196 ^VNNUAL REPORT SMITHSONIAN INSTITUTION, 1936 



plants Glossopteris and Gangamopteris. Associated Avith tliem are 

 peculiar groups of fossil shells and fossil amphibia and reptiles. 

 Similar i-ocks, with similar associations and contents, in Peninsular 

 India have been named the Gondwana formation. Comparable for- 

 mations also occupy large regions in Australia, Tasmania, and New 

 Zealand, in Madagascar, in the Falkland Islands and Brazil, and in 

 Antarctica. 



The correspondence between these areas is so close that Suess 

 supposed they must at that date have been connected together by 

 lands, now^ sunk beneath the sea, and he named the continent thus 

 formed Gondwanaland after the Indian occurrences. The break-up 

 of this land can be followed from a study of the rocks, and it was a 

 slow^ process, its steps occupying much of Mesozoic time. Dr. A. L. 

 du Toit's comparison of South African rocks with those of Brazil 

 and elsewhere in South America favors even a closer union than this 

 between the units now scattered. 



One of the most remarkable features shown by these rocks in all 

 the areas mentioned, but to varying extents, is the presence of con- 

 glomerates made of far-traveled boulders, scratched like those borne 

 by the modern ice sheets of Greenland and the Antarctic, associated 

 w^ith other deposits of a glacial nature, and often resting upon 

 typical glaciated surfaces. There is no possible escape from the con- 

 clusion that these areas, now situated in or near the Tropics, suffered 

 an intense glaciation. This was not a case of mere alpine glaciers, 

 for the land was of low relief and not far removed from sea level, but 

 of extensive ice sheets on a far larger scale than the glaciation of the 

 northern parts of the new and old worlds in the Pleistocene ice age. 

 I have never seen any geological evidence more impressive or con- 

 vincing than that displayed at Nooitgedacht, near Kimberley ; while 

 the illustrations and other evidence published by David and Howchin 

 from Australia are equally striking. 



Du Toit's work on these glacial deposits brings out two remaikable 

 facts; first, that the movement of the ice was southerly, polfiward 

 and away from the Equator, the opposite to what would be expected, 

 and to the direction of the Pleistocene ice movement; secondly, 

 that the ice in Natal invaded the land from what is now sea to the 

 northeast. 



When it is realized that at this period there is no evidence of glacial 

 action in northern Europe or America, but a climate in which grew 

 the vegetation that formed the coal seams of our Coal Measures, 

 it is clear that we are not dealing with any general refrigeration of 

 the globe, even if that would produce such widespread glaciation; 

 we are face to face with a special glaciation of Gondwanaland. 



