112 M. Scheerer's Observations on Elccolitli and Nepheline. 



Silica 43-36 



Alumina 33-49 



Peroxide of iron (with some manganese) . 1-50 



Lime 0-90 



Soda 13-36 



Potash 7'13 



Water 1'39 



101-13 



I likewise have analysed a portion of this variety. It was 

 however impossible, notwithstanding every precaution, to ob- 

 tain perfectly pure crystals, as the very crystals inclosed mi- 

 nute parts of the rock in their interior. Their nucleus is 

 frequently the most impure, often exhibiting a 6- or 4-sided 

 opacity, which is environed by a more transparent frame. I 

 found it to be composed of 



13. 



Silica 43-70 



Alumina 32-31 



Peroxide of iron . . 1-07 



Lime 0'84 



Soda 15-83 



Potash ..... 5-60 



Water 1-39 



100-74 



The amount of water is here adopted from Prof. Gm elm's 

 determination, as a sufficient quantity of pure mineral was not 

 at my disposal. The silica separated by hydrochloric acid left 

 about 2-5 per cent, residue, insoluble in carbonate of soda, which 

 evidently consisted of minute parts of the rock. A consider- 

 able difference between M. Gmelin's analysis and mine will 

 be found solely in the relative quantities of potash and soda. 

 The source of this is without doubt to be sought in the mode 

 of determining these two alkalies employed by L. Gmelin. 

 He weighed the heated chlorides and determined by a 

 solution of silver the quantity of chlorine contained in them, 

 and thence calculated the requisite quantities of chloride of 

 potassium and chloride of sodium which must be present 

 to give this quantity of chlorine in the mixture of the two 

 salts. On closer consideration, it is evident that every error 

 of weight is greatly increased by calculation, as a small 

 defect in the quantity of chloride of silver corresponds to a 

 considerable difference in the relative quantity of potash and 

 soda. 



