Microsatellites, which have rapidly become the preferred marker in population genetics, reliably assign individual fish to the Sacramento River winter vs. fall, late-fall, or spring chinook runs of California's Central Valley (Banks et al. 2000 CHAS 57:915). A substantial proportion of this discriminatory power comes from Ots-2, a simple CA repeat, which is expected to evolve rapidly under the stepwise mutation model (SMM). We have sequenced a 300 by region around this locus and typed 668 microsatellite-flanking sequence haplotypes, in order to explore further the basis of this microsatellite divergence. Three sites of nucleotide polymorphism in the Ots-2 flanking sequence define five haplotypes that are shared by California and Canadian populations. The Ots-2 alleles are non-randomly distributed among these five haplotypes, in a pattern of gametic disequilibrium that is also shared among populations. Retention of phylogenetic information at the microsatellite locus causes differences, among haplotypes, in the variance of Ots-2 allele-size that are inconsistent with simulations based on SMM. Divergence between winter run and other Central Valley stocks is caused by a combination of surprisingly static evolution at Ots-2 within a context of more rapidly evolving haplotype frequencies. |