DISINFECTION STUDIES 197 



plicate bottles giving parallel curves show but insignificant devia- 

 tions in pH from each other. On the other hand, a wide diver- 

 gence of the curves is coincident with appreciable fluctuations 

 in pH. 



These findings apply alike to double-distilled water and to 

 Washington tap water. In the case of the distilled water, which 

 contained a minimum of dissolved salts, the possible contami- 

 nants are the atmospheric gases of the laboratory which include, 

 among others, ammonia and sulfur dioxide. The tap water 

 contained in addition to the gases, about 163 parts per million 

 of dissolved mineral matter. Both waters are poorly buffered, 

 with their pH equiUbria situated on the steep portion of a curve, 

 so that a trace of strong base or acid will cause a relatively large 

 change in the hydrogen ion concentration. This circumstance 

 makes it difficult to estimate the pH accurately, for the hydro- 

 gen electrode cannot be used, and the addition of free indicators, 

 which are themselves strong acids, may x-itiate the results. We 

 have, however, made a colorhnetric estimation of the pH by 

 utihzing the sodium salts of the Clark and Lubs indicators and 

 comparing the results with those obtained with the free acids. 



These pH deviations fall within the physiological range of 

 our organisms, as will be sho\vn later, and it could scarcely be 

 urged, therefore, that these insignificant shifts were the cause 

 of the large variations in mortality. It may however be that the 

 slight pH changes observed were indicative of large and more 

 profound local fluctuations in the unbuffered waters surrounding 

 the bacteria ; but it is worthless to pursue these thoughts further 

 for they are purely speculative, and we are ignorant of all the 

 forces operating in the present case. Yet the obvious inference 

 occurs, that the control of the pH with buffers might stabiUze a 

 possible factor in the problem and lead to more consistent results. 

 Later experiments are concerned with this aspect of the problem. 



Our observations suggest that unbuffered solutions may not 

 be depended upon to retain a constant and uniform hydrogen 

 ion concentration for long, so that we must conclude that results 

 of like experiments in the past may have been vitiated by this 

 factor. Falk (1920) mentions the importance of pH in the study 

 of salt effects upon the viability of bacteria. 



