The authors review the literature on subarachnoid haemorrhage of unknown aetiology (SAHUE) and analyze a personal series of 212 patients diagnosed as SAHUE. These patients represent 30% of all cases of primary SAH admitted over a 14.5 year period. The age, sex, antecedents and initial clinical presentation of patients with SAHUE were indistinguishable from those of patients with subarachnoid haemorrhage due to ruptured aneurysm (SAHRA). However, the present series of SAHUE compare favourably with both a personal and a previously reported series of SAHRA insofar as clinical grade on admission (94% of patients in grades I-II of Botterell), presence of blood on CT (51%), vasospasm (5%), ischaemic deficits (3.3%), persistent hydrocephalus (3.5%), rebleeding (6%) and fatal result (3.9%) are concerned. The amount of blood on CT scan in our patients with SAHUE was associated with a significantly higher incidence of brain ischaemia and hydrocephalus but did not correlate with the Botterell grade on admission or final outcome, which were good in the majority of cases regardless of the presence or not of visible cisternal haemorrhage. The results of the present series confirm that the final prognosis of patients with primary SAH showing normal four-vessel cerebral angiography is essentially favourable.