BIOLOGY AND ZOOLOGY 135 



" pro-mammalia " may be applied. The Multituberculata must be regarded as related 

 with the marsupials, but were not diprotodont. The Triconodonts, excluding Peralestes, 

 are also to be regarded as having marsupial affinities and to be distinct from the Multi- 

 tuberculata. The Trituberculata, although showing Creodont and Insectivore resem- 

 blances, are also more nearly related with the Metatheria. The primitive Marsupial 

 skull on the whole approaches the Cynodont type more nearly than does that of the 

 Monotremes, and the Marsupio-Placental remnant still retained some of these reptilian 

 characters after the branching off of the Monotremes, but, owing to the assumption of 

 semi-arboreal habits, it lost the Monotreme characters of the shoulder girdle. At such 

 a point the divergence of the Marsupials and Placentals began, the former retaining 

 and perfecting their arboreal characters. He gives a conception which he admits to be 

 somewhat speculative, as the fossil history is obscure, of the ancestral Insectivores. 

 These would be ordinally related with the contemporary ancestors of the Creodonts, 

 but would be distinguished chiefly by smaller size and less robust structure, more 

 elongate snout and prem-axillaries, smaller canines, smaller and more sharply cusped 

 molars, and perhaps by habits that were more arboreal, less predatory and more in- 

 sectivorous-frugivorous. He removes the Tillodonts from their supposed relations 

 with the Rodents, and places them near the Insectivore-Creodont stock. 



The Creodonts are to be derived from Mesozoic Insectivora, and their relationship 

 with polyprotodont Marsupialia rests chiefly on convergence. The Sparassodonts 

 were very early separated from the Creodonts. The Fissipede Carnivora are descended 

 from Creodonts by way of the Miacidae. The Fissipede Carnivora, as shown especially 

 by the close relation between seals and bears, are derivatives of Arctoid Fissipedes. 



The Menotyphla (Tupaiidae, Macroscelididae), Dermoptera (Galeopithecidae), 

 Chiroptera and Primates are associated in a super-order to which Gregory gives the 

 name Archonta ("Apx wi; > chief, cf. German Herrenthiere). The separation of lemurs 

 from apes and monkeys may date back to the Lower Eocene, but the close relationship 

 of the two groups is accepted. The whole group Archonta is to be taken as a derivative 

 from large-brained arboreal Insectivores resembling Tupaia in many ways. 



The Rodentia form an extremely ancient order, perhaps going back to the Creta- 

 ceous, and it seems possible that their stem forms were Mesozoic Placentals, allied to the 

 ancestors of the modern Insectivora and possibly to the contemporary ancestors of some 

 of the Edentates. The extremely diverse character of the animals which are grouped as 

 Edentata for convenience, is accepted, and Gregory thinks that there is a balance of 

 evidence for the view that the Xenarthra, Tubulidentata and Pholidota are the highly 

 modified descendants of some Mesozoic Placentals which branched off before the differ- 

 entiation of Insectivores, Creodonts and Proto-ungulates. The various suborders of 

 hoofed mammals are connected by " a tangled web of resemblances," a web which 

 early caused the " order " Ungulata to be regarded as a natural group. It is probable, 

 however, that convergent and parallel evolution accounts for many of the resemblances. 

 The Artiodactyla are widely removed not only from the Perissodactyla, but from all 

 other ungulate groups, and may be an offshoot from the ancestors of the Creodont 

 Mesonychidae, or even from some Lemuroid-Insectivore group. Even in the lower 

 Eocene, Artiodactyls and Perissodactyls are separated by deep-seated characters. All 

 the remaining ungulate suborders may trace their ancestry back to a varied order of 

 Cretaceous Proto-ungulates, which were primitive Placentals of small size, related with 

 the Creodont-Insectivore stock, but falling under the original definition of the Con- 

 dylarthra. By Basal Eocene times they had spread, probably from some Boreal 

 Holarctic centre as yet unknown, into North America, South America and Europe, and 

 had already split up into well marked families, such as Phenacodonts, Meniscotheres, 

 Periptychids and Pantolambdids in N.. America, Condylarths and Amblypods in Europe, 

 and Condylarths, Periptychids and Homalodotheria in S. America. Later on the 

 Amblypoda may have sprung from the N. American Pantolambdids, the Hyraces, 

 Embrithopods, Barutheres, Proboscidea and Sirenia from a Lower Eocene African stock; 

 the Perissodactyls and Chalicotheres from Holarctic Condylarths allied to the Phenaco- 



