The Reflex Mechanism of the Insect Leg

In attempts to resolve animal behaviour into simpler terms much weight has always been given to the concept of the reflex as a unit. Since the classical work of Sherrington and others (Creed, etc., 1932) on the responses of the spinal vertebrate, many authors have supposed it to be merely a matter of time before all behaviour would be interpreted in such terms. In recent years, however, another school of thought seems to be gaining ground ; in the works of Coghill (1929) and v. Holst (1936) the idea of patterns of behaviour, inherent in the nervous system and influenced only to a limited degree by external events, is given increasing prominence, while the reflex is relegated to second place. Gray & Lissmann (1938,EXBIO_17_1_8C6a, EXBIO_17_1_8C7b), in studies of the locomotion of the earthworm and leech, discuss at some length the relative merits of the two concepts and come to the conclusion that neither can be excluded from the general picture, though one may be more important than the other in special cases. It is the purpose of this paper to enquire how far reflexes in the true sense of the word are present in the make-up of a running insect (cockroach).

Previous evidence for the existence of reflexes in the insect leg may be divided into subjective and objective classes. Many authors have found it necessary to postulate reflex mechanisms in order to explain their observations of the grosser aspects of insect behaviour, but few have produced any objective evidence for their existence. Of the former only those which have a direct bearing on the present work will be considered here.

Hoffmann (1933) described a righting reflex in the cockroach which was initiated by absence of contact of the legs with the ground. Although he did not manage to decide between tactile and pressure stimuli as the normal means of inhibiting this reflex, he claimed that the fact that the pawing movements are immediately stopped in any leg that comes in contact with a rigid object is evidence that there must be some local response system for each leg, in fact what may be called a reflex.

Fraenkel (1932) and v. Holst (1935) describe more elaborate effects of contact of the legs with the ground on the regulation of flight and walking movements respectively. Fraenkel showed that the movements of flight are inhibited immediately by contact stimuli to the legs, while v. Holst maintained that it is the proprioceptive influences from the legs that regulate the general co-ordination. But the most important work, where the assumption of local responses becomes essential for the hypothesis put forward, is that of Crozier & Stier (1927-8, 1928-9) on geotropic orientation. These authors showed that, in spite of the absence of any special gravity-sensitive organ, insects are capable of accurate geotropic orientation, and postulated instead elaborate proprioceptive reflexes, working as they thought from sense organs in the muscles, and enabling the force exerted by each leg to be related directly to the proportion of the body weight borne by that leg. No actual proof of the existence of such reflexes was, however, given.

Actual evidence of the presence of reflexes in the leg system was provided by the work of Rijland (1932a, b), who recorded electrical ‘‘impulses “in the leg muscles of a number of insects and showed that their frequency was influenced by movement of the joints, and also by changes in the position of the centre of gravity of the body during respiratory movements. His results will be more fully discussed during the course of the paper in the light of the observations to be described.

The electrical changes in the muscles of the legs of the cockroach (Periplaneta americana L.) have been recorded with the aid of a condenser-coupled amplifier and Matthews oscillograph. As was shown in a previous paper (Pringle, 1939), most of the muscles of the cockroach leg are innervated by only two nerve fibres, impulses from one producing a small electrical effect in the muscle and a tonic contraction, those from the other a larger electrical effect and a rapid twitch contraction. From the oscillograph records from the muscles it is easy to distinguish the two types of electrical effect, and the frequency of the spikes may be taken as an indication of the state of excitation of a single unit in the central nervous system.

For the experiments the cockroach is held on its back on a paraffin block by pins through the sides of the thoracic tergites. When intact insects were used the legs other than the one under study (usually one of the metathoracic legs) were allowed to make contact with a strip of paper across the body. Unless this is done the animal makes periodic attempts to right itself and these upset the results. For experiments on the isolated metathoracic ganglion the connectives were cut in front and behind with the aid of a pair of scissors, and in some experiments the whole of the ganglion was exposed for electrical stimulation of the nerves.

The main coxal extensor trochanteris and the extensor tibiae muscles have been used respectively as examples of depressors and levators. The former is innervated from the large, the latter from the most anterior of the small nerves to the leg. Each has one “quick” and one “slow” nerve fibre (Pringle, 1939). The variation in the frequency of the slow fibre discharge and the incidence of bursts of impulses in the quick fibre were followed under various types and intensities of sensory stimulation.

In the intact preparation a record from the extensor trochanteris muscle (depressor) of the metathoracic leg usually shows a steady rhythm in the slow fibre at about 5-20 per sec., even when the leg is free. If the contraction of the depressor muscles is resisted by contact of any part of the leg with a solid object, an increase in frequency is to be observed. Rijland (1932 b), in the course of a study of this effect in Hydrophilus, concluded that forced movements of flexion or extension were necessary to produce the reflex, and that a lasting increase in frequency in response to flexion occurred only in the isolated ganglion preparation, the intact animal becoming “adapted” to the new position. His results did not seem to show exactly what were the conditions for the reflex response.

A number of experiments were therefore performed to elucidate more clearly which sense organs in the leg were responsible for the effect, and exactly what relation held between the movement or position of the limb and the frequency of the tonic discharge. The results are given below in summary form.

• (1) Flexion at either the coxo-trochanteral, femoro-tibial or tibio-tarsal joints produces an increase in frequency, as also does any bending of the leg at the trochantero-femoral hinge. In order to simplify the problem the leg was therefore amputated at the distal end of the femur, and §§ (2) and (3) below refer to such a preparation.

• (2) Fig. 1 A, B, show records taken during a sudden decrease and increase respectively of the resistance to extension of the femur. The femur was moved by a metal rod attached to a solenoid and wired in series with the marker on the record in order to give the exact moment of the stimulus. In this way while movement of the rod in one direction flexed the femur and increased the tension on the extensor trochanteris muscle, movement the other way merely released the tension and did not forcibly extend the leg, which took up its new position gradually as the muscle relaxed. Results of similar experiments are shown graphically in Fig. 2A, B, C.

  In general, a sudden reduction in the resistance to extension causes the motor impulses to cease momentarily and then restart at a lower frequency; with a sudden increase there is a short high-frequency burst, sometimes followed by a “silent period “, and then the discharge continues at a higher level than before. With slower changes in pressure the frequency changes gradually to its new level; only over a very long period is there any “adaptation” to the new conditions.

• (3) Experiments were performed to determine the nature of the sense organs responsible for the reflex. Three possibilities were considered :

  • (A) Internal endings in the coxa—muscle endings or chordotonal organs.

  • (B) Position receptors—the trochanteral hair plate (Pringle, 1938b).

  • (C) Stress receptors—the trochanteral groups of campaniform sensilla (Pringle, 1938a).

The results are summarized below

• (a) Cautery of the nerves in the trochanter stops the reflex. After this treatment the muscle could still be excited normally by other types of reflex stimuli from the body and endings in the coxa should not have been harmed.

• (b) The apodeme of the extensor muscle was cut and attached to a thread. Tension on this muscle, provided the rest of the leg did not move, had no effect on the frequency of the discharge.

  These two experiments eliminate (A) above.

• (c) It can be seen in this preparation that the frequency of the discharge is quite independent of the initial position of the leg. Moreover, in the experiments described above it is to be noted that the cessation of the motor discharge occurs immediately on removing the resistance to extension of the femur although the femur does not instantaneously take up its new position owing to the muscle viscosity. There is thus no real correlation between the changes in frequency of the discharge and the actual position of the joint.

• (d) It was noted by Pringle (1938a) that the most intense excitation of the campaniform sensilla could be obtained by pressure of the cuticle of the segment in which they lie. Pressure on the cuticle of the trochanter has a profound influence on the frequency of the motor discharge to the extensor muscle, modifying it in the same way as movement of the leg (Fig. 2D). With strong steady stimuli the frequency of the reflex discharge declines at a rate which corresponds closely with the adaptation rate of these campaniform sensilla.

• This response to pressure can be obtained when there is no movement of the trochanter on the coxa and also when the apodeme of the extensor muscle is cut. The conclusion is therefore reached that it is the campaniform sensilla on the trochanter that are responsible for the reflex.

• (4) As was also pointed out by Pringle (1938a), most of the campaniform sensilla on the leg of the cockroach are arranged in such a way as to be excited when the insect is standing on the ground in the normal position. The evidence that these sense organs are responsible for the reflex influence on the tone of the depressor muscles of the leg completes the objective proof of the existence in the insect of a proprioceptive mechanism analogous to that of vertebrates.

Under this heading may be considered two reflex effects that have been observed in the extensor tibiae and other levator muscles.

Records from the levator muscles during stimulation of the campaniform sensilla show a variation in the frequency of the slow fibre impulses that bears a close inverse relation to that in the depressor set (cp. Rijland, 1932 b). Fig. 3 A shows the effect on the frequency of the slow fibre impulses to the extensor tibiae of short duration electric shocks applied to the central end of the large leg nerve of the same side. This nerve contains other fibres besides those from the campaniform sensilla, but the reflex inhibitory effect is well shown (isolated ganglion preparation).

With light stimulation of the campaniform sensilla by pressure on the trochanter the extensor tibiae slow fibre discharge may merely be reduced in frequency for the duration of the stimulus, returning thereafter to its normal level, but after strong stimulation there is a marked rebound, which may affect also the quick fibre. Fig. 3 B shows this result in a preparation where there was no resting discharge. This rebound corresponds with the silent period in the depressor reflex and may play an important role in normal locomotion (see below).

Tactile stimuli when applied to the upper surface of the tarsus or to the tibial spines lead to a momentary lifting of the leg so that it is placed on top of the stimulating object. With light stimuli only the slow fibres of the levator system are involved ; stronger stimuli bring in the quick fibre also and may initiate the rhythmic pawing movements of the “Suchenreflex” (Hoffmann, 1933), particularly in the decapitated animal. In either case excitation of the levator fibre is accompanied by a corresponding inhibition of the depressor sets.

The antagonism of the depressor and levator systems is well brought out by the balance that can be struck between them. Simultaneous stimulation of sense organs on the leg tending to elicit opposite responses may cancel out. The most usual case of this is when the leg is pressing against a solid object with one of the backwardly directed tibial spines. Stimulation of these spines alone elicits the levator response, but the force of resistance to depression of the leg also excites the campaniform sensilla of the trochanter and the balance is held on the depressor side. Further excitation of the spines can ultimately produce the levator response in spite of the simultaneous excitation of the campaniform sensilla (cp. the results of Morson & Phillips (1936) on the reflex antagonism between the A and B endings in the cat vasto-crureus muscle).

There is a slight but definite effect of stimulation of the campaniform sensilla of one side, on the frequency of the slow fibre impulses of the other side. Fig. 1C shows a record from the right metathoracic extensor trochanteris (depressor) while the leg on the opposite side was moved passively with rhythmic extension and flexion. The frequency is regularly reduced during the forced flexion movements.

In the isolated ganglion preparation electrical stimulation of the central end of the large sensory nerve while recording from the extensor tibiae of the other side produces a slight increase in the frequency of the discharge, followed always after an interval of about half a second by a silent period. This may be the accompaniment in the levator muscles of a rebound excitatory burst in the depressors.

In contrast to these results is the complete failure of many attempts to show reflex effects transmitted along the cord from one leg to another. In the intact preparation, handling or electrical stimulation of any part of the body produces a temporary rise in both the depressor and levator discharges, but no regular effect on the depressor-levator balance of one leg could be observed in response to any form of stimulation of the sense organs of the legs of another segment. The three pairs of thoracic legs appear to be completely independent as far as their reflex systems are concerned.

As has been stated above, any stimulation of the body of the insect tends to increase the frequency of both the depressor and levator fibres. Some types of stimuli, however, have more specific effects.

Touch on the side of the abdomen decreases the discharge to the depressor muscles of the same side and increases those of the opposite side. This would have the effect in the normal standing position of moving the abdomen away from the stimulus.

A marked effect is produced by draughts of air on the anal cerci (cp. Pumphrey & Rawdon-Smith, 1937). Each puff of air excites the depressor muscles of all the legs and evokes also a burst of impulses in the quick fibres, producing a co-ordinated jump forwards.

Section of the connectives in different places has very constant effects on the reflex behaviour of the preparation.

Removal of the head produces a state of hyperactivity of all the slow fibres, but particularly the depressors. The animal stands in a position reminiscent of decerebrate rigidity in the mammal. The leg reflexes are all present but levator responses tend to merge very readily into states of disco-ordinated activity in which all the legs are moved violently to and fro. The response to draughts is increased and the insect appears generally to be in a hypersensitive condition.

Isolation of the ganglia (experiments have mostly been done on the metathoracic ganglion) has a very different effect, the general tone being much lowered and the emphasis now rather on the levator sets. Some such preparations will show a slow resting discharge to the levator muscles and none to the depressors; others have no tonic activity at all. Nevertheless the reflex effects are still present and normal, and the preparation appears merely to be somewhat inexcitable.

Two points in particular arising out of the foregoing account seem to need further discussion; namely, the question of the excitation of the quick and slow fibres, and the origin of rhythmic movements.

In general it seems that the quick fibre is called into action by the same types of reflex stimuli that excite the slow fibre, but that it requires a greater intensity of excitation. In particular the quick fibre seems to be involved in rebound phenomena and rhythmic movements, while the slow fibre responds to prolonged reflex stimulation. There are no facts that would not fit a scheme in which the two fibres are supposed to have the same central connexions, but a difference in threshold and accommodation. A sudden rise and fall of central excitation will then excite both fibres, while a slower increase will normally bring in only the slow. Observations on peripheral nerves have shown that fibres with the greatest capacity for accommodation are also most liable to be excited by “off” effects, and this may explain the sensitivity of the quick fibre to rebound. Otherwise it is most easily called into action through the “pseudo-giant fibre” system from the anal cerci (Pumphrey & Rawdon-Smith, 1937); possibly the large size of these fibres causes a large and sudden rise in central excitation on the arrival of each impulse at the motor centre.

On the question of the nature and origin of the rhythmic movements that play such an important part in normal locomotion, it is by no means so easy to speculate. In some of the preparations studied there have been signs of rhythm in a sort of double or treble rebound after the cessation of sensory stimulation, particularly in the decapitated animal after the levator response to touching the leg. Fig. 4 shows the result of one experiment where an attempt was made to determine accurately the conditions for this phenomenon. The preparation showed a steady slow fibre discharge to the depressor muscles, and this could be momentarily inhibited by movement of a single tibial spine (levator reflex as described above). By gradually increasing the intensity of sensory stimulation it was possible to reach a point where the inhibition, instead of being followed by a rapid return to the normal frequency, was followed by a short high-frequency burst, involving also the quick fibre, followed by another silent period and then a return to normal.

A still higher intensity of stimulation would produce several of these alternating bursts and silent periods before the insect again quietened down. It was as though each excitatory burst, alternatively to the depressor and levator systems, stimulated the next to activity, but that the rebound was not quite sufficiently strong to keep up the rhythm for more than a certain period against a natural damping effect. The stronger the initial stimulus, the longer could this rhythm persist. Possibly in the normal walking movements sensory stimuli are continually arriving at the motor centres in sufficient intensity to maintain the alternation.

It should be noted in this connexion that rhythmic movements of the legs do occur even when they are out of contact with the ground (“Suchenreflex” of Hoffmann, 1933). Under these conditions there is no stimulation of tactile or campaniform sensilla and the rhythm must either be spontaneous or generated by reflexes other than those described here, possibly from the hair plate or chordotonal sensilla. Such a state of activity could, however, be produced in each of the leg systems by an increase in the central rebound phenomena. Further elucidation of the nature of this effect and of its control by higher centres may give the clue to the origin of spontaneous movements.