THE BACTERIOLOGICAL EVIDENCE 91 



species become adapted to a parasitic existence through the Natural 

 Selection of spontaneous variations. Doubtless unicellular organ- 

 isms, like the higher types, vary spontaneously. If that be admitted, 

 then, presumably, in the struggle with the cells, the microbes with 

 the weaker defensive powers, the weaker toxins, or the weaker 

 powers of producing toxins under unfavourable conditions, tend to 

 perish, while those with the stronger powers tend to survive. In 

 that case evolution must follow till the saprophytic race becomes 

 adapted to a parasitic habit of life, after which it would be maintained 

 in adaptation by the same process of selection. The facts that 

 microbic species evolve virulence only when in a position to be 

 selected by living cells, that this virulence and their other germinal 

 changes adapt them precisely to the particular species of animal 

 they attack, and that they tend invariably to lose the adaptation 

 (i.e. undergo retrogression) when no longer in a position to be so 

 selected, that these changes in both directions are continuous and 

 so gradual that many generations must come and go before they 

 are completed, and that they proceed no further than enables the 

 species to persist in the new environment, are so highly significant 

 that we have no alternative but to attribute them, like the adaptive 

 changes of the higher species, to selection on the one hand or 

 cessation of it on the other. They are too adaptive to be due to 

 any other cause. They fit the many and diverse species of uni- 

 cellular parasites too closely to their environments to be results of 

 mere coincidence, of mere accidental effects, produced by the 

 direct action of the environment. 



146. The belief that the increasing virulence of saprophytic 

 organisms, when they are becoming parasitic, results from the 

 direct action of the environment on their germ-plasm, is one of those 

 'obvious' inferences which, when all the evidence is taken into 

 account, are found to rest on a series of impossible assumptions. 

 Here, as always when considering any complex and difficult 

 problem, we cannot link together all the evidence unless we make 

 a rigorous deductive inference of consequences and follow it by an 

 appeal to reality. The generalization that microbic species change 

 adaptively when placed under right conditions in changed environ- 

 ments is really ' obvious ' ; it is an ' immediate ' inference. The 

 generalization that they change because their germ-plasm is directly 

 affected by the environment is not really obvious, and must be 

 tested. When the test is made, when we do not shirk the hard 

 thinking necessary, it is seen to be obviously inaccurate. 



147. Selection affords an explanation of the adaptive changes 



