THE COLON GROUP OF BACILLI 1396 



mentation with gas-formation had taken place in only 

 one or two of the three varieties of media, cross-inocu- 

 lation was made into the three media again and into 

 dextrose broth. Streaks were also made on litmus 

 lactose agar. Incubation was again maintained for 

 24 hours. It was found that organisms which would 

 ferment all four media after cross-inoculation always 

 proved to be B. coli, and only rarely was there any 

 failure to confirm B. coli when aU the original three 

 media gave positive presumptive tests. These failures 

 were ascribed to anaerobic organisms, but no isolations 

 were made. They were so rare, however, as to be 

 neghgible, and Yoimg recommends the use of but 

 three media, discarding the dextrose broth. 



The method of cross-inoculation as above outlined 

 appears to have several advantages with western 

 waters at least, and probably with all waters, in that 

 positive presmnptive tests in all three media either 

 directly or on cross-inoculation make it unnecessary 

 to apply confirmatory tests. Furthermore, , the results 

 are obtained in 48 hours in case all media respond 

 positively, while if only a portion respond, the con- 

 firmatory test is already started by the streak culture 

 on the lactose plate. Experience seems to show that 

 this method gives more reliable results than the use 

 of a single presumptive medium and that the inhibitive 

 efifect of the bile and the overgrowths in lactose broth 

 are counterbalanced. 



The table on page 139c shows the comparative results 

 of 140 analyses of samples that showed fermentation in 

 less than 50 per cent of the first fermentation tubes. 



