Origin of the Angiospermous Flora of South Africa. 341 



exchange types of animals and plants with Tropical America- 

 up to early Tertiary times, and that in this way a second con- 

 nection between the South African Flora and the Flora of America 

 was established. 



In formulating this theory I was aware of the great objections- 

 which Wallace has urged against the assumption that the distribu- 

 tion of sea and land has been radically changed during the Tertiary 

 and Secondary periods, and I have even, some ten or twelve years 

 ago, expressed the same view under the influence of his writings. 

 I well remember his warning that it is " so easy and pleasant to 

 speculate on former changes of land and sea with which to cut the 

 Gordian knot offered by anomalies of distribution, that we still con- 

 tinually meet with suggestions of former continents stretching in 

 every direction across the deepest oceans, in order to explain the 

 presence in remote parts of the globe of the same genera even of 

 plants or of insects — organisms which possess such exceptional 

 facilities both for terrestrial, aerial, and oceanic transport, and of 

 whose distribution in early geological periods we generally know 

 little or nothing." 



But the facts to which I referred, and others which might have^ 

 been added, seem imperatively to call for the theory which I have 

 just mentioned. The survival of an ancient Flora of the Southern 

 Hemisphere seems to be generally admitted now (see, e.g., Diels, 1906) 

 and its existence cannot be explained with the aid of existing land- 

 connections, nor are many of the elements of this Flora particularly 

 well adapted for easy terrestrial, aerial, and oceanic transport, and 

 objections which may be raised against arguments derived from the 

 distribution of insects cannot be raised when we consider animals 

 such as Perijjatus and earthworms. 



There seems to be no way of escaping the conclusion that in late 

 mesozoic times, possibly up to the Cretaceous period, there was still 

 a more or less direct land-connection between South Africa and 

 Australia. H. F. Blandford (1875, p. 540) first clearly stated that in 

 Permian times India, South Africa, and Australia were connected by 

 an Indo-oceanic continent. There is no need to inquire here into 

 the evidence of this hypothesis ; it may suffice to say that, apart' 

 from purely geological considerations, it rests chiefly on the known 

 distribution of the Glossopteris-'Flor^ (compare Arber, 1905). The 

 general statement that such a continent existed in Permian times is 

 now pretty universally admitted. Direct land-connection between 

 South Africa, Madagascar, and India through Ceylon is admitted to 

 have persisted right into the middle of the Tertiary period (Stanley 

 Gardiner, 1906, p. 321, and Gadow in Stanley Gardiner, 1906, p. 469).. 



