Functional morphology of tegmina-based stridulation in the relict species Cyphoderris monstrosa (Orthoptera: Ensifera: Prophalangopsidae)

Within the orthopteran suborder Ensifera, males of most bush crickets (Tettigoniidae), field crickets (Gryllinae) and a few closely related species, generate acoustic signals through forewing stridulation, primarily to attract females. In all cases, at least one forewing bears a row of teeth, the file, which overlaps the other forewing, itself endowed with a scraper (or plectrum), located along the posterior margin (Pierce, 1948). The plectrum is scraped over the file in sequence, and the resulting vibrations cause particular wing cells (e.g. harp and mirror) to oscillate and radiate sound (Pierce, 1948; Broughton, 1964; Bailey, 1970; Sales and Pye, 1974; Bennet-Clark, 2003). This elaborate system, known from the Triassic period (ca. 220 million years ago; Béthoux, 2012), has been subject to tremendous evolution, resulting in a wide array of wing morphologies, body sizes, and behaviour. In some cases forewing-based sound radiation is enhanced by particular shapes of the pronotum (Morris and Mason, 1995), holes in plant leaves acting as acoustic baffles (Prozesky-Schulze et al., 1975; Forrest, 1991), particular tegminal inflations (Hemp, 2001; Montealegre-Z and Mason, 2005), or burrows purposely shaped for enhancing sound radiation (Bailey et al., 2001; Forrest, 1991).

Field and bush crickets represent two main lines of evolution. Field crickets (subfamily: Gryllinae; ∼1100 species according to Eades et al., 2016), produce low-frequency calls (∼3–8 kHz) (Walker, 1973; Hoy et al., 1982) from visually symmetrical (but functionally asymmetrical) wings with multiple, well delimited, radiating cells, the largest of which is the harp. The harp is the primary radiator for the fundamental components of the call, while a secondary area, the mirror, also contributes to sound production for the high-frequency components, mostly obvious in the courtship calls (Bennet-clark, 2003; Montealegre-Z et al., 2011a). The resonance of each of these cells is not affected by manipulation of the physical properties of the other membranous areas (Bennet-Clark, 2003). Field crickets predominantly have a ‘right-wing-on-top’ arrangement (Masaki et al., 1987), which is required for stridulation resulting in tonal purity in the produced signals (Elliott and Koch, 1983). The carrier frequency of these calls is dictated by an escapement mechanism (Elliott and Koch, 1985; Koch et al., 1988; Prestwich et al., 2000) and thus reliant on the natural frequency of the harps. Furthermore, coherent vibration of both wings at this resonant frequency is facilitated by a phase-shifting mechanism, which allows oscillatory synchrony between both tegmina (Montealegre-Z et al., 2009).

Most noticeable about most modern bush crickets (family: Tettigoniidae, ∼7000 species according to Eades et al., 2016) is the bilateral asymmetry between the forewings (Dumortier, 1963; Montealegre-Z and Postles, 2010; Gu et al., 2012; Chivers et al., 2014; Sarria-S et al., 2016; Cole and Chiang, 2016). The major radiating wing cell is named the mirror and is found primarily on the right wing, with the functioning file being on the left wing. The mirror is very well delimited, generally composed of the cross-vein free mirror (purple in Fig. 1D), sometimes in association with a second cell (light green in Fig. 1D) (Bailey and Broughton, 1970; Montealegre-Z and Postles, 2010). In many cases, there exists a mirror on the left wing, and a file on the right wing, but the former is usually partially damped and filled with cross-veins, and the latter greatly atrophied (Montealegre-Z and Mason, 2005; Montealegre-Z and Postles, 2010; Chamorro-Rengifo et al., 2014). These morphological wing characteristics result in an obligatory wing arrangement of ‘left-over-right’ for stridulatory sound production. Specific to bush crickets is the widespread use of ultrasonic frequencies in the calls; indeed ∼70% of species use ultrasounds in the range of 20–150 kHz (Morris et al., 1994; Montealegre-Z, 2009; Sarria-S et al., 2014), with some exceptions singing in the sonic range as low as 600 Hz (Heller, 1995). It has been suggested that the production of tonal calls at high ultrasonic frequencies is facilitated by the single radiating structure in this group, whereby there is no need to synchronize radiating cells on two wings for coherent sound generation (Montealegre-Z et al., 2009; Gu et al., 2012).

In summary, both lineages exhibit particular modifications with respect to the assumed ground plan. Only a handful of extant species of grigs (family: Prophalangopsidae, eight extant species according to Eades et al., 2016) seem to have retained a more plesiomorphic wing structure (Gorochov, 2003; Bethoux, 2012). Among these species, Cyphoderris spp. are the most easily accessible, and comparatively well documented. Recent molecular-based phylogenies of the Orthoptera concluded that Cyphoderris spp. are more closely related to bush crickets than to field crickets (Zhou et al., 2014; Song et al., 2015; Cole and Chiang, 2016). However, similar to field crickets, tegminal stridulation in Cyphoderris spp. occurs from largely symmetrical forewings (Morris et al., 2002). In addition, unlike both field and bush crickets, stridulation in Cyphoderris spp. can be performed with either wing on top, and also the wings lack some of the well-delimited wing cells, largely free of cross-veins (Morris and Gwynne, 1978; Fig. 1A,B). Furthermore, Cyphoderris spp. sing at ∼12–15 kHz (Morris and Gwynne, 1978; Morris et al., 2002), a range higher than that in field crickets (3–8 kHz; Hoy et al., 1982; Otte, 1992) but lower than that in most bush crickets (>20 kHz; Montealegre-Z, 2009). In summary, Cyphoderris spp. are of major importance to investigate the early evolutionary stages of a critical innovation, at the root of a tremendous diversity. However, little is known about wing function in Cyphoderris spp., and a detailed knowledge of stridulatory mechanics in this group will allow a better understanding of the forewing biophysics and communication capabilities in both extant relatives and related fossil taxa.

Prior to functional analysis we reconsider conjectures of topographic homology (THC) in the wing venation of grigs and bush crickets. This is the primary step to determine which areas became specialized as sound radiators. Then we investigate the functionality of the sound generators in the extant prophalangopsid Cyphoderris monstrosa Uhler 1864. Using biomechanical evidence and advanced experimental analysis, this study aims to characterize the radiating cells, and associated veins, on the forewings. This will test the hypothesis that the mirror in C. monstrosa is the primary oscillating membrane and contributes to dictating the carrier frequency of the male call. Vibratory function of the wings of C. monstrosa will then be considered in the context of the evolution of the stridulatory structures between groups of ensiferans.

The debate on the insect wing venation ground plan and/or the actual topographic homology conjecture (THC) to be applied to Orthoptera (Gorochov, 2005; Béthoux, 2007; Rasnitsyn, 2007; Béthoux, 2008) is only remotely related to the present contribution. Any ground plan (i.e. ‘serial’ or ‘M5’) can be applied to address THC aspects within the Ensifera (Béthoux, 2012). We favour the serial ground plan for reasons stated earlier (Béthoux, 2008).

The terminology and colour-coding already developed and used for Grylloptera (a formal taxon encompassing ensiferans possessing a file) in Béthoux (2012) is followed, with some modifications. General wing venation terminology is repeated for convenience, as follows: R, radius; RP, posterior branch of radius; M, media; Cu, cubitus; CuA, anterior cubitus; CuP, posterior cubitus; CuPa, anterior branch of CuP; CuPaα, anterior branch of CuPa; CuPaβ, posterior branch of CuPa; CuPb, posterior branch of CuP. Béthoux (2012) identified the ‘column’ (as in grigs and field crickets) in bush crickets. Herein we argue that the corresponding structure in bush crickets is not homologous to the column; therefore we propose to refer to it using a new term, ‘string’. This also has consequences regarding the harp area (turquoise in Fig. 1; brown in Béthoux, 2012), which can then be divided into three parts (Fig. 1B): the anterior portion (h1) is posteriorly delimited by the string (or its approximate position if it does not occur, as in field crickets); the median portion (h2) is delimited anteriorly by the string and posteriorly by the column (or its approximate position if it does not occur, as in bush crickets); and the posterior portion (h3) is anteriorly delimited by the column (or its approximate position if it does not occur, as in bush crickets). The mirror is closed by a specialized (set of) cross-vein(s) that we propose to refer to as the ‘frame’. Finally, we propose to complete the terminology of particular areas as follows: the area in grey in Fig. 1B (same colour coding as in Béthoux, 2012) is termed the ‘pre-mirror’; the area in light blue (Fig. 1B,D) is termed the ‘post-mirror’ (by definition it is delimited basally by the handle and distally by CuPaα2; therefore it does not occur in field crickets, in which CuPaα2 is partially fused with the handle; Fig. 1G,H); the area in orange [Fig. 1B,D,H; delimited anteriorly by (M+)CuA and posteriorly by CuPa/CuPaα] is termed the ‘neck’ (cf. terminology of harp pieces). An area of putative importance, as far as function is concerned, is anteriorly delimited by R/RP, posteriorly by M/MA, and distally either by MA or a cross-vein connecting RP and MA. This area will be referred to as the ‘Larunda’ area (see Discussion).

To ease comparison we indicate in Fig. 1 the CuPaα–CuPaβ fork by a white arrow (delimited in black) and the CuPaα1–CuPaα2 fork by a black arrow (delimited in white); the course of CuPaβ is indicated by a blue arrow following the corresponding vein (also for CuPaα2 in Fig. 1E).

We examined forewings of specimens of the three known species of Cyphoderris (three males of C. monstrosa; three males of C. buckelli Hebard 1934; and three males of C. strepitans Morris and Gwynne, 1978). The forewing venations of the three species show no major differences, but a comparatively high level of intra-specific and intra-individual variability (see Discussion). This variability does not affect the validity of our interpretations. As for bush and field crickets, we selected a representative species (Tettigoniidae: Copiphora brevirostris and Gryllinae: Gryllus bimaculatus), and intact preparations, from a small sample of each group belonging to the private collection of one of the authors (O.B.; acronym IWC OB). Wing preparation follows the method of Béthoux and Wieland (2009; see also Béthoux, 2012), except for the specimen IWC OB 862 (Fig. 1C–F), mounted in ‘White Euparal’ (Asco Laboratories, Manchester, UK).

Photographs were taken using a Canon EOS 5D Mark III digital camera coupled to a Canon MP-E 65 mm macro lens. Resulting photographs were dusted off manually (stamp tool) and optimized using Adobe Photoshop CS6. Venation schemes, areas and associated labels were prepared using Adobe Illustrator CS6.

How an optimal set of THCs can be identified among competing ones has been already clarified by Béthoux (2012). In short, an optimal set of THCs is the one implying the lesser amount of transformation to explain the observed patterns. In other words, it maximizes correspondences, given the observed elements.

Male specimens of C. monstrosa were collected as adults from Paul Lake Provincial Park, outside Kamloops, British Columbia (50°45′15.6″N, 120°07′08.8″W). Specimens were transported to the University of Toronto Scarborough, and subsequently to the University of Lincoln, where they were maintained in individual containers at ∼9°C, with a 12 h:12 h light:dark cycle, and fed daily with fresh apple and water-soaked cotton.

The calls of the singing males were recorded using a wide-bandwidth response 1/8-inch microphone (Brüel & Kjaer, 4138-A-015, with pre-amplifier model 2670, Brüel & Kjaer, Nærum, Denmark), and Polytec software (PSV 9.2, Waldbronn, Germany). Recordings were made at a sampling frequency of >256,000 samples s⁻¹. All experiments were performed in a sound attenuating chamber, on an anti-vibration table, at temperatures of 25.5±1.4°C.

This method of micro-scanning Doppler vibrometry follows previous works (Montealegre-Z et al., 2009, 2011a; Montealegre-Z and Postles, 2010; Sarria-S et al., 2016). Specimens were immobilized by exposure to a triethylamine-based mix (FlyNap, Carolina Biological Supply Company, Burlington, NC, USA) for 3–5 min. The specimens were placed on a block of BluTack (Bostik, La Défense, Paris, France) and their legs were gently clamped to the block with small staple clamps. The surface of the block was flat except for the front end where the surface angled downwards. The specimen was positioned so that the head and thorax of the animal were on the angled surface, with a clamp over the pronotum to maintain the positioning of the specimen. This position forces the prothorax to bend downwards, allowing free manipulation of the forewings. The forewings were then separated from each other in a raised position and fixed with a mix of beeswax (Fisher Scientific, Loughborough, UK) and Colophonium (Sigma-Aldrich, Dorset, UK) (50:50). The BluTack block was affixed to a brass plate which itself was attached to an articulated aluminium rod allowing the specimen to be manoeuvred into the required position.

Vibration-compliant areas of grig forewings, and associated frequency characteristics, were measured using a micro-scanning laser Doppler vibrometer (Polytec PSV-500; Waldbronn, Germany) fitted with a close-up attachment. The mounted specimens were positioned so that the extended wings were perpendicular to the lens of the laser unit. A loudspeaker was positioned above the laser unit and facing the animal to broadcast the sound stimulus. The acoustic stimulus used was periodic chirps, generated by the Polytec software (PSV 9.2), passed to an amplifier (A-400, Pioneer, Kawasaki, Japan), and sent to the loudspeaker (Ultrasonic Dynamic Speaker Vifa, Avisoft Bioacoustics, Glienicke, Germany). The periodic chirps spanned frequencies between 2 and 50 kHz, and the stimulus was flattened so all frequencies were represented at 60±1.5 dB (SPL re. 20 µPa) at the position of the wings. A Brüel & Kjaer 1/8-inch condenser microphone was placed at the position of the wings to monitor and record the acoustic stimulus at the position of the wings as a reference. The laser system was used in scan mode. A scan of the entire extended wings was performed using 670–1300 scan points. Within the frequency domain setting of the vibrometer, a frequency spectrum was calculated for each point using a fast Fourier transform (FFT) with a rectangular window, at a sampling rate of 128 kHz, 128 ms sampling time, and with a frequency resolution of 7.8125 Hz. A high-pass filter of 1 kHz was applied to both the vibrometer and reference microphone signals during the scanning process, with an average of three samples taken at each point. The data of one specimen are included in which each forewing was dissected in turn at the attachment by cutting the pteralia of the wing base, immediately sealing the cut with wax, and affixing it to a short length of wire by the wax seal. The wire was attached to a clamp arm and clamp stand, damped to vibrations with BluTack, and the wing was scanned using the same protocol as before, within 20 min of dissection. To facilitate comparison of vibrational response between groups, a wing scan from the bush cricket Copiphora brevirostris and the field cricket Gryllus bimaculatus are presented (see Discussion). These scans were made with the same experimental set-up.

Vibrometer scan data were analysed using Polytec software (9.2). Frequency spectra of the vibrometry data were normalized to those of the reference signal by computing the transfer function of the two (Windmill et al., 2005). The magnitude-squared coherence between the vibrometer and microphone signals was also computed for each data point to estimate the amount of unrelated noise (Windmill et al., 2007). Coherence values can range between zero and one, with a value of one indicating the absence of unrelated and external noise. The parameter Q is a dimensionless index indicating the sharpness of tuning in resonant systems (Bennet-Clark, 1999b). Here, for the acoustic data, Q was calculated as the peak frequency divided by the bandwidth at 3 dB below the peak amplitude. Q was also calculated from the frequency spectra of the vibrometry data by dividing the peak frequency by the bandwidth at values equalling 0.707 times the peak amplitude, corresponding to 3 dB below peak (Fletcher, 1992). Data were tested for normality, and peak frequencies and magnitude of vibrations between left wing and right wing were compared with a paired sample t-test. Q-values of the wing resonances were compared with a Wilcoxon signed rank test. Analysis of vibration data, acoustic data and statistical analysis was performed in MATLAB (R2015b, The MathWorks Inc., Natick, MA, USA) and SPSS (version 21, IBM Corporation, Armonk, NY, USA).

Here, we present new THCs for the wing venation pattern, and the associated areas, for Cyphoderris spp. (Fig. 1A,B) and bush crickets (Fig. 1C,D). To ease comparison with previous accounts, we also provide the challenged THC for bush crickets (Fig. 1E,F), and one we consider valid for field crickets (Fig. 1G,H; according to Béthoux, 2012). The most recent account on the topic (Béthoux, 2012) assumed a partial fusion of CuPaα2 and CuPaβ in bush crickets (Fig. 1E). Both veins were assumed to run along the handle for some distance, with CuPaα2 diverging first. As a consequence, the mirror cell was found to be bordered distally by CuPaα2 and basally by CuPaβ, as in field crickets. Moreover, the ‘column’, a particular cross-vein bridging CuPaβ and CuPb, was identified in both field and bush crickets (Fig. 1).

The emergent consensus on the position of Cyphoderris spp. as more closely related to bush crickets than to field crickets (Desutter-Grandcolas, 2003; Zhou et al., 2014; Song et al., 2015), as well as the data from this study, prompted us to revise THCs proposed earlier. Indeed, in Cyphoderris spp., the mirror cell is not bordered distally by CuPaα2, but instead by a series of curved cross-veins (‘frame’ in Fig. 1A). Moreover, although there is not a well-individualized column or string in these species, the positions of the existing cross-veins show that the ‘premises’ of both co-occur. One can then legitimately challenge the homology of the ‘column’ as identified by Béthoux (2012) for bush crickets (Fig. 1E) and field crickets (Fig. 1G).

The THC we propose herein for Cyphoderris spp. is not essentially different from that proposed by Béthoux (2012, plate 1E,F): the pre-mirror (grey in Fig. 1B) is large, as in other Prophalangopsidae and as in stem-Grylloptera. As for bush crickets, we propose that: (1) CuPaα splits (into CuPaα1 and CuPaα2) in a more distal position than assumed by Béthoux (2012); (2) the mirror is bordered distally and posteriorly by a specialized cross-vein (‘frame’ in Fig. 1C) homologous to the set of unspecialized cross-veins in Cyphoderris spp.; and (3) the cross-vein referred to as ‘column’ by Béthoux (2012; herein ‘string’, Fig. 1C) is not homologous to the ‘column’ of field crickets, but instead to the set of weakly specialized cross-veins bridging CuPaβ and CuPb opposite the handle in Cyphoderris spp. The bush cricket ‘harp’ according to Béthoux (2012; Fig. 1F) represents the anterior portion of the (partly differentiated) harp in Cyphoderris spp. only. Under the favoured THC presented herein, in bush crickets, CuPaβ runs backwards for some distance (Fig. 1C), along the handle, an assumption proposed earlier (Béthoux, 2012).

The new THC for bush crickets supposes fewer transformations than that proposed by Béthoux (2012). Notably, a CuPaα1–CuPaα2 fork located in a distal position, close to the fusion of CuA and CuPaα, maximizes correspondences with the THC for Cyphoderris spp. and stem-Grylloptera. In addition, the new THC handles the fact that both ‘string’ and ‘column’ co-occur in Cyphoderris spp., and therefore cannot be homologous.

Acoustic analysis agrees with previously reported results for the calling song of C. monstrosa (Fig. 2) (Morris and Gwynne, 1978; Spooner, 1973; Morris et al., 2002). Mean peak frequency of the call is 13.08±0.1 kHz (N=5). These calls are highly resonant, with the mean Q of the calling song being 57.3±18.39 (Fig. 2). Scanning laser vibrometry experiments to reveal the vibration-compliant areas of the forewings of C. monstrosa were successfully achieved in five specimens, and we hereby present an initial treatment of wing mechanics in this species. Vibrations in response to sound stimulus (at a band around the calling frequency) are limited to the mirror and also the areas herein termed the neck, the pre-mirror and the anterior portion of the harp (Fig. 3). Vibrations of the mirror and adjacent area occur in phase (‘as-one’ vibrations; Fig. 4), in a basic mode corresponding to the dominant resonant frequency. Although the mirror and adjacent areas are weakly delimited, the vibrating surface is confined within the surrounding veins. This vibration pattern was observed in both wings of all specimens scanned (N=5) and further symmetry of wing function is exhibited. Peak vibration amplitudes are observed on the mirror area (Fig. 4) with the average vibration amplitude of the mirror (as calculated by averaging all scan points on the mirror in displacement) being 182.4±77.5 nm Pa⁻¹ for the left wing and 138.4±52.8 nm Pa⁻¹ for the right wing, and this difference was not significant (paired t-test, t=0.874, d.f.=4, P=0.432). Mean resonant frequency of the mirror was 14.2±1.08 kHz (N=5) for the right wing and 14.6±2.2 kHz (N=5) for the left wing (Fig. 3), and this difference was also not significant (paired t-test, t=0.618, d.f.=4, P=0.569). High coherence of vibration of the mirror shows that the observed response is reliable, with coherence approaching one around the frequencies of resonance (Fig. 3). Thus the mirrors exhibit a natural tuning at frequencies close to the calling song (Fig. 5) and this matching suggests the natural frequency of the wings is functionally tuned to a specific frequency, which is in turn being exploited for efficient, resonant sound production. The areas adjacent to the mirror (pre-mirror, neck and h1) vibrate at a lower amplitude than the mirror, and exhibit no sharp resonance (Fig. 6). The Q of the mirror's resonance, measured from the peak on the spectra, is much lower than that of the calling song, with Q of the mirror area being 27.8±24.8 for left wing and 12.7±3.7 for the right wing (this difference being not significant: Wilcoxon, Z=−0.674, P=0.5).

Multiple conjectures on the evolution of stridulatory apparatus exist in the literature (Gwynne, 1995; Desutter-Grandcolas, 1997, 2003; Jost and Shaw, 2006; Bethoux, 2012 and references therein) with unequivocal homologies between groups of orthopterans remaining elusive. This work presents a revised THC of wing venation patterns within the Orthoptera, building on previous efforts (Béthoux, 2012). The new THC is considered superior to that proposed by Béthoux (2012) because it requires fewer transformations to explain the observed patterns. An important implication of the new THC is that, in both Cyphoderris spp. and bush crickets, there is a large area composed of the neck, h1 and the pre-mirror that is essentially cross-vein free. Together with the mirror, laser micro-scanning vibrometry has revealed these cells as the main areas functioning for resonant sound production in Cyphoderris monstrosa. The reduction of the pre-mirror, complete in bush crickets thanks to the migration of the first fork of CuPa (white arrow in Fig. 1C) in a distal position, results in an optimal reduction of the partitioning of the corresponding area. Moreover, it is noticeable that CuPa is weakened in both taxa, a point indicative of the vibrational compliance of the corresponding area (a point demonstrated by the functional comparison; Fig. 7A,B). Together with the mirror this large area acts as a single vibrating structure for sound production in both groups (Bailey, 1970; Montealegre-Z and Postles, 2010; Sarria-S et al., 2016).

This situation contrasts with that observed in the field cricket lineage (Fig. 7C). Although the pre-mirror was reduced or lost in this lineage (Béthoux, 2012), and the mirror is maintained, the main radiating structure is composed of an area encompassing h1 and h2 (Bennet-Clark, 2003; Montealegre-Z et al., 2011a; Robillard et al., 2013), which have a minor or no contribution in sound production in Cyphoderris spp. (Fig. 7A).

The wings of C. monstrosa have a natural tuning close to the frequency of the call at ca. 13 kHz, a requirement for resonant sound production, and a feature often reported in resonantly calling crickets and bush crickets (Koch et al., 1988; Bennet-Clark 1999a; Montealegre-Z and Postles, 2010; Montealegre-Z et al., 2011a). A mismatch of ∼2 kHz between wing resonant frequency and call frequency is apparent in two specimens. This difference is likely to be a result of changes in the physical properties of the wings, the oscillating areas, and associated delimiting veins during stridulation when the wings are actively engaged and in motion (Montealegre-Z et al., 2011a,b). A similar mismatch is previously reported in bush crickets (Montealegre-Z and Postles, 2010; Sarria-S et al., 2016). Vibrating areas and natural tuning at the calling frequency is apparent in both wings of C. monstrosa. High Q-values as seen in C. monstrosa forewing vibrational response are similar to those previously reported for the wing vibrations of other ensiferans exploiting resonant sound production (Bennet-Clark, 2003; Montealegre-Z and Postles, 2010).

The mirror alone is responsible for dictating the main spectral energy of the call, with the adjacent neck, pre-mirror and h1 having no intrinsic tuning. These adjacent areas probably form a lightly damped area of wing, oscillating in-phase with the resonant frequency of the nearby mirror during sound production. The innovation of a lightly damped sounding board, which indiscriminately amplifies any input frequencies, has been reported in bush crickets (Morris and Pipher, 1967; Bailey, 1970) as well as in-phase vibrations of the mirror and adjacent area (Montealegre-Z and Postles, 2010; Sarria-S et al., 2016). In C. monstrosa, the areas of vibration are minimally delimited, with narrow veins impinging on the area of the mirror, and the bounding veins being non-specialized (i.e. not thicker than other veins) (Fig. 1). This study demonstrates that sharply tuned structures for resonant sound production can exist outside the maxims of derived wing morphology exhibited in both modern field and bush crickets. Given this non-specialization of vibrating areas, the question of how such resonant tuning is achieved, through the morphology of the wing, presents itself. The mirror of bush crickets is usually bounded by a thick frame (Morris and Pipher, 1967; Bailey, 1970; Montealegre-Z and Postles, 2010; Bethoux, 2012; Chivers et al., 2014; Sarria-S et al., 2016) (Fig. 1C). This frame is reported to act as a cantilever during oscillation, being clamped either along the file bearing vein (Morris and Pipher, 1967; Bailey, 1970) or at the region of the plectrum (Montealegre-Z and Postles, 2010; Sarria-S et al., 2016). Properties of length, mass, stiffness, as well as membrane structure and attachment are implied as contributing to the resultant resonant frequency of this system (Morris and Pipher, 1967; Bailey, 1970; Keuper et al., 1988; Bennet-Clark, 2003). In C. monstrosa, under the cantilever model, the frame of the mirror is clamped along the distal region (where CuPaβ meets the ‘frame’; Fig. 1). In bush crickets, clamping occurs by the thickness of the framing vein at a single region (Bailey, 1970; Montealegre-Z and Postles, 2010; Sarria-S et al., 2016). However, the mirror frame in C. monstrosa is formed by a vein of standard thickness (Fig. 1A). Clamping of this frame can be attributed to its attachment to multiple small cells of the distal region of the wing and, possibly, a depressed area sub-parallel to the frame and reaching the posterior wing margin, here referred to as the ‘gorge’. Areas of wing with such hexagonal structures are associated with increased stiffness (Montealegre-Z et al., 2009, 2011a), and indeed do not vibrate in response to acoustic stimuli as seen in this study (Fig. 3). This condition thus may represent a transitional stage between resonantly tuned wing cells on unspecialized forewings, and the highly modified systems seen in modern bush crickets.

One particular area was scrutinized for a putative functional role: being located at the widest part of the area bordered anteriorly by R/RP and posteriorly by M/MA (brown in Fig. 8A), herein named ‘Larunda’ (Larunda was a nymph whose tongue was cut out by Jupiter for her talkativeness). As far as morphology is concerned, the Larunda area is characterized by cross-veins, locally weakened (Fig. 8B–D). Importantly, variation of these cross-veins occurs, including cases where no weakening can be observed (Fig. 8E). It is tempting to interpret the Larunda as a relictual vibration-compliant area. Indeed, in several ancient Grylloptera the corresponding area is broadened, filled with few and/or narrow cross-veins and distally sealed by a secondary structure (Sharov, 1968, 1971; their fig. 29) (Fig. 8F). The morphological variation observed in Cyphoderris spp. could then be compared with that of the mirror on the non-functional forewings of bush crickets. However, our analyses revealed no particular vibrational compliance of the Larunda area in forewings of C. monstrosa. It is interesting to note that in certain Phaneropterine bush crickets, the corresponding area of the here-defined Larunda area is thin and transparent, thus being suggestive of playing a role in sound radiation (Heller et al., 2015). This area merits further investigation across groups.

The calls of C. monstrosa are highly resonant (Q of ca. 57), and worth noting is the indication that during stridulation both wings must be excited at their resonant frequency by the appropriate tooth strike rate, and that the vibrating areas of both wings must oscillate in phase (Bennet-Clark, 1999a; Montealegre-Z et al., 2011a). In field crickets, during stridulation, the different areas of both wings (mirror, harp) all oscillate coherently for resonant sound production (Montealegre et al., 2011a,b). The use of a single oscillating area in the grigs, and the derived extreme of a single and well-delimited mirror in bush crickets, would reduce the effect of destructive interference of multiple cells, and facilitate resonant sound production at higher frequencies (Montealegre-Z, 2005). Furthermore, coherent sound production from both wings has been shown to rely on a phase-shift mechanism, as seen in field crickets (Bennet-Clark, 2003; Montealegre-Z et al., 2009), due to the varying spatial position of energy input during each closing stroke. A similar phase-shifting mechanism has been observed on the right wing of certain bush crickets (Bailey, 1970; Montealegre-Z and Postles, 2010); however, the function of such a mechanism in a system in which there is no need to synchronize two wings is unknown. No such mechanism is apparent in C. monstrosa from our data (e.g. out-of-phase scraper and mirror vibration at the resonant frequency) and yet such a system is conceivably required for coherent oscillations of both wings during stridulation under an ‘either-wing-on-top’ arrangement. This question deserves further attention among both the grigs and the bush crickets, with focused investigations into the transmission, and relative phase of, vibrations across areas of each wing in relation to the position of energy input, i.e. at the scraper or along the length of the file (after Montealegre-Z et al., 2009).

Finally, this raises the question of how Cyphoderris spp. are regulating the frequency of energy input to excite the forewings at their natural frequency. Field crickets, which largely produce resonant calls, achieve this by an escapement mechanism of the file and scraper, with the frequency of energy input regulated by the natural resonant frequency of the wings (Elliott and Koch, 1985; Koch et al., 1988; Prestwich et al., 2000; Bennet-Clark and Bailey, 2002). This escapement is associated with the above-mentioned phase-shift mechanism for pure-tone sound production (Bennet-Clark, 2003; Montealegre-Z et al., 2009). In contrast to the escapement system of frequency regulation, excitation of the natural frequency of the wings in bush crickets relies on an association between wing closure velocity and arrangement of teeth on the file (Montealegre-Z et al., 2006). In this way a scraper passed over consistently distanced teeth at a stable velocity produces a consistent tooth strike rate (Montealegre-Z et al., 2011b). Alternatively, an increase of tooth spacing can be associated with an associated increase in wing closure velocity, thus keeping the TSR consistent during each wing closure (Montealegre-Z and Mason, 2005). The lack of a phase-shift mechanism in the forewings of Cypoderris spp. suggests that frequency regulation is similar to that of bush crickets, possibly due to the requirement for tonal signals at higher frequencies (Montealegre-Z et al., 2011a,b). A proper treatment of the mechanism of frequency regulation during stridulation in grigs is beyond the scope of this study, but certainly deserves further attention.

This study highlights the importance of considering the relationship between the functional and evolutionary aspects of morphology. Consideration of just one of these aspects, while being robust in the chosen field, may lead to unreliable interpretation of data (e.g. Ower et al., 2016). Some attempts have been made to infer basic singing parameters from Jurassic cricket species (Gu et al., 2012), but inferences are currently limited by a lack of knowledge on the characteristics of weakly delimited vibration-compliant areas. Our comparative analysis demonstrates that radiating structures observed in extant groups relate to a single ‘ancestral’ pattern. The mirror area is the morphological and functional feature shared by grigs, bush crickets and field crickets, although its contribution to sound production, in the latter, is more modest. Yet, as the field and bush cricket lineages strongly diverged, through similar trends such as cross-vein specialization, the progressive reduction of selected areas occurred, becoming less and less functional and hence more prone to reduction. However, it must be acknowledged that grigs, in which this reduction and specialization is nascent, are only partly relevant for inferences on the functional morphology of the earliest stem-Grylloptera. In the former, the radiating area is composed of several cells, but they already are relatively large and cross-vein-free areas compared with the same areas in the latter (Sharov, 1968, 1971) (Fig. 8F). Therefore attempts to infer basic singing parameters from Jurassic stem-Grylloptera (Gu et al., 2012) must be regarded as a preliminary effort, given the yet limited knowledge on the vibrational characteristics of areas with a higher density of uniform cross-veins. Other approaches, such as computer-assisted models, will allow a refinement of our understanding of the early stages of tegmina-based stridulation in Grylloptera. The current data on grigs will allow a validation of the models before application to fossil species.

We would like to thank G.K. Morris, J. Lapeyrie and N.W. Bailey for their generous donation of specimens to O.B. This work is part of a thesis partly supported by the School of Life Sciences, University of Lincoln, UK.