41 8 PATAGONIAN EXPEDITIONS I PALAEONTOLOGY. 



notch has disappeared and the protoconid and paraconid are of the same 

 height, forming an elevated trenchant blade. 



Owing to the absence of transitional forms in the Santa Cruz fauna, so 

 far as known, it is less easy to trace the development of the peculiar 

 notched and fluted sectorials of Abderites, which are to be regarded as 

 highly specialized structures adapted to a piercing habit, suggesting that 

 the animal fed on the eggs of birds. The loss of the metaconid is a 

 further adaptation toward the perfection of the piercing function. The 

 derivation of the sectorial teeth in Abderites from the tuberculo-sectorial 

 type of molar characteristic of the Caenolestinae is indicated by the broad 

 heel, and by the additional fact that the remaining molars, although some- 

 what less lophodont than in the Palaeothentinae, retain both the paraconid 

 and metaconid as distinct cusps. 



A satisfactory discussion of the derivation of the upper molar patterns 

 in the Caenolestidae is at present impossible, owing to a lack of material 

 illustrating the upper dentition in many of the genera, especially in the 

 more primitive forms. 



Less uncertainty exists regarding the lower molars. In the Palaeo- 

 thentinae, lophodont molars have been developed from teeth of the primi- 

 tive tuberculo-sectorial type, shown in Halmarhiphtis, by the formation of 

 cross crests uniting the cusps of the talonid and heel. In the buno-loph- 

 odont molars of Abderites, all the cusps of the original tuberculo-sec- 

 torial crown have been retained, except in My. The loss of cusps in this 

 tooth has already been discussed. 



So little is known of the skull in the majority of the Caenolestidae that 

 any attempt at a discussion would resolve itself into a repetition of 

 Thomas's excellent description of the skull of Ccenolestes (1895). A de- 

 scription of an incomplete skull of Palceothentes will be found on a later 

 page, to which, and to the accompanying illustrations of the skull of 

 Ccenolestes (PI. LXIII, figs. 14-14^) the reader is referred. 



Little is known of the podial structure of the Caenolestidae. Ameghino 

 (1894, pp. 80, 81) describes the feet of the Santa Cruz representatives of 

 his suborder Paucituberculata (= the Caenolestidae) as follows: "The four 

 limbs were almost equal in length, but the hind feet were longer than the 

 fore. They were plantigrade, with five toes on the hind feet and probably 

 also on the fore feet, with all the toes well developed and without the 

 least trace of syndactyly." 



