Population genetics, larval dispersal, and connectivity in marine systems

Population connectivity plays significant roles on both evolutionary and ecological time-scales; however, quantifying the magnitude and pattern of exchange between populations of marine organisms is hindered by the difficulty of tracking the trajectory and fate of propagules. We explored biophysical correlates of population substructure to determine how well pelagic larval duration (PLD) correlates with population genetic estimates of connectivity in a sample of 300 published studies drawn pseudo-randomly from about 1600 hits on electronic searches. In direct contrast to the general expectation of a strong correlation, we find that average PLD is poorly correlated (r(2) < 0.1) with genetic structure (F-ST). Furthermore, even this weak correlation is anchored by non-pelagic dispersal, because removal of the zero PLD class (direct developers) from the analysis resulted in a non-significant relationship between FST and PLD. For species in which minimum, maximum, and mean PLDs were available, it is noteworthy that both minimum and maximum PLDs are better correlated with FST than the mean larval duration, which has been used in all such previous studies. A 3-way ANCOVA reveals that genetic marker class (allozymes, microsatellites, and mitochondrial DNA sequences), as opposed to habitat or swimming ability, explain most of the variation in F-ST (F = 7.113, df = 2, p = 0.001), with higher values Of FST obtained from mtDNA than with either microsatellites or allozymes (which were not significantly different). Our meta-analysis refutes recent reviews and conventional wisdom that PLD is a good predictor of the magnitude of gene flow and geographic scale of population structure in marine systems.