Ondansetron (1-3 mg/kg), granisetron (0.3-1 mg/kg) and dexamethasone (0.3-1 mg/kg), administered at 12-h intervals, were investigated for their potential to prevent cisplatin (30 mg/kg, i.p.)-induced emesis during a 72-h observation period. Ondansetron appeared more active than granisetron to antagonise the emetic response occurring in the first 4-h (P<0.05) period, but none of the regimens significantly antagonised emesis during the 0-24- and 24-72-h periods (P>0.05). However, ondansetron was more active to antagonise emesis on day 1 using a more frequent drug administration, whereas bilateral vagotomy only reduced emesis for 2 h, and 5-HT, 2-methyl-5-HT and 1-m-chloro-phenylbiguanide (up to 20-30 mg/kg, i.p.) were not emetic. The combination of ondansetron 1 mg/kg and dexamethasone 1 mg/kg, both administered every 12 h, significantly delayed the onset of emesis (P<0.05) but failed to reduce the total numbers of retches+vomits over the 3-day period (P>0.05). Results are discussed in relation to the clinical situation.