Ethanol, like other drugs of abuse, has motivating properties that can be developed as animal models of self-administration. A major strength of the operant approach where an animal must work to obtain ethanol is that it reduces confounds due to palatability and controls for nonspecific malaise-inducing effects. In the domain of opioid peptide systems, limited access paradigms have good predictive validity. In addition, animal models of excessive drinking-either environmentally or genetically induced-also appear sensitive to blockade or inactivation of opioid peptide receptors. Ethanol availability can be predicted by cues associated with positive reinforcement, and these models are sensitive to the administration of opioid antagonists. Perhaps most exciting are the recent results suggesting that the key element in opioid peptide systems that is important for the positive reinforcing effects of ethanol is the mu-opioid receptor. How exactly ethanol modulates mu-receptor function will be a major challenge of future research. Nevertheless, the apparently critical role of the mu receptor in ethanol reinforcement refocuses the neuropharmacology of ethanol reinforcement in the opioid peptide domain and opens a novel avenue for exploring medications for treating alcoholism.