92 ANNALS NEW YORK ACADEMY OF SCIENCES 



As an example of the imperfection of the fossil record may be given 

 Cryptobranclius, an amphibian, living in Japan and the Mississippi Val- 

 ley. The only known fossil relative is Andrias, Scheuzer's Homo diluvii 

 testis of the Miocene of Europe. Its distribution, in a way, is like that 

 required for a northern dispersal of the Gonclwana flora. As an example 

 of the imperfection of exploration may be given the two new marine 

 horizons in the Conemaugh series of western Pennsylvania, found by 

 Raymond in a region which has been explored by many geologists. As I 

 have already stated, however, there is little hope of finding the necessary 

 antecedent types of the Gangamopteris flora in North America, because 

 the necessary highland deposits which produced this stunted flora have 

 apparently disappeared, or have not yet been found. 



The idea of a continuous Gondwana Land has little or no support 

 other than indecisive statistical data derived from the distribution of 

 living and extinct animals and plants. Furthermore, it appears to me 

 that even such a vast amount of indecisive data which admit of a variety 

 •of interpretations can never outweigh the fact that the Permian reptiles, 

 the ooze and shark teeth, etc., of the great ocean depths, such as exist 

 hetween Africa, Australia and the Antarctic islands, the geology of the 

 Brazilian coast, the absence of giant east and west trend lines of South 

 America and the Antarctic islands offer strong positive evidence that a 

 continuous Gondwana Land did not exist. Until this vast array of posi- 

 tive evidence in favor of the persistence of the great ocean depths and 

 the continental shelves has been satisfactorily accounted for, it is just as 

 inconceivable to explain the existence of a continuous Gondwana Land 

 as to conceive either a northern distribution of the Gangamopteris flora 

 or that this flora developed orthogenetically from some unknown north- 

 ern ancestors which evolved from the Devonian cosmopolitan flora. This 

 is all the more true when we have to acknowledge that the actual point of 

 family origin of the Gangamopteris flora is unknown. 



Before, however, the reader forms an opinion concerning the existence 

 of a continuous Gondwana Land, as is indicated by the distribution of 

 the Gondwana flora, he should consider still two other possible means of 

 distribution of the Permian plants which require no continuous Gond- 

 wana Land. 



The distribution of the Gondwana flora by strong winds like those 

 which blow dust from the pampas of Argentina to Africa is possible but 

 perhaps not probable. The Gangamopteris plants are now generally 

 considered to be seed-bearing, but in most cases no seeds are known; 

 hence wind distribution would be out of the question, unless the seeds 

 were of very light-winged type. Then they might have been blown for 



