Saillant and Gold Population structure and variance effective size of Lutianus campechonus in ttie norttiern Gulf of Mexico 



141 



note the following. First, it is possible that gene flow 

 among present-day red snapper in the northern Gulf is 

 limited but there has been insufficient time for semi- 

 isolated lineages to completely sort into monophyletic 

 assemblages. Pruett et al. (2005), on the basis of re- 

 sults of nested-clade analysis of mtDNA haplotypes ob- 

 tained from representative samples of the same cohorts 

 (and localities) studied in the present study, hypoth- 

 esized that semi-isolated assemblages of red snapper 

 in the northern Gulf may exist over the short term, 

 yet over the long term comprise a larger metapopu- 

 lation tied together by periodic gene flow. Similarity 

 in allele frequencies of genetic markers (such as used 

 here and in previous studies of red snapper) presumed 

 to be neutral to natural selection in theory could be 

 maintained in such a metapopulation during periods 

 when gene flow was limited or even absent. Second, all 

 the genetic markers studied to date are presumed to 

 be selectively neutral and to be affected primarily by 

 the interaction(s) between gene flow and genetic drift. 

 Genes affecting life-history and other traits that are in- 

 fluenced by natural selection need not necessarily follow 

 the same pattern(s), and geographic differences in adap- 

 tively useful alleles at such genes can be maintained 

 even in the face of substantial gene flow (Conover et 

 al., 2005). It is thus not implausible that red snapper 

 across the northern Gulf could differ in allele frequency 

 at adaptively useful genes yet be homogeneous at selec- 

 tively neutral ones. 



Contemporaneous effective size (N^^y) and present-day 

 demographic dynamics 



Estimates of contemporaneous or variance effective 

 size {N^y) for the Texas and Alabama localities (-1100) 

 were essentially the same, but were at least an order of 

 magnitude less than the N^,y estimate (>75,000) for the 

 Louisiana locality. These estimates reflect differences 

 in effective population size under the assumption that 

 no immigration into a locality has occurred during the 

 study interval, an assumption at odds with the general 

 absence of allele-frequency heterogeneity among locali- 

 ties as well as the low estimates of F>,.j. between pairs 

 of samples. Short-term immigration (within the time 

 interval of the study) could increase the variance in 

 allele frequencies, thus resulting in an overestimate of 

 A^^;. (Wang and Whitlock, 2003); whereas longer-term 

 immigration at a (more-or-less) constant rate from a 

 source population would have the opposite effect. The 

 observed differences among localities could thus reflect 

 differences in effective sizes, differences in patterns and 

 intensity of immigration, or both. Temporal variation in 

 allele frequencies also could occur if only a fraction of 

 potential spawners at a locality actually contributed to 

 recruitment and if such "temporal" subpopulations dif- 

 fered in allele frequencies between years. Regardless, 

 the differences in 7V,^• may indicate that different demo- 

 graphic dynamics currently exist among localities. 



Wang and Whitlock (2003) recently extended previous 

 maximum-likelihood methods to allow simultaneous es- 



timation of N^.y and m (rate of migration), provided data 

 from multiple loci were available and all sources of im- 

 migrants into a focal population were known. Because 

 of the latter, we were able to generate estimates of N^.y 

 and m only for the sample from the Louisiana locality 

 (focal population), using the samples from the Texas 

 and Alabama localities as source populations. Surpris- 

 ingly, the estimate of N^,y for the Louisiana sample 

 (4887, 95% confidence intervals of 1543 and 31,254) was 

 at least -15 times smaller than the estimate based on 

 no migration; m was estimated to be 0.0097 (95% confi- 

 dence intervals of <0.001 and 0.0355). Clearly, more ex- 

 tensive sampling across the northern Gulf is warranted 

 to obtain estimates of N^.y and ;?; at other localities and 

 to place this finding into perspective. 



Effective size (/Vp)/census size(AV) ratios 



Estimates of N^,y for all three sample localities were two 

 or more orders of magnitude less than the current, pre- 

 liminary estimates of adult census size (7.8-11.7 million) 

 across the northern Gulf (Cowan-^; Porch^). Given that 

 empirically derived N^JN ratios from a variety of verte- 

 brates are 0.10-0.11 on average (Frankham, 1995), this 

 result is somewhat surprising in that red snapper have a 

 long reproductive life-span and overlapping generations 

 (Wilson and Nieland, 2001), life-history features that 

 are expected to increase N^./N by limiting variance in 

 lifetime reproductive success among individuals (Jorde 

 and Ryman, 1995; Waite and Parker, 1996). The issue 

 is of importance in that census sizes of many commer- 

 cially exploited marine fish populations are generally 

 orders of magnitude larger than sizes where genetic 

 resources might be lost (Franklin, 1980; Schultz and 

 Lynch, 1997). However, species or populations with 

 exceedingly small N^JN ratios potentially could be in 

 danger of losing genetic resources, resulting in reduced 

 adaptation and population productivity (Hauser et al., 

 2002). In addition, low N^JN ratios may explain in part 

 why there often is a poor relationship between spawn- 

 ing stock size and recruitment (Hauser et al., 2002). 

 To date, N^,/N ratios smaller than 10" ' have been found 

 for four other exploited marine fish species (Hauser et 

 al., 2002; Turner et al., 2002; Hutchinson et al., 2003; 

 Gomez-Uchida and Banks^M. 



Factors that theoretically can lower genetic effec- 

 tive size with respect to census size include fluctuating 

 adult number and year-class strength (Hedgecock, 1994; 

 Vucetich et al., 1997), and variance in reproductive suc- 

 cess. The latter can arise from biased sex ratio, high 

 variance in male or female reproductive success, vari- 

 ance in productivity among habitats, or any combina- 

 tion of these factors (Nunney, 1996, 1999; Whitlock and 

 Barton, 1997). Virtually any of these factors could lower 

 N./N ratios in red snapper. Biased sex-ratio, however. 



" Gomez-Uchida. D., and M. A. Bank.s. 2003. Oregon State 

 Univ., Hatfield Marine Science Center, Department of Fish- 

 eries and Wildlife, 20.30 SE Marine Sci. Dr. Newport, OR 

 97365. 



