The authors have prospectively examined the occurrence of postoperative wound infection following clean neurosurgery in 936 patients. Fewer than 1% received perioperative antibiotic prophylaxis. The overall rate of deep wound infection was 2.6%; no deaths were directly attributable to these infections. Deep wound infections occurred significantly more frequently following craniotomy (4.3%) than following spinal (0.9%) or other clean neurosurgery. Among craniotomies, the deep wound infection rate varied significantly from 11% following repeat operations for recurrent gliomas to 2.5% following non-tumor surgery. Risk of deep wound infection varied more than 11-fold depending on the type of clean neurosurgical operation. It is most feasible to demonstrate the potential efficacy of perioperative antibiotics in clean neurosurgical procedures with the greatest risk of postoperative wound infection. The potential benefit from such prophylaxis would be greatest for patients undergoing these high-risk operations.