816 ON THE STRUCTURE OF THE MAMMAL GALIDICTIS. [NoV. 12, 



in position. There is no " fissui-e anterior" (anterior ectosylvian) 

 such as occurs in Nandinia. 



The foregoing description of anatomical facts relating to Gali- 

 dictis may be summarised for the purpose of shortly presenting 

 its characters and of comparing them with other Viverrid Oar- 

 nivora ; they are as follows : — 



(1) Prescrotal (i.e. postvulvar) glands are present. 



(2) The anus opens directly on to tlie surface of the body and 



not into a superficial cutaneous depression. There are a 

 pair of anal glands. 



(3) The stomach is not particularly elongated; it is more 



globular in form. 



(4) The csecum is long (for aYiverrid) and jjointed at the apex. 



(5) The convolutions of the brain are partly Herpestine and 



partly Viverrine in their characters. 



(6) The uterus is completely double. 



(7) The tongue has an anterior patch of conical papillae. 



These characters collectively distinguish Galidictis from all 

 other genera of Viverridse the anatomy of which is known, and I 

 believe that the last but one mentioned character — the completely 

 double uterus — is new to the Oarnivora. The specimen may of 

 course be abnormal, it being the only one that has been dissected, 

 though the probabilities ax-e against this. 



The division of the Viverridse by Mivart into a number of 

 subfamilies, viz. : Yiverrinse, Hei-pestinse, Cryptoproctinee, Eu- 

 plerinse, and Galidictinse, is not used by every one. For example, 

 in the most recent ' List of Vertebrate Animals,' * published by 

 the Zoological Society, there is no such division adopted. If,, 

 however, this division is finally allowed, I should confirm from 

 rny own experience the justice of separating Galidictis in a sub- 

 family apart from some others. But whether it will be found 

 to agree with Galidia is a matter which cannot as yet be decided. 

 For the latter genus has not been thoroughly examined. It is 

 clear, however, that the form of the csecum is more alike in the 

 two genera than is that of either of them to that of other forms. 

 But I rather gather from Mivart's classificatory scheme that 

 Galidia has not the scent-glands of Galidictis. In this the former 

 genus resembles Eupleres, as it does by the possession of four 

 instead of three {Galidictis) premolars. 



The relationship in fact which Galidictis bears to other Viver- 

 ridse is quite analogous to that which Ezipleres bears to other 

 Yiverridfe. Carlsson has justly commented, after describing the 

 principal facts in the anatomy of that form, upon the deduction 

 that Eiipleres'f seems to be a more ancient type of Viverrine 

 than the other genera, excepting only Nandinia, which the same 



* 9tli ed. 1896. Carlsson (Zool. Jalirb. xvi. Syst. Theil, p. 217) only allows 

 Herpestinee and Viverrinse. 



t Zool. Jalii-b., Abth. f. Syst. xTi. 



