Low doses of the selective 5-HT1A agonist 8-OH-DPAT increase ethanol intake in a limited access paradigm following peripheral injection. This may be due to a reduction in 5-HT neurotransmission following activation of raphe somatodendritic autoreceptors. In order to test this hypothesis, and to determine the effects of selective reductions in raphe 5-HT activity, experiments examined the effects of injecting 8-OH-DPAT into the dorsal raphe (0, 0.02, 0.1, 1 and 2.5 micrograms) or the median raphe (0, 0.1, 1 and 5 micrograms) in rats trained to drink 12% ethanol for 40 min each day. The effects of the GABAA agonist, muscimol, on ethanol intake were also examined. Ethanol intake was increased at the highest dose of 8-OH-DPAT following injection into either site, with no change in water intake. Thus, the effects of 8-OH-DPAT are selective for ethanol. The selective 5-HT1A antagonist, (+)-WAY100135 (0.3, 1 and 3 mg/kg), blocked the effect of 8-OH-DPAT, showing that activation of 5-HT1A receptors underlies the ethanol drinking induced by 8-OH-DPAT. These results are consistent with the idea that reduced 5-HT function increases ethanol intake. Several behavioral mechanisms for this effect are discussed. Muscimol (50-100 ng) also increased ethanol drinking. Following injection into the median raphe, muscimol also stimulated water intake. These effects are probably due to non-specific behavioural activation induced by this treatment. However, the effect of muscimol in the dorsal raphe was specific for ethanol since water intake was not altered.(ABSTRACT TRUNCATED AT 250 WORDS)