I'JIO.] The Rodentia. 325 



1849. Gervais sets apart the Lagomorphs as a suborder, under lUiger's 

 term " Duplieidentata." 



1851-1855. Brandt proposes his four subordinal divisions "Lago- 

 uiorpha," "Seiuromorpha," " Myomorpha," "Hystrieomor[)ha." 



1899. Tullberg monographs the entire group. He divides the SimpH- 

 eidentata into two great series "Sciurognathi" and "Hystricognathi." The 

 former divides into the "Myomorphi" and " Sciuromorphi " the latter into 

 the " Bathyergomorphi " and "Hystricomorphi." The Myomorphi again 

 inchide the "Ctenodaetyloidei," "Anomahiroidei" (Pedetidje + x\nomahu'i- 

 dfe), "Myoidei." The Sciuromorphi include the " Geomyoidei," "Castor- 

 oidei," "Sciuroidei." 



Naturalness of the Order. 



Before taking up the problem of the general relations of this order it is 

 necessary to inquire whether the t^^'o suborders, Duplieidentata and Simpli- 

 cidentata, which are separated by many important characters and between 

 which no fossil linking forms have been discovered, are in truth related or 

 whether the numerous resemblances that tend to connect them may not be 

 due largely to convergent evolution. 



Mr. Gidley has suggested (1906, p. 99) that the oldest known type of 

 Duplicidentate molars, that of the Oligocene Paloeolagus, might have been 

 derived from the Mesozoic triconodont type by the intervention of a hypo- 

 thetical intermediate stage. This hypothesis seems, however, to rest on very 

 scanty evidence. The Triconodonts, if their inflected angle proves anything, 

 are Marsupials (Metatherians) allied remotely to the Polyprotodonts; but 

 it would be difficult to point to any definitely Marsupial characters in the 

 Duplicidentates except perhaps the perforation of the transverse process of 

 the seventh cervical vertebra by the vertebral artery; while very typical 

 Placental characters abound. Dr. Forsyth Major, although an opponent 

 of the theory of trituberculy, shows (1899, pp. 433-520) that the upper molars 

 of certain Miocene Duplicidentates retain clear evidence of derivation from 

 a sharply triangular type which is wholly unlike the Triconodont type; 

 and that the deep groove which finally sinks in and divides the Leporid 

 molars into distinct anterior and posterior halves is a secondary character, 

 initiated perhaps by the hypertrophy of the postero-internal ridge or hypocone 

 (see also Osborn, 1907, p. 149). 



The Duplieidentata (Weber, 1904, p. 493) possess many deep seated 

 characters in common with the Simplicidentata, such as: a typically com- 

 plete uterus duplex, a discoidal deciduate placenta, a small allanto-chorion, 

 a large cup-shaped invaginate yolk sack in which the embryo lies, nineteen 



