GEOGRAPHY OF THE MIGRATIONS. 165 
of such a folk movement should most distinctly appear in the region 
Awalama-Tavara-Suau, with extension to Sariba and in a less degree 
to Tubetube. We have already seen in the Melanesian chart of the 
South Cape how slight is the affiliation between Suau and Sariba with 
Tavara. In the corresponding chart of the east coast we discover at 
once that Tavara is the terminus of a very consistent series of affilia- 
tion which extends northward with much uniformity as far as Wedau 
and in only a slightly less strong phase into the northern subdivision 
of the province. Between Tavara and Awalama essential unity is 
discernible in a little less than half the vocables available for our study. 
This value undergoes accretion as we run over the languages of the 
north shore. The group Awalama-Taupota-Wedau shows such unity 
in an even half of the vocabulary data; in the hamlets of the shore of 
Goodenough Bay, Wedau and Galavi establish a two-thirds unity 
which extends as far as Boniki at the tip of Cape Vogel, the possible 
point of subdivision of this province. From Tavara there are ves- 
tiges of direct affiliation with Taupota and Wedau which has passed 
outside of Awalama. In amount this is a mere trace, and it does not 
extend into the Goodenough Bay country. The Goodenough Bay 
group is established by Wedau-Galavi-Boniki, but we find a trace of 
affiliation as between Wedau and Mukawa at the beginning of the 
northern group, excluding Boniki, and on the Galavi-Mukawa line 
we have considerable confirmatory data. In the northern group we 
find a somewhat strongly drawn line of affiliation between Boniki 
and Mukawa; thence with about half the weight it is prolonged from 
Mukawa to Kubiri, and finally with a reduction of one unit from 
Kubiri to Kiviri. In the Collingwood Bay region we find a fairly 
well-defined interrelation between Kwagila and Kubiri. Of the four 
languages which dominate this bay we find two clearly defined pairs, 
Raga and Oiun as one, the inner pair, Kubiri and Kiviri pairing around 
the other pair. The exterior relations of this province reach in but 
one direction—to Dobu, the nearest record station of the Dentre- 
casteaux province, a trace from Collingwood Bay, a trifle more from 
Goodenough Bay, and somewhat more than one-tenth of the available 
data from the Wedau-Taupota-Awalama-Tavara filiation. ‘The Poly- 
nesian chart shows some interesting points of diversity in addition 
to a marked reduction of the common elements. From Suau as rep- 
resenting the South Cape province we find the filiation with Tavara 
very slight. Between Tavara and Awalama the affiliation is no more 
than half of that seen in the Melanesian language group. From 
Awalama by Taupota to Wedau the amount of affiliation is denoted 
by the same Tavara-Awalama line instead of increasing, as is the case 
upon the Melanesian chart. Above Wedau the affiliation is less than 
between Tavara and Awalama, instead of greater, as in the Melanesian 
chart; and this curve continues out of Goodenough Bay and around 
