go Elements of Water Bacteriology. 



suggested in the classic researches of Theobald Smith 

 (Smith, 1892), this method may be made approximately 

 quantitative by the inoculation of a series of tubes with 

 measured portions of the water. If, for example, of ten 

 tubes inoculated each with x^jy of a cubic centimeter, 

 four show B. coli, we may assume that some 40 of these 

 organisms were present to the cubic centipieter. Irons 

 (Irons, 1901), in a comparative study of various methods 

 jFor the isolation of B. coli, showed that the preliminary 

 enrichment frequently gave positive results when the 

 results of the direct use of the agar plate were negative, 

 and concluded that " where the amount of B. coli is 

 small and the colony count large, the lactose plate for 

 .plating water direct is inferior to the dextrose fermen- 

 tation-tube." Gage obtained similar results (Gage, 1902). 



The medium ordinarily used for the preliminary enrich- 

 ment is ordinary broth to which i.o per cent of dextrose 

 has been added, and the reaction brought to the neutral 

 point. Into each of a number of fermentation-tubes of 

 this medium a measured quantity of the water to be 

 examined is inoculated, and the culture is incubated for 

 twenty-four hours at 37.5° C. At the end of this time the 

 tubes are examined for gas formation. If gas is found, a 

 small amount of the culture should be added, after suit- 

 able dilution, to litmus lactose agar and plated. 



With polluted waters it will be found advantageous to 

 plate out on the first appearance of gas (4-8 hours). It 

 has been shown by one of us (Prescott, 1902'') that a very 



