714 EEPORT— 1893. 



break up the union of the two elements, hut rather cause the ultimate formaiion 

 of the aiumino-silicates which are so abundant in most of our rocks. 



Yiewed in the light of tbe facts already cited and the inferences we have drawn 

 from them as to the nitrogen-like relationship of aluminium to silicon, I am 

 disposed to regard the natural aiumino-silicates as products of final oxidation of 

 sometime active silico-aluminium analogues of carbo-nitrogen compounds, rather 

 than ordinary double salts. It is generally taken for granted that they are double 

 salts, but recent work on the chromoxalates by E. A. Werner has shown that this 

 view is not necessarily true of all such substances. 



Without going into undue detail we can even form some conception of the 

 general course of change from simple aluminium silicide to an alumino-silicate, if 

 ■we allow the analogies already traced to lead us further. 



We recognise the existence of silico-formyl in Friedel and Ladenburg's silico- 

 formic anhydride ; hence silico-triformamide is a compound whose probable formation 

 we can admit, and, on the basis of our aluminium-nitrogen analogy, an aluminium 

 representative also. Thus — 



.COH .SiOH .SiOH .SiO.R' 



N— COH : N— SiOH : Al— SiOH : Al— SiO,B' 



\C0H ^SiOH \siOH ^SiO-.K.' 



Triformamide. Silico-triformamide. Silico-alumino- Salt of an alumino- 



triformamide. silicic acid. 



Now, oxidation of triformamine would lead to complete resolution into nitrogen 

 gas, carbondioxide gas and water rendering it an extremely unstable body ; under 

 similar conditions silico-triformamide would probably alford nitrogen gas and 

 silicic acid (or silicon dioxide and water) ; while the third compound, instead of 

 breaking up, would (owing to the fixity of aluminium as compared with nitrogen) 

 be likely at first to afford a salt of an alumino-silicic acid, in presence of much 

 basic material. 



The frequent recurrence of the ratios SijAl, SijAl.,, &c., in the formulse of 

 natural aiumino-silicates, suggests that some at least of "these minerals are derived 

 from oxidation products of the above triformic type. Without stopping to trace 

 all the possible stages in the oxidation of the primary compound Al(SiO^Il)3, or 

 variations in basicity of the products, I may cite the four following examples out 

 of many others which might be given of resulting representative mineral groups : — ■ 



.SiOjR' .SiO^R'3 /SiO^E'g .SiO^R"" 



Al— SiOjR' : Al— SiO^R'., : Al— SiO^R'" : Al— SiO^R'" 



\siOsR' \siO,R"' ^SiO^R'" ^SiO.R'" 



Beryl tjrpe (hemi-). Garnet type. Muscovite type. Xenolite type. 



Five years ago Professor F. W. Clarke, of the United States Geological 

 Survey, published a most interesting paper on the structure of the natural sili- 

 cates. In this he adopts the view that the mineral xenolite, SiaAljO,,, is the 

 primary from which all other aiumino-silicates may be supposed to arise by various 

 substitutions. Nature, however, seems to teach us that such minerals as xenolite, 

 fibrolite, and the related group of ' clays' are rather to be regarded as end-products 

 of a series of hydrolytic changes of less aluminous silicates than primary substances 

 themselves ; hence the sketch which I have ventured to give above of the probable 

 genesis of aiumino-silicates seems to provide a less arbitrary basis for Clarke's 

 interesting work, without materially disturbing the general drift of his subsequent 

 reasoning. 



We may now consider for a moment in what direction evidence can be sought 

 for the existence in nature of derivatives of the hypothetical intermediate products 

 of oxidation between a primary silicide and its fully oxidised silicate. 



' In these cases where R"' = A1 it is, of course, assumed that the latter is acting 

 only as a basic radical. 



