Five hundred and twenty faecal samples or rectal swabs, collected over one year from children mainly under 24 months, were tested for the presence of human rotavirus (HRV) in enzyme linked immunosorbent assays (ELISA). Fifty-three (23.6 pc) of 225 diarrhoea samples and 25 (8.5 pc) of 295 control samples were HRV positive. The association between diarrhoea and detection of HRV in stools was statistically significant with an overall odds ratio of diarrhoea patients of 3.32 (0.001 greater than p). No significant difference in the incidence of HRV infection between populations at a communal farming location and a high density suburb community was proved. The highest HRV incidence was found during the dry cool season in diarrhoea patients between four and 24 months of age. Overall proportions of 16.7 pc and 28.6 pc HRV positive in groups of 0-3 and 4-6 months old diarrhoea patients respectively was remarkably high, suggesting influence of close human contact, crowding in residences and insufficient sanitary facilities on the transmission of the HRV. Approximately 10 pc of 4-18 months old control patients were HRV positive.