Graves et al,: Stock structure of Pomatomus saltatnx along the mid-Atlantic coast 



709 



conspecific populations has been noted between Atlan- 

 tic and Pacific blue marlin (Graves and McDowell, 

 unpubl. data). The striking lack of variation within the 

 Australian sample could be the result of a smaller 

 effective population size of females resulting from 

 population bottlenecks, or may simply reflect a period 

 of isolation sufficient for the sorting of gene trees (Nei 

 1987, Avise et al. 1988, Chapman 1990, Bowen and 

 Avise 1990). 



We found little evidence to support the hypothesis 

 that genetically distinct stocks of bluefish exist along 

 the mid-Atlantic coast. Although appreciable mean 

 nucleotide sequence divergences were found between 

 sampling locations (Table 4), when corrected for within- 

 group variation the values became extremely small, in- 

 dicating that most of the observed differentiation could 

 be accounted for by variation within the samples. The 

 lack of population structuring was also supported by 

 the homogeneous distribution of all genotypes and the 

 fact that the level of genetic divergence among sam- 

 pling locations was not appreciably greater than the 

 level of divergence among samples taken at any one 

 location in different years. 



The extent of gene flow among populations can also 

 be inferred from the frequency distribution of rare 

 alleles (Slatkin 1989). An inspection of Table 2 indicates 

 that almost all mtDNA genotypes that occurred more 

 than once were found in different collections, sug- 

 gesting significant gene flow among sampling loca- 

 tions. For example, the genotype AAAABAABA, 

 which was present in three individuals, occurred in the 

 VA89, NC88, and NJ90-Su collections. An exception 

 to this pattern was presented by the genotype 

 AAAAAAAAD, which occurred seven times: in six 

 individuals of the VA88 sample and one individual of 

 the VA89 sample. However, an examination of bluefish 

 mtDNA genotypes not included in this analysis— be- 

 cause the individuals were greater than one year old, 

 or because they came from a sample that was too small 

 for inclusion in this analysis— suggests that the ob- 

 served distribution of the AAAAAAAAD genotype 

 may be an artifact of sampling error. The genotype was 

 present in two bluefish collected in 1988 (one in New 

 York and one in Connecticut) and in six bluefish col- 

 lected in 1989 (two in New York, two in Virginia, and 

 two in North Carolina). 



In contrast to the genetic similarity among mid- 

 Atlantic samples, a large, consistent genotypic dif- 

 ference was noted between the mid-Atlantic bluefish 

 and a conspecific population in Australia. The corrected 

 mean nucleotide sequence divergence of almost 2% is 

 more than an order of magnitude larger than the values 

 detected among mid-Atlantic samples, and is similar 

 to values reported between northwest Atlantic and 

 Barents Sea capelin populations (Dodson et al. 1991) 



or among populations of freshwater fishes of different 

 river systems (Bermingham and Avise 1986). 



While significant genetic differentiation was found 

 between mid- Atlantic and Australian bluefish, no 

 major differences were detected between mid- Atlantic 

 bluefish and a small sample from the Gulf of Mexico. 

 Consistent restriction-site differences have been 

 reported between Gulf of Mexico and mid-Atlantic 

 populations of a number of marine organisms, including 

 horshoe crabs Limulus polyphemus (Saunders et al. 

 1986), oysters Crassostrea virginica (Reeb and Avise 

 1990), and black sea bass Centropristis striata (Bowen 

 and Avise 1990). These preliminary results suggest that 

 bluefish from the Gulf of Mexico and the mid- Atlantic 

 are not as genetically isolated as many other coastal 

 marine species, although much larger samples will have 

 to be surveyed to determine if significant mtDNA 

 genotypic frequency differences exist between the two 

 areas. Considering the high vagility of bluefish and 

 their continuous distribution around Florida, this result 

 is not unexpected. 



The lack of significant genetic differentiation be- 

 tween spring- and summer-spawned bluefish is consis- 

 tent with the results of Chiarella and Conover (1990), 

 who found no correlation between the season in which 

 an adult bluefish spawned and the hatch-date of an in- 

 dividual. These data suggest that the bimodal distribu- 

 tion of YOY bluefish in mid-Atlantic estuaries results 

 from two major spawning events of the same popula- 

 tion of bluefish, rather than the participation of dif- 

 ferent stocks. The morphological differences found 

 between spring- and summer-spawned bluefish are 

 probably ecophenotypic, resulting from early-life- 

 history development in appreciably different environ- 

 ments. Similar morphological plasticity has been dem- 

 onstrated in many other marine fishes (Barlow 1961). 



The high degree of genetic homogeneity detected 

 within mid- Atlantic bluefish is also consistent wath the 

 results of tag and recapture studies. While many 

 bluefish return to the same site for several years (Lund 

 and Maltezos 1970), migratory habits appear to change 

 with age (Wilk 1977). Thus, the potential exists for con- 

 siderable interchange, and it is important to note that 

 even small levels of exchange can prevent the accumu- 

 lation of genetic differentiation (Hartl 1988). 



The results of this study cannot disprove the null 

 hypothesis that bluefish along the mid- Atlantic coast 

 share a common gene pool. There appears to be suffi- 

 cient gene flow to prevent the accumulation of even 

 slight genetic differences. Determining the magnitude 

 of exchange between geographic regions would require 

 an extensive tag and recapture program. Until such 

 data are available, the resource should be managed as 

 assumed in the Fishery Management Plan for the 

 Bluefish— as a single, genetically homogeneous stock. 



