618 CARNIVORA. 



These divisions, though very convenient when appHed to Uving forms, 

 break down when the extinct species are considered. Thus, as already 

 mentioned, in the older extinct Felidae the alisphenoid canal and post- 

 glenoid foramen are both present, the condylar foramen opens behind the 

 foramen lacerum posterius, and the molars are more numerous. More- 

 over some of the older Viverridae seem to approach so close to the Mustelidae 

 both in their cranial and dental characters that it is difficult to separate the 

 two families. The inference has been drawn from these facts that the 

 Arctoidae are not a natural group and that the resemblances between the 

 living Mustelidae and Ursidae have been independently acquired. This 

 conclusion is not in our opinion justified. To say that features of resem- 

 blance which undoubtedly exist between two families A and B, have 

 been independently acquired simply because A is found to present features 

 of resemblance to a third family C, which B does not resemble so closely, 

 seems to us an unsatisfactory position for a systematic zoologist to hold. 

 We should prefer to leave out all speculations as to descent and to place the 

 matter in this way. Certain extinct animals (e.g. Stenoplesictis, etc.) must 

 be placed with the living Felidae and Viverridae because of certain charac- 

 ters, but they differ from these living forms in also possessing characters 

 which belong to Mustelidae. These however are not strong enough to 

 separate them from the living Felines, but are only sufficient to show a 

 closer affinity of Felines to Mustelidae than was at first thought possible. 

 In the same way the Mustelidae have certain Ursine features which induce 

 us to class them with bears. These are on the whole more important than 

 the resemblances to the extinct Felines, so we leave the Mustehdae with 

 the Ursidae, merely noting the fact that they have affinities to the extinct 

 Felines. What we have here is merely an example of the principle, to 

 which we have often before called attention in this work, that the more 

 closely any given group of animals is studied, the more complex are the 

 mutual relations between its different members foimd to be. For an 

 example of this we may refer the reader to vol. 1, p, 410. We have there 

 given a rough diagram showing the interconnections existing between 

 different groups of the Nudibranchiata. A similar diagram might be 

 constructed in the present case. 



AELUROroEA. 



Fam. 1. Felidae.* Dentition i i c \ p ^-^m -^ ; canines very strong ; 

 upper carnassial with three lobes on the blade and an inner cusp (Fig. 322), 

 lower with two outer cusps (Fig. 323), a weak or absent inner cusp, and 

 weak cutting or absent talon. U. molar very small and transverse. 

 Premolars reduced in number. Auditory bulla inflated, with an internal 

 septum and short bony auditory meatus, the paroccipital processes 

 flattened against the bulla ; without alisphenoid canal f ; carotid canal 

 minute ; condylar foramen opens witliin foramen lacerum posterius. The 

 humerus has an entipecondylar foramen. Digitigrade, manus with 5, pes 

 usually with 4 digits. Os penis small, Cowper's glands present, prostate 



* Efliott, Monograph of the Felidae, London, 1878-83. Mivart, The 

 Cat, London, 1881. Cope, The extinct cats of America, Amer. Nat., 

 1880, p. 833. Adams, Extinct FeUdae of N. America, Amer. Journ. 

 Sci.,'l, 1896, p. 419. 



■j- Present in some extinct forms. 



