FOYAITE-IJOLITE SERIES OF MACAET COVE 669 



tion, /. r., solidification of the magma, and any such process is, 

 of course, impossible in a solid mass of rock. 



This difficulty, however, does not seem to be insuperable. 

 The evolution of heat on the solidification of molten lava is a 

 well-known phenomena, having been actually observed, and the 

 same has been experimentally verified, and the amount of heat 

 evolved on the solidification of diabase has been determined by 

 Barus.' Therefore the crystallization of minerals from a molten 

 magma is an exothermic change. 



It is therefore conceivable that the solidification of a lac- 

 colithic mass may give rise to sufficient heat to remelt por- 

 tions of it, which might easily remain liquid long enough for a 

 secondary differentiation to take place. That this was actually 

 the case at Magnet Cove cannot be definitely shown, but it is at 

 least an explanation which has a certain degree of probability 

 in its favor. 



In regard to problems of interpolation and extrapolation, by 

 which Pirsson obtained such close agreement between calculated 

 and observed chemical composition, we are not in a favorable 

 position, since the analyses which I have made exhaust the 

 known main types of abyssal rocks at the locality. A compari- 

 son of the analyses of the dike rocks given by Williams with 

 the curves in Diagram 3 confirms the supposition expressed in 

 my former paper that the tinguaites and the nephelite-porphyry 

 are aschistic, and that the fourchites and ouachitites like the 

 covite, are diaschistic. 



It was expected that the analyses of the Fourche Mountain 

 pulaskites would fall in with the curves as laid down. But, 

 determining the abscissal position by silica, it is found that 

 AI3O3, NagO, and K^O are below, and (FeO), MgO and CaO 

 are above, what would be their "normal" positions, as will be 

 seen on reference to the diagram. 



Although these rocks are unquestionably derived from the 

 same general magma, yet, as their distance from Magnet Cove 

 is about forty miles, it is clear that we need not be surprised to 



'C. Barus, Am. Jour. Sci., Vol. XLIII, p. 56, 1892. 



