A presumptive diagnosis of heartwater in the living animal can be based on clinical and epidemiologic observations. In Guadeloupe, heartwater can be confused with haemonchosis in goats or cerebral babesiosis in cattle. Confirmation of the clinical diagnosis by brain biopsy is useful in experimental infections but is hardly applicable in the field. Positive results were obtained from 92% of animals 16 to 18 days after experimental infection. In febrile animals the best results were obtained between the 3rd and 6th days of the thermal response. Diagnosis can also be supported by serological tests. These are useful for monitoring experimental infections and for checking recovered animals in the field. Nineteen goats out of 27 were negative on Day 1 of the febrile reaction but positive a week later. The remaining 8 goats were positive on Day 1 and had greatly increased antibody titres a week later. Confirmation of a diagnosis can also be achieved by subinoculating into susceptible animals either blood or suspensions of ticks collected from suspect animals and then homogenized. Ticks that have engorged on a suspect animal can be allowed to moult and then fed on a susceptible animal to test their infectivity. These methods are time consuming but useful for heartwater surveys.