The influence of food groups on oesophageal cancer risk was analysed using data from a case-control study conducted between 1992 and 1999 in the Swiss Canton of Vaud on 101 incident, histologically confirmed cases (92 squamous cell, 9 adenocarcinomas) and 327 controls admitted to hospital for acute, non-neoplastic conditions. Multivariate odds ratios (OR) were computed after allowance for age, sex, tobacco, alcohol and non-alcohol energy. Significant increased risks emerged for red meat (OR = 1.7 for an increase of one serving per day), pork and processed meat (OR = 1.6), and eggs (OR = 1.5), whereas inverse associations were observed for milk (OR = 0.7), raw and cooked vegetables (OR = 0.5), citrus and other fruits (OR = 0.5), as well as for a more varied diet (OR = 0.5). Most associations were apparently stronger in heavy alcohol drinkers, suggesting an interaction between poor diet and alcohol drinking in oesophageal carcinogenesis.