MATTHEW, CLIMATE AND EVOLUTION 307 



cies and faunas independently created. Such a view was held by Agassiz 

 and most of his predecessors, but it is unnecessary to consider it in the 

 present state of scientific belief. 



If, on the other hand, we accept the belief that the successive species 

 of each phylum are genetically related, how are we to explain the fact 

 that these phyla are usually approximate and not direct, and that where 

 the evidence is most complete, the fact that they are not in a direct line 

 of structural evolution stands out most clearly. Take for example the 

 ancestry of the horse, the most striking, easily recognizable, widely known 

 and thoroughly studied illustration of mammalian evolution. It was 

 possible, when the "documents" were few and imperfect, to trace a sup- 

 posedly direct line of ancestry through European predecessors. Later, 

 when the fossil fields of the western United States were first explored, a 

 much more direct line of ancestry was found in this country, and the 

 European series was recognized as not being the direct line. But the 

 further progress of exploration in America, and the discovery of complete 

 skeletons of the supposed ancestral stages known at first only from frag- 

 mentary specimens, has demonstrated that this line too is an indirect and 

 approximate series so far as the succession of the known species is con- 

 cerned. This has been recognized in recent years by American students, 

 and variously phrased or interpreted. The most probable explanation of 

 the facts is to suppose that the known phylum is approximate, not direct ; 

 that the direct line of descent leads through unknown or imperfectly 

 known species, and that those known to us are offshoots of varying close- 

 ness. The direct line is, then, admittedly through hypothetical species, 

 and the only question is whether the habitat of these species was in the 

 regions where we have searched vainly for their remains, or in the much 

 greater intervening region where we have not searched. Horses are found 

 throughout the Tertiary in central and western Europe on the one hand, 

 on the Western Plains of America on the other. There is every reason to 

 believe that they inhabited all or parts of the intervening region and we 

 have no right at all, in weighing the evidence, to refuse to take this re- 

 gion into consideration, on the plea that it has furnished no "documents" 

 as yet. To place such limitations on our theories would hardly tend to 

 solving our problems, however much it might seem to simplify them. It 

 is merely to prefer a conclusion that we know to be false to a conclusion 

 that we cannot prove by direct evidence to be true. 



What I have stated in regard to the fossil ancestry of the horse applies 

 to most mammalian phyla, in greater or less degree according to the per- 

 fection and number of our "documents." Where these are few and frasr- 

 mentary, it is still possible to build up phyla which cannot be proven to 



