CUCURBITACEAE.—The tendrils in this family have been ranked by competent judges as modified leaves, stipules, or branches; or as partly a leaf and partly a branch. De Candolle believes that the tendrils differ in their homological nature in two of the tribes. {29} From facts recently adduced, Mr. Berkeley thinks that Payer's view is the most probable, namely, that the tendril is "a separate portion of the leaf itself;" but much may be said in favour of the belief that it is a modified flower-peduncle. {30}

Echinocystis lobata.—Numerous observations were made on this plant (raised from seed sent me by Prof. Asa Gray), for the spontaneous revolving movements of the internodes and tendrils were first observed by me in this case, and greatly perplexed me. My observations may now be much condensed. I observed thirty-five revolutions of the internodes and tendrils; the slowest rate was 2 hrs. and the average rate, with no great fluctuations, 1 hr. 40 m. Sometimes I tied the internodes, so that the tendrils alone moved; at other times I cut off the tendrils whilst very young, so that the internodes revolved by themselves; but the rate was not thus affected. The course generally pursued was with the sun, but often in an opposite direction. Sometimes the movement during a short time would either stop or be reversed; and this apparently was due to interference from the light, as, for instance, when I placed a plant close to a window. In one instance, an old tendril, which had nearly ceased revolving, moved in one direction, whilst a young tendril above moved in an opposite course. The two uppermost internodes alone revolve; and as soon as the lower one grows old, only its upper part continues to move. The ellipses or circles swept by the summits of the internodes are about three inches in diameter; whilst those swept by the tips of the tendrils, are from 15 to 16 inches in diameter. During the revolving movement, the internodes become successively curved to all points of the compass; in one part of their course they are often inclined, together with the tendrils, at about 45 degrees to the horizon, and in another part stand vertically up. There was something in the appearance of the revolving internodes which continually gave the false impression that their movement was due to the weight of the long and spontaneously revolving tendril; but, on cutting off the latter with sharp scissors, the top of the shoot rose only a little, and went on revolving. This false appearance is apparently due to the internodes and tendrils all curving and moving harmoniously together.

A revolving tendril, though inclined during the greater part of its course at an angle of about 45 degrees (in one case of only 37 degrees) above the horizon, stiffened and straightened itself from tip to base in a certain part of its course, thus becoming nearly or quite vertical. I witnessed this repeatedly; and it occurred both when the supporting internodes were free and when they were tied up; but was perhaps most conspicuous in the latter case, or when the whole shoot happened to be much inclined. The tendril forms a very acute angle with the projecting extremity of the stem or shoot; and the stiffening always occurred as the tendril approached, and had to pass over the shoot in its circular course. If it had not possessed and exercised this curious power, it would infallibly have struck against the extremity of the shoot and been arrested. As soon as the tendril with its three branches begins to stiffen itself in this manner and to rise from an inclined into a vertical position, the revolving motion becomes more rapid; and as soon as the tendril has succeeded in passing over the extremity of the shoot or point of difficulty, its motion, coinciding with that from its weight, often causes it to fall into its previously inclined position so quickly, that the apex could be seen travelling like the minute hand of a gigantic clock.

The tendrils are thin, from 7 to 9 inches in length, with a pair of short lateral branches rising not far from the base. The tip is slightly and permanently curved, so as to act to a limited extent as a hook. The concave side of the tip is highly sensitive to a touch; but not so the convex side, as was likewise observed to be the case with other species of the family by Mohl (p. 65). I repeatedly proved this difference by lightly rubbing four or five times the convex side of one tendril, and only once or twice the concave side of another tendril, and the latter alone curled inwards. In a few hours afterwards, when the tendrils which had been rubbed on the concave side had straightened themselves, I reversed the process of rubbing, and always with the same result. After touching the concave side, the tip becomes sensibly curved in one or two minutes; and subsequently, if the touch has been at all rough, it coils itself into a helix. But the helix will, after a time, straighten itself, and be again ready to act. A loop of thin thread only one-sixteenth of a grain in weight caused a temporary flexure. The lower part was repeatedly rubbed rather roughly, but no curvature ensued; yet this part is sensitive to prolonged pressure, for when it came into contact with a stick, it would slowly wind round it.

One of my plants bore two shoots near together, and the tendrils were repeatedly drawn across one another, but it is a singular fact that they did not once catch each other. It would appear as if they had become habituated to contact of this kind, for the pressure thus caused must have been much greater than that caused by a loop of soft thread weighing only the one-sixteenth of a grain. I have, however, seen several tendrils of Bryonia dioica interlocked, but they subsequently released one another. The tendrils of the Echinocystis are also habituated to drops of water or to rain; for artificial rain made by violently flirting a wet brush over them produced not the least effect.

The revolving movement of a tendril is not stopped by the curving of its extremity after it has been touched. When one of the lateral branches has firmly clasped an object, the middle branch continues to revolve. When a stem is bent down and secured, so that the tendril depends but is left free to move, its previous revolving movement is nearly or quite stopped; but it soon begins to bend upwards, and as soon as it has become horizontal the revolving movement recommences. I tried this four times; the tendril generally rose to a horizontal position in an hour or an hour and a half; but in one case, in which a tendril depended at an angle of 45 degrees beneath the horizon, the uprising took two hours; in half an hour afterwards it rose to 23 degrees above the horizon and then recommenced revolving. This upward movement is independent of the action of light, for it occurred twice in the dark, and on another occasion the light came in on one side alone. The movement no doubt is guided by opposition to the force of gravity, as in the case of the ascent of the plumules of germinating seeds.

A tendril does not long retain its revolving power; and as soon as this is lost, it bends downwards and contracts spirally. After the revolving movement has ceased, the tip still retains for a short time its sensitiveness to contact, but this can be of little or no use to the plant.

Though the tendril is highly flexible, and though the extremity travels, under favourable circumstances, at about the rate of an inch in two minutes and a quarter, yet its sensitiveness to contact is so great that it hardly ever fails to seize a thin stick placed in its path. The following case surprised me much: I placed a thin, smooth, cylindrical stick (and I repeated the experiment seven times) so far from a tendril, that its extremity could only curl half or three-quarters round the stick; but I always found that the tip managed in the course of a few hours to curl twice or even thrice round the stick. I at first thought that this was due to rapid growth on the outside; but by coloured points and measurements I proved that there had been no sensible increase of length within the time. When a stick, flat on one side, was similarly placed, the tip of the tendril could not curl beyond the flat surface, but coiled itself into a helix, which, turning to one side, lay flat on the little flat surface of wood. In one instance a portion of tendril three-quarters of an inch in length was thus dragged on to the flat surface by the coiling in of the helix. But the tendril thus acquires a very insecure hold, and generally after a time slips off. In one case alone the helix subsequently uncoiled itself, and the tip then passed round and clasped the stick. The formation of the helix on the flat side of the stick apparently shows us that the continued striving of the tip to curl itself closely inwards gives the force which drags the tendril round a smooth cylindrical stick. In this latter case, whilst the tendril was slowly and quite insensibly crawling onwards, I observed several times through a lens that the whole surface was not in close contact with the stick; and I can understand the onward progress only by supposing that the movement is slightly undulatory or vermicular, and that the tip alternately straightens itself a little and then again curls inwards. It thus drags itself onwards by an insensibly slow, alternate movement, which may be compared to that of a strong man suspended by the ends of his fingers to a horizontal pole, who works his fingers onwards until he can grasp the pole with the palm of his hand. However this may be, the fact is certain that a tendril which has caught a round stick with its extreme point, can work itself onwards until it has passed twice or even thrice round the stick, and has permanently grasped it.

Hanburya Mexicana.—The young internodes and tendrils of this anomalous member of the family, revolve in the same manner and at about the same rate as those of the Echinocystis. The stem does not twine, but can ascend an upright stick by the aid of its tendrils. The concave tip of the tendril is very sensitive; after it had become rapidly coiled into a ring owing to a single touch, it straightened itself in 50 m. The tendril, when in full action, stands vertically up, with the projecting extremity of the young stem thrown a little on one side, so as to be out of the way; but the tendril bears on the inner side, near its base, a short rigid branch, which projects out at right angles like a spur, with the terminal half bowed a little downwards. Hence, as the main vertical branch revolves, the spur, from its position and rigidity, cannot pass over the extremity of the shoot, in the same curious manner as do the three branches of the tendril of the Echinocystis, namely, by stiffening themselves at the proper point. The spur is therefore pressed laterally against the young stem in one part of the revolving course, and thus the sweep of the lower part of the main branch is much restricted. A nice case of co-adaptation here comes into play: in all the other tendrils observed by me, the several branches become sensitive at the same period: had this been the case with the Hanburya, the inwardly directed, spur-like branch, from being pressed, during the revolving movement, against the projecting end of the shoot, would infallibly have seized it in a useless or injurious manner. But the main branch of the tendril, after revolving for a time in a vertical position, spontaneously bends downwards; and in doing so, raises the spur-like branch, which itself also curves upwards; so that by these combined movements it rises above the projecting end of the shoot, and can now move freely without touching the shoot; and now it first becomes sensitive.

The tips of both branches, when they come into contact with a stick, grasp it like any ordinary tendril. But in the course of a few days, the lower surface swells and becomes developed into a cellular layer, which adapts itself closely to the wood, and firmly adheres to it. This layer is analogous to the adhesive discs formed by the extremities of the tendrils of some species of Bignonia and of Ampelopsis; but in the Hanburya the layer is developed along the terminal inner surface, sometimes for a length of 1.75 inches, and not at the extreme tip. The layer is white, whilst the tendril is green, and near the tip it is sometimes thicker than the tendril itself; it generally spreads a little beyond the sides of the tendril, and is fringed with free elongated cells, which have enlarged globular or retort-shaped heads. This cellular layer apparently secretes some resinous cement; for its adhesion to the wood was not lessened by an immersion of 24 hrs. in alcohol or water, but was quite loosened by a similar immersion in ether or turpentine. After a tendril has once firmly coiled itself round a stick, it is difficult to imagine of what use the adhesive cellular layer can be. Owing to the spiral contraction which soon ensues, the tendrils were never able to remain, excepting in one instance, in contact with a thick post or a nearly flat surface; if they had quickly become attached by means of the adhesive layer, this would evidently have been of service to the plant.

The tendrils of Bryonia dioica, Cucurbita ovifera, and Cucumis sativa are sensitive and revolve. Whether the internodes likewise revolve I did not observe. In Anguria Warscewiczii, the internodes, though thick and stiff, revolve: in this plant the lower surface of the tendril, some time after clasping a stick, produces a coarsely cellular layer or cushion, which adapts itself closely to the wood, like that formed by the tendril of the Hanburya; but it is not in the least adhesive. In Zanonia Indica, which belongs to a different tribe of the family, the forked tendrils and the internodes revolve in periods between 2 hrs. 8 m. and 3 hrs. 35 m., moving against the sun.

VITACEAE.—In this family and in the two following, namely, the Sapindaceae and Passifloraceae, the tendrils are modified flower- peduncles; and are therefore axial in their nature. In this respect they differ from all those previously described, with the exception, perhaps, of the Cucurbitaceae. The homological nature, however, of a tendril seems to make no difference in its action.

Vitis vinifera.—The tendril is thick and of great length; one from a vine growing out of doors and not vigorously, was 16 inches long. It consists of a peduncle (A), bearing two branches which diverge equally from it. One of the branches (B) has a scale at its base; it is always, as far as I have seen, longer than the other and often bifurcates. The branches when rubbed become curved, and subsequently straighten themselves. After a tendril has clasped any object with its extremity, it contracts spirally; but this does not occur (Palm, p. 56) when no object has been seized. The tendrils move spontaneously from side to side; and on a very hot day, one made two elliptical revolutions, at an average rate of 2 hrs. 15 m. During these movements a coloured line, painted along the convex surface, appeared after a time on one side, then on the concave side, then on the opposite side, and lastly again on the convex side. The two branches of the same tendril have independent movements. After a tendril has spontaneously revolved for a time, it bends from the light towards the dark: I do not state this on my own authority, but on that of Mohl and Dutrochet. Mohl (p. 77) says that in a vine planted against a wall the tendrils point towards it, and in a vineyard generally more or less to the north.

The young internodes revolve spontaneously; but the movement is unusually slight. A shoot faced a window, and I traced its course on the glass during two perfectly calm and hot days. On one of these days it described, in the course of ten hours, a spire, representing two and a half ellipses. I also placed a bell-glass over a young Muscat grape in the hot-house, and it made each day three or four very small oval revolutions; the shoot moving less than half an inch from side to side. Had it not made at least three revolutions whilst the sky was uniformly overcast, I should have attributed this slight degree of movement to the varying action of the light. The extremity of the stem is more or less bent downwards, but it never reverses its curvature, as so generally occurs with twining plants.

Various authors (Palm, p. 55; Mohl, p. 45; Lindley, &c.) believe that the tendrils of the vine are modified flower-peduncles. I here give a drawing (fig. 10) of the ordinary state of a young flower-stalk: it consists of the "common peduncle" (A); of the "flower-tendril" (B), which is represented as having caught a twig; and of the "sub- peduncle" (C) bearing the flower-buds. The whole moves spontaneously, like a true tendril, but in a less degree; the movement, however, is greater when the sub-peduncle (C) does not bear many flower-buds. The common peduncle (A) has not the power of clasping a support, nor has the corresponding part of a true tendril. The flower-tendril (B) is always longer than the sub-peduncle (C) and has a scale at its base; it sometimes bifurcates, and therefore corresponds in every detail with the longer scale-bearing branch (B, fig. 9) of the true tendril. It is, however, inclined backwards from the sub-peduncle (C), or stands at right angles with it, and is thus adapted to aid in carrying the future bunch of grapes. When rubbed, it curves and subsequently straightens itself; and it can, as is shown in the drawing, securely clasp a support. I have seen an object as soft as a young vine-leaf caught by one.

The lower and naked part of the sub-peduncle (C) is likewise slightly sensitive to a rub, and I have seen it bent round a stick and even partly round a leaf with which it had come into contact. That the sub-peduncle has the same nature as the corresponding branch of an ordinary tendril, is well shown when it bears only a few flowers; for in this case it becomes less branched, increases in length, and gains both in sensitiveness and in the power of spontaneous movement. I have twice seen sub-peduncles which bore from thirty to forty flower- buds, and which had become considerably elongated and were completely wound round sticks, exactly like true tendrils. The whole length of another sub-peduncle, bearing only eleven flower-buds, quickly became curved when slightly rubbed; but even this scanty number of flowers rendered the stalk less sensitive than the other branch, that is, the flower-tendril; for the latter after a lighter rub became curved more quickly and in a greater degree. I have seen a sub-peduncle thickly covered with flower-buds, with one of its higher lateral branchlets bearing from some cause only two buds; and this one branchlet had become much elongated and had spontaneously caught hold of an adjoining twig; in fact, it formed a little sub-tendril. The increasing length of the sub-peduncle (C) with the decreasing number of the flower-buds is a good instance of the law of compensation. In accordance with this same principle, the true tendril as a whole is always longer than the flower-stalk; for instance, on the same plant, the longest flower-stalk (measured from the base of the common peduncle to the tip of the flower-tendril) was 8.5 inches in length, whilst the longest tendril was nearly double this length, namely 16 inches.

The gradations from the ordinary state of a flower-stalk, as represented in the drawing (fig. 10), to that of a true tendril (fig. 9) are complete. We have seen that the sub-peduncle (C), whilst still bearing from thirty to forty flower-buds, sometimes becomes a little elongated and partially assumes all the characters of the corresponding branch of a true tendril. From this state we can trace every stage till we come to a full-sized perfect tendril, bearing on the branch which corresponds with the sub-peduncle one single flower- bud! Hence there can be no doubt that the tendril is a modified flower-peduncle.

Another kind of gradation well deserves notice. Flower-tendrils (B, fig. 10) sometimes produce a few flower-buds. For instance, on a vine growing against my house, there were thirteen and twenty-two flower-buds respectively on two flower-tendrils, which still retained their characteristic qualities of sensitiveness and spontaneous movement, but in a somewhat lessened degree. On vines in hothouses, so many flowers are occasionally produced on the flower-tendrils that a double bunch of grapes is the result; and this is technically called by gardeners a "cluster." In this state the whole bunch of flowers presents scarcely any resemblance to a tendril; and, judging from the facts already given, it would probably possess little power of clasping a support, or of spontaneous movement. Such flower- stalks closely resemble in structure those borne by Cissus. This genus, belonging to the same family of the Vitaceae, produces well- developed tendrils and ordinary bunches of flowers; but there are no gradations between the two states. If the genus Vitis had been unknown, the boldest believer in the modification of species would never have surmised that the same individual plant, at the same period of growth, would have yielded every possible gradation between ordinary flower-stalks for the support of the flowers and fruit, and tendrils used exclusively for climbing. But the vine clearly gives us such a case; and it seems to me as striking and curious an instance of transition as can well be conceived.

Cissus discolor.—The young shoots show no more movement than can be accounted for by daily variations in the action of the light. The tendrils, however, revolve with much regularity, following the sun; and, in the plants observed by me, swept circles of about 5 inches in diameter. Five circles were completed in the following times:- 4 hrs. 45 m., 4 hrs. 50 m., 4 hrs. 45 m., 4 hrs. 30 m., and 5 hrs. The same tendril continues to revolve during three or four days. The tendrils are from 3.5 to 5 inches in length. They are formed of a long foot-stalk, bearing two short branches, which in old plants again bifurcate. The two branches are not of quite equal length; and as with the vine, the longer one has a scale at its base. The tendril stands vertically upwards; the extremity of the shoot being bent abruptly downwards, and this position is probably of service to the plant by allowing the tendril to revolve freely and vertically.

Both branches of the tendril, whilst young, are highly sensitive. A touch with a pencil, so gentle as only just to move a tendril borne at the end of a long flexible shoot, sufficed to cause it to become perceptibly curved in four or five minutes. It became straight again in rather above one hour. A loop of soft thread weighing one-seventh of a grain (9.25 mg.) was thrice tried, and each time caused the tendril to become curved in 30 or 40 m. Half this weight produced no effect. The long foot-stalk is much less sensitive, for a slight rubbing produced no effect, although prolonged contact with a stick caused it to bend. The two branches are sensitive on all sides, so that they converge if touched on their inner sides, and diverge if touched on their outer sides. If a branch be touched at the same time with equal force on opposite sides, both sides are equally stimulated and there is no movement. Before examining this plant, I had observed only tendrils which are sensitive on one side alone, and these when lightly pressed between the finger and thumb become curved; but on thus pinching many times the tendrils of the Cissus no curvature ensued, and I falsely inferred at first that they were not at all sensitive.

Cissus antarcticus.—The tendrils on a young plant were thick and straight, with the tips a little curved. When their concave surfaces were rubbed, and it was necessary to do this with some force, they very slowly became curved, and subsequently straight again. They are therefore much less sensitive than those of the last species; but they made two revolutions, following the sun, rather more rapidly, viz., in 3 hrs. 30 m. and 4 hrs. The internodes do not revolve.

Ampelopsis hederacea (Virginian Creeper).—The internodes apparently do not move more than can be accounted for by the varying action of the light. The tendrils are from 4 to 5 inches in length, with the main stem sending off several lateral branches, which have their tips curved, as may be seen in the upper figure (fig. 11). They exhibit no true spontaneous revolving movement, but turn, as was long ago observed by Andrew Knight, {31} from the light to the dark. I have seen several tendrils move in less than 24 hours, through an angle of 180 degrees to the dark side of a case in which a plant was placed, but the movement is sometimes much slower. The several lateral branches often move independently of one another, and sometimes irregularly, without any apparent cause. These tendrils are less sensitive to a touch than any others observed by me. By gentle but repeated rubbing with a twig, the lateral branches, but not the main stem, became in the course of three or four hours slightly curved; but they seemed to have hardly any power of again straightening themselves. The tendrils of a plant which had crawled over a large box-tree clasped several of the branches; but I have repeatedly seen that they will withdraw themselves after seizing a stick. When they meet with a flat surface of wood or a wall (and this is evidently what they are adapted for), they turn all their branches towards it, and, spreading them widely apart, bring their hooked tips laterally into contact with it. In effecting this, the several branches, after touching the surface, often rise up, place themselves in a new position, and again come down into contact with it.

In the course of about two days after a tendril has arranged its branches so as to press on any surface, the curved tips swell, become bright red, and form on their under-sides the well-known little discs or cushions with which they adhere firmly. In one case the tips were slightly swollen in 38 hrs. after coming into contact with a brick; in another case they were considerably swollen in 48 hrs., and in an additional 24 hrs. were firmly attached to a smooth board; and lastly, the tips of a younger tendril not only swelled but became attached to a stuccoed wall in 42 hrs. These adhesive discs resemble, except in colour and in being larger, those of Bignonia capreolata. When they were developed in contact with a ball of tow, the fibres were separately enveloped, but not in so effective a manner as by B. capreolata. Discs are never developed, as far as I have seen, without the stimulus of at least temporary contact with some object. {32} They are generally first formed on one side of the curved tip, the whole of which often becomes so much changed in appearance, that a line of the original green tissue can be traced only along the concave surface. When, however, a tendril has clasped a cylindrical stick, an irregular rim or disc is sometimes formed along the inner surface at some little distance from the curved tip; this was also observed (p. 71) by Mohl. The discs consist of enlarged cells, with smooth projecting hemispherical surfaces, coloured red; they are at first gorged with fluid (see section given by Mohl, p. 70), but ultimately become woody.

As the discs soon adhere firmly to such smooth surfaces as planed or painted wood, or to the polished leaf of the ivy, this alone renders it probable that some cement is secreted, as has been asserted to be the case (quoted by Mohl, p. 71) by Malpighi. I removed a number of discs formed during the previous year from a stuccoed wall, and left them during many hours, in warm water, diluted acetic acid and alcohol; but the attached grains of silex were not loosened. Immersion in sulphuric ether for 24 hrs. loosened them much, but warmed essential oils (I tried oil of thyme and peppermint) completely released every particle of stone in the course of a few hours. This seems to prove that some resinous cement is secreted. The quantity, however, must be small; for when a plant ascended a thinly whitewashed wall, the discs adhered firmly to the whitewash; but as the cement never penetrated the thin layer, they were easily withdrawn, together with little scales of the whitewash. It must not be supposed that the attachment is effected exclusively by the cement; for the cellular outgrowth completely envelopes every minute and irregular projection, and insinuates itself into every crevice.

A tendril which has not become attached to any body, does not contract spirally; and in course of a week or two shrinks into the finest thread, withers and drops off. An attached tendril, on the other hand, contracts spirally, and thus becomes highly elastic, so that when the main foot-stalk is pulled the strain is distributed equally between all the attached discs. For a few days after the attachment of the discs, the tendril remains weak and brittle, but it rapidly increases in thickness and acquires great strength. During the following winter it ceases to live, but adheres firmly in a dead state both to its own stem and to the surface of attachment. In the accompanying diagram (fig. 11.) we see the difference between a tendril (B) some weeks after its attachment to a wall, with one (A) from the same plant fully grown but unattached. That the change in the nature of the tissues, as well as the spiral contraction, are consequent on the formation of the discs, is well shown by any lateral branches which have not become attached; for these in a week or two wither and drop off, in the same manner as does the whole tendril if unattached. The gain in strength and durability in a tendril after its attachment is something wonderful. There are tendrils now adhering to my house which are still strong, and have been exposed to the weather in a dead state for fourteen or fifteen years. One single lateral branchlet of a tendril, estimated to be at least ten years old, was still elastic and supported a weight of exactly two pounds. The whole tendril had five disc-bearing branches of equal thickness and apparently of equal strength; so that after having been exposed during ten years to the weather, it would probably have resisted a strain of ten pounds!

SAPINDACEAE.—Cardiospermum halicacabum.—In this family, as in the last, the tendrils are modified flower-peduncles. In the present plant the two lateral branches of the main flower-peduncle have been converted into a pair of tendrils, corresponding with the single "flower-tendril" of the common vine. The main peduncle is thin, stiff, and from 3 to 4.5 inches in length. Near the summit, above two little bracts, it divides into three branches. The middle one divides and re-divides, and bears the flowers; ultimately it grows half as long again as the two other modified branches. These latter are the tendrils; they are at first thicker and longer than the middle branch, but never become more than an inch in length. They taper to a point and are flattened, with the lower clasping surface destitute of hairs. At first they project straight up; but soon diverging, spontaneously curl downwards so as to become symmetrically and elegantly hooked, as represented in the diagram. They are now, whilst the flower-buds are still small, ready for action.

The two or three upper internodes, whilst young, steadily revolve; those on one plant made two circles, against the course of the sun, in 3 hrs. 12 m.; in a second plant the same course was followed, and the two circles were completed in 3 hrs. 41 m.; in a third plant, the internodes followed the sun and made two circles in 3 hrs. 47 m. The average rate of these six revolutions was 1 hr. 46 m. The stem shows no tendency to twine spirally round a support; but the allied tendril-bearing genus Paullinia is said (Mohl, p. 4) to be a twiner. The flower-peduncles, which stand up above the end of the shoot, are carried round and round by the revolving movement of the internodes; and when the stem is securely tied, the long and thin flower- peduncles themselves are seen to be in continued and sometimes rapid movement from side to side. They sweep a wide space, but only occasionally revolve in a regular elliptical course. By the combined movements of the internodes and peduncles, one of the two short hooked tendrils, sooner or later, catches hold of some twig or branch, and then it curls round and securely grasps it. These tendrils are, however, but slightly sensitive; for by rubbing their under surface only a slight movement is slowly produced. I hooked a tendril on to a twig; and in 1 hr. 45 m. it was curved considerably inwards; in 2 hrs. 30 m. it formed a ring; and in from 5 to 6 hours from being first hooked, it closely grasped the stick. A second tendril acted at nearly the same rate; but I observed one that took 24 hours before it curled twice round a thin twig. Tendrils which have caught nothing, spontaneously curl up to a close helix after the interval of several days. Those which have curled round some object, soon become a little thicker and tougher. The long and thin main peduncle, though spontaneously moving, is not sensitive and never clasps a support. Nor does it ever contract spirally, {33} although a contraction of this kind apparently would have been of service to the plant in climbing. Nevertheless it climbs pretty well without this aid. The seed-capsules though light, are of enormous size (hence its English name of balloon-vine), and as two or three are carried on the same peduncle, the tendrils rising close to them may be of service in preventing their being dashed to pieces by the wind. In the hothouse the tendrils served simply for climbing.

The position of the tendrils alone suffices to show their homological nature. In two instances one of two tendrils produced a flower at its tip; this, however, did not prevent its acting properly and curling round a twig. In a third case both lateral branches which ought to have been modified into tendrils, produced flowers like the central branch, and had quite lost their tendril-structure.

I have seen, but was not enabled carefully to observe, only one other climbing Sapindaceous plant, namely, Paullinia. It was not in flower, yet bore long forked tendrils. So that, Paullinia, with respect to its tendrils, appears to bear the same relation to Cardiospermum that Cissus does to Vitis.

PASSIFLORACEAE.—After reading the discussion and facts given by Mohl (p. 47) on the nature of the tendrils in this family, no one can doubt that they are modified flower-peduncles. The tendrils and the flower-peduncles rise close side by side; and my son, William E. Darwin, made sketches for me of their earliest state of development in the hybrid P. floribunda. The two organs appear at first as a single papilla which gradually divides; so that the tendril appears to be a modified branch of the flower-peduncle. My son found one very young tendril surmounted by traces of floral organs, exactly like those on the summit of the true flower-peduncle at the same early age.

Passiflora gracilis.—This well-named, elegant, annual species differs from the other members of the group observed by me, in the young internodes having the power of revolving. It exceeds all the other climbing plants which I have examined, in the rapidity of its movements, and all tendril-bearers in the sensitiveness of the tendrils. The internode which carries the upper active tendril and which likewise carries one or two younger immature internodes, made three revolutions, following the sun, at an average rate of 1 hr. 4 m.; it then made, the day becoming very hot, three other revolutions at an average rate of between 57 and 58 m.; so that the average of all six revolutions was 1 hr. 1 m. The apex of the tendril describes elongated ellipses, sometimes narrow and sometimes broad, with their longer axes inclined in slightly different directions. The plant can ascend a thin upright stick by the aid of its tendrils; but the stem is too stiff for it to twine spirally round it, even when not interfered with by the tendrils, these having been successively pinched off at an early age.

When the stem is secured, the tendrils are seen to revolve in nearly the same manner and at the same rate as the internodes. {34} The tendrils are very thin, delicate, and straight, with the exception of the tips, which are a little curved; they are from 7 to 9 inches in length. A half-grown tendril is not sensitive; but when nearly full- grown they are extremely sensitive. A single delicate touch on the concave surface of the tip soon caused one to curve; and in 2 minutes it formed an open helix. A loop of soft thread weighing one thirty- second of a grain (2.02 mg.) placed most gently on the tip, thrice caused distinct curvature. A bent bit of thin platina wire weighing only fiftieth of a grain (1.23 mg.) twice produced the same effect; but this latter weight, when left suspended, did not suffice to cause a permanent curvature. These trials were made under a bell-glass, so that the loops of thread and wire were not agitated by the wind. The movement after a touch is very rapid: I took hold of the lower part of several tendrils, and then touched their concave tips with a thin twig and watched them carefully through a lens; the tips evidently began to bend after the following intervals—31, 25, 32, 31, 28, 39, 31, and 30 seconds; so that the movement was generally perceptible in half a minute after a touch; but on one occasion it was distinctly visible in 25 seconds. One of the tendrils which thus became bent in 31 seconds, had been touched two hours previously and had coiled into a helix; so that in this interval it had straightened itself and had perfectly recovered its irritability.

To ascertain how often the same tendril would become curved when touched, I kept a plant in my study, which from being cooler than the hot-house was not very favourable for the experiment. The extremity was gently rubbed four or five times with a thin stick, and this was done as often as it was observed to have become nearly straight again after having been in action; and in the course of 54 hrs. it answered to the stimulus 21 times, becoming each time hooked or spiral. On the last occasion, however, the movement was very slight, and soon afterwards permanent spiral contraction commenced. No trials were made during the night, so that the tendril would perhaps have answered a greater number of times to the stimulus; though, on the other hand, from having no rest it might have become exhausted from so many quickly repeated efforts.

I repeated the experiment made on the Echinocystis, and placed several plants of this Passiflora so close together, that their tendrils were repeatedly dragged over each other; but no curvature ensued. I likewise repeatedly flirted small drops of water from a brush on many tendrils, and syringed others so violently that the whole tendril was dashed about, but they never became curved. The impact from the drops of water was felt far more distinctly on my hand than that from the loops of thread (weighing one thirty-second of a grain) when allowed to fall on it from a height, and these loops, which caused the tendrils to become curved, had been placed most gently on them. Hence it is clear, that the tendrils either have become habituated to the touch of other tendrils and drops of rain, or that they were from the first rendered sensitive only to prolonged though excessively slight pressure of solid objects, with the exclusion of that from other tendrils. To show the difference in the kind of sensitiveness in different plants and likewise to show the force of the syringe used, I may add that the lightest jet from it instantly caused the leaves of a Mimosa to close; whereas the loop of thread weighing one thirty-second of a grain, when rolled into a ball and placed gently on the glands at the bases of the leaflets of the Mimosa, caused no action.

Passiflora punctata.—The internodes do not move, but the tendrils revolve regularly. A half-grown and very sensitive tendril made three revolutions, opposed to the course of the sun, in 3 hrs. 5 m., 2 hrs. 40 m. and 2 hrs. 50 m.; perhaps it might have travelled more quickly when nearly full-grown. A plant was placed in front of a window, and, as with twining stems, the light accelerated the movement of the tendril in one direction and retarded it in the other; the semicircle towards the light being performed in one instance in 15 m. less time and in a second instance in 20 m. less time than that required by the semicircle towards the dark end of the room. Considering the extreme tenuity of these tendrils, the action of the light on them is remarkable. The tendrils are long, and, as just stated, very thin, with the tip slightly curved or hooked. The concave side is extremely sensitive to a touch—even a single touch causing it to curl inwards; it subsequently straightened itself, and was again ready to act. A loop of soft thread weighing one fourteenth of a grain (4.625 mg.) caused the extreme tip to bend; another time I tried to hang the same little loop on an inclined tendril, but three times it slid off; yet this extraordinarily slight degree of friction sufficed to make the tip curl. The tendril, though so sensitive, does not move very quickly after a touch, no conspicuous movement being observable until 5 or 10 m. had elapsed. The convex side of the tip is not sensitive to a touch or to a suspended loop of thread. On one occasion I observed a tendril revolving with the convex side of the tip forwards, and in consequence it was not able to clasp a stick, against which it scraped; whereas tendrils revolving with the concave side forward, promptly seize any object in their path.

Passiflora quadrangularis.—This is a very distinct species. The tendrils are thick, long, and stiff; they are sensitive to a touch only on the concave surface towards the extremity. When a stick was placed so that the middle of the tendril came into contact with it, no curvature ensued. In the hothouse a tendril made two revolutions, each in 2 hrs. 22 m.; in a cool room one was completed in 3 hrs., and a second in 4 hrs. The internodes do not revolve; nor do those of the hybrid P. floribunda.

Tacsonia manicata.—Here again the internodes do not revolve. The tendrils are moderately thin and long; one made a narrow ellipse in 5 hrs. 20 m., and the next day a broad ellipse in 5 hrs. 7 m. The extremity being lightly rubbed on the concave surface, became just perceptibly curved in 7 m., distinctly in 10 m., and hooked in 20 m.

We have seen that the tendrils in the last three families, namely, the Vitaceae, Sapindaceae and Passifloraceae, are modified flower- peduncles. This is likewise the case, according to De Candolle (as quoted by Mohl), with the tendrils of Brunnichia, one of the Polygonaceae. In two or three species of Modecca, one of the Papayaceae, the tendrils, as I hear from Prof. Oliver, occasionally bear flowers and fruit; so that they are axial in their nature.

The Spiral Contraction of Tendrils.

This movement, which shortens the tendrils and renders them elastic, commences in half a day, or in a day or two after their extremities have caught some object. There is no such movement in any leaf- climber, with the exception of an occasional trace of it in the petioles of Tropaeolum tricolorum. On the other hand, the tendrils of all tendril-bearing plants, contract spirally after they have caught an object with the following exceptions. Firstly, Corydalis claviculata, but then this plant might be called a leaf-climber. Secondly and thirdly, Bignonia unguis with its close allies, and Cardiospermum; but their tendrils are so short that their contraction could hardly occur, and would be quite superfluous. Fourthly, Smilax aspera offers a more marked exception, as its tendrils are moderately long. The tendrils of Dicentra, whilst the plant is young, are short and after attachment only become slightly flexuous; in older plants they are longer and then they contract spirally. I have seen no other exceptions to the rule that tendrils, after clasping with their extremities a support, undergo spiral contraction. When, however, the tendril of a plant of which the stem is immovably fixed, catches some fixed object, it does not contract, simply because it cannot; this, however, rarely occurs. In the common Pea the lateral branches alone contract, and not the central stem; and with most plants, such as the Vine, Passiflora, Bryony, the basal portion never forms a spire.

I have said that in Corydalis claviculata the end of the leaf or tendril (for this part may be indifferently so called) does not contract into a spire. The branchlets, however, after they have wound round thin twigs, become deeply sinuous or zigzag. Moreover the whole end of the petiole or tendril, if it seizes nothing, bends after a time abruptly downwards and inwards, showing that its outer surface has gone on growing after the inner surface has ceased to grow. That growth is the chief cause of the spiral contraction of tendrils may be safely admitted, as shown by the recent researches of H. de Vries. I will, however, add one little fact in support of this conclusion.

If the short, nearly straight portion of an attached tendril of Passiflora gracilis, (and, as I believe, of other tendrils,) between the opposed spires, be examined, it will be found to be transversely wrinkled in a conspicuous manner on the outside; and this would naturally follow if the outer side had grown more than the inner side, this part being at the same time forcibly prevented from becoming curved. So again the whole outer surface of a spirally wound tendril becomes wrinkled if it be pulled straight. Nevertheless, as the contraction travels from the extremity of a tendril, after it has been stimulated by contact with a support, down to the base, I cannot avoid doubting, from reasons presently to be given, whether the whole effect ought to be attributed to growth. An unattached tendril rolls itself up into a flat helix, as in the case of Cardiospermum, if the contraction commences at the extremity and is quite regular; but if the continued growth of the outer surface is a little lateral, or if the process begins near the base, the terminal portion cannot be rolled up within the basal portion, and the tendril then forms a more or less open spire. A similar result follows if the extremity has caught some object, and is thus held fast.

The tendrils of many kinds of plants, if they catch nothing, contract after an interval of several days or weeks into a spire; but in these cases the movement takes place after the tendril has lost its revolving power and hangs down; it has also then partly or wholly lost its sensibility; so that this movement can be of no use. The spiral contraction of unattached tendrils is a much slower process than that of attached ones. Young tendrils which have caught a support and are spirally contracted, may constantly be seen on the same stem with the much older unattached and uncontracted tendrils. In the Echinocystis I have seen a tendril with the two lateral branches encircling twigs and contracted into beautiful spires, whilst the main branch which had caught nothing remained for many days straight. In this plant I once observed a main branch after it had caught a stick become spirally flexuous in 7 hrs., and spirally contracted in 18 hrs. Generally the tendrils of the Echinocystis begin to contract in from 12 hrs. to 24 hrs. after catching some object; whilst unattached tendrils do not begin to contract until two or three or even more days after all revolving movement has ceased. A full-grown tendril of Passiflora quadrangularis which had caught a stick began in 8 hrs. to contract, and in 24 hrs. formed several spires; a younger tendril, only two-thirds grown, showed the first trace of contraction in two days after clasping a stick, and in two more days formed several spires. It appears, therefore, that the contraction does not begin until the tendril is grown to nearly its full length. Another young tendril of about the same age and length as the last did not catch any object; it acquired its full length in four days; in six additional days it first became flexuous, and in two more days formed one complete spire. This first spire was formed towards the basal end, and the contraction steadily but slowly progressed towards the apex; but the whole was not closely wound up into a spire until 21 days had elapsed from the first observation, that is, until 17 days after the tendril had grown to its full length.

The spiral contraction of tendrils is quite independent of their power of spontaneously revolving, for it occurs in tendrils, such as those of Lathyrus grandiflorus and Ampelopsis hederacea, which do not revolve. It is not necessarily related to the curling of the tips round a support, as we see with the Ampelopsis and Bignonia capreolata, in which the development of adherent discs suffices to cause spiral contraction. Yet in some cases this contraction seems connected with the curling or clasping movement, due to contact with a support; for not only does it soon follow this act, but the contraction generally begins close to the curled extremity, and travels downwards to the base. If, however, a tendril be very slack, the whole length almost simultaneously becomes at first flexuous and then spiral. Again, the tendrils of some few plants never contract spirally unless they have first seized hold of some object; if they catch nothing they hang down, remaining straight, until they wither and drop off: this is the case with the tendrils of Bignonia, which consist of modified leaves, and with those of three genera of the Vitaceae, which are modified flower-peduncles. But in the great majority of cases, tendrils which have never come in contact with any object, after a time contract spirally. All these facts taken together, show that the act of clasping a support and the spiral contraction of the whole length of the tendril, are phenomena not necessarily connected.

The spiral contraction which ensues after a tendril has caught a support is of high service to the plant; hence its almost universal occurrence with species belonging to widely different orders. When a shoot is inclined and its tendril has caught an object above, the spiral contraction drags up the shoot. When the shoot is upright, the growth of the stem, after the tendrils have seized some object above, would leave it slack, were it not for the spiral contraction which draws up the stem as it increases in length. Thus there is no waste of growth, and the stretched stem ascends by the shortest course. When a terminal branchlet of the tendril of Cobaea catches a stick, we have seen how well the spiral contraction successively brings the other branchlets, one after the other, into contact with the stick, until the whole tendril grasps it in an inextricable knot. When a tendril has caught a yielding object, this is sometimes enveloped and still further secured by the spiral folds, as I have seen with Passiflora quadrangularis; but this action is of little importance.

A far more important service rendered by the spiral contraction of the tendrils is that they are thus made highly elastic. As before remarked under Ampelopsis, the strain is thus distributed equally between the several attached branches; and this renders the whole far stronger than it otherwise would be, as the branches cannot break separately. It is this elasticity which protects both branched and simple tendrils from being torn away from their supports during stormy weather. I have more than once gone on purpose during a gale to watch a Bryony growing in an exposed hedge, with its tendrils attached to the surrounding bushes; and as the thick and thin branches were tossed to and fro by the wind, the tendrils, had they not been excessively elastic, would instantly have been torn off and the plant thrown prostrate. But as it was, the Bryony safely rode out the gale, like a ship with two anchors down, and with a long range of cable ahead to serve as a spring as she surges to the storm.

When an unattached tendril contracts spirally, the spire always runs in the same direction from tip to base. A tendril, on the other hand, which has caught a support by its extremity, although the same side is concave from end to end, invariably becomes twisted in one part in one direction, and in another part in the opposite direction; the oppositely turned spires being separated by a short straight portion. This curious and symmetrical structure has been noticed by several botanists, but has not been sufficiently explained. {35} It occurs without exception with all tendrils which after catching an object contract spirally, but is of course most conspicuous in the longer tendrils. It never occurs with uncaught tendrils; and when this appears to have occurred, it will be found that the tendril had originally seized some object and had afterwards been torn free. Commonly, all the spires at one end of an attached tendril run in one direction, and all those at the other end in the opposite direction, with a single short straight portion in the middle; but I have seen a tendril with the spires alternately turning five times in opposite directions, with straight pieces between them; and M. Leon has seen seven or eight such alternations. Whether the spires turn once or more than once in opposite directions, there are as many turns in the one direction as in the other. For instance, I gathered ten attached tendrils of the Bryony, the longest with 33, and the shortest with only 8 spiral turns; and the number of turns in the one direction was in every case the same (within one) as in the opposite direction.

The explanation of this curious little fact is not difficult. I will not attempt any geometrical reasoning, but will give only a practical illustration. In doing this, I shall first have to allude to a point which was almost passed over when treating of Twining-plants. If we hold in our left hand a bundle of parallel strings, we can with our right hand turn these round and round, thus imitating the revolving movement of a twining plant, and the strings do not become twisted. But if we hold at the same time a stick in our left hand, in such a position that the strings become spirally turned round it, they will inevitably become twisted. Hence a straight coloured line, painted along the internodes of a twining plant before it has wound round a support, becomes twisted or spiral after it has wound round. I painted a red line on the straight internodes of a Humulus, Mikania, Ceropegia, Convolvulus, and Phaseolus, and saw it become twisted as the plant wound round a stick. It is possible that the stems of some plants by spontaneously turning on their own axes, at the proper rate and in the proper direction, might avoid becoming twisted; but I have seen no such case.

In the above illustration, the parallel strings were wound round a stick; but this is by no means necessary, for if wound into a hollow coil (as can be done with a narrow slip of elastic paper) there is the same inevitable twisting of the axis. When, therefore, a free tendril coils itself into a spire, it must either become twisted along its whole length (and this never occurs), or the free extremity must turn round as many times as there are spires formed. It was hardly necessary to observe this fact; but I did so by affixing little paper vanes to the extreme points of the tendrils of Echinocystis and Passiflora quadrangularis; and as the tendril contracted itself into successive spires, the vane slowly revolved.

We can now understand the meaning of the spires being invariably turned in opposite directions, in tendrils which from having caught some object are fixed at both ends. Let us suppose a caught tendril to make thirty spiral turns all in the same direction; the inevitable result would be that it would become twisted thirty times on its own axis. This twisting would not only require considerable force, but, as I know by trial, would burst the tendril before the thirty turns were completed. Such cases never really occur; for, as already stated, when a tendril has caught a support and is spirally contracted, there are always as many turns in one direction as in the other; so that the twisting of the axis in the one direction is exactly compensated by the twisting in the opposite direction. We can further see how the tendency is given to make the later formed coils opposite to those, whether turned to the right or to the left, which are first made. Take a piece of string, and let it hang down with the lower end fixed to the floor; then wind the upper end (holding the string quite loosely) spirally round a perpendicular pencil, and this will twist the lower part of the string; and after it has been sufficiently twisted, it will be seen to curve itself into an open spire, with the curves running in an opposite direction to those round the pencil, and consequently with a straight piece of string between the opposed spires. In short, we have given to the string the regular spiral arrangement of a tendril caught at both ends. The spiral contraction generally begins at the extremity which has clasped a support; and these first-formed spires give a twist to the axis of the tendril, which necessarily inclines the basal part into an opposite spiral curvature. I cannot resist giving one other illustration, though superfluous: when a haberdasher winds up ribbon for a customer, he does not wind it into a single coil; for, if he did, the ribbon would twist itself as many times as there were coils; but he winds it into a figure of eight on his thumb and little finger, so that he alternately takes turns in opposite directions, and thus the ribbon is not twisted. So it is with tendrils, with this sole difference, that they take several consecutive turns in one direction and then the same number in an opposite direction; but in both cases the self-twisting is avoided.

Summary on the Nature and Action of Tendrils.

With the majority of tendril-bearing plants the young internodes revolve in more or less broad ellipses, like those made by twining plants; but the figures described, when carefully traced, generally form irregular ellipsoidal spires. The rate of revolution varies from one to five hours in different species, and consequently is in some cases more rapid than with any twining plant, and is never so slow as with those many twiners which take more than five hours for each revolution. The direction is variable even in the same individual plant. In Passiflora, the internodes of only one species have the power of revolving. The Vine is the weakest revolver observed by me, apparently exhibiting only a trace of a former power. In the Eccremocarpus the movement is interrupted by many long pauses. Very few tendril-bearing plants can spirally twine up an upright stick. Although the power of twining has generally been lost, either from the stiffness or shortness of the internodes, from the size of the leaves, or from some other unknown cause, the revolving movement of the stem serves to bring the tendrils into contact with surrounding objects.

The tendrils themselves also spontaneously revolve. The movement begins whilst the tendril is young, and is at first slow. The mature tendrils of Bignonia littoralis move much slower than the internodes. Generally, the internodes and tendrils revolve together at the same rate; in Cissus, Cobaea, and most Passiflorae, the tendrils alone revolve; in other cases, as with Lathyrus aphaca, only the internodes move, carrying with them the motionless tendrils; and, lastly (and this is the fourth possible case), neither internodes nor tendrils spontaneously revolve, as with Lathyrus grandiflorus and Ampelopsis. In most Bignonias, Eccremocarpus Mutisia, and the Fumariaceae, the internodes, petioles and tendrils all move harmoniously together. In every case the conditions of life must be favourable in order that the different parts should act in a perfect manner.

Tendrils revolve by the curvature of their whole length, excepting the sensitive extremity and the base, which parts do not move, or move but little. The movement is of the same nature as that of the revolving internodes, and, from the observations of Sachs and H. de Vries, no doubt is due to the same cause, namely, the rapid growth of a longitudinal band, which travels round the tendril and successively bows each part to the opposite side. Hence, if a line be painted along that surface which happens at the time to be convex, the line becomes first lateral, then concave, then lateral, and ultimately again convex. This experiment can be tried only on the thicker tendrils, which are not affected by a thin crust of dried paint. The extremities are often slightly curved or hooked, and the curvature of this part is never reversed; in this respect they differ from the extremities of twining shoots, which not only reverse their curvature, or at least become periodically straight, but curve themselves in a greater degree than the lower part. In most other respects a tendril acts as if it were one of several revolving internodes, which all move together by successively bending to each point of the compass. There is, however, in many cases this unimportant difference, that the curving tendril is separated from the curving internode by a rigid petiole. With most tendril-bearers the summit of the stem or shoot projects above the point from which the tendril arises; and it is generally bent to one side, so as to be out of the way of the revolutions swept by the tendril. In those plants in which the terminal shoot is not sufficiently out of the way, as we have seen with the Echinocystis, as soon as the tendril comes in its revolving course to this point, it stiffens and straightens itself, and thus rising vertically up passes over the obstacle in an admirable manner.

All tendrils are sensitive, but in various degrees, to contact with an object, and curve towards the touched side. With several plants a single touch, so slight as only just to move the highly flexible tendril, is enough to induce curvature. Passiflora gracilis possesses the most sensitive tendrils which I have observed: a bit of platina wire 0.02 of a grain (1.23 mg.) in weight, gently placed on the concave point, caused a tendril to become hooked, as did a loop of soft, thin cotton thread weighing one thirty-second of a grain (2.02 mg.) With the tendrils of several other plants, loops weighing one sixteenth of a grain (4.05 mg.) sufficed. The point of a tendril of Passiflora gracilis began to move distinctly in 25 seconds after a touch, and in many cases after 30 seconds. Asa Gray also saw movement in the tendrils of the Cucurbitaceous genus, Sicyos, in 30 seconds. The tendrils of some other plants, when lightly rubbed, moved in a few minutes; with Dicentra in half-an- hour; with Smilax in an hour and a quarter or half; and with Ampelopsis still more slowly. The curling movement consequent on a single touch continues to increase for a considerable time, then ceases; after a few hours the tendril uncurls itself, and is again ready to act. When the tendrils of several kinds of plants were caused to bend by extremely light weights suspended on them, they seemed to grow accustomed to so slight a stimulus, and straightened themselves, as if the loops had been removed. It makes no difference what sort of object a tendril touches, with the remarkable exception of other tendrils and drops of water, as was observed with the extremely sensitive-tendrils of Passiflora gracilis and of the Echinocystis. I have, however, seen tendrils of the Bryony which had temporarily caught other tendrils, and often in the case of the vine.

Tendrils of which the extremities are permanently and slightly curved, are sensitive only on the concave surface; other tendrils, such as those of the Cobaea (though furnished with horny hooks directed to one side) and those of Cissus discolor, are sensitive on all sides. Hence the tendrils of this latter plant, when stimulated by a touch of equal force on opposite sides, did not bend. The inferior and lateral surfaces of the tendrils of Mutisia are sensitive, but not the upper surface. With branched tendrils, the several branches act alike; but in the Hanburya the lateral spur-like branch does not acquire (for excellent reasons which have been explained) its sensitiveness nearly so soon as the main branch. With most tendrils the lower or basal part is either not at all sensitive, or sensitive only to prolonged contact. We thus see that the sensitiveness of tendrils is a special and localized capacity. It is quite independent of the power of spontaneously revolving; for the curling of the terminal portion from touch does not in the least interrupt the former movement. In Bignonia unguis and its close allies, the petioles of the leaves, as well as the tendrils, are sensitive to a touch.

Twining plants when they come into contact with a stick, curl round it invariably in the direction of their revolving movement; but tendrils curl indifferently to either side, in accordance with the position of the stick and the side which is first touched. The clasping movement of the extremity is apparently not steady, but undulatory or vermicular in its nature, as may be inferred from the curious manner in which the tendrils of the Echinocystis slowly crawled round a smooth stick.

As with a few exceptions tendrils spontaneously revolve, it may be asked,—why have they been endowed with sensitiveness?—why, when they come into contact with a stick, do they not, like twining plants, spirally wind round it? One reason may be that they are in most cases so flexible and thin, that when brought into contact with any object, they would almost certainly yield and be dragged onwards by the revolving movement. Moreover, the sensitive extremities have no revolving power as far as I have observed, and could not by this means curl round a support. With twining plants, on the other hand, the extremity spontaneously bends more than any other part; and this is of high importance for the ascent of the plant, as may be seen on a windy day. It is, however, possible that the slow movement of the basal and stiffer parts of certain tendrils, which wind round sticks placed in their path, may be analogous to that of twining plants. But I hardly attended sufficiently to this point, and it would have been difficult to distinguish between a movement due to extremely dull irritability, from the arrestment of the lower part, whilst the upper part continued to move onwards.

Tendrils which are only three-fourths grown, and perhaps even at an earlier age, but not whilst extremely young, have the power of revolving and of grasping any object which they touch. These two capacities are generally acquired at about the same period, and both fail when the tendril is full grown. But in Cobaea and Passiflora punctata the tendrils begin to revolve in a useless manner, before they have become sensitive. In the Echinocystis they retain their sensitiveness for some time after they have ceased to revolve and after they have sunk downwards; in this position, even if they were able to seize an object, such power would be of no service in supporting the stem. It is a rare circumstance thus to detect any superfluity or imperfection in the action of tendrils—organs which are so excellently adapted for the functions which they have to perform; but we see that they are not always perfect, and it would be rash to assume that any existing tendril has reached the utmost limit of perfection.

Some tendrils have their revolving motion accelerated or retarded, in moving to or from the light; others, as with the Pea, seem indifferent to its action; others move steadily from the light to the dark, and this aids them in an important manner in finding a support. For instance, the tendrils of Bignonia capreolata bend from the light to the dark as truly as a wind-vane from the wind. In the Eccremocarpus the extremities alone twist and turn about so as to bring their finer branches and hooks into close contact with any dark surface, or into crevices and holes.

A short time after a tendril has caught a support, it contracts with some rare exceptions into a spire; but the manner of contraction and the several important advantages thus gained have been discussed so lately, that nothing need here be repeated on the subject. Tendrils soon after catching a support grow much stronger and thicker, and sometimes more durable to a wonderful degree; and this shows how much their internal tissues must be changed. Occasionally it is the part which is wound round a support which chiefly becomes thicker and stronger; I have seen, for instance, this part of a tendril of Bignonia aequinoctialis twice as thick and rigid as the free basal part. Tendrils which have caught nothing soon shrink and wither; but in some species of Bignonia they disarticulate and fall off like leaves in autumn.

Any one who had not closely observed tendrils of many kinds would probably infer that their action was uniform. This is the case with the simpler kinds, which simply curl round an object of moderate thickness, whatever its nature may be. {36} But the genus Bignonia shows us what diversity of action there may be between the tendrils of closely allied species. In all the nine species observed by me, the young internodes revolve vigorously; the tendrils also revolve, but in some of the species in a very feeble manner; and lastly the petioles of nearly all revolve, though with unequal power. The petioles of three of the species, and the tendrils of all are sensitive to contact. In the first-described species, the tendrils resemble in shape a bird's foot, and they are of no service to the stem in spirally ascending a thin upright stick, but they can seize firm hold of a twig or branch. When the stem twines round a somewhat thick stick, a slight degree of sensitiveness possessed by the petioles is brought into play, and the whole leaf together with the tendril winds round it. In B. unguis the petioles are more sensitive, and have greater power of movement than those of the last species; they are able, together with the tendrils, to wind inextricably round a thin upright stick; but the stem does not twine so well. B. Tweedyana has similar powers, but in addition, emits aerial roots which adhere to the wood. In B. venusta the tendrils are converted into elongated three-pronged grapnels, which move spontaneously in a conspicuous manner; the petioles, however, have lost their sensitiveness. The stem of this species can twine round an upright stick, and is aided in its ascent by the tendrils seizing the stick alternately some way above and then contracting spirally. In B. littoralis the tendrils, petioles, and internodes, all revolve spontaneously. The stem, however, cannot twine, but ascends an upright stick by seizing it above with both tendrils together, which then contract into a spire. The tips of these tendrils become developed into adhesive discs. B. speciosa possesses similar powers of movement as the last species, but it cannot twine round a stick, though it can ascend by clasping the stick horizontally with one or both of its unbranched tendrils. These tendrils continually insert their pointed ends into minute crevices or holes, but as they are always withdrawn by the subsequent spiral contraction, the habit seems to us in our ignorance useless. Lastly, the stem of B. capreolata twines imperfectly; the much-branched tendrils revolve in a capricious manner, and bend from the light to the dark; their hooked extremities, even whilst immature, crawl into crevices, and, when mature, seize any thin projecting point; in either case they develop adhesive discs, and these have the power of enveloping the finest fibres.

In the allied Eccremocarpus the internodes, petioles, and much- branched tendrils all spontaneously revolve together. The tendrils do not as a whole turn from the light; but their bluntly-hooked extremities arrange themselves neatly on any surface with which they come into contact, apparently so as to avoid the light. They act best when each branch seizes a few thin stems, like the culms of a grass, which they afterwards draw together into a solid bundle by the spiral contraction of all the branches. In Cobaea the finely- branched tendrils alone revolve; the branches terminate in sharp, hard, double, little hooks, with both points directed to the same side; and these turn by well-adapted movements to any object with which they come into contact. The tips of the branches also crawl into dark crevices or holes. The tendrils and internodes of Ampelopsis have little or no power of revolving; the tendrils are but little sensitive to contact; their hooked extremities cannot seize thin objects; they will not even clasp a stick, unless in extreme need of a support; but they turn from the light to the dark, and, spreading out their branches in contact with any nearly flat surface, develop discs. These adhere by the secretion of some cement to a wall, or even to a polished surface; and this is more than the discs of the Bignonia capreolata can effect.

The rapid development of these adherent discs is one of the most remarkable peculiarities possessed by any tendrils. We have seen that such discs are formed by two species of Bignonia, by Ampelopsis, and, according to Naudin, {37} by the Cucurbitaceous genus Peponopsis adhaerens. In Anguria the lower surface of the tendril, after it has wound round a stick, forms a coarsely cellular layer, which closely fits the wood, but is not adherent; whilst in Hanburya a similar layer is adherent. The growth of these cellular out-growths depends, (except in the case of the Haplolophium and of one species of Ampelopsis,) on the stimulus from contact. It is a singular fact that three families, so widely distinct as the Bignoniaceae, Vitaceae, and Cucurbitaceae, should possess species with tendrils having this remarkable power.

Sachs attributes all the movements of tendrils to rapid growth on the side opposite to that which becomes concave. These movements consist of revolving nutation, the bending to and from the light, and in opposition to gravity, those caused by a touch, and spiral contraction. It is rash to differ from so great an authority, but I cannot believe that one at least of these movements—curvature from a touch—is thus caused. {38} In the first place it may be remarked that the movement of nutation differs from that due to a touch, in so far that in some cases the two powers are acquired by the same tendril at different periods of growth; and the sensitive part of the tendril does not seem capable of nutation. One of my chief reasons for doubting whether the curvature from a touch is the result of growth, is the extraordinary rapidity of the movement. I have seen the extremity of a tendril of Passiflora gracilis, after being touched, distinctly bent in 25 seconds, and often in 30 seconds; and so it is with the thicker tendril of Sicyos. It appears hardly credible that their outer surfaces could have actually grown in length, which implies a permanent modification of structure, in so short a time. The growth, moreover, on this view must be considerable, for if the touch has been at all rough the extremity is coiled in two or three minutes into a spire of several turns.

When the extreme tip of the tendril of Echinocystis caught hold of a smooth stick, it coiled itself in a few hours (as described at p. 132) twice or thrice round the stick, apparently by an undulatory movement. At first I attributed this movement to the growth of the outside; black marks were therefore made, and the interspaces measured, but I could not thus detect any increase in length. Hence it seems probable in this case and in others, that the curvature of the tendril from a touch depends on the contraction of the cells along the concave side. Sachs himself admits {39} that "if the growth which takes place in the entire tendril at the time of contact with a support is small, a considerable acceleration occurs on the convex surface, but in general there is no elongation on the concave surface, or there may even be a contraction; in the case of a tendril of Cucurbita this contraction amounted to nearly one-third of the original length." In a subsequent passage Sachs seems to feel some difficulty in accounting for this kind of contraction. It must not however be supposed from the foregoing remarks that I entertain any doubt, after reading De Vries' observations, about the outer and stretched surfaces of attached tendrils afterwards increasing in length by growth. Such increase seems to me quite compatible with the first movement being independent of growth. Why a delicate touch should cause one side of a tendril to contract we know as little as why, on the view held by Sachs, it should lead to extraordinarily rapid growth of the opposite side. The chief or sole reason for the belief that the curvature of a tendril when touched is due to rapid growth, seems to be that tendrils lose their sensitiveness and power of movement after they have grown to their full length; but this fact is intelligible, if we bear in mind that all the functions of a tendril are adapted to drag up the terminal growing shoot towards the light. Of what use would it be, if an old and full-grown tendril, arising from the lower part of a shoot, were to retain its power of clasping a support? This would be of no use; and we have seen with tendrils so many instances of close adaptation and of the economy of means, that we may feel assured that they would acquire irritability and the power of clasping a support at the proper age—namely, youth- -and would not uselessly retain such power beyond the proper age.



CHAPTER V.—HOOK AND ROOT-CLIMBERS.—CONCLUDING REMARKS.



Plants climbing by the aid of hooks, or merely scrambling over other plants—Root-climbers, adhesive matter secreted by the rootlets— General conclusions with respect to climbing plants, and the stages of their development.

Hook-Climbers.—In my introductory remarks, I stated that, besides the two first great classes of climbing plants, namely, those which twine round a support, and those endowed with irritability enabling them to seize hold of objects by means of their petioles or tendrils, there are two other classes, hook-climbers and root-climbers. Many plants, moreover, as Fritz Muller has remarked, {40} climb or scramble up thickets in a still more simple fashion, without any special aid, excepting that their leading shoots are generally long and flexible. It may, however, be suspected from what follows, that these shoots in some cases tend to avoid the light. The few hook- climbers which I have observed, namely, Galium aparine, Rubus australis, and some climbing Roses, exhibit no spontaneous revolving movement. If they had possessed this power, and had been capable of twining, they would have been placed in the class of Twiners; for some twiners are furnished with spines or hooks, which aid them in their ascent. For instance, the Hop, which is a twiner, has reflexed hooks as large as those of the Galium; some other twiners have stiff reflexed hairs; and Dipladenia has a circle of blunt spines at the bases of its leaves. I have seen only one tendril-bearing plant, namely, Smilax aspera, which is furnished with reflexed spines; but this is the case with several branch-climbers in South Brazil and Ceylon; and their branches graduate into true tendrils. Some few plants apparently depend solely on their hooks for climbing, and yet do so efficiently, as certain palms in the New and Old Worlds. Even some climbing Roses will ascend the walls of a tall house, if covered with a trellis. How this is effected I know not; for the young shoots of one such Rose, when placed in a pot in a window, bent irregularly towards the light during the day and from the light during the night, like the shoots of any common plant; so that it is not easy to understand how they could have got under a trellis close to the wall. {41}

Root-climbers.—A good many plants come under this class, and are excellent climbers. One of the most remarkable is the Marcgravia umbellata, the stem of which in the tropical forests of South America, as I hear from Mr. Spruce, grows in a curiously flattened manner against the trunks of trees; here and there it puts forth claspers (roots), which adhere to the trunk, and, if the latter be slender, completely embrace it. When this plant has climbed to the light, it produces free branches with rounded stems, clad with sharp- pointed leaves, wonderfully different in appearance from those borne by the stem as long as it remains adherent. This surprising difference in the leaves, I have also observed in a plant of Marcgravia dubia in my hothouse. Root-climbers, as far as I have seen, namely, the Ivy (Hedera helix), Ficus repens, and F. barbatus, have no power of movement, not even from the light to the dark. As previously stated, the Hoya carnosa (Asclepiadaceae) is a spiral twiner, and likewise adheres by rootlets even to a flat wall. The tendril-bearing Bignonia Tweedyana emits roots, which curve half round and adhere to thin sticks. The Tecoma radicans (Bignoniaceae), which is closely allied to many spontaneously revolving species, climbs by rootlets; nevertheless, its young shoots apparently move about more than can be accounted for by the varying action of the light.

I have not closely observed many root-climbers, but can give one curious fact. Ficus repens climbs up a wall just like Ivy; and when the young rootlets are made to press lightly on slips of glass, they emit after about a week's interval, as I observed several times, minute drops of clear fluid, not in the least milky like that exuded from a wound. This fluid is slightly viscid, but cannot be drawn out into threads. It has the remarkable property of not soon drying; a drop, about the size of half a pin's head, was slightly spread out on glass, and I scattered on it some minute grains of sand. The glass was left exposed in a drawer during hot and dry weather, and if the fluid had been water, it would certainly have dried in a few minutes; but it remained fluid, closely surrounding each grain of sand, during 128 days: how much longer it would have remained I cannot say. Some other rootlets were left in contact with the glass for about ten days or a fortnight, and the drops of secreted fluid were now rather larger, and so viscid that they could be drawn out into threads. Some other rootlets were left in contact during twenty-three days, and these were firmly cemented to the glass. Hence we may conclude that the rootlets first secrete a slightly viscid fluid, subsequently absorb the watery parts, (for we have seen that the fluid will not dry by itself,) and ultimately leave a cement. When the rootlets were torn from the glass, atoms of yellowish matter were left on it, which were partly dissolved by a drop of bisulphide of carbon; and this extremely volatile fluid was rendered very much less volatile by what it had dissolved.

As the bisulphide of carbon has a strong power of softening indurated caoutchouc, I soaked in it during a short time several rootlets of a plant which had grown up a plaistered wall; and I then found many extremely thin threads of transparent, not viscid, excessively elastic matter, precisely like caoutchouc, attached to two sets of rootlets on the same branch. These threads proceeded from the bark of the rootlet at one end, and at the other end were firmly attached to particles of silex or mortar from the wall. There could be no mistake in this observation, as I played with the threads for a long time under the microscope, drawing them out with my dissecting- needles and letting them spring back again. Yet I looked repeatedly at other rootlets similarly treated, and could never again discover these elastic threads. I therefore infer that the branch in question must have been slightly moved from the wall at some critical period, whilst the secretion was in the act of drying, through the absorption of its watery parts. The genus Ficus abounds with caoutchouc, and we may conclude from the facts just given that this substance, at first in solution and ultimately modified into an unelastic cement, {42} is used by the Ficus repens to cement its rootlets to any surface which it ascends. Whether other plants, which climb by their rootlets, emit any cement I do not know; but the rootlets of the Ivy, placed against glass, barely adhered to it, yet secreted a little yellowish matter. I may add, that the rootlets of the Marcgravia dubia can adhere firmly to smooth painted wood.

Vanilla aromatica emits aerial roots a foot in length, which point straight down to the ground. According to Mohl (p. 49), these crawl into crevices, and when they meet with a thin support, wind round it, as do tendrils. A plant which I kept was young, and did not form long roots; but on placing thin sticks in contact with them, they certainly bent a little to that side, in the course of about a day, and adhered by their rootlets to the wood; but they did not bend quite round the sticks, and afterwards they re-pursued their downward course. It is probable that these slight movements of the roots are due to the quicker growth of the side exposed to the light, in comparison with the other side, and not because the roots are sensitive to contact in the same manner as true tendrils. According to Mohl, the rootlets of certain species of Lycopodium act as tendrils. {43}

Concluding Remarks on Climbing Plants.

Plants become climbers, in order, as it may be presumed, to reach the light, and to expose a large surface of their leaves to its action and to that of the free air. This is effected by climbers with wonderfully little expenditure of organized matter, in comparison with trees, which have to support a load of heavy branches by a massive trunk. Hence, no doubt, it arises that there are so many climbing plants in all quarters of the world, belonging to so many different orders. These plants have been arranged under four classes, disregarding those which merely scramble over bushes without any special aid. Hook-climbers are the least efficient of all, at least in our temperate countries, and can climb only in the midst of an entangled mass of vegetation. Root-climbers are excellently adapted to ascend naked faces of rock or trunks of trees; when, however, they climb trunks they are compelled to keep much in the shade; they cannot pass from branch to branch and thus cover the whole summit of a tree, for their rootlets require long-continued and close contact with a steady surface in order to adhere. The two great classes of twiners and of plants with sensitive organs, namely, leaf-climbers and tendril-bearers taken together, far exceed in number and in the perfection of their mechanism the climbers of the two first classes. Those which have the power of spontaneously revolving and of grasping objects with which they come in contact, easily pass from branch to branch, and securely ramble over a wide, sun-lit surface.

The divisions containing twining plants, leaf-climbers, and tendril- bearers graduate to a certain extent into one another, and nearly all have the same remarkable power of spontaneously revolving. Does this gradation, it may be asked, indicate that plants belonging to one subdivision have actually passed during the lapse of ages, or can pass, from one state to the other? Has, for instance, any tendril- bearing plant assumed its present structure without having previously existed as a leaf-climber or a twiner? If we consider leaf-climbers alone, the idea that they were primordially twiners is forcibly suggested. The internodes of all, without exception, revolve in exactly the same manner as twiners; some few can still twine well, and many others in an imperfect manner. Several leaf-climbing genera are closely allied to other genera which are simple twiners. It should also be observed, that the possession of leaves with sensitive petioles, and with the consequent power of clasping an object, would be of comparatively little use to a plant, unless associated with revolving internodes, by which the leaves are brought into contact with a support; although no doubt a scrambling plant would be apt, as Professor Jaeger has remarked, to rest on other plants by its leaves. On the other hand, revolving internodes, without any other aid, suffice to give the power of climbing; so that it seems probable that leaf-climbers were in most cases at first twiners, and subsequently became capable of grasping a support; and this, as we shall presently see, is a great additional advantage.

From analogous reasons, it is probable that all tendril-bearers were primordially twiners, that is, are the descendants of plants having this power and habit. For the internodes of the majority revolve; and, in a few species, the flexible stem still retains the capacity of spirally twining round an upright stick. Tendril-bearers have undergone much more modification than leaf-climbers; hence it is not surprising that their supposed primordial habits of revolving and twining have been more frequently lost or modified than in the case of leaf-climbers. The three great tendril-bearing families in which this loss has occurred in the most marked manner, are the Cucurbitaceae, Passifloraceae, and Vitaceae. In the first, the internodes revolve; but I have heard of no twining form, with the exception (according to Palm, p. 29. 52) of Momordica balsamina, and this is only an imperfect twiner. In the two other families I can hear of no twiners; and the internodes rarely have the power of revolving, this power being confined to the tendrils. The internodes, however, of Passiflora gracilis have the power in a perfect manner, and those of the common Vine in an imperfect degree: so that at least a trace of the supposed primordial habit has been retained by some members of all the larger tendril-bearing groups.

On the view here given, it may be asked, Why have the species which were aboriginally twiners been converted in so many groups into leaf- climbers or tendril-bearers? Of what advantage has this been to them? Why did they not remain simple twiners? We can see several reasons. It might be an advantage to a plant to acquire a thicker stem, with short internodes bearing many or large leaves; and such stems are ill fitted for twining. Any one who will look during windy weather at twining plants will see that they are easily blown from their support; not so with tendril-bearers or leaf-climbers, for they quickly and firmly grasp their support by a much more efficient kind of movement. In those plants which still twine, but at the same time possess tendrils or sensitive petioles, as some species of Bignonia, Clematis, and Tropaeolum, it can readily be observed how incomparably better they grasp an upright stick than do simple twiners. Tendrils, from possessing this power of grasping an object, can be made long and thin; so that little organic matter is expended in their development, and yet they sweep a wide circle in search of a support. Tendril-bearers can, from their first growth, ascend along the outer branches of any neighbouring bush, and they are thus always fully exposed to the light; twiners, on the contrary, are best fitted to ascend bare stems, and generally have to start in the shade. Within tall and dense tropical forests, twining plants would probably succeed better than most kinds of tendril-bearers; but the majority of twiners, at least in our temperate regions, from the nature of their revolving movement, cannot ascend thick trunks, whereas this can be affected by tendril-bearers if the trunks are branched or bear twigs, and by some species if the bark is rugged.