The mechanism for long-term maintenance of St. Louis encephalitis (SLE) virus in California is unknown. Two possibilities are 1) that the virus is maintained locally in discrete enzootic foci by one or more reservoir mechanisms, and/or 2) that the foci are ephemeral in nature and virus is reintroduced periodically from other enzootic areas by migratory birds or movement of vectors. We have investigated these epidemiologic alternatives by studies of genetic variation within a 277 nucleotide portion of the envelope-encoding region among 17 strains of SLE virus isolated since 1952 from different geographic locations in California. Three lineages of virus were detected. One lineage, Group A, consisted of four SLE virus strains isolated in California since 1972 from the Coachella, Sacramento, and San Joaquin Valleys. The group A strains were closely related to strain MSI-7 of SLE virus isolated in Mississippi in 1975. The 13 other strains formed the second and third lineages (Groups B1 and B2) that had geographically overlapping distributions. Group A (BFN 4585) and Group B2 (BFN 4820) appeared to be sympatric in the Sacramento Valley in 1972. Strains from the San Joaquin Valley isolated prior to 1989 (Groups B1 and B2) differed markedly from a 1989 isolate from the same location, Kern 373 (Group A). These results suggest that virus introduction(s) led to changes in genotype, or alternatively that the enzootic virus was subjected to selective pressure leading to rapid emergence of a new genotype. Nucleotide sequences of the envelope and 5' untranslated region of the viral genome of these virus strains did not correlate with virulence as measured by mortality in weanling mice, nor viremia levels and duration in chickens.