EARTH S CRUST WASHINGTON. 



299 



promise of fruitful application in the future. A similar " order of 

 affinity " as regards alumina is also true of the same elements. 



Magnesia and the iron oxides tend to go together, or to replace each 

 other in many minerals, which seems to be of much the same import, 

 and these oxides are, as we have seen, generally opposed to soda, 

 potash, and lime. 



Among the more interesting of such correlations are those of soda 

 and iron on the one hand, and of potash and magnesia on the other, 

 these two pairs tending to go together. This is shown by many min- 

 erals, the details concerning which it is unnecessary to give here, 

 though there may be mentioned the sodic pyroxenes, which contain 

 much iron and little if any magnesia, and the potassic micas, which 



n* 2 o 



w 



Fig. 1. — Relation of Na and Fe to K and Mg. 



generally contain more magnesia than iron along with the potash. 

 Study of many analyses of igneous rocks also brings this relation 

 out very clearly, and it is expressed in the above figure (fig. 1) pub- 

 lished some years ago. 26 In this the abscissas represent the relative 

 amounts of soda and potash, and the ordinates those of iron oxide 

 and magnesia. The general drift of high soda coincident with high 

 iron, and conversely of high potash with high magnesia (though 

 such rocks are comparatively few), is clearly shown, and, as the data 

 are derived from numerous analyses, and are substantiated by many 

 others more recently made, the general " drift " may be considered as 

 fairly well established. That the points fall in a rather broad zone, 



* Washington, H. S., Proc. Natl. Acad. Sci., i, p. 574, 1915, 



