The genetic stock structure of larval and juvenile winter flounder larvae in Connecticut waters of eastern Long Island Sound and estimations of larval entrainment

Variability at six microsatellite loci was examined among 536 winter flounder Pseudopleuronectes americanus larvae collected from three locations in eastern Long Island Sound shown to be nursery areas for the species. Substantial genetic differences were seen among the putative source populations and thus appeared to be geographically based. These differences were used to characterize the most likely sources of winter flounder larvae entrained at the Millstone Power Station as well as recruitment to juvenile winter flounder collected in the Niantic River. Samples were classified to the most likely geographical source population both by a conditional maximum likelihood method and by a multi‐layer feed‐forward neural net trained on the differences in microsatellite allele frequencies. The classification of samples by both methods is compared and discussed in the context of winter flounder management.