his studies in isolysins which inaugurate a physiology peculiar to the individual as opposed to the normal physiology of the species; his idea of a canonical study of the blood-serum in health and disease, so that a norm or "blood-canon" for the investigation of new or unknown conditions may be available; his demonstration that cancer may be changed into sarcoma by successive inoculations; that the growth of cancer in an animal body depends upon the presence of certain food-stuffs in that body; finally the gigantic labors involved in building up hundreds of new compounds and testing them as remedies for the two great groups of parasitic diseases, the spirilloses and the trypanosomioses, his prospective success being one of the greatest triumphs of the method of "trial and error" on record. Surely a career in scientific medicine only matched in recent times by those of Pasteur, Helmholtz, Koch and Lister.
To the orthodox chemist, who works by rule and formula, Ehrlich's experimental methods might seem mere haphazard "test-tubing," and he himself has jestingly referred to his laboratory procedure as "Spielchemie," an epithet which well describes the experimentation that results from the free play of a singularly acute mind. Some of his admirers have even gone so far as to say that he declines to work quantitatively. Although his own statement (in the Harben lectures) is just to the opposite effect—and all experiment that seeks the general law behind related facts is obviously quantitative in its intention—it is quite true that Ehrlich has steadfastly declined to follow Arrhenius in applying the quantitative methods of physical chemistry to the unknown entities of immunity reactions. "I have always emphasized the chemical nature of the reaction," he says, only "the formulas devised by Arrhenius and Madsen for the reactions of toxins and antitoxins explain absolutely nothing. Even in particularly favorable cases they can merely represent experimental results in the form of interpolation formulas."[1] In Ehrlich's view, the mistake made by the distinguished Swedish physicist lies in the assumption that toxins and antitoxins can be treated mathematically as simple indivisible substances, whereas there is unimpeachable evidence of their dual and multiple nature—that they can be split up into labile components of such extreme complexity as to have, so far, defied ultimate analysis. Under these conditions assumed constants become inevitably dependent variables and the physical chemist is dealing with the shifting evanescent aspects of substance in the labile state. The careful quantitative work of such a competent experimenter as Dr. W. H. Manwaring[2] has shown that, with the knowledge at present available, physico-chemical measure-
