232 L. A. ROGERS, W. M. CLARK AND H. A. LUBS 



have been ascribed to the uncertainties of the fermentation itself 

 rather than to the crudeness of the method. The work of 

 Keyes (1909) and of Keyes and Gillespie (1912) suggested the 

 possibilities in the application of exact methods to the measure- 

 ment of bacterial fermentations and the value of results so ob- 

 tained in taxonomic problems. 



Our own work has included a study of a collection of colon 

 organisms from milk, one from bovine feces and one from grains. 

 In these investigations we have endeavored to use methods which 

 are as exact as the large number of determinations necessary 

 for statistical treatment will reasonably permit. We first applied 

 exact gas analysis. This has now been supplemented by more 

 rational methods of studying acid fermentation and it is hoped 

 that in time improvements in other cultural tests will become 

 available. 



In this paper we give the results of a study of a collection of 

 lactose fermenting bacteria from human feces. This was under- 

 taken primarily to establish the characters of colon cultures 

 from the human intestine. 



The failure in the past to secure satisfactory classifications on 

 the basis of fermentation tests has been largely because the 

 characters selected for primary divisions were not of a funda- 

 mental nature. 



The gaseous fermentation appears to be a basic function of the 

 organisms in question. Even if it is not, it furnishes a basis 

 for primary classification which has established a considerable 

 degree of order, not only among the results of other cultural tests 

 but in the correlation between these tests and the sources of the 

 cultures. We have, therefore, made exact gas determinations 

 on all of the cultures isolated. In addition, we have determined 

 the fermentation of various sugars and alcohols, the formation of 

 indol from tryptophane, the formation of pigment, the lique- 

 faction of gelatin, the carbinol or Voges-Proskauer test and the 

 Clark and Lubs methyl red test. 



The isolation of cultures. The cultures were all secured by 

 plating fresh samples of feces. Twenty-one samples were used. 

 These came from 18 subjects and furnished 176 cultures. The 



