General anaesthesia can be induced by a wide variety of structurally dissimilar molecules. Consequently, the mechanism must involve some rather nonspecific interactions as the target site, generally held to be in nerve membranes. The primary site of action has been postulated to be either lipid or protein or both. Although recent work has cast doubt on the lipid hypotheses, protein models still flourish. In particular, Seeman and his co-workers have shown that biological membranes expand when anaesthetic molecules are added, and that this expansion is far greater than that which occurs with lipid bilayers. It has been suggested that this difference is due to extensive conformational changes in the membrane proteins, and several mechanisms have been proposed to explain this large expansion of proteins. We now report the first direct measurements of the volumes occupied by general anaesthetic molecules in both biological membranes and lipid bilayers. We show that, in fact, biological membranes expand much less and lipid bilayers expand more than previously reported. The volume that a general anaesthetic molecule occupies is essentially the same in biological membranes, lipid bilayers and water. Our results lead us to question all generalized membrane expansion hypotheses for the mechanism of general anaesthesia, in favour of hypotheses would include more specialized target sites.