148 SISSANO. 
In the present stage of Polynesian speech we find only the surd 
of the palatal mutes, and we observe that a tendency is strongly 
marked to drop this k in many of the languages. That the kappa- 
tion of the t is in many cases succeeding this extinction of the k may 
be taken as evidence of the long existence of this movement to drop 
the palatal. If the Proto-Polynesian vocable had been vaka we 
should expect to find in later forms only vaka and vaa, for this 
consonant entails no difficulty to those who have retained it. In this 
comprehensive survey of the history of the word in the three language 
areas the Fijian now for the first time affords a clue to the past. It 
employs the form wangga, in which we see the sonant mute g pre- 
faced by the nasal of its own series, a mechanical device with which 
we have become quite familiar in the course of these studies. The 
Viti possesses the surd k and finds no difficulty in its employment. 
Accordingly if the vocable in the form in which it was brought to 
Fiji had been vaka we should find waka in Viti. But the Fijian 
has shown himself particularly sedulous in preserving the sounds of 
speech importations, even when they involved consonants quite 
unfamiliar to the usage of his own phonetics. Here we have evidence 
that by his characteristic method of employing the serial nasal pre- 
face the Fijian has done his best to preserve the sonant g alien to his 
speech and has succeeded with most interesting results. This shows 
clearly that the word when brought to Fiji was vaga. From this 
we may argue that Proto-Polynesian was in possession of the sonant 
g, that finding it objectionable the early members of the family 
allowed it to become surd k, and that in later times the same objec- | 
tion to the palatal mute led to its extinction, a process which is still 
operative in living tongues of the family. With this clue from Viti 
we can trace the same movement in Indonesia (46-51) and in Mel- 
anesia (14-15, 17, 19-20, 34-35, and 39). ‘The original stem vaga is 
preserved in Indonesia in 52, and in Melanesia is of very frequent 
occurrence. 
CHECK-LIST OF INDONESIAN LANGUAGES. 
The subdivisions of Indonesia are in no sort in this finding list governed by any scheme 
of linguistic provinces, for the determination of such a distinction has yet to be made. 
Here no object is sought but that of geographical convenience. I have accordingly made 
a division into five general parts as indicated by the following initials. The western 
division denoted by W includes Sumatra, Java, and Bali with the Malay language. The 
central region, C, comprises Borneo and Celebes. ‘The eastern division, E, includes the 
islands between Celebes and the coast of New Guinea, principally Sula, Ceram, Banda 
and Gilolo or Halmaheira. ‘The northern division, N, consists of the Philippines with the 
Sulu Archipelago as a purtenance. The southern division, S, includes the chain from 
Lombok to Timor Laut. 
