Mr. G. R. Waterhouse on the Rodentia. 83 



line in the longitudinal direction, being but slightly arched, and 

 the plane of the occiput is vertical ; whereas in B. Capensis 

 the plane of the occiput is slightly oblique, receding from the 

 upper part. 



Loiver Jaio, — The lower jaw has the descending ramus (a) 

 Plate II, thrown out from the outer side of the alveolus of the 

 inferior incisor {h). 



In Bathyeryus Capensis the descending ramus approaches 

 somewhat to a semicircular form : its greatest extent is in the 

 longitudinal direction of the jaw, and its greatest depth is be- 

 neath the condyle ; in front it is curved suddenly upwards 

 and inwards so as to form an obtuse angle at (c). 



In Orycterus maritimus the descending ramus is thrown 

 much more boldly out from the alveolar portion of the jaw, 

 and is of very great extent ; the lower part is cun^ed inwards, 

 and the posterior part is produced far beyond the line of the 

 condyle, becoming gradually narrower towards the extremity. 



The coronoid process is small in these two genera, and situ- 

 ated in a line with the last molar. The cond\'loid process is 

 short, and the articular surface is large and rounded. 



As regards the characters furnished by the skull and lower 

 iaw, the present group is very isolated. In the structure of 

 the molar teeth and in the contracted form of the palate be- 

 tween them, Bathyergus approaches most nearly to Spalax^y 

 but in other cranial characters there is a wide difference. 



The skull of the animal figured by Dr. Ruppell under the 

 name Bathyerfjus splendens agrees in most of its characters 

 with that of SpaJax, and in some respects links that genus 

 with Rhizomys ; here the ant-orbital foramen is of moderate 



* I was incliicecl, owing to the remai'kaWe form of the lower jaw, to place 

 Spalax near to Geomys, but upon ve-exaniination I fear I have given too 

 much weight to that character. This genus I have no doubt will prove an 

 aberrant form of the Mitrida;, and I think it will be more correct to regard 

 the Arvlcolidte as constituting a subfamily of that group than as a section of 

 equal importance. Since the publication of the former parts of this paper I 

 have had an opportunity of examining several skulls belonging to species of 

 these sections which I had not before seen. The skull of lihixomys I was 

 only acquainted with through Temminck's figure ; the skull itself I have now 

 examined, and I am quite satisfied that the animal belongs to the family 

 Miiridce, The skull of Ascomys mexicanus I find agrees in all essential 

 characters with that of Geomys (see fig. 71. p. 59G, Mag. Nat. Hist. vol. 

 iii. New Series). The diOerences between Ascomys, Geomys, Diplosloma 

 (Richardson) and Pseudostoma are, it appears to me, not of sufficient im- 

 portance to entitle them to rank as distinct genera. These genera or sub- 

 genera form a very natural little section of the Arvicolidce, differing from the 

 more typical species of that group in having ~ molars, and these small 



and of very simple structure, and also in the absence of any ant-orbital fo- 

 ramen for the transmission of a portion of the masseter uuiscle : the very 

 small slit found at the root of the zygoma in these animals evidently serves 

 only to admit the passage of the ant-orbital nerve, 



G 2 



