On the Excitation of Crustacean Muscle: II. Neuromuscular Facilitation

While the physiology of nervous conduction in Crustacea does not seem to differ fundamentally from conduction in vertebrate nerve, the relation between stimulus and muscular response is much more complicated (Pantin, 1934). The response varies in a peculiar manner with both the frequency and the intensity of the stimulus. The work of Barnes (1934) and of Wiersma (1933) suggests that these peculiarities are due partly to the tendency for a strong stimulus to give rise to repetitive impulses in crustacean nerve. This factor must be taken into account in any analysis of neuromuscular mechanism. Some workers, particularly Lucas (1917a), have brought evidence of separate “quick” and “slow” contraction mechanisms in crustacean muscle. But apart from these complications, the contraction of Crustacea skeletal muscle is governed by factors such as frequency of stimulation in ways not normally found in vertebrate skeletal muscle. The response increases enormously with the number and frequency of the stimuli; and, as has been known since Richet’s (1879) time, there is great neuromuscular summation. But the relation of the response to frequency has not been systematically analysed.

In considering the experiments of many earlier workers there is noticeable discrepancy between the frequencies which have generally been employed for experiment and those which one may consider as normal for crustacean nerve. Thus the impulses recorded by Barnes in the nerve are of the general order of 100 per sec., and Beresina and Feng’s (1933) analysis of the heat production suggests that the nerve will carry impulses up to a frequency of 300 per sec. The work of Lucas and others has shown how greatly neuromuscular summation depends upon the interval between stimuli. It is therefore perhaps scarcely surprising that anomalous results are obtained when single stimuli and stimulation frequencies of less than 10 per sec. are employed to elucidate a neuromuscular mechanism which normally responds to frequencies of perhaps several hundred per sec. A systematic investigation of the response of the limb muscles of Carcinus maenas was therefore made, extending to the highest frequencies.

When working with crustacean nerve-muscle preparations it was found by experience that to obtain reproducible experimental results certain precautions have to be taken. The most important of these concerns the freshness of the material. The limbs of crabs in poor condition may show responses to stimuli for long periods, but they are subject to a physiological decay expressed in such ways as progressive shifts of the threshold of excitation. In the following experiments fresh crabs were always used and the limbs were taken off with as little damage as possible. Even in healthy crabs the neuromuscular mechanism is very delicate, and dissection to expose the nerve is sufficient to alter its physiological state completely. In the experiments the leg was separated at the natural point of fracture, and pin-shaped silver electrodes were pushed through the meropodite above and below the nerve with minimal damage to the tissue. The neuromuscular system fatigues easily. There is a fall in the rate of contraction, a depression of neuromuscular summation, and particularly a prolonged delay of relaxation (Fig. 1). Because of this tendency batteries of stimuli should be as few and brief as possible if reproducible controls are to be obtained.

Most of the experiments which have so far been done upon crustacean muscle have been performed upon the chelae. But it will be shown presently that the chela mechanism is specialised. In these experiments, the thoracic walking legs of Carcinus maenas were used and the chelae subsequently compared with them.

An important precaution which was found necessary was the employment of stimuli which were as short or shorter than the true chronaxie of the nerve. This minimises the tendency to repetitive discharge of impulses. The most general method of stimulation was a neon-lamp stimulator after the type of Briscoe and Leyshon (1930). A 300-volt battery charged a condenser usually of 0.02µF capacity. This was allowed to discharge periodically through a neon lamp, and a resistance of 5000 Ω. A potentiometer was in series with the resistance, and the electrodes were led from it. In other experiments induction shocks were used, in some cases with the make shocks short-circuited so as to repeat the experiments of Blaschko, Cattell and Kahn (1931). This method is, however, unsuitable for the very high frequencies to which the rate of stimulation has to be carried. The frequency of stimulation was determined stroboscopically.

The amputated leg was generally used without perfusion. The tendon of the extensor or flexor muscle was cut where it joined the dactylopodite, thus leaving a simple muscle preparation of the other muscle attached to the dactylopodite which was in turn attached to the recording lever. All contractions were registered by isometric levers unless otherwise stated. Most of the experiments were performed on the flexor muscle, and reference is made to this unless some other muscle is specially mentioned. In all its essential features the behaviour of the extensor is identical with that of the flexor. The points of difference between these muscles will be considered in a later paper. The experiments were done at temperatures of from 14 to 18 ° C.

The muscular response which follows excitation of the nerve in the limbs of Carcinus varies both with the frequency and the intensity of the stimulus. Let us first consider the relation of the response to frequency. In the following experiments the response of the flexor of the dactylopodite of the leg was determined on stimulation of the nerve in the meropodite. The intensity of the stimuli was 10-20 per cent, above threshold value at frequencies of 100-200 per sec. The threshold does not vary greatly over a wide range of frequencies in this neighbourhood. In a typical experiment it rose from an arbitrary value of 11·4 at a frequency of 45 per sec. to a value of 12·0 at 175 per sec.

Variation in frequency has an enormous effect upon the rate of contraction. The rate of contraction increases greatly with the frequency till the very high value of about 300 shocks per sec. This maximum rate of contraction with high-frequency stimuli is as great as any rate of contraction that can be obtained by excitation of the nerve or muscle by any means whatever. The rate of contraction can be lowered to any extent simply by decreasing the rate of stimulation. Fig. 2 A illustrates a series of superimposed isometric contractions of the flexor muscle. The duration of stimulation was constant for each (0·45 sec.). The frequency of stimulation was progressively lowered from 250 to 33 per sec. Repeated stimulation is very apt to fatigue the crustacean nerve-muscle unit. To check this the superimposed series is followed by a single response to a stimulus at 156 per sec. of 0·59 sec. duration. It is evident that fatigue is very slight, and does not greatly enter into the recorded effects.

From Fig. 2,A it is evident that the contraction rate varies with its frequency and does so in an absolutely continuous manner for this muscle. There is no abrupt change at any point from a slower to a more rapid contraction. At the upper limit the contraction rate is as great as any of which the muscle is capable. At the lower limit, the continuity of the series can be demonstrated far beyond the 33 per sec. of Fig. 2. A visible contraction can be recorded down to about 10 shocks per sec., the response remaining continuous and not breaking up into a series of discrete contractions to each shock. By microscopic observation of the free dactylopodite diminishing contraction can be traced well below this frequency. Ultimately even single stimuli can sometimes produce a microscopic excursion of the free dactylopodite, though no more.

Though the response can be traced downwards in this way to very low frequencies the whole range of normal contraction rates is covered by the range 50-200 per sec.; a range of remarkably high frequency.

The continuous relation of frequency and contraction rate gives no evidence for more than one muscular system. Blaschko, Cattell and Kahn (1931) found in the chela of Maia squinado and Cancer pagurus a discontinuous transition from a slow to a quick type of contraction with increase in frequency. In the leg flexor of Carcinus there is no such discontinuity : but the range of frequency over which the contraction rate increases from a scarcely perceptible movement to a maximal twitch is narrow, so that unless intermediate frequencies are carefully followed a discontinuity might seem apparent. We shall also see that the chela can present special features.

As the maximum contraction rate is approached at a frequency of about 300 per sec., the intensity of stimulus required for a complete response rises rapidly. This is partly due to the rise of threshold and can be corrected by increasing intensity of the stimulus, for unless this is done every stimulus will not be successful. In Table I the isometric tension developed for a constant duration of stimulation at increasing intensities is shown. At a stimulation interval of 2.8 σ, the effect described above is well marked. But apart from this, there is a real reduction of contraction rate at very high frequencies, possibly owing to a Wedensky effect. Alternatively, it might be due to uncompleted recovery of some of the muscle fibres from the absolute refractory period, a condition which apparently is found in certain muscles of the Actinozoa (Pantin, 1935). But stimuli of these high frequencies have a very fatiguing and even irreversibly damaging effect upon the neuromuscular mechanism.

These experiments show that for stimuli which are not far above threshold value a simple relation exists between the rate of contraction and the frequency. The rate of contraction increases rapidly until the interval between stimuli is very short. It is natural to suggest that the limit which is then reached is set by the refractory period : the frequency at which the maximum response is attained corresponding to the shortest interval at which each shock can initiate an impulse by falling just outside the refractory period of its predecessor. This upper limit of frequency corresponds to some 300, and in exceptional cases almost 400, stimuli per sec., that is, we would expect a refractory phase some 30σ in length.

Experiments were conducted to determine the refractory period of the motor nerve of Carcinus walking leg, after the manner of Adrian and Lucas (1912). The stimuli were break induction shocks made through contacts of a Lucas spring rheotome. Ordinary iron-cored physiological induction coils were used with the complicated relationship is found. Above the threshold, the rate of contraction remains constant at first, but when the intensity has been raised to 50 or 100 per cent., the contraction rate undergoes rapid diminution to a lower value (Fig. 4A), and this lower rate is maintained even though the intensity is increased several hundred per cent. This diminution is due to the excitation of an inhibitor nerve, the presence of which has been substantiated by a succession of workers, particularly Biedermann (1887), Hardy (1894), Hoffmann (1914), and Knowlton and Campbell (1929). The physiological evidence in favour of this view was discussed in an earlier paper (Pantin, 1934), and experiments to be described fully in a later paper confirm it.

The threshold intensity at which inhibition begins to affect the response varies with different preparations. When the threshold of inhibition is unusually high the relation of the response of the muscle to the intensity of the stimulus can be examined free from the complication of superimposed inhibition. A fundamental independence between response and intensity becomes evident except near the threshold. Fig. 4B shows a case where the threshold of inhibition was not reached until 10 times the threshold of excitation. The identity of the rate of contraction over the whole range is clear. The stimuli were induction shocks at a rate of 37 per sec., the make shocks being short-circuited. It is, in fact, the incidence of inhibition which is primarily responsible for variation of response with intensity of stimulus, and were it not for this the independence of intensity would be immediately apparent.

Although over a wide range there is a significant independence between intensity of stimulus and response, there is almost always a region near the threshold where the response to a battery of stimuli gradually increases with the intensity. This gradation of response is at least partly due to a shift of threshold during prolonged stimulation so that not all the stimuli of a battery are strong enough to be effective. This is particularly evident when the muscle only responds to the first few stimuli of a battery of threshold strength and then relaxes ; though a very slight increase in intensity enables the contraction to be maintained. Such effects become more evident at high frequencies. They can be diminished until the response takes place almost in an all-or-nothing manner by restricting the stimuli to very short batteries and by ensuring that the preparation is undamaged.

Nowhere between the threshold of excitation and that of inhibition does the rate of contraction change abruptly from a lower to a higher rate. But if the intensity of a stimulus of low or moderate frequency is increased to high values a point is reached, usually at about 10-20 times the threshold value, at which the rate of contraction suddenly increases. The increase in rate may be considerable. This apparent passage from a “slow” to a “quick” contraction with increase in intensity has been noted by various workers, notably Blaschko, Cattell and Kahn (1931). Lucas (1917 a) brought evidence to show the existence of two separately excitable contractile systems, a slow and a quick, with different thresholds in the claw of Astacus. It is natural to suggest that the sudden increase in response found in Carcinus leg at high intensities may be due to the presence of a second excitable contractile system with a higher threshold. But it is also possible that the increased response is due to repetitive excitation by the strong stimuli. The evidence is in favour of this and against the presence of two excitable systems. First, the rate of this quick contraction never exceeds that which can be achieved at low intensities, merely by raising the frequency. This is hard to explain on the basis of two excitable systems whereas it is a natural consequence of repetitive excitation because this is tantamount to an increase in frequency. Secondly fatigue of the slower contraction also fatigues the quick, so that these are not physiologically separate. Finally strong stimulation frequently shows a marked after-discharge, and it does not seem easy to account for this except on the basis of repetitive discharge of the nerve.

The development of a large rapid contraction in response to strong stimuli has been a source of confusion in the past. Thus the small tensions and their smooth development in the muscles of Carcinus (Fig. 2A)are very different from those obtained by Richet (1879) in the claw of Astacus. To a series of shocks at a frequency of 10 per sec. or less, Richet obtained decided contractions. These did not resemble the responses we have just described. Instead of a smooth rise in tension, each shock caused a separate contraction, so that the whole response resembled a staircase. The first few shocks of a series might produce no response, but each succeeding one after that caused an individual contraction of increasing extent. Though these individual contractions partly fuse, the clonic nature of the whole response differentiates it from the smooth development of tension seen in Fig. 2A.

The responses of Richet were obtained with relatively strong induction shocks, and similar responses can be produced in Carcinus at similar low frequencies if the intensity is raised to 10 or 20 times the true threshold value. This kind of response is explicable if it is assumed that these strong shocks cause repetitive discharge from the nerve, as such stimuli are known to do (cf. Barnes, 1934; Wiersma, 1933). If each shock gives rise to a brief high-frequency train of impulses, each will initiate a sharp tetanic contraction of brief duration. Partial fusion will give just such a clonic response as is shown in Fig. 2B.

Richet’s experiments have been taken to show the existence of neuromuscular summation in crustacean muscle (Lucas, 1917b). But while they may well involve this, there is a very strong probability that they are due primarily to repetitive excitation of the nerve.

We have concluded that the rapid contraction of the dactylopodite flexor of the leg in response to very strong stimuli is probably not due to a second excitable system. But various authors have attributed a double mechanism to other crustacean muscles and we must consider them more fully. The most certain evidence is that presented by Lucas (1917a). He found in Astacus and Homarus that some chelae showed the existence of two excitable systems connected with two separate contractile elements, a “quick” and a “slow”, distinguished by very different rate of contraction, different refractory period and different excitation time. Thus in the intensity-duration curve, long stimuli of low intensity often gave the slow type of contraction, while short stimuli of higher intensity gave rise to the quick. The strength-duration curves of the quick and of the slow systems often overlapped, so that the compound curve of the two systems taken together often showed a decided discontinuity. The intensity-duration relations of the dactylopodite flexor were therefore examined for such discontinuities.

The neon-lamp stimulator provides a simple means for determining an intensity-duration curve under the same conditions as those used in causing normal responses in crustacean muscle. Intensity-duration relations found in this way refer to repetitive stimuli, and may therefore differ from single-shock methods involving the use of rheotomes. If the resistance of a neon-lamp discharge circuit is kept constant, the duration of each shock depends upon the capacity of the condenser. By varying this any duration of shock is obtainable. The relation of capacity to duration is approximately linear. It was determined empirically by observing the duration of the flash in the neon lamp with the aid of a rotating mirror. The relation of discharge time to capacity for the circuit employed will be seen in Fig. 5 on the scales of the lower abscissa. In the neon-lamp circuit the condenser was charged by a battery of 300 volts through a 10-megohm resistance. It discharged through the neon lamp, a resistance of 5000 ohms, and a potentiometer of 300 ohms, in series. From the potentiometer arose silver-silver-chloride electrodes similar to those already used. The threshold was usually determined for frequencies of 80, 40, 20, and 10 per sec. Of necessity the higher frequencies could not be used for shocks of very long duration. The threshold varies little at moderate frequencies, as already noted (Table I), though generally it increased slightly with frequency.

Fig. 5 (black triangles A) shows the results of a typical experiment. The threshold intensity increases in a regular way with decrease in flash duration. In terms of flash duration, there is a “chronaxie” of 0·4-0·6σ, This value is obtained quite regularly. Over the range of flash durations employed there is no evidence of any marked discontinuity.

The discharge of the neon-lamp circuit does not correspond in form to a rectilinear current as obtained from a direct current and a rheotome, so that “chronaxies” determined by the two methods do not correspond. By employing the same electrodes, intensity-duration curves by both methods can be compared successively. When using a Lucas spring rheotome, only a single stimulus can be sent into the nerve at one time. The response to such a single stimulus is minute, but it can be satisfactorily observed by viewing the dactylopodite under a microscope with a micrometer. Intensity-duration curves obtained with a rheotome in this way gave chronaxies of the order of 0·2-0·4σ at 19° C. This is half the value obtained in terms of flash duration for repetitive stimulation.

In Fig. 5, curve A, the results of both methods used in an experiment are plotted together. The values obtained with a rheotome (crosses) are plotted on a time scale (upper abscissa) which is double that used for the results of the neon-lamp method (lower abscissa). On these scales the whole intensity-duration curves by the two methods superpose. A neon-lamp discharge thus appears to be equivalent to a rectilinear pulse of half its duration over the whole range.

The chronaxie to which these intensity-duration curves correspond is extremely short and in agreement with the short refractory period and the high effective frequency for maximal stimulation. No evidence was obtained by either method of a discontinuity in the intensity-duration curve similar to that found by Lucas; nor was there any change in the form of the response to repetitive excitation of a given frequency, whatever the duration of the individual shocks.

We have seen that if its intensity is greatly raised, a stimulus of low frequency may cause a contraction of considerable size, apparently owing to repetitive excitation. Indeed at frequencies of 3 or 4 per sec. this is the first response visible to the naked eye. The contractions differ greatly from the even development of tension caused by stimuli near threshold value. Fig. 2B shows such an enhanced response to stimuli at a frequency of 3 per sec., but at 17 times the true threshold intensity.

The relation of the intensity to the duration of shock required to produce these enhanced responses follows a very regular curve which simulates an ordinary chronaxie curve. An example is shown in Fig. 5, curve B, where the threshold is traced for the enhanced response to stimulation of the nerve in the same limb, for which the true threshold intensity-duration curve is recorded in the lower figure (A, black triangles). Neon-lamp stimuli were used at a frequency of 4 per sec. These apparent chronaxie curves are very misleading, because single or very low-frequency stimuli of true threshold value produce only a microscopic response, so that if a visible response is taken to indicate the threshold, only the intensityduration curve for the enhanced response is disclosed. These curves give an apparent chronaxie of the order of 10-30σ, but we have already shown that the entire muscle can be made to respond maximally to shocks sent in at over 300 a sec., that is, at intervals far shorter than this apparent chronaxie. It is difficult to account for this except on the basis of our previous conclusion that the enhanced response is due to repetitive excitation of the nerve. The resemblance of its intensity-duration relationship to a true “chronaxie” curve is superficial. There exist in the literature records of chronaxies of 6σ and upwards for crustacean muscle (Lapicque, 1926). It is highly probable that these long chronaxies refer in fact to the abnormal enhanced responses to stimuli of high intensity, and therefore throw no direct light upon the true chronaxie of the neuromuscular system.

The difference between the true threshold and that of the enhanced response is much greater for stimuli of very short duration. This has a practical consequence of considerable importance, because by keeping the duration of shocks as short as possible we are thus enabled to render minimal the possibility of repetitive discharge of the nerve. For this reason the duration of all stimuli was kept less than the chronaxie.

The experiments have shown that though the effect of increased intensity of stimulus is complicated by inhibition and by repetitive stimulation, there is no evidence for the existence of more than one kind of excitable system in the muscle. It is even possible that this system may correspond to a single motor unit. In Crustacea Hardy (1894), Hoffmann (1914) and others have shown that a muscle is supplied by an extraordinarily small number of motor axons or even by a single axon. In vertebrate skeletal muscle a multiplicity of motor units allows a graded increase of the response with the intensity of stimulus applied to the nerve. The poverty of motor units in the Crustacea renders this impossible. Here the motor units are so few that if more than one were present an increase in intensity must be followed by an increase of response in a succession of a few abrupt steps. Such steps are not found.

This poverty of motor units renders impossible any gradation of response on the vertebrate plan. But our experiments have shown that both the rate and extent of contraction of the flexor muscle can be very completely controlled by the frequency and number of stimuli. How this can take place may be explained most simply in the way indicated by Lucas (1917 a) in his work on the crustacean claw. He demonstrated well-marked summation of nervous impulses at the neuromuscular junction. A single impulse might be ineffective, but it facilitated the transmission of a second impulse, suitably timed, to the muscle fibres. The second impulse will thus cause some muscle fibres to respond through facilitation. A succession of such impulses might well involve increasing numbers of muscle fibres through progressive neuromuscular facilitation in this way. Evidence supporting this was put forward in an earlier paper (Pantin, 1934). By examining the redevelopment of tension following a quick release, during slow contractions produced by low-frequency stimulation, it was shown that increasing numbers of contractile units came into action as the contraction developed.

It can be shown that at all frequencies increasing numbers of muscle fibres come into action during a succession of stimuli. Experiments were performed to show this by measuring the isometric tension developed in response to known numbers of stimuli. Fig. 6 illustrates a typical experiment. A neon-lamp stimulator was adjusted to give shocks 10 per cent, above the threshold strength and the frequency was carefully determined. The number of shocks was controlled by a calibrated clockwork motor which made a mercury contact followed by a short circuit in the electrode circuit. The ordinates in Fig. 6 give the maximum tension developed to known numbers of shocks at the frequency indicated. At all frequencies the isometric tension developed increases progressively with the number of shocks, in a characteristically sigmoid manner. A few stimuli produce scarcely any response. With more stimuli the tension rises to increasingly large values. With still more stimuli, this increase falls off as a limiting tension is approached. At higher frequencies this limiting tension corresponds to the maximum possible tension that the muscle can produce. At lower frequencies the limit may be far below the maximum however long the stimulation continues.

These experiments support the view that the development of tension in the muscle takes place as follows. The first impulse of a series activates only a very few muscle fibres. Each succeeding impulse brings in more and more muscle fibres, at first in rapidly increasing numbers, and then more slowly as the full contraction is attained. This means that there is a statistical distribution of the sensitivity of the muscle fibres about a mode. The majority require a certain number of impulses before they are brought into action. Some are more sensitive, and so require fewer impulses to facilitate them, while some are less sensitive and require more. Such a statistical distribution of sensitivity is evidently a factor of the highest importance, since it controls the relation of the rate of development of tension to the number of stimuli sent in. Without its aid neuromuscular facilitation alone would be unable to cause the graded rate of response with frequency.

One is tempted to suppose that the isometric tensions shown in Fig. 6 are approximate measures of the number of fibres brought into action. The danger of this can be appreciated when it is remembered that a single isometric twitch in vertebrate skeletal muscle may produce only some 25 per cent, of the maximal tension developed in a tetanus. Nevertheless the very large number of shocks required to produce maximum tension in the Crustacea muscle, the sigmoid increase of tension with number of shocks, and the minute tensions developed to the first few shocks distinguish this muscle altogether from Vertebrate skeletal muscle. Comparison of Fig. 2 A with the effect of frequency on the genesis of tetanus in vertebrate skeletal muscle also shows how different is the method of development of tension in the two cases.

We have shown that in a battery of stimuli each successive shock activates more and more muscle fibres. It follows that increase in frequency must of itself increase the rate of contraction simply by permitting more contraction increments in a given space of time. Did this factor operate alone, we would expect the contraction rate to be roughly proportional to the frequency of stimulation. But Fig. 2A shows that the response increases much too rapidly with the frequency for this to be true. Lucas, in his experiments using pairs of stimuli, found that as the stimulation interval was increased from a certain short optimum value, the response caused by the second stimulus diminished. From this it appears that the facilitating effect of the first stimulus diminishes with time, so that fewer and fewer muscle fibres remain able to respond. Such decay in the facilitating power of previous stimuli may take place in the large responses caused by batteries of many stimuli as well as for the minute ones caused by a pair only. Fig. 6 shows that this is true. The tension developed in response to a battery of 100 shocks is about 80 per cent, of the maximum when the stimulation interval is 5−6σ, but it is less than 5 per cent, when the interval is 22σ. Indeed, if the stimuli are given at sufficiently long intervals no facilitated response at all is obtained. This must mean that neuromuscular facilitation decays after each stimulus.

We may suppose that after each impulse has passed down a nerve it produces some change between nerve and muscle which tends to permit a second impulse to activate the muscle fibres. In time this effect gradually dies away. This decay of facilitation can be traced by observing how the tension developed in response to a known number of stimuli falls off as the stimulation interval is lengthened from the optimum. Fig. 7 shows that facilitation falls off most rapidly at first.

These experiments show that the development of tension in the flexor muscle is due to the progressive bringing into action of more and more muscle fibres with each successive impulse. This is brought about by the facilitated transmission of excitation to muscle fibres of varying sensitivity. The most sensitive respond to a single stimulus. The majority require many, while more than 100 shocks are required to activate the least sensitive muscle fibres. Finally, the facilitating power of each impulse rapidly decays, so that the contraction increment which a subsequent stimulus can produce falls off rapidly as the time interval between the impulses is lengthened. The functional result of this is that the rate and extent of contraction increase with great rapidity over a comparatively narrow zone of frequencies, as shown in Fig. 2A.

At low frequencies the rate of development of tension is governed by the numbers of stimuli received in a given time and by the tension increment following each stimulus. Both these factors increase with the frequency but, as we have already seen, the rate of contraction does not pass an optimum value corresponding to a frequency of some 300 per sec. There is a similar optimum for the isometric tension developed, as shown in Fig. 7. It has been shown earlier that one of the factors limiting the rate of contraction is the refractory period. But the refractory period is not the only factor involved. In the experiments upon which Fig. 6 is based the rate of contraction at a stimulation interval of 2. 8 σ exceeded that at 5·6 σ. But the rate of contraction is determined by the increase in tension in a given time, and in a given time there are twice as many shocks at an interval of 2·8 σ as there can be at an interval of 5·6 σ. If instead of this we plot, as in Fig. 7, the maximum tension produced by a given number of shocks, the tension developed at 2·8 σ interval is seen to fall well below that developed at 5·6 σ interval. This means that the average tension increment following each shock is already falling before the contraction rate has actually achieved its maximum. This fall-off in tension increment per shock at high frequencies may be due to a real decay of facilitation. There are, however, several other possible explanations for it.

Whatever causes determine the optimum contraction rate, there is one factor which must certainly limit it if the rate is increased far enough. This is the natural speed of contraction of the whole muscle as controlled by its physical properties such as viscosity and elasticity. In the muscles of Carcinus this limit is actually approached. At low frequencies the duration of the rise in tension in response to a battery of stimuli is about equal to the duration of the battery itself. But at higher frequencies the duration of the battery of stimuli only occupies a fraction of the time required for the development of tension by the muscle, even under the most rigidly isometric conditions. This is shown in Table II.

When this evidence is added to the fact that the rate of contraction at the optimal frequency is almost identical with the redevelopment of tension following a sudden release during a tetanus, it seems certain that the contraction rate is ultimately limited by the physical properties of the muscle.

From the functional point of view it is significant that several factors in this way naturally limit the rate of contraction in the neighbourhood of frequencies of about 300 per sec. The stimulation interval is here close to the beginning of the relative refractory period. The tension increment caused by each shock in the series has already begun to fall off at this frequency, and finally the duration of stimulus is now becoming so short compared with the natural period of contraction of the muscle that further increase in the frequency is of no service in increasing the speed of contraction. It is a matter of some interest that despite the very different methods by which the skeletal muscles of crustaceans and vertebrates are called into action the limiting factor in both cases appears to be set by the physical properties of the muscle itself.

The essential hypothesis deduced from the foregoing experiments is as follows. The flexor muscle of the dactylopodite of the leg, and such other muscles as have been investigated, behave as a single excitable system. The response is governed entirely by the frequency and number of stimuli. This is brought about through neuromuscular facilitation, each shock in a series bringing more and more muscle fibres into action. At low and moderate frequencies the rate of development of tension is governed primarily by the rate at which impulses reach the muscle and the extent of the tension increment following each impulse owing to facilitation. At the highest frequency, a limit is set by the physical properties of the muscles.

At low and intermediate frequencies, it is evident that on this hypothesis the smooth development of tension, as shown in Fig. 2A, must be the result of mechanical summation. The stimulation intervals with which we have been concerned in the present experiments are of the order of 10 σ, and these are about the same as those employed by Lucas in his experiments. Compared with this interval, the rate of relaxation, especially during its initial stages, is very slow. It also increases greatly with fatigue (Fig. 1). It is therefore not surprising that stimuli of the frequencies with which we have been concerned develop smooth increases in tension and do not show clonic twitches with each shock.

The method of development of tension in crustacean muscle seems to differ rather strikingly from that in vertebrate skeletal muscle. In the latter the development of tension to a series of maximal shocks is governed primarily by the physical properties of the muscle, while in the crustacean the rate of increase in tension is controlled by progressive activation of fresh muscle fibres during a battery of stimuli. The tension in vertebrate skeletal muscle can indeed be caused to develop in a similar manner to that of the crustacean if the motor units are not caused to respond in unison, but are progressively brought into action in increasing numbers during a battery of stimuli. Such a mode of activation of the muscle is frequently met with in reflex responses (Fulton, 1926). The process is referred to as “recruitment” of successive motor units. Recruitment here is a central phenomenon. The processes in crustacean muscle are exactly comparable to it except that there is a peripheral recruitment between the endings of the motor axon and the individual muscle fibres.

While there is an indirect parallel between the peripheral neuromuscular action of crustacean muscle and some central phenomena in the vertebrate, there is a similarity or perhaps even an identity of mechanism between crustacean neuromuscular mechanism and the activation of muscles by the nerve nets found in representatives of the simpler Metazoa. The interpretation of the nature of the neuromuscular response in crustacean muscle which has been given is exactly the same as that which has elsewhere been shown to apply to the neuromuscular organisation of certain coelenterates (Pantin, 1935). Thus in the anemone Callactis parasitica, a whole muscle may behave in certain circumstances as a single motor unit. All its responses are graded by the number and frequency of the stimuli alone through the agency of neuromuscular facilitation. Owing to the low frequencies involved and the special properties of the muscles, the mechanism in the case of the Actinozoa can be demonstrated with unusual simplicity.

The resemblance of the mechanism in Crustacea to that in coelenterates and the distinction from that in vertebrates is probably significant. The vertebrate mechanism rests upon a multiplicity of motor units. This does not appear to be commonly found in the animal kingdom. In most animals the motor nerves are restricted to a few large axons which serve their muscles by repeated division (Gaskell, 1920). It is possible that muscular control through neuromuscular facilitation in a restricted number of motor units may be the most widespread.

Finally, so far as the present experiments go, there is no evidence of two distinct kinds of contractile neuromuscular mechanism. In the flexor muscle of the dactylopodite in Carcinus there is only evidence for a single excitable system. Apart from inhibition, the effects can even be accounted for if the entire muscle is served by a single motor neurone conducting in an all-or-nothing manner, but with facilitation between nerve and muscle. Two phenomena may have given rise in the past to the supposition of the existence of more than one contractile mechanism in crustacean muscle. These are the tendency to repetitive discharge and the very rapid increase in rate of response with frequency over the critical range. But these will not account for all observed cases of differentiation into a quick and slow system, and we will consider the rest of the evidence for these in the next paper.

Part of this work was done while holding the Cambridge University Table and the Bidder Fund at the Stazione Zoologica, Naples. I wish to express my sincere thanks to Dr R. Dohrn and his staff for their hospitality and for the great facilities they gave me. I also wish to thank my wife for valuable help during many of these experiments. The cost of some of the apparatus was defrayed by a grant from the Government Grant Committee of the Royal Society.