90 H. S. H. WARDLAW. 
by the slow addition of the ammonium molybdate. This. 
effect is ascribed to the occurrence of local excess of the 
reagent when the latter is added quickly. The following 
are the results obtained when two, three, five, and ten 
minutes were the respective times taken to add the pre- 
cipitant to the phosphate solution. The solution was 
thoroughly shaken while the precipitant was being added. 
Table VIT. 
Influence of rate of addition of precipitant on error. 
: : Precipitant Error, 
Estimation. aed P,O, found. | P.O, present. per cent. 
to 
39 10 min. 22°8 mg. +1°2 
40 a TOG 5°38 
4] De 23°83 3 22°5 mg. 58 
42 Os ee DOO i 3°0 
poured in DOs i 5:0 
As long as the time taken to add the precipitant remains. 
less than five minutes, therefore, no certain diminution of 
the positive error is observed. When the precipitant is 
added, in ten minutes, there isa considerable diminution of 
the error, but as the solution has by this time cooled down 
about 20° O., this diminution is due to incomplete precipita- 
tion and not to any decrease in the excess of molybdenum 
due to the slow addition of the ammonium molybdate (vide 
infra). 
3. Length of time between filtration and precipitation. 
—Baxter and Griffin concluded from their experiments 
that the excess of molybdenum in the precipitate of am- 
monium phosphomolybdate obtained by them was due to 
the occlusion of a mixture of ammonium molybdate and 
molybdic acid, and showed that this occlusion apparently 
took place in two stages, (1) during the precipitation, (2) 
while the precipitate, already formed, lay in contact with 
the mother liquor. They therefore advise that the precipi- 
tate be filtered off as soon as is compatible with complete 
