We examined the effect of systemic administration of the endogenously occurring progesterone metabolite, allopregnanolone, on oral self-administration of ethanol by male rats. Rats were trained to perform an operant response for presentation of 0.1 ml of a solution of 10% ethanol in water using the sucrose fading technique. After acquisition of stable lever-press responding on a fixed-ratio 4 schedule, subjects received subcutaneous injections of 1, 3, or 10 mg/kg of allopregnanolone, or vehicle, 20 min prior to the self-administration session. Pretreatment with 3 mg/kg, but not 1 or 10 mg/kg, increased the mean total number of lever press responses made to obtain ethanol, and therefore increased the mean total number of ethanol presentations. The number of responses and response rate were examined as a function of the number of "runs" within the 30-min session; a "run" was defined as a series of consecutive responses with an interresponse interval of <1 min. The increase in total responses after 3 mg/kg was due in part to an increased number of responses for the first run of the session, with no effect on response rates. However, the higher dose of 10 mg/kg decreased response rates within the first run. Thus, allopregnanolone alters ethanol-reinforced responding at concentrations lower than those that depress rates of responding. The effects of administration of the benzodiazepene, diazepam, were determined for comparison with those of the neurosteroid. The subcutaneous injection of 0.3, 1.0, or 3.0 mg/kg of diazepam did not produce any clear dose-dependent changes in measures of ethanol-reinforced operant responding, supporting the suggestion of differences in the contribution of the benzodiazepene and neurosteroid binding sites to GABA(A) receptor function. The results indicate that exogenous administration of allopregnanolone dose-dependently alters ethanol-reinforced operant responding, and suggest that this endogenously occurring neurosteroid could mediate some of the reinforcing effects of ethanol.