120 CLUB TYPES OF NUCLEAR POLYNESIA. 
The type of socketing is most clearly displayed in 3362 (Plate IV, 
fig. 2), in which we have our best view of the true socket as between 
the single knob and the pair of knobs, extending from the shaft to the 
blade. As in the obsidian spears of the Admiralty Islands, we inter- 
pret the transverse band as the solid central block of the socket sup- 
plying the individual strength which it must possess in order to carry 
its double engagement with shaft and with blade. On the proximal 
side of this block we regard the first unit in its double appearance— 
once on each face of the piece, as slotted for engagement with a broad 
and thin tenon on the end of the shaft, and this tenon I find contin- 
ued in gradually decreasing sharpness of outline in the three pieces; 
viewed in 3362 it is that lateral wing which extends downward from the 
haft to the end of the socket, where it 
is cut off with a sharp angle to the 
blade. From this point onward we 
shall find no difficulty in reconstruct- 
ing the hypothetical tenon as shown in 
the figure (a) drawn from 3362. The 
principal lines of tenon and socket are 
readily identifiable; the pair of dotted 
lines forming approximately a right 
angle opening downward toward ‘the pic. 2.—Reconstruetion of tenoa and 
left are in the hypothesis that bearing socket. 
of the tenon which engages with the slot of the socket; the dotted circles 
represent the positions of two of the knobs which appear on the club. 
By the same process of subtraction, in this instance dissecting the tenon 
from the combined unit, we are able to reconstruct the socket-piece 
(fig. b). The central transverse area is solid, the upper and lower 
units are slotted in the plane of the sketch, and the right-hand edge 
carries the slot around so as to admit upon two bearings the tenon 
within the upper unit and the downward extension of its edge, the 
blade within the lower unit as far as the dotted line. From this very 
clear picture of the two constructive pieces we shall have no difficulty 
in discovering the same elements similarly situated in respect to one 
another in the much ruder pieces; the ruder of the two, 3361, has indeed 
given the clue to the hypothesis of this structural tenon, for it is only 
thus that we can comprehend that initial element which in the detailed 
description of the piece I have characterized as an extra unit. The 
faces of the socket in 3362 are uniformly treated in twine patterns, and 
with this we rehearse from Moseley ‘‘the upper turns of twine are 
arranged in diagonals, etc.’’; the tenon areas which show outside the 
socket, the upper panel within the ridged angle, and the inner wing are 
treated in parallel lines suggestive of twine wrapping; the same treat- 
ment is repeated on the left wing of the blade-socket, probably through 
the symmetrizing principle, for this can not be considered part of the 

