240 



SCIENCE. 



[N. S. Vol. VIII. No. 191. 



may act as a whole ; the second that V„0- 

 may form combinations with a number of 

 molecules of tungstic or molybdic teroxide 

 analogous to phosphotungstates or phos- 

 phomolybdates. These assumptions, taken 

 separately or together as the case may re- 

 quire, in many cases at least reduce com- 

 plex to simple formulas perfectly com- 

 parable to those of salts containing phos- 

 phoric or arsenic pentoxide. Thus we have 



5W03.VA-4(l^HJ,0.2Hp + IIH.O. 

 5Mo03.V,05.3(NH,),0.3H,0 -|- 4H,0 and 

 SMoOj.VA-SBaO + SMoOj-SBaO^H^O. 



In addition I have described two vanadio- 

 tungstic acids having respectively the form- 

 ulas: 



10WO3.V,O6.6H,O +16 H,0. 



ISWO^VA-SHjb + 30 H,b. 



The existence of these compounds has 

 been denied,* but I see no sufficient reason 

 for rejecting the formulas, which at least 

 closely correspond to the analyses. The 

 same remark applies to the acid 



20VA-P2O5-6H,O + 53 H,0. 

 or 



20V,OA-PA-('iH:p + 53 H,0. 



which I would compare with 



20WO3.PA-6H3O. 



It must be admitted, however, that in 

 view of the extraordinary number and va- 

 riety of the combinations of vanadic pentox- 

 ide with bases it cannot be denied that 

 there are very numerous compounds of this 

 oxide with the teroxides of tungsten and 

 molybdenum, which, in the present state of 

 our knowledge, must be classed as double 

 salts. The empirical composition of these 

 salts may be expressed by the general 

 formula : 



mWOj.nVA-PI^O 

 or m'MOjU'VjOsP'EO. 



*See Friedheim. Beriolite, Band XXIII., p. 1505 

 and 1530 (and Rosenheim Dissertation, Berlin, 1888); 

 Zeitsehrift fiir amorgan. Chemie, Band V., p. 437 ; 

 Berichte, Band 24, p. 1173. 



in which m, m' represent respectively the 

 number of molecules of tungsten or molyb- 

 denum and n, n', p, p' are whole numbers. 

 Friedheim assumes that in many of these 

 compounds vanadic pentoxide acts as a base 

 and that in others mixtures of isomorphous 

 salts are present. The pentoxide dissolves 

 readily in the stronger acids,as, for instance, 

 in sulphuric and phosphoric acids, but I am 

 not aware that well defined crystalline salts 

 of this oxide have been obtained. There are 

 also, I believe, no known instances of iso- 

 morphism between tungstic and vanadic or 

 molybdic and vanadic oxides. In spite, 

 therefore, of all that has been done, it seems 

 necessary to admit that our knowledge of the 

 vanadio-tungstates and vanadiomolybdates 

 is still very imperfect, but this remark ap- 

 plies also to the very large group of alka- 

 line vanadates.* Of course, the theory that 

 in the salts in question we are dealing, 

 ■chiefly at least, with double salts is that 

 which would naturally present itself first, 

 but we meet with the difficulty that the 

 constituents of these double salts are 

 themselves double salts, so that we must 

 deal in fact with triple and quadruple salts. 



Id my view the acids or salts which are 

 usually termed complex may be divided 

 into two classes. 



Class first embraces those in which the 

 union of the determinant and sub-deter- 

 minant is so complete that the special char- 

 acters of the constituents are nearly or 

 wholly lost. The salts of this class are not 

 decomposed by solution. They may be 

 compared to certain so-called double cya- 

 nides, as, for example, to ferrocyanides and 

 cobalticyanides. The acids 12 WOj.PO.H, 

 and 12 MoOj-PO^H, are examples. 



Class second embraces double salts proper 

 which are decomposed by solution alone 

 and which contain two different constitu- 

 ents. Salts of this class are very numerous, 

 and there are probably various sub-classes. 



*See Proceedings of Am. Acad., Vol. XVIII., p. 74. 



