run in the manner shown in Fig, 3. In this figure the darker portion (n) represents a cross-section of the "sheath," the two lighter pieces (b and d) are cross-sections of the lancets, one (d) in situ. Running the entire length of the "sheath" we find the T-rail projections (g g) along which the lancets slide, being channeled to fit, as seen at b, Fig. 3. Thus we see that the main "sheath" is not a proper term for this part of the apparatus, for the lancets are entirely outside of it and run along the rails. The poison-gland (P) empties into the cylindrical part of the sheath, and keeps it constantly full of virus, at least when the bee is excited. A, B, and C, are broad chitinous pieces, to which muscles are attached. They form a peculiar combination of levers, too complicated in their action to be described here in detail, but they serve to thrust out the "sheath" and the lancets, giving to the former a powerful thrust, and to the latter a movement of great rapidity.
We are now prepared to understand the operation of stinging. The two lancets (K K) when in position lie close against the "sheath," as already described, and their ends reach just to the point of the latter. When the insect stings, the palpi (E E), which are drawn away from their proper place in the figure, serve to direct the organ to the most vulnerable point of attack. Then, with a sudden, powerful motion, the "sheath" is forced out and produces the puncture, penetrating as far as the point d, where the expansion begins. Instantly the two lancets are then forced out together, increasing the depth of the wound made by the "sheath." It has generally been supposed that the lancets were the organs that made the puncture, but this is not the fact. The lancets are thrust out until the stop-valves (p p) strike against the shoulder d (Fig. 1). This closes the cylindrical part of the sheath, which is full of virus, and this virus, being under pressure either from the sudden stoppage of the free outlet by the stop-valves or the contractions of the poison-gland (P), or both these causes combined, makes its way into the tubular lancets through the openings already mentioned just back of the stop-valves, and enters the wound through the branch-tubes b b b b b (Fig. 2). Thus we see that the injection of the poison into the wound is fairly comparable to the working of an hydraulic ram.
When the honey-bee stings, it is well known that the sting is not withdrawn from the wound. The sharp barbs on the lancets make it impossible for the bee to withdraw them, but more than these may be left behind. By allowing the insect to sting a piece of soft leather, not only the lancets but also the sheath and poison-gland will be beautifully dissected out, the bee apparently not suffering from their loss.
It will be seen that the lancets are curved at their attachment with the levers which move them. This curved portion is flexible, while the points are brittle. The poison-gland is provided with a muscular coat. It has been previously supposed that the virus was expelled from the gland by the pressure of other parts. There are several interesting points connected with the mechanism of the sting, which have been
