A Locust Wind and Ocellar Brain Neurone

Although the structures of a few of the largest neurones in the brains of locusts have been described (Williams, 1975) the functions of most of them are unknown. Approaches which may be employed to establish a function for a particular brain neurone are to list the sensory stimuli which excite it ; to search for output connexions from it to other identified neurones such as motoneurones; and to observe how it responds when a locust is allowed to move in different behavioural situations. These approaches have been applied to two large descending brain neurones, the descending contralateral movement detector (DCMD) and the tritocerebral commisure giant (TCG). A DCMD is excited by an abrupt movement of a small object in front of one compound eye and by a loud noise (Rowell, 1971 ; O’Shea, 1975), and it makes excitatory connexions to some of the motoneurones which move the hind legs and the fore-and hind wings (Burrows & Rowell, 1973; Pearson & Goodman, 1979; Simmons, 1980). Burrows & Rowell (1973) have suggested that it plays a role in the initiation of a jump, and it may also help to extend the wings prior to a jump and to execute rapid changes in flight course to avoid collisions (Simmons, 1980). A TCG is excited by mechanical stimulation of some of the wind-sensitive hairs on the head or of an antenna and by alterations in light intensity (Bacon & Tyrer.içyS) and it makes excitatory connexions with some flight motoneurones (Bacon and Tyrer, 1979).

This paper deals with the stimuli which influence the frequency of spikes in another of the large descending brain interneurones in one side of the nervous system of a locust, and with the output connexions which it makes with some of the thoracic motoneurones. It is excited by a current of air flowing over the head and by a reduction in the intensity of light falling on the median ocellus. An increase in the intensity of light falling on the median ocellus causes a reduction in the frequency of wind-induced spikes.

The pathway that the interneurone provides between an ocellus and flight motoneurones is interesting because ocelli have been thought for some time to be associated with the control of flight (e.g. Demoll & Scheuring, 1912; Link, 1909; Kalmus, 1945), and M. Wilson (1978 a) has suggested that they help to stabilize flight by monitoring changes in the orientation of the visual horizon. He points out that a number of features make the large second-order neurones of ocelli suitable for detecting changes in the orientation of the horizon. These features are: (1) they are very sensitive to ultra violet light, which makes them well able to detect a change in the position of the border between the sky and the land ; (2) their large axons ensure the rapid transmission with little attenuation of signals from the receptors of an ocellus into the brain; (3) they have very large spatial sensitivity; and (4) the three ocelli of a locust are disposed in a way that allows them to detect the horizon right around the animal. Changes in the intensity of light reaching the median ocellus of a flying locust would signal changes in orientation in the pitch plane. The connexions that the interneurone described here makes with flight motoneurones are suitable for altering the angle of pitch of a locust’s flight course, and I shall suggest that one of its functions is to ensure the stability of flight in this plane.

Experiments were performed on 65 adult Schistocerca americana gregaria (Dirsh), of both sexes, taken from cultures. An incision was made along the middle of the dorsal surface of a locust from the second abdominal segment to the head. The wings were gently pulled sideways and secured with plasticene to spread the thorax laterally. The gut, fat, some small muscles and parts of the endoskeleton which lie over the thoracic central nervous system were removed. A stainless steel, plastic-coated platform was placed beneath the meso-and metathoracic ganglia by manipulating it through the abdominal connectives. In some experiments the prothoracic ganglion was stabilized in a similar manner. Hook electrodes placed beneath each pro-meso-thoracic connective recorded extracellular spikes from some interneurones.

Microelectrodes for recording from central neurones were filled with 2 M potassium acetate and had d.c. resistances of 50–80 MΩ. In most experiments the perineurium of ganglia and connectives was softened by applying a 1 % solution of protease in saline for 2 min. No differences were found between animals which were treated in this way and those which were not. The identity of a flight motoneurone penetrated by a microelectrode was established by correlating spikes in it, induced by applying a depolarizing current, with twitches seen in a muscle. By cutting through the elevator muscles halfway along their lengths, parts of all the flight muscles were made visible. This method proved more reliable than employing extracellular leads to record from and stimulate motoneurone terminals in flight muscles. Most intracellular recordings from motoneurones were made in the regions of nerve roots, where the most likely neuronal process for an electrode to penetrate is the axon of a motoneurone or sensory neurone. The muscles are named and numbered from Snodgrass (1929).

To stain interneurones electrodes were filled with 1 M cobalt nitrate, and 500 ms pulses of positive current not exceeding 15 nA were applied for up to 2 h at 1 Hz. After treatment in 2 % ammonium sulphide in saline and fixation for 30 min in buffered 10% formalin ganglia were intensified as whole mounts (Bacon & Altman, 1977).

The experiments to establish the connexions that an interneurone makes with motoneurones were performed in Cambridge. Here the ocelli were illuminated by light from 2-2 V lens-end torch bulbs powered by 3 V batteries. Switching of lights was achieved manually, by relays operated from a pulse generator, or by a rotor spun by an electric motor. A light-sensitive transistor was used to monitor the switching. Light was led to each ocellus by a plastic light guide (of ‘Croton’, du Pont). Each light guide had a diameter the same as the lens of an ocellus, and was positioned with its end less than 1 mm from its ocellus lens. This method of illumination is inexpensive, but delivers no ultra violet light. Puffs of air were delivered to the locust’s head from a glass tube with diameter 3 mm, using either a compressed air supply or a rubber pipette bulb.

The experiments which established the stimuli which excite the interneurone and how they interact were performed in Berlin. Here the light source was a 900 W Xenon Arc Lamp, and the light path contained quartz optical elements. Interference and neutral density filters were used to vary the colour and intensity of the light beam. It was restricted to the median ocellus by a hypodermic tube which just fitted over its lens. Air was delivered to the whole head by a tube of internal bore slightly greater than it and placed 5 cm from it. By employing two solenoid-operated valves to switch between the stimulating jet and an exhaust jet, an almost constant velocity of air was achieved during a stimulus.

The interneurone described in this paper has an axon which is one of the largest in the nerve cord and occupies a dorso-lateral position within it. Spikes are recorded from it both in response to currents of air directed over the head and to a reduction in the intensity of light falling on the median ocellus. During this study, no other neurone with its axon in the dorso-lateral part of the nerve cord was found to spike in response to both of these stimuli. From its anatomy in the brain, it is most likely that the interneurone is the neurone termed ‘03’ by Williams (1975-see also C. S. Goodman, 1976), and it will be referred to as 03 throughout this paper.

An 03 neurone has its cell body on the posterior face of the protocerebrum, adjacent to that of a DCMD (O’Shea, Rowell & Williams, 1974), near to where one of the major tracheal trunks enters the brain (Fig. 1). From the cell body, a neurite runs to meet the thickest process of 03 in the brain, which runs for 200 μm, parallel to the midline of the brain and 150 μm away from it. From this process, most branches of 03 within the brain originate. None of them cross the midline. Some extend laterally, towards the optic lobe. One very extensive branch extends anteriorly, almost reaching the antennal lobe, and ventrally as far as the tritocerebrum. Ramifications from it intermingle with those from some branches which originate in the tritocerebrum. The thickest process of 03 runs close to some large, second order neurones of the median and ipsilateral ocellus, and primary wind-sensitive neurones have branches near those of 03 in the tritocerebrum (Bacon & Tyrer, 1978).

The axon of an 03 passes down the connective ipsilateral to its cell body and has processes in the ipsilateral sides of the suboesophageal and all three thoracic ganglia (Fig. 1). It extends as far as the abdomen, but its projections here have not been traced. In each connective it is one of the largest axons and occupies a dorso-lateral position. It runs in the latero-dorsal tract in each ganglion (Gregory, 1974).

In the suboesophageal ganglion there are five or six major branches from each 03 axon, all originating posterior to nerve 2. In the pro-and mesothoracic ganglia, the structure of an 03 is similar. There are two major branches, one originating level with the root of nerve 3 and passing postero-laterally, and the other originating just posterior to the root of nerve 5 and passing antero-laterally. Other branches extend laterally and medially from the axon. There are also two major branches within the metathoracic ganglion, both anterior to the root of nerve 3.

In a well-lit laboratory, in still air, the axon of an 03 is normally silent. It responds, usually with a single spike, to a reduction in the intensity of light falling on the median ocellus (Fig. 2 a). A spike from an 03 induced by this, or any other stimulus cannot be recognized in extracellular recordings from a whole connective unless accompanied by an intracellularly-recorded spike, as in Fig. 2(b). Most recordings from 03s in different locusts were made just anterior to the mesothoracic ganglion, and their spikes reach this position 25–30 ms after a reduction in the intensity of light to the median ocellus.

In response to an air current flowing over the whole of the front of a locust’s head, an 03 spikes continually (Figs. 2c, d). These spikes continue for as long as an air current is applied. The optimal direction for an air current to elicit spikes is from directly in front of the head. When an air current is directed at a small area of the head from a jet of internal diameter 1 mm and positioned 5 mm from it an 03 usually responds only with a brief burst of spikes, implying that it is excited from windsensitive hairs situated on widely separated locations over the head.

Increasing the intensity of light falling on the median ocellus causes a decrease in the frequency of spikes induced by an air current (Fig. 2d). When the light applied to the median ocellus is switched off there is usually a transient increase in the frequency of these spikes.

Occasionally a spike is recorded from an 03 axon when a bright light applied to the ocellus ipsilateral to it is switched off. The intensity change required is several log. units greater than that required for light applied to the median ocellus. The possibility that light applied to the ipsilateral ocellus stimulates receptors of the median ocellus due to scattering within the head capsule cannot be ruled out, but stimulation of the ocellus contralateral to an 03 axon has never been found to elicit spikes in it. No spikes have been recorded from an 03 axon in response to stimuli applied to the compound eyes, antennae, ears, wings or legs.

An 03 conducted spikes with a velocity of 2·3 m/s between two electrodes placed in its axon 1·54 mm apart and either side of the mesothoracic ganglion. Measurement of the conduction velocity between hook electrodes placed beneath a connective and a microelectrode placed in various locations along the axon did not show any difference in the velocity of spikes travelling along a connective and spikes travelling through a ganglion. For comparison, Burrows & Rowell (1973) report that a DCMD axon has a conduction velocity of 3·1 m/s. The slower conduction velocity of an 03 axon is consistent with its smaller diameter.

A spike in an 03 axon mediates a unitary excitatory postsynaptic potential (EPSP) in the motoneurones of the ipsilateral mesothoracic subalar muscle (no. 99), and probably also in those of the ipsilateral metathoracic subalar muscle (no. 129) (Fig. 3). The following observations establish the existence of the connexion to a right mesothoracic subalar motoneurone. First, following the extinction of a light applied to the median ocellus there is a one-to-one correspondence between a spike in the axon of the right 03 axon and an EPSP in the motoneurone (Fig. 3,a), and when the interneurone spike fails so does the EPSP in the motoneurone (Fig. 3b). Second, when a pulse of current applied to the axon of an 03 elicits a burst of spikes in it, a series of summating EPSPs is recorded from the interneurone (Fig. 3 c). A stream of air applied to the locust’s head elicits spikes in the 03 and EPSPs in the motoneurone (Fig. 3d). Some of these EPSPs are probably mediated by other wind-sensitive, descending brain interneurones.

In a metathoracic subalar motoneurone, an EPSP follows extinction of a light applied to the median ocellus by 30 ms, about the time required for a spike in an 03 to reach this ganglion (Figs. 3 e-g). In a recording from the axon of the motoneurone at the edge of the neuropile this EPSP does not always stand out clearly from other EPSPs, some of which may be mediated by interneurones other than 03 which are excited from the median ocellus.

Each subalar muscle is innervated by two motoneurones, which can be distinguished because one of them innervates the anterior part and the other the posterior part. In one experiment I made recordings successively from both motoneurones of a right mesothoracic subalar muscle, and found that they both received an EPSP 30 ms after extinction of a light applied to the median ocellus.

Each spike in a right 03 mediates an EPSP in all of the motoneurones of the right dorsal longitudinal muscles of a locust (Fig. 4). This EPSP is clearly seen in recordings from cell bodies of dorsal longitudinal motoneurones (Figs. 4 d-f) as well as from the neuropile (Figs. 4,a-c ). Its latency, following arrival of an 03 spike at the anterior of a ganglion, is 1 ms (Fig. 1 c). This is consistent with the existence of a chemical synapse between the interneurone and motoneurones. Sometimes a single spike in an 03 can elicit a spike in a dorsal longitudinal motoneurone (Fig. 4J). When two spikes occur in the interneurone in response to stimulation of the medial ocellus, two EPSPs occur in a dorsal longitudinal motoneurone (Fig. 4b), and they may summate to produce a motoneurone spike (Fig. 4e).

The five motoneurones which innervate each dorsal longitudinal muscle can be distinguished because they each innervate a distinct block of fibres within it. Four of the motoneurones innervating the right metathoracic dorsal longitudinal muscle have their cell bodies in the posterior part of the right side of the mesothoracic ganglion, and the remaining motoneurone has its cell body on the left side of the metathoracic ganglion (Fig. 4). The innervation of the mesothoracic dorsal longitudinal muscle is similar. An EPSP occurs in all the motoneurones of a dorsal longitudinal muscle when there is a spike in the 03 axon ipsilateral to the muscle. When a recording is made from two motoneurones that innervate the same dorsal longitudinal muscle, an EPSP is seen in both 30 ms after light to the median ocellus is switched off (Fig. 4f).

In the single motoneurone in the mesothoracic ganglion which innervates the right mesothoracic dorsal longitudinal muscle, interneurone s in both left and right connectives mediate EPSPs after light applied to the median ocellus is reduced (Fig. 5). When both connectives are intact, a 3 mV EPSP is recorded from the cell body of this motoneurone and from that of its contralateral partner following this stimulus (Fig. 5 a). When the right connective is cut, the EPSP in the motoneurone to the muscle on the right is reduced to about 0·5 mV and the EPSP in the motoneurone to the muscle on the left is only slightly reduced in amplitude (Fig. 5b). When this experiment was repeated for the two dorsal longitudinal motoneurones which have their cell bodies in the metathoracic ganglion, the same result was obtained. The most obvious explanation for these results is that both left and right 03s connect with these four motoneurones. Further recordings are required to show whether this is so. These motoneurones have extensive branches on both sides of their ganglia, whereas the branches from other flight motoneurones are mostly restricted to one side only (Tyrer & Altman, 1974), making it most unlikely that an 03 would connect directly with other flight motoneurones contralateral to its axon.

In order to discuss the function of an 03 in the control of a locust’s behaviour it is necessary to consider how it might be excited in natural circumstances and what effects its known outputs to motoneurones have on the locust’s movement. Because an 03 is excited by a current of air flowing over the head and because it connects with flight motoneurones, it is probably involved in the control of flight. The way in which the frequency of spiking in it is affected by the intensity of light falling on the median ocellus and the way in which the muscles whose motoneurones receive EPSPs from it move the wings both suggest that an 03 plays a role in regulating the angle of pitch of a flying locust. This is illustrated in Fig. 6.

In a locust flying along a horizontal course, each 03 would probably spike continually in response to the stimulation of the wind-sensitive hairs on the head. If the locust pitched upwards, the intensity of light falling on the median ocellus would increase and the frequency of the spikes in the 03s would decrease. Conversely, if the locust pitched downwards, the intensity of light falling on the median ocellus would decrease, and the frequency of spiking in the 03s would increase.

The subalar and dorsal longitudinal muscles, whose motoneurones receive EPSPs from the 03s, are capable of regulating the lift produced by the wings and the angle of pitch of a locust. (Fig. 6 a shows the arrangement of the flight muscles on the right side of a locust, and the subalar and dorsal longitudinal muscles are stippled.) The. lift generated by a locust wing depends upon the power applied to it and upon its angle of attack and velocity relative to the air stream moving past the body. The subalar muscles are important in regulating the twisting motions of the wings around their bases, and their angles relative to the long axis of the body (D. M. Wilson and Weis-Fogh, 1962). The dorsal longitudinal muscles are powerful depressors of the wings, and alteration of the force generated by them will alter the power delivered to the wings and the lift they generate. When a locust is flying along a horizontal flight path, the angle of attack of the wings is slightly positive (Fig. 6 b). If the locust pitches upwards, the angle of attack of the wings (relative to the air stream) will initially increase and the locust may be in danger of stalling (Fig. 6c). However, the excitation of the subalar and dorsal longitudinal motoneurones from the 03s will decrease. The resulting decrease in force generated by the subalar muscles will allow the wings to twist around their hinges towards their former angles of attack. The twisting of the wings is indicated by the curved arrows in Fig. 6c. Reduction in the force generated by the dorsal longitudinal muscles will help reduce the danger of stalling. If a locust pitches downwards, the angles of attack of the wings will decrease, which could precipitate a nose-dive (Fig. 6d). The increase in frequency of spiking of the 03s will cause an increase in force generated by the subalar muscles, which will twist the wings towards their former angles of attack (Fig. 6d). Increased force from the dorsal longitudinal muscles will assist the locust to reattain its former flight path.

Because it is continually excited by air currents in a flying locust, an 03 can register both increases and decreases in the pitch of flight. It provides a rapid pathway for transmission of information from the median ocellus to flight motoneurones, enabling the locust to adjust its flight quickly to overcome disturbances. Other interneurone s connected with the lateral ocelli may function in a similar way to stabilize flight in the roll plane. No doubt several different sensory pathways register changes in flight directions, but the ocelli are equipped to register changes in the orientation of the locust relative to the horizon very rapidly, as summarized in the introduction. Experiments on the flight of tethered locusts could provide useful information about their function in stabilizing flight. Experiments are also required to demonstrate the nature of air flow around a freely flying locust’s head, so that we can predict more accurately than is presently possible the type of activity that occurs in wind-sensitive neurones when a locust flies and turns.

Study of the mechanisms of excitation of an 03 in the brain would be worthwhile. Some of its processes overlap processes from large second-order ocellar neurones (C. S. Goodman, 1974, 1976; L. J. Goodman, Patterson & Mobbs, 1975; Guy, Goodman & Mobbs 1977), and the second order neurones may synapse directly with it. It should be relatively easy to make intracellular recordings from the two types of neurone since they are among the largest in the brain of a locust. Such recordings would provide useful information about the processing of information by brain neurones, and about the physiology of synapses, since the large second-order ocellar neurones exhibit both graded potentials and spikes (M. Wilson, 1978 b).

I am grateful for the help given me in performing this work from Malcolm Burrows, Jochen Erber, Melody Siegler, Frank Gwilliam, Randolph Menzel, Jürgen Milde and Claire Rind. My visit to Berlin was supported from a grant from the Deutsche Forschungs-gemeinschaft, no. Me 365/6 to Dr J. Erber and Professor R. Menzel of the Free University.