GEOGRAPHY OF THE MIGRATIONS. 169 
its putative source. If this objection inheres in the argument for 
derivation from the north, New Guinea migration and reversal of 
current into the Gulf of Papua, regard the hypothesis of a Torres 
Strait migration eastward. In this conception of the case Motu and 
its neighboring coast lies as our nearest datum point to an uncon- 
taminated source in proximate southern Indonesia; the degradation 
in quality is progressive eastward in the line of traverse as the wander- 
ing Polynesians tarried with an alien and inferior folk and left their 
impress upon minds poorly equipped to assume the superior culture. 
On this assumption of the independent migration stream eastward 
through Torres Strait we find our difficulties reduced to a minimum; 
we find confirmation in the line of affiliation wherealong quality is 
maintained—a line clearly marked from the Gulf Coast province to 
Suau, to Sariba, to Tubetube, to Panaieti with almost unimpaired 
weight. On this assumption it is not particularly necessary to explain 
away the progressive degradation of Polynesian material on the east 
coast; it may even have come to pass aside from direct Polynesian 
contact, but at second hand through Melanesian interchanges; and 
it is immaterial if we credit the affiliations of Kiriwina-Murua-Nada 
to drift from the nearest coast of Neu-Pommern. 
We are now in a better situation to consider the general probabilities 
of migration through Melanesia as set forth to view on Chart XVI. 
When we examine the possibilities of folk movement in this region we 
encounter three interlacing hypotheses: 
A. 
Assuming the migration eastward of the Proto-Polynesians out of 
Indonesia under a sufficient stress of expulsion (for in the nature of 
the case a draft of attraction to happier lands in the Pacific could not 
exist), if the migrants touched at uninhabited islands we should expect 
to find the laggards of the migration scattered along the course and 
speaking a pure, albeit archaic, Polynesian. Such we find in a clearly 
drawn and interesting track upon the chart. The region nearest to 
Indonesia in which we first identify this migration track is north of 
the equator (not included in this chart), at Kapingamarangi and 
Nukuoro. Other record-points in a generally southeasterly course 
are Tauu, Nuguria, Nukumanu, Sikaiana, the Reef Islands, Ticopia, 
and Anuda, whence the course is clear to Rotuma, Futuna, Uvea, and 
Samoa. Attention has already been directed upon the probability 
of a deflection of some of this migration through the southern Solomons 
to a clear record-point at Moiki-Moava. In the two channels between 
Malanta and Guadalcanar and Malanta and San Cristoval, respec- 
tively, we find more or less high degrees of intermingling with strictly 
Melanesian material, and the same is true of the Reef Islands, Ticopia, 
and Anuda; for in these cases we have other reason to predicate the 
