50 THE AGE OF MAMMALS 



Sources of error in correlation. — Evidence of these five kinds as the 

 basis of the correlation of formations contains several sources of error. 

 First, we should always be on our guard against imperfections in records 

 and should keep in mind the possible presence, while a given formation 

 was being deposited, of mammals which perhaps escaped fossilization or 

 whose fossil remains have not yet been discovered. We must not too 

 hurriedly assume the absence of a mammal from an entire continent or 

 even from the geographic region of a certain formation simply because it 

 has not yet been discovered in that formation. Many mammals long con- 

 sidered absent from the entire American Eocene, for example, the peculiar 

 armadillo-hke forms of South America, have recently been discovered in 

 the Bridger Formation above mentioned. Again, some mammals living 

 near the larger streams or along the shore lines are much more apt to be 

 caught and entombed in certain formations than others living at a dis- 

 tance, in the forests or out on the uplands, for example. 



Most formations are limited in geographic extent, and we must always 

 keep in our imagination the life of the vast outside areas which were also 

 thickly populated, with their differences of habitat, of longitude, or eastern 

 and western distribution, of latitude, or northern and southern distribu- 

 tion, of altitude, or vertical distribution, such as on mountain ranges and 

 in the valleys; in short, there were always in past times such differences 

 of distribution as exist among mammals to-day, which render it improl)- 

 able that the restricted area of a given 'formation' will give us an ade- 

 quate picture of the entire contemporary life of a continent. 



Progressive Correlation 



European Correlations. — The foundation for the correlation of Euro- 

 pean formations with each other naturally began with the early work of 

 Cuvier and advanced with the progress of mammalian palseontology on 

 the continent. In France, Gervais ('59, '69), Gaudry ('62, '73, '78, '86, 

 '88, etc.), Filhol ('77, '79, '81, '88, '91), Lemoine ('78, '80, '82, '85, '87, '88), 

 Boule ('83, '88, '93, '96, etc.), and especially Deperet ('87, '90, '92, '93, '00, 

 '05, '06) have successively described typical horizons or formations and 

 the mammals characteristic of them. Parallels between the formations 

 of England and France were early set forth by Owen ('60), followed by 

 Sir Joseph Prestwich ('88), and William Boyd Dawkins ('80, '94). Par- 

 allels with Germany have been especially treated by Von Zittel, Schlosser 

 ('88, '83-'97, '90, '95, '02), and Deperet ('85, '87, '90, '92, '93, '05, '06, etc.).^ 



The first step in correlation through faunal parallelism, or similar life 

 zones, is naturally to assemble as full a list as possible of the character- 

 istic species and genera of mammals. Valuable tables of such European 

 faunal parallels are those given by Von Zittel in his great Handbuch der 



' Principal titles are given in the Biljliography. 



