An attempt has been made to fit the amino acid sequence of bacteriorhodopsin to the three-dimensional density map of the molecule. First, seven segments of the sequence were selected as being probable transmembrane alpha helices. Then each of the 5040 possible ways of fitting these seven segments into the seven regions of helical density in the map were evaluated based on the criteria of connectivity of the nonhelical link regions, charge neutralization, and total scattering density per helix. A single model that may be experimentally tested emerged as the most probable.