122 PROCEEDINGS OF THE AMERICAN ACADEMY 



oxides of the types ^D^, ROo, and Tl.,0^ (coiTesi:)ondiiig to chlorides 

 RCls). We may assume jjrovisioually that the corapoueut 



enters into the structural formulas which represent this class of com- 

 pounds. Struve states that he has obtained similar compounds con- 

 taining tungsten, but these have not been described. 



Klein* has carefully studied four series of boro-tungstates, embraced 

 respectively under the general formulas 



7 WO3 BA . 2 R,0 



9 WO3 . B2O3 . 2 R.O 

 12 WO3 . BA . 4 R,0 

 14 WO3 . B,03 . 3 RA 



Since boron is the lowest term in the third column of the periodic 

 series, aluminum being the next higher, it would seem probable that 

 the oxides which correspond to still higher terms in the same column 

 will form similar compounds. Chromium, manganese, iron, nickel, 

 and cobalt may be regarded as higher harmonics of aluminum, and at 

 least three of these form analogous series. It still remains to study 

 the relations of the oxides of the cerium and yttrium groups to mo- 

 lybdic and tun gs tic oxides, since these correspond to chlorides of the 

 type RClo, and yet differ in a very marked degree from oxides like 

 AS2O3, SbgOs, etc. I have made some progress in this investigation 

 and will in another paper give ray results in detail. » 



GENERAL CONCLUSIONS. 



The formulas and details which I have given in the present instal- 

 ment of my work appear to justify the following general conclusions. 



1. Vanadic pentoxide unites with phosphoric or arsenic pentoxides 

 in various proportions to form well-defined complex acids. In some, 

 at least, of these, we seem to have series in which vanadic pentoxide 



* Bull, de la Societe Chim., [2,] xxxvi. 547, xxxvii. 202. 



