the skin of a reptile, probably the "iguana." The maker
of a drum must climb up the tree from the wood of which he is about
to make it, and there, until the drum is finished, he must remain
sitting among the branches, or, if these are inconvenient for the
purpose, he may erect a scaffold around the trunk of the tree, with
a platform on the top of it, and work upon that. Whilst working,
he must always keep the upper or tympanic end of his drum facing
the wind, the idea of this being that the wind gets into the drum,
and makes it musical. His food is brought to him, whilst in his tree,
by some woman, probably his mother if he is a bachelor, or his wife
if he is married, and he lets down a string by which he hauls it up;
but he is under no special restriction as to the food he may eat. There
is no superstition, such as is found among the Roro and Mekeo people,
compelling him, in the event of his seeing a woman during the making
of the drum, to throw it away and begin a new one.

The jew's-harp (Plate 20, Fig. 2), though seen in Mekeo, is, I was
told, as regards its manufacture, an instrument of the mountains. It
is made out of bamboo or palm, or some other tree having a hollow
or soft interior, from which is cut a piece about 8 or 10 inches
long. A portion of this piece is cut away longitudinally, leaving
for the making of the instrument only two-thirds or half, or even
one-third, of the convex outside stem circumference on one side and
the flat surface of the cut-away part on the other, and the latter
is then hollowed out, leaving, however, a solid head an inch or two
long at one end. The hollow piece thus produced is cut into three
longitudinal sections or strips, of which the two outside ones are
longer than the central one. The two outside strips are left at their
full width from the head downwards to a distance of 2 or 3 inches
from the other end, from which point they are cut away, very much
as one would cut away the divided nib of a quill pen, so that the
actual tips of these two strips are quite slender, being no broader
than their thickness. These two ends are tied together with fine
vegetable fibre. The centre strip, which is generally narrower than
the other two at its commencement by the head, is further reduced in
width by a more immediate and gradual process of paring down, and so
becomes a very slender vibrating tongue or reed, the tip of which goes
almost up to, but does not quite reach, the point at which