38 The Philippine Journal of Science i9i6 



known content of organic matter, in some cases (benzoic acid) 

 there was little or (for urea) no action whatever. Cavel,* work- 

 ing with known concentrations of various organic substances, 

 including alcohols, aldehydes, acids, ketones, amides, phenols, 

 sugars, indole, urea, skatole, etc., came to the conclusion that 

 the determination as applied to water gave no indication of 

 the nature or quantity of organic matter present. 



The determination of oxygen consumed not only fails to 

 measure accurately the amount or kind of organic matter pres- 

 ent, but it is also greatly affected by the presence of different 

 foreign substances. Unoxidized mineral substances, such as 

 ferrous salts, sulphides, and nitrites,^ introduce errors which 

 it is generally sought to overcome by digesting the sample for 

 three minutes in the cold with sulphuric acid and direct titration 

 with permanganate. M. Dittrich ^° recommends that any sul- 

 phur in water should be precipitated before the regular proce- 

 dure for oxygen consumption is begun. 



As early as 1886, Dupre ^^ called attention to the fact that 

 a high content of soluble chlorides greatly affected the oxygen- 

 consuming capacity when the digestion was made at boiling tem- 

 perature, owing to the decomposition of the permanganate in 

 the presence of sulphuric acid and chlorides, chlorine being set 

 free. He found that as little as 15 parts per million of chlorine 

 (1.05 grains per gallon) would appreciably affect the results. 

 It is probaly due to his work that, in the English standard 

 methods,^^ digestion is carried on at 27° C. (80° F.), a cor- 

 respondingly longer digestion period being used than when 

 digestion is carried on at boiling temperature. 



In short, the determination of the oxygen-consuming capacity 

 gives comparable results only when a fixed method is strictly 

 adhered to, and even in this case certain errors may creep in 

 which go far toward destroying the value of the analysis. 



We have attempted to study quantitatively some of the factors 

 which influence the determinations. All the data recorded are 

 the average of two or more concordant determinations. Dup- 

 licate titrations which did not check each other within 0.5 cubic 

 centimeter were rejected. 



^Rev. gen. Chem. (1912), 15, 73. 



" Standard Methods of Water Analysis, 2d ed. Am. Pub. Health Assoc, 

 New York (1912) 28. 



"Zeitschr. f. analyt. Chem. (1912), 50, 697. 



^Analyst (1886), 10, 188. 



" Cf . Sutton, Volumetric Analysis, 10th ed. P. Blakiston's Son & Co., 

 Philadelphia (1911), 485. 



