This study identifies sources of variation in the prevalences of paras
itized and hyperparasitized Paralomis spinosissima, a lithodid found a
round South Georgia. The parasite is Briarosaccus callosus, a rhizocep
halan; the hyperparasite is an undescribed liriopsinine. Generalized a
dditive models were used to model data collected during 1992 and 1995.
Parasites were less prevalent on female crabs and most prevalent in s
ubmarine canyons. The former effect may be due to sex-specific differe
nces in gill cleaning behavior, and the latter effect may result from
reduced gill cleaning efficiency in areas where the seafloor is covere
d with fine sediment. The size-prevalence curve was dome-shaped; this
probably indicates that B, callosus reduces the growth rate and increa
ses the mortality rate of P. spinosissima. Parasite prevalence decreas
ed with increasing crab density but increased with increasing parasiti
zed crab density. These results suggest that parasitized crabs do not
aggregate with unparasitized males. Habitat was a significant source o
f variation in hyperparasite prevalence during 1992 but not 1995. Betw
een-year comparisons of parasite and hyperparasite prevalences were eq
uivocal and failed to provide convincing evidence against parameter st
ationarity and stability in the P. spinosissima-B, callosus-liriopsini
ne system.