THE KARROO SYSTEM 235 



well-defined species as the hard parts of marine animals 

 that are used for correlation purposes in the marine 

 beds of Carboniferous and Permian age. 



The Witteberg beds which were laid down immedi- 

 ately before the lowest beds of the Karroo formation in 

 the south of the Colony afford very little help, for they 

 are singularly barren of fossils ; the few casts of plant 

 stems which they have yielded are of no real use in 

 determining the age of the beds in which they occur. 



The important fact in attempting to correlate the 

 lower beds of the Karroo formation is that a peculiar 

 flora, called the Glossopteris flora l from the name of its 

 most widespread genus, occurs in them and in associa- 

 tion with the Dwyka glacial beds. In India, South 

 America, and the Australian region the same kind of 

 plants have been found associated with a similar glacial 

 deposit. 



In India 2 the glacial boulder-beds (Talcher) lie uncon- 

 formably upon old.er rocks in the Peninsula and are 

 followed by shales and sandstones containing the Glos- 

 sopteris flora, as in the case of the Dwyka tillite in the 

 north of Cape Colony ; in the Salt Kange similar boulder- 

 beds lie unconformably upon older rocks, but in this 

 case they are overlain by strata containing Permian or 

 Upper Carboniferous marine fossils, some of which are 

 identical with Australian species; this marine fauna 

 gives place to a group of unfossiliferous rocks which are 



1 For a very useful and complete account of this flora see the British 

 Museum Catalogue of the Glossopteris Flora, by E. A. Newell Arber, 

 London, 1905. Dr. David White prefers the name Gangamopteris Flora. 



2 Holland, T. H., in the Imperial Gazetteer of India, vol. i., 1907, 

 pp. 70-71. 



