Since the first reports appeared more than fifty years ago, using the computer to analyze structureâ activity relationships has grown into a vibrant field. The original work recognized the key role of the optimum hydrophobicity in the biological properties of molecules. It also put hydrophobicity on a scale to be used in a multivariate equation of substituent effects, also characterized with Hammett sigma constants and parameters for steric properties. Subsequent work led to computer programs to calculate octanolâ water log P from structure alone, as well as methods to treat categorical data. In addition, 3D QSAR has become a popular alternative strategy. Non-statistical methods for analyzing the data have been implemented and strategies to investigate and improve the predictivity of QSARs have been developed.