80 THE ANATOMY 01<' THE IIOJS'EV BEE. 



tion of either gland alone do not die for a long time even in spite of 

 the necessary mutilation, while (3) successive inoculations of the 

 same fly first from one gland and then from the other produce death 

 in a much shorter time than when inoculated from one gland alone — 

 presumably as soon as the two liquids mix within the body. 



The two secretions, one acid and the other alkaline, are poured 

 together into the base of the sting bulb and mix within the cavity 

 of the latter. The resulting poison is then driven through the chan- 

 nel in the shaft to near the tip of the latter, where it makes its exit 

 into the wound. Since the large poison sac is not muscular, the poison 

 is not forced through the sting by it, as is often supposed. A glance 

 at figure 57 (see p. 135) w^ill show that the accessory plates of the sting- 

 support several very comj^act sets of muscles on their inner faces. 

 These muscles so act during the process of stinging that the triangular 

 plates (figs. 3() and 37, Tri) turn upon their hinge-joint articulations 

 with the oblong plates {Oh). By this motion of the triangular 

 plates the attached lancets {Let) are moved back and forth along 

 the tracks on the lower edges of the sheath and its arms (*S7<.i), 

 Each of these tracks consists of a ridge with a constricted base which 

 dovetails into a correspondingly shaped groove on the dorsal surface 

 of the lancet. This structure, as seen in cross sections through the 

 shaft and bulb of the sting, is shown by fig. 40 A, B, and C. The 

 lancets are thus held firmly in place, while at the same time they may 

 slide back and forth with perfect freedom. The figures show also 

 that all three parts of the sting are hollow, each containing a pro- 

 longation {I)() of the bodj^ cavity. Between them, however, is in- 

 closed another cavity through which the poison flows. This is the 

 poison canal {PsnC). In the bulb (fig. 40 C) the body cavity is 

 reduced to a narrow cleft {be) by the great size of the invaginated 

 poison canal {PsnC). 



It will noAv be most convenient to describe the apparatus b}' means 

 of which the poison is ejected from the sting. As before pointed out, 

 the large poison sac can have no functions in this connection because 

 its walls are entirely devoid of muscle fibers. On the other hand, 

 there is an actual pumping apparatus situated within the bulb. This 

 consists of two pouchlike lobes, having their concavities directed 

 posteriorly, attached to the upper edges of the lancets (fig. 40 D and 

 G, VI r) on the anterior ends of the parts of the latter which slide 

 within the loAver edges of the bulb chamber. The lobes lie side by 

 side within the bulb (fig. 40 C, VI v), when the lancets are in the same 

 position, and each has an accessory lamiiui against its own inner wall. 

 When the lancets are pushed backward the walls of the lobes flare 

 apart against the poison contained in the bulb and drive this liquid 

 before them into the channel of the shaft, while at the same time they 

 suck more poison into the front of the bulb from the glands. When, 



