In many flowers, the stamens, as they ripen, move so as to place themselves directly before the entrance to the nectary, where they are necessarily struck by any insect searching for honey; after the pollen is shed, they move aside or bend downward, and their place is taken by the pistil, so that an insect which has come from a younger flower will strike the part of the body previously dusted with pollen against the stigma, and deposit the pollen upon it. This arrangement is very beautifully seen in the nasturtium and larkspur (Fig. 99, _J_). The tubular flowers of the _Sympetalæ_ are especially adapted for pollination by insects with long tongues, like the bees and butterflies, and in most of these flowers the relative position of the stamens and pistil is such as to ensure cross-fertilization, which in the majority of them appears to be absolutely dependent upon insect aid. The great orchid family is well known on account of the singular form and brilliant colors of the flowers which have no equals in these respects in the whole vegetable kingdom. As might be expected, there are numerous contrivances for cross-fertilization among them, some of which are so extraordinary as to be scarcely credible. With few exceptions the pollen is so placed as to render its removal by insects necessary. One of the simpler contrivances is readily studied in the little spring-orchis (Fig. 89) or one of the _Habenarias_ (Fig. 90, _G_). In the first, the two pollen masses taper below where each is attached to a viscid disc which is covered by a delicate membrane. These discs are so placed that when an insect enters the flower and thrusts its tongue into the spur of the flower, its head is brought against the membrane covering the discs, rupturing it so as to expose the disc which adheres firmly to the head or tongue of the insect, the substance composing the disc hardening like cement on exposure to the air. As the insect withdraws its tongue, one or both of the pollen masses are dragged out and carried away. The action of the insect may be imitated by thrusting a small grass-stalk or some similar body into the spur of the flower, when on withdrawing it, the two pollen masses will be removed from the flower. If we now examine these carefully, we shall see that they change position, being nearly upright at first, but quickly bending downward and forward (Fig. 89, _D_, ii, iii), so that on thrusting the stem into another flower the pollen masses strike against the sticky stigmatic surfaces, and a part of the pollen is left adhering to them.