Two systems of general-practitioner obstetric care currently in use in England are compared. In one (Oxford) a large number of doctors, working with community midwives, have autonomous responsibility for the management of patients whose condition conforms with defined booking criteria. In the other (Dulwich) two practitioners, working with hospital midwives, operate a minimal selection policy, but all their patients are reviewed at 36 weeks' gestation by a specialist obstetrician. Despite the differing selection policies which created incongruities in the groups in respect of age, social class, and birthweight distribution, the outcomes in terms of mode of delivery showed no significant differences. The perinatal mortality rates were also similar and appeared to be better than those achieved nationally.