Review of cerebral arteriograms of all our patients [19] with epidural hematoma over a 2 year period revealed evidence of localized extrinsic compression of the cerebral cortex in all cases. This finding, although not as specific as those previously described for epidural hematoma, was strongly suggestive. A 2 year prospective study was then initiated employing internal, rather than common carotid arteriography, because opacification of the external carotid artery is not necessary for evaluation of cortical compression. In this study, which consisted of 21 patients, epidural hematoma was correctly diagnosed in all, relying primarily on signs of localized extrinisic cerebral cortical compression. No epidural hematoma was misinterpreted as being a subdural hematoma or an intracerebral injury, and all epidural hematomas that were present were diagnosed correctly. We conclude that internal carotid arteriography is a sensitive method for diagnosing epidural hematoma. The signs of extrinsic cortical compression should be emphasized in the interpretation of arteriograms of patients with head injuries whether the internal or the common carotid artery is injected.