In the Broom (Sarothamnus scoparius) the flowers also explode. If, however, the bee alights on a newly-opened blossom, the shorter stamens only emerge and dust the abdomen of the insect. If, on the contrary, the flower is a day or two older, the pistil and longer stamens spring out, and the hairs on the pistil deposit pollen on the bee's back. The pistil gradually curls, and the stigmatic surface turns up, so as to stand close to the anthers of the shorter stamens. In this position it is so placed that it would come in contact with the abdomen of the bee. "Thus," says Mr. Darwin, "both the upper and lower surface of the bee get dusted with pollen, which will be transferred to the stigma at two different periods." (Linn. Jour. v. ix. p. 358.)

In Medicago sativa, as in Genista and the Broom, the flowers open once for all; but the elastic power is confined to the upper stamens. In the Broom and in Genista, the resistance is obtained by the union of the upper edges of the keel. These are also united in Medicago. but even if they are separated no explosion takes place; the flower being still locked together by four processes, two of which point forwards and two backwards. These fit so beautifully that the proboscis of a humble bee is sufficient to unlock them, and release the stamens; though, according to Henslow, (Linn. Jour. 1866, p. 328) the hive bee is unable to do so. Hildebrand, however, has observed that in the absence of insects, it fertilises itself. In M. hipulina the elasticity is much less than in M. sativa. Medicago is a honey-bearing genus.

In the Leguminosae hitherto mentioned, when the keel is forced open, both stamens and pistil emerge from it. In Lathyrus (the Pea), however, this is not the case. In L. pratensis, for instance, the stamens do not leave the keel, but the pistil is provided with a brush of hairs, which sweep the pollen before them.

In the Scarlet Runner (Phaseolus communis) which has been described by Farrer (Ann. and Mag. of Nat. His. 1868, p. 255), the keel is spiral, as well as the stamens and pistil. The former are weak, and never protrude; while the pistil, on the contrary, is stout, strong, and very elastic. In the natural position, the stigma just protrudes out of the mouth of the keel, while the~ terminal portion of the style within the tube is covered with fine hairs. When, therefore, the bee alights on a flower, and inserts her proboscis into it, the stigma will come in contact with the base of the proboscis, and will sweep off any pollen which may be adhering to it. As, however, the bee presses more on the flower, in its efforts to get the honey, the pistil comes further out of the flower; the stigma turns upwards, away from the insect, and the brush of hairs, which has swept the sticky honey out of the anthers, and is consequently covered with it, rubs against the head of the bee and the base of the proboscis, on which it deposits a certain quantity of the pollen, to be again transferred to the stigma of the next flower which the bee visits.

The Common Pea (Pisum sativum) is said not to be well adapted to our British bees. Its structure, probably, has reference to some of the larger southern species.

In Vicia cracca each wing is united to the keel in two places. Though the parts of the flower fit closely to one another, still from the smallness of its size the honey is accessible to most bees; and, owing to the conspicuousness of its bunches, it is much visited by them. From their arrangement and elasticity, the various parts of the flower resume their original position after each visit.

Vicia scpiiun, in general characters, agrees with V. cracca, though the arrangement of the hairs on the pistil is very different. The insects by which it is visited are, however, much fewer. Its larger size, coupled with other minor differences, excludes flies, Lepidoptera, and the smaller bees. Even Bombus ter-rcstris (the Common Humble Bee) does not attempt to suck it, but bites a hole through the side. In V. faba the wings and keel are less closely united, and the honey is more easily accessible. The flower also is less elastic, and if opened widely does not again resume its original form.

It appears then that the Leguminosae are all adapted to fertilisation by bees, and, as Delpino has pointed out, the flowers fall into four series.

1. Those in which the pressure of the bee pumps out, as it were, a certain quantity of pollen; the flower resuming* its original form when the pressure is removed. (Lotus, Anthyllis, Ononis, and Lupinus.)

2. Those in which not only the pollen, but also some of the stamens are pressed out; the flower resuming its form on the removal of the pressure, as in the first division. (Melilotus, Trifolium, Onobrychis.)

3. Those in which the flower bursts on pressure and ejects the pollen. (Medicago, Genista, Sarothamnus.)

4. Those in which, on the pressure of the bee, the pollen is swept out by a brush of hairs situated on the pistil. (Lathyrus, Vicia, Pisum, Phaseolus.)

The power of self-fertilisation seems to be lost in some species of Phaseolus, Onobrychis, and Sarothamnus; and to be much diminished in others, as in Trifolium rep ens and Vicia faba.