FRAGMENTATION BARRELL 297 



several thousand feet and forming the Western Ghats. The flows 

 and ash beds reach a maximum thickness of as much as 10,000 feet 

 along the coast of Bombay. It would appear as if they once extended 

 far seaward (perhaps across to Arabia), but that this western por- 

 tion had broken off and sunk into the Arabian Sea. The date of 

 the basaltic outpourings is fixed as latest Mesozoic, transitional to 

 the Eocene. 



Far to the south, in the early Eocene, Madagascar is thought 

 to have been still connected Avith India, forming the hypothetic 

 land of Lemuria, the eastern portion of Gondwana. The evidence 

 here lies in the nature of the Madagascan fauna, which contains a 

 primitive assemblage of mammals and is devoid of that more power- 

 ful and advanced placental fauna whose evolution dates from the 

 later Eocene. The foundering of Lemuria and the spread of the 

 Indian Ocean westward and northward may therefore be connected 

 with the basic igneous intrusions of the previous geologic period, 

 of which the Deccan traps are remnants. 



On the other hand, the great Columbia Plateau of basaltic 

 rocks in Idaho, Washington, and Oregon, which rivals the Deccan 

 traps in magnitude, is not known to be connected with any adjacent 

 subsiding area. The presence of basaltic flows does not in itself 

 prove that there has been a rise of new basic magmas in great 

 volume into the outer crust. All great cycles of igneous activity 

 begin normally with the outflow of basaltic or the slightly more 

 siliceous andesitic lavas. Throughout the Cordilleran belt the great 

 intrusive masses which followed the cxtrusives are in part exposed by 

 later erosion and are seen to be nearly as siliceous as the true granites. 

 They are therefore called granocliorites. Their specific gravity is 

 close to that of granite and very different from that of the gabbros. 

 Hence the Columbia Plateau basalts are seemingly related to the 

 granodiorite magmas and are probably underlain deeper in the crust 

 by them. 



The causes of continental fragmentation which have so far been 

 suggested involve several postulates which should now be brought 

 clearly into view. 



First. There is involved the conception of an initial granitic crust 

 grading down through an intermediate or dioritic one into a denser 

 basaltic shell. This is of course not open to direct proof, but is an 

 inference reached by a number of petrologists who have dealt with 

 the profounder aspects of their subject.^** This hypothesis of the 

 layered nature of the earth is also in accordance with a number of 

 lines of inference developed in the present work. 



'' See especially the comprehensive work by R. A. Daly, Igneous Rocks and Their Origin, 

 Ch. VIII, 1914. 



