Washington — Constitution of Some Salic Silicates. 567 



The composition of natural nephelite here proposed is much 

 the same as that of Schaller and of Bowen, hexagonal (KSi) 

 molecules being in isomorphons mixture with pure nephelite 

 and kaliophilite. The result is the same if we write: 

 2ab + 9ne + 21c. The former is preferred, as potassium 

 shows greater affinity for silica than sodium. The composition 

 of either mixture, which is supposed to have the ratio 

 Na 2 : K s O = 5 : 1, is : Si0 2 = 43-43, Al A = 34-08, Na 2 = 

 17*26, K 2 = 5-23. This agrees well with the best analyses 

 cited by Foote and Bradley, the ratio Si0 2 : A1. 2 3 : (Na,K) 2 

 being 217 : 1 : 1. As the natural mineral is thus an isomor- 

 phons mixture, and not a molecular compound like leucite, its 

 composition will vary somewhat, but can always be expressed 

 in terms of the three molecules given above. 



The constitution of cancrinite is uncertain, though undoubt- 

 edly complex, but can be expressed rationally as above. It 

 does not seem possible to harmonize the reported compositions 

 of microsommite and davyne with these formulas, but these 

 rare minerals have been very imperfectly investigated. 



As regards attack by acids we must note that all the minerals 

 of this group, distinguished by the presence of (B/ 2 A1) and the 

 total absence or very small amount of the (B/Si) integral, are 

 readily soluble in acids with gelatinization. 



Scapolite Group, tetragonal. 



Marialite, (NaSi)„AlSi 3 O s Meionite, (CaAl) 2 AlSi 3 8 



Mizzonite, (NaSi) 3 (CaAl) (AlSi 3 8 ) 2 



Wernerite, (NaSi) (CaAl)AlSi 3 8 



As mentioned above, under this interpretation the acid of 

 the scapolites (scapolitic acid) differs somewhat from that of 

 the feldspars and lenads, being H 10 AlSi 3 O 8 . This general 

 resemblance and the joint presence of the integrals (NaSi) and 

 (CaAl) would explain the relationships and analogies between 

 the feldspars and the scapolites. In the formulas given above 

 the intermediate members, ideal mizzonite and wernerite, are 

 shown with stoichiometric ratios, though in nature they are 

 really somewhat indefinite mixed crystals of the two end 

 members, like the plagioclase series, which are of identical 

 type of formula and therefore capable of isomorphous replace- 

 ment and admixture, agreeing also crystallographically. 



The function of CI and S0 3 , which are often present, espec- 

 ially toward the marialite end, is uncertain. It is worthy of 

 remark, however, that sodium chloride is driven off on heating, 

 just as it is from sodalite. It is possible to assume the presence 

 of a quadrivalent integral (A1C1) replacing silicon and write 

 the formula of chlorine-bearing marialite and of meionite 

 thus : 



