Temporal genetic changes between cohorts in a natural population of a marine fish, Diplodus sargus

Temporal changes at 17 allozyme loci in the Diplodus sargus population of Banyuls sur Mer (Mediterranean Sea, France) were monitored within a single population among ten year‐classes (cohorts) sampled over a 6‐month period. The genetic survey was combined with evaluation of the demographic structure of the population by determining variation of abundance between cohorts. The population showed variation in abundance among cohorts ranging from 16 to 214 individuals. Significant divergences in genetic structure were observed between cohorts (P < 0.0001) despite very low values of F_ST (multilocus F_ST over all cohorts = 0.0018). The heterozygosity of each cohort, as well as the F_IS values, was significantly correlated with the abundance of each cohort, with abundant cohorts showing lower heterozygosity and a significant deficit of heterozygotes (positive F_IS values). Finally, multilocus temporal genetic variance (F_k) computed between successive cohorts was higher in low abundance cohorts. Change of heterozygosity between cohorts, distribution of year‐class genetic structure, and change in the genetic structure within a cohort appear to be affected mostly by the abundance of the cohort and are therefore driven by genetic drift. We propose that the Diplodus sargus cohorts are built up from the mixing of families during the pelagic stage or later during recruitment, and that the decrease in heterozygosity leading to a deficit of heterozygotes is characteristic of a Wahlund effect. Such a Wahlund effect would derive from the mixing of the progeny of families made up of few individuals, but exhibiting high fecundity and high variability of reproductive success. Therefore, although cohorts derived from poor recruitment would only group a few families and would exhibit limited deficit of heterozygotes (higher heterozygosity values), they would lead to high genetic drift and appear more divergent (higher mean temporal genetic variance) than cohorts with high abundance. While not demonstrating directly the family structure of marine populations, our survey provides evidence of highly structured populations. © 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 76 , 9–20.