XIII, A, 1 Behrman: Determination of Hardness of Water 25 



least possible that the reason may be found in an acid fermen- 

 tation of the glucose, in which a portion of the potassium pal- 

 mitate is decomposed, with the separation of free palmitic acid. 



It is possible that this objection may be overcome by the 

 choice of & material more suitable than glucose. However, using 

 the pellets made with glucOse, it has been found very satisfactory 

 to standardize the tablets weekly in the central laboratory in 

 Manila and to supply the worker in the field with these data. 

 A typical series of such standardizations is given in Table I. 



Table I. — Reading values of "whole" potassium palTuitate pellets. 

 (Milligrams of calcium carbonate per pellet.) 



Date. Reacting value. 



April 4 1.68 



April 12 1.59 



April 27 1.52 



April 30 1.49 



May 7 1.43 



May 14 1.33 



May 21 . 1.22 



May .28 1.06 



June 5 0.89 



The figures in Table I indicate a gradual and fairly uniform 

 loss of strength in the pellets. In field investigations extending 

 over comparatively short periods of time, reacting values may 

 be obtained without serious error by interpolation. Where an 

 extensive field study is planned, however, the method of periodic 

 standardization in a central laboratory is preferable. 



ACCURACY OF RESULTS AND COMPARISON OF METHODS 



The first method described is essentially a field modification 

 of Blacher's laboratory procedure and, therefore, has as its 

 maximum accuracy that of the latter manipulation. From this 

 must be subtracted the errors accruing from the field technic. 

 Here the possible sources of error are the presence of a relatively 

 large amount of glucose, the lack of uniformity of the pellets, 

 the inaccuracy in reading volumes, and the fact that an excess 

 of reagent is employed to obtain the end point. 



The accuracy of the Blacher laboratory method has been found 

 by a number of workers ^ to be about 2 to 3 per cent. That 

 glucose does not introduce an error, in the quantities used, was 



°Zink, J., and Hollandt, F., Zeitschr. f. angew. Chem. (1914), 27, 439. 

 Nochmann, E., Pharm. Zentralh. (1914), 55, 436-7. Behrman, A. S., loc. 

 cit. 



