ALLEN: mammalia: canid^. 153 



The "Antarctic Wolf," or "Falkland Island Dog," a species restricted 

 to the Falkland Islands and apparently approaching extinction, is the only 

 recognizable species of Hamilton Smith's "section" Diisicyon (emended 

 to Dasicyon by most later writers who have referred to it) ; although named 

 by Shaw, in 180Q, as Cauis autarcticus, it was first really made known by 

 Waterhouse and Darwin (Zool. Voy. Beagle, Mamm., 1839, pp. 7-10, pi- 

 iv, animal) ; its skull, described by Mivart (Mon. Canidae, 1890, pp. 26-29, 

 pi. viii, animal) in 1890, appears not to have been figured, and its claims 

 to subgeneric distinction are not clear. 



The " Azara Fox," the "Magellanic Dog," and their allies fall into Ham- 

 ilton Smith's "section" Cerdocyon {= Lyca/opex Burmeister, 1854, part), 

 and seem at least subgenerically distinct from the foregoing and from the 

 larger, broader-headed Canids of Guiana, Venezuela and Colombia. For 

 the latter J. E. Gray proposed, in 1868, the genus T/iotis, which has been 

 adopted as a subgenus by several later writers. This name, however, is 

 untenable in this connection, having been applied exclusively by its 

 proposer, Hamilton Smith, in 1839, to certain African and other species 

 of Jackal (type, Caiiis aiitlms F. Cuv.). Thoiis of Gray includes the 

 "Crab-eating Dog" [Cauis cancrivorus Desm.) and its allies of northeast- 

 ern South America, and in the absence of any synonym to replace Thoits 

 the group may be called Carcinocyon with Cauis thotts^ Linn. (= Cauis 

 caucrivorns Desm.) as type, to comprise the C. thous group, C aquilus 

 Bangs, and probably other forms, including Cauis sclateri" (nom. nov.). 



The two species of Canids known certainly to occur in southern Pata- 

 gonia belong to Cerdocyou, as recorded below. C uiageUauicns is a well- 

 marked species, quite distinct from any of its northern allies. The relation 

 of C. griseus to the C. azarce group is evidently close, but in the ab- 

 sence of specimens from Paraguay, the type region of C. azara;, and of 

 Uruguay specimens representing C eutrcriauus (Burmeister), or of C. 

 vetulus (Lund) from Minas Geraes, Brazil, it is impracticable to attempt to 

 express an opinion as to their relationships. They are evidently closely 

 allied, and may probably rank, with C griseus, as only more or less well- 

 marked subspecies of C azarcs. All are apparently quite different from 



• Mr. Oldfield Thomas has recently adopted (Ann. & Mag. Nat. Hist. (7), XII, Oct., 1903, p. 

 460, and footnote), the Linnaean name Canis thous (ex Guiana) in place of cancrivorus Desm., on 

 seemingly reasonable grounds. 



^ Canis microtis ?)c\a.i&r, 1882, not Canis microtus Reichenbach, 1834, = Canis ve/oxSay. 



