The abundance of snail hosts and the rates of infection with Schistosoma mansoni were monitored monthly for four years in two representative localities subjected to repeated chemotherapy of infected persons. Snail abundance varied from 1.0 to 4.4 collected per person/minute/station for Biomphalaria straminea and from 0.1 to 7.0 for B. glabrata. Infection rates of snails in nature varied from 0% to 15% for the former and from 0% to 70% for the latter species. Human infection increased from 35.5% to 61.9% in the locality occupied by B. straminea, and decreased from 40.3% to 20.8% in that occupied by B. glabrata. No relationship could be detected between human infection and the snail variables. Despite seasonal variations, natural infection persisted throughout the monitoring period in both snail species. It reached remarkably high levels in B. straminea when compared to those obtained by other authors probably because of differences in methodology. It is recommended that longitudinal studies should be carried out focally and periodically to avoid underestimating the prevalence of schistosome infection in snails.