GULF STREAM SYSTEM LASAEA
putative ancestral haplotype for both regional western North Atlantic populations.
Phylogenetic trees represent inferred historical relation- ships, and our CO I dataset, when rooted with BDA1, reveals a countercurrent topology of source (Bermuda) and founder (Florida) populations. Countercurrent gene-flow patterns have also been recently inferred for a variety of marine invertebrates (Palumbi et al., 1997; Palumbi, 1997; Benzie, 1999; Lessios et al., 1999; O´ Foighil and Jozefo- wicz, 1999) and may indicate that contemporary gene flow is insufficient in these cases to obscure genetic structuring produced under previous current regimes (Benzie, 1999). However, it is difficult to envisage a plausible countercur- rent dispersal mechanism for the Floridian and Bermudan study populations. Although it has fluctuated in intensity, the Florida Current has apparently been a persistent feature of the North Atlantic Gyre (Keffer et al., 1988; Lynch- Stieglitz et al., 1999; Duplessy, 1999). In addition, unstud- ied Caribbean populations are much more plausible sources of Floridian lineages, and diagnostic genetic differences among island and mainland populations for two mitochon- drial genes (COI and 16S [O´ Foighil and Jozefowicz, 1999]) rule out historical (anthropogenic) countercurrent gene flow.
a population founded by one ancestral type (presumably BDA1) that has experienced a phase of rapid growth, a process associated with lowered stochastic elimination of novel and rare lineages (Avise et al., 1984; Slatkin and Hudson, 1991). Further support for this demographic inter- pretation of the Bermudan population was obtained when a lineages-through-time analysis (Nee et al., 1996) (not shown) yielded profiles consistent with a historically grow- ing population. However, nucleotide mismatch distribu- tion analysis of the Bermudan population using Arlequin (Schneider et al., 1999) failed to meet the sudden-expansion model (C. Cunningham, Duke University, pers. comm.), implying that the primary wave of growth for this oceanic island population may have occurred soon after the initial founding event. Differential historical rates of loss/gain of novel haplotypes in the two regional populations could act to displace ancestral haplotypes from central topological positions in the western Atlantic clade. For instance, if we were to assume that the mainland lineage FL1 represents the “true” ancestral haplotype in our western Atlantic data set, its noncentral position in the tree (Fig. 2b) might result from the repeated pruning of rare novel Floridian mt genotypes, one step removed from FL1, by sequential episodes of mainland population constrictions.
It may be pertinent that the low values observed for genetic diversity are consistent with an evolutionarily recent origin for both regional Lasaea populations. The most de- tailed fossil-calibrated estimates of molluscan gene diver- gence are provided by Collins et al. (1996) for the marine snail Nucella (2%/myr/lineage for [predominantly synony- mous] third codon transitional differences). The central mitochondrial genotype, BDA1, is a maximum of two nu- cleotide substitutions removed from all of the other mem- bers of the western Atlantic Lasaea clade. Although complicated by phylogenetic and other potential biases, application of the Collins et al. (1996) rate yields a crude maximum age estimate of 0.229 Ma for the entire western Atlantic clade and also for the Bermudan population. In- trapopulational allelic relationships are modulated by demo- graphic history (Slatkin and Hudson, 1991; Nee et al., 1996; Lavery et al., 1996), and it is possible that the observed patterns of CO I gene variation stem from divergent demo- graphic processes experienced by these two evolutionarily recent regional populations, rather than from a countercur- rent pattern of gene flow.
Application of coalescence theory assumes that the study populations are at equilibrium. However, this assumption may not be met by our data set. Domination of the Floridian regional population by a single haplotype (FL1) is consis- tent with either a recent founder event from an unstudied source population, or else with a history of population size bottlenecking; a process that promotes stochastic elimina- tion of novel and rare haplotypes. In contrast, the starlike allelic topology of the Bermuda samples is characteristic of
The demographic distinctions implicit in the mitochon- drial genetic structure of Floridian and Bermudan Lasaea populations ultimately reflect environmental differences in the intertidal crevice habitat of this organism among the oceanic island and mainland locations. In Bermuda, Lasaea are easily sampled wherever such habitats exist (O´ Foighil, pers. observ.). In Florida, and in other Caribbean locations we investigated, Lasaea was respectively much more spo- radic or nonexistent. This small clam seems to be a rela- tively rare component of the Caribbean rocky shore fauna. Indeed, apart from two lots from the Bahamas and a single individual from Belize, all museum records of western North Atlantic Lasaea examined in a previous survey (O´ Foighil, 1989) were from Bermuda and southern Florida.
Because of their reduced biotic diversity, oceanic islands may allow some taxa to escape competitive interactions that severely reduce survivorship in other parts of their geo- graphic range (Paulay, 1994). Although we cannot rule out a contributing role for physical environmental factors, Ber- muda may well represent such an ecological release for western Atlantic Lasaea lineages. If this interpretation is correct, future characterization of the genetic structure of Caribbean Lasaea populations, though likely to uncover novel haplotypes, is predicted to yield depauperate genetic profiles similar to that exhibited by our Floridian samples.
Our thanks to P. Mikkelsen, A. Frias-Martins, R. Bullock, and C. Simmons for facilitating our sampling efforts. The