in flower. So close was the resemblance that, after getting the photograph, I went back to the spot on which they were growing to assure myself that there was no mistake. It cannot be doubted that the true nettle is protected by its power of stinging; and, that being so, it is scarcely less clear that the deadnettle must be protected by its likeness to the other. Moreover, though I was fortunate in lighting on so good an illustration as that shown in the figure just when I had the opportunity of photographing it, still every one must have observed that the two species are very commonly found growing together. Assuming that the ancestor of the deadnettle had leaves. possessing a faint resemblance to those of the true nettle, those in which the likeness was greatest would have the best chance of survival, and consequently of ripening seeds. There would be a tendency, therefore, according to the well-known principles of Mr. Darwin, to a closer and closer resemblance. I am disposed to suggest whether these resemblances may not serve as a protection, not only from browsing quadrupeds, but also from leaf-eating insects. On this part of the subject we have as yet, however, I think, no sufficient observations on record.

Ajuga chamæpitys, the yellow bugle, has leaves crowded and divided into three linear lobes, the lateral ones sometimes again divided. They differ, therefore, greatly from those of its allies, and this puzzled me much until one day I found it growing abundantly on the Riviera among Euphorbia cyparissias, and I was much struck by the curious likeness. The Euphorbia has the usual acrid juice of the genus, and it struck me that the yellow ajuga was perhaps protected by its resemblance.

Leaves which float on the surface of still water tend to be orbicular. The water-lilies are a well-known illustration. I may also mention Limnanthenum nymphæoides, which, indeed, is often taken for a water-lily, though it really belongs to the family of Gentians, and Alisma natans, a species allied to the Plantains. In running water, on the contrary, leaves tend to become more or less elongated.

Subaqueous leaves of fresh-water plants have a great tendency either to become long and grass-like or to be divided into more or less hair-like filaments. I might mention, for instance, Myriophyllum; Hippuris, or mares-tail, a genus which among English plants comes next to Circæa, the enchanter's nightshade; Ranunculus aquatilis, a close ally of the buttercup; and many others.

Some, again, which, when mature, have rounded, floating leaves, have long, narrow ones when young. Thus in Victoria regia the first leaves are filiform, then come one or more which are sagittate, and then follow the great orbicular leaves.

Another interesting case is that in which the same species has

two forms of leaf (fig. 22)—namely, more or less rounded ones on the surface, and a second series which are subaqueous and composed of more or less linear or finely divided segments.

Mr. Grant Allen has suggested that this tendency to subdivision in subaqueous leaves is due to the absence or paucity of carbonic acid. I have ventured to suggest a different explanation. Of course it is important to expose as large a surface as may be to the action of the water. We know that the gills of fish consist of a number of thin plates, which while in water float apart, but have not sufficient consistence to support even their own weight, much less any external force, and consequently collapse in air. The same thing happens with thin, finely cut leaves. In still water they afford the greatest possible extent of surface with the least expenditure of effort in the formation of skeleton. This is, I believe, the explanation of the prevalence of this form in subaqueous leaves.

FIG. 22.

Again, in still air the conditions, except so far as they are modified by the weight, would approximate to those of water; but the more the plant is exposed to wind the more would it require strengthening. Hence, perhaps, the fact that herbs so much oftener have finely cut leaves than is the case with trees. In the Umbellifers, for instance, almost all the species have the leaves much divided-more, I need hardly say, than is the case with trees. Shrubs and trees are characterized by more or less entire leaves, such as those of the laurel, beech, hornbeam, lime, or by similarly shaped leaflets as in the ash, horse-chestnut, walnut.

There are, however, many groups of plants which, while habitually herbaceous, contain some shrubby species, or vice versa. Let us take some groups of this description in which the herbaceous species have their leaves much cut up, and see what is the character of the foliage in the shrubby species.

The vast majority of Umbellifers, as I have just observed, are herbaceous, and with leaves much divided, the common carrot being a typical example. One European species, however, Bupleurum fructicosum, is a shrub attaining a height of more than six feet, and has the leaves (fig. 23) coriaceous, and oblong lanceolate.

The common groundsel (fig. 24), again, is a low herb with much cut leaves. Some species of Senecio, however, are shrubby, and their leaves assume a totally different character, Senecio laurifolius and S. populifolius having, as their specific names denote, leaves respectively resembling the laurel and poplar. In the genus Oxalis,

again, to which the shamrock belongs, there is a shrubby species, O. laureola, with leaves like those of a laurel.

I would venture, then, to suggest these considerations as throwing

light on the reason why herbaceous plants so often have their leaves much cut up.*

Next let me say a few words on the reasons why some plants have broad and some narrow leaves. Both are often found within the limits of a single genus. I have ventured to indicate the distance

between the buds as a possible reason in certain cases. It would not, however, apply to herbaceous genera such as Plantago or Drosera. Now, Drosera rotundifolia (fig. 25) has the leaves nearly orbicular,

Mr. Grant Allen, who had been also struck by the fact that herbaceous plants so often have their leaves much cut up, has suggested a different explanation, and thinks it is due to the fierce competition that goes on for the carbon of the air between the small matted undergrowth of every thicket and hedgerow."

Plantago while in D. anglica (fig. 26), they are long and narrow. media (fig. 27) has ovate leaves, while in P. lanceolata (fig. 28) More or less they are lanceolate, and in P. maritima nearly linear. similar cases occur in Ranunculus.

These differences depend, I believe, on the attitude of the leaf, for it will be found that the broad-leaved ones are horizontal, forming a rosette more or less like that of a daisy, while the species with

FIG. 27.

narrower leaves carry them more

FIG. 28.

or less erect. In the Daisy the rosette lies on the ground, but in other cases, as in Daphne (fig. 29), it is at the end of a branch.

Any one who has looked with an observant eye at the vegetation of hot, dry countries must have noticed how much the general char

FIG. 29.

acter of the vegetation differs from that which prevails in a climate like

ours.

*

There is a marked increase of prickly, leathery, and aromatic species. The first two characteristics evidently tend to protect the leaves. As regards the third, Mr. Taylor,* in his charming book on Flowers, has pointed to the power which, as Tyndall has shown, the spray of perfume possesses to bar out the passage of heat rays, and has suggested that the emission of essential

oils from the leaves of many plants which live in hot climates may serve to protect themselves against the intensely dry heat of the desert sun.

* Page 311.

I am rather disposed to think that the aromatic character of the leaves protects them by rendering it less easy for animals to eat them.

In still dryer regions, such as the Cape of Good Hope, an unusually large proportion of species are bulbous. These, moreover, do not belong to any single group, but are scattered among a large number of very different families: the bulbous condition cannot, therefore, be explained by inheritance, but must have reference to the surrounding circumstances. Moreover, in a large number of species the leaves tend to become succulent and fleshy. Now in organisms of any given form the surface increases as the square, the mass as the cube, of the dimensions. Hence, a spherical form, which is so common in small animals and plants, and which in them offers a sufficient area of surface in proportion to the mass, becomes quite unsuitable in larger creatures, and we find that both animals and plants have orifices leading from the outside to the interior, and thus giving an additional amount of surface. But in plants which inhabit very dry countries it is

necessary that they should be able to absorb moisture when opportunity offers, and store it up for future use. Hence, under such circumstances fleshy stems and leaves are an advantage, because the surface exposed to evaporation is smaller in proportion than it would be in leaves of the ordinary form. This is, I believe, the reason why succulent leaves and stems are an advantage in very dry climates, such as the Canaries, Cape of Good Hope, &c.

The genus Lathyrus, the wild pea, contains two abnormal and interesting species in which the foliaceous organs give the plant an appearance very unlike its congeners. Fig. 30 represents L. niger, with leaves of the ordinary type. In the yellow pea (L. aphaca, fig. 31),