might mention Lathyrus, Genista, Passiflora, Acacia, Spiraea, Saxifraga, Rosa, Berberis, &c.-in which some species have stipules while others have none.

While, then, in most families of plants the species are all either stipulate or exstipulate, there are a good many families in which the genera differ in this respect; some genera in which the species differ; and, lastly, there are some cases, even within the limits of a single species, in which certain leaves are said to develop stipules, and others to be exstipulate, as for instance Helianthemum guttatum, Ipomoea pendula,1 Exochorda Alberti, &c.

Let us now return to Vaucher's problem-Why should some species of Helianthemum have stipules, and others not?-and see whether we can find the answer to it.

Our common Helianthemum (H. vulgare) has stipules; so have H. polifolium, H. tomentosum, H. ægyptiacum, H. rhodanthum, H. ciliare, H. lavandulæfolium, and H. rosmarinifolium. On the other hand, there are a number of species which have no stipules: H. œlandicum, H. lasianthum, H. ocymoides, H. formosum, H. Libanotis, and others. Now, in H. œlandicum (fig. 23) the plant is dwarf and prostrate; the leaves are opposite, oblanceolate or spathulate, tapering to a broad base, sessile, ciliate, especially at the base, which also widens so as closely to sheathe the young bud, thus effectually Choisy, in De Candolle's Prodromus, ix. p. 387.

protecting it. So, also, in the other exstipulate species the leaf-stalks are all more or less widened at the base, and the bud is thus protected.

On the other hand, our common Helianthemum (H. vulgare) (fig. 22) has the leaf-stalk narrowed at the base, and glabrous or nearly so. The bud, therefore, would be exposed if it were not for the stipules. So, also, in the other species which have stipules the leaf-stalks are narrowed at the base, and the buds are protected by the stipules.

Hence we have a very strong presumption that the answer to Vaucher's problem is that some species of Helianthemum have stipules in order to protect the buds; and that in others, where this function is performed by the dilated bases of the leaf-stalk, stipules would be unnecessary, and do not occur.

Helianthemum guttatum is particularly interesting in this respect. The upper leaves have stipules, while the lower ones have none. Now the lower leaves have broad leaf-bases, which effectually protect the bud, and they are exstipulate; the upper leaves, on the contrary, are narrow at the base, but they are provided with stipules.

The reason for the presence or absence of stipules seems, then, quite obvious, so far as the Rock Roses (Helianthemum) are concerned.

Let us now see whether the conformation of other species supports this view. In the allied genus Cistus,

which has no stipules, the petiole is always widened as in the species of Helianthemum without stipules.

This is shown, for instance, in the following figure of Cistus cyprius (fig. 49).

In Cistus vaginatus (fig. 50) the petioles are connate for some distance, so as to form a sheath.

It may, however, be objected that this is hardly a case in point, because none of the species of Cistus have stipules, and the petioles are all widened at the base.

FIG. 49.-SHOOT OF CISTUS CYPRIUS.

Reduced one-half.

S, S, connate sheaths of leaves; A, A, axillary shoots.

There are, however, as already mentioned, some other families of plants in which some species have stipules and others have not.

For instance, in the Caryophyllaceae (the Pink family), the genera Spergula and Spergularia have stipules, while the others are exstipulate. Now Spergula and Spergularia have narrow petioles, while in the other generaDianthus, Tunica, Gypsophila, Saponaria, Silene, Lychnis, Cerastium, Arenaria, Sagina, &c.-containing a very large number of species, the basal part of the leaves is

always more or less connate and sheathing, or the petioles are widened so as to protect the bud.

Again, the Guttiferce have broad petioles and no stipules. Quiina, however, is an exception, for it has stipules, and here also the petiole is narrowed at the

a a, axis; 1, portion of leaf; sh, sheath formed by the dilated and connate
bases of the petioles; ab, ab, axillary buds.

The Rosacea generally have stipules, but in Exochorda some species are exstipulate. In Exochorda Alberti (figs. 53, 54), for instance, the leaves have welldeveloped foliaceous stipules.

On the other hand, Exochorda grandiflora (figs. 51,52) is exstipulate. The petiole, however, is concave and amplexicaul, and the winter-bud is covered with a number of broadly triangular brown scales, some of which show

indications of being tridentate at the apex. These lateral teeth are probably evidence of the former exist

LEAVES OF EXOCHORDA.

FIG. 51.-Leaf of Exochorda grandiflora, nat. size.

FIG. 52.-Transverse section of petiole at 3 mm. from base, x12; US, upper

surface; W, wing.

FIG. 53.-Leaf of Exochorda Alberti, nat. size; St, stipule.

FIG. 54.-Transverse section of petiole at 3 mm. from base, x12; US, upper surface.

ence of stipules. The petiole of the lower leaves also sometimes shows short teeth at its upper end, which also probably represent the last trace of stipules.