The share of general-practitioner units in the provision of maternity services, though minor, has increased since 1956, but their facilities have consistently been used less intensively than those of consultant obstetric hospitals. Their fetal and neonatal mortality rates, however, have consistently been much lower than in obstetric hospitals, a disparity which the higher proportion of births in hospital, recorded as being at above average risk, is not nearly enough to explain. These facts should be important considerations in any review of maternity services occasioned by changes in the birthrate.