Abstract
Direct development in benthic marine invertebrates is usually associated with narrow geographical range, low rates of colonization and low levels of gene flow. Paradoxically, the small brittle star Amphipholis squamata broods its larvae to a crawl-away juvenile stage, yet has a cosmopolitan distribution.
Using scanning electron microscopy and sequence and RFLP analysis of nuclear and mitochondrial DNA (COI, 16S, 18S and ITS) from 52 coastal populations, I tested whether A squamata populations world-wide are indeed morphologically and genetically homogeneous to warrant single species status. If A. squamata is a poor disperser, as may be expected from its brooding habit, I predicted that local and regional populations would be genetically highly structured and possibly composed of cryptic sibling species, as has recently been shown to be the case in a number of widespread marine invertebrates.
Genetic analyses revealed seven extremely divergent and reproductively isolated lineages, which can occur in syntopy, however, vary in abundance, haplotype diversity and overall geographic distribution. Two lineages are widespread and abundant and occur in allopatry: lineage A has an anti-tropical distribution throughout temperate areas of the northern and southern hemispheres, while lineage E occurs throughout tropical and subtropical areas, world-wide. All other lineages are considerably less abundant, although some are apparently widespread. The geographic distribution of lineages is correlated with ocean current flows and temperature.
At a regional scale, the partitioning of genetic variation is consistent with known marine biogeographic zones and appears to be linked to the regime of oceanic circulation. These results indirectly support the contention that dispersal in this brooding species is achieved by passive transport by drifting or rafting and emphasize that direct development does not necessarily lead to poor dispersal ability.
Paradoxically, extreme genetic divergence among lineages does not appear to be linked to morphological differentiation. Although lineages may have been separated since the Cretaceous, up to 100 million years ago, morphometric analysis found no significant differences between lineages or geographic groupings. I also show that disk colour, which is linked to bioluminescence behaviour, is not a phylogenetically useful character.