Births occurring at home in England and Wales declined from 85 per cent of all births in 1927 to 1.2 per cent in 1980. During this period successive committees recommended further reductions and eventually the phasing out of home births on the grounds that hospital deliveries were safer. Perinatal and infant mortality rates for place of delivery have only been available for England and Wales since 1975. These data showed that the perinatal mortality rate for home births was rising and from 1977 onwards had risen to a level above that for hospital births. It was suggested that this apparent increase in the risk of perinatal death for babies born at home was artifactual; the result of an increase in the number of unplanned births at home (carrying a high risk of perinatal death) relative to the number of planned home deliveries. Having described the historical background this thesis reports on a survey which was carried out to establish the intended place of delivery of the 8856 births which occurred at home in 1979 as these data were not collected routinely. The results of the survey showed that 67 per cent of these births had been booked for delivery at home, 21 per cent had been booked for delivery in hospital, 3 per cent were unbooked and for 9 per cent the intended place of delivery was unknown. Perinatal mortality varied almost 50 fold according to the intended place of delivery. The rate for those planning to give birth at home was 4.1 per 1000 while for those who were unbooked the rate was 196.6 per 1000. For mothers booked for delivery in consultant units the rate was 67.5 per 1000. Births booked for delivery at home included the smallest proportion of babies weighing 2500 grams or less at birth: 2.5 per cent compared with 18 per cent for those booked for delivery in consultant units. Within the low birthweight band perinatal mortality was lowest for planned home births. Among babies who weighed more than 2500 grams at birth perinatal mortality was uniformly low irrespective of the intended place of delivery; the only exception was unbooked deliveries. In all groups perinatal mortality was significantly higher for nulliparous than for parous women. When data derived from the survey were used to standardise the overall perinatal mortality rate for births occurring at home, the results suggested that perinatal mortality among home births declined at about the same rate as that for all births. The results of the survey showed clearly that women planning to give birth at home were a select group, and indeed some may have transferred to hospital during labour and thus were not included in the survey. Nevertheless, these data suggest that planned delivery at home is compatible with a low risk of perinatal death.