By comparing the structure of the leaves, their degree of complication, and their rudimentary parts in the six genera, we are led to infer that their common parent form partook of the characters of Drosophyllum, Roridula, and Byblis. The leaves of this ancient form were almost certainly linear, perhaps divided, and bore on their upper and lower surfaces glands which had the power of secreting and absorbing. Some of these glands were mounted on pedicels, and others were almost sessile; the latter secreting only when stimulated by the absorption of nitrogenous matter. In Byblis the glands consist of a single layer of cells, supported on a unicellular pedicel; in Roridula they have a more complex structure, and are supported on pedicels formed of several rows of cells; in Drosophyllum they further include spiral cells, and the pedicels include a bundle of spiral vessels. But in these three genera these organs do not possess any power of movement, and there is no reason to doubt that they are of the nature of hairs or trichomes. Although in innumerable instances foliar organs move when excited, no case is known of a trichome having such [page 359] power.* We are thus led to inquire how the so-called tentacles of Drosera, which are manifestly of the same general nature as the glandular hairs of the above three genera, could have acquired the power of moving. Many botanists maintain that these tentacles consist of prolongations of the leaf, because they include vascular tissue, but this can no longer be considered as a trustworthy distinction.** The possession of the power of movement on excitement would have been safer evidence. But when we consider the vast number of the tentacles on both surfaces of the leaves of Drosophyllum, and on the upper surface of the leaves of Drosera, it seems scarcely possible that each tentacle could have aboriginally existed as a prolongation of the leaf. Roridula, perhaps, shows us how we may reconcile these difficulties with respect to the homological nature of the tentacles. The lateral divisions of the leaves of this plant terminate in long tentacles; and these include spiral vessels which extend for only a short distance up them, with no line of demarcation between what is plainly the prolongation of the leaf and the pedicel of a glandular hair. Therefore there would be nothing anomalous or unusual in the basal parts of these tentacles, which correspond with the marginal ones of Drosera, acquiring the power of movement; and we know that in Drosera it is only the lower part which becomes inflected. But in order to understand how in this latter genus not only the marginal but all the inner tentacles have become capable of movement, we must further assume, either that through the principle of correlated development this * Sachs, ‘Traité de Botanique’ 3rd edit. 1874, p. 1026.