The traditional belief about ethanol's mechanism of action is based on ethanol's lipophilicity and capability to penetrate and disorder lipid bilayers. This traditional belief is now being supplanted by growing evidence that ethanol has relatively selective actions on certain synaptic receptors, such as those for NMDA, serotonin, and GABA. It was recently argued that these receptor specificities are secondary to a preferential ability of ethanol to displace membrane bound water in the domains of certain receptors. The data obtained in this study are consistent with the original hypothesis: any disorganization of cellular water by ethanol will be detectable by proton nuclear magnetic resonance (NMR) spectroscopy. In particular, the relaxation times of water hydrogen protons reflect how constrained water molecules are by the macromolecules within cells. The relaxation time of "bulk" water is lengthened relative to water molecules that are under the influence of electromagnetic fields of macromolecular surfaces within cells. Here, we tested this hypothesis in living fish, which dosed themselves by swimming in water that had added ethanol. Estimates of brain alcohol at 5 min after initial exposure revealed that the brain concentration was only about 1/3 that of the water in which they were swimming. The average value of the NMR relaxation time T1, but not T2, was decreased at 5 min (when brain concentrations were on the order 100 mM) and reached statistical significance at 10 and 30 min after initial exposure.(ABSTRACT TRUNCATED AT 250 WORDS)