ELIMINATION OF SPURIOUS TESTS FOR B. COLI 349 



in the standard broth, but totally negative in the presence of 

 gentian violet. According to the Bureau of Sanitary Engineer- 

 ing B. coll was originally isolated in 20 cc. only and was there- 

 fore considered to have died out from the sample. 



Again we can call attention to the disproportion between 1st 

 and 2nd successful isolations of B. coli favoring the use of gentian 

 violet in the presumptive test, although in this series of tests the 

 total number of samples yielding B. coli in the dye test was 

 exceeded by one in the standard test, due to the fortuitous cir- 

 cumstances above mentioned. 



TIME OF FIRST APPEARANCE OF GAS 



Standard methods permit a negative report on presumptive 

 tests failing to show gas in forty-eight hours. In the beginning 

 of our work we adopted a somewhat longer period, namely five 

 days, in order to isolate as large a proportion as possible of colon 

 bacilli present. It is interesting to note that in no case was 

 B. coli isolated from a presumptive test first showing gas after 

 four days. With gentian violet 77 per cent of the tests from 

 which B. coli was isolated showed gas within forty-eight hours, 

 without gentian violet, 88 per cent. The figures are not quite 

 representative, however, as we find on analysis that the presence 

 of gentian violet in the presumptive test does not generally delay 

 appearance of gas. The somewhat unfavorable percentage is 

 due almost entirely to a small group of samples tabulated in 

 table 5. These were originally very badly contaminated, but 

 had almost eliminated B. coli on standing; if table 5 were ex- 

 cluded the result would stand 88 per cent positive within forty- 

 eight hours with gentian violet, to 87 per cent without. The 

 data for all the tests are tabulated in table 6. The main point 

 seems to be that we are justified in waiting at least four days 

 before calling the presumptive test negative. 



The positive tests from which it was impossible for us to 

 isolate B. coli are also tabulated, in table 7. 



The figures include 1st trials many of which failed to yield 

 B. coli but from the subcultures of which in a second presump- 



