A quality-control phantom is described which is capable of testing the performance of inhalation rebreathing systems for measurements of regional cerebral blood flow (rCBF). The phantom is designed so that data similar to those from a patient study are obtained. The standard rCBF data-processing algorithm is used for the calculation of phantom study results. Malfunctions were induced to simulate instrument malfunction, and their effects on detector and air curves were evaluated. The use of this phantom is recommended prior to a patient study if the rCBF instrument is used infrequently or after service for routine maintenance or malfunction.