GEOGRAPHY OF THE MIGRATIONS. 173 
least three subdivisions of Melanesia, and underlying the present 
Melanesian in certain of the larger islands a persistence of an earlier 
and yet more primitive folk. But we can not yet prove the limits of 
these subdivisions; perhaps it will be found too late; and though it 
is probable that in more than one region we are dealing with Poly- 
nesian material mutilated by second-hand transmission, we may not 
differentiate it from direct accumulations along the migration sweep. 
Upon this chart I continue the division of the Samoa stream along 
the northern New Guinea coast, the Bismarck Archipelago (and now 
I attach greater importance to the establishment of the line eastward 
of Neu-Mecklenburg on the strength of the Lihir record-point), thence 
through the Solomon Islands and eastward to Nuclear Polynesia. Set 
off from this I draw the Viti stream out through Torres Strait to the 
New Hebrides and thence by the more southerly course to Nuclear 
Polynesia. 
When we take up the investigation of the Polynesian material per- 
sisting in the varied languages of the Malay Archipelago as shown on 
Chart XVII, we feel at once that the problems are somewhat of a 
different order than those which are drawn out from the Melanesian 
nexus. There we have to deal with a folk movement under an impulse 
outward from the rear, led onward along courses whose direction is 
made clear by the most casual study of the charts. Here we find 
ourselves confronted with a notably different set of factors. 
In respect of the mere fact of the persistence of a certain scanty 
yet abundantly recognizable element in Indonesian languages which 
we must recognize as a borrowing from Polynesian, we find no great 
difficulty. We are concerned with roughly about 250 vocables; very 
few indeed are represented at all generally in Indonesian languages. 
Thus we conclude that the borrowing was not at a single center and 
thence carried by the onward movement of the peoples who first 
assumed this material. Rather do we find that this assumption has 
taken place in various proportion at several or many spots where 
Indonesians came into somewhat enduring contact with persons who 
spoke some Polynesian tongue. Furthermore, from the essential con- 
cord of the borrowed material with itself and with the present Poly- 
nesian of the Pacific we hold the opinion that at the time of such con- 
tact the Polynesian spoken in Indonesia was to all intents and pur- 
poses a single speech. 
At this point we require a brief statement of the history of this 
period of contact of the two peoples. Our postulate of the presence 
of Proto-Polynesians in the Indonesian area deduces from the persist- 
ence of this loan material. Against the possibility that Polynesian 
material has been communicated to Indonesia by westward drift we 
set two determinant facts. One is the uniformity of Polynesian 
tradition-history that the original home of the race was in the west, 
