406 PROF. P. MARSHALL OX THE SEQUENCE OF [Dec. I9M, 



same as in the island of Tahiti, where there is a similar assem- 

 blage of alkaline and basic lavas. 



(14) In Tahiti erosion has now exposed the reservoir, and the 

 dominant type of rock seen in it is essexite, although the lavas 

 of the island are remarkably free from the hornblende which is 

 the most conspicuous mineral of the essexite. 



-•(15) It is suggested that the frequent occurrence of alkaline 

 and basic rocks at Moorea, Huaheine, Eaiatea, and Rarotonga is 

 due to differentiation from an essexitic magma. 



EXPLANATION OF PLATES LILT & LIV. 



Plate Lin. 



Diagrams showing the chemical composition of the lavas and the molecular 

 proportions of the oxides in the series of rocks. 



Plate LIV. 



Diagram showing the specific gravities ; and variation- diagram of the molecular 

 proportions of the metallic oxides compared with silica, which in each 

 case equals 100. 



Discussiox. 



Dr. J. W. Evans expressed his pleasure at this valuable addition 

 to the evidence bearing on the subject of the differentiation of 

 igneous magmas, although he agreed with Prof. Brogger that the 

 variations among volcanic rocks did not afford such definite informa- 

 tion as the order of intrusion among intrusive rocks, of different 

 composition but proceeding from a common source. He remarked 

 on the predominance of basaltic lavas in the succession described, 

 and expressed his opinion that the magma from which the intrusions 

 proceeded must have had the composition of a basalt somewhat 

 rich in soda. In a communication to the Twelfth International 

 Geological Congress at Toronto, he had expressed the opinion that 

 in the presence of the elements of water in considerable amount, a 

 clear differentiation was possible into a magma corresponding to a 

 granite in composition (except that it contained a much larger 

 amount of water), and a basic magma with little water. Where water 

 was deficient, the separation was less complete, and the basic magma 

 was richer in soda. At the same time, there was evidence that an 

 alkaline magma tended to separate into layers of different com- 

 position when cooled to a temperature but little above that at 

 which crystallization commenced. In the present instance, this 

 appeared to have resulted in the separation of a comparatively small 

 amount of more acid rocks from the main magma. 



He deprecated the use of very special terms like kaiwekite, 

 side by side with general terms such as basalt. It was regrettable 

 that something like the distinction made in biology between family, 

 generic, and specific names could not be introduced into petrology. 



