42 



glucose in vnrviiig iinunitits, tho composition of this gas varving from 0% CO, to 

 100% CO,. 



Bacill.ii^s coli communis and two otliers, much less common hut also sewage hac- 

 teria, )>roduce 0.4 to 0.7 of a tul)e full of gas, of which 0.5 to 0.7 is COj. .\11 

 others ohserved produce amounts of gas and of COj, readily distinguishable from 

 those, and are, therefore, easily dropped from consideration. 



The process in examining a sample of water is to prepare a sufficient num- 

 ber of tubes of sterile bouillon, and an incubator. The sample of water is dis- 

 tributed among the tubes, an etpial amount in each, the amount varying according 

 to the impurity of the water. With a pure water 1 i-.c. may be used. With .m*w- 

 age 7r^5 c.c. may be found to be too much. The tubes are placed in the incubator 

 and left at the constant temperature of !*8° for thirty-six hours or a little more. 

 They are then taken out and the proportion of gas in each tube is determined. 

 Those promising to contain Bacillus colt communis or its companions are treated 

 with an alkali to absorb the COj, and the propor ion of CO2 is thus determined. 

 From the two determinations the number of Bacillus coli communis in the sample is. 

 derived, and thence the number in a c.c If much more than half the tubes in- 

 oculated from a sample contain " coli" the amount of water used in a tube has 

 been too large to produce a result which will compare closely with other determi- 

 nations. Likewise, if the sample has been diluted with sterilized water before 

 inoculating the tubes, too small a number of tubes with coli shows too great dilu- 

 tion to produce results that will check up with others from the same sample. 



Many results have been obtained by this method in the last three or four 

 years which seem to give closer determinations of the amount of sewage pollution 

 than any method heretofore used. The method has not yet had wide enough ap- 

 plication to demonstrate its value under various conditions, but we feel certain 

 that it has great value in the examination of streams used or proposed as sources 

 of water supply for cities. 



I have not time in the limits of this paper to give the results, and give the 

 methods only, as perhaps the more suitable for the purposes of this association at. 

 this time. 



Psychological Laboratory of Indiana University. By W. L. Bryan. 



