The relation between ovarian cancer and coffee drinking habits was evaluated in a case-control study of 247 histologically confirmed epithelial ovarian cancers and 494 agematched controls, admitted to hospital for acute conditions apparently unrelated to coffee consumption. Compared to rates for women who had never drunk coffee, the crude relative risk estimates for those who drank less than two, two or three, and four or more cups per day were 1.3, 1.5 and 1.4 respectively; however, when allowance was made for smoking habits, these risk estimates became 1.3, 1.7 and 1.8 respectively, and a significant linear trend of increasing risk with more elevated coffee consumption was evident. These results were not explained by various other potential confounding factors, including the major risk factors for ovarian cancer, but we had no information on dietary variables. The relative risk, however, did not increase with increasing duration of use. The findings of this study give apparent support in favour of the hypothesis that coffee consumption, or related dietary variables, may be associated with the risk of epithelial ovarian cancer. Further studies in different settings, however, are required in order to establish whether this association is real, and if so, whether it is causal.