226 W. D. Mattheiv — Fossil Vertehrates and 



rest so largely on unpublished material. It should be 

 added that the methods of interpretation and conclusions 

 are my own; other vertebratists have interpreted the 

 evidence differently^ Professor Osborn, who has studied 

 the correlation thoroughly, adheres to the classic view 

 that the extinction of the dinosaurs marks the close of 

 the Mesozoic; Mr. Gidley agrees with the palgeobotanists. 

 This divergence of conclusions, to a certain type of mind, 

 demonstrates the worthlessness of vertebrate evidence. 

 But that is not its real significance ; it is caused by the 

 attempt to interpret and evaluate the evidence, to dis- 

 cover its real meaning rather than to rely upon super- 

 ficial relations and rule-of- thumb methods. 



I have elsewhere pointed out-^ that owing to the com- 

 paratively rapid, obvious and w^ell understood changes 

 which vertebrate races, and especially mammals, undergo 

 in the teeth and characteristic parts of the skeleton, they 

 are able to afford much more precise and sure evidence 

 in problems of exact correlation than is furnished by the 

 fossil remains of more slowly changing groups of animals 

 or plants. Presumably all races of highly complex 

 organisms alter in various respects w^ith time, but if the 

 parts preserved as fossils are of relatively simple struc- 

 ture we may not be able to observe the changes in many 

 of them, or to distinguish fully between the results of 

 migration, of time (evolution in loco), or of varying facies 

 in the faunas compared. We are then reduced to the 

 crude statistical methods formerly universal, and still 

 widely employed in palaeontology, especially in palgeobot- 

 any. These statistical methods will bring approximately 

 correct results through the laws of averages, provided 

 the material be ample and no migration or differences of 

 facies be involved; but so long as they fail to differ- 

 entiate adequately the three elements of change — time, 

 facies and migration — in the faunas or floras compared, 

 so long will their results be untrustworthy and more or 

 less at variance with the evidence where these elements 

 can be and have been more fully distinguished. In prac- 

 tice, fossil vertebrates, and especially fossil mammals are 

 capable of affording much more precise and exact results 

 than invertebrates or plants, when there is sufficient evi- 

 dence at hand. Where they are very rare or fragmentary 



^'Matthew, 1914, 1. c, p. 390; 1915, in Problems of Amer. Geol., p. 406. 



