254 



Fishery Bulletin 90(2), 1992 



(GsL 0.0060, Glt 0.0050). The remaining 98.9% (Hg/ 

 Hx) was due to within-sample heterozygosity. 



Allele-frequency heterogeneity among samples was 

 significant (F<0.05) at several loci and was due to both 

 spatial and temporal components of variation (Table 

 4). Temporal variation at Ballast Key (P<0.001) and 

 Coffins Patch (P<0.01) was due primarily to variation 

 at MDH-2* and MDH-1*, respectively. On the other 

 hand, the heterogeneity among years at Craig Key 

 (P<0.05) was due primarily to the cumulative effects 

 of variation at IDH-2* and PGM-1*. Significant allele- 

 frequency variation also existed among localities, but 

 this latter variation did not exceed the temporal varia- 

 tion within localities as measured by F-ratio com- 

 parisons (P>0.05) at each locus. 



Comparisons with Caribbean Sea 

 and Bermuda populations 



Allele frequencies dX DPEP-1*, MDH-2*, PGM-1*, and 

 PGDH* for the Florida Keys and Bimini populations 

 of S. gigas can be compared directly with those for 

 populations sampled by Mitton et al. (1989). In that 

 previous study, queen conch were collected from 16 

 sampling sites representing eight major localities 

 throughout the Caribbean Sea area (Fig. 1). In addi- 

 tion, conch were collected from one site in Bermuda. 

 Patterns of genetic variation among populations in 

 the Caribbean Sea and Bermuda were similar to those 



for populations in the Florida Keys and Bimini (Table 

 5). Total gene diversities (Hj) for the two groups of 

 populations were essentially equal (0.355 and 0.354, 

 respectively). However, the diversity within and among 

 localities was somewhat greater for Caribbean Sea and 

 Bermuda populations (Glt 1-69%, Gsl 1.14%) than for 

 populations from the Florida Keys and Bimini (Glt 

 0.39%, Gsl 0.68%). This latter result might be ex- 

 pected considering the relative geographic scales over 

 which populations were sampled in the two studies 

 (Fig. 1). In this context, summing Glt ^^d Gsl for 

 Florida Keys and Bimini populations yields a percent- 

 age of gene diversity (1.07%) that is approximately 

 equal to Gsl (sites within localities) for the Caribbean 

 Sea and Bermuda populations (1.14%). 



A dendrogram based on Nei's index of gene identity 

 clearly reflected the high genetic similarity among 

 populations of S. gigas (Fig. 2). The average gene iden- 

 tity (based on the four aforementioned loci) among 

 populations sampled by Mitton et al. (1989) was 0.984, 

 among those sampled here was 0.993, and between 

 populations (samples) of the two studies was 0.988. 

 Twenty-three of these populations clustered together 

 at the 0.99 level or above. The Bermuda population and 

 the 1987 Ballast Key population (sample) formed a 

 separate subcluster, due primarily to divergent allele 

 frequencies at MDH-2*. The Vieux Fort (St. Lucia) and 

 Six Hill Cay (Turks and Caicos Islands) populations also 

 clustered separately, due primarily to slightly divergent 



