4«4 
BRITISH CENTRAL AFRICA 
up vocabularies I have passed beyond the political limits of British Central 
Africa and have, for the sake of comparison, included tongues spoken in 
other and adjoining spheres. 
The Manyema language which comes first on my list is illustrated by a 
vocabulary supplied to me by Manyema slaves in the Arab settlements on the 
south coast of Tanganyika. Manyema is a remarkable tongue. Its locale is the 
country west.of Tanganyika, on the Congo versant, and its nearest allies are the 
languages spoken on the Upper Congo below (£<?., north of) the Stanley Falls, 
on the Aruwimi River, and south of the Central Congo (Bololo). Manyema 
is an extremely corrupted and worn-down language as will be seen by my 
vocabulary. Some of the sixteen Bantu prefixes are apparently dropped. 
The few that remain are abbreviated almost beyond recognition. 
Ki-guha is a most interesting form of Bantu speech. It is spoken on the west 
and north-west coast of Tanganyika and is allied to Ivi-rega (a language spoken 
in the countries on the Congo Watershed, somewhere to the south of the 
Albert Nyanza), and to Lu-nyoro and Lu-ganda. 
Ki-wemba or Ki-emba is a tongue of very archaic features, especially that 
dialect which is spoken in Itawa, on the south-west coast of Tanganyika. 
The Ki-wemba of the Awemba country further south does not retain so 
many interesting primitive features. Ki-wemba offers points of resemblance 
to the remarkable Nkonde dialects on the north and north-west of Lake 
Nyasa and on a portion of the Nyasa-Tanganyika plateau. Their influence 
extends as far to the N.E. as the south shore of Lake Rukwa, where the Wungu 
language obviously belongs to the Nkonde group. Perhaps Ivi-bisa and other 
languages further to the west are also allied. All these tongues are remarkable 
for retaining the full form of the prefixes when the latter are used as definite 
articles— Umuntu , a man ; Abantu or Awantu , men, and so on. The Ki-lungu 
and Ki-mambwe (with the allied Ki-fipa) languages of South and South-East 
Tanganyika and'the Nyasa-Tanganyika plateau are disappointing in that they 
present few, if any, archaic features; though Ki-lungu is less altered than 
Ki-mambwe and may be connected with the Ki-emba stock. 
The Ki-kese of North-East Nyasa is interesting as offering points of 
resemblance to the Yao as well as to the Nkonde dialects. 
The Bisa language while belonging to the same stem as the Ki-wemba 
evidently represents somewhat closely the original stock from which the 
Tumbuka group (Ci-tumbuka, Ci-henga, Ci-tonga, etc.) and the Ci-senga and 
Ci-nyanja dialects were derived. The Tumbuka and Tonga tongues exhibit 
a phonetic feature almost unheard of elsewhere within the Bantu Family— 
the approximation of .two consonants (s and k) neither of which is a nasal 
or semi-vowel. This however really arises from the gutturalising of a “y” 
sound—what should be pronounced sya becomes sya and ska. 
Ci-nyanja is the dominant language of Nyasaland. It is represented by 
the Ci-nyanja, Ci-cewa, Ci-cipeta, Ci-maravi, Ci-makanga, Ci-mananja, and 
Ci-mbo dialects, but the Senga, Nyungwi (Tete), Sena, and Mazaro languages 
of the Luangwa Valley and the Lower Zambezi are closely allied. In true 
Ci-nyanja the second prefix shrinks to a-, though in some of the north-western 
dialects it is wa-, and in the Machinjiri country south of the River Ruo, 
especially in the hills, there is a suspicion of an aboriginal ba- (in the 
language of Tete it is va-). The Ci-nyanja language is further remarkable 
for the curious changes of the eighth prefix ($/-). In one or two dialects 
this becomes vi-, here and there in remote corners it is byi- or pyi -, but in 
