A combined experimental and theoretical study is performed on binary dilauroylphosphatidylcholine/distearoylphosphatidylcholine (DLPC/DSPC) lipid bilayer membranes incorporating bacteriorhodopsin (BR). The system is designed to investigate the possibility that BR, via a hydrophobic matching principle related to the difference in lipid bilayer hydrophobic thickness and protein hydrophobic length, can perform molecular sorting of the lipids at the lipid-protein interface, leading to lipid specificity/selectivity that is controlled solely by physical factors. The study takes advantage of the strongly nonideal mixing behavior of the DLPC/DSPC mixture and the fact that the average lipid acyl-chain length is strongly dependent on temperature, particularly in the main phase transition region. The experiments are based on fluorescence energy transfer techniques using specifically designed lipid analogs that can probe the lipid-protein interface. The theoretical calculations exploit a microscopic molecular interaction model that embodies the hydrophobic matching as a key parameter. At low temperatures, in the gel-gel coexistence region, experimental and theoretical data consistently indicate that BR is associated with the short-chain lipid DLPC. At moderate temperatures, in the fluid-gel coexistence region, BR remains in the fluid phase, which is mainly composed of short-chain lipid DLPC, but is enriched at the interface between the fluid and gel domains. At high temperatures, in the fluid phase, BR stays in the mixed lipid phase, and the theoretical data suggest a preference of the protein for the long-chain DSPC molecules at the expense of the short-chain DLPC molecules. The combined results of the experiments and the calculations provide evidence that a molecular sorting principle is active because of hydrophobic matching and that BR exhibits physical lipid selectivity. The results are discussed in the general context of membrane organization and compartmentalization and in terms of nanometer-scale lipid-domain formation.