3W FUliMKll LAND COXXKCTIUXS. 



Africa : except far the regions just mentioned, tiie genus Hyla 

 itself is almost world-wide. This is exactly as we should expect 

 of a moderately recent group on our theory of a centre of dispersal 

 for the Procoela in or through Antarctica. There is one peculiar 

 record which 1 do not profess to understand, namely, a species 

 (jf Hyla in Abyssinia. Of tliis species, Mr. Noble says: — " It 

 has very little in common with any Asiatic Hyla and but little 

 more with any Australian or East Indian species. It agrees 

 entirely with the maxima group of species from northern tSouth 

 America." It does not show close affinities with the species 

 found in Syria and Palestine. The interpretation offered is as 

 follows: — " It is very possible that the peculiar arrangement of 

 the vomerine teeth characteristic of the maxima group (South 

 American) could have had a parallel evolution in two such widely 

 separated areas as Abyssinia and South America ' ' ; and further, 

 " If this specimen actually came from Abyssinia, and we have 

 no reason to doubt it, the genus Hyla must have existed in 

 Africa for a very long time since H. waclici shows no close 

 affinity to the eastern representatives of the genus." Thus we 

 see how easy it is to explain the most intimate relationships 

 between animals so remotely situated if only we are willing to 

 make use of parallel evolution. My own inchnation, in this case, 

 is to suspect the premises, especially as the species is only 

 known from the type specimen. 



Conclusion. 



The distribution data just considered seem most readily 

 explained on the assumption of direct land connections between 

 Africa, South America and Antarctica. On evidence which is 

 similar, yet certainly more complete, rests the accepted theory 

 of the Behring Straits connection. Yet, in view of the latter, 

 some will prefer to assume a former world-wide dispersal followed 

 by extinction over large areas : others will virge the claims of 

 convergent or parallel evolution to ex])lain particular resem- 

 blances : failing such explanations, accidental agencies of dispersal 

 isuch as floating rafts niay sometimes be invoked. 



The credibility of the hypothetical Antarctic connection 

 depends on the improbability of the total extinction of an original 

 cosmopolitan stock throughout the world, leaving persisting 

 remnants only in the extreme south. If it could be shown that 

 the southern " remnants " are represented in the northern 

 hemisphere by more highly-evolved members of the same stock, 

 a northern centre of dispersal might be the most reasonable 

 explanation. But such evidence is lacking in the cases we have 

 considered. The extinction of all the northern elements of a 

 fauna still living in the tropics is a known fact, being attributed 

 to the frigid conditions which culminated in an extensive northern 

 glaciation. But such conditions did not extend to the tropics at 

 high altitudes. 



The most important evidence for a direct trans-Atlantic con- 

 nection between Africa and South America lies in the fact that, 

 in several instances, the distribution in South America, Africa and 

 Southern Asia is exactly what we should expect in a single 



