490 
THE KILIM A -NJARO EXPEDITION. 
very far from proving to be the Sanskrit of the Bantu. 
Again, other authors, speak of the concord (a feature 
by no means peculiar to the Bantu, but present in 
many other radically different languages, as I have 
pointed out on page 459) as “ euphonic alliteration.” 
This is rubbish. Its origin is clearly not euphony, 
bat lies in the laborious reiteration which in so many 
savage languages is used to obtain certainty of mean¬ 
ing ; and as to alliteration, why in most languages it 
does not exist because the nominal and other prefixes 
no longer correspond in sound. For instance, in 
Lu-ganda, one of the most archaic of Bantu languages, 
we should say— 
Mu 2 ^' gnli mu vundo gu tuka ; tugisale. 
Tree that rotten falls ; let us it cut down. 
= That rotten tree is falling; let us cut it down. 
If it were a question purely of euphonic alliteration 
we should expect to find mu as the concord-prefix 
applied to “Mu ti,” a noun of the third class. But, 
as you see, it is sometimes mu, sometimes gu, and 
sometimes gi. This is explained to the careful student 
who finds by research that the original form of the 
third prefix was ngu, which became changed in some 
situations to mu, and in others to gu or gi. This is not 
only the case in nearly all the Bantu languages, but it 
applies to many of the prefixes, and shows that any¬ 
thing like alliteration does not often occur. 
Another generally enunciated dictum about the 
Bantu languages was that all their roots were poly¬ 
syllabic. This is, however, not the case. While 
some of the simple concepts expressed in words of 
more than one syllable cannot be dissolved any farther 
into their component elements, at the same time many 
most important ideas are expressed by monosyllabic 
