CEREBROSPINAL FLUID SHUNTING is an important method of treating hydrocephalus. It sometimes has been difficult to achieve the optimum CSF pressure and ventricular size after shunt placement. Hydraulic and mechanical mismatching may occur, creating either overdrainage, as manifested by slit ventricle syndrome and/or subdural hygromas; or underdrainage, as manifested by a failure of the ventricles to change in size, with persistence of symptoms. A variable pressure valve (Codman-Medos Programmable Hakim valve system) has been developed to allow non-invasive increase or decrease in cerebrospinal fluid pressure in differentials of 10 mm of water within a range of 30–200 mm of water. This allows the neurosurgeon to adjust the pressure to the patient's particular clinical needs at any given moment. As part of a pilot study of the valve's safety, 13 patients in this center were treated with this programmable shunting system from October 1991 to January 1993. Twelve patients were over age 18 and one was younger than this. Six patients had previously been shunted with other valve systems, and the remaining seven had never been shunted before. The etiology of hydrocephalus included idiopathic normal pressure hydrocephalus (6 patients), aqueductal stenosis (3 patients), Chiari malformation (2 patients), meningitis (1 patient), and unknown etiology (1 patient). This paper describes the use of this valve in five illustrative cases; slit ventricle syndrome in an adult, chronic ventriculomegaly with aqueductal stenosis (2 cases), and idiopathic normal pressure hydrocephalus (2 cases). This programmable valve has been particularly useful in gradually decreasing ventricle size in idiopathic normal pressure hydrocephalus. In one case a subdural hematoma occurred as the pressure was diminished several months after surgery; it was treated by drainage and increasing the valve pressure for a period of several months without tying the shunt off. It has also been particularly useful in slit-ventricle syndrome management, and in gradually lessening the pressure in the huge ventricles of aqueductal stenosis. In this study several patients required a wide range of pressure changes during the period of observation, in some as much as 110 mm H2O. Several others were sensitive to changes as little as 10 mm H2O. All of the patients have required several pressure reprogrammings.