Neurophysiological, neurochemical and behavioral studies of the effects of ethanol on the nervous system have so far failed to identify specific, direct, primary mechnisms of action that may account for the typical pattern of alcohol intoxication in vivo. Electroencephalogram and evoked response studies indicate biphasic effects in the intact subject, which may correlate better with the level of arousal than with a specific drug action. Effects on spinal reflexes are also biphasic, probably representing the net result of direct influence on resting membrane potential, primary afferent depolarization, and neurotransmitter release. With the exception of its inhibitory effect on release of oxytocin, vasopressin and possibly other hypothalamic peptides, ethanol does not appear notably different in its spectrum of effects from a wide range of other hypnotics, anesthetics and minor tranquilizers. Interpretation of the findings is complicated by the fact that functional alteration of any given neuronal system by ethanol in vivo may reflect a) direct local action of ethanol on the cells under study, b) change in the input to those cells because of an action elsewhere in the nervous system, c) effects of ethanol metabolites, or d) indirect consequences of decreased blood flow, oxygen or metabolite supply, hormonal action, or hypothermia, due to disturbances of homeostasis in the whole body as a result of deep intoxication. To date, attempts to circmvent b, c and d by the study of brain tissue in vitro have shown consistent effects of ethanol only at concentrations well above those that are meaningful in vivo. Relatively specific patterns of action of different drugs in vivo may prove to be largely dependent on their customary rates and routes of administration, and on summation of minor differences in the dose-response curves with different types of neuron, even though the basic types of molecular action may be essentially similar.