1910.] The Triconodonta, the Trituberculata. 177 



Conclusions. The Triconodonta and iheir bearing nn fJie oriejin of the mam- 

 malian molar. 



(1) The order Triconodonta, after the exchision of Peralestes, is prob- 

 ably a natural group, allied to the Marsupials and ordinally distinct from the 

 Trituberculata (as here understood), but inheriting in common with that 

 group several characters {e. g., dental formula, distinction between premolars 

 and molars, p^ higher than nij) that may run back even to the Cynodont 

 reptiles. 



(2) Nothing is known unfavorable to Osborn's hypothesis that the 

 triconodont molar type was derived from the Protodont type. 



(3) The hypothesis of the origin of the triangular pattern of the molars 

 by rotation of the cusps is supported by such evidence as there is only when 

 strictly limited to the Triconodonta. 



(4) The upper teeth of Peralestes (p. 174) belong more probably to the 

 Trituberculata, and their presumed connection with the lower molar teeth of 

 Spalacotherium is discredited (p. 175); hence the only cited evidence of 

 morphological and genetic transition from the triconodont to the tritubercular 

 type seems to fail at the critical point. 



The Trituberculata (Pantotheria ix part). 



The order Trituberculata as defined by Osborn (1907, pp. 22-24) in- 

 cludes the families Amphitheriidse, and Amblotheriidt"e (Stylacodontidje), 

 Paurodontidcie and doubtfully the Dicrocynodontidte. To this assemblage 

 Peralestes should probably be added (p. 174). Amphitherium a,nd Amphi- 

 tylus both come from the Stonesfield Slate (Middle Jurassic), the other 

 genera date from the Upper Jurassic (or possibly Lower Cretaceous) i. e., 

 from the Purbeck of England and Morrison of Wyoming. Most of the 

 genera are founded on lower jaws with teeth; upper teeth are known only 

 in Dicrocynodon , Kurtodon, Dryolestes and Peralestes. 



Generalized characters retained among Trituhercidates. 



The order is of especial interest because it appears to be remotely related 

 on the one hand to the Metatherian order Triconodonta and on the other 

 hand to the common stem of the more generalized Placental orders Insec- 

 tivora and Creodonta. The Trituberculata present so many analogies in 

 the dentition to the Insectivora that certain families were at first named 

 Insectivora Primitiva by Osborn (1888, p. 247) ; but without further evi- 



