150 Elements of Water Bacteriology. 



It is certain that the bile medium sometimes shows gas 

 when typical B. coli are absent; it is certain that it some r 

 times gives entirely negative results when B. coli may be 

 isolated in other ways. Neither of these errors is com- 

 monly of great magnitude, however, and in general the 

 lactose bile results correspond fairly well with those 

 obtained by the complete isolation of B. coli. 



No merely presumptive test of this sort should be sub- 

 stituted for the complete demonstration of B. coli in the 

 detailed sanitary study of a special source of supply. For 

 extensive routine surveys of considerable series of samp- 

 ling stations, on the other hand, the bile test offers a satis- 

 factory approximation to the truth. 



The litmus-lactose-agar plate furnishes a presumptive 

 test of considerable value as indicated in Chapter IV, 

 although it is probably less delicate than the fermentation 

 methods. With polluted waters, however, comparative 

 studies of the agar plate and the bile method are much to 

 be desired. 



Other special media have been suggested for rapid 

 routine water analysis of which those containing " neutral 

 red," one of the safranine dyes, have been most fully studied. 

 Rothberger (Rothberger, 1898) first pointed out that B. 

 coli reduces solutions of this substance, the color chang- 

 ing to canary-yellow accompanied by green fluorescence. 

 Makgill (Makgill, 1901), Savage (Savage, 1901), and 

 other English observers, as well as Braun (1906), in 

 France, report favorable results from the use of this test, 



