30 



CEXOZOIC MAMMAL HOBIZO^vS 



time scale for the entire Tertiary and (2) to speak with precision re- 

 garding the time of successive migrations, and it is possible that we 

 shall be able (3) to describe our subdivisions in the terms of the stages 

 or etages employed by our European confreres. These are results 

 toward which the writer has worked for many years in cooperation 

 with many colleagues in this country and in Euroj^e." 



METHODS OF CORRELATION. 



Bast s. — The faunistic bases which the writer laid down** for Euro- 

 pean and American correlation were: 



1. Percentages of common genera and species. To this families 

 should now be added. 



2. Similar stages of detailed evolution in related forms, e. g., 

 Eocene Equidie. Miocene Rhinocerotidie. 



3. Simultaneous introduction of new forms by migration, e. g., 

 Mastodontin^e in middle Pliocene. 



4. Predominance or abundance of certain forms, e. g.. Promery- 

 cocharm in all lower Pliocene deposits. 



5. Convergence and divergence of fauna" in comparison with Europe 

 and Asia, e. g., lower Eocene, Oligocene, middle 2kIiocene, middle 

 Pleistocene. 



6. ExtLQCtion of certain forms, e. g., Steneofiher, iu lower Pliocene. 

 Similarly the correlation of the Tertiaries of northwestern America, 



iuter se. should be based on — 



1. Presence of similar specific and generic stages. 



2. Evidence of similar local evolution. 



•3. Dominance or scarcity of similar animals in the fauna. 



4. Disappearance or apparent extinction of similar forms. 



5. First appearance of similar forms. 



It should always be kept in mind that the appearance of a new 

 mammal in any of the Plains or Moimtaiu deposits may occur by 

 migration from one of several possible sources, namely: 



1. From the unknown Plains or Mountain regions of eastern North 

 America or of northern North America. 



2. From Europe or Eurasia as a whole. 



3. From South America. 



4. From Africa. 



5. From AustraUa via Antarctica. 



Sources of error. — This kind of evidence as to the bases of correla- 

 tion is subject to two great limitations: 



First, imperfections in our records, or the possible presence of 

 animals which have existed in contemporaneous, earlier, or later 



« Correlation IjetTreen Tertiary ynaTDTT^ai horizoiis of Europe and America: Ann. Xew York Acad. Sci-, 

 ToL 13. 1900, pp. .>-14. Also. Trial siieets of typical and homotaxial Tertiary horizons, issoed In 1SE>7. 

 1S9& and 1900. 



