those of the Wallflower? Because, as I have already mentioned, the seeds of the Wallflower are thinner than those of the Mustard, so that there would not be room for the four folds of the doubled leaves.1 In illustration of the influence which the form of the bud exercises on the shape of the leaf I may refer to my paper in the Contemporary Review' for May, 1885, to several memoirs in the Journal of the Linnean Society,' and to my book on Fruits, Flowers, and FIG. 16.-SECTIONS OF SEED OF WALLFLOWER. Transverse R, radicle; C, cotyledon. Leaves.' I will here only mention one or two cases. The leaves of the Tulip Tree (Liriodendron tulipifera) have long attracted attention from the peculiarity of their form. They are saddle-shaped, abruptly truncate at the end; or, in the words of Bentham and Hooker, 'sinuato-4-loba.' I long wondered what could be the purpose or the advantage to the tree of this remarkable 1 This interesting subject is dealt with more fully in my book on Seedlings. shape. One idea which occurred to me was that the difference of form might enable insects to perceive the tree at some distance, just as the colours of flowers are an advantage in rendering them more conspicuous. I then looked closely to see whether the peculiar forms could in any way be explained by the position of the leaves on the tree. I believe, however, that the cause is of a different nature, and has reference to the peculiar character of the bud. a FIG. 17.-DIAGRAM SHOW- sor. Each young leaf is, as in the family Magnoliaceae generally, originally enclosed in and sheltered by the stipules of its predecesThese are in Liriodendron oval, or in form resembling a shallow dish or spoon, so that when placed face to face they form a hollow almond-shaped box. Inside this lies the next younger pair of stipules; and the rest of the space is occupied by the young leaf, which is conduplicate, or folded on itself down the middle, like a sheet of notepaper, and also turned back towards the base of the bud. This unusual position is probably due to the early development of the petiole. The young stages in the development of the leaf are shown in figs. 17 and 18. Here it seems obvious that the peculiar form of the leaf is due to the form and arrangement of the bud. 1 M. Emery has made some criticisms on my suggestions, which, however, I need not notice, for, having merely seen a short and incorrect abstract, and not having taken the trouble to refer to the original paper, he misunderstood my view, as M. Devaux pointed out at the FIG. 18.-ARRANGEMENT OF YOUNG LEAVES OF LIRIODENDRON. A corresponds with leaf No. 6 in C; B, next younger leaf; C, ground-plan of a bud, showing successive leaves (1-7), with stipules (St 5, &c.) time. I should gather also from his remarks that he can never have examined a bud in a very early state. Or take, again, the case of the Oak and Beech. Both have a conical bud, and the young leaves are protected by stipules. I shall presently refer to them again from this point of view. For the moment I am dealing 1 Bull. Soc. Bot. France, 1888, p. 327. with the form of the leaf only. In both species the leaves, like those of Liriodendron, are conduplicate. The leaves are in both species of about equal length, or those of the Oak are the longer; but the buds of the Beech (fig. 19) are much longer than those of the Oak (fig. 20). The young leaves of the Beech are able to lie straight in the bud; those of the Oak have not room to do so, and are, consequently, bent somewhat like a bow along the midrib. Now, if the outer edge were straight, the result would be that when the young leaf emerged and straightened itself the edge must tear. This, however, is avoided by its being thrown into folds; and this, I believe, accounts for the lobes and bays so characteristic of the Oak leaf. DORMANT AND ADVENTIVE BUDS As a rule each leaf has a single bud in its axil; but there are two other classes of buds about which I must say a few words. In some species not merely one, but several buds arise in the axil. They appear to have been In such cases as, for in first noticed by Tristan (3). stance, the Robinia ('Acacia') all but one of these, after attaining a certain stage of growth, become quiescent, and are gradually covered over by the bark. Such buds are known as 'dormant.' They may remain in the same state for years, but if the growing shoots are injured in the struggle for existence these reserves are called into the field. Their presence is often connected with that of thorns, as in Robinia, Gleditschia, Colletia, Genista, &c.; or with tendrils, as in Passiflora. Another class of buds has been termed by Du Petit Thouars' adventive' buds. They appear, especially in certain plants, on the leaves, internodes, and roots. They may be divided into two classes-natural and accidental. Natural adventive buds are formed by Cardamine pratensis, Dentaria bulbifera, Bryophyllum calycinum, C |