GEOGRAPHY OF THE MIGRATIONS. 157 
social area to Melanesian Fiji is not so clearly established as is the case 
in its northern and central determining points, but by inference from 
many custom survivals we are amply warranted in drawing the bound- 
ary as presented upon the accompanying chart. 
With far less accuracy of determination I have felt justified in 
delineating the line of demarcation of the pottery culture. In general 
we inay feel confident that fictile art is a Melanesian possession and is 
quite lacking to the Polynesians. But the presence of pottery is 
quite irregularly spaced within the Melanesian region. ‘The line on 
the chart is provisional only; there is a great paucity of data, and for 
many communities the evidence must be characterized as inconclu- 
sive. It is, therefore, quite certain that this boundary will require 
extensive readjustment. 
In an even higher degree this uncertainty characterizes an important 
culture determinant, the employment of the bow and arrow. I have 
not yet succeeded in differentiating the use of the bow as an offensive 
weapon (in general Melanesian) from the bow as an implement of 
sport or toy of the young (in general Polynesian but with a possible 
exception in the Tuamotu which upsets all calculations). For this 
reason I have not essayed to draw upon the culture chart the curve 
of the bow-and-arrow custom. An excellent beginning of the study 
has been published quite recently by Friederici (Ein Beitrag zur 
Kenntnis der Trutzwaffen der Indonesier, Siidseevélker und Indianer, 
Baessler-Archiv, Beiheft VII, 1915). He presents two charts covering 
Indonesia, Melanesia, and Polynesia, one for the bow and the other 
for the arrow; for the bow he outlines the geographical extension of 
bows of three types, or rather of three linguistic stems, and for the 
arrow he shows two types. 
A wholly anomalous and alien culture element makes a brief but 
distinct and positive entrance into the area under present discussion. 
On this account the curve of the loom culture is of peculiar interest. 
The art of the webster is quite absent from Polynesia; in Melanesia 
it is clearly established in Ndeni and in the northernmost of the Banks 
Group. The occurrence of the loom in islands of the Polynesian 
Verge need not be taken to militate against its absence from Poly- 
- nesian culture in general, for the weaving of Liuaniua and Nukumanu 
is to be regarded as merely establishing a migration stage of a foreign 
art. ‘The nearest locus of the textile art is in Micronesia north of the 
Line, where, in the Caroline Islands, it has attained a high develop- 
ment within certain material limits. For this reason I have decided 
to express the curve of the Verge and Melanesian occurrences of the 
loom and of woven fabrics as opening toward the north in suggestion 
of a probable derivation from the Carolines. We lack any data upon 
which to postulate communication between the Verge islands and the 
Melanesian islands, in which woven fabrics persist from such modern 
