Oxalis sensitiva.—The leaflets, as in the last species, bend vertically down at night, without becoming folded. The much elongated main petiole rises considerably in the evening, but in [page 328] some very young plants the rise did not commence until late at night. We have seen that the cotyledons, instead of sinking like the leaflets, rise up vertically at night.

Oxalis bupleurifolia.—This species is rendered remarkable by the petioles being foliaceous, like the phyllodes of many Acacias. The leaflets are small, of a paler green and more tender consistence than the foliaceous petioles. The leaflet which was observed was .55 inch in length, and was borne by a petiole 2 inches long and .3 inch broad. It may be suspected that the leaflets are on the road to abortion or obliteration, as has actually occurred with those of another Brazilian species, O. rusciformis. Nevertheless, in the present species the nyctitropic movements are perfectly performed. The foliaceous petiole was first observed during 48 h., and found to be in continued circumnutation, as shown in the accompanying figure (Fig. 130). It rose during the day and early part of the night, and fell during the remainder of the night and early morning; but the movement was not sufficient to be called sleep. The ascending and descending lines did not coincide, so that an ellipse was formed each day. There was but little zigzagging; if the filament had been fixed longitudinally, we should probably have seen that there was more lateral movement than appears in the diagram.

Fig. 130. Oxalis bupleurifolia: circumnutation of foliaceous petiole, filament fixed obliquely across end of petiole; movements traced on vertical glass from 9 A.M. June 26th to 8.50 A.M. 28th. Apex of leaflet 4 inches from the glass, so movement not much magnified. Plant 9 inches high, illuminated from above. Temp. 23 1/2o - 24 1/2o C.

A terminal leaflet on another leaf was next observed (the petiole being secured), and its movements are shown in Fig. 131. During the day the leaflets are extended horizontally, and at night depend vertically; and as the petiole rises during the day the leaflets have to bend down in the evening [page 329] more than 90o, so as to assume at night their vertical position. On the first day the leaflet simply moved up and down; on the

Fig. 131. Oxalis bupleurifolia: circumnutation and nyctitropic movement of terminal leaflet, with filament affixed along the midrib; traced on a vertical glass from 9 A.M. on June 26th to 8.45 A.M. 28th. Conditions the same as in the last case.

second day it plainly circumnutated between 8 A.M. and 4.30 P.M., after which hour the great evening fall commenced. [page 330]

Averrhoa bilimbi (Oxalidae).—It has long been known,* firstly, that the leaflets in this genus sleep; secondly, that they move spontaneously during the day; and thirdly, that they are sensitive to a touch; but in none of these respects do they differ essentially from the species of Oxalis. They differ, however, as Mr. R. I. Lynch** has lately shown, in their spontaneous movements being strongly marked. In the case of A. bilimbi, it is a wonderful spectacle to behold on a warm sunny day the leaflets one after the other sinking rapidly downwards, and again ascending slowly. Their movements rival those of Desmodium gyrans. At night the leaflets hang vertically down; and now

Fig. 132. Averrhoa bilimbi: leaf asleep; drawing reduced.

they are motionless, but this may be due to the opposite ones being pressed together (Fig. 132). The main petiole is in constant movement during the day, but no careful observations were made on it. The following diagrams are graphic representations of the variations in the angle, which a given leaflet makes with the vertical. The observations were made as follows. The plant growing in a pot was kept in a high temperature, the petiole of the leaf to be observed pointing straight at the observer, being separated from him by a vertical pane of glass. The petiole was secured so that the basal joint, or pulvinus, of one of the lateral leaflets was at the centre of a graduated arc placed close behind the leaflet. A fine glass filament was fixed to the leaf, so as to project like a continuation of the

* Dr. Bruce, 'Philosophical Trans.,' 1785, p. 356.

** 'Journal Linn. Soc.,' vol. xvi. 1877, p. 231. [page 331]

midrib. This filament acted as an index; and as the leaf rose and fell, rotating about its basal joint, its angular movement

Fig. 133. Averrhoa bilimbi: angular movements of a leaflet during its evening descent, when going to sleep. Temp. 78o - 81o F.

could be recorded by reading off at short intervals of time the position of the glass filament on the graduated arc. In order [page 332] to avoid errors of parallax, all readings were made by looking through a small ring painted on the vertical glass, in a line with the joint of the leaflet and the centre of the graduated arc. In the following diagrams the ordinates represent the angles which the leaflet made with the vertical at successive instants.* It follows that a fall in the curve represents an actual dropping of the leaf, and that the zero line represents a vertically dependent position. Fig. 133 represents the nature of the movements which occur in the evening, as soon as the leaflets begin to assume their nocturnal position. At 4.55 P.M. the leaflet formed an angle of 85o with the vertical, or was only 5o below the horizontal; but in order that the diagram might get into our page, the leaflet is represented falling from 75o instead of 85o. Shortly after 6 P.M. it hung vertically down, and had attained its nocturnal position. Between 6.10 and 6.35 P.M. it performed a number of minute oscillations of about 2o each, occupying periods of 4 or 5 m. The complete state of rest of the leaflet which ultimately followed is not shown in the diagram. It is manifest that each oscillation consists of a gradual rise, followed by a sudden fall. Each time the leaflet fell, it approached nearer to the nocturnal position than it did on the previous fall. The amplitude of the oscillations diminished, while the periods of oscillation became shorter.

In bright sunshine the leaflets assume a highly inclined dependent position. A leaflet in diffused light was observed rising for 25 m. A blind was then pulled up so that the plant was brightly illuminated (BR in Fig. 134), and within a minute it began to fall, and ultimately fell 47o, as shown in the diagram. This descent was performed by six descending steps, precisely similar to those by which the nocturnal fall is effected. The plant was then again shaded (SH), and a long slow rise occurred until another series of falls commenced at BR', when the sun was again admitted. In this experiment cool air was allowed to enter by the windows being opened at the same time that the blinds were pulled up, so that in spite of the sun shining on the plant the temperature was not raised.

The effect of an increase of temperature in diffused light is

* In all the diagrams 1 mm. in the horizontal direction represents one minute of time. Each mm. in the vertical direction represents one degree of angular movement. In Figs. 133 and 134 the temperature is represented (along the ordinates) in the scale of 1 mm. to each 0.1 degree C. In Fig. 135 each mm. equals 0.2o F. [page 333]

shown in Fig. 135. The temperature began to rise at 11.35 A.M. (in consequence of the fire being lighted), but by 12.42 a marked fall had occurred. It may be seen in the diagram that when the temperature was highest there were rapid oscillations

Fig. 134. Averrhoa bilimbi: angular movements of leaflet during a change from bright illumination to shade; temperature (broken line) remaining nearly the same.

of small amplitude, the mean position of the leaflet being at the time nearer the vertical. When the temperature began to fall, the oscillations became slower and larger, and the mean position of the leaf again approached the horizontal. The rate of oscillation was sometimes quicker than is represented in the above diagram. Thus, when the temperature was between 31o and [page 334]

Fig. 135. Averrhoa bilimbi: angular movement of leaflet during a change of temperature; light remaining the same. The broken line shows the change of temperature. [page 335]

32o C., 14 oscillations of a few degrees occurred in 19 m. On the other hand, an oscillation may be much slower; thus a leaflet was observed (temperature 25o C.) to rise during 40 m. before it fell and completed its oscillation.

Fig. 136. Porlieria hygrometrica: circumnutation and nyctitropic movements of petiole of leaf, traced from 9.35 A.M. July 7th to about midnight on the 8th. Apex of leaf 7 inches from the vertical glass. Temp. 19 1/2o - 20 1/2o C.

Porlieria hygrometrica (Zygophylleae).—The leaves of this plant (Chilian form) are from 1 to 1 inch in length, and bear as many as 16 or 17 small leaflets on each side, which do not stand opposite one another. They are articulated to the petiole, and the petiole to the branch by a pulvinus. We must premise that apparently two forms are confounded under the same name: the leaves on a bush from Chili, which was sent to us from Kew, bore many leaflets, whilst those on plants in the Botanic Garden at Wrzburg bore only 8 or 9 pairs; and the whole character of the bushes appeared somewhat different. We shall also see that they differ in a remarkable physiological peculiarity. On the Chilian plant the petioles of the younger leaves on upright branches, stood horizontally during the day, and at night sank down vertically so as to depend parallel and close to the branch beneath. The petioles of rather older leaves did not become at night vertically depressed, but only highly inclined. In one instance we found a branch which had grown perpendicularly downwards, and the petioles on it moved in the same direction relatively to the branch as just stated, and therefore moved upwards. On horizontal branches the younger petioles likewise move at night in the same direction as before, that is, towards the branch, and are consequently then extended horizontally; but it is remarkable that the older petioles on the [page 336] same branch, though moving a little in the same direction, also bend downwards; they thus occupy a somewhat different position, relatively to the centre of the earth and to the branch, from that of the petioles on the upright branches. With respect to the leaflets, they move at night towards the apex of the petiole until their midribs stand nearly parallel to it; and they then lie neatly imbricated one over the other. Thus half of the upper surface of each leaflet is in close contact with half of the lower surface of the one next in advance; and all the leaflets, excepting the basal ones, have the whole of their upper surfaces and half of their lower surfaces well protected. Those on the opposite sides of the same petiole do not come into close contact at night, as occurs with the leaflets of so many Leguminosae but are separated by an open furrow; nor could they exactly coincide, as they stand alternately with respect to one another.

The circumnutation of the petiole of a leaf 3/4 of an inch in length, on an upright branch, was observed during 36h., and is shown in the preceding diagram (Fig. 136). On the first morning, the leaf fell a little and then rose until 1 P.M., and this was probably due to its being now illuminated through a skylight from above; it then circumnutated on a very small scale round the same spot until about 4 P.M., when the great evening fall commenced. During the latter part of the night or very early on the next morning the leaf rose again. On the second day it fell during the morning till 1 P.M., and this no doubt is its normal habit. From 1 to 4 P.M. it rose in a zigzag line, and soon afterwards the great evening fall commenced. It thus completed a double oscillation during the 24 h.

The specific name given to this plant by Ruiz and Pavon, indicates that in its native arid home it is affected in some manner by the dryness or dampness of the atmosphere.* In the Botanic Garden at Wrzburg, there was a plant in a pot out of doors which was daily watered, and another in the open ground which was never watered. After some hot and dry weather there was a great difference in the state of the leaflets on these two plants; those on the unwatered plant in the open ground remaining half,

* 'Systema Veg. Florae Peruvianae et Chilensis,' tom. i. p. 95, 1798. We cannot understand the account given by the authors of the behaviour of this plant in its native home. There is much about its power of foretelling changes in the weather; and it appears as if the brightness of the sky largely determined the opening and closing of the leaflets. [page 337]

or even quite, closed during the day. But twigs cut from this bush, with their ends standing in water, or wholly immersed in it, or kept in damp air under a bell-glass, opened their leaves though exposed to a blazing sun; whilst those on the plant in the ground remained closed. The leaves on this same plant, after some heavy rain, remained open for two days; they then became half closed during two days, and after an additional day were quite closed. This plant was now copiously watered, and on the following morning the leaflets were fully expanded. The other plant growing in a pot, after having been exposed to heavy rain, was placed before a window in the Laboratory, with its leaflets open, and they remained so during the daytime for 48 h.; but after an additional day were half closed. The plant was then watered, and the leaflets on the two following days remained open. On the third day they were again half closed, but on being again watered remained open during the two next days. From these several facts we may conclude that the plant soon feels the want of water; and that as soon as this occurs, it partially or quite closes its leaflets, which in their then imbricated condition expose a small surface to evaporation. It is therefore probable that this sleep-like movement, which occurs only when the ground is dry, is an adaptation against the loss of moisture.

A bush about 4 feet in height, a native of Chili, which was thickly covered with leaves, behaved very differently, for during the day it never closed its leaflets. On July 6th the earth in the small pot in which it grew appeared extremely dry, and it was given a very little water. After 21 and 22 days (on the 27th and 28th), during the whole of which time the plant did not receive a drop of water, the leaves began to droop, but they showed no signs of closing during the day. It appeared almost incredible that any plant, except a fleshy one, could have kept alive in soil so dry, which resembled the dust on a road. On the 29th, when the bush was shaken, some leaves fell off, and the remaining ones were unable to sleep at night. It was therefore moderately watered, as well as syringed, late in the evening. On the next morning (30th) the bush looked as fresh as ever, and at night the leaves went to sleep. It may be added that a small branch while growing on the bush was enclosed, by means of a curtain of bladder, during 13 days in a large bottle half full of quicklime, so that the air within must have been intensely dry; yet the leaves on this branch did not suffer in the [page 338] least, and did not close at all during the hottest days. Another trial was made with the same bush on August 2nd and 6th (the soil appearing at this latter date extremely dry), for it was exposed out of doors during the whole day to the wind, but the leaflets showed no signs of closing. The Chilian form therefore differs widely from the one at Wrzburg, in not closing its leaflets when suffering from the want of water; and it can live for a surprisingly long time without water.

Tropaeolum majus (?) (cultivated var.) (Tropaeoleae).—Several plants in pots stood in the greenhouse, and the blades of the leaves which faced the front-lights were during the day highly inclined and at night vertical; whilst the leaves on the back of the pots, though of course illuminated through the roof, did not become vertical at night. We thought, at first, that this difference in their positions was in some manner due to heliotropism, for the leaves are highly heliotropic. The true explanation, however, is that unless they are well illuminated during at least a part of the day they do not sleep at night; and a little difference in the degree of illumination determines whether or not they shall become vertical at night. We have observed no other so well-marked a case as this, of the influence of previous illumination on nyctitropic movements. The leaves present also another peculiarity in their habit of rising or awaking in the morning, being more strongly fixed or inherited than that of sinking or sleeping at night. The movements are caused by the bending of an upper part of the petiole, between and 1 inch in length; but the part close to the blade, for about 1/4 of an inch in length, does not bend and always remains at right angles to the blade. The bending portion does not present any external or internal difference in structure from the rest of the petiole. We will now give the experiments on which the above conclusions are founded.

A large pot with several plants was brought on the morning of Sept. 3rd out of the greenhouse and placed before a north-east window, in the same position as before with respect to the light, as far as that was possible. On the front of the plants, 24 leaves were marked with thread, some of which had their blades horizontal, but the greater number were inclined at about 45o, beneath the horizon; at night all these, without exception, became vertical. Early on the following morning (4th) they reassumed their former positions, and at night again became vertical. On the 5th the shutters were opened at 6.15 A.M., and [page 339] by 8.18 A.M., after the leaves had been illuminated for 2 h. 3 m. and had acquired their diurnal position, they were placed in a dark cupboard. They were looked at twice during the day and thrice in the evening, the last time at 10.30 P.M., and not one had become vertical. At 8 A.M. on the following morning (6th) they still retained the same diurnal position, and were now replaced before the north-east window. At night all the leaves which had faced the light had their petioles curved and their blades vertical; whereas none of the leaves on the back of the plants, although they had been moderately illuminated by the diffused light of the room, were vertical. They were now at night placed in the same dark cupboard; at 9 A.M. on the next morning (7th) all those which had been asleep had reassumed their diurnal position. The pot was then placed for 3 h. in the sunshine, so as to stimulate the plants; at noon they were placed before the same north-east window, and at night the leaves slept in the usual manner and awoke on the following morning. At noon on this day (8th) the plants, after having been left before the north-east window for 5 h. 45 m. and thus illuminated (though not brightly, as the sky was cloudy during the whole time), were replaced in the dark cupboard, and at 3 P.M. the position of the leaves was very little, if at all, altered, so that they are not quickly affected by darkness; but by 10.15 P.M. all the leaves which had faced the north-east sky during the 5 h. 45 m. of illumination stood vertical, whereas those on the back of the plant retained their diurnal position. On the following morning (9th) the leaves awoke as on the two former occasions in the dark, and they were kept in the dark during the whole day; at night a very few of them became vertical, and this was the one instance in which we observed any inherited tendency or habit in this plant to sleep at the proper time. That it was real sleep was shown by these same leaves reassuming their diurnal position on the following morning (10th) whilst still kept in the dark.

The pot was then (9.45 A.M. 10th) replaced, after having been kept for 36 h. in darkness, before the north-east window; and at night the blades of all the leaves (excepting a few on the back of the plants) became conspicuously vertical. At 6.45 A.M. (11th) after the plants had been illuminated on the same side as before during only 25 m., the pot was turned round, so that the leaves which had faced the light now faced the interior of the room, and not one of these went to sleep at night; [page 340] whilst some, but not many, of those which had formerly stood facing the back of the room and which had never before been well illuminated or gone to sleep, now assumed a vertical position at night. On the next day (12th) the plant was turned round into its original position, so that the same leaves faced the light as formerly, and these now went to sleep in the usual manner. We will only add that with some young seedlings kept in the greenhouse, the blades of the first pair of true leaves (the cotyledons being hypogean) stood during the day almost horizontally and at night almost vertically.

A few observations were subsequently made on the circumnutation of three leaves, whilst facing a north-east window; but the tracings are not given, as the leaves moved somewhat towards the light. It was, however, manifest that they rose and fell more than once during the daytime, the ascending and descending lines being in parts extremely zigzag. The nocturnal fall commenced about 7 P.M., and the leaves had risen considerably by 6.45 A.M. on the following morning.

Leguminosae.—This Family includes many more genera with sleeping species than all the other families put together. The number of the tribes to which each genus belongs, according to Bentham and Hooker's arrangement, has been added.

Crotolaria (sp.?) (Tribe 2).—This plant is monophyllous, and we are informed by Mr. T. Thiselton Dyer that the leaves rise up vertically at night and press against the stem.

Lupinus (Tribe 2).—The palmate or digitate leaves of the species in this large genus sleep in three different manners. One of the simplest, is that all the leaflets become steeply inclined downwards at night, having been during the day extended horizontally. This is shown in the accompanying figures (Fig. 137), of a leaf of L. pilosus, as seen during the day from vertically above, and of another leaf asleep with the leaflets inclined downwards. As in this position they are crowded together, and as they do not become folded like those in the genus Oxalis, they cannot occupy a vertically dependent position; but they are often inclined at an angle of 50o beneath the horizon. In this species, whilst the leaflets are sinking, the petioles rise up, in two instances when the angles were measured to the extent of 23o. The leaflets of L. sub-carnosus and arboreus, which were horizontal during the day, sank down at night in nearly the same manner; the former to an angle of 38o and the latter of 36o, beneath the horizon; but their petioles [page 341] did not move in any plainly perceptible degree. It is, however, quite possible, as we shall presently see, that if a large number of plants of the three foregoing and of the following species

Fig. 137. Lupinus pilosus: A, leaf seen from vertically above in daytime; B, leaf asleep, seen laterally at night.

were to be observed at all seasons, some of the leaves would be found to sleep in a different manner.

In the two following species the leaflets, instead of moving downwards, rise at night. With L. Hartwegii some stood at noon at a mean angle of 36o above the horizon, and at night at 51o, thus forming together a hollow cone with moderately steep sides. The petiole of one leaf rose 14o and of a second 11o at night. With L. luteus a leaflet rose from 47o at noon to 65o above the horizon at night, and another on a distinct leaf rose from 45o to 69o. The petioles, however, sink at night to a small extent, viz., in three instances by 2o, 6o, and 9o 30 seconds. Owing to this movement of the petioles, the outer and longer leaflets have to bend up a little more than the shorter and inner ones, in order that all should stand symmetrically at night. We shall presently see that some leaves on the same individual plants of L. luteus sleep in a very different manner.

We now come to a remarkable position of the leaves when asleep, which is common to several species of Lupines. On the same leaf the shorter leaflets, which generally face the centre of the plant, sink at night, whilst the longer ones on the opposite side rise; the intermediate and lateral ones merely twisting on their own axes. But there is some variability with respect to which leaflets rise or fall. As might have been expected from such diverse and complicated movements, the [page 342] base of each leaflet is developed (at least in the case of L. luteus) into a pulvinus. The result is that all the leaflets on the same leaf stand at night more or less highly inclined, or even quite vertically, forming in this latter case a vertical star. This occurs with the leaves of a species purchased under the name of

Fig. 138. Lupinus pubescens: A, leaf viewed laterally during the day; B, same leaf at night; C, another leaf with the leaflet forming a vertical star at night. Figures reduced.

L. pubescens; and in the accompanying figures we see at A (Fig. 138) the leaves in their diurnal position; and at B the same plant at night with the two upper leaves having their leaflets almost vertical. At C another leaf, viewed laterally, is shown with the leaflets quite vertical. It is chiefly or exclusively the youngest leaves which form at night vertical stars. But there [page 343] is much variability in the position of the leaves at night on the same plant; some remaining with their leaflets almost horizontal, others forming more or less highly inclined or vertical stars, and some with all their leaflets sloping downwards, as in our first class of cases. It is also a remarkable fact, that although all the plants produced from the same lot of seeds were identical in appearance, yet some individuals at night had the leaflets of all their leaves arranged so as to form more or less highly inclined stars; others had them all sloping downwards and never forming a star; and others, again, retained them either in a horizontal position or raised them a little.

We have as yet referred only to the different positions of the leaflets of L. pubescens at night; but the petioles likewise differ in their movements. That of a young leaf which formed a highly inclined star at night, stood at noon at 42o above the horizon, and during the night at 72o, so had risen 30o. The petiole of another leaf, the leaflets of which occupied a similar position at night, rose only 6o. On the other hand, the petiole of a leaf with all its leaflets sloping down at night, fell at this time 4o. The petioles of two rather older leaves were subsequently observed; both of which stood during the day at exactly the same angle, viz., 50o above the horizon, and one of these rose 7o - 8o, and the other fell 3o - 4o at night. We meet with cases like that of L. pubescens with some other species. On a single plant of L. mutabilis some leaves, which stood horizontally during the day, formed highly inclined stars at night, and the petiole of one rose 7o. Other leaves which likewise stood horizontally during the day, had at night all their leaflets sloping downwards at 46o beneath the horizon, but their petioles had hardly moved. Again, L. luteus offered a still more remarkable case, for on two leaves, the leaflets which stood at noon at about 45o above the horizon, rose at night to 65o and 69o, so that they formed a hollow cone with steep sides. Four leaves on the same plant, which had their leaflets horizontal at noon, formed vertical stars at night; and three other leaves equally horizontal at noon, had all their leaflets sloping downwards at night. So that the leaves on this one plant assumed at night three different positions. Though we cannot account for this fact, we can see that such a stock might readily give birth to species having widely different nyctitropic habits.

Little more need be said about the sleep of the species of Lupinus; several, namely, L. polyphyllus, nanus, Menziesii, speciosus, [page 344] and albifrons, though observed out of doors and in the greenhouse, did not change the position of their leaves sufficiently at night to be said to sleep. From observations made on two sleeping species, it appears that, as with Tropaeolum majus, the leaves must be well illuminated during the day in order to sleep at night. For several plants, kept all day in a sitting-room with north-east windows, did not sleep at night; but when the pots were placed on the following day out of doors, and were brought in at night, they slept in the usual manner. the trial was repeated on the following day and night with the same result.

Some observations were made on the circumnutation of the leaves of L. luteus and arboreus. It will suffice to say that the leaflets of the latter exhibited a double oscillation in the course of 24 h.; for they fell from the early morning until 10.15 A.M., then rose and zigzagged greatly till 4 P.M., after which hour the great nocturnal fall commenced. By 8 A.M. on the following morning the leaflets had risen to their proper height. We have seen in the fourth chapter, that the leaves of Lupinus speciosus, which do not sleep, circumnutate to an extraordinary extent, making many ellipses in the course of the day.

Cytisus (Tribe 2), Trigonella and Medicago (Tribe 3).—Only

Fig. 139. Medicago marina: A, leaves during the day; B, leaves asleep at night.

a few observations were made on these three genera. The petioles on a young plant, about a foot in height, of Cytisus fragrans rose at night, on one occasion 23o and on another 33o. The three leaflets also bend upwards, and at the same time [page 345] approach each other, so that the base of the central leaflet overlaps the bases of the two lateral leaflets. They bend up so much that they press against the stem; and on looking down on one of these young plants from vertically above, the lower surfaces of the leaflets are visible; and thus their upper surfaces, in accordance with the general rule, are best protected from radiation. Whilst the leaves on these young plants were thus behaving, those on an old bush in full flower did not sleep at night.

Trigonella Cretica resembles a Melilotus in its sleep, which will be immediately described. According to M. Royer,* the leaves of Medicago maculata rise up at night, and "se renversent un peu de manire presenter obliquement au ciel leur face infrieure." A drawing is here given (Fig. 139) of the leaves of M. marina awake and asleep; and this would almost serve for Cytisus fragrans in the same two states.

Melilotus (Tribe 3).—The species in this genus sleep in a remarkable manner. The three leaflets of each leaf twist through an angle of 90o, so that their blades stand vertically at night with one lateral edge presented to the zenith (Fig. 140). We shall best understand the other and more complicated movements, if we imagine ourselves always to hold the leaf with the tip of the terminal leaflet pointed to the north. The leaflets in becoming vertical at night could of course twist so that their upper surfaces should face to either side; but the two lateral leaflets always twist so that this surface tends to face the north, but as they move at the same time towards the terminal leaflet, the upper surface of the one faces about N.N.W., and that of the other N.N.E. The terminal leaflet behaves differently, for it twists to either side, the upper surface facing sometimes east and sometimes west, but rather more commonly west than east. The terminal leaflet also moves in another and more remarkable manner, for whilst its blade is twisting and becoming vertical, the whole leaflet bends to one side, and invariably to the side towards which the upper surface is directed; so that if this surface faces the west the whole leaflet bends to the west, until it comes into contact with the upper and vertical surface of the western lateral leaflet. Thus the upper surface of the terminal and of one of the two lateral leaflets is well protected.

The fact of the terminal leaflet twisting indifferently to either

* 'Annales des Sc. Nat. Bot.' (5th series), ix. 1868, p. 368. [page 346]

side and afterwards bending to the same side, seemed to us so remarkable, that we endeavoured to discover the cause. We imagined that at the commencement of the movement it might be determined by one of the two halves of the leaflet being a little heavier than the other. Therefore bits of wood were gummed on one side of several leaflets, but this produced no effect; and they continued to twist in the same direction as

Fig. 140. Melilotus officinalis: A, leaf during the daytime. B, another leaf asleep. C, a leaf asleep as viewed from vertically above; but in this case the terminal leaflet did not happen to be in such close contact with the lateral one, as is usual.

they had previously done. In order to discover whether the same leaflet twisted permanently in the same direction, black threads were tied to 20 leaves, the terminal leaflets of which twisted so that their upper surfaces faced west, and 14 white threads to leaflets which twisted to the east. These were observed occasionally during 14 days, and they all continued, with a single exception, to twist and bend in the same direction; for [page 347] one leaflet, which had originally faced east, was observed after 9 days to face west. The seat of both the twisting and bending movement is in the pulvinus of the sub-petioles.

We believe that the leaflets, especially the two lateral ones, in performing the above described complicated movements generally bend a little downwards; but we are not sure of this, for, as far as the main petiole is concerned, its nocturnal movement is largely determined by the position which the leaf happens to occupy during the day. Thus one main petiole was observed to rise at night 59o, whilst three others rose only 7o and 9o. The petioles and sub-petioles are continually circumnutating during the whole 24 h., as we shall presently see.

The leaves of the following 15 species, M. officinalis, suaveolens, parviflora, alba, infesta, dentata, gracilis, sulcata, elegans, coerulea, petitpierreana, macrorrhiza, Italica, secundiflora, and Taurica, sleep in nearly the same manner as just described; but the bending to one side of the terminal leaflet is apt to fail unless the plants are growing vigorously. With M. petitpierreana and secundiflora the terminal leaflet was rarely seen to bend to one side. In young plants of M. Italica it bent in the usual manner, but with old plants in full flower, growing in the same pot and observed at the same hour, viz., 8.30 P.M., none of the terminal leaflets on several scores of leaves had bent to one side, though they stood vertically; nor had the two lateral leaflets, though standing vertically, moved towards the terminal one. At 10.30 P.M., and again one hour after midnight, the terminal leaflets had become very slightly bent to one side, and the lateral leaflets had moved a very little towards the terminal one, so that the position of the leaflets even at this late hour was far from the ordinary one. Again, with M. Taurica the terminal leaflets were never seen to bend towards either of the two lateral leaflets, though these, whilst becoming vertical, had bent towards the terminal one. The sub-petiole of the terminal leaflet in this species is of unusual length, and if the leaflet had bent to one side, its upper surface could have come into contact only with the apex of either lateral leaflet; and this, perhaps, is the meaning of the loss of the lateral movement.

The cotyledons do not sleep at night. the first leaf consists of a single orbicular leaflet, which twists at night so that the blade stands vertically. It is a remarkable fact that with M. Taurica, and in a somewhat less degree with M. macrorrhiza and petitpierreana, all the many small and young leaves produced during [page 348] the early spring from shoots on some cut-down plants in the greenhouse, slept in a totally different manner from the normal one; for the three leaflets, instead of twisting on their own axes so as to present their lateral edges to the zenith, turned upwards and stood vertically with their apices pointing to the zenith. They thus assumed nearly the same position as in the allied genus Trifolium; and on the same principle that embryological characters reveal the lines of descent in the animal kingdom, so the movements of the small leaves in the above three species of Melilotus, perhaps indicate that this genus is descended from a form which was closely allied to and slept like a Trifolium. Moreover, there is one species, M. messanensis, the leaves of which, on full-grown plants between 2 and 3 feet in height, sleep like the foregoing small leaves and like those of a Trifolium. We were so much surprised at this latter case that, until the flowers and fruit were examined, we thought that the seeds of some Trifolium had been sown by mistake instead of those of a Melilotus. It appears therefore probable that M. messanensis has either retained or recovered a primordial habit.

The circumnutation of a leaf of M. officinalis was traced, the stem being left free; and the apex of the terminal leaflet described three laterally extended ellipses, between 8 A.M. and 4 P.M.; after the latter hour the nocturnal twisting movement commenced. It was afterwards ascertained that the above movement was compounded of the circumnutation of the stem on a small scale, of the main petiole which moved most, and of the sub-petiole of the terminal leaflet. The main petiole of a leaf having been secured to a stick, close to the base of the sub-petiole of the terminal leaflet, the latter described two small ellipses between 10.30 A.M., and 2 P.M. At 7.15 P.M., after this same leaflet (as well as another) had twisted themselves into their vertical nocturnal position, they began to rise slowly, and continued to do so until 10.35 P.M., after which hour they were no longer observed.

As M. messanensis sleeps in an anomalous manner, unlike that of any other species in the genus, the circumnutation of a terminal leaflet, with the stem secured, was traced during two days. On each morning the leaflet fell, until about noon, and then began to rise very slowly; but on the first day the rising movement was interrupted between 1 and 3 P.M. by the formation of a laterally extended ellipse, and on the second day, at the same time, by two smaller ellipses. The rising movement then [page 349] recommenced, and became rapid late in the evening, when the leaflet was beginning to go to sleep. The awaking or sinking movement had already commenced by 6.45 A.M. on both mornings.

Trifolium (Tribe 3).—The nyctitropic movements of 11 species were observed, and were found to be closely similar. If we select a leaf of T. repens having an upright petiole, and with the three leaflets expanded horizontally, the two lateral leaflets will be seen in the evening to twist and approach each other, until their upper surfaces come into contact. At the same time they bend downwards in a plane at right angles to that of their former position, until their midribs form an angle of about 45o with the upper part of the petiole. This peculiar change of position requires a considerable amount of torsion in the pulvinus. The terminal leaflet merely rises up without any twist-

Fig. 141. Trifolium repens: A, leaf during the day; B, leaf asleep at night.

ing and bends over until it rests on and forms a roof over the edges of the now vertical and united lateral leaflets. Thus the terminal leaflet always passes through an angle of at least 90o, generally of 130o or 140o, and not rarely—as was often observed with T. subterraneum—of 180o. In this latter case the terminal leaflet stands at night horizontally (as in Fig. 141), with its lower surface fully exposed to the zenith. Besides the difference in the angles, at which the terminal leaflets stand at night in the individuals of the same species, the degree to which the lateral leaflets approach each other often likewise differs.

We have seen that the cotyledons of some species and not of others rise up vertically at night. The first true leaf is generally unifoliate and orbicular; it always rises, and either stands vertically at night or more commonly bends a little over so as to expose the lower surface obliquely to the zenith, in the same manner as does the terminal leaflet of the mature leaf. But it does not twist itself like the corresponding first simple leaf of Melilotus. [page 350] With T. Pannonicum the first true leaf was generally unifoliate, but sometimes trifoliate, or again partially lobed and in an intermediate condition.

Circumnutation.—Sachs described in 1863* the spontaneous up and down movements of the leaflets of T. incarnatum, when kept in darkness. Pfeffer made many observations on the similar movements in T. pratense.** He states that the terminal leaflet of this species, observed at different times, passed through angles of from 30o to 120o in the course of from 1 to 4 h. We observed the movements of T. subterraneum, resupinatum, and repens.

Trifolium subterraneum.—A petiole was secured close to the base of the three leaflets, and the movement of the terminal leaflet was traced during 26 h., as shown in the figure on the next page.

Between 6.45 A.M. and 6 P.M. the apex moved 3 times up and 3 times down, completing 3 ellipses in 11 h. 15 m. The ascending and descending lines stand nearer to one another than is usual with most plants, yet there was some lateral motion. At 6 P.M. the great nocturnal rise commenced, and on the next morning the sinking of the leaflet was continued until 8.30 A.M., after which hour it circumnutated in the manner just described. In the figure the great nocturnal rise and the morning fall are greatly abbreviated, from the want of space, and are merely represented by a short curved line. The leaflet stood horizontally when at a point a little beneath the middle of the diagram; so that during the daytime it oscillated almost equally above and beneath a horizontal position. At 8.30 A.M. it stood 48o beneath the horizon, and by 11.30 A.M. it had risen 50o above the horizon; so that it passed through 98o in 3 h. By the aid of the tracing we ascertained that the distance travelled in the 3 h. by the apex of this leaflet was 1.03 inch. If we look at the figure, and prolong upwards in our mind's eye the short curved broken line, which represents the nocturnal course, we see that the latter movement is merely an exaggeration or prolongation of one of the diurnal ellipses. The same leaflet had been observed on the previous day, and the course then pursued was almost identically the same as that here described.

* 'Flora,' 1863, p. 497.

** 'Die Period. Bewegungen,' 1875, pp. 35, 52. [page 351]

Fig. 142. Trifolium subterraneum: circumnutation and nyctitropic movement of terminal leaflet (.68 inch in length), traced from 6.45 A.M. July 4th to 9.15 A.M. 5th. Apex of leaf 3 7/8 inches from the vertical glass, and movement, as here shown, magnified 5 1/4 times, reduced to one-half of original scale. Plant illuminated from above; temp. 16o - 17o C.

Trifolium resupinatum.—A plant left entirely free was placed before a north-east window, in such a position that a terminal leaflet projected at right angles to the source of the light, the sky being uniformly clouded all day. The movements of this leaflet were traced during two days, and on both were closely similar. Those executed on the second day are shown in Fig. 143. The obliquity of the several lines is due partly to the manner in which the leaflet was viewed, and partly to its having moved a little towards the light. From 7.50 A.M. to 8.40 A.M. the leaflet fell, that is, the awakening movement was continued. It then rose and moved a little laterally towards the light. At 12.30 it retrograded, and at 2.30 resumed its original course, having thus completed a small ellipse during the middle of the day. In the evening it rose rapidly, and by 8 A.M. on the following morning had returned to exactly the same spot as on the previous morning. The line representing the nocturnal course ought to be extended much higher up, and is here abbreviated into a short, [page 352] curved, broken line. The terminal leaflet, therefore, of this species described during the daytime only a single additional ellipse, instead of two additional ones, as in the case of T. subterraneum. But we should remember that it was shown in the fourth chapter that the stem circumnutates, as no doubt does the main petiole and the sub-petioles; so that the movement represented in Fig. 143 is a compounded one. We tried to observe the movements of a leaf kept during the day in darkness, but it began to go to sleep after 2 h. 15 m., and this was well pronounced after 4 h. 30 m.

Fig 143. Trifolium resupinatum: circumnutation and nyctitropic movements of the terminal leaflet during 24 hours.

Trifolium repens.—A stem was secured close to the base of a moderately old leaf, and the movement of the terminal leaflet was observed during two days. This case is interesting solely from the simplicity of the movements, in contrast with those of the two preceding species. On the first day the leaflet fell between 8 A.M. and 3 P.M., and on the second between 7 A.M. and 1 P.M. On both days the descending course was somewhat zigzag, and this evidently represents the circumnutating movement of the two previous species during the middle of the day. After 1 P.M., Oct. 1st (Fig. 144), the leaflet began to rise, but the movement was slow on both days, both before and after this hour, until 4 P.M. The rapid evening and nocturnal rise then commenced. Thus in this species the course during 24 h. consists of a single great ellipse; in T. resupinatum of two ellipses, one of which includes the nocturnal movement and is much elongated; and in T. subterraneum of three ellipses, of which the nocturnal one is likewise of great length.

Securigera coronilla (Tribe 4).—The leaflets, which stand opposite one another and are numerous, rise up at night, come into close contact, and bend backwards at a moderate angle towards the base of the petiole. [page 353]

Fig. 144. Trifolium repens: circumnutation and nyctitropic movements of a nearly full-grown terminal leaflet, traced on a vertical glass from 7 A.M. Sept. 30th to 8 A.M. Oct. 1st. Nocturnal course, represented by curved broken line, much abbreviated.

Lotus (Tribe 4).—The nyctitropic movements of 10 species in this genus were observed, and found to be the same. The main petiole rises a little at night, and the three leaflets rise till they become vertical, and at the same time approach each other. This was conspicuous with L. Jacoboeus, in which the leaflets are almost linear. In most of the species the leaflets rise so much as to press against the stem, and not rarely they become inclined a little inwards with their lower surfaces exposed obliquely to the zenith. This was clearly the case with L. major, as its petioles are unusually long, and the leaflets are thus enabled to bend further inwards. The young leaves on the summits of the stems close up at night so much, as often to resemble large buds. The stipule-like leaflets, which are often of large size, rise up like the other leaflets, and press against the stem (Fig. 145). All the leaflets of L. Gebelii, and probably of the other species, are provided at their bases with distinct pulvini, of a yellowish colour, and formed of very small cells. The circumnutation of a terminal leaflet of L. peregrinus (with the stem secured) was traced during two days, but the movement was so simple that it is not worth while to give the diagram. The leaflet fell slowly from the early morning till about 1 P.M. It then rose gradually at first, but rapidly late in the evening. It occasionally stood still for about 20 m. during the day, and sometimes zigzagged a little. The movement of one of the basal, stipule-like leaflets was likewise traced in the same manner and at the same time, and its course was closely similar to that of the terminal leaflet.

In Tribe 5 of Bentham and Hooker, the sleep-movements of species in 12 genera have been observed by ourselves and [page 354] others, but only in Robinia with any care. Psoralea acaulis raises its three leaflets at night; whilst Amorpha fruticosa,* Dalea alopecuroides, and Indigofera tinctoria depress them. Ducharte** states that Tephrosia caribaea is the sole example of "folioles couches le long du ptiole et vers la base;" but a

Fig. 145. Lotus Creticus: A, stem with leaves awake during the day; B, with leaves asleep at night. SS, stipule-like leaflets.

similar movement occurs, as we have already seen, and shall again see in other cases. Wistaria Sinensis, according to Royer,*** "abaisse les folioles qui par une disposition bizarre sont inclines dans la mme feuille, les suprieures vers le

* Ducharte, 'Elments de Botanique', 1867, p. 349.

** Ibid., p. 347.

*** 'Ann. des Sciences Nats. Bot.' (5th series), ix. 1868. [page 355]

sommet, les infrieures vers la base du petiole commun;" but the leaflets on a young plant observed by us in the greenhouse merely sank vertically downwards at night. The leaflets are raised in Sphaerophysa salsola, Colutea arborea, and Astragalus uliginosus, but are depressed, according to Linnaeus, in Glycyrrhiza. The leaflets of Robinia pseudo-acacia likewise sink vertically down at night, but the petioles rise a little, viz., in one case 3o, and in another 4o. The circumnutating movements of a terminal leaflet on a rather old leaf were traced during two days, and were simple. The leaflet fell slowly, in a slightly zigzag line, from 8 A.M. to 5 P.M., and then more rapidly; by 7 A.M. on the following morning it had risen to its diurnal position. There was only one peculiarity in the movement, namely, that on both days there was a distinct though small oscillation up and down between 8.30 and 10 A.M., and this would probably have been more strongly pronounced if the leaf had been younger.

Coronilla rosea (Tribe 6).—the leaves bear 9 or 10 pairs of opposite leaflets, which during the day stand horizontally, with

Fig. 146. Coronilla rosea: leaf asleep.

their midribs at right angles to the petiole. At night they rise up so that the opposite leaflets come nearly into contact, and those on the younger leaves into close contact. At the same time they bend back towards the base of the petiole, until their midribs form with it angles of from 40o to 50o in a vertical plane, as here figured (Fig. 146). The leaflets, however, sometimes bend so much back that their midribs become parallel to and lie on the petiole. They thus occupy a reversed position to what they do in several Leguminosae, for instance, in Mimosa [page 356] pudica; but, from standing further apart, they do not overlap one another nearly so much as in this latter plant. The main petiole is curved slightly downwards during the day, but straightens itself at night. In three cases it rose from 3o above the horizon at noon, to 9o at 10 P.M.; from 11o to 33o; and from 5o to 33o—the amount of angular movement in this latter case amounting to 28o. In several other species of Coronilla the leaflets showed only feeble movements of a similar kind.

Hedysarum coronarium (Tribe 6).—The small lateral leaflets on plants growing out of doors rose up vertically at night, but the large terminal one became only moderately inclined. The petioles apparently did not rise at all.

Smithia Pfundii (Tribe 6).—The leaflets rise up vertically, and the main petiole also rises considerably.

Arachis hypogoea (Tribe 6).—The shape of a leaf, with its two pairs of leaflets, is shown at A (Fig. 147); and a leaf asleep,

Fig. 147. Arachis hypogoea: A, leaf during the day, seen from vertically above; B, leaf asleep, seen laterally, copied from a photograph. Figures much reduced.

traced from a photograph (made by the aid of aluminium light), is given at B. The two terminal leaflets twist round at night until their blades stand vertically, and approach each other until they meet, at the same time moving a little upwards and backwards. The two lateral leaflets meet each other in this same manner, but move to a greater extent forwards, that is, in a contrary direction to the two terminal leaflets, which they partially embrace. Thus all four leaflets form together a single packet, with their edges directed to the zenith, and with their lower surfaces turned outwards. On a plant which was not growing vigorously the closed leaflets seemed too heavy for the [page 357] petioles to support them in a vertical position, so that each night the main petiole became twisted, and all the packets were extended horizontally, with the lower surfaces of the leaflets on one side directed to the zenith in a most anomalous manner. This fact is mentioned solely as a caution, as it surprised us greatly, until we discovered that it was an anomaly. The petioles are inclined upwards during the day, but sink at night, so as to stand at about right angles with the stem. The amount of sinking was measured only on one occasion, and found to be 39o. A petiole was secured to a stick at the base of the two terminal leaflets, and the circumnutating movement of one of these leaflets was traced from 6.40 A.M. to 10.40 P.M., the plant being illuminated from above. The temperature was 17o - 17 1/2o C., and therefore rather too low. During the 16 h. the leaflet moved thrice up and thrice down, and as the ascending and descending lines did not coincide, three ellipses were formed.

Fig. 148. Desmodium gyrans: leaf seen from above, reduced to one-half natural size. The minute stipules unusually large.

Desmodium gyrans (Tribe 6).—A large and full-grown leaf of this plant, so famous for the spontaneous movements of the two little lateral leaflets, is here represented (Fig. 148). The large terminal leaflet sleeps by sinking vertically down, whilst the petiole rises up. The cotyledons do not sleep, but the first-formed leaf sleeps equally well as the older ones. The appearance presented by a sleeping branch and one in the day-time, copied from two photographs, are shown at A and B (Fig. 149), and we see how at night the leaves are crowded together, as if for mutual protection, by the rising of the petioles. The petioles of the younger leaves near the summits of the shoots rise up at night, so as to stand vertical and parallel to the stem; whilst those on the sides were found in four cases to have risen respectively 46 1/2o, 36o, 20o, and 19.5o above the inclined positions which they had occupied during the day. For instance, in the first of these four cases the petiole stood in the day at 23o, and at night at 69 1/2o above the horizon. In the evening the rising of the petioles is almost completed before the leaflets sink perpendicularly downwards. [page 358]

Circumnutation.—The circumnutating movements of four young shoots were observed during 5 h. 15 m.; and in this time each completed an oval figure of small size. The main petiole also circumnutates rapidly, for in the course of 31 m. (temp. 91o F.) it changed its course by as much as a rectangle six times, describing a figure which apparently represented two ellipses.

Fig. 149. Desmodium gyrans: A, stem during the day; B, stem with leaves asleep. Figures reduced.

The movement of the terminal leaflet by means of its sub-petiole or pulvinus is quite as rapid, or even more so, than that of the main petiole, and has much greater amplitude. Pfeffer has seen* these leaflets move through an angle of 8o in the course of from 10 to 30 seconds.

A fine, nearly full-grown leaf on a young plant, 8 inches in height, with the stem secured to a stick at the base of the leaf, was observed from 8.30 A.M. June 22nd to 8 A.M. June 24th.

* 'Die Period. Beweg.,' p. 35. [page 359]

In the diagram given on the next page (Fig. 150), the two curved broken lines at the base, which represent the nocturnal courses, ought to be prolonged far downwards. On the first day the leaflet moved thrice down and thrice up, and to a considerable distance laterally; the course was also remarkably crooked. The dots were generally made every hour; if they had been made every few minutes all the lines would have been zigzag to an extraordinary degree, with here and there a loop formed. We may infer that this would have been the case, because five dots were made in the course of 31 m. (between 12.34 and 1.5 P.M.), and we see in the upper part of the diagram how crooked the course here is; if only the first and last dots had been joined we should have had a straight line. Exactly the same fact may be seen in the lines representing the course between 2.24 P.M. and 3 P.M., when six intermediate dots were made; and again at 4.46 and 4.50. But the result was widely different after 6 P.M.,—that is, after the great nocturnal descent had commenced; for though nine dots were then made in the course of 32 m., when these were joined (see Figure) the line thus formed was almost straight. The leaflets, therefore, begin to descend in the afternoon by zigzag lines, but as soon as the descent becomes rapid their whole energy is expended in thus moving, and their course becomes rectilinear. After the leaflets are completely asleep they move very little or not at all.

Had the above plant been subjected to a higher temperature than 67o - 70o F., the movements of the terminal leaflet would probably have been even more rapid and wider in extent than those shown in the diagram; for a plant was kept for some time in the hot-house at from 92o - 93o F., and in the course of 35 m. the apex of a leaflet twice descended and once ascended, travelling over a space of 1.2 inch in a vertical direction and of .82 inch in a horizontal direction. Whilst thus moving the leaflet also rotated on its own axis (and this was a point to which no attention had been before paid), for the plane of the blade differed by 41o after an interval of only a few minutes. Occasionally the leaflet stood still for a short time. There was no jerking movement, which is so characteristic of the little lateral leaflets. A sudden and considerable fall of temperature causes the terminal leaflet to sink downwards; thus a cut-off leaf was immersed in water at 95o F., which was slowly raised to 103o F., and afterwards allowed to sink to 70o F., and the sub-petiole of the terminal leaflet then curved downwards. The water was afterwards [page 360]

Fig. 150. Desmodium gyrans: circumnutation and nyctitropic movement of leaf (3 3/4 inches in length, petiole included) during 48 h. Filament affixed to midrib of terminal leaflet; its apex 6 inches from the vertical glass. Diagram reduced to one-third of original scale. Plant illuminated from above. Temp. 19o - 20o C. [page 361]

raised to 120o F., and the sub-petiole straightened itself. Similar experiments with leaves in water were twice repeated, with nearly the same result. It should be added, that water raised to even 122o F. does not soon kill a leaf. A plant was placed in darkness at 8.37 A.M., and at 2 P.M. (i.e. after 5 h. 23 m.), though the leaflets had sunk considerably, they had by no means acquired their nocturnal vertically dependent position. Pfeffer, on the other hand, says* that this occurred with him in from 3/4 h. to 2 h.; perhaps the difference in our results may be due to the plant on which we experimented being a very young and vigorous seedling.

The Movements of the little Lateral Leaflets .—These have been so often described, that we will endeavour to be as brief as possible in giving a few new facts and conclusions. The leaflets sometimes quickly change their position by as much as nearly 180o; and their sub-petioles can then be seen to become greatly curved. They rotate on their own axes, so that their upper surfaces are directed to all points of the compass. The figure described by the apex is an irregular oval or ellipse. They sometimes remain stationary for a period. In these several respects there is no difference, except in rapidity and extent, between their movements and the lesser ones performed by the large terminal leaflet whilst making its great oscillations. The movements of the little leaflets are much influenced, as is well known, by temperature. This was clearly shown by immersing leaves with motionless leaflets in cold water, which was slowly raised to 103o F., and the leaflets then moved quickly, describing about a dozen little irregular circles in 40 m. By this time the water had become much cooler, and the movements became slower or almost ceased; it was then raised to 100o F., and the leaflets again began to move quickly. On another occasion a tuft of fine leaves was immersed in water at 53o F., and the leaflets were of course motionless. The water was raised to 99o, and the leaflets soon began to move; it was raised to 105o, and the movements became much more rapid; each little circle or oval being completed in from 1 m. 30 s. to 1 m. 45 s. There was, however, no jerking, and this fact may perhaps be attributed to the resistance of the water.

Sachs states that the leaflets do not move until the surrounding air is as high as 71o - 72o F., and this agrees with our

* 'Die Period. Beweg.,' p. 39. [page 362]

experience on full-grown, or nearly full-grown, plants. But the leaflets of young seedlings exhibit a jerking movement at much lower temperatures. A seedling was kept (April 16th) in a room for half the day where the temperature was steady at 64o F., and the one leaflet which it bore was continually jerking, but not so rapidly as in the hot-house. The pot was taken in the evening into a bed-room where the temperature remained at 62o during nearly the whole night; at 10 and 11 P.M. and at 1 A.M. the leaflet was still jerking rapidly; at 3.30 A.M. it was not seen to jerk, but was observed during only a short time. It was, however, now inclined at a much lower angle than that occupied at 1 A.M. At 6.30 A.M. (temp. 61o F.) its inclination was still less than before, and again less at 6.45 A.M.; by 7.40 A.M. it had risen, and at 8.30 A.M. was again seen to jerk. This leaflet, therefore, was moving during the whole night, and the movement was by jerks up to 1 A.M. (and possibly later) and again at 8.30 A.M., though the temperature was only 61o to 62o F. We must therefore conclude that the lateral leaflets produced by young plants differ somewhat in constitution from those on older plants.

In the large genus Desmodium by far the greater number of the species are trifoliate; but some are unifoliate, and even the same plant may bear uni- and trifoliate leaves. In most of the species the lateral leaflets are only a little smaller than the terminal one. Therefore the lateral leaflets of D. gyrans (see Fig. 148) must be considered as almost rudimentary. They are also rudimentary in function, if this expression may be used; for they certainly do not sleep like the full-sized terminal leaflets. It is, however, possible that the sinking down of the leaflets between 1 A.M. and 6.45 A.M., as above described, may represent sleep. It is well known that the leaflets go on jerking during the early part of the night; but my gardener observed (Oct. 13th) a plant in the hot-house between 5 and 5.30 A.M., the temperature having been kept up to 82o F., and found that all the leaflets were inclined, but he saw no jerking movement until 6.55 A.M., by which time the terminal leaflet had risen and was awake. Two days afterwards (Oct. 15th) the same plant was observed by him at 4.47 A.M. (temp. 77o F.), and he found that the large terminal leaflets were awake, though not quite horizontal; and the only cause which we could assign for this anomalous wakefulness was that the plant had been kept for experimental purposes during [page 363] the previous day at an unusually high temperature; the little lateral leaflets were also jerking at this hour, but whether there was any connection between this latter fact and the sub-horizontal position of the terminal leaflets we do not know. Anyhow, it is certain that the lateral leaflets do not sleep like the terminal leaflets; and in so far they may be said to be in a functionally rudimentary condition. They are in a similar condition in relation to irritability; for if a plant be shaken or syringed, the terminal leaflets sink down to about 45o beneath the horizon; but we could never detect any effect thus produced on the lateral leaflets; yet we are not prepared to assert positively that rubbing or pricking the pulvinus produces no effect.

As in the case of most rudimentary organs, the leaflets are variable in size; they often depart from their normal position and do not stand opposite one another; and one of the two is frequently absent. This absence appeared in some, but not in all the cases, to be due to the leaflet having become completely confluent with the main petiole, as might be inferred from the presence of a slight ridge along its upper margin, and from the course of the vessels. In one instance there was a vestige of the leaflet, in the shape of a minute point, at the further end of the ridge. The frequent, sudden and complete disappearance of one or both of the rudimentary leaflets is a rather singular fact; but it is a much more surprising one that the leaves which are first developed on seedling plants are not provided with them. Thus, on one seedling the seventh leaf above the cotyledons was the first which bore any lateral leaflets, and then only a single one. On another seedling, the eleventh leaf first bore a leaflet; of the nine succeeding leaves five bore a single lateral leaflet, and four bore none at all; at last a leaf, the twenty-first above the cotyledons, was provided with two rudimentary lateral leaflets. From a widespread analogy in the animal kingdom, it might have been expected that these rudimentary leaflets would have been better developed and more regularly present on very young than on older plants. But bearing in mind, firstly, that long-lost characters sometimes reappear late in life, and secondly, that the species of Desmodium are generally trifoliate, but that some are unifoliate, the suspicion arises that D. gyrans is descended from a unifoliate species, and that this was descended from a trifoliate one; for in this case both the absence of the little lateral leaflets on very young seedlings, and their sub- [page 364] sequent appearance, may be attributed to reversion to more or less distant progenitors.*

No one supposes that the rapid movements of the lateral leaflets of 'D. gyrans' are of any use to the plant; and why they should behave in this manner is quite unknown. We imagined that their power of movement might stand in some relation with their rudimentary condition, and therefore observed the almost rudimentary leaflets of Mimosa albida vel sensitiva (of which a drawing will hereafter be given, Fig. 159); but they exhibited no extraordinary movements, and at night they went to sleep like the full-sized leaflets. There is, however, this remarkable difference in the two cases; in Desmodium the pulvinus of the rudimentary leaflets has not been reduced in length, in correspondence with the reduction of the blade, to the same extent as has occurred in the Mimosa; and it is on the length and degree of curvature of the pulvinus that the amount of movement of the blade depends. Thus the average length of the pulvinus in the large terminal leaflets of Desmodium is 3 mm., whilst that of the rudimentary leaflets is 2.86 mm.; so that they differ only a little in length. But in diameter they differ much, that of the pulvinus of the little leaflets being only 0.3 mm. to 0.4 mm.; whilst that of the terminal leaflets is 1.33 mm. If we now turn to the Mimosa, we find that the average length of the pulvinus of the almost rudimentary leaflets is only 0.466 mm., or rather more than a quarter of the length of the pulvinus of the full-sized leaflets, namely, 1.66 mm. In this small reduction in length of the pulvinus of the rudimentary leaflets of Desmodium, we apparently have the proximate cause of their great and rapid circumnutating movement, in contrast with that of the almost rudimentary leaflets of the Mimosa. The small size and weight of the blade, and the little resistance opposed by the air to its movement, no doubt also come into play; for we have seen that these leaflets if immersed in water, when the resistance would be much greater, were prevented from jerking forwards. Why, during the reduction of the lateral leaflets of Desmodium, or during their reappearance—if they owe their origin to reversion—the pulvinus should have been so much less affected than the blade, whilst with the

* Desmodium vespertilionis is closely allied to D. gyrans, and it seems only occasionally to bear rudimentary lateral leaflets. Duchartre, 'Elments de Botanique,' 1867, p. 353. [page 365]

Mimosa the pulvinus has been greatly reduced, we do not know. Nevertheless, it deserves notice that the reduction of the leaflets in these two genera has apparently been effected by a different process and for a different end; for with the Mimosa the reduction of the inner and basal leaflets was necessary from the want of space; but no such necessity exists with Desmodium, and the reduction of its lateral leaflets seems to have been due to the principle of compensation, in consequence of the great size of the terminal leaflet. Uraria (Tribe 6) and Centrosema (Tribe 8).—The leaflets of Uraria lagopus and the leaves of a Centrosema from Brazil both sink vertically down at night. In the latter plant the petiole at the same time rose 16 1/2o.

Amphicarpoea monoica (Tribe 8).—The leaflets sink down vertically at night, and the petioles likewise fall considerably.

Fig. 151. Amphicarpoea monoica: circumnutation and nyctitropic movement of leaf during 48 h.; its apex 9 inches from the vertical glass. Figure reduced to one-third of original scale. Plant illuminated from above; temp 17 1/2o - 18 1/2o C.

A petiole, which was carefully observed, stood during the day 25o above the horizon and at night 32o below it; it therefore fell 57o. A filament was fixed transversely across the terminal leaflet of a fine young leaf (2 1/4 inches in length including the [page 366] petiole), and the movement of the whole leaf was traced on a vertical glass. This was a bad plan in some respects, because the rotation of the leaflet, independently of its rising or falling, raised and depressed the filament; but it was the best plan for our special purpose of observing whether the leaf moved much after it had gone to sleep. The plant had twined closely round a thin stick, so that the circumnutation of the stem was prevented. The movement of the leaf was traced during 48 h., from 9 A.M. July 10th to 9 A.M. July 12th. In the figure given (Fig. 151) we see how complicated its course was on both days: during the second day it changed its course greatly 13 times. The leaflets began to go to sleep a little after 6 P.M., and by 7.15 P.M. hung vertically down and were completely asleep; but on both nights they continued to move from 7.15 P.M. to 10.40 and 10.50 P.M., quite as much as during the day; and this was the point which we wished to ascertain. We see in the figure that the great sinking movement late in the evening does not differ essentially from the circumnutation during the day.

Glycine hispida (Tribe 8).—The three leaflets sink vertically down at night.

Erythrina (Tribe 8).—Five species were observed, and the leaflets of all sank vertically down at night; with E. caffra and with a second unnamed species, the petioles at the same time rose slightly. The movements of the terminal leaflet of E. crista-galli (with the main petiole secured to a stick) were traced from 6.40 A.M. June 8th, to 8 A.M. on the 10th. In order to observe the nyctitropic movements of this plant, it is necessary that it should have grown in a warm greenhouse, for out of doors in our climate it does not sleep. We see in the tracing (Fig. 152) that the leaflet oscillated twice up and down between early morning and noon; it then fell greatly, afterwards rising till 3 P.M. At this latter hour the great nocturnal fall commenced. On the second day (of which the tracing is not given) there was exactly the same double oscillation before noon, but only a very small one in the afternoon. On the third morning the leaflet moved laterally, which was due to its beginning to assume an oblique position, as seems invariably to occur with the leaflets of this species as they grow old. On both nights after the leaflets were asleep and hung vertically down, they continued to move a little both up and down, and from side to side.

Erythrina caffra.—A filament was fixed transversely across [page 367]

a terminal leaflet, as we wished to observe its movements when asleep. The plant was placed in the morning of June 10th under a skylight, where the light was not bright; and we do not know whether it was owing to this cause or to the plant having been disturbed, but the leaflet hung vertically down all day; nevertheless it circumnutated in this position, describing a figure which represented two irregular ellipses. On the next day it circumnutated in a greater degree, describing four irregular ellipses, and by 3 P.M. had risen into a horizontal position. By 7.15 P.M. it was asleep and vertically dependent, but continued to circumnutate as long as observed, until 11 P.M.

Fig. 152. Erythrina crista-galli: circumnutation and nyctitropic movement of terminal leaflet, 3 3/4 inches in length, traced during 25 h.; apex of leaf 3 inches from the vertical glass. Figure reduced to one-half of original scale. Plant illuminated from above; temp. 17 1/2o - 18 1/2o C.

Erythrina corallodendron.—The movements of a terminal leaflet were traced. During the second day it oscillated four times up and four times down between 8 A.M. and 4 P.M., after which hour the great nocturnal fall commenced. On the third day the movement was equally great in amplitude, but was remarkably simple, for the leaflet rose in an almost perfectly straight line from 6.50 A.M. to 3 P.M., and then sank down in an equally straight line until vertically dependent and asleep. [page 368]

Apios tuberosa (Tribe 8).—The leaflets sink vertically down at night.

Phaseolus vulgaris (Tribe 8).—The leaflets likewise sink vertically down at night. In the greenhouse the petiole of a young leaf rose 16o, and that of an older leaf 10o at night. With plants growing out of doors the leaflets apparently do not sleep until somewhat late in the season, for on the nights of July 11th and 12th none of them were asleep; whereas on the night of August 15th the same plants had most of their leaflets vertically dependent and asleep. With Ph. caracalla and Hernandesii, the primary unifoliate leaves and the leaflets of the secondary trifoliate leaves sink vertically down at night. This holds good with the secondary trifoliate leaves of Ph. Roxburghii, but it is remarkable that the primary unifoliate leaves which are much elongated, rise at night from about 20o to about 60o above the horizon. With older seedlings, however, having the secondary leaves just developed, the primary leaves stand in the middle of the day horizontally, or are deflected a little beneath the horizon. In one such case the primary leaves rose from 26o beneath the horizon at noon, to 20o above it at 10 P.M.; whilst at this same hour the leaflets of the secondary leaves were vertically dependent. Here, then, we have the extraordinary case of the primary and secondary leaves on the same plant moving at the same time in opposite directions.

We have now seen that the leaflets in the six genera of Phaseoleae observed by us (with the exception of the primary leaves of Phaseolus Roxburghii) all sleep in the same manner, namely, by sinking vertically down. The movements of the petioles were observed in only three of these genera. They rose in Centrosema and Phaseolus, and sunk in Amphicarpaea.

Sophora chrysophylla (Tribe 10).—The leaflets rise at night, and are at the same time directed towards the apex of the leaf, as in Mimosa pudica.

Caesalpinia, Hoematoxylon, Gleditschia, Poinciana.—The leaflets of two species of Caesalpinia (Tribe 13) rose at night. With Haematoxylon Campechianum (Tribe 13) the leaflets move forwards at night, so that their midribs stand parallel to the petiole, and their now vertical lower surfaces are turned outwards (Fig. 153). The petiole sinks a little. In Gleditschia, if we understand correctly Duchartre's description, and in Poin- [page 369] ciana Gilliesii (both belonging to Tribe 13), the leaves behave in the same manner.

Fig. 153. Haematoxylon Campechianum: A, branch during daytime; B, branch with leaves asleep, reduced to two-thirds of natural scale.

Cassia (Tribe 14).—The nyctitropic movements of the leaves in many species in this genus are closely alike, and are highly complex. They were first briefly described by Linnaeus, and since by Duchartre. Our observations were made chiefly on C. floribunda* and corymbosa, but several other species were casually observed. The horizontally extended leaflets sink down vertically at night; but not simply, as in so many other genera, for each leaflet rotates on its own axis, so that its lower surface faces outwards. The upper surfaces of the opposite leaflets are thus brought into contact with one another beneath the petiole, and are well protected (Fig. 154). The rotation and other movements are effected by means of a well-developed pulvinus at the base of each leaflet, as could be plainly seen when a straight narrow black line had been painted along it during the day. The two terminal leaflets in the daytime include rather less than a right angle; but their divergence increases greatly whilst they

* I am informed by Mr. Dyer that Mr. Bentham believes that C. floribunda (a common greenhouse bush) is a hybrid raised in France, and that it comes very near to C. laevigata. It is no doubt the same as the form described by Lindley ('Bot. Reg.,' Tab. 1422) as C. Herbertiana. [page 370]

sink downwards and rotate, so that they stand laterally at night, as may be seen in the figure. Moreover, they move somewhat backwards, so as to point towards the base of the petiole.

Fig. 154. Cassia corymbosa: A, plant during day; B, same plant at night. Both figures copied from photographs.

In one instance we found that the midrib of a terminal leaflet formed at night an angle of 36o, with a line dropped [page 371] perpendicularly from the end of the petiole. The second pair of leaflets likewise moves a little backwards, but less than the terminal pair; and the third pair moves vertically downwards, or even a little forwards. Thus all the leaflets, in those species which bear only 3 or 4 pairs, tend to form a single packet, with their upper surfaces in contact, and their lower surfaces turned outwards. Lastly, the main petiole rises at night, but with leaves of different ages to very different degrees, namely some rose through an angle of only 12o, and others as much as 41o.

Cassia calliantha.—The leaves bear a large number of leaflets, which move at night in nearly the same manner as just described; but the petioles apparently do not rise, and one which was carefully observed certainly fell 3o. Cassia pubescens.—The chief difference in the nyctitropic

Fig. 155. Cassia pubescens: A, upper part of plant during the day; B, same plant at night. Figures reduced from photographs.

movements of this species, compared with those of the former species, consists in the leaflets not rotating nearly so much; [page 372] therefore their lower surfaces face but little outwards at night. The petioles, which during the day are inclined only a little above the horizon, rise at night in a remarkable manner, and stand nearly or quite vertically. This, together with the dependent position of the leaflets, makes the whole plant wonderfully compact at night. In the two foregoing figures, copied from photographs, the same plant is represented awake and asleep (Fig. 155), and we see how different is its appearance.

Cassia mimosoides.—At night the numerous leaflets on each leaf rotate on their axes, and their tips move towards the apex of the leaf; they thus become imbricated with their lower surfaces directed upwards, and with their midribs almost parallel to the petiole. Consequently, this species differs from all the others seen by us, with the exception of the following one, in the leaflets not sinking down at night. A petiole, the movement of which was measured, rose 8o at night.

Cassia Barclayana.—The leaflets of this Australian species are numerous, very narrow, and almost linear. At night they rise up a little, and also move towards the apex of the leaf. For instance, two opposite leaflets which diverged from one another during the day at an angle of 104o, diverted at night only 72o; so that each had risen 16o above its diurnal position. The petiole of a young leaf rose at night 34o, and that of an older leaf 19o. Owing to the slight movement of the leaflets and the considerable movement of the petiole, the bush presents a different appearance at night to what it does by day; yet the leaves can hardly be said to sleep.

The circumnutating movements of the leaves of C. floribunda, calliantha, and pubescens were observed, each during three or four days; they were essentially alike, those of the last-named species being the simplest. The petiole of C. floribunda was secured to a stick at the base of the two terminal leaflets, and a filament was fixed along the midrib of one of them. Its movements were traced from 1 P.M. on August 13th to 8.30 A.M. 17th; but those during the last 2 h. are alone given in Fig. 156. From 8 A.M. on each day (by which hour the leaf had assumed its diurnal position) to 2 or 3 P.M., it either zigzagged or circumnutated over nearly the same small space; at between 2 and 3 P.M. the great evening fall commenced. The lines representing this fall and the early morning rise are oblique, owing to the peculiar manner in which the leaflets sleep, as already described. After the leaflet was asleep at 6 P.M., and whilst the glass filament hung [page 373] perpendicularly down, the movement of its apex was traced until 10.30 P.M.; and during this whole time it swayed from side to side, completing more than one ellipse.

Fig 156. Cassia floribunda: circumnutation and nyctitropic movement of a terminal leaflet (1 5/6 inch in length) traced from 8.30 A.M. to same hour on following morning. Apex of leaflet 5 inches from the vertical glass. Main petiole 3 3/4 inches long. Temp. 16o - 17 1/2o C. Figure reduced to one-half of the original scale.

Bauhinia (Tribe 15).—The nyctitropic movements of four species were alike, and were highly peculiar. A plant raised from seed sent us from South Brazil by Fritz Mller, was more especially observed. The leaves are large and deeply notched at their ends. At night the two halves rise up and close completely together, like the opposite leaflets of many Leguminosae. With very young plants the petioles rise considerably at the same time; one, which was inclined at noon 45o above the horizon, at night stood at 75o; it thus rose 30o; another rose 34o. Whilst the two halves of the leaf are closing, the midrib at first sinks vertically downwards and afterwards bends backwards, so as to pass close along one side of its own upwardly inclined petiole; the midrib being thus directed towards the stem or axis of the plant. The angle which the midrib formed with the horizon was measured in one case at different hours: at noon it stood horizontally; late in the evening it depended vertically; then rose to the opposite side, and at 10.15 P.M. stood at only 27o beneath the horizon, being directed towards the stem. It had thus travelled through 153o. [page 374] Owing to this movement—to the leaves being folded—and to the petioles rising, the whole plant is as much more compact at night than during the day, as a fastigiate Lombardy poplar is compared with any other species of poplar. It is remarkable that when our plants had grown a little older, viz., to a height of 2 or 3 feet, the petioles did not rise at night, and the midribs of the folded leaves were no longer bent back along one side of the petiole. We have noticed in some other genera that the petioles of very young plants rise much more at night than do those of older plants.

Tamarindus Indica (Tribe 16).—The leaflets approach or meet each other at night, and are all directed towards the apex of the leaf. They thus become imbricated with their midribs parallel to the petiole. The movement is closely similar to that of Haematoxylon (see Fig. 153), but more striking from the greater number of the leaflets.

Adenanthera, Prosopis, and Neptunia (Tribe 20).—With Adenanthera pavonia the leaflets turn edgeways and sink at night. In Prosopis they turn upwards. With Neptunia oleracea the leaflets on the opposite sides of the same pinna come into contact at night and are directed forwards. The pinnae themselves move downwards, and at the same time backwards or towards the stem of the plant. The main petiole rises.

Mimosa pudica (Tribe 20).—This plant has been the subject of innumerable observations; but there are some points in relation to our subject which have not been sufficiently attended to. At night, as is well known, the opposite leaflets come into contact and point towards the apex of the leaf; they thus become neatly imbricated with their upper surfaces protected. The four pinnae also approach each other closely, and the whole leaf is thus rendered very compact. The main petiole sinks downwards during the day till late in the evening, and rises until very early in the morning. The stem is continually circumnutating at a rapid rate, though not to a wide extent. Some very young plants, kept in darkness, were observed during two days, and although subjected to a rather low temperature of 57o - 59o F., the stem of one described four small ellipses in the course of 12 h. We shall immediately see that the main petiole is likewise continually circumnutating, as is each separate pinna and each separate leaflet. Therefore, if the movement of the apex of any one leaflet were to be traced, the course described would be compounded of the movements of four separate parts. [page 375] A filament had been fixed on the previous evening, longitudinally to the main petiole of a nearly full-grown, highly-sensitive leaf (four inches in length), the stem having been secured to a stick at its base; and a tracing was made on a vertical glass in the hot-house under a high temperature. In the figure given (Fig. 157), the first dot was made at 8.30 A.M. August 2nd, and the last at 7 P.M. on the 3rd. During 12 h. on the first day the petiole moved thrice downwards and twice upwards. Within the same length of time on the second day, it moved five times downwards and four times upwards. As the ascending and descending lines do not coincide, the petiole manifestly circumnutates; the great evening fall and nocturnal rise being an exaggeration of one of the circumnutations. It should, however, be observed that the petiole fell much lower down in the evenings than could be seen on the vertical glass or is represented in the diagram. After 7 P.M. on the 3rd (when the last dot in Fig. 157 was made) the pot was carried into a bed-room, and the petiole was found at 12.50 A.M. (i.e. after midnight) standing almost upright, and much more highly inclined than it was at 10.40 P.M. When observed again at 4 A.M. it had begun to fall, and continued falling till 6.15 A.M., after which hour it zigzagged and again circumnutated. Similar observations were made on another petiole, with nearly the same result.

Fig. 157 Mimosa pudica: circumnutation and nyctitropic movement of main petiole, traced during 34 h. 30 m.

On two other occasions the movement of the main petiole [page 376] was observed every two or three minutes, the plants being kept at a rather high temperature, viz., on the first occasion at 77o - 81o F., and the filament then described 2 ellipses in 69 m. On the second occasion, when the temperature was 81o - 86o F., it made rather more than 3 ellipses in 67 m. therefore, Fig. 157, though now sufficiently complex, would have been incomparably more so, if dots had been made on the glass every 2 or 3 minutes, instead of every hour or half-hour. Although the main petiole is continually and rapidly describing small ellipses during the day, yet after the great nocturnal rising movement has commenced, if dots are made every 2 or 3 minutes, as was done for an hour between 9.30 and 10.30 P.M. (temp. 84o F.), and the dots are then joined, an almost absolutely straight line is the result.

To show that the movement of the petiole is in all probability due to the varying turgescence of the pulvinus, and not to growth (in accordance with the conclusions of Pfeffer), a very old leaf, with some of its leaflets yellowish and hardly at all sensitive, was selected for observation, and the plant was kept at the highly favourable temp. of 80o F. The petiole fell from 8 A.M. till 10.15 A.M., it then rose a little in a somewhat zigzag line, often remaining stationary, till 5 P.M., when the great evening fall commenced, which was continued till at least 10 P.M. By 7 A.M. on the following morning it had risen to the same level as on the previous morning, and then descended in a zigzag line. But from 10.30 A.M. till 4.15 P.M. it remained almost motionless, all power of movement being now lost. The petiole, therefore, of this very old leaf, which must have long ceased growing, moved periodically; but instead of circumnutating several times during the day, it moved only twice down and twice up in the course of 24 h., with the ascending and descending lines not coincident.

It has already been stated that the pinnae move independently of the main petiole. The petiole of a leaf was fixed to a cork support, close to the point whence the four pinnae diverge, with a short fine filament cemented longitudinally to one of the two terminal pinnae, and a graduated semicircle was placed close beneath it. By looking vertically down, its angular or lateral movements could be measured with accuracy. Between noon and 4.15 P.M. the pinna changed its position to one side by only 7o; but not continuously in the same direction, as it moved four times to one side, and three times to the opposite side, [page 377] in one instance to the extent of 16o. This pinna, therefore circumnutated. Later in the evening the four pinnae approach each other, and the one which was observed moved inwards 59o between noon and 6.45 P.M. Ten observations were made in the course of 2 h. 20 m. (at average intervals of 14 m.), between 4.25 and 6.45 P.M.; and there was now, when the leaf was going to sleep, no swaying from side to side, but a steady inward movement. Here therefore there is in the evening the same conversion of a circumnutating into a steady movement in one direction, as in the case of the main petiole.

It has also been stated that each separate leaflet circumnutates. A pinna was cemented with shellac on the summit of a little stick driven firmly into the ground, immediately beneath a pair of leaflets, to the midribs of both of which excessively fine glass filaments were attached. This treatment did not injure the leaflets, for they went to sleep in the usual manner, and long retained their sensitiveness. the movements of one of them were traced during 49 h., as shown in Fig. 158. On the first day the leaflet sank down till 11.30 A.M., and then rose till late in the evening in a zigzag line, indicating circumnutation. On the second day, when more accustomed to its new state, it oscillated twice up and twice down during the 24 h. This plant was subjected to a rather low temperature, viz., 62o - 64o F.; had it been kept warmer, no doubt the movements of the leaflet would have been much more rapid and complicated. It may be seen in the diagram that the ascending and descending lines do not coincide; but the large amount of lateral movement in the evening is the result of the leaflets bending towards the apex of the leaf when going to sleep. Another leaflet was casually observed, and found to be continually circumnutating during the same length of time.

The circumnutation of the leaves is not destroyed by their being subjected to moderately long continued darkness; but the proper periodicity of their movements is lost. Some very young seedlings were kept during two days in the dark (temp. 57o - 59o F.) except when the circumnutation of their stems was occasionally observed; and on the evening of the second day the leaflets did not fully and properly go to sleep. The pot was then placed for three days in a dark cupboard, under nearly the same temperature, and at the close of this period the leaflets showed no signs of sleeping, and were only slightly sensitive to a touch. On the following day the stem was cemented to a [page 378] stick, and the movements of two leaves were traced on a vertical glass during 72 h. The plants were still kept in the dark, excepting that at each observation, which lasted 3 or 4 minutes,

Fig 158. Mimosa pudica: circumnutation and nyctitropic movement of a leaflet (with pinna secured), traced on a vertical glass, from 8 A.M. Sept. 14th to 9 A.M. 16th.

they were illuminated by two candles. On the third day the leaflets still exhibited a vestige of sensitiveness when forcibly pressed, but in the evening they showed no signs of sleep. Nevertheless, their petioles continued to circumnutate distinctly, [page 379] although the proper order of their movements in relation to the day and night was wholly lost. Thus, one leaf descended during the first two nights (i.e. between 10 P.M. and 7 A.M. next morning) instead of ascending, and on the third night it moved chiefly in a lateral direction. The second leaf behaved in an equally abnormal manner, moving laterally during the first night, descending greatly during the second, and ascending to an unusual height during the third night.

With plants kept at a high temperature and exposed to the light, the most rapid circumnutating movement of the apex of a leaf which was observed, amounted to 1/500 of an inch in one second; and this would have equalled 1/8 of an inch in a minute, had not the leaf occasionally stood still. The actual distance travelled by the apex (as ascertained by a measure placed close to the leaf) was on one occasion nearly 3/4 of an inch in a vertical direction in 15 m.; and on another occasion 5/8 of an inch in 60 m.; but there was also some lateral movement.

Mimosa albida.*—The leaves of this plant, one of which is here figured (Fig. 159) reduced to 2/3 of the natural size, present some

Fig. 159. Mimosa albida: leaf seen from vertically above.

interesting peculiarities. It consists of a long petiole bearing only two pinnae (here represented as rather more divergent than is usual), each with two pairs of leaflets. But the inner

* Mr. Thiselton Dyer informs us that this Peruvian plant (which was sent to us from Kew) is considered by Mr. Bentham ('Trans. Linn. Soc.,' vol. xxx. p. 390) to be "the species or variety which most commonly represents the M. sensitiva of our gardens." [page 380]

basal leaflets are greatly reduced in size, owing probably to the want of space for their full development, so that they may be considered as almost rudimentary. They vary somewhat in size, and both occasionally disappear, or only one. Nevertheless, they are not in the least rudimentary in function, for they are sensitive, extremely heliotropic, circumnutate at nearly the same rate as the fully developed leaflets, and assume when asleep exactly the same position. With M. pudica the inner leaflets at the base and between the pinnae are likewise much shortened and obliquely truncated; this fact was well seen in some seedlings of M. pudica, in which the third leaf above the cotyledons bore only two pinnae, each with only 3 or 4 pairs of leaflets, of which the inner basal one was less than half as long as its fellow; so that the whole leaf resembled pretty closely that of M. albida. In this latter species the main petiole terminates in a little point, and on each side of this there is a pair of minute, flattened, lancet-shaped projections, hairy on their margins, which drop off and disappear soon after the leaf is fully developed. There can hardly be a doubt that these little projections are the last and fugacious representatives of an additional pair of leaflets to each pinna; for the outer one is twice as broad as the inner one, and a little longer, viz. 7/100 of an inch, whilst the inner one is only 5/100 - 6/100 long. Now if the basal pair of leaflets of the existing leaves were to become rudimentary, we should expect that the rudiments would still exhibit some trace of their present great inequality of size. The conclusion that the pinnae of the parent-form of M. albida possessed at least three pairs of leaflets, instead of, as at present, only two, is supported by the structure of the first true leaf; for this consists of a simple petiole, often bearing three pairs of leaflets. This latter fact, as well as the presence of the rudiments, both lead to the conclusion that M. albida is descended from a form the leaves of which bore more than two pairs of leaflets. The second leaf above the cotyledons resembles in all respects the leaves on fully developed plants.