Epidemiologic evidence on the relation between nutrition and pancreatic cancer is reviewed. A number of epidemiologic studies of diet and cancer of the pancreas have been reported including descriptive, case-control, and cohort studies. Overall, fairly consistent patterns of positive associations with the intake of meat, carbohydrates, and dietary cholesterol have been observed. Consistent inverse relationships with fruit and vegetable intakes and, in particular, with two markers of such foods, namely fiber and vitamin C, also have been noted. However, the methodologic limitations of these studies, particularly the descriptive and case-control studies, are such that causal inferences regarding these empirical associations currently are not warranted. Future follow-up of existing dietary cohorts should enable more precise assessment of the possible role of diet in the etiology of cancer of the pancreas.