The present paper suggests that the difficulty in inducing animals to orally self-administer quantities of ethanol sufficient to produce dependence is related to the schedule of ethanol access. If the variables that control the various patterns of ethanol consumption can be identified, then the patterns of ethanol consumption leading to excessive intake and the development of dependence will be able to be identified and generated experimentally. Evidence is presented which shows that the amount of ethanol consumed per bout, the daily frequency of ethanol bouts, as well as the total amount of ethanol consumed, is a function of the frequency of ethanol access periods, feeding condition, ethanol concentration, and the temporal proximity of ethanol access periods.In the attempt to discover the variables responsible for the development of human alcoholism, many investigators have attempted to induce excessive ethanol consumption in nonhuman organisms. Modeling pathology in animals has proved enormously useful in other areas of science and medicine, and there is little reason to believe, a priori, that an animal model of human alcoholism would not contribute to our understanding of human alcoholism. Not only could the many presumed psychological causes of alcoholism, such as stress, for example, be examined under controlled conditions, but also treatment techniques having the potential to decrease consumption could be designed andtested. An adequate animal model would also help determine the physiological effects of prolonged ethanol consumption uncontaminated by the possible nutritional and other health-eare deficiencies usually observed in human alcoholics.Unfortunately, nonhuman organisms do not spontaneously consume ethanol in amounts that readily lead to intoxication or the development of physical dependence. The failure of procedures to induce oral consumption of ethanol for its pharmacological effect (