A cluster of four deaths in late December 1993, marked the onset of an outbreak of disease of African elephants (Loxodonta africana) in the Kruger National Park (KNP) in South Africa, which has an estimated population of 7,500 elephants. Mortalities peaked in January 1994, with 32 deaths, and then declined steadily to reach pre-outbreak levels by September, but sporadic losses continued until November. During the outbreak altogether 64 elephants died, of which 53 (83%) were adult bulls. Archival records revealed that, in addition to the usual losses from known causes such as poaching and intraspecific fighting, sporadic deaths from unexplained causes had, in fact, occurred in widely scattered locations from at least 1987 onwards, and from that time until the perceived outbreak of disease there had been 48 such deaths involving 33 (69%) adult bulls. Carcases had frequently become decomposed or had been scavenged by the time they were found, but seven of eight elephants examined early in 1994 had lesions of cardiac failure suggestive of encephalomyocarditis (EMC)-virus infection, and the virus was isolated from the heart muscles of three fresh carcases. The results of tests for neutralizing antibody on 362 elephant sera collected for unrelated purposes from 1984 onwards and kept frozen, indicated that the virus had been present in the KNP since at least 1987. Antibody prevalences of 62 of 116 (53%) 18 of 139 (13%) and seven of 33 (21%) were found in elephants in three different regions of the KNP in 1993 and 1994. Studies had been conducted on myomorph rodents in the KNP for unrelated purposes since 1984, and trapping attempts were increased during the perceived outbreak of disease in elephants. There was a striking temporal correlation between the occurrence of a population explosion (as evidenced by markedly increased catch rates per trap-night) and a surge in prevalence of antibody to EM virus in rodents, and the occurrence of the outbreak of disease in elephants.