294 



Fishery Bulletin 102(2) 



growth rate, reproductive capacity, and maturity sched- 

 ules related to temperature (Brander, 1994). Cod within 

 our study zone generally avoid water temperatures greater 

 than 10°C, but Nantucket Shoals cod are abundant in 

 temperatures as warm as 15°C (Klein-MacPhee, 2002). 

 This differential thermal tolerance may support genetic 

 structuring of Nantucket Shoals cod by selecting against 

 individuals from other areas. 



Closely related gadid species such as cod and haddock 

 may exhibit similar patterns of population genetic struc- 

 turing associated with similar life histories, selective 

 pressures, and ecological constraints. Our results are 

 concordant with a previous study suggesting that had- 

 dock from Browns Bank and Georges Bank are genetically 

 similar and that haddock from Nantucket Shoals are dis- 

 tinct (Lage et al., 2001). However, Ruzzante et al. (1998) 

 observed significant genetic differentiation between cod 

 from Browns Bank and Georges Bank. Our results do not 

 agree with this previously observed heterogeneity between 

 Browns Bank and Georges Bank and may be due to the 

 examination of different loci, different sampling compari- 

 sons, or small sample sizes used in both studies (or to a 

 combination of these variables) (Ruzzante, 1998; Smouse 

 and Chevillon, 1998). 



Among loci, the greatest genetic differentiation was 

 observed at locus Gmol32. Indeed, observed statistical 

 significance of population differentiation and F ST depends 

 entirely on Gmol32. Length variation at Gmol32 is a func- 

 tion of mutations in the repetitive array and of an indel in 

 a flanking region (Ruzzante et al., 1998) causing bimodal 

 allele distributions in some populations. When compared to 

 other microsatellite loci, Gwol32 has shown the greatest 

 differentiation among other Northwest Atlantic cod popu- 

 lations (Bentzen et al., 1996; Ruzzante et al, 1998, 2001) 

 and among Northwest Atlantic haddock populations (Lage 

 et al., 2001) by an order of magnitude. Other loci examined 

 have not shown similarly strong measures of population 

 structuring. Observed genetic structuring may be due to 

 forces currently determining regional larval and adult 

 distributions, including bathymetry and oceanographic 

 patterns. However, because similar genetic structuring 

 is not observed at all loci, another potential explanation 

 is that structuring at Gmol32 is due to forces that acted 

 during the formation of populations rather than to forces 

 presently maintaining strong reproductive isolation. Once 

 genetic structure was generated during the formation of 

 these populations subsequent to the last ice age. biological 

 and oceanographic forces may have maintained such struc- 

 ture; other loci may show an absence of structure simply be- 

 cause it may not have been present when populations were 

 formed. Pogson et al. (2001) reported that the recent age 

 of populations, rather than extensive gene flow, may be re- 

 sponsible for weak population structure in Atlantic cod, and 

 that interpreting limited genetic differences among popula- 

 tions as reflecting high levels of ongoing gene flow should 

 be made with caution. This suggests that the observed lack 

 of heterogeneity between Browns Bank and Georges Bank 

 may not be due to high levels of ongoing gene flow, but to 

 similarities between recently generated populations main- 

 tained by small but adequate levels of gene flow. 



Alternatively, significant structuring associated with 

 Gmol32 in both cod and haddock may suggest that selec- 

 tion is acting at this or at a linked locus. Although micro- 

 satellites themselves may be generally considered neutral, 

 there is, in theory, potential for physical linkage or drift- 

 generated linkage disequilibrium between microsatellite 

 and functional loci. There is however, recent evidence of 

 selection acting directly on microsatellite loci in tilapia in 

 high-salinity environments. Streelman and Kocher ( 2002 ) 

 found a strong functional genotype-environment interac- 

 tion and suggested that microsatellite repeats of varying 

 length might induce promoter conformations that differ in 

 their capacity to bind transcriptional regulators. A poten- 

 tial selective mechanism to support the observed genetic 

 structuring of Nantucket Shoals cod (and haddock) may 

 be differential thermal tolerance, although this hypothesis 

 remains untested. 



There is strong evidence for an unusual mix of balanc- 

 ing and directional selection at the pantophysin (Pa«I ) lo- 

 cus in cod but no evidence of stable geographically varying 

 selection among North Atlantic populations ( Pogson, 2001; 

 Pogson et al., 2001). In the present study, the Paul locus 

 showed little variation and no significant genetic structur- 

 ing (Table 1). The observed lack of geographic structuring 

 at Panl provides no evidence for local adaptation. How- 

 ever, our observations may be due to strong balancing 

 selection among the geographically proximate populations 

 examined or, if Panl is not under selection, insufficient 

 variation to resolve genetic structure. Alternatively, this 

 observed lack of genetic divergence at Panl could be due 

 to similarities among recently generated populations of 

 North Atlantic cod. 



Our research suggests that the cod spawning on Nan- 

 tucket Shoals are genetically differentiated from cod 

 spawning on Browns Bank and Georges Bank. Managers 

 may wish to consider Nantucket Shoals cod as a separate 

 stock for assessment and management purposes in light of 

 current practices that combine Georges Bank with regions 

 to the south as one management unit. Cod from within the 

 Gulf of Maine can potentially migrate along the coast to 

 Nantucket Shoals where there is little geographic barrier 

 to adult movement. If this is true, the Nantucket Shoals 

 sample that we analyzed may actually be representative of 

 a mixed Gulf of Maine and Nantucket Shoals population. 

 Additional analyses are needed to evaluate the hypothesis 

 that Nantucket Shoals cod are genetically distinct from 

 cod spawning within the Gulf of Maine. Further studies 

 should address the issues of temporal stability and robust 

 sampling and should incorporate cod samples from within 

 the Gulf of Maine. 



Acknowledgments 



We thank three anonymous reviewers for their insightful 

 comments. We thank the captains, crews, and scientists 

 from the Canadian Department of Fisheries and Oceans 

 and the United States. National Marine Fisheries Service. 

 In particular, we would like to thank Chris Taggart. Nina 

 Shepard, and Holly McBride for assistance in obtaining 



