8o4 WHITMAN CROSS 



marked off by the percentages of silica shown on analysis, the lines 

 being drawn at 66 and 52 per cent. These Hmiting percentages are 

 admitted by Hatch to be "quite arbitrary" but permitting "a 

 convenient separation in accordance with existing records of rock 

 types. "^ The meaning of the latter phrase is not clear to me. 

 Now it must be self-evident that no chemical consideration of 

 importance, to say nothing of a principle, requires or suggests a 

 division of igneous rocks by any special silica percentages what- 

 ever. Calling the group with more than 66 per cent SiOa "acid," 

 and designating all the rocks of this group as the " Granite Family," 

 implies that only rocks of this silica content may contain important 

 amounts of quartz. The application of the name "Syenite Family " 

 to alkalic rocks of the intermediate group implies that they must 

 be quartz-free, or nearly so. And the designation of the group 

 with less than 52 per cent SiOa as "basic" implies the generali- 

 zation that quartz is absent and that sihcates of such rocks must 

 be less rich in silica than those of the intermediate group. 



The well-known facts are that many rocks having less than 52 

 per cent SiOz contain metasilicates and several per cent of free 

 siUca; many intermediate rocks carry 25 to 30 per cent of quartz; 

 and many of 66 per cent carry little or no quartz. It is a funda- 

 mental principle, ignored by Hatch, that the petrographic impor- 

 tance of the silica contents of a rock, as influencing the development 

 of its minerals on crystalKzation, can be determined only by a 

 study of the relative amounts of the associated bases and recognition 

 of their influence in the magmatic solution. This principle is 

 illustrated by the normative calculations on which the Quantitative 

 System is based. 



The fact that the silicity^ represented by 66 per cent SiOz has 

 no relation to the presence of free quartz is illustrated by three 

 analyses quoted by Hatch in tables representing the Calc-Alkali 

 Series.^ In Washington's tables it is shown that the granodiorite 

 of SiOa 68.65 has 24.2 per cent normative quartz; the tonalite of 



' Science Progress, p. 2. 



- A usefiil term of self-evident meaning recently proposed by Washington (this 

 Journal, XXII [1914], 16). 



3 Science Progress, p. 9. 



