Respiration in the Desert Locust: I. The Control of Ventilation

The abdomen of the desert locust, Schistocerca gregaria Forskal, makes regular and nearly continual pumping movements which ventilate the larger tracheal trunks. The increase in amplitude and frequency of these movements in response to carbon dioxide or oxygen lack is well known (Krogh, 1941), but the mechanisms by which the reaction is controlled have remained largely unexamined.

Matula (1911) believed that the head of Aeschna nymphs contained carbondioxide receptors which stimulated ventilation, but the idea of regulation via the head has been discarded by later authors. Fraenkel (1932) described sensitive secondary centres in the thoracic ganglia of Schistocerca. Roeder (1953) detected a rhythmical discharge from the isolated ganglia of Melanoplus in 5% carbon dioxide, and he suggested that the ganglia themselves might respond to carbon dioxide. Recent observations by Weis-Fogh (1960) on Schistocerca in flight have again suggested that control via the head may be important.

The nature of earlier experiments, in which one part of the central nervous system was removed and the remainder tested for carbon-dioxide sensitivity, made it desirable to carry out localized tests on the intact system as far as possible, and the present work was undertaken to re-examine the ventilatory centres and if possible to locate the carbon-dioxide receptors, with this in mind.

Schistocerca gregaria were supplied as week-old adults by the AntiLocust Research Centre. They were kept in cages as described by Hunter-Jones (1956), and fed with sprouting wheat shoots. Locusts at all stages of maturity and of both sexes were used in the following experiments.

Ventilation was recorded on a revolving smoked drum with a light lever attached by a thread to the third abdominal sternum. A second lever attached to the head recorded neck and prothoracic ventilation. This method allows accurate recordings of the frequency and an estimate of changes in amplitude to be made. A more elaborate technique for recording the actual volume of air pumped was considered unnecessary for the present work where only qualitative information was required. The preparation was placed in a Perspex gassing box (capacity 500 ml.) and the threads to the levers passed through holes in the lid. The box was perfused with mixtures of oxygen, nitrogen and carbon dioxide supplied from three calibrated flowmeters. Analyses of the gas were made periodically to check its composition (Scholander, 1951). Low carbon-dioxide tensions in air give rise to a much more immediate and vigorous ventilatory response than oxygen lack, so they were used as test stimuli in the experiments on the location of the ventilation regulators. Unless otherwise stated low concentrations of carbon dioxide are always in air.

Nerve impulses were recorded by lifting the nerve into air on two hooked platinum and 10% iridium electrodes, insulated down to the hooks with ‘Araldite’, and mounted on two Zeiss micromanipulators. The following Tektronix equipment was used to amplify and display the recordings: two low-level a.c. preamplifiers, type 122; a plug-in, dual trace, d.c. preamplifier, type 53/54C; a cathode ray oscilloscope, type 532. Photographic recordings were made with a Shackman camera, AC 2/25.

Respiratory movements were recorded on the oscilloscope by means of a manually operated tapping key which substituted small high-frequency pulses for the 50 cyc./sec. wave of the time marker (‘buzzer’).

Experiments were carried out at 18–20° C. : in a second series the temperature was raised to 28–30°C., but the results obtained were similar.

Muscles are numbered according to the scheme used by Snodgrass (1935). The spiracles are numbered 1–10 from the anterior, regardless of their position on the thorax or abdomen. The separate ganglia of the abdomen are numbered 1–5 ; it should be remembered that the metathoracic ganglion is fused with the first three abdominal ganglia.

Four types of rhythmical ventilation movement make their appearance in the non-flying locust:

1. Active raising and lowering of the abdominal sterna by the primary respiratory muscles (Snodgrass, 1935). This represents type 2 in the scheme of Plateau (1884).

2. Longitudinal telescoping movements of the abdominal segments by the secondary respiratory muscles. Both movements are again active.

3. Protraction and retraction of the head, which I have termed ‘neck ventilation’ (Fig. 1). Retraction, in phase with abdominal expiration, is brought about by the contraction of muscles 49, 54 and 57, while protraction (inspiration) is by muscles 50, 51, 52 and 53 which straighten the cervical sclerites.

   Fig. 1.

   

   View largeDownload slide

   A, lateral view of the muscles responsible for neck and prothoracic ventilation in the locust. Broken lines indicate positions at the end of expiration. B, the position of the cervical sclerites at the end of expiration. C, ventral view of the pro- and mesothorax showing the outward movements of the pronotal flanges (broken lines) during expiration.

   Fig. 1.

   

   View largeDownload slide

   A, lateral view of the muscles responsible for neck and prothoracic ventilation in the locust. Broken lines indicate positions at the end of expiration. B, the position of the cervical sclerites at the end of expiration. C, ventral view of the pro- and mesothorax showing the outward movements of the pronotal flanges (broken lines) during expiration.

4. Protraction and retraction of the prothorax, which is termed ‘prothoracic ventilation’. Retraction is brought about by the contraction of muscles 59 and 60, and protraction results from the elastic return of the ventral parts of the pronotal flanges, which during expiration ride over the anterior part of the mesothorax (Fig. 1).

Du Buisson (1924) noted neck ventilation in Stenobothrus and, prior to flight, in Melolontha. It has been observed during the present investigation in the following additional Acrididae: Locusta migratoria, Anacridium aegyptium, Nomadacris septemfasciata and Eyprepocnemis plorans.

The first type of ventilation is more or less continuons in the locust under normal conditions, although subject to great variation in amplitude and frequency. The other types are auxiliary ventilating mechanisms called upon for short periods after great activity: longitudinal telescoping by the abdomen first appears, then neck, and finally prothoracic ventilation.

In immature adults, or in locusts kept below 10° C., normal ventilation may be regularly interrupted by pauses when little or no movement occurs. Fig. 2 is a continuous record from a mature locust at 10°C., where approximately 2 min. bursts of ventilation are interrupted by 1 min. pauses.

Weis-Fogh (1960) has shown that vigorous abdominal pumping (dorso-ventral and longitudinal movements) can provide a maximum of 300 1. air/kg./hr. (167 mm³ per ventilatory stroke for a locust weighing 2 g.). The following measurements were undertaken to determine what extra volume could be pumped by the neck and prothorax.

Spiracles 1 were sealed and the anterior part of the pterothorax was waxed into one end of a tube (5·0 × 1·5 cm.) so that the head and prothorax could move freely inside it. The tube was filled with coloured water to which a very small amount of detergent had been added, and a narrow tube (internal diameter 2 mm.) was corked into the far end. The preparation, with both tubes almost horizontal, was perfused with carbon dioxide. The resulting neck and prothoracic ventilation moved the meniscus up and down in the narrow tube ; the distance it travelled was measured under a binocular microscope. Simultaneous recordings of the abdominal ventilation frequency were made on a kymograph. The volume of water moved in the narrow tube was taken to represent the additional volume of air pumped by the neck and prothorax. Further measurements were made after the removal of the abdomen, but the amplitude of neck and prothoracic ventilation did not alter appreciably. The results from a male and a female locust are shown in Table 1. Altogether twenty locusts have been tested and they have shown that neck and prothoracic ventilation can together contribute a maximum of about 14% of the total volume of air pumped by a non-flying locust, neck ventilation providing 11% and prothoracic ventilation 3%. The significance of these auxiliary forms may be greater than the figures suggest, since they ventilate primarily the head through spiracle 1.

It has been known for a long time that each abdominal ganglion can initiate independent ventilating movements in its own segment (Baudelot, 1864), and that one or more thoracic ganglia contain a pacemaker which controls the rhythm in the whole insect (Fraenkel, 1932).

In the following experiments the central nervous system was sectioned at particular sites under carbon-dioxide anaesthesia and the cuticular wounds were sealed with wax. The locusts were observed periodically for several days and in some cases for 3 or 4 weeks after the operation.

After section between the meta- and mesothoracic ganglia all sign of rhythmical movements disappears from the segments anterior to the operation ; the synchronized movements of spiracles 3–10 and abdominal ventilation continue with only a slight reduction in amplitude and frequency. If the nerve cord is then sectioned between the metathoracic and the first abdominal ganglia the rhythm persists unchanged in the first three abdominal segments and in spiracles 3–5 (supplied by nerves from the metathoracic ganglion), while it drops to a low frequency and amplitude in the remaining abdominal segments. Likewise, after section between the metathoracic and first abdominal ganglia of an intact locust, the rhythm anterior to the operation persists with little change while that posterior falls to a low level.

The results of sectioning between abdominal ganglia have confirmed that all are able to initiate at least weak pumping movements.

The rhythm initiated by the metathoracic ganglion is nearly always faster and more vigorous than that by the abdominal ganglia, and this ganglion probably contains a pacemaker which drives the slower rhythms of the abdominal ganglia and initiates synchronized movements in the more anterior segments.

Hoyle (1959) describes a rhythm in spiracle 2 of the locust which arises from the isolated mesothoracic ganglion: for reasons discussed elsewhere (Miller, 1960), this rhythm may have no relation to that occurring in the intact insect. There appears to be no rhythmical centre anterior to the metathorax, and after section between meso- and metathorax it is impossible to evoke neck or prothoracic ventilation. Hyperventilation by the abdomen then produces passive movements of the head. Moreover, the suboesophageal ganglion must be in communication with the metathoracic for these auxiliary forms to appear.

By nervous regulation is meant an overriding control from higher centres. It is apparent, for example, during handling, when extremely high frequency (180–220/min.) and shallow amplitude ventilation appears, which cannot be evoked by any combination of high carbon dioxide, low oxygen tension and increased temperature. Alternatively, it appears as a complete cessation of ventilatory movements, again during handling or at the start of flight. It has not been observed after decapitation.

A number of experiments was undertaken to locate the carbon-dioxide receptors. They provided the following additional observations:

1. 1% carbon dioxide in air provokes a considerable increase in ventilation frequency and amplitude, while over 10% causes a reduction in frequency and a great increase in amplitude. At lower concentrations much individual variation occurs as to whether hyperventilation is achieved more by increasing the frequency or the amplitude.

2. Oxygen lack is much less effective in producing hyperventilation than carbon dioxide, 10% oxygen in nitrogen producing approximately the same hyperventilation as 1–2% carbon dioxide.

3. In 1 % carbon dioxide neck and abdominal longitudinal ventilation are discernible, and in 2% prothoracic movements start.

4. After 1 hr. in 5 % carbon dioxide no type of ventilation shows any diminution as a result of fatigue or sensory adaptation.

To test the head for carbon dioxide sensitivity, a small hole was burnt through the cuticle into the air-sac of each mandible and cannulae were waxed in (Fig. 3 A). In some experiments all the tissues between the head and the prothorax were removed except for the nerve cord, and a Perspex shield was placed over the open end of the prothorax. Alternatively, the tissues were left intact, but the longitudinal ventral trunks and the tracheae from spiracle 1 to the head were ligatured with fine hair, 0·1–0·5 ml- doses of various gas mixtures were introduced from a syringe into one cannula and they escaped from the other. Modifications of the ventilatory rhythm were recorded on a kymograph. Since the normal air passages to the head were blocked, it was necessary to inject air into the mandibular air-sacs at frequent intervals.

1 % carbon dioxide gave rise to a detectable increase in ventilation and more than 4% produced almost immediate hyperventilation (Fig. 4). The injection of air produced either no effect or a decrease in the frequency. After squashing the nerve cords in the neck, similar injections had no effect.

Confirmatory experiments were carried out by cutting out part of the irons, opening the exposed air-sacs and then removing all the neck tissue except for the nerve cord. The head was placed inside a small gas box (capacity 1·5 ml.) and the nerve cord led out through a nick which was subsequently sealed with petroleum jelly (Fig. 3 B). Perfusion of the gas box gave similar results after a slight delay. Cauterization of the supra-oesophageal ganglion did not abolish the reaction, whereas after cauterization of the suboesophageal it disappeared.

The muscles for neck ventilation are innervated from the prothoracic ganglion, and those for prothoracic ventilation from the mesothoracic ganglion. In spite of this it is not possible to evoke either form of ventilation after decapitation. Apparently carbon-dioxide receptors in the head, perhaps the same as those which give rise to abdominal hyperventilation, must be stimulated: they relay to the metathoracic ganglion, which then brings about neck and prothoracic ventilation via the pro- and mesothoracic ganglia.

To conclude, the head contains carbon-dioxide receptors which modify abdominal ventilation. In addition the same or other receptors in the head must be stimulated for neck and prothoracic ventilation to appear. The drop in ventilation which follows decapitation is therefore attributable both to a reduction of the general level of excitation in the central nervous system and to the loss of head receptors.

Several hours after its complete denervation, except for the posterior connectives, the prothoracic ganglion, attached to part of the longitudinal ventral trachea, was lifted clear of surrounding tissue and placed on a pad of Ringer-moistened filter-paper in the small gas box. The connectives were led out through the nick, which was then sealed with petroleum jelly. The base and sides of the box were positioned with two Zeiss micromanipulators. As with the head, perfusion of the gas box with more than 2 % carbon dioxide gave rise to almost immediate hyperventilation which was recorded on a kymograph.

Similar results were obtained from testing the mesothoracic ganglion after removing the prothoracic, and from the metathoracic after removing the mesothoracic ganglion. Further tests, made on each ganglion without removing the more anterior ganglia, gave similar results. They show that each thoracic ganglion can stimulate ventilation when it alone is treated with carbon dioxide, and that this holds true when the nerve cord is intact or after the removal of the more anterior ganglia.

After cutting away the superficial tracheae and air-sacs from each ganglion the response was delayed but not reduced, so that possible receptors must fie on the surface or within the ganglion.

If a gentle stream of carbon dioxide is directed into the open end of the longitudinal ventral trunk attached to the ganglion, or into opened air-sacs on the ganglion surface, hyperventilation follows often in less than 0·5 sec. When the cut ends of the trunk are closed, however, and the air-sacs intact, hyperventilation does not follow in less than 5–10 sec. This suggests that the receptors lie within the ganglion.

At times the faster rhythm induced by carbon dioxide was seen to be superimposed on a slower rhythm (Fig. 4G, H), the latter being due perhaps to the abdominal ganglia alone.

Tests made on the abdominal ganglia show that in the presence of the thoracic they do not modify ventilation when they alone are treated with carbon dioxide. The tests were carried out by placing all the thoracic ganglia of a decapitated locust in the small gas box and perfusing it with air (Fig. 3C). The preparation, including the small box, was then placed in a larger gas box and perfused with carbon-dioxide mixtures. In this way the whole animal was treated with carbon dioxide, except for the thoracic ganglia. No hyperventilation was produced, and this showed in addition that no gas reached the thoracic ganglia. After removal of the thoracic ganglia, ventilation frequency and amplitude fall to a low level. Carbon dioxide does then cause a slight increase in frequency with occasional ventilations of large amplitude (coughs).

The foregoing tests show that the head and thoracic ganglia are each able to modify ventilation in response to carbon dioxide. (Tests were not made on denervated cephalic ganglia, but by analogy with the thorax it seems probable that the ganglia themselves mediate the reaction.) Stahn (1928) concluded that the receptors which modify ventilation in Dixippus are situated in the tracheae close to the spiracles. To test for the presence of additional receptors in Schistocerca, small amounts of 10% carbon dioxide were injected from a silicone-lined pipette into various parts of the thoracic tracheal system and the delay before the onset of hyperventilation was measured. Injections into the longitudinal ventral trunks, which supply the ganglia, produce almost immediate hyperventilation; injections into other tracheae produce a response after a longer delay or frequently no response at all. Intravital methylene-blue staining has failed to reveal nerves associated with the longitudinal trunks, and it is probable that all carbon-dioxide reception takes place within the ganglia.

Recordings from various parts of the isolated thoracic and abdominal nerve cord have provided further evidence for some of the foregoing conclusions.

Records from any part of the cord anterior to the metathoracic ganglion have failed to demonstrate a rhythmical discharge, although it seems clear that such must exist. Records from the posterior connectives of the metathoracic ganglion or from the lateral nerves to the first three abdominal segments, after they are squashed distally, usually show a rhythmical discharge : it can be seen to correspond to the ventilatory frequency if the lateral nerves are not squashed or if the more anterior ganglia are left intact (Fig. 5). The frequency of impulses and of the rhythm increases after treatment with 5 % carbon dioxide.

A rhythmical discharge has been detected in the lateral nerve stumps of each isolated abdominal ganglion. From 10 to 15 bursts occur per minute and the frequency of impulses, although not that of the rhythm, increases in 5 % carbon dioxide (Fig. 5). Recording simultaneously from the two lateral nerves of one ganglion shows a similar, but not identical, pattern of impulses.

If the bursts of impulses in fact comprise the motor excitation which normally brings about ventilation, then these recordings show that the bursts are initiated centrally and can be modified in response to carbon dioxide in the absence of any extra-ganglionic sensory mechanism.

The sites of the centres initiating and regulating the rhythm are summarized in Fig. 6.

The response to partial anoxia has not been investigated in detail, but the evidence suggests that it acts at the same sites as carbon dioxide and produces similar modifications of ventilation. It was pointed out that the effect of 10% oxygen in nitrogen is comparable to that of 1–2 % carbon dioxide. Analysis of the composition of the gas in the thoracic air-sacs during flight (Weis-Fogh, 1960) has shown that there is commonly about 5 % carbon dioxide and 15 % oxygen present. If these values can be taken as representative of the concentrations at the gassensitive sites within the ganglia, then carbon dioxide would appear to provide the more important ventilatory stimulus in the intact insect. Since carbon dioxide diffuses through animal tissues faster than oxygen (Krogh, 1919) such analyses may be misleading. However, assuming that the flight muscles are the prime source of carbon dioxide and users of oxygen during flight, then gas tensions will probably be little different in the ganglia and in the air-sacs.

The occurrence of pauses in the ventilatory rhythm of resting locusts and the occasional failure to detect a rhythmical discharge from the isolated metathoracic ganglion suggest that a low tension of carbon dioxide is necessary to initiate the rhythm. The pauses in ventilation are reminiscent of Cheyne-Stokes ventilation in man and may have a similar origin, namely, the periodic washing out from the blood of chemical stimuli necessary for ventilation.

The excitatory effect of carbon dioxide on some nerves and ganglia has been demonstrated by Boistel, Coraboeuf & Guérin (1957). This effect may be responsible for the ventilatory response, and possibly carbon dioxide acts directly at the synapses of the motor neurones which supply the ventilatory muscles. Hoyle (1960) has shown that carbon dioxide has a direct effect on the process of neuromuscular transmission in the closer muscle of spiracle 2, reducing the electrical responses and the tension developed. The effect of carbon dioxide on neuromuscular transmission is inhibitory, therefore, compared with the excitatory action on the ganglion. Both effects serve to increase ventilation.

The direct action of carbon dioxide at the spiracle muscle and its possible direct action on the motor neurone synapse both lead to an economy in sensory nerves and perhaps in connexions within the central nervous system. Whether this has any significance for the insect is unknown, but it is worthwhile recalling the small total number of neurones in the insect nervous system, which is available to perform the many observed behaviour patterns (Wiersma, 1952).

I would like to thank Prof. V. B. Wigglesworth under whose supervision this work was carried out. I am most grateful to Prof. T. Weis-Fogh for much encouragement and help, and for permission to quote his unpublished results. My thanks are due also to Mr F. Darwin and Mr J. S. Edwards for reading the manuscript. I am grateful to the Agricultural Research Council for financial support.