1. The tension reflex in a number of isolated intact segments of the earthworm exhibits a definite threshold to weight. The value of the threshold in the series of preparations tested, varies between 0·1 and 1·0 g., measured as direct tension applied to the preparation. There is also a threshold for the duration of application of the weight. In the figure shown this lies above 0·5 sec.

  2. After an applied stimulus has been removed the peristaltic response of a preparation may continue for a finite length of time which seldom exceeds 3 min. The period for which the movement persists after the stimulus has been removed depends on the magnitude of the stimulus applied but not on the height of the contractions occurring. This fact is taken to support the view that the continuation of the response after the stimulus has been removed is a true after-discharge, due to a persistence of excitement in some region of the reflex arc.

  3. The after-discharge consists of a co-ordinated movement of both muscle sets which may be perfectly rhythmic. The after-discharge may be inhibited by suitable stimuli. The frequency of the after-discharge may be increased by the application of a stimulus of a different type from that which elicited the original response.

  4. When the preparation is submitted to a continuous tension, the response dies away within a period which does not normally exceed 25 min. The rate of beat in these circumstances shows three phases : a phase of acceleration, a plateau phase and a phase of decline. After the disappearance of a reflex under continuous stimulation, the response can be elicited once more by the addition of further tension.

  5. The frequency of beat in any one preparation shows a well-defined upper limit which, over the series of experiments carried out, was never observed to exceed 1 beat in 2·0 sec. The maximum frequency of beat is obtained with tensions intermediate between those of threshold value and those near the upper limit of endurance of the preparation. Lower and higher tensions than that producing maximum frequency produce submaximal rates of beat. The range of variation between rates of beat produced by a series of different tensions is less than the range of tensions used.

  6. After a tension reflex has been evoked, an increase in the weight applied may accelerate a submaximal frequency of beat, but cannot force the rhythm above its maximum. Large increments in weight lower the rate of beat. The rate of beat of a tension reflex may be accelerated by the application of a tactile stimulus and the frequency of a response to tactile stimulation may be accelerated by the application of tension. By whatever stimulus the acceleration of beat of a reflex is carried out, it is not accompanied by any break in the rhythm.

Some of the responses of a series of complete segments excised from the body of the earthworm have been described in the preceding paper (Collier, 1939). It was there shown that the rhythmic peristaltic reflexes are not initiated in the absence of the nerve cord. While the responses, of an isolated region of the worm, to touch are somewhat variable and exhibit several different components, the response to tension possesses definite and regular properties. Moreover, from evidence presented in this and in the preceding paper and fully discussed on p. 309, it would seem that the rhythm of the tension reflex is maintained by a reaction of the nervous system to the original stimulus, rather than by subsidiary reflexes set up by the peristaltic movements themselves. This simplicity and regularity in the behaviour of a number of segments submitted to longitudinal tension renders their response a suitable subject for the study of reflex action in the earthworm. It is the purpose of the present paper to analyse the tension reflex as far as possible along quantitative lines. The material and methods used are those described in the preceding paper.

The threshold to weight

The tension reflex is elicitable only by weights which exceed a definite threshold. The weight threshold is taken as the smallest magnitude of weight which, applied for an indefinite length of time, will evoke regular peristaltic movements. The value of this can be found by simple trial and error. Over a number of preparations the weight threshold was found to vary between 0·1 and 1·0 g., measured as tension applied direct to the segments. During the measurement of threshold these preparations were already under direct tension of up to 0·25 g., produced by the balancing weights necessary to prevent the preparation from sinking in the Ringer’s solution.

In any one preparation the weight threshold does not remain constant, but varies considerably over the period of about 6 hr. which is the normal experimental life of the preparation. The following experiment illustrates the range of variation.

Experiment on 20 July 1938. Temp. 18·5° C. Preparation set up at 11.30 a.m. Balanced in Ringer’s solution under 0·2 g. direct tension.

The variability of the threshold renders it unsuitable for use in examining quantitatively the properties of the tension reflex.

The threshold to the duration of stimulus

It is necessary to apply a weight for more than a minimum duration of time if a peristaltic response is to be obtained. In the case illustrated in Fig. 1 a, b, it will be seen that a direct tension of 0·5 g. must be applied to the preparation for a period exceeding 0·5 sec. in order to evoke a peristaltic response.

Fig. 1.

a and 1 b. Responses of a preparation to applications of tension for brief periods. Denervated suspension. Time in these and all subsequent tracings except Fig. 9 a marked in seconds. ↓ indicates application, ↑ removal of 0·5 g. tension.

Fig. 1.

a and 1 b. Responses of a preparation to applications of tension for brief periods. Denervated suspension. Time in these and all subsequent tracings except Fig. 9 a marked in seconds. ↓ indicates application, ↑ removal of 0·5 g. tension.

Fig. 2.

After-discharge following the application of 1·0 g. for 10 sec.

Fig. 2.

After-discharge following the application of 1·0 g. for 10 sec.

The characteristic rhythmic response to certain stimuli exhibited by the earthworm preparation may persist for some minutes after the stimulus has ceased. In the case illustrated in Fig. 2 the contractions elicited by a tension stimulus continue to occur for a period of nearly 3 min. after the weight has been removed. The period for which contractions persist is in no way related to the latency of the response, but is many hundred times longer. While the persistence of peristaltic responses for periods of up to 3 min. after removal of the stimulus is frequently met with, the contractions nevertheless definitely cease within a finite period of time.

The continuation of contractions after removal of the stimulus may be explained in two ways : (i) there may be a persistence, in some region of the reflex arc, of the excitement caused by the external stimulus ; (ii) reflexes arising from the peristaltic movements themselves may excite new movements and these in turn excite still more. That explanation (i) is the correct one is shown by a study of the relation of the duration of the persistence of movements to the magnitude of the stimulus applied. If tensions of different magnitude are applied to a preparation for a constant period of time, the duration of the persisting movement is directly dependent on the magnitude of the tension applied, but not on the height of the beat. This is illustrated in Fig. 3,a, b, c. It will be seen that the maximum height of contraction in Fig. 3 b, c, disregarding in each case the upstroke during which the weight was removed, is nearly the same, while the persistence of after-discharge is of longer duration with the more intense stimulus. We can thus speak of the persistence of movement after the stimulus has ceased as after-discharge in the same sense as is used in referring to the similar phenomenon in vertebrate reflexes.

Fig. 3.

The relation of after-discharge in a single preparation to tension applied. The pairs of arrows mark the application of different tensions for a constant period of 15 sec. a,0·5 g. ; b, 1·0 g.; C, 1·5 g-

Fig. 3.

The relation of after-discharge in a single preparation to tension applied. The pairs of arrows mark the application of different tensions for a constant period of 15 sec. a,0·5 g. ; b, 1·0 g.; C, 1·5 g-

The after-discharge exhibits almost all the same features as the response itself. The beats of the after-discharge can be seen to consist of contractions of the circular muscle alternating in a co-ordinated way with contractions of the longitudinal muscle. The amplitudes of the beats are of the same order of size in the reflex and in the after-discharge. The rhythm of the after-discharge may resemble the reflex proper in its regularity (Fig. 4), but more usually the rhythm of the after-discharge resembles that sometimes seen in a fatigued reflex. In this case it consists of occasional short bursts of rhythmic contractions or of isolated beats at odd intervals of time. The rhythm of the after-discharge further resembles the rhythm of the reflex in that it can be accelerated and inhibited in the same manner and by the same means. In Fig. 5 a slow rhythmic after-discharge has been set up by a tactile stimulus. During the movements of the preparation a weight is applied. The rhythm is accelerated without a break, after the preparation has passively extended under the weight. The inhibition of the after-discharge which follows tactile stimuli is illustrated in Fig. 6 a, b. Both tactile and vibrational stimuli are capable of inhibiting the after-discharge.

Fig. 4.

Rhythmic after-discharge following the application (↓) and removal (↑) of 1·0 g.

Fig. 4.

Rhythmic after-discharge following the application (↓) and removal (↑) of 1·0 g.

Fig. 5.

Acceleration of the rhythm of after-discharge by a tension stimulus. At ↑ a stream of oxygen bubbles which is being played on the preparation is turned off and a rhythmic after-discharge begins. At ↓ a tension of 1·5 g. is applied.

Fig. 5.

Acceleration of the rhythm of after-discharge by a tension stimulus. At ↑ a stream of oxygen bubbles which is being played on the preparation is turned off and a rhythmic after-discharge begins. At ↓ a tension of 1·5 g. is applied.

Fig. 6.

Inhibition of the after-discharge, a, The middle signal indicates the application of a camel’s hair brush to the dorsal surface. The after-discharge is inhibited by thumping on the table (lower signal), b, The middle signal indicates the application of a camel’s hair brush to the dorsal surface. The after-discharge is inhibited by application of a brush to the ventral surface (lower signal).

Fig. 6.

Inhibition of the after-discharge, a, The middle signal indicates the application of a camel’s hair brush to the dorsal surface. The after-discharge is inhibited by thumping on the table (lower signal), b, The middle signal indicates the application of a camel’s hair brush to the dorsal surface. The after-discharge is inhibited by application of a brush to the ventral surface (lower signal).

The close resemblance of the after-discharge to the reflex proper leads to the conclusion that, whatever excitement causes the reflex response, the after-discharge is simply due to a persistence of the same excitement in the same regions of the reflex arc.

If a tension be applied for an indefinite period to a series of intact segments of an earthworm, the peristaltic reflex takes place continuously for a period which may reach but seldom exceeds 25 min. When a stimulus is applied for an indefinite period of time, the rate of response passes through three phases. After a period of latency, the beat begins with a phase of acceleration, during which its rate rises to a maximum. The frequency of the beat then continues for a plateau phase, in which it remains constant at the maximum rate, or declines slowly from the maximum. Finally, there is a phase of decline, during which the frequency falls rapidly until the beat ceases altogether. In some preparations persistent cessation of movement is not ultimately reached until one or more periods of quietude have alternated with further brief periods of rhythmic peristaltic movement. The normal course of a tension reflex under a constant stimulus is shown in Fig. 7.

Fig. 7.

Response to a continuous tension. At ↓a tension of 1 g. is applied.

Fig. 7.

Response to a continuous tension. At ↓a tension of 1 g. is applied.

What are the factors determining the duration of the reflex? If a preparation is submitted to a series of different tensions for indefinite periods of time, it is found that differences occur between the reflexes elicited by tensions of different magnitudes. In Fig. 8 frequency of beat is plotted against time from the application of the tension in three reflexes elicited separately from one preparation. In this figure the frequency at any one beat is expressed as the average of the time intervals between that beat and the immediately preceding and the immediately subsequent ones. It will be seen that the frequency-time curves obtained from direct tensions of 0·25, 1·0 and 2·5 g. are similar in general form, but differ in details. The plateau phase is shortest and the phase of decline most abrupt in the reflex evoked by the lowest tension. The plateau is of longest duration and the decline most gradual in the reflex elicited by the greatest magnitude of weight used. The properties of the plateau and of the decline are intermediate in the reflex obtained from the intermediate weight. As with the after-discharge, in any one preparation under constant conditions, the duration of the reflex depends directly on the magnitude of the tension applied, though duration does not vary with magnitude in a simple arithmetical manner.

Fig. 8.

The course of the frequency of beat in three tension reflexes where the stimulus is applied indefinitely.

Fig. 8.

The course of the frequency of beat in three tension reflexes where the stimulus is applied indefinitely.

The factors that bring the reflex to an end may be one or more of the following: (i) decline of the sensory excitement to a low level through the adaptation of the receptors ; (ii) a break in the reflex arc through fatigue at a synapse ; (iii) the fatigue of the muscles concerned. It is possible to show that, at the time when the response ceases, muscular fatigue has not developed to a degree sufficient to stop the contractions. If movement comes to a standstill during the continuous application of a stimulus, the addition of a further weight will evoke a fresh burst of contractions, although the load on the muscle is increased (Fig. 9,a). Likewise, if the movement has ceased during the application of a tactile stimulus, such as that caused by playing a stream of oxygen bubbles on the surface of the preparation, the addition of a weight will evoke movement once more (Fig. 9 b). It is unlikely that the disappearance of the response to a series of bubbles is due to a process of sensory adaptation. Hoagland (1933) has shown that application of a series of discontinuous tactile stimuli can continue to evoke a response from the touch receptors of the frog’s skin for an indefinite period of time. If the tactile receptor mechanism of Lwnbricus resembles that of the frog in principle, as Prosser’s work suggests, then a series of discontinuous tactile stimuli, such as that produced by a jet of oxygen bubbles, is unlikely to produce a state of inactivity of the receptors through a process of sensory adaptation. If that is so, then the disappearance of the response to a continuous tension is due to a process of synaptic fatigue, taking place somewhere between sense organ and muscle.

Fig. 9.

Re-excitement of the peristaltic reflex after it has ceased under the continuous application of a stimulus, a, At each ↓1g. is applied. Time is marked in 3 sec. b, At the first ↓. a jet of oxygen bubbles ia directed on the surface. At the second ↓ 0·1 g. is applied.

Fig. 9.

Re-excitement of the peristaltic reflex after it has ceased under the continuous application of a stimulus, a, At each ↓1g. is applied. Time is marked in 3 sec. b, At the first ↓. a jet of oxygen bubbles ia directed on the surface. At the second ↓ 0·1 g. is applied.

The rate of beat in any preparation shows a well-marked upper limit, and no increase of stimulation, whether of the same type of receptors or of a different type of receptor, is capable of forcing the rate above the maximum. In Fig. 8 the maximum frequency of beat under the series of weights applied is a rate of 1 beat in 4·0 sec. This rate is attained under a direct tension of 1·0 g. Under higher and lower tensions the frequency is slightly lower—1 beat in 4·5 sec. under a direct tension of 2·5 g. and 1 beat in 5·1 sec. under a tension of 0·25 g. The maximum rate recorded in this whole series of experiments was 1 beat in 2’0 sec. ; the usual range lay between i beat in 7·5 sec. and 1 beat in 3·0 sec.

It will be seen from Figs. 8 and 10 b that the conditions under which the maximum frequency of beat occurs are those produced by an intermediate degree of tension. Consequently, if a weight is added to a preparation already beating under a low tension at a submaximal rate, the rhythm can be accelerated (Fig. 10,a, b). If a further weight is added, however, the frequency may again be lowered (Fig. 10b). In Fig. 10 b the frequency of beat under the weight of 1·0 g. first placed in the pan is 1 beat in 8·6 sec. After the addition of a second gram weight to the pan, the frequency rises to 1 beat in 7·0 sec., while on the addition of a third gram weight the frequency sinks to 1 in 8·0 sec. It will be noted that, in these changes in the rate of rhythm induced by the addition of weights, no break in the rhythm occurs, but only on occasions a passive extension due to the extra load.

Fig. 10.

Responses of preparations beating under a tension to the application of further tensions. a, To a preparation that ia beating under a tension of 1·5 g. a further 1·0 g. is added (↓) and removed (↑). b, A series of 0·5 g. tensions is applied to a preparation. Each application is indicated by ↓.

Fig. 10.

Responses of preparations beating under a tension to the application of further tensions. a, To a preparation that ia beating under a tension of 1·5 g. a further 1·0 g. is added (↓) and removed (↑). b, A series of 0·5 g. tensions is applied to a preparation. Each application is indicated by ↓.

If tension stimuli are superimposed on tactile stimuli or vice versa, the effects on the frequency of the rhythm are similar to those observed when only one kind of stimulus is used. A submaximal rhythm due to light tactile stimulation may be accelerated by application of tension (Fig. 11,a) and vice versa (Fig. 11 b). A maximal rhythm, however, is not accelerated by a further stimulus of another kind. Thus the addition of 5 g. to the pan when a preparation is beating under a strong tactile stimulus results in a slightly slower rhythm (Fig. 11 c). As when the rhythm is accelerated by increments in a single type of stimulus, the acceleration of rhythm induced by a second stimulus of different type from the first is accompanied by no break in the rhythm. Since there is strong evidence that tactile and tension stimuli excite separate end-organs, their converging effect on the rhythmicity throws considerable light on its source of origin. The topic of the site of origin of the rhythmicity will be dealt with in the discussion.

Fig. 11.

The effect of a second stimulus of different kind from the original stimulus on the rhythm of a peristaltic reflex, a. The preparation is beating under the stimulus of a stream of oxygen bubbles. Between the arrows a tension of 0·5 g-is applied, b, The preparation is beating under a tension of 1·0 g. Between the first pair of arrows a stream of Ringer’s solution is played on the surface of the preparation. Between the second pair of arrows a tension of 1·0 g. is applied, c, At the first ↓a stream of oxygen bubbles is directed on to the surface of the preparation. At the second ↓a tension of 2·5 g. is applied; at ↑ it is removed.

Fig. 11.

The effect of a second stimulus of different kind from the original stimulus on the rhythm of a peristaltic reflex, a. The preparation is beating under the stimulus of a stream of oxygen bubbles. Between the arrows a tension of 0·5 g-is applied, b, The preparation is beating under a tension of 1·0 g. Between the first pair of arrows a stream of Ringer’s solution is played on the surface of the preparation. Between the second pair of arrows a tension of 1·0 g. is applied, c, At the first ↓a stream of oxygen bubbles is directed on to the surface of the preparation. At the second ↓a tension of 2·5 g. is applied; at ↑ it is removed.

Gray & Lissmann (1938) describe the effects of applying a transitory longitudinal tension to a piece of earthworm in the course of its rhythmical movement. If the tension be applied at the correct point in the rhythmic cycle, it will elicit contraction of the longitudinal muscle ; if the tension be applied at another point in the cycle, it will accelerate the relaxation of the longitudinal muscle and the rate of onset of the circular contraction. The experiments described in the present paper show the effect of a continuous stimulus applied to a beating preparation to be similar to the above effect of a brief tension.

The application of a continuous tension does not produce any visible specific muscular contraction in the beating preparation; on the contrary, movement continues, after application of the stimulus, in exactly the same sense as would be expected if no stimulus had been applied. The frequency of the rhythm, however, may be accelerated or retarded. It can be concluded, therefore, that the order in which the two muscle sets in a series of intact segments of the earthworm come into action is determined by internal factors. The problem of the factors responsible for the initiation and maintenance of their rhythmical activity is discussed below.

Mode of origin of the rhythmicity

The question arises as to how the rhythmicity of the peristaltic reflexes is maintained. Two general explanations, not mutually exclusive, are possible:

  • That the rhythmicity of movement follows a rhythmic waxing and waning of excitement created at some point in the reflex arc by the stream of sensory impulses resulting mainly from the external stimulus applied. The after-discharge which sometimes occurs for a short period after the stimulus ceases would, in this event, be in the nature of a persistence of excitation in some region of the reflex arc.

  • That each beat occurring in the course of the response generates reflexes that elicit the subsequent beat. The rhythmicity would in this case be due to a chain of reflexes. The continuation of movement after removal of the external stimulus would be due to the carrying on of the chain of reflexes and not to an after-discharge from an excited region of the reflex arc.

Although it is impossible, owing to the mixing of sensory and motor fibres in each lateral nerve root, to distinguish by operational techniques between explanations (i) and (ii) above, there are a number of facts which overwhelmingly favour the first hypothesis. The fact that the rhythmicity is only set up by an external stim ulus and ceases within a finite period after removal of the stimulus shows that the rhythmicity is primarily dependent on external stimulation. The fact that the duration of the after-discharge is dependent on the magnitude of stimulus applied and independent of the height of the contraction is evidence for the existence of a region in the reflex arc which can be excited to give persistent rhythmic discharge.

When an external tension is applied to a quiet preparation, the first movement seen to result is a contraction of the circular musculature, which is in turn followed by a contraction of the longitudinal muscles. If the rhythmicity is to be ascribed to the operation of a chain of reflexes, any mechanism postulated to describe the process must take into account this observable response to tension. It may be supposed therefore, that in the active preparation, the longitudinal tension or extension produced during one beat will be the operative stimulus for evoking the next beat. Several facts, however, suggest that tensions created during the movement of the preparation cannot play a determining part in the maintenance of peristalsis. In the first place, if a single beat of the response is capable of initiating another beat, one would expect that a beat would never occur singly in an unfatigued preparation. In the after-discharge, however, even in fresh preparations, single full-sized beats may be observed to occur alone. A second objection to the chain reflex theory is provided by the behaviour of a beating preparation to which a weight is added. If a longitudinal tension is applied at any point in the rhythmic cycle, the contraction which follows is that which would be expected at that point in the cycle and is not a specific response to the tension (see Figs. 5, 10,a, b, 11 a, b, c). A third fact difficult to reconcile with the chain reflex theory is the execution of rhythmic movements in preparations where the longitudinal tensions created during movement must be very small. If the circular musculature is transected by a longitudinal incision through the body wall, the tension which the circular muscles can exert on the longitudinal must be greatly reduced. Nevertheless, a perfectly clear rhythmicity can be obtained in response to tactile stimulation of such a preparation (Collier, 1939, Fig. 5 c).

A number of observations agree and none disagree with the hypothesis that the rhythmicity originates internally, and that it develops when excitement reaches a sufficiently high level. On the other hand, it is difficult to reconcile several facts with the view that the rhythmicity is maintained by a chain of reflexes created by the movements themselves. The type of explanation that the experiments favour is similar to that recently applied to the swimming rhythm of fishes and of the leech (see Gray et al. 1938). It is supposed that the reflex arcs concerned with the peristaltic reflexes in the earthworm are capable of translating the sensory stream into alternate rhythmic excitement of two sets of motor centres—those of the longitudinal and those of the circular muscles. The sensory stream will be mainly derived from the external stimulus applied, though it may well be to some extent reinforced by the movements of the muscles themselves.

Site of origin of the rhythmicity

The view is adopted that the rhythmicity originates internally in response to continuous external stimulation. The question of the region of origin of the rhythmicity is best approached by a process of elimination. Although it is known that the sense organs of the earthworm may respond with a rhythmic series of impulses to tension of the body wall (Prosser, 1935), the maximum rate of sensory response is 18 impulses per sec., which is 36 times faster than the maximum rate of beat observed in the preparation used in these experiments. This numerical difference eliminates the possibility that the sense organs are the source of the rhythm. Their elimination is confirmed by the evidence provided by the interaction of two types of stimulus.

Moore (1922) concluded from the results of anaesthetizing the surface of the earthworm that the reception of tension and of touch was carried out by different organs. Consequently the fact that stimuli of tension and touch are capable of interacting together to accelerate the rhythm in a regular way confirms the conclusion that the rhythm of movement is not simply a product of rhythmic discharges in the sense organs.

The fact that the response with which we are dealing is a complicated one, consisting of co-ordinated movements of two muscle sets, makes it difficult to conceive of any means whereby it could originate in the muscles themselves or in the neuro-muscular junctions. We are thus led to the conclusion that the nerve cord of the earthworm is capable of translating a sensory stream into alternate rhythmic excitement of two muscle sets. In this capacity it resembles that part of the mammalian spinal cord involved in carrying out the scratch reflex. A continuous pinch on the flank of a spinal dog produces an alternate rhythmic excitation of the motoneurone pools of the flexor and the extensor muscles of the ipselateral hind leg, accompanied by reciprocal inhibition of the antagonistic musculature (see Sherrington, 1911, pp. 42–69). If we carry the analogy of the mammalian scratch reflex further, many points of comparison with the peristaltic reflexes of the earthworm appear. Among the most striking points of resemblance may be noted the existence of an upper limit to the frequency of the rhythm, the absence of any break in the rhythm when increments of stimulation are applied, the ability of different types of stimulus to evoke the same rhythmic movement;, and the existence of after-discharge which is both rhythmic and co-ordinated.

Collier
,
H. O. J.
(
1939
).
J. exp. Biol
.
16
,
286
.
Gray
,
J.
&
Lissmann
,
H. W.
(
1938
).
J. exp. Biol
.
15
,
506
.
Gray
,
J.
,
Lissmann
,
H. W.
&
Pumphrey
,
R. J.
(
1938
).
J. exp. Biol
.
15
,
408
.
Hoagland
,
H.
(
1933
).
J. gen. Physiol
.
16
,
911
.
Moore
,
A. R.
(
1922
).
J. gen. phytsiol
.
5
,
327
.
Prosser
,
C. L.
(
1935
).
J. exp. Biol
.
12
,
95
.
Sherrington
,
C. S.
(
1911
).
The Integrative Action of the Nervous System
.
London
:
Constable
.