876 
Fishery Bulletin 96(4), 1 998 
200 400 600 800 
Depth (m) 
Figure 5 
Catch per haul by 100-m depth intervals for pink snailfish 
(solid) and red snailfish (dotted). 
dance of the three developmental stages (uneyed 
eggs, eyed eggs, and larvae) did not differ between 
snailfish species (chi square test, P=0.163), indicat- 
ing that the timing of the egg deposition events and 
subsequent embryonic and larval developmental 
rates were about the same for both species. Summed 
over both species, the relative abundance of uneyed 
and eyed embryos was about equal, whereas the 
abundance of larvae was less than 20% that of ei- 
ther embryonic stage (Table 1). 
Mean egg diameter was 4.83 mm (SD = 0.29, 
eggs=4207, egg masses=18) and mean egg number 
was 688 (SD=84.0, egg masses=15). By comparison, 
the mean diameter of eggs taken from the ovarian 
lumen of a single pink snailfish was 4.9 mm 
(SD=0.124, n=208) and the total number of eggs was 
790 with a displacement volume of 48.2 mL. Based 
Table 1 
Incidence of Careproctus egg masses and larval clusters 
by species and stage of development. 
Uneyed 
embryos 
Eyed 
embryos 
Larvae 
Total 
Red snailfish 
36 
45 
6 
87 
Pink snailfish 
24 
14 
3 
41 
Total 
60 
59 
9 
128 
on this specimen, mean egg volume was 0.061 mL 
(SD=0.017) and mean egg mass was 0.064 gm 
(SD=0.020). 
Volume of single egg masses increased asymptoti- 
cally with branchial volume (Fig. 7), indicating that 
up to some crab sizes, egg mass volume is limited by 
the available space in the branchial chambers. Up 
to 65% of the available volume within the branchial 
chambers (i.e. branchial volume - gill volume) was 
filled by a single egg mass. 
Incidence of snailfish eggs and 
larvae in golden king crab 
Before developing a statistical model describing the 
1982 incidence of snailfish eggs and larvae as a func- 
tion of crab sex, carapace length, and sampling depth, 
we performed statistical tests to determine whether 
the two snailfish species could be combined in the 
analysis. Incidence of pink and red snailfish eggs did 
not differ by sex of host (chi-square test, df=l, 
P=0.28), nor by size of host (Utest, df=128, P= 0.58), 
but incidence did differ by depth. Pink snailfish oc- 
currences declined with increasing depth until none 
were encountered at depths >400 m (Fig. 8). In con- 
trast to this, red snailfish masses increased in abun- 
dance with depth until reaching a peak at between 
300-400 m, then declined until none were encoun- 
tered at depths >600 m. Despite the difference in depth 
distribution for each species, the species were combined 
to simplify analysis. Because the combined incidence 
of eggs and larvae increased with depth to a peak at 
250 m, then subsequently declined (Fig. 8), the depth 
effect in the model included a quadratic term. The 
best fit of a logistic model to the combined incidence 
data indicated that sex (P<0.001), length (P=0.003), 
and depth (P<0. 001) were all highly significant (Table 
2 ). The fitted model predicts 1 ) incidence increases with 
size (Fig. 9), 2) incidence is greater for males them fe- 
males (Fig. 9), and 3) incidence is greater at middepth 
(i.e. significant quadratic depth effect, Table 2). 
The best fit of a logistic model to the 1996 inci- 
dence in commercial males indicated that area 
(P<0.001) and whether the crabs were landed dead 
or alive (P=0.046) were both significant (Table 2). 
Incidence in the Bering Sea was over three times 
greater than in the Aleutian Islands. Incidence in 
dead crabs was 1.9 times greater than in live crabs 
in the Bering Sea and 1.6 times greater in the Aleu- 
tian Islands (Table 3). Incidences in live commercial 
males from the Aleutian Islands were 2.5 times 
greater than in sublegal males and 12.5 times greater 
than in females (Table 3). Incidence in 1996 of live 
commercial males in the Bering Sea was less than 
20% of the incidence in 1982 commercial size males 
