226 Theobald Smith 



of the typhoid bacillus. In a brief article on the uses of the 

 fermentation tube published in 1890', I incidentally called at- 

 tention to this difference as a valuable means of diagnosis. 

 The fact, however, remained unnoticed and in 1891 Chante- 

 messe and Widal™ brought forth the same test as new, using 

 lactose in place of glucose in the bouillon. Their method 

 consisted in observing gas bubbles rising and forming a light 

 froth on the surface of the culture fluid in ordinary flasks. 

 This publication induced me to defend my priority in a second 

 article in which I quoted the original announcement of the 

 test". But even this has been largely overlooked by subse- 

 quent writers. 



The publication of Chantemesse and Widal first called gen- 

 eral attention to the gas test as the older differential characters 

 were melting away and something more definite was urgently 

 needed in this very practical field. They were opposed at 

 once by Dubief "' who regarded the differences between these 

 species as merely quantitative. Recently a number of writers 

 (Tavel-", G. W. Fuller"'", W. Dunbar^ Germano and Maurea'\ 

 Ferrati™, and Pane^',) have contributed long articles on this 

 subject and all of them confirm the gas test and give it the 

 most important place among the means of diagnosis between 

 B. typhosus and the colon group. Dunbar in ignorance of my 

 second article * naively recommends the bent tube, closed at 

 one end, as the simplest means of determining gas production. 

 The same thing had been suggested by G. W. Fuller in a 

 prior publication as a substitute for the more expensive fer- 

 mentation tube. Dunbar further recommends simple bouillon 

 {Fletschwasset'), a recommendation likely to lead astray as I 

 have pointed out above. Since gas production in bouillon de- 

 pends solely on the muscle glucose the test would fail when 

 this is absent. The use of lactose, as suggested by Chante- 

 messe and Widal is not so trustworthy as that of glucose, for 

 we have a large group of pathogenic bacilli, the hog cholera 

 group, easily confounded with B. typhosus because neither act 

 on lactose and hence do not coagulate milk. The use of glu- 

 cose bouillon would clear up the difficulty at once. 



* This must liave appeared before the couclusioa of his work for he 

 refers to a publication subsequent to miue. 



