes, larvae, juveniles, and
even adults would seem almost inevitable during the course of
normal commercial bivalve aquaculture. The water outfall from
the Trident Seafarms facility would be expected to carry veligers
from uncontrolled broodstock spawning and early juveniles from
the nursery. "Seed" clams were certainly lost in transfer to pens,
and older individuals were lost as the result of pen failure, espe-
cially during Hurricane Hugo. To the contribution by direct escape
would be added gametic gene flow. ARC clams may mature at 20
mm, well before normal harvest, and seem to spasm under con-
ditions similar to native populations, as temperatures reach 25°C.
Thus, one might reasonably expect hybridization between caged
and native clams to be commonplace.
I was, however, almost entirely unable to detect elevated fre-
TABLE 2.
The frequency of the notata phenotype in clams less than 50 mm at
the four sites, the gene frequencies implied, and the incidence of
marker GPI alleles.
notatas
Gene showing
Site/Location N notatas Frequency GPI 60 or 70
1 . Clam Farm
351
8
0.0114
0
2. Kiawah Island
300
5
0.0083
0
3. Little Oak Island
309
6
0.0097
2a
4. Cole Island
314
7
0.0111
0
a Significant linkage disequilibrium.
quencies of ARC marker alleles in the Mercenaria population
currently inhabiting the estuaries around Charleston, over all sites
and size classes examined. Either the ARC genome has generally
proved to be disadvantageous in the wild, or it has by now become
diluted beyond my ability to detect it.
1 did obtain one significant result. The apparent linkage dis-
equilibrium between notata and GPI 70 at site 3 (Little Oak Island)
constitutes some indirect evidence of ARC contribution to natural
populations. However, it should be remembered that I examined
by data 12 times in my effort to detect this phenomenon: four
direct tests for GPI 60 + 70, four direct tests for notata, and four
tests for disequilibrium between the two markers. Because Bon-
ferroni correction would require an adjusted value of p = 0.05/12
= 0.004 for significance, my single significant result (p =
0.0005) may be attributable to type I statistical error.
Some further survey of the Mercenaria population around Lit-
tle Oak Island would seem warranted in the future. However,
viewed over the entire Charleston-area population of M. merce-
naria, the genetic effect of aquaculture from 1980 to 1989 would
seem to have been negligible.
ACKNOWLEDGMENTS
Major contributions of time and intellect were provided by
R. T. Dillon, P. M. Gaffney. T. Roop, E. W. McMillan, N. H.
Hadley. N. S. Dayan, and A. R. Wethington. Others providing
assistance: J. Bennett, C. Chollet, D. D'Emeck, J. Dwyer, C.
Hall, J. J. Manzi, P. Powers, D. Reynolds. C. Sears. M. Shif-
man, C. Walton, and anonymous reviewers. This research was
funded by grants from the Slocum Lunz Foundation and the Con-
chologists of America. It is based on a thesis submitted in partial
fulfillment of the requirements for an M.S. in Marine Biology
from the University of Charleston, SC.
LITERATURE CITED
Adamkewicz. L . S. R. Taub & J. R. Wall. 1984. Genetics of the clam
Mercenaria mercenaria. 1. Mendelian inheritance of allozyme varia-
tion. Bioehem. Genet. 22:215-219.
Brown, J. W., J. J. Manzi, H. Q M. Clawson & F. S. Stevens. 1983.
Moving out the learning curve: An analysis of hard clam, Mercenaria
mercenaria. nursery operations in South Carolina. Mar. Fish. Rev.
45:10-15.
Chanley, P. E. 1961. Inheritance of shell markings and growth in the hard
clam, Venus mercenaria. Proc. Nail. Shellfish Assoc. 50:163-169.
Dillon, R. T., Jr. 1985. Correspondence between the buffer systems suit-
able for electrophoretic resolution of bivalve and gastropod isozymes.
Comp. Bioehem. Physiol. 82B:64 3-645.
Dillon, R. T., Jr. 1988. Evolution from transplants between genetically
distinct populations of freshwater snails. Genetica 76:1 1 1-1 19.
Dillon, R. T , Jr. 1992. Electrophoresis IV: nuts and bolts. World Aqua.
23:48-51.
Dillon, R. T , Jr. & J. J. Manzi. 1987. Hard clam, Mercenaria merce-
naria, broodstocks: genetic drift and loss of rare alleles without reduc-
tion in heterozygosity. Aquaculture 60:99—105.
Dillon. R. T, Jr. & J. J. Manzi. 1988. Enzyme heterozygosity and growth
Aquaculture Effects on Merc en aria
491
rate in nursery populations of Mercenaria mercenaria (L). J. Exp.
Mar. Biol. Ecol. 116:79-86.
Dillon. R T.,Jr. &J. J Manzi. 1989a. Genetics and shell morphology of
hard clams (Genus Mercenaria) from Laguna Madre. Texas. Nautilus
103:73-77.
Dillon. R. T., Jr. & J. J. Manzi. 1989b. Genetics and shell morphology in
a hybrid zone between the hard clams Mercenaria mercenaria and M .
campechiensis. Mar. Biol. 100:217-222
Eldndge. P. J„ W. Waltz & H. Mills 1976. Relative abundance of Mer-
cenaria mercenaria nolala in estuaries in South Carolina. Veliger 18:
396-397.
Gaffney, P. M. & S. K. Allen. Jr 1992 Genetic aspects of introduction
and transfer of molluscs. J. Shellfish Res 1 1:535-538.
Groves. R. H. & J. J. Burdon. eds. 1986. Ecology of Biological Inva-
sions. Cambridge University Press. Cambridge. England
Helle. J. H. 1981. Significance of the stock concept in artificial propaga-
tion of salmonids in Alaska. Can. J. Fish Aquat. Sci. 38:1665-1671.
Hindar. K . N. Ryman & F. Utter. 1991. Genetic effects of cultured fish
on natural fish populations. Can J Fish. Aquat. Sci. 48:945-957.
Humphrey. C. M. & R. L. Walker. 1982. The occurrence of Mercenaria
mercenaria form Nolala in Georgia and South Carolina: calculation of
phenotypic and genotypic frequencies. Malacologia 23:75-79.
Krueger. C. C. A. J. Gharrett. T. R. Dehnng & F. W. Allendorf. 1981.
Genetic aspects of fisheries rehabilitation programs Can. J. Fish.
Aquat. Sci. 38:1877-1881.
Manzi. J. J.. N. H. Hadley. C. Battey. R. Haggerty. R. Hamilton & M.
Carter. 1984. Culture of the northern hard clam Mercenaria merce-
naria (Linne'l in a commercial-scale, upflow, nursery system. J. Shell-
fish Res. 4:119-124
Manzi, J. J.. N. H. Hadley & M. B. Maddox. 1986. Seed clam. Merce-
naria mercenaria, culture in an experimental-scale upflow nursery sys-
tem. Aquaculture 54:301-31 1
McCay, B. J. 1990. Social and cultural dimension of the movement of
mollusks. J. Shellfish Res. 8:466.
McMillan, E. W. 1993. Comparison of genetic vanation among local
demes of three bivalve species: Crassostrea virginica, Mercenaria
mercenaria. and Geukensia demissa. Masters Thesis, University of
Charleston. SC. 42 pp.
Mooney. H. A. & J. A. Drake, eds. 1986. Ecological Studies: Analysis
and Synthesis, vol. 58. Ecology of Biological Invasions of North
America and Hawaii. Springer- Verlag. New York.
Stevens. F. S.. I. J. Manzi & H. Clawson. 1984. Development of field
growout techniques for the northern hard-shell clam Mercenaria mer-
cenaria (Linnel in South Carolina. J. Shellfish Res. 4:101.
Strand. I. E. & E. A. Lavan. 1990. The economics of mollusc introduc-
tion and transfer: history and future of private and public decisions. J.
Shellfish Res. 8:466.
Strayer. D. L. 1991 Projected distribution of the zebra mussel. Dreissena
polymorpha. in North America. Can. J. Fish. Aquat. Sci. 48:1389-
1395.
Verspoor, E. 1988. Reduced genetic variability in first-generation hatchery
populations of Atlantic salmon tSulmo salar). Can. J . Fish. Aquat.
Sci. 45:1686-1690.
Weir, B. S. 1990. Genetic Data Analysis. Sinauer Associates, Sunder-
land, MA.
Journal of Shellfish Research, Vol. 13, No. 2.493-501, 1994.
INTERNAL SHELL STRUCTURE AND GROWTH LINES IN THE SHELL OF THE ABALONE,
HALIOTIS M1DAE
J. ERASMUS, P. A. COOK, AND N. SWEIJD
Zoology Department
University of Cape Town
South Africa
ABSTRACT The internal shell structure of the South African abalone. Haliolis midae, was studied using scanning electron micros-
copy. Growth lines were counted. The shell consists of the periostracum, prismatic, and nacreous layers. The outer, noncalcified
periostracum consists of a single, honeycombed layer. Calcified layers include a simple prismatic and a nacreous layer, the latter being
composed of sheet nacre, which is uncharacteristic of gastropods. Growth lines were deposited in the nacreous layer of animals, even
under controlled laboratory conditions. In animals of known age, three lines were deposited in the first year and one in each subsequent
year. Rate of deposition of lines was not related to diet, temperature, or photoperiod.
KEY WORDS: abalone. shell structure, growth lines
INTRODUCTION
Age and growth rate of commercially exploited molluscan
stocks are important factors in the determination of the exploita-
tion status of such stocks (Ricker 1977). Growth rate is also an
important consideration in molluscan mariculture because of the
relationship between growth and production costs (Day and Flem-
ing 1992). Although aging techniques and growth rates of haliotid
populations have been extensively studied (e.g., Sainsbury 1982,
Prince et al. 1988, Day and Fleming 1992, Nash 1992), no com-
pletely reliable and validated aging techniques have been pub-
lished (Ward 1986).
Growth checks appearing on the external shell surface have
been used to estimate age in various abalone species (Forster 1967,
Poore 1972, Kojima et al. 1977, Hayashi 1980. Saito 1981, Kim
and Cheung 1985, Prince et al. 1988), but age estimates resulting
from such studies have been extremely variable. Although several
of the listed studies reported difficulty in interpreting external
growth marks because of variations in the timing of the formation
of such marks or because of shell damage caused by boring or-
ganisms (Forster 1967). the majority concluded that growth
checks, at least in larger animals, were deposited annually. Mu-
noz-Lopez (1976). however, stated that external growth checks
did not always occur in abalone, whereas Shepherd and Hern
(1983) noted that checks were not necessarily annual.
An alternative method of aging abalone shells was developed
by Munoz-Lopez ( 1 976) , who ground the tip of the spire to expose
alternate light and dark layers of concholin and aragonite. He
found that, in the Mexican abalone Haliotis corrugata and Haliotis
fulgens, concholin was formed during winter whereas aragonite
was formed during the summer. Although he concluded that layers
were annual, no independent growth data were presented to vali-
date this assumption. Prince et al. (1988) related similar layers to
independent growth data for a Tasmanian population of Haliotis
rubra and concluded that three minor growth lines were deposited
during the first 16 months of life and that, thereafter, lines were
deposited approximately annually. This aging method has not,
however, found universal acceptance, and Prince et al. (1988), for
example, noted that it could not be applied to populations of H.
rubra from warmer waters of New South Wales. Beamish and
McFarlane (1983) stressed the importance of validating all aging
techniques with independent growth data for each species and each
area studied, whereas McShane and Smith (1992) stated that the
use of shell growth checks as a method for age determination of//.
rubra was not reliable because of variation in frequency of growth
checks between populations and the unreliability of growth checks
in the shells of individuals that had been infested with boring
organisms. Schiel and Breen (1991) also cast doubts on the use-
fulness of growth lines, but Shepherd et al. ( 1994) suggested that,
despite the variation that occurred in the number of growth lines
deposited annually, counts could still be a useful aid to aging
shells, particularly if loss of lines from abrasion could be taken
into account.
Working on ocean quahogs and using methods developed by
Kummel and Raup (1965), Ropes (1984) suggested that some of
the ambiguities associated with counting molluscan shell lines
could be reduced by cutting complete longitudinal sections of
shells and counting lines from acetate peels. Thompson et al.
(1980) used the same method to age shells of ocean quahogs and
provided independent growth data to validate the technique.
Mutvei et al. (1985) also commented on the contradictory data
present in the literature on the interpretation of molluscan growth
lines and suggested that, in the case of haliotids, this could be
partly explained by extremely variable rates of shell mineraliza-
tion. There data showed that the pallial layer of the abalone mantle
was capable of secreting aragonite and calcite simultaneously and
that rates of mineralization could change rapidly during growth.
Because shell ultrastructure can vary markedly between species,
each species requires its own description of shell structure before
shell checks can be used to estimate age or growth rates.
Growth rate of the South African abalone Haliotis midae was
estimated previously by Newman (1968), who measured growth,
over a 2 year period, in animals that had been tagged, returned to
the sea, and then recaptured. Tarr ( 1990 and 1993), having listed
several criticisms of this work, reassessed growth rates for this
species, using a more reliable mark-recapture method. No attempt
has, however, been made to validate the use of growth lines to age
shells of H. midae. The aim of this study, therefore, was to ex-
amine the ultrastructure of the shells of//, midae, using scanning
electron microscopy (SEM) in order to evaluate the use of acetate
peels to count growth lines as an aging technique for these shells.
MATERIALS AND METHODS
With the exception of the 37 month old animals, all shells used
in this investigation were from animals reared in an experimental
493
494
Erasmus et al.
hatchery in Sea Point. Cape Town. Therefore, the exact age of the
animals was known. In the case of the 37 month old shells, ani-
mals were collected from the wild at a mean length of 10 mm and
an estimated age of 6 months (Newman 1968) and were thereafter
transferred to the experimental hatchery for the rest of the exper-
iment. Shells were used to investigate the ultrastructure of shell
layers, to elucidate the relationship between age and the number of
growth lines, and to investigate the possible influence of some
environmental variables (e.g., photoperiod, food availability) on
the deposition of growth lines. All hatchery animals were kept at
a constant temperature of 18 ± 0.5°C. Other experimental condi-
tions for each group, listed in Table 1 , were maintained throughout
the life of each group from the time of settlement to the time of
sampling.
The area of the shell used in this study was the nacreous layer
(Fig. la), because according to Lutz (1976). preservation of
growth lines is better in the nacreous than in the prismatic layer.
The area used within the nacre of H . midae was the section from
the apex of the shell to the outer lip (X-Y. Fig. lb). This area was
chosen because lines near the margin of the shell (Z, Fig. lb) often
condense, making interpretation difficult (Lutz 1976), and because
the shell grows radially from the apex so that all growth lines are
contained within the area X-Y (Lutz and Rhoads 1980).
Acetate Peels
Sections of the shell were cut to give a flat surface along the
axis from X to Z (Fig. lb). The procedure used to produce acetate
peels was that proposed by Ropes (1984) for the ocean quahog
Arctica islandica, except that polished abalone shells were etched
for 25 seconds with 1% HC1. Acetate peels were viewed under a
light microscope to examine the internal structure of the whole
shell and the growth lines. Growth line counts were accepted only
after at least two of three separate investigators counted the same
number of lines.
Because peels were made from an entire season of shell, they
provided a range means to determine the location and arrangement
of the different structural layers (Wise and Hay 1968). Specific
areas were selected for detailed SEM study on the basis of the
acetate peels.
Scanning Electron Microscopy
SEM was used to investigate the fine ultrastructure of the
growth lines, nacre, prismatic, and periostracum. The perios-
tracum was not clearly visible in acetate peels because it is a thin,
TABLE 1.
Experimental conditions of groups.
Periostracum
Group
Age
(months)
Light (L):Dark (D) Food" Nh R (SD)'
Prismatic
Nacre
Figure la
Lip-X
A
B
C
D
E
F
G
13
23
37
37
37
37
Ambient
E
4
1 (0)
Ambient
E
10
3.2 (0.4)
Ambient
E
4
4.5 (0.6)
Shell Structure
24 h dark
12 h L:12hD
E
E
7
7
5.1 (1.2)
5.7 (0.5)
Acetate Peels
Ambient
Ambient
E
L
5
5
5.2 (0.8)
6.0(1.0)
An acetate
:incx ( Y) (Fit'
Figure lb
Figure 1. (a) Layers of the shell of H. midae and area in nacre where
growth lines were counted, (b) Regions of the shell to show where each
section was cut.
noncalcified layer. SEM also shows more relief and spatial rela-
tions of the shell crystals than the acetate peels (Wise and Hay
1968).
Shells that had been used previously for the preparation of
acetate peels were cleaned in an ultrasonic bath and prepared for
SEM following the procedure described by Wise and Hay ( 1968).
Shells were mounted onto stubs and coated with 20 u.m of gold
palladium using a Polaron sputter coater (Polaron Equipment.
Ltd.. Watford, U.K.) Cambridge S-200 microscope (Cambridge
Instruments, Cambridge, U.K.) S2 100 sputter coater. They were
then scanned and photographed under a Cambridge S-200 micro-
scope .
RESULTS
' E, diet of Ecklonia maxima; L. diet of Laminaria pallida.
h Number of abalone per treatment.
1 Mean number of lines (±standard deviation |SD]).
An acetate peel (Fig. 2) of the region from the lip (X) to the
apex (Y) (Fig. lb) of the shell, shows the three shell layers that are
common in molluscs. The aragonitic crystals of the nacre are
deposited parallel to each other between X and Y (Fig. 2) and
obliquely between Y and Z (Fig. 3a and b). The change in depo-
Shell. Structure and Growth Lines
495
Figure 2. Acetate peel between X and Y (Fig. lb) of a shell showing three layers. Aragonite crystals, in nacre, are deposited in parallel layers.
Pe, periostracum; P, prismatic; N, nacre; G, growth line. Original magnification, x40.
sition could possibly be to strengthen the shell (R. W. Day. per-
sonal communication). Occasionally, a parallel layer of aragonite
was seen and it is possible that this was deposited after a predator
attempted to bore into the shell. The difference in the deposition of
the aragonite indicates that the nacre is growing differently be-
tween XY and YZ. The aragonite grows progressively thicker
from the lip to the apex as the parallel layers are deposited on top
of one another. The nacre grows wider from the apex to the grow-
ing edge (YZ) as the oblique layers are deposited.
Scanning Electron Microscopy
Further details of shell structure were apparent from the SEM.
The nacre is composed of tabular blocks of aragonite (Fig. 4). This
type of nacre is known as sheet nacre (Clark 1974), because the
aragonitic tablets resemble a brick wall. Spaces between the
blocks of aragonite are filled with organic matrix; this was not
examined here. The growth line has a somewhat different structure
from the nacre (Fig. 4), the tabular blocks of aragonite appearing
to be interspersed with deposits of calcium (Lutz and Rhoads
1980). The growth lines are continuous through the nacre and have
a higher topography than the surrounding nacre.
The lip of the shell is also composed of tabular aragonite (Fig.
5), similar to the nacre. Growth lines in the lip have a higher
topography than the surrounding aragonite, probably a result of
differential etching. Growth lines in the lip are not continuous and
could be a reflection of tides or daily rhythms (Lutz and Rhoads
1980).
The prismatic layer in H. midae has a typical prism structure
(Fig. 6) and is classified as simple prismatic because calcitic
prisms are composed of a stack of disc-shaped crystals deposited
on top of one another (Watabe 1984). The calcitic crystals develop
from spherulites seen on the undersurface of the periostracum
(Fig. 7). The periostracum itself is a tanned protein layer which
has a honeycomb appearance (Fig. 8).
Validation of the Acetate Peel Aging Technique
Acetate peels were used to count the number of growth lines in
each shell. Growth lines on acetate peels are delimited by dark
lines that run from the lip (X) of the shell to the apex (Y) (Fig. lb).
A summary of the mean number of lines for shells of known age
is given in Table 1 . Shells that were 8 months old had one growth
line, whereas 13 month old shells had an average of 3.2 lines. By
26 months, shells had deposited an average of 4.5 lines, and by 37
months, an average of 5.5 lines had been laid down.
Data on the number of growth lines counted in the shells of the
37 months old animals were tested for a normal distribution, and
a one-way analysis of variance was performed to determine wheth-
er there were significant differences between the number of growth
lines deposited under the different experimental conditions (groups
D, E, F, and G), each group being compared with every other
group. The result showed that there were no significant differences
between groups (95% confidence limits). This showed that growth
lines were deposited under conditions of constant temperature and
that differences in diet and photoperiodicity did not affect the
number of growth lines formed.
DISCUSSION
Mollusc shells vary in structure and composition, differences
being most evident when bivalves are compared with gastropods
496
Erasmus et al.
Figure 3. (a) Acetate peel of a shell from Y to Z (Fig. lb). Aragonite is deposited in oblique layers. N, nacre; A, oblique layer of aragonite.
Original magnification, x40. (b) Higher magnification of layers of aragonite shown in panel a. The oblique deposition can be clearly seen.
Original magnification, X60.
(Watabe 1984). The mollusc shell has two major components,
namely, the noncalcified periostracum and the calcified inner lay-
ers. This SEM study has shown that the periostracum of H. midae
has a honeycomb appearance, similar to that described by Saleud-
din and Petit (1983) for other species. Of the five molluscan cal-
cified layers described by Clark (1974), only two, the prismatic
and the nacreous layers, were observed in H. midae (Fig. 4).
Mutvei et al. (1985) examined the shell structure of eight spe-
Shell Structure and Growth Lines
497
Figure 4. Nacre of shell and growth line. Nacre is composed of tabular blocks of aragonite, and growth line is a deposit of calcium. A, tabular
aragonite; G. growth line. SEM, original magnification, x895.
cies of abalone and found the prismatic layer of each of the eight
species to be different, being composed of calcite, aragonite, or a
combination of the two. The fact that the pallial epithelium in the
mantle, which secretes the prismatic layer, is capable of secreting
calcite and aragonite simultaneously and that the rates of miner-
alization can change rapidly during growth (Mutvei et al. 1985)
could partially explain why the prismatic layer could be so differ-
ent among species of abalone.
In the gastropod family Liltorinidae , variations in the structure
of the prismatic layer have been attributed to differences in water
temperature, which could influence the physiological control of
shell mineralogy (Taylor and Reid 1990). Those authors suggested
that prismatic shell could be calcific at lower temperatures and
aragonitic at higher temperatures. Whether the same reasoning can
be applied to explain differences in prismatic composition in ab-
alone is unknown at present. The structure of the prismatic layer of
H. midae is consistent with the structure of the simple prismatic
described by Watabe (1984).
There are three different types of nacre found in mollusc shells,
all of which are composed of aragonite (Clark 1974) and consist of
numerous horizontal lamellae deposited on top of one another
(Watabe 1984). Generally, gastropods have a columnar nacre and
bivalves have a sheet nacre, but an exception to this is the gas-
tropod Cirrarium pica, which has a sheet nacre (Wise and Hay
1968). Columnar nacres were described in an unidentified haliotid
species by Mutvei ( 1978) and in Haliotis rufescens by Nakahara et
al. (1982). This study shows that the nacre of H. midae, being a
sheet nacre, is uncharacteristic of the usual gastropod pattern (Fig.
4). The spaces around the tabular blocks of aragonite are filled
with organic matrix, and such spaces can also be seen in Figure 4.
The nacre of Mollusca may contain growth lines that, accord-
ing to Koike (1980), are zones of high calcium concentration and
low sodium concentration. Although Koike (1980) was working
on growth bands of Meretrix lusoria that resulted from daily or
tidal rhythms, it is possible that the bands she described are similar
in structure to the annual rings described here, that is. raised strips
of shell matter that were left by HC1 etching, the surface material
being more easily eroded than the growth lines. The growth lines
in the lip of the shell of H. midae are higher than the surface relief
of the nacre (Fig. 7). suggesting that they are also less easily
etched than the surface material. This could be because of a higher
concentration of organic matter in the growth lines.
The width of the growth lines in abalone shells is variable
(Watabe 1984), being controlled by the rate of deposition of the
nacre, which precedes the formation of the line. Watabe (1984)
attributed the change in width to seasonal variation in the envi-
ronment. This suggestion is supported by results of this study, in
which the widths of growth lines in shells from animals grown at
constant temperature in the experimental hatchery were uniform.
This study has shown that growth lines in the nacre of H. midae
can be visualized and counted by using the acetate peel method.
Studies on counting growth rings, in acetate peels, have been
validated previously by comparison with tagging data, but the
disadvantage of this method is that, as suggested by Day and
Fleming ( 1992), tags on shells could influence the growth rate of
the animals or add disturbance rings. In this study, individual
Figure 5. Lip of a shell, composed of aragonite, and growth line of calcium. A, aragonite; G, growth line. SEM, original magnification, x 5,790.
Figure 6. Prisms in prismatic region. Transverse striatums in the crystals. P, prismatic; T, transverse striatums. SEM, original magnification. x807.
Figure 7. Under surface of periostracum. Antrum (A), on which spherulites (S) form. SEM, original magnification, x5,670.
Figure 8. Honeycomb structure of periostracum. SEM, original magnification, x 1,370.
500
Erasmus et al.
shells were not tagged but were grown under controlled conditions
in an experimental hatchery and were, therefore, of known age.
The influence of variations in temperature, diet, and light on the
rate of deposition of growth lines could, therefore, be directly
assessed. Results showed that none of these factors influenced the
rate at which growth lines were deposited. The results are similar
to those described by Richardson (1988), who although working
on microgrowth bands of possibly daily or tidal origin, found that
deposition of growth bands in the clam Tapes philippinarum was
not related to environmental factors. The clams were placed under
different constant conditions (temperature, water flow, light, and
diet), and none of the factors influenced band formation. Richard-
son ( 1988) concluded that the bands were deposited because of an
innate rhythm, which was controlled by the growth of the clam.
A possible factor that could influence formation of growth lines
in abalone is related to the spawning season. Sakai (1960) and
Forster (1967) found that growth lines in Haliolis discus and Ha-
liotis tuberculata were deposited when the animals spawned.
Thompson et al. (1980) suggested that this resulted from energy
being channeled into gonad production and not growth. In this
study, however, growth ring deposition cannot have been related
to spawning because all animals were subadults and none had
reached spawning age.
From the results obtained in this study, the most likely pattern
of growth line deposition in H. midae appears to be very similar to
that described by Prince et al. (1988) for H. rubra, that is, the
deposition of three lines during the first year, followed by annual
deposition thereafter. Growth line deposition was not, however,
directly related to environmental variables, and it seems likely,
therefore, that it could be controlled by an endogenous rhythm
related to the growth cycle of the animal. Carter ( 1980) reviewed
the relation between shell microstructure and mechanical proper-
ties, and it is possible that the laminated shell structure that results
from the deposition of growth lines and from the alternate layers of
calcite and aragonite may enhance shell strength and help to lo-
calize or deflect shell fractures.
LITERATURE CITED
Beamish. R. J. & G. A. McFarlane. 1983. The forgotten requirement for
age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-
743.
Carter, J. G. 1980. Environmental and biological controls of bivalve shell
mineralogy and microstructure. pp. 69-113. In: D C. Rhoads and
R. A. Lutz (eds.). Skeletal Growth of Aquatic Organisms. Plenum
Press, New York.
Clark, G. R. 1974. Seasonal variations in shells of recent growth. Am.
Mus. Nat. Hist. Anthropoid. 56:161-179.
Day. R. W. & A. E. Fleming. 1992. The determinants and measurement
of abalone growth, pp. 141-164. In: S. A. Shepherd. M. J. Tegner
and S. A. Guzman del Proo (eds.). Abalone of the World. Blackwell
Scientific Publications, Oxford.
Forster, G. R. 1967. The growth of Haliotis tuberculata: results of tagging
experiments in Guernsey 1963-65. J . Mar. Biol. Assoc. U.K. 60:431-
437.
Hayashi, 1. 1980. Structure and growth of a shore population of the ornier.
Haliotis tuberculata. J. Mar. Biol. Assoc. U.K. 60:431-437.
Kim. J. W. & S. C. Cheung. 1985. On the growth of abalones, Sttlculus
diversicolor diversicolor ( Reeve I and S . diversicolor aquatilis (Reeve ) .
in Cheju Island Bull. Mar. Resour. Res. Inst. Jeju. Natl. Univ. 9:39-
50.
Koike. H. 1980. Seasonal dating by growth line counting of the clam.
Meretrix lusoria. Univ. Mus. Univ. Tokyo Bull. 18:1-25.
Kojima, H., Y. Nakahisa, H. Tanimoto & K. Isibasi. 1977. A study of the
stock of Japanese black abalone, Haliotis discus discus, in Tokushima
Prefecture. 1. Growth of shells. Bull. Tokai Reg. Fish. Res. Lab.
90:25-37.
Kummel, B. & D. Raup (eds.). 1965. Handbook of Paleontological Tech-
niques. W. H. Freeman, San Fransisco. 852 pp.
Lutz, R. A. 1976. Annual growth patterns in the inner shell layer of
Mytilus eduhs L. J. Mar. Biol. Assoc. U.K. 56:723-731.
Lutz, R. A. & D. C. Rhoads. 1980. Growth patterns within the molluscan
shell: an overview, pp. 203-254. In: D. C. Rhoads and R. A. Lutz
(eds.). Skeletal Growth of Aquatic Organisms. Plenum Press, New
York.
McShane, P. E. & M. G. Smith. 1992. Shell growth checks are unreliable
indicators of age in the abalone Haliotis rubra (Mollusca: Gastropodal.
Aust. J Mar. Freshwater Res. 43:1215-1219.
Munoz-Lopez, T. 1976. Preliminary results on a method for age determi-
nation in the abalone (Haliotis spp) from Baja California, pp. 1-301.
In: Proceedings of the First National Symposium on the Main Fishery
Resources of Mexico. Special volume on abalone and lobster. Inst.
Nacl. Pesca. Ensenada. Mexico.
Mutvei, H. 1978. Ultrastructural characteristics of the nacre in some gas-
tropods. Zool. Scripla. 7:287-296.
Mutvei. H.. Y. Dauphin & J. Cuif. 1985. Observations sur l'organisation
de la couche externe du test Haliotis (Gastropoda): un cas exceptionnel
de variabilite mineragique et microstructurale. Bull. Mus. Nam. Hist.
Nat. Pans 1:73-91.
Nakahara. H.. G. Bevelander & M. Kakei. 1982. Electron microscopic
and amino acid studies of the shell layers of Haliotis rufescens. Venus
{Jap. Jour. Malac). 41:33-46.
Nash, W. J. 1992. Age determination of blacklip abalone [Haliotis rubra)
in Tasmania, pp. 88-92. In: The Measurement of Age and Growth of
Fish and Shellfish. Australian Society of Fish Biology Workshop,
Lome. Bureau of Rural Resources Proceedings Number 12.
Newman. G. G. 1968. Growth of the South African abalone Haliotis
midae. Invest. Rep. Div. Sea. Fish. S. Afr., 64:1-24.
Poore, G. C. B. 1972. Ecology of New Zealand abalones Haliotis species
(Mollosca: Gastropoda) 3. Growth. N. Z. J. Mar. Freshwater Res.
6:534-539.
Prince, J. D., T. L. Sellers, W. B. Ford & S. R. Talbot. 1988. A method
for ageing the abalone Haliotis rubra (Mollusca: Gastropoda). Aust. J.
Mar. Freshwater Res. 39:167-175.
Richardson, C. A. 1988. Exogenous and endogenous rhythms of band
formation in the shell of the clam Tapes philippinarum (Adams et
Reeve. 1850). J. Exp. Mar. Biol. Ecol. 122:105-126.
Ricker. W. E. 1977. The histological development, pp. 1-12. In: J. A.
Gulland (ed.). Fish Population Dynamics. Wiley. New York.
Ropes, J. W. 1984. Procedures for preparing acetate peels and evidence
validating the annual periodicity of growth lines formed in the shells of
ocean quahogs, Arctica islandica. Mar. Fish. Res 46:27-35.
Sainsbury. K. J. 1982. Population dynamics and fishery management of
the paua, Haliotis iris. 1. Population structure, growth, reproduction
and mortality. N. Z J Mar. Freshwater Res. 16:147-162.
Saito, K. 1981. The appearance and growth of 0-year-old Ezo abalone.
Bull. Jap. Soc. Sci. Fish. 47:1393-1400.
Sakai, S. 1960. On the formation of the annual ring on the shell of the
abalone. Haliotis discus var. hannoi Ino. Tohoko J . Agricult. Res.
11:239-244.
Saleuddin. A. S. M. & H. P. Petit. 1983. The mode of formation and the
structure of the periostracum. pp. 227-231. In: A. S. M. Saleuddin
and K. M. Wilbur (Eds). The Mollusca. Vol. 4. Physiology. Part 1.
Academic Press, Inc., New York.
Shell Structure and Growth Lines
501
Schiel, D. R. & P. A. Breen. 1991. Population structure, ageing and
fishing mortality of the New Zealand ahalone Haliotis iris. Fish Bull.
U.S. 89:681-691.
Shepherd, S. A. & W. S. Heam. 1983. Studies on .southern Australian
abalone (genus Haliotis) IV. Growth of H. laevigata and H. ruber.
Aust. J. Mar. Freshwater Res. 34:461—475.
Shepherd. S. A., D. Al-Wahibi & A. R. Al-Azri. 1994. Recruitment and
growth of the Omani abalone. Haliotis mariae. Aust. J. Mar. Fresh-
water Res. (in press).
Tarr. R. J. Q. 1990. Aspects of reproductive biology of Haliotis midae.
pp. 2-6. In: Cook, P. (ed). Perlemoen Farming in South Africa. The
Manculture Association of Southern Africa, Cape Town, South Africa.
Tarr, R. J. Q. 1993. Stock assessment and aspects of the biology of the
South African abalone, Haliotis midae. MSc. thesis. University of
Cape Town.
Taylor, J. D. & D. G. Reid. 1990. Shell microstructure and mineralogy of
the Littorinidae: ecological and evolutionary significance. Hydrobio-
logia 193:199-215.
Thompson, I.. D. S. Jones & D. Dreibelbis. 1980. Annual internal growth
banding and life history of the ocean quahog Arctica islandica (Mol-
lusca: Bivalvia). Mar. Biol. 57:25-34.
Ward. T. J. 1986. Abalone biology and fisheries in Australia: a review of
research. Aust. Fish. Serv. Canberra. Aust. Fish. Paper No. 86 (12).
Watabe. N. 1984. Shell, pp. 448-472. In: J. Bereiter-Hahn. A. G. Ma-
toltsy and K. S. Richards (eds.). Biology of the Integument, vol. 1.
Invertebrates. Springer- Verlag, Berlin.
Wise, S. W. & H. Y. Hay. 1968. Scanning electron microscopy of mol-
luscan shell ultrastructures. Techniques for polished and etched sec-
tions. Trans. Am. Microsc. Soc. 87:411-418.
Journal of Shellfish Research. Vol. 13. No. 2. 503-505, 1994.
SURVIVAL, GROWTH, AND GLYCOGEN CONTENT OF PACIFIC OYSTERS, CRASSOSTREA
GIGAS (THUNBERG, 1793), AT MADEIRA ISLAND (SUBTROPICAL ATLANTIC)
M. J. KAUFMANN,'* M. N. L. SEAMAN,' C. ANDRADE,2 AND
F. BUCHHOLZ3
'instil itt fur Meereskunde
24105 Kiel, Germany
~Direccao Regional das Pescas
9000 Funchal, Portugal
3Biologische Anstalt Helgoland
27498 Helgoland, Germany
ABSTRACT Pacific oysters, Crassostrea gigas. were introduced to Madeira Island (subtropical Atlantic) and grown at depths of 10
to 25 m. There were no significant differences between oysters grown at 10, 15, and 25 m depth for any of the parameters analyzed.
Live weight and shell size increased significantly, whereas dry meat weight and condition declined significantly by 45 and 70%,
respectively. Glycogen content decreased by 90% within 5 weeks; this was not linked to gametogenesis. Overall mortality attained
73% in 5 months. The oysters' poor performance is attributed to a combination of stress factors.
KEY WORDS: bivalve, oyster, Crassostrea, introduction, aquaculture, condition
INTRODUCTION
The Pacific oyster, Crassostrea gigas (Thunberg), is being
grown with commercial success in a variety of environments
worldwide. In temperate zones, its growth decreases in winter
(e.g.. Askew 1972, Heral and Deslous-Paoli 1991). but under
subtropical conditions, continuous growth can be achieved during
the entire year (e.g., Hughes-Games 1977). At Madeira Island,
water temperatures are favorable at 17 to 23°C year around. Oys-
ters (Ostrea species) have been found occasionally (Abreu, per-
sonal communication), but there has never been any local fishery
or culture. A small-scale grow-out trial with imported Pacific oys-
ters, however, led to encouraging results (Waschkewitz. personal
communication) and motivated this present study.
MATERIALS AND METHODS
Healthy half-grown oysters, C. gigas, of British origin, that
had been grown in Flensburg Fjord (Western Baltic) from spat
size, were introduced to Madeira at a mean live weight of 9.6 g.
They were grown at 10, 15. and 25 m depth in six lantern nets with
1 cm mesh that were anchored at a depth of 27 m in a small bay
near Funchal Harbour. Each net was stocked with about 320 oys-
*Present Address: Universidade da Madeira, Largo do Colego, 9000 Fun-
chal, Portugal.
ters, at a density of less than 1 g cm 2 (as recommended by
Spencer 1990).
Every 2 or 3 weeks from July until December 1991. the water
temperature at the site was determined and 30 oysters from each
depth were sampled randomly by SCUBA diving. Dry weight was
determined after drying for 24 h at 60°C. and ash weight was
determined after incineration for 24 hours at 550°C. Glycogen was
determined after Keppler and Decker (1984); pieces of mantle
tissue of about 100 mg were excised from six randomly selected
live oysters in each sample, stored in micro-test tubes at -60°C,
and later transported on dry ice to IfM Kiel for analysis. To assess
the oysters' overall performance, the condition index (CI) of
Lawrence and Scott (1982), recommended by Bodoy et al. (1986)
and Crosby and Gale (1990), was calculated as: CI = (dry meat
weight x l,000)/(live weight - dry shell weight).
RESULTS
Sea water temperature varied between 24°C in August and
September and 19°C in December. There were no significant dif-
ferences between oysters grown at all three depths (Mann- Whitney
U test, 0.95 level), and the results for 15 and 25 m have been
pooled; the lantern nets at 10 m disappeared 8 weeks after the
beginning of the experiment and have been excluded from further
analysis.
There was an increase in live weight and shell size and a
simultaneous decrease in dry and ash-free dry meat weight of the
TABLE 1.
Growth (mean ± standard deviation) and mortality of oysters (C. gigas) during grow out at Madeira Island. Pooled data for oysters grown
at 15 and 25 m depth. N = 60 (week 21, N = 30).
Wet Meat
Dry Meat
Ash-Free Dry
Dry Shell
Cumulative
Weeks in
Shell Height
Live Weight
Weight
Weight
Meat Weight
Weight
Mortality
Culture
(mm)
(g)
(g)
(g)
(g)
(%)
0
41 ± 3
9.6 ± 1.7
2.4 ± 0.6
0.37 ± 0.14
0.31 ± 0.13
5.5 ± 1.1
7
43 ± 4
10.4 ± 1.8
2.2 ± 0.5
0.23 ± 0.08
0.16 ± 0.07
5.8 ± 1.1
39
14
47 ± 6
12.6 ± 3.3
2.5 ± 0.7
0.23 ± 0.07
0.15 ± 0.05
7.6 ± 2.1
55
21
47 ± 6
13.5 ± 4.0
2.4 ± 0.8
0.20 ± 0.06
0 12 ± 0.04
7.8 ± 2.4
73
503
504
Kaufmann et al.
a> ©■
m
■a
c
CO
c
a>
en
o
o
.>-
O
50
en 40
■
S in
Ash^--—
^
of dry me.
D O (
Condition
■
^
Glycogen
o
JUL
1 AUG
SEP ' OCT '
NOV ' DEC
100
80
60
40
20
0
■o
c
o
O
Growout period
Figure 1. Glycogen and ash levels (left scale) and condition index
(right scale) in oysters (C. gigas) during grow out at Madeira Island.
Pooled data for oysters grown at 15 and 25 m depth. N = 12 for
glycogen; N = 60 for ash and condition index (except December, where
N = 6 and N = 30, respectively).
oysters during the rearing period; wet meat weight remained con-
stant. The mortality exceeded 70% (data summarized in Table 1).
The ash content of the meat increased insignificantly in absolute
terms, but it more than doubled as a percentage of the dry meat
weight. The condition index and the glycogen level in the meat
declined, respectively, to less than 40% and less than 10% of the
initial values (Fig. 1).
DISCUSSION
We have no indications that the stock might have been dis-
eased, nor has there been any record of disease or mortality in the
original Baltic stock in the years before, during, or after the in-
troduction to Madeira (the source of the oysters was once intended
as a quarantine station for imports to Germany). The oysters'
initial condition index was in the normal range, and the values
found at the end of the experiment are not unusually low (cf. Table
2), except that they did not result from spawning activity.
Oysters are usually cultured in the intertidal zone or in shallow
depths, but cultivation at depths of 10 m and more is often advan-
tageous (Marteil 1979); we have not found any reports on oyster
cultivation at 15 to 25 m, but the absence of any depth correlation
in our data makes negative effects seem unlikely.
The temperature and salinity stress brought about by the trans-
fer from the Baltic to Madeira can by itself hardly account for the
oysters' poor performance. It brought an increase in salinity (by 15
to 20 ppt) and in temperature (by about 5°C) for the oysters.
Oysters from the identical Baltic stock were, however, subjected
to the same stress by Waschkewitz and performed better. Salinities
TABLE 2.
Comparison of condition (determined after Lawrence and Scott
1982) in oysters (C. gigas and Crassostrea virginica).
Source
High Low
Values Values
Remarks
61
41 C. virginica; sampled Jan-June
Lawrence and
Scon (1982)
Bodoy et al 60-120 20-30 C. gigas: 4-year study of 21
(1986) condition indices
Crosby and Gale 79-88 C. virginica; sampled in
(1990) May-June
This study 90 35 C. gigas; sampled July-Dec.
at Madeira are high (36 to 37 ppt year around; IMP 1982), but
Hughes-Games (1977) obtained excellent growth of C. gigas at 42
ppt and at summer temperatures of 20 to 34°C.
The facilities and equipment available for this investigation
were insufficient for the analysis of phytoplankton and nutrients,
but the waters of Madeira are oligotrophic (Chi. values of 0.05 to
0.15 mg m~2; INIP 1982), and starvation may have been an
important additional stress. In a 6-month starvation experiment
with C. gigas, Riley (1976) also found an important decline in dry
meat weight; carbohydrate levels declined more than lipid and
protein, but in the mantle (its main storage site), carbohydrate
declined only from 28 to 23% (28 to 2% glycogen in this study).
Seaman ( 1991 ) kept oysters from the same stock without food for
a similar period of time with similar results as Riley (cf. Table 3).
The steep decline in glycogen content that followed the oysters'
transfer to Madeira was not associated with gametogenesis. This
indicates that the animals were more stressed than in the examples
cited above and were thus forced to consume their reserves entirely
(cf. Gabbott 1983). We think that the oysters' poor performance
resulted from high metabolic demand (due to salinity and temper-
ature stress) in combination with nutritive stress. Better results
might be obtained if the oysters were introduced at another time of
year, as had been the case with Waschkewitz' introduction.
ACKNOWLEDGMENTS
We thank R. Waschkewitz (Canico, Madeira) for financial and
material support, the Direc^ao Regional da Agricultura for the use
of its laboratory facilities at Camacha (Madeira), and H.
Rosenthal, R.-A. Vetter, and R. Saborowski (IfM Kiel) for dis-
cussions and for help with the glycogen determination.
TABLE 3.
Performance of oysters (C. gigas) after 20 to 25 weeks of starvation, as compared with this study.
Source
Duration of
Experiment
(Weeks)
Decline in
Meat Weight
(% of Initial Value)
Decline in Carbohydrate"
(% of Initial Value)
Mantle
Whole Body
Cumulative
Mortality
(%)
Remarks
Riley (1976)
25
39
20
38
40
Starvation at 13.5°C
Seaman (1991)
20
30
20-40
20-48
Air storage at 7°C
This study
21
46
93
73
Same stock as Seaman (1991)
1 Total carbohydrate in Riley (1976); glycogen in Seaman (1991) and in this study
Grow Out of Oysters at Madeira Island
505
LITERATURE CITED
Askew, C. 1972. The growth of oysters Ostrea edulis and Crassostrea
gigas in Emsworth Harbour Aquaculture 1:237-259.
Bodoy, A.. J. Prou & J. P. Berthome. 1986. Etude comparative de dif-
ferents indices de condition chez 1'huitre creuse (Crassostrea gigas).
Haliotis 15:173-182.
Crosby, M. P. & L. D. Gale. 1990. A review and evaluation of bivalve
condition index methodologies with a suggested standard method. J.
Shellfish Res. 9:233-237.
Gabbott. P. A. 1983. Developmental and seasonal metabolic activities in
marine molluscs, pp. 165-217. In: K. M. Wilbur (ed.). The Mollusca.
vol. 2. Environmental Biochemistry and Physiology (ed. by P. W.
Hochachka). Academic Press. New York.
Heral. M. & J.-M. Deslous-Paoli. 1991 . Oyster culture in European coun-
tries, pp. 153-190. In: W. Menzel (ed.). Estuarine and Marine Bivalve
Mollusk Culture. CRC Press, Boca Raton. FL.
Hughes-Games, W. L. 1977. Growing the Japanese oyster [Crassostrea
gigas) in subtropical seawater fish ponds. I. Growth rate, survival and
quality index. Aquaculture 11:217-229.
Instituto Nacional de Investiga$ao das Pescas. 1982. Programa de apoio as
pescas na Madeira. II. Cruzeiro de reconhecimento de pesca e ocean-
ografia 020170680. Cruzeiro de reconhecimento de pesca e ocean-
ografia 020241 180. Relatonos No 1 1 . Instituto Nacional de Investiga-
c,ao das Pescas, Lisbon. Portugal, 220 pp.
Keppler. D. & K. Decker. 1984. Glycogen, pp. 1 1-18. In: J. Bergmeyer
and M. Grassl (eds.). Methods of Enzymatic Analysis. Vol. VI. Verlag
Chemie, Weinheim.
Lawrence, D. R. & G 1. Scott. 1982. The determination and use of con-
dition index of oysters. Estuaries 5:23-27.
Marteil. L. 1979. La Conchyliculture Franchise, vol. 3. L'ostreMculture.
Rev. Trav. Inst. Peches Mant. 45:5-130, I. S. T. P. M.. Nantes,
France .
Riley, R. T. 1976. Changes in the total protein, lipid, carbohydrate and
extracellular body fluid free amino acids of the Pacific oyster. Cras-
sostrea gigas, during starvation. Proc. Nat. Shellfish Assoc. 65:84-90.
Seaman. M. N. L. 1991. Survival and aspects of metabolism in oysters.
Crassostrea gigas, during and after prolonged air storage. Aquaculture
93:389-395.
Spencer. B. E. 1990. Cultivation of Pacific oysters. Laboratory Leaflet,
Ministry of Agriculture, Fisheries and Food Directorate of Fisheries
Research, Lowestoft. U.K. 47 pp.
Journal of Shellfish Research. Vol. 13. No. 2. 507-511, 1994.
THE BREEDING AND SECONDARY PRODUCTION OF THE FLAT TREE-OYSTER
1SOGNOMON ALATUS (GMELIN 1791) IN TROTTS POND BERMUDA
MARTIN L. H. THOMAS AND JOYCE C. DANGEUBUN
Marine Research Group
University of New Brunswick
P.O. Box 5050, Saint John
New Brunswick. Canada
ABSTRACT Isognomon alatus (Gmelin), the flat mangrove oyster, is very abundant in Trotts Pond, Bermuda. The oyster occurs
mainly on the submerged prop roots of the red mangrove Rhizophora mangle (L.), averaging 250 oysters root" ' or 2,700 m~2. The
sex ratio found was 68 female:26 male:6 indeterminate. Mature oysters of both sexes were present throughout the year, but maximum
breeding activity, shown by maximum gamete volume fraction followed by a fall and by a peak in spatfall. occurred in the spring. No
larvae were detected in the plankton during the course of the study. The mean caloric content of the oysters was 5.23 cal mg~ '. The
mean dry weight biomass was 399.3 g m :, and the total of somatic gonadal and shell production was 863.9 g m~~ per year giving
a productiombiomass ratio of 2.2:1.
KEY WORDS: oyster, mangrove, breeding, production, ecology
INTRODUCTION
Isognomon alatus (Gmelin) [Isognomonidae], the flat man-
grove oyster, is common in tropical and subtropical coastal waters
of the western Atlantic Ocean (Abbot 1974, Siung 1980. Rehder
1981, Morton 1983). Its typical habitat is in clumps on the sub-
merged prop roots of the red mangrove Rhizophora mangle (L.),
although it also occurs on rocks and manmade structures. In Ber-
muda, it was formerly common on coastal mangroves (Sterrer
1986) but is now confined to the mangrove habitat of the two
largest anchialine ponds, where it is very abundant (Thomas et al.
1992. Thomas 1993). The species is harvested only in Jamaica but
is of potential commercial importance elsewhere (Siung 1980).
/. alatus has been the subject of only sporadic research. True-
man and Lowe (1970) looked at the relation between temperature
and heart rate, and Kolehmainen et al. (1973) investigated the
response of the species to thermally elevated discharge water,
finding the oyster resistant to elevated temperature. Sutherland
(1980) discussed the epibenthic community dominated by /. alatus
and drew attention to its capacity both to colonize new surfaces
and. through its shells, to provide a large area of substrate for
associated species. Siung (1980) investigated at the breeding bi-
ology and its potential for cultivation in Jamaica. She found that
peak spawning occurred at a time of decreasing water salinity at
the onset of the rainy season in the autumn, but that some repro-
duction took place throughout the year. She also found that the
presence of eggs and/or sperm in the water acted as a stimulus to
spawning.
Trotts Pond is an anchialine pond with an area of 3.85 ha, a
low-tide volume of 103.565 m~3, a tidal range of 1.5 cm, a
peripheral mangrove swamp with an area of 0.84 ha, and sedi-
ments consisting of a matrix of organic mud with large quantities
of/, alatus shell; it has a small connection to the ocean at about
mean pond surface level (Thomas et al. 1991 ). Temperatures have
been reported to range from about 16 to 31°C, and salinities range
from 23 to 34%f (Thomas et al. 1991).
The occurrence of this species in the anchialine ponds of Ber-
muda is of particular interest. Thomas et al. (1991, 1992) have
pointed out that the pond environment is an extreme one and that
the diversity of life therein is low. Dissolved oxygen shows an
extreme range between anoxia and slight supersaturation. In sum-
mer, the combination of high temperature and low dissolved ox-
ygen results in difficult conditions for many aquatic organisms.
Nevertheless. /. alatus are the dominant attached organism in the
two largest ponds. Mangrove Lake and Trotts Pond, colonizing
most of the available mangrove roots. This study was designed to
investigate their reproductive and secondary production ecology in
this extreme environment.
MATERIALS AND METHODS
Studies on /. alatus took place from February 1991 to May
1992. Environmental conditions in Trotts Pond were monitored
from February 1991 to August 1993.
Temperature was monitored at two sites in the pond. A Branck-
ner XL- 100 (R. Branckner Research, Ottawa, Canada) electronic
data logger, sampling at hourly intervals, was installed to hang at
50 cm depth in the center of the mangrove swamp on the north side
of the pond (Fig. 1 ). It was operated from February 1991 to May
1992 and downloaded to a computer at 6 month intervals. On the
west side of the pond, temperature and dissolved oxygen were
monitored continuously from an electrode array, suspended at a
depth of 40 cm, in the center of the mangrove swamp (Fig. 1).
Two temperature sensors were used, one within a Rosemount
(Rosemount Analytical, Irvine, CA) dissolved oxygen analyzer
equilibrium electrode and the other a Cole-Parmer (Cole-Parmer
Instrument Co.. Niles, IL) thermistor sensor attached to the out-
side of the electrode. The thermistor sensor data were recorded on
a Cole-Parmer vertical chart recorder; the electrode data could be
displayed on the oxygen meter as a check. They were always
within 0.2°C of each other. The dissolved oxygen meter displayed
either percent saturation or ppm but recorded in ppm on the same
chart as temperature. The oxygen meter-electrode system was
recalibrated at 3 month intervals and reconditioned after the first
year. Readings showed linear drift with time and were corrected
on this basis. Salinity was monitored with a Biomarine Refrac-
tometer (Biomarine Inc., Hawthorn, CA) during biological oper-
ations.
The reproductive condition of oysters was determined from
four samples of randomly chosen oysters, 25 for each sampling
period. The gonad of each measured specimen was removed,
weighed, and then embedded in Paraplast for sectioning. Sections
507
508
Thomas and Dangeubun
METERS
Figure 1. Trotts Pond, Bermuda, showing the locations of the tem-
perature logger (TL), temperature and dissolved oxygen monitor
(TOM), the connection to the ocean (<-•), biomass and demography
sampling satations (N, W, E, S) and spat collector stations (1 to 10).
The shaded area is the mangrove swamp.
were cut as follows. The first section was 8 |xm thick followed by
10 10 (jLm thick sections, then a second 8 u,m section, and so on.
The 8 (j.m sections were mounted on slides and stained with Mal-
lory's stain. The 10 u.m sections were discarded. Sectioning was
continued until the entire gonad was sampled. The sex and stage of
gonad development were determined from the sections by refer-
ence to similar previous studies on other molluscan species
(Williams 1964, Fish 1972, Chase 1991). The gamete volume
fraction within the gonad was determined by stereology (Steer
1981).
Plankton in the pond were sampled weekly during study peri-
ods with a 10 cm mouth Hensen (Kahl Scientific Instrument
Corp., San Diego, CA) egg net fitted with 64 |xm mesh. The net
was used for 5 minute tows in the pond at both 5 cm deep and 20
to 50 cm deep, being towed behind a small Zodiac (Zodiac Corp.,
Mississauga, Canada) inflatable boat outside the aquatic margin of
the mangrove swamp. It was also deployed for 5 to 10 minute
periods in the inflowing current of the connection to the ocean
(Fig. 1).
Spat settlement was monitored on collectors prepared in Feb-
ruary 1991. Two types of collector were used, 21 mm diameter,
600 mm long polyvinyl chloride (PVC) pipes roughened with sand
paper and hung vertically from 10 cm below the surface and nat-
ural, living red mangrove prop roots carefully lifted from the wa-
ter, scrubbed free of all living macroorganisms and returned to the
water. One of each collector type was situated at each of 10 points
around the margin of the pond (Fig. 1). each site being marked
with fluorescent flagging tape. Collectors were sampled at each
subsequent sampling period, and all spat were counted and measured.
Biomass was sampled in four plots, one each at the centers of
the W, N, S, and E sides (Fig. 1) of the pond, having areas of
23.6, 35.8, 26.3, and 67.2 m~2, respectively, and containing
41 1, 515, 301, and 388 red mangrove prop roots (counted at the
surface), respectively. At each sampling period, nine prop roots
were selected in a stratified random pattern, three each from the
inner, middle, and outer portions of the aquatic part of the man-
grove swamp. Each root was stripped of all aquatic organisms,
which were sorted to species and weighed fresh. Oysters were
counted and then weighed wet and dry with the bodies and shells
separated; finally, samples of both bodies and shells were ashed at
435°C to give organic content for production calculations. Bio-
mass was calculated from biomass root ' and root abundance.
Demography was studied from five samples of oysters, one
from each sampling period. Each sample consisted of pooled oys-
ters from the centers of each of the W, N, S, and E shores (Fig. 1)
derived from randomly selected natural clumps. The samples con-
tained 553, 363, 628, 754, and 875 /. alatus respectively. The
length of each specimen was measured with Mitutoyo (M.T.I.
Corp., Paramus, NJ), digital Vernier calipers, which automati-
cally accumulated the data in a computer. Length frequency his-
tograms were plotted with SYSTAT (Systat, Evanston, IL), and
the polymodal distributions were separated into cohorts by the
method of Harding (1949). incorporating modifications of Cassie
(1954) and Cerrato (1980). Secondary production was estimated
by the removal summation method originally described by Boy-
sen-Jensen (1919) but modified by many authors (e.g.. Downing
and Rigler 1984). Production was considered as the sum of so-
matic, gonadal, and shell production following MacDonald and
Thompson (1985). Calorimetry was carried out on about every
10th oyster collected for demographic analysis (319 in all). Pow-
dered, mixed material from whole-body samples was used to make
pills burned in a Phillipson Microbomb calorimeter. Standardiza-
tion with benzoic acid was performed after every 10th sample.
RESULTS
The maximum water temperature recorded during the study
period was 34.7°C in August 1991 , and the minimum was 12.9°C
in February 1992. The daily mean temperature was 28.4°C in
summer and 18.2°C in winter. The diurnal range in summer av-
eraged 1.9°C (max., 3.7°C), whereas in winter, the average was
1.1CC (max., 2.9°C). Surface water salinities ranged from 27 to
34%o, with a mean of 29.5%o. Dissolved oxygen levels in Trotts
Pond were exceedingly variable, with a maximum of 9.93 ppm
(139% sat.) in December 1991 and minima of 0 ppm on numerous
summer nights from July to October 1991 . Daytime levels usually
rose above zero, but on four separate days in mid-August 1991,
continuous anoxia prevailed day and night at the meter site.
The overall sex ratio (femaleimale) of/, alatus in Trotts Pond
was 2.6:1, with individual samples ranging from about 4:1 to 2:1 .
Six percent of the specimens examined could not be sexed. The
mean gonad weight was 0. 16 ± 0. 12 g; there were no significant
differences in gonad weight among samples or between sexes.
Stages of gonad development were identified as: I, immature; II.
developing; III, mature 2; and IV, spent. All four stages could be
identified in females but only the last three in males. Figure 2
shows the distribution of the four stages in males, females, and the
population at the five sampling times. Mature individuals of both
sexes were present throughout the year but peaked early in the
year. This suggests that spawning was continuous, with a maxi-
mum in the early spring. Gamete volume fraction (GVF) data (Fig.
Mangrove Oysters in Bermuda
509
May 91
Aug 91
SAMPLING TIMES
May 92
FFFB STAGE I B STAGE II ^ STAGE III □ STAGE IV
May 91
Aug 91
SAMPUNG TIMES
May 92
| STAGE II ^g STAGE III fZ] STAGE IV
Feb 91
May 91 Aug 91 Dec 91 May 92
SAMPLING TIMES
Figure 2. Breeding cycle for (A) female and (B) male /. alatus and (C) GVF for both sexes in Trotts Pond, Bermuda.
2) supported this conclusion, showing high values at all times and
including an early spring peak.
Plankton samples failed to show the presence of/, alatus eggs
or larvae in the water at any time. The main constituents of the
plankton in order of decreasing frequency of occurrence were:
cyanobacteria, dinoflagellates, diatoms, copepods. cladocerans,
mysids, and chaetognaths.
The total numbers of spat that settled on the 10 natural prop-
root collectors by the end of each observation period were 1 8 1 , 73,
92, and 89 for May 1991, August 1991, December 1991, and May
1992, respectively. For the PVC pipes, the totals were 6. 1.2. and
1, respectively. These data support the continuous nature of
spawning with a spring peak suggested by the gonad studies.
The population of mangrove oysters on the prop roots of the red
mangrove was dense all around Trotts Pond. Mean values were
250 ± 1 1 root" ' or 2,696 ± 896 m-2, with a mean dry weight
biomass of 399.3 ± 89.6 g m~2. Length-frequency data for each
sampling period failed goodness of fit tests for normal distribu-
tions in all but one case (May 1991), showing the polymodal
nature of the population; however, a large overlap between cohorts
prevented their separation by length frequency alone. Cumulative
frequency plots analyzed by the modified Harding ( 1949) method
suggested the presence of five cohorts at all sampling times. It is
assumed that these represent annual groups dominated by the
spring spatfall but also containing oysters from earlier and later
settlement. It is concluded that the maximum lifespan is 51/: years
510
Thomas and Dangeubun
at this location. In all but one case (February 1991), the majority
of the population fell in cohort 3, suggesting a dominant spatfall in
the spring of 1988. Assuming tha' a spring spatfall dominates each
cohort, the mean lengths attained each year are about 17.5, 23,
32.5, 40.5, and 47.5 mm. with a maximum mean length of 52.5
mm. Individual oysters attain up to about 65 mm maximum
length. Caloric contents of individual oysters ranged 2.85 to 8.02
cal mg~'. Means for each sampling period were 4.98 ± 0.64,
4.70 ± 0.03, 4.64 ± 0.53, 5.70 ± 0.70, and 5.73 ± 0.77 cal
mg~' for February, May, August, and December 1991 and May
1992, respectively. Tukey's multiple range test confirmed that the
data fell into two significantly different groups comprising the first
three and last two samples, respectively, and showing higher lev-
els in the winter and spring of 1991 to 1992.
Table 1 gives the values calculated for somatic, shell, gamete,
and total production for each cohort during the period from 19
February 1991 to 28 May 1992, the grand total being 2.31 gm"
per day. The majority of the production was contributed by cohort
3. Annual production varied somewhat according to how the year
was defined within the total sampling time, but averaged 843 g
m-2 per year, yielding an annual production :biomass (P:B) ratio
of about 2.2:1.
DISCUSSION
The physical data for the pond collected in this study and
combined with that of Thomas et al. (1991) show a subtropical
anchialine pond with a pronounced annual temperature range of
about 13 to 35°C that varies little from year to year and a rather
constant salinity ranging from 24 to 34%o. Thomas et al. (1991)
attributed the low salinity range to the fact that the connection to
the sea at about mean pond surface level allowed fresh water from
rainfall and drainage to drain off the surface. Nevertheless, the
pond is somewhat hyposaline with respect to the coastal ocean at
about 36%f (Morris et al. 1977). Dissolved oxygen, however,
showed extreme variation from slight supersaturation to anoxia.
The latter condition occurred regularly in summer, when temper-
atures were close to a maximum, particularly frequently at night,
and lasting up to 3 days continuously at the sensor site. Observa-
tions suggested that anoxia was a patchy condition, never involv-
ing the entire pond during the period of study. However, residents
close to the pond have reported past conditions when a smell of
H,S was associated with the pond and when there were massive
mortalities of pond biota. However, total anoxia has never oc-
curred from 1 980 to the end of this study and may therefore be
considered an infrequent occurrence. However, because the pond
has a tidal range of less than 2 cm, the attached biota of the roots
are always submerged. The mangrove oysters and their associated,
TABLE 1.
Total production (gm2) of the population of /. alalus in Trotts Pond,
Bermuda during the period from 19 February 1991 to 26 May 1992.
Somatic
Shell
Gamete
Total
Cohort
(Pg)
(Ps)
(Pg)
(P)
Cohort 1
7.0632
5.1901
0.8802
13.1335
Cohort 2
79.7561
60.1829
12.8977
152.8367
Cohort 3
298.5551
174.7528
53.5776
526.8854
Cohort 4
177.4929
102.2235
38.2069
317.9233
Cohort 5
47.0579
25.5166
11.8924
84.4669
Total
609.2952
367.8658
117.4548
1095.2457
Daily
1.2687
0.7926
0.2531
2.3144
sessile biota are therefore exposed to anoxic conditions for days at
a time during the height of summer, and it is assumed that the
oysters shift to anaerobic respiration at such times. Such an ability
has been demonstrated for other oyster species but is usually a
response to winter inactivity rather than summer environmental
conditions (e.g., Wilbur and Yonge 1966). Where this species is
exposed by the tide, the valves gape slightly in air, allowing a
period of normal respiration (Littlewood 1994). The ability of this
species to exist without dissolved oxygen for periods in excess of
a day at over 30°C in Trotts Pond shows remarkable adaptation to
these harsh environmental conditions.
The preponderance of females over males in the ratio of over
2:1 is similar to that recorded for Crassostrea rhizophorae by
Littlewood and Gordon (1988) and suggests that sex reversal is
taking place in this species. Such a process is quite common
among oysters (Andrews 1979). Examination of the gonads from
a point of view of both maturity and GVF suggested that spawning
was a continuous process in Trotts Pond but that there was a
pronounced spring peak as temperatures rose from the winter min-
imum. This was confirmed by a similar pattern in the settlement of
spat. Siung ( 1980) also reported continuous spawning of this spe-
cies in Jamaica, but there, a spawning peak coincided with a
period of low salinity, falling to a minimum of about 15%o in
September to November as temperatures declined from their peak.
In Jamaica, the temperature range was only 26 to 30°C compared
with the 13 to 35°C range in Bermuda. In contrast to Jamaica,
salinity is fairly constant, without seasonal trends. Evidently,
spawning occurs under different environmental conditions in the
two locations. The absence of larvae in the plankton is curious.
Siung (1980) described pelagic larvae of this species in Jamaica
but was unable to determine the time spent in the plankton. In
Bermuda, we know that breeding occurred and spat settled, but
pelagic larvae were absent both in the open water of the pond and
in the tidal inflow water. Two explanations of this are possible.
The pelagic larval period may be very short, and settlement may
be immediate. However. Siung's observations suggest that it is at
least several days. Alternatively or additionally, the larvae may be
strongly photonegative and remain under the dense canopy of the
mangrove swamp. The lack of any significant tidal or wind-driven
currents in the very sheltered environment of Trotts Pond would
aid in any swimming behavioral adaptation in the larvae. Photo-
negativity in sessile benthic invertebrate larvae is a common and
widespread phenomenon (e.g.. Keough and Downes 1982).
The interpretation of demography from length-frequency data
is always difficult and in tropical populations often impossible.
However, in the case of the mangrove oyster in Trotts pond, it was
felt that the pronounced seasonal variation in temperature should
result in annual periodicity of phenomena related to growth and
reproduction. This assumption proved to be correct, but the con-
tinuous nature of spawning inevitably attenuated the length-
frequency range of annual groups. The conclusion from the length-
frequency data that there are five annual age groups present is
supported by the spawning maximum and maximum spatfall in
spring as well as by data from thin sections of shells, which
showed a maximum of five annual growth interruptions (Dan-
geubun 1994). Production calculations were carried out on the
basis of the five identified cohorts, but the total would not be
significantly affected if the assumption of five cohorts were incor-
rect because production would just be allocated differently. Both
total production and the P:B ratio seem rather low for tropical
invertebrates. However, it is very likely that total production is
underestimated, particularly for the reproductive component. This
Mangrove Oysters in Bermuda
511
aspect of production was calculated from data on the seasonal
reduction in the mean weight of the gonads; it would therefore
only include the spring maximum. The steady spawning through-
out the rest of the year would not be included by this method. It is
not possible to apply a correction for this, but if spat settlement is
a good index of production, then reproductive output is at about
half the spring value in the other three seasons. On this basis,
reproductive production would rise by a factor of about 1.5 to
about 276 g m~2, and the overall P:B ratio would rise to about
3:1. This is somewhat lower than the average for univoltine zoo-
benthos reported by Waters ( 1977) but higher than the trend shown
by Warwick ( 1980) for benthos with a 5 year lifespan. The crowd-
ing of the oysters and the environmental instability would both
tend to reduce secondary productivity (Mann 1967). The signifi-
cantly higher caloric content of oysters from the last two samples
is difficult to explain because it is not a seasonal trend, but it is
probably related to feeding conditions. Further work on the larval
life of this species is certainly needed, but it is evidently a suc-
cessful and productive species in very severe environmental con-
ditions.
ACKNOWLEDGMENTS
Financial support for this project was provided by a Natural
Sciences and Engineering Research Council of Canada operating
grant to the first author and by an Eastern Indonesia University
Development Project fellowship to the second author. Additional
support for the second author was by a grant in aid from the
Bermuda Biological Station for Research. Logistic support was
provided by the Bermuda Aquarium, Natural History Museum and
Zoo. The Mid Ocean Golf Club of Bermuda kindly granted access
to Trotts Pond . This article is Contribution # 1 370 of the Bermuda
Biological Station for Research.
LITERATURE CITED
Abbott, T. T. 1974. Sea Shells of North America. Van Nostrand Co. Inc.,
New York, 633 pp.
Andrews, J. D. 1979. Pelecypoda: Ostreidae. pp. 293-341. In: A. C.
Giesse and J. S. Pearse. (eds). Reproduction of Marine Invertebrates.
Academic Press. New York.
Boysen-Jensen, P. 1919. Valuation of the Limfjord. 1 Studies on the
fish-food in the Limfjord 1901-1917, its quantity, variation and annual
production. Rep. Danish Biol. Sta. 26:3-44.
Cassie, R. M. 1954. Some uses of probability paper in the analysis of
size-frequency distributions. J. Ausl. Mar. Freshwater Res. 5:513-
522.
Cerrato. R. M. 1980. Demographic analysis of bivalve populations, pp.
417-465. In: D. C. Rhoads and R. A. Lutz (eds). Skeletal Growth of
Aquatic Organisms. Plenum Press, New York.
Chase, M. E. 1991 The reproductive energetics of the common periwin-
kle, Litlorina littorea, on Pendleton Island, New Brunswick, M.Sc.
Thesis, University of New Brunswick. Saint John, 154 pp.
Dangeubun, J. C. 1994. Settlement, growth, mortality and reproductive
biology of the flat mangrove oyster, Isognomon alalus, in Trotts Pond
Bermuda. M.Sc. Thesis, Univ. New Brunswick, Saint John, 220 pp.
Downing, J. A. & F. H. Rigler. 1984. A Manual for the Assessment of
Secondary Production in Freshwaters. IBP. Handbook #17. Black-
well Scientific Publishers. Oxford. 501 pp.
Fish, J. D. 1972. The breeding cycle and growth of the open ocean and
estuanne populations of Liltorina littorea. J. Mar. Biol Assoc. U.K.
52:1011-1019.
Harding. J. P. 1949. The use of probability paper for the graphic analysis
of polymodal frequency distributions. J. Mar. Biol. Assoc. U.K. 28:
141-153.
Keough, M. J. & B. J. Downes. 1986. Recruitment of marine inverte-
brates: the role of active larval choices and early mortality. Oecologia
54:348-352.
Kolehmainen, S. T., T. Morgan & R. Castro. 1973. Mangrove root com-
munities in a thermally altered area in Gyanilla Bay, Puerto Rico. pp.
330-360. In: The effect of air-gaping behaviour on extrapallial fluid
pH in the tropical oyster, Crassostraea rhizophorae . Aquirre Power
Project Environmental Studies, 1972 Ann. Rept. USAEC rept, PRNC-
162. Puerto Rico Nuclear Centre.
Littlewood, D. T. J. 1994. Gibbons, J. W. & R. R. Shantz (eds.) Comp.
Biochem. Physiol 107A:l-6.
Littlewood, D T. J. & C. M. Gordon. 1988. Sex ratio, condition and
glycogen content of raft cultivated mangrove oysters Crassostrea
rhizophorae. J. Shellfish Res. 7:395-399.
Macdonald. B. A. & R. J. Thompson. 1985. Influence of temperature and
food availability on the ecological energetics of the giant scallop Pin-
copecten magellanicus . II. Reproductive output and total production.
Mar. Ecol. Prog. Ser. 25:293-303.
Mann. K. H. 1967. The cropping of the food supply, pp. 237-257. In:
S. D. Gerking (ed.) The Biological Basis of Freshwater Fish Produc-
tion. IBP Symposium on the Productivity of Freshwater Communities,
Reading, U.K. 1966. Wiley, New York.
Morns. B., J. Bames, F. Brown & J. Markham. 1977. The Bermuda
marine environment. Bermuda Biol. Sta Spec. Publ. #15. Bermuda
Biological Station, 120 pp.
Morton, B. 1983. Mangrove bivalves, pp. 77-130. In: W. D. Russell-
Hunter (ed.). The Mollusca. vol. 6. Ecology. Academic Press Inc..
New York.
Rehder, H A. 1981. The Audubon Society Field Guide to the North
American Sea Shells. A. A. Knopf Inc. New York, 894 pp.
Siung, S. M. 1980. Studies on the biology of Isognomon alatus Gmelin
(Bivalvia: Isognomonidae) with notes on its potential as a commercial
species. Bull. Mar. Sci. GulfCaribb. 30:90-101.
Steer, M. W. 1981. Understanding Cell Structure. Cambridge University
Press, Cambridge, U.K. 126 pp.
Sterrer. W. 1986. Marine fauna and flora of Bermuda. Wiley-Interscience,
New York, 742 pp.
Sutherland. J. P. 1980. Dynamics of the epibenthic community on roots of
the mangrove Rhizophora mangle, at Bahia de Buche, Venezuela.
Mar. Biol. 58:75-84.
Thomas, M. L. H. 1993. Mangrove swamps in Bermuda. Atoll Res. Bull.
386:1-17
Thomas, M. L. H., K. E. Eakins & A. Logan. 1991. Physical character-
istics of the anchiahne ponds of Bermuda. Bull. Mar. Sci. 48:125-136.
Thomas, M. L. H., A. Logan. K E. Eakins & S. M. Mathers. 1992.
Biotic characteristics of the anchiahne ponds of Bermuda. Bull. Mar.
Sci. 50:133-157.
Trueman, E. R. & G. A. Lowe. 1970. The effect of temperature and
littoral exposure on the heart rate of a bivalve mollusc, Isognomon
alatus. in tropical conditions. Comp. Biochem. Physiol. 38:555-564.
Warwick, R. M. 1980. Population dynamics and secondary production of
benthos, pp. 1-24 In: K R. Tenore & B. C. Coull (eds). Marine
Benthic Dynamics. The Belle W. Baruch Library in Marine Science,
Iniversity South Carolina Press. Columbia, SC.
Waters. T. F. 1977. Secondary production in inland waters, pp. 91-164.
In: A. Macfadyen (ed). Advances in Ecological Research, vol. 10.
Academic Press, New York.
Wilbur, K. M. & C. M. Yonge. 1966. Physiology of Mollusca. vol. 1.
Academic Press. New York, 645 pp.
Williams, E. E. 1964. The growth and distribution of Litlorina littorea
(L.) on a rocky shore in Wales. J. Anim. Ecol. 337:413-432.
Journal of Shellfish Research. Vol. 13. No. 2. 513-519, 1994.
LABORATORY STUDY OF FOOD CONCENTRATION AND TEMPERATURE EFFECT ON THE
REPRODUCTIVE CYCLE OF ARGOPECTEN VENTRICOSUS
JANZEL R. VILLALAZ G.
Departamento de Biologia Acudtica
Facultad de Ciencias Naturales y Exactas
Universidad de Panama
Republica de Panama
ABSTRACT A laboratory study was carried out in Lewes, Delaware, to assess, first, the relationship of gametogenesis in Argopecten
ventricosus to seasonal changes in temperature and phytoplankton densities, and second, the correlation of the gametogenic cycle with
relative changes in the dry weight of digestive gland, adductor muscle, and mantle-gills. The laboratory study showed that A.
ventricosus increased significantly in total weight by 30 days in high phytoplankton densities. Gonadal dry weight increased signif-
icantly after 40 days at high food ration, but gonadal index (dry weight of gonad/dry weight of whole animal x 100) declined. The
digestive gland declined sharply in dry weight under high and low phytoplankton densities, possibly suggesting that this organ was
providing energy for reproduction The adductor muscle index (dry weight of muscle/dry weight of whole animal X 100) was higher
at a high than at a low food ration. These laboratory studies suggest that gametogenesis in A. ventricosus is developed primarily during
the wet season at low phytoplankton densities, and somatic growth (especially of the adductor muscle) is enhanced during either dry
or wet seasons at high phytoplankton densities.
KEY WORDS: scallop, Argopecten ventricosus. reproduction
INTRODUCTION
Environmental Factors Affecting Bivalve Reproduction
Reproduction in the bivalve is coupled to seasonal changes of
environmental factors, such as water temperature and food avail-
ability. Less dramatic seasonal changes in environmental factors
have been observed in coastal tropical zones as compared with
those in temperate areas. Despite less dramatic seasonal changes in
the tropics, temperature and phytoplankton densities still affect the
reproduction of some bivalves as Wilson and Hodgkin (1967)
report in Amigdalum glaberrimu: Brachidontes cf. variabilis and
Septifer bilocularis, by Carvajal (1969). Lunetta (1969), and
Berry (1978); Velez and Epifanio (1981) in Perna perna, and by
Velez (1976, 1985) in Crassostrea rhtzophorae and Donax den-
ticulatus.
Meteorological and Oceanographic Conditions in Panama
The climate on the Isthmus of Panama is characterized by
pronounced seasonal changes in rainfall and wind velocity. Sea-
sonality is driven by the passage of the intertropical convergence
zone across the equator. This passage is related to alternation of
wet (April-December) and dry season (January-March). A dis-
tinct upwelling event in the Gulf of Panama on the Pacific side of
the isthmus is observed at the dry season. Life cycles of phyto-
plankton, anchovies, tunas, and shrimps are related to upwelling.
Annual gross production rate in the Gulf of Panama ranges from
255 to 280 g of C m~" (Forsberg 1969), and net primary produc-
tion is 180 g of C m~2, of which 90 is fixed during upwelling
(S may da 1966).
The major development of Panamanian fisheries during late
1985 and 1986 was an increase in scallop catches. Shipments for
1985 totaled 41 .0 t and for the first 6 months of 1986 were 1 .700
t. valued at $10 million (Marine Fisheries Review 1987). Deple-
tion of the scallop population in the Bay of Panama after 1986 was
probably due to a combination of high temperatures, overfishing,
and predation.
Reproductive Biology in Bivalves
Reproduction requires energy that is partitioned between sev-
eral physiological processes. The literature indicates the process
by which a bivalve can obtain energy either directly from food or
from storage substrates in organs and tissues, such as the digestive
gland (Taylor and Venn 1979). the adductor muscle (Ansell 1974.
Epp et al. 1988), or the mantle (Barber and Blake 1981, Lowe et
al. 1982).
Reproductive cycles in tropical bivalves have been studied ei-
ther by using histological techniques by such authors as Velez
( 1976) in C. rhtzophorae and Joseph and Madhyastha ( 1984) in C.
madrasensis or by using gonadal weights in Argopecten purpura-
tus by Wolff (1988). This points to the need for a comprehensive
study over time of the reproductive cycle and the composition of
organs and tissues of tropical bivalves relative to major environ-
mental factors.
Objectives
The objectives of this study were, first, to determine in con-
trolled experiments in the laboratory the effect of environmental
factors on reproduction in A. ventricosus. and second, under con-
trolled experiments in the laboratory, to relate seasonal changes in
the gametogenic cycle of A. ventricosus to relative changes in the
digestive gland, adductor muscle, mantle-gill, and gonad.
MATERIALS AND METHODS
Reproductive Condition
Morphometric data were obtained from dry weights of diges-
tive gland, adductor muscle, mantle-gill, and gonad. Indexes for
each organ were defined as dry weight of body organ/dry weight
of whole animal x 100.
Culture
Scallops were collected off of Farallon Beach, Panama Bay,
Panama, with a 1 m dredge, in 10 m depth (Fig. 1). Animals were
513
514
VlLLALAZ
Caribbean
San Bias
Gulf of Chiriqui
Coiba Is
Pacific Ocean
Figure 1. Map of study area in the Isthmus of Panama, including Farallon Beach on the Gulf of Panama. Is, island.
kept in fiberglass tanks in running sea water in the marine labo-
ratory of the University of Panama on Naos Island. Scallops were
transported by airplane to the College of Marine Studies in Lewes,
Delaware, in September 1988.
In Lewes, 40 scallops were conditioned and held in a 200 1
recirculating seawater tank at 19°C and 307cc salinity. A com-
bined algal diet (50:50) of Isochrysis species (C-1SO) and Chae-
toceros calcitrans was provided daily. In February 1989, 20 scal-
lops were induced to spawn by thermal stimulation (26°C) and
frequent water changes. Larvae were reared in a 200 1 tank at
24°C, and a combined algal diet was provided daily.
In October 1989, 540 8 month old scallops from the broodstock
group spawned in February 1989 were separated into 18 groups
(30 scallops each), and their sizes and weights were determined.
Initial condition was analyzed for 10 individuals by determining
wet and dry weights of the digestive gland, adductor muscle,
mantle-gills, and gonad.
Thirty of the 540 scallops were placed in each of 18 aquariums
(40 1 each) (six treatments x three replicates) (Table 1 ) containing
aerated filtered seawater (seawater was passed through a sand filter
and a diatomaceous-earth filter to remove particles 3=5 u,m). A
combination by cell (50:50) of C-ISO and Chaetoceros gracilis
was added daily. A high ration (5 x 104 cells ml~ ' per day) was
added to nine aquariums, while a low ration (1.25 x 104 cells
TABLE 1.
Experimental design for two rations (high and low phytoplankton
densities) and three temperatures (20, 24, and 28°C) in the culture
of A. ventricosus.
Phytoplankton density
Temperature (°C)
High Low
20
24
28
X X
X X
X X
ml " ' per day) was added to the other nine aquariums; these rations
were calculated from data from Panama Bay during the dry and
wet seasons, respectively. In addition, as a control, I placed sev-
eral scallops in the normal densities of phytoplankton observed
during the dry and the wet seasons, and the phytoplankton cleared
by scallops in each condition was determined. This experiment
was performed every other week to determine if sufficient food
was being provided to the experimental scallops. Quantities of
phytoplankton consumed by experimental scallops were deter-
mined daily by direct count with a hemocytometer and weekly
with a Coulter counter.
Six of the 18 aquariums (3 with high and the other 3 with low
phytoplankton concentrations) were maintained at 20°C, 6 other
aquariums were maintained at 24°C, and the last 6 aquariums were
maintained at 28°C (Table 1) (these temperatures were those ob-
served in the field in Panama during the dry, prewet. and wet
seasons, respectively). Water temperature was maintained in a
circulating bath (Forma Scientific). The pH of the water was re-
corded daily. Salinity was measured daily with a refractometer.
In each of the 18 aquariums. 10 scallops were sacrificed at 40
and again at 58 days. The condition of each scallop was deter-
mined by the dry weight and index of the digestive diverticulum,
adductor muscle, mantle-gills, and gonad. Morphometric mea-
surements of the scallops were compared with analysis of variance
(ANOVA). To satisfy the assumption of normality, weight data
were transformed (In x ) and index data were transformed (arcsine
x) before ANOVA. Symbols in figures represent means, and
vertical bars represent standard errors. Confidence intervals for
means are 95%.
RESULTS
pH and salinity are presented in Table 2. In general, pH and
salinity were similar to those recorded in the Gulf of Panama.
Morphometric A nalysis
Scallops significantly increased in shell height and total weight
after 31 days on high phytoplankton densities compared with those
A. VENTRICOSUS REPRODUCTIVE CYCLE
515
TABLE 2.
Data of environmental factors on aquariums during the experiment.
Parameter
Mean ± SE
Range
PH
Salinity (%o)
7.94
33
0.26
1.6
6.66-8 1 1
30-37
maintained on low phytoplankton densities (Figs. 2 and 3) (/? <
0.01).
Dry weights and indices of gonads showed a significant inter-
action between time, food ration, and temperature (Figs. 4 and 5)
(Tables 3 and 4) (p < 0.01). The greatest increase in gonadal dry
weight was detected in animals exposed to high levels of food at
a temperature of 20°C for 58 days. High phytoplankton densities
and low temperatures are conditions found in Panama Bay during
the upwelling season.
On day 40, at 24°C in low phytoplankton densities, A. ventri-
cosus had a higher gonadal dry weight and gonadal index than at
20°C or at 28°C in low phytoplankton densities (Figs. 4 and 5).
Gonadal dry weight was higher at high phytoplankton densities
than at low phytoplankton densities at the 20 and 28°C treatment
Only on day 40. did scallops have a higher gonadal index at low
phytoplankton densities than at high densities, at all temperatures
(Fig. 5).
Temperature, as a single factor, used in the experiment seemed
to have no effect on dry weights of the digestive diverticulum,
adductor muscle, and mantle-gills (Table 3). Rather, temperature
seemed to be related with time and food rations. The digestive
gland in the scallop declined sharply in dry weight under high and
low phytoplankton densities (Fig. 6).
A. ventricosus exposed to high phytoplankton densities in-
creased significantly in shell and total weight (Figs. 2 and 3)
(especially dry weights of adductor muscle and mantle-gills) (Figs.
7 and 8). Indices of the digestive diverticulum and adductor mus-
cle were significantly affected by interaction of time, temperature,
and food ration (Figs. 9 and 10) (Table 4). Scallops had a higher
20
18
X
o
S
w
DC
16
14
12
I
1 1 1 1
I
•
-
HIGH RATION
•
-
-
•
LOW RATION
— v — '
_— 2
1
i i i i
i
i
10
20 30
DAYS
40
50
60
70
Figure 2. Comparison of shell height of A. ventricosus fed two rations
during 58 days of experiment.
o
3
,— I
<
E-
O
O
_!
1.0
0.9
0.7 -
0.6
0.5
0 4
HIGH RATION
LOW RATION
0
10
50
60
70
20 30 40
DAYS
Figure 3. Comparison of natural logarithm (LN) of total wet weight
versus two different rations during 58 days of experiment.
digestive gland index at low phytoplankton densities than at high
densities, at any temperature (Fig. 9). The digestive gland index
was higher on day 8 compared with days 40 and 58, at any tem-
perature. The adductor muscle index was higher at a high food
ration than at a low food concentration. Scallops at 24°C and high
phytoplankton density had the highest adductor muscle index (Fig
10).
DISCUSSION
Temperature
Temperature probably has a greater influence on reproduction
in temperate than in tropical bivalves, because thermal fluctuations
there are greater. The effect of temperature on metabolism is doc-
umented by Sastry ( 1968) in Argopecten irradians, by Keck et al.
( 1975) in Mercenaria mercenaha, by Kennedy and Krantz (1982)
in Crassostrea virginica, and by MacDonald and Thompson
(1986) in Placopecten magellanicus.
Less dramatic seasonal changes in temperature have been ob-
served in coastal tropical zones, but they may still affect the re-
production of some bivalves in the tropics. Studies by Carvajal
(1969) in P. perna and Velez (1976) in C. rhizophorae show a
seasonal pattern of gametogenesis related to temperature. This
effect of temperature is expected in bivalves in Panama Bay, be-
cause pronounced seasonal upwelling occurs in this region, ac-
companied by changes in water temperature.
In the laboratory study, a relationship between reproductive
condition and water temperature per se was not observed in A.
ventricosus. Temperature, as a single factor, in A. ventricosus
could be important in the initiation of gametogenesis and as a
spawning stimulus. In the laboratory, temperature interacted with
food ration and time, promoting changes in scallop gonads. The
adaptive response of A. ventricosus to a wide range of tempera-
tures could explain the broad distribution of this species, from
California to Peru (Keen. 1971).
Food
This experimental study did demonstrate a significant differ-
ence between animals at high and those at low phytoplankton
516
VlLLALAZ
a
<
z
o
o
o
o
2
Q
1
i ■ ii-
-1
- 20
■ c o -
-
HIGH RATION /
-
i
=rH~ LOW RATION"
• V -V
1 1 1 1 1
6 -
6 -
10 20 30 40 50 60 70
1
24 •
i
C
I i i i
-"-V\ ^*
-
2>c
'
10 20 30 40 50 60 70
-\ 1 r
28 • C
_i i u
10
20 30 40 50
DAYS
60 70
X
w
Q
z
-J
<
<
Z
o
o
X
w
Q
Z
<
Q
<
Z
o
o
X
w
Q
z
<
Z
o
o
12
10
8
6
4
2
0
12
10
8
6
4
2
0
1 1 1
20 • C
T LOW RATION-
-«
F
Source of
Digestive
Adductor
Mantle-
Variation
Gonad
Diverticulum
Muscle
Gills
Time
0.707
<0.001
0.125
0.206
Temperature
0.441
<0.001
0.282
0.276
Food
<0.001
0.011
<0.001
<0.001
Interactions
0.002
<0.001
0.003
0.122
Time x Temperature
0.281
0.093
0.556
0.623
Time x Food
0.001
0.005
<0.001
0.007
Temperature x Food
0.030
0.001
0.183
0.026
A. VENTRICOSUS REPRODUCTIVE CYCLE
517
TABLE 4.
Study of environmental factors. Results of three-factor ANOVA on
days 8. 40. and 58, among three different temperatures and
between two food rations for indices of gonad, digestive
diverticulum, adductor muscle, and mantle-gills (arcsine
transformation) of .4. ventricosus.
be obtained directly from food in the water or from energy stored
in different organs. In scallops, energy for reproduction can come
from the digestive gland (Sastry 1970) or adductor muscle (Taylor
and Venn 1979, Barber and Blake 1981).
I found that the digestive gland decreased significantly in dry
weight, possibly suggesting that this organ was providing energy
for reproduction. This agrees with Sastry's findings (1970) in
Massachusetts that A . irradians had a statistically significant neg-
ative correlation between digestive gland index and gonadal index.
These findings may also support the idea that A. ventricosus is a
subtropical species where reproductive strategies are similar to
those of scallops living in cooler ecosystems.
Interaction of Food and Temperature
In the experimental laboratory study, A. ventricosus exposed to
high phytoplankton densities at all temperatures (20, 24, and
M
Q
<
O
o
E-
O
i — i
W
OS
Q
o
3.0
2.5
2 0
1.5
1 1
i i i i
\
HIGH RATION
_
\ J>
i i
LOW RATION
i i i i
0
10
20
50
60
70
30 40
DAYS
Figure 6. Comparison of natural logarithm (LN) of digestive gland
dry weight versus two rations during 58 days of the experiment.
tiO
p > I
s
Source of
Digestive
Adductor
Mantle-
o
E-i
Variation
Gonad
Diverticulum
Muscle
Gills
Time
0.197
<0.001
0.003
<0.001
K
Temperature
0.670
0.266
0.827
0.026
o
Food
0.001
<0.001
<0.001
0.178
IY1
Interactions
0.003
0.030
0.053
0.090
£
Time X Temperature
0.319
0.428
0.867
0.306
Time X Food
<0.001
0.002
0.003
0.368
OS
Q
Temperature x Food
0.409
0.693
0.248
0.029
o
J
-
1
1 1
HIGH RATION
i
i i
i i
i
LOW RATION
i i
0
•
20
50
60
70
30 40
DAYS
Figure 7. Comparison of natural logarithm (LN) of adductor muscle
dry weight versus two rations during 58 days of the experiment.
Means of all temperatures.
28°C) developed only moderate gonadal and digestive gland in-
dexes, while building considerable tissue in the adductor muscle.
Also, high phytoplankton densities appeared to inhibit the alloca-
tion of energy from the digestive gland to the gonad, but enhanced
adductor muscle growth. This would suggest that, during high
phytoplankton densities during the dry or the wet season, scallops
allocate energy primarily for growth of the adductor muscle.
on
w
•z
<
S
o
E-h
X
o
I — I
W
>-
OS
Q
o
3.0
2.5
2.0
1.5
HIGH RATION
LOW RATION
0
10
20
50
60
70
30 40
DAYS
Figure 8. Comparison of natural logarithm (LN) of mantle-gill dry
weight versus two rations during 58 days of the experiment. Means of
all temperatures.
518
VlLLALAZ
0 20 30 40 50 60 70
X
a
u
50
40
30
20 ■ C
HIGH RATION
S * -
LOW RATION
10 20 30 40 50 60 70
X
w
a
Q
2
20 30 40 50 60 70
20 30 40 50 60 70
DAYS
Figure 9. Comparison of digestive gland index and two different ra-
tions during 58 days of the experiment at three temperatures. Means
of all temperatures.
The laboratory study showed that scallops exposed to low phy-
toplankton densities at 24 and 28°C developed a maximal gonadal
index. Energy to support gonadal growth was provided mostly by
the digestive gland under these conditions. A decline in adductor
muscle index was also observed. This suggests that low phyto-
plankton densities and temperatures higher than 24°C promoted
allocation of energy from the digestive gland to the gonad. Thus,
maximal gonadal condition could have related to a decrease in
adductor muscle growth. In the Gulf of Panama, low phytoplank-
ton densities and temperatures above 24°C occur during the wet
season.
0 10 20 30 40 50 60 70
. . 60
X
W
Q
5 50 -
Ed
y <°
30
1
28 °
i
C
i i
i i
a —
— 5
T
'
0 10 20 30 40 50 60 70
DAYS
Figure 10. Comparison of the adductor muscle index, two phyto-
plankton concentrations, and three temperatures during 58 days of the
experiment.
Sastry and Blake (1971) observed that the transfer of reserves
from the digestive gland to the gonad is regulated by temperature
and the stage of gametogenesis. Reserves from the digestive gland
are transferred to gonads and adductor muscle, depending on the
activity of gonads or the temperature (Sastry and Blake 1971).
Gametogenesis in this species is thus developed primarily during
the wet season at low phytoplankton densities, and somatic growth
(especially adductor muscle) is enhanced during either dry or wet
seasons at high phytoplankton densities.
LITERATURE CITED
Ansell, A. 1974. Seasonal changes in biochemical composition of the
bivalve Chlamys seplemradiala from the Clyde Sea area. Mar. Biol
25:85-99.
Barber, B & N. Blake. 1981 . Energy storage and utilization in relation to
gametogenesis in Argopecten irradians concentricus (Say). J. Exp.
Mar. Biol. Ecol. 52:121-134.
Berry. P. F. 1978 Reproduction, growth and production in the mussel,
Perna perna (Linnaeus), on the east coast of South Africa. South
Africa Association for Marine Biological Research. Oceanographic
Research Institute. Investigational Report No 48.
Carvajal, R. J. 1969. Fluctuacidn mensual de las larvas y crecimiento del
mejillon Perna perna (L) y las condiciones ambientales de la ensenada
de Guatapanare, Venezuela. Bol. Inst. Oceanogr. Univ. Oriente Ven-
ezuela 8:13-20.
Epp, J., V. M. Bnceh, & R. Malouf. 1988. Seasonal partitioning and
utilization of energy reserves in two age classes of the bay scallop
Argopecten irradians irradians Lamarck. J. Exp. Mar. Biol. Ecol.
121:113-136.
Forsberg, E. 1969. On the climatology, oceanography and fisheries of the
Panama Bight. Inter-Amencan Tropical Tuna Commission, vol. 14,
no. 2. 385 pp.
lohansson, R. J. O. & T. L. Hopkins. 1972. Primary productivity, pp.
48-59. In: Anclote Environmental Project Report 1972, Contnbution
41. Department of Marine Science, University of South Florida, St.
Petersburg, Florida.
Joseph, M. M. & M N Madhyastha. 1984. Annual reproductive cycle
and sexuality of the oyster C. madrasensis. (Preston). Aquaculture
40:223-231.
A. VENTRICOSUS REPRODUCTIVE CYCLE
519
Keck, R., D. Maurer & H. Lind. 1975. A comparative study of the hard
clam gonad development cycle. Biol, Bull. 148:243-258.
Keen, A. M. 1971. Sea shells of tropical West America. Stanford Uni-
versity Press. Stanford, California, 1064 pp.
Kennedy V. & L. Krantz. 1982. Comparative Gametogenic and Spawning
patterns of the Oyster C. virginica (Gmelinl in Central Chesapeake
Bay. J. Shellfish Res. 2:133-140.
Lowe. D. M.. M. N. Moore & B. L. Bayne. 1982. Aspects of gameto-
genesis in the marine mussel Mytilus edulis L. J. Mar. Biol. Assoc.
U.K. 62:133-145.
Lunetta. J. E. 1969. Reproductive physiology of the mussel Mynlus
perna. Boletin de la facultad de Filosofia, Ciencias e Letras. Univer-
sidad de Sao Paulo. Zool. Biol. Mar. (Sao Paulo) 26:33-1 1 1
MacDonald, B. A. & R. J. Thompson. 1986. Production, dynamics and
energy partitioning in two populations of the giant scallop Placopecten
magellanicus (Gmelin). J. Exp. Mar. Biol. Ecol. 101:285-299.
Marine Fisheries Review. 1987. The Latin American Scallop Fisheries,
1980-86. 49:72-76.
Sastry, A. N. 1968. The relationships among food, temperature, and go-
nad development of the Bay Scallops, Aequipecten irradians Lamarck
Physiol. Zool. 41:44-53.
Sastry, A. N. 1970. Reproductive physiological variation in latitudinally
separated populations of the bay scallop, Aequipecten irradians La-
marck. Biol. Bull. 138:56-65.
Sastry. A. N. & N. I. Blake. 1971. Regulation of gonad development in
the bay scallop. Aequipecten irradians Lamarck Biol. Bull. 140:274-
282.
Smayda, T. 1966. A quantitative analysis of the phytoplankton of the Gulf
of Panama. Inter-American Tropical Tuna commission, vol. 1 1 . no. 5.
583 pp.
Taylor, A. C. & T. J. Venn. 1979. Seasonal variation in weight and
biochemical composition of the tissues of the queen scallop, Chlamys
opercularis. from the Clyde Sea area. J. Mar. Biol. Assoc. U.K. 59:
605-621.
Thayer. G. W. 1971. Phytoplankton production and the distribution of
nutrients in a shallow unstratified estuarine system near Beaufort, N.C.
Chesapeake Sci. 12:240-253.
Velez, A. 1976. Crecimiento, edad y madurez sexual del ostion Crassos-
trea rhizophorae de Bahia de Mochima y Laguna Grande. Bol. Inst.
Oceanogr. Univ. Oriente Venezuela 11:39-43.
VeTaz. A. & C. E. Epifanio. 1981. Effects of temperature and ration on
gametogenesis and growth in the tropical mussel Perna perna (L).
Aquaculture 22:21-26.
Velez. A. 1985 Reproductive biology of the tropical clam Donax dentic-
ulatus in Eastern Venezuela. Carih. J Sci. 21:125-136.
Wilson. B. R. & E P. Hodgkin. 1967. A comparative account of five
species of marine mussels (Bivalvia: Mytilidae) in the vicinity of Fre-
mantle. Western Australia. Aust. J. Mar. Freshwater Res. 18:175-
203.
Wolff, M. 1988. Spawning and recruitment in the Peruvian scallop A.
purpuratus Mar. Ecol. Prog. Ser. 42:213-217.
Journal of Shellfish Research, Vol 13. No. 2, 521-527, 1994.
COMPARISON OF INFECTIVITY AND PATHOGENICITY OF MERONT (TROPHOZOITE) AND
PREZOOSPORANGIAE STAGES OF THE OYSTER PATHOGEN PERKINSUS MARINUS IN
EASTERN OYSTERS, CRASSOSTREA VIRGINICA (GMELIN, 1791)
ASWANI K. VOLETY AND FU-LIN E. CHU
School of Marine Science
Virginia Institute of Marine Science
College of William & Mary
Gloucester Point, Virginia 23062
ABSTRACT Two experiments were conducted to compare the infectivity and pathogenicity of two life stages of the parasite
Perkinsus marinus, meronts (trophozoites) and prezoosporangia (hypnospores), in eastern oysters, Crassosirea virginica. Oysters were
inoculated with 5 x 104 meronts or prezoosporangia per oyster by injection into the shell cavity. Prevalence and intensity of P. marinus
infections, condition index, serum protein concentrations, and lysozyme activities were measured in oysters after 15, 25, 40, and 65
days in experiment 1 and after 20, 40, 50, 65, and 75 days postchallenge by P. marinus cells in experiment 2. Controls were injected
with filtered York River water. In the first experiment, P. marinus infections were initially detected in oysters exposed to prezoo-
sporangia after 15 days postchallenge. In the second experiment, infection was not detected in oysters until 40 days postchallenge with
either meronts or prezoosporangia. Intensity and prevalence of P marinus infection were significantly higher (p < 0.002) in oysters
challenged by meronts compared with prezoosporangia-challenged oysters at the end of both experiments. In experiment 1, a
significant decrease (p < 0.05) was observed in serum protein in infected oysters challenged by prezoosporangia compared with
uninfected oysters. Condition index was higher in uninfected oysters compared with infected oysters challenged by prezoosporangia.
The differences in condition index and protein were insignificant between oysters infected by meronts or prezoosporangia. Lysozyme
activities were significantly lower {p < 0.05) in infected oysters than in uninfected oysters challenged with meronts. No significant
differences were observed in condition index, protein concentrations, and lysozyme activities between oysters challenged by meronts
and prezoosporangia in experiment 2. Lower condition index and protein concentrations in the groups of oysters infected with
prezoosporangia compared with the groups infected by meronts and nonchallenged at the end of experiment 1 suggest a higher
energetic demand on these oysters.
KEY WORDS: Perkinsus marinus, Crassosirea virginica, oyster disease, lysozyme, condition index, protein
INTRODUCTION
The once-thriving oyster industry in the Chesapeake Bay and
the East Coast of United States has been threatened by overfishing
and diseases caused by two protistan parasites, Haplosporidium
nelsoni (MSX) and Perkinsus marinus (Dermo). The effects of the
diseases caused by the two protists have been well documented
(Andrews 1988. Barber et al. 1988, Ford 1988, Ford and Figueras
1 988 , Chu et al . 1 993 , Chu and La Pey re 1 993a and b , Pay nter and
Burreson 1991). Since 1986, P. marinus has reportedly caused
greater oyster mortalities in lower Chesapeake Bay than H. nelsoni
(Andrews 1988).
The life history of P. marinus was studied in detail by Perkins
(1966). Three life stages were identified, namely, merozoites.
prezoosporangia, and the biflagellated zoospores. Immature
meronts (merozoites), usually found in the phagosomes of hemo-
cytes are 2 to 4 u.m in size and coccoid, with a fibrogranular wall.
As the cells mature, they enlarge to about 10 to 20 p.m with an
eccentrically placed vacuole, which often contains a refringent
vacuoplast. The mature meronts, on repeated karyokinesis and
cytokinesis, yield sporangia (schizont, 10 to 40 p.m in size), an 8
to 32 cell stage enclosed within a mother cell wall (Perkins 1966).
Enlargement of meronts to form prezoosporangia is achieved by
incubating the meronts in fluid thioglycollate medium (FTM) (Ray
1952). The prezoosporangia are characterized by an extremely
large vacuole, which compresses the cytoplasm into a thin layer
against the cell wall. On enlargement, the vaculoplast disappears
and the nucleus attains a sausage shape, with numerous small
lipoid droplets dispersed inside the cell.
Numerous field (Soniat 1985. Craig et al. 1989. Soniat and
Gauthier 1989, Crosby and Roberts 1990, Gauthier et al. 1990,
Burreson 1989 and 1990) and laboratory studies (Mackin 1951,
1956 and 1962, Andrews and Hewatt 1957, Perkins 1966, Chu
and La Peyre 1989, Ragone 1991, Ragone and Burreson 1993)
have investigated the effects of temperature and salinity on the
disease processes off. marinus in eastern oysters. Other previous
laboratory experiments induced P. marinus infection through ex-
posure of oysters to meronts, merozoites, and schizonts contained
in unpurified or partially purified infected oyster tissue (Chu & La
Peyre 1993a, Hewatt & Andrews 1956, Mackin 1962). For con-
venience, the cellular stages found in oyster tissue will hereafter be
termed meronts, with the recognition that merozoites and schiz-
onts are also present.
In nature, the meronts (3 to 15 p.m) rarely enlarge to a size of
15 to 100 u.m in moribund oysters and when enlarged are called
prezoosporangia (Perkins 1966). Prezoosporangia, when placed in
seawater, divide by successive bipartitioning and form biflagel-
lated zoospores (Perkins 1966. Chu and Greene 1989). Whereas
slightly enlarged cells, believed to be prezoosporangia, can be
found in moribund oyster, such cells have never been isolated and
induced to form zoospores. The presumption is that they have the
capability to zoosporulate. Because the exposure of oysters to
minced oyster tissue containing meronts or freshly isolated and
partially purified meronts results in a high prevalence of P. mari-
nus infection, Perkins (1966) suggested that meronts and merozo-
ites may be the primary infective agents transmitting disease
among oysters in the field, with the recognition that zoospores also
can induce infections. However, similar infection rates were found
by exposing oysters to prezoosporangia and biflagellated zoo-
spores in our laboratory (Chu et al., unpublished results). These
521
522
VOLETY AND CHU
results suggest that all three life stages, namely, meronts, prezoo-
sporangia. and biflagellated zoospores, are capable of inducing
infection in oysters, although some of the previous studies have
used minced infected oyster tissue (Hewatt and Andrews 1956) or
minced infected oyster tissue incubated in thioglycollate medium
in 1 day (Mackin 1962). Therefore, the infective cells used in
previous studies would mostly be meronts with some prezoospo-
rangia. None of the previous studies have examined purified pre-
zoosporangia as an infective agent, nor were the physiopatholog-
ical effects investigated. This article reports the results of exper-
iments in which the infectivity and pathogenicity of meronts and
prezoosporangia were compared. The physiological responses of
oysters challenged by these two infective stages were also deter-
mined.
MATERIALS AND METHODS
P. marinus Diagnosis
P. marinus infections were diagnosed using hemolymph and
tissue assays (Gauthier and Fisher 1990, Ray 1952 and 1966). The
hemolymph assay was as follows: 300 u.1 of hemolymph contain-
ing hemocytes was obtained and incubated in FTM containing
antibiotics (penicillin and streptomycin) for 4 days. After incuba-
tion, the thioglycollate medium was separated by centrifugation at
800 x g and incubated with IN NaOH for 1 hour to remove tissue
debris and hemocytes. The suspension was then washed twice with
water, and prezoosporangia were stained with Lugol's iodine and
counted. Disease intensity was ranked from 1 to 5 (light to heavy).
At the end of each experiment, infections were also diagnosed
according to the method of Ray (1952) by incubating pieces of
rectal and mantle tissue in FTM. Infection intensities were rated as
light to heavy (1 to 5), and weighted indices were calculated based
on Ray (1954) and Mackin (1962).
Lysozyme Activity
Lysozyme activity (L) was determined spectrophotometrically
according to Shugar (1952) and modified by Chu and La Peyre
( 1989). Briefly, 0. 1 ml of cell-free oyster serum was added to 1 .4
ml of bacterial {Micrococcus lysodiekticus) suspension. The de-
crease in absorbance at 450 nm on a Shimadzu UV 600 spectro-
photometer was measured after 1 minute. Results are expressed as
units per ml of oyster serum. One unit is defined as decrease in
absorbance of 0.001 in the bacterial suspension per minute at room
temperature (22-23°C).
Serum Protein Concentration
The concentrations of serum protein (P) were measured spec-
trophotometrically according to Lowry et al. (1951) with bovine
albumin as a standard.
Experiments
Two experiments were conducted to compare the pathogenic
effects of meronts and prezoosporangia.
Experiment I
Eastern oysters were collected from the Ross' Rock area of the
Rappahannock River, Virginia (ambient salinity, 6 ppt; ambient
temperature, 19CC). Oysters from this location have the lowest
prevalence of ''. marinus infection of any oyster bed in Virginia
(Ragone Calvo and Burreson 1994). Oysters were gradually ac-
climated over a period of o weeks to the test conditions (temper-
ature, 25.6 ± 1.3°C; salinity. 20.7 ± 1.04) in a 200 1 tank.
Ninety-six oysters were then randomly placed in aerated individual
chambers with flowing I (x filtered York River water (YRW).
Oysters were fed daily during the acclimation and the experimental
period with algal paste (0. 1 g per oyster, using a mixture of Iso-
chrysis galbana, Pavlova lutheri, and Tahitian /. galbana), and
water was changed every other day. Meronts were partially puri-
fied from infected oyster tissue according to Chu and La Peyre
( 1993a). Prezoosporangia were cultured on the basis of the method
described by Chu and Greene (1989). One hundred microliters of
filtered YRW containing 5 x 104 meronts or prezoosporangia cells
(meronts cultured in FTM and enlarged to size range of > 100 u.m)
was injected into the shell cavity of each oyster. Controls were
injected with 1 u,m filtered YRW. There were three treatments:
control, meront-challenged, and prezoosporangia-challenged oys-
ters. To monitor infection development, eight oysters were ran-
domly sampled from each treatment at 15, 25, 40. and 65 days
postchallenge. Hemolymph samples were withdrawn from the an-
terior adductor muscle of individual oysters with a syringe with a
27 gauge needle. Serum L and P concentration were measured.
Hemolymph was also assayed to evaluated P. marinus infection
(Gauthier and Fisher 1990). After withdrawal of hemolymph sam-
ples, oysters were sacrificed and condition index (CI) (dry meat
weightydry shell weight X 100; Lucas and Beninger 1985) was
determined. P. marinus infections in oysters were also diagnosed
using rectal and mantle tissue according to the tissue assay de-
scribed by Ray (1952).
Experiment 2
The experimental conditions were similar to those of experi-
ment 1 , with the exception that oysters were collected from the
Damarsicotta River, Maine, a region out of the range of P. mari-
nus (ambient salinity and temperature. 32 to 35 ppt and 12 to
14°C. respectively). As in experiment 1. oysters were gradually
adjusted to the test conditions (temperature, 21.78 ± 0.84°C; sa-
linity, 20.5 ± 1.19 ppt) in 6 weeks, and then, 135 oysters were
randomly placed in individual chambers with 1 u.m filtered aerated
YRW. Nine oysters from each treatment were sampled at the end
of 20. 40. 50. 65. and 75 days after being challenged with infec-
tive particles. Measurements of CI and serum L and P were con-
ducted in individual oysters as indicated above.
Statistical Analyses
A one-factor analysis of variance (ANOVA) followed by a
Tukey-Kramer test was used to determine the differences in CI, L.
and P among treatments. The data were first analyzed for differ-
ences among treatments and sampling times. Some of the oysters
were not infected after they were challenged with meronts or
prezoosporangia. CI. L, and P data from challenged oysters at all
sampling times from experiment 1 were split into infected and
uninfected oysters. Data from uninfected oysters from each treat-
ment at all sampling times were pooled with the controls. This
resulted in three groups, namely, uninfected, meront infected, and
prezoosporangia infected. Data were then reanalyzed with one-
way ANOVA to determine differences among groups. In experi-
ment 2, CI, L, and P data were analyzed with one-way ANOVA
without being split into infected and uninfected groups. Logistic
regression (Agresti 1990) was used to determine differences in
prevalence of infection between treatments and sampling times in
both experiments.
RESULTS
In experiment 1. infection first appeared in oysters 15 days
after being challenged with prezoosporangia and 25 days after
Infectivity of Two Life Stages of Perkinsus mar/nus
523
challenge with meronts (Fig. la). Prevalence, at 65 days postchal-
lenge, was higher in oysters challenged by meronts (87.5%) com-
pared with oysters challenged by prezoosporangia (43r/r ) (Fig. la).
Prevalences of both groups significantly increased with time (p <
0.05). Prevalence was not significantly different between meront-
challenged and prezoosporangia-challenged oysters. Intensities of
infections ranged from light to heavy (1 to 5) in meront-challenged
oysters, whereas no heavy infections were detected in prezoospo-
rangia-challenged oysters (Fig. 2a). When intensity of infection
was expressed as weighted incidence (sum of disease code num-
ber/total number of oysters examined), it showed a trend similar to
that of prevalence. Weighted incidence (Table 1 ) at the end of the
experiment was higher in oysters challenged with meronts (2.13)
compared with oysters challenged with prezoosporangia (0.86).
In experiment 2, the first infections appeared after 40 days in
both meront- and prezoosporangia-challenged oysters. Prevalence
(Fig. lb) was significantly (/? < 0.002) higher in meront-
challenged oysters (77.5%) compared with prezoosporangia-
challenged oysters (57.2%). As in experiment I , infection in both
groups increased with time (p < 0.0001 ). Intensities of infections
ranged from light to moderate heavy (1 to 4) in oysters challenged
with meronts, whereas only light infections (1) were observed in
prezoosporangia-challenged oysters (Fig. 2b). Weighted incidence
(Table 1 ) at the end of the experiment was higher in meront-
challenged oysters (0.86) as compared with prezoosporangia-
challenged oysters (0.5).
There were no differences in CI. L. and P among treatments at
different sampling times in experiment 1 (p > 0.05). In experiment
1. within the prezoosporangia-challenged group. CI of infected
oysters was lower than that of uninfected oysters (Fig. 3). The CI
of infected oysters from the group challenged by meronts was not
different from that of infected oysters from the group challenged
by prezoosporangia (p > 0.05). Serum P concentrations in infected
1a 100
80
§. 60
o
CD
£ 40
20
0
1b too
80
w
CD
I 60
o
a>
J 40
c
20
0
B I
■ Sporangia
■ Meront
□ Control
25. 40. 65.
■ Sporangia
■ Meront
□ Control
20.
Nimber of days (post chaltenge)
Figures la and lb. P. marinus prevalence in oysters after 15, 25, 40,
and 65 days (a) and 20, 40, 50, 65, and 75 days postchallenge (b) by
meronts or prezoosporangia.
/LIGHT
MODERATE
HEAVY
MERONTS SPORANGIA
TREATMENT
MERONTS SPORANGIA
TREATMENT
Figures 2a and 2b. P. marinus infection intensity in oysters from ex-
periment 1 (a) after 65 days and experiment 2 (b) after 75 days post-
challenge by meronts and prezoosporangia.
oysters challenged with prezoosporangia were significantly lower
(p < 0.05) than those in the uninfected oysters (Fig. 4). However,
no significant difference in P concentrations was observed be-
tween infected and uninfected oysters in the group of oysters chal-
lenged with meronts. No differences (p > 0.05) were observed in
serum P concentrations between meront- and prezoosporangia-
challenged oysters. Also, no significant difference in P concen-
trations was observed between infected oysters challenged with
meronts or with prezoosporangia. In oysters challenged by
meronts, L was significantly higher {p < 0.05) in uninfected than
in infected oysters (Fig. 5). No such differences were observed
between infected and uninfected oysters challenged with prezoo-
sporangia.
In experiment 2. CI and serum P concentrations significantly
decreased (p < 0.05) in all treatments with time. The CI of
oysters at the end of 20 days was significantly higher than the
524
VOLETY AND CHU
TABLE 1.
Weighted incidence of P. marinus infection and experimental conditions.
Weighted Incidence
Experimental Conditions*
Infective Cell
Experiment 1 Experiment 2
Experiment 1 Experiment 2
Meront
Prezoosporangia
2.13 1.33
0.86 0.5
T = 25.6 ± 1.33°C T = 21.78 ± 0.84°C
S = 20.7 ± 1.04 ppt S = 20.5 ± 1.19 ppt
1 T, temperature; S. salinity.
CI of the oysters at the end of 50 and 75 days (Fig. 6). P concen-
trations in oysters from all treatments decreased with time (Fig. 7).
P concentrations at the end of 20, 40, and 50 days were signifi-
cantly (p < 0.05) higher than at the end of 65 and 75 days post-
challenge (Fig. 7). No significant differences were observed in L
between treatments at any sampling time.
DISCUSSION
Results of this study show that both meronts and prezoospo-
rangia infect oysters, with meronts being more infective than
prezoosporangia. This supports the hypothesis (Perkins 1966) that
meronts are the primary agents of disease transmission of P. mari-
nus in oysters. The higher prevalence of infection in oysters chal-
lenged with meronts might have been due to the higher virulence
of meronts. The meronts may multiply rapidly in oysters at warm
temperatures, such as those (Table 1) used in this study. The cause
for the lower infection rate of prezoosporangia is not clear. Al-
though the prezoosporangia injected into the oysters were >95%
viable at the time of infection, viability may drop after injection
into the oyster tissue, resulting in lower infections. The prezoo-
sporangia used in this study have been cultured in FTM, which
may have affected their infectivity. In the field, the infectivity of
prezoosporangia could be different. Oysters challenged with cells
from pure cultures of P. marinus (meronts, merozoites, and
schizonts) did not exhibit as heavy infections as those obtained
with meronts in homogenized oyster tissue (Volety and Chu, un-
published results, Bushek et al. 1993). Culture of P. marinus in
artificial media may reduce virulence of the cell stages.
Division of prezoosporangia into meront-like structures by
schizogony has been observed in culture (La Peyre 1993; Perkins,
O
Control
Meront
Treatment
Sporangia
personal communication). Although sporangia divide and release
biflagellated zoospores in seawater (Perkins 1976. Chu and
Greene 1989), the production of zoospores by meronts or prezoo-
sporangia in oyster tissue or in cells isolated from oyster tissue
without FTM treatment has not been documented. Indeed, the
production of biflagellated zoospores and their subsequent release
into seawater may not take place in oyster tissue. Furthermore, the
fate of inoculated prezoosporangia in oyster tissue is not known.
The lower prevalence in oysters challenged with prezoosporangia
may be the result of a long lag time in the division of sporangia
into meronts and/or the high mortality rate of cells induced to form
prezoosporangia.
Dittman (1993) reported insignificant differences in CI be-
tween lightly infected and uninfected oysters. However, in the
same study, significantly lower CI values were observed in heavily
Fig 4
Fig 5
Uninfected
Meront
Treatment
Sporangia
Figure 3. Mean CI (±SE) in uninfected, meront-, and prezoosporan-
gia-challenged oysters.
Figures 4 and 5. Mean serum P concentration (±SE) (Fig. 4) and
mean serum L (±SE) (Fig. 5) in uninfected and infected oysters chal-
lenged by meront and prezoosporangia. Bars with similar letters are
not significantly different (p > 0.05).
Infectivity of Two Life Stages of Perkinsus marinus
525
o
O
O 2
Time In days (post challenge)
Figures 6 and 7. Mean CI (±SE) (Fig. 6) and serum P concentration (±SE) (Fig. 7) in oysters at the end of 20, 40, 50, 65, and 75 days
postchallenge. Bars with similar letters are not significantly different (p > 0.05).
infected oysters compared with uninfected ones. Lower CI in in-
fected oysters challenged by prezoosporangia compared with un-
infected oysters in experiment 1 . although not statistically signif-
icant (Fig. 3), may be because only a few of the oysters were
heavily infected. The decrease in CI of oysters with time in ex-
periment 2 may be due to the stress in the confined environment.
The results from experiment 1 indicated that infected oysters
challenged by prezoosporangia had significantly lower P concen-
trations than did uninfected oysters. Lower tissue and hemolymph
protein have been observed in oysters heavily infected by H. nel-
soni (Ford 1986a and b. Barber et al. 1988, Ling 1990). However,
no significant differences in P concentrations were observed in
oysters lightly infected by P. marinus as compared with uninfected
oysters (Chu and La Peyre 1993a).
Lysosomal enzymes are believed to play a role in defense in
both vertebrates and invertebrates (Ingram 1980, Jolles and Jolles
1984), including molluscs (McDade and Tripp 1967a and b.
Cheng 1981 and 1983, Huffman and Tripp 1982, Moore and
Gelder 1985, Chu 1988). L in oysters was observed to be nega-
tively correlated with P. marinus infection and temperature (Chu
and La Peyre 1993a). L of uninfected oysters in experiment 1 had
significantly higher activities than infected oysters challenged with
meronts (Fig. 5). Lysozyme is hypothesized to be an important en-
zyme in resistance to P. marinus infection (Chu et al. 1993). The
absence of P. marinus infection in some of the oysters may have
been as a result of higher serum L. which may explain the signif-
icantly higher L in uninfected oysters. However, no difference in
L was observed between meront-challenged and prezoosporangia-
challenged oysters.
The higher prevalence, intensity, and weighted indices of P.
marinus infections in experiment 1 compared with experiment 2
may be due to the higher temperature in the former experiment
(Table 1). Temperature is one of the two most important factors
(the other being salinity) influencing the geographic distribution of
P. marinus in oysters. Chu and La Peyre (1993a) reported that
prevalence and intensity of P. marinus infection increased with
increasing temperature. In their study, the prevalence off. mari-
nus in oysters was 23, 46, 91, and 100% at 10, 15, 20, and 25°C
respectively. P. marinus infection is positively correlated with
temperature in the field (Soniat 1985, Craig et al. 1989, Soniat and
Gauthier 1989, Crosby and Roberts 1990, Gauthier et al. 1990).
The batches of P. marinus meronts used for challenging the oys-
ters in the two experiments were isolated from different infected
oysters. Their relative infectivity and virulence could differ, con-
tributing to the different infection rates. The difference in the
source of oysters may also have been one of the factors for the
lower incidence of P. marinus infection. Differences in the sus-
ceptibility of oysters from different populations to P. marinus
infection have been reported (Chu and La Peyre 1993b, La Peyre
1993). Their studies have shown differences in the prevalence of
P. marinus infection in oysters from three locations in the Ches-
apeake Bay and between Chesapeake Bay and Gulf oysters. Hab-
526
VOLETY AND CHU
itat and genetic dissimilarities were suggested as the reasons for
the differences in prevalence of infection.
Because only light infections were detected in experiment 2 in
both oysters challenged with meronts and prezoosporangia, the
insignificant differences noted in CI, L, and P between different
treatments were not surprising. These results agree with the find-
ings by Dittman (1993) and Chu and La Peyre (1993a). Neither
found differences in CI, L, and P concentrations between lightly
infected and uninfected oysters.
In summary, meronts are more infective than prezoosporangia
and are possibly the principal agents of disease transmission in
the field. The lower CI and P values in the treatment of infec-
ted oysters challenged with prezoosporangia, compared with un-
infected and meront-challenged oysters, suggest that prezoo-
sporangia may be exerting a higher energetic demand on the host
than do meronts. Further studies are needed to examine the causes
for the lower P concentrations in prezoosporangia-challenged oys-
ters.
ACKNOWLEDGMENTS
This study was supported by Grant NA16FL0402-01 from
NOAA through the oyster disease program. We thank Drs. Robert
Hale, Peter Van Veld. Frank Perkins, Richard Lee, and the anon-
ymous reviewers for critical review and helpful comments. The
authors thank Dr. Roger Mann for his kindness in providing the
spectrophotometer in his laboratory for lysozyme measurement.
Contribution No. 1904 from the Virginia Institute of Marine Sci-
ence, College of William & Mary.
LITERATURE CITED
Agresti, A. 1990. Categorical Data Analysis. John Wiley & Sons, New
York, pp 79-129.
Andrews. J. D. 1988. Epizootiology of the disease caused by the oyster
pathogen Perkinsus marinus and its effects on the oyster industry. Am.
Fish. Spec. Publ. 18:47-63.
Andrews, J. D. & W. G. Hewatt. 1957. Oyster mortality studies in Vir-
ginia. II. The fungus disease caused by Dermocystidium marinum in
oysters of the Chesapeake Bay. Ecol. Monogr. 27:1-25.
Barber, J. B., S. E. Ford & H. H. Haskin. 1988. Effects of the parasite
MSX [Haplosporidium nelsoni) on oyster (Crassostrea virginica) en-
ergy metabolism: II. Tissue biochemical composition. Comp. Bio-
chem. Physiol. 91A:603-608.
Burreson. E. M. 1989. Prevalence of the major oyster diseases of Virginia
waters — 1988. A summary of the annual monitoring program. Maine
Resource Report, 89-1, Virginia Institute of Marine Science. Glouces-
ter Point, Virginia.
Burreson, E. M. 1990. Status of the major oyster diseases in Virginia-
1989. A summary of the annual monitoring program. Marine Resource
Report, 90-1, Virginia Institute of Marine Science. Gloucester Point.
Virginia.
Bushek. D., S. K Allen. Jr., K A. Alcox. R Gustafson. & S. Ford.
1993. Dose response of the eastern oyster. Crassostrea Virginica, to
cultured cells of Perkinsus marinus. the agent of Dermo disease. J.
Shellfish Res. 13:313 (Abstract).
Cheng, T. C. 1981. Bivalves, pp. 233-300. In: N A. Ratchffe and A. F
Rowley, (eds.). Invertebrate Blood Cells, vol 1 . Academic Press, New
York.
Cheng, T. C. 1983. Triggering of immunologic defense mechanisms of
molluscan shellfish by biotic and abiotic challenge and its applications.
Mar. Tech. Soc. J. 17:18-25.
Chu. F.-L. & K. H. Greene. 1989. Effect of temperature and salinity on
in vitro culture of the oyster pathogen, Perkinsus marinus (Apicom-
plexa: Perkinsea). J. Invertebr. Pathol. 53:260-268.
Chu, F.-L. E. & J. F. La Peyre. 1989. Effect of environmental factors and
parasitism on hemolymph lysozyme and protein of American oysters
(Crassostrea virginica). J. Invertebr. Pathol. 54:224-232.
Chu, F.-L E. 1988. Humoral defense factors in marine bivalves. Am.
Fish. Soc. Spec Publ 18:178-188.
Chu, F.-L. E. & J. F. La Peyre. 1993a. Perkinsus marinus susceptibility
and defense related activities in eastern oysters Crassostrea virginica:
temperature effects. Dis. Aquat. Org. 16:223-234.
Chu, F.-L E. :. La Peyre. 1993b. Development of the disease
caused by the parasite, Perkinsus marinus and defense-related he-
molymph factors in three populations of oysters from Chesapeake Bay.
USA. J. Shellfish Res. 12:21-27.
Chu. F.-L. E., J. F. La Peyre & C. S. Burreson. 1993. Perkinsus marinus
infection and potential defense-related activities in eastern oysters,
Crassostrea virginica: salinity effects. J. Invert. Pathol. 62:226-232.
Craig, A, E. N.Powell. R. R.Fay&J. M.Brooks. 1989. Distribution of
Perkinsus marinus in Gulf Coast oyster populations. Estuaries 12:82-
91.
Crosby, M. P. & C. F. Roberts. 1990. Seasonal infection intensity cycle
of the parasite Perkinsus marinus (and an absence of Haplosporidium
spp.) in oysters from a South Carolina salt marsh. Dis Aquat. Org.
9:149-155.
Dittman. D E. 1993 The quantitative effects of Perkinsus marinus on
reproduction and condition index in the eastern oyster, Crassostrea
virginica. J. Shellfish Res. 12:127.
Ford, S. E. 1986a. Comparison of hemolymph proteins in resistant and
susceptible oysters, Crassostrea virginica, exposed to the parasite
Haplosporidium nelsoni (MSX). J. Invertebr. Pathol. 47:283-294.
Ford, S. E. 1986b. Effect of repeated hemolymph sampling on growth,
mortality, hemolymph protein and parasitism of oysters, Crassostrea
virginica. Comp. Biochem. Physiol. 85A:465— 470.
Ford, S. E. 1988. Host-parasite interactions in eastern oysters selected for
resistance to Haplosporidium nelsoni (MSX) disease: survival mecha-
nisms against a natural pathogen. Am. Fish. Soc. Spec. Publ. 18:206—
224.
Ford, S. E. & A. J. Figueras. 1988. Effects of sublethal infection by the
parasite, Haplosporidium nelsoni (MSX) on gametogenesis, spawning.
and sex ratios of oysters in Delaware Bay, USA. Dis. Aquat. Org.
4:121-133.
Gauthier. J. D. & W. S. Fisher. 1990. Hemolymph assay for diagnosis of
Perkinsus marinus in oysters Crassostrea virginica (Gmelin, 1791). J.
Shellfish Res. 9:367-372.
Gauthier, J. D . T. M Soniat & J S Rogers. 1990. A parasitologic^
survey of oysters along salinity gradients in coastal Louisiana. J World
Aquacul. Soc. 21:105-115.
Hewatt, W. G. & J. D. Andrews. 1956. Temperature control experiments
on the fungus disease, Dermocystidium marinum, of oysters. Proc.
Natl. Shellfish Assoc. 46:129-133.
Huffman, J. E. & M. R Tnpp. 1982. Cell types and hydrolytic enzymes
of soft shell clam (Mya arenaria) hemocytes J. Invert Pathol. 40:
68-74.
Ingram. G. A. 1980. Substances involved in the natural resistance of fish
to infection — a review. Fish. Biol. 16:23-60.
Jolles, P. & J Jolles 1984. What is new in lysozyme research? Always a
model system, today as yesterday. Mol. Biochem. 63:165-189.
La Peyre, J. F. 1993. Studies on the oyster pathogen Perkinsus marinus
(Apicomplexa): interactions with host defenses of C. virginica and C
gigas, and in vitro propagation. PhD dissertation. The College of
William & Mary, Williamsburg, pp 110-135.
Ling. W.-J. 1990. Cellular and humor responses of resistant and suscep-
tible oysters. Crassostrea virginica. to the infection of Haplosporidium
nelsoni (MSX). M.S. Thesis, University of Connecticut. 102 pp.
Lowry, O H , N. I. Rosebrough. A. L. Farr & R. J. Randall. 1951.
Infectivity of Two Life Stages of Perkinsus marinus
527
Protein measurement with the Fohn phenol reagent J Biol. Chem.
193:265-275,
Lucas, A. & P. G. Beninger. 1985. The use of physiological condition
index in marine bivalve aquaculture. Aquaculture 44:187-200.
Mackin, J. G. 1951. Histopathology of infection of Crassostrea virginica
(Gmelin) by Dermocyslidium marinum Mackin, Owen and Collier.
Bull. Mar. Sci. GuIfCaribb. 1:72-87.
Mackin. J. G. 1956. Dermocyslidium marinum and other microorganisms
in Louisiana. Proc. Natl. Shellfish Assoc. 46:116-133.
Mackin. J. G. 1962. Oyster disease caused by Dermocyslidium marinum
and other microorganisms in Louisiana. Puhl. Inst. Mar. Sci. Univ.
Texas 7:132-229.
McDade, J. E. &M. R.Tripp. 1967a. Lysozyme in the hemolymph of the
oyster, Crassostrea virginica. J. Invert. Pathol. 9:531-535.
McDade, J. E. & M. R. Tripp. 1967b. Lysozyme in oyster mantle mucus
J. Invert. Pathol. 9:581-582.
Menzel. R. W. & S. H. Hopkins. 1955. The growth of oysters parasitized
by the fungus Dermocyslidium marinum and by the trematode Buce-
phalus cuculus. J. Parasitol. 41:333-342.
Moore, C. A. & S. R. Gelder. 1985. Demonstration of lysosomal en-
zymes in hemocytes of Mercenaria mercenaria (Mollusca:bivalvia).
Trans. Am. Microscopical. Soc. 104:242-249.
Paynter. K. T. & E. M. Burreson. 1991. Effects of Perkinsus marinus
infection in the eastern oyster. Crassostrea virginica: II. Disease de-
velopment and impact on the growth rate at different salinities. J.
Shellfish Res. 10:425-431.
Perkins, F. O. 1966. Life history studies of Dermocyslidium marinum, an
oyster pathogen. Dissertation. Florida State University, 273 pp.
Perkins. F. O. 1976. Zoospores of the oyster pathogen. Dermocystidium
marinum. I. Fine structure of the conoid and other sporozoan-like
organelles. J. Parasitol. 62:959-974.
Ragone. L. M 1991. The effect of low salinity on established infections
of Perkinsus marinus (Apicomplexa: Perkinsasida) in the eastern oys-
ter, Crassostrea virginica. M.S. Thesis, College of William and Mary,
Williamsburg, 52 pp.
Ragone, L. M. & E. M. Burreson. 1993. Effect of low salinity on infec-
tion progression and pathogenicity of Perkinsus marinus in the eastern
oyster, Crassostrea virginica (Gmelin, 1791 ). J. Shellfish Res. 12:1-7.
Ragone Calvo, L. M. & E. M. Burreson. 1994. Characterization of over-
wintering infections of Perkinsus marinus (Apicomplexa) in Chesa-
peake Bay oysters. J. Shellfish Res. 13:123-130.
Ray, S. M. 1952. A culture technique for the diagnosis of infections with
Dermocyslidium marinum Mackin, Owen and Collier in oysters. Sci-
ence 116:360-361
Ray, S. M. 1954. Biological studies of Dermocystidium marinum, a fun-
gus parasite of oysters. Rice Institute pamphlet, 1 14 pp. (monograph in
Biological Special Series Issue).
Ray, S. M. 1966. A review of a culture method for detecting Dermocys-
tidium marinum with suggested modifications and precautions. Proc.
Natl. Shellfish Assoc. 54:55-80.
Shugar, D. 1952. The measurement of lysozyme activity and the ultra-
violet inactivation of lysozyme. Biochim. Biophys. Acta 8:302-309.
Soniat, T. M. 1985. Changes in levels of infection of oysters infected by
Perkinsus marinus, with special reference to the interaction of temper-
ature and salinity upon parasitism. Northeast Gulf Sci. 7:171-174.
Soniat, T. M. & J. D. Gauthier. 1989. The prevalence and intensity of
Perkinsus marinus from the mid northern Gulf of Mexico, with com-
ments on the relationship of the oyster parasite to temperature and
salinity. Tul. Stud. Zool. Boi. 27:21-27.
Journal of Shellfish Research, Vol. 13, No. 2, 529-537. 1994.
MORPHOLOGICAL AND GENETIC VARIATION AMONG THREE POPULATIONS OF CALICO
SCALLOPS, ARGOPECTEN GIBBUS
M. K. KRAUSE,1* W. S. ARNOLD,2 AND W. G. AMBROSE JR.3t
Department of Ecology and Evolution
SUNY Stony Brook
Stony Brook, New York 11794
'Florida Department of Environmental Protection
Florida Marine Research Institute
St. Petersburg, Florida 33701-5095
^Department of Biology and
Institute for Coastal and Marine Resources
East Carolina University
Greenville, North Carolina 27858
ABSTRACT We surveyed morphological and genetic variation among three populations of Argopecten gibbus from the Marquesas
Keys. Florida; Cape Canaveral, Florida; and Cape Lookout, North Carolina; in order to determine the extent of the genetic isolation
of these populations and to examine the hypothesis for larval transport between populations. Bumaby size-adjusted principal com-
ponent analysis of 14 morphological characters revealed significant differences among sites for the third principle component, which
explained 6% of the total variation. Analyses of electrophoretic loci showed significant allele frequency heterogeneity among sites for
one of seven polymorphic loci. The magnitude of the morphological and genetic differences between the Marquesas Keys sample and
both of the Atlantic coast samples was generally greater than between the geographically more distant Cape Canaveral and Cape
Lookout samples, although overall levels of variation among sites were small. Estimates of gene flow suggest that relatively frequent
migration, sufficient for panmixia in the absence of historical divergence, may occur between populations. Our results suggest that
oceanographic processes play a critical role in the transport of A. gibbus larvae between populations. We recommend that management
of the A. gibbus fishery should include consideration of these processes.
KEY WORDS: Argopecten, calico scallop, population structure
INTRODUCTION
The calico scallop. Argopecten gibbus (Linnaeus), occurs in
open coastal waters throughout the Gulf of Mexico and off the
western Atlantic Coast from Florida to Cape Hatteras, North Caro-
lina, in depths of 1 to 370 m (Waller 1969, Blake and Moyer
1991). During the 1940s and 1950s, large "beds" of calico scal-
lops were discovered near Cape San Bias and Cape Canaveral,
Florida, and Cape Lookout, North Carolina (Allen and Costello
1972). Those beds have been commercially exploited on a more or
less continuous basis since the late 1950s and remain the most
productive harvest areas. Since its inception, the commercial fish-
ery has followed a trend of increasing landings, but as is typical of
many short-lived scallop species, there has been considerable vari-
ability in landings among years. For example, annual landings
from the Cape Canaveral calico scallop beds have ranged between
39,000 and 11,000.000 pounds of meats over the last 5 years
(Blake and Moyer 1991). The Cape San Bias and Cape Lookout
beds are even more transient, typically producing commercial
quantities every 2 to 4 years.
Management of the calico scallop fishery has proved difficult
because of the extreme year-to-year variability in landings and
because the extent to which populations are self-sustaining or are
dependent on allochthonous larval recruitment is unknown. Al-
though the species' range is broad, dense concentrations of calico
*Current address: Natural Science Division, Southampton College, LIU,
Southampton, NY 11968-4198
^Current address: Biology Department, Bates College, Lewiston, Maine
04240
scallops are primarily associated with coastal prominences, includ-
ing the capes mentioned above and, to a lesser extent, the Florida
Keys (Cummins 1971, Allen and Costello 1972). This discontin-
uous distribution suggests that oceanographic processes may play
an important role in structuring calico scallop populations. Capes
and headlands, such as those found near large beds of scallops, are
capable of generating downstream eddies, which can cause aggre-
gations of zooplankton (Alldredge and Hamner 1980, Wolanski
and Hamner 1988). High local abundances of other scallop spe-
cies, including Patinopecten yessoensis, Amusium bolloti,
Chlamys tehuelcha, and Chlamys patagonica, have been attributed
to larval retention in the presence of such eddies (Yamamoto 1964.
Heald and Caputi 1981, Orensanz et al. 1991a). Eddying in the
vicinity of Cape Canaveral has also been suggested to enhance the
retention of calico scallop larvae until settlement (Bullis and Cum-
mins 1961. Allen 1979). North Carolina stocks of calico scallops,
in particular, may be strongly influenced by oceanographic pro-
cesses that affect larval transport. Kirby-Smith (1970) proposed
that a considerable proportion of the North Carolina stock may
result from larvae originating from the Cape Canaveral, Florida,
beds, carried by the Gulf Stream, and retained by eddying in
coastal bays. Similar transport and retention mechanisms may be
responsible for the concentrations of calico scallops that occasion-
ally occur off of the Georgia coast near Savannah (Anderson and
Lacey 1979). Although the 14 to 16 day planktonic period esti-
mated for A. gibbus larvae (Costello et al. 1973) provides the
potential for considerable transport by offshore currents, the hy-
pothesis of such larval transport has not been tested.
Extreme annual variation in commercial landings of A. gibbus
supports the assertion that oceanographic processes are important
529
530
Krause et al.
in controlling the distribution and abundance of calico scallop
populations. Orensanz et al. (1991b) observed that many scallop
species have irregular fluctuations in abundance when recruitment
is strongly dependent on hydrographic conditions. If processes
such as larval retention by eddies or transport by currents are
important for calico scallop recruitment, successful recruitment
will depend on the concurrence of such events with spawning. The
relatively short lifespan of A. gibbus ( 18 to 24 months. Roe et al.
1971) makes this species particularly vulnerable to fluctuations in
recruitment success, which may also contribute to interannual
variability in stock abundance (Orensanz et al. 1991b). Failure of
a single recruitment event can thus potentially cause the local
extinction of scallops. North Carolina calico scallops spawn once
a year, from January to April (Singhas 1992), whereas Florida
populations of calico scallops have two spawning periods: the
main spawn occurs in late spring when the majority of spawning
scallops are small, which is followed by a smaller event in the fall
when most spawning scallops are large (Blake and Moyer 1991).
In 1984, the failure of the fall recruitment in the Cape Canaveral
population apparently resulted in the collapse of the fishery in
1985, because the spring 1985 spawning stock was limited to the
few large scallops that survived the winter (Moyer and Blake
1986). The extent to which larval recruitment from other popula-
tions mitigates such local population crashes and local interannual
variability in abundance is unknown.
The relative importance of biological and oceanographic fac-
tors in structuring scallop populations, particularly calico scallops,
remains largely undetermined. Quantification of the degree of ge-
netic and morphological variation along the scallop's range may
provide important clues concerning the extent of larval transport
and gene flow between stocks. Waller (1969) characterized the
morphology of fossil and extant calico scallops from the entire
geographic range of the species. He found no evidence for exten-
sive geographic variation; however, samples from the Florida
Keys were slightly distinct morphologically from scallops in other
locations in Florida and North Carolina. With the exception of that
study, there are no descriptions of morphological, physiological,
or genetic variation among A. gibbus populations. We present here
a preliminary attempt to describe the degree of morphological and
genetic variation among three populations of calico scallops sam-
pled from the Gulf of Mexico and Atlantic Coasts of North Amer-
ica. The purposes of this study were to define the extent of isola-
tion among stocks from the eastern and western Florida coasts and
from North Carolina and to examine the hypothesis that extensive
larval transport occurs between these stocks.
METHODS
Collections
Calico scallops were collected from the vicinity of Marquesas
Keys, Florida, in the Gulf of Mexico (24°43'N, 82°10'W) on July
7, 1990; from the vicinity of Cape Canaveral, Florida (28°20'N,
80°10'W), on January 13, 1990, and February 15, 1990; and from
the vicinity of Cape Lookout, North Carolina (34°35'N,
76°35'W), on February 2, 1990. In all cases, scallops were col-
lected with modified otter trawls in depths of 20 to 40 m. Samples
were either dissected immediately after capture or were returned
alive to the laboratory for dissection. Subsequent to dissection,
tissue samples were held in a freezer at - 70°C. Frozen samples of
adductor muscle and digestive gland were lyophilized for storage
before electrophoretic analyses.
Electrophoretic Analysis
Samples were prepared for electrophoresis by homogenization
in 0.1 M Tris-HCl buffer, pH 8.0, with 20% glycerol w/v and
centrifuged for 10 minutes at 5,000 x g at 4°C. Electrophoresis of
the supernatant was performed on 280 mm horizontal starch gels.
Four buffer systems were used to resolve the enzymes. The 1 1
enzymes studied and the buffer and staining systems used to detect
them are shown in Table 1 . Two isozymes were present for MDH,
AAT, ARK, and IDH, although we were only able to include the
most cathodal isozyme in our analyses from AAT, ARK. and IDH
because of inconsistent resolution. Alleles were designated ac-
cording to approximate relative mobility on the gel to that of the
most common allele, which was arbitrarily assigned a value of
100.
Statistical analyses of genetic data were performed with pro-
grams from the Statistical Analysis System software (SAS Insti-
tute, Inc. 1982). For polymorphic loci (frequency of the common
allele =£0.95), we compared allele frequencies between samples
and genotype frequencies with Hardy-Weinberg equilibrium ex-
pectations by pooling rare alleles with the electrophoretically clos-
est common allele to obtain allelic class frequencies with N & 5.
Observed genotypic frequencies were compared with those ex-
TABLE 1.
List of enzymes assayed, their abbreviations and enzyme number, and the bufTer and staining protocol used for their detection in the
electrophoretic study of A. gibbus.
Enzyme
E.C. Number
Abbreviation
Buffer"
Staining Reference
Aspartate aminotransferase
2.6.1.1
AAT
1
Argininc kinase
2.7.3.3
ARK
1
Isocitrate dehydrogenase
1.1.1.42
IDH
1
Glycerol-3-] hosphate dehydrogenase
1.1.1.8
GPD
2
Pyruvati
2.7.1.40
PK
1
Glucose-6- phosphate isomerase
5.3.1.9
GPI
3
Phosphoglucomutaxe
2.7.5.1
PGM
4
Octopine dehydrogenase
1.5.1.11
ODH
3
6-Phosphogluconale dehydrogenase
1.1.1.44
6PGD
2
Malic dehydrogenase
1.1.1.37
MDH
2
Mannose phosphate dehydrogenase
5.3.1.8
MPI
1
Johnson and Utter (1973)
Scopes (1968)
Schaal and Anderson (1974)
Shaw and Prasad (1970)
Hams and Hopkinson ( 1976)
Schaal and Anderson (1974)
Schaal and Anderson (1974)
Dandoet al. (1981)
Schaal and Anderson (1974)
Schaal and Anderson (1974)
Schaal and Anderson (1974)
' Buffers were as follows: 1. Tris-citrate pH 7.5 (Rodhouse and Gaffney 1984); 2, Tris-citrate pH 7.0 (Rodhouse and Gaffney 1984); 3, discontinuous
lithium hydroxide (Koehn et al. 1976); 4, Tris-maleate pH 7.0 (Koehn et al. 1984).
Calico Scallop Morphology and Genetics
531
pected under Hardy-Weinberg equilibrium using the G-test and.
when necessary, William's correction for small sample sizes
(Sokal and Rohlf 1981 ). Alpha levels were corrected to avoid type
I errors from multiple tests of a single hypothesis by use of the
sequential Bonfcrroni technique (Holm 1979. Rice 1989). Hetero-
zygote deficiencies or excesses were calculated using the D sta-
tistic = (H0 - Hc)/Hc (Selander 1970). Allele frequencies among
sampling sites were compared using the R x C test of indepen-
dence and the G statistic (Sokal and Rohlf 1981 ), with significance
levels adjusted by use of the sequential Bonferroni technique
(Holm 1979. Rice 1989). Effective numbers of alleles were cal-
culated as nc = 1/2 p,2. where p, is the frequency of the ith allele
in the population (Hartl and Clark 1989). We determined the fix-
ation index, FST. for all polymorphic loci between pairs of samples
and across all samples using the formula of Wright, with correc-
tion for limited sample sizes (Wright 1978). Gene flow was esti-
mated from this measure of population differentiation by the re-
lation Nm = [(1/FST) - l]/4 (Slatkin 1985a, 1987). The degree
of population subdivision was also examined by calculating stan-
dard genetic distances between populations using the formulae of
both Nei (1978) and Rogers (1972) applied to all polymorphic and
monomorphic loci scored.
Morphometric Analysis
Morphometric parameters were selected following the criteria
of Waller (1969). Only mensural parameters suitable for use in
subsequent statistical treatments were included (Table 2). Mea-
sures of the pallia] scar and assorted muscle scars were not used
because of the difficulty of resolution of those features in many of
our shell samples. In all cases, we used measurements collected
from the right valve of the specimen. A total of 52, 40, and 52
scallops were morphometrically analyzed from the Marquesas
Keys. Cape Canaveral, and Cape Lookout, respectively.
Morphological data were quantitatively compared by the use of
Burnaby size-adjusted principal component analysis (Burnaby
1966), which removes the influence of variation in the first eigen-
vector (principal component 1 [ PC 1 ] ) of the variance-covariance
matrix from each population (Rohlf and Bookstein 1987). This
approach was taken to minimize the influence of animal size on
subsequent statistical analyses. Although the first principal com-
ponent only gives the direction of maximum variation within the
populations, this will be a size-related component if there is a large
difference in size among populations (Rohlf and Bookstein 1987);
the remaining principal components are then primarily shape re-
lated. We next plotted principal component 2 (PC2) against prin-
cipal component 3 (PC3) to determine the relative distribution of
the dominant shape-related principal components among popula-
tions. Finally, for each of the second and third principal compo-
nents, adjusted shape vector scores (ASVS) were subjected to
Model 1 one-way analysis of variance ( ANOVA) to compare mean
ASVS among populations (Sokal and Rohlf 1981). Significant
differences among means were further tested with the Student-
Newman-Keuls (SNK) multiple range test (Sokal and Rohlf 1969).
Unless otherwise stated, a type I error rate of 0.05 was used for all
statistical tests.
RESULTS
Genetics
On the basis of the examination of the 12 electrophoretic loci,
the proportion of polymorphic loci, P, in the three samples of
Argopecten gibbus was 0.64. Allele frequencies for the seven
polymorphic loci and sample sizes for electrophoretic analyses
are shown in Table 3. Ark- 1 , Aal-1 . Pk, and 6Pgd were completely
monomorphic. whereas Idhl exhibited very low levels of poly-
morphism (frequency of the common allele >0.95) in all of the
sampled populations. The average level of heterozygosity (H) over
all samples and loci was 0.103. whereas the effective number of
alleles per locus ranged from 1.00 to 4.53 (Table 3).
In general, isozyme allele frequencies agreed with Hardy-
Weinberg equilibrium expectations (Table 3). With one exception.
TABLE 2.
Description of parameters used in the morphological analysis of A. gibbus (from Waller 1969) and the mean and standard deviation (SD) for
each parameter for each population.
Site-
Parameter
Description
MK
cc
CL
Mean
SD
Mean
SD
Mean
SD
20.4
2.1
44.2
1.9
46.5
1.6
10.3
1.2
24.1
1.1
26.2
1.1
10.4
1.2
22.7
1.2
23.1
1.3
11.0
1.0
23.2
1.6
24.1
1.6
10.1
1.1
21.0
1.5
22.4
1.6
6.5
0.8
13.6
0.8
13.0
0.8
3.0
1.1
7.2
0.5
7.5
0.6
5.4
1.3
10.7
0.7
10.3
0.8
1.3
0.2
2.8
0.2
3.1
0.2
0.6
0.1
1.3
0.2
1.3
0.1
1.0
0.1
2.1
0.2
2.3
0.1
1.3
0.2
2.6
0.3
2.9
0.3
5.8
2.3
14.3
1.3
13.8
1.4
5.8
1.8
13.8
0.9
13.9
1.4
4.6
1.7
12.4
0.9
13.1
1.2
AM Height of disk (linear)
AD Posterior half-diameter of disk
DG Anterior half-diameter of disk
AK Posterior dorsal half-diameter of disk
GP Anterior dorsal half-diameter of disk
LO Convexity
EI Height of anterior auricle
BJ Height of posterior auricle
PW Plical width
IW Inlerplical width
ad Height of resilial insertion
ce Length of resilial insertion
DF Length of anterior outer ligament
CD Length of posterior outer ligament
DE Distance between two lines perpendicular to outer ligament, one
passing through origin of growth and the other passing through
anterior auricular notch
a MK, Marquesas Keys, Florida; CC, Cape Canaveral, Florida; and CL, Cape Lookout, North Carolina.
532
K.RAUSE ET AL.
TABLE 3.
A. gibbus allele frequencies (f), heterozygosities (H), and effective
numbers of alleles (ne) at seven polymorphic enzyme loci from three
sampling sites and G-tests (G) for fit to Hardy-Weinberg
equilibrium expectations. For the G-tests, rare alleles were pooled
with the electrophoretically closest common allele so that f 3 5.
N = sample size.
Locus Allele
6Pgd
Mdhl
Mdh2
Pgm
Odh
Mpi
Gpi
Marquesas
Keys
f G
Site
Cape
Canaveral
f G
Cape
Lookout
f
106
104
100
96
rare
H
n„
110
104
100
90
80
rare
H
ne
104
100
94
rare
H
ne
102
100
96
95
92
90
88
rare
H
nc
106
100
94
rare
H
ne
100
96
rare
H
ne
100
rare
H
n„
0.025
0.275
0.625
0.050
0.025
0.275
2.13
0.400
0.075
0.150
0.050
0.325
0.000
0.400
3.38
0.125
0.850
0.000
0.025
0.050
1.35
0.075
0.625
0.175
0.075
0.025
0.000
0.000
0.025
0.225
2.31
0.000
0.950
0.025
0.025
0.050
111
0.650
0.350
0.000
0.200
1.83
0.950
0.050
0.050
ill
40
0.104
0.601
0.623
0.032
0.055
0.292
0.055
0.059
0.118
0.794
0.029
0.000
0.118
1.54
0.044
0.088
0.647
0.044
0.162
0.015
0.206
2.19
0.132
0.794
0.059
0.015
0.206
1.53
0.044
0.382
0.221
0.118
0.088
0.029
0.059
0.059
0.294
4.53
0.073
0.853
0.029
0.045
0.147
1.36
0.368
0.618
0.014
0.235
1.93
0.971
0.029
0.029
1.06
68
0.103
7.204"
1.612
2.285
0.111
0.031
0.089
0.031
0.011
0.109
0.826
0.054
0.000
0.073
1.43
0.000
0.042
0.583
0.135
0.229
0.011
0.240
2.43
0.135
0.813
0.031
0.021
0.240
1.47
0.052
0.458
0.219
0.125
0.021
0.052
0.031
0.042
0.365
3.56
0.021
0.865
0.042
0.072
0.146
1.33
0.531
0.458
0011
0.229
2.03
0.948
0.052
0.052
III
96
0.106
13.716"
0.028
1.546
0.397
0.145
0.709
0.145
* p < 0.05 after adjustment with the sequential Bonferroni technique
b p < 0.01 after adjustment with the sequential Bonterroni technique.
c Insufficient data for analysis.
there was no tendency for loci to exhibit significant, consistent
heterozygote deficiencies or excesses. The 6Pgd locus, however,
showed significant heterozygote deficiency in both the Cape
Canaveral and Cape Lookout populations (D = 0.460 and 0.546,
respectively).
Allele frequency comparisons revealed relatively little genetic
differentiation among the sampled populations of calico scallops
(Table 4). Of seven polymorphic loci, only Mdhl differed signif-
icantly among all three populations. The observed differences
were generally greater between Cape Canaveral and Marquesas
Keys scallops, which differed significantly in allele frequencies
for both Mpi and Mdhl , than between Cape Canaveral and Cape
Lookout populations, which showed statistically homogeneous al-
lele frequencies for all loci. These relationships among sites were
confirmed by the genetic distances between the samples calculated
using all scored loci (Table 5). In order to further assess the degree
of genetic differentiation among populations, FST values deter-
mined for each locus including Idhl were used to derive a rough
estimate of gene flow, Nm, which may be interpreted as the num-
ber of diploid individuals exchanged among populations per gen-
eration. Values of Nm varied considerably among loci, ranging
from 1.7 to 1935.0 (Table 6), with an estimate across loci and
populations of 5.5. The locus that had significant allele frequency
differences among populations, Mdhl . contributed disproportion-
ately to the overall FST estimate. Exclusion of Mdhl from the
calculations of FST and Nm resulted in a twofold higher overall
estimate of gene flow of 12.25. Pairwise FST calculations sug-
gested that gene flow was, in general, much greater between the
Cape Lookout and Cape Canaveral populations than between the
Marquesas Keys and either of the Atlantic Coast samples (Ta-
ble 6).
Morphology
Fifteen mensural characteristics were deemed suitable for in-
clusion in the morphometric analysis of calico scallop shells (Ta-
ble 2), including measures of disc shape (AM. AD, DG, AK, GP,
and LO), auricular shape (El and BJ), plical shape (PW and 1W),
TABLE 4.
G-test statistic comparisons of .4. gibbus allele frequencies at seven
polymorphic enzyme loci from three sampling sites: Marquesas
Keys (MK), Cape Canaveral (CC), and Cape Lookout (CL). For
frequency comparisons, rare alleles were pooled with the
electrophoretically closest common allele so that f & 5.
MK x CC x CL CC x CL CC x MK CL x MK
Locus
df
df
df
df
6Pgd
Mdhl
Mdh2
Pgm
Odh
Mpi
Gpi
7.480
52.321"
1.337
15.95
3.395
7.812
1.911
4
6
4
10
4
1
1.499
7.958
0.315
5.641
0.481
3.565
1.767
3.757
28.931b
1.313
8.714
2.744
7.308°
1.112
6.475
45.227"
0.595
7.819
2.410
1.642
0.003
a Degrees of freedom.
"p < 0.01 after correction of a levels with the sequential Bonferroni
technique.
c p < 0.05 after correction of a levels with the sequential Bonferroni
technique.
Calico Scallop Morphology and Genetics
533
TABLE 5.
Nei's (1978) genetic distance (upper diagonal) and Rogers (1972)
genetic distance (lower diagonal) between the three populations of
calico scallops, calculated from electrophoretic data.
North
Carolina
Cape
Canaveral
Marquesas
Keys
North Carolina
Cape Canaveral
Marquesas Keys
0.061
0.105
0 009
0.136
0.050
0.070
dimensions of resilial insertion (ad and ce), and ligament shape
(DF. CD. and DE). All characters were normally distributed.
The mean disc height (AM) of the three populations was sig-
nificantly different (type I ANOVA, p < 0.001). The Marquesas
Keys specimens were generally less than half the size of the Cape
Canaveral and Cape Lookout populations, and the mean disc
height of the Cape Canaveral population was 2.3 mm less than that
of the Cape Lookout population (Table 2). Mean disc height of
each population was significantly different from that of the other
two populations (SNK, a = 0.05). The observed difference in
overall size necessitated the application of the size-adjusted Bur-
naby analysis, which legitimized the interpretation of PC2 and
PC3 as the primary shape vectors.
Of the three principal components considered in this analysis,
PCI contributed 39.0%, PC2 contributed 21.3%. and PC3 con-
tributed 6.9% to the total variance of the model. Within PC2, the
model was dominated by the influence of the auricular shape vari-
able BJ. which contributed 81.2% of the observed variance. The
ligament shape variable DE contributed 7.2% to the model, and
each of the other variables contributed less than 2.5% to the re-
maining variance. Within PC3. the model was influenced by the
ligament shape variables DF (40.1% of total variance) and CD
(23.3% of total variance). The variables GP (9.5%). AD (6.3%),
PW (4.0%). AK (4.0%), LO (2.9%), and IW (2.5%) all contrib-
uted at least 2.5% to the total variance.
Results of the Bumaby size-adjusted principal component anal-
ysis provided no clear discrimination among the Cape Canaveral,
Cape Lookout, and Marquesas Keys populations, particularly
along the PC2 axis, although the spread of the Marquesas Keys
population along the PC2 axis was considerably greater than the
spread for the Cape Canaveral and Cape Lookout populations (Fig.
1). There was a significant difference (p = 0.0001) among mean
ASVS along the PC3 axis; SNK test results indicate that the mean
ASVS of each of the three populations was significantly different
from that of the other two populations. The Marquesas Keys pop-
ulation showed the largest spread along the PC3 axis and the least
overlap with the other two populations. However, mean ASVS for
Cape Canaveral and Cape Lookout were also significantly differ-
ent from one another, although considerable overlap among ASVS
for the Cape Canaveral and Cape Lookout populations was ob-
served.
DISCUSSION
The study reveals relatively low overall levels of morphologi-
cal and genetic variation among three geographically distant A.
gibbus populations, although significant heterogeneity was present
among samples for a small proportion of the characters examined.
In general, the magnitude of differentiation was greater in com-
parisons of the Marquesas Keys population with both the Cape
Canaveral and Cape Lookout populations than in comparisons be-
tween these two East Coast samples. These data appear to indicate
that the Florida peninsula serves as a partial barrier to gene flow
between the Gulf Coast and the Atlantic Coast A. gibbus popula-
tions, although relatively frequent migration may occur between
all populations.
The samples of calico scallops measured showed only slight
morphological differentiation, with the exception of the small size
of the Marquesas Keys scallops; however, the Marquesas scallops
may have been juveniles. This agrees with earlier observations by
Waller (1969) that calico scallops generally lack the geographic
morphological variation exhibited by other pectinid species such
as Argopecten irradians and Placopecten magellanicus. After cor-
rections for size differences, A. gibbus populations were only dis-
tinguished by one of two shape-related principal components
(PC3), and this difference was greatest between the Marquesas
Keys and Atlantic Coast samples. Because this component con-
tributed only 69c to the total morphological variance, these results
suggest that A. gibbus populations are similar in their overall mor-
phological composition, although there are subtle differences
among populations, particularly between the Gulf and the Atlantic
Coasts, within a small suite of morphological characters. The spe-
cific characters that differed among populations were principally
measures of ligament shape, which often exhibit geographical
variation in members of the genus Argopecten (Waller 1969).
Waller (1969) suggested that such variation may represent local
environmental influences on morphology, but this has yet to be
confirmed.
In agreement with our results based on shell morphology, ge-
netic differences among the three A. gibbus populations were
small, although the relative magnitude of differentiation was
greater between the Marquesas Keys population and Cape Canav-
eral population than between the Cape Canaveral and Cape Look-
out populations, which are more geographically separated. Ge-
netic distances between A. gibbus populations fall within the
ranges of those calculated for several other scallop species (Kijima
et al. 1984, Beaumont and Zouros 1991), which typically reveal
moderate levels of genetic differentiation but no evidence for ex-
tensive population isolation. Concordantly. although significant
allele frequency heterogeneity was observed among all samples for
Mdhl , and additionally for Mpi between the Marquesas Keys and
Cape Canaveral samples, our estimates of gene flow indicate that
relatively frequent migration occurs, or has historically occurred,
among all of the populations. In an island model of equilibrium
population structure, frequencies of neutral alleles will not diverge
among populations when Nm S> 1 . where N is the effective pop-
ulation size, and m is the migration rate among populations (Slat-
kin 1985b). Our overall and pairwise values of Nm show that
migration around the Florida Peninsula occurs at a reduced rate
compared with that along the Atlantic Coast, but all Nm values are
well above one, suggesting that gene flow is sufficient for pan-
mixia in the absence of historical divergence. However, the sig-
nificant heterogeneity among populations revealed by a small pro-
portion of the morphological and genetic characters examined in-
dicates that, on an evolutionary time scale, stochastic or selective
factors have allowed for some degree of population differentiation.
One possible hypothesis to account for these observations is that
historical biogeographical processes may have led to some degree
of divergence among geographically distant populations, particu-
larly between the Gulf and Atlantic Coasts, that is partially coun-
teracted by migration between these populations. Results from
534
Krause et al.
TABLE 6.
Pairwise and overall FST values and estimated numbers of migrants (Nm) between populations per generation on the basis of calculated FST
values for A. gibbus. See Table 2 for population abbreviations.
MK x
"C x CL
CC
x CL
CC
x MK
CL x
MK
Locus
Fsx
Nm
Fst
Nm
FST
Nm
FST
Nm
6Pgd
0.027
9.01
0.001
226.4
0.021
11.52
0.032
7.6
Mdhl
0.118
1.9
0.004
62.3
0.131
1.7
0.116
1.9
Mdh2
0.001
199.1
0.001
288.1
0.002
115.8
0.001
428.1
Pgm
0.014
18.0
0.001
222.4
0.020
12.3
0.009
28.8
Odh
0.009
26.5
0.002
153.3
0.015
169
0.007
33.8
Mpi
0.043
5.6
0.020
12.4
0.066
3.6
0.005
52.3
Gpi
0.002
101.4
0.001
180.2
0.001
1935.0
0.001
205.8
Idhl
0.012
20.2
0.008
30.3
a
— a
0.008
30.3
Average
0.043
5.5
0.005
50.1
0.053
4.5
.038
6.4
a Both populations were monomorphic for the same allele
studies of a number of coastal species whose ranges span the Gulf
and Atlantic Coasts of North America also show genetic discon-
tinuities in the region of the Florida Peninsula, although often to a
much greater extent than the oceanic A. gibbus. These observa-
tions support the hypothesis of historical isolation resulting from
climatic and sea level changes (Saunders et al. 1986, A vise et al.
1987, Reeb and Avise 1990, Karl and Avise 1992). Alternatively,
various selective scenarios could be advanced in order to account
for the disparity between the population heterogeneity indicated by
a small proportion of our genetic and morphological characters and
the homogeneity indicated by the majority of characters examined.
Although the data presented in this initial study are inadequate to
rigorously examine such hypotheses, it is clear that overall levels
of divergence among A. gibbus populations are small, particularly
along the Atlantic Coast, and that migration from Florida popula-
tions to more northern populations may occur at a relatively fre-
quent rate.
The influence of oceanographic processes may be particularly
critical to the dispersal of A. gibbus larvae and may maintain the
relative genetic and morphological homogeneity among popula-
tions for the majority of traits we examined. This species is found
primarily in deeper shelf water from depths of 10 to 400 m (Waller
1969. Allen and Costello 1972). which contrasts with the shallow-
water habitat occupied by Argopecten irradians and Crassostrea
■4-1
c
CD
c
o
Q_
E
o
o
"5
'o
c
Q_
0.700
0.600
0.500
0.400
0.300 -
0.200 -
0.100
^ Marquesas Keys
go* S
D
D
o Cape Canaveral
□ Cape Lookout
OnO0iS)D
□
o o^0
A
A* O
AA A A A
*u f >
AA
A A
/fAA
A
* A
-
A
*£
A
l l I
1
1 1 1
-0.600 -0.500 -0.4-00 -0.300 -0.200 -0.100
0.000
0.100
0.200
Principal Component 2
Figure 1. Bivariate plot of adjusted shape vector scores for PC3 against PC2 from results of Burnaby (1966) size-adjusted principal component
analysis of t gibbus morphology.
Calico Scallop Morphology and Genetics
535
virginica along the East and Gulf Coasts of North America. This
oceanic distribution places A. gibbus more directly under the in-
fluence of the Loop Current, Florida Current, and Gulf Stream
than those estuarine species. Several specific features of this cur-
rent system may enhance larval migration between the studied
areas near the Marquesas Keys. Cape Canaveral, and Cape Look-
out, supporting our hypothesis for high rates of gene flow among
these populations. In particular, the Florida Current exhibits oc-
casional wavelike meanders along its western edge (Boudra et al.
1987). and the Gulf Stream also shows characteristic wavelike
meanders as well as large frontal eddies (Lee and Atkinson 1983,
Lee et al. 1985 and references therein). These eddies and mean-
ders prevail, are strengthened in regions near capes and shoals
(Blanton et al. 1981) that are associated with calico scallop con-
centrations, and induce upwelling of nutrient-rich waters through
intrusions of cold bottom water over shelf regions (Blanton et al.
1981, Hofmann et al. 1981, Lee and Atkinson 1983). In order to
spawn, calico scallops require relatively cool waters (<22°C)
compared with the typical summer and fall coastal water temper-
atures (Blake and Moyer 1991). The upwelling caused by mean-
ders and eddies provides the opportunity for shelf waters to reach
these lower temperatures (Miller et al. 1981 ) and allows for larvae
to become associated with these water masses. Scallop spawning
near Cape Canaveral is known to be associated with such intru-
sions of cold water (Allen and Costello 1972. Miller et al. 1981 ).
Water masses associated with the meanders and eddies of the
Florida Current and Gulf Stream maintain their integrity over large
distances and for extended periods of time (Leming 1979, Hof-
mann et al. 1981, Lee et al. 1981. Tester et al. 1991). After
scallop spawning, these water masses may transport and concen-
trate large numbers of larvae from the west to the east coasts of
Florida and to North Carolina. Direct evidence of the capability of
Florida Current and Gulf Stream meanders to transport plankton
from the west coast of Florida to North Carolina comes from
extensive observations of a recent red tide (Gymnodinium breve)
event (Tester et al. 1991 ). A red tide bloom was observed off of
the southwest coast of Florida in September 1987. Red tide cells
from this bloom were apparently transported by the Florida Cur-
rent-Gulf Stream system to North Carolina, where a large bloom
occurred in October 1987. The occurrence of the bloom in North
Carolina coastal waters was associated with a large Gulf Stream
meander that persisted and apparently concentrated G. breve cells
over this region for a period of almost 3 weeks. Subsequent ob-
servations have revealed that the transport of red tide cells from
western Florida to eastern Florida and North Carolina may occur
more often than previously recognized and suggest that the occa-
sional transport of calico scallop larvae by this system may be
quite plausible (Murphy et al. 1975, Tester et al. 1991). In fact, on
the basis of a range of flow rates from 40 to 100 cm s - ' in the Gulf
Stream (Brooks and Bane 1983), it is possible for larvae to travel
the 500 km from Cape Canaveral to Cape Lookout in approxi-
mately 8 to 23 days. Costello et al. ( 1973) reported 14 to 16 days
as the planktonic larval period of calico scallops in the laboratory,
so it is not unreasonable to hypothesize that competent larvae
spawned off Cape Canaveral could successfully settle off of the
North Carolina coast. Similarly, estimated flow in the Florida
Current averages approximately 60 cm s~ ' (Brown et al. 1989).
Although the time required for larvae to enter the Loop Current-
Florida Current system from the shelf may vary considerably be-
cause of midshelf currents and the presence of local gyres (Lee et
al. 1992), larvae entrained in this system would be able to traverse
the distance between the Marquesas Keys to Cape Canaveral in
approximately 5 to 6 days.
One of the assumptions of using allozyme frequencies for es-
timates of gene flow between populations, as presented in this
study, is that they are neutral with respect to selection, reflecting
only the stochastic forces of drift and migration (mutation is as-
sumed to be negligible). A number of recent studies of geographic
variation in marine organisms that used either restriction fragment
length polymorphism analysis of mitochondrial DNA (mtDNA) or
single-copy nuclear DNA (scnDNA) call into question the gener-
ality of this assumption, and hence, the utility of allozyme data for
such estimates (e.g. Karland Avise 1992). Allozyme surveys of
geographic variation in American oysters (C. virginica) revealed
relatively little population subdivision and high rates of gene flow
throughout the range of the species along the Gulf of Mexico and
Atlantic Coasts (Buroker 1983). In contrast, surveys of both
mtDNA and scnDNA demonstrated sharp discontinuities between
Gulf and Atlantic oyster populations, with the genetic "break"
occurring along the eastern coast of Florida (Rceb and Avise 1990,
Karl and Avise 1992). The pattern of differentiation determined
from mtDNA and scnDNA was ascribed to vicariant biogeo-
graphic processes that initiated the current population structure and
is maintained by the Florida Peninsula serving as an effective
barrier to gene flow between populations in the Gulf of Mexico
and those along the Atlantic Coast (Reeb and Avise 1990. Karl and
Avise 1992), whereas Karl and Avise ( 1992) attributed the relative
genetic homogeneity described by allozyme surveys to balancing
selection that maintains allele frequencies despite constraints on
gene flow. Obviously, the results of this study of A. gibbus pop-
ulations should be interpreted with caution until supported by anal-
yses with other types of genetic markers. On the other hand, the
oceanic distribution of this species, as well as the intimate inter-
action of oceanographic processes with calico scallop life-history
attributes, suggests that the Florida Peninsula may serve only as a
partial barrier to gene flow for A. gibbus and that the relative
overall uniformity of allozyme frequencies and morphology may
indeed reflect high rates of migration between populations com-
pared with more estuarine species.
With these caveats in mind, the apparent genetic and morpho-
logical similarity between populations of North Carolina and Flor-
ida A. gibbus. particularly those on the east coast of Florida,
implies that migrants from Florida populations may be important
for sustaining or reestablishing North Carolina stocks. Proper
management of the North Carolina fishery may require consider-
ation of this process. Although the data presented here can only
support the hypothesis that migration occurs at a sufficient level to
maintain relative genetic and morphological homogeneity on an
evolutionary time scale, the actual frequency and intensity of
present-day migration from Florida to North Carolina remain to be
quantified. Additional work combining recruitment studies with
more sensitive genetic techniques will be needed to define more
clearly the forces that maintain the geographically isolated North
Carolina calico scallops and the extent of the contribution (via
larval dispersal) of Florida scallops.
ACKNOWLEDGMENTS
Roger Jones facilitated the collection of North Carolina scal-
lops and K. A. Sanddy helped with their dissection. We also thank
Richard Darden for the collection of Florida scallops, Steve Fisk
for support with morphological analyses, and Dan Marelli for help
536
Krause et al.
with statistical and morphological analyses. Comments from Dan
Marelli, Terry Bert, Eric Holm, Jim Quinn. and Tom Perkins
greatly improved the manuscript. Funds for the N.C. portion of
this research were provided to W.A. by the University of North
Carolina Marine Sciences Coordinating Committee for Marine
Programs Support for M.K.K. was provided by an NSF disser-
tation improvement grant #BSR-90 15991 and an NSF grant
#BSR-8918027 to R. K. Koehn. This is contribution No. 928
from the Department of Ecology and Evolution, SUNY Stony
Brook.
LITERATURE CITED
Alldredge, A. L. & W. M. Hamner. 1980. Recurring aggregation of zoo-
plankton by a tidal current. Est. Coast. Mar. Sci. 10:31-37.
Allen, D. M. 1979. Biological aspects of the calico scallop. Argopecten
gibbus, determined by spat monitoring. The Nautilus 94:107-1 19.
Allen, D. M. & T. J. Costello. 1972. The Calico Scallop, Argopecten
gibbus. NOAA Technical Report NMFS SSRF-656. 19 pp.
Anderson, W. D. & W. Lacey. 1979. South Carolina's first calico scallop
fishery. Proc. Natl. Shellfish Assoc. 69:201.
Avise, J. C C. A. Reeb & N. C. Saunders. 1987. Geographic population
structure and species differences in mitochondrial DNA of mouth-
brooding marine catfishes (Ariidae) and demersal spawning toadfishes
(Batrachoididae). Evolution 41:991-1002.
Beaumont, A. R. & E. Zouros. 1991. Genetics of scallops, pp. 585-623.
In: S. E. Shumway (ed.). Scallops: Biology, Ecology, and Aquacul-
ture. Elsevier, New York.
Blake, N. J. & M. A. Moyer. 1991. The calico scallop, Argopecten gib-
bus, fishery of Cape Canaveral, Florida, pp. 899-91 1 . In: S. E. Shum-
way (ed.). Scallops: Biology, Ecology, and Aquaculture. Elsevier.
New York.
Blanton, J. O., L. P. Atkinson, L. J. Pietrafesa & T. N. Lee. 1981 The
intrusion of Gulf Stream water across the continental shelf due to
topographically-induced upwelling. Deep Sea Res. Pt. A 28:393^105.
Boudra. D B , R Bleck & F Schott. 1987. Study of transport fluctua-
tions and meandering of the Florida Current using an isopycnic coor-
dinate numerical model, pp. 149-168. In: J. C. J. Nihoul and B. M.
Jamart (eds.l. Three-Dimensional Models of Marine and Estuarine
Dynamics. Elsevier, New York.
Brooks, D. A. & J. M. Bane. Jr. 1983. Gulf Stream meanders off North
Carolina during winter and summer 1979. J. Geophys. Res. 88:4633-
4650.
Brown, J., A. Collingh, D. Park. J. Phillips. D Rothery & J. Wright.
1989. Ocean Circulation. The Open University and Pergamon Press,
Oxford. 238 pp.
Bullis. H. R. & R. Cummins. 1961 . An interim report of the Cape Canav-
eral scallop bed. Commer. Fish. Rev. 23:1-8.
Burnaby, P. P. 1966. Growth-invariant disciminant functions and gener-
alized distances. Biometrics 22:96-110.
Buroker. N. E. 1983. Population genetics of the American oyster Cras-
sostrea virginica along the Atlantic coast and the Gulf of Mexico. Mar.
Biol. 75:99-112
Costello. T. J., J. H. Hudson, J. L. Dupuy & S. Rivkin. 1973. Larval
culture of the calico scallop, Argopecten gibbus. Proc. Natl. Shellfish.
Assoc. 63:72-76.
Cummins. R. 1971. Calico scallops of the southeastern United States,
1959-1969. NOAA Technical Report NMFS SSRF-627. 22 pp.
Dando, T. R., K. B. Storey, P. W. Hochachka & J M. Storey. 1981.
Multiple dehydrogenase in marine molluscs; electrophoretic analysis of
alanopine dehydrogenase, strombine dehydrogenase, octopine dehy-
drogenase and lactate dehydrogenase. Mar. Biol. Lett. 2:249-257.
Harris, H. & D A. Hopkinson. 1976. Handbook of Enzyme Electropho-
resis in Human Genetics. North Holland, Amsterdam. 259 pp.
Hard, D. L. & A G. Clark. 1989. Principles of Population Genetics. 2nd
ed. Sinauer Associates, Inc., Sunderland. MA, 682 pp.
Heald, D. I. & N. Caputi. 1981. Some aspects of growth, recruitment and
reproduction in the southern saucer scallop, Amusium halloti (Ber-
nards 1861 ) in Shark Bay, Western Australia. Dept. Fish. Wildl. (West
Australia) Fish Res. Bull. 25:1-33.
Hofmann. E. E., L. J. Pietrafesa & L. P.Atkinson. 1981 . A bottom water
intrusion in Onslow Bay. North Carolina. Deep Sea Res. 28A:329-
345.
Holm. S. 1979. A simple sequentially rejective multiple test procedure.
Scand. J. Stat. 6:65-70.
Johnson, A. G. & F. M. Utter. 1973. Population differences of aspartate
amino-transferase and peptidase in the bay mussel. Mytilus edulis.
Anim. Blood Grps. Biochem. Genet. 6:71-80.
Karl, S. A. & J. C. Avise. 1992. Balancing selection at allozyme loci in
oysters: implications from nuclear RFLPs. Science 256:100-102.
Kijima. A., K. Mori & Y. Fujio. 1984. Population differences in gene
frequency of the Japanese scallop Patinopecten yessoensis on the
Lkhotsk Sea Coast of Hokkaido Bull. Japan. Soc. Sci. Fish. 50:241-
248.
Kirby-Smith, W. W. 1970. Growth of the scallops, Argopecten irradians
concentricus (Say) and Argopecten gibbus (Linne), as influenced by
food and temperature. Ph.D. Dissertation. Duke University. Durham.
N.C. 127 pp.
Koehn, R. K ., J G Hall, D J Innes & A. J. Zera. 1984. Genetic dif-
ferentiation of Mytilus edulis in eastern North America. Mar. Biol.
79:117-126.
Koehn, R. K., R. Milkman & J. B. Mitton. 1976. Population genetics of
marine pelecypods. IV. Selection, migration and genetic differentia-
tion in the blue mussel Mytilus edulis. Evolution 30:2-32.
Lee, T. N. & L. P. Atkinson. 1983. Low-frequency current and temper-
ature variability from Gulf Stream frontal eddies and atmospheric forc-
ing along the southeast U.S. outer continental shelf. J . Geophys. Res.
88:4541-4567.
Lee, T. N., L. P. Atkinson & R. Legeckis. 1981 . Detailed observations of
a Florida current spinoff eddy on the Georgia continental shelf. Deep
Sea Res. 28:347-378.
Lee, T. N., V. Kourafalou. J. D. Wang. W. J. Ho. J. O. Blanton, L. P.
Atkinson & L. J. Pietrafesa. 1985. Shelf circulation from Cape Canav-
eral to Cape Fear during winter, pp. 33-62. In: L. P. Atkinson. D. W.
Menzel, and K. A. Bush (eds.). Oceanography of the Southeastern
U.S. Continental Shelf. American Geophysical Union, Washington,
DC.
Lee, T. N., C. Rooth, E. Williams, M. McGowan, A. F. Szmant& M. E.
Clarke. 1992. Influence of Florida current, gyres and wind-driven cir-
culation on transport of larvae and recruitment in the Florida Keys coral
reefs. Com. Shelf Res. 12:971-1002.
Leming, T. D. 1979. Observations of temperature, current, and wind vari-
ations off the central eastern coast of Florida during 1970 and 1971.
NOAA Technical Mem. NMFS-SEFC-6. 172 pp.
Miller, G. C, D. M. Allen & T. J. Costello. 1981. Spawning of the
calico scallop Argopecten gibbus in relation to season and temperature.
J. Shellfish Res. 1:17-21.
Moyer, M. A. & N. J. Blake. 1986. Fluctuations in calico scallop pro-
duction (Argopecten gibbus). Proceedings of the Eleventh Annual
Tropical and Subtropical Fisheries Conference, pp. 45-58.
Murphy. E. B , K. A. Steidinger, B. S. Roberts, J. Williams & J, W.
Jolley. Jr. 1975. An explanation of the Florida east coast Gymnodinium
breve red tide of November. 1972. Limnol. Oceanogr. 20:481^196.
Nei, M. 1978. Estimation of average heterozygosity and genetic distance
from a small number of individuals. Genetics 89:583-590.
Orensanz. J. M., A. M. Parma & O. O. Inbarne. 1991b. Population dy-
namics and management of natural stocks, pp. 625-713. In: S. E.
Shumway (ed.). Scallops: Biology, Ecology, and Aquaculture.
Elsevier, New York.
Calico Scallop Morphology and Genetics
537
Orensanz, J. M . M. Pascual & M. Fernandez. 1991a. Fisheries and aqua-
culture: Argentina, pp. 981-999. In: S. E. Shumway led. ). Scallops:
Biology. Ecology, and Aquaculture. Elsevier, New York.
Reeb, C. A. & J. C. Avise. 1990. A genetic discontinuity in a continu-
ously distributed species: mitochondrial DNA in the American oyster.
Crassoslrea virginica. Genetics 124:397—406.
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-
225.
Rodhouse. P. G. & P. M. Gaffney. 1984. Effect of heterozygosity on
metabolism during starvation in the American oyster Crassoslrea vir-
ginica. Mar. Biol. 80:179-187.
Roe. R. B.. R. Cummins. Jr. & H R. Bullis, Jr. 1971 Calico scallop
distribution, abundance, and yield off eastern Florida. 1967-1968.
NMFS Fish. Bull. 69:399-409.
Rogers, J. S. 1972. Measures of genetic similarity and genetic distance,
pp. 145-153. In: Studies in Genetics VII. Univ. Texas. Publ. 7213,
Austin. TX.
Rohlf. F. J. & F. L. Bookstein. 1987. A comment on shearing as a method
for "size correction." Syst. Zool. 36:356-367.
SAS Institute. Inc. 1982. SAS Users Guide: Statistics. SAS Institute.
Cary, N.C. 584 pp.
Saunders. N. C. L. G. Kessler & J. C Avise. 1986. Genetic variation
and geographic differentiation in mitochondrial DNA of the horseshoe
crab . Limulus polyphemus . Genetics 112:613-627.
Schaal. B. A. & W. W. Anderson. 1974. An outline of techniques for
starch gel electrophoresis of enzymes from the American oyster, Cras-
soslrea virginica (Gmelin). Technical Report Series No. 74-3. Georgia
Marine Sciences Center, Skidaway Island. GA. 17 pp.
Scopes. R. K. 1968. Methods for starch gel electrophoresis of sarcoplas-
mic proteins. Biochem. J. 107:139-150.
Selander. R. K. 1970. Behaviour and genetic variation in natural popula-
tions. Am. Zool. 10:53-66.
Shaw. C. R. & R. Prasad. 1970. Behaviour and genetic variation in nat-
ural populations. Am. Zool. 10:53-66.
Singhas, L. S. 1992. Reproductive periodicity and gonadal development
of the calico scallop Argopecten gibbus in North Carolina. Masters
Thesis, East Carolina University. Greenville. N.C, 104 pp.
Slatkin. M. 1985a. Gene flow in natural populations. Annu. Rev. Ecol.
Syst. 16:393-430.
Slatkin. M. 1985b Rare alleles as indicators of gene flow. Evolution
39:53-65.
Slatkin, M. 1987. Gene flow and the geographic structure of natural pop-
ulations. Science 236:787-792.
Sokal, R. R. & F. J. Rohlf. 1969. Biometry. 1st ed. W. H. Freeman and
Company, San Francisco, 776 pp.
Sokal, R R. & F. J. Rohlf. 1981. Biometry. 2nd ed. W. H. Freeman and
Company, San Francisco, 859 pp.
Tester, P. A.. R. P. Staumpf. F. M. Vukovich. P. K. Fowler & J. T.
Turner. An expatriate red tide bloom: transport, distribution, and per-
sistence. Limnol. Oceanogr. 36:1053-1061.
Waller. R. R. 1969 The evolution of the Argopecten gibbus stock (Mol-
lusca: Bivalvial, with emphasis on the Tertiary and Quaternary species
of eastern North America. J. Paleontol. 43:1-125.
Wolanski. E. & W. M. Hamner. 1988. Topographically controlled fronts
in the ocean and their biological influence. Science 241:177-181.
Wright, S. 1978. Evolution and the Genetics of Populations, vol. 4. Vari-
ability Within and Among Natural Populations. University of Chicago
Press, Chicago. 580 pp.
Yamamoto, G. 1964. (Studies on the propagation of the scallop, Pati-
nopecten yessoensis (Jay), in Mutsu Bay). Japan Mar. Res. Prol.
Assoc, Booklet 6, 77 pp. In Japanese. Translation No. 1054 of the
Fish. Res. Bd. of Canada.
Journal of Shellfish Research, Vol. 13, No. 2. 539-546, 1994.
GROWTH AND MORTALITY OF THE TROPICAL SCALLOPS: ANNACHLAMYS FLABELLATA
(BERNARDI), COMPTOPALLWM RADULA (LINNE) AND MIMACHLAMYS GLORIOSA (REEVE)
IN SOUTHWEST LAGOON OF NEW CALEDONIA
Y. LEFORT
Lycee Francois de Koweit
P.O. Box 9450
22 095 Salmiya, Kuwait
ABSTRACT Major growth relationships were computed for Annachlamys flabellata. Comptopallium radula. and Mimachlamys
gloriosa. The size-weight relationship for M. gloriosa is isometric; for other species studied, the relationship is decreasingly allo-
metric. The edible part (muscle + gonad) represents about 50% of the fresh weight for the three species studied. The estimated
parameters of Von Bertalanffy's equation on N capture-recapture measurements were estimated. Growth rate, during the year, is
greatest when the temperature is increasing; it is reduced at maximum temperatures and remains stable during the cool season. Pectinids
can be separated into two groups according to growth rate. Natural mortality was estimated for two species (C. radula and M. gloriosa)
of pectimdae in Southwest Lagoon.
KEY WORDS: Growth, mortality, scallops. New Caledonia, lagoon
INTRODUCTION
MATERIALS AND METHODS
Observation of the growth of bivalves is usually carried out by
recording shell measurements: the criterion most widely used is
the greater dimension ( = height, dorsoventral distance), because it
has the advantage of being easy to record without damaging the
specimens subjected to examination. Other allometric character-
istics can then be derived from such measurements.
A mathematical description of such growth revolves around
two major concerns: on the one hand, one wants to find a rela-
tionship that best describes the observed growth, and on the other
hand, one wants to define as precisely as possible the parameters
of a growth function, which then can be included into a population
dynamics model. Hoepe (1959) mentioned about 100 functions
that had been used to describe growth. Von Bertalanffy's (1938)
model is still the most widely used when describing the growth
of bivalves in general and that of Pectinidae in particular (Thei-
sen 1973, Broom 1976, Ralph and Maxwell 1977. Taylor and
Venn 1978, Heald and Caputi 1981, Williams and Dredge 1981.
MacDonald and Bourne 1987). This model is convenient in two
ways, because it is both heuristic and simple: knowing the K
(growth coefficient) and H, (maximum height estimated) param-
eters is enough to thoroughly reconstruct the geometry of a growth
curve.
On the soft bottoms of New Caledonia lagoons, the molluscs
form a large and very diverse group and 30 Pectinid species are
recognized so far (Dijkstra et al. 1989). The more common species
are Mimachlamys gloriosa (Reeve 1853) (=M. subgloriosa. Ire-
dale 1939), Bractechlamys vexillum (Reeve 1853), Comptopal-
lium radula (L. 1758). and Juxtamusium coudeini (Bavay 1803).
Only the first three species are large enough for human consump-
tion. Most of these filter feeders live on the substrate, covered by
a thin sediment layer. Only M. gloriosa is fixed to various sub-
strates by a byssus. In this article, allometric relationships repre-
senting the relative growth were defined for three Pectinidae in the
Southwest Lagoon of New Caledonia (Annachlamys flabellata
[ = A. leopardus. Iredale 1939], C. radula, and M. gloriosa). To-
tal growth and natural mortality were estimated for the two last
species.
Relative Growth
Samples of scallops were collected monthly during 1 year by
SCUBA divers in the vicinity of Noumea, New Caledonia
(22°15'S, 166°25'E) (Fig. 1). Fifty specimens per species, per
sample, were examined. Shells were weighed, measured and dis-
sected as quickly as possible (at the latest three hours following
collection). Shells were cleaned, and their epibionts (numerous on
M. gloriosa) were removed with a wire brush. The height of shells
was measured with a slide caliper (±0. 1 mm), from the umbo to
the ventral margin. Drained shells were weighed on a precision
balance (±0.01 g). The viscera, mantle, hepatopancreas. muscle,
and gonad were dissected out. Gonad, muscle, and viscera were
drained on filter paper for a constant time before being weighed.
The dry weight was measured after oven drying at 60°C for 48
hours, and ash-free dry weights (AFDW) were obtained after 3
hours of combustion at 550°C. The amount of organic matter
(OM) for any one organ was obtained by dry weight minus ash
weight.
Relative growth can then be defined as the relationship between
the measurements of two organs. The basic law of growth was
expressed in the following way: y = a • xb where Y represents the
dimension of the organ examined and X is that of the reference
organ; a and b are constants.
This relationship is usually turned into a linear function by
means of a logarithmic relation
Ln(y) = b • Ln(x) + Ln(a) or Y = b • X + A.
The slope (b) of such a line corresponds to the rate of relative
growth of examined organ as opposed to that of the reference
organ. According to the values of b, three types of growth were
defined: isometric (b = 1 or 3), decreasingly allometric (b < 1 or
3), and increasingly allometric (b > 1 or 3).
Total Growth
Two experimental sites were built in order to ( 1 ) follow the
growth of Pectinidae in a natural environment and (2) estimate
their natural mortality rate. These sites were set up in biotopes that
539
540
Lefort
164° E
166°
Figure 1. Location of study site.
were characteristic for the various species studied. One site was
set up in "Rocher a la Voile" (near Noumea) for M. gloriosa and
the other was located in Dumbea Bay for C. radula (Fig. 1), at
depths of 12 and 6 m, respectively. A wire mesh size of 2 cm was
selected so that it would not foul too rapidly and yet Pectinidae
could not escape. Respective dimensions of the two experimental
areas were 5 x 5 m for the "Rocher a la Voile" and 10 x 10 m
for the one in Dumbea Bay. The sizes of these experimental areas
have been imposed by the deepness and its structure. The sites
were monitored weekly; the parks were cleared (of seaweed) every
2 weeks.
The natural population (all benthic species) was removed from
the site. A label was glued on the shells of the experimental spec-
imens with cyanoacrylate glue after cleaning and drying the shells.
Scallops were labeled, measured, and returned to their respective
sites: C. radula in Dumbea Bay and M. gloriosa in the "Rocher a
la Voile." Scallops were measured every 2 months, and dead
individuals were removed. Bimonthly measurements were chosen
Growth and Mortality of Tropical Scallops
541
as a compromise between frequent measurements, which could
stress the pectinids and retard their growth, and infrequent mea-
surements, which could reflect the effects of external events, such
as storms (cyclones) or diseases.
The mathematical expression used almost everywhere to de-
scribe the overall growth of molluscs is Von Bertalanffy's equa-
tion (1938), which is written out as followed: H, = H« - [Hx -
H0) ■ e-k,]orH, = H»- (1
-k(l -t<,K
where H, = height at
time t; FU = maximum asymptotic size; H0 = height at t„; k =
constant; t0 = theoretical time where height is equal to 0.
Values of Von Bertalanffy's equation parameters were deter-
mined by using the SAS/STAT software (Cary, NC) and the NLIN
(Nonlinear regression least squares iterative method) Newton pro-
cedure, which allows the model parameters to be estimated in an
iterative way using the least squares method.
Data on growth were obtained from "mark and recapture"
experiments. The initial equation had to be modified in order to
express the measurements as a function of At (time interval be-
tween two measurements: 2 months) rather than t.
At time t, one finds: H, = FIJI - e"k"_t"'). At recapture
time, the equation becomes:
H([H
=
H^
(1 - e"
k([-
to + aK
=
H, -
- H, • e
-k(l
-to)
• e
ka
i,
=
H« -
- [H« ■ e
-ka
■ e
-kit
-to)'
- H,
. + [H«
■ e~
-k(t-
■t«>i
= [H, • e^"-'01] • (1 - e~ka)
but we have:
H, • e-k 0.05); NS, not significant; r.
correlation coefficient.
542
1.1 [-OR I
TABLE 2.
Allometric relationships for C. radula.
Relation
N
r
a
b
*obs
TW = a > Hb
632
0.96
21.2
10 "5
2,896
NS
FW = a x Hb
632
0.91
1.92
10"5
3,072
—
MW = a > Hh
45
0.86
0.14
10 "5
3,395
—
FW = a x (TW)b
632
0.93
16 2
10";
1,045
NS
MW = a x (TW)b
45
0.89
3.6
10 ~-
1,128
—
Note: For abbreviations, see footnote to Table 1.
the total annual growth. Growth is important when temperature is
increasing, but it reduces when temperature reaches a maximum
and then remains stable during the cool season (Figs. 3 and 4).
Mortality
Labeled specimens were prevented from leaving the experi-
mental site by a wire mesh. After a first diver had searched, a
second would come and collect individuals overlooked by the first;
capture efficiency was estimated at 100%.
Two hundred fourteen M. gloriosa were labeled on 30 May
1989; nine of them had died on 2 June 1989. On 6 December 1989
(after 27 weeks of freedom), a count allowed us to recover 156 live
specimens that were numbered and 1 1 live individuals found with-
out labels but whose shells showed evidence of wire brushing.
For C. radula two experiments were carried out:
Length of trial (weeks) 16 17
Number of labeled scallops 86 128
Scallops dead in the first week 5 6
Live scallops recovered with label 54 91
Live scallops recovered without label 15 13
If we discount the individuals that died from handling and those
that lost their tags, a total of 205 (214 - 9) M. gloriosa were
present at the start of the experiment, thus giving a 167/208 =
81.46% survival rate over a 192 day period (38 scallops dead),
i.e., a weekly death rate of 6.76 x 10~3. Therefore, the natural
mortality rate for M. gloriosa was 0.35. Analogous calculations
applied to the first case ( 16 weeks) give a natural mortality rate of
0.48 for C. radula, and in the second case (17 weeks) give a
coefficient of 0.45; thus, the average for C. radula was equal to
0.47.
DISCUSSION
The main interest of Von Bertalanffy's parameter estimations
becomes obvious only if their use is other than descriptive, for
example, for productivity calculations or biomass evolution study
TABLE 3.
Allometric relationships for A. flabellata.
Relation
N
r
a
b
'obs
TW = a x Hb
213
0.97
67.3
10"5
2,708
NS
FW = a x Hb
213
0.94
8.51
irr5
2,916
—
MW = a x Hb
51
0.99
7.68
icr5
2.686
—
FW = a x (TW)b
213
0.95
23.2
10~2
1,064
NS
MW = a x (TW)b
51
0.99
5.40
10~2
1.176
—
Note: For abbreviations, see footnote to Table 1.
for fishery management. The total growth of M. gloriosa and C.
radula cannot be completely represented by Von Bertalanffy's
model, because it can only correctly express the growth of species
after sexual maturity has been reached. This is not peculiar to New
Caledonian Pectinidae. The growth rate of New Caledonia Pec-
tinidae is very quick: sexual maturity is achieved during the second
year for M. gloriosa and during the third year forC. radula (Lefort
and Clavier 1994).
The growth of Pectinidae. and more generally that of bivalves,
cannot be described by any single model. Theisen (1973) also
found that Von Bertalanffy's model was well adapted to shellfish,
such as Pecten maximus, as long as these were mature. Williams
and Dredge (1981) reached the same conclusion, i.e.. Von Ber-
talanffy's classic model can correctly describe the growth of Amu -
slum balloti, but only when the size is larger than 50 mm. Many
models of growth have been improved without success; however.
Von Bertalanffy's model was the most adjusted to the studied
pectinids (Lefort 1991). Even when an adequate model has been
chosen, parameters still need to be evaluated. For Von Berta-
lanffy's model, numerous methods have been worked out, the best
known of which are those of Walford (1946). Beverton (1954),
Gulland and Holt (1959), Fabens (1965), and Allen (1966). Cal-
culating growth parameters with these various methods can lead to
significantly different results. The choice of a model and the
choice of an estimation method for parameters are crucial (and
often interconnected).
Within the Pectinidae. there is a great deal of diversity in
growth rates (Table 4; Fig. 3) and longevity. Growth parameters
vary from one species to the next. They can also vary within the
same species, depending on the geographical site (Antoine et al.
1979; Liana 1988) and/or depending on depth (MacDonald and
Thompson 1988). However, an examination of published values
of k (Table 4) made it possible to compare species on the basis of
growth rate. Thus, the Pectinidae listed in Table 4 can be subdi-
vided arbitrarily into two groups (k > 1 and k < 1). Individuals
displaying a high rate of growth (k = 1 to 2.8) are: A. balloti, B.
vexillum, and M . gloriosa. The second group is composed of the
great majority of Pectinidae where k < 1 .
It is generally agreed that the greater the growth rate (k), the
lower the longevity rate. This was particularly true for species
such as A. balloti, where longevity is only 3 years (Heald and
Caputi 1981). On the contrary, P. maximus (Antoine 1979) and
Crassadoma gigantea (MacDonald et al. 1991) live for 15 years or
more; as for Placopecten magellanicus (which has the lowest
growth coefficients with k = 0.156), longevity is 20 years (Mac-
Donald and Thompson 1988). New Caledonian Pectinidae have
relatively high growth coefficients. B. vexillum (k = 1.41 to 1.85)
has a growth rate clearly superior to that of M. gloriosa (k =
1.015), but for the former, longevity is 3 to 4 years whereas it is
8 years for the latter C. radula has a lower growth coefficient (k
= 0.35), but this is offset by a high longevity rate compared with
that of the two other species. This is in agreement with the gen-
erally accepted theory of the relation between longevity and the
growth rate.
Weekly relative growth was calculated for M. gloriosa (Fig.
2). As was the case with A. balloti, no interruption of growth was
observed in that species (Heald and Caputi 1981). There are sev-
eral cases of substantial variations from month to month or from
year to year (Borkowski 1974, Kojima 1975, Lewis et al. 1979).
Williams and Dredge (1981) observed seasonal variations in
growth among A. balloti, especially in the adductor muscle, which
Growth and Mortality of Tropical Scallops
543
K
3-
2.5-
2-
1.5-
1 -
0.5 ..
10
•11
• 13
.7
• 3
17
,.18
12
•20
23
• 19
14
•21
*25
• .22
*24 .16
*20
.15
•28 .
26**29
27
30
60
90
120
150
180
H«
Figure 2. Relative rate of weekly growth for M. gloriosa.
I I I
0) 0)
1— (—
0)
01
1— 1—
0)
H—
01
h- 1—
0)
1-4—
0
M—
0
h-f-
o
t—r-
D
1— t—
D
1— 1—
o
o
» 00
CO
CO
CO
CO
CO
0)
0)
0)
0)
0)
01
0)
01 01
0)
01
01
01
01
01
0)
0)
01
01
0)
0)
>*
^
•x
^
X
V
^
X
•x
^
•x
^
X
N
N
CO
CO
0)
a
r-
CM
T-
CM
CO
m
v
in
CO
o
D
0
0
r-
T-
r-
D
n
n
0
n
o
n
X
>■.
N
•X
X
X.
>*
x
^
N
N
v
s
N
T-
01
CO
n
□
CO
CM
is.
T
r-
i\
v
CM
o
0
cm
CM
CM
CM
I-
i-
0
0
m
CM
CM
CM
MONTHS
Figure 3. Temperature during the study period.
(.10 I
6-
45
s s
ib
j j
A S ' O
1989
M A M J
1990
MONTHS
Figure 4. Comparison of growth parameters for Von Bertalanffy's
equation.
544
Lefort
TABLE 4.
Comparison of growth parameters for von Bertalanffy's equation.
Species
Locality
References
A. ballon
103.8
1.42
98.7
1.20
104.9
2.83
B. vexillum
49.4
1.41
47.4
4.86
Chlamys opercularis
73.5
0.75
74.5
0.48
60.3
1 10
Chlamys varia
78.0
0.41
52.0
0.46
52.3
0.44
C. radula
92.4
0.35
M. gloriosa
73.9
1.02
P. caurinus
106.0
0.44
106.4
0.34
Paiinopecten yessoensis
148.1
0.40
Peclen sulcicostalus
86.7
0.32
95.5
0.38
107.1
0.52
106.7
0.66
P. maximus
124.6
0.56
139.8
0.56
135.4
0.58
138.9
0.47
123.5
0.53
158.4
0.16
166.0
0.21
P. magellanicus
155.9
0.22
162.0
0.16
146.4
0.35
Shark Bay (Australia)
North lagoon (New Caledonia)
Queensland (Australia)
South-west lagoon (New Caledonia)
South-west lagoon (New Caledonia)
Aberdeen (Scotland)
Kilmore (Irish sea)
Jersey (English channel)
Brest Harbour (France)
Brest Harbour (spring cohort)
Brest Harbour (autumn cohort)
South-west lagoon (New Caledonia)
South-west lagoon (New Caledonia)
Washington coasts (U.S.A.)
Georgia Strait (Canada)
Saroma Lake (Japan)
Mossel Bay (Bonne esperance)
False Bay (Bonne esperance)
Armen (France)
Brest Harbour (France)
St. Brieuc Bay (France)
Seine Bay (France)
Seine Bay (France)
Seine Bay (France)
Dieppe (France)
Sunnyside (Newfoundland)
St. Andrews (New Brunswick)
New Jersey
Bay of Fundy
Georges Bank
Heald and Caputi 1981
Clavier personal communication
Williams and Dredge 1981
Luro 1985
Clavier personal communication
Antoine 1979
Antoine 1979
This study
This study
Antoine 1979
Antoine 1979
Antoine 1979
Antoine et al 1979
MacDonald and Thompson 1988
Antoine 1979
went through its maximum growth during the summer months,
whereas the gonad developed little. Seasonal variations in shell
growth for M. gloriosa are tied to the temperature of seawater, the
highest rate being observed during warmer months. However, B.
vexillum underwent much greater seasonal changes in growth,
which were correlated with temperature and with particulate or-
ganic carbon (POC) (Clavier et al. in preparation).
Seasonal variations of growth are frequent among Bivalvia (P.
maximus, Mason 1958; Mizuhopecten yessoensis, Golikov and
Scarlato 1970. Pickett and Franklin 1975; Argopecten irradians.
Broom 1976). Such variations were related to abiotic factors such
as depth (MacDonald and Thompson 1988), type of substrate
(Gruffydd 1974). temperature (Theisen 1973), and currents
(Kirby-Smith 1972, Kirby-Smith and Barber 1974). A maximum
growth rate is observed in M. gloriosa when seawater temperature
goes up (October to January); it then drops when temperature
reaches its maximum. In contrasts with A. irradians, maximum
growth (6 to 8 mm per month) occurs when temperature is at an
optimal level; growth at 10°C is three times as high as that
achieved at 1()°C (Castagna and Duggan 1971).
A potential index of growth 0(0 = loglcl(k • Wj (Pauly
1982) was calculated for the three New Caledonian Pectinidae.
According to Pauly. this index must be more or less constant for
species of similar taxonomy; it should therefore be possible to
subdivide the Pectinidae family according to this index. The high-
est value was obtained for M . gloriosa (O = 1 .64), the next was
forfl. vexillum (Jl = 1 .60), and the lowest value (fl = 1.57) was
found for C. radula. The potential index of growth for A. balloti
in the northern lagoon of New Caledonia was il = 1.89 (Clavier
unpublished data).
The estimates for natural mortality rate reported in this article
are lower than those usually reported. Our estimations of natural
mortality are approximately one and are valid only for the Noumea
area. M. gloriosa is a subtidal bivalve living on muddy or muddy-
sand bottoms, covered by sponges. Sponges could be a good pro-
tection against predators (Dijkstra et al. 1989). C. radula lives on
muddy substrata in shallow water (about 5 to 15 m) and particu-
larly likes the inner part of bays. C. radula lives under heads of
living coral or in their immediate proximity, lightly covered so that
only the ring of sensorial tentacles can be seen by the predators.
These species live inside a lagoon, i.e., a semiclosed environment,
and this may protect them from predation.
The estimation of the natural mortality of species raised for
commercial purposes is usually more difficult than that for non-
commercial species. Most methods used to do so require knowl-
edge of the extent of fishing (Beverton and Holt 1956) or the
annual yield of fisheries (Csirke and Caddy 1983). In many cases,
empirical formulae were used, notably that of Pauly (1982) and of
Rikhter and Efanov (1976). Natural mortality estimation of non-
commercial species was much easier, especially if regular records
of population density can be kept, and it was possible for the two
species in the Southwest Lagoon of New Caledonia. However, it
seemed interesting to calculate such mortality rate with some more
commonly used formulae with our data.
Growth and Mortality of Tropical Scallops
545
Species
Pauly
Kik liter and Kfanov
fi. vexillum
M. gloriosa
C. radula
M = 2.19
M = 1.31
M = 0.61
M = 1.0
M = 1.8
M = 3.5
Empirical patterns give variable results and differ from the
applied results. Some authors have already mentioned this prob-
lem and advised the use of these methods only for an approximate
estimate of natural death. The results achieved through these pat-
terns are sometimes too remote from the experimental results; they
have to be used only in a limited way with extreme caution. The
result obtained with Pauly's formula was the closest to that found
for C. radula (M = 0.47). In the case of B. vexillum, the formula
of Rikhter and Efanov corresponded best to the values found ex-
perimentally (M = 0.8; Clavier, personal communication). How-
ever, for M. gloriosa, none of the mathematical results obtained
could approach the value we had found. Natural mortality varies
with the number of predators and probably with the geographic
situation. The main predators we could identify were starfish,
gasteropoda (Muricidae species) fish (Lethrinus species), octo-
puses, and crustaceans. It is common knowledge that pectinids
clap their valves; doing so enables the bivalve to move around. M .
gloriosa and C. radula can make "leaps" when they are disturbed
by predators.
CONCLUSION
The growth rate of New Caledonia Pectinidae is very quick; the
stage of first sexual maturity is achieved during the second year for
M . gloriosa and during the third year for C. radula. The annual
increase is then very low and is more or less constant in 1 year's
time. Their growth is mainly achieved during the first 2 years. The
growth rate is higher when the seawater is temperature increases,
it is lower when the seawater temperature is at its maximum, and
it is steady during the cool season.
Our estimates of natural mortality rate reported in this article
are lower than those usually reported (valid only for Noumea
area). This can be explained by the ecology of the species studied
and because they live in a semiclosed environment, which pro-
vides them with effective protection against some predators and
the external environment.
LITERATURE CITED
Allen. K. R. 1966. A method of fitting growth curves of the von Berta-
lanffy type to observed data. J. Fish. Res. Board Canada 23: 163-179.
Antoine, L. 1979. La Croissance de la Coquille Saint-Jacques (Pecten
maximus) et ses Variations en Mer Celtique et en Manche. These 3emt
cycle; University of Bretagne Occidentale, 148 pp.
Antoine. L., P. Arzel, A. Laurec & E. Morize. 1979. La croissance de la
coquille Saint-Jacques (Pecten maximus L.), dansles divers gissements
francais. Rapport p-v. Reunion du Conseil International pour l'Ex-
ploration de la Mer, 175:85-90.
Beverton. R. J. H. 1954. Notes on the use of theorical models in the study
of the dynamics of exploited fish population. U.S. Fisheries Labora-
tory Beaufort. N.C. Misc. Contrib.. 2.
Beverton, R. J. H. & S. J. Holt. 1956. A review of methods for estimat-
ing mortality rates in exploited fish populations, with special reference
to sources of bias in catch sampling. Rapport P. V. Reunion du Conseil
Permanent International pour ['Exploration de la Mer, 140:67-83.
Borkowski, T. V. 1974. Growth, mortality and productivity of south flo-
ridian Littorinidae (Gastropoda: Prosobranchia). Bull. Mar. Sci. 24:
409-438.
Broom, M. J. 1976. Synopsis of biological data on scallops Chlamys
(Aequipecten) opercularis (Linnaeus), Argopecten irradians (La-
marck) and Argopecten gibbus (Linnaeus). FAO Fish. Synop. No. 114.
Castagna, M. & W. P. Duggan. 1971. Rearing of the bay scallop Ae-
quipecien irradians. Proc. Nail. Shellfish. Assoc. 61:80-85.
Clavier, J. 1987. Contribution a la mise en valeur des populations d'or-
meaux [Haliotis tuberculata L.). These de Doctorat.. Univ. Paris VI.
Oceanologie biologique, 319 pp.
Clavier, J., Y. Lefort & P. Chardy. (in preparation) Growth, mortality and
the role of environmental factors in seasonal growth in the tropical
scallop Bractechlamys vexillum (Mollusca: Pectinidae).
Conan. G. & M. S. Shafee. 1978. Growth and biannual recruitement of
the black scallop Chlamys varia (L. ) in Lanveoc area. Bay of Brest. J.
Exp. Mar. Biol. Ecol. 35:59-71
Csirke, J. & J. F. Caddy. 1983. Producing modeling using mortality es-
timates. Can. J. Fish. Aqua. Sci. 40:43-51.
Dijkstra, H. H., B Richer de Forges, J. Clavier & Y. Lefort. 1989.
Pectinidae found on the soft bottoms of the New Caledonian and Ches-
terfield lagoons. Part I. Rossiniana 45:21-24.
Fabens. A. J. 1965. Properties and fitting of von Bertalanffy growth
curve. Growth 29:265-289.
Golikov. A. N. & O. A. Scarlato. 1970. Abundance, dynamics and pro-
duction properties of edible bivalves Mizuhopecten yessoensis and
Spisula sachalinensis related to the problem of organisation of control-
lable submarine farms at the western shores of the Sea of Japan. Hel-
golander Wiss. Meeresunlers 20:498-513.
Gruffydd, L. I. D. 1974. The influence of certain environmental factors
on the maximum length of the scallop Pecten maximus. J. Exp. Mar.
Biol. Ecol. 35:300-302.
Gulland. J. A. & S J. Holt. 1959. Estimation of growth parameters for
data at unequal time intervals. J. Conseil 25:47—48.
Heald. D. I. & N. Caputi. 1981. Some aspects of growth, recruitement
and reproduction in the saucer scallop, Amusium balloli (Bernardi.
1861) in Shark bay, western Australia. Fish. Res. Bull. West. Aust.
25:1-33.
Hoepe, K. 1959. Das reationskinetische grundgesetz auf das wachstum
bezogen und andersen wachstum formulierungen gegenubergestellt.
Diss. Gie-en
Kirby-Smith. W. W. 1972. Growth of the bay scallop: the influence of
experimental water currents. J. Exp. Mar. Biol. Ecol. 8:7-18.
Kirby-Smith, W. W. & R. T. Barber. 1974. Suspension-feeding aquacul-
ture systems: effects of phytoplankton concentration and temperature
on growth of the bay scallop. Aquaculture 3:135-145.
Kojima, H. 1975. A study on the growth of the shell of Haliotis discus
discus (Gastropoda: Haliotidae)-l: age and growth. Aquaculture 23:
61-66.
Lefort, Y. 1991 . Etude des populations de pectinides du lagon sud-ouest de
Nouvelle-Caledonie. These Doct. Univ. Bordeaux I. 234 pp.
Lefort. Y. & J. Clavier. 1994. Reproduction of Annachlamys flabel lata
(Bernardi), Comptopallatm radula (L.) and Mimachlamys gloriosa
(Reeve) (Mollusca: Pectinidae) in the south-west lagoon of New Cale-
donia. Aqua. Living Res. 7:39^16.
Lewis, J. B., F. Axelsen, J. Goodbroy. C. Page, G. Chislett & M. Chaud-
houry. 1979. Latitudinal differences in growth rates of some intertidal
marine mollusks in the Carribean. Mar. Sci. Manuscript Rep. 12,
McGill Univ., Montreal: 88 pp.
Liana. M. E. G. 1988. Growth mortality and recruitement of the Asian
moon scallop (Amusium pleuronectes) in the Visayan sea, Philippines.
FAO Fish. Rep. No. 389:16-24.
Luro. C. 1985. Etude des communautes benthiques du lagon sud-ouest de
la Nouvelle-Caledonie et d'une population de pectinidae: Comptopal-
Hum vexillum. Rapp. de stage: Bevet de Techn. Sup. Mer. Noumea:
ORSTOM. 40 pp.
546
Lefort
MacDonald, B. A. & N. F. Bourne. 1987. Growth, reproductive output,
and energy partitioning in weathervane, Patinopeaen caurinus. from
British Colubmia. Can. J. Fish. Aqua. Sci. 44:152-160.
MacDonald. B. A. & R. J. Thompson. 1988. Intraspecific variation in
growth and reproduction in latitudinally differentiated populations of
the giant scallop Placopecien magellanicus (Gmelin). Biol. Bull. 175:
361-371.
MacDonald. B. A., R. J. Thompson & N. F. Bourne. 1991. Growth and
reproductive energetics of three scallop species from British Columbia
(Chlamys hastata, Chlamys rubida and Crassadoma giganlea). Can. J.
Fish. Aqua. Sci. 48:215-221.
Mason, M. 1958. The breeding of the scallop Pecten maximus L. in Manx
waters. J. Mar. Biol. Assoc. U.K. 36:13-26.
Pauly, Y. D. Une selection de methodes simples pour 1'estimation des
stocks de poissons tropicaux. F.A.O. Circ. Peches. 729:63 pp.
Pickett, G. D. & A. Franklin. 1975. The growth of the queen scallop
(Chlamys opercularis), in cages off Plymouth, southwest England.
ICES CM. 1975/K:20 pp.
Ralph, R. & J. G. H. Maxwell. 1977. Growth of two Antarctic lamelli-
branchs: Adamussium colbecki and Laternula elliptica. Mar. Biol. 42:
171-175.
Ricker, W. E. 1980. Calcul et interpretation des statistiques biologiques
des populations de poissons. Bull. I' Office Recherches Pecheries Can-
ada 409 pp.
Rikhter, V. A. & V. N. Efanov. 1976. On one of the approaches to
estimation of natural mortality of fish populations. I. C.N. A. F. Res.
Doc, 76/Vl/8:12 pp.
Taylor, A. C, & T. J. Venn. 1978 Growth of the queen scallop. Chlamys
opercularis from the Clyde Sea area. J. Mar. Biol. Assoc. U.K. 58:
687-700.
Theisen, B. F. 1973. The growth of Mytilus edulis (L.) from Disko and
Thule District, Greenland. Ophelia 12:59-77.
Von Bertalanffy, L. 1938. A quantitative theory of organic growth. Hum
Biol. 10:181-213.
Walford, L. A. 1946. A new graphic method of describing the growth of
animals. Biol. Bull. Mar. Biol. Lab. Woods Hole 90:141-147
Williams, M. J. & M. C. L. Dredge. 1981. Growth of the saucer scallop
Amusium japonicum ballon Habe in Central Eastern Queensland. Aust.
J. Mar. Freshwater Res. 32:657-666.
Journal of Shellfish Research, Vol. 13, No. 2. 547-553. 1994.
THE USE OF RANDOM AMPLIFIED POLYMORPHIC DNA MARKERS IN GENETIC STUDIES
OF THE SEA SCALLOP PLACOPECTEN MAGELLANICUS (GMELIN, 1791 )t
MOHSIN U. PATWARY,1 ELLEN L. KENCHINGTON,2
CAROLYN J. BIRD,3 AND ELEUTHERIOS ZOUROS4
l1094 Wellington St., #403
Halifax, Nova Scotia. Canada B3H 2Z9
2Molluscan Fisheries & Aquaculture Section
Biological Sciences Branch
Department of Fisheries & Oceans
PO Box 550
Halifax, Nova Scotia, Canada B3J 2S7
Institute for Marine Biosciences
National Research Council of Canada
1411 Oxford St.
Halifax. Nova Scotia, Canada B3H 3Z1
4Biology Department
Dalhousie University
Halifax, Nova Scotia, Canada B3H 4J1
ABSTRACT This study represents the first published application of the random amplified polymorphic DNA (RAPD) technique to
bivalve DNA. A total of 222, 10 base (10 mer) primers were screened against one DNA sample from the sea scallop. Placopecten
magellanicus , under predetermined optimal reaction conditions. One hundred thirty RAPD primers were found to be positive (59%).
When 40 of these positive primers were randomly selected and used to compare RAPD profiles among 24 individuals collected from
different scallop beds, at least 15 primers revealed clear polymorphisms. Inheritance of RAPD alleles was examined by the analysis
of banding patterns from a pair-mated family, in which almost all alleles segregated in a Mendelian fashion. Although none of the
RAPD markers was unique to a single population in our small samples, the frequencies of polymorphic bands at different loci varied
between populations. Thus, genetic similarity based on allele frequencies can be estimated and used as an additional tool for
understanding the genetic structure of sea scallop populations.
KEY WORDS: RAPD, polymorphism, inheritance, genetic segregation. Placopecten
INTRODUCTION 1991 ) in contrast to the shell shape analyses. However, for certain
enzyme systems, a pronounced heterozygote deficiency was ob-
The sea scallop (Placopecten magellanicus) supports a valu- serve(J suggesting that allozyme-specific selection may be acting
able commercial fishery on the coasts of Atlantic Canada and the Qn mis spedes (Gartner-Kepkay and Zouros 1985). With the avail-
northeastern United States. This species is contiguously distnb- abi,ity of new molecular bioi0gicai techniques, there has been an
uted within its range, forming spatially discrete aggregates or lncreasing emphasis on the use of DNA characteristics as markers,
"beds" (Bourne 1964, Sinclair et al. 1985). In Atlantic Canada, DNA variation may provide answers t0 the questions of genetic
there are a number of scallop beds (Fig. 1 ) that are commercially relatedness between scallop stocks, which are of direct concern to
exploited (Black et al. 1993). and these are each managed as resource management
separate "stocks" (Kenchington and Lundy 1992, Robert et al. Mos[ methods for tne study of DNA variation rely on the use of
1993). In order to better understand the relationsh.ps between nonspeciflc probes that reveal multiallelic profiles unique to each
these beds, morphological (Kenchington and Full 1994) and en- individuai ( minisatellite DNA fingerprints) or on the use of spe-
zyme electrophoretic data (Volckaert and Zouros 1989; Beaumont dfic probes mat reCQver locus.specific multiallelic variation (e.g.,
and Zouros 1991; Volckaert et al. 1991) have been collected and microsateilites). Minisatellites are not suitable for population stud-
analyzed. The morphological characters identified scallops on St. ies because they do nQt a,low quantificatlon of genetic differences
Pierre Bank and in the Bay of Fundy (Fig. 1) as being signifi- and measure of heterozygosities and genetic distances. Microsat-
cantly different from those on the Georges Bank and Sable Island e„ites me sujtab,e for those purposes but ^ difficult and eXpen-
beds in a suite of separate analyses using Fourier shape descriptors sjve tQ deve,op and app,y m large.scaie population studies,
of the scallop upper shell (Kenchington and Full 1994). However, A reiativeiy new technique, using markers referred to as ran-
morphological differences may be the result of genetic and envi- dom amplified polymorphic DNA. or RAPD, has been described
ronmental differences that cannot be studied separately. (We|sh and McClelland 1990, Williams et al. 1990, Caetano-
Enzyme electrophoresis of scallops from these same beds iden- Ano„6s e, a] ,991a) ,, Js a technique based on the polymerase
tified a high degree of genetic similarity (Beaumont and Zouros chain reaction (pcR. Sai|d et a, 19gg) m wnlch slng,e short
oligonucleotide primers of arbitrary sequence are used to amplify
anonymous regions of genomic DNA. The amplification products
+NRCC 38025. are then resolved by gel electrophoresis, and amplified polymor-
547
548
47 *•
Patwary et al.
45 -
43
41
NEW
BRUNSWICK
ISt. Pierre Bank
Sable ls!sui::::;|
Western- Bank .
Lurcher Shoal
Brawn's '"
^vi.Barik''.'
.Georges Bank '.
^00 &
65'
60
55'
Figure 1. Locations of commercial sea scallop (P. magellanicus) beds along the coast of Atlantic Canada.
phic products are used as genetic markers. Compared with other
DNA-based techniques, the detection and use of RAPD markers is
faster and less expensive. It requires a small quantity of template
DNA per reaction; many reactions can be done simultaneously in
commercially available thermocyclers. and the reaction products
can be resolved and documented easily. The technique has been
applied successfully in a variety of genetic, phylogenetic. and
population genetic studies (Caetano-Anolles et al. 1991b. Bassam
et al. 1992, Hadrys et al. 1992, Tingey and del Tufo 1993). It has
been used to study genetic segregation (Carlson et al. 1991, Du-
rand et al. 1993, Kazan et al. 1993. Yu and Pauls 1993), charac-
terize hybrids (Baird et al. 1992, Patwary and van der Meer 1994),
identify strains (Goodwin and Annis 1991. Hu and Quiros 1991,
Welsh et al. 1991a), assess genetic diversity in culture collections
(Akopyanz et al. 1992. Patwary et al. 1993), obtain information
on parentage (Welsh et al. 1991b, Hadrys et al. 1993, Scott and
Williams 1993), identify markers linked to a disease resistance
gene (Martin et al. 1991, Paran et al. 1991), and study speciation
(Arnold et al. 1991. Crawford et al. 1993, Crossland et al. 1993)
and pollination (Philbrick 1993), and in phylogenetic (Halward et
al. 1992, Strongman and MacKay 1993), genetic mapping
(Williams et al. 1990, Reiter et al. 1992), and DNA fingerprinting
(Welsh and McClelland 1990, Wilde et al. 1992, Eskew et al.
1993) studies.
This article represents the first application of the RAPD tech-
nique to bivalve DNA. We identify a number of 10 mcr RAPD
primers that show polymorphisms in the sea scallop, P. magel-
lanicus, and test their suitability for genetic studies of this species.
MATERIALS AND METHODS
The sea scallops [P. magellanicus (Gmelin)] used in population
comparisons were collected by the Department of Fisheries and
Oceans, Canada, from five commercial scallop beds (Fig. 1):
Browns Bank. Sable Island Bank, Western Bank (offshore).
Lurcher Shoal, and Digby Gut (Bay of Fundy. inshore). Adductor
muscles were dissected from the scallop at sea and immediately
frozen in liquid nitrogen. Some scallops were brought back alive
and maintained in seawater tanks until used for DNA extraction
from fresh muscle.
Mendelian inheritance of the RAPD bands was tested by com-
paring samples from a pair-mated family (Marine Gene Probe
Laboratory, Dalhousie University, Halifax). The parents and 12
randomly selected offspring were analyzed.
DNA was extracted from individual adductor muscle tissue.
Approximately 0.2 g of tissue was ground to a fine powder in the
presence of liquid nitrogen in 10 ml Falcon culture tubes (Elkay
Products Inc., Boston, MA) with a steel plunger as a pestle. The
frozen powder was immediately mixed with 750 p.1 of lysis buffer
containing 10 mM Tris (pH 8.2), 1 mM NaEDTA, and 400 mM
NaCl. To the lysate. solutions of sodium dodecyl sulfate (SDS)
and proteinase K were added to final concentrations of 0.8% and
100 |igmr', respectively. The samples were mixed thoroughly,
incubated at 37°C for an hour, and then digested overnight at
55°C. Protein present in the lysate was precipitated by adding and
vortexing 250 p.1 of saturated NaCl and was pelleted by centrifu-
gation at room temperature at 800 rpm for 20 minutes. The nucleic
RAPD Markers in Sea Scallop
549
kb
2.69-
1.49-
093-
0.42-
Cvj
m
o
O
o o
CD 00
O O O
O
TABLE 2.
A list of primers used in this study.
Figure 2. The effect of template DNA concentration on RAPD ampli-
fication of sea scallop DNA using a single primer (I'BC No. 287). The
RAPD products were resolved by electrophoresis through 1.4% aga-
rose gel and stained with ethidium bromide. The numerals above the
lanes indicate nanograms of template DNA included in respective 25
|i I reactions. The lane marked C is a negative control reaction in
which no template DNA was added. The lane labeled M contains DNA
size marker bacteriophage \ DNA digested by the restriction enzyme
Sfyl with the size of the X DNA fragments indicated as the number of
kilobase (kb) pairs.
TABLE I.
RAPD primer composition and the results of screening the primers
against the DNA from one individual sea scallop.
G + C
No. of Primers
Screened
No. of Positive
Primers
% Positive
50
54
60
61
70
58
80
39
90
1
40
25
27
28
1
80
41
47
71
100
kb
1.49-
M a b c d e
0.93-
0.42-
::.: IS::z .-.:: ;==
A B C D
Figure 3. The effect of nonionic detergents on RAPD amplification of
sea scallop DNA using a single primer (L'BC No. 208). The detergents
applied in subpanels are (A) Nonidet P-40, (B) Tween 20, (C) Triton
X-100, (D) 1:1 mixture of Nonidet P-40 and Tween 20, (E) 1:1 mixture
of Nonidet P-40 and Triton X-100, (F) 1:1 mixture of Tween 20 and
Triton X-100, (G) 1:1:1 mixture of all three detergents. The percent-
age of detergents used in reactions is: lane a, 0.0; b, 0.5; c, 1.0; d, 2.0:
e, 4.0; f, 6.0. The lane marked "g" is a negative control reaction.
Lane M contains size markers as in Fig. 2. kb, kilobase.
UBC Primer
Number
15
25
63
67
71
77
81
83
84
98
100
101
102
104
105
106
110
116
129
134
137
145
146
149
150
153
164
166
167
171
174
181
186
190
198
202
205
208
210
211
212
287
Nucleotide
Sequence
CCTGGGTTTG
ACAGGGCTCA
TTCCCCGCCC
GAGGGCGAGC
GAGGGCGAGG
GAGCACCAGG
GAGCACGGGG
GGGCTCGTGG
GGGCGCGAGT
ATCCTGCCAG
ATCGGGTCCG
GCGGCTGGAG
GGTGGGGACT
GGGCAATGAT
CTCGGGTGGG
CGTCTGCCCG
TAGCCCGCTT
TACGATGACG
GCGGTATAGT
AACACACGAG
GGTCTCTCCC
TGTCGGTTGC
ATGTGTTGCG
AGCAGCGTGG
GGAGGCTCTG
GAGTCACGAG
CCAAGATGCT
ACTGCTACAG
CCAATTCACG
TGACCCCTCC
AACGGGCAGC
ATGACGACGG
GTGCGTCGCT
AGAATCCGCC
GCAGGACTGC
GAGCACTTAC
CGGTTTGGAA
ACGGCCGACC
GCACCGAGAG
GAAGCGCGAT
GCTGCGTGAC
CGAACGGCGG
Study
segregation
segregation
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
segregation3
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
popu
ation
ation"
ation
ation,
ation"
ation
ation
ation
ation
ation
ation
ation
ationa
ation
ationa
ationa
ation
ation
segregation3
segregation
segregation"
ation, segregation
ation
ation"
ation. segregation"
ation"
ation
ation"
ation,
ation,
ation
ation
ation
ation
ation
ation
ation, segregation"
ation
ation
ation
ation, segregation"
ation
J Primers revealed clear polymorphisms between individuals.
acids present in the supernatant were precipitated for 30 minutes at
-20°C by the addition of an equal volume of isopropanol, col-
lected by centrifugation at 14,000 rpm for 10 minutes, washed in
1 ml of 70% ethanol. dried in a vacuum for 5 minutes, and dis-
solved in 500 u.1 of TE (10 mM Tris-HCl, 1 mM EDTA). The
RNA present in the nucleic acid preparation was digested by add-
ing 10 u.g of RNAse A and incubating at 37°C for 30 minutes. The
DNA solution was first extracted with an equal volume of phenol-
chloroform [lphenol:chloroform:isoamyl alcohol = 50:49:1).
0.1M Tris (pH 7.5), 0.2% (3-mercaptoethanol] and then with an
equal volume of chloroform-isoamyl alcohol (24:1). The DNA in
the aqueous phase was precipitated by adding one-third volume of
2.5 M NH4OAc and 2.5 vol of 95% ethanol. The DNA pellet
550
Patwary et al.
obtained after centrifugation was washed with 70% ethanol to
remove excess salts, dried, dissolved at 37°C in 200 to 300 u,L of
TE (10 mM Tris-HCl, 0.1 mM EDTA), and quantified with a
Beckman DU-64 spectrophotometer that calculates nucleic acids
concentrations based on Warburg and Christian coefficients (War-
burg and Christian 1942).
A number of PCR using varying quantities of template DNA
(2.5 to 200 ng). Taq DNA polymerase (0.5 to 2.5 U), MgCk (0.5
to 6.0 mM). and nonionic detergents (Nonidet P-40, Tween-20,
and Triton X-100 at concentrations of 0.5 to 6.0%) and several
temperature profiles were conducted to determine optimal reaction
conditions for our preparation of scallop DNA. The final prepa-
ration of RAPD reactions was the same as described by Patwary et
al. (1993), except that the Mg2 + concentration was increased to
2.0 mM and a 1:1 mixture of nonionic detergents (Tween-20 and
Triton X-100; Sigma Chemical Co., St. Louis, MO) was added to
a final concentration of 2%. The final primer concentration in the
reaction mix, in this protocol, is 0.2 jjlM (Patwary et al. 1993).
The detergents were mixed thoroughly in the reaction mix by
vortexing for 20 to 30 s before adding DNA and Taq polymerase,
to obtain optimal results.
A total of 222 primers were screened against one DNA sample
(from a single individual) using the optimal reaction conditions
described above. These RAPD primers were obtained from Dr.
John E. Carlson, Biotechnology Laboratory. University of British
Columbia UBC, Canada. The amplification was performed in a
Gene Amp PCR System 9600 (Perkin Elmer Cetus. Norwalk, CT)
thermocycler programmed for 40 cycles: four initial cycles of 3
minutes each at 94, 36, and 72°C; followed by 35 cycles of 30
seconds at 94°C. 1 minute at 36°C, and 2 minutes at 72°C; and a
final amplification of 1 cycle of 30 seconds at 94°C, 1 minute at
36°C, and 10 minutes at 72°C. One-third of each reaction product
was separated by electrophoresis at 3 to 4 Vcm~ ' in 1.4% aga-
rose/TBE (Tris borate EDTA) gels and stained by mixing ethidium
bromide with TBE electrophoresis buffer to a final concentration
of 0.2 |i,g/ml~ '. Assessment of the RAPD markers was subjective
in that information contained in positive primers, that is, ones
producing visible banding patterns, was reduced to those bands
that were highly visible and readily scoreable (e.g. , bands depicted
with arrows on Fig. 5).
RESULTS
Amplification products were obtained in all reactions contain-
ing 2.5 to 200 ng of template DNA per 25 jjlI of reaction, but the
number and amount of products were greater in reactions contain-
ing over 40 ng of DNA (Fig. 2). Concentrations of MgCL between
2.0 and 3.0 mM were found to be most satisfactory; those below
1.0 mM failed to amplify the DNA and concentrations over 3.0
mM produced one or more secondary bands. The addition of either
one or a mixture of two or more nonionic detergents (Nonidet
P-40. Tween-20, and Triton X-100) to the reaction mix signifi-
cantly improved the yield of RAPD products (Fig. 3). Amplifica-
tion occurred in all detergent concentrations (0.5 to 6.0%) exam-
ined. Amplification was not inhibited significantly by the addition
of detergents even at 6.0%, but the yield was generally better in
reactions with 1.0 or 2.0% detergents. Vortexing the detergents
after addition to the reaction mix was necessary to obtain satisfac-
tory results. The optimal Taq concentrations were found to be 0.75
or 1.0 unit per reaction. The yield of amplified products was
insufficient when less Taq was used, whereas more than 1 .0 U of
Taq per reaction caused the accumulation of nonspecific back-
ground products. The reproducibility of RAPD profiles was im-
proved when four initial cycles with longer denaturation. anneal-
ing, and extension periods were used before the regular cycles.
One hundred thirty RAPD primers were found to yield visible
amplification products. The G + C content of these primers is
given in Table 1 . Forty of these positive primers were arbitrarily
kb
2.69-
1.49-
0.93-
0.42
SIB
I WB |
BB I
DG1 I
DG2 I
LS
SIB |
WB |
BB I
DG1 I
DG2
I LS
;
m5
Wmm
^Mnref
kh SIB I WB I BB I DG1 I DG2 I LS
SIB | WB | BB 1 DG1 i DG2 I LS
2.69- ■
1.49-H
0.93- ■
0 42-BMUiL
_-r-~ :
1
i
5
Figure 4. Variable RAPD profiles in the sea scallop. Each lane represents a different animal. The beds from which animals were collected are
labelled as SIB (Sable Island Bank), WB (Western Bank), BB (Browns Bank), DG1 (Digby Gut, Bay of Fundy), DG2 (Digby Gut, Bay of Fundy),
and LS (Lurcher Shoal, Bay of Fundy). The reactions were performed with primer (A) UBC No. 67, (B) UBC No. 81, (C) UBC No. 171, and
(D) UBC No. 205. The left most lane of each set A-D contains size markers as in Fig. 2; kb, kilobase.
RAPD Markers in Sea Scallop
551
M f m progeny
M t m progeny M
093
M f m
M f m
Figure 5. Segregation of RAPD markers (polymorphic bands) gener-
ated by single primers (A) UBC No. 25, (B) UBC No. 77, (C) UBC No.
15, and (D) UBC No. 145. The lanes labelled "f" and "m" identify
DNA from the female and male parent, respectively. The remaining
lanes in each subpanel represent progeny obtained from the mating of
"f" and "m." The arrows indicate the positions of segregating loci.
Lane M contains size markers as in Fig. 2; kb. kilobase.
selected and used to compare RAPD profiles among 24 individuals
from different scallop beds (Table 2). At least 15 primers revealed
clear polymorphisms. An example of polymorphism revealed with
four of these primers is given in Figure 4. The reactions were
repeated, and the polymorphisms were found to be stable and
characteristic of each individual animal.
Mendelian inheritance of RAPD alleles was examined by the
analysis of banding patterns from a pair-mated family. Thirty of
the original positive primers, including the 15 that revealed poly-
morphisms above, were screened against the parents of this family
in order to identify polymorphisms between the parents that could
then be examined in the progeny. Twenty of the primers either
produced bands that did not differ between the two parents or were
too close in size to allow for reliable identification. Ten primers
were suitable for progeny analysis. These 10 primers revealed
polymorphisms in a total of 21 loci. Almost all alleles in these
loci segregated in a Mendelian fashion (Fig. 5; Table 3). Het-
erozygosity cannot be determined from the banding pattern itself,
because RAPD amplification produces one band when amplifying
from one or both chromosomes. However, analysis of the progeny
identified families in which one parent was homozygous negative
and the other was positive/negative heterozygous (e.g.. 15-1.2)
and families in which both parents were heterozygous (e.g., 77-
0.7).
DISCUSSION
One of the problems associated with the RAPD technique has
been the reproducibility of bands, particularly, the minor or faint
bands (Penner et al. 1993). We have been able to improve the
reproducibility of scallop RAPD profiles by optimizing the con-
centration of reaction ingredients and cycling parameters. The sat-
isfactory application of this technique needs considerable care in
preparing reactions to reduce the possibility of contamination from
extraneous DNA. The presence of impurities, such as polysaccha-
rides and phenolic compounds, also seriously impairs RAPD per-
formance. Our initial attempts to obtain reproducible RAPD pro-
files of scallop DNA were mostly unsuccessful. After optimizing
reaction conditions and adding nonionic detergents to the reaction
mix, the quality of RAPD profiles improved significantly. The
addition of detergents appears to enhance the specificity of primer
binding with template DNA and at least partially suppresses the
effect of impurities. Using these reaction conditions, we have also
obtained reproducible RAPD profiles in mussels and lobsters.
In order for the markers identified here to be useful in genetic
studies, it must be demonstrated that the bands are heritable and
segregate from generation to generation in a Mendelian fashion as
in other organisms (Carlson et al. 1991 . Echt et al. 1992, Kazan et
al. 1993). This has been demonstrated for 10 of the RAPD markers
(21 loci) used on sea scallop DNA in this study.
One disadvantage of RAPD polymorphisms compared with al-
lozymes. complementary DNA (cDNA) probes, and microsatel-
lites is that the heterozygote cannot be distinguished from the
positive homozygote. The PCR will produce the same product
from the DNA of an individual that has twice one site to which the
primer binds (a positive homozygote) and from an individual that
has the binding site once (a positive/negative heterozygote). The
negative homozygote is identifiable by the absence of the band. In
pair-mating crosses, this presents no major problem, because the
genotypes of the parents can be inferred from the genotype ratios
among the offspring. In population surveys, frequencies for the
TABLE 3.
Segregation of RAPD alleles in a pair-mating progeny and \2 values
for expected Mendelian ratios.
Markers in
RAPD
Allele
Phenot"
ipe
Offspring
Expected
Ratio"
Female
Male
Present
Absent
x2b
15-1.2
-
+
7
5
1:1
0.33
15-0.8
15-0.5
-
+
+
5
3
7
9
1:1
1.1
0.33
3.00
25-1.88
77-1.2
+
+
—
6
6
6
6
1:1
1:1
0.00
0.00
77-0.8
77-0.7
+
+
+
2
10
10
2
1:1
3:ld
5.33c
0.44
77-0.6
+
-
4
8
1:1
1.33
137-0.8
+
-
8
3
1:1
2.16
145-0.94
+
-
7
4
1:1
0.83
145-0.92
145-0.9
+
+
4
5
7
6
1:1
1:1
0.83
0.17
150-0.65
150-0.55
167-0.4
171-1.2
171-1.1
+
+
+
+
+
4
12
7
3
9
8
0
5
9
3
1:1
1:0'
1:1
1:1
1:1
1.33
0.00
0.33
3.00
3.00
171-0.9
205-0.6
205-0.5
+
+
+
6
8
12
6
4
0
1:1
1:1
1:0*
0.00
1.33
0.00
212-0.7
212-0.6
+
+
8
4
3
7
1:1
1:1
2.16
0.83
a Expected ratios for progeny on the basis of dominant marker phenotype.
b X2 at p = 0.05 and df (degrees of freedom) = 1 is 3.84.
c Significant deviation from the expected ratio.
d Both parents are assumed to be heterozygous at this locus.
e Female assumed to be either homozygous dominant or the allele is sex
linked.
552
Patwary et al.
positive and negative allele at a RAPD "locus" have to be esti-
mated by assuming Hardy-Weinberg equilibrium.
Repeated applications of the RAPD method to a number of
individuals from different populations showed that polymorphisms
are stable. Although we were not able to detect any RAPD markers
that were unique to a population, the results indicate that the
frequency of polymorphic bands at different loci can be estimated
and used as an additional tool to understand the genetic structure
of sea scallop populations. As in most other species, it appears that
no single set of markers will suffice for the genetic characteriza-
tion of stocks of scallops. Allozymes tend to inflate the genetic
similarity of populations, because stabilizing selection tends to
establish the same equilibrium frequencies in all populations (Karl
and Avise 1992). Multibanded minisatellite profiles are not ame-
nable to population genetic analysis because they distinguish vari-
ation at the level of the individual. Mitochondrial variation can be
very useful for population discrimination (Moritz et al. 1987), but
normally, mitochondrial DNA restriction fragment length poly-
morphism is not abundant and one has to combine PCR amplifi-
cation and sequencing of the amplified product. Microsatellite
assays are expensive to develop and produce polymorphisms that
are more useful for the characterization of individuals rather than
populations. cDNA probes (Pogson and Zouros 1994) are also
expensive to develop but, once developed, have the potential to be
very useful tools for population discrimination. RAPDs are much
less expensive, but each primer alone produces only a limited
amount of information, so that a multiple set of primers must be
used (Welsh and McLelland 1990). Experimentation with ampli-
fication using smaller primers (5 mers) may produce more vari-
ability, but at the cost of a more elaborate detection protocol
(Caetano-Anolles et al. 1991a). The final discrimination of scallop
populations may have to rely on a combination of such tools. The
most useful combination can be determined only through pilot
analyses of representative scallop populations.
ACKNOWLEDGMENTS
We thank Mr. J. Angel (Department of Fisheries & Oceans,
Halifax, Nova Scotia. Canada) and Dr. J. P. van der Meer (In-
stitute for Marine Biosciences, National Research Council of Can-
ada. Halifax, Nova Scotia) for their assistance in the development
of this project. We thank Mr. D. Cook (Marine Gene Probe Lab-
oratory, Department of Biology, Dalhousie University) for gener-
ously providing tissue samples of the pair-mated family. We are
also grateful to Dr. M. Ball for giving us a number of DNA
samples and to J. Williams and B. Gjetvaj for laboratory assis-
tance.
LITERATURE CITED
Akopyanz, N., N. O. Bukanov, T. U. Westblom, S. Kresovich & D. E.
Berg. 1992. DNA diversity among clinical isolates of Helicobacter
pylori detected by PCR-based RAPD fingerprinting. Nucleic Acids
Res. 20:5137-5142.
Arnold, M. L., C. M. Buckner & J. J. Robinson. 1991. Pollen-mediated
introgression and hybrid speciation in Louisiana irises. Proc. Natl.
Acad. Sci. USA 88:1398-1402.
Baird, E., S. Cooper-Bland, R. Waugh, M. DeMaine & W Powell. 1992.
Molecular characterization of inter- and intra-specific somatic hybrids
of potato using randomly amplified polymorphic DNA (RAPD) mark-
ers. Mol. Gen. 233:469-475.
Bassam, B. J., G Caetano-Anolles & P. M. Gresshoff. 1992. DNA am-
plification fingerprinting and its potential application for genome anal-
ysis, pp. 8-16. In: P. M. Gresshoff. (ed.). Current Topics in Plant
Molecular Biology, vol. 1. Chapman and Hall, New York.
Beaumont, A. R. & E. Zouros. 1991. Genetics of scallops, pp. 585-624.
In: S. E. Shumway. (ed.). Scallops: Biology, Ecology and Aquacul-
ture. Elsevier, New York.
Black, G A. P., R. K. Mohn, G Robert & M. J. Tremblay. 1993. Atlas
of the biology and distribution of the sea scallop Placopecten magel-
lanicus and Iceland scallop Chlamys islandica in the Northwest Atlan-
tic. Can. Tech. Rep. Fish. Aqua. Sci. No. 1915:1-34.
Bourne, N. 1964. Scallops and the offshore fishery of the Maritimes. Fish.
Res. Board Canada Bull. No. 145:1-60.
Caetano-Anolles, G., B. J. Bassam & P M. Gresshoff. 1991a. DNA
Amplification Fingerprinting using very short arbitrary oligonucleotide
primers. Bio/Technology 9:553-557.
Caetano-Anolle's, G., B. J. Bassam & P. M. Gresshoff. 1991b. DNA
Amplification Fingerprinting: A strategy for genome analysis. Plant
Mol. Biol. Rep. 9:294-307.
Carlson, J. E., L K. Tulsieram, J. C. Glaubitz. V. W K Luk, C.
Kauffelddt & R Rutledge. 1991. Segregation of random amplified
DNA markers in F, progeny of conifers. Theoret. Appl. Genet. 83:
194-200.
Crawford, D. J.,S. Brauner, M. B.Cosner&T. F. Stuessy. 1993. Use of
RAPD markers to document the origin of the intergeneric hybrid x
Margyracaena skottsbergii (Rosaceae) on the Juan Fernandez islands
Am. J. Botany 80:89-92.
Crossland, S.. D. Coates, J. Grahame & P. J. Mill. 1993. Use of random
amplified polymorphic DNAs (RAPDs) in separating two sibling spe-
cies of Littorina. Mar. Ecol. Prog. Ser. 96:301-305.
Durand, N., P. Reymond & M. Fevre. 1993. Randomly amplified poly-
morphic DNAs assess recombination following an induced parasexual
cycle in Penicillum roqueforti. Curr. Genet. 24:417—420.
Echt, C. S., L. A. Erdahl & T. J. McCoy. 1992. Genetic segregation of
random amplified polymorphic DNA in diploid cultivated alfalfa. Ge-
nome 35:84-87.
Eskew, D. L.. G. Caetano-Anolles, B. J. Bassam & P. M. Gresshoff.
1993. DNA amplification fingerprinting of the Azolla-Anabaena sym-
biosis. Plant Mol. Biol. 21:363-373.
Gartner-Kepkay, K. & E. Zouros. 1985. Influence of environment and
human selection of the genetic structure of some economically impor-
tant marine animal species. Government of Canada Department of
Supply and Services Contract Report No. 08sc. FP101-3-0301, 73 pp.
Goodwin, P. H., & S. L. Annis. 1991. Rapid identification of genetic
variation and pathotype of Leptosphaeria maculans by random ampli-
fied polymorphic DNA assay. Appl. Environ. Microbiol. 57:2482-
2486.
Hadrys, H., M. Balick & B. Schierwater. 1992. Applications of random
amplified polymorphic DNA (RAPD) in molecular ecology. Mol.
Ecol. 1:55-63.
Hadrys, H., B. Schierwater, S. L., Dellaporta, R. Desalle & L. W. Buss.
1993. Determination of paternity in dragonflies by random amplified
polymorphic DNA fingerprinting. Mol. Ecol. 2:79-87.
Halward, T , T. Stalker. E. LaRue & G. Kochert. 1992. Use of single-
primer DNA amplifications in genetic studies of peanut (Arachis hy-
pogaea L). Plant Mol. Biol. 18:315-325.
Hu, J. & C. F Quiros. 1991. Identification of broccoli and cauliflower
cultivars with RAPD markers. Plant Cell Rep. 10:505-511.
Karl, S. A. & J. C. Avise. 1992 Balancing selection at allozyme loci in
oysters: implications from nuclear RFLPs. Science 256:100-102.
Kazan, K, J. M. Manners & D. F. Cameron. 1993. Inheritance of ran-
dom amplified polymorphic DNA markers in an interspecific cross in
the genus Stylosanthes. Genome 36:50-56.
Kenchington, E. & W. E. Full. 1994. Fourier analysis of scallop shells
RAPD Markers in Sea Scallop
553
(Placopecten magellanicus) in determining population structure. Can.
J Fish- Aqua- Sci. (in press).
Kenchington. E. & M. J. Lundy. 1992. 1991 (Digby) Bay of Fundy scal-
lop stock assessment. Canadian Atlantic Fisheries Scientific Advisory
Committee Research Document 92/41, 27 pp.
Martin, G B.. J G. K. Williams & S. D. Tanksley. 1991. Rapid iden-
tification of markers linked to a Pseudomonas resistance gene in to-
mato by using random primers and near isogenic lines. Proc. Natl.
Acad. Sci. USA 88:2336-2340.
Moritz, C . T. E. Dowling & W. M. Brown. 1987. Evolution of animal
mitochondrial DNA: relevance for population biology and systematics.
Annu. Rev. Ecol. Syst. 18:269-292.
Paran, I., R. Kesseli & R. Michelmore. 1991. Identification of restriction
fragment length polymorphism and random amplified polymorphic
DNA markers linked to downy mildew resistance genes in lettuce,
using near-isogenic lines. Genome 34:1021-1027.
Patwary. M. U., & J P. van der Meer. 1994. Application of RAPD mark-
ers in an examination of heterosis in Gelidnim vagum (Rhodophyta). J.
Phycol. 30:91-97.
Patwary, M. U.. R. M. MacKay & J. P. van der Meer. 1993. Revealing
genetic markers in Geltdium vagum (Rhodophyta) through the random
amplified polymorphic DNA (RAPD) technique. J Phycol. 29:216-
222.
Penner, G. A.. A. Bush. R Wise. W. Kim, L. Domier, K. Kasha, A.
Laroche, G. Scoles. S. J. Molnar & G. Fedak. 1993. Reproducibility
of random amplified polymorphic DNA (RAPD) analysis among lab-
oratories. PCR Methods Appl. 2:341-345.
Philbnck, C. T. 1993. Underwater cross-pollination in Callitnche her-
maphroditica (Callitnchaceae): Evidence from random amplified poly-
morphic DNA markers. Am. J. Botany 80:391-394
Pogson, G. H. & E. Zouros. 1994. Allozyme and RFLP heterozygosities
as correlates of growth rate in the scallop Placopecten magellanicus: a
test of the associative overdominance hypothesis. Genetics (in press).
Reiter, R. S., J. G. K. Williams, K. A. Feldmann, J. A. Rafalski, S. V.
Tingey & P. A. Scolnik. 1992. Global and local genome mapping in
Arabidopsis thaliana by using recombinant inbred lines and random
amplified polymorphic DNAs. Proc. Natl. Acad. Sci. USA 89:1477-
1481.
Robert, G . G. A. P Black & M A. E. Butler. 1993. Georges Bank
scallops stock assessment- 1992. Department of Fisheries and Oceans.
Canada Atlantic Fisheries Research Document 93/15, 33 pp.
Saiki, R. K.. D H. Gelfand, S. Stoffel, S. J. Scharf, R Higuchi, G. T.
Horn & H. A. Erlich. 1988. Pnmer-directed enzymatic amplification
of DNA with a thermostable DNA polymerase. Science 239:487^191 .
Scott, M. P. & S. M Williams. 1993. Comparative reproductive success
of communally breeding burying beetles as assessed by PCR with
randomly amplified polymorphic DNA. Proc. Natl. Acad. Sci. USA
90:2242-2245.
Sinclair, M . R K. Mohn, G. Robert & D L. Roddick. 1985. Consider-
ation for the effective management of Atlantic scallops. Can. Tech.
Rep. Fish. Aqua. Sci. No. 1382.
Strongman, D. B. & R. M MacKay. 1993. Discrimination between Hir-
sutella longicolla var. longicolla and Hirsutella colla var. cornuta us-
ing random amplified polymorphic DNA fingerprinting. Mycologia
85:65-70.
Tingey, S. V. & J. P. del Tufo. 1993. Genetic analysis with random
amplified polymorphic DNA markers. Plant Physiol. 101:349-352.
Volckaert, F. & E. Zouros. 1989. Allozyme and physiological variation in
the scallop Placopecten magellanicus and a general model for the
effects of heterozygosity on fitness in marine molluscs. Mar. Biol.
103:51-61.
Volckaert, F., S. E. Shumway & D. F. Schick. 1991. Biometry and pop-
ulation genetics of deep- and shallow-water populations of the sea
scallop Placopecten magellanicus (Gmelin. 1791) from the Gulf of
Maine, pp. 156-163. In: S. E. Shumway and P. A. Sandifer. (ed.).
Scallop Biology and Culture. World Aquaculture Workshops, 1.
World Aquaculture Society, Louisiana State University, Baton Rouge.
LA
Warburg, O. & W. Christian. 1942. Isolierung und Knstallisation des
Garungsferments Enolase. Biochem. Z. 310:384-^123.
Welsh, J. & M. McClelland. 1990. Fingerprinting genomes using PCR
with arbitrary primers. Nucleic Acids Res. 18:7213-7218.
Welsh. J., C. Peterson & M. McClelland. 1991a. Polymorphisms gener-
ated by arbitrarily pnmed PCR in the mouse: application to strain
identification and genetic mapping. Nucleic Acids Res. 19:303-306.
Welsh. J . R J Honeycutt. M. McClelland & B. W. S. Sobral. 1991b.
Parentage determination in maize hybrids using the arbitrarily primed
polymerase chain reaction (AP-PCR). Theoret. Appl. Genet. 82:473-
476.
Wilde, J., R. Waugh & W. Powell. 1992. Genetic fingerprinting oiTheo-
broma clones using randomly amplified polymorphic DNA markers.
Theoret. Appl. Genet. 83:871-877.
Williams, J G. K., A. R. Kubelik, K. J. Livak, J. A. Rafalski & S. V.
Tingey. 1990. DNA polymorphisms amplified by arbitrary primers are
useful as genetic markers. Nucleic Acids Res. 18:6531-6535.
Yu. K. & K. P. Pauls. 1993. Identification of a RAPD marker associated
with somatic embryogenesis in alfalfa. Plant Mol. Biol. 22:269-277.
Journal of Shellfish Research. Vol. 13, No. 2. 555-564, 1994.
SEASONAL GROWTH MODELS FOR GREAT SCALLOPS (PECTEN MAXIMUS (L.)) AND
QUEEN SCALLOPS (AEQUIPECTEN OPERCULARS (L.))
EDWARD H. ALLISON1
Port Erin Marine Laboratory
University of Liverpool
Port Erin, Isle of Man, British Isles
ABSTRACT The growth in shell length of two northeast Atlantic scallop species. Pecten maximus and Aequipeelen operculars,
oscillates seasonally. The weights of the marketable parts of these scallops (the gonad and adductor muscle) are also seasonally
variable; the weight of the adductor muscle decreases during the winter, and the gonad weight decreases during the protracted summer
spawning period. \nP. maximus > 4 years old and A. operations > 2 years old. the amplitude of seasonal change in total edible yield
from each scallop exceeds mean annual increments. The growth in shell length of both species is modeled with the von Bertalanffy
growth function (VBGF). incorporating a sine-wave function to account for the seasonal fluctuations. The growth of the adductor
muscle, gonad, and total edible yield is described by combining the seasonalized VBGF for length with seasonalized length-weight
relationships.
KEY WORDS: Pecten maximus, Aequipeelen operations, seasonal growth, von Bertalanffy. length-weight
INTRODUCTION
The growth of most bivalves from temperate waters, including
Pecten maximus (L.) and Aequipeelen opercttlaris (L.), shows
strong seasonality (Comely 1974, Antoine et al. 1979, Taylor and
Venn 1979). In the North Irish Sea, the shells of these pectinids
show annual growth checks that are associated with the cessation
of growth during the winter months (November-March: Mason
1958. Soemodihardjo 1974. Dare and Deith 1991. Allison et al.
1994). Seasonal changes in weight are particularly evident in the
gonad, due to spawning and gametogenesis, and in the adductor
muscle, which acts as an energy store during the winter months
when food supplies are limited (Le Pennec et al. 1991. Thompson
and MacDonald 1991). The gonad and adductor muscle are the
commercially marketable portions of these pectinids. so that the
yield of "meat" from each scallop will vary according to the
season. Because these species are of considerable commercial im-
portance in the North Irish Sea and elsewhere in Europe (see
reviews by Ansell et al. 1991 and Brand et al. 1991a), explicit
consideration of seasonal growth patterns is required to refine
yield-based stock assessments and to determine optimal timing of
seasonal fishery closures with regard to maximizing the yield from
the fisheries.
Seasonal cycles of the somatic and reproductive tissues in pec-
tinids have normally been described with reference to a "standard-
ized animal," with shell length being the standardizing variable
(e.g., Ansell 1974, Comely 1974, Taylor and Venn 1979); there
have been no previous attempts to incorporate seasonal cycles into
growth curves for the edible portions of scallops. Seasonal fluc-
tuations have, however, been previously accounted for in model-
ing the growth in shell length of French stocks of P. maximus
(Buestel and Laurec 1976, Antoine et al. 1979).
The von Bertalanffy growth function (VBGF) is the basic
model chosen to describe the growth in both length and weight.
The VBGF is still the standard model in fishery-orientated growth
studies and has previously been shown to give a good fit for shell
'Present address: Renewable Resources Assessment Group. Imperial Col-
lege of Science. Technology and Medicine, 8 Prince's Gardens. London
SW7 1NA, United Kingdom.
growth of both P. maximus and A. opercularis (Orensanz et al.
1991 . Ansell et al. 1991 . for reviews). The VBGF for weight has
also been shown to provide a good fit to the annual growth of the
adductor muscle and the total edible yield in these populations
(Allison 1993).
The linkage between seasonal temperature fluctuation and
growth was pioneered by Ursin ( 1963), who incorporated temper-
ature in the VBGF and found that the sinusoidal seasonal temper-
ature fluctuations corresponded with the seasonality of the growth
pattern. The seasonality of growth has since been simulated by
incorporating a sine-wave function in the VBGF (e.g.. Pitcher and
MacDonald 1973, Cloern and Nichols 1978, Pauly and Gaschultz
1979, Antoine et al. 1979, Hanumara and Hoenig 1987, Somers
1988. Soriano and Jarre 1988, Hoenig and Hanumara 1990).
The recent literature abounds with studies of the fitting of
seasonally oscillating VBGFs for length, thanks to the incorpora-
tion of this growth curve in recent versions of the widely used
ELEFAN routine for extracting growth parameters from length-
frequency data (Gayanilo et al. 1987). However, there is little
work on seasonal patterns of growth in weight. This is despite the
fact that the seasonal patterns of growth in weight are of poten-
tially greater importance. Only Shul'man (1974) has presented
growth functions that incorporate seasonal patterns of weight loss
and gain. Recent interest in this problem has led Sparre ( 1991 ) to
present a method of yield per recruit (Y/R) analysis (Thompson
and Bell, 1934; Beverton and Holt. 1957) that incorporates sea-
sonality of growth and mortality. Sparre (1991) suggests using
empirically determined weights at age. collected at monthly inter-
vals, in a Thompson-Bell type Y/R equation, or combining a sea-
sonalized VBGF for length with season-specific length/weight
equations to calculate weights.
In this article, I aim to generate model-based estimates of
weight at age for incorporation into subsequent stock assessments.
To achieve this aim, I have fitted seasonalized VBGFs to scallop
length-at-age data and used a seasonalized model of the length-
weight relationship to fit a growth model to the weight at age of the
edible portions of the scallop. This growth model accounts for the
seasonal changes in yield that occurs as the result of spawning and
to the use of the adductor muscle as an energy store during the
winter.
555
556
Allison
MATERIALS AND METHODS
Sample Collection
Samples of P. maximus and A. operculars were collected with
"Newhaven" spring-loaded toothed dredges fished from the Uni-
versity of Liverpool's 15 m stern-trawler, the R.V. "Cuma."
Additional samples of A. operations were collected from on-
board samples of the catches of commercial vessels fishing with
both dredges and otter trawls. The samples of P. maximus were
taken from an inshore (Bradda Head) and offshore (South-east
Douglas) fishing ground (Fig. 1). from January 1987 to October
1988 at approximately monthly intervals. Monthly samples of A.
opercularis from SE Douglas were also taken during this period.
A. opercularis were also sampled from commercial vessels fishing
the East Douglas ground during the summer of 1988.
The shell length (anterioposterior axis) of all P. maximus in
each sample was measured to the nearest millimeter with a scallop
measuring board. A. opercularis catches were generally subsam-
pled if >400 animals were caught. Both species were aged by
means of the growth rings on the shell (see Allison et al. 1994).
Weight at age was measured by randomly subsampling approx-
imately 100 to 200 scallops and/or queens from each sample. In
addition to measuring shell length and age, adductor muscle and
gonad weights (to the nearest 0. 1 g) were determined with a Met-
tler 3000 PC electronic balance, after removing surface moisture
by blotting with absorbent paper and leaving the scallops to air dry
at room temperature for 1 hour.
Modeling Growth
The growth of the shell of both species was modeled using the
von Bertalanffy growth equation (von Bertalanffy 1938 and 1964),
modified to incorporate a term representing the seasonal pattern of
growth.
LA\
-*(! — »„) +S(f)\
(1)
where:
Bradda
Head
54° N
S.E. Douglas
4°W
L, = length at age t
L^ = asymptotic length
k = Brody/Bertalanffy growth coefficient
t0 = age at zero length
Sit) is the function describing the seasonal perturbation, de-
rived from the equation for a simple sine wave (e.g., see Batsche-
let, 1981, p 159):
Y = M + A cos((2ir/D ) (2)
where Y = the variable which is oscillating
M = mesor. or mean value of Y
A = amplitude of oscillation of Y
T = period of oscillation
4> = acrophase (/ - t„ where ts defines the beginning of
the sine wave).
There are several formulae for 5(0 in the literature. The version
used in this study, developed by Pauly and Gaschultz (1979),
modified by both Somers (1988) and Hoenig and Hanumara
(1990). gives the best description of the seasonal oscillation with
respect to the unbiased estimation of t„:
kC kC . „
SU) = — sin2iT(r - t j) - — sin2Tr(f„
2tt 2tt
(3)
Figure 1. The North Irish Sea Fishing grounds from which P. maximus
and A. opercularis were sampled.
where C = amplitude of seasonal perturbation
t, = time of start of oscillation relative to t„
k = the Brody/Bertalanffy growth coefficient
In fitting the model to the data, an ordinary VBGF was first
fitted to the raw length-at-age (in months) data to obtain initial
estimates of the parameters L,, k, and t0. The model was fitted to
the data by the use of the nonlinear regression (NLR) procedure in
SPSSX, which fits data to user-defined models iteratively using
the Levenberg-Marquadt algorithm (SPSSX, 1988). The proce-
dure requires initial estimates of the parameters L^, k, and t0, but,
in simple models, is robust with respect to them, converging to the
same solution from a wide range of starting values. The fitted
ordinary VBGF was then subtracted from the data, and the resid-
uals of the fit of the data to the ordinary VBGF, which should lie
on a sine wave of constant mesor and period of 1 year, were
examined to obtain starting values for the parameters C and ts. The
seasonal model [equation 1, where SU) is defined in equation 3]
was then fitted to the raw length-at-age data by the use of the
SPSSX-NLR procedure. Good initial estimates of the parameters
are required to fit this more complex model. The youngest age
class in the P. maximus samples was not used to fit the equation,
because it is likely that only the larger fraction of the length dis-
tribution would be retained by the size-selective fishing gear.
The data set for A. opercularis from the East Douglas ground
was limited. Seasonal oscillations were readily discernible, al-
though the many missing months made it difficult to fit a seasonal
curve without the prior assumption of growth cessation during the
winter. The seasonal VBGF was fitted by constraining C to 1 to
simulate the winter cessation of growth. The curve was fitted using
the constrained nonlinear regression procedure (CNLR) in SPSSX.
Because the two procedures (NLR and CNLR) use different algo-
rithms and convergence criteria, direct comparison of parameter
estimates from the SE Douglas and E Douglas grounds should be
undertaken with caution.
The seasonal patterns of growth of the edible portions of the
scallop (gonad and adductor muscle) were modeled by converting
the seasonally oscillating growth curves for length to weight using
Seasonal Growth Models for Scallops
557
season-specific length-weight relationships. Age was expressed in
months (entered as decimal fractions of a year). The computational
steps were as follows:
1. For each sample date, geometric mean (GM) regression
parameters of the relationship between shell length and adductor
muscle and between shell length and gonad weight were calculated
from the linearized allometric length-weight equation (Ricker 1975a).
InH' = a + b{\nL)
(4)
GM regression, rather than ordinary least-squares regression,
was chosen for computation of the relationship between these two
variables, because they have a functional, rather than a dependent
relationship (see Ricker 1973 and 1975a, Saila et al. 1988, for
methods of fitting and calculating errors on the model coefficients,
and Ricker 1975b. Laws and Archie 1981. for discussion).
2. In order to smooth variability in the data and calculate
length/weight relationships for times of year not sampled, a model
was then fitted to the calculated values of the exponent (b) and the
intercept (a) of the length-weight relationship. The equation for a
sine wave (equation 2) was chosen to model the seasonality of the
regression parameters and was fitted by least-squares regression
using the Quasi-Newtonian search algorithm in the NLR module
of SYSTAT (Wilkinson 1990). Starting values for mesor, ampli-
tude, and ts are easily obtained by visual inspection of the data.
No model was fitted to the length/gonad weight regression
parameters for A. opercularis from SE Douglas, because the pa-
rameters did not show an obvious seasonal cycle. The analysis for
these data therefore terminated at this step.
3. Model parameters from step 2 provide smoothed length-
weight regression parameters. These were used to convert monthly
length-at-age values (predicted from the seasonally oscillating
VBGF for length) to predicted weight-at-age values for gonads and
adductor muscles from:
W, = a + AaCOSid-n/T^JL;
(i> + /t(,COS(2TT/rXt>(,)
(5)
where a = the intercept of the functional length-weight regres-
sions equations
Aa = the amplitude of seasonal fluctuation of a (from
equation 2)
4>a = acrophase of seasonal pattern of a (from equation 2),
and b,Ah,^h are the equivalent parameters for the
slope of the length- weight regression equations.
This model was found to fit the growth pattern of the adductor
muscles in all cases and was also applicable to the growth of A.
opercularis gonads in samples from the SE Douglas ground. The
annual increase in weight of gonads was slow in P. maximus from
Bradda Head beyond the age of 5 years: no annual increase in the
weight of gonads was observed in P. maximus more than 8 years
old from SE Douglas, whereas both shell length and adductor
muscle weight continue to increase slowly over the whole age
range sampled. This led to predicted gonad weights that fit ob-
served data well in younger scallops, but both mean weight and the
amplitude of annual fluctuations were overestimated in older scal-
lops. In order to correct these errors and obtain realistic values of
predicted total yield, post hoc modifications to models of gonad
weight growth were applied. In order to simulate the changes in
weight of the gonads of the older scallops on the SE Douglas
ground, a sine-wave equation of constant amplitude (equation 2)
and constant mesor (A/) was fitted directly to data on weight at age
of these gonads. In the case of P. maximus from Bradda Head,
gonad weight continued to increase, approximately linearly, over
the age range from 5 to 1 1 years, so the mesor (W) in equation 2
was modified to give:
GW1 = (c + dt) + A cos(2ir/7")<|> (6)
where c and d are the intercept and slope of the linear regression
of mean gonad weights on age, between ages 5 and 1 1 .
4. To obtain predicted total edible yields, model-predicted go-
nad and adductor muscle weights were summed.
Weight data and length-weight relationships calculated from
samples taken during 1987 and 1988 were pooled to give a single
series of data covering most months of the year. Mean gonad
weight and mean adductor muscle weights at age were plotted with
model-predicted growth curves to provide a visual assessment of
the fit of the model.
RESULTS
Growth curves fitted to shell length at age of monthly samples
of P. maximus from the Bradda and SE Douglas grounds show a
seasonal pattern, with growth occurring in a series of steps: a
plateau corresponding to the winter cessation in shell growth (Oc-
tober to March), exponentially increasing length in spring/early
summer, and slowing down through August/September to reach
another asymptote the following winter.
The pattern is only clearly visible in scallops aged 2 to 5 years
in the Bradda Head samples (Fig. 2a). Between-sample variability
masks the seasonal pattern in older age classes, where the annual
increment is much less and sample sizes are smaller. The SE
Douglas samples show visible seasonal growth fluctuations up to
age 8+ (Fig. 2b), reflecting the different growth pattern on this
ground, where successive growth increments decrease in size more
slowly (lower von Bertalanffy k value). The fit of the seasonal
VBGF is good on both grounds, and the pronounced seasonal
fluctuation is reflected in the high values of the seasonality pa-
rameter C (Table 1). The fact that C exceeds unity produces an
apparent slight decrease in length during the winters. This is be-
cause of the mathematical form of the curve, which simulates a
waveform rather than a series of flat-topped discontinuities.
A. opercularis from SE Douglas monthly samples showed
strong seasonality (Fig. 3a). The data set from the East Douglas
ground was more limited but less variable, and the seasonal os-
cillations are also readily discernible (Fig. 3b). although the many
missing months make it difficult to fit a seasonal curve without the
prior assumption of growth cessation during the winter.
The slopes of the length-weight relationships (b) show consid-
erable variability and large standard errors (Figs. 4 and 5). This is
not unexpected; the weight of individual scallops is variable, and
wet weight is difficult to measure accurately and consistently.
Short-term variability, especially of gonads during summer
spawning periods, also accounts for inconsistencies between suc-
cessive samples. The pooling of 2 years' data may also increase
variability, but was considered necessary in order to obtain data
for as many months of the year as possible and to increase the
likelihood that values were generally applicable, rather than spe-
cific to a single year. Despite the variability, an underlying pattern
of seasonal fluctuation can be detected, and separate sine- wave
models have been fitted to the monthly values of b from each set
of samples: the intercept a is simply the inverse correlate of b. The
models provide reasonable fit to all but the calculated length/gonad
weight parameters for A. opercularis from SE Douglas (Fig. 4).
For P. maximus, the model predicts the maximum values of the
558
Allison
~i 1 r
6 8 10
Age (years)
6 8 10 12 14 16 IX
Age (years)
Figure 2. VBGF with seasonal oscillation fitted to P. maximus length-
at-age data from (a) the Bradda Head fishing ground, and (b) the SE
Douglas ground. Ages (to the nearest month) were calculated from the
number of growth rings and date of sampling, assuming a July 1st
"birthday."
exponent {b) of the length/adductor muscle weight and length/
gonad weight relationships to occur in midwinter (early December
in scallops from SE Douglas, early January in scallops from
Bradda Head), with the minima there occurring in May and June,
respectively (Fig. 4a). A. opercularis from the SE Douglas ground
have a maximum exponent of the length/adductor muscle weight
relationship in late February (Fig. 4b) and a minimum in August.
The parameters of the length/gonad weight equation for ,4. oper-
cularis from SE Douglas do not show clear seasonality (Fig. 4b),
so no sine-wave model has been fitted.
Seasonal data for A . opercularis from the East Douglas ground
are limited, but a sine-wave model can be fitted to provide
monthly estimates of the length/weight parameters and to enable a
seasonally oscillating model to be fitted to the observed weight-
at-age data (Fig. 4b). The parameters of the fitted sine-wave mod-
els shown in Figure 4 are given in Table 2.
The growth of P. maximus gonads and the growth of adductor
muscles in samples taken from Bradda Head (Fig. 5a) and SE
Douglas (Fig. 5b) both show strong seasonality. Model-predicted
values generally showed good agreement with observed mean
weight-at-uge data for adductor muscles. There are slight overes-
timations of adductor muscle weights of scallops aged 2 to 2.5
years from the Bradda Head ground, and underestimation of the
weight of adductor muscles of scallops from SE Douglas of ages
4.9 to 5.3 years. Data for the mean weight of muscles and gonads
from older year classes show more scatter than for younger year
classes, and the fit of the predicted curves is poorer for ages
beyond 6 in Bradda Head scallops and 10 in SE Douglas scallops.
Although fitting the models to growth of adductor muscles was
relatively straightforward, fitting the model to gonad data was
more problematic, because the exponent of the length-weight
equation is generally high (5 to 9), reflecting rapid increase in
gonad weight with little increase in length as the scallop or queen
scallop attains maturity, when further growth of gonads proceeds
more slowly. Beyond the age of 5 years, the growth of the gonads
in P. maximus from Bradda Head is slow (Fig. 5a), and no annual
increase in the weight of gonads is observed in P. maximus more
than 8 years old from SE Douglas (Fig. 5b).
The gonad shows clear seasonality in weight gain and loss,
with annual weight maxima being reached in March/April and
minima in September, after the autumn spawning. The adductor
muscle weights reach annual maxima in October and minima in May.
The seasonal cycles of weight change in adductor muscles and
gonads vary out of phase (approximately 160°), and the seasonal-
ity of total yield is thus dampened to some extent (Fig. 5a and b).
For P. maximus from Bradda Head, the annual maxima occur in
February to March, while gonad weights are increasing and close
to the maxima and adductor muscle weights are decreasing. Low-
est yields are found in September, when the gonad is generally
spent and the adductor muscle is still increasing in weight. Be-
tween the ages of 3.0 and 4.6 years, when the yield is increasing
most rapidly and the scallop is becoming vulnerable to exploita-
tion, there is little or no seasonal decrease in yield. However, a
period of cessation in growth of yield does occur in scallops of
between 3.2 and 3.75 years old. In older scallops, there is marked
seasonality in total edible yield per scallop.
P. maximus from SE Douglas show a similar seasonal pattern
in growth of the edible yield portions (Fig. 5b). Maxima occur in
March; minima occur in September. Seasonality of edible yield is
damped up to age 6 years, but still shows cessation of growth over
periods of up to 6 months in each year. Above age 6, there are
significant annual decreases in yield. These are worthy of consid-
eration on this fishing ground, where, at the time of sampling,
over 50% of the exploitable population was aged 6 years or over
(Allison 1993).
The limited data set for A. opercularis from the East Douglas
fishing grounds (Fig. 6a) indicates an increase in both muscle
weight and gonad weight during the late spring and early summer.
Gonad weights reach maxima in July to August, and adductor
muscle weights peak in September to October. Total yield is max-
imized in September. Predicted gonad weights are at their minima
in January, although the data suggest that this minima is in fact
reached 3 months earlier. Gonad weights probably do not increase
and decrease sinusoidally, which will account for the inability of
the model to simulate rapid build-up of gonads and sharp decrease
due to spawning. Adductor muscle weights are better predicted by
the model for A. opercularis from both the East and SE Douglas
fishing grounds (Fig. 6).
A. opercularis from the SE Douglas ground (Fig. 6b) show
Seasonal Growth Models for Scallops
559
TABLE 1.
Parameters of seasonally oscillating von Bertalanffv growth models fitted to P. maximus and A. opercularis length-at-age data by nonlinear
least-squares regression.
Species and Area
n
r2
Lx (mm)
*
'«
ts
C
P. maximus, Bradda
4120
0.807
133.68
0.466
0.280
0.524
1.192
P. maximus, SE Douglas
2758
0.805
133.92
0.329
0.009
0.558
1.049
A. opercularis, SE Douglas
3153
0.637
75.42
0.696
-0.149
0.428
1.356
A. opercularis, E Douglas'
4927
0.856
77.05
0.678
-0.189
0.665
1.000
Seasonality parameter (C) constrained to 1 for model fitting-
Strong seasonal cycles in adductor muscle weight, the seasonal
change in weight exceeding the annual growth increment in queen
scallops older than 24 months. Maxima occur in September to
October; minima occur in February to March. The model tends to
underestimate muscle weights of A. opercularis aged 2 to 2.5. The
gonad cycle is more variable than that of P. maximus, possibly
because of multiple spawning peaks and rapid recovery of gonads
in this species. It was not possible to simulate seasonal changes in
gonad weight (and therefore in total yield I from these data by the
3
3
2 3 4 5
Age (years)
Figure 3. VBGF with seasonal oscillation fitted to .4. opercularis
length-at-age data from (a) the Southeast Douglas fishing ground, and
(b) the East Douglas ground. Ages (to the nearest monthl were calcu-
lated from the number of growth rings and date of sampling, assuming
a July 1st "birthday."
modeling approach applied to P. maximus and to A. opercularis
from the East Douglas ground. Gonad weight is generally higher
in spring and early summer, with lower adductor muscle weights
during the same period. The gonad makes up a relatively small
proportion of the yield, and the yield will therefore follow the
pattern of adductor muscle growth closely. Maxima occur in Au-
gust and September; minima occur in February.
DISCUSSION
The seasonality of growth in pectinids. correlated with annual
water temperature cycles, is well known (Ansell 1974. Comely
1974, Taylor and Venn 1979). Growth of the shell and increase in
body mass cease during the winter months, due either to the effects
of lower temperature on metabolic rate or to insufficient food
during the winter to maintain metabolic requirements (Broom and
Mason 1978). The resumption of growth of the shell and increase
in total weight each spring is triggered by rising temperature or
increase in food supply after the spring phytoplankton bloom
(Broom and Mason 1978. Vahl 1980).
The VBGF with seasonal perturbation provides a good fit to
length-at-age data for both species. A slight decrease in length
over winter, rather than an extended period of growth cessation, is
predicted by the form of the curve, which incorporates a sine
wave. Sine-wave models of this type have previously been used to
model growth of P. maximus in French stocks (Buestel and Laurec
1976, Antoine et al. 1979). Pauly et al. (1992) have developed a
seasonal growth model that incorporates a period of zero growth,
rather than a simple sine wave; their model may provide a slightly
improved fit to scallop length-at-age data than the standard sine-
wave model used here.
Models for seasonal growth cycles of the gonad and adductor
muscle also fit well. The two constituents of the yield vary 160°
out of phase; whereas the gonad weight decreases because of
spawning in the summer months, the adductor muscle is increasing
in weight. During the winter, the adductor muscle loses weight
and the gonad gradually fills, reaching a maximum weight in the
spring. The large adductor muscle acts as the main storage site for
metabolic reserves in P. maximus, and weight loss in winter oc-
curs as the result of mobilization of reserves to contribute to met-
abolic requirements and development of the gonad (Le Pennec et
al. 1991, Faveris and Lubet 1991, Barber and Blake 1991, for
reviews). A. opercularis adductor muscle weights show the same
pattern of seasonal fluctuation, and it is likely that the adductor
muscle performs the same function as in P. maximus.
Spawning of P. maximus takes place in the summer months
(June to August). Only one major spawning per year has been
observed in this study, and the gonad growth models account for
only one, although a more detailed study in the same area (Mason
1958) has indicated two spawning peaks may sometimes occur.
560
Allison
a) Pecten maximus
Bradda Head
L:GWT
t 1 1 r
120 180 240 300 360
Julian Day
100
90
80 -
70
60 -
50
36
b 32
SE Douglas
L:GWT
L : AMWT
~i i i i i r
60 120 180 240 300 360
Julian Day
b) Aequipecten opercularis
SE Douglas
120
11 0
100
90 -
80
70
60
50
40
30 -J
38
36
34
32
30
28
26
2 4
22
L:GWT
♦ *
AMWT
180
jlian Day
Figure 4. The seasonal variation in the slope (ft) of functional regression equations between shell length-adductor muscle weight (L:AMWT) and
shell-length:gonad weight (L:GWT) of (a) P. maximus from the Bradda Head and SE Douglas Ashing grounds and (b) A. opercularis from the
E and SE Douglas fishing grounds. Standard errors of the functional regression parameters are shown, and sine-wave models are fitted to each
data set, with the exception of L:GWT of A. opercularis from SE Douglas, where data were too variable.
Seasonal Growth Models for Scallops
561
a) Bradda Head
60
["■•it i — i — i — i — i — i — i — i
0123456789 10 11
Age (years)
b) SE Douglas
50 -|
0 12 3 4 5 6
Age (years)
Figure 5. Mean gonad (closed symbols) and adductor muscle (open
symbols) weights at age of P. maximus from (a) the Bradda Head
fishing ground and lb) the SE Douglas ground, with fitted models
incorporating seasonal oscillations ( 1 — gonad weights, 2 = adductor
muscle, 3 = total edible yield). Models for adductor muscle weights
were VBGFs incorporating seasonal fluctuations (equation 5), as were
models for gonad growth up to age 5 on the Bradda Head ground and
age 8 on the SE Douglas ground. Gonad weights above age 5 on the
Bradda Head ground were modeled by fitting equation 2 empirically,
and gonads of scallops of age 8+ on the SE Douglas ground were
modeled by fitting equation 6 empirically. Ages were calculated to the
nearest month, assuming a July 1st "birthday."
Spawning in A. opercularis appears to occur sporadically, with
rapid recovery of the gonad. There may be two or three peaks of
spawning activity each year in A . opercularis from the North Irish
Sea, all of which occur over the summer months (Aravindakshan
1955, Soemodihardjo 1974, Paul 1978, Duggan 1987). The sim-
ple sine-wave model used here is unable to simulate multiple
spawning peaks occurring at irregular time intervals, but is able to
describe the basic summer-winter cycle. In populations with more
than one distinct spawning period, more complex models would be
required to simulate the pattern of weight loss and gain of the
gonad.
Beyond the age of 4 years, the seasonal fluctuation in the
weight of both the adductor muscle and the gonad of P. maximus
exceeds the annual growth increment; it is the dominant feature of
the growth pattern of older scallops. The amplitude of seasonal
weight fluctuations in the edible tissues of A. opercularis exceeds
the annual growth increment after the second growth season. Be-
cause the minimum legal landing size of P. maximus in the North
Irish Sea (110 mm shell length) and the minimum commercially
acceptable size of A. opercularis (55 mm) are attained at the ages
of 4 and 2 years, respectively (Brand et al. 1991a). these seasonal
cycles in yield are of great importance to the fishery.
The models described here serve to highlight the importance of
the seasonal cycle in the study of growth in scallops, but should be
regarded as preliminary, rather than exemplary, approaches to the
problem. Although the incorporation of seasonal cycles into
growth curves for length is relatively straightforward by the use of
established methods (Hoenig and Hanumara 1990, for review),
hypothesis and goodness-of-fit testing suffer from the questionable
assumptions made in the calculation of confidence limits after
NLR analysis (Donaldson and Schnabel 1987. Cerrato 1990). The
techniques used here to model seasonal weight changes require
additional development. The models lack an error structure, so no
objective goodness-of-fit criteria can be established. Direct fit of a
seasonalized VBGF for weight by nonlinear least squares proved
difficult; the high value of the exponent of the length/adductor
muscle and length/gonad weight relationships in scallops caused
successive iterations to diverge so widely that optimal solutions
were seldom found. In choosing to fit the models in stages through
calculation of length-weight relationships, complexity is in-
creased, and the resultant number of parameters in the models is
TABLE 2.
Parameters of sine-wave models describing seasonal variability in the exponent and intercept of functional length-weight regression
relationships for scallops [P. maximus) and queens (A. opercularis).
Exponent (b)
Intercept (a)
Species & Area
M„
A„
t„*,
r2
y
M„
A,
«««)
r2
V
Length: adductor muscle weight
Scallops. Bradda
3.24
0.247
8
0.640
13
-12.24
1.161
195
0.626
13
Scallops. SE Douglas
3.21
0.176
347
0.507
10
-12.05
0.751
172
0.438
10
Queens, SE Douglas
2.88
0.265
26
0.680
12
- 10.34
1.234
215
0.711
12
Queens, E Douglas
2.61
0.713
326
0.974
4
-9.25
2.889
149
0.998
4
Length: gonad weight
Scallops. Bradda
8.52
1.619
9
0.615
12
-38.63
7.525
187
0.605
12
Scallops, SE Douglas
6.74
0.989
342
0.717
10
-30.40
4.975
152
0.746
10
Queens, SE Douglas
Queens. E Douglas
4.51
1.102
361
0.993
4
-18.93
5.571
179
0.999
4
Note: M. mesor; A. amplitude; t„. = time at which sine wave begins.
562
Allison
certainly higher than would be required to fit a purely empirical
model to the same data. The direct fit of a seasonalized VBGF for
weight may be easier with data from whole finfish. where the
length- weight exponent (b) can either be fixed at 3, or assumed to
approximate it in entering starting values for the parameters. In
this study b was found to show considerable seasonal variation,
and fixing b and describing seasonality by only allowing a (the
length:weight "condition factor") to fluctuate were not possible.
Seasonal fluctuations in b imply that the degree of seasonal fluc-
tuation is size dependent, and examination of the data shows that
the seasonal fluctuations in the edible yield are greater in larger,
sexually mature scallops.
The seasonal growth and reproductive cycles described in this
article are of great importance in determining both yield and fish-
ing strategy in these fisheries. P. maximus can only be fished from
November 1st to May 31st under current legislation. This analysis
b) SE Douglas
12
T
3 4 5
Age (years)
Figure 6. Mean weights at age of gonads (closed symbols) and adduc-
tor muscles (open symbols I of A. opercularis from (a) E Douglas and
(b) SE Douglas. Model-predicted weight at age of gonads ( I ), adductor
muscles (2), and total yield (3) are shown for A. opercularis from the
E Douglas fishing grounds. Predicted weight at age of adductor mus-
cles only ( 1 ) are shown for samples from the SE Douglas ground. No
model has been fitted to SE Douglas gonads or total yield, because
these data were too variable.
of seasonal growth patterns suggests that this period represents the
optimal time of year for fishing this species with a 7 month fishing
season. The previous year's growth has been completed, and total
yield remains relatively constant, despite relative changes in gonad
weight and adductor muscle weight. The gonads of >80% off.
maximus are full enough to be acceptable to the market throughout
the season (Allison 1993), and muscle weight is high until the
latter part of the season. Provided that natural mortality does not
show strong seasonality, the current fishing season optimizes Y/R
for any given age at first capture. The fact that this is the case is
largely coincidental; the closed season was originally introduced in
the late 1930s to reduce spoilage of scallops transported to markets
during the warmer summer months. The closed season has subse-
quently remained in place with the objective of reducing effort and
closing the grounds during the period of spawning and spat set-
tlement in this species (Brand et al. 1991b).
The seasonality of natural mortality is potentially important in
determining optimal fishing times. More detailed examinations of
the effect of seasonality on yield to the fishery could be performed
if monthly or quarterly fishing and natural mortality rates could be
calculated. This would help to determine the optimal time of year
for harvesting of restocked scallops or closed areas (Brand et al.
1991b), where shorter periods of harvesting may be legislated.
Maximum yields are attained just over half way through the cur-
rent fishing season, in February or March. If a later start to the
fishing season is contemplated, or if particular grounds are opened
for short periods only, the gains due to increase in weight must be
balanced against the losses due to mortality during the period
before allowing fishing, bearing in mind that no new recruits enter
the fishery over the winter, because shell growth takes place only
during the summer.
A. opercularis tend to be fished in the closed season for P.
maximus. June to October (Allison 1993), and the yield from an
individual A. opercularis is maximized in the latter part of this
season (late August to September), when the adductor muscle is in
peak condition. The shell of A. opercularis grows rapidly during
the early summer (June to July), and the 1 + age class does not
recruit to the fishery until the latter part of the season on most
grounds. On the E Douglas fishing ground. A. opercularis attains
the minimum commercially acceptable size of approximately 55
mm shell length during the second growth season (age 1 + ; Fig.
3b). but not until the third growth season (age 2+ ) in A. opercu-
laris from SE Douglas (Fig. 3a). Low meat yields for a given shell
length and apparently lower catchability of trawled A. opercularis
during spring and early summer (due to reduced swimming ability
correlated with seasonal energy storage cycles; A. R. Brand un-
published data) mean that fishing for A. opercularis in the early
part of the season is a barely viable concern. The option of legis-
lating an extension to the scallop season into June or July to cover
this unprofitable period is not tenable, because scallop gonads are
spent or partially spent during this period and are therefore unac-
ceptable to the market; meat weights are also low.
In North American Placopecten magellanicus (L.) fisheries,
which are regulated by control of the meat count (maximum num-
ber of meats per pound weight), the variability in growth rates and
shell length/adductor muscle weight relationships has far-reaching
implications for management and has consequently received con-
siderable attention (Worms and Davidson 1986, Shumway and
Schick 1987. for reviews), although seasonal growth models have
not been constructed. This work highlights the magnitude of sea-
sonal variability in yield from pectinids in temperate waters. In-
Seasonal Growth Models for Scallops
563
corporation of seasonal growth patterns into yield-based stock as-
sessments is recommended for the North Irish Sea scallop fisher-
ies. Investigations of the importance of growth seasonality in other
scallop fisheries would be worthwhile.
ACKNOWLEDGMENTS
This work was carried out as part of a research program on
scallop fisheries in the North Irish Sea. funded by the Isle of Man
Department of Agriculture and Fisheries. 1 am grateful to Dr.
A. R. Brand, the program director, for his guidance and for his
comments on the manuscript. S. Lawrence assisted with data col-
lection. U. A. W. Wilson prepared Figures 1 and 2, and Dr. D.
Pauly made a number of helpful suggestions on analytical methods
as did an anonymous referee. The manuscript was prepared during
the tenure of an award under the Associate Professional Officers
Scheme of the UK's Overseas Development Administration.
LITERATURE CITED
Allison. E. H. 1993. The dynamics of exploited populations of scallops
(Pecten maximus L. ) and queens (Chlamys opercitlaris L. ) in the North
Irish Sea. Ph.D. Thesis. University of Liverpool, 254 pp.
Allison, E. H.. U. A. W. Wilson & A. R. Brand. 1994. Age-deter-
mination and the first growth ring in North Irish Sea populalions of the
scallop, Pecten maximus L. J. Moll. Stud. 60:91-95.
Ansell, A. D. 1974. Seasonal changes in biochemical composition of the
bivalve Chlamys septemradiata from the Clyde Sea area. Mar. Biol.
25:85-99.
Ansell, A. D., J-C. Dao & J. Mason. 1991. Three European scallops:
Pecten maximus. Chlamys {Aequipecten) operations and C. (Chlamys)
varia. pp. 715-750. In: S. E. Shumway. (ed.). Scallops: Biology.
Ecology and Aquaculture. Elsevier. Amsterdam, The Netherlands.
Antoine, L.. P. Arzel. A. Laurec & E. Morize. 1979 La croissance de la
coquille Saint-Jacques (Pecten maximus (L.I) dans les divers gisements
Francais. Rapp. P-v. Reun. Cons. Int. E.xplor. Mer 175:8-19.
Aravindakshan. I. 1955. Studies on the biology of the queen scallop
Chlamys opercitlaris (L). Ph.D. Thesis. University of Liverpool, 79
PP
Barber. B. J. & N. J. Blake. 1991. Reproductive physiology, pp. 377-
420. In: S. E. Shumway. (ed.). Scallop: Biology, Ecology and Aqua-
culture. Elsevier. Amsterdam, The Netherlands.
Batschelet, E. 1981. Circular Statistics in Biology. Academic Press, New
York. 371 pp.
Beverton, R. J. H. & S. J. Holt. 1957. On the dynamics of exploited fish
populations. Fishery Investigations. MAFF. U.K. (Ser. 2), 19, 533 pp.
Brand. A. R., E. H. Allison & E. J. Murphy. 1991a. North Irish Sea
scallop fisheries: a review of changes, pp. 204-21X. In: S. E. Shum-
way & P. A. Sandifer. (eds.). An International Compendium of Scal-
lop Biology and Culture. World Aquaculture Workshops 1 . World
Aquaculture Society. Baton Rouge, LA.
Brand. A. R., U. A. W. Wilson. S. J. Hawkins, E. H. Allison & N. A.
Duggan. 1991b. Pectinid fisheries, spat collection and the potential for
stock enhancement in the Isle of Man. ICES Mar. Sci. Symp. 192:79-
86.
Broom. M. J. & J. Mason. 1978. Growth and spawning of the pectinid
Chlamys opercitlaris in relation to temperature and phytoplankton con-
centration. Mar. Biol. 47:277-285.
Buestel, D. & A. Laurec. 1976. Croissance de la coquille Saint-Jacques
(Pecten maximus L.) en Rade de Brest et en Baie de Saint-Brieuc.
Haliotis 5:279-283.
Cerrato, R. M. 1990. Interpretable statistical tests for growth comparisons
using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat.
Sci. 47:1416-1426.
Cloem, J. E. & F. H. Nichols. 1978. A von Bertalanffy growth model
with a seasonally varying coefficient. J . Fish. Res. Bd. Can. 35:1479-
1482.
Comely, C. A. 1974. Seasonal variations in the flesh weights and bio-
chemical content of the scallop (Pecten maximus L. ) in the Clyde Sea
area. J. Cons. Int. E.xplor. Mer 35:281-295.
Dare, P. J. & M. R. Deith, 1991. Age determination of scallops, Pecten
maximus (Linnaeus 1758) using stable oxygen isotope analysis, with
some implications for fisheries management in British waters, pp. 1 18-
133. In: S. E. Shumway & P. E. Sandifer. (eds.). An International
Compendium of Scallop Biology and Culture. World Aquaculture
Workshops 1, World Aquaculture Society. Baton Rouge, LA.
Donaldson. J. R. & R. B. Schnabel. 1987. Computational experience with
confidence regions and confidence intervals for nonlinear least squares.
Technometrics 29:67-82.
Duggan, N. A. 1987. Recruitment in North Irish Sea scallop stocks. Ph.D.
Thesis, University of Liverpool, U.K., 178 pp.
Faveris. R. & P. Lubet 1991. Energetic requirements of the reproductive
cycle in the scallop Pecten maximus (Linnaus. 1758) in Baie de Seine
(Channel), pp. 67-73. In: S. E. Shumway & P. A. Sandifer. (eds.).
World Aquaculture Workshops 1, World Aquaculture Society. Baton
Rouge. LA
Gayanilo, F. C, Jr., M. L. Soriano & D. Pauly. 1987. A draft guide to
the Compleat ELEFAN. ICLARM Software 2, 70 pp.
Hanumara. R. C. & N. A. Hoenig. 1987. An empirical comparison of the
fit of linear and non-linear models for seasonal growth in fish. Fish.
Res 5:359-381.
Hoenig, N. A. & R. C. Hanumara. 1990. An empirical comparison of
seasonal growth models. Fishhyte 8:32-34.
Laws. E. A.&J. W.Archie. 1981 . Appropriate use of regression analysis
in Marine Biology. Mar. Biol. 65:13-16.
Le Pennec. M., P. G. Beninger. G. Dorange & Y-M. Paulet. 1991.
Trophic sources and pathways to the developing gametes of Pecten
maximus (Bivalvia: Pectinidae). J. Mar. Biol. Assoc. U.K. 71:451-
463.
Mason. J. 1958. The breeding of the scallop. Pecten maximus (L.), in
Manx waters. J. Mar. Biol. Assoc. U.K. 37:653-671.
Mason. J. 1983. Scallop and Queen Fisheries in the British Isles. Fishing
News Books, Farnham. U.K., 144 pp.
Orensanz, J M.. A. M. Parma & O. O. Inbane. 1991. Population dy-
namics and management of natural stocks, pp. 625-689. In: S. E.
Shumway (ed.). Scallops: Biology, Ecology and Aquaculture.
Elsevier, Amsterdam. The Netherlands.
Paul, J. D. 1978. The biology of the queen scallop. Chlamys opercitlaris
in relation to its potential for cultivation. PhD. Thesis. University of
Liverpool.
Pauly, D. & G. Gaschultz. 1979. A simple method for fitting oscillating
length growth data, with a program for pocket calculators. ICES CM
1979/G24. 26 pp.
Pauly. D., M. Soriano-Bartz, J. Moreau & A. Jarre -Teichmann. 1992. A
new model accounting for seasonal cessation of growth in fishes. Aust.
J. Mar. Freshwater Res. 43:1151-1156.
Pitcher. T. J. & P. D. M. MacDonald. 1973. Two models for seasonal
growth in fishes. J. Appl. Ecol. 10:599-606.
Ricker, W. E. 1973. Linear regression in fisheries research. J. Fish. Res.
Bd. Can. 30:409-434.
Ricker. W. E. 1975a. Computation and interpretation of biological statis-
tics of fish populations. Bull. Fish. Res. Bd. Can. 191:382 pp.
Ricker. W. E. 1975b. A note concerning Professor Jolicoer's comments
[on linear regression in fishery research]. J. Fish. Res. Bd. Can. 32:
1494-1498.
Saila. S. B.. C. W. Recksiek & M. H. Prager. 1988. Bivanate regression
analysis including orthogonal, least squares, functional and Bartlett's
regressions, pp. 131-139. In: Basic Fishery Science Programs. Devel-
564
Allison
opments in Aquaculture and Fisheries Science No. 18, Elsevier, Am-
sterdam, The Netherlands.
Shul'man. G. E. 1974. Life cycles of fish; physiology and biochemistry.
Israeli Progress in Science Translations. Wiley & Sons, New York.
Shumway, S. E. & D. F. Schick. 1987. Variability of growth, meat count
and reproductive capacity in Placopecten magellanicus: are current
management policies sufficiently flexible? ICES CM 1987/K:2, 8 pp.
Soemodihardjo. S. 1974. Aspects of the biology of Chlamys opercularis
(L.) (Bivalvia) with comparative notes on four allied species. PhD
Thesis. University of Liverpool, U.K., 1 10 pp.
Somers, I. F. 1988. On a seasonally oscillating growth function. Fishbyte
6:8-11.
Soriano, M. L. & A. Jarre. 1988. On fitting Somers' equation for season-
ally oscillating growth, with emphasis of T-subzero. Fishbyte 6:13-14.
Sparre, P. 1991 . Estimation of yield per recruit when growth and mortality
oscillate seasonally. Fishbyte 9:40-44.
SPSSXInc. 1988. Nonlinear regression: NLR and CNLR. Chapter 36, pp.
666-721.
Taylor, A. C. & T. J. Venn. 1979. Seasonal variation in weight and
biochemical composition of the tissues of the queen scallop, Chlamys
opercularis. from the Clyde Sea area. J Mar. Biol. Assoc. UK 59:
605-621.
Thompson, R. J. & B. A. MacDonald. 1991. Physiological integrations
and energy partitioning, pp. 347-372. In: S. E. Shumway. (ed.). Scal-
lops: Biology, Ecology and Aquaculture. Elsevier, Amsterdam, The
Netherlands.
Thompson, W. F. & F. H. Bell. 1934. Biological statistics of the Pacific
halibut fishery. 2. Effect of changes in intensity upon total yield and
yield per unit of gear. Rep. Int. Comm. Pacific Halibut Fishery 8:49pp.
Ursin, E. 1963. On the incorporation of temperature in the von Bertalanffy
growth equation. Medd. Dan. Fisk. Havund. 4:1-15.
Vahl, O. 1980. Seasonal variations in seston and in the growth rate of the
Iceland scallop. Chlamys islandica (O. F. Muller) from Balsfjord, 70
N. J. Exp. Mar. Biol. Ecol. 48:195-204.
von Bertalanffy, L. 1938. A quantitative theory of organic growth (Inqui-
ries on growth laws II) Human Biol. 10:181-213.
von Bertalanffy, L. 1964. Basic concepts in quantitative biology of me-
tabolism. Helgolander wiss. Meeresunters. 9:5-37.
Wilkinson, L. 1990. SYSTAT: The system for statistics. SYSTAT Inc.,
Evanston, IL, pp. 342-387.
Worms, J. & L. A. Davidson, 1986. The variability of Southern Gulf of
St Lawrence sea scallop meat weight-shell height relationships and its
implication for resource management. ICES CM 1986/K:24, 33p.
Journal of Shellfish Research. Vol. 13, No. 2, 565-570. 1994.
DISPERSAL AND MORTALITY OF SEA SCALLOPS, PLACOPECTEN MAGELLANICUS
(GMELIN 1791), SEEDED ON THE SEA BOTTOM OFF ILES-DE-LA-MADELEINE
G. CLICHE,' M. GIGUERE,2 AND S. VIGNEAU3
]Ministere de I Agriculture , des Pickeries el de
V Alimentation du Quebec
Direction de la recherche scientifique et technique
P.O. Box 658
Cap-aux-Meules, Quebec, Canada GOB IBO
2Department of Fisheries and ocean
Maurice Lamontagne Institute
P. O. Box 1000
Mont-Joli, Quebec. Canada G5H 3Z4
iRoche Ltee Groupe Conseil
Ste-Foy, Quebec, Canada G1W 4Y4
ABSTRACT In 1992. a total of 8,980 sea scallops were tagged and placed on a 30 x 30 m seeding site situated at the center of a
150 x 150 m sampling site. The sampling site was divided into nine squares, 50 x 50 m each. Systematic sampling was carried out
by video and diving surveys. The density of tagged scallops on the seeding site dropped over 44 days from 10 to 0.29 scallops m"2.
On the central square, including seeding site, density change was less dramatic, declining from 3.6 to 1.0 scallops m-2 after 25 days.
On eight surrounding squares, the density increased from 0 to 0. 13 scallop m " 2 between the beginning and the end of the experiment.
Scallop dispersal was extensive and rapid. Movement after 44 days was more than 60 m for 49% of the seeded scallops. Short-term
movements seemed to be preferentially toward the south. A diving survey 44 days after the seeding revealed that 12.8% of the observed
individuals on the seeding site were dead or tagged shell debris. Predation by crabs appeared to be an important cause of mortality,
because about 90% of the dead scallops had broken shells. The most probable causes of dispersal in the scallops were their high initial
density and the presence of predators.
KEY WORDS: Placopecten magellanicus. scallop. Gulf of St. Lawrence, enhancement, movement, predation
INTRODUCTION
Dispersal and predation of sea scallops, Placopecten magel-
lanicus (Gmelin). are two important factors in culturing operations
and the restocking of natural beds. Experimental seedings carried
out in 1990 and 1991 off Iles-de-la-Madeleine indicated rapid
dispersal and heavy mortality of the scallops (Picard and Vigneau,
unpublished data). Many scallop species are capable of swimming
spontaneously or in reaction to stimuli such as the presence of
predators, unsatisfactory substrate, or physical changes in the en-
vironment (e.g., salinity, pressure) (Orensanzet al. 1991). Posgay
(1981) and Melvin et al. (1985) demonstrated that sea scallop
movement for the size examined (3=50 mm) could reach around 10
km year"1. Parsons et al. (1992) obtained mean movements of
only 3.3 m for juvenile sea scallops over 3 to 4 months.
Most scallop species are not well adapted for making extensive
migrations (Brand 1991). Sinclair et al. (1985) maintained that sea
scallop aggregations along the Atlantic Coast tend to be in precise
geographical locations that are persistent over time. This is sup-
ported by the fact that the scallop beds off tles-de-la-Madeleine
have remained spatially stable, even after 20 years of fishing.
As for mortality, Scheibling et al. (1991) observed a high rate
of predation on small scallops by the sea star Asterias vulgaris.
Rock crabs (Cancer irroratus), lobsters (Homarus americanus),
and moon snails (Lunatia hews) can also be natural predators of
sea scallops (Dickies and Medcof 1963, Caddy 1973, Jamieson et
al. 1982, Orensanz et al. 1991). The objectives of this study were
to assess the importance of movements of seeded scallops in re-
lation to their size, their survival, and changes in the abundance of
predators on a spatial and temporal scale.
MATERIALS AND METHODS
The sampling site was located at the center of a small scallop
bed off Iles-de-la-Madeleine in the Gulf of St. Lawrence (Quebec,
Canada) that has been closed to fishing since 1991 (Fig. 1). The
sampling site was between 20 and 22 m deep on a uniform sub-
strate of gravel and sand. A recording current meter (Aanderan).
which also recorded temperature and sampled once every 30 min-
utes, was placed on the sampling site for the duration of the study.
The average current speed on the bottom was 3 to 5 cm s ~ ' , and
the temperature was between 9 and 13°C during the sampling
period. A progressive vector diagram simulating the path of a
particle was produced from the current meter data. The vector was
traced after low pass filtering the data to remove variations of a
factor of 100 due to diurnal and semidiurnal tides, and only the
residual current data remained. Because only one current meter
was used, the vector did not take into account spatial variations in
the current.
The sampling site covered a surface area of 22,500 m2 and was
divided into nine squares of 50 x 50 m. The seeding site was
situated at the center of the central square (#5) and was 30 x 30
m. At the corner of each square, there was a cement block with a
surface buoy attached. The attaching rope had as little slack as
possible to reduce horizontal movement of the buoy. A system of
ropes joining the buoys also aided the location of the squares from
the surface.
The 8,680 seeded scallops (=£70 mm) were acquired from
Newfoundland aquaculture operations at Port-au-Port, and 300
scallops (>70 mm) were collected from Jles-de-la-Madeleine fish-
ing grounds. Approximately 11% of the scallops were measured
with an electronic calliper before seeding. In order to distinguish
565
566
Cliche et al.
Gulf of
St Lawrence
»
1
2
3
4
. :io m -
8
f
o
r
5
7
B
9
Figure 1. Location of the sampling site and the sampling plan for the
video camera during the sea scallop seeding operation carried out in
tles-de-la-Madeleine in 1992.
the seeded scallops from the indigenous population and to follow
movements through time, the seeded scallops were marked with
tags. The plastic disc tag was attached near the hinge with cy-
anoacrylate glue. The scallops were kept in net cages (pearl nets)
for about 15 days to verify mortality associated with manipulation
and to check the resistance of the glue. Seeding took place on 4
July 1992 by means of a 10 cm diameter flexible tube from the
surface. The release of the scallops on the seeding site was con-
trolled by SCUBA divers in order to obtain a uniform distribution
(—10 scallops m-2).
A video camera deployed from the surface was used to monitor
the movements of the seeded scallops and to estimate the number
of predators present. The camera, enclosed in a waterproof case
Seeding site
Central square
Surrounding squares
20 30
Number of days after seeding
Figure 2. Density of live tagged sea scallops on the seeding site, on the
central square (#5), and on the surrounding squares (#1 to 4, #6 to
9). Densities obtained from video sampling and SCUBA diver survey.
and fixed to a support, was adjusted so that the square metal base
of the support, the sides of which measured 0.5 m. was seen.
Thus, each sampling frame was 0.25m2. Camera adjustments were
done from the surface.
The video sampling was conducted on a systematic basis (Fig.
1). Five hundred frames along 10 transects were recorded per
sampling period for the central square (#5). and 100 frames on 4
transects were recorded per sampling period for each of the sur-
rounding squares ( # 1 to 4, #6 to 9). For each sample, 5% of the
central square and 1% of the surrounding squares were sampled,
totaling 5,000 frames recorded on VHS videocassettes. During
analysis, the recordings were played back on a video tape recorder
(Panasonic VTR) with stop action. For each video frame, tagged
scallops, untagged scallops, and predators were identified and
counted; tagged scallops were identified as live or dead. Both
cluckers (empty with the two valves attached) and shell fragments
with a tag were identified as dead scallops. The shell height of
tagged scallops on the central square was measured with the mor-
phometry software Bioquant system IV (R&M Biometrics Inc),
which grabs frames from video playbacks for analysis. This
method required that scallops be completely visible and correctly
oriented toward the camera for precise measurements.
The sampling periods were often spread over several days be-
TABLE I.
Sea scallop seeding operation carried out in iles-de-la-Madeleine. Experimental procedure in 1992: type of activity, date, sampling method,
and area sampled
Activity
Date
Sampling Method
Area Sampled
Site characterization
Seeding
Sampling period 1
Sampling period 1
Sampling period 2
Sampling period 2
Sampling period 3
Sampling period 4
Sampling period 4
1 July-day — 3
4 July-day 0
7 July-day 3
8 July-day 4
13 July-day 9
17 July-day 13
29 July-day 25
17 August-day 44
1 7 August-day 44
Video camera
Video camera
Video camera
Video camera
Video camera
Video camera
Video camera
SCUBA diver
All squares
Seeding site
Central squares (#5)
Eight surrounding squares (#1 to 4, #6 to 9)
Central square
Eight surrounding squares
Central square
Eight surrounding squares
Seeding site
Dispersal and Mortality of Seeded Sea Scallops
567
cause calm seas were required for clear sampling frames. The
video data were grouped into distinct blocks from 1 July to 17
August 1992 (Table 1). In addition, on 17 August, the seeding site
was sampled by SCUBA divers and included 30 quadrats of 3 m2
each distributed systematically on three transects of 30 m. This
manual sample was done to check the video results and to evaluate
scallop mortality, in particular, tagged scallops with broken shells.
RESULTS AND DISCUSSION
t
Days 3 and 4
Days 9 and 13
Day 25
Day 44
Scallops/1. 25m2
0
I 1 1 - 4
IBBBBBBBBmH O i. y
■H > 20
V777A no data
1 seeding site
I
Figure 3. Density and distribution of live tagged sea scallops deter-
mined by contour analysis with Axum Data Software (Trimetix Inc.)
on sampling site (nine squares) at each sampling period. Data obtained
by video sampling and SCUBA diver survey.
Dispersal
The density of indigenous scallops (untagged) was estimated to
be 0.24 scallop m~2 during the study site characterization a few
days before the seeding (day - 3) and 0.42, 0.54, 0.87, and 0.69
scallop m~2, respectively for the subsequent sampling periods.
The changes in density of indigenous scallops were insignificant
and possibly due in part to lack of tag visibility (scallop turned
over or partially hidden). The density of live tagged scallops
changed over time on the seeding site, on all of the central square
(#5), and on the rest of the surrounding squares sampled (Fig. 2).
During the first 10 days, seeded scallop movements were impor-
tant but limited to the central square. At the end of the study, 44
days after seeding, the density on the seeding site decreased from
10 to 0.29 scallops m 2. In the central square, the change was less
dramatic, dropping from 3.6 to 1 .0 scallops m2 after 25 days. On
the surrounding squares (#1 to 4. #6 to 9), the density of tagged
scallops increased noticeably from 0 to 0.13 scallop m-2 between
the beginning and the end of the study. Dispersal of the tagged
scallops was outward from the seeding site toward the periphery,
and it occurred gradually with a tendency toward uniform density
over the sampling site after 44 days. At the end of the study, the
total density of scallops (tagged and untagged) over the sampling
site was approximately 1 scallop m~2, with between 0.13 and
0.29 tagged scallop m~2 and 0.69 untagged scallop m~2. Kalash-
nikov (1991) noted that, with scallops in culture (Mizuhopecten
yessoensis), high density stimulated the movement of the scallops,
which brought about reduced density. According to Orensanz et
al. (1991), scallops adjust their density toward an equilibrium
level that reduces competition for a resource in limited supply
(most likely, food) without reducing chances of cross-fertilization.
In the last sampling period, the number of tagged scallops
remaining alive on the whole sampling site was estimated to be
3,404, which was 38% of the initial number. This estimate is
North
15
Path
• Positions at 5 day intervals
10
- / \ ; \
Day 50 \\ / \
Day 10
\
5
\ /
A
0
SX
4 July
-20 -15 -10 -5 0 5
Distance (km)
Figure 4. Simulation of a path of a particle originating at the sea
scallop seeding site (0,0) on 4 July 1992, according to a temporal series
of residual current speed. North indicates magnetic north.
568
Cliche et al.
based on SCUBA diver observations on the seeding site and by
video observations on the eight surrounding squares. After 44
days, only 2.9% of the seeded scallops remained alive on the
seeding site. Thus, 44 days after the seeding, 49% of the tagged
scallops were no longer within the sampling site. The strong de-
crease in the number of tagged scallops on the sampling site cannot
be attributed to tag loss. Another experiment demonstrated that the
retention rate of tags was 99.5% over 36 days. It appears that a
significant number of the seeded scallops moved more than 60 m
N = 984
N = 39
-i — ' — i — > — i — ' — i — ' — i
10 20 30 40 50 60 70 80 90 100 110
Shell height (mm)
Figure 5. Shell height distribution of live tagged sea scallops on the
central square (#5), grouped by size class of 5 mm. Scallops were
measured with electronic callipers on day 0 (before seeding) and by
morphometric image analysis for the three other sampling periods.
in 44 days. These are greater movements than those found in
Passamaquoddy Bay, which averaged 3.3 m over 4 months for
scallops =£25 mm (Parsons et al. 1992), but our scallops were
larger.
The distribution of tagged scallops on the central square
changed noticeably with each sampling period (Fig. 3). On day 3,
70% of the scallops were still concentrated on the seeding site. On
day 9, only 32% of the individuals remained and were mainly
concentrated in the southern part of the central square, suggesting
a movement in that direction. On day 25, the majority of the
scallops had left the central square, and the distribution of the
remaining scallops was fairly uniform. In the last sampling period,
tagged scallops were present over the whole sampling site. The
southern direction of scallop movements in the first few days after
seeding corresponded with the southeast geostrophic currents for
the month of July in lles-de-la-Madeleine, as described by El-Sabh
(1976). However, the 1992 current meter data provided a better
indication of the currents on a smaller scale. A progressive vector
Sea star
Rock crab
Day -3
Days 3 and 4
J&j^"
Days 9 and 13
Day 25
i
Day 44
I 1
Predators/1 .25 m2
I I 0
1 - 3
4-6
> 7
No data
Figure 6. Density and distribution, determined by contour analysis
(Axum, TriMetrix Inc.), of predators on the sampling site (nine
squares) for each sampling period. Densities obtained by video sam-
pling.
Dispersal and Mortality of Seeded Sea Scallops
569
diagram simulating the path of a particle, originating at the seeding
site in the bottom layer, indicated a northeast movement of water
after 10 days and a general northwest direction after 50 days (Fig.
4). Thus, the general direction of tagged scallop movements did
not seem to correspond with the movement of the water mass.
The shell height distributions for scallops measured during the
different sampling periods (day 3 to 25) were similar (Fig. 5).
However, they differed from the size structure obtained before
seeding; small scallops <35 mm and scallops between 50 and 70
mm appeared to be underrepresented during the sampling periods.
This is difficult to explain because the densities estimated during
the first two sampling periods were similar to the initial density.
Limits imposed by experimental conditions and techniques may
partially explain the observed difference. Among other things,
about 50% of the scallops, mainly small, were unmeasurable in the
video frames because they were difficult to distinguish properly.
In addition, live and dead scallops were not measured during the
SCUBA diver survey 44 days after seeding. It was thus impossible
to evaluate the effect of size on scallop dispersal because of the
lack of information on the size of live and dead scallops during the
sampling periods.
Mortality and Predators
On the basis of video observations, the number of dead tagged
scallops on the central square 25 days after seeding was estimated
to be 126. which is 1.4% of the original number seeded. On the
surrounding squares, the estimated number of dead tagged scallops
varied from 0 to 206 between the sampling periods, representing
between 0 and 2.3% of the original number seeded. Very few
broken shells were observed in the video samples. The diving
survey 44 days after seeding estimated 1 , 1 50 dead scallops or parts
of tagged shells on the seeding site, which is 12.8% of the total
number released. Short-term mortality of seeded scallops was
therefore quite significant, especially on the seeding site. Mortal-
ity estimates based on diving observations were much higher than
those based on video observations, suggesting that the video was
inadequate in estimating numbers of dead tagged scallops. It was
not always evident whether closed shells were empty in video
playbacks, and the shells of dead scallops were often disarticulated
and fragmented, making it difficult to detect tags.
The fluctuations in the abundance of the principal predators
(sea stars and rock crabs) on the sampling site before seeding and
during each successive sampling period were marked (Fig. 6).
Between days - 3 and 44, the predator density on the whole study
site increased from 946 to 1 ,454 sea stars 1 ,000 m ~ 2 and from 0
to 123 rock crabs 1,000 m~2. The density of rock crabs was most
likely underestimated because sampling took place during the day,
when they are less active and often bury themselves in the sub-
strate (Stehlik et al. 1991, Gendron and Cyr 1994).
Our results showed that the distribution of the sea stars was
relatively uniform, whereas the rock crabs seemed to have a
patchy distribution on the sampling site. There was no apparent
relationship between the distribution of these two predators and
that of the scallops. Although the sea stars were the most abundant
predators on the study site, they were probably not the principal
cause of scallop mortality. Diving observations revealed that 90%
of the dead tagged scallops on the seeding site had broken shells,
which is consistent with crab predation (Elner and Jamieson 1979,
Jamieson et al. 1982). Thus, it appears that crab and/or lobster
predation is significant. The main predator is probably the rock
crab because very few lobsters and spider crabs (Hyas species)
were observed on the sampling site.
As summarized by Orensanz et al. (1991). several factors may
induce scallop movement, including the presence of predators,
unsatisfactory substrate, or physical changes in the environment
(e.g., salinity, pressure). The presence of indigenous scallops on
the sampling site seems to indicate that the substrate and environ-
mental conditions were satisfactory for sea scallops. Therefore, it
appears that displacements of the seeded scallops were principally
caused by: (1) an initial density that was too high and beyond the
optimal levels for growth and reproduction; or (2) the increasing
density of predators, which triggered escape responses in scallops
(Thomas and Gruffydd 1971. Winter and Hamilton 1985). In gen-
eral, our results emphasize the major impact of predation on
seeded scallops. The tendency of scallops to disperse and reduce
their density greatly affects harvesting yields and the economic
profitability of experimental seedings. It is therefore important to
chose seeding sites and periods carefully in order to minimize
dispersal and mortality of seeded scallops.
ACKNOWLEDGMENTS
We thank Sylvie Brulotte, Marc Lanteigne, Roberta Miller,
and Marcel Roussy for their cooperation in this study. We are
particularly grateful to Dr. David Booth for the analysis of the
current meter data. We also thank Dr. Marcel Frechette and the
reviewers for their judicious recommendations.
LITERATURE CITED
Brand. A. R. 1991. Scallop ecology: distribution and behavior, pp 517-
584. In S. E. Shumway (ed.l. Scallops: Biology. Ecology and Aqua-
culture. Developments in Aquaculture and Fisheries Science, vol. 21.
Elsevier. Amsterdam.
Caddy. J. F. 1973. Underwater observations on tracks of dredges and
trawls and some effects of dredging on a scallop ground. J . Fish. Res.
Board Can. 30:173-180.
Dickies. L. M. & J. C. Medcof. 1963. Causes of mass mortalities of
scallops (Placopecten magellanicus) in the southwestern Gulf of St.
Lawrence. J. Fish. Res. Board Can. 20:452^482.
Elner. R. W. & G. S. Jamieson. 1979. Predation of sea scallops. Pla-
copecten magellanicus. by the rock crab. Cancer irroratus. and the
American lobster, Homarus americanus. J. Fish. Res. Board Can.
36:537-543.
El-Sabh, M. I. 1976. Surface circulation pattern in the Gulf of St.
Lawrence. J. Fish. Res. Board Can. 33:124-138.
Gendron. L. & C. Cyr. 1994. Distribution bathymetnque et saisonniere du
crabe commun. Cancer irroratus, au large d'Anse-a-Beaufils. Quebec.
Can. Tech. Rep. Fish. Aquat. Sci. 2014:IX + 53 p.
Jamieson. G S., H. Stone & M. Etter. 1982. Predation of sea scallops
(Placopeclen magellanicus) by lobsters {Homarus americanus) and
rock crabs (Cancer irroratus) in underwater cage enclosures. Can. J.
Fish. Aquat. Sci. 39:499-505.
Kalashnikov, V. Z. 1991. Soviet Union, pp 1057-1082. In: S. E. Shum-
way. (ed.). Scallops: Biology, Ecology and Aquaculture. Develop-
ments in Aquaculture and Fisheries Science, vol. 21. Elsevier, Am-
sterdam.
Melvin, G. D.. M. J Dadswell & R. A. Chandler. 1985. Movement of
scallops (Placopeclen magellanicus) on Georges Bank. Can. Atlant.
Fish. Sci. Adv. Comm. Research Document 85/30. 29 p.
Orensanz. J. M., A. M. Parma & O. O. Iribame. 1991. Population dy-
namics and management of natural stocks, pp 625-713. In: S. E.
570
Cliche et al.
Shumway. (ed.). Scallops: Biology, Ecology and Aquaculture. Devel-
opments in Aquaculture and Fisheries Science, vol. 21. Elsevier, Am-
sterdam.
Parsons, G. J, C. R. Warren-Perry & M. J. Dadswell. 1992. Movements
of juvenile sea scallops Placopecten magellanicus (Gmelin, 1791) in
Passamaquoddy Bay, New Brunswick. J. Shellfish Res. 11:295-297.
Posgay, J. A. 1981 . Movement of tagged sea scallops on Georges Bank.
Mar. Fish. Rev. 43:19-25.
Scheibling, R. E., B. G. Hatcher, L. Taylor & M. A. Barbeau. 1991.
Seeding trial of giant scallops {Placopecten magellanicus) in Nova
Scotia. Proceedings 8th International Pectinid Workshop (in press).
Sinclair, M., R. K. Mohn. G. Robert & D. L. Roddick. 1985. Consider-
ations for the effective management of Atlantic scallops. Can. Tech.
Rep. Fish. Aquat. Sci. 1382: 113 p.
Stehlik, L. L., C. L. MacKenzie Jr. & W. W. Morse. 1991. Distribution
and abundance of four brachyuran crabs on the northwest Atlantic
shelf. Fish. Bull. 89:473-492.
Thomas, G. E. & LI. D. Gruffydd. 1971. The types of escape reactions
elicited in the scallop Pecten maximus by selected sea-star species.
Mar. Biol. 10:87-93.
Winter, M. A. & P. V. Hamilton. 1985. Factors influencing swimming in
bay scallop, Argopecten irradians (Lamarck. 1819). J. Exp. Mar.
Biol. Ecol. 88:227-242.
Journal of Shellfish Research. Vol. 13. No. 2. 571-579. 1994.
TECHNICAL EFFICIENCY, BIOLOGICAL CONSIDERATIONS, AND MANAGEMENT AND
REGULATION OF THE SEA SCALLOP, PLACOPECTEN MAGELLANICUS (GMELIN,
1791) FISHERY
JAMES E. KIRKLEY AND WILLIAM D. DUPAUL
School of Marine Science
Virginia Institute of Marine Science
College of William and Man'
Gloucester Point, Virginia 23062
ABSTRACT Achieving social and economic efficiency in a fishery requires that production be technically efficient. Yet, technical
efficiency (TE) is rarely examined for a fishery. By the use of detailed trip-level data and information about resource conditions
obtained from routine sampling, a stochastic frontier production model relating landings to days at sea, crew size, and resource
conditions is specified and estimated for 10 Mid-Atlantic sea scallop {Placopecten magellanicus) dredge vessels. TE is shown to
depend partly on the mix of controllable inputs such as days at sea and crew size but possibly more on uncontrollable factors such as
resource conditions and biological characteristics. Last, we illustrate that two regulations recently implemented by the management
authorities should increase TE in the U.S. sea scallop fishery.
KEY WORDS: Technical efficiency, stochastic frontier, biological conditions
INTRODUCTION
Management and regulation of commercial fisheries primarily
focus on efficient resource utilization and resource conservation.
Efficient resource utilization, however, requires that production be
technically efficient or maximized, given input levels and the tech-
nology (Fare et al. 1985). That is. technical efficiency (TE) is a
necessary, but not sufficient, condition for economic and social
efficiency. Unfortunately, achieving TE in fisheries may be diffi-
cult because of varying resource and environmental conditions.
These are important determinants of landings, which unlike days
at sea and crew size, captains cannot easily change. In addition,
achieving TE may be complicated by inadequate information
about the parameters that define and influence TE.
Studies on TE in fisheries have been limited (Comitini and
Huang 1967, Hannesson 1983, Hilbom 1985). TE has most often
been analyzed in terms of landings per unit effort (LPUE) or with
respect to a TE parameter, the constant term in a multiplicative
regression model relating landings to nominal fishing effort, as in
Strand et al. (1981). TE has also been analyzed in terms of the
proportion of fish that could be harvested by a particular gear type
(e.g., a catch of 25 out of 100 possible fish yields an efficiency
estimate of 25%). Although these types of analyses provide useful
information, they, nevertheless, do not provide adequate informa-
tion about TE or the ability of a producer to produce the maximum
output (frontier) possible from a given set of inputs and production
technology.
A serious limitation of previous studies on TE is the use of a
deterministic measure of TE. A deterministic measure does not
accommodate noise, measurement error, or random shocks be-
yond the control of the production unit. A deterministic measure,
in fact, attributes all noise, random shock, and measurement error
to inefficiency in production (Fare et al. 1985). A deterministic
measure may. therefore, be a seriously biased measure of TE for
a fishery and lead to erroneous policies by fishery managers.
In this article, we illustrate a stochastic approach for estimating
and examining TE in a fishery. By the use of a panel data set
reflecting production activities for a sample of sea scallop dredge
vessels operating in the Mid-Atlantic, a stochastic production
function or frontier is specified and estimated. The estimated fron-
tier is used to calculate TE for each trip and vessel. Estimates of
TE are subsequently summarized and examined relative to input
levels — days at sea and crew size — and biological conditions —
stock abundance, meat yield, and reproductive activities. Last, TE
is examined relative to two regulations — an annual days at sea
restriction per vessel and crew size limits — recently implemented
by the New England Fishery Management Council (NEFMC), the
U.S. agency responsible for managing the fishery.
MATERIALS AND METHODS
Methods
TE is a measure of the ability of a producing unit to produce the
maximum output, given the level of inputs and the technology. A
producer that produces in the interior of a maximum output bound-
ary or frontier or requires more inputs than necessary to produce a
maximum output is technically inefficient. Consider the hypothet-
ical frontier depicted in Figure 1 . Output or landings is measured
along the vertical axis, and fishing effort or days at sea is measured
along the horizontal axis. The darkened line represents the maxi-
mum output possible or the production frontier, given the produc-
er's technology and level of inputs. A producer is said to be TE if
production is on the frontier and technically inefficient if produc-
tion is below the frontier output. A measure of TE, however, is a
relative measure in that the production performance of one pro-
ducing unit is compared with the production performance of a
best-practice input-output relationship (Squires and Tabor 1991).
TE is, therefore, measured in terms of the deviations of individual
vessels from this best-practice frontier.
It is important to recognize, however, that TE is not the same
as economic and social efficiency. Economic efficiency is a mea-
sure of the ability of a firm or industry to use inputs to produce
output, given input and output prices; that is, economic efficiency
indicates whether or not production is consistent with the input and
output levels required to minimize cost or maximize profit (Fare et
al. 1985). The mathematical product of economic or allocative
efficiency and TE provides a measure of overall efficiency (Corbo
and de Melo 1986). Social efficiency is a measure of how well
571
572
KlRKLEY AND DuPAUL
by calculating the distribution of e2 conditional on e
(Jondrow et al. 1982):
+ e,
TE
1 + E[e2 | e, + e2]
equation 2
Fishing Effort
Figure 1. Hypothetical production frontier or maximum output
boundary
firms use inputs to produce outputs, such that net benefits to so-
ciety are maximized; this is also known as the long-run competi-
tive equilibrium output (Fare et al. 1985 and 1994). This study
only considers TE.
We assume the best-practice frontier is stochastic, rather than
deterministic, and consists of two error terms. One error term is
assumed to be normally distributed with a mean of zero and a
constant variance; this is the familiar normally distributed error
term assumed in conventional regression. This error term captures
measurement error and random exogenous shocks beyond the con-
trol of the producing unit. Aigner et al. (1976 and 1977) Meeusen
and van den Broeck (1977a and b) and Green ( 1980) have shown,
however, that another error term that is one sided may be intro-
duced to represent technical inefficiency.
Following Aigner et al . ( 1 976 and 1 977 ) and Meeusen and van
den Broeck (1977a and b), the stochastic frontier production func-
tion may be written as follows:
Y{ = h(X,.X2 XN;A)u,, i = 1,2 M
equation 1
where Y, is the output of the ith vessel, X, (j = 1 N) is the
jth of N inputs, A is a vector of parameters, and u, equals the
exponential value , 2.71828, raised to the power of the sum of two
independent error terms — en and e2,.
The sum of the two independent error terms equals the distur-
bance term e,. We let e, be a symmetric, normally distributed error
term; e, is less than, equal to, or greater than zero. This error term
allows for random variation of the production function across ves-
sels and reflects statistical noise, measurement error, and exoge-
nous shocks beyond the control of the producer. The error term e2
is one sided and nonpositive and represents technical inefficiency
relative to the stochastic frontier (Jondrow et al. 1982). The tech-
nical inefficiency error term may follow a half-normal, exponen-
tial, or truncated normal distribution. In the examination of the TE
of the sea scallop fishery, we assume the half-normal distribution.
The value of the error term e2 must be less than or equal to zero
(Fare et al. 1985; Jondrow et al. 1982). If e, = 0, production lies
on the stochastic frontier and is TE (TE = 1); if e, < 0, produc-
tion lies below the frontier and is technically inefficient (0 =£ TE
=£ 1). In applied work, TE for each observation may be measured
where E is the expectations operator.
The symmetric error e, is independently and identically dis-
tributed as N(0,CTel2). For estimation purposes, the nonpositive
error e2 is assumed to be distributed as the absolute value of the
normal distribution, |N(0,ae22)| (i.e., half-normal). Use of the
absolute value of €2 requires that the + signs in equation 2 be
changed to - signs in order to calculate TE. The variance of e
equals the sum of the variances of e, and e2. The variances of the
two error terms play an important role in assessing TE. We let 8
= o-e2"/ff€l". The more 8 exceeds one in value, the more produc-
tion is dominated by technical inefficiency, and the closer 8 is to
zero in value, the greater is the likelihood that differences between
the observed and frontier output are primarily associated with ran-
dom factors beyond the control of the captain.
Estimating TE for the individual firm (TE,) requires estimation
of the error term (e2) and decomposing the error e, for the ith
observation into the individual components — 6, and e2. Jondrow
et al. (1982) suggests a decomposition method from the condi-
tional distribution of e2, given €,. Given the normal distribution of
e, and the half-normal distribution of e2, the conditional mean of
e2l, given e, for each observation i, is as follows:
E[e2,|e,] = (ffe2o-£l/a) [(f(e,8/a)/(l - F(e,8/(j))
- (6,8/ct)] equation 3
where f( ) and F( ) are the values of the standard normal density
function and the standard normal distribution function estimated at
e,8/a and 8 = o-e22/ael2. The measure of individual TE may thus
be calculated for each observation (i) as e( - E[e2l|e,]), where e, is
replaced by its estimate and 0 =s TE, =£ 1 . The expected value
indicates the TE of firm i relative to the practices of the best
fishing vessels; the closer TE, lies to 1 (0), the closer (further) the
TE of firm i lies to the best-practice frontier.
A Second-Order Flexible Functional Form Specification of
the Frontier
It is advantageous to have estimable production relationships
that place relatively few restrictions on the technology or the na-
ture of the relationship between output and inputs (Chambers
1988). Alternatively, it is desirable to have generalized functions
or flexible-functional forms so that various hypotheses about pro-
duction can be examined. A flexible form providing a second-
order numerical or differential approximation to the transform of
an arbitrary function is particularly well suited for specifying the
production technology.
The production frontier may be specified by any well-behaved
second-order function. A well-behaved second-order function re-
quires that its parameters may be chosen such that the value of the
function, its gradient, and Hessian equal the corresponding mag-
nitudes for any arbitrary second-order function evaluated at any
point. We specify a translog production frontier, which is a natural
log transform of the generalized second-order quadratic, G(X) =
a„ + 2j S, g, (X,) + (1/2) I, I} By g,(X,) gj (Xj), at the level of
the individual fishing trip for 10 mid- Atlantic sea scallop dredge
vessels operating between 1987 and 1990:
P. MAGELLAN1CVS FISHERY
573
In Y„ = ot0 + a. In DA„ + a: In L„ + a. In S„
+ 2 Pi d> + P4 DR + t, (In DA„): + t: (In L„):
;'=i,3
+ t3 (In S,,)2 + t12 In DA„ In L„ + t13 In DA„ In S„
+ t14 In L„ In S„ + ^ rk Dk In S„
k = 2.12
+ 2j 0k Dk (In Stl)2 + e„ equation 4
k = 2.12
where i and t index individual scallop vessels and trips, respec-
tively. The variables are landed scallop meat weight (Y„) by the
ith vessel on the tth fishing trip, days at sea per trip (DA„), crew
size (L„), resource stock size (S„), annual dummy variables for
1988 through 1990 (D,). dummy variable for dredge size (DR =
1 for 3.96 m dredge and zero otherwise), and dummy variables for
the months of February through December (Dk). The parameters to
TABLE 1.
Characteristics of 10 mid-Atlantic sea scallop dredge vessels.
3), pj (j = 1,2.3). Tk (k =
2 12), and 0m. The
be estimated are a, (i = 0,
1,2,3), t„ (1 = 1.2.3). Tm (m
disturbance term e„ is assumed to be composed of a normally
distributed error (e,) and a half-normal distributed error (e,).
The translog function is flexible in that it imposes few restric-
tions on the underlying relationship between landings and the fac-
tors of production (e.g., days at sea, crew size, and resource
conditions). Direct estimation of the translog, however, is often
difficult because of multicollinearity problems, particularly when
there are more than two distinct right-hand side variables. Thus,
empirical results obtained from estimates of equation 4 should be
evaluated relative to the potential problems of multicollinearity. In
addition, not all right-hand side variables used in equation 4 are
appropriate measures of factor or input usage. For example, days
at sea embodies electronics, fuel, gear, and other inputs and, thus,
is assumed to be a composite or aggregate input. Crew size is a
stock and not the flow of labor services; labor services, however,
are assumed to be proportional to crew size. Despite these limi-
tations, equation 4 offers a convenient framework for examining
TE in the mid-Atlantic sea scallop fishery. Moreover, the speci-
fication is consistent with biological and bioeconomic specifica-
tions typically used to examine the relationship between catch and
effort.
Data
Information on production activities and vessel performance
was obtained from vessel owners of 10 scallop dredge vessels
operating in the mid-Atlantic between 1987 and 1990. Settlement
sheets or trip-level financial summaries provided detailed data on
landings, days at sea, and crew size. The 10 vessels were rela-
tively homogeneous in vessel characteristics (Table 1). All vessels
were constructed between 1979 and 1987, and all were steel
hulled. Vessel size ranged from 24.4 to 27.4 m (length overall).
Three of the 10 vessels, however, had lower horsepower engines
and, therefore, pulled smaller dredges (3.96 vs. 4.57 m). All
vessels had two radars, Loran C, and plotters. The 10 vessels
made 581 trips between January 1987 and December 1990.
Information on stock size was obtained from data regularly
collected as part of a Virginia Institute of Marine Science moni-
toring program established to determine the gametogenic cycle and
resource conditions of sea scallops in the mid-Atlantic resource
area; the data collection program is discussed in detail in
Schmitzer et al. (1991) and Kirkley and DuPaul (1991). For this
study, stock abundance per vessel per trip was calculated in terms
Length Overall
Dredge Width
Vessel
Year Built
(m)
Horsepower
(m)
1
1985
27.4
620
4.57
2
1980
24.4
500
3.96
3
1979
25.9
500
3.96
4
1987
27.4
620
4.57
5
1980
25.9
500
3.96
6
1981
25.9
520
4.57
7
1984
26.5
520
4.57
8
1980
25.9
520
4.57
9
1981
26.5
520
4.57
10
1981
25.9
520
4.57
of the geometric mean of the number of baskets of scallops caught
per hour by approximately 36 vessels fishing the same area during
the same period of time and using the same dredge size; baskets of
scallops caught per hour were only for the last tow that a group of
vessels regularly make for research purposes.
There is considerable debate about the validity of using catch
per unit effort (CPUE) or LPUE to indicate stock abundance for
any species (Ricker 1940, Dickie 1955, Paloheimo and Dickie
1964, Westrheim and Foucher 1985, Pennington 1986, Richards
and Schnute 1986. Hilborn and Walters 1992). An anonymous
referee suggested that CPUE and LPUE may provide a good es-
timate of density within a scallop bed but not likely a good mea-
sure of resource abundance. The referee suggested that CPUE or
LPUE may not be an adequate indicator of stock abundance be-
cause scallops are generally sedentary and patchily distributed.
There are also several other reasons why LPUE and CPUE may
not provide valid measures of stock abundance (e.g., the func-
tional form for the short-run catch-effort model may be different
than the traditional multiplicative model or the effort may not be
properly measured). Recognizing these limitations, we, neverthe-
less, used the geometric mean of CPUE from the last tow of
several vessels fishing the same area during the same periods of
time as, at least, a crude indicator of stock abundance. Moreover,
we note that Dickie (1955, pp. 805-807) suggested that average
catch per vessel per trip may be a valid indicator of the relative
abundance of sea scallops.
In this study, all biological parameters other than stock abun-
dance were defined relative to 90 to 94 mm shell height sea scal-
lops. This size was commonly observed in our scallop monitoring
program and is a size that is fully recruited. The reproductive or
spawning cycle was defined in terms of wet gonadal weight. Meat
yields were calculated as the average weight of meats obtained
from 90 to 94 mm shell height scallops.
RESULTS
Equation 4 was estimated by maximum likelihood procedures
available in LIMDEP 6.0 (Green 1992). Most of the parameters
were statistically different than zero at the 5% level of significance
(Table 2). For the purpose of assessing the estimated stochastic
frontier model, the adjusted R2 was calculated for the ordinary
least-squares regression (R2 = 0.845). The model was further
estimated subject to several structural restrictions; specifically, the
structure of the traditional multiplicative model and the lack of
technical interactions between crew and fishing effort, crew and
574
KlRKLEY AND DuPaUL
TABLE 2.
Parameter estimates of stochastic production frontier.
Variable"
Final Model
Intercept
Days at sea
Labor
Stock abundance
1988 constant
1989 constant
1990 constant
Dredge size
Days at sea squared
Labor squared
Stock abundance squared
Days at sea * labor
Days at sea * stock abundance
Labor * stock abundance
February * stock
March * stock
April * stock
October * stock
November * stock
December * stock
February * stock squared
March * stock squared
April * stock squared
October * stock squared
November * stock squared
December * stock squared
(r(u)/o-(v)
O-2 = (T~(U) + Q-2(V)
-2.97
(2.79)b
3.18c
(0.50)
5. 00'
(2.32)
0.10
(0.68)
-0.17c
(0.05)
-0.26c
(0.04)
-0.20'
(0.05)
-0.11'
(0.03)
-0.18'"
(0.06)
-0.82
(0.51)
0.21'
(0.05)
-0.38
(0.23)
-0.12
(0.07)
0.13
(0.29)
0.22
(0.19)
0.92c
(0.24)
0.73
(0.49)
-0.24
(0.27)
-0.37'
(0.13)
0.04
(0.15)
-0.03
(0.18)
-0.65c
(0.22)
-0.46
(0.35)
0.04
(0.28)
0.02
(0.13)
-0.61'
(0.17)
1.28'"
(0.14)
0.41'
(0.02)
a All variables except intercept and dummy variables are in natural loga-
rithms
b Numbers in parentheses .ire standard errors.
c Statistically significant at the 5% level of significance.
* indicates product of variables
stock size, and effort and stock size were examined. Calculated
chi-square values were 71.12 for the multiplicative model and
12.234 for no interactions; corresponding critical chi-square val-
ues for 12 and 3 restrictions were 21.0 and 7.81, respectively.
Additional likelihood-ratio tests suggested that all monthly dummy
variables except those for February through March and October
through December could be omitted from the model (chi-square
for 10 restrictions was 17.2; the critical chi-square value was 18.3
at the 5% level of significance).
The ratio 5 = ui2/till was 1.28 and statistically significant at
the 1% level of significance (/-statistic was 8.98). Therefore, the
discrepancy between observed and frontier output was dominated
by technical inefficiency rather than by random factors beyond the
control of the captains. This result suggests that there are oppor-
tunities for expanding production and increasing TE. Alterna-
tively, sources of technical inefficiency can. at least, be identified.
Following the procedures of Jondrow et al. (1982), TE per
vessel per trip was calculated and subsequently summarized over
vessels, trips, months, and years. Estimates of TE were compared
with input levels, meat yields, and stock abundance. Unfortu-
nately, estimates of TE could not be easily compared with the
same variables used to estimate equation 4. Conventional regres-
sion in which estimates of TE would be regressed against days at
sea, crew size, and stock abundance would be biased because of
simultaneous equation bias, and the associated parameters would
be inefficient. This is because regressions in which the value of the
dependent variable is determined by the same variables against
which it is being regressed yield biased, inefficient, and inconsis-
tent parameter estimates.
The usual analysis of variance (ANOVA) also could not be
used to assess possible differences in mean TE between vessels,
months, and years. TE was nonnormally distributed and censored
at 0 and 1; therefore, the normality assumptions required for
ANOVA were violated. A tobit or Tobin's (1958) tobit, limited
dependent variable regression model could have been used to as-
sess differences in mean TE or to approximate ANOVA proce-
dures but was rejected as being unnecessary because nonparamet-
ric procedures were available. Pairwise Kruskal-Wallis or Mann-
Whitney tests were, thus, used to further examine TE.
Kruskal-Wallis tests suggested that TE per trip was equal be-
tween most vessels. These results are omitted from this article
because 45 tests were required to compare the equality of TE
among vessels. Differences in mean efficiency between some ves-
sels, however, could not be rejected. The overall Kruskal-Wallis
test results indicated that differences primarily occurred in 1988
(chi-square for 1987 to 1990 with 9 degrees of freedom equaled
14.53). No differences in TE were detected between the most
efficient vessel (#1 in Table 1; TE = 0.78), which dragged 4.57
m dredges, and the less efficient vessels (2, 3, and 5; TE#2 =
TE#3 = TE#3 = 0.74), which dragged 3.96 m dredges.
Equality of mean TE between months was tested and rejected
by Kruskal-Wallis tests (1987 to 1990. chi-square with 1 1 degrees
of freedom equaled 36.87; 1987, chi-square with 11 degrees of
freedom equaled 41.51; 1988, chi-square = 33.32; 1989, chi-
square = 27.26; 1990. chi-square = 34.65). Pairwise Kruskal-
Wallis or Mann-Whitney tests of the equality of efficiency be-
tween months did not indicate a clear consistent pattern in TE.
These results are not presented because 320 tests were conducted
(64 tests per year and over all years). The most significant differ-
ences were detected between May and September (chi-square =
25.3), May and December (chi-square = 10.14), May and No-
P. MAGELLANICUS FISHERY
575
veniber (chi-square = 20.58), June and October (chi-square =
10.14), June and November (chi-square = 8.17), and June and
December (chi-square = 8.19). Significant differences were also
found between January. March. April. May. July. August. Octo-
ber, and September (i.e., January vs. September and October vs.
September).
Nonparametric Kruskal-Wallis tests were also used to test the
equality of mean TE relative to number of days at sea and crew
size. Differences in mean TE per trip relative to number of days at
sea could not be rejected by the Kruskal-Wallis test (chi-square
with 24 degrees of freedom equaled 54.57). Additional Kruskal-
Wallis tests, however, failed to reject equality of mean efficiency
for trips between 2 and 9 days (chi-square with 7 degrees of
freedom equaled 13.03). A Kruskal-Wallis test of mean TE for
crew sizes between 6 and 15 suggested no differences in TE (chi-
square for 8 degrees of freedom equaled 6.85). The null hypoth-
esis that mean efficiencies for crew sizes of 6 to 15 were equal for
trips between 15 and 20 days, however, was rejected (chi-square
with 8 degrees of freedom equaled 17.58).
A clear, concise relationship between efficiency and biological
conditions was not depicted by the data. Alternatively, high values
of TE occurred for low and high values of resource abundance and
meat yield and during all stages of reproduction. A simple regres-
sion of monthly mean efficiency against monthly mean gonadal
weights, meat yields, and stock abundance was not significant at
the 5% level of significance (F344 = 2.68 vs. critical value of
F, 44 = 3.21). Further examination of efficiency and biological
conditions on a year-by-year basis, however, suggested a possible
relationship. Regressions of TE against stock abundance, gonad
weight, meat yield, and these variables squared were found to be
statistically significant for 1989 and 1990 (1989, F6 ,, = 7.23;
1990, F611 = 6.56).
In general, TE was found to increase as stock abundance and
meat yield increased; gonadal weight did not appear to be a sig-
nificant explanatory variable except during 1990 (Table 3). This
latter result may be caused by the possibility that gonadal weight
and meat weight may be redundant variables (meat weight changes
as gonad weight changes). Kirkley and DuPaul (1989) and
Schmitzer et al. (1991) have shown, in fact, that meat weight
changes in response to changes in spawning events. Alternatively,
changes in meat yields are consistent with gonadal weights only
during parts of a year (e.g.. Kirkley and DuPaul [1991] demon-
strated that meat yields change during spring and fall spawns but
not during other periods of the year).
DISCUSSION
In evaluating TE, it is important to recognize that TE must be
evaluated relative to the technology or the technical constraints
imposed on the underlying technology. That is, differences in
vintage of capital, vessel size, and resource conditions should be
considered in evaluating TE. TE, for example, may be equal for
low- and high-resource conditions if vessel captains use the proper
levels of inputs. Alternatively, the TE of a 3.96 m dredge may
equal the TE of a 4.57 m dredge, even though the larger dredge
realizes higher catches per trip. The critical issue is what changes
in the input mix and production strategies can be made to improve
TE in the scallop fishery?
Unfortunately, the analysis does not provide sufficient infor-
mation for making precise prescriptions to improve TE. There are
simply too many variables to consider (e.g., days at sea, crew
size, dredge size, stock abundance, meat yield, reproductive ac-
tivities, and temporal weather patterns). We can. however, offer
qualitative prescriptions on the basis of examining TE of the more
highly efficient trips relative to the least efficient trips. Alterna-
tively, we can offer a "monkey-see, monkey-do" prescription
based on examining the general pattern of TE.
Over all trips and fishing vessels, TE, in general, positively
varied with input and output levels and resource conditions (Table
4). Higher TEs were associated with higher landings, input levels,
stock abundance, and meat yields. There were, however, some
substantial differences from these patterns. For example, one of
the relatively lowest average TEs (0.38) was associated with the
highest average stock abundance (3.19); this group of trips also
had a low average number of days at sea and a high average crew
size per trip. The maximum or highest average TE (0.91 ) per trip
between 1987 and 1990 was associated with moderately few days
(10.6 days) at sea per trip, the maximum average crew size (11.4
TABLE 3.
Statistical results of regressing TE on gonad weight, meat yield, and stock abundance, 1987 to 1990
Parameter Estimates"
Meat
Meat
Gonad
Gonad
Stock
Stock2
Yield
Yield2
Weight
Weight2
Constant
(baskets per hour)
(g)
(g)
(g)
R2
1987
8.37446
0.18789
-0.0196
-1.0957
0.0371
0.00056
0.0015
0.74
[1.96]
[2.81]
[1.31]
[1.78]
[2.77]
[0.008]
[0.23]
1988
4.4022
-0.38831
0.0751
-0.4896
0.0195
-0.05625
0.0074
0.47
[1.70]
[1.60]
[1.65]
[1.21]
11.23]
[0.82]
[0.85]
1989
-4.4842
0.06073
-0.0087
0.8876
-0.0371
-0.09342
0.0143
0.89
[2.04]
[0.31]
[0.24]
[3.14]
[4.17]
[0.72]
[0.98]
1990
-6.4098
-0.27606
0.0497
1.0967
-0.3873
-0.10038
0.0109
0.89
[2.84]
[2.50]
[2.19]
[3.31]
13.12]
[2.81]
[2.89]
Numbers in brackets are (-statistics.
576
KlRKLEY AND DuPAUL
TABLE 4.
TE and average catch per trip, days at sea per trip, crew size, stock abundance, and weight of individual scallops, 1987 to 1990.
Yield of
Efficiency
Catch
per Trip
(mt)
Days at Sea
per Trip
Crew Size
(No. of People)
Stock
Abundance
(Baskets per Hour)
90 to 94 mm
Scallops
Gonad Weight
90 to 94 mm
(g)
0.90-0.94
4.80
10.60
11.40
3.08
14.51
3.57
[5]a
0.85-0.89
4.77
14.30
9.50
2.15
13.61
4.31
[54]
0.80-0.84
4.83
15.69
9.46
2.96
13.63
3.93
[151]
0.75-0.79
4.43
15.78
9.66
3.03
13.49
3.91
[165]
0.70-0.74
4.23
16.28
9.87
3.07
13.44
3.53
[90]
0.65-0.69
3.36
16.85
9.50
2.66
13.35
3.46
[40]
0.60-0.64
2.63
14.43
10.48
2.69
12.78
2.94
[23]
0.55-0.59
2.34
15.47
9.42
2.04
13.13
4.00
[19]
0.50-0.54
1.92
14.57
9.29
1.89
12.96
4.13
[V]
0.40-0.49
1.31
12.46
8.77
2.11
12.78
4.02
[13]
0.30-0.39
2.05
13.33
10.33
3.19
13.20
4.50
[3]
0.10-0.29
0.58
9.80
8.88
1.95
12.93
3.75
[5]
0.00-0.09
0.33
9.17
9.83
2.52
12.68
3.30
[6]
Numbers in brackets indicate number of observations.
individuals), relatively high stock abundance (3.08), and the high-
est average meat yield (14.51 g).
Further examination of TE by days at sea and crew size group-
ings offered some possible prescriptions for increasing TE in the
sea scallop fishery (Table 5). Higher TEs given relatively low-
resource (=£2.00 baskets per hour) conditions were associated with
crew sizes of fewer than 8 individuals; larger crew sizes for low-
resource conditions were generally associated with inefficient
trips. For higher stock abundances (e.g., S3. 00 baskets per hour),
high TEs per trip were generally associated with crew sizes of 1 1
or more individuals. Over all observations, maximum average TEs
generally occurred for trips between 14 and 22 days. Inefficient
trips between 14 and 22 days per trip were generally associated
with low abundance (=s2.00 baskets per hour) and large crew sizes
(&9 individuals).
A consistent linear association between efficiency, input lev-
els, and resource conditions, however, was not evident from the
analysis. Neither Pearson nor Spearman rank correlation coeffi-
cients indicated a linear relationship. The lack of a consistent
linear pattern suggests that captain's skill and unknown biological
and environmental factors are likely to be important determinants
of TE in the sea scallop fishery. Alternatively, the relationship
between TE, input levels, and resource conditions may be highly
nonlinear because of technical, environmental, and biological con-
straints. For example, a large number of days at sea may be in-
efficient because of inadequate resource conditions or too few
workers, or efficient production for trips with few days requires a
large number of workers.
The most discernable pattern between efficiency, catch, days at
sea, number of crew, and resource conditions was exhibited on a
monthly basis (Table 6). Average TE closely followed vessel per-
formance and resource conditions (i.e., high stock abundance and
high TE and low abundance and low efficiency). Maximum effi-
ciency occurred in June, when scallops have just completed their
spring spawn (Schmitzer et al. 1991). TE was also high between
March and May and in August when meat yields are high. TE
declined in July when meat yields also are high; the decline may
reflect a product wafering problem, which typically occurs in July
because of high air temperatures (DuPaul et al. 1990).
Captain's skill in determining labor needs and other levels of
inputs may be responsible for obtaining maximum efficiency in
June. Knowledgeable captains recognize that stock abundance is
typically high and relatively constant in June, and individual meat
yields are predictably low relative to many other periods of the
year. Captains can. therefore, more easily predict their labor needs
and make changes in their input mix (e.g., days at sea and crew
size). In late summer and early fall, resource abundance and
spawning appears to be erratic; captains cannot, therefore, easily
predict these conditions and change crew size.
Minimum TE primarily occurred between September and Feb-
ruary. Low TE in September may have been associated with
weather. This is a period of intense storms in the mid-Atlantic
P. MAGELLANICUS FISHERY
577
TABLE 5.
TE and biological characteristics, and input utilization per trip,
1987 to 1990.
Days
at
Meat Weight
Stock Abundance
TE
(g)
Baskets per
Hour
sea
«8
9-10
5=11
s8
9-10
sll
«8
9-10
sll
5
0.80
0.71
0.77
13.8
13.7
15.0
1.66
2.61
2.81
6
0.89
0.72
0.83
9.6
13.6
14.2
1.12
2.36
4.08
7
0.72
0.74
0.74
14 4
12.9
14.2
1.44
1.92
3.16
8
0.00
0.75
0.75
14.5
13.9
14.7
2.54
1.89
3.04
9
0.75
0.71
0.63
12.6
12.9
12.9
1.56
2.62
3.13
10
0.84
0.72
0.79
13.6
13.5
15.0
2.89
2.76
3.66
11
0.46
0.70
0.68
13.2
12.9
14.1
1.44
2.41
4.03
12
0.68
0.78
0.76
12.6
13.1
14.7
1.38
2.63
3.50
13
0.60
0.76
0.54
12.9
13.0
13.7
1.65
2.54
3.29
14
0.89
0.74
0.71
10.5
12.3
13.1
1.80
2.28
2.70
15
0.79
0.78
0.78
12.7
13.3
14.6
1.81
2.24
3.93
16
0.78
0.77
0.66
11.3
13.0
12.7
1.64
2.55
3.11
17
0.77
0.79
0.76
12.6
13.4
14.8
2.12
2.64
4.00
18
0.79
0.77
0.80
12.9
13.5
15.0
2.09
2.83
4.06
19
0.76
0.76
0.76
12.5
13.2
14.3
2.15
2.93
4.00
20
0.76
0.75
0.74
13.0
13.8
13.7
2.62
3.16
3.19
21
0.77
0.74
0.84
12.6
13.8
13.0
2.39
2.73
3.44
22
0.80
0.78
NAa
10.5
14 1
NA
2.11
3.13
NA
23
0.00
0.56
0.50
NA
11.7
9.6
NA
2.01
1.17
24
0.74
0.71
0.75
11.7
15.0
13.4
2.76
2.35
3.98
<14
0.66
0.72
0.74
12.9
13.2
14.2
1.66
2.50
3.27
14-22
0.77
0.77
0.76
12.6
13.4
14.2
2.20
2.84
3.77
>22
0.70
0.67
0.62
13.2
13.6
11.5
2.98
2.28
2.58
Note: results are summarized relative to three groups of crew size (=s8,
9-10, and si 1 individuals).
a NA indicates no observations for days at sea and crew size group.
resource area, and vessels must often ride out the storms at sea or
cut trips short and return to port. Low efficiency between October
and November likely was associated with reproductive activities
and reduced meat yields. The fall spawn, although erratic in the
mid-Atlantic resource area, typically occurs between October and
November (DuPaul et al. (1989) Kirkley and DuPaul 1991). The
minimum average efficiency of December was likely associated
with low abundance and meat yields.
Low efficiencies in September. January, and February, how-
ever, also may have been caused by poor judgment about crew size
by captains. Stock abundance during these 3 months was relatively
low but highly variable, as indicated by the coefficient of varia-
tion, whereas changes in crew size were infrequent. If captains fail
to change crew size in response to changes in resource conditions,
TE might be low because of too much or too little labor. In fact,
higher mean TEs per trip in each month were associated with
lower average crew sizes. For example, mean crew size for TE <
0.8 = 9.55 and mean crew size for TE ^ 0.8 = 8.91 in February;
differences were similar for September and January. It is, thus,
quite probable that technical inefficiency during September, Feb-
ruary, and January is at least partly caused by excess labor.
In comparison, resource abundance and crew size were both
highly variable in December, but the monthly mean TE per trip
was the lowest relative to all other months. The coefficients of
variation for crew size and stock abundance indicated that captains
changed crew size in response to changes in resource conditions.
Low efficiencies during December, however, were also associated
with high crew sizes (e.g.. mean crew size for TE < 0.8 = 9.56,
whereas mean crew size for TE s= 0.8 = 7.96). It is, thus, likely
that captains misjudge their labor needs and use too much labor in
December. Alternatively, it may be possible that vessel operating
requirements cause excess labor relative to resource conditions in
December (e.g.. safe operation of a scallop vessel requires seven
to nine individuals, and eight or more crew may be excessive,
given resource abundance).
Short of curtailing fishing activities between September and
February, what prescriptions does the analysis offer for improving
TE in the mid-Atlantic sea-scallop. P. magellanicus, fishery? In
general, vessel captains can consistently achieve high TE by tak-
ing trips between 14 and 22 days. High TE may also be achieved
with shorter trips, but economic conditions seriously limit net
returns for short trips; maximum profit occurs for trips between 15
and 22 days (Kirkley and DuPaul 1992). When resource abun-
dance is relatively low, crew size should be restricted to eight or
fewer individuals; when abundance is relatively high, crew size
should be sufficient to provide the labor services of nine or more
individuals. In essence, the best qualitative prescription for im-
proving TE is for captains to become more aware of changes in
resource and environmental conditions and accordingly change
input levels.
Thus far, analyses indicate that TE is affected by controllable
factors of production and relatively uncontrollable, but not neces-
sarily unpredictable, biological and environmental conditions.
Relatively high levels of TE can be maintained over a wide range
of resource conditions and input and output levels. Managers de-
signing policies to promote efficient resource conservation, there-
fore, need to consider the potential interaction between TE, re-
source conditions, known temporal patterns in variables that might
affect efficiency, and input levels.
The NEFMC has required that crew size be restricted to nine
individuals to limit total harvest and prevent excess harvesting of
small scallops. NEFMC has also proposed restrictions on days at
sea per year per vessel to limit total production and size of scal-
lops. Assessing the likelihood of these regulations to accomplish
their goals will require considerable detailed knowledge of TE and
the importance of biological and economic conditions. For exam-
ple, what are the interactions between resource conditions and the
proposed regulations? Will TE be increased because of the regu-
lations?
By the use of information obtained from eight at-sea experi-
ments in which crew sizes or shucking capacities required for
given harvest levels were estimated, it was determined that pro-
duction levels for only 63 of 581 trips between 1987 and 1990
would have been affected by restrictions on crew size. Crew,
however, would have to work more hours per day (i.e., increase
the number of hours per individual from 12 up to 16 hours).
Relative to production during 1990, a restriction on crew size
would have reduced landings for only 4 of 132 trips. A nine-man
crew limit would, therefore, only marginally restrict total catch
and reduce fishing mortality relative to observed levels; it would,
however, improve TE in the mid-Atlantic sea scallop fishery.
On the basis of an assumed proportional relationship between
fishing mortality and fishing effort (days at sea), the NEFMC
restricted the annual number of days at sea a vessel may fish.
Analyses in Kirkley and DuPaul ( 1992) have shown that an annual
restriction on days at sea, depending on the total number of days
allowed, will likely cause vessels to stop fishing between October
and January, when economic returns are low. This also corre-
578
KlRKLEY AND DuPaUL
TABLE 6.
Average monthly technical efficiency, 1987 to 1990.
Weight of
90 to 94 mm
Gonad Weight
Stock
Sea Scallops
90 to 95 mm
Month
Efficiency
Effort
Crew Size
Abundance
(g)
(g)
January
0.74
14.57
9.03
1.73
13.22
3.98
[21.4]b
[31.8]
[9.4]
[34.7]
[6.5]
[21.0]
February
0.74
12.89
9.41
2.24
13.77
5.86
[18.3]
[48.9]
[10.8]
[37.6]
[5.1]
[12.1]
March
0.76
14.53
10.04
2.52
14.46
6.79
[15.2]
[36.8]
[10.7]
[35.7]
[6.2]
[6.8]
April
0.77
17.02
10.71
3.54
14.13
6.07
[7.5]
[27.2]
[15.4]
[24.8]
[4.9]
[24.6]
May
0.77
15.57
10.24
3.77
13.85
3.76
[20.5]
[33.8]
[13.1]
[33.0]
[9.7]
[17.2]
June
0.79
16.71
9.92
3.78
13.09
2.43
[10.3]
[23.6]
[14.5]
[28.3]
[7.0]
[29.1]
July
0.75
15.25
10.06
3.53
13.97
2.47
[17.4]
[35.5]
[16.6]
[33.8]
[6.1]
[17.1]
August
0.77
16.23
9.80
3.29
14.09
2.31
[9.6]
[29.9]
[13.3]
[34.3]
[8.5]
[11.1]
September
0.72
16.02
9.84
2.94
13.88
3.08
[12.0]
[35.6]
[10.7]
[36.7]
[10.0]
[13.1]
October
0.74
15.70
9.05
2.41
12.89
3.96
[19.4]
[29.8]
[15.4]
[35.4]
[14.0]
[27.6]
November
0.71
15.91
8.89
2.07
12.19
2.59
[20.0]
[31.3]
[13.8]
[37.8]
[9.3]
[13.2]
December
0.69
14.60
8.56
1.67
11.76
2.41
[25.7]
[35. 0J
[16.4]
[35.7]
[11.7]
[20.3]
Note: effort is the number of days at sea per trip, crew size is the number of individuals aboard the vessel per trip, and stock abundance is number of
baskets per hour.
a Numbers in brackets are coefficients of variation.
sponds to the time period when scallops spawn, resource abun-
dance and meats yields are low, and TE is minimum. Given that
vessels curtail fishing activities during this period, average TE per
trip can be expected to increase in response to a restriction on days
at sea per vessel per year. Average TE over all 10 vessels and
months between 1987 and 1990 was 0.74. Mean TE per trip in-
creased to 0.77 or by approximately 4.1% when fishing activity
was restricted to February through September. A restriction on the
number of days at sea per vessel should, therefore, increase TE
while overall harvest levels decline and spawning occurs.
This study demonstrated that TE in the mid-Atlantic sea scallop
dredge fishery from 1987 through 1990 was maintained over a
wide range of input levels and resource conditions. Production was
most efficient for trips of 14 to 22 days and made between March
and August. Minimum efficiency occurred for trips in excess of 22
days, or in general, less than 14 days, and taken between Septem-
ber and February. Inadequate input levels, given resource condi-
tions or incorrect decisions by captains about input levels, were
determined to contribute to technical inefficiency. It was also de-
termined, however, that efficiency more closely followed the tem-
poral patterns of resource abundance, meat yields, and the repro-
ductive cycle. Gains in TE are, thus, highly likely if captains can
correctly match input levels to resource conditions.
Efficacious regulatory policy will have to consider the effects
on input and output levels and TE, as well as the relationship
between TE, input levels, and resource and environmental condi-
tions. Failure to recognize these linkages could result in inade-
quate regulatory policies, particularly those policies designed to
promote social and economic efficiency. Relative to the restric-
tions on days at sea and crew size, TE should increase while
harvest levels decline.
A comprehensive evaluation of TE in the sea scallop fishery
must consider stock conditions, input levels, vessel characteris-
tics, and managerial ability. This study did not address managerial
ability, and thus, it is possible that levels of TE may have been
inappropriately attributed to biological and economic conditions.
Alternatively, estimates of TE may be biased because of the omis-
sion of captain's skill from the analysis. Managers concerned with
TE in the sea scallop fishery must be cognizant of changes in TE,
economic performance, and biological and environmental condi-
tions when determining regulations. Although TE is necessary for
social efficiency, it is not a sufficient condition. Evaluation of TE
in the sea scallop fishery and other fisheries will require consid-
erably more information than is usually available. Additional
information on vessel captain, gear, vessel characteristics, envi-
ronmental conditions, food abundance, and the gametogenic cy-
cle must be obtained to better understand TE in the sea scallop
fishery.
ACKNOWLEDGMENTS
Partial funding for this research was provided by Virginia Sea
Grant. We are grateful to the vessel owners that provided detailed
information on production activities. This is VIMS Contribution
No. 1903.
P. MAGELLANICUS FISHERY
579
LITERATURE CITED
Aigner. D. J., T. Amemiya & D. J. Poner 1976. On the estimation of
production frontiers: Maximum likelihood estimation of the parameters
of a discontinuous density function Inter. Econ. Rev. 17:377-396.
Aigner. D. J.. C. A. K. Lovell & P. Schmidt. 1977. Formulation and
estimation of stochastic frontier production function models. J. Econo-
metrics 6:21-37.
Chambers. R. G. 1988. Applied Production Analysis: A Dual Approach.
Cambridge University Press, New York.
Comitini. S. & D. S. Huang. 1967. A study of production and factor
shares in the halibut fishing industry. J. Political Econ. 75:366-372.
Corbo. V. & J. de Melo. 1986. Measuring technical efficiency; a com-
parison of alternative methodologies with census data. pp. 1-39. In: A.
Dogramaci. (ed.). Measurement issues and behavior of productivity
variables. Kluwer Nijhoff Publishing. Boston.
Dickie. L. M. 1955. Fluctuations in abundance of the giant scallop. Pla-
copecten magellanicus (Gmelin), in the Digby area of the Bay of
Fundy. J. Fish. Res. Board Can. 12:797-854.
DuPaul. W D . J E. Kirkley & A. C. Schmitzer. 1989. Evidence of a
semiannual reproductive cycle for the sea scallop {Placopecten magel-
lanicus) (Gmelin) in the mid-Atlantic region. J. Shellfish Res. 8:173-
178.
DuPaul. W. D . R. A. Fisher & J. E. Kirkley. 1990. An evaluation of
at-sea handling practices:effects on sea scallop meat quality, volume.
and integrity. Virginia Institute of Marine Science. VIMS Marine Re-
source Report. 76 pp.
Fare. R.. S. Grosskopf & C. A. K. Lovell. 1985. The Measurement of
Efficiency of Production. Kluwer-Nijhoff Publishing. Boston.
Fare. R . S Grosskopf & C. A. K. Lovell. 1994. Production Frontiers.
Cambridge University Press, New York.
Green, W. 1980. Maximum likelihood estimation of econometric frontier
functions. J. Econometrics 13:27-56.
Green. W. 1992. Limdep. Version 6.0. User's Manual and Reference
Guide. Econometric Software. Inc., Bellport, NY.
Hannesson. R. 1983. Bioeconomic production function in fisheries: theo-
retical and empirical analysis. Can. J. Fish. Aquat. Sci. 40:968-982.
Hilborn. R. 1985. Fleet dynamics and individual variation: Why some
people catch more fish than others. Can J. Fish. Aquat. Sci. 42:2-13.
Hilborn. R. & C. J. Walters. 1992. Quantitative fisheries stock assess-
ment: choice, dynamics, and uncertainty. Chapman and Hall. New
York.
Jondrow. J., K. Lovell, I. Materov & P. Schmidt. 1982. On the estimation
of technical inefficiency in the stochastic frontier production model. J
Econometrics 19:233-238.
Kirkley. J. E & W. D. DuPaul. 1989. Science, commercial practices, and
the determination of fishery regulations. J Shellfish Res. 8:139—149.
Kirkley. J. E. & W. D. DuPaul. 1991. Temporal variations in spawning
behavior of sea scallops (Placopecten magellanicus) (Gmelin, 1791 ) in
the Mid-Atlantic resource area. J. Shellfish Res. 10:389-394.
Kirkley. J. E. & W. D. DuPaul. 1992. Preliminary assessment of the
economic impacts of various catch and effort restrictions on the U.S.
sea scallop fishery. Virginia Institute of Marine Science. VIMS Marine
Resource Report. 28 pp.
Meeusen, W. & van den Broeck. 1977a. Efficiency estimation from Cobb-
Douglas production functions with composed error. Inter. Econ. Rev.
18:435-444.
Meeusen. W. & van den Broeck. 1977b. Technical efficiency and dimen-
sion of the firm: some results on the use of frontier production func-
tions. Emp. Econ. 2:109-122.
Paloheimo, J. E. & L. M. Dickie. 1964. Abundance and fishing success.
Rapp. P.-V. Reun. Cons. Int. Explor. Mer 155:152-163.
Pennington, M. 1986. Some statistical techniques for estimating abun-
dance indices from trawl surveys. Fish. Bull. 3:519-525.
Richards, L. J. & J. T. Schnute. 1986. An experimental and statistical
approach to the question: is CPUE an index of abundance? Can. J.
Fish. Aquat. Sci. 43:1214-1227.
Ricker, W. E. 1940. Relation of "catch per unit effort" to abundance and
rate of exploitation. J. Fish. Res. Board Can. 5:43-70
Schmitzer. A. C, W. D. DuPaul & J. W. Kirkley. 1991. Gametogemc
cycle of sea scallops (Placopecten magellanicus) (Gmelin 17911 in the
Mid-Atlantic Bight. J. Shellfish Res. 10:221-228.
Squires, D. E. & S. Tabor. 1991. Technical efficiency and future produc-
tion gains in Indonesian agriculture. Dew Econ. 29:258-270.
Strand, I. E., J. E. Kirkley & K. E. McConnell 1981 Economic analysis
and the management of Atlantic surf clams. In: L. G. Anderson, (ed.).
Economic analysis for fishery management plans. Ann Arbor Science.
Ann Arbor.
Tobin. J. 1958. Estimation of relationships for limited dependent vari-
ables. Econometrica 26:24—36.
Westrheim. S. J. & R. P. Foucher. 1985. Relative fishing power for Ca-
nadian trawlers landing Pacific cod (Gadus macrocephalus) and im-
portant shelf cohabitants from major offshore areas of western Canada,
1960-81. Can J. Fish. Aquat. Sci. 42:1614-1626.
Journal of Shellfish Research, Vol. 13, No. 2. 581-585. 1994.
AGE, GROWTH RATE, SEXUAL DIMORPHISM AND FECUNDITY OF KNOBBED WHELK
BUSYCON CARICA (GMELIN, 1791) IN A WESTERN MID-ATLANTIC LAGOON
SYSTEM, VIRGINIA
MICHAEL CASTAGNA' AND JOHN N. KRAEUTER2
1 Virginia Institute of Marine Science
College of William and Mary
Wachapreague. VA 23480
Haskin Shellfish Research Laboratory
Institute of Marine and Coastal Studies
Cook College, Rutgers University
Port N orris, NJ
ABSTRACT Growth, onset of sexual maturity, and sexual reversal in laboratory-reared Busycon carica have been examined.
Animals first matured at 9 years of age. The first sign of maturity in all animals we reared was the presence of the penis. At 12.4 years
of age, one of the animals laid an egg case that did not contain embryos. This animal, and all others, still retained a penis. At 13.5
years, three egg cases were laid and over half the animals had undergone sex reversal (loss of the penis). Field studies have shown
that egg strings are laid in the fall on tidal and intertidal flats and over winter to hatch in the spring. Organisms that require a relatively
long time to mature, that lay few eggs per spawning season, and that are vulnerable for a long time are difficult to manage for a
sustained yield fishery.
KEY WORDS: Busycon carica, whelk, age. growth, sex reversal, sexual dimorphism, fecundity
INTRODUCTION
Busycon carica is a large, predacious gastropod that is com-
mercially harvested along the east coast of the United States and
marketed as conchs or, more properly, as whelks. Over 200,000
lbs were landed in Virginia in 1986. The frozen meats are used
in salads and chowders or sold in ethnic markets as squingelli
(DiCosimo 1986, Kaplan and Boyer 1992).
There are few published data on growth rates, age of onset of
sexual dimorphism, or sex ratios of busyconine whelks (Frank
1969, Powell and Cummings 1985, Kraeuter et al. 1989), and no
information on the age at which females become sexually mature.
Although there are relatively large numbers of studies on epifaunal
snails, other than growth rates, little information is available on
size at age or the age of sexual maturity of long-lived predaceous
infaunal gastropods. Growth information is available for Polinices
duplicatus (Say 1822) (Edwards and Huebner 1977). Gendron
(1992) reported growth rate and the size of sexually mature Buc-
cinum undatum (Linne 1758), whereas Santarelli and Gros (1985)
examined age structure in B. undatum on the basis of opercular
striae. Heller (1990) compared longevity throughout the entire
mollusca phylum to reveal common patterns of reproduction. This
was based on data gleaned from the existing literature on the life
durations of 547 species from marine, freshwater, or terrestrial
habitats. Gastropods are the second most long-lived of mollusks,
after bivalves. Short-lived mode of life is often correlated with:
lack of external shell or an external shell that is semitransparent.
dwelling in a harsh microenvironment that is exposed to high solar
radiation and high temperatures, dwelling in an environment in
which reproduction occurs at least once a year, and very minute
size. Powell and Cummings ( 1985) compiled data on longevity of
bivalves and gastropods and found that a higher-than-average
number of long life spans coincide with periods of long-term cy-
cles in marine communities. These cycles could affect longevity in
one of two ways: cyclic phenomena that produce environmental
changes beyond the species tolerance limit or cycles that might
affect recruitment success and thereby exert selective pressure for
longevities longer than the cycle affecting recruitment. Interest in
culture of queen conch, Strombus gigas (Linne 1758), has pro-
vided substantial information on the growth, size-specific mortal-
ity, and ecology of that species (Wefer and Killingley 1980, Ap-
peldorn 1988).
Magalhaes (1948) studied B. carica growth rate in the field
near Beaufort, North Carolina, and Sisson (1972) measured
growth of Busycotypus {Busycon) canaliculars (Linne 1758) in
Narragansett Bay, Rhode Island. The general ecology of whelks is
best known from Magalhaes (1948), Peterson (1982) in North
Carolina, and from Menzel and Nichy (1958), Paine (1962 and
1963), and subsequent studies by Kent (1983) in Florida. Davis
and Sisson (1988) have increased the available information on
whelks in Rhode Island, Massachusetts, and Georgia (Walker
1988). The work of Magalhaes ( 1948) remains the most complete
ecological study of B. carica to date. She reports egg laying in
Beaufort, North Carolina, from May to June and again from Sep-
tember to November. Numbers of egg capsules per string ranged
from 9 to 156 (mean, 80), and the total numbers of egg per string
ranged from 4.000 to 6.000 (Magalhaes, 1948). Ram ( 1977) found
that an extract from the nervous system would cause mature ani-
mals to lay egg capsules.
We report on the continuation of a 14 year study of growth
rates in which B. carica were reared in the laboratory from hatch-
ing to sexual maturity to egg laying. Kraeuter et al. (1989) exam-
ined the growth rate of a B. carica population in Virginia, using
three methods: ( 1 ) measurement of individuals marked and recap-
tured in the field; (2) examination of growth lines in the opercu-
lum; and (3) measurement of laboratory-reared individuals. Most
whelks that were marked and recaptured were larger than 170 mm.
The smallest individual tagged and recaptured grew from 138 to
151 mm (0.098 mm/day"'). Growth rates for males were not
calculated because too few were obtained to make accurate esti-
mates. Most of the males were in the 170 to 209 mm size classes.
Two series of marked individuals were released into the field.
581
582
Castagna and Kraeuter
Males were 7.8% (N = 23) of the 190 individuals in the first
group and 9.0% (N = 167) of the 1 ,859 individuals of the second
group. No indications of sex reversal were noted in the field stud-
ies. Information on numbers of eggs per capsule, numbers of
capsules per string, seasonal changes in gonad and nidamental
gland precursor (precapsulin; see Goldsmith et al. 1978), and time
of hatching for a Virginia population of B. carica is presented in
this article.
METHODS
Sexual Maturity and Growth at Age
B. carica were hatched from egg cases collected from an in-
tertidal sand flat behind Cedar Island, Virginia, in the winter of
1976 to 1977. Nine egg strings were returned to the Virginia
Institute of Marine Science Laboratory in Wachapreague, Vir-
ginia, and maintained in a running seawater system. Seawater is
drawn from a nearby channel where salinity ranges from 25 to 32
ppt, with short excursions below 20 ppt after extreme rain storms.
Hatching from the strings was complete by May 1977. Attempts to
keep newly hatched whelks in glass dishes, plastic trays, or shal-
low trays filled with sand from the intertidal flat failed because the
animals continually climbed out of the water and desiccated. Some
individuals were hatched in a 2.4 x 0.6 m fiberglass flowing
seawater tray in which a miniature sand beach had been made.
Some newly hatched individuals from several egg strings were
reared inside a polypropylene filter bag that received flowing am-
bient seawater. When the animals reached approximately 20 mm
shell length, they were transferred to the running seawater trays
with beach sand substrate and supplied with small, live juvenile
clams (Mulinia lateralis, Mercenaria mercenaria. or Mya are-
naria). As the B. carica grew, larger size clams were furnished.
M. mercenaria was the most commonly fed species and was used
exclusively after the third year. Dietary supplements may have
entered with the flowing seawater.
Length measurements were made on individual whelks re-
moved from the substrate five times during the first year, then two
or three times in subsequent years. Measurements were made on a
minimum of 25 individuals per sampling period, until 1985, when
the number was reduced to 20.
Initially, over 2,000 individuals were maintained but these
were reduced to 26 1 through attrition and sacrifice for other stud-
ies by the fall of 1977. Mortality had reduced the total numbers
being maintained to 93 after 5 years (February 1982). In the fall of
1989, the remaining whelks were divided into groups of six and
placed in four sand-filled 0.6 x 0.6 m boxes with a standpipe
supplied with ambient flowing seawater. The organisms were
maintained in these boxes until April 1990. when a number of
individuals had developed a boring sponge (Cliona) infestation in
their shells. All individuals were given a 10 second dip in a sat-
urated salt solution and air dried for 1 hour before being returned
to flowing seawater. These same individuals were used for the
sexual maturation studies.
The cultured whelks were sexed by observing the presence or
absence of a penis. The observations were made in three ways.
Animals were placed in a shallow plastic tray containing 1 to 3 cm
of flowing ambient seawater until they were firmly affixed to the
bottom (about 1 hour) The penis, a C-shaped organ, could be
observed by gently tilting the shell clockwise. In another method,
the whelks' shells were attached dorsal side down and held in
place with a small ball of clay or with malleable lead wire. The
tray was left in a partially darkened room for 0.5 to 2 hours.
allowing time for the animals to attempt to right themselves by
extending their foot. The extension of the foot exposed the penal
area. A third method was sometimes used for older intractable
specimens. With a 1 ml syringe and a 23 gauge needle, 0.4 ml of
2 mM serotonin (5-hydroxytryptamine, creatinine sulfate com-
plex; Sigma Chemical Co., St. Louis, MO) was injected into the
foot just posterior to the operculum. The animals reacted within 2
hours by extending the foot and rolling the operculum away from
the aperture, exposing the phallus.
Gonadal Analysis
About 10 specimens were randomly collected in 1976 to 1977
at approximately 1 1 monthly intervals (2 N = 105) from the study
site behind Cedar Island, returned to the laboratory, and frozen.
Shell length and width were measured before the removal of the
flesh. The animal was extracted from the shell, sexed (presence or
absence of eggs or a penis), and weighed to the nearest 0.1 g. The
tissue was dissected into four fractions: meat, viscera, nidamental
gland (females only), and gonad, and each was weighed to the
nearest 0.1 g.
Egg Strings and Egg Capsules
Several studies examined the timing of egg laying, numbers of
capsules per egg string, and time and number of egg hatching. Egg
laying and paired individuals (copulation) were recorded during
mark recapture field studies (approximately monthly during 1976
and 1977). Numbers of egg capsules per egg string were recorded
in the field, the location was noted, and the string was marked.
These were followed on subsequent visits until hatching was re-
corded.
To determine hatching success, the shoreline was examined for
old egg strings in the fall and winter of 1976 and these were
removed. Any egg string that was found on the beach in the spring
(late March to May 1977) was returned to the laboratory, and
numbers of juveniles per capsule were determined. These data
formed the basis for an estimate of the numbers of hatchlings per
string and the percent mortality. For these estimates, the total
number of eggs per string was estimated on the basis of the largest
numbers of juveniles found in capsules from the midportion of the
egg string.
A number of strings were returned to the laboratory for more
detailed study of the number of embryos per capsule and the num-
ber of capsules per string. The number of capsules before the first
egg-bearing capsule (anchoring part of string) was determined
from these strings. Capsules were randomly removed from the
length of the string and measured (height, diameter, and volume).
The number of embryos per capsule was determined by removing
the top of the capsule and counting the eggs or developing em-
bryos.
Hatching studies were performed on capsules removed from
the same strings as those used to determine the distribution of the
number of embryos per capsule. Ten capsules from each of four
egg strings were placed in baths containing water of three temper-
ature regimens: cold (2 to 5°C) ambient (10 to 15°C), and warm
(20 to 25°C). Capsules were examined daily, and the day of hatch-
ing for various capsules was recorded.
RESULTS
Sexual Maturity and Growth at Age
Growth rate for individuals maintained in the laboratory was
highest in the first year, when the animals grew from 4 to 36.5 mm
Characteristics of B. carica
583
shell length. The average size after 10 years of growth was 144
mm, and by 14 years, the average size was 168.7 mm shell length
(Fig. 1). All 20 cultured individuals in 1986 at 9 years of age were
considered to be males, because 9 had a relatively large, well-
developed penis and 1 1 had a smaller penis. By October 1987, all
20 had either a large ( 10) or a moderately developed ( 10) penis,
and by April 1989, all had a well-developed penis.
On September 25, 1989. 12 years 5 months after hatching (ca.
13 years after eggs were laid), one whelk laid an egg case, but it
contained no embryos. The animal producing the egg case was 172
mm long and 95 mm wide and had a 25 mm long penis in April
1989. By April 6, 1990, 5 animals were female with the penis
reduced to a rounded protuberance. 10 were males, and 5 had
died, including the original female.
By May 20, 1991, 9 of the remaining 15 individuals were
females. Rounded protuberances remained, but the individuals
that appeared to have most recently become female had a small,
vestigial, flap-like penis. This flap was approximately 5 to 7 mm
across the widest point. Busycon under 166 mm length had a
larger. C-shaped penis and were considered to be males.
In September 1991, 3 of the remaining 14 Busycon laid egg
cases containing viable embryos. Five whelks were now presump-
tive males (under 166 mm), 7 were definite females with round,
button-like protuberances, and 2 were apparently transitional with
neither a flap, a well-shaped button, or a C-shaped phallus.
The three egg cases were maintained in running seawater from
September to April, while temperatures ranged from 9.2 to 14.5°C
and salinity averaged 30 ppt. During the first 3 weeks in April
most of the B. carica hatched from the egg cases. Some of the
early hatchlings climbed out of the tank and died. The others were
removed from the tray and grown in the laboratory for nearly 60
days, during which they approximately doubled in size before
being released in early June on the sand flats where the egg strings
were collected in 1976.
Gonadal Analysis
Of the 105 individuals used for gonadal analysis, only 8 were
males (sex ratio = 0.082), so males were not included in the
analysis. Females were collected on June 30, July 29, August 26.
October 10, and November 11, 1976, and on January 11, 1977.
and a second series was collected on April 15, May 1 1. June 12,
Growth Rate of Busycon carica
150
— 125
E
E, 100
1? 75 -
u
J 50
25
o
•
. . ••
• •
•
*
>•
(
20 40 60 80 100 120 140 160
10 30 50 70 90 110 130 150 17
Age (months)
i
Figure 1. Growth rate of B. carica.
July 7. and August 8, 1979. The range in gonadal weight was 0.3
to 27.4 g. Average gonadal wet weight ranged from 4. 1 to 12.2 g
and average nidamental gland weight ranged from 21.9 to 52.1 g.
The percentage of gonad or nidamental gland to total weight and
meat weight are given in Table 1 . The data indicate a decrease in
these percentages in October, the time of maximum observed egg
laying in the field. Although the data suggest a spring spawning,
no such spawning was seen in over 5 years of field observations.
Egg Strings and Egg Capsules
The only months in which copulating individuals were ob-
served on the intertidal flats were June and July. Eggs were laid in
the field from mid-August to November, with most egg laying
from mid-September to mid-October. The largest number of egg
strings found on the intertidal flat was 66 in 1976. Capsules re-
mained closed until spring, and hatching (on the basis of open
capsules) was observed from mid-March through early May. A
few of the earliest laid strings (late August) began hatching by the
end of October. In general, strings were present on the intertidal
flat throughout the winter, but by April or May, most had disap-
peared. Some of these egg strings washed ashore in those 2
months, and juveniles that had not hatched were dead.
Capsules per string ranged from 42 to 121. The average num-
ber of capsules above the anchor point per string was 99.7 (stan-
dard error [SE], 4.68; N = 16) in the winter of 1977 to 1978, 89.4
(SE, 2.44; N = 66) in the winter of 1978 to 1979, and 92.4 ±
4.93 in the spring of 1979. The number of capsules in the anchor
portion of the string ranged from 8 to 22 (mean, 13.1; SE, 1.21;
N = 16).
On the basis of the random selection of capsules along the
length of four strings, the average number of eggs per capsule is
less in the first and last 10% of egg-bearing capsules (Table 2).
Seventeen egg strings were found washed up on the beach in April
to May 1976. These cases were covered with fouling organisms
(algae, lUyanassa eggs, Corophium tubes) and had apparently bro-
ken off from their anchor. One case had not hatched (escape plugs
were intact); development had not occurred. The percent hatch
(hatching success) ranged from 18 to 86% (Table 3). Thirteen of
the 16 strings analyzed held hatching rates of more than 60%.
Fouling by corophid amphipods or mud snail (lllyanassa obtusata)
egg capsules may block the escape of whelks in the field.
Laboratory studies on hatching yielded a positive correlation of
hatching with temperature. With increasing temperatures, hatch-
ing times were reduced; at 2 to 5°C, no hatching occurred; at 10 to
15°C, hatching began at 65 to 78 days, and at 20 to 25°C, it began
at 22 to 30 days. One string was excluded because of a drop in
water temperatures.
DISCUSSION
Sexual Maturity and Growth at Age
The growth rates and timing of the sex changes in these labo-
ratory studies may not mimic those from the natural habitat; how-
ever, the whelks always had a surplus of clams of varied sizes, so
food should not have been a limiting factor. The temperatures and
salinities were typical of those in the field because of the contin-
uously supplied ambient seawater. The containers in which the
whelks were grown, although relatively uncrowded. would cer-
tainly not compare with the low-density habitats found in nature.
584
Castagna and Kraeuter
TABLE 1.
Seasonal gonadal indices (%) for female B. carica in Virginia. Indices are all wet weight
Month and day
J
J
A
O
N
J
A
M
J
J
A
Item measured*
30
3.1
29
26
10
11
11
15
11
12
7
8
G/TW
2.4
2.7
1.4
3.0
2.0
3.1
11
1.8
2.5
2.2
N/TW
12.9
10.3
12.3
8.1
12.7
11.3
11.1
7.3
8.6
12.5
11.2
G + N/TW
15.9
12.4
13.3
9.5
15.7
13.3
14.2
8.4
10.4
15.0
13.4
G/N
24.3
23.8
21.7
17.4
23.6
18.0
28.3
15.1
20.9
20.3
19.4
G + N/MW
22.1
16.8
20.5
12.8
21.7
18.2
20.6
11.3
15.1
22.0
19.1
G/MW
4.4
3.2
3.7
1.9
4.1
2.8
4.5
1.8
2.6
3.7
2.8
N/MW
17.7
lental gland;
13.6
TW, total
16.8
10.9
17.6
15.4
16.0
9.5
12.5
18.3
16.4
a G, gonad; N, nidan
weight; MW.
meat weight
Gonadal Analysis
Egg Strings and Egg Capsules
The sex ratio (0.082) from this study is within the range 7.8 to
9.0% reported by Kraeuter et al. (1989) for whelks from this
location and is almost identical to that reported for intertidal pop-
ulations of B. carica in Georgia (Walker 1988). Magalhaes (1948)
suggests that eggs are laid in the spring both at Beaufort, North
Carolina, and in Connecticut, but her data from farther north are
based on "fresh" capsules and not observed egg laying. On the
basis of gonadal and nidamental gland data and the small amount
of information available from the males, there appear to be two
spawning seasons in Virginia. In over 5 years of field study,
however, we did not find eggs being laid on an intertidal flat
during the spring, although spawning might occur in deeper water.
Ram (1977) used categories of gonadal weights ranging from
0.0 to 0.9 g to >3.6 g. The gonads of his animals were signifi-
cantly smaller than those on our animals; his largest class is
smaller than the average weight in any month in our study and far
less than our maximum of 27.4 g. It does not seem plausible that
the large percent body weight reduction from egg laying could be
recovered within a month, as our data suggest. The rapid recovery
of the ratios after May and October could be the result of cohorts
of snails moving on and off the intertidal flat. This rapid move-
ment is consistent with mark-recapture studies done on animals
from this same location (unpublished data). Similar rapid emigra-
tions of marked snails were reported by Weil and Laughlin ( 1984)
for tagged queen conchs on subtidal tropical grass flats. This ex-
planation is further supported by the evidence that there was never
a sample in which all females had spent gonads. It is possible that,
in addition to fairly rapid immigration and emigration, not all of
the females lay eggs every year.
TABLE 2.
Numbers of B. carica eggs per capsule along the length of strings
collected from Cedar Island, Virginia, December 1976 to
January 1977
Percent
distance
No. of capsules
Average no.
along string
sampled
of eggs
SE
0-10
6
11-30
5
31-60
8
61-90
9
91-100
5
20.0
8.53
37.4
5.98
51.5
4.20
37.5
3.64
10.2
3.08
Magalhaes (1948) reported finding copulating individuals of
Busycon in March, June, August, and September, but her article
does not distinguish between B. carica and B. canaliculaia. Kent
(1983) found copulating Busycon spiratum and Busycon contrar-
ium in October and January on tidal flats in northwest Florida.
Walker (1988) found whelks mating in the spring and fall on
Georgia intertidal flats. Our data fit the general pattern that cop-
ulation may be seasonal, but not necessarily at the same time as
egg laying.
The range in numbers of egg capsules above the anchor from
Beaufort, North Carolina (9 to 156) (Magalhaes 1948) spans the
number found in Virginia. The mean number of capsules in North
Carolina (80) is slightly less than the average number produced in
2 years on Cedar Island, VA. The numbers of individuals poten-
tially hatching from a single string appear to range up to 5,000 to
6.000 in both North Carolina and Virginia, although 2,000 to
3,000 would be more typical.
The timing of egg laying (September) in the laboratory-reared
animals matches that which we have observed in field studies on
intertidal flats. Hatching in the laboratory in spring also mirrors
what we have observed in both the field and from controlled-
temperature laboratory studies. Egg strings washing ashore during
winter storms or, more typically, during April and May, were
destroyed along with unhatched juveniles. The mechanism that
causes the egg strings to weaken and wash ashore around hatching
time is not known.
Several factors of importance to management are apparent from
this study. Clearly, the length of time required for maturity has
far-reaching implications. The presence of a penis does not unam-
biguously define whether a particular individual is male or female,
and thus, attempting to regulate harvests on the basis of sex or sex
ratios would be difficult. The length of time before females ap-
TABLE 3.
Egg cases of B. carica washed up on the bay beach of Cedar Island,
Virginia, in April and May 1977
Category
Range
No. of
srings
Average
SE
Whelks/string
Whelks/capsule
% Hatched
1.945-5,508
30-51
18-86
16
16
16
3770
41.79
206.50
1.54
Characteristics of B. carica
585
peared in the cohort was very long (12 years), and their length
(172 mm) was large. Harvest selection by size would probably be
specific for females. Egg cases remain vulnerable to storms or
other forms of bottom disturbance for 7 months before they hatch.
This long development further restricts the amount of time an area
could be exploited by dredges or other gear that disrupt the bot-
tom. Most egg cases were concentrated in a relatively small por-
tion of the flat. Hatching was controlled by temperature, and some
factor also positively associated with temperature allows the egg
strings to weaken so that they are more apt to wash away.
Although we did not make estimates of posthatch mortality, the
rapid loss of individuals in the laboratory suggests that field mor-
tality of small individuals is significant. We were never able
to discover small individuals in the field. Laboratory observat-
ions suggest that they remain buried most of the time (as do the
adults).
This species, which requires a long time to mature, lays rela-
tively few eggs per spawning season, has a low survival rate of its
young, and appears to be extremely vulnerable to harvest pressure.
Management for this resource for an optimal sustainable yield will
be difficult.
ACKNOWLEDGMENTS
The authors thank the staff members of the College of William
and Mary, Virginia Institute of Marine Science, Eastern Shore
Laboratory, for the care of the whelks cohort used in this study. A
special thanks to Jean Watkinson, Robert Bisker, and Doug Ayres
for their efforts in caring for the whelks and furnishing bivalves for
their food. We thank Nancy Lewis for preparing the manuscript,
Ms. Elizabeth Keane, Drs. George Grant, Joseph Loesch, Mark
Luckenbach and Roger Mann for their review of the manuscript,
and Dr. Edwards and an unknown editor for greatly improving the
manuscript. Contribution No. 1902 from Virginia Institute of
Marine Science. This is contribution no. 94-31 from the In-
stitute of Marine and Coastal Sciences, Rutgers University, and NJ
Agricultural Experiment Station Publication No. D-32406-1-93.
Support for some of this work has been provided by NJ state funds
and the NJ Commission on Science and Technology.
LITERATURE CITED
Appeldorn, R. S. 1988. Age determination, growth, mortality, and age of
first reproduction in the adult queen conch. Strombus gigas L., off
Puerto Rico. Fish. Res. 6:363-378.
Davis, J. P. & R. T. Sisson. 1988. Aspects of the biology relating to the
fisheries management of the whelks, Busycotypus canaliculatus and
Busycon carica. J Shellfish Res. 7:453—160.
DiCosimo, J. 1986. Virginia conch fishery: no longer crawling at snails
pace. Comm. Fish. Newslett. 6:
Edwards, D. C. & J. D. Huebner. 1977. Feeding and growth rates of
Polinices duplicalus preying on Mya arenaria Barnstable Harbor.
Massachusetts. Ecology 58:1218-1236.
Frank. P W. 1969. Growth rates and longevity of some gastropod mol-
luscs on the coral reef at Heron Island. Oecologia 2:232-250.
Gendron, L. 1992. Determination of the size at sexual maturity of the
waved whelk Buccinum undatum Linnaeus, 1758. in the Gulf of St.
Lawrence, as a basis for the establishment of a minimum catchable
size. J. Shellfish Res. 11:1-7.
Goldsmith. L. A., H.-M. Hanigan, J. M. Thorpe & K. A. Lindberg
1978. Nidamental gland precursor of the egg capsule protein of the
gastropod mollusc Busycon carica. Comp. Biochem. Physiol. 59B:
133-138.
Kaplan, I. M. & B. C. Boyer. 1992. Monitoring marine resources: eco-
logical and policy implications affecting the scientific collecting and
commercial value of New England conch {Busycon). Biol. Bull. 183:
Kent. B. W. 1983. Natural history observations on the busyconine whelks
Busycon contrarium (Conrad) and Busycon spiralum (LaMark). J.
Moll. Stud. 49:37-42.
Kraeuter. J. N., M. Castagna & R. Bisker. 1989. Growth rate estimates
for Busycon carica (Gmelin, 1791) in Virginia. J. Shellfish Res. 8:
219-225.
Magalhaes, H. 1948. An ecological study of snails of the genus Busycon
at Beaufort, North Carolina. Ecol. Monogr. 18:379-409.
Menzel, R. W. & F. E. Nichy. 1958. Studies of the distribution and
feeding habits of some oyster predators in Alligator Harbor. Florida.
Bull. Mar. Sci. GulfCarib. 8:125-145.
Paine, R. T. 1962. Ecological diversification in sympatric gastropods of
the genus Busycon. Evolution 16:515-523.
Paine. R. T. 1963. Trophic relationships of 8 sympatric predatory gastro-
pods. Ecology 44:63-73.
Peterson. C. H. 1982. Clam predation by whelks (Busycon spp.): Exper-
imental tests of the importance of prey size, prey density, and seagrass
cover. Mar. Biol. 66:159-170.
Powell, E. N. & H. Cummings. 1985. Are molluscan maximum life spans
determined by long-term cycles in benthic communities? Oecologia
(Bert. 167:177-182.
Ram, J. L. 1977. Hormonal control of reproduction in Busycon: laying of
egg capsules caused by nervous system extracts. Biol. Bull. 152:221-
232.
Santarelli, L. & P. Gros. 1985. Determination de 1'age et de la croissance
de Buccinum undatum L. (Gasteropoda: Prosobranchia) a l'aide des
isotopes stables de la coquille et de l'ornementation operculaire.
Oceanol. Acta 8:221-229.
Sisson, R. T. 1972. Biological and commercial fisheries related research
on the channeled whelk Busycon canaliculatum (Linne) in Narragansett
Bay, Rhode Island. Master's Thesis. University of Rhode Island.
Walker, R. L. 1988. Observations on intertidal whelk {Busycon and Busy-
cotypus) populations in Wassaw Sound, Georgia. J. Shellfish Res.
7:473-478.
Wefer, G. & J. S. Killingley. 1980. Growth histories of strombid snails
from Bermuda recorded in their 0-18 and C- 13 profiles. Mar. Biol.
60:129-135.
Weil, E. M. & R. G. Laughlin. 1984. Biology, population dynamics, and
reproduction of the queen conch Strombus gigas Linne in the Archi-
pielago de los Roques national park. J. Shellfish Res. 4:45-62.
Journal of Shellfish Research, Vol. 13, No. 2. 587-593, 1994.
THE OCCURRENCE OF DOMOIC ACID IN RAZOR CLAMS (SILIQUA PATULA), DUNGENESS
CRAB (CANCER MAGISTER), AND ANCHOVIES (ENGRAUL1S MORDAX).
JOHN C. WEKELL, ERICH J. GAUGLITZ, JR.,
HAROLD J. BARNETT, CHRISTINE L. HATFIELD, AND
MEL EKLUND.
National Oceanic and Atmospheric Administration, National Marine
Fisheries Senile, Northwest Fisheries Science Center, Utilization
Research Division, 2725 Montlake Blvd. East,
Seattle, Washington 98112, U.S.A.
ABSTRACT In September 1991, water fowl died in Monterey Bay, CA, after eating anchovies (Engraulis mordax) contaminated
with domoic acid. Analysis revealed that the anchovies contained up to 485 ppm domoic acid in their viscera. This was the first
reported incidence of domoic acid-related mortality of any organism in the United States. After this reported outbreak we obtained
frozen samples of anchovies that were harvested near Newport. CA. in April 1991 and found they contained 270 ppm domoic acid
in their viscera. By May, average domoic acid levels in frozen anchovy samples from this same area were less than 1 ppm. In October
1991 . domoic acid was detected in razor clams {Siliqua patula) from Oregon and Washington and appeared to peak (an average of 106
ppm for all Washington State beaches) in the first week of December 1991. The averages then declined to less than 20 ppm without
6 months. However, domoic acid was still present at low levels (averages <5 ppm) in razor clams from Washington state beaches in
December 1993. Dungeness crab [Cancer magisler) in Washington and Oregon were also found to contain domoic acid, but only in
their viscera. Domoic acid concentrations in the raw viscera of individual crabs from Washington state in December 1991 averaged
13 ppm and ranged from 0.8 to 90 ppm. The highest average levels of domoic acid in Washington state crabs were in the Grays Harbor
and Willapa Bay samples. 32 and 31 ppm. respectively. By 1992 domoic acid level averages were <5 ppm in preseason samples of
Dungeness crab taken along the Oregon and Washington coasts, ranging from 0 to 71 ppm The highest levels of domoic acid in 1992
(36-71 ppm) were recorded in samples taken early in that year (January through April).
KEY WORDS: domoic acid. Dungeness crab. Cancer magisler, anchovies, Engraulis mordax, razor clams. Siliqua patula
INTRODUCTION
The first outbreak of domoic acid poisoning in the United
States was reported from Monterey Bay, CA, in September 1991
(Worketal. 1991, 1993). In that outbreak, a large number of sick
and dead pelicans (Pelecanus occidentals) and cormorants (Pha-
lacrocorax penicillatus) were observed. Analyses of the stomach
content of the affected birds revealed they had consumed ancho-
vies {Engraulis mordax). Chemical and biological assays of the
stomachs of pelican and anchovies taken from the bay revealed
high levels of domoic acid (T.M. Work pers. coram.; anonymous
1991; Work et al. 1991; Fritz et al. 1992). Examination of water
column samples and gut contents of the anchovies from Monterey
Bay revealed large numbers of a diatom, later identified as Pseu-
donitzschia australis (Fryxell 1992; Buck et al. 1992). When sam-
ples of recovered diatoms were grown in the laboratory, they were
shown to produce domoic acid (Garrison 1992; Garrison et al.
1992; Buck et al. 1992; Villac et al. 1993).
Because of the tragic consequences of the 1987 domoic acid
outbreak in Canada (Todd 1990. 1993), a rapid response to the
threat of domoic acid poisoning was mounted in the United States.
The U.S. Food and Drug Administration (FDA) declared an action
limit of 20 ppm domoic acid in seafoods entering interstate com-
merce, and health regulatory agencies in Washington, Oregon,
and California began monitoring for the presence of domoic acid
in a variety of marine species. As a result of the early warning and
rapid response by state and federal agencies, no confirmed human
illnesses were reported from Washington, Oregon, or California.
In order to manage rationally the risks associated with domoic
acid poisoning, it was necessary to determine both the biological
and geographical distribution of the outbreak. Because anchovies
were implicated in Monterey Bay, a survey of commercial frozen
stocks of anchovies intended for human consumption was per-
formed. Shellfish that consume P. australis and their predators
were also examined. Finally, it was necessary to determine the
longevity or persistence of domoic acid in certain target species. In
November 1991, our laboratory began analyses of several marine
species that included razor clams (Siliqua patula), mussels
(Mytilus edulis). anchovies (E. mordax), and Dungeness crab
(Cancer magisler).
Wekell et al. (1994) reported that domoic acid levels in Wash-
ington state razor clams seemed to peak in December 1991 and
then declined below the regulatory limit of 20 ppm by June 1992.
In addition, we continued to monitor and observe low levels (<5
ppm) of domoic acid in these clams along the Washington coast
well into December 1993 more than 2 years after the initial out-
break was reported in 1991. However, in the late fall (November
through December) of 1992 and 1993, we observed brief increases of
domoic acid levels in some samples of Washington state razor clams.
As potential predators of razor clams and other contaminated
sources, it was clear that Dungeness crab might also become con-
taminated with domoic acid. We report here some information
about the geographic distribution of the domoic acid in Dungeness
crabs and razor clams along the Washington and Oregon coasts
from December 1991 to December 1993. In addition, we present
data on mussels and anchovies from Monterey Bay, CA. The
anchovies were taken at the time of the initial observations of
domoic acid poisoning in September 1991 . while the mussels were
taken in November 1991. We also include domoic acid data on
commercially frozen anchovies from southern California landed in
April and May of 1991. This work was part of a combined effort
that included health and regulatory agencies in the states of Cali-
fornia. Oregon, and Washington, the U.S. Food and Drug Ad-
ministration, and the National Marine Fisheries Service.
587
588
Wekell et al.
MATERIALS AND METHODS
Reagents
Methanol (MeOH) and acetonitrile (MeCN) were HPLC grade
(Baxter Healthcare Corp., Burdick and Jackson Division,
Muskegon, MI 49442). Sodium chloride was ACS reagent grade.
Water to be used for extractions and HPLC analyses was prepared
by passing distilled water through a Milli-Q water system (Milli-
pore Products Division, Bedford, MA 01730). The Milli-Q water
used as an eluent in HPLC analyses was filtered through a 0.45-
p.m HA filter (Millipore Products Division, Bedford, MA 01730)
and mixed with 100 mL MeCN in a 1 -liter volumetric flask. To
this mixture was added 1.0 mL trifluoroacetic acid (TFA) (Sigma
Chemical Co., St. Louis. MO 63178) and then diluted to mark.
Immediately before chromatography, solvents were degassed with
helium for 5 min.
Standards
A working domoic acid standard solution was prepared from
the certified standard DACS-1 (Canadian National Research
Council, Institute of Marine Biosciences, 1411 Oxford Street,
Halifax, N.S.. Canada B3B 3Z1; Hardstaff et al., 1990) by dilut-
ing it with 10% aqueous MeCN to a final concentration of 4.45
ppm. A 1000 ppm stock solution of tryptophan was prepared by
dissolving 100 mg L-tryptophan (Sigma Chemical Co., St. Louis.
MO 63178) in 10% MeCN and diluting to 100 mL. A working
standard solution of tryptophan was prepared by diluting the stock
solution to 100 ppm using 10% MeCN. The tryptophan standard
was used only to set initial chromatographic parameters, i.e., res-
olution of tryptophan and domoic acid, when chromatographic
conditions, columns, or guard columns were changed.
Anchovy
Anchovies from Monterey Bay were obtained from Dr. Thierry
Work, California State Fish and Game, Wildlife Investigations
Laboratory, Rancho Cordova, CA. The samples were frozen and
held in storage for about 3 months before we received them. Com-
mercially frozen anchovies from southern California were received
from Mr. Glenn Kiel, National Marine Fisheries Service, Seafood
Inspection Branch, Bell. CA. To facilitate analysis, the fish were
allowed to thaw only partially so that the frozen viscera could be
removed as an intact unit for domoic acid analysis. After eviscer-
ation, the head and fins were removed from the remaining body.
The body and viscera were homogenized and analyzed separately.
Mussels
Mussels from Monterey Bay were obtained from Mr. Kenneth
Hansgen, California Health Services, Environmental Health Ser-
vices Section, Sacramento, CA. Mussels were removed from their
shells and the whole animal was taken for analysis. For sampling,
a sufficient number of mussels were selected so that about 50 g of
shucked meats were obtained, usually 20 to 30 individual animals.
Razor Clams
Razor clams were furnished by the Washington State Depart-
ment of Fisheries (WDF). Clams were collected from an area that
starts at the mouth of the Columbia River and extends north along
the Washington coast to the Moclips River. This area is divided
into four major management beaches: Long Beach, Twin Harbors,
Copalis, and Mocrocks (see Fig. 3). In most cases, six clams were
taken from each management site at each sampling period, usually
once a month at low tide, kept cool with "gel-ice" packets, and
transported to our laboratory for shucking the next day. If clams
could not be processed the next day, the specimens were frozen
whole in the shell. Razor clams were first rinsed quickly in run-
ning water to remove sand and other debris and then shucked. If
needed, the clam meat was rinsed briefly in running water to
remove any further sand or debris, and then divided into edible
meat and viscera. This procedure was similar to the steps that
recreational and commercial processors employ for preparing ra-
zor clams for consumption or retail sale (Washington State De-
partment of Fisheries pamphlet, no publication date). Clams were
analyzed individually and an arithmetic mean determined for each
sampling.
Dungeness Crab
Dungeness crab samples from Washington and Oregon were
provided by the Washington State Department of Agriculture and
Oregon State Department of Agriculture, respectively. In Wash-
ington state, crabs were collected in December 1991 by commer-
cial fishermen from areas designated by the WDF representing the
crab fishery off the Washington state coast. In 1992 and 1993
crabs were collected in both states as part of a monitoring program
to determine the suitability of opening the commercial crabbing
season. For a given site, a minimum of six crabs were used for
analysis. Crabs were analyzed individually for domoic acid and an
unweighted mean determined for each sampling site. Visceral tis-
sues removed for analyses included the hepatopancreas, stomach,
and intestinal tract. Heart and the epidermal membrane were also
removed and included with the visceral tissues for analysis. Gills,
which are not part of the visceral sac, were removed separately and
not included in the visceral sample.
Sample Preparation
Tissues were homogenized using a common household type
blender or food processor depending on volume of the sample. If
sample sizes were too small for the blender, the tissue was mixed
with distilled water, in a 1:1 ratio (w/w), to physically increase the
volume and ensure thorough homogenization. Allowances were
made for this added water in the final calculations of domoic acid
content.
Domoic Acid Analysis
All tissue homogenates were extracted according to the method
of Quilliam et al. (1991) with modification to the solid phase
extraction (SPE) cleanup procedure (Hatfield et al. 1994). It was
necessary to subject samples to the SPE procedure to remove
compounds that interfered with the domoic acid HPLC analyses.
All analyses were performed on a Hewlett-Packard 1090 HPLC.
equipped with a Vydac 201TP column (Reversed phase C18, 2.1
mm x 25 cm, Separations Group, Hesperia, CA 92345) and a
diode array detector set at 242 nm with a 10-nm bandwidth and
reference signal at 450 nm with 10-nm bandwidth. The domoic
acid was chromatographed isocratically at 40°C with water MeCN
TFA (90:10:0.1) (v/v/v) at a flow rate of 0.300 mL/min. Chro-
matographic conditions (solvent polarity and flow rate) were ad-
justed so that domoic acid was separated from tryptophan and had
retention times between 6 and 9 min. Sample injections, using a
25-p.L sample loop, ranged from 5 to 20 p.L. A certified domoic
Domoic Acid in Clams, Crab, and Fish
589
acid standard (DACS-1) was included before, after, and some-
times within each set of samples for calibration and quantitation.
RESULTS AND DISCUSSION
Anchovies
Chronologically, domoic acid poisoning was first reported in
the United States in mid-September of 1991 from Monterey Bay,
CA (Work et al. 1991). Our analyses of anchovies taken at this
time in Monterey Bay found an average of 275 ppm domoic acid
in the viscera and 77 ppm in the body meat (Fig. 1). Levels in the
viscera of six fish ranged from 177 to 485 ppm, whereas the body
content ranged from 51 to 105 ppm. Whether domoic acid is
naturally distributed in the body tissues of these fish is question-
able because samples provided to our laboratory for analysis had
marked decomposition of the visceral tissues and indication of
enzymatic deterioration of the surrounding body tissue. It is pos-
sible that the detection of domoic acid in the body tissues may be
due to contamination by the postmortem migration of the toxin
from the catabolized visceral tissues into the immediate surround-
ing body tissues.
In November 1991 we analyzed frozen anchovies from south-
em California, caught in the area between Point Vincente and
Newport and east of Santa Catalina Island. In anchovy samples
taken on April 21, 1991, 4 months before the domoic acid poi-
soning was reported in Monterey Bay, we found 192 ppm domoic
acid in the viscera and 28 ppm in the body meant (Fig. 1). The
body concentration of domoic acid was less in these Southern
California anchovies than was found in the Monterey Bay samples
(i.e., 77 ppm); however, less decomposition of the body tissue
surrounding the viscera was noted in these samples, perhaps re-
ducing migration of domoic acid from the visceral tissue into the
surrounding body cavity tissue. Further samplings of anchovies
from southern California in May 1991, about 1 month later,
showed levels of domoic acid of less than 1 ppm.
Mussels
Mussels taken from sites within Monterey Bay (Fig. 1) on
November 25. 1991 showed domoic acid levels of 0.6 ppm at
Mussels
Monterey Bay.
25-Nov-91
Anchovies
Monterey Bay
Mid-Sept 91
body: 77(51-105)
viscera: 275(117-485)
[San Francisco
V
iSanta Cruz
rMoss Landing
|Monterey/Pt. Pinos
Oregon
California
Southern California
Mussel and Anchovy
Sampling-1991
Anchovies C^Santa Barbara
Southern California T^jjfr* Los Angeles
21-Apr-91
192 (Viscera)
28 (body)
14-May-91
0.3 (Viscera)
0.2 (Body)
24-May-91
0.6 (Viscera)
0.5 (Body)
Figure 1. Average and range of domoic acid content (ppm) of mussels
and anchovies taken from southern California sampling sites.
Monterey (Point Pinos), 9.1 ppm at Moss Landing (Elkhorn
Creek), and 1 .0 ppm at Santa Cruz (Natural Bridges). We note that
the highest levels were reported in the inner most part in the Bay
at Moss Landing, whereas the lower levels were seen at the more
seaward sites. Because mussels are known to depurate domoic
acid rapidly (Novaczek et al. 1992), this would suggest that the
domoic acid source was still resident in the Bay 2 months after the
initial observation in September 1991.
Razor Clams
By October 1991 . domoic acid was found in razor clams from
Oregon and Washington states. Wekell et al. (1994) reported that
razor clams in Washington state apparently acquired their highest
levels of domoic acid in the first week of December 1991. In that
study, razor clams were separated into edible and nonedible (vis-
ceral tissue) portions as described earlier in this paper. The highest
average concentration of domoic acid observed in edible razor
clam tissue was 147 ppm. in samples taken from the Twin Harbors
area on December 3, 1991. whereas the lowest average (73 ppm)
on this date occurred in samples from the Long Beach area. In the
study, the twin Harbors and Long Beach areas represented the two
extremes in the collected data. We found that domoic acid levels
in samples from Twin Harbors always exceeded the average for all
Washington state beaches and the levels in the Long Beach area
samples always fell below that overall average. The nature of this
phenomenon is not understood at this time. Copalis and Mocrocks
areas had values for domoic acid that fell between the two ex-
tremes of the Twin Harbors and Long Beach areas. We began our
observations of domoic acid concentrations in razor clams in No-
vember 1991 and have continued through January 1994. From our
observations, it seems that domoic acid achieves its highest levels
in razor clams during the late fall (October through December).
However, levels in 1992 and 1993 were considerably lower than
those observed in December 1991 (Fig. 2). The lowest levels of
domoic acid in razor clam samples from all Washington state
beaches were observed in the late summer of 1993, with averages
below 5 ppm. The late fall domoic acid season in 1993 seemed to
produce higher levels of the toxin than was observed in 1992.
It also seems that domoic acid was not a new phenomenon in
1991 but has been present on the west coast of the United States
since at least 1985. In 1992, we examined both frozen and canned
razor clams dug from 1985 to 1990 in Washington state and found
traces of domoic acid (Wekell et al. 1994).
120
1992
Collection Date
Figure 2. Average domoic acid levels (ppm) in the edible tissue of
razor clams from all Washington state management areas, September
1991 through December 1993.
590
Wekell et al.
Dungeness Crab
The commercial Dungeness crab season along the Washington
and Oregon coastal areas normally begins on December 1 and can
extend through the summer months of the following year. The
season is usually closed for a brief period as the crab enter their
moulting stage, sometime between June and September. However.
the highest fishing activity and the major portion of the total land-
ings occur in the first 6 to 8 weeks of the season, i.e., December
and the following January, weather permitting. When domoic acid
concentrations in razor clams reached a maximum in the first week
of December 1991, a limited survey of Dungeness crabs off the
Washington coast was undertaken because, as potential predators
of razor clams, they might accumulate harmful levels of domoic
acid.
Dungeness crab is marketed in several forms, e.g., live, whole
cooked (fresh or frozen), picked meat, and sections. Although
most people consume only the cooked leg and body meat, a sig-
nificant portion of the population also consumes crab viscera either
directly, usually the hepatopancreas (crab "butter" or "mus-
tard"), or indirectly as a soup, stock, flavoring ingredient, or
condiment. Whole cooked Dungeness crab is one of the more
common forms and is prepared by placing the crabs in large bas-
kets and submerging them in boiling salted water for varying
lengths of time, usually 20-30 min depending on the processor.
After the crabs are cooked, they are allowed to cool, either in
cooling water baths or in air. During this cooling process, the
crabs may drain or retain any entrapped cooking or cooling water
that has entered the visceral compartment beneath the carapace.
Because domoic acid is readily soluble in boiling water (Quilliam
et al. 1991), it was not clear how much domoic acid may be lost
by draining or diffusion into other parts of the cooked crab. The
uptake, retention, loss, or diffusion of fluids from the crab will
depend on their orientation in the basket during processing and the
integrity of the crab's body compartment. Therefore, cooking crab
may potentially lower its domoic acid content.
In unpublished studies on raw Dungeness crab, we could not
detect domoic in the gills, hemolymph (blood), or in any of the
edible body and leg meats of the crab. Domoic acid was found to
be confined exclusively to the viscera of these crabs, with the
largest portion in the hepatopancreas. a part of the digestive sys-
tem. Because of the possible influences of the cooking process and
other handling procedures and our findings on live crab, all results
presented in this report represent visceral concentrations of domoic
acid in uncooked Dungeness crab.
When it was noted that domoic acid in razor clams reached
very high levels in the first week of December, concern was ex-
pressed about similar levels in Dungeness crab, a possible predator
of razor clams. We began our survey of Dungeness crabs with
samples taken on December 11, 1991. During the following year,
our laboratory analyzed more than 400 raw crab viscera samples
(Tables 1 and 2).
Although the Dungeness crab season usually extends into the
summer months, the major fishing effort of the 1991-1992 season
occurred between December and April; therefore we confined our
analyses for the season to the period December 11, 1991 to April
3, 1992 (Table 1). The high values of domoic acid, first recog-
nized in razor clams in 1991, were also reflected in the 1991-1992
season sample weighted average for Dungeness crab, i.e., 17
ppm, with an average range of 4.7 to 35 ppm. However, by the
latter part of 1992 (Table 2). the preseason weighted average
dropped to 2.5 ppm with an average range of 0.4 to 1 1 ppm. The
highest values of domoic acid during the 1991-1992 season were
found in samples from Grays Harbor, WA with 90 ppm (Decem-
ber 17, 1991) and a sample from Willapa Bay, WA with values of
78 (December 12, 1991). Both of these samples were collected
approximately a week after our estimate of the peak domoic acid
content was reached in razor clams. During this period the 1991
commercial crab fishing season in Washington was closed. Later
in the spring, a high level of 71 ppm (March 1, 1992) was ob-
served in Willapa Bay.
For the 1992-1993 preseason sampling study, we surveyed
TABLE 1.
Summary of domoic acid levels in Dungeness crab in NMAFS survey: December 1991 through April 1992.
Collection Site
Collection
Average"
Standard
Minimum
Maximum
Location
Date
ppm
Deviation
ppm
ppm
N
Willapa Bay, WA
ll-Dec-91
0.8
NO"
0.8
0.9
2
Willapa Bay. WA
12-Dec-91
30
20
11
78
12
Ocean Shores. WA
17-Dec-9I
6.2
7.7
1.4
22
6
Grays Harbor, WA
17-Dec-91
28
36
1.5
90
7
Off Grays Harbor, WA
17-Dec-91
6.1
3.0
2.7
9.5
4
Pacific Beach, WA
17-Dec-91
10
4.0
5.7
16
6
Destruction Island, WA
17-Dec-91
17
5.9
8.1
26
6
North Head Light, WA
17-Dec-91
26
NO
24
29
2
Long Beach, WA
17-Dec-91
6.2
6.5
1.9
19
6
Sea View, WA
I7-Dec-91
2.7
NO
1.4
4.0
2
Klipsan Beach, WA
17-Dec-91
9.7
NO
4.4
15
2
Willapa Bay. WA
19-Dec-91
14
10
2.8
29
8
Willapa Bay, WA
20-Jan-92
17
13
2.0
36
6
Willapa Bay, WA
l-Mar-92
20
26
4.5
71
6
Pt. Grenville, WA
27-Mar-92
16
13
3.3
37
6
Willapa Bay, WA
3-Apr-92
31
26
0
70
6
1991-92 Average
17"
14"
4.7"
35a
87
' arithmetic mean
b sample weighted mean
c Not Calculated, since N =
2
Domoic Acid in Clams, Crab, and Fish
591
TABLE 2.
Summary of domoic acid levels in Dungeness crab in NMFs pre-season survey: July 1992 through November 1992.
Collection Site
Collection
Average"
Standard
Minimum
Maximum
Location
Date
ppm
Deviation
ppm
ppm
N
North Head Light, WA
20-Jul-92
0.3
0.3
0
0.7
6
Grays Harbor, WA
24-Jul-92
1.3
2.2
0
8.2
17
Grays Harbor, WA
18-Aug-92
0.3
0.7
0
2.5
12
Columbia River. WA
20-Aug-92
0.9
1.1
0
3.4
12
Ocean Shores, WA
M-Aug-92
3.7
4.9
0
22
18
Alsea Bay. OR
28-Sep-92
0.8
1.5
0
3.8
6
Yaquina Bay, OR
30-Sep-92
0
0
0
0
6
Coos Bay, OR
30-Sep-92
2.8
4.9
0
16
12
Columbia River, OR
l-Oct-92
0
0
0
0
12
Brookings, OR
2-Oct-92
2.9
6.3
0
23
12
Tillamook Bay, OR
3-Oct-92
0.2
0.6
0
2.1
12
Bellmgham Bay. WA
l3-Oct-92
0.3
0.4
0
0.8
6
Long Beach. WA
26-Oct-92
1.4
1.0
4.0
4.0
12
Brookings, OR
26-Oct-92
6.4
10
1.2
38
12
Grays Harbor. WA
28-Oct-92
1.8
1.5
0
3.4
12
Off Willapa Bay, WA
28-Oct-92
3.1
2.2
1.1
6.6
8
Astoria, OR
28-Oct-92
0.7
0.9
0
2.4
12
N. of Jetty, Coos Bay, OR
28-Oct-92
11
20
1.2
74
12
Tillamook Bay. OR
31-Oct-92
0.2
0.4
0
1.0
6
2mi.S. of Columbia River, OR
9-Nov-92
1.0
1.5
0
4.8
12
3mi.S. of Cape Blanco, OR
10-Nov-92
2.3
2.4
0
7.2
12
Brookings. OR
ll-Nov-92
4.6
6.0
0.9
20
12
Newport, OR
12-Nov-92
3.9
5.5
0
15
9
Garibaldi, OR
12-Nov-92
1.9
1.9
0
6.2
12
Coos Bay, OR
12-Nov-92
3.7
5.3
0
19
12
Grayland. WA
15-Nov-92
2.9
1.7
1.0
5.9
6
Grays Harbor, WA
15-NOV-92
2.3
3.0
0
8.0
6
Copalis Rock, WA
15-Nov-92
3.6
2.9
1.6
7.9
4
Off Willapa Bay, WA
15-Nov-92
2.0
1.4
1.2
4.8
6
Ocean Shores, WA
15-Nov-92
1.1
0.7
0
2.1
6
Newport, OR
15-Nov-92
3.9
5.2
0
16
12
1992 PreSeason Average
2 5b
3.1'
0.4a
11"
314
Combined 1991-1992 Average
5.7b
6.2ac
1.9"
19a
401
■ arithmetic mean
b sample weighted meanc excludes standard deviation data where N = 2
Dungeness crabs taken between July 20, 1992 and November 15,
1992. Although domoic acid levels declined in late 1992, occa-
sional high levels samples were observed. For example, we ob-
served one sample containing 74 ppm from the Coos Bay area in
Oregon on October 28, 1992. Nevertheless, for our whole sam-
pling effort from December 1 1, 1991 to November 15, 1992, the
overall combined sample-weighted average* was 5.7 ppm (Table
2), with an averaget high value of 19 ppm and an average low
value of 1 .9 ppm.
Washington
Samples of raw Dungeness crabs taken from all sites along the
Washington coast in December 1991 were found to contain some
domoic acid (Fig. 3) with a weighted average level below 20 ppm.
*Sample weighted mean = (N * [DA]mean)/(2 N), where N is the number
of samples in the group. (DA]mcan is the average domoic acid concentration
of the group, and S N is the total number of individual samples,
t Arithmetic mean = £ [DA]hlgh or loJl Nslles, where [DA]hlgh or low is the
highest or lowest value of domoic acid concentration for each sample
group and £ Nsllcs is the total number of sites sampled.
Three of these samples (Willapa Bay, Grays Harbor, and North
Head Light) exceeded 20 ppm. The highest average levels of do-
moic acid were found in crab taken from inside both Grays Harbor
and Willapa Bay. 28 and 30 ppm. respectively (Table 1). How-
ever, considerable variation was found within the samples from
each area; domoic acid concentrations in individual crab ranged
from 1 .5 to 90 ppm for samples from Grays Harbor and 0.8 to 78
ppm for crabs from Willapa Bay (Table 1). Lower levels of do-
moic acid were observed in samples collected just outside of Grays
Harbor: 6.2 ppm (range: 1 .4-22 ppm) for the northern side and 5.9
ppm (range: 2.7-9.5 ppm) for the southern side (Fig. 3). Destruc-
tion Island, about 40 km north of the entrance to Grays Harbor,
produced Dungeness crab samples that averaged 17 ppm domoic
acid and ranged from 8 to 26 ppm (Fig. 3). Domoic acid levels in
viscera from Dungeness crab harvested along the Long Beach
Peninsula averaged 9.6 ppm and ranged from 1 .4 to 29 ppm. Crab
samples from Long Beach were broken down into six sample
subsites, with only two to three crabs from each of these sites.
Figure 3 shows the results of this subsampling. In the Long Beach
area, the highest level observed were in crabs taken near the mouth
of the Columbia River (average of 26 ppm); however, because of
the small sample size (N = 2 or 3) and the wide range of con-
592
Wekell et al.
\
17(8-26) iMocllpsR. \^l
Destruction Island V^ j%. y-. £=*
Copalls R
Mocrocks
Copalis
5.9(2.7-9.5)
Twin Harbors 3i