August 26, 1898.] 



SCIENCE. 



237 



36W03.3SiO,.2A],03 + 93H,0 and + 81H,0 

 seWOjSSiO.'.SAlACISrHJp + 75H,0 

 36W03.3SiO,.2Mb5.8]Sra.,0 + 45H,0. 



These salts may all be considered as con- 

 taining three molecules of the complex : 



ISWOj.SiO,. 



The two first may be, respectively : 



2(A1A3H,0) + 3(W03SiO,4Hp)+ 75H,0 



2(A1,0.3H,0) 



+ 2(12W03SiO,.3(ISrHJ,OH,0+ 63H,0 



and the third : 



2(N",05N"aP) + 3(12W03SiO,2]Srap2H,0) 

 + 39'Hp. 



The following also deserve notice in the 

 same connection : 



12W03SiO,3K,O.H,0 



+ 12WO,SiO,.2K,032H,0 + 26H,0. 



12W03SiO,.3CaO.H,0 ' 



+ 12W03SiO,.2Ca0^2Hp + 46H,0. 



The simplest view which can be taken of 

 all these compounds is that they contain 

 the complex i2WO3.SiO2.4RO and that they 

 are double and not quadruple salts or acids. 

 And it seems to me that the weight of evi- 

 dence is in favor of the view that the com- 

 plex W03SiO,.4RO or 10WO3SiO,.4EO is 

 single or unitary and not a double salt of a 

 tungstate and silicate. 



Marignac did not attempt to explain the 

 isomerism of silico-tungstic and tungsto- 

 silicic acids, and the problem is one of much 

 interest, since there are no allotropic modi- 

 fications of tungsten or silicon. 



To obtain a further insight into the con- 

 stitution of the simpler complex acids we 

 may, with advantage, also consider the 

 cases of the phosphotungstates and phos- 

 phomolybdates. As normal we may take 

 the acids 12W03.PO,H3 or 12Mo03.PO,H3. 

 These correspond to the twelve molecule 

 silico tungstates or tungsto-silicates of Ma- 

 rignac, and the reasoning applied to these 



cases will apply also to the phosphotung- 

 states and phosphomolybdates. Only in 

 these last the number of known compounds 

 is much greater, acids in which the propor- 

 tion of tungstic or molybdic oxide to phos- 

 phoric oxide is as 24:1, 22:1, 20:1, 18:1, 16:1, 

 5:1, being known in their corresponding 

 salts. Besides these we have arsenic com- 

 pounds in which the ratio of tungstic or 

 molybdic oxide to arsenic oxide is as 6:1 

 and 7:1, as well as double salts in the ordi- 

 nary acceptation of the term. Instances 

 of these last are : 



(1) 22Mo03.P,0,.3K,0 



4- 22Mo03.PA-2K,O.H,0 + 21 H,0. 



(2) 2(20MoO3.PA-3(N"HJ20) 



+ 20MoO3.PA-2(N'HJ,O.H,O + 11 H^O. 



(3) 24Mo03P,05.3(NH,),0 



+ (24Mo03.PA-2(NH,)p.Hp -f- 16 H^O. 



Upon any other than the view which I 

 have taken the phosphotungstates and phos- 

 phomolybdates of the types nW03Pj056EO 

 and nMo3.Pj056EO are salts sui generis 

 analogous to metatungstates and meta- 

 molybdates. The compounds expressed by 

 these last formulas are quaternary. 



We may compare these salts (1), (2) and 

 (3) with the double tungstate 



12 WOjSNa.O + 10 W03.4Na,0 + 51 aq. 



They behave like salts of meta-tungstates 

 or meta-molybdates which have taken into 

 their structure molecules of phosphoric or 

 arsenic oxides without materially changing 

 their general character. The precise mode 

 of combination of the tungstic and molybdic 

 molecular masses with silica, phosphoric or 

 arsenic oxides is not at present known, but 

 it is not difiicult to show that the units of 

 afi&nity of all the elements present may be 

 accounted for and illustrated by structural 

 diagrams. In like manner, the condensa- 

 tion of n molecules of tungstic or molybdic 

 teroxide is easily explained, the number of 

 unsaturated or residual units of affinity 



