The physiology of sound reception by tympanal organs, long hair sensillae on the anal cerci, and hair sensillae on the thorax and abdomen has been studied by electrophysiological methods in four species of Acrididae, namely Stenobothrus lineatus, Omocestus viridulis, Chorthippus parallelus and C. brunneus.
The discharge in the tympanal nerve of all four species on stimulation by pure tones was asynchronous and showed no equilibration or fatigue. Comparative data are given for threshold sensitivity to pure tones for the four species over the range 0-20 kc./s., and shows the sensitivity to be very similar throughout the group.
The tympanal organs of all four species responded to pulsed sound with synchronous volleys of spikes up to pulse repetition frequencies between 90 and 100 pulses per second ; at higher rates synchronism broke down.
Long-hair sensillae on the anal cerci responded to gross air movements with bursts of spikes and to pure tones up to 300 c./s. with a synchronous discharge. The cereal nerves synapse with fibres in the ventral nerve cord, and the postganglionic discharges on continuous maximal stimulation of the cereal sensillae show accommodation. It is concluded that these cereal sensillae are similar in response and innervation to those on the cerci of the cricket, cockroach and migratory locust, but that owing to their small number and restricted frequency range they are probably not concerned with reception of natural stridulation.
Discharges were evoked in segmental nerves of the third thoracic ganglion in the four species in response to vibration of the substratum and stimulation by airborne sounds. The sensillae mediating the vibration response are hair sensillae situated on the abdominal stenites.
The receptors mediating response to air-borne stimuli may be hair sensillae on the abdominal pleura and stenites, or segmental chordotonal sensillae. The present work indicates that the former are probably the receptors concerned, but further work to elucidate the roles played by the two types of organ is required. Data are given on the response and threshold of the hair organs.
The responses of all the sound receptors investigated were independent of age and sex and also, in females, of the state of the ovaries.
The basic physiology of sound reception in orthopteroid insects, as mediated by tympanal organs and various types of hair receptor, was described by Pumphrey and Rawdon-Smith in a series of papers between 1936 and 1940. In 1940 Autrum published identical conclusions in Germany. Since then, no further physiological data on these organs have been published. Thus at the Paris conference on the acoustics of Orthoptera in 1954, one of the chief drawbacks to discussion was the fact that the physiology of reception was only partially known in half a dozen species, and had never been clearly related to acoustic behaviour in these insects, since neither Pumphrey and Rawdon-Smith nor Autrum had investigated the nervous discharge consequent on stimulation of the sound receptors of their experimental insects by stridulation of the orthopteroid type.
The present papers describe work on the physiology of reception carried out during an investigation of stridulation and associated behaviour in certain Orthoptera.
MATERIALS AND METHODS
A group of four closely related Orthoptera was examined in order to obtain comparative data on sound reception; the insects were adult grasshoppers of the group Acrididae, namely Stenobothrus lineatus (Panzer), Omocestus viridulus (Linnaeus), Chorthippus parallelus (Zetterstedt), and C. brunneus (Thunberg) (= C. bicolor (Charp)). These insects occur in intermingled ‘colonies’ on grassland at Imperial College Field Station, Silwood Park, Sunninghill, Berkshire, and their occurrence, general biology and population dynamics have been worked out by Richards & Waloff (1954).
The organs investigated were the tympanal organs, long hair sensillae on the anal cerci and hair sensillae on the thorax and abdomen; the method used was an electro-physiological one similar to that of Pumphrey & Rawdon-Smith (1936 A). A double-walled copper box, lagged to prevent echo, was used to house the nerve preparation; a short copper tube 5 in. in diameter directed sound stimuli at the preparation through an opening in the side of the box.
The nerve preparation used for each organ is described under the appropriate heading. The recording system was the conventional a.c.-coupled amplifier and double-beam oscilloscope with photographic recording.
Pure tone stimuli were derived either from a B.S.R. oscillator type LO 800A or a specially built beat frequency oscillator (Haskell & Haskell, 1955), the sound being directed at the preparation by loudspeaker as mentioned above. No electrical interference was experienced from the stimulus; the waveform and frequency of the stimulus was monitored by a crystal microphone inside the preparation box feeding an audio-amplifier, whose output could be displayed on an oscilloscope. The intensity of the stimulus was measured by a Dawe sound level meter type 1400 C, the microphone of which was mounted inside the preparation box; this enabled the sound pressure at the preparation to be determined in decibels relative to a reference level where o db. was equivalent to 0-0002 dyne/cm.2. All measurements of intensity given in this paper are relative to this reference level. The meter gave readings for sounds in the frequency range 50 C./S.-13 kc./s., down to a level 24 db. above the reference level. Sound levels at frequencies between 13 and 20 kc./s. (the limit of the oscillators) were derived by calculation. Although not so accurate as figures from the level meter, these calculated figures agreed well with the curve of frequency/sensitivity as given by the data derived from the meter and so have been allowed to stand.
Pulse stimuli, whose intensity, duration, and rate of repetition could be varied, were obtained by passing the output of a square wave generator through a transformer coupled to the stimulus loudspeaker.
(a) Tympanal organs
The nerve preparation for tympanal organs described by Pumphrey & RawdonSmith (1936b) for Locusta was first tried but proved to be unsatisfactory owing to the small size of the present insects, and the following method was used. The insect was secured in a block of plasticine ventral side down and the tegmina and wings were cut off. A small sagittal dorsal cut was made in the integument backwards from the pronotum and held open by small wire hooks attached to pins. The section of gut underlying this cut was removed and fat and connective tissue were cleared away to uncover the metathoracic ganglion; the tympanic nerve was then picked up on the electrodes and the preparation and apparatus were checked by giving a short burst of stimulus. The nerve was then severed from the ganglion and the screening box closed. Such preparations maintained their initial sensitivity for long periods, often up to 5 hr. and longer if occasionally moistened with a suitable saline (Pringle, 1938). Decapitation of the insect was not carried out; it caused no change in the nervous discharge and generally led to more rapid desiccation with consequent loss of sensitivity. Care was taken not to puncture or disarrange the air sacs surrounding the tympanal organs more than necessary; but in fact even complete removal had little effect on sensitivity and none on the pattern of nervous discharge.
Determination of the threshold sensitivity at various frequencies was carried out as follows: beginning with low frequencies, the intensity of the stimulus was adjusted until a response was just obtained from the preparation, when the sound level was read on the meter. The frequency of the stimulus was then increased and the process repeated, until the frequency limit of the stimulating oscillators, 20 kc./s., was reached. It is virtually certain, in view of the work of Auger & Fessard (1928), that the tympanal organs in Acrididae respond to sound over a much greater frequency range than 0-20 kc./s. ; but this range covers all the frequencies found in the natural stridulation of the species, which was all that was necessary for the present study. When one run over the whole frequency range had been completed, the external noise level was measured to ensure that this was too low to interfere with the response of the preparation. This procedure was carried out three times for each individual insect, and the levels for each frequency were then averaged; the readings at all frequencies were found to agree within 5 db. in every case.
Fig. 1 shows graphs relating the threshold intensity in decibels to frequency of a pure tone stimulus for the four species ; each graph is the average for three specimens.
The nervous response to a pure tone stimulus between o and 20 kc./s. was asynchronous in all cases. Fatigue, equilibration or adaptation were not observed in any case of prolonged stimulation. Under conditions of zero stimulation, a resting discharge was present.
The nervous response to sounds consisting of trains of pulses was also studied.
Fig. 2 shows a series of oscillograms of action potentials in the tympanic nerve of a male Stenobothrus lineatus due to stimulation with pulsed sound from the square wave generator ; at pulse repetition frequencies between 1 and 90 per second the nerve discharge is synchronous with the stimulus, although falling in amplitude as the top of the range is approached. Above 90 pulses per second synchronism breaks down, and in all four species this “was found to be the case at rates between 90-100 pulses per second.
By measuring the interval between pulses at repetition rates at which asynchronous responses begin to appear, figures for what might be called the ‘refractory period for the system’ were obtained. These were of the order of 20-30 msec, for all four species, and represent the least interval that must occur between successive sound stimuli to allow them to produce a synchronous discharge in the tympanic nerve.
Although Pumphrey (1940) reports that in Locusta the synchronous discharge of volleys at the modulation frequency of the stimulating sound showed the phenomena of ‘alternation’ and ‘equilibration’ these were not noted in the present experiments with the frequencies and stimulation times used. Finally it may be stated that continuous stimulation at subliminal level does not produce discharges by temporal summation.
These experiments were repeated with males and females of all four species, results being identical in all cases. Females were also tested on maturation of the ovaries and just prior to and after oviposition, but the nervous responses at these times were the same as previously described. This is of importance since observations on behaviour associated with stridulation in these species (Haskell, in preparation) shows that at these times important changes in acoustic responses occur.
(b) Long hair sensillae on anal cerci
The anal cerci of the present species are small, generally of the order of 1 mm. in length, and the number of long hair sensillae on them is also small. Although different in size, the receptors appear identical in external morphology with the long hair sensillae on the anal cerci of the domestic cricket which have been described by Sihler (1924). No histological work on these organs was carried out, but in view of the fact that they exhibit responses similar to those described by Pumphrey & Rawdon-Smith (1936a, b) for the sensillae of the cricket it is probable they have a similar structure.
The nerve preparation was made by sinking the insect in a block of plasticine, ventral side down, cutting off the tegmina and wings and making a dorsal sagittal cut at the rear end of the abdomen; gut and connective tissue were then removed to reveal the last abdominal ganglion. The nerve to one of the cerci was picked up on the electrodes, the preparation tested, and the nerve cut at the ganglion. In the present species the nerve is very fine and the preparation rapidly loses sensitivity (presumably due to desiccation, although this could not be arrested by the application of a suitable saline).
Fig. 3 A shows the response of these hair sensillae in a male Omocestus viridulus to a puff of air directed at them from a small pipette one foot away ; this response was common to all four species. The receptors in all species also showed a synchronous discharge to stimulation with pure tones up to 300 c./s. Both these types of response were abolished when the hairs on the cerci were entangled with one another by smearing them with vaseline. No clear evidence of equilibration or frequency halving or doubling was obtained with these preparations.
That the cereal fibres synapse with fibres in the ventral nerve cord was shown by recording from the latter between abdominal ganglia 7 and 8, when spikes of the type associated with giant fibres were produced by stimulation of the sensillae on the cerci. Continuous maximal stimulation of the sensillae produced accommodation in the post-synaptic fibres, and Fig. 3 B shows this effect in a male O. viridulus.
On comparing these results with the findings of Pumphrey & Rawdon-Smith as regards the acoustic functions of cereal sensillae in the cricket (1936b) and the transmission of impulses from the cereal nerve to the ventral nerve cord in the cockroach (1937), it is considered that the evidence is sufficient to justify the assumption that the long hair sensillae on the anal cerci of the present insects respond to sounds in the same manner as the sensillae on the anal cerci of crickets and cockroaches and that a similar arrangement of synapses with giant fibres in the ventral cord exists in all these insects. This assumption is supported by the work of Cook (1951) which confirms the existence of a similar system in Locusta migratoria, another acridid.
(c) Hair sensillae on the thorax and abdomen
Pumphrey & Rawdon-Smith (1936a) showed that some unidentified end organs situated in the abdomen of Locusta migratoria, with afferent fibres in abdominal segmental nerves, were sensitive to air-borne sounds. In the present species preparations of the second and third tergal nerves of the metathoracic ganglion (aTg Nv, 3Tg Nv, Albrecht’s (1953) notation), when severed from the ganglion, gave responses when the insects were exposed to stimulation by air-borne sounds.
Fig. 4 shows electrical responses from the third tergal nerve, left side, of a female Chorthippus parallelus, to the recorded song of the male of the same species ; trace A shows responses in the normal insect, which were greatly reduced (trace B) when the thorax and second and third abdominal segements were smeared with vaseline. The nature and meaning of the residual activity in the nerves after the sensillae were vaselined will be referred to in the discussion.
It proved impossible to obtain a preparation with one fibre only firing, and thus threshold measurements refer to the sensitivity of a group of sensillae. Because of this, only a few measurements were made at frequencies between 2 and 4 kc./s. when it was found that over this range, a minimum pressure of 7 dynes/cm2 at the sensillae was necessary to produce a response. The discharge due to the hair receptors was asynchronous with regard to the stimulating sound, but increases in the number of active fibres and in the frequency of discharge followed increases of stimulus intensity. No fatigue was noted after prolonged stimulation.
The hair sensillae on the ventral side of the abdomen are in a position to mediate responses to vibration of the substratum, and discharges were elicited in the tergal nerves of all species in response to vibrations of the preparation platform produced by a gentle tap with a rubber-coated glass rod. The vibrations were picked up by a Rothermel V.P. 5 vibration pick-up and displayed on one trace of the oscilloscope. A crystal microphone was placed near the preparation during this experiment, feeding a loudspeaker through an audio-amplifier, to allow the observer to check that air-borne sounds were not being produced by the tapping of the substrate. The response of the hairs to vibration was completely abolished when the abdominal stenites were smeared with vaseline.
The results show that in the experimental insects at least three types of receptors exist which are sensitive to sound. Of the three, it seems dubious, however, if the long hair sensillae can be greatly involved in the reception of stridulation, in view of the poor development of the anal cerci, the few hair sensillae found thereon and the restricted frequency range of these receptors. It has been suggested above, that the arrangement of central representation of these sensillae is similar to that of analogous receptors in the cricket, cockroach and migratory locust. In the cockroach stimulation of the cereal sensillae by a puff of air produces the ‘evasion response’, a burst of rapid locomotor activity, but in the locust this response is poor or absent (Cook, 1951). The anal cerci of Locusta are larger and the number of sensillae greater than in any of the species here studied, which suggests that their effect, if any, in the latter will be small, and they will therefore not be considered further as organs concerned with the reception of stridulation. It is possible that the true function of these organs in seen in the act of mating, during which they are mechanically stimulated by copulatory movements, when they may influence the state of central excitation. Huber (1952) has demonstrated that the cereal sensillae in Gryllus campestris must be stimulated for full mating behaviour to be realized in this species.
The remaining receptors sensitive to sound are the tympanal organs and the thoracic and abdominal hair sensillae. The basic physiology of the response of acridid tympanal organs to sound has long been established (Pumphrey & RawdonSmith, 1936a, b, 1939), and in all cases the responses of these organs in the present species conformed to their findings. The sensitivity of these organs is high for such simple structures, and Pumphrey (1950) has pointed out that the resting discharge, present in most invertebrate hearing organs so far investigated, may indicate a regenerative system, whose sensitivity would be greater than any non-regenerative counterpart. This latter system has recently been found to exist in certain lepidopterous tympanal organs (Haskell & Belton, 1956), but no comparative data on sensitivity as between the two types are yet available. The threshold graphs for the four species show there to be little difference in sensitivity between them ; this is to be expected considering the similarity in size and structure of the tympanal organs throughout the group.
It must be pointed out that data for threshold sensitivity derived from the graphs of Fig. 1 must be treated with caution. One reason for this is that, if the tympanic system is regenerative, the threshold will change in an irregular manner and any measurement only represents the average sensitivity at the moment of recording. But the most important reason is the fact that of all physical quantities sound intensity is perhaps the most difficult to measure accurately. In experiments such as one described above, done without benefit of a soundproof room, allowances of ±10% at least must be made on all figures. If this figure seems large it must be remembered that the main desideratum in sound measurement is a completely homogeneous medium to avoid refraction, reflexion and absorption, and owing to convection and conduction of heat by air it is very difficult to obtain such conditions in a laboratory. However, the fact that three trials were made in these experiments, in which readings at particular frequencies all agreed within 5 db., gives some guarantee that conditions were reasonably constant over the experimental period.
In the light of the findings of Pumphrey & Rawdon-Smith (1939) on the response of the tympanal organs of Locusta migratoria to modulated sound it was to be expected that the nervous discharge consequent on stimulation of the tympanal organs in the present species with pulsed sound would be volleys of spikes synchronous with the pulse repetition frequency of the sound. Fig. 3 clearly shows this; as the limiting pulse frequency is reached the spike height of the resultant discharge is considerably reduced. In the species studied, synchronism broke down at pulse repetition frequencies between 90 and 100 per sec. ; this figure is far in excess of any pulse repetition rates found in the natural stridulation of the species.
Turning now to consider the receptors which mediate responses in the segmental nerves we find differing opinions as to the type of sense organ concerned. Pumphrey & Rawdon-Smith (1936a) at first considered the responses to originate in the hair sensillae distributed over the body of the locust. Pumphrey, however, in a later paper (1940) inclined to the view that the organs concerned were segmental chordotonal sensillae, the evidence for this modified view being that there only appeared to be two active fibres in the segmental nerves, and the receptors did not fatigue on constant stimulation as do most hair organs. In the present work the pattern of nervous discharge has indicated a variable number of active fibres in the various preparations, ranging from three to five. In some preparations the number of active fibres depended on the intensity of the stimulus, a higher intensity increasing the number of firing fibres as well as increasing the frequency of discharge, which supports the view that the receptors are hair organs. The nearly complete abolition of response in the segmental nerves when test insects were smeared with vaseline also points to hair organs as being the receptors concerned. However, some residual activity was recorded from the nerves after the application of vaseline, and furthermore the receptors did not fatigue with a constant stimulus, all of which evidence supports the view that the responses were due to chordotonal organs. Hughes (1952), recording from the segmental nerves of Locusta, found discharges due to acoustic stimuli and also bursts of impulses during respiratory movements of the insect ; he concludes that ‘some sense organs respond to both the respiratory movements and auditory stimulation, while others are excited by either one or the other’. It thus seems possible that residual activity after application of vaseline in the present insects might be due to the effects of respiratory movements on chordotonal organs, and that hair organs act as the main receptors of acoustic stimuli.
Examination of the abdominal chaetotaxy in the present species supports the view. The 2nd and 3rd tergal nerves of the metathoracic ganglion, from which recordings were made, supply the 2nd and 3rd abdominal segments respectively. On the sterna of these segments there is a fairly large number of hairs approximately 250µ in length; in the normal resting position of the insect these would press against the substratum, and while suited to receive vibrational stimuli would be unlikely to respond to air-borne sounds. The pleura of these segments have short hair sensillae, approximately 50µ in length, distributed over them; these hairs are fairly stiffly articulated. However, at the sternal borders of the pleura a very few of the longer (250 µ) hairs are found, with a considerably looser articulation, which move when a current of air is passed over them. These hairs are in a position to respond to air-borne acoustic stimuli, and their relative scarcity may explain the small number of active fibres in the preparations. Although this evidence is suggestive, more detailed work is required to place the matter beyond doubt.
If it be accepted that these hair organs do act as receptors for acoustic stimuli, then their use in nature for mediating responses to natural stridulation in the species under discussion must be greatly restricted. If their maximum sensitivity is of the order of 70-80 db., as suggested by the present findings, then the range over which they respond to natural stridulation in the field must be measured in centimetres, since data on the sound intensity of stridulation in the present species (Haskell, 1955) indicates that only very close to the emitting insect are such levels reached. It is possible that these organs may be called into play during the courtship behaviour of these acridids, when males and females manoeuvre within a centimetre or so of one another.
Whether or no these hair sensillae respond to air-borne stimuli, the hairs on the stenites certainly mediate the responses to vibration recorded from the segmental nerves, since this response is completely abolished on the application of vaseline to the appropriately situated sensillae.
I am grateful to Prof. J. W. Munro for encouraging this work at its inception and to Prof. O. W. Richards for his criticism and advice. I am indebted to the following for the loan of apparatus : Mr Toombs, Department of Electrical Engineering, Imperial College; Dawe Instruments, Ltd., Brentford; Mr Bayard, Director of the Acoustics Section, Admiralty Research Laboratory. I have also profited by advice from and discussion with Dr B. P. Uvarov, Director, Anti-Locust Research Centre, Dr N. Waloff and Mr J. W. Siddom, Department of Zoology and Applied Entomology, Imperial College of Science and Technology, and Prof. R. J. Pumphrey, F.R.S., Department of Zoology, Liverpool University, who kindly read and criticized the manuscript. Some of the apparatus used in this work was purchased with a grant from the Central Research Fund of the University of London, and the work comprises part of a thesis for the Ph.D. degree of that University.