1910.] Principles of Mammalian Phylogeny. 107 



Reconstruction of synthetic types. — For these reasons the phvlogenist 

 must attempt a sort of survey by triangulation : the characters of known 

 forms merely affording variously situated points for the backward projection 

 of assumed lines of development to their intersection in undiscovered syn- 

 thetic types. 



Persistent primitive forms. — The very factors wdiich complicate the 

 problem, however, serve also to elucidate it, because the immigrant elements 

 in one fauna furnish examples of a fauna originally bred elsewhere; wdiile 

 in each age the persistent primitive forms and structures carry us back to 

 the stem forms of antecedent ages. 



Relative phylogenetic value of characters of the dentition and of the feet. — 

 As showai above (pp. 42S 80) this question is an historic one: Which set 

 of characters is more often adaptive and more likely to exliibit misleading 

 homoplastic resemblances betw^een ordinally separated forms, the characters 

 of the dentition or the characters of the feet ? 



Before attempting to answer this question we note the predominant 

 place of the dentition in mammalian palaeontology. The majority of fossil 

 mammalian species, especially in Europe, are knowai chiefly through the 

 dentition, and dental characters enter very prominently into the classification 

 of the species, genera and families. Even many orders take their names 

 from characters relating either to the teeth or to the food and manner of 

 using the teeth, as " Protodonta," "Triconodonta," "Trituberculata," 

 "Insectivora," "Creodonta," ''Carnivora," "Rodentia," "Tillodontia," 

 "Simplicidentata," "Duplicidentata," "Tubulidentata," and "Toxo- 

 dontia." Generally in cases w-here the ordinal position of an animal is 

 known on other grounds, the teeth offer a safe criterion of its specific, generic, 

 family and subordinal affinities. 



Characters of the dentition often inferior in value to characters of the feet. — 

 Where the ordinal or superordinal affinity is in doubt the teeth are often, 

 but not always, inferior in value to the feet. 



(1) Great differences in the dentition may occur even among members 

 of the same family, in cases where the mutual affinity is clearly indicated in 

 the feet. 



(a) Contrast, for example, the cat-like cheek teeth of Cryptoprocta with 

 the bunoid cheek teeth of Arctictis; yet both genera belong in the Viverridse. 

 Still greater differences may occur within the limits of the same suborder or 

 order. 



(6) How great is the difference between the many-columned last molar 

 of Phacochoerus and the two-ridged molar of Dicotyles; yet the resemblances 

 in the feet clearly indicate the superfamily alliance between the genera in 

 question. 



