Experiment 7.—A bit of half roasted meat (not measured) and a bit of gelatine were placed on the two ends of a leaf, which [page 303] opened spontaneously after eleven days; a vestige of the meat was left, and the surface of the leaf was here blackened; the gelatine had all disappeared.

Experiment 8.—A bit of half roasted meat (not measured) was placed on a leaf which was forcibly kept open by a clip, so that it was moistened with the secretion (very acid) only on its lower surface. Nevertheless, after only 22 1/2 hrs. it was surprisingly softened, when compared with another bit of the same meat which had been kept damp.

Experiment 9.—A cube of 1/10 of an inch of very compact roasted beef was placed on a leaf, which opened spontaneously after twelve days; so much feebly acid secretion was left on the leaf that it trickled off. The meat was completely disintegrated, but not all dissolved; there was no mould. The little mass was placed under the microscope; some of the fibrillae in the middle still exhibited transverse striae; others showed not a vestige of striae; and every gradation could be traced between these two states. Globules, apparently of fat, and some undigested fibro-elastic tissue remained. The meat was thus in the same state as that formerly described, which was half digested by Drosera. Here, again, as in the case of albumen, the digestive process seems slower than in Drosera. At the opposite end of the same leaf, a firmly compressed pellet of bread had been placed; this was completely disintegrated, I suppose, owing to the digestion of the gluten, but seemed very little reduced in bulk.

Experiment 10.—A cube of 1/20 of an inch of cheese and another of albumen were placed at opposite ends of the same leaf. After nine days the lobes opened spontaneously a little at the end enclosing the cheese, but hardly any or none was dissolved, though it was softened and surrounded by secretion. Two days subsequently the end with the albumen also opened spontaneously (i.e. eleven days after it was put on), a mere trace in a blackened and dry condition being left.

Experiment 11.—The same experiment with cheese and albumen repeated on another and rather torpid leaf. The lobes at the end with the cheese, after an interval of six days, opened spontaneously a little; the cube of cheese was much softened, but not dissolved, and but little, if at all, reduced in size. Twelve hours afterwards the end with the albumen opened, which now consisted of a large drop of transparent, not acid, viscid fluid.

Experiment 12.—Same experiment as the two last, and here again the leaf at the end enclosing the cheese opened before the [page 304] opposite end with the albumen; but no further observations were made.

Experiment 13.—A globule of chemically prepared casein, about 1/10 of an inch in diameter, was placed on a leaf, which spontaneously opened after eight days. The casein now consisted of a soft sticky mass, very little, if at all, reduced in size, but bathed in acid secretion.]

These experiments are sufficient to show that the secretion from the glands of Dionaea dissolves albumen, gelatine, and meat, if too large pieces are not given. Globules of fat and fibro-elastic tissue are not digested. The secretion, with its dissolved matter, if not in excess, is subsequently absorbed. On the other hand, although chemically prepared casein and cheese (as in the case of Drosera) excite much acid secretion, owing, I presume, to the absorption of some included albuminous matter, these substances are not digested, and are not appreciably, if at all, reduced in bulk.

[Effects of the Vapours of Chloroform, Sulphuric Ether, and Hydrocyanic Acid.—A plant bearing one leaf was introduced into a large bottle with a drachm (3.549 ml.) of chloroform, the mouth being imperfectly closed with cotton-wool. The vapour caused in 1 m. the lobes to begin moving at an imperceptibly slow rate; but in 3 m. the spikes crossed, and the leaf was soon completely shut. The dose, however, was much too large, for in between 2 and 3 hrs. the leaf appeared as if burnt, and soon died.

Two leaves were exposed for 30 m. in a 2-oz: vessel to the vapour of 30 minims (1.774 ml.) of sulphuric ether. One leaf closed after a time, as did the other whilst being removed from the vessel without being touched. Both leaves were greatly injured. Another leaf, exposed for 20 m. to 15 minims of ether, closed its lobes to a certain extent, and the sensitive filaments were now quite insensible. After 24 hrs. this leaf recovered its sensibility, but was still rather torpid. A leaf exposed in a large bottle for only 3 m. to ten drops was rendered insensible. After 52 m. it recovered its sensibility, and when one of the filaments was touched, the lobes closed. It began [page 305] to reopen after 20 hrs. Lastly another leaf was exposed for 4 m. to only four drops of the ether; it was rendered insensible, and did not close when its filaments were repeatedly touched, but closed when the end of the open leaf was cut off. This shows either that the internal parts had not been rendered insensible, or that an incision is a more powerful stimulus than repeated touches on the filaments. Whether the larger doses of chloroform and ether, which caused the leaves to close slowly, acted on the sensitive filaments or on the leaf itself, I do not know.

Cyanide of potassium, when left in a bottle, generates prussic or hydrocyanic acid. A leaf was exposed for 1 hr. 35 m. to the vapour thus formed; and the glands became within this time so colourless and shrunken as to be scarcely visible, and I at first thought that they had all dropped off. The leaf was not rendered insensible; for as soon as one of the filaments was touched it closed. It had, however, suffered, for it did not reopen until nearly two days had passed, and was not even then in the least sensitive. After an additional day it recovered its powers, and closed on being touched and subsequently reopened. Another leaf behaved in nearly the same manner after a shorter exposure to this vapour.]

On the Manner in which Insects are caught.—We will now consider the action of the leaves when insects happen to touch one of the sensitive filaments. This often occurred in my greenhouse, but I do not know whether insects are attracted in any special way by the leaves. They are caught in large numbers by the plant in its native country. As soon as a filament is touched, both lobes close with astonishing quickness; and as they stand at less than a right angle to each other, they have a good chance of catching any intruder. The angle between the blade and footstalk does not change when the lobes close. The chief seat of movement is near the midrib, but is not confined to this part; for, as the lobes come together, each curves inwards across its whole breadth; the marginal spikes however, not becoming curved. This move- [page 306] ment of the whole lobe was well seen in a leaf to which a large fly had been given, and from which a large portion had been cut off the end of one lobe; so that the opposite lobe, meeting with no resistance in this part, went on curving inwards much beyond the medial line. The whole of the lobe, from which a portion had been cut, was afterwards removed, and the opposite lobe now curled completely over, passing through an angle of from 120o to 130o, so as to occupy a position almost at right angles to that which it would have held had the opposite lobe been present.

From the curving inwards of the two lobes, as they move towards each other, the straight marginal spikes intercross by their tips at first, and ultimately by their bases. The leaf is then completely shut and encloses a shallow cavity. If it has been made to shut merely by one of the sensitive filaments having been touched, or if it includes an object not yielding soluble nitrogenous matter, the two lobes retain their inwardly concave form until they re-expand. The re-expansion under these circumstances—that is when no organic matter is enclosed—was observed in ten cases. In all of these, the leaves re-expanded to about two-thirds of the full extent in 24 hrs. from the time of closure. Even the leaf from which a portion of one lobe had been cut off opened to a slight degree within this same time. In one case a leaf re-expanded to about two-thirds of the full extent in 7 hrs., and completely in 32 hrs.; but one of its filaments had been touched merely with a hair just enough to cause the leaf to close. Of these ten leaves only a few re-expanded completely in less than two days, and two or three required even a little longer time. Before, however, they fully re-expand, they are ready to close [page 307] instantly if their sensitive filaments are touched. How many times a leaf is capable of shutting and opening if no animal matter is left enclosed, I do not know; but one leaf was made to close four times, reopening afterwards, within six days, On the last occasion it caught a fly, and then remained closed for many days.

This power of reopening quickly after the filaments have been accidentally touched by blades of grass, or by objects blown on the leaf by the wind, as occasionally happens in its native place,* must be of some importance to the plant; for as long as a leaf remains closed, it cannot of course capture an insect.

When the filaments are irritated and a leaf is made to shut over an insect, a bit of meat, albumen, gelatine, casein, and, no doubt, any other substance containing soluble nitrogenous matter, the lobes, instead of remaining concave, thus including a concavity, slowly press closely together throughout their whole breadth. As this takes place, the margins gradually become a little everted, so that the spikes, which at first intercrossed, at last project in two parallel rows. The lobes press against each other with such force that I have seen a cube of albumen much flattened, with distinct impressions of the little prominent glands; but this latter circumstance may have been partly caused by the corroding action of the secretion. So firmly do they become pressed together that, if any large insect or other object has been caught, a corresponding projection on the outside of the leaf is distinctly visible. When the two lobes are thus completely shut, they

* According to Dr. Curtis, in 'Boston Journal of Nat. Hist,' vol. i 1837, p. 123. [page 308]

resist being opened, as by a thin wedge driven between them, with astonishing force, and are generally ruptured rather than yield. If not ruptured, they close again, as Dr. Canby informs me in a letter, "with quite a loud flap." But if the end of a leaf is held firmly between the thumb and finger, or by a clip, so that the lobes cannot begin to close, they exert, whilst in this position, very little force.

I thought at first that the gradual pressing together of the lobes was caused exclusively by captured insects crawling over and repeatedly irritating the sensitive filaments; and this view seemed the more probable when I learnt from Dr. Burdon Sanderson that whenever the filaments of a closed leaf are irritated, the normal electric current is disturbed. Nevertheless, such irritation is by no means necessary, for a dead insect, or a bit of meat, or of albumen, all act equally well; proving that in these cases it is the absorption of animal matter which excites the lobes slowly to press close together. We have seen that the absorption of an extremely small quantity of such matter also causes a fully expanded leaf to close slowly; and this movement is clearly analogous to the slow pressing together of the concave lobes. This latter action is of high functional importance to the plant, for the glands on both sides are thus brought into contact with a captured insect, and consequently secrete. The secretion with animal matter in solution is then drawn by capillary attraction over the whole surface of the leaf, causing all the glands to secrete and allowing them to absorb the diffused animal matter. The movement, excited by the absorption of such matter, though slow, suffices for its final purpose, whilst the movement excited by one of the sensitive filaments being touched is rapid, and this is indis- [page 309] pensable for the capturing of insects. These two movements, excited by two such widely different means, are thus both well adapted, like all the other functions of the plant, for the purposes which they subserve.

There is another wide difference in the action of leaves which enclose objects, such as bits of wood, cork, balls of paper, or which have had their filaments merely touched, and those which enclose organic bodies yielding soluble nitrogenous matter. In the former case the leaves, as we have seen, open in under 24 hrs. and are then ready, even before being fully-expanded, to shut again. But if they have closed over nitrogen-yielding bodies, they remain closely shut for many days; and after re-expanding are torpid, and never act again, or only after a considerable interval of time. In four instances, leaves after catching insects never reopened, but began to wither, remaining closed—in one case for fifteen days over a fly; in a second, for twenty-four days, though the fly was small; in a third for twenty-four days over a woodlouse; and in a fourth, for thirty-five days over a large Tipula. In two other cases leaves remained closed for at least nine days over flies, and for how many more I do not know. It should, however, be added that in two instances in which very small insects had been naturally caught the leaf opened as quickly as if nothing had been caught; and I suppose that this was due to such small insects not having been crushed or not having excreted any animal matter, so that the glands were not excited. Small angular bits of albumen and gelatine were placed at both ends of three leaves, two of which remained closed for thirteen and the other for twelve days. Two other leaves remained closed over bits of [page 310] meat for eleven days, a third leaf for eight days, and a fourth (but this had been cracked and injured) for only six days. Bits of cheese, or casein, were placed at one end and albumen at the other end of three leaves; and the ends with the former opened after six, eight, and nine days, whilst the opposite ends opened a little later. None of the above bits of meat, albumen, &c., exceeded a cube of 1/10 of an inch (2.54 mm.) in size, and were sometimes smaller; yet these small portions sufficed to keep the leaves closed for many days. Dr. Canby informs me that leaves remain shut for a longer time over insects than over meat; and from what I have seen, I can well believe that this is the case, especially if the insects are large.

In all the above cases, and in many others in which leaves remained closed for a long but unknown period over insects naturally caught, they were more or less torpid when they reopened. Generally they were so torpid during many succeeding days that no excitement of the filaments caused the least movement. In one instance, however, on the day after a leaf opened which had clasped a fly, it closed with extreme slowness when one of its filaments was touched; and although no object was left enclosed, it was so torpid that it did not re-open for the second time until 44 hrs. had elapsed. In a second case, a leaf which had expanded after remaining closed for at least nine days over a fly, when greatly irritated, moved one alone of its two lobes, and retained this unusual position for the next two days. A third case offers the strongest exception which I have observed; a leaf, after remaining clasped for an unknown time over a fly, opened, and when one of its filaments was touched, closed, though rather slowly. Dr. Canby, [page 311] who observed in the United States a large number of plants which, although not in their native site, were probably more vigorous than my plants, informs me that he has "several times known vigorous leaves to devour their prey several times; but ordinarily twice, or, quite often, once was enough to render them unserviceable." Mrs. Treat, who cultivated many plants in New Jersey, also informs me that "several leaves caught successively three insects each, but most of them were not able to digest the third fly, but died in the attempt. Five leaves, however, digested each three flies, and closed over the fourth, but died soon after the fourth capture. Many leaves did not digest even one large insect." It thus appears that the power of digestion is somewhat limited, and it is certain that leaves always remain clasped for many days over an insect, and do not recover their power of closing again for many subsequent days. In this respect Dionaea differs from Drosera, which catches and digests many insects after shorter intervals of time.

We are now prepared to understand the use of the marginal spikes, which form so conspicuous a feature in the appearance of the plant (fig. 12, p. 287), and which at first seemed to me in my ignorance useless appendages. From the inward curvature of the lobes as they approach each other, the tips of the marginal spikes first intercross, and ultimately their bases. Until the edges of the lobes come into contact, elongated spaces between the spikes, varying from the 1/15 to the 1/10 of an inch (1.693 to 2.54 mm.) in breadth, according to the size of the leaf, are left open. Thus an insect, if its body is not thicker than these measurements, can easily escape between the crossed spikes, when disturbed by the closing lobes and in- [page 312] creasing darkness; and one of my sons actually saw a small insect thus escaping. A moderately large insect, on the other hand, if it tries to escape between the bars will surely be pushed back again into its horrid prison with closing walls, for the spikes continue to cross more and more until the edges of the lobes come into contact. A very strong insect, however, would be able to free itself, and Mrs. Treat saw this effected by a rose-chafer (Macrodactylus subspinosus) in the United States. Now it would manifestly be a great disadvantage to the plant to waste many days in remaining clasped over a minute insect, and several additional days or weeks in afterwards recovering its sensibility; inasmuch as a minute insect would afford but little nutriment. It would be far better for the plant to wait for a time until a moderately large insect was captured, and to allow all the little ones to escape; and this advantage is secured by the slowly intercrossing marginal spikes, which act like the large meshes of a fishing-net, allowing the small and useless fry to escape.

As I was anxious to know whether this view was correct—and as it seems a good illustration of how cautious we ought to be in assuming, as I had done with respect to the marginal spikes, that any fully developed structure is useless—I applied to Dr. Canby. He visited the native site of the plant, early in the season, before the leaves had grown to their full size, and sent me fourteen leaves, containing naturally captured insects. Four of these had caught rather small insects, viz. three of them ants, and the fourth a rather small fly, but the other ten had all caught large insects, namely, five elaters, two chrysomelas, a curculio, a thick and broad spider, and a scolopendra. Out of these ten insects, no less than eight [page 313] were beetles,* and out of the whole fourteen there was only one, viz. a dipterous insect, which could readily take flight. Drosera, on the other hand, lives chiefly on insects which are good flyers, especially Diptera, caught by the aid of its viscid secretion. But what most concerns us is the size of the ten larger insects. Their average length from head to tail was .256 of an inch, the lobes of the leaves being on an average .53 of an inch in length, so that the insects were very nearly half as long as the leaves within which they were enclosed. Only a few of these leaves, therefore, had wasted their powers by capturing small prey, though it is probable that many small insects had crawled over them and been caught, but had then escaped through the bars.

The Transmission of the Motor Impulse, and Means of Movement.—It is sufficient to touch any one of the six filaments to cause both lobes to close, these becoming at the same time incurved throughout their whole breadth. The stimulus must therefore radiate in all directions from any one filament. It must also be transmitted with much rapidity across the leaf, for in all ordinary cases both lobes close simultaneously, as far as the eye can judge. Most physiologists believe that in irritable plants the excitement is transmitted along, or in close connection with, the fibro-vascular bundles. In Dionaea, the course of these vessels (composed of spiral and ordinary vascular

* Dr. Canby remarks ('Gardener's Monthly,' August 1868), "as a general thing beetles and insects of that kind, though always killed, seem to be too hard-shelled to serve as food, and after a short time are rejected." I am surprised at this statement, at least with respect to such beetles as elaters, for the five which I examined were in an extremely fragile and empty condition, as if all their internal parts had been partially digested. Mrs. Treat informs me that the plants which she cultivated in New Jersey chiefly caught Diptera. [page 314]

tissue) seems at first sight to favour this belief; for they run up the midrib in a great bundle, sending off small bundles almost at right angles on each side. These bifurcate occasionally as they extend towards the margin, and close to the margin small branches from adjoining vessels unite and enter the marginal spikes. At some of these points of union the vessels form curious loops, like those described under Drosera. A continuous zigzag line of vessels thus runs round the whole circumference of the leaf, and in the midrib all the vessels are in close contact; so that all parts of the leaf seem to be brought into some degree of communication. Nevertheless, the presence of vessels is not necessary for the transmission of the motor impulse, for it is transmitted from the tips of the sensitive filaments (these being about the 1/20 of an inch in length), into which no vessels enter; and these could not have been overlooked, as I made thin vertical sections of the leaf at the bases of the filaments.

On several occasions, slits about the 1/10 of an inch in length were made with a lancet, close to the bases of the filaments, parallel to the midrib, and, therefore, directly across the course of the vessels. These were made sometimes on the inner and sometimes on the outer sides of the filaments; and after several days, when the leaves had reopened, these filaments were touched roughly (for they were always rendered in some degree torpid by the operation), and the lobes then closed in the ordinary manner, though slowly, and sometimes not until after a considerable interval of time. These cases show that the motor impulse is not transmitted along the vessels, and they further show that there is no necessity for a direct line of communication from the filament which is [page 315] touched towards the midrib and opposite lobe, or towards the outer parts of the same lobe.

Two slits near each other, both parallel to the midrib, were next made in the same manner as before, one on each side of the base of a filament, on five distinct leaves, so that a little slip bearing a filament was connected with the rest of the leaf only at its two ends. These slips were nearly of the same size; one was carefully measured; it was .12 of an inch (3.048 mm.) in length, and .08 of an inch (2.032 mm.) in breadth; and in the middle stood the filament. Only one of these slips withered and perished. After the leaf had recovered from the operation, though the slits were still open, the filaments thus circumstanced were roughly touched, and both lobes, or one alone, slowly closed. In two instances touching the filament produced no effect; but when the point of a needle was driven into the slip at the base of the filament, the lobes slowly closed. Now in these cases the impulse must have proceeded along the slip in a line parallel to the midrib, and then have radiated forth, either from both ends or from one end alone of the slip, over the whole surface of the two lobes.

Again, two parallel slits, like the former ones, were made, one on each side of the base of a filament, at right angles to the midrib. After the leaves (two in number) had recovered, the filaments were roughly touched, and the lobes slowly closed; and here the impulse must have travelled for a short distance in a line at right angles to the midrib, and then have radiated forth on all sides over both lobes. These several cases prove that the motor impulse travels in all directions through the cellular tissue, independently of the course of the vessels.

With Drosera we have seen that the motor impulse [page 316] is transmitted in like manner in all directions through the cellular tissue; but that its rate is largely governed by the length of the cells and the direction of their longer axes. Thin sections of a leaf of Dionaea were made by my son, and the cells, both those of the central and of the more superficial layers, were found much elongated, with their longer axes directed towards the midrib; and it is in this direction that the motor impulse must be sent with great rapidity from one lobe to the other, as both close simultaneously. The central parenchymatous cells are larger, more loosely attached together, and have more delicate walls than the more superficial cells. A thick mass of cellular tissue forms the upper surface of the midrib over the great central bundle of vessels.

When the filaments were roughly touched, at the bases of which slits had been made, either on both sides or on one side, parallel to the midrib or at right angles to it, the two lobes, or only one, moved. In one of these cases, the lobe on the side which bore the filament that was touched moved, but in three other cases the opposite lobe alone moved; so that an injury which was sufficient to prevent a lobe moving did not prevent the transmission from it of a stimulus which excited the opposite lobe to move. We thus also learn that, although normally both lobes move together, each has the power of independent movement. A case, indeed, has already been given of a torpid leaf that had lately re-opened after catching an insect, of which one lobe alone moved when irritated. Moreover, one end of the same lobe can close and re- expand, independently of the other end, as was seen in some of the foregoing experiments.

When the lobes, which are rather thick, close, no trace of wrinkling can be seen on any part of their upper [page 317] surfaces, It appears therefore that the cells must contract. The chief seat of the movement is evidently in the thick mass of cells which overlies the central bundle of vessels in the midrib. To ascertain whether this part contracts, a leaf was fastened on the stage of the microscope in such a manner that the two lobes could not become quite shut, and having made two minute black dots on the midrib, in a transverse line and a little towards one side, they were found by the micrometer to be 17/1000 of an inch apart. One of the filaments was then touched and the lobes closed; but as they were prevented from meeting, I could still see the two dots, which now were 15/1000 of an inch apart, so that a small portion of the upper surface of the midrib had contracted in a transverse line 2/1000 of an inch (.0508 mm.).

We know that the lobes, whilst closing, become slightly incurved throughout their whole breadth. This movement appears to be due to the contraction of the superficial layers of cells over the whole upper surface. In order to observe their contraction, a narrow strip was cut out of one lobe at right angles to the midrib, so that the surface of the opposite lobe could be seen in this part when the leaf was shut. After the leaf had recovered from the operation and had re-expanded, three minute black dots were made on the surface opposite to the slit or window, in a line at right angles to the midrib. The distance between the dots was found to be 40/1000 of an inch, so that the two extreme dots were 80/1000 of an inch apart. One of the filaments was now touched and the leaf closed. On again measuring the distances between the dots, the two next to the midrib were nearer together by 1 to 2/1000 of an inch, and the two further dots by 3 to 4/1000 of an inch, than they were before; so that the two extreme [page 318] dots now stood about 5/1000 of an inch (.127 mm.) nearer together than before. If we suppose the whole upper surface of the lobe, which was 400/1000 of an inch in breadth, to have contracted in the same proportion, the total contraction will have amounted to about 25/1000 or 1/40 of an inch (.635 mm.); but whether this is sufficient to account for the slight inward curvature of the whole lobe, I am unable to say.

Finally, with respect to the movement of the leaves, the wonderful discovery made by Dr. Burdon Sanderson* is now universally known; namely that there exists a normal electrical current in the blade and footstalk; and that when the leaves are irritated, the current is disturbed in the same manner as takes place during the contraction of the muscle of an animal.

The Re-expansion of the Leaves.—This is effected at an insensibly slow rate, whether or not any object is enclosed. One lobe can re-expand by itself, as occurred with the torpid leaf of which one lobe alone had closed. We have also seen in the experiments with cheese and albumen that the two ends of the same lobe can re-expand to a certain extent independently of each other. But in all ordinary cases both lobes open at the same time. The re-expansion is not determined by the sensitive filaments; all three filaments on one lobe were cut off close to their bases; and the three

* Proc. Royal Soc.' vol. xxi. p. 495; and lecture at the Royal Institution, June 5, 1874, given in 'Nature,' 1874, pp. 105 and 127.

Nuttall, in his 'Gen. American Plants,' p. 277 (note), says that, whilst collecting this plant in its native home, "I had occasion to observe that a detached leaf would make repeated efforts towards disclosing itself to the influence of the sun; these attempts consisted in an undulatory motion of the marginal ciliae, accompanied by a partial opening and succeeding collapse of the lamina, which at length terminated in a complete expansion and in the destruction of sensibility." I am indebted to Prof. Oliver for this reference; but I do not understand what took place. [page 319]

leaves thus treated re-expanded,—one to a partial extent in 24 hrs.,—a second to the same extent in 48 hrs., and the third, which had been previously injured, not until the sixth day. These leaves after their re-expansion closed quickly when the filaments on the other lobe were irritated. These were then cut off one of the leaves, so that none were left. This mutilated leaf, notwithstanding the loss of all its filaments, re-expanded in two days in the usual manner. When the filaments have been excited by immersion in a solution of sugar, the lobes do not expand so soon as when the filaments have been merely touched; and this, I presume, is due to their having been strongly affected through exosmose, so that they continue for some time to transmit a motor impulse to the upper surface of the leaf.

The following facts make me believe that the several layers of cells forming the lower surface of the leaf are always in a state of tension; and that it is owing to this mechanical state, aided probably by fresh fluid being attracted into the cells, that the lobes begin to separate or expand as soon as the contraction of the upper surface diminishes. A leaf was cut off and suddenly plunged perpendicularly into boiling water: I expected that the lobes would have closed, but instead of doing so, they diverged a little. I then took another fine leaf, with the lobes standing at an angle of nearly 80o to each other; and on immersing it as before, the angle suddenly increased to 90o. A third leaf was torpid from having recently re-expanded after having caught a fly, so that repeated touches of the filaments caused not the least movement; nevertheless, when similarly immersed, the lobes separated a little. As these leaves were inserted perpendicularly into the boiling water, both surfaces and the filaments [page 320] must have been equally affected; and I can understand the divergence of the lobes only by supposing that the cells on the lower side, owing to their state of tension, acted mechanically and thus suddenly drew the lobes a little apart, as soon as the cells on the upper surface were killed and lost their contractile power. We have seen that boiling water in like manner causes the tentacles of Drosera to curve backwards; and this is an analogous movement to the divergence of the lobes of Dionaea.

In some concluding remarks in the fifteenth chapter on the Droseraceae, the different kinds of irritability possessed by the several genera, and the different manner in which they capture insects, will be compared. [page 321]



CHAPTER XIV.

ALDROVANDA VESICULOSA.

Captures crustaceans—Structure of the leaves in comparison with those of Dionaea— Absorption by the glands, by the quadrifid processes, and points on the infolded margins— Aldrovanda vesiculosa, var. australis—Captures prey—Absorption of animal matter— Aldrovanda vesiculosa, var. verticillata—Concluding remarks.

THIS plant may be called a miniature aquatic Dionaea. Stein discovered in 1873 that the bilobed leaves, which are generally found closed in Europe, open under a sufficiently high temperature, and, when touched, suddenly close.* They re-expand in from 24 to 36 hours, but only, as it appears, when inorganic objects are enclosed. The leaves sometimes contain bubbles of air, and were formerly supposed to be bladders; hence the specific name of vesiculosa. Stein observed that water-insects were sometimes caught, and Prof. Cohn has recently found within the leaves of naturally growing plants many kinds of crustaceans and larvae. Plants which had been kept in filtered water were placed by him in a vessel con-

* Since his original publication, Stein has found out that the irritability of the leaves was observed by De Sassus, as recorded in 'Bull. Bot. Soc. de France,' in 1861. Delpino states in a paper published in 1871 ('Nuovo Giornale Bot. Ital.' vol. iii. p. 174) that "una quantit di chioccioline e di altri animalcoli acquatici" are caught and suffocated by the leaves. I presume that chioccioline are fresh-water molluscs. It would be interesting to know whether their shells are at all corroded by the acid of the digestive secretion.

I am greatly indebted to this distinguished naturalist for having sent me a copy of his memoir on Aldrovanda, before its publication in his 'Beitrge zur Biologie der Pflanzen,' drittes Heft, 1875, page 71. [page 322]

taining numerous crustaceans of the genus Cypris, and next morning many were found imprisoned and alive, still swimming about within the closed leaves, but doomed to certain death.

Directly after reading Prof. Cohn's memoir, I received through the kindness of Dr. Hooker living plants from Germany. As I can add nothing to Prof. Cohn's excellent description, I will give only two illustrations, one of a whorl of leaves copied from his work, and the other of a leaf pressed flat open, drawn by my son Francis. I will, however, append a few remarks on the differences between this plant and Dionaea.

Aldrovanda is destitute of roots and floats freely in the water. The leaves are arranged in whorls round the stem. Their broad petioles terminate in from four to six rigid projections,* each tipped with a stiff, short bristle. The bilobed leaf, with the midrib likewise tipped with a bristle, stands in the midst of these projections, and is evidently defended by them. The lobes are formed of very delicate tissue, so as to be translucent; they open, according to Cohn, about as much as the two valves of a living mussel-shell, therefore even less than the lobes of Dionaea; and this must make the capture of aquatic animals more easy. The outside of the leaves and the petioles are covered with minute two-armed papillae, evidently answering to the eight-rayed papillae of Dionaea.

Each lobe rather exceeds a semi-circle in convexity, and consists of two very different concentric portions; the inner and lesser portion, or that next to the midrib,

*There has been much discussion by botanists on the homological nature of these projections. Dr. Nitschke ('Bot. Zeitung,' 1861, p. 146) believes that they correspond with the fimbriated scale-like bodies found at the bases of the petioles of Drosera. [page 323]

is slightly concave, and is formed, according to Cohn, of three layers of cells. Its upper surface is studded with colourless glands like, but more simple than, those of Dionaea; they are supported on distinct footstalks, consisting of two rows of cells. The outer

FIG. 13. (Aldrovanda vesiculosa.) Upper figure, whorl of leaves (from Prof. Cohn). Lower figure, leaf pressed flat open and greatly enlarged.

and broader portion of the lobe is flat and very thin, being formed of only two layers of cells. Its upper surface does not bear any glands, but, in their place, small quadrifid processes, each consisting of four tapering projections, which rise from a common [page 324] prominence. These processes are formed of very delicate membrane lined with a layer of protoplasm; and they sometimes contain aggregated globules of hyaline matter. Two of the slightly diverging arms are directed towards the circumference, and two towards the midrib, forming together a sort of Greek cross. Occasionally two of the arms are replaced by one, and then the projection is trifid. We shall see in a future chapter that these projections curiously resemble those found within the bladders of Utricularia, more especially of Utricularia montana, although this genus is not related to Aldrovanda.

A narrow rim of the broad flat exterior part of each lobe is turned inwards, so that, when the lobes are closed, the exterior surfaces of the infolded portions come into contact. The edge itself bears a row of conical, flattened, transparent points with broad bases, like the prickles on the stem of a bramble or Rubus. As the rim is infolded, these points are directed towards the midrib, and they appear at first as if they were adapted to prevent the escape of prey; but this can hardly be their chief function, for they are composed of very delicate and highly flexible membrane, which can be easily bent or quite doubled back without being cracked. Nevertheless, the infolded rims, together with the points, must somewhat interfere with the retrograde movement of any small creature, as soon as the lobes begin to close. The circumferential part of the leaf of Aldrovanda thus differs greatly from that of Dionaea; nor can the points on the rim be considered as homologous with the spikes round the leaves of Dionaea, as these latter are prolongations of the blade, and not mere epidermic productions. They appear also to serve for a widely different purpose. [page 325]

On the concave gland-bearing portion of the lobes, and especially on the midrib, there are numerous, long, finely pointed hairs, which, as Prof. Cohn remarks, there can be little doubt are sensitive to a touch, and, when touched, cause the leaf to close. They are formed of two rows of cells, or, according to Cohn, sometimes of four, and do not include any vascular tissue. They differ also from the six sensitive filaments of Dionaea in being colourless, and in having a medial as well as a basal articulation. No doubt it is owing to these two articulations that, notwithstanding their length, they escape being broken when the lobes close.

The plants which I received during the early part of October from Kew never opened their leaves, though subjected to a high temperature. After examining the structure of some of them, I experimented on only two, as I hoped that the plants would grow; and I now regret that I did not sacrifice a greater number.

A leaf was cut open along the midrib, and the glands examined under a high power. It was then placed in a few drops of an infusion of raw meat. After 3 hrs. 20 m. there was no change, but when next examined after 23 hrs. 20 m., the outer cells of the glands contained, instead of limpid fluid, spherical masses of a granular substance, showing that matter had been absorbed from the infusion. That these glands secrete a fluid which dissolves or digests animal matter out of the bodies of the creatures which the leaves capture, is also highly probable from the analogy of Dionaea. If we may trust to the same analogy, the concave and inner portions of the two lobes probably close together by a slow movement, as soon as the glands have absorbed a slight amount of [page 326] already soluble animal matter. The included water would thus be pressed out, and the secretion consequently not be too much diluted to act. With respect to the quadrifid processes on the outer parts of the lobes, I was not able to decide whether they had been acted on by the infusion; for the lining of protoplasm was somewhat shrunk before they were immersed. Many of the points on the infolded rims also had their lining of protoplasm similarly shrunk, and contained spherical granules of hyaline matter.

A solution of urea was next employed. This substance was chosen partly because it is absorbed by the quadrifid processes and more especially by the glands of Utricularia—a plant which, as we shall hereafter see, feeds on decayed animal matter. As urea is one of the last products of the chemical changes going on in the living body, it seems fitted to represent the early stages of the decay of the dead body. I was also led to try urea from a curious little fact mentioned by Prof. Cohn, namely that when rather large crustaceans are caught between the closing lobes, they are pressed so hard whilst making their escape that they often void their sausage-shaped masses of excrement, which were found within most of the leaves. These masses, no doubt, contain urea. They would be left either on the broad outer surfaces of the lobes where the quadrifids are situated, or within the closed concavity. In the latter case, water charged with excrementitious and decaying matter would be slowly forced outwards, and would bathe the quadrifids, if I am right in believing that the concave lobes contract after a time like those of Dionaea. Foul water would also be apt to ooze out at all times, especially when bubbles of air were generated within the concavity.

A leaf was cut open and examined, and the outer [page 327] cells of the glands were found to contain only limpid fluid. Some of the quadrifids included a few spherical granules, but several were transparent and empty, and their positions were marked. This leaf was now immersed in a little solution of one part of urea to 146 of water, or three grains to the ounce. After 3 hrs. 40 m. there was no change either in the glands or quadrifids; nor was there any certain change in the glands after 24 hrs.; so that, as far as one trial goes, urea does not act on them in the same manner as an infusion of raw meat. It was different with the quadrifids; for the lining of protoplasm, instead of presenting a uniform texture, was now slightly shrunk, and exhibited in many places minute, thickened, irregular, yellowish specks and ridges, exactly like those which appear within the quadrifids of Utricularia when treated with this same solution. Moreover, several of the quadrifids, which were before empty, now contained moderately sized or very small, more or less aggregated, globules of yellowish matter, as likewise occurs under the same circumstances with Utricularia. Some of the points on the infolded margins of the lobes were similarly affected; for their lining of protoplasm was a little shrunk and included yellowish specks; and those which were before empty now contained small spheres and irregular masses of hyaline matter, more or less aggregated; so that both the points on the margins and the quadrifids had absorbed matter from the solution in the course of 24 hrs.; but to this subject I shall recur. In another rather old leaf, to which nothing had been given, but which had been kept in foul water, some of the quadrifids contained aggregated translucent globules. These were not acted on by a solution of one part of carbonate of ammonia to 218 of water; and this negative result [page 328] agrees with what I have observed under similar circumstances with Utricularia.

Aldrovanda vesiculosa, var. australis.—Dried leaves of this plant from Queensland in Australia were sent me by Prof. Oliver from the herbarium at Kew. Whether it ought to be considered as a distinct species or a variety, cannot be told until the flowers are examined by a botanist. The projections at the upper end of the petiole (from four to six in number) are considerably longer relatively to the blade, and much more attenuated than those of the European form. They are thickly covered for a considerable space near their extremities with the upcurved prickles, which are quite absent in the latter form; and they generally bear on their tips two or three straight prickles instead of one. The bilobed leaf appears also to be rather larger and somewhat broader, with the pedicel by which it is attached to the upper end of the petiole a little longer. The points on the infolded margins likewise differ; they have narrower bases, and are more pointed; long and short points also alternate with much more regularity than in the European form. The glands and sensitive hairs are similar in the two forms. No quadrifid processes could be seen on several of the leaves, but I do not doubt that they were present, though indistinguishable from their delicacy and from having shrivelled; for they were quite distinct on one leaf under circumstances presently to be mentioned.

Some of the closed leaves contained no prey, but in one there was a rather large beetle, which from its flattened tibiae I suppose was an aquatic species, but was not allied to Colymbetes. All the softer tissues of this beetle were completely dissolved, and its chitinous integuments were as clean as if they had been [page 329] boiled in caustic potash; so that it must have been enclosed for a considerable time. The glands were browner and more opaque than those on other leaves which had caught nothing; and the quadrifid processes, from being partly filled with brown granular matter, could be plainly distinguished, which was not the case, as already stated, on the other leaves. Some of the points on the infolded margins likewise contained brownish granular matter. We thus gain additional evidence that the glands, the quadrifid processes, and the marginal points, all have the power of absorbing matter, though probably of a different nature.

Within another leaf disintegrated remnants of a rather small animal, not a crustacean, which had simple, strong, opaque mandibles, and a large unarticulated chitinous coat, were present. Lumps of black organic matter, possibly of a vegetable nature, were enclosed in two other leaves; but in one of these there was also a small worm much decayed. But the nature of partially digested and decayed bodies, which have been pressed flat, long dried, and then soaked in water, cannot be recognised easily. All the leaves contained unicellular and other Algae, still of a greenish colour, which had evidently lived as intruders, in the same manner as occurs, according to Cohn, within the leaves of this plant in Germany.

Aldrovanda vesiculosa, var. verticillata.—Dr. King, Superintendent of the Botanic Gardens, kindly sent me dried specimens collected near Calcutta. This form was, I believe, considered by Wallich as a distinct species, under the name of verticillata. It resembles the Australian form much more nearly than the European; namely in the projections at the upper end of the petiole being much attenuated and covered with [page 330] upcurved prickles; they terminate also in two straight little prickles. The bilobed leaves are, I believe, larger and certainly broader even than those of the Australian form; so that the greater convexity of their margins was conspicuous. The length of an open leaf being taken at 100, the breadth of the Bengal form is nearly 173, of the Australian form 147, and of the German 134. The points on the infolded margins are like those in the Australian form. Of the few leaves which were examined, three contained entomostracan crustaceans.

Concluding Remarks.—The leaves of the three foregoing closely allied species or varieties are manifestly adapted for catching living creatures. With respect to the functions of the several parts, there can be little doubt that the long jointed hairs are sensitive, like those of Dionaea, and that, when touched, they cause the lobes to close. That the glands secrete a true digestive fluid and afterwards absorb the digested matter, is highly probable from the analogy of Dionaea,—from the limpid fluid within their cells being aggregated into spherical masses, after they had absorbed an infusion of raw meat,—from their opaque and granular condition in the leaf, which had enclosed a beetle for a long time,—and from the clean condition of the integuments of this insect, as well as of crustaceans (as described by Cohn), which have been long captured. Again, from the effect produced on the quadrifid processes by an immersion for 24 hrs. in a solution of urea,—from the presence of brown granular matter within the quadrifids of the leaf in which the beetle had been caught,—and from the analogy of Utricularia,—it is probable that these processes absorb excrementitious and decaying animal matter. It is a more curious fact that the points on [page 331] the infolded margins apparently serve to absorb decayed animal matter in the same manner as the quadrifids. We can thus understand the meaning of the infolded margins of the lobes furnished with delicate points directed inwards, and of the broad, flat, outer portions, bearing quadrifid processes; for these surfaces must be liable to be irrigated by foul water flowing from the concavity of the leaf when it contains dead animals. This would follow from various causes,—from the gradual contraction of the concavity,—from fluid in excess being secreted,- -and from the generation of bubbles of air. More observations are requisite on this head; but if this view is correct, we have the remarkable case of different parts of the same leaf serving for very different purposes—one part for true digestion, and another for the absorption of decayed animal matter. We can thus also understand how, by the gradual loss of either power, a plant might be gradually adapted for the one function to the exclusion of the other; and it will hereafter be shown that two genera, namely Pinguicula and Utricularia, belonging to the same family, have been adapted for these two different functions. [page 332]



CHAPTER XV.

DROSOPHYLLUM—RORIDULA—BYBLIS—GLANDULAR HAIRS OF OTHER PLANTS— CONCLUDING REMARKS ON THE DROSERACEAE.

Drosophyllum—Structure of leaves—Nature of the secretion—Manner of catching insects— Power of absorption—Digestion of animal substances—Summary on Drosophyllum—Roridula- -Byblis—Glandular hairs of other plants, their power of absorption—Saxifraga—Primula— Pelargonium—Erica—Mirabilis—Nicotiana—Summary on glandular hairs—Concluding remarks on the Droseraceae.

DROSOPHYLLUM LUSITANICUM.—This rare plant has been found only in Portugal, and, as I hear from Dr. Hooker, in Morocco. I obtained living specimens through the great kindness of Mr. W.C. Tait, and afterwards from Mr. G. Maw and Dr. Moore. Mr. Tait informs me that it grows plentifully on the sides of dry hills near Oporto, and that vast numbers of flies adhere to the leaves. This latter fact is well-known to the villagers, who call the plant the "fly-catcher, " and hang it up in their cottages for this purpose. A plant in my hot-house caught so many insects during the early part of April, although the weather was cold and insects scarce, that it must have been in some manner strongly attractive to them. On four leaves of a young and small plant, 8, 10, 14, and 16 minute insects, chiefly Diptera, were found in the autumn adhering to them. I neglected to examine the roots, but I hear from Dr. Hooker that they are very small, as in the case of the previously mentioned members of the same family of the Droseraceae.

The leaves arise from an almost woody axis; they [page 333] are linear, much attenuated towards their tips, and several inches in length. The upper surface is concave, the lower convex, with a narrow channel down the middle. Both surfaces, with the exception of the channel, are covered with glands, supported on pedicels and arranged in irregular longitudinal rows. These organs I shall call tentacles, from their close resemblance to those of Drosera, though they have no power of movement. Those on the same leaf differ much in length. The glands also differ in size, and are of a bright pink or of a purple colour; their upper surfaces are convex, and the lower flat or even concave, so that they resemble miniature mushrooms in appearance. They are formed of two (as I believe) layers of delicate angular cells, enclosing eight or ten larger cells with thicker, zigzag walls. Within these larger cells there are others marked by spiral lines, and apparently connected with the spiral vessels which run up the green multi-cellular pedicels. The glands secrete large drops of viscid secretion. Other glands, having the same general appearance, are found on the flower-peduncles and calyx.

FIG. 14. (Drosophyllum lusitanicum.) Part of leaf, enlarged seven times, showing lower surface.

Besides the glands which are borne on longer or shorter pedicels, there are numerous ones, both on the upper and lower surfaces of the leaves, so small as to be scarcely visible to the naked eye. They are colourless and almost sessile, either circular or oval in outline; the latter occurring chiefly on the backs of the leaves (fig. 14). Internally they have exactly the same structure as the larger glands which are supported on pedicels; [page 334] and indeed the two sets almost graduate into one another. But the sessile glands differ in one important respect, for they never secrete spontaneously, as far as I have seen, though I have examined them under a high power on a hot day, whilst the glands on pedicels were secreting copiously. Nevertheless, if little bits of damp albumen or fibrin are placed on these sessile glands, they begin after a time to secrete, in the same manner as do the glands of Dionaea when similarly treated. When they were merely rubbed with a bit of raw meat, I believe that they likewise secreted. Both the sessile glands and the taller ones on pedicels have the power of rapidly absorbing nitrogenous matter.

The secretion from the taller glands differs in a remarkable manner from that of Drosera, in being acid before the glands have been in any way excited; and judging from the changed colour of litmus paper, more strongly acid than that of Drosera. This fact was observed repeatedly; on one occasion I chose a young leaf, which was not secreting freely, and had never caught an insect, yet the secretion on all the glands coloured litmus paper of a bright red. From the quickness with which the glands are able to obtain animal matter from such substances as well-washed fibrin and cartilage, I suspect that a small quantity of the proper ferment must be present in the secretion before the glands are excited, so that a little animal matter is quickly dissolved.

Owing to the nature of the secretion or to the shape of the glands, the drops are removed from them with singular facility. It is even somewhat difficult, by the aid of a finely pointed polished needle, slightly damped with water, to place a minute particle of any kind on one of the drops; for on withdrawing the [page 335] needle, the drop is generally withdrawn; whereas with Drosera there is no such difficulty, though the drops are occasionally withdrawn. From this peculiarity, when a small insect alights on a leaf of Drosophyllum, the drops adhere to its wings, feet, or body, and are drawn from the gland; the insect then crawls onward and other drops adhere to it; so that at last, bathed by the viscid secretion, it sinks down and dies, resting on the small sessile glands with which the surface of the leaf is thickly covered. In the case of Drosera, an insect sticking to one or more of the exterior glands is carried by their movement to the centre of the leaf; with Drosophyllum, this is effected by the crawling of the insect, as from its wings being clogged by the secretion it cannot fly away.

There is another difference in function between the glands of these two plants: we know that the glands of Drosera secrete more copiously when properly excited. But when minute particles of carbonate of ammonia, drops of a solution of this salt or of the nitrate of ammonia, saliva, small insects, bits of raw or roast meat, albumen, fibrin or cartilage, as well as inorganic particles, were placed on the glands of Drosophyllum, the amount of secretion never appeared to be in the least increased. As insects do not commonly adhere to the taller glands, but withdraw the secretion, we can see that there would be little use in their having acquired the habit of secreting copiously when stimulated; whereas with Drosera this is of use, and the habit has been acquired. Nevertheless, the glands of Drosophyllum, without being stimulated, continually secrete, so as to replace the loss by evaporation. Thus when a plant was placed under a small bell-glass with its inner surface and support thoroughly wetted, there was no loss by evaporation, and so much [page 336] secretion was accumulated in the course of a day that it ran down the tentacles and covered large spaces of the leaves.

The glands to which the above named nitrogenous substances and liquids were given did not, as just stated, secrete more copiously; on the contrary, they absorbed their own drops of secretion with surprising quickness. Bits of damp fibrin were placed on five glands, and when they were looked at after an interval of 1 hr. 12 m., the fibrin was almost dry, the secretion having been all absorbed. So it was with three cubes of albumen after 1 hr. 19 m., and with four other cubes, though these latter were not looked at until 2 hrs. 15 m. had elapsed. The same result followed in between 1 hr. 15 m. and 1 hr. 30 m. when particles both of cartilage and meat were placed on several glands. Lastly, a minute drop (about 1/20 of a minim) of a solution of one part of nitrate of ammonia to 146 of water was distributed between the secretion surrounding three glands, so that the amount of fluid surrounding each was slightly increased; yet when looked at after 2 hrs., all three were dry. On the other hand, seven particles of glass and three of coal-cinders, of nearly the same size as those of the above named organic substances, were placed on ten glands; some of them being observed for 18 hrs., and others for two or three days; but there was not the least sign of the secretion being absorbed. Hence, in the former cases, the absorption of the secretion must have been due to the presence of some nitrogenous matter, which was either already soluble or was rendered so by the secretion. As the fibrin was pure, and had been well washed in distilled water after being kept in glycerine, and as the cartilage had been soaked in water, I suspect that these substances must [page 337] have been slightly acted on and rendered soluble within the above stated short periods.

The glands have not only the power of rapid absorption, but likewise of secreting again quickly; and this latter habit has perhaps been gained, inasmuch as insects, if they touch the glands, generally withdraw the drops of secretion, which have to be restored. The exact period of re-secretion was recorded in only a few cases. The glands on which bits of meat were placed, and which were nearly dry after about 1 hr. 30 m., when looked at after 22 additional hours, were found secreting; so it was after 24 hrs. with one gland on which a bit of albumen had been placed. The three glands to which a minute drop of a solution of nitrate of ammonia was distributed, and which became dry after 2 hrs., were beginning to re-secrete after only 12 additional hours.

Tentacles Incapable of Movement.—Many of the tall tentacles, with insects adhering to them, were carefully observed; and fragments of insects, bits of raw meat, albumen, &c., drops of a solution of two salts of ammonia and of saliva, were placed on the glands of many tentacles; but not a trace of movement could ever be detected. I also repeatedly irritated the glands with a needle, and scratched and pricked the blades, but neither the blade nor the tentacles became at all inflected. We may therefore conclude that they are incapable of movement.

On the Power of Absorption possessed by the Glands.—It has already been indirectly shown that the glands on pedicels absorb animal matter; and this is further shown by their changed colour, and by the aggregation of their contents, after they have been left in contact with nitrogenous substances or liquids. The following observations apply both to the glands supported on [page 338] pedicels and to the minute sessile ones. Before a gland has been in any way stimulated, the exterior cells commonly contain only limpid purple fluid; the more central ones including mulberry-like masses of purple granular matter. A leaf was placed in a little solution of one part of carbonate of ammonia to 146 of water (3 grs. to 1 oz.), and the glands were instantly darkened and very soon became black; this change being due to the strongly marked aggregation of their contents, more especially of the inner cells. Another leaf was placed in a solution of the same strength of nitrate of ammonia, and the glands were slightly darkened in 25 m., more so in 50 m., and after 1 hr. 30 m. were of so dark a red as to appear almost black. Other leaves were placed in a weak infusion of raw meat and in human saliva, and the glands were much darkened in 25 m., and after 40 m. were so dark as almost to deserve to be called black. Even immersion for a whole day in distilled water occasionally induces some aggregation within the glands, so that they become of a darker tint. In all these cases the glands are affected in exactly the same manner as those of Drosera. Milk, however, which acts so energetically on Drosera, seems rather less effective on Drosophyllum, for the glands were only slightly darkened by an immersion of 1 hr. 20 m., but became decidedly darker after 3 hrs. Leaves which had been left for 7 hrs. in an infusion of raw meat or in saliva were placed in the solution of carbonate of ammonia, and the glands now became greenish; whereas, if they had been first placed in the carbonate, they would have become black. In this latter case, the ammonia probably combines with the acid of the secretion, and therefore does not act on the colouring matter; but when the glands are first subjected to an organic [page 339] fluid, either the acid is consumed in the work of digestion or the cell-walls are rendered more permeable, so that the undecomposed carbonate enters and acts on the colouring matter. If a particle of the dry carbonate is placed on a gland, the purple colour is quickly discharged, owing probably to an excess of the salt. The gland, moreover, is killed.

Turning now to the action of organic substances, the glands on which bits of raw meat were placed became dark-coloured; and in 18 hrs. their contents were conspicuously aggregated. Several glands with bits of albumen and fibrin were darkened in between 2 hrs. and 3 hrs.; but in one case the purple colour was completely discharged. Some glands which had caught flies were compared with others close by; and though they did not differ much in colour, there was a marked difference in their state of aggregation. In some few instances, however, there was no such difference, and this appeared to be due to the insects having been caught long ago, so that the glands had recovered their pristine state. In one case, a group of the sessile colourless glands, to which a small fly adhered, presented a peculiar appearance; for they had become purple, owing to purple granular matter coating the cell-walls. I may here mention as a caution that, soon after some of my plants arrived in the spring from Portugal, the glands were not plainly acted on by bits of meat, or insects, or a solution of ammonia—a circumstance for which I cannot account.

Digestion of Solid Animal Matter.—Whilst I was trying to place on two of the taller glands little cubes of albumen, these slipped down, and, besmeared with secretion, were left resting on some of the small sessile glands. After 24 hrs. one of these cubes was found [page 340] completely liquefied, but with a few white streaks still visible; the other was much rounded, but not quite dissolved. Two other cubes were left on tall glands for 2 hrs. 45 m., by which time all the secretion was absorbed; but they were not perceptibly acted on, though no doubt some slight amount of animal matter had been absorbed from them. They were then placed on the small sessile glands, which being thus stimulated secreted copiously in the course of 7 hrs. One of these cubes was much liquefied within this short time; and both were completely liquefied after 21 hrs. 15 m.; the little liquid masses, however, still showing some white streaks. These streaks disappeared after an additional period of 6 hrs. 30 m.; and by next morning (i.e. 48 hrs. from the time when the cubes were first placed on the glands) the liquefied matter was wholly absorbed. A cube of albumen was left on another tall gland, which first absorbed the secretion and after 24 hrs. poured forth a fresh supply. This cube, now surrounded by secretion, was left on the gland for an additional 24 hrs., but was very little, if at all, acted on. We may, therefore, conclude, either that the secretion from the tall glands has little power of digestion, though strongly acid, or that the amount poured forth from a single gland is insufficient to dissolve a particle of albumen which within the same time would have been dissolved by the secretion from several of the small sessile glands. Owing to the death of my last plant, I was unable to ascertain which of these alternatives is the true one.

Four minute shreds of pure fibrin were placed, each resting on one, two, or three of the taller glands. In the course of 2 hrs. 30 m. the secretion was all absorbed, and the shreds were left almost dry. They [page 341] were then pushed on to the sessile glands. One shred, after 2 hrs. 30 m., seemed quite dissolved, but this may have been a mistake. A second, when examined after 17 hrs. 25 m., was liquefied, but the liquid as seen under the microscope still contained floating granules of fibrin. The other two shreds were completely liquefied after 21 hrs. 30 m.; but in one of the drops a very few granules could still be detected. These, however, were dissolved after an additional interval of 6 hrs. 30 m.; and the surface of the leaf for some distance all round was covered with limpid fluid. It thus appears that Drosophyllum digests albumen and fibrin rather more quickly than Drosera can; and this may perhaps be attributed to the acid, together probably with some small amount of the ferment, being present in the secretion, before the glands have been stimulated; so that digestion begins at once.

Concluding Remarks.—The linear leaves of Drosophyllum differ but slightly from those of certain species of Drosera; the chief differences being, firstly, the presence of minute, almost sessile, glands, which, like those of Dionaea, do not secrete until they are excited by the absorption of nitrogenous matter. But glands of this kind are present on the leaves of Drosera binata, and appear to be represented by the papillae on the leaves of Drosera rotundifolia. Secondly, the presence of tentacles on the backs of the leaves; but we have seen that a few tentacles, irregularly placed and tending towards abortion, are retained on the backs of the leaves of Drosera binata. There are greater differences in function between the two genera. The most important one is that the tentacles of Drosophyllum have no power of movement; this loss being partially replaced by the drops of viscid [page 342] secretion being readily withdrawn from the glands; so that, when an insect comes into contact with a drop, it is able to crawl away, but soon touches other drops, and then, smothered by the secretion, sinks down on the sessile glands and dies. Another difference is, that the secretion from the tall glands, before they have been in any way excited, is strongly acid, and perhaps contains a small quantity of the proper ferment. Again, these glands do not secrete more copiously from being excited by the absorption of nitrogenous matter; on the contrary, they then absorb their own secretion with extraordinary quickness. In a short time they begin to secrete again. All these circumstances are probably connected with the fact that insects do not commonly adhere to the glands with which they first come into contact, though this does sometimes occur; and that it is chiefly the secretion from the sessile glands which dissolves animal matter out of their bodies.

RORIDULA.

Roridula dentata.—This plant, a native of the western parts of the Cape of Good Hope, was sent to me in a dried state from Kew. It has an almost woody stem and branches, and apparently grows to a height of some feet. The leaves are linear, with their summits much attenuated. Their upper and lower surfaces are concave, with a ridge in the middle, and both are covered with tentacles, which differ greatly in length; some being very long, especially those on the tips of the leaves, and some very short. The glands also differ much in size and are somewhat elongated. They are supported on multicellular pedicels.

This plant, therefore, agrees in several respects with [page 343] Drosophyllum, but differs in the following points. I could detect no sessile glands; nor would these have been of any use, as the upper surface of the leaves is thickly clothed with pointed, unicellular hairs directed upwards. The pedicels of the tentacles do not include spiral vessels; nor are there any spiral cells within the glands. The leaves often arise in tufts and are pinnatifid, the divisions projecting at right angles to the main linear blade. These lateral divisions are often very short and bear only a single terminal tentacle, with one or two short ones on the sides. No distinct line of demarcation can be drawn between the pedicels of the long terminal tentacles and the much attenuated summits of the leaves. We may, indeed, arbitrarily fix on the point to which the spiral vessels proceeding from the blade extend; but there is no other distinction.

It was evident from the many particles of dirt sticking to the glands that they secrete much viscid matter. A large number of insects of many kinds also adhered to the leaves. I could nowhere discover any signs of the tentacles having been inflected over the captured insects; and this probably would have been seen even in the dried specimens, had they possessed the power of movement. Hence, in this negative character, Roridula resembles its northern representative, Drosophyllum.

BYBLIS.

Byblis gigantea (Western Australia).—A dried specimen, about 18 inches in height, with a strong stem, was sent me from Kew. The leaves are some inches in length, linear, slightly flattened, with a small projecting rib on the lower surface. They are covered on all sides by glands of two kinds [page 344] —sessile ones arranged in rows, and others supported on moderately long pedicels. Towards the narrow summits of the leaves the pedicels are longer than elsewhere, and here equal the diameter of the leaf. The glands are purplish, much flattened, and formed of a single layer of radiating cells, which in the larger glands are from forty to fifty in number. The pedicels consist of single elongated cells, with colourless, extremely delicate walls, marked with the finest intersecting spiral lines. Whether these lines are the result of contraction from the drying of the walls, I do not know, but the whole pedicel was often spirally rolled up. These glandular hairs are far more simple in structure than the so-called tentacles of the preceding genera, and they do not differ essentially from those borne by innumerable other plants. The flower-peduncles bear similar glands. The most singular character about the leaves is that the apex is enlarged into a little knob, covered with glands, and about a third broader than the adjoining part of the attenuated leaf. In two places dead flies adhered to the glands. As no instance is known of unicellular structures having any power of movement,* Byblis, no doubt, catches insects solely by the aid of its viscid secretion. These probably sink down besmeared with the secretion and rest on the small sessile glands, which, if we may judge by the analogy of Drosophyllum, then pour forth their secretion and afterwards absorb the digested matter.

Supplementary Observations on the Power of Absorption by the Glandular Hairs of other Plants.—A few observations on this subject may be here conveniently introduced. As the glands of many, probably of all,

* Sachs, 'Trait de Bot.,' 3rd edit. 1874, p. 1026. [page 345]

the species of Droseraceae absorb fluids or at least allow them readily to enter,* it seemed desirable to ascertain how far the glands of other plants which are not specially adapted for capturing insects, had the same power. Plants were chosen for trial at hazard, with the exception of two species of saxifrage, which were selected from belonging to a family allied to the Droseraceae. Most of the experiments were made by immersing the glands either in an infusion of raw meat or more commonly in a solution of carbonate of ammonia, as this latter substance acts so powerfully and rapidly on protoplasm. It seemed also particularly desirable to ascertain whether ammonia was absorbed, as a small amount is contained in rain-water. With the Droseraceae the secretion of a viscid fluid by the glands does not prevent their absorbing; so that the glands of other plants might excrete superfluous matter, or secrete an odoriferous fluid as a protection against the attacks of insects, or for any other purpose, and yet have the power of absorbing. I regret that in the following cases I did not try whether the secretion could digest or render soluble animal substances, but such experiments would have been difficult on account of the small size of the glands and the small amount of secretion. We shall see in the next chapter that the secretion from the glandular hairs of Pinguicula certainly dissolves animal matter.

[Saxifraga umbrosa.—The flower-peduncles and petioles of the leaves are clothed with short hairs, bearing pink-coloured glands, formed of several polygonal cells, with their pedicels divided by partitions into distinct cells, which are generally colourless, but sometimes pink. The glands secrete a yellowish viscid fluid, by

*The distinction between true absorption and mere permeation, or imbibition, is by no means clearly understood: see Mller's 'Physiology,' Eng. translat. 1838, vol. i. p. 280. [page 346]

which minute Diptera are sometimes, though not often, caught.* The cells of the glands contain bright pink fluid, charged with granules or with globular masses of pinkish pulpy matter. This matter must be protoplasm, for it is seen to undergo slow but incessant changes of form if a gland be placed in a drop of water and examined. Similar movements were observed after glands had been immersed in water for 1, 3, 5, 18, and 27 hrs. Even after this latter period the glands retained their bright pink colour; and the protoplasm within their cells did not appear to have become more aggregated. The continually changing forms of the little masses of protoplasm are not due to the absorption of water, as they were seen in glands kept dry.

A flower-stem, still attached to a plant, was bent (May 29) so as to remain immersed for 23 hrs. 30 m. in a strong infusion of raw meat. The colour of the contents of the glands was slightly changed, being now of a duller and more purple tint than before. The contents also appeared more aggregated, for the spaces between the little masses of protoplasm were wider; but this latter result did not follow in some other and similar experiments. The masses seemed to change their forms more rapidly than did those in water; so that the cells had a different appearance every four or five minutes. Elongated masses became in the course of one or two minutes spherical; and spherical ones drew themselves out and united with others. Minute masses rapidly increased in size, and three distinct ones were seen to unite. The movements were, in short, exactly like those described in the case of Drosera. The cells of the pedicels were not affected by the infusion; nor were they in the following experiment.

Another flower-stem was placed in the same manner and for the same length of time in a solution of one part of nitrate of ammonia to 146 of water (or 3 grs. to 1 oz.), and the glands were discoloured in exactly the same manner as by the infusion of raw meat.

Another flower-stem was immersed, as before, in a solution of one part of carbonate of ammonia to 109 of water. The glands, after 1 hr. 30 m., were not discoloured, but after 3 hrs. 45 m. most of them had become dull purple, some of them blackish-

*In the case of Saxifraga tridactylites, Mr. Druce says ('Pharmaceutical Journal, ' May 1875) that he examined some dozens of plants, and in almost every instance remnants of insects adhered to the leaves. So it is, as I hear from a friend, with this plant in Ireland. [page 347]

green, a few being still unaffected. The little masses of protoplasm within the cells were seen in movement. The cells of the pedicels were unaltered. The experiment was repeated, and a fresh flower-stem was left for 23 hrs. in the solution, and now a great effect was produced; all the glands were much blackened, and the previously transparent fluid in the cells of the pedicels, even down to their bases, contained spherical masses of granular matter. By comparing many different hairs, it was evident that the glands first absorb the carbonate, and that the effect thus produced travels down the hairs from cell to cell. The first change which could be observed is a cloudy appearance in the fluid, due to the formation of very fine granules, which afterwards aggregate into larger masses. Altogether, in the darkening of the glands, and in the process of aggregation travelling down the cells of the pedicels, there is the closest resemblance to what takes place when a tentacle of Drosera is immersed in a weak solution of the same salt. The glands, however, absorb very much more slowly than those of Drosera. Besides the glandular hairs, there are star-shaped organs which do not appear to secrete, and which were not in the least affected by the above solutions.

Although in the case of uninjured flower-stems and leaves the carbonate seems to be absorbed only by the glands, yet it enters a cut surface much more quickly than a gland. Strips of the rind of a flower-stem were torn off, and the cells of the pedicels were seen to contain only colourless transparent fluid; those of the glands including as usual some granular matter. These strips were then immersed in the same solution as before (one part of the carbonate to 109 of water), and in a few minutes granular matter appeared in the lowercells of all the pedicels. The action invariably commenced (for I tried the experiment repeatedly) in the lowest cells, and therefore close to the torn surface, and then gradually travelled up the hairs until it reached the glands, in a reversed direction to what occurs in uninjured specimens. The glands then became discoloured, and the previously contained granular matter was aggregated into larger masses. Two short bits of a flower-stem were also left for 2 hrs. 40 m. in a weaker solution of one part of the carbonate to 218 of water; and in both specimens the pedicels of the hairs near the cut ends now contained much granular matter; and the glands were completely discoloured.

Lastly, bits of meat were placed on some glands; these were examined after 23 hrs., as were others, which had apparently not long before caught minute flies; but they did not present any [page 348] difference from the glands of other hairs. Perhaps there may not have been time enough for absorption. I think so as some glands, on which dead flies had evidently long lain, were of a pale dirty purple colour or even almost colourless, and the granular matter within them presented an unusual and somewhat peculiar appearance. That these glands had absorbed animal matter from the flies, probably by exosmose into the viscid secretion, we may infer, not only from their changed colour, but because, when placed in a solution of carbonate of ammonia, some of the cells in their pedicels become filled with granular matter; whereas the cells of other hairs, which had not caught flies, after being treated with the same solution for the same length of time, contained only a small quantity of granular matter. But more evidence is necessary before we fully admit that the glands of this saxifrage can absorb, even with ample time allowed, animal matter from the minute insects which they occasionally and accidentally capture.

Saxifraga rotundifolia (?).—The hairs on the flower-stems of this species are longer than those just described, and bear pale brown glands. Many were examined, and the cells of the pedicels were quite transparent. A bent stem was immersed for 30 m. in a solution of one part of carbonate of ammonia to 109 of water, and two or three of the uppermost cells in the pedicels now contained granular or aggregated matter; the glands having become of a bright yellowish-green. The glands of this species therefore absorb the carbonate much more quickly than do those of Saxifraga umbrosa, and the upper cells of the pedicels are likewise affected much more quickly. Pieces of the stem were cut off and immersed in the same solution; and now the process of aggregation travelled up the hairs in a reversed direction; the cells close to the cut surfaces being first affected.

Primula sinensis.—The flower-stems, the upper and lower surfaces of the leaves and their footstalks, are all clothed with a multitude of longer and shorter hairs. The pedicels of the longer hairs are divided by transverse partitions into eight or nine cells. The enlarged terminal cell is globular, forming a gland which secretes a variable amount of thick, slightly viscid, not acid, brownish-yellow matter.

A piece of a young flower-stem was first immersed in distilled water for 2 hrs. 30 m., and the glandular hairs were not at all affected. Another piece, bearing twenty-five short and nine long hairs, was carefully examined. The glands of the latter contained no solid or semi-solid matter; and those of only two [page 349] of the twenty-five short hairs contained some globules. This piece was then immersed for 2 hrs. in a solution of one part of carbonate of ammonia to 109 of water, and now the glands of the twenty-five shorter hairs, with two or three exceptions, contained either one large or from two to five smaller spherical masses of semi-solid matter. Three of the glands of the nine long hairs likewise included similar masses. In a few hairs there were also globules in the cells immediately beneath the glands. Looking to all thirty-four hairs, there could be no doubt that the glands had absorbed some of the carbonate. Another piece was left for only 1 hr. in the same solution, and aggregated matter appeared in all the glands. My son Francis examined some glands of the longer hairs, which contained little masses of matter, before they were immersed in any solution; and these masses slowly changed their forms, so that no doubt they consisted of protoplasm. He then irrigated these hairs for 1 hr. 15 m., whilst under the microscope, with a solution of one part of the carbonate to 218 of water; the glands were not perceptibly affected, nor could this have been expected, as their contents were already aggregated. But in the cells of the pedicels numerous, almost colourless, spheres of matter appeared, which changed their forms and slowly coalesced; the appearance of the cells being thus totally changed at successive intervals of time.

The glands on a young flower-stem, after having been left for 2 hrs. 45 m. in a strong solution of one part of the carbonate to 109 of water, contained an abundance of aggregated masses, but whether generated by the action of the salt, I do not know. This piece was again placed in the solution, so that it was immersed altogether for 6 hrs. 15 m., and now there was a great change; for almost all the spherical masses within the gland-cells had disappeared, being replaced by granular matter of a darker brown. The experiment was thrice repeated with nearly the same result. On one occasion the piece was left immersed for 8 hrs. 30 m., and though almost all the spherical masses were changed into the brown granular matter, a few still remained. If the spherical masses of aggregated matter had been originally produced merely by some chemical or physical action, it seems strange that a somewhat longer immersion in the same solution should so completely alter their character. But as the masses which slowly and spontaneously changed their forms must have consisted of living protoplasm, there is nothing surprising in its being injured or killed, and its appearance wholly changed by long immersion in so strong a solution of the carbonate as that [page 350] employed. A solution of this strength paralyses all movement in Drosera, but does not kill the protoplasm; a still stronger solution prevents the protoplasm from aggregating into the ordinary full-sized globular masses, and these, though they do not disintegrate, become granular and opaque. In nearly the same manner, too hot water and certain solutions (for instance, of the salts of soda and potash) cause at first an imperfect kind of aggregation in the cells of Drosera; the little masses afterwards breaking up into granular or pulpy brown matter. All the foregoing experiments were made on flower-stems, but a piece of a leaf was immersed for 30 m. in a strong solution of the carbonate (one part to 109 of water), and little globular masses of matter appeared in all the glands, which before contained only limpid fluid.