BACTERIA IN WATER 697 



Rosenau states that a ground water should be condemned even if 

 only a few colon bacilli are found, for, as he puts it, "these bacteria have 

 no business in a soil-filtered and properly protected well or spring- 

 water." Surface waters, however, may easily contain a few colon bacilli 

 without necessarily having been exposed to contamination by human 

 forces. The limit of safety, Rosenau states, is one colon bacillus per c.c. 

 If more are present the water should be regarded as suspicious. If more 

 than 10 per c.c. are found the water must be regarded as dangerous 

 and unqualifiedly condemned. 



Presumptive Colon Tests. For this purpose, a large number of 

 methods have been devised. In examining sewage or other polluted 

 waters in which the number of colon bacilli is comparatively large, the 

 direct use of lactose litmus agar plates yields excellent results. The 

 method advised by the American Public Health Association is as follows: 



"Add the quantities of water or sewage to be tested to fermentation tubes 

 holding at least 40 cubic centimeters of lactose bile, 1 incubate at 37 C. and 

 note the production of gas. The standard time for observing gas production 

 is 48 hours. Small numbers of somewhat attenuated B. coli may require three 

 days to form gas. Attenuated B. coli does not represent recent contamination 

 and all B. coli not attenuated grows readily in lactose bile. No other organism 

 except B. Welchii gives such a test in lactose bile. B. Welchii is of rather rare 

 occurrence in water, is of fecal origin, is almost invariably accompanied by B. 

 coli, and the sanitary significance is the same. 



"A comparison of the positive results obtained in the various dilutions of 

 the water or sewage planted into the lactose bile gives a good idea of the rela- 

 tive amount of contamination in the various samples examined. 



" Quantities of Water Tested. For ordinary waters, 0.1, 1.0 and 10.0 c.c. shall be 

 used for the colon test. For sewage and highly polluted surface waters, smaller 

 quantities shall be used; and for ground waters, filtered waters, etc., the quan- 

 tities shall be larger, if necessary to obtain positive results. The quantities 

 shall vary preferably in the tenfold manner indicated. Single tests with quan- 

 tities which give ordinarily a positive result or ordinarily a negative result are 

 in themselves of but little account for quantitative determinations. The range 

 in quantities studied shall be sufficient to allow the quantities needed for both 

 a positive and a negative result to be recorded for each sample. When this is 

 done, the results of several tests allow an approximate estimate of the number 

 of B. coli per c.c." 



The identification of colon bacilli so obtained should then be under- 

 taken. The following table, again taken from the report of the A. P. H. 

 A., will be of help: 



1 Prescott, Science, xvi, 1902. 



