In this study 18 rural hospitals in the Cape Province were prospectively surveyed in respect of the number of deliveries, perinatal deaths, low-birth-weight babies, and the hospital facilities available for routine antenatal, intrapartum and neonatal care. During the study period (1 January - 31 December 1985) there were a total of 19,274 deliveries (birth weight greater than or equal to 1,000 g). The mean perinatal mortality rate was 27,3/1,000 deliveries (range 15.3-45.3). The prevalence of low-birth-weight babies ranged from 13.3% to 34.4% with a mean of 17.7%. By plotting the perinatal mortality rate against the prevalence of low-birth-weight babies a perinatal care index was established. A low perinatal mortality coupled with a high prevalence of low-birth-weight infants, for example, signifies proper care. The perinatal index was then used to compare the neonatal outcome of different hospitals. Outcome varied considerably, as did facilities and care. Serological tests for syphilis and blood groups were not known in the majority of patients at 4 hospitals involving 2,387 deliveries. No cervical cytology was done at 6 hospitals involving 5,462 deliveries. One hospital had no neonatal resuscitation equipment. There were no facilities for serum bilirubin measurements at 10 hospitals involving 8,867 deliveries. However, most problems identified are easily and cheaply rectifiable.