GEOGRAPHY OF THE MIGRATIONS. Bis 
which they were forced to lay up for the growing of new crops as 
victual for their voyages. Alite, however, being on the mainland, 
could offer at best but an insecure halting-place, and whatever Poly- 
nesian influence might have made a lodgment would tend to oblitera- 
tion in the greater numbers of the population of the larger island, which 
would resume its sway after the passage of the intruders. 
sufficiently near to be considered in connection with this region of 
strongly marked Polynesian influence in languages distinctly Melane- 
sian, we note the occurrence of almost pure Polynesian settlement in 
Moiki and Moava, the Rennel and Bellona of the charts. We can 
not establish the origin of this settlement. If we accept the possibility 
of a migration track through Torres Strait south of New Guinea it is 
quite consistent therewith to regard Moiki and Moava as a station 
thereupon; the chief objection thereto inheres in the problems of 
navigation and canoe-sailing, and to this objection full weight must 
be given. During the trade-wind season the winds over that tract of 
sea blow most commonly from east and southeast and generally within 
that sector. From any point of departure in the mouth of Torres 
strait or at the nearest of the Louisiade Islands either of these winds 
would be a head-wind and no canoe could lay a course which would 
fetch these two islands. The same objection holds as to a possible 
migration track along the north shore of New Guinea and through 
the Dampier-Vitiaz exit with Kiriwina or Nada as a point of departure. 
No difficulties of seamanship appear when we investigate the possi- 
bility of a Moiki settlement from a highly distinct Polynesian track in 
another direction. Parallel with the northeastern aspect of the chain 
of the Solomon Islands we find at no very great distance offshore the 
islands of the Polynesian Verge, a clearly marked trace of Polynesian 
folk movement established north of the Line in Nukuoro and Kapi- 
ngamarangi, and south of the equator in Tauu, Nuguria, Nukumanu, 
Liuaniua, Sikaiana, Ticopia. We see upon the chart a very attrac- 
tive suggestion. If a straight line be laid down with one end at 
Nukumanu-Liuaniua and the other at Moiki-Moava, it will be seen 
to pass through the Solomon Islands in the fairway between Guadal- 
canar and Malanta, and this fairway is the region in which we have 
found such distinct evidence of Polynesian loan material in the lan- 
guages of Ngela, Vaturanga, and Wango. 
The mere drafting of straight lines upon charts is idle delineation 
unless the lines may be brought into correlation with known forces of 
nature. In this sea area the rhumb of the trade wind is predominantly 
east. This and other meteorological constants are available for study 
in the Pilot Charts of the South Pacific for each quarter of the year. 
It scarcely needs argument to establish the period of canoe voyages 
in the Pacific as conditioned by the prevailing fair-weather season of 
the trades and as avoiding the discomfort and peril of the rainy season 
