Cyphoderris monstrosa Uhler, 1864

Males of all three species produce a succession of short musical trills, beginning in late evening and continuing well past midnight if weather permits. C. buckelli invariably sing near the ground from low shrubs (knee-height), the bases of tree trunks or on the forest floor itself. The same is true of C. strepitans. Only C. monstrosa climb high into the trees as the night's signalling progresses. At Monck Park singing heights in excess of 5 rn were common and an hour after sunset collection without climbing trees becomes impossible.

The calling songs are generated by tegrninal stridulation. As in Gryllidae the tegmina are morphological mirror-images, both left and right bearing a functional file and scraper. Unlike gryllids however, which maintain a characteristic 'right above' forewing overlap, the overlap of a Cyphoderris male may change during his lifetime and both files take part in his stridulation.

Certain Tettigoniidae also have mirror-image tegrnina and two functional files: Megatympanon speculaturn Piza (Listroscelinae) (Riek, 1976), Neduba macneilli Rentz & Birchirn, Neduba sierranus Rehn & Hebard (Decticinae) (Morris et al., 1975). Most tettigoniids have structurally distinct left and right forewings and overlap them 'left above'. In the Neduba species some individuals show left above, some right above. Unlike Cyphoderris they appear to maintain their particular overlap as individuals through life. Both overlaps were represented by Riek's two (pinned) specimens of M. speculaturn.

Spooner (1973) analysed the calling song of C. rnonstrosa and describes it as a trill of grylloid (sinusoidal) pulses at a carrier frequency of 13 kHz. He noted substantial variation in the intensity and frequency of pulses and suggested that these changes "reflect irregular switching of tegrnina from top to bottom position". He refers to this habit as "switch-wing singing" and regards it as occurring several to many times in the course of a single trill.

Overlap at rest (i.e. between singing bouts) is very infrequently changed in C. buckelli. The overlap of 16individually-caged males was monitored by examining them once a day during almost 2 weeks. Of 141 checks, only 4 reversals from the immediately previous overlap were observed; the incidence of resting overlap reversal was less than 3%. Thirteen of these males never showed an overlap reversal.

Four C. monstrosa males checked over 5 days, gave similar results: two were never found with reversed overlap (checked respectively 5 and 6 times), one was reversed once in 5 checks and one twice in 6 checks. If Cyphoderris alter overlap several times within a single trill, it is strange that individuals end up so consistently at the same overlap with which they began.

We recorded the calling song of a C. monstrosa specimen (Figure 6, 75-6) before and after damaging with a scalpel, several teeth in the central region of his right tegrnen file. In oscillograrns of post- mutilation recorded song, his use of the damaged file (i.e. right above overlap) was apparent as a drastic mid-pulse drop in arnpli- tude. In one oscillograrn, a portion of which makes up Figure 6 (second trace from bottom), 20 pulses in succession were 'right above'.

Switch-wing singing as suggested by distinctive pulse envelopes within the same trill was only evident in our records on one occasion. A male of C. monstrosa had been released in the immediate vicinity (i.e. within antenna1 range) of a mature female on the observer's hand. He began to sing while walking about on the hand and directing his attention toward the female. His song was recorded and on analysis found to be a trill in which every other pulse was identical in envelope and distinctly different from the intervening pulse i.e. there were two pulse types occurring in alternation without break in the sequence of the trill (Figure 6, bottom trace). This was apparently a courtship song.

It is clear that pulse envelopes are highly variable in the genus, though usually quite consistent for a particular recording session of a particular individual. Switch-wing stridulation is probably not an everyday feature of C. monstrosa calling song but it may occur under special circumstances such as courtship.

Oscillograrns of normal calling songs are given in Figure 6. The pulses of C. strepitans and C. buckelli are apparently indistinguish- able. They are usually wedge-shaped: each begins with a steep rise to maximum amplitude, then falls steadily to the pulse's end. The pulses of C. monstrosa also have a steep onset but are usually of longer duration. They are drawn out in an uneven envelope near their maximum amplitude before dropping away to silence.

Carrier frequency spectra of all three species are highly similar. Specimens were analysed 'live' (i.e. without tape-recording) by directing the output of a Bruel & Kjaer sound level meter (2204) fitted with a Gffmicrophone (4135) into a Tektronix 3L5 spectrum analyser. This system will detect ultrasonic frequencies up to 100 kHz. No substantial sound energy exists in the ultrasonic range for any of the Cyphoderris species. The sinusoidal nature of the waveform is apparent in the narrowness of the dominant frequency peak, suggesting the operation of a sharply-tuned (high Q) tegrninal resonator (Sales & Pye, 1974).

In the figured C. strepitans male (Figure 7), the dominant peak centers on 12.7 kHz and there are very weak second and third harmonics near 25 and 38 respectively. The C. buckelli specimen has its principal peak near 13.3 kHz and a lesser peak occupies the range 28-30 kHz. Like Spooner (1973) we obtained 13 kHz as the dominant carrier frequency of C. monstrosa.

Sound level measurements were obtained with the '/^microphone and the 2204 meter, the latter on 'linear, fast' setting. At 5 crn dorsal aspect, the sound level of C. strepitans (76-7) was between 100.5 and 101.0 dB. A specimen of C. buckelli (76-3) was 102k 2 dB at 6.5 crn dorsal.

Pulse rate varies linearly with temperature (Figure 8) as in other acoustic Ensifera (Walker 1962, 1975). Both field and laboratory recordings of calling song contributed to the regression lines. One C. monstrosa plotted point is from Spooner (1973) (S in Figure 8); 5 different males provide the other 6 points. C. buckelli's regression is based on 12 different individuals, 3 at two different temperatures each. Each of the 13 C. strepitans pulse rates derives from a different individual; all those at temperatures of 8OC and below are field recordings. Pulse rates were calculated from an oscilloscope display in which a single beam sweep embraced 3-13 pulses. Successive, single-sweep samples (3-6), were averaged to obtain each plotted value. The coefficients of determination indicate a very good fit to the calculated regression lines. Although the C. monstrosa regres- sion line is different from the strepitans and buckelli lines the slopes and Y intercepts of the latter two species are not significantly different.

C. strepitans males stridulate at very low temperatures. Previous reports cite minimal singing temperatures for acoustic Orthoptera of about 7OC [e.g. Fulton (1925) for the tree cricket Oecanthus fultoni (under the name of 0. niveus) and Frings and Frings (1957) for the katydid Neoconocephalus ensiger]. On May 17, 1977, at the holotypic site, one of us (D.T.G.) heard three of four males singing from branches and logs near the ground when the air temperature at waist level was -0.5' C. On June 4 and 5, 1978, tape recordings were made of males singing at temperatures as low as 2OC (see Figure 8). Following the recording the thermometer bulb was placed close to the singing male's perch. There is a suggestion in the plotted rates in Figure 8 of a departure from linearity at very low temperatures.

In conclusion, the song of C. monstrosa differs from the other two species in both the shape of the pulse amplitude envelope and in pulse rate, both these parameters being useful diagnostic features. C. buckelli and C. strepitans, however, have virtually identical calling songs: song intensities, carrier frequencies, and pulse arnpli- tude envelopes provide no basis for human discrimination; the pulse rates, especially, are indistinguishable at any given temperature.

It is interesting to note that Alexander (1969) has questioned the traditional interpretation that reproductive isolating mechanisms evolved to prevent "mating mistakes" between species. He suggested (citing evidence from acoustical insects) that species differences have most likely arisen as a result of the different selection pressures operating on populations while they are in allopatry. He reasoned that if this is so, among other things, we should rarely find identical pair forming signals among allopatric or allochronic species. C. strepitans and C. buckelli are allopatric (Figure 5) and by the above reasoning their songs should have diverged yet this is not the case. Any difference between these two species (including the above mentioned habitat differences) apparently have not affected their pair forming signals. [1]

Referenzen