ELIMINATION OF SPURIOUS TESTS FOR B. COLI 339 



isolation of B. coli we appreciate of course that the method is 

 essentially the same as that used by Drigalski and Conradi 

 (1902) in the isolation of B. typhosus, only the aim is slightly 

 different. 



Isolated coliform colonies were stained by Gram's method as 

 modified by Stovall and Nichols (1916) and if typical Gram 

 negative coccobacilli were present, they were tested in lactose 

 broth for gas formation and in gelatin for non-liquefaction. 

 These confirmatory tests were made at 37° C. for forty-eight 

 hours, gelatin being tested for solidification by immersion in 

 ice- water. In case onl}^ non-acidifying colonies appeared, at 

 least two were similarly isolated and if otherwise coliform were 

 tested in lactose broth to detect so-called ''attenuated" B. coli, 

 as required by the 1917 Standard Methods for the Examination 

 of Water and Sewage. 



Since our hypothesis involved the demonstration of B. coli 

 in as many cases as possible where gas was produced in the 

 presence of gentian violet, we adopted a uniform procedure of 

 repeating the above tests with a subculture into a new tube of 

 lactose broth from the original presumptive, if the first plate 

 failed to show typical coliform acidifiers, or if the identification 

 of B. coli failed, in either of the tests of the unheated sample. 



The tabulation of the complete data occupies so much space 

 that it seems best only to summarize the findings for the sake 

 of economy, as in table 2. 



Unquestionably the most interesting feature in table 2 is 

 the behavior of the heated samples. In only one case was a 

 positive presumptive test secured in the broth containing the 

 dye as against 18 positive tests without it; in fact practically 

 all of the tubes with dye appeared to be sterile. Aerobic spores 

 as well as anaerobic spores were inhibited. The single tube 

 mentioned showed slight gas only after four days' inoculation 

 and this was found to be due, neither to B. coli nor necessarily 

 to anaerobes; a lactolytic acid and gas forming sporulating bacil- 

 lus was secured from the plates. This culture in fact was the 

 starting point of our use of gentian violet in the lactose agar 

 plate inasmuch as repeated tests from the broth failed to show 



