Analysis of Drymopsalta songs

Terminology. The songs of each species comprise the same basic pulse structures, and are constructed into four basic forms. These are: (1) single ticks, each comprising a set of simple double pulses; (2), chirps, comprising groups of multiple ticks, typically 2–16 in number, but more in some song variants. In the normal songs, most commonly one or two single ticks occur between a pair of chirps, but up to nine for the D. crepitum song. (3) Phrase, comprising a chirp and associated, intervening tick(s); (4), buzzing echemes, extended and continuous emission of ticks, lasting from 0.1->30seconds, and sounding aurally as a continuous buzzing sound. These song types are illustrated in Figs 9–11, 13–16, 18–19.

A. Normal calling songs

Each of the Drymopsalta normal songs follows a similar pattern and structure, specifically the emission of chirps, each comprising between 2 to 16 distinct ‘ticks’ (Tables 1–3). Between the chirps occur 1 ( D. wallumi, D. hobsoni and D. acrotela), 2 ( D. daemeli), or 1 to 9 single ticks ( D. crepitum). Each tick, either singly or as occurring within the chirps or echemes, comprises a single dominant primary pulse followed by a weaker secondary pulse (e.g. Figs 9 B, D, and especially clear in Figs 16 and 19 B–D). In detail, the duration between the two pulses varies slightly between species, the following values measured in ms (means±σ, number of measurements, minimum–maximum range): D. wallumi 1.81±0.43, 141, 1.0–2.6; D. hobsoni 2.28±0.27, 133, 1.7–2.8; D. acrotela 2.39±0.11, 84, 2.2–2.7; D. daemeli 3.44±0.47, 96, 2.6–4.6; D. crepitum 3.12±0.12, 93, 2.8–3.4.

PLATE 5. Drymopsalta crepitum Ewart. A. male, north side of Pennefather River, western Cape York Peninsula, PS1957; B, female, Heathlands Station, northern Cape York Peninsula, PS1956; body lengths 11.5 and 12.9 mm, respectively.

D. wallumi, D. hobsoni and D. acrotela .

The songs of these three species are treated together as each has only a single tick between chirps. A useful diagnostic song parameter that distinguishes D. wallumi from the other two species is the timing of the tick emission relative to the adjacent chirps. This is defined by the ratio R1/R2 where R1 is the duration between the tick and the preceding chirp, while R2 is the duration between the tick and the following chirp (Figs 9 A, 13C). For D. wallum i, this ratio is normally>0.8, with mean values>1. For D. hobsoni and D. acrotela, the ratio is <0.75, and in fact completely overlaps between these two species. As discussed below, these two species have indistinguishable songs, an observation that we find to be very unusual for Australian cicadas.

Summaries of the song parameters of the three species are listed in Tables 1 and 2. Chirp repetition rates and pulse repetition rates tend to be slightly lower in D. hobsoni and D. acrotela compared to D. wallumi . Pulses per chirp are comparable between the three species. The overall similarities between these three songs is seen in Figs 9, 10, 13 and 16, the latter emphasising the close similarity between the fine scale pulse structures (and therefore mechanisms) as observed in the envelope curves.

Amplitude spectra for D. wallumi (Fig. 12) from the five studied localities show comparable dominant frequencies, lying between 19.7–21.5 kHz, effectively in the ultrasonic range. Each spectrum also shows broad band frequencies, overall 4.5 to 6.9 kHz. It is noted that both of these parameters could be slightly underestimated due to possible reduction of sensitivity response of recording equipment at the highest measured frequencies. Apparent sidebands show multiple values, the highest frequencies most likely reflecting the pulse repetition rates within the chirps and ticks (~140->300 Hz), the lower frequencies corresponding to the chirp repetition rates (~3–6 Hz), while the chirp phrase durations may account for the intermediate sideband frequencies, these ranging between approximately 13 to 47ms (21–77 Hz; data from Table 1).

Amplitude spectra for D. hobsoni and D. acrotela (Fig. 17) indicate dominant frequencies of near 22 kHz and 20 kHz, respectively, with band widths of approximately 4 to>6 kHz. In the case of D. hobsoni, the spectra indicate that the dominant song frequency envelopes do continue beyond the recording limit of 24 kHz (possibly extending to 28 kHz based on bat detector readings), and that the dominant frequencies and bandwidths are likely underestimated. The distribution of the sideband frequencies are explained as previously.

D. daemeli (Table 3, Figs 18, 20 A, B).

The song consists of short chirps, each normally consisting of between 2 to 4 ticks, each with the same pulse structures as described for the previous species. Between the chirps are most commonly 2 ticks (less often 1 tick as per the species above), each separated, and which are here specified as the ratios R1/R2 applied to the first tick, and R3/R4 applied to the second tick. For each of the two ticks individually, these define the ratio of the duration from the preceding chirp phrase to the duration of the following chirp phrase (Fig.18 A). It is noted that the R1/R2 ratios are quite similar to the R1/R2 ratios observed within the D. hobsoni and D. acrotela songs. The timing of the second tick within each phrase is clearly more variable. Chirp repetition rates are similar to the D. wallumi song, while the pulse repetition rates in the chirps are distinctly lower than measured in the D. wallumi, D. hobsoni and D. acrotela songs, this also evidently reflected in the longer inter-pulse durations within the ticks of the D. daemeli song, and the significantly lower pulse repetition rates within the chirps.

The amplitude spectra (Figs. 20 A, B) indicate dominant frequencies of 19.3–19.4 kHz and relatively high bandwidths, comparable to the previously described species although with slightly lower dominant frequency. The highest frequency sidebands (104–105 Hz) appear to correlate with the pulse repetition rates, the lower frequency sidebands at 3–5 Hz with the chirp repetition rates. Chirp durations lie in the general range 15–45 ms, equivalent to frequencies of 22–67 Hz accounting for at least some of the intermediate sideband frequencies.

D. crepitum (Table 3, Figs 19, 20 C)

The remarkably diverse song structure of this species is described in Ewart (2005). The most commonly emitted song, referred to as the normal calling song, is illustrated in Fig. 19 based on a previously unpublished field recording by D. Marshall and K. Hill. This shows the characteristic repeated chirps, each consisting of most commonly 3–8 ticks, with multiple single ticks emitted between chirps, most commonly 3–5, but ranging between 1–9 in the range of song variants described in Ewart (2005). The structures of the ticks (Figs 19 B–D), with the primary and secondary pulses, is comparable to those described for the other Drymopsalta species. The measured inter-pulse durations (3.12±0.12) are between the values observed in D. daemeli and D. hobsoni . Thus, although the D. crepitum normal song exhibits the same basic structures as the previously described songs, it clearly differs in respect to the larger number and timing of the intervening ticks. This results in chirp repetition rates that are significantly lower than observed in the previous species. Pulse repetition rates within the chirps are intermediate between those of D. daemeli and D. wallumi (Tables 1, 3).

The amplitude spectrum (Fig. 20 C) has the dominant frequency at 17.6 kHz, lower than the other described Drymopsalta songs, but again a broadband song (>6 kHz). The highest sideband frequencies also correlate with the pulse repetition rates in the chirps, the lowest with the chirp repetition rates and the inter-tick durations (between chirps and to nearest chirp; Table 3).

Ratio of single click distance from preceding and following chirps (see Fig. 9 A) (1) Ratio of single click distance from preceding and following chirps (see Fig. 13 C)

B. Buzzing echemes

The emission of extended buzzing, interspersed within the normal and ‘anomalous’ (see below) songs, is characteristic of four of the described species, the exception being D. wallumi . In Tables 2 and 3, the durations and properties of the buzzing component are based on available electronic recordings, but field observations indicate that the durations of buzzing phases do in some cases extend to beyond 1 minute. Figs 14 A, 15A and 17A, C–D illustrate the common patterns of the buzzing echemes. These occur as continuous buzzing, buzzing interspersed within normal and ‘anomalous’ chirps (e.g. Fig. 14 A), and sometimes as quasi-continuous buzzing (e.g. Fig.15 A). The time expanded plots of the buzzing show them to comprise regularly repeated strings of ticks, each identical to those of the single ticks and chirps (i.e. the double pulses as described above; Figs 14 B). The buzzing song of D. daemeli, when examined in time expanded plots, show that the buzzes are not actually continuous, but occur as closely spaced sets of shorter ‘echemes’ (or chirps), each ranging from 210–890 (mean 460) ms in duration, separated by gaps with durations of 41–74 (mean 48) ms. The pulse repetition rates within each buzzing component are comparable to those in the chirps, sometimes slightly lower. The emission of the buzzing echemes appears to occur most commonly during the hottest periods of the day, and also seems to also coincide with high population densities. Amplitude spectra are similar to the normal calling songs (Figs 17 A, C, 20B).

D. daemeli D. crepitum (2) (1) Definitions of R1 to R4, the ratios of the durations between the ticks and the adjacent chirps, are presented in Fig. 18 A. (2) Field recording by D. Marshall and K. Hill.

(3) Based on recordings from Normanton, N. Queensland.

C. Anomalous chirping songs

These occur in the song repertoires of four of the described Drymopsalta species, the exception being D. crepitum (although this species does exhibit variability of song phases; Ewart, 2005). Anomalous songs differ from the normal chirping songs in two distinct ways: (a) The absence of single ticks between the chirps, and, (b) the chirps vary considerably in duration and the therefore in the number of constituent ticks. Therefore, the chirp lengths vary quite erratically, as illustrated in Figs 11 A, 14C–D, 15B). In certain respects, the discontinuous echemes of the D. daemeli buzzing songs (Fig. 18 C) are a more regular development of the anomalous songs, with longer and more consistent echeme durations. The inter-echeme durations are also more regular. However, in the D. daemeli songs, the anomalous chirping songs are also observed, comparable to those of the other Drymopsalta species. The detailed structures of the anomalous chirps are identical to those of the chirps within the normal songs, and pulse repetition rates in the ‘anomalous’ chirps are comparable to those of the normal songs, although in some cases slightly lower. Dominant frequencies, as seen in the amplitude spectra are also similar (e.g. Fig. 12 C).

The reasons for the emission of the anomalous songs are not understood. Individual cicadas can often be heard emitting the anomalous song with surrounding cicadas singing the normal chirping songs. Anomalous songs are commonly associated with buzzing echemes and sometimes occur within the normal songs. It is noted, however, that although D. wallumi does not emit extended buzzing echemes, they do sporadically emit anomalous songs,