This study assessed whether there is any variation in the incidence of haematological malignancies between geographical areas of differing water supplies in the South West peninsula of the United Kingdom (1984 to 1988 inclusive). The possibility of correlations existing between variation in water quality and variation in the incidence of haematological malignancies was examined. Haematological incidence data, taken from the Leukaemia Research Fund's Data Collection Study, were mapped into 46 geographical areas of differing water supply. The distribution of the mapped cases was then tested for homogeneity using the Potthoff and Whittinghill (1966) test score. The age-adjusted incidence ratios calculated during the heterogeneity testing were examined for correlations with water quality indicators using correlation and stepwise regression. Significant heterogeneity in the incidence rates among water supply areas was observed for two groups of disease - acute leukaemias and myeloproliferative disorders. Three water quality indicators - pH, nitrate concentration and aluminium concentration - varied considerably over the study period. Significant correlations were observed between the standardized incidence ratios of five disease categories and some water quality indicators, especially aluminium and trihalomethane concentrations. The standardized incidence ratios of some haematological malignancies differed between geographical areas of water supply in South West England, and the evidence suggests that this variation may be associated with variation in water quality indicators. Although this lends support to similar findings in the United States of America, the pattern of correlations are affected by disease latency and statistical methodology.