The direct vasomotor effects of a number of agents used in the therapy of cerebral circulatory and metabolic disease were studied in isolated segments of cat pial vessels in vitro. Studies were repeated eight times generally, thus permitting the calculation of the mean maximum response—termed EAm (in dynes)—and of the molar agonist concentration required to effect the half-maximal response, termed EC50. The results obtained with various agonists were compared to a well-documented vasodilator, acetylcholine ( EAm = −780 dyn; EC50 = 0,029 μm) and a known vasoconstrictor, 5-hydroxytryptamine ( EAm = + 1630 dyn; EC50 = 0.036 μm). Papaverine and its derivatives effected a relaxation of the pial arteries with the following order of potency: YC-93 > papaverine > naftidrofuryl > viquidil. The EC50 for papaverine was 1,5 μm. Methyl xanthine derivatives (aminophylline, pentoxifylline, and theophylline) were essentially inactive, In contrast, drugs that are known to be capable of decreasing the volume of an experimental infarction, many of which are described as α-adrenolytic agents, contracted the isolated cerebrovascular smooth muscle, Their order of efficacy, based on the mean EAm values, was ifenprodil > vincamine > nicergoline > dihydroergotoxine > raubasine, In addition, it was considered worthwhile to determine whether the ifenprodil-induced vasoconstriction occurred when human, rather than cat, pial vessels were studied, Ifenprodil and vincamine contracted the human vessels in a biphasic manner; the EC50 was calculated to be 2,0 and 30 μm. Based on the above observations, the following comments would appear justified, Firstly, an increase in cerebral blood flow does not necessarily mean a compound is a direct vasodilator. Secondly, the vasoconstrictory actions of some agents correlate well with their anti-ischaemic properties (e.g., ifenprodil, vincamine, and nicergoline), Lastly, one action of effective anti-ischaemic agents might be to reduce flow, by vasoconstriction, in hyperaemic tissue: the inverse “steal” effect.