On the variation of toxic effects over species, its cause, and analysis by "structure-selectivity relations".

The intractable problems of how to select and to develop a toxicant against a particular target species while leaving most other species unaffected are analyzed. It is proposed to study the mean and variation over species in general, as tolerance is concerned, as a function of the structure in much the same way that tolerances of target species are analyzed by quantitative structure-activity relations (QSAR). This approach enables the dissection of the complex problem into three parts, which together provide rational information on how it may be solved or at least on how decisions should be made. It is argued that variation in tolerance over species, in general, must be due to variation in binding energy over macromolecular binding sites in general. The latter variation likely depends on the surface area or, better, on the "complexity" of the toxicant. It is however, not clear as to how to express complexity in an appropriate way. To illustrate the approach, it was applied to a limited series of toxicity data for seven species, from five taxonomic classes, to seven organophosphorus insecticides. A quantitative measure of the degree of variation over species correlates significantly with the connectivity index, as developed by Randić (1975, J. Amer. Chem. Soc. 97, 6609-6615), a known correlate of the surface area of the molecule. Due to several limitations, this finding mainly serves to stimulate further work on more extended data series.