THE ANTARCTIC WOLF. 391 



came from Mar del Plata and died in the Gardens, the above- 

 described crests on the upper carnassial are not better developed 

 than in G. antarctic iis, and the masseteric ridge on the malar bone 

 shows a decided approximation to the condition seen in that 

 species. This latter character is still better marked in the skull 

 of another South- American dog, the exact locality of which is 

 unknown, but which was a different species * from the Mar del 

 Plata example, and the crests on the carnassial exhibit the same 

 feebleness of development. But it may be noted that in both 

 these skulls the positions of the cusps on the upper and lower 

 carnassials are more latransAike than antarcticus-like, so that in 

 this respect at least they serve to bridge over the difference 

 between those two species ; a fact in keeping with the idea that 

 C. antarcticus is a specialised form of the group of South- 

 American dogs above alluded to, but specialised in a direction 

 away from that taken by 0. latrans and its allies. 



The external characters of C. antarcticus and C. latraiis also 

 afford no justification for the claim of close relationship between 

 them. In the first place the ears of G. antarcticus are very small, 

 smaller indeed comparatively, I believe, than in any wild species 

 of the dog family, with the exception perhaps of G. sclateri, 

 C. {Nyctereutes) procj/onoides, and Vulpes (Alopex) lagopics. In 

 G. latrans, on the contrary, they are as large as in most, at all 

 events, of the species of Ganis. An idea of their length in the 

 two species maybe gathered from tlie measurement of a specimen 

 of each of approximately the same size given by Mivart, the ear 

 of G. antarcticus being 6*5 cm. (= about 2| inches) and that of 

 G. lattrms 14 cm. ( = about 5| inches), or more than twice as long. 

 It may be added that the measurement of 2 inches 9 lines assigned 

 by Waterhouse to the ear of G. antarcticus confirms Mivart 's 

 statement. 



As regards colour G. latrans varies from grey to greyish fawn, 

 mixed with black above, and shows the charactei'istic clouded or 

 patchy coloration caused by the running together of the bands of 

 the individual long coarse hairs of the back and sides seen in so 

 many of the so-called wolves and jackals. One of these long 

 coarse hairs, pulled at random from a skin, measured about 

 3 inches long, the black tip being | inch (12 mm.) and the 

 whitish area below it 1 inch (25 mm.). The whole of the ventral 

 surface from the chin to the root of the tail is usually white or 

 whitish, and always apparently markedly paler than the back and 

 sides, though sometimes the continuity of the light tint is inter- 

 rupted on the throat by an infusion of fawn. Thei-e is no dark 

 patch above the hock on the hind leg, and the tail matches the 

 back approximately in colour throughout, the tip and the gland- 

 spot being blacker than the rest. 



In G. antarcticus the coat is thick and soft, and comparatively 

 shoi't, with none of the long coarse hair seen in G. latrans. One 



* The Sontli-Amevicaii dogs of tliis gvoun are in siicli a systematic nindclle that 

 it is very difficult to identity specimens without a complete revision of the whole 

 series. 



