Techniques are reviewed for the experimental feeding of alcohol, including a liquid diet procedure invented 25 years ago. This technique results in much higher ethanol intake than with other approaches. As a consequence, various complications observed in alcoholics can be reproduced in animal models. These include fatty liver, hyperlipemia, various metabolic and endocrine disorders, tolerance to ethanol and other drugs, physical dependence and withdrawal and, in the baboon, liver fibrosis and cirrhosis. Variations of the liquid diet formulation are compared, and adequacy of nutrition in terms of minerals, vitamins, lipotropes, carbohydrates and proteins is discussed. The importance of selecting proper controls is emphasized. The respective advantages of three standardized basic rat formulas are reviewed: (i) an all-purpose (35% fat) diet, comparable to the diet previously referred to as the "Lieber-DeCarli formula" and suitable for most experimental applications, particularly those intended to mimic the clinical situation in which the various effects of alcohol occur in the setting of hepatic changes characterized by a fatty liver; (ii) a low-fat diet comparable in all respects to the preceding diet but with a lower fat content, intended to minimize the hepatic changes, and (iii) a high-protein formula particularly useful in those circumstances in which an oversupply of dietary protein might be recommended (i.e. pregnancy). Variations of this technique, including continuous intragastric infusion, are also discussed. It is concluded that, for most experimental studies of chronic alcohol consumption, the liquid diet technique provides one of the most efficient tools to study the effects of ethanol under controlled nutritional conditions because it allows for alcohol consumption of clinical relevance and offers flexibility to adjust to special experimental or physiologic needs by allowing for various substitutions required for a particular experimental design, including changes in lipids, proteins or other dietary constituents. The technique also facilitates the comparison with controls by simplifying the pair feeding and is the best procedure available for the study of the toxic effects of alcohol and their interactions with deficiency or excess of various nutrients.