GEOGRAPHY OF THE MIGRATIONS. 167 
Dobu again is related with Suau at the first unit of the scale which 
we have been employing, with Sariba at the second, with Tubetube 
at the third unit. When, now, we bring into comparison the Poly- 
nesian chart, we find the picture much simpler and the dominant lines 
reduced greatly in weight. We have the same even curve of Kiri- 
wina-Murua-Nada affiliation, but all the other lines are of the first 
unit and represent mere traces, except the Dobu-Tubetube affiliation 
of equal weight with that which extends through the northern group. 
This graphic delineation of the affiliations of Polynesia and Melane- 
sian languages at this extreme southeastern tip of New Guinea suggests 
any number of interesting problems of small scope which in time will 
undoubtedly repay investigation. In relation to the present work we 
shall rest satisfied with the discussion of the major problem of the 
direction of the movement of speech around this southeastern promon- 
tory and of its connection with the establishment of the larger sweep 
of Polynesian migration into the Pacific. 
We have in the first place to set down a memorandum of the char- 
acter of that element in these charts which we, following our few 
predecessors, find it convenient to designate as Melanesian. The most 
which we may venture to posit of the vocables in this group is that 
they occur in the keeping of peoples whom we for the present elect 
to denominate Melanesian and that they are not identified in any 
language of Polynesia. There exists a certain possibility that some 
of these vocables may belong to the Polynesian language as it was con- 
stituted at the period of the sojourn of Polynesian migrants in the 
place where these words are now found; they may be survivors. That 
such a chance of remote survival conditions the comparison of these 
widely distributed languages is shown in some of our earlier studies. 
In the Visayan of the southern Philippines we find the word alimango 
as the name of a crab; we find the same word in Samoa in the same 
sense; yet if the word had dropped out of Samoan we should have no 
means of knowing that the Visayan word is Polynesian and it must 
have gone on record as purely Indonesian. ‘Therefore it is clear that 
we are not warranted in the statement that this element is wholly 
non-Polynesian merely because it is not now identifiable in the Poly- 
nesian of the present and of which the dictionaries are notably incom- 
plete. According, upon this series of charts we shall distinguish 
between Polynesian and Melanesian only those traces of language 
movement which markedly differ, and where the traces follow the 
same course we shall feel justified in refraining from attaching too 
much importance to the different designation of the two elements. 
In the discussion of language movement as an index of folk move- 
ment around this critical New Guinea point, earlier investigators have 
taken the position more or less distinctly that the movement was 
from the north down the east coast and thence westerly into the Gulf 
