MAMMALS OF AXGLO-EGYPTL^" SUDAN — SZ7ZZ?. 457 



tail generally longer, hind foot longer, more pronounced posiauricular 

 spots, and the general dorsal color is darker. The skull has a wider, 

 blunter rostrum, a wider interorbital region, braincase more inflated, 

 zygomatic arches more widely flaring, upper toothrow shorter and the 

 auditory bullae more inflated. 



E. r. hoogstmali differs from the type and two topo types of E. r. 

 ddicatu-s in: Color darker: the hairs of the belly white and not washed 

 with buff; in both kinds the hairs of the median line are white to the 

 base instead of being plumbeous. The skuU is somewhat larger in 

 over-all measurements, bullae are markedly more inflated, wings of 

 mesopterygoid more concave than convex, and the rostrum is 

 generally wider. 



From the type of E. r. mariakana^ . E. r. h^K^g^iraali differs in: 

 Color hghter, postauiicular spots markedly lighter in color, white 

 hairs of belly white to bas€ in midventral line and not plumbeous. 

 In the skull the rostrum is narrower, nasals narrower, and the 

 zygomatic arches more rounded and less angular. 



E. r. hoogstraaii differ? from E. r. pha^u^- in: Lifter dorsal color, 

 hairs of belly white to base and not plumbeous. The skull has the 

 rostrum shorter and wider, zygomatic arches less angular, and the 

 upper toothrow is less crowded. 



Remakks: This series of 40 specimens from the Sudan is remark- 

 ably imiform in color and in cranial characters when the sexes are 

 separated and animals of like age are studied. The series was 

 obtained between the middle of Xovember and the end of March. 

 AU of the specimens were taken in a savanna habitat of tall grass. 



It is apparent when studying the types of the various kinds of 

 Elephantulu-s from eastern Africa that the species referred to as 

 E. dunda-i-i is in reality only a subspecies of the earlier E. rufescfns. 

 There are no characters by which this species can be dintinguished. 

 either craniafly or from the skin, from specimens of rufe-^c^n^s from 

 any part of its range. The characters by which it can be distin- 

 guished are no greater than exist between any of the known sub- 

 species. Therefore, although no actual intergradaiion can be 

 demonstrated, the name should stand as Elf phantulu-s rufescen-s 

 dundasi. 



Family Soricid.a^ 



Genus Crocidura Wa^er 



From examining specimens of Crocidura. especially the types, in 

 the British Museum and in the L'. S. National Museum. I am led to 

 beUeve that the species groups of Dollman \,19i5 a-f. 1916 ' are no 

 more than races of the oldest name in each group. There are, cer- 

 tainly, exceptions to the above statement, such as the case of Crocidura 



