580 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
sand stuck on the upper surface of the plank. The rotan projects 
beyond the clay to a distance somewhat greater than the length of 
the cylinder. The mould, bound together with split rotan, is placed 
centrally and vertically over the projecting rotan, thus forming a 
box closed below with clay, open at the top, and having a rotan in 
the centre. Into this the molten metal is poured. When cool the 
rotan is withdrawn, the mold open, and the cylinder is complete. A 
good mould will make three or four castings, but, as a rule, the first 
destroys it. The measurements of the cylinder are: Length, 3} 
inches; width, 4 inch; bore 2 inch. This is the average size; larger 
ones do not work well; smaller ones are of no use.” 
British North Borneo.—The only actual specimen which I have 
from British North Borneo (fig. 37 on pl. v) was sent in 1890 by 
Mr. L. P. Beaufort, who collected it on the west coast, to Sir R. 
Biddulph Martin, who has very kindly given it to me. It is quite a 
remarkable and specialized form, unlike any other which I have seen 
from any part of the east. As in the last-mentioned examples from 
Sarawak, the cylinder is of lead, or possibly lead and tin, cast in a 
bivalve bamboo mold, and decorated at the lower end with faintly 
raised, foliated designs, and at the upper end with punched or incised 
zigzags. The great peculiarity of this example lies in the form of the 
lower end of the cylinder. The base, instead of being flat or rounded, 
is of unsymmetrical form and concave, and just above this is a broad, 
rounded notch on one side. From this notch a perforated duct com- 
municates with the bottom of the bore of the cylinder, very much 
after the fashion of the touchhole and fire duct of an early muzzle- 
loading cannon. The presence of this duct is a most peculiar feature, 
and its raison d’étre is not readily accounted for. It certainly recalls 
to one’s mind those early European and English forms, in which the 
air is violently driven through holes, to which I have already referred, 
and it has occurred to both Mr. Miller Christy and myself that, pos- 
sibly, the tinder was held in the outside notch against the small 
orifice, through which the air was violently driven in a compressed 
state bv the piston, the friction due to passing through the small duct 
being largely responsible for the production of heat. At the same 
time I am disinclined to think that this was the case. The duct is, 
to my mind, far too large for the purpose, and it does not appear to 
have been enlarged at all since it was first made; through such a duct 
the air would escape so easily and quickly when forced through by 
the piston that there would be very little compression or friction, and, 
consequently, very little rise of temperature. The tinder, moreover, 
would almost certainly be blown away. It seems to me more likely 
that the tinder was placed, as usual, on the end of the piston (which 
is, indeed, hollowed out, cup-wise, in the usual manner, evidently 
with this intention), and that when the piston was driven forcibly 
