In a two-stage screening procedure using the Edinburgh Postnatal Depression Scale and Goldberg's Standardised Psychiatric Interview, 232 women six months after delivery were compared with control women individually matched for age, marital status and number of children, obtained from general practitioner lists, who were not pregnant nor had had a baby in the previous 12 months. No significant difference in the point prevalence of depression at six months was found between the postnatal (9.1%) and control women (8.2%) nor in the six-month period prevalence (13.8% postnatal, 13.4% controls), but a threefold higher rate of onset of depression was found within five weeks of childbirth. The possible explanations relate to the long duration of depression in women with young children, and the stressful effect of childbirth and its psychosocial sequelae.