Gold et al.: Genetic studies of Scomberomorus cavalla in Florida 



497 



Tables 4 and 5). Fjg values (after Weir and Cockerham, 

 1984) for these four samples were all positive, indicating 

 a deficit of heterozygotes and the possible presence of a 

 null allele. However, probability values for tests of Hardy- 

 Weinberg equilibrium at Sca-23 among the remaining 16 

 samples were nonsignificant and averaged (±SE) 0.385 

 ±0.076. Although the possibility of a null allele at Sca-23 

 cannot be dismissed unequivocally, genotypes at Sca- 

 23 appear overall to be distributed in accordance with 

 Hardy-Weinberg equilibrium expectation. Finally, tests 

 of genotypic equilibrium between pairs of microsatellites 

 (samples pooled) yielded only one significant value {Sca- 

 44 X Sca-61) following Bonferroni correction (Table 3). 

 Probability values of tests involving these two microsatel- 

 lites, i.e. Sca-44 x Sca-61, carried out within each of the 20 

 samples were all nonsignificant and averaged (±SE) 0.338 



±0.067. Of the remaining (pairwise) tests carried out with- 

 in samples (420 tests in all), only four significant prob- 

 ability values were obtained: Sca-37 and Sca-44 (CCN), 

 Sca-23 X Sca-44 (PCY), Sca-37 x Sca-49 (BCG), and Sca-23 

 X Sca-44 (SEB^). These results indicate that genotypes at 

 the seven microsatellites are independent and hence are 

 not linked genetically. 



Tests for independence of allele distributions at each of 

 the seven microsatellites versus both sex and age of indi- 

 viduals were carried out 1 ) within each of the 20 samples, 

 and 2) among individuals (pooled) sampled from the At- 

 lantic, the Florida Keys, the Gulf, and overall. In tests for 

 independence with sex, eight significant probability values 

 (P<0.05) were found prior to Bonferroni correction for si- 

 multaneous tests (data available from first author). None 

 of these, however, were significant after Bonferroni correc- 



