138 SUPPOSED INSTANCES [CH. ix 



the tubes and both pairs of plates remained sterile. In the 

 other four tubes growth took place with gas formation and 

 the four pairs of plates all showed colonies. Broth cultures 

 were made from these colonies and yielded strains which gave 

 the sugar reactions of B. proteus. 



The writer failed to demonstrate the presence in any case of 

 organisms of the paratyphoid group. The experiment however 

 was successful in demonstrating that bacteria which failed to 

 give evidence of their presence in the faeces of a healthy 

 guineapig might make their appearance in the faeces of the 

 same animal after it had been given for a few days unwhole- 

 some and irritating food. 



This conclusion lends weight to the suggestion already 

 made that the results obtained by Schmitt, and also by 

 Miihlens, Dahm and Fiirst, might possibly be explained by the 

 presence of a secondary invader. 



Their experiments are, in both instances, open to one 

 further criticism. The identification of the paratyphoid or- 

 ganisms was made to depend solely upon their agglutination 

 reactions. If it is admitted that the power to form and to 

 absorb specific agglutinins on the part of an organism is 

 subject to variation it must be recognised that such tests 

 alone are insufficient to establish the identity of the organism. 

 In other words, it is within the bounds of possibility that only 

 one type of organism was actually present, but that its agglu- 

 tination properties varied. 



Such a contingency would be likely to arise in the case of 

 two organisms so closely allied as B. Aertryck and B. Gaertner. 

 An elaborate investigation into the agglutination properties 

 of these two organisms was conducted by Sobernheim and 

 Seligmann (1910). Pure colonies of numerous strains were 

 secured by the Indian ink method. They found that colonies 

 derived from the same strain and growing side by side differed 

 in their agglutination reactions. The same strain differed at 

 different times. The agglutination reactions, in some instances, 

 became altered after passage through the mouse and after a 

 culture had been heated. In some instances the injection of 

 living bacilli yielded a serum which was much more variable 



