Patients considered to have a possible disorder of the circulation of the cerebrospinal fluid (CSF) were prospectively randomized on clinical and computed tomographic grounds into one of four categories: low probability of a disorder of circulation of the CSF (n = 14); high probability of a disorder of the circulation of the CSF (n = 10); low probability of shunt malfunction (n = 10); and high probability of shunt malfunction (n = 9). Patients with possible disorders of the circulation of the CSF who did not meet the entry criteria for each of these categories were excluded from this study. A ventricular catheter connected to a subgaleal Rickham reservoir not in continuity with a shunt system (if this was present) either was inserted into each patient or was in place from previous surgery. Infusion studies were conducted by the infusion of 1 ml/min of normal saline through a 25-gauge needle inserted through the skin into the Rickham reservoir. A separate 25-gage needle was inserted into the Rickham reservoir for continuous recording of intracranial pressure. There were 43 infusion studies that were included in the four categories. Recordings of baseline intracranial pressure for a high probability of a disorder of the circulation of the CSF did not differ significantly from a low probability of a disorder of the circulation of the CSF. Baseline intracranial pressure for a high probability of shunt malfunction was significantly higher than the low probability of a disorder of the circulation of the CSF group and the low probability of shunt malfunction group; there was, however, marked overlap in values between these groups.(ABSTRACT TRUNCATED AT 250 WORDS)