1. The impulses from the tympanic organ are transmitted at the prothoracic ganglion to a central neuron, the auditory T large fibre, which lies in the cord between the brain and the metathoracic ganglion. The impulses in the T large fibre are conducted rostrally and caudally with the same discharge pattern. Information is sent up to the brain, and down to the metathoracic ganglion, after a delay of about 12 msec.

  2. The impulses from the cereal hair sensilla are transmitted to two similar auditory C large fibres which lie in the cord between the metathoracic and last (6th) abdominal ganglia and are then sent up to the mesothoracic ganglia by other auditory large fibres.

  3. Central inhibitory interaction between the impulses from the tympanic nerves of the two sides are shown by a marked increase of impulses in the T large fibre following section of one of the tympanic nerves. No inhibitory interaction is found between the impulses from the two cereal nerves.

  4. The auditory T large fibre receives not only the excitatory effect from the ipsilateral tympanic nerve at the prothoracic ganglion, but also the inhibitory and weak excitatory effects from the contralateral one.

  5. The response range of the T large fibre is narrower than the threshold curve of the tympanic nerve and corresponds with one type of response range in the tympanic neurons. The response ranges of the C large fibres correspond closely with the threshold curve of the cereal nerve.

  6. A large difference in threshold between the two T large fibres is found in the response to sound incident from the side. The number of impulses in the T large fibre nearer to the sound source is greater than in that farther from the source.

  7. The difference in the number of impulses between the two T large fibres is most marked in the response to sound of the frequency which is dominant in stridulation. This difference is due to the mutual inhibitory interaction of neurons which modifies the number of impulses without changing the threshold of the tympanic large fibre.

  8. It is suggested that the central inhibitory interaction increases the information about a sound source and plays an important role in the mechanism of the directional sense.

  9. The stridulation of the group activates the tympanic nerve and evokes synchronized discharge in the T large fibre, but scarcely at all in the primary C large fibre. The tympanic organ and its neural network seem well adapted to reception of stridulation.

  10. It is concluded that though neither of the two sound receptive organs—the tympanic organ and the cereal hair sensilla—can perform frequency analysis, the insect may be able to do so by making use of both organs, since they have different frequency ranges and are served by different auditory large-fibre tracts.

The electrical responses to sound stimuli have already been recorded from the auditory nerve bundle in several kinds of insect, in Orthoptera by Pumphrey & Rawdon-Smith (1936) and Haskell (1956, 1957), in Lepidoptera by Haskell & Belton (1956) and Roeder & Treat (1957), and in Hemiptera by Pringle (1953). The central mechanism of hearing, however, has not so far been much explored. Quite recently the present authors (Katsuki & Suga, 1958, 1960) studied electrophysiologically the problems of directional sense and frequency analysis in the tympanic organ of an insect by recording activities of the peripheral auditory neurons. The central mechanism has been further studied, and this paper is concerned with the experimental results. Three problems have particularly been posed: frequency analysis of sound, directional sense and central inhibition.

The experiments were performed on Gampsocleis buergeri (Tettigoniidae), because of its large size and ready availability.

The insect was pinned on its back on a cork board and the ventral exoskeleton covering the nerve cord was removed. The tracheae distributing along the nerve cord were separated from the latter and the non-auditory inputs were also severed. The operated animal was placed about 50 cm from the loud-speakers and the sound was delivered from its left side in a sound-proofed room which was air-conditioned at about 26°C.

The impulses in response to sound stimuli were recorded from the connectives which were hooked up in the air with a 200 μ. silver wire electrode mounted on a micromanipulator. In order to trace the auditory tract in the nerve cord, two different parts of the cord were simultaneously hooked up with two recording silver wire electrodes and the electrical response of each part was led through an amplifier to two beams of an oscilloscope. The sound wave and the time signal were indicated on a third beam simultaneously. The indifferent electrode was a silver wire placed on wet cotton on the abdominal segments from which the exoskeleton was removed.

Most records were photographed on a running film. By such a recording method, the difference in the response pattern and the time delay between the responses recorded from two different parts could be measured with some accuracy and the functional disposition of the auditory tract in the cord could be explored. In order to study the responses from the tympanic and cereal nerves separately, one or other of them was cut in most experiments.

The stimulating and the recording equipment used in the present work was the same as that described in the previous papers (Katsuki, Watanabe & Suga, 1959; Katsuki & Suga, 1960).

(1) Response in the nerve cord

When the recording was made from the thoracic connectives of Gampsocleis buergeri, spontaneous discharges of several units were always observed, their spike heights being various. The responses to tone bursts were seen only at the onset of sound among spontaneous discharges. Interfering impulses from regions other than the tympanic nerve were eliminated by cutting the rostral and caudal parts of the hooked-up connectives and the other peripheral nerves except the tympanic. Thus only the responses to tone bursts remained, the size of impulses ranging between 1 and 3 mV. In the tympanic nerve, the train of impulses lasted as long as the stimulus sound continued. In contrast, in the thoracic connectives the responses were evoked only at the onset of sound, that is the ‘on’-type response. The conduction velocity measured at the suboesophageal-prothoracic connective was found to be about 6 m./sec. Pumphrey & Rawdon-Smith (1937) and Roeder (1948) have already reported that the conduction velocities of the cereal nerve and the giant fibre in a cockroach range from 2 to 3 m./sec. and 6 to 7 m./sec. respectively. Therefore, from the phasic discharge pattern, the large spike height, and the conduction velocity, it may be reasonable to conclude that the impulses originate from the large fibre in the connective.

The hair sensilla on the cerci can, as is already known, respond readily to low-frequency sound and the impulses evoked in the cereal nerve by sound are transmitted to the abdominal cord through the last (6th) abdominal ganglion. Our records, which were obtained from the abdominal nerve cord, always showed the distinct discharges of two units, the sizes of which ranged between 2 and 4 mV. The response pattern of the output of the last abdominal ganglion was more phasic as compared with that of the input. The conduction velocities of the fibres were about 6 m./sec. It was thus confirmed electrophysiologically that there were two auditory large fibres in the abdominal nerve cord.

(2) Auditory large fibre

The auditory large fibres in the nerve cord are divided into two : the auditory T large fibre and the auditory C large fibre.

(a) T large fibre

The patterns of the responses recorded from the connectives between the brain and the metathoracic ganglion were very similar. The impulses immediately evoked by the activity of the tympanic nerve at the prothoracic ganglion seemed to be conducted to the rostral and caudal ganglia through one and the same large fibre. In the hope of confirming this the impulses were recorded simultaneously from the connectives rostral and caudal to the prothoracic ganglion, that is to say, one recording electrode hooked up the suboesophageal-prothoracic connective and the other the ipsilateral meso-metathoracic connective. The descending impulses from the brain and the ascending impulses from the cereal nerve were excluded by cutting the rostral and, caudal parts of the connectives beyond the electrodes. The contralateral thoracic connectives and all the peripheral nerves except the ipsilateral tympanic were cut so as to leave only the simple system consisting of the unilateral tympanic nerve and the ipsilateral thoracic connective. The discharge pattern of the ascending impulses from the ganglion was exactly the same as that of the descending ones for any sound stimulus. The impulse sent to the suboesophageal ganglion is always delayed by 0·7 msec., compared with that directed to the mesothoracic ganglion. Thus it was confirmed that the impulses immediately evoked at the prothoracic ganglion were conducted rostrally and caudally on one and the same fibre.

Further attempts were made to discover whether an auditory large fibre of this type extends from the brain to the metathoracic ganglion. One of the electrodes hooked up the brain-suboesophageal connective and the other the meso-metathoracic connective. Fig. 1 represents an example of such simultaneous recordings. In each record of the figure, the upper and middle traces represent respectively the impulses sent up to the brain and down to the metathoracic ganglion, and the lower trace represents the sound stimulus. The frequency of the sound stimulus was changed as shown by the figures at the left side of each column. Good coincidence in the discharge patterns as between the upper and the middle traces is seen for the responses to all frequencies. The time delay between the impulses on the upper and middle traces was always 0·3 msec, and the spike heights were the same. This result shows that the auditory T large fibre runs in the cord as far as from the brain to the metathoracic ganglion and that the impulses in this fibre are initiated at the prothoracic ganglion, probably only by the activity of the tympanic nerve, and are conducted to the brain and to the metathoracic ganglion in exactly the same manner. Thus the information about the sound which stimulated the tympanic organ is sent up to the brain after about 12 msec, and down to the metathoracic ganglion after almost the same time from the arrival of the sound at the tympanic organ. The conduction velocity of this fibre was about 6 m./sec., measured at any part of the nerve cord.

The T large fibre does not extend to the abdominal ganglia, because no response originating from the tympanic organ was found beyond the metathoracic ganglion.

(b) C large fibre

The discharge pattern of the response evoked by impulses in the cereal nerve was found to be similar at all connectives from the last abdominal to the mesothoracic ganglion. The simultaneous records obtained from the connectives rostal and caudal to the metathoracic ganglion, however, did not show one-to-one correspondence of impulses. The responses recorded from the two connectives indicated that they consisted of the impulses of two large fibres. The number of impulses was always less at the thoracic connective than at the abdominal one, but the number of spontaneous discharges was just the reverse. From these results it may be said that these large fibres do not extend beyond the metathoracic ganglion, where they have synapses.

The impulses recorded from the metathoracic-abdominal connective, however, showed a perfect one-to-one correspondence with those from the V-VI abdominal connective in responses to sound and also in spontaneous discharges. The impulses at the rostral end of the abdominal cord were always delayed by 2·2 msec, relative to those at the caudal end. The conduction velocities of these fibres were almost the same, about 5·8 m./sec. Thus the abdominal nerve cord has two pairs of similar auditory large fibres which run to the metathoracic ganglion from the last abdominal ganglion. Those auditory large fibres are called the primary auditory C large fibres in this paper.

A pair of auditory large fibres were, as described above, found in the meso-meta-thoracic connective. These fibres were activated at the metathoracic ganglion by the primary C large fibres. For this reason the auditory large fibres between the meso-and metathoracic ganglia are called the secondary auditory C large fibres. In the herve cord rostral to the mesothoracic ganglion, the C large fibre tract was not found electrophysiologically, but the activities evoked by low-frequency tone bursts were observed on several small fibres.

(3) Interaction between impulses from the tympanic nerves of opposite sides

The tympanic nerve in Orthoptera consists of about 100 nerve fibres (Vogel, 1921). When the tympanic nerve was hooked up with a silver wire electrode, the grouped discharges of those fibres were observed as long as the sound lasted. However, in each fibre a sigmoid relation was found between the sound intensity on a decibel scale and the number of impulses per sec. The information about the sound intensity can be thus signalled to the central nerve cord. On the other hand, the responses of the T large fibres were phasic, so it may be that the information about the intensity is not sent to the brain in the same form as it has in the tympanic nerve.

In order to study in detail the responses of the T large fibres, a pair of connectives between the suboesophageal and prothoracic ganglia was hooked up on the electrodes, the ascending impulses from the rear ganglia being eliminated by cutting the connectives. There remained both the tympanic nerves, the prothoracic ganglion, and both the ascending connectives. In Fig. 2 the upper and middle traces show the impulse discharges of the tympanic large fibres of the right and left sides respectively. The lower trace shows the wave form of the delivered sound, the frequency of which is 13 kcyc./sec. A and B represent respectively the responses before and after cutting the left tympanic nerve. The T large fibre of the left side, which was activated by the impulses sent up from the tympanic nerve nearer to the loud-speakers, sent more impulses than that of the right side (A). Here a very interesting phenomenon was found as a result of cutting the left tympanic nerve (B): no impulse discharge was found on the left side but in response to the same sound stimulus a remarkable increase in the number of impulses was found on the right side. When the right tympanic nerve was cut, the reverse effect was observed. When there were many impulses in a response to sound, the increase in number of impulses after cutting the nerve was not so marked, whereas when the number of impulses in the response was fewer, the increase after the cut was more marked. It was found that the impulses delayed by more than 4·6 msec, after the first impulse in the response were suppressed by the impulses of the contralateral tympanic nerve. This phenomenon shows that impulses of the tympanic nerve on one side have an inhibitory effect on the contralateral T large fibre.

In three cases out of twenty-six the impulses in the T large fibre remained even after cutting the ipsilateral tympanic nerve, though the number was small. Disappearance of these remaining impulses after cutting the contralateral tympanic nerve proved that they were evoked by the activity of the contralateral nerve. They were delayed by several milliseconds as compared with the ipsilaterally evoked impulses. The shortest delay was 3·3 msec. When the contralaterally evoked impulse was observed, an attempt was made to discover whether the contralaterally and ipsilaterally evoked impulses converged on the same T large fibre. No difference in the spike height and no summation of spikes were found among them. The delay of the first spike transmitted from the contralateral tympanic nerve was always about 10 msec. rom the onset of the sound stimulus, but the spike with this delay was not found regularly in the response of the T large fibre before cutting the ipsilateral tympanic nerve. These facts suggest that the ipsilaterally and the contralaterally conducted impulses in the tympanic nerves activate one and the same T large fibre. However, there still remains the question of whether the inhibitory effect is exerted upon the tympanic nerve. By recording the electrical activity from the latter it was confirmed that the inhibitory effect of the tympanic nerve on one side does not extend to the tympanic nerve of the opposite side.

Therefore it is highly probable that the T large fibre receives not only the excitatory effect from the ipsilateral tympanic nerve at the prothoracic ganglion but also the inhibitory and weak excitatory effects from the contralateral nerve. The irregularity of the contralaterally evoked impulses tells us that the inhibitory effect may be varied by intrinsic factors.

The discharge pattern of the T large fibre was phasic as described above, but it was due to the inhibitory effect of fibres from the opposite side. When the contralateral tympanic nerve was cut, though the T large fibre responded with increased impulses, the response adapted so that the train of impulses lasted at most only for the first 0·1 or 0·2 sec. of continuous tones of o db. on our scale (corresponding to about 100 db. above average human threshold for 1 kcyc./sec.), the frequencies being higher or lower than the characteristic frequency of the tympanic organ (Katsuki & Suga, 1960) ; the discharge continued for about 1 sec. in the response to sound of the characteristic frequency at the same intensity.

On the other hand, such an inhibitory effect was not disclosed in the primary C large fibres. When one of the cereal nerves was cut, only a few impulses in response to sound remained in the ipsilateral large fibres and many others disappeared. On the contralateral large fibres, a decrease of a few impulses was observed as compared with the original response. Therefore the interaction between the impulses from the two cereal nerves was only excitatory.

(4) Response ranges of the auditory large fibres

As already reported (Katsuki & Suga, 1958, 1960), the tympanic organ of Gampsocleis buergeri responds to the sounds of 0·6−75 kcyc./sec. and among them most sensitively to 10 kcyc./sec. sound. The thresholds of the T large fibre for sounds of various frequencies were measured at the brain-suboesophageal connective after cutting off the brain. The frequency range responded to was found to be from 2 to 60 or 70 kcyc./sec. and the most effective frequency from 10 to 20 kcyc./sec. In the previous paper (1960), the authors reported two ranges of response in the tympanic nerve of this insect. The response range of the T large fibre was found to be almost the same as one of them, i.e. that in response to higher-frequency sounds. The threshold of this fibre for the sound of the most effective frequency was not higher than, but almost the same as, those of the tympanic neurons. The responses which might be evoked by the impulses of neurons of the other type in the response range were not found in the central nerve cord, at least not in the large fibre.

On the other hand, the threshold curve of the cereal hair sensilla covers the range of sounds from below 30 up to 2000 cyc./sec. Sounds from 400 to 800 cyc./sec. were most effective to this sense organ. The frequency range of the primary C large fibre was the same as that of the peripheral cereal nerve. No difference was found in the threshold curve between the cereal nerve and the primary C large fibre.

Coming now to directional sense, it was to ultrasonic waves that the largest difference in sensitivity between a pair of tympanic organs of the locust* was found (Katsuki & Suga, 1960). Therefore the difference in the response range between the two T large fibres of Gampsocleis buergeri can be anticipated.

Simultaneous records were made from a pair of the suboesophageal-prothoracic connectives. The threshold of the left T large fibre , directed towards the sound source, was lower for all frequencies as compared with that of the opposite one , as shown in Fig. 3. After cutting the left tympanic nerve the response range of the right T large fibre did not change in spite of the marked increase in the number of impulses. This fact shows that some of the tympanic nerve fibres can operate to modify the content of the information sent by the contralateral T large fibre without changing its response range, because the inhibitory effect from the contralateral tympanic nerve suffers some time delay.

(5) Modification of information in the T large fibres

The extent of modification of information in a pair of the T large fibres by the inhibitory interaction already described was next studied. The tone bursts used had a constant duration of 70 msec, and either their intensity or their frequency was varied. In A of Fig. 4 the ordinate and the abscissa represent respectively the number of impulses in the response and the frequency of the sound in kcyc./sec. The intensity of the sound was kept constant at 0 db. A difference in the number of impulses was found between the responses of both left and right fibres only to sounds of from 10 to 20 kcyc./sec., and the most marked difference at 17 kcyc./sec. When the left tympanic nerve (nearer to the loud-speakers) was cut, the response of the left fibre (responding with more impulses before cutting) disappeared completely, and on the opposite right fibre a marked increase in the number of impulses immediately appeared as a result of the release of the inhibitory effect from the left tympanic nerve. The frequency range of sounds which activated the T large fibre was from 2 to 65 kcyc./sec. The increase of impulse discharge after the elimination of the inhibitory effect was found in the response to sounds of from 5 to 30 kcyc./sec., but the marked difference in the number of impulses between the two large fibres under inhibitory effect from the opposite side was found in the narrower range as described above. This range almost coincides with the frequency range most effective to the tympanic organ and also with the dominant frequencies involved in the stridulatory sound of the group.

In B of Fig. 4 the ordinate and abscissa show respectively the number of impulses in the response and the intensity of the sound in decibels. As the change in the number of impulses was small in the frequency regions both lower and higher than the most effective region, the curves were plotted for the responses to the most effective frequency, at which the change was very marked. The right T large fibre , situated farther from the sound source, responded with only one spike from o to −50 db., while the left one responded with several spikes even to a weak sound. When the left tympanic nerve was cut, the right T fibre showed many impulses, (the response evoked by the right tympanic nerve alone) but no change of the threshold

(6) Response to the stridulation of the group

The responses of the tympanic nerve to the stridulation of the group show, as already known (Haskell, 1956, 1957; Katsuki & Suga, 1958, 1960), good synchronization with the pulsatory sounds which compose the stridulation. What information about the stridulation is sent through the nerve cord to the brain, however, is still obscure.

The responses were simultaneously recorded from the T large fibre and also from the primary C. Several stridulating males (Gampsocleis buergeri) were placed in a bamboo cage at a distance of about 50 cm. from the insect being recorded. The T large fibre discharged synchronously with the pulsatory sounds, but the primary C did not. From this fact it can be said that auditory communication in the group is mainly mediated by the tympanic organ.

When the responses to stridulatory sounds were simultaneously recorded from a pair of the T large fibres, the impulses of the T large fibre closer to the source showed a distinct one-to-one correspondence with the pulsatory sound, but those of the opposite fibre did not (A of Fig. 5). When the source was moved from the side to a position in front of the insect being recorded, the impulse discharges of both connectives became almost the same. After one of the tympanic nerves was cut, the contralateral T large fibre responded to the pulsatory sound with a volley of impulses. Only three or four impulses were synchronized with the initial weaker pulsatory sounds while a train of impulses was observed to the later stronger ones (B of Fig. 5).

As described above, the impulses arising in the tympanic nerve in response to sound evoke spikes in the T large fibre in the cord. However, the records obtained from the suboesophageal-prothoracic connective showed that those impulses were transmitted not only to the large fibre, but also to a small fibre, the size of whose impulses, whenever found, was less than half of that of the large fibre. The response range of the small fibre was nevertheless the same as that of the large one.

On the other hand, besides two C large fibres described above, one or two units were found in the abdominal nerve cord. In this paper, however, the description is limited only to the auditory T and C large fibres which were observed very clearly and constantly.

(1) Auditory large fibres

The functional analysis of the nerve cord in Arthropods has been thoroughly carried out in the crayfish by many authors (Wiersma, Ripley & Christensen, 1955 ; Wiersma, 1958; Furshpan & Potter, 1959a, Kennedy & Preston, 1960; Preston & Kennedy, i960; Hughes & Wiersma, 1960). A T-shaped interneuron was found anatomically in the abdominal nerve cord by Allen (1894). Hughes & Wiersma (1960) reported that certain abdominal peripheral nerves evoked simultaneously ascending and descending impulses in one and the same abdominal nerve fibre. The anatomical study of the nerve cord of Gampsocleis buergeri by the present authors has revealed that there are T-shaped neurons and a pair of large fibres (about 27 μ in diameter) which fie in the cord between the brain and the metathoracic ganglion. The suboesophageal-prothoracic connective has 8 or 9 fibres larger than 25 μ in diameter and the largest of these is 38 μ. In the abdominal nerve cord five large fibres are found with a diameter of 20−27μ. But it is difficult at present to determine which of these is the auditory large fibre.

(2) Directional sense

The inhibitory interaction between the tympanic nerves of opposite sides has no effect on the difference between the thresholds of left and right T large fibres, in spite of the large modification of information carried by those fibres. From the point of view of the threshold the difference in responses between the large fibres therefore follows that between the two tympanic nerves. If the tympanic large fibres were released from the mutual inhibitory interaction, each of them could send a train of impulses and the information about the intensity difference would simply correspond to that in the tympanic nerves. In fact, one tympanic nerve firing with many impulses will activate strongly the ipsilateral T large fibre while strongly suppressing the activity of the contralateral one. By such a mechanism, the information about a sound source is increased and the information is sent quickly to the brain through a pair of T large fibres. Such an inhibitory interaction may be one of the most important factors in the mechanism of directional sense. The inhibitory interaction was also clearly observed in the nerve cord of Homoeocoryphus lineosus.

(3) Frequency analysis

It is difficult to understand that about 100 tympanic neurons connect with only two of the central neurons. However, during the course of the present experiments there was no sign that some of the tympanic nerve fibres or any central small fibres (except for two fibres described above) sent impulses to the brain beyond the prothoracic ganglion. The response ranges of the T large fibres were narrower than those of the tympanic nerve (quite recently, the same phenomenon in the locust and cricket was noticed by Horridge (1960)) and coincided with one type of response range of single tympanic neurons. The characteristic frequency was always found to be between 10 and 20 kcyc./sec. There was also no evidence that the central neurons activated by the tympanic nerve sent impulses into the brain with different response patterns for different frequencies of sounds. Neurons having different characteristic frequencies were not found in the peripheral nerve of a locust, notwithstanding that the sensilla of the tympanic organ are divided into three groups (Suga, 1960). However, from his observation that the characteristic frequency of an ascending central neuron is shifted by a continuous pure tone, Horridge (1960) has suggested the existence of two groups of sensory neurons in the tympanic organ or in the ganglion, that is to say, the existence of a mechanism of frequency analysis. In Homoeogryllus japonicus, the present authors measured the thresholds for tone bursts with various frequencies before and after a continuous pure tone of a certain frequency was delivered and found that the thresholds became relatively high for tone bursts with a background tone. The authors think that Horridge’s suggestion is quite reasonable that there are at least two groups of sensory neurons, but the fact that impulses from two groups are transmitted in only one central neuron does not confirm the possibility of frequency analysis even if the response range of the neuron was shifted by a continuous tone. If each of the central neurons which are activated by the tympanic neurons has a different frequency range and characteristic frequency, we can entertain the possibility of frequency analysis in the tympanic organ and its neural pathway. As things stand now, we shall have to look over the whole body surface of an insect if we believe in the possibility of frequency analysis in insects.

The large fibre activated by sound reception of the cereal hair sensilla sends up the information to the mesothoracic ganglion through the relay at the metathoracic ganglion. The information sent up to the mesothoracic ganglion is transmitted to some fine fibres in the ganglion and finally sent into the brain. The response ranges of these C large fibres correspond with the threshold curve of the whole cereal nerve.

The difference in the frequency range between the two sound receptive organs, the tympanic organ and the cereal hair sensilla is very remarkable, and they send impulses into different auditory large fibres. Therefore it will be concluded that the insect can analyse sound frequencies to some extent, probably by the spatial and temporal pattern of impulses. The tympanic organ together with its neural network seems to be elaborated to receive the stridulatory sound of the group.

In order to support the interpretation of the central interaction described above, several pharmacological agents which have a specific action on inhibitory or excitatory synapses were applied to the prothoracic ganglion and the change in impulse discharges of the auditory T large fibres was examined. The results afforded evidence that the central interactions described above were synaptic events. The details of the pharmacological experiment will be reported in a separate paper.

We are indebted to the Ministry of Education of Japan and the Rockefeller Foundation (GA MNS 59115) for the financial support of this work. We would also like to thank Dr B. P. Uvarov for indicating the correct name of a locust.

Allen
,
E. D.
(
1894
).
Studies on the nervous system of Crustacea, 1. Some nerve elements of the embryonic lobster
.
Quart. J. Micr. Sci
.
36
,
461
82
.
Furshpan
,
E. J.
&
Potter
,
D. D.
(
1959a
).
Transmission at the giant motor synapses of the crayfish
.
J. Physiol
.
145
,
289
325
.
Furshpan
,
E. J.
&
Potter
,
D. D.
(
1959b
).
Slow post-synaptic potentials recorded from the giant motor fiber of the crayfish
.
J. Physiol
.
145
,
326
35
.
Haskell
,
P. T.
(
1956
).
Hearing in certain Orthoptera. I, II
.
J. Exp. Biol
.
33
,
756–66
,
767
76
.
Haskell
,
P. T.
(
1957
).
The influence of flight noise on behaviour in the desert locust Schistocerca gregaria
.
J. Insect. Physiol
.
1
,
52
75
.
Haskell
,
P. T.
&
Belton
,
P.
(
1956
).
Electrical responses of certain Lepidopterous tympanal organs
.
Nature, Lond
.,
177
,
139
40
.
Horridge
,
G. A.
(
1960
).
Pitch discrimination in Orthoptera (Insecta) demonstrated by responses of central auditory neurons
.
Nature, Lond
.,
185
,
623
4
.
Hughes
,
G. M.
&
Wiersma
,
C. A. G.
(
1960
).
Neuronal pathways and synaptic connexions in the abdominal cord of the crayfish
.
J. Exp. Biol
,
37
,
291
307
.
Katsuki
,
Y.
&
Suga
,
N.
(
1958
).
Electrophysiologial studies on hearing in common insects in Japan
.
Proc. Jap. Acad
.
34
,
633
8
.
Katsuki
,
Y.
&
Suga
,
N.
(
1960
).
Neural mechanism of hearing in insects
.
J. Exp. Biol
,
37
,
279
90
.
Katsuki
,
Y.
,
Watanabe
,
T.
&
Suga
,
N.
(
1959
).
Interaction of auditory neurons in response to two sound stimuli in cat
.
J. Neurophysiol
.
22
,
603
23
.
Kennedy
,
D.
&
Preston
,
J. B.
(
1960
).
Activity pattern of interneurons in the caudal ganglion of the crayfish
.
J. Gen. Physiol
.
43
,
655
70
.
Preston
,
J. B.
&
Kennedy
,
D.
(
1960
).
Integrative synaptic mechanism in the caudal ganglion of the crayfish
,
J. Gen. Physiol
.
43
,
671
81
.
Pringle
,
J. W. S.
(
1953
).
Physiology of song in cicadas
.
Nature, Lond
.,
172
,
248
.
Pumphrey
,
R. J.
&
Rawdon-Smith
,
A. F.
(
1936
).
Hearing in insects: The nature of the response of certain receptors to auditory stimuli
.
Proc. Roy. Soc. B
,
121
,
18
27
.
Pumphrey
,
R. J.
&
Rawdon-Smith
,
A. F.
(
1937
).
Synaptic transmission of nervous impulses through the last abdominal ganglion of the cockroach
.
Proc. Roy. Soc. B
,
122
,
106
18
.
Roeder
,
K. D.
(
1948
).
Giant fiber system, roach
.
J. Exp. Zool
.
108
,
243
62
.
Roeder
,
K. D.
&
Treat
,
A. E.
(
1957
).
Ultrasonic reception by the tympanic organ of noctuid moths
.
J. Exp. Zool
.
134
,
127
58
.
Suga
,
N.
(
1960
).
Peripheral mechanism of hearing in locust
.
Jap. J. Physiol
,
10
,
533
46
.
Vogel
,
R.
(
1921
).
Bericht über ein Gehdrorgan bei Singzikaden
.
Naturwissenschaften
,
32
,
427
31
.
Wiersma
,
C. A. G.
(
1958
).
On the functional connexions of single units in the central nervous system of the crayfish, Procambarus clarhii, Girard
.
J. Comp. Neurol
,
110
,
421
72
.
Wiersma
,
C. A. G.
,
Ripley
,
S. H.
&
Christensen
,
E.
(
1955
).
The central representation of sensory stimulation in the crayfish
.
J. Cell. Comp. Physiol
.
46
,
307
26
.
*

Locuita migratoria danica in the previous paper was erroneous. It should be L. migratoria minilensis