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



distinct preference to either expression. A careful study of 

 the genesis and associations of margarite may determine which 

 of the two formulae is the better. Possibly the occurrence of 

 the mineral in such mixtures as euphyllite, and the probable 

 existence of a similar barium salt in oellacherite points to the 

 first of the two formulae as the better. The common occur- 

 rence of margarite with diaspore also points in the same direc- 

 tion. 



The Clintonite grozip. 



These minerals, the so-called " brittle micas, " have also been 

 discussed by Tschermak,* who includes under this heading 

 seybertite, brandisite, xanthophyllite, chloritoid, masonite, 

 ottrelite, sismondine, and sapphirine. Physically, they are 

 closely related to the micas proper, and to margarite in partic- 

 ular ; but chemically they are much more basic. The first 

 three species Tschermak regards as mixtures of the hypothetical 

 compounds H 2 CaMg 4 Si 3 12 and H„CaMgAl 6 12 ; to chloritoid 

 he assigns the composition H 2 Si 2 Fe // 2 1 + H !! Al 4 7 , and sap- 

 phirine he represents by the formula Si 2 Mg 2 6 + Al 6 Mg 2 O n . 



Using the analyses cited by Tschermak, I find that all these 

 minerals, with the possible exception of sapphirine, may be 

 represented by the general expression 



Al— >1X 

 \SiO=R' 3 , 



which is clearly and directly related to the formulae already 

 assigned to the micas and to margarite, and in which the mode 

 of union of R" with Al, when R ;/ =Mg, suggests the common 

 association of members of this group with spinel. In seybertite 

 we have a mixture of A10 2 R // Si0 4 (MgOH) 3 with AlO.R^SiO, 

 (A10) 3 , R" being partly Ca and partly Mg. In brandisite we 

 have a similar constitution, with about one fourth of R/' 

 replaced by H 2 . In chloritoid, R'^Fe and B' 3 =H 3 (A10); 

 and sismondine is similar, with R/ 3 possibly replaced in part by 

 Al. The compound A10 2 MgSi0 4 (A10) 3 , found, in seybertite, 

 is a rough approximation to sapphirine, which mineral possibly 

 has this composition plus some impurities as yet unidentified. 

 A wider range of analyses is needed in order to establish these 

 formulae completely ; but they seem to have distinct advantages 

 over the formulae proposed by Tschermak. 



Co?iclusions. 



All the micas, vermiculites, chlorites, margarite, and the 

 clintonite group may be simply represented as isomorphons 



* Zeitschr. Kryst. Min., iii, 496. 



