280 PROCEEDINGS OF THE AMERICAN ACADEMY. 



It is certainly difficult to decide between these two formulas, only 

 it must be observed that the higher formula is derived from the mean 

 of four analyses, which agree well with each other, and has therefore 

 the weight of analytical evidence in its favor until further research 

 shall prove its inaccuracy. Still another formula has been proposed 

 by Von Knorre,* who writes 12 WO3 . 3 (NH4)2 . 2 Na.O + 13 aq., 

 which requires 84.41% WO3, 4.73% (NH4)A 3.76% NaA and 7.10% 

 HgO. The correspondence between the results of the analyses and 

 the data required by the formula is much less than with the formula 

 which I first gave and which Von Knorre rejects, and of course still 

 less than with the new formula. 



New Phospho- Tungstate. 



Pure normal sodic tungstate was mixed in solution with sodic ortho- 

 phosphate, P04Na2H + 12 aq., in the proportion of twelve molecules 

 of the first to one of the second salt, and chlorhydric acid added in 

 small excess. A phospho-tungstate crystallized from the solution, 

 and was redissolved and twice recrystallized. The new salt was in 

 fine colorless crystals, less soluble than the now well known salt 

 which has the formula, 24 WO3 . P0O5 . 2 NajO + 27 aq. The salt 

 was dried on paper for analysis. It effloresced or became opaque in 

 dry air. Mr. Charles D. Howard obtained the following results on 

 analysis. 



1.4941 grams lost on ignition with WOiNa. 0.1078 gram = 7.21% \\0. 

 1.4502 grams lost on ignition with WO^Nao 0.1046 gram = 7.21% HgO. 

 1.0950 grams gave 1.0019 grams WO3 + PA = 91.49%. 

 1.1950 grams gave 1.0975 grams WO, + PoOj = 91.58%. 

 1.6782 grams gave 0.0720 gram WO, + PA = 2.74% PA5. 



The analyses lead to the formula, 



20 WO3 . PA • NasO . 2 II.O + 19 aq., 

 which requires: 



* Berichte der deutschen chem. Gesellschaft, XIX, 823. 



