A nested case-control study was conducted to investigate the effect of alcohol consumption and tobacco smoking on cancers of various sites. The study population was based on 887 cases and 1774 controls, selected from a cohort of 17,200 male participants of a gastric mass survey in 1984, who were followed up for 9 years by the Chiba Cancer Registry, Japan. The odds ratio (OR) of colon cancer was significantly elevated in alcohol drinkers of one cup of sake-equivalent (27 ml ethanol) per day (OR = 3.5), and three cups of sake-equivalents per day (OR = 3.2) compared with nondrinkers, but its dose-response was not clear since two cups of sake-equivalents per day had an OR of 1.9, which was nonsignificant. Cancer risk elevation was especially predominant in the proximal colon, again showing no dose-response: OR = 30.7 for one cup of sake-equivalent per day, OR = 12.4 for two or more cups per day. Lung cancer showed a dose-response relationship with alcohol consumption, independent of tobacco smoking. A synergistic effect of alcohol intake and tobacco smoking was observed for upper aerodigestive tract and bladder cancer. Both alcohol drinking and tobacco smoking were found not to be associated with stomach cancer.