ORIGIN BY DIFFERENTIATION 459 



have contained at least a very considerable percentage of liquid of suffi- 

 cient fluidity to allow the crystals to sink through, and this residual 

 magma is represented by the present anorthosite, which was at least 

 largely molten as such. Also it is clear that my hypothesis, which requires 

 sinking of femic minerals only, in not very great quantities, through a 

 moderate thickness of magma, is much more plausible than Bowen's, 

 which requires sinking of femic crystals through a much greater thickness 

 of magma, followed by a long process of wholesale settling of plagioclase 

 crystals through a great thickness of the remaining magma. Could the 

 magma have possessed sufficient fluidity long enough to have permitted 

 this repeated process of crystal sinking, especially in view of the fact that 

 no evidence points to more than a moderate degree of fluidity of the 

 original gabbroid laccolithic magma ? 



Bo wen strongly emphasizes two points : First, the failure to find anor- 

 thosite which consists of almost pure plagioclase in dike form, and, second, 

 the failure to find an efl'usive equivalent of such anorthosite. The first 

 statement needs some modification, as I have already shown, but I believe 

 Bowen is correct in emphasizing these points. I do not, however, agree 

 with Bowen when he maintains that these facts prove the anorthosite 

 never to have been, to a considerable degree at least, molten as such. 

 When we consider that the original gabbroid magma in the Adirondack 

 region was probably not at a high enough temperature to be very free- 

 flowing ; that the marginal chilled border f acies relatively soon developed 

 to protect the still liquid interior portion, and that the anorthosite was 

 not intruded as such, but w^as slowly formed by differentiation in situ — 

 that is to say, by crystallization of the magma left after and during the 

 precipitation of many of the femic minerals — it is easy to understand 

 why the purer Marcy anorthosite has never been found in the form of 

 dikes or small intrusive tongues. 



But the fact should not be overlooked that some dikes of highly feld- 

 spathic border f acies anorthosite (Whiteface type) do occur in the Adi- 

 rondack region. In such cases, evidently, enough differentiation, prob- 

 ably by settling of femic minerals, must have taken place in the upper 

 levels of the magmatic chamber to give rise to portions of magma possess- 

 ing sufficient fluidity to be forced into the country rock in the form of 

 dikes. There is, however, no reason to think that such dikes ever reached 

 the surface of the earth, because the rather pure anorthosite magma thus 

 formed did not, no doubt, possess sufficient fluidity long enough to pene- 

 trate very far into the thick country rock over the laccolithic magma. If 

 anorthosites in general have originated by some such process as that here 



