14 ANNALS NEW YORK ACADEMY OF SCIENCES 



My studies on the distribution of South American animals have also 

 led me to place more emphasis on negative evidence than is usually 

 granted by most writers. This difference of view is primarily due to the 

 fact that a specialist in any one group of animals places too much weight 

 on his positive evidence. Such emphasis at first sight appears to be 

 absolutely correct, but on closer analysis it is usually contradicted by 

 positive evidence from other groups of animals. For example, one au- 

 thor working on Crustacea and mollusca finds that the alleged connection 

 between South America and Africa had already disappeared in the early 

 Cretaceous. If, however, we consider other groups, we find evidence 

 which does not confirm this view. Thus, the affinity existing between 

 the South American and African characinid and cichlid fishes is as close 

 as that of the mollusca, yet there is no evidence that either of these 

 families of fishes existed during the early Cretaceous. 



In my conclusions, therefore, I have been led to balance the positive 

 and negative evidence in the cases of many different groups. This bal- 

 ancing has been attempted not only by considering long lists of species, 

 but by taking into due account the influence of various environments on 

 the ancestral stocks (whose points of origin are usually unknown). 



In fact, we should not, in such considerations, lose sight of the fact 

 that our knowledge of the existing species of any group of South Ameri- 

 can and African animals is still very imperfect. Many species are still 

 to be caught, many are exact synonyms and many are without doubt 

 local somatic changes which are not always inherited. Therefore any 

 positive evidence derived from such lists is, in my opinion, entirely in- 

 adequate to warrant the reconstruction of the earth's surface, unless 

 supported by strong geological evidence. This is all the more true when 

 there are other means of distribution which do not involve great topo- 

 graphical changes. 



In point of fact, the inadequacy of the existing data is at once re- 

 flected by the number of alleged land-bridges and seas required during 

 various past geological epochs. Each writer has constructed his new set 

 of barriers, seas and lancl-bridges in his effort to explain the distribution' 

 of the fauna or flora in which he is recognized as an authority. 



Notwithstanding the diversity of views concerning the time of exist- 

 ence and the location of the alleged lancl-bridges to the South xYmerican 

 continent, we may roughly consider them under the following two 

 groups : 



1. The Gondwana Land of Suess and others — a late paleozoic conti- 

 nent traversing the greater part of the southern hemisphere and connect- 

 ing India, Australia, South Africa and South America. The last re- 

 mains of this old land-mass connecting Africa and South America have 



