1. The purpose of this investigation was to locate the site of fatigue in the giant fibre reflex of the earthworm.

  2. The following sites do not show rapid fatigue on repetitive stimulation : contractile mechanism of muscle, neuromuscular junctions, junctions in the course of the motor neurone tracts.

  3. Rapid failure of transmission (accommodation) occurs between the sensory neurones and the giant fibre, and between the giant fibre and the motor neurones.

When the anterior end of an earthworm is repeatedly prodded the rapid contractions of the longitudinal muscle, extensive at first, rapidly get smaller and finally cease to be evoked altogether although slow locomotory movements continue to take place. A fast contraction is rarely given after ten successive stimuli even when they are separated by as much as 5 or 10 sec. This rapid fatigue, so characteristic of the escape response, might be the result of :

  • (1) Adaptation of the sensory endings in the skin.

  • (2) Transmission failure (accommodation) at junctions on the afferent side of the reflex, i.e. between successive sensory neurones or between sensory neurones and giant fibre (‘sensory-to-giant’ junctions).

  • (3) Accommodation at junctions between the giant fibre and motor neurones (‘giant-to-motor’ junctions).

  • (4) Accommodation at junctions in the course of the motor neurone pathways.

  • (5) Failure at the neuromuscular junctions.

  • (6) Fatigue of the longitudinal muscle fibres.

In order to determine its cause an analysis was made of the fatigue properties of the afferent and efferent sides of the giant fibre reflex.

Mature specimens of the earthworm Lumbricus terrestris were used in all experiments. Since a number of different preparations were used they will be described in the accounts of the individual experiments in which they featured. The stimulating and recording techniques are described in a previous paper (Roberts, 1962a).

Sensory adaptation

To test whether fatigue of the rapid response might be the result of adaptation of the sensory endings in the skin, impulses were recorded from the nerve cord in response to repetitive tactile stimulations of the body wall. It was found that long after giant fibre impulses had ceased to be produced small slow impulses were recorded and continued to be evoked for as long as stimulation was continued. In order to explain fatigue of the rapid response on the basis of sensory adaptation it would be necessary to assume that adaptation of those sensory endings which ultimately connect with the giant fibre system occurs much more rapidly than those which connect with the other, slowly conducting, components of the nerve cord. The alternative explanation is that failure occurs either at junctions in the course of the sensory neurone tracts or between the sensory neurones and the giant fibre.

Sensory-to-giant junction

The transmission properties of this junction were investigated by directly stimulating sensory fibres at the anterior end and recording the giant fibre impulses produced. Attempts were made to stimulate sensory neurones through a pair of micro-electrodes placed in contact with individual segmental nerves which were exposed by dissection, but in the majority of cases no impulses were evoked in the giant fibre, accommodation having presumably occurred as a result of the operation. Another preparation, involving less dissection, was therefore used. The worm was quickly pinned down with its lateral surface uppermost at the anterior end, and the ventral surface uppermost in the region immediately posterior to the clitellum (Fig. 1). The cord was exposed for a short distance behind the clitellum and hooked over a pair of recording electrodes. Two pairs of stimulating electrodes were now inserted through the lateral body wall at the anterior end, the first pair 2 mm. and the second pair 7 mm. from the nerve cord. The preparation was earthed between the first pair of electrodes and the nerve cord in order, as far as possible, to avoid direct stimulation of the giant fibres. It is known that the segmental nerves, on leaving the cord, run out laterally, pass through the longitudinal muscle layer and continue between this and the circular muscle towards the mid-dorsal surface (Hess, 1925). Stimuli delivered through the lateral body wall would be expected to excite both the sensory and motor neurones of the peripheral nervous system. In the present experiment the junctions between the motor neurones and the giant fibre were accommodated when the preparation was set up so that we may regard giant fibre impulses recorded as being the result of excitation of the giant fibre through its sensory connexions.

Fig. 1.

Lumbricus terrestris. Preparation for recording action potentials from the exposed nerve cord in response to excitation of the giant fibres through their sensory connexions. Worm twisted so that lateral surface is uppermost at the anterior end, ventral surface uppermost behind clitellum. First pair of stimulating electrodes (S1) positioned 2 mm. from cord, second pair (S2) 7 mm. from cord. Preparation earthed between the stimulating and recording electrodes and between Si and the nerve cord.

Fig. 1.

Lumbricus terrestris. Preparation for recording action potentials from the exposed nerve cord in response to excitation of the giant fibres through their sensory connexions. Worm twisted so that lateral surface is uppermost at the anterior end, ventral surface uppermost behind clitellum. First pair of stimulating electrodes (S1) positioned 2 mm. from cord, second pair (S2) 7 mm. from cord. Preparation earthed between the stimulating and recording electrodes and between Si and the nerve cord.

Fig. 2 a shows the result of stimulating the preparation with a single shock delivered through the first (more ventral) pair of electrodes. 15 msec, after the stimulus a smedian giant fibre impulse is recorded which is suspected to be the result of excitation of the giant fibre through its sensory connexions. That this is so is confirmed by the fact that when the shock is delivered through the second pair of electrodes the interval between the stimulus and the giant fibre impulse is increased to 25 msec. (Fig. 26). This increased delay is the result of the extra conducting distance involved in stimulating the sensory fibres at a greater distance from the nerve cord.

Fig. 2.

L. terrestris. Action potentials produced by excitation of the median giant fibre through its sensory connexions (preparation Fig. i). a, single shock delivered through first pair of electrodes,S 1 ;b, single shock delivered through second pair of electrodes, S2. Stimuli arrowed. Tracing of original oscillograph record. Further explanation in text.

Fig. 2.

L. terrestris. Action potentials produced by excitation of the median giant fibre through its sensory connexions (preparation Fig. i). a, single shock delivered through first pair of electrodes,S 1 ;b, single shock delivered through second pair of electrodes, S2. Stimuli arrowed. Tracing of original oscillograph record. Further explanation in text.

The fatigue characteristics of the sensory-to-giant junction were investigated by stimulating through the second pair of electrodes at a frequency of about 2 shocks per sec. In most preparations giant fibre impulses were given to the first 1 or 2 shocks, in a few cases to the first 3, and thereafter no further impulses were recorded from the giant fibre for as long as stimulation was continued (Fig. 3). It will be noticed that slow fibre activity which is also evoked by peripheral stimulation does not accommodate rapidly and continues to be recorded.

Fig. 3.

I., terrestris: accommodation at sensory-to-giant junctions. Action potentials recorded from exposed nerve cord in response to repetitive stimulation of segmental nerves through second pair of electrodes (S2 in Fig. 1). First three stimulus artifacts arrowed. Giant fibre responds to first stimulus only. Further explanation in text.

Fig. 3.

I., terrestris: accommodation at sensory-to-giant junctions. Action potentials recorded from exposed nerve cord in response to repetitive stimulation of segmental nerves through second pair of electrodes (S2 in Fig. 1). First three stimulus artifacts arrowed. Giant fibre responds to first stimulus only. Further explanation in text.

These experiments were repeated using different frequencies of stimulation in order to find the minimum period required to elapse between successive shocks in order that each one might produce an impulse in the median giant fibre. This was found to vary between preparations from 1·0 to 2·5 sec. and in one case it was over 3 sec.

These recovery times were consistently longer than those found from parallel kymograph experiments ; this may be due to operational trauma resulting from the dissection required for oscillograph recording.

Longitudinal muscle

That a site of fatigue is not be to located in the longitudinal muscle fibres was shown by inducing efferent fatigue of the giant fibre reflex by repetitively stimulating the giant fibres in the nerve cord and then stimulating the longitudinal muscle direct. It was found that although the efferent side of the giant fibre reflex had been fatigued the longitudinal muscle responded to direct stimulation by giving normal contractions (Fig. 4), indicating that the muscle fibres were unaffected and in an apparently normal physiological condition.

Fig. 4.

L. terrestris. Kymograph recording of longitudinal contractions obtained by direct repetitive stimulation of muscle after the development of fatigue on the efferent side of the giant fibre reflex. Frequency: a shocks per sec. Upper trace: 1 sec. time signals. Middle trace: stimulus marks. Lower trace: muscle records.

Fig. 4.

L. terrestris. Kymograph recording of longitudinal contractions obtained by direct repetitive stimulation of muscle after the development of fatigue on the efferent side of the giant fibre reflex. Frequency: a shocks per sec. Upper trace: 1 sec. time signals. Middle trace: stimulus marks. Lower trace: muscle records.

Neuromuscular junction

The features of neuromuscular transmission were investigated by stimulating the central ends of motor neurones in individual segmental nerves and recording the action potentials produced in the longitudinal muscle. The preparation was made by transecting the nerve cord on either side of the segmental nerve and lifting the piece of nerve cord on to a pair of stimulating electrodes (Fig. 5). Thus impulses have to traverse any junctions in the course of the motor neurone pathways as well as the neuromuscular junctions before reaching the muscle. Neon lamp pulses were delivered to the segmental nerve at frequencies ranging between 4 and 10 shocks per sec. Action potentials, registered by placing a pair of recording electrodes in contact with the surface of the muscle, were fed into a pre-amplifier and oscilloscope.

Fig. 5.

L. terrestris. Segmental nerve-muscle preparation for recording action potentials from the surface of the longitudinal muscle in response to stimulation of motor neurones in segmental nerve I. Further explanation in text.

Fig. 5.

L. terrestris. Segmental nerve-muscle preparation for recording action potentials from the surface of the longitudinal muscle in response to stimulation of motor neurones in segmental nerve I. Further explanation in text.

On repetitive stimulation at all frequencies tried each shock evoked an action potential (Fig. 6) : the first two or three invariably showed a slight increase in size after which they maintained a steady level for about 80 shocks (at 4 per sec.) and then gradually decreased in size. However, the rate at which they declined was extremely slow, and even after 700 shocks at 4 per sec. the potentials were reduced to only half the size of the first one. It was found to require several hundred more shocks at this frequency to abolish the response altogether. The recovery time was short, and after a rest period of 2 min. the muscle potentials were found to return to their original size and to respond to the first shock. However, the decline in magnitude invoked by repetitive stimulation was now more rapid.

Fig. 6.

L. terrestris. Longitudinal muscle potentials in response to repetitive stimulation of motor neurones in segmental nerve I (preparation Fig. 5). Frequency of shocks 4 per sec. Each muscle potential preceded by stimulus artifact inked over and dotted. The numbers indicate how many shocks have been already delivered.

Fig. 6.

L. terrestris. Longitudinal muscle potentials in response to repetitive stimulation of motor neurones in segmental nerve I (preparation Fig. 5). Frequency of shocks 4 per sec. Each muscle potential preceded by stimulus artifact inked over and dotted. The numbers indicate how many shocks have been already delivered.

These results were obtained from preparations whose longitudinal muscle no longer responded when excited through the giant fibres and they indicate that neither neuromuscular fatigue, nor failure at junctions in the course of the motor neurone pathways supplying the longitudinal muscle, are responsible for fatigue of the giant fibre reflex,

Giant-to-motor junctions

These junctions whose existence has not yet been confirmed histologically are assumed to be situated between the giant fibres and their motor connexions. Their transmission properties were investigated by repetitively stimulating the giant fibres and recording action potentials from the longitudinal muscle. Impulses set up in the giant fibres would be expected to have to cross the giant-to-motor junctions, if such exist, before reaching the longitudinal muscle.

The most useful preparation, one in which operational trauma was reduced to a minimum, was set up as shown in Fig. 7. The worm was pinned to a wax block with the ventral surface uppermost at the anterior end and the dorsal surface uppermost in the region posterior to the clitellum. The longitudinal muscle and a short length of the nerve cord were quickly exposed behind the clitellum and a pair of recording electrodes placed in contact with each. Stimulating electrodes were inserted through the mid ventral body wall at the anterior end and placed on either side of the nerve cord, in which position a single shock of sufficient intensity was found to evoke in both giant fibre systems between one and three impulses, probably as a result of stimulating the giant fibres directly and also through some of their sensory connexions. The intensity was kept below threshold for the lateral giants so that impulses were evoked in the median giant fibre only. Each pair of recording electrodes was connected to its own preamplifier and thence to an oscilloscope where the muscle responses were arranged to appear on the upper trace and the nerve impulses on the lower trace.

Fig. 7.

L. terrestris. Preparation for recording action potentials from the surface of the longitudinal muscle (Rec. 1) and from the exposed nerve cord (Rec. 2) in response to stimulation of the giant fibres in the cord (Stim.). Worm twisted so that ventral surface uppermost anterior to clitellum, dorsal surface uppermost behind clitellum. Further explanation in text.

Fig. 7.

L. terrestris. Preparation for recording action potentials from the surface of the longitudinal muscle (Rec. 1) and from the exposed nerve cord (Rec. 2) in response to stimulation of the giant fibres in the cord (Stim.). Worm twisted so that ventral surface uppermost anterior to clitellum, dorsal surface uppermost behind clitellum. Further explanation in text.

The muscle responses were found to vary greatly from preparation to preparation and also from one region of the muscle to another. In some cases the junctions had already failed as a result of the dissection, and no muscle potentials were initiated by giant fibre impulses. However, in favourable preparations repetitive stimuli at a frequency of about 1 per sec. produced very clear muscle potentials after a delay (20–25 msec.) which was so short that they could only have arisen from the giant fibre impulses (Fig. 8). Such muscle responses were produced for a short while in an irregular manner and thereafter no further responses could be recorded. The muscle potentials varied in size even when evoked by the same number of median giant fibre impulses and this suggests that in each segmental nerve there are several motor neurones each with its independent giant-to-motor junction. Increasing the intensity of stimulation so that the lateral giants are excited showed that, here also, rapid failure occurs with repetitive stimulation.

Fig. 8.

L. terratris. Longitudinal muscle potentials (upper traces) and median giant fibre impulses Gower traces) recorded simultaneously in response to repetitive stimulation of nerve cord (preparation Fig. 7). Frequency of shocks 3 per sec. Short, downward-projecting stimulus artifacts appear on upper traces (first one arrowed).

Fig. 8.

L. terratris. Longitudinal muscle potentials (upper traces) and median giant fibre impulses Gower traces) recorded simultaneously in response to repetitive stimulation of nerve cord (preparation Fig. 7). Frequency of shocks 3 per sec. Short, downward-projecting stimulus artifacts appear on upper traces (first one arrowed).

The experiments described above indicate that rapid fatigue of the earthworm’s escape response is not due to sensory adaptation or to intramuscular fatigue or to neuromuscular fatigue, or to failure at junctions in the course of the motor neurone tracts ; but is caused by rapid failure at sensory-to-giant and giant-to-motor junctions.

Rapid fatigue is characteristic of the giant fibre response of the earthworm and there is evidence that it characterizes the escape responses of other annelids.

Rapid habituation to repetitive tactile stimulation, probably as a result of accommodation at sensory-to-giant junctions, has been demonstrated in Myxicola and appears to occur in Branchiomma and other tubicolous polychaetes (Roberts, unpublished observations). Accommodation at sensory-to-giant junctions is also known to occur in Nereis and Harmothöe, though it develops slightly less rapidly than in Lumbricus, giant fibre impulses being recorded to the first 2 or 3 shocks delivered to the sensory fibres (Horridge, 1959).

Accommodation at giant-to-motor junctions does not occur in Myxicola (Roberts, 1962b) and this is correlated with the fact, demonstrated histologically, that the giant fibre has motor branches which directly innervate the longitudinal muscle fibres (Nicol, 1948). Rapidly accommodating giant-to-motor junctions occur in Nereis and Harmothöe (Horridge, 1959), though in both cases the process is less rapid than in Lumbricus.

The physiological junctions discussed in this paper have not yet been analysed histologically though in the past several attempts have been made to identify them. Concerning the sensory-to-giant junctions, Smallwood & Holmes (1927), working on Lumbricus terrestris and Eisema foetida, describe nerve cells situated in the lateral region of the cord, each of which sends a main branch to the lateral giant fibre of the opposite side and numerous branches to the neuropile of the same side. As the latter is closely associated with the segmental nerve roots the suggestion is made that the neuropile has synaptic connexions with sensory neurones and that the main branch of the nerve cell ensures a route to the effector mechanism of the other side. A more detailed picture is given by Ogawa (1939) in Pheretima. Contra-and ipsi-lateral sensory fibres are described which enter the cord from all three segmental nerves and divide into anterior and posterior branches which fie in close proximity to the giant fibres and give off numerous collaterals en route. It is possible that the sensory-to-giant junctions are to be located in these regions.

Concerning possible giant-to-motor junctions, large bipolar nerve cells, situated ventrolaterally in the posterior region of the nerve cord of Lumbricus, are described by Friedlander (1888) as sending the outer branch through a sinusoidal path to the lateral giant fibre of the same side. However the course of the inner branch is not described and the picture is misleading since a relationship is described between the median and lateral giant fibres which more recent physiological evidence has shown does not exist (Rushton, 1945). Evidence of a connexion between the median giant and a ‘dorsal nerve bridge’ in the middle region of the nerve cord is also presented. Since the dorsal bridge is associated with the segmental nerves it has been suggested that this may represent a possible motor pathway for the median giant fibre. A large bipolar motor cell, situated in the ventrolateral region of the cord is described by Smallwood & Holmes (1927). One of its processes connects with the lateral giant fibre of the same side whilst the other one crosses the cord and joins the outgoing fibres in the segmental nerve of the opposite side. It is possible that the point of contact between the first process and the giant axon represents the giant-to-motor junction. A detailed account of the neurone anatomy of the nerve cord of Pheretima is given by Ogawa (1939) from methylene blue whole-mount preparations. Branches of both the median and lateral giant fibres are described. Those of the median giant were not followed very far but the branches of the lateral giants, having given off many collaterals, are said to pass to the other side of the cord and enter the posterior trunk of the ‘double nerve ‘(segmental nerves II and III). It is impossible at the moment to say whether the relationship between the giant fibre and its branch involves a continuity of axoplasms or a structural synapse (Ogawa, personal communication).

Failure at sensory-to-giant junctions appears to occur in most annelids that possess a giant fibre system and its function seems to be that it prevents the animal from being dominated by its escape response as a result of excessive excitation of the giant fibres. The value of giant-to-motor accommodation is less obvious. In an active burrowing animal it is essential that the effector mechanism should not be liable to over-excitation and subsequent fatigue for the muscle is always being called upon for the performance of crawling and so on. The value of giant-to-motor accommodation might therefore be to protect the muscle from excessive excitation through the giant fibres. Accommodation at the sensory-to-giant junctions might be thought to render this unnecessary and perhaps giant-to-motor accommodation should be regarded as a safety mechanism whose function is to prevent over-activity on the efferent side of the giant fibre reflex should this not be ensured by the development of an afferent block.

Sensory-to-giant and giant-to-motor accommodation indicate the existence of morphological junctions in these regions of the reflex. This confirms the suggestion made by Bullock (1945) that the giant fibres of Lumbricus probably represent internuncial rather than efferent neurones.

I acknowledge with gratitude the receipt of a grant from the Department of Scientific and Industrial Research. Some of this work was done at the Gatty Laboratory, St Andrews, and I owe grateful thanks to the Director, Dr G. A. Horridge, for putting the facilities of his laboratory at my disposal. I am also indebted to Prof. C. F. A. Pantin, F.R.S., who supervised this research, for his help and encouragement.

Bullock
,
T. H.
(
1945
).
Functional organisation of the giant fiber system of Lumbricus
.
J. Neurophysiol
.
8
,
55
71
.
Friedlander
,
B.
(
1888
).
Beitrfige zurKenntnis des Centralnervensystems von Lumbricus
.
Z. wiss. Zool
47
.
47
84
.
Hess
,
W. N.
(
1935
).
Nervous system of the earthworm, Lumbricus terrestris L
.
J. Morph
.
40
,
335
59
.
Horridob
,
G. A.
(
1959
).
Analysis of the rapid responses of Nereis and Harmothne (Annelida)
.
Proc. Roy. Soc. B
,
150
,
245
63
.
Nicol
,
J. A. C.
(
1948
).
The giant nerve fibres in the central nervous system of Myxicola (Polychaeta, Sabellidae)
.
Quart. J. Mier. Sci
.
89
,
1
45
.
Ogawa
,
F.
(
1939
).
The nervous system of the earthworm (Pheretima comrnunissima) in different ages
.
Sci. Rep. Tohoku Imp. Univ. (Series 4)
,
13
,
395
488
.
Roberts
,
M. B. V.
(
1962a
).
The giant fibre reflex of the earthworm, Lumbricus terrestris L. I. The rapid response
.
J. Exp. Biol, (in the Press
.)
Roberts
,
M. B. V.
(
1962b
).
The rapid response of Myxicola infundibulum (Grube)
.
J. Mar. Biol. Ass. (in the Press)
.
Rushton
,
W. A. H.
(
1945
).
Action potentials from the isolated nerve cord of the earthworm
.
Proc. Roy. Soc. B
,
133
,
433
37
.
Smallwood
,
W. M.
&
Holmes
,
M. T.
(
1927
).
The neurofibrillar structure of the giant fibres in Lumbricus terrestris and Eisenia foetida
.
J. Comp. Neurol
.
43
,
327
45
.