The prevalence of human infection by Trypanosoma cruzi was assessed using an enzyme-linked immunosorbent assay (ELISA) in a serological survey in 1998 of 2 rural communities (SMH and PS) in Guatemala. In SMH (Department of Zacapa), where Rhodnius prolixus was the principal vector, the seroprevalence amongst 373 people tested was 38.8%. In PS (Department of Santa Rosa), where the main vector was Triatoma dimidiata, 8.9% of the 428 people tested were seropositive. The overall prevalence of seropositivity was higher in females than in males in both SMH (40% vs 36%) and PS (11.9% vs 4.9%), although this difference was significant only in PS. Historical seroconversion rates, estimated retrospectively by fitting a transmission model to the age-prevalence curves, were 3.8% per year in SMH and 0.5% per year in PS. There was some indication of a recent reduction in incidence in both villages. In PS, but not in SMH, both the observed prevalence and the estimated incidence rates were significantly higher in females than in males.