ABSTRACT
The prothoracic sternal spine of Schistocerca is described. The second paired nerve innervates the sensilla of the sternum, coxae, and femora, and in subimaginal stages supplies chordotonal organs registering movement of the cervical membrane. Each nerve fibre serves the axons of several sensilla.
When the sensilla are waxed, control of the legs is disturbed.
When the prothoracic ganglion is isolated from the CNS stimulation of the sensilla produces complex, oriented, reflex cleaning movements by the legs.
Recordings were made from nerve 2 and from the commissures. Co-ordination is mainly segmental; the signal in the commissures is much reduced. At normal levels of stimulation there is no activity in nerve 2 unless several sensilla are excited.
Among related genera there is some correlation between manipulative ability and the number of sensilla on the spine. It is suggested that these sensilla contribute to the co-ordination of movements of the front legs.
INTRODUCTION
The family Catantopinae (= Cyrtacanthacrinae) of the acridid grasshoppers have a raised tubercle or spine on the sternum of the prothorax, between the prothoracic legs. Its shape is typically a short, thick, slightly backwardly directed spine, with many hair sensilla; this is the condition in Schistocerca (Pl. 1, A, B). In other genera, however, it is relatively larger (e.g. Tropidacris) or nearly absent (Mesopsera), blunt (Tropidopola) or very sharp (Opshomala) : in some it is of entirely different shape, e.g. bifurcate in Goniae australasis, or mushroom-shaped in Noliba elegantula.
Uvarov (1928) commented that, while no investigations had been made, the structure had no obvious biological significance. Chauvin (1939) described the gross anatomy of the spine in Schistocerca, and did some extirpation experiments, but reached no conclusions as to its function.
MATERIAL AND METHODS
Schistocercagregaria from laboratory culture has been used for most of the work herein des cribed, but the much larger Omithacris turbida has also been used, particularly for anatomical work. The nucleus of the latter culture was kindly supplied by the Anti-locust Research Centre.
Detailed dissection of the nervous system, especially of adult Schistocerca, is hindered by a tough sheath of connective tissue, tracheae and fat body around nerves and ganglia. For this reason dissection of fine sensory nerves was only practicable in material which had been fixed for some months in Bouin’s alcoholic fluid to which had been added a little glycerin. This dissolved the fatty material, leaving only a thin fibrous sheath, and did not make the tissues too brittle.
The prothoracic spine and neighbouring structures are heavily cuticularized. Later instars were therefore killed soon after moulting, and embedded in ester wax, which produces minimal hardening. Sections were silver-stained by the methods of Holmes (Carleton & Drury, 1957), Samuel (1953), or Betchaku (1960), after fixation in Camoy’s, Davenport’s (Samuel, 1953) or Bouin’s aqueous or alcoholic fluids, or one of various formol mixtures. Some material was stained in bulk by a modification of Holmes’s technique. All these combinations gave successful results on occasion, and none was completely reliable. The most consistent method was fixation in Davenport’s solution for several days, followed by Samuel’s method with 4 hr. in the impregnating solution.
For fibre counts, fixation was in 2% osmium tetroxide saturated with picric acid, 3 days at 0° C., and subsequent staining in iron haematoxylin.
Electrical activity in the ventral nerve cord was recorded with hooked silver/silver chloride electrodes and in smaller nerves with hooked stainless steel electrodes, 20 μ in diameter, set in Araldite about 100 μ apart. Saline was removed from the nerve/ electrode junction with a fine pipette, and replaced when recording was satisfactory by a drop of liquid paraffin. As long as this did not enter the tracheal system, the preparation lived for up to 2 hr. Sensilla were stimulated either by a fine glass needle attached to the armature of a relay, or by puffs of air controlled in duration and intensity by relay-operated taps. The movements of the former were calibrated with a phototransistor, and the puffs of air measured by directing the jet at a diaphragm microphone, both feeding into the oscilloscope amplifier.
STRUCTURE AND DEVELOPMENT
The sternum of the prothoracic segment carries a roughly Y-shaped area of dense cuticle (Text-fig. 1). The tail of the Y is the spina-, near its anterior end the sternal apophyses run sideways and upwards as flying buttresses to the pleural walls, and from their point of attachment to the sternum two ridges run to the left and right anterior comers of the segment, where they fuse with the transverse anterior ridge, which forms the anterior margin of the segment and joins the tips of the Y. At the junction of the arms, at the tip of the spina, is a depression in the sternal floor which leads backwards and downwards to form the sternal spine. The outside of the spine bears (in Schistocerca) about 100–140 symmetrically distributed, long, articulated setae, up to o-8 mm. long, and a few much smaller setae, about 0·05 mm. long. No other sensilla are present. Similar hair sensilla are present on the rest of the sternum and proximal leg joints (Pl. 1, A, B).
Under the sternal cuticle lies a single layer of ectodermal cells, and on this a layer of small, flattened, irregular air-sacs, bound together by connective tissue—this layer) extends to the tip of the spine. On top of these small air-sacs lie the longitudinal ventral tracheae (LVT). Posteriorly they run with the posterior nerve commissures between spina and apophyses, and on the inner surface at the anterior end they feed two pairs of large air-sacs, which end blindly on the sternal floor anterior to the spine. The more posterior (second) pair lie under the segmental ganglion, which is supported on a cushion of minor air-sacs and tracheae fed from the dorsal surfaces of the LVT (Text-fig. 2).
The nerves leaving the ganglion are shown in Text-fig. 3. The sternal sensilla are supplied solely by nerve 2. The nerve emerges from the ganglion together with a small trachea, and both run vertically downward to the inner surface of the LVT, at the junction of air-sacs 1 and 2 (Text-figs. 2, 3). Here the nerve enters a patch of connective tissue attached to the tracheal wall. From this patch an anterior fibrous strand runs forward to an insertion on a specially thickened patch in the midline of the cervical membrane. The nerve divides inside the connective tissue. The larger branch turns under the LVT, and emerges between the main tracheal trunks supplying the legs. It runs to the anterior ridge, where it splits up, supplying the sensilla along the anterior and lateral margins of the sternum. A subsidiary branch runs backwards to the leg, and supplies hair sensilla on the coxa and femur. The smaller branch runs downwards and slightly backwards from the LVT, together with a posterior fibrous strand, which is inserted distally on the cuticle at the anterior base of the sternal spine. From this point the nerve branches freely over the ectoderm to the sensilla of the spine and mid sternum.
The spine is absent in the first instar, and develops progressively at each moult. In the first two instars the tracheal system is rudimentary, but the nerve has the appropriate relations with the tissue blocks inside which the tracheae develop. By the third instar (Text-fig. 4) the main tracheae are complete, but the second nerve is still not divided into two main branches as in the adult. From the point of contact with the LVT the nerve branches many times and supplies all areas of the sternum. The nerve also supplies a large chordotonal organ, containing about fifteen sense cells. This is attached by connective tissue to the point of contact of nerve and LVT, and is inserted on the cuticle at the anterior base of the sternal spine. In the first instar the chordotonal organ is fully formed and innervated, but as the LVT are not yet formed, its origin is diffuse. In the fifth instar the chordotonal sense cells often appear to be degenerating, and in the adult all that remains is the posterior fibrous strand, which preserves the origin, relations and insertion of the original organ. Exceptionally the organ does not degenerate, and remains apparently functional in the adult. The anterior fibrous strand is by contrast fully developed and non-nervous throughout life.
As the insertion of the chordotonal organ is on rigid cuticle, it can only monitor movements of the point of origin, i.e. the LVT, relative to the insertion. Such movement could possibly occur during moulting (perhaps a reason for degeneration after the final moult), but it is more likely that the normal function is to register movement of the cervical membrane, transmitted to the LVT by the anterior fibrous strand.
The second nerve is shown in transverse section in Pl. 1, C, immediately before it enters the ganglion. It contains some thirty-seven fibres of 5·9μ diameter, and seventeen smaller ones down to 1 μ. The paired second nerves, with their total of 108 fibres, supply in all about 600–700 sensilla, so it is clear that in many cases the axons of several sense cells must fuse to form a single fibre. This is confirmed by silver-stained sections. The lateral branch of the nerve supplies only its own side of the segment, but in the midline each nerve receives fibres from both sides of the spine and sternum. It has not been possible to decide whether axons from both large and small sensilla ever fuse.
The path of the fibres from nerve 2 within the prothoracic ganglion is shown diagrammatically in Text-fig. 5. They appear to synapse soon after entering the neuropile with internuncial neurones which are associated with the motor fibres of, at least, nerve 3, and also make connexion with a vertical tract of fibres which continue into the anterior connective and thus to the suboesophageal ganglion. No connexion with fibres running to the posterior connectives was found. These established connexions, though certainly incomplete, would fit in well with the behavioural and physiological results described below.
BEHAVIOURAL EXPERIMENTS ON FUNCTION
Extirpation of the sternal spine by cautery under anaesthesia in the second or later instars produced no obvious effect on the subsequent behaviour of the animals; this agrees with the results of Chauvin (1939). If, however, the spine of an adult locust was covered with low melting-point paraffin (40° C.) without anaesthesia, similar disturbances were seen in a significant number of cases. The movements of the prothoracic legs appeared to be disturbed; in extreme cases the legs would ‘windmill’ when walking, making very exaggerated stepping movements (Text-fig. 6A). When stationary, these animals would often stand with one or both prothoracic legs off the ground. (Text-fig. 6B). (These effects are also seen more commonly in animals which have one or both commissures sectioned between the prothorax and head ; the disturbances are then permanent.)
These effects soon vanished, usually within a few hours, sometimes within 30 min. If the waxing was performed under CO2, or less drastically, N2 anaesthesia, the effect was masked by the general lack of co-ordination and disability during recovery, and was gone by the time recovery was complete. Direct recording from nerve 2 confirmed that the wax coating stopped the sensilla functioning in response to normally effective stimuli.
The results suggest that the animal is deriving proprioceptive information relating to the prothoracic legs from the spine sensilla, but that other sources are available, as the animal quickly adapts to its loss. The long sensilla on the spine touch hairs on the coxae and mesothorax, and can presumably signal movement of these parts relative to the prosternum (Pl. 1, A, B).
SEGMENTAL REFLEXES
Stimulation of the sensilla of the sternum in the intact animal does not produce very useful results because the response of the animal varies greatly with circumstances, especially the degree to which the animal is disturbed by experimental conditions. If much disturbed, as when restrained and strongly illuminated, it would ignore the most violent stimulation, e.g. cautery, and continue its attempts to escape. In more favourable conditions light stimulation with a fine glass needle produced grooming movements of the prothoracic legs, or movement away from.the needle.
When the ventral nerve cord is sectioned, both anterior and posterior to the prothoracic ganglion, a precise and strictly repeatable system of reflex responses is seen. Movement with a needle of one or more sensilla on the spine produces an effective grooming movement of the legs, in which the tarsal claws are scraped across the area of stimulation. Either the ipsi- or contralateral leg may be used; continued stimulation brings both legs into action. This is true of all sensilla near the midline. Towards the lateral edge of the sternum, the ipsilateral leg is always brought first into action ; this agrees with the anatomical connexions described. Stimulation of each part of the sternum and accessible parts of the pleuron produces different, correctly oriented grooming responses; stimulation of the leg itself produces a movement of that leg to withdraw the stimulated area from the needle. A leg never attempts to remove a stimulus from a contralateral leg, as it does from the sternum, unless it is the inside base of the coxa which is stimulated.
Complete grooming responses are given to stimulation of a single hair, provided the displacement is considerable (cf. results of direct recording, below).
Thus a complex series of reflex arcs exist within the prothoracic segment involving input along nerve 2 and motor output along the leg nerves. The system reacts differently to different areas of stimulation, which indicates some form of spatial representation in the input along nerve 2. Similar systems must also exist around the sensory nerves from other joints of the leg.
Stimulation of spine hairs in this way always produced grooming movements, and not the type of response one would expect if the stimulation corresponded to normal proprioception. A different quality of stimulation would be expected in the latter situation, as input will be received from the hairs which engage the spine hairs as well as from the spine hairs themselves; but the fact that proprioceptive disturbances are produced by section of the commissures to the head might imply that the head ganglia are involved in complete proprioceptive responses. Certainly it can be shown by direct recording that sensory information from both spine and legs ascends to the head.
DIRECT RECORDINGS FROM NERVES
Preparation
The prothoracic sternal floor, together with a small flap of mesothoracic sternum at its posterior margin and the ventral cervical membrane at its anterior margin, was cut from the animal. In most experiments this piece was pinned to a pedestal of plasticine to allow dissection of the nerves from above, and maximum access to the sensilla of legs and sternum. As, however, nerve 2 derives from the sensilla distributed over the entire sternum, and the three proximal leg joints, this method continually excites some sensilla. For more critical work, the preparation was suspended from two horizontal Nichrome wires passing under the sternal apophyses, with only the tarsi in contact with the plasticine base.
RECORDS FROM NERVE 2
Owing to the small size of the nerve it was not possible to record with external electrodes from the larval instars or from separate branches of the adult nerve. Records were made from the nerve of adults immediately before it enters the ganglion ; here it runs free for some 300 μ, and can be lifted on fine hooked electrodes.
No response was detected to radiant heat, or to the vapour of xylene, benzene or pyridine, or to pure tones from 20 eye./sec. to 10 kilocyc./sec. The hairs are probably too long and flexible to be adequately displaced by sound waves.
Slight or medium deflexion of a single hair on the spine or elsewhere produced no discernible activity. An angular deflexion of 70° or more produces an initial burst of large spikes lasting about 0·05 sec., and then there follows a rhythmic sequence of short bursts each of about five large spikes in 0·02 sec., at a constant rhythm of about 30 per second. The frequency varies from hair to hair—25 to 36 per second have been recorded—but is constant for any given hair. Under the experimental conditions, no adaptation took place during a maintained displacement of 8 min. The activity ceases immediately the hair is replaced—there is no after-discharge (Text-fig. 7, A).
When a group of hairs, three or more, was moved simultaneously, a small discharge occurred. This was a short burst of 10–20 large spikes and a longer lasting burst of small ones, of about 1/5 the amplitude of the former. The same effect was seen on replacement. The frequency of the larger spikes depended on the rate of movement of the hairs (Text-fig. 8B). Similar results were obtained from several other groups of hairs, including both ipsi- and contra-surfaces of the spine. Raising or lowering of the flap of mesosternum relative to the preparation caused hairs present along the anterior margin of the former (which are innervated from the mesothoracic ganglion) to move some of the hairs on the prosternal spine. The activity recorded in nerve 2 resembled that obtained by moving a group of hairs with a needle as described above. A complex burst of large and small spikes occurred at both displacement and replacement; faster movement gave a higher frequency of impulses (Text-fig. 7B). The activity did not persist more than 0·2 sec. after the end of movement in either case. When some hairs of the lateral surface of the spine were stimulated by moving a thin glass rod horizontally towards the midline, simulating an observed movement of the coxa, another complex burst appeared, but of different form. Here the large spikes predominated during displacement and the small during replacement.
When the majority of the sensilla were moved synchronously, as by a puff of air, a large and complex burst was recorded. Under these conditions a very small movement of the individual hairs suffices for a complete response. The number of fibres involved increased a little with strength of stimulation, but large, medium and small spikes were always present. The system is very sensitive to change in intensity of air currents; a weak stimulus, 0-2 sec. long, gave quite a large burst but little after-discharge, while a strong one of the same length produced a burst not many times larger but with an after-discharge which was still twice the resting level 3 sec. after stimulation had ended (Text-fig. 7D, 8A).
The properties of the sternal and coxal hairs are the same as those of the spine. The sensilla of the femur are mainly of the short type, which are not easily manipulated, with a needle, but they give the same response to air currents as do the larger ones.
RECORDS FROM COMMISSURES
If the sternal or coxal sensilla are stimulated and activity recorded in the anterior or posterior commissures, a great burst of impulses is recorded. These, however, are almost all correlated with the reflex movement of the legs. If the leg nerves are cut, the response in the anterior commissure is greatly reduced and that in the posterior commissure abolished. When the sensilla are covered with wax, or when the second nerves are also cut, this activity vanishes from the anterior commissure, leaving only the spontaneous activity of the ganglion.*
The results described below were obtained from ganglia which were completely de-afferented except for the second nerves.
Medium deflexion of a single hair produces no detectable activity in the commissure. The same is true of extreme deflexion of a single hair, which gives continuous rhythmic discharge in nerve 2. Bending a group of hairs, which gives complex activity in nerve 2, produces activity in a single fibre in the commissure; this fires first at the moment of maximum deflexion, and then at irregular, progressively increasing intervals of about o-2 sec. while the displacement is maintained. No regular activity could be correlated with replacement, though this also gives a burst in nerve 2 (Text-fig. 7D).
A puff of air to spine or coxa, which produced a complex burst and a long after-discharge in nerve 2, gave a burst of a few large spikes in the commissure. The burst did not usually last longer than the stimulus, and never longer than twice the length of stimulus, although activity in nerve 2 continues for a much longer time (Text-fig.7E).
When one of the second nerves was cut, and recording made from both the ipsi- and contra-commissures after a puff of air to the spine or coxa, closely similar records were obtained. This shows that in addition to receiving information from sensilla on both sides of the animal, each nerve connects with both ipsi- and contra-commissures.
If the cervical membrane is lightly stretched forward, as in life, and then slightly displaced by a relay-operated needle, a response can regularly be recorded in the anterior commissure of fourth and fifth instar nymphs, and occasionally from adults. It is probable that this displacement is registered and transmitted by thé chordotonal organs in the manner suggested above. Activity is confined to a single commissural fibre, and commences before displacement has reached its maximum (Text-fig. 7F); it appears to be very sensitive to small movements. The activity in nerve 2 during this stimulation was not successfully recorded, owing to the difficulty of dissecting far enough without damaging or stretching the chordotonal organs.
The results of these recordings are summarized in Table 1.
DISCUSSION
The hair sensilla of the sternum respond to displacement caused by movement of neighbouring parts of the body, and by air currents. The sensilla of the sternal spine do not appear to differ in any way from those of the sternum, or from some of those of the leg, and it seems that the structure serves mainly to increase the amount of proprioceptive information available, by increasing the number of sensilla in the most effective area. The hairs on the spine do transmit detailed information about the relative position of neighbouring parts, and loss of this information impairs, at least temporarily, the animal’s control over them.
It is not clear why the Catantopinae in particular should supplement their proprioceptive input in this way; most members of the family are active and manipulative, but not conspicuously more so than, say, the Oedipodinae, which produce convergent forms in most habitats. Certainly within the Catantopinae there seems to be a correlation between a sternal spine well supplied with sensilla and manipulative ability of the forelegs. Schistocerca is an active animal, travelling from one food plant to another, picking food up and inspecting it when it is found, conveying small pieces of leaf to the mouth, turning them edgewise to go between the mandibles; in captivity it will rake together with the forelegs pieces of bran from a rather watery gruel, and then lift them to the mouth. Ornithacris, a much larger, more sluggish animal, lives in reed-beds. The leaves it prefers are too large to be handled, and it often merely gnaws thick stems. The spine is well developed, but with comparatively few sensilla. Romalea, yet more sedentary, has hardly any sensilla in the sternal region.
The interest of the system described lies chiefly in the organization of the nervous system which handles the tactile information. Two levels of integration of the signal from the sensillum are seen—at the nerve fibre and at the ganglion. It is common in insects for several sensilla to be served by one sensory nerve fibre, and this is well seen in the present material. At all but the highest levels of stimulation, it seems that more than one sensillum must be excited before a propagated impulse occurs in the nerve. Only when a single sensillum is stressed to the maximum does it pass a signal to the ganglion (perhaps as a warning signal to avoid damage?). If the area of excitation includes sensilla connected to different nerve fibres, and the properties and/or connexions of the nerve fibres differ, then the system offers an economical method of distinguishing spatial patterns of stimulation. These properties are present here in the considerable variation in diameter of fibres all deriving from similar sensilla, and in the very different types of activity produced by stimulating different groups of hairs.
Of all the complex information which arrives at the ganglion through the second nerves, a strikingly small proportion, and that greatly simplified, is transmitted to the brain. Only when a number of sensilla are excited simultaneously, as by air currents, is activity produced in one or very few fibres of the commissure. Apparently the majority of decisions are made at a segmental level, and this is confirmed by the complexity of the segmental reflexes which stimulation produces. The performance of the reflex, however, results in a great deal of activity in the commissure ; it is to be expected that this information about the movement of effectors would be necessary to the highest co-ordinating systems, which are cerebral.
ACKNOWLEDGMENT
This work was undertaken during the tenure of a D.S.I.R. Research Fellowship.
REFERENCES
EXPLANATION OF PLATE
A and B. The sternal spine of Schistocerca, showing the profusion of sensilla, and the way that those located on different parts contact each other.
C. Transverse section of nerve 2, near ganglion. Picric-osmium/haematoxylin.
While the wax used for this control is warm, a small regular discharge occurs in the anterior commissure, decreasing and vanishing as the wax reaches room temperature. Presumably some sensilla are sensitive to temperature changes of this sort, though not to radiant heat (above).