A rational approach to the design of centrally acting agents is presented, based initially upon a comparison of the physicochemical properties of three typical histamine H2 receptor antagonists which do not readily cross the blood-brain barrier with those of the three brain-penetrating drugs clonidine (6), mepyramine (7) and imipramine (8). A good correlation was found between the logarithms of the equilibrium brain/blood concentration ratios in the rat and the partition parameter, delta log P, defined as log P (1-octanol/water)-log P (cyclohexane/water), which suggests that brain penetration might be improved by reducing overall hydrogen-bonding ability. This model has been employed as a guide in the design of novel brain-penetrating H2 antagonists by the systematic structural modification of representatives of different structural types of H2 antagonists. Although marked increases in brain penetration amongst congeners of cimetidine (1), ranitidine (9), and tiotidine (10) were achieved, no compound was found with an acceptable combination of H2 antagonist activity (-log KB in the guinea pig atrium greater than 7.0) and brain penetration (steady-state brain/blood concentration ratio greater than 1.0). Conversely, structural modification of N-[[(piperidinyl-methyl)phenoxy]propy]acetamide (30) led to several potent, novel compounds which readily cross the blood-brain barrier. One of these, zolantidine (SK&F 95282, 41), whose -log KB is 7.46 and steady-state brain/blood ratio is 1.4, has been identified for use in studying histaminergic H2 receptor mechanisms in brain. Comparison of delta log P values with the logarithms of the brain/blood ratios for 20 structurally diverse compounds for which data became available confirms a highly significant correlation and supports the general validity of this model.