556 N. L. BOWEN 



The efifect was indeed very similar, for the sillimanite-rich inclusions 

 are found to be changed by the magma to masses rich in corundmn, 

 spinel, and cordierite, with other related minerals. It is inclusions 

 that have been thus affected that constitute the emery deposits. 

 They represent the reaction between the magma and inclusions 

 arrested midway. No doubt many inclusions completely disap- 

 peared, becoming an integral part of the igneous rock in virtue of 

 reactions involving adjustment of composition of existing phases. 

 (See p. 549.) The feldspathic emery and the noritic emery may, 

 from this point of view, be regarded as inclusions approaching their 

 final disappearance. 



These transient states in which inclusions may be very rich in 

 certain substances are no doubt of some importance in differentia- 

 tion. Localized masses in process of reaction with the general mass 

 may move, say in response to gravity, and their accumulation may 

 give rise to bodies rich in minerals formed during the reactions. 

 All up and down the Appalachian Mountain system of Eastern 

 North America there are intrusive masses showing ultrabasic dif- 

 ferentiates with dunite as the extreme and with associated perido- 

 tites and pyroxenites frequently rich in rhombic pyroxene, saxonite, 

 websterite and enstatolite itself. These are usually, perhaps always, 

 intrusive into slates and mica schists that were originally aluminous 

 sediments, the Farnham slates of Quebec, the Savoy and Rowe 

 schists of New England, the Manhattan of New York, the Wissa- 

 hickon of Pennsylvania and Maryland, and the Carolina gneiss of 

 the Southern States. These have perhaps had an important influ- 

 ence in emphasizing differentiates of the types mentioned above. 

 There are, moreover, corundum deposits, either as emery or in 

 purer forms, in frequent association with these ultra-basic rocks 

 and in some cases the origin and accmnulation of corundum may be 

 referred to processes outlined above. Gordon has demonstrated a 

 different origin for some of them,' but the action here described 

 seems unquestionable for the Cortlandt emery and the gangue 

 minerals of some of the Carolina deposits strongly suggest a similar 

 origin. These gangue minerals are basic plagioclase, sillimanite and 



' Formation by reaction between pegmatite and serpentine. S. G. Gordon, 

 Proc. Acad. Natural Sci. Philadelphia, Part I (192 1), p. 169. 



