Several homologous series of 9-anilinoacridines, each bearing a different pKa modulating acridine substituent, have been synthesized and screened in the L1210 leukemia system, to examine if measures of agent lipophilic-hydrophilic balance utilized in regression analysis should be corrected for the effects of changing base strength. The measure of tumor selectivity modeled was the maximum increase in life span in L1210 tests (ILSmax), and dose potency was gauged as D40, the molar drug dose necessary to provide 40% extension in life span. Agent lipophilic-hydrophilic balance was measured as chromatographic Rm values, and the pKa modulating factors examined were log (1 - ±) and log ±, where ± is the fraction of drug ionized at physiological pH. Regression analyses of data from 78 L1210-active compounds show that the measured Rm values, unmodified by pKa correction factors, furnish superior correlation equations. An equation in Rm2, with indicator variables denoting the presence of acridine 3-NO2 or 4-CONRR substituents, successfully models ILSmaj. To develop successful regression equations for D40 it was necessary to restrict attention to close structural congeners which are likely to be metabolized by similar routes. Results for a series of 3-nitroacridine derivatives which may be reduced in vivo to more dose-potent and/or more hydrophilic compounds could not be incorporated. Acceptable equations developed for D40 contain terms in Rm, Rm2, and pKB or t1/2. The latter provides a measure of the rate of agent thiolytic cleavage, a prominent contributor to drug decay in vivo. © 1979, American Chemical Society. All rights reserved.