1010.] Ancestral Rodents. 331 



Rodent astragalus, as represented by that of Parainijs (described above, p. 

 327), differs in details from the Insectivore types seen in Solenudon,Centetes, 

 Erinaceus, Talpa, and Tupaia, especially in its very broad trochlea; but on 

 the whole it agrees much better with the Insectivore than with the Creodont 

 types. The Rodent astragalus (Fig. 30, p. 454) also resembles the Tubuli- 

 dentate and Xenarthrous ty})es (especially those of the Santa Cruz Edentates) 

 in many respects. The average humerus and femur of Rodents appear to be 

 less primitive in details than those of the lowest Insectivora, especially with 

 regard to the frequent reduction of the entepicondylar foramen and third 

 trochanter. Perhaps the most noteworthy feature of the vertehroe is that in 

 Lepus, Cavia and others the transverse process of the seventh cervical is 

 pierced as it is in Marsupials. The pseudo-sacral vertebne in Rodents often 

 have large transverse processes which tend to unite with the ischium. 



There are i)retty clear indications (especially in the Duplicidentata) 

 that the dental formula of Rodents has been derived by reduction from Ixjif;, 

 and that the dental replacement was formerly normal (Weber, 1904, p. 480). 

 The evidence for the derivation of the Rodent cheek teeth from the trituber- 

 cular-tuberculosectorial type has been cited above (p. 327). 



Hypothetical Reconstruction of the ancestral Rodents. — I'he ancestral 

 Rodents probably exhibited the following characters: mandibular condyle 

 narrow; post-glenoid process very small (as in Tupaia), or even absent (as 

 in Macroscelides); premaxillaries elongate, anterior upper incisors (li) 

 somewhat enlarged {cf. Solenodo7i), lower incisors (l^^) procumbent (cf. 

 Tupaia) ; canines and anterior premolars reduced ; cheek teeth adapted for 

 frugivorous-omnivorous diet. These conditions are all realized in the 

 Basal Eocene Mixodectidse, which Osborn and Matthew were formerly 

 inclined to regard as ancestral Rodents ("Proglires" Osborn, 1902, pp. 

 203-213); but the upper molars in this family contain no suggestion what- 

 ever of the peculiar features seen in the Lower Eocene Paramys; the lower 

 molars likewise were normally tuberculo-sectorial, without the hypertrophy 

 and internal displacement of the antero-internal cusp, or the reduction of the 

 protoconid and asymmetry of the talonid, which are characteristic of the most 

 primitive Rodents; so that in spite of the resemblances in the astragalus 

 it seems more likely that the Mixodectidse are not ancestral Rodents but 

 aberrant Insectivores (Matthew, 1909). 



The ancestral Rodents may also have been adapted for semi-arboreal life, 

 since, so far as the evidence goes, they were not distinguished by strongly 

 marked fossorial, ac^uatic, or cursorial adaptations; they were probably 

 claviculate, well clawed, pentadactyl, with well developed pollex and hallux, 

 with small body and long tail, free tibia and fibula, unreduced lumbar para- 

 pophyses; the foregoing characters together with the assumed frugivorous- 



