For a given body mass, hominoids have longer clavicles than typical monkeys, reflecting the lateral reorientation of the hominoid glenoid. Relative length of the clavicle varies among hominoids, with orangutans having longer clavicles than expected for body mass and gorillas and chimpanzees having shorter clavicles than expected. Modern humans conform to the general hominoid distribution, but Neandertals and Upper Paleolithic Homo sapiens have longer clavicles than expected for their size and exhibit marked positive allometry in clavicle length. Relative to clavicle length, adult and newborn humans have broader shoulders (biacromial breadths) than comparable apes, because the reduced elevation of the human shoulder swings the acromion laterally downward away from the head. Since broadened shoulders yield an increased risk of maternal and neonatal injury and/or death from shoulder dystocia during birth, we might expect hominins to manifest trends toward reduction in shoulder breadth and clavicle length. They do not, presumably because of countering selection pressures favoring a long clavicle in the adults. The marked sexual dimorphism seen in patterns of clavicular growth and static adult allometry in humans suggests that those selection pressures have disproportionately affected the males.