ABSTRACT
The majority of male cicadas do appear to have a similar anatomy and similar relative body dimensions, even though their body lengths range from below 15 mm to over 50 mm. We may refer to these as ‘typical’ cicadas, following Young (1990). There are a few notable exceptions to this common design, such as the bladder cicada Cystosoma saundersii, which has a distended thin-walled abdomen (Simmons and Young, 1978).
Bennet-Clark and Young (1992) found that the dimensions of the abdominal cavity and of the tympana of three species of cicada, Cyclochila australasiae, Macrotristria angularis and Magicicada cassini, when applied to equation 1, give resonant frequencies that agree well with the song frequencies measured for these species. However, the cicada Magicicada septendecim, which has anomalously thick tympana, did not fit the model well. We now report the relationship between body length and the dominant song frequency for a series of typical cicadas.
For several species studied by Young and Josephson (1983a), we used tape recordings and specimens from that earlier work. For Cicadetta quadricincta, we used recordings made by H.C.B.-C. on a Sony V90 camcorder using its own microphone. These records were analysed using a Kay DSP Sonagraph model 5500; the song frequency is taken as that at which the peak energy is seen. Body length was measured from dried preserved specimens of each species; these may be slightly shorter than live insects in the singing posture (Young, 1990) but, since the body lengths quoted by others have been measured in the same way, the various data may reasonably be regarded as comparable.
For the other species, data for carrier frequency and/or body dimensions have been obtained from the literature (Pierce, 1948; Pringle, 1954; Popov, 1989, 1991), supplemented by measurements of museum specimens. Some of the older published measurements of song carrier frequency were made by direct measurement from oscillograms and so they may be slightly less precise than those made using more recent methods of analysis.
The data for the dominant song frequency and body length for 16 cicada species are shown in Table 1. From these data, a plot of the reciprocal of body length against song frequency is shown in Fig. 1. The correlation coefficient of the calculated linear regression given on Fig. 1 is 0.875 (r2=0.766, 30 d.f.), which is highly significant (P<0.001). Such a correlation suggests that body size is acting as a constraint on the sound frequencies that are produced by typical cicadas. This is perfectly understandable if all, or nearly all, these species are employing a similar type of Helmholtz resonator as an acoustic load for their sound-generating tymbals. The dominant song frequency will then be an inevitable consequence of the dimensions of their abdominal cavities and tympana, as indicated in equations 1 and 2.
Another consequence of employing this common design of resonant structure in different species is that the system allows similar insect to air impedance matching at all sizes. The physical acoustics of sound production would suggest that, for good impedance matching, the linear dimensions of the sound-radiating structure should vary in direct proportion to the sound wavelength (see, for example, Olson, 1957; Seto, 1971). In cicadas which are known to radiate sound through the tympana (Cyclochila australasiae, Macrotristria angularis), this sound source is evidently large enough to provide good impedance matching between the insect and the surrounding air. Since it appears that the wavelength of the song of cicadas scales with body length, good matching should also be possible for smaller species. Certainly, cicadas are, for their size, extremely noisy.