Classification of existing Felidae. 331 



question would be profitless. So also with Trouessart^s 

 classification. The truth is that neither of these authors 

 had sufficient acquaintauce with the forms dealt with to 

 allocate them otherwise than by conjecture. Moreover, 

 their choice of generic and subgeneric terms was by no 

 means in accordance with the rules of nomenclature. 



In a series of five papeis * on the existing Felidae, I have 

 recently discussed various cliaracters, both cranial and 

 externa], which — for the most part, at all events — have not 

 been employed hitherto in the discrimination of species and 

 genera. Particular attention was drawn to the structure 

 of the hyoid apparatus, of the tympanic bulla, of the feet 

 arid the rhinariura, and it was shown (1) that F. leo, tigris, 

 onc'i, pardus, and uncia, which differ from the rest of the 

 family in having the suspensorium of the hyoid imperfectly 

 ossified, constitute a little group of Felidae containing two 

 genera, Panthera and Uucia ; (2; that Acinonyx may be 

 distinguished by the complete absence of cutaneous sheaths 

 guarding the claws ; and (3) that in the remaining species, 

 all provisionally referred to Felis, there is very considerable 

 variation in the structure of the feet, the size and shape of the 

 rhiuarium, and the structure of the auditory bullae. These 

 characters and others have been used in the following 

 attempt to classify the existing species of Feli I se ; but my 

 main purpose in publishing what folljws has beeu to show 

 the true relatiousliip of the species to one another, so far as 

 it can be determined, and to di-^pose of such prevalent but 

 fictitious groupings as those wliicli imply that the lion (/eo) 

 and the puma [concolor) are closely allied and that the 

 Ivnx {lynx) differs more from the domestic cat {cat us) than 

 the latter differs from the tiger {tigris) f. 



The consideration of generic names, although of sub- 

 ordinate importance, has been inevitable. Probably no two 

 authors will be quite in accord on the delimitation of the 

 genera. In the present state of our knowledge and with the 

 rapidly shifting conception of the value of the terms 

 "genera^' and "species," it would be idle to claim finality 

 on this subject^ and I do not pretend in all cases to have 

 been consistent in the admission of species to generic rank. 

 In some cases I may have gone too far, as in the severance 

 of ZibethaUurus from Prionailurus ; in others not far enough, 



♦ Ann. & Ma-. Nat. Hist. (8) xviii. pp. 2^22-229, 1916 (Aup:.) ; 

 pp. 306-316 ( Sept. ) ; pp. 326-334 (Oct.) ; pp. 419-129 (Nov.), and xix. 

 ],p. 113-130, 1917 (Jan.). 



t Perhaps the most important point connected with correctness of 

 view ou'thej^e and similar matters is that erroneous affiliation of species 

 may be honelessly misleading to students of geographical distribution. 



22* 



