600 THE WHITE COFFEE-LEAF MINER. 
The perfect insect (Fig. 130) is honey yellow, except the eyes, 
ocelli, and sometimes the ocellar space, which are black. The 
antennæ ‘are also black, and consist of twelve slender, sub-equal, 
uniformly cylindrical joints beyond the scape,-which seems to 
consist of two short, stout joints, making fourteen in all. The 
head is transverse; the front projects slightly beyond the eyes; 
the hind margin of the vertex is emarginate, the emargination 
filled with the upper edge of the occiput. The three ocelli are 
arranged triangularly ; in some examples they are approximate, 
in others distant, which may be a sexual difference. Then is 
distinct; the back of the mesothorax sub-hexagonal, rounded, 
tapering anteriorly. In some of my examples the abdomen 
is elongate, subclavate; in others it is rotundate. I think the 
shape given in the figure is the most lifelike. In some examples 
the veins near the extremity of the wings are very feebly devel- 
oped. The wings are fringed. The legs, abdomen, thorax, head, 
antennæ and wings are pubescent. Length 1:9 millimeters; 
expanse 4:7 millimeters. The immature example, or pupa of the 
Bracon, which I found as stated above, bad wing pads instead 
of wings. These were dark, smoky black, 0°5 millimeter long. 
The antennz are honey yellow, instead of black, as in the devel- 
oped specimens. They lie along the breast, and reach to the end 
of the posterior femora, which is about the length of the whole 
body. The first two joints are retracted within a ca oe ce - 
front of the head, which seems to be the result of a doubling ™ 
of the front. Only the two posterior ocelli are visible, distant, 
and though enough of the front remains to contain the me 
ocellus, it seems as if it must be still farther forward a? go 
imago. The hind pair of legs is stuck straight out behind. ie 
abdomen is broken off.) ‘The first and second pair of ke D 
the femur folded forwards; the tibia and tarsus lie pane a 
end of the body. This example lay within the skin of the lav 
with its head towards the head of the larva. 
A possible enemy is a greenish-yellow spider 
the edge of a coffee-leaf on the under side, and spin 
from this edge to the surface of the leaf, leaving eac 
nest open. “ie 
which draws dow? 
ga light web . 
pend 
Geographical Distribution. — As we have stated, M. P inthe 
met these insects in the Antilles; M. Madinier found "a 
island of Martinique; and I observed them in the 
