ACCURACY OF NEUMANN’S METHOD. 83 
the present stage of the work we have no grounds for | 
stating whether the composition of the precipitate is that 
assumed by Neumann or not (see later). 
2. (a) The sufficiency of the washing was proved by the 
neutrality to litmus of the wash-water. (b) By evaporating 
down the washings extremely slight amounts of the pre- 
cipitate were found to have dissolved, but the quantity 
was too smail to introduce any appreciable error. (c) It 
has been shown by Hundeshagen, and confirmed by Neu- 
mann, that large amounts of water cause the two loosely 
ound molecules of nitric acid to split off from the molecule 
of ammonium phosphomolybdate, a fact that was also 
noticed in the present work. This, however, is indicated 
by the washings becoming acid again, and washing was 
always stopped as soon as the washings became neutral. 
3. (a) The freedom of the alkaline solution from ammonia 
was judged by testing the issuing steam with litmus paper. 
It was found necessary to boil the solution for considerably 
over an hour, instead of for twenty minutes, as stated by 
Neumann, before the issuing steam no longer reacted with 
litmus paper. The ammonia is therefore not readily got 
rid of. Bang considers this stage to be the weak point of 
the whole method, and recommends the removal of the 
ammonia as hexamethylene tetramine by the addition of 
formaldehyde. (b) We are not ina position to say whether 
this boiling, latterly with rather concentrated caustic soda, 
may not modify the substances present in some Way. 
4. (a) The acid solutions were vigorously boiled for ten 
minutes, so that little carbon dioxide could have remained 
in solution, and any error arising from this source must be 
very small. (b) With regard to any possible effect on the 
nature of the substances present we can again say nothing, 
5. (a) The end-point of the titration left little to be 
desired, being determinable to one drop without difficulty. 
