The relation of episiotomy to third-degree perineal tears was investigated in 21 278 singleton deliveries. The incidence of episiotomy was 28.4% (n = 6041). Third-degree tears occurred in 1.4% (85) of the deliveries with episiotomy and in 0.9% (132) of the deliveries without episiotomy (P less than 0.01). To avoid the effect of confounding factors, we analysed a sub-sample that included only vertex presentations with spontaneous occipitoanterior vaginal deliveries. After stratification for birthweight and parity, no relation between episiotomy and third-degree tear was found.