GULF STREAM SYSTEM LASAEA
Figure 1. Location of the Lasaea sampling sites in the western North Atlantic, plotted onto a satellite image of sea-surface temperatures (mod- ified from Thurman, 1997). The body of warm water flowing through the Florida Straits represents the Florida Current portion of the Gulf Stream System. Past Cape Hatteras, the Gulf Stream proper leaves the continental slope and develops a series of meanders that pinch off persistent mesoscale eddies. Cold eddies migrate southward and approach Bermuda.
nary (Nunn, 1994) and, although late Pleistocene (last 130,000 years) sea-level changes ranged up to 25 m, the island group was not submerged during this period (Harmon et al., 1978, 1981).
The Bermudan shallow-water marine fauna encompasses a low level (2.4%) of endemic species and represents a moderately impoverished oceanic extension of the Carib- bean fauna (Sterrer, 1986, 1998). Twenty-eight of a total of 49 common nearshore molluscs found throughout the Ca- ribbean also occur on this oceanic island platform (Warmke and Abbott, 1961), and there is some evidence for a partial turnover of the Bermudan marine malacofauna on ecologi- cal timescales (Abbott and Jensen, 1967; Sterrer, 1986). The nearest source populations for Bermudan marine taxa of Caribbean origin are about 1500 km to the southwest, in southern Florida and the Bahamas. Based on net transport of drift bottles, this distance represents a minimum passage of 21–30 days for passive pelagic transport of larvae or rafted propagules (Jackson, 1986). Such a time frame does not pose a serious obstacle for species with long-lived pelagic larvae, but may approach the upper limit for many species with faster developing pelagic larval stages. The few data available on genetic distances of Bermudan marine inver- tebrates from Caribbean conspecifics are consistent with this view (Mitton et al., 1989; Hateley and Sleeter, 1993). A
number of Bermudan marine faunal constituents are known to have been introduced by human agency (Sterrer, 1986, 1998), a process that is independent of current polarity or geographic distance.
The genus Lasaea is composed of minute, crevice-dwell- ing, intertidal clams that have a near-cosmopolitan distribu- tion on rocky shores; its North Atlantic lineages are exclu-
´ sively composed of direct-developers (O Foighil, 1989).
Molluscan systematists generally recognize one North At- lantic species, Lasaea adansoni (Gmelin) (5L. rubra Mon- tagu; L. bermudensis Bush) (Sterrer, 1986). This systematic conclusion is based solely on shell characters and has not been corroborated by population genetic (Tyler-Walters and
´ Crisp, 1989) and phylogenetic (O Foighil and Smith, 1995;
Foighil and Jozefowicz, 1999) studies. O Foighil and
Jozefowicz (1999) constructed molecular phylogenetic trees for North Atlantic Lasaea lineages based on mitochondrial (mt) 16S gene variation in two continental putative source populations (Florida, the northern limit of distribution in North America; Iberia) and two oceanic island populations (Bermuda, Azores). No amphi-Atlantic genotypes were de- tected; Bermudan lineages co-clustered exclusively with Floridian congeners to form a western Atlantic clade and Azorean samples formed an exclusive clade with Iberian haplotypes. The western Atlantic lineages showed low lev- els of genetic diversity (a single Florida haplotype differed by one diagnostic nucleotide substitution from three Ber- mudan haplotypes), formed a shallow polytomy, and (re- markably) were sister to polyploid Australian clonal conge- ners in global phylogenetic trees (O´ Foighil and Jozefowicz, 1999).
Jackson (1986) attributed the presence of Caribbean spe- cies lacking an extended pelagic larval stage on Bermuda to colonization via rafting events from continental source pop- ulations. Based on present-day circulation patterns in the western North Atlantic, the hypothesis that Bermudan La- saea populations were established by rafting events assumes that intermittent gene flow has occurred, over evolutionarily significant timeframes, from western continental margin source populations—for example, from south Florida. Is- land lineages are predicted to represent a subsample of mainland genetic diversity, and to contain endemic geno- types that cluster in the terminal tips of a western Atlantic clade. O´ Foighil and Jozefowicz’s (1999) preliminary 16S data set had some features that were consistent with the rafting hypothesis: the three Bermudan haplotypes had close phylogenetic links to, but were genetically distinct from, the Floridian haplotype. However, these workers were unable to distinguish among ancestral and derived relationships among the four polytomous western Atlantic mt 16S geno- types, and it was unclear how representative the sole Flo- ridian haplotype, sampled from a single population, was of mainland regional genetic diversity.
The aim of the present study was to perform a compre-