(1) Three main barrier layers at the interface between blood and tissue protect the central nervous system (CNS): the endothelium of brain capillaries, and the epithelia of the choroid plexus (CP) and the arachnoid. The classical work on these barriers in situ until the 1970s laid the foundations for modern understanding. Techniques for brain endothelial cell isolation and culture pioneered by Ferenc Joó in the 1970s opened up new fields of examination, enabling study of mechanisms at the cellular and molecular level. (2) Astrocytic glial cells are closely associated with the brain endothelial barrier. During evolution the barrier appears to have shifted from the glial to the endothelial layer, in parallel with the increasing importance of the microvasculature and its regulation. Vestiges of the barrier potential of glia remain in the modern mammalian CNS. (3) Evolutionary evidence suggests that the advantage derived from ionic homeostasis around central synapses was the major selective pressure leading to refinement of CNS barrier systems. This is one element of the modern 'multitasking' barrier function. (4) While epithelia are constitutively able to form barriers at appropriate interfaces, the 'default' condition for endothelia is more leaky; inductive influences from associated cells especially astrocytes are important in generating the full blood-brain barrier (BBB) phenotype in brain capillaries. The underlying mechanisms are being elucidated at the molecular and genomics level. (5) The barrier layers of the nervous system can be modulated by a number of receptor-mediated processes, involving several signal transduction pathways, both calcium dependent and independent. Some agents acting as 'inducers' in the long term can act as 'modulators' in the short-term, with some overlap of signaling pathways. Modulating agents may be derived both from the blood and from cells associated with cerebral vessels. Less is known about the modulation of the CP. (6) The challenge for the next era of CNS barrier studies will be to apply new knowledge from proteomics and genomics to understanding the in vivo condition in physiology and pathology.