RESISTANCE OF BACTERIA 219 



for laboratory animals is higher. Gratia^*' has noted the same thing. 

 Davison^^' has shown that a resistant Shiga dysentery bacillus kills 

 rabbits in a quantity one-fifth of that required with susceptible bacteria 

 of the same strain. Insofar as B. pestis is concerned, working with 

 a slightly virulent strain recently isolated from a bubo in a benign case, 

 I have shown that while with the susceptible strain it was necessary to 

 inject 0.1 cc. of a 24-hour bouillon culture in order to kill a guinea-pig 

 of 400 grams in 52 to 58 hours, a dose of 0.0002 cc. of tiacilli of the same 

 origin, but rendered experimentally refractory to the action of a very 

 virulent bacteriophage, regularly caused death in 46 to 50 hours. 



It is, however, quite important to bear in mind that all of these new 

 characters associated with the acquisition of resistance to the bacterio- 

 phage by a bacterium are not uniformly distributed among the different 

 colonies derived from a single secondary culture. This is particularly 

 true as regards the loss of agglutinability and the increase in virulence. 

 For example, with B. pestis (taken as an illustration because of its high 

 virulence for laboratory animals), associated with colonies of increased 

 virulence I have found others completely avirulent. We are not dealing, 

 therefore, with an absolute rule. What actually happens is that through 

 the influence of the resistance to the bacteriophage sudden mutations 

 occur. These are almost always of a temporary nature, disappearing 

 when the resistance is lost. But sometimes they are irreversible, as 

 we will see. These mutations do not involve all of the characters of 

 the bacterium, but only a certain number of them, varying from one 

 bacterium to another, even in a single culture. With one bacterium the 

 character "agglutinability" will be modified, with another, the character 

 "virulence," and with a third, both of these will be changed at the same 

 time. All that may be predicted is that, usually, there is a reduction in 

 agglutinability and an increase in virulence. 



The changes in the bacterium endowed with an acquired resistance 

 extend to its morphology also. The bacilli frequently assume a cocco- 

 bacillary form; they may even take the coccus form. The loss of resist- 

 ance is accompanied by a return to normal form. I have obtained this 

 reversion on several occasions, with resistant colonies constituted 

 solely in the beginning of cocci, as is shown by the following experiment: 



A Petri dish is heavily seeded from an agar culture of B. dysenteriae 

 and is placed in the incubator at 37°C. for about four hours. A drop of 

 the bacteriophage filtrate is then placed in the centre of the plate. The 

 race of bacteriophage should be one of average activity, that is, one 

 capable of regularly causing complete dissolution of a bacterial suspen- 



