Bioacoustics of Eneoptera nigripedis

Description. see Robillard & Desutter-Grandcolas (2005) . Calling song. ( Figs 4. 10 , Table 1 ) Only one recording is available on iNaturalist that could be accounted to this species. Based on the morphology of the stridulatory file showing a double teeth section, the species’ call likely shares most of the particularities of the call of E. guyanensis : it consists of a continuous trill including sections made of two types of syllables alternating regularly. The syllables in these sections have different spectral properties, durations and amplitude. The alternating sections are also coupled with a strong amplitude modulation, likely generated by a complex calling behavior as in E. guyanensis ( Robillard & Desutter-Grandcolas, 2011 ; Robillard et al. , 2015 ). The comparison of the call of the two species ( Table 1 ) reveals that the song of E. nigripedis has a slower pace of alternating sections: the group of low frequency (LF) and high frequency (HF) sections is almost twice longer than in E. guyanensis . This difference in duration is mainly due to a longer LF section (2973 ms vs. 588 ms in E. guyanensis ). The LF section is likely produced by the simple stridulatory teeth; it includes a larger number of syllables in E. nigripedis (81 syllables vs. 39 in E. guyanensis ), which are also slightly longer in duration as the syllables in E. surinamensis (4.8 vs. 3.6 ms), with similar low dominant frequencies (3.6 vs 3.3 kHz). However, the sampling rate of the E. nigripedis recording is insufficient to determine whether its low-frequency (LF) syllables also exhibit a dominant high-frequency peak, as observed in E. guyanensis . The HF section of the call of E. nigripedis is slightly longer than in E. guyanensis (c. 248 vs. 200 ms), and involves more syllables (29 vs 20 syllables in E. guyanensis ) with longer durations (4.9 vs. 2.7 ms). Contrary to E. guyanensis , the frequency spectrum of the HF syllable of E. nigripedis is not characterized by a broadband frequency, but by two, almost codominant frequency peaks, at c. 3 kHz. It may correspond to the fundamental frequency of the HF syllables, similar to that of the LF syllables. The second frequency peak, slightly dominating the spectrum, is at 17.5 kHz, but given the low sampling rate of the recording, this peak is likely under-sampled and may correspond to a more clearly dominant frequency; it may correspond to a high fundamental frequency generated by the crests of the double teeth, as in E. guyanensis , or to an upper harmonic of the 3-kHz peak. Note that, contrary to E. guyanensis , the LF sections of the call show less amplitude than the HF sections, which is also likely an artifact of the low sampling rate. Sound-production mechanism. ( Fig. 11D , Table 2 ) The stridulatory file of E. nigripedis is divided into two well differentiated sections, as in E. guyanensis . The anal section includes double-crest teeth, each crest likely behaving as a separate tooth during stridulation. The basal file section includes 40 simple teeth. The anal section shows 54 double-crests teeth, each tooth of the anal section being subdivided by a longitudinal furrow over its whole length, which separates one tooth into two parallel crests. The consequence is that the anal region consists of a set of equidistant crests likely acting as secondary teeth, which doubles the number of stridulatory crests implied during the stridulation process. We measured and plotted teeth/crests numbers and distances ( Fig. 11D ). The mean inter-tooth distance for simple teeth is 21.07±3.5 μm; the mean inter-crest distance for the double teeth is 8.30±1.4 μm. The functional parts for each region of the file are estimated as follows: the basal part with normal teeth is 0.42 μm long, while the anal part with double teeth is 0.6405 μm long. The global speed of stridulation derived from the syllable duration and file length is c. 87.7 mm /s for the simple teeth section, which matches the tooth strike rate that generates the fundamental peak of the spectrum of the LF syllables. For the double tooth section, the estimated global speed is 130.7 mm /s. It matches the instantaneous speed considering the HF peak as the fundamental frequency of the HF section of the call. It suggests that the low frequency observed in the HF section would correspond to a ghost frequency or to a subproduct of the tooth impact, while the dominant peak would result from the crest impact.