OBJECTIVE To compare use of working hours, patient turnover, and patient characteristics in two general practitioner groups which differed in contractual arrangements and income levels (partly fee-for-service vs. salary). METHODS Cross-sectional study of all patient contacts during one week. METHODS All general practitioners (GPs) in Oslo (N = 263) with contracts with the municipality. 83% participated. METHODS Time spent in various types of work and patient turnover; the number of patients seen per hour of consultation time. RESULTS Fee-for-service GPs worked almost exclusively with patients in their own practice, while salaried GPs spent more time out of office (consultation-based patient work constituting 55% of the working hours). The fee-for-service group registered on average 2.68 consultations per hour of patient contact (95% confidence interval: 2.51-2.84). The corresponding numbers for the salaried group were 2.37 (2.24-2.50). Salaried GPs had somewhat more telephone consultations per hour (2.01 vs 1.66 for fee-for-service GPs), leaving a picture of two GP groups with small differences in patient turnover. The two patient populations were remarkably similar and so was the general practitioners' own evaluation of the consultations. CONCLUSIONS Over a number of years the GPs in Oslo have practised under different contractual arrangements. As a consequence of different expectations and incentives it was expected that consultation activity and patient populations had evolved differently. This proved not to be the case.