396 Ml. 0. Thomas 07i 



evcrsmanni. Tail with more black in it than in tlie other 

 lorms, the top more or less mixed with black to tiie base, and 

 even the sides and lower surface with an intermixture ot" 

 bkick hairs. 



ISkull as in the largest specimens of erythrouruSj larger 

 than in any of our Samarkand series of eversmanni. Inter- 

 parietal more developed in its antero-posterior diameter, its 

 transverse extent being about the same as in erythrourus. 

 Ijullaj very large, decidedly larger than in eversmanni. 



Dimensions of the type (measured in the flesh) : — 



Head and body 145 mm. ; tail 144; hind foot 35 ; ear 19. 



Skull: greatest length 42*5 ; condylo-incisive length 37*6; 

 zygomatic breadth 33; length of nasals 16"5; interorbital 

 breadth 6*7; interparietal 6x8 7; palatal foramina 7*7; 

 greatest horizontal diameter of bulla 15'5 ; upper molar 

 series 5'8. 



Hab. &, Type as above. 



I can find no previous record of the occurrence of a gerbil 

 of the erythrourus type east of the central mountain-system 

 of Asia ; but there is no doubt that the present animal is 

 allied to that form, though it may be considered as repre- 

 senting a s])ecial subspecies of it. The Transca^ipian M. evers- 

 manni I sliould also regard as a third subspecies of the 

 Atglian form. 



lA. Meriones near kozlovi, Sat, 



cJ . 238. Tur-kul Plateau, Harai Mts., Chinese Turkestan. 

 70UU'. 



1 he type locality of J/, kozlovi is the Lower Kobdo. 



15. Rhomhomys oplmus giganteus, Biichn. 



(J. 233 (immature). Ilami, N.E. Chinese Turkestan. 

 25UU'. 



? . 260, 261, 262. Manass R., Central Dzungaria. 

 1200'. 



c? . 296 ; ? . 294, 295. Barlik Mts. S., N.W. Dzungari;.. 

 5000'. 



Type locality Ebi-Nor, W. Dzungarin, the specimens 

 from the Barlik Mts. being therefore almost topotypical. 



'"'' Gerhillus " turfanensis and "(r." opiums nigresceus are 

 two lurther members of this group described by Dr. Satunin 

 from localities by no means very distant from the present. 

 The evidence of Mr. Carruthers's series does not appear lo 

 increase the probability of these being distinct. 



