AbstractGeometry optimization for each member of a series of 15 substituted phenethylamines with reported psychotomimetic activity was performed using MM2 and MOPAC93 (AM1) methods. This was followed by calculation of previously developed, numerically unitary structure indices, viz., the molecular transform indices FTm, FTe, and FTc, indicating general, electronic, and charge properties respectively, and the analogous normalized molecular moment indices Mn, Me, and Mc. Each of the indices was used independently with the pharmacological data in a procedure to generate clusters containing three or more data points; linear correlations within each cluster were then determined. The Mn index performed best under the criteria, giving three clusters containing a total of 10 compounds; the other indices gave clusters accounting for totals of five to eight compounds for each index. Analysis of the clusters indicated that the 2‐ and 5‐positions on the phenyl ring should have substituents and that substitution at the 4‐position, preferably with a halogen, methyl, or thioether group, was critical for maximum pharmacological activity. The latter results suggest that ease of metabolic hydroxylation at the 4‐phenyl position (or the substituent group at that position, e.g., methyl, ethyl, etc.) is necessary for maximum activity. The performance of the Mn index appears to indicate more of a dependence on 3D spatial requirements for bioactivity, with molecular electronic and charge properties, respectively, being of secondary importance. Further, clustering and correlation regression of the biological data on the calculated logarithm of the octanol/water partition coefficient (CLogP) showed that the latter may be an important factor influencing pharmacological activity in this series. © 2004 Wiley Periodicals, Inc. Int J Quantum Chem, 2004