The influence of selected variables on the prescribing volume of eighty family practitioners in a large HMO was quantified through multiple regression analysis. Over 70% of the variation was explained. The patient panel size and age composition were the most important determinants and accounted for 45.7% of the variance. The number of patients seen per clinic hour explained 14.3% and interclinic differences explained another 8.1% of the variance. Further, the data indicated two processes that might have affected results of earlier studies. First, female patients tend to select female physicians, as indicated by a high correlation between the sex of a physician and the proportion of females in the panels. Second, older physicians tend to have older patients. This implies that physician age does not explain higher prescribing rates, but is merely a proxy for older panels of patients who generally have more chronic illnesses and need more drugs.