E UTECTICS IN ROCK MA GMA S 85 



as he himself has pointed out, there are no more alkalis in most 

 rocks than alumina and silica to go with them. In the second 

 place, there are very few rocks with high silica percentage down 

 about to the point representing the average rock in which the 

 ratio of alkali-silica is less than 0.04. As the amount of alkalis 

 increases, the number of analyses becomes more numerous. 

 Somewhere between the ratios 0.08 to 0.09 they are most abun- 

 dant. After that they diminish somewhat and become less 

 up to a ratio between 0.12 and 0.16, and then they become more 

 numerous once more. If we suppose this not to be accidental, 

 we must assume a te7ide?icy on the part of a certain group of magmas 

 toward an alkali-silica ratio betzveen 0.08 and o.og. I fixed on 0.083 

 from the diagram. This, I noted, was the ratio of the average 

 rock. Expressing this in fractions, the alkali-silica ratio would 

 be 1:12. This suggested to me at once a combination of ortho- 

 clase or albite and quartz with equal molecular proportions of 

 silica. This again reminded me of Rosenbusch's spherulite-form- 

 ing microfelsite, which he supposes to be of the composition 

 KgO • AI3O3 ■ X SiOg, with x greater than 6. It is also the com- 

 position of micropegmatite where the quartz and feldspar are 

 equal. All these facts harmonize with the supposition that the 

 combination (Na^O, K^O) • AlgOg • 6Si02+6Si02 is the eutectic 

 ratio of alkali and silica. If such magmas are composed 

 mainly of alkalis, alumina, and silica, these will tend to crystal- 

 lize out whichever of these components happens to be in excess, 

 and the analyst will be likely in the long run to obtain more 

 analyses near this ratio than either above or below. This suppo- 

 sition is obviously compatible with the idea that the igneous 

 rocks are derived from fusion either of pre-existing sedimentaries 

 or of planetesimals. 



We have remarked that down to where silica equals 0.59 and 

 0.57 there seems to be a belt of analyses having about this 

 ratio, with comparatively few as the ratio dropped. Now, 

 of course, these rocks are not all simply feldspar and quartz. 

 How do we explain this belt or line? Simply by supposing 

 that this eutectic relation holds, even though other elements 

 were combined with the silica, that the 6 silica may be more or 



