A population-based case-control study was designed to examine if the risk of developing intra-oral squamous-cell carcinoma in Denmark was associated with occupation, marital status, residence, dental status, and exposure to coffee, tea, tobacco, and alcohol. Cases consisted of 161 consecutively-admitted incident patients with histologically verified, primary, intra-oral squamous-cell carcinoma treated at the Aarhus University Hospital from January 1986 to November 1990. For each case, three controls of the same gender and age were selected randomly from among nonhospitalized residents in the hospital's catchment area (some 1.4 m inhabitants). Four hundred of the selected 483 controls participated in the study. Risk was associated significantly with marital status, residence, dental status, alcohol consumption, and exposure to tobacco. When correcting for tobacco and alcohol consumption, only marital status and dental status remained significant. The association between risk and marital status was particularly prominent among divorced compared with married persons (odds ratio [OR] = 2.3, 95 percent confidence interval [CI] = 1.1-4.6). Persons with less than five teeth had an OR of 2.4 (CI 1.3-4.1) compared with persons with 15 or more teeth. Tobacco and alcohol exposure were the strongest individual risk-indicators in both lifetime and current consumption estimates, and their composite effect was particularly strong. Compared with nonusers, OR for tobacco (> 20 g/d) adjusted for alcohol = 5.8 (CI = 3.1-10.9); OR for alcohol (> 5 drinks/d) adjusted for tobacco = 8.4 (CI = 4.0-17.6). The OR for heavy users of tobacco and alcohol (> 20 g tobacco/d and > 5 drinks/d) was 80.7 (CI = 21.8-298.8). These results confirm that tobacco and alcohol contribute significantly to the risk of developing oral cancer. There were no significant differences between the risk estimates for the two genders or young and old persons. Two simulation studies indicate that the observed risk associated with tobacco and alcohol consumption cannot be explained reasonably by a high consumption among the 83 nonrespondents.