448 



Fishery Bulletin 101(2) 



and Georgia 1997, was significant after correction for mul- 

 tiple testing (a<0.001). Single-locus population pairwise 

 Ffj.]. values for the RP2 locus were also low (mean=0.006), 

 ranging from F^-^ < 0.000 for most of the comparisons to 

 a high of 0.050 between Georgia 1997 and Delaware Bay 

 1997. None of the exact F permutation tests were signifi- 

 cant after correction for multiple testing. 



From our results we were unable to reject the null hy- 

 pothesis that weakfish comprise a single, genetically ho- 

 mogeneous stock. These results are consistent with those 

 based on allozymes (Crawford et al., 1989) and RFLP 

 analysis of mtDNA (Graves et al., 1992) and illustrate the 

 point that increased genetic variability in microsatellites 

 in relation to more traditional markers will not always 

 provide greater stock resolution (Seeb et al., 1998). The 

 amount of genetic exchange necessary to prevent the ac- 



cumulation of significant genetic divergence between fish 

 from different locations may be as little as a few individu- 

 als per generation (Allendorf and Phelps, 1981). Weakfish 

 tagging data indicate that low levels of exchange occur 

 between geographically distant populations of weakfish 

 (Bain et al., 1998). Estimates of natal homing in yearling 

 weakfish, calculated by Thorrold et al. (2001) using geo- 

 chemical signatures in the otoliths of the same weakfish 

 used in the present study, indicated spawning-site fidelity 

 ranging from SI'S to 81%, suggesting exchange rates suf- 

 ficient to prohibit genetic divergence between locations. 



The inclusion of nontarget species in our weakfish sam- 

 ples illustrates the advantages in using multiple marker 

 systems. If only a single microsatellite locus had been 

 used, or if the study had been restricted to nuclear intron 

 markers alone, it is very likely that the sand seatrout and 



