54 Pratt — Origin of the Corundum associated with 



increase of magnesia in the vein material as the dunite is 

 approached, and that there is an approximate decrease in the 

 percentage of alumina. 



From these analyses, the series from the section across the 

 vein are divided into three groups, aluminum silicates, 

 aluminum-magnesium silicates, and magnesium silicates. The 

 middle term of this chemical series is also the middle number 

 of the held series. He regards the corundum as an accessory 

 mineral, frequently not being found at all in the vein, and 

 sometimes but in small amount, and therefore to be considered 

 as the result of a certain balance between the magnesium and 

 aluminum silicates, which have by their union produced the 

 chlorite and vermiculites. 



In describing the occurrence of corundum at Chester County, 

 Pa., Mr. J. P. Lesley* says that it seems difficult to imagine 

 its excessive compact hardness as produced in any other way 

 than by heat. 



J. V. Lewis'f report on " Corundum and the Basic Magne- 

 sian Rock in North Carolina " is mainly a report of field 

 observations and does not take up any discussion of facts bear- 

 ing on the origin of the corundum ; though many points of 

 interest and importance are brought out regarding the altera- 

 tions of the dunite and of the character of the vein and vein 

 materials and also concerning the alteration of corundum. 



The corundum deposits of Georgia are similar, almost iden- 

 tical, with those in North Carolina. F. P. King,:}: in his report 

 on the corundum deposits of Georgia, which is principally de- 

 voted to the description of localities, occurrences and varieties 

 of corundum, etc. comes to the conclusion that corundum is an 

 accessory mineral and that its presence is occasioned by an 

 excess of alumina present in the rock masses, " chrysolite, 

 gneiss and hornblende gneiss." This is explained by the 

 alteration of these rocks into magnesium silicates, alkaline salts 

 and ferro-silicates, which in conjunction with carbonic acid of 

 percolating waters would dissolve the combined alumina, and 

 produce on crystallization all the minerals associated with co- 

 rundum ; and when the alumina is in excess this would produce 

 corundum itself. 



As has been shown above, a number of theories concerning 

 the origin of the corundum have been advanced. With one 

 exception, however, they are all prior to the numerous and 

 elaborate experiments that have been made by different inves- 

 tigators on the solubility of alumina in a molten basic glass 

 and the separating out of corundum and spinel crystals from 



* Penn. Geol. Survey, C4, p. 352, 1883. 

 f N. C. Geol. Survey, Bulletin No. 11. 

 % Georgia Geol. Survey, Bulletin 2. 



