ruit they chew and spit out. I do not know what the tree is,
but I do not think it is the Pandanus, whose fruit is, I believe,
used for body-staining. The yellow stain is obtained from the root of
a plant which I understand to be rather like a ginger. They dry the
root in the sun, and afterwards crush it and soak it in water, and the
water so coloured becomes the pigment to be used. The black stain is
obtained in the same way as that used for face-staining. These dyes are
put on to the cloth with the fingers, which the men dip into the dye,
or with feathers. In making a design they do not copy from a pattern
placed before them, nor do they first trace the design on the cloth.

Netting.

In dealing with netting, I should begin with the making of the string;
but, as I think the method adopted is not confined to the mountains,
it is perhaps sufficient to refer to my previous description of
thread-making in connection with the manufacture of leg-bands;
though in most netting the strings are necessarily very much thicker
and stronger than are the threads used for leg-bands, and they are
three-stranded.

Hunting and fishing nets are made by men in a simple open form of
netting, worked on the common principle of the reef knot, and having
diamond-shaped holes, with a knot at each corner of each hole. I shall
refer to this form of netting as "ordinary network." The nets are made
of thick, strong material, except as regards the hand fishing nets,
which are made of the fine material used for making leg-bands. These
nets are never coloured.

Hammocks are made by men. They are sometimes done entirely with
ordinary network, and are then, I think, similar to Mekeo-made
hammocks; but often only two or three lines of netting are done in
this way, the rest of the net being made in a closer and finer pattern
of interlacing knotless network, which is never adopted on the coast
and Mekeo plains (all nets of this description found there having come
down from the mountains) and which I will call "Mafulu network." [90]
I have watched the making of one of these nets, and will endeavour
to describe the process. The ultimate result of the Mafulu network
part of this is shown in Plate 81.

The maker first formed a base line of three strands of native
string stretched out horizontally. This base line is marked _a b_
in Fig. 8. He then wound a long length of netting string round a
rough piece of stick to be used as a sort of netting shuttle.