390 F. W. Clarice— Theory of the Mica Group. 



represented fairly well by the subjoined formulae, in which 

 ferric iron, present in small quantities, has been reckoned with 

 the alumina. 



/SiO=(MgOH) 3 /SiO = l M /SiO =MgH 



Al— SiO =Al Al— SiO = \ Mg3 Al— SiO =MgH 



\SiO=Al \SiO=Al \Si0 4 =Al 



+3H a O. +3H a O. + 3H,0. 



Dudleyite. Culsageelte. Jefferisite. 



It is by no means certain that the vermiculites are so simple in 

 composition or so definite as these symbols would seem to 

 indicate; but the formulae are decidedly suggestive, and they 

 show how clearly the relations between the micas and their 

 derivatives may be expressed. 



The ferro-magnesian micas. 



This group of micas, which includes biotite, lepidomelane, 

 annite, haughtonite, siderophyllite, and other supposed species, 

 is apparently quite complex. In place of magnesia, ferrous 

 iron is often predominant ; in the lepidomelanes, ferric iron 

 replaces aluminum, and in forty-four out of the fifty-six 

 analyses discussed, the oxygen was in excess of Si0 4 . Only 

 seven of these micas appeared to be phengitic ; and only six- 

 teen of the analyses reported fluorine in small quantities. In 

 six instances AlO and A1F 2 were both absent. 



For thirty-four of the micas in this group, formulae could be 

 easily computed upon the lines already followed. That is, in 

 each case the composition was represented by a mixture of 

 AlX 3 E // 5 E / 3 with A1X S R' 9 and A],X S R'„ the second or third of 

 these molecules being occasionally absent. Among the lepi- 

 domelanes there was an approximation to a distinctively mus- 

 covitic type ; and in two cases formula No. 4, AlgXgE^E, 

 seemed to apply. With each mica an attempt was made to 

 determine the proportions of the several admixed molecules ; 

 but the results, although numerically conformable to the gen- 

 eral theory, were not absolutely conclusive. The chief diffi- 

 culty lay in the uncertainty attaching to the water determina- 

 tions, upon which the question of oxygen excesses depends. 

 Since water has a low molecular weight, a small error in its 

 estimation becomes relatively large in the molecular ratios ; 

 and two sorts of errors are presumable. First, an excess of 

 water may be enclosed mechanically in the material analyzed ; 

 and secondly, a determination by simple ignition is likely to be 

 too low because of the oxidation of ferrous iron. If these 

 errors occur together, they obviously tend to compensate each 

 other; but either one alone seriously affects the coefficient of 

 E', and appears in the ratio between AlgXjE', and AIX3EV 



