The outflow conductance (C(out)) of the cerebrospinal fluid (CSF) system is a parameter considered to be predictive in selection for hydrocephalus surgery. C(out) can be determined through an infusion test. A new apparatus for performing infusion tests in a standardized and automated way was developed. The objective was to evaluate repetitiveness as well as to propose and evaluate a method for real time estimation of the reliability of individual C(out) investigations. Repeated investigations were performed on an experimental model simulating the CSF system, and on 14 patients with hydrocephalus. DeltaC(out), calculated as the 95% confidence interval of C(out), was introduced as an estimate of the reliability of individual C(out) investigations. On the model, no significant difference was found between DeltaC(out) and the actual C(out) variation in repeated investigations (p = 0.135). The correlation between the first and the second patient investigation was high (R = 0.99, p < 0.05), although there was a significant difference between the investigations (p < 0.05). The standard deviation of difference was 2.60 microl (s kPa)(-1). The repetitiveness of C(out) with the new apparatus was high, and DeltaC(out) reflected the reliability of each investigation. This feature has to be taken into account in every individual case, before making a decision or performing research based on measurements of C(out) in the future.