CHAP. XV.] THE PEDIGREE AND OlilGIN OF THE CAT. 517 



than that the allaiitois should first appear as a_ functionless rudi- 

 ment. It also seems more probable that the habit of forming milk 

 teeth is one which has been lost or almost lost in certain mammals, 

 than that the j^rocess should first have arisen through the replace- 

 ment of a single, relatively unimportant, tooth, by a vertical successor 

 to it — a condition which has been found* to be the case in the 

 marsupials of our own day. 



§ 7. The succession of mammalian carnivorous life — the mammalian 

 portion of the cat's pedigree — may then be represented as follows : — 



From unknown Insectivora-like mammals, two diverging series of 

 forms may have started, one soon leading to Ardocyon (as the 

 Insectivora-like root of the placental Carnivora), the other series 

 developing such forms as Provkerra, Ilijcenodon, and Pterodon, and 

 continuing on the main stem through Gymnura-like creatures to the 

 modern InsccUvora. From this insectivorous stem we may imagine 

 a side-shoot to be given off leading through Palceonictis to fomis 

 like certain existing marsupials, and diverging into the American 

 Didelphys and the analogous Australian Bcmjiwus. From Ardocyon 

 we may conceive the great carnivorous branch — destined to quite 

 surpass and overshadow the insectivorous stem — to divide into 

 cynoid and arctoid branches. The former continuing on through 

 Cynodon and Cynodidis, would lead up through GaJecymis and 

 forms like the existing Otocyon to the typical Canidce. The great 

 arctoid branch may have given ojff a limb leading through Amphicyon, 

 Ilymiardos and Idndred forms, to the existing Ursidw, and then 

 continued on through Procelarus and Lutridis, as the 3InsteHdce. 

 From some such form as Proceltirus the great .^luroid sub-order 

 may have started, and before continuing on, as the Virenidce, have 

 given off a great branch to be developed, by bifurcating, into the 

 JIyre)iidce,Cryptoprodidfe SindFelidce. The first family is the culmina- 

 tion of one division which passes through Iditlierunn, and which 

 gives off Proteks as a one-sided branchlet. The other division into 

 which the ^Fluroid branch bifurcates, continues on as the cats, first 

 giving off however, near the bifurcation, the branchlet ending in 

 Cnjptoproda. The proper feline branch then continues on through 

 Ardudurus, Dinidis, Nimmvus andPseudcdanis, and then bifurcates. 

 It ends in the typical genus Fdis on one side— an aberrant twig 

 being given off for Cymehmis — while on the other side it continues 

 on though Hoplophoneus, Pogonodon and Madicerodus f to the very 

 specialized aberrant form Eusmiliis. 



This hypothetical genealogy is only offered as a speculation, especi- 

 ally that part of it which represents conditions anterior to the evolution 

 of the viverrine branch. It reposes mainly upon dental characters, 

 and teeth are organs which not only might be expected to vary with 



* By Professor Flower. See his paper 

 in tlie Philosophical Transactions, vol. 

 clvii., 1867, p. 631. 



+ The small lamina of bone which em- 

 braces the external carotid and so forms 



the "ali-sphenoid canal " may -well have 

 independently disappeared and again, by 

 reversion, reappeared in either sub- 

 division of the feline branch. 



