A new classification for caesarean section was developed in a two-part study conducted at six hospitals. Initially, 90 anaesthetists and obstetricians graded ten clinical scenarios according to five different classification methods--visual analogue scale; suitable anaesthetic technique; maximum time to delivery; clinical definitions; and a 1-5 rating scale. Clinical definitions was the most consistent and useful, and this method was then applied prospectively to 407 caesarean sections at the same six hospitals. There was close agreement (86%) between anaesthetists and obstetricians for the five-point scale (weighted kappa 0.89), increasing to 90% if two categories were combined (weighted kappa 0.91). We suggest that the resultant four-grade classification system--(i) immediate threat to life of woman or fetus; (ii) maternal or fetal compromise which is not immediately life-threatening; (iii) needing early delivery but no maternal or fetal compromise; (iv) at a time to suit the patient and maternity team--should be adopted by multidisciplinary groups with an interest in maternity data collection.