146 



SCIENCE. 



[N. S. Vol. VII. No. 162. 



and Scott, on paleontological grounds, and 

 Hatcher, on geological grounds, have dem- 

 onstrated, this fauna is by no means of the 

 great antiquity assigned to it by Ameghino. 

 It is rathern modern and specialized than 

 central and ancestral. The sources of this 

 fauna should be sought possibly in an over- 

 flow of primitive Marsupials from Australia 

 or elsewhere, and partly in an early emi- 

 gration of Condylartha from North Amer- 

 ica. The latter may have constituted the 

 origin of the Litopterna, but they give us 

 no light upon the Toxodontia. At the 

 same time the general principle of North 

 American origin is strongly reinforced by 

 the demonstrated relationship of the Gano- 

 donta to the Edentata. 



As regards the remaining Ungulates of 

 the world, the origin of the Proboscidia and 

 Hyracoidea is still wholly unknown. The 

 Sirenia also remain without known ances- 

 tors. The group of Ancylopoda, proposed by 

 Cope for Chalicotherium and other clawed 

 forms with the bodily proportions of the 

 Sloths,but many essential skeletal structures 

 of the Perissodactyls, loses its distinctness 

 from the Perissodactyl phylum, because of 

 the discovery that Agrlochoerus besides Dip- 

 lobune, both undoubted Artiodactyls, ex- 

 hibit a very marked parallel specializa- 

 tion of hoofs into claws. Going further 

 back to the Lower Eocene there still exists 

 a break between the Artiodactyla and Per- 

 issodactyla and any of the known forms of 

 Condylarthra, for none of the latter are as 

 yet proved to be directly ancestral to the 

 even or odd-toed Ungulates. The Ambly- 

 poda stand apart as a very ancient and dis- 

 tinct phylum, geologically the oldest, and 

 in structure the most archaic of all Ungu- 

 lates ; they should include the Periptychidce 

 and thus embrace the whole range of am- 

 blypods from the small arboreal Peripty- 

 chids to the huge clumsy Uintatheres. 



The most primitive type of Condylarth 

 {Euprotogoiiia) and of Amblypod (Panto- 



lambda), as recently studied by Osborn and 

 Matthew, strongly reinforces the hypothesis 

 first enunciated by Cope, that the source of the 

 Uiigidata istobe foundiiithe Greodonta. Upon 

 the other side of the great Mammalian 

 tree, the numerous branches of Unguicu- 

 lates or primitive clawed types also have 

 converged towards a Creodont ancestry, as 

 seen especially in the characters of the 

 Ganodonta,or ancestral edentates, and of the 

 Rodentia, if Matthew's supposition proves 

 to be correct ; also of the Tillodontia. Thus 

 all these groups should probably be added 

 to the Caraivora as Creodont derivatives. 

 The Carnivora extend back into Creodont 

 prototypes; but, as in the case of the Arti- 

 odactyla and Perissodactyla, the actual 

 points of contact or links between these two 

 divisions are yet to be discovered. So, again, 

 with the Primates. Recent embryological 

 evidence has tended to separate the Lemu- 

 roid and Anthropoid phyla. Hubrecht is con- 

 firmed by others in placing Tarsius near the 

 parting of these phyla (although not in his 

 separation of this genus from the Lemurs), 

 and he makes a very radical break between 

 Lemurs and monkeys upon grounds of pla- 

 centation. The point of contact of the Pri- 

 mates with the Creodonta is still entirely 

 wanting, but their relations appear to be 

 here rather than with the Insectivora. 



In spite, therefore, of the many remain- 

 ing deficiencies or absence of links in our 

 paleontological evidence, it has none the 

 less come about that the Creodont typ>e takes 

 the central position which ivas assigned by Hux- 

 ley in 1880 to the Insectivora, for the known 

 Creodonta are more generalized and more 

 central than any other of the known In- 

 sectivora, fossil or living, the known Insec- 

 tivora showing a very considerable special- 

 ization, especially in their dental succession, 

 which places them apart as a distinct side 

 phylum. This does not aifect the deriva- 

 tion of the Creodonta themselves from stem 

 forms of unspecialized Insectivora existing in 



