and its ch sest Allies. 281 



bat«. The taxononiic and zoogeographical facts, as derived 

 from an examination of this material, may 1)C l)ri(;Hy 

 epitomized as follo\ys : — 



(1) A small tootlied and nai'row j iwcd form, //. nijj'ii' 

 ty/iiciis, occurs fron> j-lrythn a and Kordofan in the north, 

 thron;^h Hritisli Kast Africa, German East Africa, Nyasalaiid, 

 and Lower Zambesi, to Transvaal, Zululand, and Pondoland ; 

 a |)crfe("ly coiitinnons area, comprisini; the eastern side of 

 the continent. I know of no record of any bat of the caj/'er 

 type north of Erythrca or sonth of Tondoland. — From the 

 sonthern part of this area, no doul^t throiij^h the Zambesi 

 Valley, this form has spread to Angola. Tlu-re are parallels 

 to this among other Ethiopian species of llorseshoe-Hats : 

 //. Commersuin, essentially East-African, bnt occurring also 

 in Angola*; Hhinolojj/tua Darli/if/i, distributed from Mazoe 

 to An;:ola f. 



(2 1 A large-toothed and broad-jawed form, H. c. centralis, 

 inliabits the i^piatorial region of the continent, from the 

 Congo Estuary in the uest, tlirough the whole of tlie Congo 

 Valley and Uganda, to British and German East Africa. 



It will be ol)>eived that the geographical areas of these 

 two loruis overlap ea<h other in the east; from the southern 

 ])art of British East Africa to Zomba, i. e. w ithin the area 

 where /f c. /i/picus is the predominant form, we also find 

 H. c. centralis. \Vhen the two forms occur together we 

 might anticipate, in view of their very close relationship, 

 that intermediate s| ceimens would prove to be rather 

 common. Such is, however, not the ease; the two races 

 preserve, also in these circumstances, their peculiarities, so 

 well indeed that, with very rare exceptions, they are distin- 

 guishable at a glance by their cranial and dental characters, 

 I know of only two intermediate examples, the one in the 

 Uritish ^Inseum (from Zanzibar , the other in the Berlin 

 Mn»eum (from German East Afiiea; the type of P/(. rut/ra). 



The reason why //. c. centralis inhal)its a (lart of the area 

 occupied hy //. c. ttjincns is pr d)ab!y this : — All the examples 

 1 have seen I'rom Uganda and the Congo Valley are perfectly 

 clearly pronounced //. c. centralis ; it is theiefore but 

 reasonable to suppose that this '' cttan area," Uganda and 

 the Con(jn Valleij, is the true home of H. c. centralis, and 

 that from there it has s[)read eastwards into British and 

 German I'iist Africa, south-eastwards to Zomba. 



♦ Knud An ler.^en, Ann. it Mag. N. II , Jan. I'JOG, p. 41, footnote, and 

 p. 47. 



t Id., Ann. vt Mag. N. II., Jan. 1005, pp. 71-72; P. Z. P. LiOo, ij. 

 p. 118. 



