The relationship between sugar added to coffee and other hot beverages--as an indicator of taste for sugar and sugar intake outside main meals--and the risk of colorectal cancer was investigated using data from a case-control study conducted in Northern Italy on 953 cases of histologically confirmed colon cancer, 633 of rectal cancer and 2845 controls admitted to hospital for acute, non-neoplastic, non-digestive tract disorders. Compared with subjects who reported adding no sugar to their beverages, the multivariate relative risks (RR) of colon cancer were 1.4 for those adding one spoonful of sugar, 1.6 for those adding 2 spoonful, and 2.0 for those adding 3 or more. The corresponding RRs for rectal cancer were 1.3, 1.5 and 1.4. For combination of colorectal cancer the RRs were 1.4, 1.5 and 1.8. All the trends in risk were significant, and the results were consistent across strata of study centre, sex and age, and were not appreciably modified by allowance for a number of major identified potential distorting factors, including an estimate of total calorie intake. These findings, if confirmed, would suggest that taste for sugar is a relevant indicator of colorectal cancer risk, and could be interpreted either in terms of a role of sugar in colorectal carcinogens, or of a specific influence of even limited amounts of sugar taken outside meals, which may stimulate the proliferation of the bowel epithelium, and hence enhance colorectal carcinogenesis.