ELIMINATION OF SPURIOUS TESTS FOR B. COLI 331 



it is clear that Jackson did not appreciate the importance of 

 the anaerobes when he said, ''no gas producer or mixture of 

 gas producing bacteria will give results as high as 25 per cent 

 gas except B. coli, even when three days' incubation is em- 

 ployed." Further experiences have shown the necessity of not 

 relying simply upon the presumptive test as a criterion of 

 pollution; Fuller (1915) found a proportion of positive presump- 

 tive tests in which B. coli could be confirmed ranging from only 

 60 to 90 per cent in 1 cc. samples, and Gumming (1916) found a 

 percentage of B. coli to total gas formers ranging from 47.5 to 

 96.3. Moreover the amount of gas produced is no criterion, as 

 is suggested, even in the most recent Standard Methods, since 

 many of the sporulating anaerobes completely fill the closed arm 

 of the fermentation tube with gas. 



A recent experience by the Bureau of Sanitary Engineering of 

 the State Board of Health of Cahfornia (1916) at Sacramento is 

 pertinent, where doses of chlorine so excessive as to cause dis- 

 agreeable tastes and odors were being used by the local authori- 

 ties to destroy all lactose fermenting bacteria in the water in 

 the behef that this was necessary to insure its freedom from ty- 

 phoid bacilli. It was indeed shown that the dosage of chlorine 

 could be safely reduced from 2.6 pounds per 1,000,000 gallons to 

 1.3 pounds in fairly clear water (30 parts per million turbidity— 

 for the dosage has to be varied according to turbidity), elimi- 

 nating B. coli and presumably the typhoid bacillus, if present, 

 without leaving any trace of free chlorine. With water so 

 treated, however, presumptive tests were frequently positive, 

 due to certain sporulating anaerobes from among which Mr. 

 Frank Bachman succeeded in isolating B. welchii, a culture of 

 which he has kindly given us. 



SELECTIVE INHIBITION IN THE PRESUMPTIVE TEST 



It is the purpose of this paper to describe a method to pre- 

 vent gas formation by these anaerobic organisms, so that a 

 greater proportion of positive tests is referable to coliform 

 bacilh alone. This we have proposed to accomphsh through 



