The Development of the Water-Glands in the Leaf of Saxifraga Crustata

—oblong obovate, with a crenate margin—are arranged in a rosette on the axis, the internodes of which are hardly at all developed.

In the centre of each of the little crenate lobes with which the leaf is fringed is seen a pit or depression, covered with a white incrustation, which appears to spread from the pit, covering the entire lobe, and even extending over other parts of the leaf as well. The incrustation consists, apparently, of calcic carbonate.

If a transparent preparation be made of the leaf (fig. 1) it is seen that the peripheral terminations of the fibro-vascular bundles which ramify through the mesophyll present the appearance of a number of dilatations. Each such dilatation is placed immediately under the bottom of the depression of each lobe, and constitutes what is known as the watergland.

The mature glands of this plant were first described by Unger, ¹ but, so far as I am aware, their development has not been investigated, and it is the object of the following paper to deal with this latter question. Since the development of the leaf and its tissues are necessarily very closely connected with that of the glands, it seems best to consider first the development of the leaf as a whole, and then that of the glands.

The punctum vegetationis of Saxifraga crustata is a hemispherical mass of meristematic cells, the external layer of which is differentiated as a dermatogen (fig. 2). The first rudiment of the leaf appears as a lateral outgrowth of this primary meristem, covered, of course, by dermatogen. At first the rudimentary leaf enlarges by rapid apical growth, and thus the first or terminal lobe is formed; with its formation apical growth ceases. The further elongation of the leaf takes place by the activity of a zone of meristematic cells at its base. Hence, its subsequent growth is basipetal, i. e. the youngest lobes are nearest the base. But although the growth of the leaf, as a whole, is basipetal, each pair of lateral lobes arise and grow from the basal meristematic zone in an apical manner, so that their development is quite similar to that of the primary lobe (fig. 3). Whilst the apical growth of the terminal lobe is still proceeding a strand of cells is differentiated in the median line of the leaf; they are small and elongated, their nuclei are very conspicuous and elongated in form, their protoplasm is granular and stains deeply with haematoxylin, and their cell walls are very delicate. The differentiation proceeds from the base towards the apex, although it may be said to be almost simultaneous. This strand of cells is evidently a procambium.

At the same time that this differentiation has been going on in the leaf, strands of cells have become differentiated in the stem, just behind the punctum vegetationis, in such a way that the apex of one of these strands of the stem is on a level with the base of the leaf, and by a differentiation in the few intervening cells the procambium of the stem comes to be continuous with that of the leaf.

Like the primary lobe, the lateral lobes consist at first of meristem covered by a dermatogen. Here, too, the procambium becomes differentiated as strands of similar cells, exactly resembling those first described. The procambium of a lobe may be directly continuous with that of the main leaf-bundle, or may join with one or more strands from contiguous lobes before so doing. The differentiation of primary meristem into procambium, in fact, does not, as a rule, follow any definite course, but starts at particular centres and in various ways, thus giving rise to the reticulate distribution of the fibro-vascular bundles in the fully-developed leaf.

The next step is the conversion of the procambium into permanent tissue. A single row of spirally thickened protoxylem cells makes its appearance in the centre of the median procambium strand. The further conversion of the meriste-tematic into permanent tissue takes place centrifugally.

There is a definite but slight increase in the number of the elements of the main fibro-vascular bundle from the apex towards the base. The zone of meristem cells at the base of the leaf is traversed by the primary fibro-vascular bundle, and as the lateral bundles are differentiated they become continuous at their proximal extremities with the peripheral portion of the primary fibro-vascular bundle, which is still of the nature of procambium. The fact that the conversion of procambium into permanent tissue takes place in a centrifugal manner allows of cell fusion so long as any procambiutn remains, and enables points of junction to be established. The last-formed fibro-vascular bundles become connected in the same manner with the main lateral branches.

An examination of the thickest part of the midrib, as seen in a transparent preparation of the leaf, enables one to make out very clearly the comparative ages of the constituent elements of which the whole bundle is composed. The vessels in which the turns of the spiral thickenings are approximated most closely can always be traced to the last-formed or youngest lobes, whereas those vessels in which they are most widely separated belong to the primary fibro-vascular bundle, and between these two extremes there is every possible gradation. This also proves conclusively that the vessels are capable of growth after the formation of the spiral thickening. The xylem and phloem are arranged on the collateral type, and the bundle is surrounded by a well-developed endodermis..

Having thus followed out the difierentation of, the tissues of the leaf, we may proceed to follow out the development of the water-glands. And since what is true of one gland is true of all, it will suffice to study the development of the first-formed or terminal gland. This can be most satisfactorily done by a series of longitudinal sections of the leaf in various stages of growth (fig 2).

As before stated the rudimentary leaf consists of undifferentiated meristematic cells covered by a dermatogen. When the differentiation of the procambium bundles has taken place, the meristematic ground-tissue cells at its apex divide, forming a mass of closely packed polygonal cells with delicate cell walls and very conspicuous nuclei. This is the first indication of what will hereafter be the gland. The further conversion of the primary meristem into gland-tissue takes place from the point of its first occurrence, towards the upper side of the leaf, finally ending beneath the dermatogen. At the inner extremity of the gland towards the fibro-vascular bundle, and at the lower third of its under surface (i. e. the surface corresponding to the lower side of the leaf) certain cells are conspicuous as becoming more elongated and fusiform than the rest and as subsequently presenting reticulate thickenings on their walls. They are joined end to end, and eventually become continuous with the spiral vessels of the fibro-vascular bundle. These cells present a series of intermediate forms between reticulate cells and spiral vessels. At the same time changes have been going on in the primary ground-tissue, for those cells immediately next the gland have become differentiated in a regular and definite manner so as to form a sheath of one or at most two layers of cells, entirely surrounding the gland with the exception of that portion of it which is covered by epidermis. This sheath is continuous with the endodermis of the bundles of the leaf.

One, two, or rarely three, of the cells of the dermatogen, where it is in immediate relation to the outer end of the gland, increase in size. Each is distinguished by a large and distinct nucleus, and each is the mother-cell of a water-pore (fig. 4).

The nucleus of this cell divides and a wall is formed, dividing the cell into two parts. The two cells thus produced separate from one another, leaving a hole or opening between them. This is the water-pore or water-stoma. It thus consists of two guard cells, which when fully developed contain chlorophyll. There is a difference in the time of formation of a water-pore as contrasted with that of the ordinary stomata. Thus the water-pores are fully developed before any trace can be detected of the ordinary stomata. This difference in point of time is exhibited in a still more striking manner by Crassula coccínea, in which the waterpores may be seen completely formed, when the cells of the epidermis are just beginning to divide to form stomata in the peculiar and complicated manner so characteristic of the Crassulacse. And not only does the water-pore differ from the stoma in its time of development, but also in size, shape, and mode of development (figs. 8 and 9). It is slightly larger than the stoma. Its contour is rounded as compared with the elliptical shape of the ordinary stomata, the breadth being greater in the former than in the latter case. The actual pore or opening of the stoma is, however, larger than the water-pore.

The water-pore is thus produced by the simple division into two equal halves of a cell of the dermatogen, which has become larger than its neighbours. This is well shown by fig. 4, which is a surface view. But the ordinary stoma is formed in a more complicated manner. One of the cells of the dermatogen divides. Of the two cells thus produced one becomes the mother-cell of the stoma. It increases in size, and after previous division of the nucleus, a wall makes its appearance in nearly every case at right angles to the plane of the first division, and thus forms the two guardcells of the stoma. The second cell of the two formed in the first instance becomes somewhat displaced in the course of subsequent growth. It never seems to grow as large as the adjoining epidermal cells, and may always be recognised in the fully developed condition as interfering with the arrangement of the four or five epidermal cells which would otherwise surround the stoma in a symmetrical manner (figs. 9 and 10).

At the same time that the mother-cells of the water-pores are making themselves apparent by their increased size, outgrowths have arisen on the surfaces of numerous dermatogen cells. These are the rudiments of the small knob-like hairs which are borne in the mature leaf on the sides of the depression, which, as has already been mentioned, marks the position of the gland. They are especially numerous on the outer side of the lobe, and when fully developed are almost perfectly spherical, with very thick and very highly refractive cell walls (fig. 5). Celldivision though not cell-growth of the water-gland tissue practically ceases with the division of the mother-cell of the water-pore; but in the leaf, cell-division still goes on, ceasing only with the formation of stomata. As a consequence of this, depressions are produced over the glands, the sides of which are fringed by the delicate highly refractive hairs.

By far the larger portion of the gland-tissue is made up of polygonal cells, slightly longer than broad, closely fitting one to the other, with no intercellular spaces. The cell-walls are thin and the cells themselves are much smaller than those of the surrounding ground tissue. The protoplasm is very granular. At first it fills the cells entirely, with the nucleus in the centre. When mature a central vacuole makes its appearance, and the protoplasm then forms a thick layer, which closely invests the cell wall; the nucleus is either imbedded in the protoplasm, or it is suspended in the vacuole by strands of protoplasm, its position being no longer central but eccentric (see fig. 7). The whole gland is roughly pearshaped, broad towards the surface of the leaf, tapering inwards towards the fibro-vascular bundle (figs. 5 and 6). It is invested by a sheath of cells continuous with the endodermis of the fibro-vascular bundle, except where it is in contact with the epidermis. The endodermis cells are long and cylindrical with moderately thick walls. They contain numerous chlorophyll granules. External to this is the mesophyll of the leaf. Free communication between the gland tissue and the external mesophyll is afforded by means of the one or more water-pores, the development and structure of which we have previously studied.

Since the aim of this paper is to regard the gland solely from a histological standpoint, the mode of action will be only glanced at in a very cursory manner. In the daytime, when the temperature is high and transpiration very vigorous, the secretion of drops of water does not take place, and even if water be secreted by the gland cells it is evaporated as fast as formed. But at night, when the temperature falls, causing decrease of transpiration, the walls of the vessels and even the vessels themselves become gorged with water, and every facility for secretion by the gland is offered. Then drops exude through the water-pores., subsequently filling the pit. Now the portion of the lobe which intervenes between the gland and the edge of the leaf slopes downwards and outwards, and it is on this side especially on which the hairs surrounding the pit are most abundant. As a consequence, when an excess of water is secreted and the pit overflows, the water will tend to collect on the margin of each lobe. The water, charged as it is with carbonic acid, holds in solution a quantity of calcic carbonate. As the water evaporates and the carbonic acid gas comes off, the chalk is precipitated, and that which is deposited in the vicinity of the pit, tends to aggregate around the hairs and becomes thus firmly held and prevented from falling into the pit and stopping up the water-pores. In spite of this special provision, however, the older glands frequently become inefficient on account of being choked, the pit becoming completely filled with the very large concretions if chalk formed.

The whole phenomenon may be easily seen by placing a bell jar over a vigorous plant. Since the air becomes saturated with aqueous vapour, transpiration is reduced and large drops of water are secreted. On removing the jar, the water rapidly evaporates and a deposit of chalk is formed.

As far as I have had an opportunity of examining water-glands, it appears that the gland of Saxifraga crustata is the most highly differentiated of all as regards the special provision made for the deposition of the calcic carbonate, the distinct differentiation of the gland tissue, the well-marked endodermis, the extreme granularity of the protoplasm, and the activity of function.

Next come the glands of the Crassulas, where there are neither hairs nor endodermis. The protoplasm is not nearly so granular, neither is the activity of function so great.

Between the glands of the Saxifragacese and Crassulacese and those occurring in the rest of the vegetable kingdom, there appears to be a great gap. We look in vain for the well-differentiated gland and the conspicuous difference between the gland-cells and the adjacent parenchyma. But the water-pore long preserves its individuality, although in many cases it seems probable that its function is taken on by an ordinary stoma, of which, indeed, even in its most differentiated form, it is but a modification.

It is interesting to note that the diminution of gland-tissue is often accompanied by an increase of fibro-vascular tissue, which appear to replace it. Comparing, for instance, Saxifraga crustata and Crassula arborescens, the reticulated cells at the base of the gland are much more numerous in the latter than in the former, and this seems to be a general tendency in the less highly differentiated glands.

As to the position of the glands, it seems to be a rule that they occur on the margin of the leaf. The exceptions to this statement are afforded by Crassula cordifolia, Crassula arborescens, and Crassula portulacea, where they occur distributed over the surface. In leaves with an entire margin, they usually occur only at the apex, i. e. the extremity of the main fibro vascular bundles, as in species of Azalea, Myosotis. In leaves whose margins are cut or indented there may be a gland at the apex of each tooth, as in Primula sinensis, Fuchsia globosa, Alchemilla vulgaris, between two teeth, in the indentation, as in Crassula spathulata, or even on the sides of the tooth, as in Senecio petasitis.

In species of Saxifraga they appear to occur only on the upper side of the leaf. In Crassula lactea and Crassula coccinea on the upper and under margin, and in Sedum Sieboldii only on the under side.

The number of water-pores in relation with each gland varies very much. In Saxifraga crustata 2—3, Bryo-phyllum calycinum 5—6, Crassula coccinea 1—3, Crassula spathulata 15—20, Crassula lactea 15—18, Primula sinensis 1—2, Hordeum vulgare 1—2.

In by far the larger number of plants the activity and life of the water-gland and structures analogous to it appear to be co-existent with that of the leaf as a whole. In certain cases, however, e.g, species of Musa, Richardia, and grasses, in which the gland is borne at the apex of the leaf, the apex soon withers and the gland becomes destroyed. Thus, in very young barley plants (Hordeum vulgare) large drops of water may be seen hanging on the apex of each leaf, even though the temperature of the surrounding air be as much as 74° Fahr. As they grow older the secretion stops, and later on each leaf-apex withers up and dies.