A central question which must be resolved before acceptable molecular descriptions of facilitated diffusion systems can be provided is the nature of the spatial and functional relationships between the transport proteins and the membrane lipids. In the work reported here, this question was addressed by investigating the dependence of the rates of glucose and uridine facilitated diffusion in human erythrocytes on membrane lipid fluidity. Two approaches were used to alter the lipid fluidity: treatment with ether, an anesthetic, and the exchange of a synthetic 3-ketosteroid, cholest-4-en-3-one, for membrane chloesterol. Both of these treatments result in a significant increase in membrane lipid fluidity, as judged by the increase in the rates of passive diffusion of uridine through cell membranes and of glucose through membrane lipid bilayer vesicles. Ether produces no change in the Km of either transport process, a slight decrease in the V for glucose transport, and no significant change in the V for uridine transport. Replacement of membrane cholesterol by cholest-4-en-3-one reduces the V for glucose transport slightly, without altering the Km, and reduces both the Km and V for uridine transport. The absence of the expected increase in the V of facilitated diffusion with increasing membrane lipid fluidity observed here with human erythrocytes is not consistent with models for the transport process which feature movement of transport proteins which are in direct contact with the bulk lipids of the membrane.