Fifteen patients harbouring intracranial aneurysms were monitored postoperatively. A thermal diffusion flow probe was placed on the cortex, in an area of interest, and continuous recording of cerebral blood flow (CBF) was carried out for 1-4 days. Vasospasm can also be detected by clinical deterioration, increased velocity by transcranial Doppler, and by angiographical evidence of decrease in vessel calibre. A comparison was made between these methods of determining vasospasm and changes in CBF. Three patients had unruptured aneurysms and no evidence of vasospasm. These patients had a range of CBF from 44 to 100 cc 100 g-1 min-1 with a daily mean of 78 cc 100 g-1 min-1. One patient was operated on, in a delayed fashion, after a period of vasospasm on post-subarachnoid haemorrhage (SAH) Day 18 and was found to have a CBF of 40 cc 100 g-1 min-1 when the pCO2 was 21 Torr and the mean blood pressure (BP) was 99 Torr. The development of vasospasm was correlated with a CBF of less than 40 cc 100 g-1 min-1. Some patients with vasospasm also develop a significant drop in CBF during the period of monitoring. Using these criteria, CBF monitoring predicted the development of vasospasm in seven patients, was compatible with the clinical course in three patients and was misleading in two patients. In the seven patients in which CBF predicted vasospasm, changes were seen in CBF prior to changes observed by clinical examination, transcranial Doppler, and/or angiography.(ABSTRACT TRUNCATED AT 250 WORDS)