to protect
the underlying region, a soft-walled region in which the probe has
its source. It is a cuirass, a lid which protects the delicate
motor-machinery during periods of inactivity but swings from back to
front and lifts when the implement has to be unsheathed and used.

We will now remove this lid with the scissors, so as to have the whole
apparatus before our eyes, and then raise the ovipositor with the point
of a needle. The part that runs along the back comes loose without the
slightest difficulty, but the part embedded in the groove at the end of
the abdomen offers a resistance that warns us of a complication which we
did not notice at first. The tool, in fact, consists of three pieces,
a central piece, or inoculating-filament, and two side-pieces, which
together constitute a scabbard. The two latter are more substantial,
are hollowed out like the sides of a groove and, when uniting, form
a complete groove in which the filament is sheathed. This bivalvular
scabbard adheres loosely to the dorsal part; but, farther on, at the tip
of the abdomen and under the belly, it can no longer be detached, as
its valves are welded to the abdominal wall. Here, therefore, we find,
between the two joined protecting parts, a simple trench in which the
filament lies covered up. As for this filament, it is easily extracted
from its sheath and released down to its base, under the shield formed
by the scale.

Seen under the magnifying-glass, it is a round, stiff, horny thread,
midway in thickness between a human hair and a horse-hair. Its tip is a
little rough, pointed and bevelled to some length down. The microscope
becomes necessary if we would see its real structure, which is much less
simple than it at first appears. We perceive that the bevelled end-part
consists of a series of truncated cones, fitting one into the other,
with their wide base slightly projecting. This arrangement produces a
sort of file, a sort of rasp with very much blunted teeth. When pressed
on the slide, the thread divides into four pieces of unequal length. The
two longer end in the toothed bevel. They come together in a very narrow
groove, which receives the two other, rather shorter pieces. These both
end in a point, which, however, is not toothed and does not project as
far as the final rasp. They also unite to form a groove, which fits into
the groove of the other two, the whole constituting a complete channel
or duct. Moreover, the two shorter piec