Mid-late Pleistocene fossil hominins such as Homo neanderthalensis and H. heidelbergensis are often described as having extensively pneumatized crania compared with modern humans. However, the significance of pneumatization in recognizing patterns of phyletic diversification and/or functional specialization has remained controversial. Here, we test the null hypothesis that the paranasal sinuses of fossil and extant humans and great apes can be understood as biological spandrels, i.e., their morphology reflects evolutionary, developmental, and functional constraints imposed onto the surrounding bones. Morphological description of well-preserved mid-late Pleistocene hominin specimens are contrasted with our comparative sample of modern humans and great apes. Results from a geometric morphometric analysis of the correlation between paranasal sinus and cranial dimensions show that the spandrel hypothesis cannot be refuted. However, visualizing specific features of the paranasal sinus system with methods of biomedical imaging and computer graphics reveals new aspects of patterns of growth and development of fossil hominins. Anat Rec, 291:1506–1516, 2008. © 2008 Wiley-Liss, Inc.