Blood-brain barrier permeability studies made in man using the indicator dilution method revealed that the extraction of the test substance increases during the upslope of the venous (outflow) dilution curve. The present study aimed to obviate the possibility that this could result from intravascular phenomena, such as interlaminar diffusion (the result of differences in molecular size) and erythrocyte carriage. Several reference substances were employed for the determination of the extraction in order that careful correction could be made for differences in intravascular behavior of the test and reference substance. The test substances studied were D-glucose, L-phenylalanine, water, propranolol, and benzodiazepines, representing both carrier-transported and lipophilic substances. In-diethylenetriamine pentaacetic acid, Na+, Cl-, L-glucose, and L-lysine were employed as reference substances. For all the substances tested, and after correction for intravascular phenomena, the extractions were found to increase during the initial part of the dilution curve. This increasing extraction can be ascribed to heterogeneity of the cerebral circulation; the higher extraction corresponds to longer contact with the blood-brain barrier and indicates a longer transit time. Signs of heterogeneity were also present when blood flow was elevated above normal. Any influence that heterogeneity might have on the mean extraction value can be minimized by using an appropriate calculation of the extraction of the test substance.