A Mitochondrial DNA Analysis of the Species Status of the Endemic Waccamaw Darter, Etheostoma perlongum
The Waccamaw Darter Etheostoma perlongum is one of three teleost fish species and several molluscs endemic to Lake Waccamaw, a large Carolina Bay in southeastern North Carolina. We assessed species status by sequencing most of the mitochondrial cytochrome b (cyt b) gene in E. perlongum and its putative sister species, E. olmstedi, from the adjoining Waccamaw River and streams in the nearby Pee Dee and Cape Fear Drainages. Phylogeographic analysis shows strongly supported river- and drainage-specific clades with divergence times dating to the Pleistocene. Haplotypes of Etheostoma perlongum are very similar and not phylogenetically distinct from sequences from the Waccamaw River population of E. olmstedi. However, strong haplotype frequency differentiation shows that free genetic exchange does not occur between the river and lake, and along with life history distinctiveness, suggests reproductive isolation despite contact between these populations. Coalescent simulations showed that incomplete lineage sorting between reproductively isolated populations is a plausible explanation for the failure of cyt b sequences to have achieved reciprocal monophyly. Gene tree/species tree discord equal to that seen in our cyt b data was produced in ≥ 5% of coalescent trees, using lake-river divergence times of T = 5–10 thousand years ago (kya), no gene flow and Ne = 105; simulations with more realistic values of Ne = 104 and T = 10–20 kya required recurrent lake-river migration to generate comparable levels of incomplete sorting of haplotypes. Studies of behavioral reproductive isolation and multilocus studies of nuclear DNA should be used to distinguish these scenarios and to further evaluate the species status of the endemic Waccamaw Darter.Abstract
Sampling locations for Etheostoma perlongum (Lake Waccamaw) and E. olmstedi (other sites). (A) Map of the sampled regions of the three watersheds. Site in the Cape Fear watershed: BR = Black River. (B) Sites in the Pee Dee watershed: PD = Pee Dee River, LPD = Little Pee Dee River, LR = Lumber River. (C) Sites in the Waccamaw watershed: WR = Waccamaw River (1: Old Dock bridge, 2: Seven Creeks, 3: Norris Lake Rd. bridge on Camp Swamp Creek, 4: Buck Creek); LW = Lake Waccamaw (1: NCWRC access, 2: North Shore, 3: near dam on Waccamaw River outlet).
Maximum likelihood phylogeny for cytochrome b haplotypes of Etheostoma perlongum and E. olmstedi cytochrome b haplotypes. The tree is one of two ML trees with equal likelihood scores (ln L = 2432.15). Numbers above branches are bootstrap support values > 70% (500 replications); upper values from fast-stepwise and lower values from NJ bootstrap. Distances are based upon the TrN+G model. Haplotype codes (Table 1) on branch termini are followed with sample size in italics when shared by multiple individuals and by site codes when present in > 1 population.
Coalescent simulations. Each figure plots the distribution of the number of deep coalescents, a measure of discord between a reciprocally monophyletic species tree and 1000 gene trees simulated under a neutral coalescent model. (A) Simulations performed over a range of divergence times between river and lake populations (T in generations), with base effective population size Ne = 100,000, and no migration. (B) Simulations performed over a range of migration rates (per individual per generation) at Ne = 10,000 and T = 10,000 generations. Arrows show observed number of deep coalescents between the species tree and the empirical cyt b tree.
Contributor Notes
Associate Editor: J. M. Quattro.