R. D. Oldham — Probable Changes of Latitude. 305 



reasoning would be derived from the doctrine of permanence of con- 

 tinents. It is, however, by no means inconceivable that the two 

 hypotheses might be quite consistent ; were the differences between 

 the continental and oceanic areas entirely due to differences in the 

 structure of the crust, the latter might shift its position relative to 

 the core to any extent without interfering with the relative positions 

 and forms of the continental and oceanic areas. 



But is this doctrine so well established that it can be used as an 

 argument against any hypothesis which is fairly supported by known 

 facts ? I think not. It is unnecessary to refer to the fact that the 

 " Oceanic " island of South Georgia has been found to consist of 

 clay-slate, 2 and not of volcanic rocks, as ex hypothesi it should, for 

 there are certain peculiarities, in the palaeontology of India and 

 South Africa, which indicate the former existence of direct land com- 

 munication between the two countries. This was first pointed out 

 by Mr. H. F. Blanford, whose paper is somewhat contemptuously 

 dismissed by Mr. Wallace with the remark that " the notion that a 

 similarity of the productions of widely separated continents at any 

 past epoch is only to be explained by the existence of a direct land 

 connexion, is entirely opposed to all that we know of the wide and 

 varying distribution of all types at different periods, and is no less 

 opposed to what is now known of the general permanency of the 

 great continental and oceanic areas." 2 This, however, implies a mis- 

 conception of the nature of the evidence, which is far from being 

 based merely on a " similarity of the productions " of the two 

 countries. There are in Africa two distinct floras of different ages ; 

 one of these, that of the Beaufort beds, has a flora consisting of five 

 distinct species, of which one is identical, two are closely allied to, 

 forms found in the Damuda beds of India, and if we accept Dr. 

 Feistmantel's 3 opinion, all belong to Damuda genera ; associated 

 with these plants is an extensive and peculiar reptilian fauna, of 

 which the most prominent genus is Dicynodon, a genus at present 

 unknown elsewhere except from the Panchet beds which overlie the 

 Damudas, and another form, Micropholis Stowei, is a near ally of 

 JBrachiops laticeps from the Kamthi beds of Mangli. 4 



At a higher horizon in South Africa, in the Uitenhage formation, 

 there is a flora consisting of 12 distinct forms, all generically different 

 from any of the Beaufort species. Of these, one is identical with, 

 four are closely allied to, species from the flora of the Rajmahal group 

 in India. The difference between the flora of the Rajmahal group 

 and of the Danudas is almost as great as in the case of the two cor- 

 responding African groups, for there are only three species in the 

 Rajmahal flora which are in any way allied to any of the Damuda 

 plants. We have then a close and continuous similarity between the 

 fauna and flora of two countries lasting through a period long enough 

 to allow of a complete specific, and almost complete generic change. 



A similar but even more conclusive argument may be derived 



1 Nature, March 27, 1884, p. 509 ; Geol. Mag. Dec. III. Vol. I. p. 225 (1884). 



2 Island Life, p. 398. 3 Palseontographica, 1878, p. 114. 

 4 Manual of the Geology of India, vol. i. p. 123. 



DECADE III. VOL. III. NO. VII. 20 



