1885.] DR. ST. G. MIVART ON THE ARCTOIDKA. 351 



limbs, compared with the length of the spine. It has also the 

 longest radius, tibia, metacarpals and metatarsals, and palate. 



In the skull the alisphenoid joins the parietal \ but the ascending 

 process of the premaxilla does not attain the descending process of 

 the frontal ". The palatine bone joins the orbito-sphenoid, and the 

 postaxial process of the premaxilla is separated mostly, if not always, 

 from the preaxially extending process of the frontal, by the junction 

 of the nasal with the premaxilla. The malar joins, at least some- 

 times ^ the lachrynal bone, and forms part of the margin of the 

 infraorbital foramen. The maxillary bone forms a floor to the 

 orbit much as it does in Canis, and therefore to a less extent than in 

 Felis. The palate extends back much beyond the last molars. The 

 postorbital processes are very small ; the subaugular process of the 

 mandible is rudimentary. The form of the basis cranii has been 

 described and figured by Professor Flower (P. Z. S. 1869, p. 9 

 fig. 3, p. 10). 



The scapula has a remarkably large supra-spinatus fossa, and 

 its metacromion is, as a rule, developed much as in the Civet. 

 The humerus is perforated by a small supra- condyloid canal. 

 The femur has the upper part of its posterior surface somewhat 

 flattened. The ungual phalanges are somewhat intermediate in 

 form between those of the Civet and those of the Dog. 



Molar formula = P. j, M. f. 



The teeth remind us rather of those of some of the Paradoxures 

 than those of any other ^luroids or Cynoids, but the non-sectorial 

 character of the Paradoxurine teeth is here carried still further. 

 The last, or fourth, upper premolar and the two upper molars have 

 each three roots. The second and third upper premolars have each 

 two roots. All the lower grinders except the first premolar have also 

 each two roots. The fourth upper premolar is not at all sectorial 

 in character. It has three external cusps (the middle one of which is 

 much the largest) continuous with a small external cingulum, and 

 two internal cusps placed opposite the interspaces of the three 

 external cusps. The first upper molar diff'ers from that of every 

 .35^1uroid or Cynoid in that it is no broader anteriorly than it is 

 posteriorly. It has four large cusps, whereof the two outer ones are of 

 equal size and rather more distant from the external cingulum than in 

 the fourth upper premolar. There is also a minute cusp developed 

 from an internal cingulum, and situated inside the postero-internal 

 cusp. The second upper molar has four cusps. Two of these are 

 external, the anterior one of the two being the larger. Of the 



'^ In this matter Proa/on agrees with Felis, Viverra, Paradoxurus, Arctictis, 

 Cynogde, Hcrpcstes, Suricata, Hyana, and Crocida, as well as many forms 

 of Canis, though I find that sometimes in Canisih^ squamosal joins the frontal, 

 as in the Wolf and Kit Fos. 



^ I have noticed this junction to occur in Hyaena, except H. hrunnea, Crocuta, 

 Herpestes, Arctictis, souietimes in Felis, and most markedly so in Suricata, where 

 the summit of either premaxilla ascends a little the inner side of the descending 

 process of the frontal. In Viverrn, Paradoxurus, Cynocjcde, and generally in 

 Felis, the premaxilla and frontals do not meet. 



^ Cuvier, Lemons d'Aiiat. Comp. vol. ii. 



