Extensions and refinements of the receptor mapping method as originally developed by Crippen are presented. In a set of newly developed algorithms measures are taken to reduce the number of required energy parameters to a statistically acceptable degree. The most important measure is the incorporation of lipophilicity as a hydrophobic bonding parameter to describe the binding of parts of the ligands to lipophilic areas on the receptor. In order to test the applicability of our set of programs, we mapped the turkey erythrocyte beta receptor using a data set of Bilezikian. It was found that the experimentally determined free energies of binding can be reasonably described using a nine-point geometrical representation of the receptor site and only six energy parameters. The deduced model predicts that the phenyl rings of phenylethanolamines and phenoxypropanolamines occupy different parts of the receptor site.