SPATIAL SCALE OF GENETIC STRUCTURE AND AN INDIRECT ESTIMATE OF GENE FLOW IN EELGRASS, ZOSTERA MARINA.

In this study, the first investigation of population structure in an aquatic angiosperm, I show that populations of a marine angiosperm (eelgrass, Zostera marina) are genetically differentiated at a number of spatial scales. I find also that there is no correspondence between geographic and genetic distances separating subpopulations, an increasingly common result in spatially stratified studies of genetic structure in marine invertebrates. F-statistics, calculated for two years from electrophoretic variation at five polymorphic allozyme loci, indicate significant genetic differentiation among sampling quadrats within each of two bays (θ = 0.064-0.208), between tide zones within a bay (θ = 0.025-0.157) and between bays (θ = 0.079). Spatial autocorrelation analysis was used to explore genetic differentiation at smaller spatial scales; estimated patch sizes (within which genetic individuals are randomly associated) indicated no appeciable genetic structure at scales less than 20 m × 20 m. Calculated values of F-statistics were a function of the spatial scale from which samples were drawn: increasing the size of the "subpopulation" included in calculation of fixation indices for the same "total" sample resulted in an increase in the magnitude of f (e.g., from 0.092 to 0.181) and a decrease in θ (e.g., from 0.186 to 0.025). On the basis of the best estimate of the spatial scale of subpopulations, the effective number of migrants per generation (N m) ranges from 1.1 to 2.8. Genetic consequences of the disturbance regime in the eelgrass habitat sampled were extreme variation between years in the allele richness and proportion of heterozygotes in a sample and a positive relationship between the extinction probability of patches and the genetic variance among them. The changes in F-statistics as a function of sampling scale and the observation that θ among sampled quadrats was positively associated with the probability of extinction among quadrats indicated that indirect estimates of gene flow (N m) calculated from θ should be cautiously interpreted in populations that may not yet be in drift-migration equilibrium.