64 ELEMENTS OF WATER BACTERIOLOGY. 



meter is inoculated into a dextrose tube. If this tube 

 shows gas formation after twenty-four hours at 37 , a lit- 

 mus-lactose-agar plate is made and the other diagnostic 

 tests applied. 



Our experience has shown that it is not specially advan- 

 tageous to apply the colon test in such large samples, 

 since the significance of B. coli when present in numbers 

 less than 1 per c.c. is extremely doubtful. On the other 

 hand the danger of overgrowth is greatly increased by 

 this process and negative results may often be obtained 

 when the organisms are really present. Hunnewell and 

 one of ourselves (Winslow and Hunnewell, ioo2 b ) found 

 that of 48 samples of certain polluted river waters 18 gave 

 B. coli when 1 c.c. was inoculated directly into dextrose 

 broth, while in only 4 cases was a positive result obtained 

 after preliminary treatment of 100 c.c. in carbol broth. 

 In 153 samples from presumably unpolluted water B. coli 

 was found 5 times in 1 c.c. and n times by the examination 

 of the larger sample. The authors, therefore, concluded 

 as follows: 



"It appears evident that the use of large samples in 

 applying the colon test to the sanitary analysis of drink- 

 ing-water is not advantageous. In comparing the results 

 of the tests in 1 c.c. and in 100 c.c, it will be noted that the 

 proportion of lactose fermenting organisms and of colon 

 bacilli in the unpolluted waters was more than doubled 

 in the latter; thus waters of good quality are more likely 

 to be condemned by the use of large samples. On the 

 other hand, in the polluted waters a considerable propor- 



