In the mouse, injection (subcutaneously) of the putative 5-HT1 agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT), produced a dose-related hypothermia (ED50:0.36 mg/kg). A maximum response was elicited by intracerebroventricular (i.c.v.) injection of 8-OH-DPAT (3 micrograms) and almost abolished by lesion of 5-HT-containing terminals in the brain with 5,7-dihydroxytryptamine (5,7-DHT; i.c.v.) or long-term treatment with p-chlorophenylalanine. The response was unaltered by a range of neurotransmitter antagonists: prazosin (alpha1-adrenoceptor), idazoxan (alpha2-adrenoceptor), metoprolol (beta1-adrenoceptor), erythro-DL-1-(7-methylindan-4-yloxy)-3-isopropylamino-but an-2-ol (beta2-adrenoceptor), (-)propranolol or (+/-)pindolol (beta-adrenoceptor), flupenthixol (dopamine) or Ro 15-1788 (benzodiazepine binding site). Classical 5-HT antagonists (methysergide, metergoline, cinanserin and methiothepin) were either without effect or facilitated the response and the 5-HT2 antagonist, ritanserin was also without effect. In contrast, quipazine and haloperidol produced a dose-related antagonism of the response. Since the response was almost abolished by a lesion induced by 5,7-DHT and was antagonised by quipazine, which is known to antagonise presynaptic 5-HT function in vitro, it is suggested that the hypothermic response is due to 8-OH-DPAT acting as an agonist at a presynaptic 5-HT receptor, which also appears to be sensitive to butyrophenones (the antagonism elicited by haloperidol but not by flupenthixol). The hypothermic response of mice to 8-OH-DPAT, therefore, may provide a simple and convenient in vivo model in which to measure the function of the presynaptic 5-HT receptor.