GIBBS. — COMPLEX INORGANIC ACIDS. 279 



with chemists, and has in fact, so far as I am aware, passed wholly 

 unnoticed. The highest term actually obtained by me * appeared to 

 have the formula : 



16 WO3 . 3 Na^O . 4 (NH4)20 + 18 aq. 



The only complete analysis made agreed well with this formula, 

 and differed very materially from that of the 12 : 5 ammonia-sodic 

 tungstate. As it has been asserted, however, that the two are iden- 

 tical, Mr. Charles D. Smith has made in my laboratory four analyses 

 with portions of the salt which had been preserved. The analyses 

 are as follows : 



0.7226 gram gave 0.6038 gram = 83.55% WO3. 



1.1598 grams gave 0.9687 gram = 83.54% WO3. 



1.3281 grams gave 1.1091 grams = 83.51% WO3. 



1.4740 grams gave 1.2309 grams == 83.50% WO3. 



1.2726 grams gave 0.0604 gram (NH4)20 = 4.75%. 



1.0901 grams gave 0.0508 gram (NH4)20 = 4.66%. 



1.1360 grams gave 0.0532 gram (NH4)20 = 4.67%. 



0.6756 gram gave 0.0318 gram (NH4)20 = 4.72%. 



4.4687 grams lost on ignition 0.5292 gram = 11.84% NH3 and H2O. 



3.0697 grams lost on ignition 0.3638 gram = 11.84% NHgandHgO. 



4.6030 grams lost on ignition 0.5444 gram = 11.82% NHgand H2O. 



3.3885 grams lost on ignition 0.4014 gram = 11.84% NH3 and HgO. 



The analyses correspond to the formula, 



24 WO3 . 5 NaaO . 6 (NH4)20 + 27 aq., 



which requires : 



6676 100.00 



The sodic oxide is determined by difference. The analyses agree 

 rather more closely with the new formula than with that formerly 

 given, which I will cite for the sake of comparison: 



16 WO3 . 4 (NH4)20 . 3 Na20 + 18 aq., 

 requires ; 



* Ibid., XVI. 76. 



