Campton et al . Genetic patchiness among Strombus gigas populations 



255 



Table 5 



Percentages of total gene diversity (H^.) among localities 

 (Glt). among samples and sites within localities (Gsl), and 

 within samples and sites (Hj/Hj) for populations of queen 

 conch Strombus gigas from the Florida Keys and Bimini (this 

 study) and from the Caribbean Sea and Bermuda (Mitton et 

 al. 1989). Data represent the means for DPEP-l*, MDH-2*, 

 PGM-1*, PGDH*. 



Gtene diversity (%) 



Populations 



Ht 



Ho/Ht. 



Florida Keys 

 and Bimini 



Caribbean Sea 

 and Bermuda 



All populations 



0.354 



0.355 

 0.354 



0.39 



1.69 

 1.24 



0.68 



1.14 

 1.01 



98.94 



97.17 

 97.75 



allele frequencies at PGDH* and DPEP-l*, respective- 

 ly (Mitton et al. 1989). 



In summary, populations of S. gigas are very similar 

 genetically and do not appear to be structured geo- 

 graphically. However, those populations cannot be con- 

 sidered totally panmictic. 



Discussion 



Population structure 



Benthic marine invertebrates with planktonic larvae 

 often exhibit spatial and temporal genetic variation 

 similar to that described here for S. gigas. For exam- 

 ple. Watts et al. (1990) found significant allele-fre- 

 quency variation among three populations of sea urchin 

 Echinometra mathaei separated by only 4 km. More- 

 over, that heterogeneity over a 4 km distance was ap- 

 proximately equal to the genetic heterogeneity among 

 populations separated by over 1300km. Those inves- 

 tigators also detected significant allele-frequency varia- 

 tion among year-classes within each of the three micro- 

 spatial sample sites. Similar patterns of heterogene- 

 ity were reported for the limpet Siphonaria jeanae 

 (Johnson and Black 1982, 1984ab) and seastar Acan- 

 thaster planei (Nash et al. 1988, Nishida and Lucas 

 1988). 



Significant microspatial genetic heterogeneity, 

 despite high macrospatial genetic similarity, has been 

 termed "genetic patchiness" (e.g., Johnson and Black 

 1984b). Such genetic patchiness could be due to either 

 postsettlement natural selection or genetic hetero- 

 geneity among groups of recruits that are spatially or 

 temporally separated (Watts et al. 1990). Under both 

 hypotheses, planktonic dispersal is believed to main- 



tain high genetic similarity among populations over 

 broad geographic areas. However, under postsettle- 

 ment natural selection, one would expect genetic varia- 

 tion among localities to be greater than the temporal 

 variation within localities because of local adaptation. 

 Conversely, under the model of presettlement genetic 

 heterogeneity, spatial and temporal components of 

 genetic variation are expected to be equal because the 

 population structure would result from presettlement 

 events that were independent of the specific localities 

 at which settlement occurred. Under this latter model, 

 spatial heterogeneity among localities would simply 

 reflect the temporal heterogeneity within localities. 



Results obtained here for S. gigas are most consis- 

 tent with the presettlement hypothesis of genetic 

 patchiness. Populations of queen conch throughout 

 their geographic range are very similar genetically, yet 

 spatial and temporal components of genetic variation 

 appear significant and approximately equal. Mitton et 

 al. (1989) obtained similar results for macrospatial 

 (among-locality) and microspatial (within-locality) com- 

 ponents of genetic variation. These results suggest a 

 dynamic population structure in which allele-frequency 

 heterogeneity may exist among groups of recruits that 

 settle in different years at the same locality or at dif- 

 ferent localities in the same year. Johnson and Black 

 (1982, 1984ab) and Watts et al. (1990) reached similar 

 conclusions regarding genetic patchiness among pop- 

 ulations of limpet and sea urchin, respectively. 



Several mechanisms can be invoked to explain genet- 

 ic patchiness due to presettlement events. Johnson and 

 Black (1984ab) and Watts et al. (1990) suggest that 



