This is a report of 33 consecutive cases of petroclival meningioma treated surgically at our institution over the last 10 years; there were 21 women and 12 men between the ages of 27 and 68 (mean age, 52). All patients were assessed by computed tomographic scans including coronal sections and bone algorithm studies; in most cases, digital subtraction angiography and magnetic resonance imaging were also done. The largest tumor diameter was between 2 and 3.5 cm in 14 cases, 3.5 to 6 cm in 15 cases, and over 6 cm in 4 cases. Dural attachment predominantly involved the clivus and apical petrous bone on one side only; in 14 cases, however, the tumor grew over the clivus midline or crossed the tentorial notch. Cranial nerve deficit was extant in all cases and was commensurate with tumor size. Cerebellar signs and somatic motor deficits were present in 60 and 30% of cases, respectively. The surgical approaches used were the retromastoid-retrosigmoid in 23 cases, subtemporal in 5 cases, and combined retromastoid subtemporal presigmoid in the remaining 5. Total removal was achieved in 26 cases (79%); incomplete removal occurred in 7 cases (21%). The extent of tumor removal and operative morbidity were not significantly related to tumor size. Brain stem indentation, arterial and cranial nerve encasement, and epidural invasion were the main factors that prevented total tumor removal and influenced operative morbidity. There was no intraoperative mortality, but three patients (9%) died perioperatively. In the postoperative period, most patients went through momentary neurological deterioration, chiefly due to new cranial nerve deficits. The average follow-up was 4.3 years in 27 patients; of these 17 were unchanged and 10 were improved. Before surgery, only 13 patients were self-sufficient; at long-term follow-up, another 6 had achieved independence. Our experience suggests that, even though real petroclival meningiomas still represent a formidable surgical challenge, such tumors can in most cases be removed completely with low attendant mortality and acceptable morbidity.