QUALITATIVE BACTERIOLOGICAL ANALYSIS. 173 



on July 25, and again on October 2, gave the following 

 results : 



These experiments are of great interest, as the dog was 

 inoculated with B. typhosus, and the results could not be in- 

 fluenced by a secondary Coli infection. When the serum of a 

 typhoid patient is tested with B. typhosus and B. coli, it is 

 sometimes found that the serum reacts in a higher dilution with 

 the B. coli than with the B. typhosus. Biberstein found that 

 in five cases of enteric fever which he investigated, the serum 

 reacted in higher dilutions with the B. coli than the B. typhosus. 

 He considered that the patients were suffering from a secondary 

 Coli infection. Stern and Widal have recorded similar results. 

 Consequently when the agglutination test is used to distinguish 

 B. coli from B. typhosus, it appears absolutely necessary that 

 the serum of an immunised animal should be used. If the 

 serum of a patient is employed, there is always the possibility 

 that he may be suffering from a mixed infection, and conse- 

 quently the serum may react to both B. coli and B. typhosus. 



Working with sera derived from immunised horses, I have 

 lately tested the reaction to the specific agglutinins of 150 

 cultures of B. coli derived partly from healthy stools, and partly 

 from the stools of typhoid patients. The results obtained with 

 the cultures derived from healthy stools are shown in the 

 following table : 



