180 



Bulletin American Museum of Natural History. [Vol. XXVII, 



me 



tfits 



-ml- 



with the conditions in Paraspalax and Paurodon (Osborn, 1907, p. 29) give 

 the appearance of having been an upgrowth of an internal basal cingvdum, 

 which is otherwise not patently represented. None of these peculiar features 

 are foreshadowed in any degree in Spalacotherium (p. 174) in which the 

 cutting trigon is symmetrical, the thorn-like paraconid and metaconid diverge 

 outward and upward and are not continuous with the strong internal basal 

 cingulum, and in which there is no suggestion of a heel. Consequently 

 there seems little evidence for the view (see also Peralestes, p. 174) that the 

 tuberculo-sectorial molar type (as represented in Amphitherium) has been 

 derived from the generalized triconodont type by way of the Spalacotherium 

 type. On the contrary some evidence will be presented (p. 185) that the 

 tuberculo-sectorial type has been evolved in an entirely different manner. 

 Probable characters' of the upper molars. The next inquiry is to what 



extent the characters of the unknown upper molars 

 of Amphitherium may be inferred from the charac- 

 ters of the known lower molars. The views de- 

 veloped below are illustrated by the accompanying 

 hypothetical reconstruction of an upper molar of 

 Amphitherium (Fig. 10). 



From the greater number of lower teeth and 

 their relatively short anteroposterior diameters it 

 is practically certain that the upper teeth, as com- 

 ]:)ared with those of Didelphis, were relatively 

 shorter antero-posteriorly and broader trans- 

 versely, and in so far approached the Dryolestes 

 type (Fig. 12, B). The talonid as in Dryolestes is 

 much lower than in Didelphis and this very likely 

 implies that its complement, the protocone, was 

 relatively high. It might be thought that a low talonid would rather imply 

 a small shelf-like internal cusp, or "protocone," as it does in the premolars 

 of Tertiary mammals but this was not the case in Dryolestes and other 

 Jurassic genera (c/. Fig. 12, B). The protocone of Didelphis acts not only 

 as a pestle to the mortar-like talonid but also as a check which (together with 

 the metacone) prevents the protoconid from piercing the roof of the mouth. 

 If, in Amphitherium the low talonid had been opposed by a very low proto- 

 cone in the upper molar the two molars coidd not have come in contact, 

 because the protoconid was high. In Amphitherium the protocone must 

 have been higher than in Didelphis and about as high as in Dryolestes. 



The interspaces between successive protocones in Amphitherium. were 

 doubtless occupied when the jaws were closed by the paraconid and meta- 

 conid of the lower molars. These cusps did not oppose anything in the 



Fig. 10. Hypotlietical re- 

 construction of an upper 

 molar of Amphitherium pre- 

 rostii. Greatly enlarged. 



