312 WATER HYGIENE 
river that lactose broth was the better medium. In 1851 parallel sets of 
lactose broth and lactose bile tubes, lactose broth gave higher results, 
confirmed as well as presumptive, than lactose bile. It was granted, 
however, by Cummings that, in the lower reaches of the river, remote 
from pollution, the bile tubes were more reliable. The opinion of other 
bacteriologists has been that for highly polluted waters the lactose 
broth was the better medium, and for slightly polluted waters lactose 
bile gives the better results. This is attributed to the presence of spore- 
forming lactose fermenting anaerobic bacteria which are not manifest 
when the number of B. coli are large. Obst (1916) also found that the 
bile inhibited about one-half of the typical B. coli. The general use, 
then of lactose broth in place of lactose bile seems to be Justified. 
Hauser (1917) has described the method used for isolating and iden- 
tifying B. colt at the Cincinnati filtration plant. This procedure is 
interesting and valuable because it is representative of the practice in 
one of the largest water-treatment plants. 
The B. coli group may be defined as aerobic non-spore forming Gram negative 
bacilli, fermenting lactose with production of gas, and not liquefying gelatin. 
At the Cincinnati filtration plant the attempt is made to carry out the examina- 
tions so that the above definition is fulfilled as completely as possible, consistent 
with the speed and facility necessitated by routine water analysis. 
The particular water samples under examination are planted into lactose 
broth. Lactose broth is used rather than lactose bile, which is the medium 
recommended in the 1912 edition of the Standard Methods of Water Analyses of 
the A. P.H.A., for the reason that, not only at this laboratory, but at the labora- 
tory of the Ohio River Investigation of the United States Public Health Service 
and other laboratories, it has given consistently a higher percentage of con- 
firmed end-results from original inoculations. In planting the water sample 
decimal dilutions are used to facilitate computations of the B. coli index. In 
routine examinations, at least three dilutions are used in order to obtain, if pos- 
sible, a negative and a positive test. The formstion of gas in any amount after 
forty-eight hours’ incubation at 37° C. is recorded as positive and considered as 
presumptive positive. The smallest quantity of water giving gas is then 
confirmed. 
A loopful of culture from the tube furnishing the presumptive test is stroked 
upon Endo’s medium previously poured in plates. The plates are incubated 
at 37° C. for twenty-four hours. If no aerobic colonies (sterile piate) develop, 
the respective gas-formation is considered due to anaerobes, and B. coliis recorded 
as not present. If aerobic colonies develop, the most typical of B. colt is trans- 
planted into lactose broth and gelatin and a microscopical examination made 
directly. As there are several varieties of typical B. coli colonies, experience 
alone will determine the choice. 
If in the lactose broth transplant, no gas is formed after the forty-eight 
