Anderson: Systematics of the North American menhadens in the genus Brevoortia 



375 



significant genetic divergence between 

 the two menhaden types, and this dis- 

 tinction is reinforced by demographic 

 characteristics shared within these 

 groups. Specifically, overall census sizes 

 (Christmas and Gunter, 1960), species 

 ranges (Christmas and Gunter, 19601, 

 and migratory behavior (Gunter, 1945; 

 Simmons, 1957; Tolan and Newstead, 

 2004) are markedly different between 

 the small-scaled and large-scaled spe- 

 cies pairs, yet similar between species 

 within these pairs. The differences in 

 the demographics of small-scaled and 

 large-scaled menhaden are reflected 

 in nuclear-based estimates of A^^;*, 

 which are four to five times larger in 

 large-scaled species. Smaller popula- 

 tion sizes, and relatively shorter overall 

 coastal ranges, are reflected in the lower 

 genetic variability of small-scaled men- 

 haden compared to large-scaled menha- 

 den. In addition, nutritional and trophic 

 differences among these species in the 

 western Gulf have previously been doc- 

 umented (Castillo-Rivera et al., 1996; 

 Castillo-Rivera and Kobelkowsky, 2000) 

 and may represent mechanisms for the 

 divergence of population demographic 

 parameters among these groups. 



In both the present study and the 

 study of Bowen and Avise (1990), high 

 levels of genetic variation were detected 

 in the mtDNA genome of menhaden. In- 

 deed, because of the high mutation rate 

 of the mtDNA region surveyed in this 

 study, saturation was a likely source of 

 bias in comparisons among divergent 

 lineages. A second similarity between 

 this study and that of Bowen and Avise 

 (1990) is the presence of two divergent 

 large-scaled haplotype groups, one that 

 is confined to B. tyrannus, and a sec- 

 ond which includes haplotypes from 

 both species. Avise (1992) suggested 

 that a likely explanation for this pat- 

 tern is recent gene flow between these species, rather 

 than incomplete lineage sorting. A Bayesian population 

 assignment of nuclear genotypes seems to support the 

 possibility of admixture between these species, with 

 individuals from both species containing genetic signa- 

 tures of the complementary group. Epperly (1989, [and 

 references therein]), suggested the existence of two 

 subpopulations of B. tyrannus in the Atlantic: one on 

 the southern Atlantic coast (below 40''N) and a subpopu- 

 lation north of Long Island, New York. Juveniles from 

 these subpopulations differed in meristic and morpho- 

 logical characteristics, and also had different biochemi- 

 cal profiles (Epperly, 1989). Epperly (1989) suggested 

 that the spawning times for these two subpopulations 



I. Brevoortia tyrannus/ 

 patronus 



II. Brevoortia rsrawms 



T 



Figure 4 



Maximum likelihood plot of unrooted menhaden mtDNA samples. The 

 numbers at the base of each of three divergent clades (99, 100, and 100) 

 are the bootstrap support values from a neighbor joining analysis. The 

 genetic distance (D) between links in the phylogeny is a modified version 

 of Kimura's distance (1981), which includes gamma-distributed rate varia- 

 tion after invariant sites are removed, as outlined in Modeltest software 

 (Posada, 1998). The roman numerals correspond to clade numbers as 

 discussed in the text, including clade I (mixture of haplotypes from Gulf 

 menhaden [Brevoortia patronus] and Atlantic menhaden [B. tyrannus]K 

 clade II (haplotypes from B. tyrannus only), and clade III (mixture of 

 haplotypes from yellowfin menhaden, B. smithi and finescale menhaden, 

 B. gunteri). 



may be different, allowing for recruitment to occur in 

 these areas at different times of the year. Variance in 

 spawning times may be a mechanism for maintenance 

 of divergence between haplotype clades for large-scale 

 menhaden recovered in this study. However, it is likely 

 that the variation in morphology and biochemistry seen 

 in northern and southern forms of Atlantic menhaden 

 is affected by the influence of recurrent gene flow be- 

 tween this species and its Gulf cognate. Avise (1992) 

 demonstrated a broad-scale demographic break between 

 Atlantic and Gulf forms of both terrestrial and marine 

 vertebrates (including menhaden); however, in many 

 species, haplotypes representative of Gulf forms were 

 present in individuals collected in the southern Allan- 



