The obstetric records of patients from Khayelitsha were examined to assess the efficiency of a system for the antenatal prevention of congenital syphilis, and to identify points of breakdown in the process. Seventy-seven (12.7%) of 607 mothers had serological evidence of syphilis, including 10 (32.3%) of 31 mothers who had received no antenatal care. Of 70 patients who required routine management, only 36 (51.4%) received 3 or more of the recommended 4 penicillin injections. Two main weaknesses in the system were identified. One was the centralisation of serological testing. This delayed results reaching the relevant unit, and was responsible for a high cumulative attrition of patients during the many stages necessitated by the centralised testing. The other was a 24.5% attrition of patients referred from the antenatal clinic to a separate sexually transmitted diseases clinic.