DISINFECTION STUDIES 193 



intervals, and at each such time the bottles were thoroughly- 

 shaken to secure an even distribution of the bacteria. 



Samples were also taken regularly for the colorimetric deter- 

 mination of pH and for the analysis of the absorbed gases. 



The determination of absorbed gases in the water was carried 

 out as follows: 5 cc. of the sample were transferred quickly to a 

 Van Slyke (1917) gasometric CO2 apparatus. The contained 

 gases were extracted, CO2 determined by means of absorption 

 with KOH, the oxygen determined by means of absorption with 

 alkaline pyrogallol and the volume of the residue noted. Pre- 

 liminary analyses of air and of air-saturated water showed the 

 method to be satisfactory for this purpose. 



By these procedures, a study w-as made of the viabiUty of 

 Bact. tijphosum and Bact. coli at 0°, 10°, 20° and 30°C. in double- 

 distilled water, in tap water and in dilute buffer solutions. 



The experimental results 



In early experiments, the mortality of the organisms was often 

 variable, and could be attributed possibly to the presence of 

 soluble constituents from the glass containers. It was first 

 suspected that some salt effect perhaps might be responsible. 

 The results of the first four experiments show clearly the marked 

 effect upon the mortality which was coincident with apparently 

 small variations in hydrogen ion concentration. 



Behavior in unbuffered solutions. The result of Experiments 

 1 to 4 may be most conveniently considered under one head. 

 In these experiments Bact. typhosum and Bact. coli were each 

 exposed to distilled water and to tap water at 0°, 10°, 20° and 

 30°C. The course of events is graphically presented in the 

 accompanying charts (figs. 1 to 4). 



All the figures are plotted with time as abscissae and the logarithms 

 of survivors as ordinates. The resulting curves show the actual rate 

 of decline in numbers. That is, the slope of a curve (the graphic equiva- 

 lent of the velocity coefficient, k) at any point is indicative of the speed 

 of disinfection and is strictly comparable with the slope of any other 

 curve on the same chart. The experimental points are connected by 

 straight hnes without effort to draw "smoothed" curves through greatly 



