Population structure in the freshwater sunfish Lepomis macrochirus was analyzed by examining genotypes at three polymorphic loci in 2,560 individuals representing 64 localities distributed evenly among eight reservoirs and two drainages in the southeastern United States. There is no evidence of inbreeding within localities, but allele frequencies among localites within a reservoir are often heterogeneous; mean standardized variance (FST) for localities of a reservoir was 0.029, consistent across loci. This magnitude of differentiation is slight, far less than for snail populations within cities or house mice on adjacent farms, and occurs despite the immense sizes of the reservoirs sampled (up to 100,000 acres and 1,200 shoreline miles). It seems realistic to regard a reservoir population as a quasi-panmictic assemblage of local populations between which differentiation is minimal and only weakly oriented to distance of geographic separation. Variances in allele frequencies among reservoirs within a drainage are much greater (F̄ST = 0.392), accounting for nearly 90 percent of the total variance observed in the study. Adjacent reservoirs are more similar in allele frequency than are those farther apart. A roughly parallel allelic cline among reservoirs of the two drainages may reflect preexisting heterogeneity among the founding populations in the linear river habitat. Allele frequencies in these reservoirs have not changed significantly during five years of observation. The adaptive significance, if any, of the alternative allelic states is unknown. Furthermore, contemporaneous processes affecting population structure, whether deterministic or stochastic, must act upon a preexisting structure derived from a unique set of historical developments which are seldom known for any species. Thus by hard criteria, definitive statements about the causal processes responsible for population structure are only rarely warranted, even in case studies such as the bluegill where the empirical results appear straightforward.