An experiment was carried out in which patients who were seeking appointments for a consultation in a general practice in south London attended consulting sessions booked at 5, 7.5, or 10 minute intervals. The particular session that the patient attended was determined non-systematically. The clinical content of the consultation was recorded on an encounter sheet and on audio-tape. At the end of each consultation patients were invited to complete a questionnaire designed to measure satisfaction with the consultation. The stress engendered in doctors carrying out surgery sessions booked at different intervals of time was also measured. At surgery sessions booked at 5 minute intervals, compared with 7.5 and 10 minute intervals, the doctors spent less time with the patients and identified fewer problems, and the patients were less satisfied with the consultation. Blood pressure was recorded twice as often in surgery sessions that were booked at 10 minute intervals compared with those booked at 5 minute intervals. There was no evidence that patients who attended sessions booked at shorter intervals received more prescriptions, were investigated or referred more often to hospital specialists, or returned more often for further consultations within four weeks. There was no evidence that the doctors experienced more stress in dealing with consultations that were booked at 5 minute intervals than at consultations booked at 7.5 and 10 minute intervals, though they complained of shortage of time more often in surgery sessions that were booked at shorter intervals.