In view of the current concern that rapid infusion of mannitol might initially aggravate intracranial hypertension, the effects of a mannitol infusion on lumbar cerebrospinal fluid pressure (CSFP) were investigated in 49 patients. The studies were performed when the patients were under general anesthesia prior to elective craniotomy for tumor resection or intracerebral aneurysm clipping. The patients were divided into two groups: 24 patients with normal CSFP (Group I, mean CSFP 10.5 mm Hg) and 25 with raised CSFP (Group II, mean CSFP 20.8 mm Hg). Measurements of CSFP, mean arterial blood pressure (MABP), and central venous pressure (CVP) were made serially during and after the infusion of 20% mannitol (1 gm.kg-1 infused over a 10-minute interval). In both groups, mannitol infusion provoked a fall in MABP and an increase in CVP. An immediate decrease [corrected] in CSFP was observed in Group II, whereas CSFP increased transiently but significantly in Group I. Analysis of the arterial and venous driving pressures which contribute to CSFP suggests that the transient increase in CSFP after mannitol in Group I was partly due to the increase in CVP. The presence of intracranial hypertension may thus alter the CSFP response to arterial and venous pressure changes. Cerebral blood volume (CBV) was measured in dogs in a separate study analogous to the human protocol. The CBV increased approximately 25% over control values after mannitol infusion both in the normal animals and in those with CSFP raised by an epidural balloon. The response of the CSFP to mannitol infusion differed between both groups in a fashion similar to that observed in the human subjects. Thus, differences in CBV changes after mannitol do not account for the difference in CSFP response between normal subjects and those with raised CSFP.