The role of alcohol consumption in oesophageal cancer in Normandy has been studied by a retrospective study of 312 male cases and 869 controls. The linear relationship between the logarithm of risk and overall daily alcohol consumption was confirmed after adjustment for tobacco. The role of each specific alcoholic beverage was further investigated by computing relative risks for individuals consuming a given beverage and for those drinking other beverages only, within each overall alcohol consumption category. It is concluded: (1) that there is a linear relationship between the logarithm of risk of oesophageal cancer and overall daily ethanol consumption, whatever the beverage; (2) that the effect is more marked for strong beverages (digestives) than for lighter beverages; (3) that there is an additional risk related to apple brandy and cider.