F. W. Clarke — Theory of the Mica Group. 391 



In twenty-two of the fifty-six analyses, however, the symbol 

 AlXgB/'jIf,, failed to account for all the bivalent metals, iron 

 and magnesium. In these cases R" was in excess of E/, and it 

 became necessary to make use of formula No. 7, Al^SiOJgR",,. 

 With the aid of this expression, all the ferro-magnesian micas 

 without a positive exception, including the phlogopites, were 

 resolvable into mixtures of A1 3 X S R' 3 , AlX 3 R/ 9 , and A^XJR/',,. 

 Upon this basis formula No. 6 becomes useless, for AlXjMggR'g 

 is evidently equivalent to a mixture, in equal molecules, of 

 A1X 3 R' 9 and Al 2 X 6 Mg 9 . All micas, then, so far as the analyses 

 are authentic, may be represented as mixtures of the molecules 

 1, 3, and 7, these being symbols of one and the same general 

 type. The magnesium (or ferrous) salt thus assumed, more- 

 over, is not absolutely hypothetical ; for, with six molecules of 

 water of crystallization added it approximates to certain indi- 

 viduals of the chlorite group. Some examples of penninite, 

 for instance, approach rather closely to the composition 

 Al 2 (Si0 4 ) 6 Mg 9 .6H 2 0, which requires 38-71 of silica, 1097 

 alumina, 38 - 71 magnesia, and 11-61 water. We have thus a 

 clue to the constitution of the chlorites, by means of which 

 they may be brought into simple relations with the micas. 

 This problem is now under investigation, and I can only state 

 as a probability that most of the chlorites, if not all, may be 

 represented as mixtures of three fundamental molecules, 

 Al a X,R"„ 6H 2 0; A1X 3 R" 3 H 3 , 3H 2 0; and Al 2 X 3 (MgOH) 6 . I 

 hope to speak more positively upon this subject before long ; 

 at present I need only point out that two of these molecules 

 are simple hydrates of Nos. 6 and 7, while the third is covered 

 by the general symbol No. 3, in which R' 6 becomes (MgOH) 6 . 

 The easy alterability of garnet, Al 2 (Si0 4 ) 3 R // 3 into mica and 

 chlorite, is a suggestive bit of evidence bearing upon these 

 expressions. 



Margarite. 



Whether this species is to be considered as a true mica or 

 not, is perhaps an open question. Its composition is rela- 

 tively simple, and is represented by the empirical formula 

 H 2 CaAl 4 Si 2 12 . Structurally, this may be written in two dis- 

 tinct ways, as follows : 



/OH .SiO=(A10) s H. 



Al— SiO=CaH or Ca( 



\ Si0 4 =( AlO), X SiO =(A10),H. 



The first of these symbols derives the compound from the 

 normal aluminum salt, and indicates a similarity of type with 

 the micas proper. The second assumes a normal calcium salt 

 as the point of derivation ; and I am unable as yet to assign a 



