882 
Fishery Bulletin 96(4), 1998 
crab. However, the variety of snailfish embryonic 
stages that occurred during the 1982 sampling 
(Table 1), coupled with the probable slow develop- 
mental rate, indicates that pink and red snailfish 
have either a protracted spawning season, or per- 
haps lack spawning seasonality. Previous studies of 
snailfish reproduction have reported that aseasonal 
spawning was typical of abyssal and slope snailfishes, 
although C. malanurus, a close relative to pink 
snailfish, has a seasonal peak in spawning off Or- 
egon (Stein, 1980). The likelihood for preferential 
selection of early molt-stage crabs is not necessarily 
diminished by aseasonal spawning in red and pink 
snailfish because golden king crab also have 
aseasonal reproduction and molting (Somerton and 
Otto, 1986). 
The incidence of snailfish eggs and larvae was 
greater in male than in female crabs and increased 
with crab size in both the 1982 survey (Table 2; Fig. 
9) and the 1996 commercial sampling (Table 3). The 
preference for male crabs as hosts over females is 
quite pronounced. For example, at the median 1982 
sampling depth, a 100-mm male is 1.9 times more 
likely to contain eggs and larvae than an equal-size 
female (Fig. 9). As in our study, male Lithodes 
tropicalis had a higher incidence of snailfish eggs 
than females (Melville-Smith and Louw, 1986), but 
the apparent sex selection was attributed to size se- 
lection and to a large sexual dimorphism in crab size. 
In our case, we believe sex itself is important in host 
choice because sex was a significant predictor of in- 
cidence even when size and depth effects were in- 
cluded in the model. Such sex selection may not be 
universal, however, because a previous study of L. 
aequispinus reported that incidence was higher in 
females (Love and Shirley, 1993). Why a preference 
for males should occur is not obvious. One possible 
explanation, based on aquarium observations (Love 
and Shirley, 1993), is that female golden king crab 
aggressively defend the embryos attached to their 
pleopods. Perhaps this aggressiveness can discour- 
age the attempts of a snailfish to extrude her eggs 
into the branchial chambers of the crab. 
The apparent preference for large crabs is also 
quite pronounced. For example, at the median 1982 
sampling depth, a 140-mm-CL male is 1.9 times more 
likely to contain snailfish eggs or larvae than a 100- 
mm-CL male (Fig. 9). The apparent preference of 
snailfish for large crabs is likely due to two distinct 
attributes associated with size in lithodid crabs. First, 
large king crabs molt less frequently than small king 
crabs. Although the molting frequency of golden king 
crabs is unknown, for male red king crabs ( Para - 
lithodes camtschaticus ), molting frequency dimin- 
ishes continuously with increasing age (McCaughran 
and Powell, 1977). Since a lower molt frequency 
would result in a lower probability of premature re- 
lease of eggs and larvae, it is an advantageous fea- 
ture for a perspective host to have. Second, larger 
crabs have larger branchial chambers to contain 
snailfish eggs. In our case, egg mass volume increased 
with branchial volume (Fig. 10), indicating that the 
size of an egg mass is limited by the size of the bran- 
chial chamber. This increase, however, diminished 
with crab size, indicating that in large crabs the size 
of an egg mass may be determined more by snailfish 
fecundity than by the availability of space. It is not 
clear what effect the space limitation would have on 
the searching behavior of snailfish. If a female 
snailfish is capable of partitioning a batch of ripe 
eggs among several spawning events, then she might 
be able to reduce the problem of space limitation by 
depositing eggs in several crabs. If instead female 
snailfish must deposit their entire batch in one 
spawning event, then considerably more searching 
would be required to find a crab with sufficient vol- 
ume. To help provide some indication of whether a 
female partitions a batch of eggs, we measured egg 
batch volume for a single 40-cm pink snailfish, con- 
sidering only eggs that were free in the ovarian lu- 
men. Because the measured volume (48 mL) is about 
equal to the median volume of the egg masses found 
in crabs (Fig. 7), it is possible that a female spawns 
an entire batch in one event. If this is true, then a 
female would have to determine, perhaps by prob- 
ing with her ovipositor, whether a prospective host 
has sufficient branchial volume for her batch of eggs. 
Besides size and sex, other aspects of crab host 
choice by snailfish have been postulated. Melville- 
Smith and Louw (1987) suggested that Careproctus 
might preferentially choose only one side of host L. 
tropicalus to deposit egg masses. In our case, nei- 
ther pink snailfish nor red snailfish chose one side 
of the crab over the other (binomial test, red snailfish, 
P= 0.39, pink snailfish, P=0.76). Love and Shirley 
(1993) and Melville-Smith and Louw (1987) sug- 
gested, after finding no crabs with egg and larvae 
masses in both branchial chambers, that snailfish 
are inhibited from depositing egg masses in the un- 
affected branchial chamber of previously infested 
crabs, presumably to reduce host mortality. In our 
case, this feature was not true because of 23 crabs 
that had at least two egg or larvae masses, 13 had 
masses on both sides. 
Because large crabs occur in shallower waters than 
small crabs and males occur in shallower waters than 
females, the preferred hosts of both snailfish species 
occur in the shallower portion of the 1982 depth 
range. Pink snailfish occur at approximately the 
same depths as their preferred hosts, but red 
