The typical leaves of acacias are pinnate, with a number of leaflets. On the other hand, many of the Australian acacias have leaves (or, to speak more correctly, phyllodes) more or less elongated or willow-like. But if we raise them from seed we find, for instance, in Acacia salicina, so called from its resemblance to a willow, that the first leaves are pinnate (fig. 16), and differ in nothing from those characteristics of the genus. In the later ones, however, the leaflets are reduced in number, and the leaf-stalk is slightly compressed laterally. The fifth or sixth leaf, perhaps, will have the leaflets reduced to a single pair, and the leaf-stalk still more flattened, while, when the plant is a little older, nothing remains except the flattened petiole. This in shape, as already observed, much resembles a narrow willow leaf, but flattened laterally, so that it

carries its edge upwards, and consequently exposes as little surface as possible to the overpowering sun. In some species the long and narrow phyllodes carry this still further by hanging downwards, and in such cases they often assume a scimitar-like form. This I would venture to suggest may be in consequence of one side being turned outwards, and therefore under more favourable conditions.

In one very interesting species (Acacia melanoxylon, fig. 17), the plant throughout life produces both forms, and on the same bough may be seen phyllodes interspersed among ordinary pinnate leaves, the respective advantages being, it would appear, so equally balanced that sometimes the one, sometimes the other, secures the predominance.

In the case of the eucalyptus, every one who has been in the South of Europe must have noticed that the young trees have a

totally different aspect from that which they acquire when older. The leaves of the young trees (fig. 18) are tongue-shaped, and horizontal. In older ones, on the contrary (fig. 19), they hang more or less vertically, with one edge towards the tree, and are scimitar-shaped, with the convex edge outwards, perhaps for the same reason as that suggested in the case of acacia. There are several other cases in which the same plant bears two kinds of leaves. species of juniper the leaves are long and pointed, in and scale-like. Juniperus chinensis has both.

Thus, in some others rounded

In the common ivy the leaves on the creeping or climbing stems are more or less triangular, while those of the flowering stems are ovate lanceolate, a difference the cause of which has not, I think, yet been satisfactorily explained, but into which I will not now

enter.

Fro. 18.

FIG. 19.

We have hitherto been considering, for the most part, deciduous trees. It is generally supposed that in autumn the leaves drop off because they die. My impression is that most persons would be very much surprised to hear that this is not altogether the case. In fact, however, the separation is a vital process, and, if a bough is killed, the leaves are not thrown off, but remain attached to it. Indeed, the dead leaves not only remain in situ, but they are still firmly attached. Being dead and withered, they give the impression that the least shock would detach them; on the contrary, however, they will often bear a weight of as much as two pounds without coming off.

In evergreen species the conditions are in many respects different. When we have an early fall of snow in autumn, the trees which still retain their leaves are often very much broken down. Hence, perhaps, the comparative paucity of evergreens in temperate regions, and the tendency of evergreens to have smooth and glossy leaves,

such as those of the holly, box, and evergreen oak. Hairy leaves especially retain the snow, on which more and more accumulates.

Again, evergreen leaves sometimes remain on the tree for several years; for instance, in the Scotch pine three or four years, the spruce and silver fir six or even seven, the yew eight, A. pinsapo sixteen or seventeen, araucaria and others even longer. It is true that during the later years they gradually dry and wither; still, under these circumstances they naturally require special protection. They are, as a general rule, tough, and even leathery. In many species, again, as is the case with our holly, they are spinose. This serves as a protection from browsing animals; and in this way we can, I think, explain the curious fact that, while young hollies have spiny leaves, those of older trees, which are out of the reach of browsing animals, tend to become quite unarmed.

In confirmation of this I may also adduce the fact that while in the evergreen oak the leaves on well-grown trees are entire and smooth-edged like those of the laurel; specimens which are cropped and kept low form scrubby bushes with hard prickly leaves.*

Mr. Grindon, in his "Echoes on Plant and Flower Life" (p. 30), says that "the occurrence of prickles only here and there among plants shows them to be unconnected with any general and ruling requirement of vegetation. We can only fall back upon the principle laid down at the outset, that they are illustrations of the unity of design in Nature, leading us away from the earth to Him who is the end of problems and the font of certainties."" Surely, however, it is obvious that the existence of spines and prickles serves as a protection.

Another point of much importance in the economy of leaves is the presence or absence of hairs. I have already observed that most evergreens are glossy and smooth, and have suggested that this may be an advantage, as tending to prevent the adherence of snow, which might otherwise accumulate and break them down.

The hairs which occur on so many leaves are of several different types. Thus, leaves are called silky when clothed with long, even, shining hairs (silver weed); pubescent or downy, when they are clothed with soft, short hairs (strawberry); pilose, when the hairs are long and scattered (herb-robert); villous, when the hairs are rather long, soft, white, and close (forget-me-not); hirsute, when the hairs are long and numerous (rose-campion); hispid, when they are erect and stiff (borage); setose, when they are long, spreading, and bristly (poppy); tomentose, when they are rather short, soft, and matted; woolly, when long, appressed, curly, but not matted (corn-centaury); velvety, when the pubescence is short and soft to the touch (foxglove); cobwebby, when the hairs are long, very fine, and interlaced

* Bunbury, "Botanical Fragments," p. 320.

like a cobweb (thistle, cobwebby house-leek). The arrangement of the hairs is also interesting. In some plants there is a double row of hairs along the stem. In the chickweed only one. This, perhaps, serves to collect rain and dew, and it is significant that the row of hairs is always opposite to the flower-stalk, which also has a single row. Now, the flower-stalk is for a considerable part of its life turned downwards, with the row of hair outwards. This, perhaps, may account for the absence of hairs on that side of the stem.

Many leaves are clothed with woolly hairs while in the bud, which afterwards disappear. Thus, in the rhododendron, horse-chestnut, and other species the young leaves are protected by a thick felt, which, when they expand, becomes detached and drops off. Many leaves are smooth on the upper side, while underneath they are clothed with a cottony, often whitish, felt. This probably serves as a protection for the sto

mata. In some cases the hairs probably tend to preserve the leaves from being eaten.

In others, as Kerner has suggested, they serve to keep off insects-apparently with the special object of preventing the flowers from being robbed of their

FIG. 20.

Fritz

honey by insects which are not adapted to fertilize them. Müller, to whom we are indebted for so many ingenious observations, gives an interesting case. The caterpillar of Eunomia eagrus, when about to turn into the chrysalis (fig. 20), breaks off its hairs and fastens them to the twig which it has selected, so as to form on each side of itself about half a dozen stiff fences, to protect it during its helpless period of quiescence.

Vaucher long ago observed, though he gave no reason for the fact, that among the Malvaceae (mallows) the species which produce honey are hairy, and those which do not are glabrous.

If we make a list of our English plants, marking out which species have honey and which have hairs, we shall find that we may lay it down as a general rule that honey and hairs go together. The exceptions, indeed, are very numerous, but when we come to examine them we shall find that they can generally be accounted for. I have made a rough list of the species in the English flora which have honey and yet are glabrous. It does not profess to be exactly correct, because there are some species with reference to which I was unable to ascertain by personal examination, or by reference to

books, whether they produced honey or not. My list, however, comprised 110 species.

Now, in the first place, of these 110 species, in sixty the entrance to the honey is so narrow that even an ant could not force its way in; twenty are aquatic, and hence more or less protected from the visits of ants and other creeping insects; thus we shall frequently find that if, in a generally hairy genus, one or more species are aquatic, they are also glabrous-as, for instance, Viola palustris, Veronica anagallis, V. beccabunga, and Ranunculus aquatilis. Polygonum amphibium is peculiarly interesting, because, as Kerner has pointed out, aquatic specimens are glabrous; while in those living on land the base of the leaf produces hairs. Half a dozen are early spring plants which flower before the ants are roused from their winter sleep; about the same number are minute ground plants to which hairs could be no protection; three or four are night flowers; there still remain a few to be accounted for, which would have to be considered individually, but probably the evidence is sufficiently complete to justify the general inference.

Lastly, I must not omit to mention the hairs which have a glandular character.

The next point to which I would call attention is the remarkable manner in which certain forms repeat themselves. In some cases, there seems much reason to suppose that one plant derives a substantial advantage from resembling another. For instance, Chrysanthemum in

odorum, the scentless mayweed, very closely resembles the chamomile in leaves, flowers, and general habit. The latter species, however, has a strong, bitter taste, which probably serves as a protection to it, and of which also, perhaps, the scentless mayweed may share the advantage. These two species, however, are nearly allied to one another, and I prefer, therefore, to take as an example of mimicry the stingingnettle (Urtica) and the common dead-nettle (Lamium album). These two species belong to totally different families; the flowers are altogether unlike, but the general habit and the form of the leaves are extremely similar.

FIG. 21.

How close the similarity is may be seen by the above illustration (fig. 21), taken from an excellent photograph made for me by Mr. Harman, of Bromley. The plants on the right are true stingingnettles; those on the left are the white deadnettle, one of which is