This research was directed at the pathophysiology of normal-pressure hydrocephalus. The experimental method consisted of accurate and simultaneous measurement of the pressure within the ventricle and over the cerebral convexity in cats with hydrocephalus but normal ventricular pressure. Hydrocephalus was induced by the intracisternal injection of kaolin. Prior to the induction of hydrocephalus, the difference between the ventricular pressure and the pressure over the convexity (the transmantle pressure) was small (0.27 +/- 0.31 cm saline, mean +/- standard deviation). After the induction of normal-pressure hydrocephalus in seven animals, there was a statistically significant elevation of the transmantle pressure to 3.4 +/- 3.9 cm saline (p less than 0.05, Student's paired t-test). There was no similar increase in animals injected with lactated Ringer's solution. This finding supports the theory that it is the transmantle pressure, and not the ventricular pressure, that is the physiological determinant of ventricular dilatation. The theory explains why hydrocephalus can develop and persist despite normal ventricular pressure.