We report the epidemiological and clinical features of unspecified subarachnoid hemorrhage by reviewing the literature published in Taiwan. Data from stroke registry from January 1988 to June 1992 showed that 4.9% of 23910 acute strokes were diagnosed as subarachnoid hemorrhage; in about half of them the cause of hemorrhage was not specified due to lack of confirmatory diagnostic examinations or to rapid deterioration of consciousness subsequent to bleeding that made angiography unapplicable during hospitalization. The age distribution, sex ratio, clinical profiles and associated risk factors in this group of patients were similar to those in the group of aneurysmal rupture. Therefore, the majority of patients categorized in the group of unspecified subarachnoid hemorrhage very likely had aneurysmal ruptures. The causes of subarachnoid hemorrhage were more frequently undetermined in hospitals where neuroradiological as well as neurosurgical facilities and staff were not immediately available, and also in patients whose clinical condition was poor on arrival at the hospital or who deteriorated rapidly after the onset. These two facts may partly explain the higher case-fatality rate and poorer prognosis in patients whose causes of subarachnoid hemorrhage were undetermined and effective surgical treatment could not be done.