OF ARTS AND SCIENCES. 87 



question upon general grounds of probability, so that our conclusions 

 are at best uncertain. Finally, both formulas being at least possible, 

 it may be that there are two isomeric modifications of each series of 

 acids represented respectively by the formulas above given. There is 

 no present evidence of the existence of such isomeric modifications in 

 the case of phospho-tungstates, phospho-molybdates, or the correspond- 

 ing arsenic series ; but Marignac has shown that there are two isomeric 

 series of silico-tungstates, which he calls respectively silico-tungstates 

 and tungsto-silicates, and it may be that the difference between these 

 depends upon differences in the mode of combination, precisely similar 

 to those which I have pointed out above. I shall return to this sub- 

 ject in the general discussion of my results. With respect to the two 

 linking terms, 



II II 







I I I / 0^ I 



3 (HO) E PO - O - OP E (0H)3 and OP — PO 



no assumption is made which is not in perfect accordance with com- 

 monly accepted views of the subject. 



We may now consider the most general case, that, namely, in which 

 there are twenty-four atoms of tungstic or molybdic, to one of phos- 

 phoric or arsenic oxide. We have for an acid of this type 



24 WO3 . P2O5 . 6 HP or 24 WO, . 0,, . PA • i^O)^ 



and in accordance with the principles above laid down the Btructural 

 formula may be written : — 



