METHODS 123 



added, .bacteridium A is little by little destroyed through 

 variations in its properties. In a phenicated bouillon 

 there is superposition of these two phenomena, that is, 

 multiplication and variation at the same time. And if you 

 transfer the bacteridia which have varied into a bouillon 

 realizing condition No. 1 for the species, they will multiply in 

 their own likeness, indeed, but with the variations acquired 

 during the period of destruction. 



Now the general fact which serves as a basis to all Dar- 

 winian explanation is as follows : Subject a species to any 

 conditions whatever allowing it to continue to live ; it 

 multiplies with variations which, as soon as the cause of 

 trouble is suppressed, are preserved identically like to them- 

 selves until some new destructive reaction intervenes. 



When bacteridia develop in a dead medium like culture 

 bouillon, we have no strictly exact means of studying the 

 variations which they undergo. To verify such variations 

 we must use the only precise reagent at our disposition 

 a living animal organism with reference to which we can 

 define a really precise property of anthrax bacteridium ; this 

 property we call the bacteridium's virulence, for that animal. 



Let us take any culture at all of bacteridia in which, with- 

 out our knowing how, variations in different senses have been 

 produced. Inject this culture into a sheep. This is a real 

 Darwinian case, for the conditions in which the bacteridia 

 varied in the previous dead media have no direct relation 

 with the sheep's fitness to survive. Therefore, even if you 

 agree with Lamarck that these variations were directly 

 adaptive to the conditions in which they were produced, they 

 were without relation to the interior medium of the sheep, 

 in which different conditions are realized. 



Here the sheep plays the part of a sieve. Those bacter- 

 idia which, after undergoing the variations, chance to be 



