ABSTRACT
Hypotheses to explain chaotic genetic structure (i.e., a surprising degree of nongeographic temporal or spatial population differentiation) include 1) variation in source of larval recruits, 2) selfrecruitment and local subdivision, 3) variance in reproductive success (sweepstakes reproduction), and 4) pre-or postsettlement natural selection. We evaluated the relative contribution of each of these four processes to the observed patterns of genetic differentiation among geographic populations of the porcelain crab, Petrolisthes cinctipes. We genotyped about 50 individuals of each size class (new recruits, subadults, adults) at each of nine sites across northern California in each of three consecutive years (N = 3602). We found significant and consistent population structure (θ ≈ 0.08) among sites from each replicate year of sampling (1997-1999). Significant population structure also occurred both among years and among sites, but the pattern of population structuring differed by size class. Among years, the differentiation of new recruits was highest (θ = 0.12) year-to-year, followed by the subadults (θ = 0.09) and finally adults (θ = 0.08). In contrast, the among-sites component was greatest among the adults (θ = 0.05), followed by the subadults (θ = 0.03), and least among new recruits (θ < 0.01, not significant). An overall hierarchical analysis of molecular variance (AMOVA) revealed significant genetic divergence among years (θ = 0.05), and size classes (θ = 0.08), but not among sites (θ < 0.01). Recruits at each site were genetically most similar to a different population of adults each year, and although some recruits were apparently self-recruits, no consistent patterns of assignment or genetic similarity to natal populations emerged across years, indicating that the source of recruitment is unpredictable. Temporal differentiation was stronger than spatial differentiation, and in the full hierarchical analysis, the among-years and among-size class components of variance explain the majority of observed population structure. Most studies of population genetic structuring are snapshots based on single collections that lack explicit temporal components, and may provide an incomplete picture of population structure as a result. Overall, these data suggest that all four hypothesized mechanisms collectively create fine-scale population genetic structure in P. cinctipes, and perhaps many other marine species with similar life histories and distributions.
