MATTHEW, CLIMATE AND EVOLUTION 191 



If the distribution of animals be interpreted along the lines here advo- 

 cated, there is no occasion for a Gondwana Land even in the Paleozoic. 

 But it is chiefly as affecting Mesozoic or Cenozoic dispersal that we are 

 here concerned with it. One may summarize the arguments for it by 

 saying that a considerable number of groups of animals and plants which 

 are absent in the northern world, either living or fossil, are found in the 

 southern continents and some of them in certain oceanic islands as well. 

 Most of the groups are unknown or almost unknown as fossils; those 

 which have any considerable fossil record are steadily being eliminated 

 from the list by the progress of discovery, showing that they or their an- 

 cestors did formerly inhabit the northern world. The remaining groups 

 agree with those southern faunal groups which have admittedly come 

 from the north, in being of primitive and archaic type and in that their 

 representatives in the different southern regions are but distantly related, 

 the remoteness being in a very direct proportion to the present isolation 

 of the region. 



There are a few instances of exclusively southern types closely related 

 (e. g., Galaxids) ; but, although they have been cited in corroboration of 

 the evidence from the groups above mentioned, they are in fact, if thus 

 interpreted, directly contradictory. For the distant relations of the one 

 series is interpreted to mean a very ancient connection, but isolation 

 since ; while the other series would indicate a very recent connection and 

 earlier isolation. The explanation here lies not in a northern ancestry, 

 but that the ocean does not form an impassable barrier to their dispersal. 

 This has been proven in the case of Galaxias; it is probably the explana- 

 tion of all similar distributions. 



The relations of the Glossopteris flora are a different and far more com- 

 plex problem of distribution. The clue to its interpretation lies perhaps 

 in its association with Permian glaciation; but it is outside the limits of 

 the present essay and will not be discussed here. 



Eegional Correlation 



The geological correlation of widely distant formations is so intimately 

 bound up with problems of geographical dispersal and migration that the 

 two series of problems must needs be studied and solved together. We 

 cannot arrive at a correct understanding of the history and causes of the 

 geographical distribution of animals, present and past, without correct 

 correlation of the geological succession in different regions. IS^or have 

 we, up to the present time, any reliable methods of exact correlation in 

 widely distant regions except the comparison of fauna and a considera- 



