EDITORIAL COMMENT: We accepted this paper for publication since everyone who practises obstetrics can profit from a revision of this information. Although shoulder dystocia is often not predictable, the editorial subcommittee, after a discussion of this article, agreed to advise readers that in their opinion there are 3 clinical warnings to be heeded. Firstly, there is the mother who has a past history of shoulder dystocia on 1 or more occasions and who appears to have an unusually large fetus; if Caesarean section is decided against, the delivery should be conducted in a suitable position (lithotomy) by suitable personnel (experienced medical practitioner plus assistant), with suitable conditions (episiotomy and analgesia/anaesthesia). Secondly, consider the possibility of shoulder dystocia when mid‐forceps delivery is performed when the fetal head arrests in the occipitoposterior or occipitolateral position, especially in an obese multipara where fetal size cannot be assessed. Thirdly, shoulder dystocia should be anticipated when a fetal head which is on view in the occipitoanterior position at the height of a contraction, recedes to the midpelvis and rotates to the occipitolateral position after the contraction and the mother's voluntary expulsive efforts cease‐ the ‘suitable conditions’ enumerated above should now be implemented. Many of these cases are associated with failure of anterior rotation of the shoulder when labour commenced with an occipitoposterior position. There are also those where the anterior shoulder has arrested above the pubic symphysis and these are the most difficult ones to deal with. Occipitoposterior positions continue to provide most of the problems encountered during labour (see previous comments on Shoulder Dystocia: Risk factors, prevention, association with macrosomia etc. Aust NZ J Obstet Gynaecol 1988; 28:107, 1989; 29:129 and 1990; 30:319). Summary: Shoulder dystocia, or impacted shoulders is an infrequently encountered obstetric emergency. Despite various risk factors identified by investigators, the occurrence of shoulder dystocia is difficult to predict. Two cases of severe shoulder dystocia managed personally by the author are presented. These cases illustrate some of the important issues regarding prediction of shoulder dystocia and the need to begin a series of well‐tested manoeuvres immediately, to successfully deliver the baby.