Changes in population allele frequencies may be driven by several forces, including selection and drift, and are revealed only by sampling over many generations. Such studies, however, are rare for protist parasites. Microsatellite allele frequencies for 4 loci were followed in a population of Plasmodium mexicanum, a malaria parasite of lizards in California USA at 1 site from 1978 to 2010. Rapid turnover of the lizards indicates the parasite was studied for a minimum of 33 transmission cycles and possibly twice that number. Sample sizes ranged from 841 to 956 scored parasite clones per locus. DNA was extracted from frozen dried blood and blood removed from stained blood smears from the earliest years, and a verification study demonstrated DNA from the blood smears provided valid genetic data. Parasite prevalence and effective population size (Ne) dropped after 2000, remaining lower for the next decade. For 2 loci, allele frequencies appeared stable for the first 2 decades of the study, but changed more rapidly after the decline in prevalence. Allele frequencies changed more gradually for the other 2 loci. Genetic drift could account for changes in allele frequencies, especially after the drop in prevalence and Ne, but the force of selection could also have driven the observed patterns.