F. W. Clarke — Theory of the Mica Group. 385 



some cases nearly up to five. In certain instances, as Tscher- 

 mak himself suggests, this excess of oxygen over the orthosili- 

 cate ratio may be due to water enclosed between micaceous 

 laminae ; and in other cases it may indicate partial alteration. 

 There are micas, however, in which neither of these explana- 

 tions appears to be satisfactory, and the weight of evidence 

 goes to show that the excess of oxygen is essential. If this be 

 true, Tschermak's theory is so far inadequate ; for to such 

 micas it cannot apply without serious modification. The ques- 

 tion can only be settled experimentally, but at present the 

 objection raised by it cannot be ignored. Some of its details 

 will be considered later. 



]STow, in order to conclusively replace Tschermak's theory by 

 something better, two fundamental conditions must be satisfied. 

 First, all micas, with the possible exception of the pseudo- 

 mica, margarite, should be reducible to one general type of 

 formula, which shall express all known relations equally well 

 with the formulae proposed by Tschermak. Secondly, hypo- 

 thetical compounds must be so far as possible avoided, and 

 admitted into consideration only when their assumption can be 

 shown to be absolutely necessary. The chief difficulties to be 

 overcome are the variations in the silicon-oxygen ratio, and the 

 presence of fluorine in many common micas. 



In a number of papers published during the past four years,* 

 I have sought to show that all orthosilicates containing alumi- 

 num may be represented as substitution derivatives of the 

 normal salt Al 4 (Si0 4 ) 3 . Muscovite can be so derived, theoreti- 

 cally, through the replacement of one aluminum atom by R' 3 ; 

 and in many other cases we have equal simplicity of expres- 

 sion. The ferro-magnesian micas, the phlogopites, and the 

 lepidolites, however, are less simply derived ; and I have in 

 certain cases suggested a linking together of two orthosilicate 

 nuclei to the group Al 2 (Si0 4 ) 6 in order to account for their 

 formation. I am now inclined to believe, however, that all the 

 true micas are referable to the same general type, and have 

 discussed nearly one hundred published analyses from that 

 point of view. If we take magnesium as a generic represen- 

 tative of the bivalent metals, and give univalent elements or 

 groups the general symbol R, we can imagine the following 

 derivatives of Al 4 (Si0 4 ) 3 as easily possible. 



1. 2. 3. 



/SiO=R s /SiO=R 3 /Si0 4 =R 3 



Al-SiO=Al Al-Si0 4 =R 3 Al-Si0 4 =R 3 



\Si0 4 =Al \SiO=Al \Si0 4 =R 3 



* See this Journal, Nov., 1886 and Aug., 1887. Also Amer. Chem. Journ., x, 

 120, March, 1888. 

 Am. Jour. Sci.— Third Series, Vol. XXXVIII, No. 227.— Nov., 1889. 

 25 



