The relationship between selected aspects of medical history and the risk of colorectal cancer was analysed using data from a case-control study of 673 cases of colon cancer, 405 of rectal cancer and 1501 controls in hospital for acute, non-neoplastic, non-digestive tract conditions, unrelated to known or suspected risk factor for large bowel cancer. Significantly elevated risks (RR) were observed for history of cholelithiasis (RR = 1.5 [95% confidence interval (CI) 1.1-2.1] for colon; 1.6 [1.2-6.4] for rectum) and diabetes (1.6 [1.1-2.3] for colon; 1.3 [0.8-2.0] for rectum), and a significant protection emerged for history of drug allergy (0.6 [0.4-0.9] for colon; 0.6 [0.5-1.0] for rectum). No significant association was found with thyroid disease, gastroduodenal ulcer, liver cirrhosis, hepatitis, pancreatitis, gastrectomy, appendicectomy, treatment with cimetidine/ranitidine, treatment with chenodesoxycholic acid or with blood transfusions. The associations with cholelithiasis, diabetes and drug allergy were not materially modified by allowance for major identified potential confounding factors, and were not restricted to the diseases diagnosed within 5 or 10 years before large bowel cancer diagnosis. Thus, the analysis of this large dataset offered further quantitative evidence suggesting a possible, however moderate, association between gallbladder disease and colorectal cancer risk, which may be related to enhanced or continuous secretion of secondary bile acids. The associations with diabetes and drug allergy were unexpected, and probably indirect, lacking previous epidemiological support or any obvious biological interpretation. Thus, they should be simply regarded as working hypotheses worthy of further consideration.