A detailed study on a modified technique of epidural analgesia (EDA) for pain relief in obstetrics has been performed. The aim of the modifications was to reduce the number of instrumental deliveries and at the same time to make the delivery as smooth as possible for the baby. This was achieved by the use of an anaesthetic with a favourable ratio between neonatal and maternal plasma levels (Bupivacaine) in low concentration (0.25%). A special technique of injection enabled us to limit the extent of the blockade. An epidural catheter was inserted between L2 and L3 and moved upwards 20 cm into the epidural space. 8-10 ml of the solution was then injected after a test dose. The blockade was continued by the repeated injection of smaller doses. As judged by the skin anaesthetic zones and by obstetric examinations, the patient was gradually positioned, during labour, from supine to half-sitting. The catheter was withdrawn at the end of the first stage of labour so that the lower sacral segments could be blocked. A group of 100 patients treated with the technique described was followed. The number of instrumental deliveries in the present series (15%) was significantly lower in similar series reported in the literature. The group of 100 patients treated with modified EDA was also compared with 100 control patients who received only conventional treatment without EDA. There was no differences in the number of abnormal presentations, while the number of caesarean sections and cases of atonic post-partum bleeding was insignificantly lower in the EDA group. The difference in the number of instrumental deliveries-although somewhat higher in the EDA group-was not statistically significant. The total length of labour was prolonged in the EDA group but EDA was not necessarily responsible for this undesired effect. The clinical status of the babies was found to be better after EDA than in control group, as judged from Apgar score. Furthermore, fetal bradycardia was significantly reduced with EDA. It is concluded that EDA, as used in this study, not only is a preferable way of achieving pain relif in the mother but also offers a means of facilitating the birth process for the infant, reducing the incidence of pre- and post-natal asphyxia.