Merkouris et al.: Genetic diversity in Chionoecetes bairdi and C opilio 
527 
Tissues (muscle, gill, hepatopancreas, and heart) 
were dissected from each individual, placed in a la- 
belled tube, which was chilled on wet ice, capped, 
and frozen at -15°C (1989-90 collections) or in liq- 
uid nitrogen (1991-93 collections). Freezing gener- 
ally occurred within 20 minutes after dissection but 
in some cases was as long as 1 hour. Tissues were 
transported to the laboratory on dry ice or liquid ni- 
trogen and stored at -80°C until analysis. Several 
locations were sampled in multiple years. 
Allozyme electrophoresis 
Procedures for horizontal starch gel electrophoresis 
followed those of Harris and Hopkinson (1976) and 
Aebersold et al. (1987). Activity reflecting 34 pre- 
sumed loci were resolved with the following buffers 
(Table 2): 1) N-(3-aminopropyl)-morpholine, citrate 
(AC, pH 6.1, 6.9) (Clayton and Tretiak, 1972); 2) Tris, 
borate, citrate, lithium hydroxide (TBCL, pH 8.7) 
(Ridgway et al., 1970); 3) Tris, citrate (TC, pH 7.0) 
(Shaw and Prasad, 1970); 4) Tris, citrate (TC, pH 
8.0) (Selander et al., 1971); and 5) Tris, borate, EDTA 
(TBE, pH 8.5) (Boyer et al., 1963). Gene nomencla- 
ture followed Shaklee et al. (1990). Allelic standards 
were used to compare relative mobilities between 
species. 
Genetic differentiation 
We estimated allele frequencies, calculated average 
observed heterozygosities, and tested conformation 
of genotype frequencies to Hardy-Weinberg expected 
frequencies with log-likelihood ratios (modified from 
Weir, 1990) for each collection (a=0.05, adjusted for 
multiple tests [Rice, 1989] ). Samples ofC. bairdi col- 
lected from the Bering Sea and Pribilof Island areas 
were pooled because sample sites were overlapping. 
Interannual heterogeneity of multiple-year collec- 
tions in Bristol Bay, Bering Sea, St. Matthew Island, 
Pribilof Islands, and Seymour Canal was tested by 
using log-likelihood statistics (Sokal and Rohlf, 1995). 
Multiple-year collections within a site were pooled 
for further analyses if no significant heterogeneity 
existed (P<0.01). We estimated degree of population 
subdivision by using Wright’s ( 1978) nonhierarchical 
F statistics and a hierarchical log-likelihood analy- 
sis. We used FSTAT (version 1.2)(Goudet, 1995) to 
