26 HOLLAND : MICA DEPOSITS OF INDIA. 



groups the general symbol R, we can imagine the following derivatives 

 of Al 4 (Si0 4 ) 3 as easily possible: — 



1 II III 



/-Si0 4 =ff 3 /Si0 4 =/? 3 /-Si0 4 =/? 3 



Al — SiO„=Al Al^— Si0 4 =R 3 Al^Si0 4 =/? 3 



\ Si0 4 =Al ' Si0 4 =A'l N Si0 4 =/? s 



IV V VI 



/Si0 4 =Mgff ■/ Si0 4 =Mgff / Si0 4 =Mg/? 



Al— Si0 4 =Al Al — Si0 4 =Mgi? Al^— Si0 4 =Mg/? 



\Si0 4 =AI \Si0 4 =Al \Si0 4 =Mg/? 



To these we may add as the bivalent analogue of III, the com- 

 pound Al 2 (Si0 4 ) 6 Mg 9 (VII). In the case of the fluorine-bearing forms 

 R 1 is represented by the univalent groups MgF and A1F 2 . Most of 

 the micas, Clarke considers, can be regarded as intermediate mixtures 

 of the above presumably isomorphous types. Thus No. I represents 

 muscovite and paragonite ; No. VI some phlogopites ; Nos. II, V and 

 VI are really unnecessary as types, being mixtures of pairs of Nos. I, 

 III, IV and VII, whilst even No. IV may be regarded from the same 

 point of view as superfluous. These formulae are satisfactorily appli- 

 cable to the micas which on analysis appear to be normal orthosilicates, 

 but there are some in which the oxygen is in excess of Si0 4 , and in 

 these Clarke assumes the existence of AlO taking the part of R l and 

 being the equivalent of Al F 2 ; in others the proportion of oxygen to 

 silicon is lower than in Si0 4 , and in these the polysilicic acid 

 H 4 Si 3 8 is supposed to replace the (also tetrabasic) orthosilicate 

 H 4 Si0 4 , giving rise to a set of circumstances paralelled amongst 

 the plagioclase felspars in which albite (NaAlSi 3 8 ) and anorthite 

 (CaAl 2 (Si0 4 ) 2 ) are apparently isomorphously replaceable. 



In these three attempts to trace a chemical isomorphism corre- 

 sponding to the evident[physical similarities of thejdifferent members of 

 the mica family, the existence of hypothetical compounds not yet found 

 separate in Nature has been postulated ; but Clarke's theory most 

 nearly escapes the dangers of this method, and is supported by partial 

 analogy with the felspar group. Clarke also asserts a similar form of 



( 16 ) 



