PROF. OSBORN ON THE ORIGIN OF MAMMALS. 417 



lucre certainly sviall, terrestrial, claived, insectivorous or omnivorous 

 forms. It is noteworthy that in the evokition of each radiation, 

 so far as we know at present, land types and organs are invariably 

 primitive, and water types and organs are secondary, exactly as we 

 find it among the Reptilia. In fact we have not found a single 

 instance in which a mammal or reptile series is known to be trans- 

 forming from a water into a land type, it is always the reverse. 

 There is certainly no evidence for a cetoid (Albrecht) stem of the 

 mammals. Again it is obvious that neither carnivorous nor herbi- 

 vorous types with highly specialized or reduced teeth and feet can 

 be so central as insectivorous and omnivorous types. In fact the 

 Insectivores among Placentals, and Opossums among Marsupials, 

 are the only animals which have preserved the dental prototype 

 close to that of the Promammal. 



The backward convergence of all Tertiary mammalia to a 

 Creodont stem, as indicated in the chart, affords the clearest 

 demonstration of the existence of an earlier insectivorous stem, 

 and it not only accords with the above principles, but we are thus 

 carried a step further to realize that the Creodont ancestor must 

 have been a generalized Insectivore. 



To reconstruct the ancestral Eutherian skeleton, therefore, we 

 should divest the Creodonta and Inscctivora, in their anatomy and 

 development, of all their secondar)' specializations, and combine 

 all their truly primitive characters. This ideal Eutherian is 

 nearer the pro- Marsupial than we supposed, it has the following 

 chief characters : — head large, body relatively small. Anterior nares 

 terminal, face elongate. Insectivorous or omnivorous habit. Molar 

 teeth pointed, tritubercular. Typical dental succession. Vertebrae 

 with intercentra. Dorso-lumbars not exceeding 20. Back arched, 

 tail long and powerful. Scapula and ilium narrow, acuminate. 

 Humerus with powerful deltoid crest, condylar crests and entepi- 

 condylar foramen. Femur with three trochanters. Feet planti- 

 grade. Centralia and tibiale. Fore-limb more or less prehensile, 

 elbows everted. 



Now the Creodonta are members of that Cretaceous radiation, 

 IV, which we find in full bloom in the Basal Eocene, mingled 

 with the dying group of Multituberculates, which belong to the 

 Jurassic radiation, V. 



In the Mid- and Uppermost Jurassic (Stonesfield, Purbeck 

 and Como Beds) however, occur all three types, which theoretically 

 constitute radiation V : first, the Triconodonts primitively but 

 typically Marsupial in structure ; second, the Inscctivora primitiva, 

 primitively Placental in type, insectivorous and without a single 

 Marsupial character either in jaw or teeth ; third, the Multituber- 

 culata, whose sub-class position is now assumed to be Monotremc. 



Always keeping in mind that our direct evidence here is of the 

 most limited character, since we have neither skulls nor skeletons, 

 only teeth and jaws, we are tempted to hypothetically connect the 



27 



