1. We studied patterns of abundance of the malarial parasite Plasmodium mexicanum in its vertebrate host, the western fence lizard Sceloporus occidentalis, over a 13-year period at 51 study sites in northern California, USA. Abundance of the vectors, sandflies in the genus Lutzomyia, was also studied among sites during a single warm season, and among nights at one site during two seasons. 2. The parasite differed in prevalence (per cent of lizards infected) in males and females; males were more often infected. Prevalence increased with body size (= older lizards were more often infected). 3. Malaria prevalence varied among sites: 0-50% of lizards were infected. Topography explains part of this variation because malaria was rare or absent at sites. > 500 m elevation. However, sites at lower elevations, even when within short distances of one another, varied in malaria prevalence. 4. Abundance of vectors was not related to prevalence of malaria at a site; sandflies were found at some sites where malaria was rare or absent, including those at high elevations. Wind speed, relative humidity and sky brightness did not affect numbers of sandflies active, but vectors were active only when air temperature was > 16°C. 5. Parasite prevalence varied among years. Environmental conditions (temperature, rainfall and plant biomass produced at the study region) were not correlated with parasite prevalence. The changes in prevalence over time resembled a cycle of long duration (10 years). 6. Classical models developed for study of malaria in humans suggest explanations for variation in prevalence of P. mexicanum among sites, and a more recent theory suggests an explanation for the possible long-duration cycle observed in this study