THE ORIGINAL idea of a blood-brain barrier arose at the turn of the century as an explanation of the properties of the living brain in relation to certain dyes and pharmacologically active compounds, injected either into the bloodstream or directly into the cerebrospinal fluid (CSF) of experimental animals (reviewed by Bradbury, 1979). This developed into the general concept of a cellular barrier between blood and the interstitial fluid of brain (Krogh, 1946), but the hypothesis had no firm structural basis and was abandoned by many when early electron micrographs showed no differences between capillaries in brain and those elsewhere, eg, in skeletal muscle (Maynard et al., 1957). In the late 1960's, the use of various hemeproteins and other markers which could be made visible under the electron microscope established unequivocally that the barrier within the brain to molecules of 1,800 daltons and above was the endothelium of cerebral capillaries (Reese and Karnovsky, 1967; Brightman and Reese, 1969). Since that time, there has been an explosive increase in experimental work on the properties of transport across the barrier. The final objective must be to elucidate the cellular and molecular mechanisms involved in this transport, and evidence about these is now beginning to appear. A number of reviews and monographs are available which deal both with general features of structure and function (Davson, 1976; Rapoport, 1976; Bradbury, 1979; Fenstermacher and Rapoport, 1984) and with specific aspects of barrier function. Particular topics for which there are recent and excellent reviews are substrate supply to the brain (Pardridge and Oldendorf, 1977; Crone, 1980; Pardridge, 1983), glucose transport (Lund-Andersen, 1979), modulation of transport (Gjedde, 1983; Gjedde and Crone, 1983), mechanism of ion distribution (Fenstermacher, 1975), volume regulation of the brain (Fenstermacher, 1984), comparative physiology (Cserr and Bundgaard, 1984), handling of monoamines (Hardebo and Owman, 1980), peptides, and the blood-brain barrier (Meisenberg and Simmons, 1983; Pardridge, 1984), and function particularly in relation to studies with isolated suspended microvessels (Goldstein and Betz, 1983; Joo, 1984). My objective here will be to consider how the ultrastructure of the cerebral endothelium and adjacent structures differs from that elsewhere and how this gives a basis for the transport properties of the barrier. Also of particular interest in relation to the physiology of the microcirculation in general are the properties of net fluid movement across the cerebral vessels and the recent demonstration of a relation between apparent permeability of the barrier and cerebral blood flow.