SYSTEMATIC PETROGRAPHY 479 



amounts of the oxides found by analysis, and in the other the 

 amounts of the elements potassium, sodium, and calcium. Here 

 again is the situation that only a part of some rocks is actually 

 under discussion, and the result can be of no real value to 

 petrography. 



F. Loeivinson-Lessing, 1890, 1897. ^ n ^90 an attempt at a 

 chemical classification of igneous rocks was made by F. Loew- 

 inson-Lessing, 1 based upon the quantitative relations of silica to 

 the various oxides of the bases, grouped under R 2 0, RO, and 

 R 3 3 , as shown in percentages by bulk analysis. By means of 

 empyrical formulae the author thought to find a way of expressing 

 regular relationships supposed to exist between the silica con- 

 tents and the various oxide groups. Rocks exhibiting the 

 following simple relationships were designated types, I to V being 

 the principal ones, and VI to IX intermediate : 



, I. Si(X= 2 (R.O + RO) + R 2 3 + O . 

 Acid ) 



Rocks "J VI. Si0 2 =^(R 2 0+ RO)+ R 2 3 +0 . 



II 



. SiQ 2 = 2 (R 2 + RO) + R 2 3 (il = I + IV ) 



Neutral 

 Rocks | VII si0 2 = ^(RX> + RO) + R 2 3 



Basic 

 Rocks 



/ II + IV\ 

 ( III. Si0 2 = R 2 + RO + R 2 3 (III = — — J . 



( VIII. Si0 2 = R 2 + RO + - R a 3 (vill = IH + IY ) 



( IV. Si0 2 =or<RO 



IX. Si0 2 = -RO 

 2 



Ultra 

 Basic 

 Rocks 



V. SiO,= 



Since percentages instead of molecular ratios were used, the 

 simple relations here adopted have no real significance as 

 expressing a regular connection between chemical and mineral 

 composition. This fault was perceived by the author and cor- 



*" Etude sur la composition chimique des roches e"ruptives," Bull. Soc. Beige de 

 Ge'ol., 1890, IV, Mem., p. 221. 



