A case-control interview study for assessing the role of dietary factors in selected cancers was undertaken in a hospital. Male patients from one community, with cancers of the oral cavity (n = 278), pharynx (n = 225), esophagus (n = 236), and larynx (n = 80) formed the case group. Patients diagnosed as not having cancer (n = 215) formed one control group, and a comparable sample of individuals from the general population (n = 177) formed another control group. All risks were adjusted for subjects' ages and habits of chewing and/or smoking tobacco, which are the two most important risk factors for cancers at these sites. A protective effect was observed with the intake of vegetables (twofold risk in nondaily vs. daily consumers) and fish (two- to threefold risk in those who did not eat at least once a week vs. those who did), and to a certain extent with pulses and buttermilk, in comparison with either one or both control groups. Intake of vegetables and fish were also observed to be risk modifiers for those who chewed and/or smoked tobacco. Lower levels of fat consumption was associated with elevated risk levels. The use of red chili powder emerged as a risk factor for all sites (two- to threefold risk with a dose-response relationship) compared with population controls. Tea drinking was also observed to be a risk factor for esophageal cancers, and to a lesser extent, for pharyngeal cancers.