We regard the increasing rate of caesarean section as a iatrogenic phenomenon of the late twentieth century, but in a thoughtful commentary (pages 1043–1045) Phil Steer persuades us that there are more fundamental reasons. The evolution of the genus Homo in the past three million years has resulted in an increase in the size of the brain and, concurrently, the adoption of the erect posture with a narrow pelvis, a competition between the need to think and the need to run. Thus we are mistaken if we think that labour and delivery evolved to the mutual advantage of mother and fetus; rather if the fetus is large there are harmful consequences to the mother, and if it is small, from prematurity or restricted growth, there are harmful consequences to the fetus. Homo supiens solves problems by technology, an example of which is caesarean section, which has developed in the last century into one of the safest operations in the world, and which resolves this conflict between mother and fetus. These fundamental considerations suggest that, at the end of the twentieth century, rates of caesarean section must always be high, for the safety of caesarean section brought about by technology far out-weighs the risks of even a slightly complicated vaginal delivery.

Evolution may even have determined the optimum age for a woman to have her baby. Adam Rosenthal and his colleagues (pages 1064–1069) noted that older women have a higher rate of obstetric intervention, and tested the hypothesis that this is a continuous phenomenon throughout the reproductive years, rather than confined to a specific age group. The authors performed a retrospective observational study of over six thousand nulliparae at term, the information for the study being derived from data collected routinely at the time of delivery. There was a linear association of age with induction of labour, epidural analgesia, instrumental vaginal delivery, emergency caesarean section and elective caesarean section; such that, for example, the annual percentage increase in the rate of emergency caesarean section in spontaneous labour is 8%. One interpretation of these findings is that obstetricians' perceptions of risk also increase in a linear fashion with maternal age; but it is equally possible that senescence adversely affects the physiology of the uterus and genital tract. This paper should stimulate basic research into the effects of age on the functions of the myometrium and cervix.

The technology of caesarean section has evolved in the past century with advances in anaesthesia, blood transfusion and surgical technique. One traditional surgical manoeuvre is eventration or exteriorisation of the uterus which in the opinion of some obstetricians assists closure of the uterus by allowing easier access to the incision in the lower segment. There have been theo- retical objections to exteriorisation of the uterus, based on reflex haemodynamic changes and an increased risk of infection introduced by exteriorisation. Edmond Edi-Osagie and his colleagues (pages 1070–1078) lay these fears to rest. The authors conducted a randomised trial of closure of the uterus when exteriorised and with the uterus in the abdomen, and found only marginal differences in haemodynamic measurements, vomiting, pain during and after the operation and puerperal morbidity. Obstetricians who find exteriorisation of the uterus helpful may be reassured that this procedure causes no harm.

A major reason for the high rate of caesarean section is repeat caesarean section, and it is often thought that if the indication for the first caesarean section was cephalo-pelvic disproportion, elective caesarean section should be performed in the future. This notion is challenged by V.H.W.M. Jongen and colleagues (pages 1079–1081), who carried out a survey of 132 consecutive women who had had a caesarean section in the second stage of labour, often following an attempt at instrumental vaginal delivery which failed. One-quarter of the women were delivered by elective caesarean section, and of those who underwent a trial of labour four-fifths delivered vaginally. Where the previous caesarean section had followed a failed attempt at vacuum extraction or forceps, three-quarters of the women delivered vaginally. There were no significant complications to the mother or her infant. The authors question the concept of cephalo-pelvic disproportion where there has been a caesarean section for delay in the second stage of labour, and their study suggests that most women in these circumstances can be safely allowed to go into labour.

But what if caesarean section and other obstetric interventions are caused by routine medical practices? This hypothesis is suggested by Ann Langer and her colleagues (pages 1056–1063), although to test this this was not the primary purpose of their study. The authors conducted a randomised trial of psychosocial support during labour and after delivery in which the primary outcome was the frequency of breast feeding at one month. The trial was rigorous: there was a clear statement of a primary hypothesis and a power calculation based on this hypothesis; a description of the process of randomisation and concealment of the allocation until the point of treatment; and the evaluation of the outcomes was made by investigators who were unaware of the treatment received by the women. The results were disappointing: although exclusive breast feeding occurred more frequently in the women who received support, the overall rate of breast feeding was low; the women felt they had more control over labour, but there were no differences in their degree of anxiety, self-esteem or feeling of satisfaction. Unlike the results of other randomised trials of social support in labour there were no differences in the rates of operative vaginal delivery or caesarean section. The trial took place in an urban maternity hospital where there was a culture of routine obstetric intervention, such that four-fifths of the women received epidural analgesia and one quarter were delivered by caesarean section. The authors suggest that obstetric interventions carried out routinely may nullify any beneficial effect of social support in labour.

The title of Philip Steer's commentary is deliberately ambiguous, for he refers not only to evolution in an anthropological sense, but also to the evolution of caesarean section, both its techniques and the indications for its performance. These fundamental concepts will affect our response to the current high rate of caesarean section, and encourage research into the pathophysiology of myometrial and cervical function, the place of routine obstetric interventions, and women's perceptions of their labour and delivery.