HASEMAN, GEOGRAPHICAL D I ST RIB UTION IN SO UTH AMERICA 89 



widely distributed but could also remain almost identical in the various 

 corners of the earth. 



Even if it be granted that a vast homogeneous Gondwana Land en- 

 tirely covered the southern hemisphere, this alone would hot explain. why 

 so many species of the typical lower Gondwana flora remained identical 

 during the greater part of tbe Permian epoch. There is positive evidence 

 from the intercalated Permian of Brazil and other Gondwana forma- 

 tions that these regions were not continuous during the entire Permian 

 epoch, and yet the flora remained identical. Hence there appears to be 

 no need for a ready and wholesale exchange of this flora, because it re- 

 mained unchanged during the most of the Permian epoch, and therefore 

 any accidental distribution of each genus and for one time only Avould be 

 adequate. In this connection, it is to be remembered that neither con- 

 tinuous land connections nor barriers nor isolation causes species (unless 

 there are only a few individuals) to evolve new forms or to remain fixed, 

 but the action of the environmental complexes on the ancestral forms, 

 which had a certain composition, and hybridization will change the spe- 

 cies. The identity of the Gangamopteris flora in such remote parts of 

 the earth as already noted may be due to the action of similar environ- 

 mental complexes upon the ancestral stock of each group of this flora. 

 It is then even possible, but not probable, that the Gangamopteris flora 

 is only an environmentally changed form of the cosmopolitan flora which 

 could not exist as such in the early Gondwana environment. At least the 

 sporaginous members of the cosmopolitan flora could have migrated into 

 the Gondwana environments, if they had been able to thrive under such 

 conditions. 



What the direct antecedent types of the different groups of the Gond- 

 wana flora were is not definitely known. David White considers that 

 Nceggerathiopis is probably of Cordaitalean origin and Gangamopteris 

 has a common origin with the neuropterid group of Cycadofilices, to 

 which the genus Glossopteris also is related; yet both are far removed 

 from the known antecedent type of the northern hemisphere. What this 

 direct ancestral stock was and where it originated are questions which 

 must be answered before we can definitely hypothesize lines of dispersion. 



I have repeatedly pointed out in the case of fishes that direct connec- 

 tions are not necessary for the production of identical species and that 

 the greater number of species in a given region is no evidence that the 

 family originated there. I will repeat here again that the fact that the 

 gangamopterids are found only in the southern hemisphere is not con- 

 clusive evidence that their ancestral stock either originated there or that 

 all of the existing Gondwana formations were directly continuous. 



