NOTE Cordes and Graves: Hyridization and stock structure in Cynosaon regalis 



447 



and two alleles 171 bp in size). All other individuals were 

 putative hybrids with weakfish mtDNA and a single small 

 SOC050 allele characteristic of silver and sand seatrout. As 

 mentioned previously, subsequent analysis of 20 weakfish 

 taken from each of the four locations outside of Georgia 

 with the 12S/16S marker revealed only weakfish mtDNA. 

 If the small SOC050 alleles found in the more northern 

 populations are not simply rare weakfish alleles shared in 

 common with the other two Cynoscion species, they may 

 indicate that introgressive hybridization is responsible for 

 the migration of the smaller alleles into northern waters 

 (although the northward movement of hybrid fish out of 

 the contact zone cannot be excluded). As a result of these 

 findings, all individuals in the 1996 and 1997 collections 

 exhibiting at least one small SOC050 allele less than 183 

 bp in length were eliminated from the population structure 

 analyses. 



Stock structure analysis 



All four microsatellite loci were polymorphic in all sampled 

 locations in both years. Allele frequency distributions for 

 each locus are available from the authors upon request. 

 Sample sizes in), number of alleles (A''), expected hetero- 

 zygosities (gene diversities), and significance test results 

 for Hardy-Weinberg equilibrium are provided in Table 4. 

 Levels of variation differed greatly among the four mic- 



rosatellite loci. The number of alleles ranged from two 

 (SOC014) to 37 (CNE612), and average expected heterozy- 

 gosities ranged from 0.085 (SOC014) to 0.928 (CNE612). 

 These values are consistent with heterozygosity ranges 

 reported in other multilocus microsatellite studies on spe- 

 cies including Atlantic cod (Bentzen et al., 1996), northern 

 pike (Miller and Kapuscinski, 1996), pink and sockeye 

 salmon (Seeb et al., 1998), and Arctic char (Brunner at 

 al., 1998). In contrast, Crawford et al. (1989) and Graves 

 et al. (19921 found very low levels of genetic variation in 

 an analysis of weakfish populations using allozymes and 

 mtDNA restriction fragment-length polymorphism (RFLP) 

 analyses, respectively. None of the genotypic distributions 

 for any of the four microsatellite loci at any of the collection 

 locations in either year differed significantly from Hardy- 

 Weinberg expectations after correcting for multiple tests 

 (Table 4). 



Digestion of actin intron (CRESIAl) amplifications with 

 the restriction endonuclease Rsa I revealed a single poly- 

 morphic restriction site that produced two alleles. Expected 

 heterozygosities ranged from 0.000 for the monomorphic 

 Georgia 1997 sample to 0.096 for the Chesapeake Bay 

 1996 sample. Digestion of the RP2 amplifications with 

 the restriction endonuclease Hinf I also resulted in two 

 alleles. Expected heterozygosities ranged from 0.194 in the 

 Delaware Bay 1997 sample to 0.370 in the Georgia 1997 

 sample. Levels of genetic variation within the two nuclear 

 gene intron regions were low in relation to three of the four 

 microsatellite loci and were more similar to those found in 

 the polymorphic allozyme loci of Crawford et al. ( 1989). In 

 another study where nuclear intron RFLP analysis was 

 used, similar levels of heterozygosity in Pacific salmon 

 were found (Moran et al., 1997), as in RFLP studies of 

 anonymous single copy nuclear (ascn) DNA loci in Atlantic 

 cod {Gadus morhua) (Pogson et al., 1995) and blue marlin 

 {Makaira nigricans) (Buonaccorsi et al., 1999). In con- 

 trast, higher heterozygosities (44-58%) were reported in 

 an ascnDNA/RFLP analysis of striped bass {Morone saxa- 

 tilis) by Leclerc et al. (1996). None of the sample genotype 

 distributions for either locus differed significantly from 

 Hardy-Weinberg expectations after correcting for multiple 

 tests (Table 4). 



To test for population structure, microsatellite loci were 

 analyzed individually and as a combined data set. AMOVA 

 results did not reveal significant differences between sam- 

 ple locations or years for any of the four loci or for the com- 

 bined data (all P>0.05). Single-locus population pairwise 

 Fgy values were relatively low, and mean Fj^.j. values ranged 

 from 0.002 (SOC050 and CNE612) to 0.018'(SOC044). Exact 

 F permutation tests were not significant for any of the four 

 loci or the combined data set after correction for multiple 

 testing. 



AMOVA results for both the actin and RP2 loci indicated 

 no significant differences between sample locations or years 

 (all P>0.05). Single-locus population pairwise F^^ values 

 for the actin locus were consistently low (mean=0.005), 

 ranging from Fgj < 0.000 for most of the comparisons to 

 an Fg-p of 0.035 between Georgia 1996 and Georgia 1997 

 and between Chesapeake Bay 1996 and Georgia 1997. 

 A single exact F permutation test, between Delaware 1996 



