1910.] The Ungulata: Conclusions. 409 



In consideration of the differences in habits between Moeritherium. and 

 Eotheriuni which are imjiHed by the possession in the former of a. well 

 developed, in the latter of a reduced, femur, it is evident that the possession 

 of so many characters in common constitute strong evidence of community 

 of origin; even although some of these characters may have been independ- 

 ently developed in the two forms. When it is further realized that apart 

 from Maritherium there is no known Artiodactyl, Perissodactyl, or represen- 

 tative of any other order that approaches Eotherium in nearly so many 

 points, and that even between existing Sirenia and Proboscidea many 

 peculiar features not obviously due to any similarity of habits are retained 

 in common, it then becomes clear that the evidence for relatively close 

 relationship between the Proboscidea and the Sirenia is very strong. Lydek- 

 ker (1892) notes that in certain Sirenia the last milk molar is molariform and 

 shows certain additional resemblances to the last milk molar of certain 

 bunoselenodont Artiodactyls. But such isolated points of similarity in the 

 dentition have often proved to be very deceptive when interpretated as indica- 

 tions of relatively near relationship. 



XII. General Conclusions. 



As shown in the preceding sections the various suborders of hoofed 

 mammals are connected by a tangled web of resemblances, a web which 

 early caused the "order" Ungulata to be regarded as a natural group — as 

 natural as, for example the Primates. It is only rather recently that it 

 has been suspected that this general stock of resemblances may have been 

 built up to a considerable extent by the processes of convergent and parallel 

 evolution. Indeed it seems almost as if certain authors had not yet become 

 aware of this possibility, for they continue to base classifications and phylo- 

 genetic conclusions upon adaptive resemblances, in a manner recalling the 

 methods of Blumenbach and Cuvier. 



The present problem in regard to the Ungulates is not whether they 

 form a wholly natural group, but rather to determine how far convergent 

 evolution has proceeded, to decide how many different primary stocks have 

 contributed to this composite "order" and especially to locate the time and 

 place, when and where these primary stocks became differentiated from 

 lower Placental orders. 



Among the provisional conclusions to which the writer has been led may 

 be mentioned the following: 



(1) The Artiodactyla are widely removed not only from the Peris- 

 sodactyla but also from all other ungulate groups. They may well be, 



