ACCURACY OF NEUMANN’S METHOD. E76 
errors to which the method may be subject will be of value. 
The present communication is an account of the errors I 
have met with in the use of this method, the effect of some 
conditions on these errors, and an attempt to locate them. 
Details of Method. 
Neumann’s method was first used in the present work for 
the analysis of milk; the modification of the acid-ashing 
process due to Plimmer and Bayliss was used and carbon 
dioxide was eliminated before the final titration. In the 
second series of experiments the estimations were per- 
formed on standard phosphate solutions. Here, of course, 
the acid-ashing process is unnecessary, but when the 
requisite amount of sulphuric acid is added to the solution 
a product is obtained closely resembling the acid-ash of an 
organic substance, and the subsequent treatment of the 
two series of estimations is the same in each case. The 
following is a summary of a typical estimation on milk:— 
Ten cc. of milk were placed in a 500 cc. round bottomed 
long necked Jena glass fiask, 10 cc. of concentrated sul- 
phuric acid and 10 cc. of concentrated nitric acid were 
added and the whole was gently heated until the nitrogen 
peroxide fumes thinned off; the mixture was then cooled, 
another 10 ce. of nitric acid were added, and the heating 
resumed till the fumes thinned off again. After four similar 
additions of nitric acid, the heating was continued until 
this had all been driven off; the mixture remained clear and 
became almost colourless after ten minutes’ further heat- 
ing, indicating completion of the oxidation. The time 
required was about four hours. After cooling, the acid-ash 
was diluted with about 30 cc. of water and boiled for five 
minutes (to get rid of the nitric oxide formed by the decom- 
position of the nitrosyl-sulphuric acid produced), then made 
up to 100 cc. with water, 35 cc. of 50% ammonium nitrate 
solution were added, the liquid was brought just to the 
