94 



expressed as a percentage. Lastly, the value — — is 



calculated. 



As a test of the truth of the calculations, in the right 

 half of the table they are set up in another way, viz, by 

 estimating the frequency of occurrence of previous influenza 

 among carriers and non-carriers of Pfeiffer's bacillus. It will 

 be seen that the two methods of calculation give identical 

 results. 



It appears that Pfeiffer's bacillus in all instances was of 

 more frequent incidence in previous cases of influenza than in 

 other subjects; the divergences are however so slight that they 

 may very well be casual. The present material may perhaps 

 be said to indicate a connection between previous cases of 

 influenza and the occurrence of Pfeiffer's bacillus, it is however 

 far from being sufficiently large to evidence such a connection. 



In the inoculations of July, Sept. and Oct. 1920, no que- 

 stions were asked about influenza because, from the experience 



