We have shown that a rebound of intracranial pressure (ICP) occurring after decompression of an intracranial mass lesion is a threshold phenomenon dependent upon the cerebral perfusion pressure (CPP) during compression and the duration of the compression. In the present study regional cerebral blood flow (rCBF) was measured during balloon compression of a degree critical for the development of a postdecompression rebound. The effects were compared with those of hydrostatically raised pressure which under similar conditions rarely produces a rebound of ICP. Disproportionately marked reductions in flow occurred in the hemisphere ipsilateral to the balloon, especially in white matter and in cortex adjacent to the balloon with flow values of, respectively, 1.1 ± 0.9 and 6.4 ± 3.4 ml 100 g–1 min–1. The differences in flow between balloon and hydrostatic compression were found to be due to an increased cerebrovascular resistance (CVR) caused by a direct compressive effect by the balloon overriding the generalized vasodilation which occurs in response to the raised ICP. Thus the increase in CVR attributable to compression by the balloon added to the reduction in CPP caused by the diffuse increase in ICP. As a consequence flow in large regions of the brain was reduced below the thresholds for structural infarction and for ischaemic damage to the blood-brain barrier.