A central belief about ethanol is that it acts mainly by partitioning into the lipid bilayer of membranes. Newer ideas focus on the neuronal synapse and suggest that ethanol can allosterically change protein conformation, as is suggested by studies on GABA-receptor-mediated chloride uptake and on (Na(+)-K+)-ATPase. Several studies from my laboratory suggest that ethanol enhances enzymatic cleavage of sialic acid (SA) from gangliosides, and perhaps also glycoproteins, but does so without stimulating enzyme activity, suggesting conformational changes that affect accessibility. I propose a new model for the cell membrane in the synaptic region, which features gangliosides surrounding membrane proteins, with an interspersed film of water creating hydrogen bonds that anchor SA moieties to membrane protein. I believe that we should consider the possibility that an important action of ethanol, and polar anesthetics, is due to hydrophilic, not hydrophobic, properties and the ability to dehydrate the cell-surface microdomain. Our laboratory has recently advanced the theory that ethanol dehydrates a "solvent regulatory site" of membrane (Na(+)-K+)-ATPase. This principle might be extended to other enzymes and receptor proteins, as well as to the accessibility of sialoglycoconjugates to sialidase (neuraminidase). Hydrogen bonding between SA and polar regions of receptor protein, and the conformation on both imposed by it, would surely be changed by minor degrees of dehydration and substitution of alcohol molecules for water. Ethanol, unlike water, can only hydrogen bond "at one end." Displacement of water by ethanol would not only "free" the SA groups and make them more vulnerable to enzymatic cleavage but also could simultaneously change the conformation of receptor protein. Similarly, ethanol may displace water that links the polar heads of phospholipids to polar portions of receptors proteins. Ethanol may have an even more important and direct effect of substituting for hydrogen-bonded water within protein itself.