There is a consensus that male physicians work more hours than female physicians, but no consensus on how or why female physicians limit hours. Drawing on a new national survey of physicians in large, multispecialty groups, we conclude that much of the hours differential is due to differences in nonpractice incomes and is likely to be permanent. However, due to reductions in family responsibilities for women and increases in nonpractice incomes for men, the differential should narrow somewhat as the current cohort of physicians ages. We also infer that the link between hours and practice choices may well account for much of the diversity of the literature.