OP ARTS AND SCIENCES. 261 



with a weighed quantity of normal sodic tungstate, in order to expel 

 the mercuric oxide completely, and obtain a constant weight. The 

 aurora-red suit may be regarded as a double compound, with the 

 possible formula 



3 {12 WO3 . P2O5 . 4 BaO . 2 11,0^ + 



" 2 {12 WO3 . VPs . 3 BaO . 3 Hp} -f- 132 aq. 



It has, I believe, the highest molecular weight, 20058, yet observed, 

 though I shall describe a little farther on a phospho-vauadio-vanadico- 

 tungstate with the formula 



60 WO3 . 3 P2O5 . V2O5 . VO2 . 18 BaO + 150 aq, 



with the molecular weight 200GG. It will be interesting to determine 

 the specific heats of bodies of this class, as well as the heat of forma- 

 tion. The formation of the aurora-red salt is very simply expressed 

 by the equation, 



3 {20 WO3 . P2O5 . G BaO} + 2 Vp. = 



GO WO3' . 3 P2O5 . 2 V2O5 . 18 BaO, 



the process being one of simple addition. 



I have mentioned above the fact that a dark red salt crystallizes 

 from the mother liquor of the aurora-red salt. The following ana- 

 lytical results were obtained with the small quantity of this salt at my 

 disposal : — 



0.7GG0 gr. gave 0.5570 gr. mixed oxides = 72.71% 

 1.0405 gr. " 0.2133 gr. SO.Ba r= 13.45% BaO 



1.0405 gr. lost on ignition 0.1451 gr. = 13.94% 



In the orange-brown salt formed by the decomposition of the aurora- 

 red salt, 



1.1919 gr. lost on ignition with WO.Na^ O.lGll gr. = 13.51% 

 1.0476 gr. gave 0.7G53 gr. mixed oxides = 73.05% 



1.4122 gr. « 0.3078 gr. SO.Ba =14.31% 



1.2372 gr. " 0.0087 gr. V^ = 0.70% 



No formulas can be deduced from these analyses, but they may serve 

 to identify the salts analyzed in future investigations. 



18 : 3 : 4 Series. — The 24 : 1 phospho-tungstate of sodium already 

 described, 24 WO3 . Vp^ . 2 Na.O . 4 H,0 -f 23 aq, is readily dis- 

 solved by boiling with a solution of ammonic meta-vanadate, and 

 gives a deep orange-red liquid. This solution does not yield crys- 



