I may here state, in passing, a point of some little interest in connection with this reinforcing action of lithocysts. When I first observed this action, it appeared to me a mysterious thing why its result was always to propagate the contraction-wave in only one direction—the direction, namely, in which the wave happened to be passing before it reached the lithocyst. For instance, suppose we have a strip A D, with a lithocyst at each of the equidistant points A, B, C, D; and suppose that the lithocyst B originates a stimulus: the resulting contraction-wave passes, of course, with equal rapidity in the two opposite directions, B A, B C (arrows b a, b c): the contraction-wave b a therefore reaches the lithocyst A at the same time as the contraction-wave b c reaches the lithocyst C, and so both A and C discharge simultaneously. What, then, should we expect to be the result? I think we should expect the wave b c to continue on its course to D, after having been strengthened at C, and a reflex wave a´ b´ to start from A (owing to the discharge at A), which would reach B at the same time as a similar reflex wave c´ b´ starting from C (owing to the discharge at C); so that by the time the original wave b c d had reached D, the point B would be the seat of a collision between the two reflex waves a´ b´ and c´ d´. And, not to push the supposed case further, it is evident that if such reflex waves were to occur, the resulting confusion would very soon require to end in tetanus. As a matter of fact, these reflex waves do not occur; and the question is, why do they not? Why is it that a wave is only reinforced in the direction in which it happens to be travelling—so that if, for instance, it happens to start from A in the above series, it is successively propagated by B C in the direction A, B, C, D, and in that direction only; whereas, if it happens to start from D, it is propagated by the same lithocysts in the opposite direction, D, C, B, A, and in that direction only—the wave in the one case terminating at the lithocyst D, and in the other case at the lithocyst A? Now, although this absence of reflex waves appears at first sight mysterious, it admits of an exceedingly simple explanation. I find that the contractile tissues of the covered-eyed Medusæ cannot be made to respond to two successive stimuli of minimal, or but slightly more than minimal intensity, unless such stimuli are separated from one another by a certain considerable interval of time. Now, when in the above illustration the contraction-wave starts from A, by the time it reaches B the portion of tissue included between A and B has just been in contraction in response to the stimulus from A, while the portion of tissue included between B and C has not been in contraction. Consequently, the stimulus resulting from a ganglionic discharge being presumably of minimal, or but slightly more than minimal intensity, the tissue included between A and B will not respond to the discharge of B; while the tissue included between B and C, not having been just previously in contraction, will respond. And conversely, of course, if the contraction-wave had been travelling in the opposite direction.

Seeing that this explanation is the only one possible, and that it moreover follows as a deductive necessity from my experiments on stimulation, I think there is no need to detail any of the further experiments which I made with the view of confirming it. But the following experiment, devised to confirm this explanation, is of interest in itself, and on this account I shall state it. Having prepared a contractile strip with a single remaining lithocyst at one end, I noted the rhythm exhibited by this lithocyst, and then imitated that rhythm by means of single induced shocks thrown in with a key at the other end of the strip. The effect of these shocks was, of course, to cause the contraction-waves to pass in the direction opposite to that in which they passed when originated by the lithocyst. Now I found, as I had expected, that so long as I continued exactly to imitate the rate of ganglionic rhythm, so long did the waves always pass in the direction B A—A being the lithocyst, and B the other end of the strip. I also found that if I allowed the rate of the artificially caused rhythm to sink slightly below that of the natural rhythm, after every one to six waves (the number depending on the degree in which the rate of succession of my induction shocks approximated to the rate of the natural rhythm) which passed from B to A, one would pass from A to B.[18]

Of course the only interpretation to be put on these facts is that every time an artificially started wave reached the terminal ganglion it caused the latter to discharge; but that the occurrence of a discharge could not in this case be rendered apparent, because of the inadequacy of that discharge to start a reflex wave. But that such discharges always took place was manifest, both à priori because from analogy we may be sure that if there had happened to be any contractile tissue of appropriate width on the other side of the ganglion, the discharge of the latter would have been rendered apparent, and à posteriori because, after the arrival of every artificially started wave, the time required for the ganglion to originate another wave was precisely the same as if it had itself originated the previous wave.

In view of these results, it occurred to me as an interesting experiment to try the effect on the natural rhythm of exhausting a ganglion thus situated, by throwing in a great number of shocks at the other end of the strip. I found that after five hundred single shocks had been thrown in with a rapidity almost sufficient to tetanize the strip, immediately after the stimulation ceased, the natural rhythm of the ganglion, which had previously been twenty in the minute, fell to fourteen for the first minute, eighteen for the second, and the original rate of twenty for the third. In such experiments the diminution of rate is most conspicuous during the first fifteen or thirty seconds of the first minute. Sometimes there are no contractions at all for the first fifteen seconds after cessation of the stimulating process, and in such cases the natural rhythm, when it first begins, may be as slow as one-half or even one-quarter its normal rate. All these effects admit of being produced equally well, and with less trouble, by faradizing the strip, when it may be even better observed how prolonged may be the stimulation, without causing anything further than such slight exhaustion of the ganglion as the above results imply.[19]

It would be impossible to imagine movements on the part of so simple an organism more indicative of physiological harmony than are the movements of Sarsia. One may watch several hundreds of these animals while they are swimming about in the same bell-jar and never perceive, as in the covered-eyed Medusæ, the slightest want of ganglionic co-ordination exhibited by any of the specimens. Moreover, that the ganglionic co-ordination is in this case wonderfully far advanced is proved by the fact of members of this genus being able to steer themselves while following a light, as previously described.[20]

In the discophorous species of naked-eyed Medusæ, however, perfectly co-ordinated action is by no means of such invariable occurrence as it is in Sarsia; for although in perfectly healthy and vigorous specimens systole and diastole occur at the same instant over the whole nectocalyx, this harmoniously acting mechanism is very liable to be thrown out of gear, so that when the animals are suffering in the least degree from any injurious conditions, often too slight and obscure to admit of discernment, the swimming movements are no longer synchronous over the whole nectocalyx; but now one part is in systole while another part is in diastole, and now several parts may be in diastole while other parts are in systole. And as in these animals very slight causes seem sufficient thus to impair the ganglionic co-ordination, it generally happens that in a bell-jar containing a number of specimens belonging to different species, numerous examples of more or less irregular swimming movements are observable.

Taking, then, the case of Sarsia first, from my previous observations on the physiological harmony subsisting between the tentacles, I was led to expect that the co-ordination of the locomotor ganglia was probably effected by means of the same tissue-tracts through which the intertentacular harmony was effected, namely, those situated in the margin of the bell. Accordingly, I introduced four short radial cuts, one midway between each pair of adjacent marginal bodies. The co-ordination, however, was not perceptibly impaired. I therefore continued the radial cuts, and found that when these reached one-half or two-thirds of the way up the sides of the inner bell (or contractile sheet), the co-ordination became visibly affected, and this for the first time.

I also tried the following experiment. Instead of beginning the radial cuts from the margin, I began them from the apex of the cone; and I found that however many of such cuts I introduced, and however far down the cone I carried them, so long as I did not actually sever the margin, so long did all the divisions of the bell continue to contract simultaneously.[21] This fact, therefore, proves that the margin of the bell is alone sufficient to maintain co-ordination.

The next experiment I tried was to make four short radial incisions in the margin as before described, and then to continue one of these incisions the whole way up the bell. By careful observation I could now perceive that all the marginal ganglia did not discharge simultaneously; for when those situated nearest to the long radial cut happened to take the initiative, the resulting contraction-wave, having double the distance to travel which it would have had if the long radial cut had been absent, could now be followed by the eye in its very rapid course round the bell. Now, the fact that in this form of section I was able to detect the passage of a wave, proves that the three short radial sections had destroyed the co-ordinated action of the marginal ganglia.

From these experiments, then, I conclude that in this genus ganglionic co-ordination, in the strict sense of the term, is effected exclusively by means of the marginal nerves. And as these experiments on Sarsia are exceedingly difficult to conduct, owing to the very rapid passage of contraction-waves in this genus, it is satisfactory to find that this conclusion is further supported by the analogy which the other species of naked-eyed Medusæ afford, and to the consideration of which we shall now proceed.

The effects of four short radial incisions through the margin of any species of Tiaropsis, Thaumantias, Staurophora, etc., are usually very conspicuous. Each of the quadrants included between two adjacent incisions shows a strong tendency to assume an independent action of its own. This tendency is sometimes so pronounced as to amount almost to a total destruction of contractional continuity between two or more quadrants of the bell; but more usually the effect of the marginal sections is merely that of destroying excitational continuity, or at least physiological harmony.

It is an interesting thing that this form of section, although in actual amount so very slight, is attended with a much more pernicious influence on the vitality of the organism than is any amount of section of the general contractile tissues. Thus, if a specimen of Tiaropsis, for example, be chosen which is swimming about with the utmost vigour, and if four equidistant radial cuts only just long enough to sever the marginal canal be made, the animal will soon begin to show symptoms of enfeeblement, and within an hour or two after the operation will probably have ceased its swimming motions altogether. The animal, however, is not actually dead; for if while lying motionless at the bottom of the vessel it be gently stimulated, it will respond with a spasm as usual, and perhaps immediately afterwards give a short and feeble bout of swimming movements. These surprisingly pernicious results are not so conspicuous in the case of Sarsia, although in this genus likewise they are sufficiently well marked to be unmistakable. I here append a table to show the comparative effects of the operation in question on different species. The cases may be regarded as very usual ones, though it often happens that a longer time after the operation must elapse before the enfeebling effects become so pronounced.

This decided effect of so slight a mutilation will not, perhaps, appear to other physiologists so noteworthy as it appears to me; for no one who has not witnessed the experiments can form an adequate idea of the amount of mutilation of any parts, other than their margins, which the Medusæ will endure without even suffering from the effects of shock. Another point worth mentioning with regard to the operation we are considering is, that not unfrequently the interruptions of the margin, which have been produced artificially, begin to extend themselves through the nectocalyx in a radial direction; so that in some cases this organ becomes spontaneously segmented into four quadrants, which remain connected only by the apical tissue of the bell. I do not think that this is due to the mere mechanical tearing of the tissues as a consequence of the swimming motions, for the latter seem too feeble to admit of their producing such an effect.

In conclusion, I may state that I have been able temporarily to destroy the ganglionic co-ordination of Sarsiæ, by submitting the animals to severe nervous shock. The method I employed to produce the nervous shock, without causing mutilation, was to take the animal out of the water for a few seconds while I laid it on a small anvil, which I then struck violently with a hammer. On immediately afterwards restoring the Medusa to sea-water, spontaneity was found to have ceased, while irritability remained. After a time spontaneity began to return, and its first stages were marked by a complete want of co-ordination; soon, however, co-ordination was again restored. But this experiment by no means invariably yielded the same result. Spontaneity, indeed, was invariably suspended for a time; but its first return was not invariably, or even generally, marked by an absence of co-ordination, even though I had previously struck the anvil a number of times in succession. I was therefore led to try another method of producing nervous shock, and this I found a more effectual method than the one just described. It consisted in violently shaking the Sarsiæ in a bottle half filled with sea-water. I was surprised to find how violent and prolonged such shaking might be without any part of the apparently friable organism, except perhaps the tentacles and manubrium, being broken or torn. The subsequent effects of shock were remarkable. For some little time after their restoration to the bell-jar, the Sarsiæ had lost, not only their spontaneity, but also their irritability, for they would not respond even to the strongest stimulation. In the course of a few minutes, however, peripheral irritability returned, as shown by responses to nipping of the neuro-muscular sheet. The animals were now in the same condition as when anæsthesiated by caffein or other central nerve-poison; but in a few minutes later central or reflex irritability also returned, as shown by single responses to single nippings of the tentacles. Last of all spontaneity began to return, and was in some few cases conspicuously marked by a want of co-ordination, all parts of the margin originating impulses at different times, with the result of producing a continuous flurried or shivering movement of the nectocalyx. After a time, however, these movements became co-ordinated; but in most cases when a swimming bout had ended and a pause intervened, the next swimming bout was also inaugurated by a period of shivering before co-ordination became established. This effect might last for a long time, but eventually it, too, disappeared, the swimming bouts then beginning with co-ordinated action in the usual way.

Dr. Eimer also affirms that the duration of the natural pauses, which in Aurelia habitually alternate with bouts of swimming, bears a direct proportion to the number and strength of the contractions that occurred in the previous bout of swimming. I observed that Sarsiæ are much better adapted than Aureliæ for determining whether any such precise relation obtains; for, in the first place, the strength of the contraction is more uniform, and, in the next place, the alternation of pauses with bouts of swimming is of a more decided character in Sarsiæ than in healthy specimens of Aureliæ. I further observed that in Sarsia no such precise relation did obtain, although in a very general way it is true, as might be expected, that unusually prolonged bouts of swimming were sometimes followed by pauses of unusual duration. As all the observations are very much the same, I shall only quote two of them:—

These observations may be taken as samples of others which it would be unnecessary to quote, as it will be seen from the above that there is no precise relation between the number of the pulsations and the duration of the pauses. Nevertheless, that there is a general relation may be seen from some cases in which unusually prolonged pauses occur. The following instance will serve to show this:—

In this case, the relation between the long pause of 380 seconds and the subsequent prolonged swimming bout of 112 pulsations is obvious; also, as the latter was then followed by a short pause of twenty seconds and another comparatively short bout of forty-five pulsations, the refreshing influence of the previous 380 seconds rest may be supposed to have been not quite neutralized by the exhausting effect of the foregoing 112 pulsations. At any rate, looking to the general nature of the previous proportions (viz. in their sum 185/211), it is certain that 380/112 leaves a large preponderance in favour of nutrition, which preponderance is not much modified by adding the next succeeding proportion, thus, (380 + 20)/(112 + 45) = 400/157. Consequently, the organism may fairly be supposed to have entered upon the next prolonged period of rest (viz. 185 seconds) with a large balance of reserve power; so that when to this large balance there was added the further accumulation due to the further rest of 185 seconds, we are not surprised to find the next succeeding swimming bout comprising the enormous number of 894 pulsations. But this great expenditure of energy seems to have been somewhat in excess of the energy previously accumulated by the prolonged rest, for this unusual expenditure seems next to have entailed an unusually prolonged period of exhaustion. At any rate, it is plainly observable that the next succeeding proportions are greatly in favour of repose; for it is not until 360 seconds have elapsed, with only twelve pulsations in the interval, that energy enough has been accumulated to cause a moderate bout of thirty pulsations. But next another long and sustained pause of 240 seconds supervenes, and, the animal being now fully refreshed with a large surplus of accumulated energy, the next succeeding swimming bout comprises two hundred pulsations. Lastly, there succeeded sixty seconds of rest, and here the observation terminated.[22]

We have next to consider Dr. Eimer's observations concerning the effects on the rhythm of Aurelia which result on cutting the animal into segments; and here, again, I much regret to say that I cannot wholly agree with this author. He says he found evidence of a very remarkable fact, viz. that by first counting the natural rhythm of an unmutilated Aurelia, and then dividing the animal into two halves, one of these halves into two quarters, and one of these quarters into two eighths; the sum of the contractions performed by these four segments in a given time was equal to the number which had previously been performed in a similar time by the unmutilated animal. And not only so, but the number of contractions which each segment contributed to this sum was a number that stood in direct proportion to the size of the segment; so that the half contracted half as many times, the quarter a quarter as many times, and the eighth parts one-eighth part the number of times that the unmutilated Aurelia had previously contracted in a period of equal duration. I am glad to observe that Dr. Eimer does not regard this rule otherwise than as liable to frequent exception; for, as already observed, I cannot say that my experiments have tended to confirm it. I am only able to say that there is general tendency for the smaller segments of an Aurelia divided in this way to contract less frequently than the larger segments.

It would be tedious and unnecessary to quote any observations in this connection; but as these observations brought out very clearly a fact which I had previously suspected, I may detail one experiment to illustrate this point. The fact in question is, that the potency of the lithocysts in any given segment of a divided Aurelia has more to do with the frequency of its pulsations than has the size of the segment. As previously mentioned, one or more lithocysts may often be observed to be permanently prepotent over the others; and I may here observe that the segmentation experiments just described have shown the converse to be true, viz. that one or more lithocysts are often permanently feebler than the others. Well, if a specimen of Aurelia exhibiting decided prepotency in one or more of its lithocysts be watched for a considerable length of time, so as to be sure that the prepotency is not of a merely temporary character, and if the animal be then divided into segments in such a way that the prepotent lithocysts shall occupy the smaller segments, it may be observed, provided time be left for the tissues to recover, that the segments containing the prepotent lithocysts, notwithstanding their smaller size, contract more frequently than do the larger segments. Conversely, if the larger segments happen to contain feeble lithocysts, their contractions will be but few. I have, indeed, seen cases in which the lithocysts appeared to be quite functionless, so far as the origination of stimuli was concerned.

The following observations were made on a healthy specimen of Aurelia having all its lithocysts in good condition, but prepotency being well marked in the case of one of them, and also, though in a lesser degree, in the case of another. I divided the animal so as to leave one of these two prepotent lithocysts in each of the eighth-part segments, and the next most powerful lithocysts in the quadrant segment. In the following description, I shall call the two eighth-part segments A and B, the former letter designating the segment containing the most powerful lithocyst. The Aurelia before being divided manifested for several hours a very regular and sustained rhythm of thirty-two per minute. After its division, the various segments contracted at the following rates, in one-minute intervals:—

Next morning, the water which contained the segments was somewhat foul, and this, as is always the case, gave rise to abnormally long pauses. This effect was much more marked in the case of some of the segments than in that of others. I therefore observed the segments over five-minute intervals, instead of one-minute intervals as on the previous day. The following is a sample of several observations, all yielding the same general result.

I now transferred all the segments to fresh sea-water, with the following results:—

Next day all the segments were dead except the largest one, in which a single lithocyst still continued to discharge at the rate of twenty-four in five minutes.

Now, with regard to these tables, it is to be observed that during the first day the prepotent lithocyst in the eighth-part segment A maintained an undoubted supremacy over all the others, and that the same is true of the comparatively potent lithocysts in the quadrant. (This is not the case with segment B; probably the degree of prepotency of the lithocyst in this case was not sufficient to counteract the antagonistic influence of the small size of the segment.) But next day the supremacy of the small segment A was not so marked; for although its rhythm was more regular in the stale water than was that of the largest segment, its actual number of contractions in a given time was just about equal to that of the largest segment. Again, after transference to fresh sea-water, the balance began to fall on the side of the larger segments; for even the quadrant, which in the stale water had ceased its motions altogether, now held a middle position between that of the half-segment and the prepotent eighth-part segment. On the next day, again, the balance fell decidedly in favour of the larger segments, and the weaker eighth-part segment died. Lastly, next day all the smaller segments were dead.

Hence the principal facts to be gathered from these observations are, that as time goes on the rhythm of all the segments progressively decreases, and that the decrease is more marked in the case of the smaller than in that of the larger segments. This lesser endurance of the smaller segments also finds its expression in their earlier death. Now as these smaller segments started with a greater proportional amount of ganglionic power than the larger segments, their lesser amount of endurance can only, I think, be explained by supposing that the process of starvation proceeds at a rate inversely proportional to the size of the segment, a supposition which is rendered probable if we reflect that the smaller the segment the greater is the proportional area of severed nutrient tubes.[23] And in this connection it is interesting to observe that, although the endurance of the smaller segments was less than that of the larger as regards the deprivation of nutriment, it was greater than that of the larger segments as regards the deprivation of oxygen. This is shown by the greater regularity of the rhythm manifested by the smaller than by the larger segments in the stale water, and the fact is presumably to be accounted for by the consideration that the ganglia in the smaller segments were more potent than those in the larger.

With regard, therefore, to the original point under consideration, I conclude that, although the size of the segments is doubtless one factor in determining the relative frequency of contraction, there are at least two other factors quite as important, viz. the relative potency of the lithocysts, and the length of time that elapses between performing the operation and observing the rhythm. Hence it is that in my experience I have found but very few examples of Dr. Eimer's rule.

The next point I have to dwell upon is one of some interest. If the manubrium of Aurelia, or of any other covered-eyed Medusa, be suddenly cut off at its base, the swimming motions of the umbrella immediately become accelerated. This acceleration, however, only lasts for a few minutes, when it gradually begins to decline, the rate of the rhythm becoming slower and slower, until finally it comes to rest at a rate considerably less than was previously manifested by the unmutilated animal. If a circular piece be now cut out from the centre of the umbrella, the rhythm of the latter again becomes temporarily quickened; but, as before, gradual slowing next supervenes. This slowing, however, proceeds further than in the last case, so that the rate at which the rhythm next becomes stationary is even less than before. If, now, another circular ring be cut from the central part of the umbrella—i.e. if the previously open ring into which this organ had been reduced by the former operation be somewhat narrowed from within—the same effects on the rhythm are again observable; and so on with every repetition of the operation, the rate of the rhythm always being quickened in the first instance, but then gradually slowing down to a point somewhat below the rate it manifested before the previous operation. It will here suffice to quote one experiment among many I have made in this connection:—

Excepting the cases where the effects of shock are apparent, some such series of phenomena as those just recorded are always sure to ensue when a covered-eyed Medusa is mutilated in the way described, and this kind of mutilation, besides producing such marked effects on the rate of the rhythm, also produces an effect in impairing the regularity of the rhythm. In some specimens the latter effect is more marked than it is in others. The following series of observations will serve to give a good idea of this effect:—

An Aurelia manifested a regular and sustained rhythm of 36. Immediately after the removal of the manubrium, the rate of rhythm in successive minutes was as follows: 40, 39, 37, 35, 32, 30, 29, 26, 24, 18, 14 (40 seconds' pause), 16, 15, 14, 15, 16 (40 seconds' pause), 22, 20, 19, 15, 16, 17, 14, 13, 13, 15, 16, 16, 17, 18, 14, 12, 13, 11, 12, 9, 15, 16, 14, 12, 9, etc., the rhythm now continuing very irregular. An hour after the operation, the following were the number of contractions given in one-minute intervals, the observations being taken at intervals of ten minutes: 15, 15, 12, 22, 14, etc.

In this experiment, therefore, as soon as the acceleration and slowing-stages had been passed, viz. about a quarter of an hour after the operation, a great disturbance was observable in regularity of the rhythm; for before the removal of the manubrium, the Medusa had been swimming for hours with perfect regularity.

Before concluding my description of these experiments, it may perhaps be as well to mention one other, which was designed to meet a possible objection to the inferences which, as I shall immediately argue, these experiments seem to sustain. It occurred to me as a remote possibility that the slowing and irregularity of the rhythm, which are observable about a quarter of an hour after the operations described, might be due to the deprivation of adequate nourishment suffered by the ganglia, in consequence of the escape of nutrient matter from the cut ends of the nutrient tubes. Accordingly, instead of cutting off the manubrium, I tried the effect of momentarily immersing it in hot water, and found that the subsequent disturbances of the rhythm were precisely similar to those which result from removal of the manubrium.

Now, to draw any inferences from such meagre facts as the above would be hazardous, unless we recognize that in so doing our inferences are not trustworthy. But, with this recognition, I think there will be no harm in briefly stating the deductions to which the facts, such as they are, would seem to point.

Physiologists are undecided as to the extent in which many apparently automatic actions may not really be actions of a reflex kind. Given any ganglio-muscular tissue which is rhythmically contracting, how are we to know whether the action of the ganglia is truly automatic, or sustained from time to time by stimuli proceeding from other parts of the organism? In most cases experiments cannot be conducted with reference to this question, but in the case of the Medusæ they may be so, and it was with the view of throwing light on this question that the experiments just described were made. Now in these experiments the fact is sufficiently obvious that mutilations of any part of the organism modify the rhythm of the marginal ganglia most profoundly. That this modification does not proceed from shock, would seem to be indicated by the facts that the first effect of the mutilation is to quicken the rhythm; that there is a sort of general proportion to be observed between the amount of tissue abstracted and the degree of slowing of the rhythm produced; and that the slowing effects continue for so long a time. All these facts seem to show that we have here something other than mere shock to deal with.

A strong suspicion, therefore, arises that the cause of the slowing of the rhythm which results from removing the manubrium, or a part of the general contractile tissue of the bell, consists in the destruction of some influence of an afferent character which had previously emanated from the parts of the organism which have been removed, and that the normal rhythm before the operation was partly due to a continuous reception, on the part of the ganglia, of this afferent or stimulating influence. In support of this view are the facts that the first effect of such an operation as we are considering is greatly to accelerate the rhythm, and that this acceleration then gradually declines through a period of about a quarter of an hour. These facts tend to support this view, because, if it is correct, they are what we might anticipate. If the manubrium, for instance, while in situ is continually supplying a gentle stimulus to the marginal ganglia, when it is suddenly cut off, the nerve-tracts through which this stimulating influence had previously been conveyed must be cut through; and as it is well known how irritable nerve-fibres are at their points of section, it is to be expected that the irritation caused by cutting these nerve-tracts, and probably also by the action of the sea-water on their cut extremities, would cause them to stimulate the ganglia more powerfully than they did before their mutilation. And here I may state that on several occasions, with vigorous specimens, I have observed a sudden removal of the manubrium to be followed, not merely with a quickening of the rhythm on the part of the bell, but with a violent and long-sustained spasm.

Again, as regards the other fact before us, it is obvious that as soon as the cut extremities of the nerves begin to die down, and so gradually to lose their irritability, the effect on the rhythm would be just what we observe it to be, viz. a gradual slowing till the rate falls considerably below that which was exhibited by the unmutilated animal. And even the irregularity which is at this stage so frequently observable is, I think, what we should expect to find if this view as to the essentially reflex character of the natural rhythm is the true one.

If this view is the true one, the question next arises as to the nature of the process which goes on in the excitable tissues, and which afterwards acts as a stimulus on the ganglionic tissues. This question, however, I am quite unable to answer. Whether the process is one of oxygenation, of chemical changes exerted by the sea-water, or a process of any other kind, further experiments may be able to show; but meanwhile I have no suggestion to offer.