BACTERIOLOGY OF WATER 151 



square of the radius of the plate by 3.1416. Then multiply this number by the 

 average per square centimeter, and we have the total number of colonies on the 

 plate. This is the principle of the Jeffers disc. 



The relative proportion between the bacterial count at 2oC. and that at 38C. 

 is of great importance from a qualitative standpoint, as will be seen later. 



2. Deliver into a series of Durham fermentation tubes containing glucose bouillon 

 and into another series containing lactose bouillon varying definite amounts of the 

 water to be examined. In tubes showing the presence of gas in both glucose and 

 lactose bouillon the evidence is presumptive that the colon bacillus is present. For 

 the positive demonstration plates must be made from such tubes as show gas. 



It is sufficient to deliver from graduated pipettes in each series quantities of 

 water varying in amount from o.i c.c. to 10 c.c. In our laboratory we inoculate 

 with o.i c.c., 0.2 c.c., 0.5 c.c., i c.c. and 10 c.c. of the suspected water. If the o.i c.c. 

 tubes show gas, we have reason to assume that the water contained at least 10 colon 

 bacilli per c.c. If only the 10 c.c. tubes showed gas those with less amounts not 

 having gas we would be in a position to state that the water contained the colon 

 bacillus in quantities of 10 c.c., but not in quantities of i c.c. or less. Many authori- 

 ties regard water as suspicious only when the colon bacillus is present in quantities 

 of 10 c.c. or less; waters of good quality frequently showing the presence of the colon 

 bacillus in quantities of 100 to 500 c.c. 



It is generally accepted that if a water shows the presence of the 

 colon bacillus in quantities of i c.c. or less, it should be regarded as 

 suspicious. 



At the present time the medium that gives the least source of error in carrying 

 out the quantitative presumptive tests is the lactose bile. It is made by adding 

 i % of the lactose and i % of peptone to ox bile, and fermentation tubes of the media 

 showing gas may be considered as very probably containing the colon bacillus. The 

 percentage of error with this method is reported to be only 11%, while with glucose 

 fermentation tubes the error is more than 50%. Gas formation is usually shown in 

 forty-eight hours, but it is advisable to continue the incubation for seventy-two hours. 

 These presumptive tests are chiefly of value in highly contaminated waters. Even 

 with this method plates should be made. 



3. As the colon and sewage streptococci ferment lactose with the production of acid 

 and hence produce pink colonies on lactose litmus agar, much information can be 

 obtained from the proportion existing between the number of pink colonies and 

 those not having such a color. Waters of fair degree of purity rarely give any pink 

 colonies. 



Qualitative Bacteriological Examination 



General Considerations. In some countries the proportion of 

 liquefying to nonliquefying colonies on gelatin plates is considered of 

 importance. Certain sewage organisms belonging to the proteus and 

 cloaca groups liquefy gelatin; consequently, if the proportion of liquefy- 

 ing to nonliquefying be greater than as i to 10, the water is considered 



