Search

The additive genetic variance-covariance matrix (G) is a concept central to discussions about evolutionary change over time in a suite of traits. However, at the moment we do not know how fast G itself changes as a consequence of selection or how sensitive it is to environmental influences. We investigated possible evolutionary divergence and environmental influences on G using data from a factorial common-garden experiment where common frog (Rana temporaria) tadpoles from two divergent populations were exposed to three different environmental treatments. G-matrices were estimated using an animal model approach applied to data from a NCII breeding design. Matrix comparisons using both Flury and multivariate analysis of variance methods revealed significant differences in G matrices both between populations and between treatments within populations, the former being generally larger than the latter. Comparison of levels of population differentiation in trait means using Q(ST) indices with that observed in microsatellite markers (F(ST)) revealed that the former values generally exceeded the neutral expectation set by F(ST). Hence, the results suggest that intraspecific divergence in G matrix structure has occurred mainly due to natural selection.

Background Small, genetically uniform populations may face an elevated risk of extinction due to reduced environmental adaptability and individual fitness. Fragmentation can intensify these genetic adversities and, therefore, dispersal and gene flow among subpopulations within an isolated population is often essential for maintaining its viability. Using microsatellite and mtDNA data, we examined genetic diversity, spatial differentiation, interregional gene flow, and effective population sizes in the critically endangered Saimaa ringed seal (Phoca hispida saimensis), which is endemic to the large but highly fragmented Lake Saimaa in southeastern Finland. Results Microsatellite diversity within the subspecies (HE = 0.36) ranks among the lowest thus far recorded within the order Pinnipedia, with signs of ongoing loss of individual heterozygosity, reflecting very low effective subpopulation sizes. Bayesian assignment analyses of the microsatellite data revealed clear genetic differentiation among the main breeding areas, but interregional structuring was substantially weaker in biparentally inherited microsatellites (FST = 0.107) than in maternally inherited mtDNA (FST = 0.444), indicating a sevenfold difference in the gene flow mediated by males versus females. Conclusions Genetic structuring in the population appears to arise from the joint effects of multiple factors, including small effective subpopulation sizes, a fragmented lacustrine habitat, and behavioural dispersal limitation. The fine-scale differentiation found in the landlocked Saimaa ringed seal is especially surprising when contrasted with marine ringed seals, which often exhibit near-panmixia among subpopulations separated by hundreds or even thousands of kilometres. Our results demonstrate that population structures of endangered animals cannot be predicted based on data on even closely related species or subspecies.