110 PROCEEDINGS OP THE AMERICAN ACADEMY 



and very small differences — often not exceeding 0.2% — are some- 

 times sufficient to change the ratio between the number of atoms of 

 phosphoric and tungstic oxides. In the case of tungstic oxide the 

 divisor which we must employ (WO3 = 232) is so large that a very 

 sensible variation in the percentage of the oxide does not sensibly 

 affect the quotient. It is consequently sometimes difficult to decide 

 between formulas in which, for example, the ratios of the two oxides 

 are as 24 : 1, as 22 : 1, or as 20 : 1. 



3. For similar reasons it is very difficult to recognize mixtures of 

 different salts. 



4. The alkaline phospho-tungstates are usually efflorescent, — fi'e- 

 quently to a remarkable degree. 



5. The salts of the different series agree so closely in physical and 

 chemical properties that, as in the case of the tungstates, distinctive 

 tests are not to be found. 



6. It is almost impossible to predict what compound will be formed 

 when tungstates and phosphates or phosphoric acid are mixed, even 

 when the mixture is made with the greatest care and in perfectly 

 definite proportions. Very small variations in the conditions of the 

 process materially affect the results. 



7. A large proportion of the salts of this series crystallize only 

 from sirupy solutions, and are consequently difficult to purify by 

 recrystallization. 



8. When new salts result from a double decomposition the phospho- 

 tungstate formed does not often correspond in composition to that from 

 which it is derived. It may be of a higher or of a lower order. 



Preparation and General Properties. — Scheibler prepared the salta 

 which he has described by boiling neutral or acid sodic tungstates with 

 ^half their weight of phosphoric acid. I have found it more convenient 

 to employ the following methods. Solutions of neutral sodic tungstate, 

 WOjNao -j- 2 aq, and of hydro-disodic phosphate, PO^Na^H -|- 12 aq, 

 in the proportion of n molecules of the former to m of the latter, are 

 to be boiled together for some time. The solution has a very strong 

 alkaline reaction. It is to be neutralized with nitric or ehlorhydric 

 acid, and then contains the sodium salt — usually acid — of one or 

 another acid of the series. As the sodium salts do not as a rule crys- 

 tallize as well as those of potassium, I have sometimes found it advan- 

 tageous to add a solution of potassic bromide or nitrate, when, after 

 some hours, crystals of a potassic salt usually separate in abundance, 

 often in fine colorless and transparent cryst;ils. If a large excess 

 of acid is added at once to the mixed solution of sodic tungstate 



