The transport of alpha-aminoisobutyric acid (AIB), N-methyl-AIB (MeAIB), and diethylenetriaminepentaacetic acid (DTPA) from blood to brain was measured over different experimental periods in eight regions of the rat brain. Unidirectional transfer rate constants were determined from multiple-time/graphical and single-time analysis of the experimental data; values of 0.0018, 0.00057, and 0.000021 ml g-1 min-1, respectively, were obtained for the thalamus by graphical analysis. The initial distribution volume of AIB and MeAIB in brain tissue was several-fold greater than that of DTPA and the tissue plasma volume, and this difference was not accounted for by red blood cell uptake. This discrepancy could be due to rapid transport of AIB and MeAIB into brain endothelial cells in addition to the relatively rapid uptake by choroidal, meningeal, and ependymal associated tissues that was demonstrated by autoradiography. Thus, it may be misleading and erroneous to consider the blood-brain barrier (BBB) to be a simple, single-membrane structure when analyzing the blood-brain transfer data of solutes such as amino acids. The data from the ventriculocisternal perfusion experiments and previously published AIB uptake data in mouse brain slices were used to estimate the transfer rate constants across brain cell membranes. These studies indicated that the transport of AIB into brain cells was approximately 110 to 265 times greater than that across normal brain capillaries per unit mass of brain tissue, and that the BBB limits blood-to-brain cell transport of this amino acid. These observations (low rate of transport across normal brain capillaries and rapid concentrative uptake by brain cells) indicate that AIB is a good marker for measuring moderate to large increases in BBB permeability by experiments that require unidirectional flux of the tracer.