IIo CLUB TYPES OF NUCLEAR POLYNESIA. 
form of the bosses in the distal and proximal cylinders, yet this piece 
is carved out of a single piece of wood, and beyond the distal head 
cylinder it reproduces so much of the shaft as in the stone pieces is 
found projecting. It was collected in Port Moresby, on the Gulf of 
Papua, and proves the existence of the wooden metamorph in the very 
region of the stone type. 
Reverting to our ula, we have now to consider the two species of 
head and the diversity of treatment of the shaft upon the proximal 
side of the head. 7 
The wheel-head consists of a series of longitudinal flanges, in num- 
ber 6, 7, 9, and 10 respectively, these flanges ovoid in.the longitudinal 
direction and therein differing from the Malekula club (fig. 20), in 
which the ovoid bosses display the major axis transversely. This 
detail of ornament found in the 5 wheel-head ula is undoubtedly 
continuous into the ornament of the 5 flanged-head rootstock clubs, 
as illustrated in Plate V, figures 5 and 6. It will be noticed that 
these much larger pieces continue likewise the detail of a distinct 
unit of distal projection beyond the flanges. Structurally we have 
seen how such a melon-head may be obtained from the root-ball of 
the sapling, but mere structural facility does not wholly account for 
its existence as a distinct species, whether of ula or of rootstock. It 
seems clear that the source is in the stone head, such as we have found 
in these New Guinea pieces, in figures 13, 14, and 15; if the work be 
stopped after the longitudinal carving has been completed and before 
the transverse carving has been begun we should have a stone ante- 
cedent of the melon type. This is hypothetical, for we have no 
examples of this form in stone. This explanation equally accounts 
for the distal projections in both ula and rootstock, but it leaves 
unexplained the disappearance in the two wooden forms of the proxi- 
mal cylinder. The haft of the wheel-headed ula generally expands 
toward the head, a structural detail which entails considerable diffi- 
culty. If I am correct in my reading of the direction of the seating 
of the stone heads in the New Guinea pieces as from the distal end 
toward the handle, we shall dispose of the difficulty so far as it relates 
to the passage of the head over an expanded portion of the shaft. We 
should then have to prove the existence of a shaft shoulder against 
which the head was seated. On the whole, when we observe the fact 
that in this species the girth of the shaft on the proximal face of the 
head corresponds very closely with the girth of the final knob, we are 
better satisfied with the structural detail shown in the Baining palau 
and the seating of the club in the direction of centrifugal force. Yet 
against this is to be set the fact that in the flanging rootstocks the 
girth of the shaft on the proximal face of the head distinctly exceeds 
the girth of the final knob. A very interesting variety is observed 
in ula 2465, attributed to New Guinea, a provenience which will be 
