Ninety-four patients with hydrocephalus following subarachnoid haemorrhage were investigated. Fourteen of these patients had temporary ventricular drains inserted, 19 had drains that were later converted to ventricular-peritoneal shunts, and 61 patients had shunts only. All patients were analysed with respect to their grade on admission, the distribution of blood on CT scan, their Glasgow Coma Score, their intracranial pressure, and the presence or absence of vasospasm on angiography. Outcome was analysed by shunt procedure and time of shunt insertion. The higher the grade on admission, the more likely it was the patient would require a shunt. Fewer patients with a good outcome required shunting as compared to those whose outcome was poor. Patients having more subarachnoid blood on CT scan tended to present with a lower Glasgow Coma Score. The outcome of patients with intraventricular haemorrhage was not obviously influenced by the insertion of a ventriculostomy. Vasospasm was not more common in patients requiring a ventriculo-peritoneal shunt, nor did early shunting in patients with hydrocephalus affect the incidence of vasospasm. The majority of patients were either shunted in the first three days or after 30 days post-SAH. The outcome was better in those patients shunted at a later date.