JiAXTU l'L.\CE-XA.Mi:.S IX AFRU A. 273 



Xtsoanatsatsi in the norih-cast their dead are buried. They then 

 >treanied across the higliland bridge between the Kalahari and 

 f.inipopo basin till they were stopped by the Drakensberg.* The 

 south-western Bantu may have come by a similar route at no 

 very remote j)eriod, perhaps by a higher crossing of the Zambesi, 

 hut the ])h()netic data are puzzling for the relatitjn of the southern 

 branches, including the easternmost or Zulu-Kanir. Thus SIxosa 

 and SI Zulu share with 'the SEsuto the S, e.g., of this prefix 

 ( Herero oiyi-), and, among other tilings., the prevalence of the 

 laterals (hi, e.g.), which are wanting in most of the other Bantu 

 dialects ; also they share the labialisation in the passive. Yet, on 

 the other hand, Zulu-Kaffir and Herero preserve the Bantu sharp 

 <m-ds K, T, and 1*, which the central group changes 

 into H. i\, and F, etc.i Neither grammatically nor 

 jjlionetically do we find much connection between Herero 

 and Siito, while, strangely enough, there is some phonetic 

 likeness between the latter and Duala in the north-west, opposite 

 Fernando Po. The Herero, again, has vowel-assimilation deve- 

 loi)ed in common with the central (Northern Zambesi) dialects, 

 and to some extent with Suaheli and the north-east Bantu dia- 

 lects. This might seem to show ( i ) a remote branching of 

 Rasuto and Duala possibly north of Tanganyika from one primi- 

 tive offshoot of Bantu, and (2) another branching of south-west 

 from south-east Bantu, possibly west of Nyassa. Thus Haddon 

 says : " The Ovaherero probably represent a branch of the cattle- 

 keeping south-eastern Bantu." but I reach the same conclusion on 

 independent grotnids. The south-eastern Bantu is also extraordi- 

 narily like Suahili, so mtich so that many words and some phrases 

 are practically identical in these extreme dialects of the Ear-t 

 Coast. The closeness of Transvaal \^enda and Tonga, midway 

 between the Suto and the Zulti-Xosa group has to be reckoned 

 with : I am not prepared to say how. Room for borrowing not 

 only of words, but of sound-system and actual dialects must, of 

 course, be left. 



This excursus is not intended even to approach a serious 

 attack on the highly questionable problem of the grouping of the 

 Bantu, but only as a general rough guide in our task of tracking 

 ])lace-names. 



In concluding this digression we may summarise the present 

 distribution of dialects by sa3'ing that in the main the Bantu 

 spread from the lakes Albert and Victoria- w^estward (e.g., the 



* Not a little light is thrown on this Volkerwandernng by the Suto- 

 Chwana names (i.e.. tliosc of the South Central group we are discussing) 

 for N. and S. Both dialects agree on BoRwa (i.e.. Bushmen', driven 

 southward before them to the mountains of Basutoland) for S. : but 

 the S. Basuto use for X. Lc-Boea, from ho boea. to return — i.e., whence 

 they came out. But among the X. Bakwena of Molcpolole, I am told 

 that this word is used for grassless country, and Kgalagali is used for N., 

 showing that these Becoana did not succeed in avoiding the desert after 

 crossing the Zambesi. 



t See author's previous paper. "The studv of .South African X'^ative 

 Languages." Re/^f. .S:A. Assn. for Adv. of Sc. Kimberley (1914), 384- 

 395- 



