120 GEOLOGY OF THE YELLOWSTONE NATIONAL PARK. 



while in the more siliceous end they vary in opposite directions. There is a 

 marked accordance between the soda and the alumina, both varying in the 

 same direction, with one exception, though not to the same extent. There 

 is a more strongly marked discordance between the alumina and the 

 magnesia, which, with one exception,, vary in opposite directions. 



These irregular variations take place not only among allied varieties of 

 rocks, but even in different parts of one and the same rock body. They 

 find expression in variations in the proportions of the essential minerals. 



The origin of this variation is undoubtedly to be sought in the chemical 

 differentiation of the molten magma. The development of the constituent 

 minerals in the solidified rock is the result of physical forces that combine 

 the chemical constituents of the magma in a variety of ways to form closely 

 analogous crystal compounds. A discussion of the possible molecular 

 condition of molten magmas will not be taken up in this place. A few 

 obvious relations between the mineral and chemical variations, however, 

 may be pointed out. The inverse variation between alumina and magnesia, 

 and the accordance between alumina and the alkalies, affect the relative 

 proportions of feldspars and ferromagnesian silicates, which vary in an 

 inverse ratio. The decrease of the alkaline earths with the increase of silica 

 and the alkalies shows itself in the diminution of the ferromagnesian silicates 

 and the calcium feldspars, which accompanies an increase in quartz and the 

 alkali feldspars. 



The reciprocal variation of the ferrous and ferric oxides indicates the 

 variable oxidation of preexisting ferrous molecules. This should naturally 

 be in accord with the development of minerals containing more or less ferric 

 oxide, the most prominent of which are hornblende and biotite. Such a 

 connection seems to be made out, but data for its complete demonstration 

 are not at hand. It is most significant on account of its bearing on the 

 question of the development of these two minerals in the coarser-grained 

 forms of rocks whose magmas may crystallize under other conditions free 

 from either mineral. It may throw light on the possible action of water 

 vapor as a mineralizing agent. 



The order in which the constituents crystallized out of the molten 

 magma to form diorite may be learned by considering the relative ages of 

 the component minerals. It has been pointed out that some crystallized 

 almost synchronously, but that they began to separate from the liquid 



