Brachial plexus injury is assumed to be associated with the traumatic delivery of a macrosomic fetus in the vast majority of cases. This study was undertaken to examine the relationship of brachial plexus injury to birth weight and trauma at delivery, to compare our incidence to the incidence in other populations, and to examine how the incidence of brachial plexus injury has changed in our institution over the last 30 years. A retrospective analysis of 14,358 births from January 1, 1987, to June 30, 1991, identified 15 cases of brachial plexus injury (all Erb's palsy, incidence 0.10%). Maternal and neonatal charts were reviewed. There were 14 cases of Erb's palsy out of 11,484 vaginal deliveries (0.12%) and one case of Erb's palsy out of 2,874 cesarean deliveries. There was birth trauma (i.e., shoulder dystocia) noted at the time of delivery in eight cases (53.3%). However, a surprising finding was that in the other seven cases (46.7%) there was no evidence of shoulder dystocia at delivery. In the group in which Erb's palsy occurred and trauma was noted at the time of delivery, the average birth weight was 4,265 ± 480 g (range 3,550–5,110 g), with seven out of eight (88%) being large for gestational age (LGA). In the group in which Erb's palsy occurred but no trauma was noted at the time of delivery, the average birth weight was 2,906 ± 745 g (range 1,590–3,950 g), with one out of seven (14%) being LGA. The infants without recognizable trauma weighed significantly less (P = 0.0009). In the group with trauma noted at delivery one out of eight (13%) received pitocin, and in the group without trauma noted at delivery one out of seven (14%) received pitocin. There was no significant difference in 5 min Apgar scores <7 (3/8 vs. 0/7), umbilical cord pH (7.27 ± 0.07 vs. 7.24 ± 0.10), or base excess (−3.1 ±1.6 vs. −5.3 ±3.3) between those with recognizable trauma and those without recognizable trauma. The incidence of brachial plexus injury in this institution from 1987 to 1991 was 0.10%, which was unchanged from the incidence of 0.12% from 1954 to 1959, even though the cesarean rate rose from 5% to 20% during this period. The appearance of Erb's palsy in the newborn may not be as closely linked to birth weight and recognizable birth trauma as has previously been thought. In this study half the cases of Erb's palsy occurred in normal‐sized infants without trauma noted at delivery. The incidence of Erb's palsy in our population is similar to that of other reported studies and has remained unchanged over the past 30 years, even as our cesarean rate has risen from 5% to 20%. J. Matern.‐Fetal Med. 6:1–5, 1997.