The incidence, time course and nature of systemic reactions to injections of bee and wasp venom during immunotherapy have been estimated in an open, prospective, single centre study. One hundred and nine survivors of moderate to severe systemic reactions to stings from hymenoptera, received courses of bee or wasp venom by monthly subcutaneous injection for up to 3 years. Systemic reactions were recorded after 7.5% of 946 weekly venom injections during the initial phase of treatment, and after 2.1% of 1789 monthly maintenance injections. In both phases of treatment, reactions were more frequent after bee (17% of initial phase, 7.8% of maintenance treatment) than after wasp (3% of initial phase, 0.3% of maintenance treatment) venom injections. The percentage of patients experiencing at least one reaction was also higher for bee (46%) than for wasp (14%) sensitive patients. Over 80% of reactions began within 30 min of injection, over 90% within 1 h and only two (2%), between 1 and 2 h, the remaining six (5.5%) starting more than 2 h after injection. Only 0.47% of venom injections produced a systemic reaction which was severe enough to require adrenaline treatment. The female patients experienced more reactions (21% of the wasp, 60% of the bee, sensitive) than the males (5.5% wasp, 20% bee). Age and atopy did not appear to be significant risk factors for systemic reactions. We conclude that wasp and bee venom immunotherapy in a conventional dosage regimen was generally well tolerated. No safety advantage was apparent in keeping the patients under observation for more than 1 h after injections. Reactions are more likely to occur during the initial phase of treatment, in women and after bee, rather than wasp, venom injections.