Relapse into problematic alcohol drinking is a serious problem in the treatment of alcoholism. Free-choice drinking rhesus monkeys show relapse-like behaviour after imposed abstinence of alcohol, by immediately reinitiating ethanol intake at an increased level. The relapse-like behaviour of the monkeys seems not induced by physical withdrawal, but rather argues for a resistance to extinction of ethanol-reinforced behaviour. It has been suggested that endogenous opioids play a role in the positive reinforcing effect of ethanol. In this study, the effect of the opiate antagonist naltrexone was investigated in eight adult male rhesus monkeys (Macaca mulatta) who had about 1 year experience with alcohol drinking, under two conditions: 1) (expt 1) during continuous and concurrent supply of drinking water and two ethanol/water solutions (16% and 32% (v/v], and 2) (expt 2) after 2 days of alcohol abstinence. In both experiments, each monkey received six doses of naltrexone (0.02, 0.06, 0.17, 0.5, 1.0, 1.5 mg.kg-1); each dose was paired with a placebo injection (im) in a cross-over design. Consumption was measured from 16.00 hours in the afternoon (30 min after injection) to 9.00 hours the next morning. In experiment 1 naltrexone reduced total net ethanol intake in a graded dose-dependent manner. The effect of naltrexone was apparent shortly after injection, and lasted until the following day. Consumption of drinking water was reduced only shortly after injection. In expt 2, reduction of net ethanol intake was largely restricted to the first few hours of reinitiation of alcohol drinking, i.e. the period in which the abstinence-induced increase was manifest. Consumption of drinking water was not affected by naltrexone. Naltrexone hardly influenced consumption of the non-preferred ethanol solution of 32%. It is postulated that the opioid modulation specifically interacted with positively reinforced behaviour. In expt 2 naltrexone reduced ethanol intake at a lower dose (0.17 mg.kg-1) compared to expt 1 (0.50 mg.kg-1), but net ethanol intakes however remained higher. It might be that alcohol abstinence resulted in altered opioid activity, leading to increased ethanol-seeking behaviour. The renewed presentation of ethanol solutions (also) might have stimulated reinitiation of alcohol drinking, representing conditioned incentive stimuli. The reported monkey model of relapse in alcohol drinking could be a useful tool to evaluate new hypotheses and experimental treatments with respect to human alcoholism.