In a prospective randomised study, the effects of an intracervical gel containing 0.5 mg PgE2 and an intravaginal tablet containing 3 mg PgE2 were compared in their efficiency to induce delivery at term. When only gel was used, induction failed in 45% of the cases, especially with an initially immature cervix score. However, these patients did show significant improvement of the Bishop score. The vaginal tablet was more effective with regard to the induction of delivery. When using the vaginal tablet alone, induction failed in only 5% of cases (p < 0.001). Best results, however, were obtained with a combination of both application forms. In this group there were no failures. Tablet application induced delivery within 24 hours in 56% of the cases, gel application in only 27% (p < 0.001). There were no significant differences observed in these groups during the course of labour. The length of the latency period was identical, whether a tablet or a gel was used in initiating labour. In 29% of the cases, beta-sympathicomimetics had to be used because of tetanic contractions or polysystolia. No significant differences in the occurrence of these contractions were seen in either the gel or the tablet group. Even in a high-risk collective with 25% pathological Doppler-flow findings, no higher rate of operative deliveries or fetal acidoses was observed after prostaglandin induction, if careful surveillance of the patients and early use of tocolytics in experienced hands were administered. Our results further confirm the concept of a differential induction of labour with prostaglandins.(ABSTRACT TRUNCATED AT 250 WORDS)