32 The Philippine Journal of Science im 



of chlorine consumption at boiling temperature, with carefully- 

 regulated constant heating, for fifteen-minute digestion periods, 

 with lime water added to the water under examination. The 

 authors obtained very concordant results by this process. The 

 figures given in Table V were obtained without addition of lime 

 water, but show that the method is accurate enough for labo- 

 ratory use. However, this method is open to objection, because 

 it does not take into account conversion of hypochlorite to chlo- 

 rate at boiling temperature. 



Digestion in the dark at room temperature is so easily carried 

 out, and gives such uniform results if care is used in keeping 

 conditions constant, that it is to be preferred to the methods 

 employing heat. 



Whether chloride of lime shows fluctuations in bactericidal 

 effect similar to those noted for its chemical decomposition 

 (Table VI, fig. 4) is not clear. The results of Glaser(5) indicate 

 that it does not, but these observations were made with such high 

 concentrations of chloride of lime that they are not conclusive. 



It is true that the data under discussion were obtained in ex- 

 periments on sewage, but apparently similar fluctuations may 

 occur with ordinary water as well and with small hypochlorite 

 concentration. Thus Stokes and Hachtel(22) in their work on 

 the treatment of Baltimore spring water by calcium hypochlorite 

 report that — 



when 1.5 parts of available chlorine per million parts of water were used 

 there is practically no residual chlorine in the water. In one case when 

 1.75 parts were used there was a large amount of residual chlorine, giving 

 an average .... of .574. On another occasion, however, when this amount 

 was used the residual chlorine was less, giving an average of 0.24 per 

 million parts of water. When 2.0 parts of available chlorine were used 

 there was an average ... of 0.206, and when 2.5 parts of available chlorine 

 were used for treatment there was an average of 0.62 parts per million 

 parts of water. These results, therefore, are somewhat variable, and it 

 is hard to explain the greater amount of residual chlorine when 1.75 parts 

 were used than when 2.0 parts were used. 



From the evidence at hand it is clear that the ordinary pro- 

 cedures for determining the amount of interaction between hy- 

 pochlorites and waters and sewage are influenced by so many 

 factors that, unless very carefully interpreted, they are very 

 likely to be misleading. If it were only the available chlorine 

 that had germicidal action, the analytical control of hypochlorite 

 sterilization would be relatively simple. Even in this case, how- 

 ever, the quality of a water would be an important consideration, 

 since waters differ widely in their ability to liberate chlorine 

 from hypochlorites. (21) Thus two different waters brought to 



