FOYAITE-IJOLITE SERIES OF MAGNET COVE 663 



In my former paper I suggested as an explanation of the 

 exceptional character of the Magnet Cove and Umptek lac- 

 coliths, in having the borders more acid than the centers, that 

 the arrangement depended on the general chemical character of 

 the undifferentiated magma. The process of differentiation was 

 conceived to be, at least for such small bodies, in great part, a 

 sort of fractional crystallization, the magma being regarded as 

 a solution, so that, in accordance with the laws of cooling solu- 

 tions, the solvent {i. e., the portion present in excess) crystal- 

 lizes out first around the borders on cooling of the mass. 



From what has been learned of the composition of the 

 magma, it is evident that, even though low in silica, it was 

 originally of a decidedly leucocratic character. In other words, 

 the potential feldspathic and feldspathoidal constituents pre- 

 dominated very largely over the calco-ferromagnesian. This is 

 seen plainly from the relative weights of the spherical shells, but 

 even the more basic composition derived from the curves shows 

 the same thing. Thus the composition with 47.24 per cent, of 

 SiOg may be obtained approximately by several different mix- 

 tures of all or some of the types analyzed, but in every case it 

 necessitates taking from six-tenths to eight-tenths of foyaite, or 

 foyaite and arkite. It seems scarcely necessary to give these 

 calculations, which are purely empirical. The same composition 

 may also be reduced to mineralogical composition in several 

 ways, according to the assumptions made, but here, also, we get 

 about two-thirds of leucocratic minerals. 



The original body of magma, then, at Magnet Cove was, 

 notwithstanding its low silica, decidedly leucocratic, as demanded 

 by the theory, so that the alumina and alkalis, with the proper 

 amount of silica for the formation of feldspar and feldspathoids, 

 playing the role of solvent, would crystallize first, and hence 

 form the outer portion of the mass. 



The latest paper by Weed and Pirrson, already cited, is of 

 great interest in this connection. Here it is shown conclusively 

 that in the well-dissected Shonkin Sag laccolith the outer mel- 

 anocratic shonkinite is present in enormously greater quantity 



