Morphological evidence of the potential for adsorptive transcytosis of protein through the mammalian blood-brain fluid barriers, first reported from this laboratory in the mouse, has been confirmed and expanded upon in rats injected intravenously or into the lateral cerebral ventricle/subarachnoid space with with exogenous lectin wheatgerm agglutinin (WGA) conjugated to horseradish peroxidase (HRP). Blood-borne WGA-HRP rapidly enters cerebral endothelia by the process of adsorptive endocytosis and labels the vascular tree throughout the CNS. At 3 h post-injection and longer, WGA-HRP occupies the perivascular clefts and labels perivascular cells and basal lamina; this suspected transendothelial transfer of the lectin conjugate from blood to brain involves specific constituents of the endothelial endomembrane system of organelles (e.g., plasmalemma, vesicles, endosomes, Golgi complex). Within 6 h, reaction product is evident in extracellular clefts beyond the perivascular basal lamina and labels endocytic vesicles, endosomes, and dense bodies within cells and processes of the neuropil. Exposure of the abluminal surface of blood-brain barrier endothelia for 1-18 h to WGA-HRP delivered into the cerebral ventricles or subarachnoid space indicates blood-brain barrier endothelia do not engage in demonstrable adsorptive endocytosis at the abluminal surface. In this preparation, no endothelial organelles comparable to those sequestering blood-borne WGA-HRP are labelled with the lectin conjugate; hence, significant adsorptive transcytosis of WGA-HRP through cerebral endothelia from brain to blood is unlikely. The demonstrable difference in membrane internalization of the luminal versus abluminal plasmalemma of blood-brain barrier endothelia suggests the blood-brain barrier is polarized regarding adsorptive endocytosis of WGA-HRP. If adsorptive transcytosis of macromolecules through the blood-brain barrier does occur, the process appears unidirectional, from blood to brain but not from brain to blood. Absence of demonstrable endocytosis at the abluminal front is an enigma in the scheme of transcytosis through the blood-brain barrier from blood to brain insofar as exocytosis and endocytosis are complementary events in the cellular secretory process. This unconventional membrane behavior associated with the abluminal plasmalemma argues against a significant transcytosis of blood-borne protein through blood-brain barrier endothelia. The potential for transcytosis of macromolecules through the blood-cerebrospinal fluid barrier of choroid plexus epithelia is not as problemmatic as that through blood-brain barrier endothelia; additional evidence is provided to suggest choroid plexus epithelia participate in adsorptive endocytosis circumferentially and adsorptive transcytosis of WGA-HRP bidirectionally between the blood and cerebrospinal fluid.