Sound Reception in the Brazilian Cicada Fidicina Rana Walk^*

Sound production is a characteristic of male cicadas, and the sound-producing mechanism has been studied by several authors (cf. Pringle, 1954; Hagiwara & Ogura, 1960; Hagiwara & Watanabe, 1956). Sound reception, however, has been little studied. Katsuki & Suga (1958) found that in four species of Cicadidae the range of sound frequencies received was from a few hundred Hertz (Hz.) to 15-20 kHz. (above 20 kHz. in one species). The frequencies of highest sensitivity were 4-5 kHz. or lower. These values coincide very well with the frequency components of the cicada song. Thus, from sonograms of the song of the same species it appears that the peak sound energy is for frequencies below 10 kHz., and sound production of frequencies above 20 kHz. is virtually absent (Hagiwara & Ogura, 1960).

During the 1967 Amazon Expedition of the R.V. Alpha Helix of the University of California we had the opportunity to study sound reception in a large cicada, Fidicina rana Walk. Sound production in the same species was studied at the same time by one of us and is reported separately (Aidley, 1969).

Seven specimens of Fidicina rana Walk (four females, three males) were used, kindly identified by Dr M. S. K. Ghauri of the British Museum. The insect was pinned with its dorsal side down on a piece of cork and the whole tympanic nerve of one side was exposed, cut and lifted on to a silver wire electrode, which in turn was connected to an A.C. amplifier and a double-beam oscilloscope. An indifferent electrode was put on the abdomen.

The tympanic organ was facing an audio loudspeaker at a distance of about 1 m. Four types of acoustic stimuli were delivered: (1) pure tones in pulses of usually 200 msec, duration; (2) trains of sound clicks produced by feeding $12$ msec, long pulses of a 4 kHz. tone into the loudspeaker (the clicks had also components of higher and lower frequencies) ; (3) play-backs of tape recordings of the cicada song ; and (4) distress calls of a live cicada.

The sound stimulus was monitored on the second beam of the oscilloscope. Sound pressures were measured by a calibrated microphone (⁠$14$ in. Brüel and Kjær, Copenhagen), placed 1-2 cm. above the tympanic organ.

The response to pure-tone pulses consisted of an initial series of a few compound action potentials of rapidly decreasing amplitude, followed by a sustained discharge of impulses, presumably unitary, but from several units (Fig. 1). The latency of the first compound potential was about 10 msec. At the higher sound frequencies the onset of the sinusoidal tone was associated with a sound click. Control experiments with click-free stimulation were performed in a primitive manner by quickly removing a foam plastic cushion held in front of the loudspeaker. No compound potential, or only a faint compound potential, was recorded under these circumstances, but the sustained response was the same as when the stimulus started with a click. Therefore, only the sustained discharge has been used in establishing the audiogram. This discharge was measured by counting the number of impulses above a certain amplitude, which was chosen so high that the pre-stimulatory, ‘spontaneous’ activity and the low-amplitude potentials of the response were counted as zero. The first 50 msec, of the response was not counted in order to exclude the initial response.

The relationship between the sustained discharge rate and the relative sound pressure is shown in Fig. 2 for a few sound frequencies. In this example the threshold was lowest for 9 kHz., and in six of the seven specimens tested the threshold minimum ranged from 6 to 9 kHz., and in one specimen it was at 4 kHz. An audiogram, based on graphs of the type shown in Fig. 2, is presented in Fig. 3. Thresholds are taken as those sound pressures necessary to elicit four action potentials (above a certain amplitude) in 100 msec. The sound pressures were measured in absolute units, and in Fig. 3 the relative pressure level of O dB. corresponds to an absolute pressure of −12 dB. ref. 1 μbar. It follows from the manner in which the responses are measured, however, that the true thresholds are much lower than −12 dB. Thus, often the sustained discharge was above the pre-stimulatory ‘spontaneous’ level without any action potentials being of sufficient amplitude to be counted. Furthermore, the responses to the initial phase of a tone pulse of the optimal frequency and responses to click stimulation were recorded for sound intensities lower than − 22 dB.

The response to a train of sound clicks consisted of a series of compound action potentials in phase with the clicks at frequencies up to about 300/sec. For higher frequencies and during prolonged stimulation with click repetition rates above 200/ sec., this synchrony broke down (Fig. 4). The amplitude of the action potentials was fairly constant at repetition rates of 100 clicks/sec. and less; at higher repetition rates the amplitude decreased with time and at a rate of 400/sec. (which is near to the click frequency in the cicada song) the amplitude was reduced to one third after four or five consecutive clicks and adapted slowly through the rest of the stimulation period (Fig. 4).

The response to play-backs of cicada song depended upon the type of the song, namely—using the terminology of Aidley (1969)—the continuous song, the distress call and the ‘zeep’ call. The response to continuous song consisted of an initial series of compound action potentials after which a rather irregular discharge appeared. The amplitude of these potentials decreased on the average, but the adaptation of the response was not nearly so marked as during artificial click stimulation of the same frequency. The distress call differs from continuous song in that the series of clicks is interrupted at intervals. Resumption of the sound after each interruption results in large compound action potentials (Fig. 5B). The response to ‘zeep’ calls, however, consisted of a series of compound action potentials all of about the same amplitude (Fig. 5 A). The ‘zeep’ call itself is a short series of clicks of steadily increasing amplitude. Thus, the tendency of the response to adapt is counter-acted by a steady increase in sound pressure throughout the call.

The purpose of the present work was to study sound reception in the Brazilian cicada Fidicina rana and to compare this with the acoustic characteristics of its own song. The two are evidently well matched to each other; it would in fact have been rather surprising if this were not so.

The song consists of a series of clicks with a dominant frequency in each click of 10-12 kHz., but there is also a considerable component at half this value, 5-6 kHz. (Aidley, 1969). These values bracket between them the most sensitive range in the audiogram shown in Fig. 3. It should be remembered that this audiogram provides a measure of the sensitivity of slowly adapting units, and it is possible that more rapidly adapting units, which would be excited by brief sound pulses but not during the later stages of a more prolonged pulse, might possess a slightly different relation between sound frequency and sensitivity. It seems unlikely, however, that such a difference would be at all large.

The auditory capabilities of Fidicina are not very different from those of the Japanese cicadas studied by Katsuki & Suga (1958). The most sensitive frequency range in Fidicina (6-9 kHz.) is somewhat higher than in the Japanese species, where it varies from 5 kHz. in Platypleura kaempferi and Graptopsaltria nigrofuscata down to 13 kHz. in Tanna japonensis.

The true auditory threshold at the most sensitive frequency could not be determined in the present study, owing both to the technique for recording of nervous activity and to the high background noise level aboard the ship. However, we obtained responses to stimuli of intensities as low as − 22 dB. ref. 1 μbar, which corresponds to 52 dB. above the standard human threshold level of 2 × 10⁻⁴μbar. This figure is not too far above the corresponding values for other insects; Suga (1966), for example, gives values of 30 dB. (ref. 2 × 10⁻⁴μbar) for one species of neotropical grasshopper and 40 dB. for three others.

The most marked feature of the nervous response to auditory stimulation is the very rapid adaptation which occurs, both in response to pure tones of constant intensity and in response to a series of clicks at high repetition rates. The possible basis for this adaptation is worthy of some discussion. Adaptation in response to a continuous pure tone is analogous to the adaptation which occurs in many sense organs in response to steady stimuli; the frequency of firing of the individual units declines more or less smoothly, and the adaptation seen in the response of the whole nerve is but the sum of the adaptational properties of the individual units. The situation during a response to a series of clicks is slightly different, however. Here, each click can elicit in a particular unit either one action potential or none, and adaptation must then consist of a reduction in the probability of firing in response to successive clicks. In addition, it is possible that the latency between click and response differs between different units in a different manner as the train of clicks proceeds, so that the individual units do not fire synchronously; this would result in a reduction in the size of the compound action potentials recorded in the whole nerve, and might lead to spuriously high estimates of the degree of adaptation.

As previously described, there is no adaptation in response to a train of clicks at repetition rates of ioo/sec. or less. This is in accordance with Pringle’s (1954) observation that there is no adaptation in the auditory nerve response of the Sinhalese cicada Platypleura capitata to clicks at repetition rates up to 93/sec. In Fidicna, however, there is very considerable adaption at repetition rates similar to those occurring in the song, a fact which is at first sight rather surprising. Why is the repetition rate of clicks in the song set at a value of 400-550/sec., when the auditory system seems to be sensitive to clicks when these occur at much lower repetition rates? The answer to this may be that the important feature of the auditory response is not the size of the response to individual clicks but the total activity over a given time; this activity will, up to a point, increase with increasing click frequency.

A rough way of measuring total activity in response to a train of clicks is to measure the size of consecutive compound action potentials and to summate these values over a given period of time. Using this procedure we found that the activity during the first 200 msec, of stimulation was about the same for the repetition rates of 85, 200, 300 and 400 clicks/sec. Setting the response to 85 clicks/sec. equal to 100%, the others were 107, 101 and 88%, respectively.

Adaptation occurs while the insect is listening to tape recordings of continuous song, although it is apparently not so pronounced as when artificially produced clicks at the same repetition are used. Two possible explanations for this effect occur to us: (1) the clicks produced during the song are not constant in amplitude—particularly if the song is recorded from one side of the animal (see Aidley, 1969)—and the intervals between them are not precisely equal ; and (2) the waveforms of our artificially produced clicks were not identical with those occurring in the song.

In view of this adaptation that takes place during continuous song, it is particularly interesting that there are two other natural calls of Fidicina that are apparently designed in such a way that the adaptational effects are circumvented, so that these calls appear to constitute an unusually effective auditory stimulus and hence probably provide an extremely effective method of intraspecific signalling. First, consider the distress call (Fig. 5B). This consists of trains of loud clicks which are frequently interrupted by silent intervals lasting 30 msec, or more. During each burst of clicking the auditory nerve response adapts, but the silent periods allow the receptor to recover so that the first few clicks following each silent period result in large compound action potentials. In view of this effect, it is perhaps not surprising that the distress call appears to function in part as an intraspecific warning signal (Aidley, 1969). Secondly, the ‘zeep’ call (Fig. 5 A) consists of a train of clicks whose amplitude increases more or less linearly with time. This increase in amplitude appears to counteract rather precisely the tendency of the auditory nerve to adapt, so that the neural response consists of a maintained high level of activity. Such a response is probably not produced by any other natural stimulus.

This work was supported by the R.V. Alpha Helix Amazon expedition, 1967. We are very grateful to Professor T. H. Bullock for inviting us to join the expedition and for the stimulating discussions of this work which we had with him.