DIMENSIONS AND STRUCTURAL DETAILS. 93 
stage or another of its employment. We sense two material considera- 
tions in this theme, one qualifying a certain type of weapon and not 
certain others, the other having a geographical element of application, 
or rather one of culture, which we may most conveniently describe in 
the terms of geography. 
The presence of the flange tends toward greater security of the grip 
in the not infrequent case that the weapon is grasped by the opponent 
and effort is made to wrest it from the hand, and a very slight increase 
in circumference is quite as effective toward this security as a great one. 
The amount of this increment of circumference is so irregular that noth- 
ing can be deduced from an examination of the figures. It ranges from 
so scant a sum as 0.25 inch to 2.5 inches; from the light and slender miss- 
ile ula toa very heavy horned club. Yet the increment exhibits no rela- 
tion to the size and weight of the 
pieces. This minimum increment 
and the scarcely differentiated 
TABLE 43. 


0.5-inch increment occur not only in | es), ign oa Disgce: 
the ula, but in clubs so heavy and 
two-handed as the billet, the lapa- 2.75 4 (sword), 6 (dagger). 
lapa, the paddle, and the mush- ey Bree ae 
room. On the other hand, while | 3.25 | 1,3 (swords), 9 (dagger). 
et ae MS 5296, 2, MS 5301 (swords). 
one-third of the billets show the 
flange, not a single one of the root- 
stocks carries it; yet they are of the same provenience and correspond 
in shape, weight, and finish. We are led to the conclusion that this 
detail arises from the cosmetic side. In the very nature of the art of 
fencing the u/a is not exposed to the risk of being wrested from the hand; 
it would have been thrown upon its deadly errand long before the con- 
testants came to grips; and if it were held until it could be seized there 
would be no purpose in clinging toit. This is confirmed by the other 
details of treatment of this piece. The simple u/a with the ball head 
show no instance of flanging; the more decorated wheel type yields 
3 and the highly ornate patterned-head type yields 4 cases of the 
flange. The greatest frequency of the flange is found in the billet, 
the lapalapa, and the paddle clubs; yet these are characteristically 
two-handed weapons and the security which a flange might offer need 
not be considered when regard is had to the greater clutch of two hands 
gripping the club, where torsion and pull may be opposed by the greater 
-leverage thus obtainable. 
When, however, the flange is studied as a culture distinction we 
come to more positive values. In Samoa it occurs in 20 out of 33 
pieces, in Tonga in 21 out of 33, in Fiji in but 13 out of 73 pieces, and 
more than half of the flanges occurring in a type of club not found 
elsewhere in Nuclear Polynesia. These figures yield 60 per cent re- 
spectively for the two Polynesian communities, and for Fiji but 18 per 
