270 F. W. Clarke — Minerals of Litchfield, Maine. 



simple general formula for cancrinite, covering all variations in 

 composition except such as are due to impurity, and correlating 

 the mineral with the allied species sodalite and nosean. For 

 the Litchfield mineral the following special formula may be 

 written, giving the theoretical composition in the column 

 below : 



Al 8 (Si0 4 ) 8 (C0 4 ) 2 CaNa 8 H 6 . 



Si0 2 



Found. 



35-83 to 37-22 



Calculated 



35-9 



Al o 3 



Na 3 



CaU 



C0 2 



28-32 to 30-12 



19-33 to 19-56 



4-27 to 5-12 



6-22 to 6-96 



30-6 



18-6 



4-2 



6-6 



H 2 



2-98 to 3-86 



4-1 



100-0 



In this case the water as found is slightly lower, and the soda 

 slightly higher, than the calculated values; which is probably 

 ascribable to the mutual replaceability of sodium and hydrogen. 



The formula commonly accepted for sodalite, and the one 

 which is certainly the simplest, is that deduced by Bamberger* 

 from his analysis of the mineral from Tiahuanuco. Written 

 empirically, this formula is Na fi Al 4 (Si0 4 ) 4 Cl, which requires 

 considerably less chlorine than has ordinarily been found in 

 the species. In Bamberger's analysis, as finally corrected, he 

 obtained 5'54 per cent, as against nearly seven per cent in 

 Whitney's determinations. The difference he ascribes to silica 

 in the chloride of silver as weighed by other analysts ; and yet 

 in my own estimation every care was taken to eliminate such 

 impurity, and my results confirm the older figures. Still, both 

 figures have theoretical interest, as will be seen further on; and 

 I am inclined to believe that the Bolivian mineral was more 

 nearly typical than that from Litchfield. To the latter we may 

 assign the empirical formula JSTa 9 Al 7 (Si0 4 ) 7 Cl 2 , which is 

 directly derivable from the formula for nephelite, and which 

 agrees quite sharply with the analyses. 



We now have three empirical formulae ready for comparison 

 side by side, as follows: 



Nephelite Al 8 (Si0 4 ),(Si0 3 ) 2 Na 8 



Cancrinite (Litchfield) Al 8 (Si0 4 ) 8 (C0 4 ) 2 Ca Na 8 H 6 



Sodalite " Al 1 (Si0 4 ),Cl 2 Na" 9 



These may easily be put into structural form by an applica- 

 tion of the principle suggested in a former paper,f that ortho- 

 silicates containing aluminum are to be represented as substi- 



* Zeit. fur Kryst, v, 581. 



f Clarke and Diller, this Journal for May, 1885. 



