120 ON VARIABILITY AND ADAPTATION 



the wild state, are said to revert to type. Now with the bacteria, 

 where such races have been developed, reversion to type may 

 still be brought about by a further process, namely, by a further 

 alteration of environment. I know of no case quite analogous 

 to this among the higher animals. 



It is only within strictly limited conditions of environment 

 that the characteristics of the usual form or of a race can be 

 preserved ; overstep these limits, make certain alterations in 

 the environment, and type and race alike are liable to vary, 

 although the liability is perhaps the greater for the race to 

 approximate towards the type than for the type to produce 

 well-marked races. The only cases that I can call to mind 

 where this law apparently does not hold good are both of modifi- 

 cations of the anthrax bacillus. It is possible so to attenuate 

 the anthrax bacillus that a race is developed which is absolutely 

 non-pathogenic even for the most susceptible of small rodents, 

 and so far no efforts have succeeded in re-establishing the 

 virulence of such attenuated forms. Roux's asporogenous race 

 also would seem thus far to remain asporogenous under all 

 conditions, although variations may be induced in its virulence. 

 Even these cases, however, will probably be found eventually 

 not to be exceptions to the rule. 



For these reasons I have throughout spoken of types and 

 races rather than of species and subspecies or varieties, for by 

 employing the more uncommon terms in this connection I have 

 attempted to guard myself against appearing to hold that among 

 the bacteria there is anything beyond a relative permanency. 

 A study of the observations that have been made up to the 

 present time does not lead to the conclusion that new species 

 are readily developed ; all that it clearly indicates is that among 

 the bacteria our employment of this term " species " must be 

 elastic — the limit to which the individual may depart from the 

 type (and from which it may again revert to type) is very far 

 removed. But granting freely that not one of the cases here 

 described records the development of a new species, this must at 

 least be admitted, namely, that the longer the action of any one 

 factor upon the bacterial growth, the longer the time requisite 

 to ensure a return to the typical condition ; the stronger the 

 impress left upon any individual microbe, the longer, again, the 

 time requisite to ensure a similar return. If, therefore, within 



