The paper describes a longitudinal field study of canine leishmaniasis in sympatric domestic dog and crab-eating fox (Cerdocyon thous) populations. Dogs were studied in house-to-house surveys, and foxes were studied by live-trapping and radio-telemetry. Because serological tests (IFAT in this case) for leishmaniasis are often of uncertain sensitivity and specificity, we draw conclusions comparatively. Both cross-sectional (age-prevalence) and longitudinal analyses indicate that incidence in dogs was highest in the dry season. Seasonal changes in the age-prevalence relationship for dogs suggest that serological conversion and recovery rates decline with prior exposure to infection, where 'recovery' may be due to loss of a positive antibody titre or death from leishmaniasis. The mean incidence in dogs was higher in the rural than in the urban population and higher in hunting dogs than pet dogs. Cross-sectional and longitudinal analyses of fox serological data suggest that foxes remain positive for longer than dogs on average, either because detectable antibody is more persistent or because they experience a lower mortality rate due to leishmaniasis.