Outbreaks of foot-and-mouth disease (FMD) from January 1982 through December 2003 were used to examine variations in serotype- and species-specific risk for three control programmes in Colombia: (1982-1983) vaccination, using an aluminium hydroxide, saponin adjuvant, required but not enforced; (1984-1996) vaccination, using an oil double-emulsion adjuvant, required but not enforced; and (1997-2003) enforced vaccination, using an oil double-emulsion adjuvant, restricted animal movement enforced, and slaughter of infected animals. Hypotheses were tested for trend, cyclicity and seasonality in FMD occurrence, and for species- and serotype-specific differences in morbidity and case-fatality. The spatial density of outbreaks was estimated by kernel smoothing. The frequency of outbreaks decreased most between 1984 and 1996 (p < 0.01) for serotype A and between 1997 and 2003 (p < 0.01) for serotype O. Outbreaks occurred in cycles of 3-4 years for both serotypes (p < 0.05). Morbidity was not significantly different in pigs from that in cattle for serotype A-associated outbreaks (p = 0.314), but was higher in pigs than in cattle (p = 0.019) for serotype O-associated outbreaks. For both serotypes, case-fatality was higher for pigs than for cattle (p < 0.009). Temporal variation in FMD incidence provided insight into the expected evolution of FMD control for countries with similar conditions and where FMD is endemic.