First song descriptions of some Anatolian species of Tettigoniidae Krauss, 1902 (Orthoptera, Ensifera)

Abstract Fourteen endemic and two sub-endemic species belonging to three subfamilies of Tettigoniidae (Tettigoniinae, Bradyporinae and Saginae) were sampled during field trips throughout the different ranges of Anatolia between the years of 2004 and 2013. Acoustic parameters of these 16 species affiliated to 8 genera (Anterastes, Apholidoptera, Gampsocleis, Parapholidoptera, Pezodrymadusa, Psorodonotus, Bradyporus and Saga) have been described for the first time in this study. Acoustical analysis showed that song characters are species-specific in the genera Saga and Psorodonotus. On the other hand, we could not find big differences among species of the genus Pezodrymadusa and Parapholidoptera castaneoviridis species-group.


Introduction
Orthoptera is one of the most well-known acoustically active insect orders (Heller 2006). The taxa of Tettigoniidae produce specific songs, which allow the recognition, location and selection of conspecific mating partners (e.g., Walker 1964, Heller 1988, Ewing 1989, Heller 1990, Robinson and Hall 2002, Heller 2006. Almost all species have a specific song structure, hence the useful and functional taxonomic character (Heller 2006) which allows the discrimination of morphologically similar species (Ingrisch 1991, Ragge and Reynolds 1998, Kolics et al. 2012). On the other hand, some genera of tettigoniids, such as Parapholidoptera (Heller 1988) and Eupholidoptera (Çıplak et al. 2009) exhibit, characteristic songs and uniform intrageneric song patterns.
The researchers of the last century were able to document many of the singing Orthoptera that are distributed in certain areas, such as North America (Walker and Moore 2004), Eastern United States (Alexander 1956) and Europe (Heller 1988, Ragge andReynolds 1998). The history of studies that include song analysis in Anatolian Orthoptera began with Stumpner and Helversen (1992) for Caelifera and Heller (1988) for Ensifera. Up to now, as many as 55 songs of endemic tettigoniids from Anatolia have been already described and these studies can be divided into three main categories: (i) single species song description (Çıplak and Heller 2001, Çıplak and Heller 2005, Sevgili et al. 2012a); (ii) song descriptions of species-group and/or groups in a genus (Heller 2004, Sevgili 2004, Heller and Sevgili 2005, Çıplak et al. 2009, Heller et al. 2008, Heller and Korsunovskaya 2009, Sevgili et al. 2010, Kaya et al. 2012, Chobanov et al. 2013; and (iii) songs of orthopteran species in a certain area (Sevgili et al. 2011). Moreover, a huge part of Anatolian tettigoniids has not been studied with regards to the song characteristics until now.
The family Tettigoniidae Krauss (1902) is the largest family of the Orthoptera and it displays species richness in Anatolia of about 360 taxa (Karabağ 1958(Karabağ , Çıplak et al. 2002(Karabağ , Ünal 2013a). More than 60% of Tettigoniidae taxa (e.g. Karabağ 1958Karabağ , Ünal 2002 recorded from Turkey are endemic to Anatolia (Çıplak et al. 1993(Çıplak et al. , Çıplak and Demirsoy 1995. A possible explanation for the richness of the Tettigoniidae species and its high endemism rate in Anatolia is that this region is one of the most important refugium in Palearctic (Hewitt 1996(Hewitt , Çıplak 2003(Hewitt , Şirin et al. 2010. However, the studies on the lineages represented in this peninsula are still far from explaining this phenomenon. In the present study, we aim (i) to obtain the first ever records of song characteristics of 14 endemic and 2 sub-endemic species belonging to 8 genera (Anterastes, Apholidoptera, Gampsocleis, Parapholidoptera, Pezodrymadusa, Psorodonotus, Bradyporus and Saga) from different parts of Anatolia and (ii) to understand the relation between the distribution and song diversity of the species under discussion.

Specimens collecting
In the present study, 16 species of 8 genera belonging to three different subfamilies of Tettigoniidae (Tettigoniinae, Bradyporinae and Saginae) were sampled during field trips throughout the different ranges of Anatolia between 2004 and 2013. Male calling songs were recorded in the field or in laboratory from live animals. Then, the recorded eral song type of the phrase is typical crescendo. The number of syllables within 100 ms is 29-40 (34.62 ± 2.21) (Figure 2d). Syllable duration varies between 2 and 5 ms (3.67 ± 0.34) with an interval of 0-1 ms (0.56 ± 0.06).
Description of song. Total of the six records from two males was analyzed. The calling song consists of several phrases in different duration ( Figure 4b). The phrases begin with thick pulse and continue with low intensity in the first part of the phrase. The following part of phrases consists of song elements with higher intensity ( Figure  4c). Phrases duration varies between 3.53-25.95 s (11.32 ± 5.53). Syllable duration varies between 33 and 40 ms (36.72 ± 1.28) with an interval of 0-3 ms (1.52 ± 0.09). Oscillographic analyses showed that each syllable contains different number of parts which are divided by the very short interval (lower than 2 ms). First and last part of a syllable generally consist of 2-4 shorter elements (each of 1 ms), while middle part consists of two longer elements (each of 6-8 ms).
Description of song. Total of the six records from two males was examined. The calling song consists of a series of regular phrases (Figure 5b) with an interval of 509-1259 ms (0.76 ± 0.10). Phrase durations vary between 219-346 ms (294 ± 0.04) and phrases consist of 4-6 (5.21 ± 0.54) syllables. The first and second syllables at the beginning of the phrase are quieter and shorter (having low amplitudes) than the following ones (Figure 5c). Syllable duration varies between 17 and 41 ms (30.74 ± 4.08) with an interval of 21-47 ms (28.52 ± 3.13). Oscillographic analyses showed that each syllable contains two parts. First part of syllables relatively short and consist of comprised song elements ( Figure 5d). First parts generally last 8-13 ms (11.71 ± 1.16) and are followed by second part after an interval of 0-5 ms (1.22 ± 0.48). The second syllable part includes several high amplitudes elements ( Figure 5d). These elements number is always 3-4 in first syllable and following respectively 9-10, 12-13, 13-15, 16-18 and 16-18 in last syllable. The second syllable part is much louder (except of the first syllable) and longer than the first part and duration varies between 9 and 27 ms (16.71 ± 3.76). Karabağ, 1961 http://species-id.net/wiki/Parapholidoptera_intermixta Distribution. Endemic for Turkey, Binboğa-Mountains ( Figure 6a) (Karabağ 1961(Karabağ , Çıplak 2000.
Description of song. Ten records from five males were examined. The calling song consists of a series of regular phrases ( Figure 6b) with an interval of 681-895 ms (810 ± 0.07). Phrase durations vary between 239-254 ms (246 ± 0.05) and phrases consist of 6-7 (6.12 ± 0.35) syllables. Syllables consisting of denser and hardly distinguishable impulses ( Figure 6d). The first and second syllables at the beginning of the phrase are quieter and shorter (having low amplitudes) than the following syllables (Figure 6c, d). Syllable duration varies between 29 and 60 ms (38.59 ± 5.78) without any silent interval [except between the first and second syllables (6-12 ms)].
Description of song. Total of five records from two males was examined. The calling song consists of polysyllabic sequences of variable duration (Figure 15b) with an interval of 5.80-8.02 s (mean 6.82). Sequences are consisting of 215-350 (262 ± 38.76) syllables. Nearly all syllables are in same amplitude ( Figure 15c); rarely syllables in begin or end point of sequences are in low amplitude. General syllables shape is a kind of crescendo (Figure 15d). Syllable period durations vary between 21 and 28 ms (25.68 ± 1.89). The number of syllables in 100 ms is approximately four (Figure 15d). Each syllable includes 14-23 impulses (16 ± 2.14).
Description of song. Total of six records from three males was examined. Oscillographic analyses showed that two possibilities (i) each syllable contains three similar elements (usually crescendo) and phrase consists of a great number of them or (ii) there are micro-phrases of three syllables in a crescending sequence and phrase consists of a great number of this micro-phrases ( Figure 16c). So, repetitive unit term was used for this situation to describe the song. The calling song consists of repetitive unit sequences of variable duration (Figure 16b). The phrase begins with 1-2 repetitive units that are quieter than the following ones. Phrase duration varies between 1.02 and 8.12 s (4.15 ± 1.29). The number of repetitive unit in 100 ms is approximately 2.5 and repetitive unit duration varies between 36 and 42 ms (39.11 ± 1.90) with an interval of 6-9 ms (6.72 ± 0.19). Each element includes 4-8 impulses and the duration of each element (Figure 16d) varies between 5 and 9 ms (7.16 ± 0.79).

Saga rhodiensis Salfi, 1929
http://species-id.net/wiki/Saga_rhodiensis Distribution. Anatolia and Rhodos (Figure 17a) (Kaltenbach 1967(Kaltenbach , 1970.  Description of song. Total of five records from two males was examined. The calling song consists of regular phrases (Figure 17b) with an interval of 3.25-4.50 s. The phrase begins with characteristic high amplitude syllable in the all phrases. After this syllable phrases continue with a quiet beginning and maximum intensity is usually reached between 1/4 and 1/3 of the phrase, however, in some of the phrases there is often a more gradual crescendo roughly up to half of the phrase (Figure 17c). Phrases contain 86-103 syllables (94.2 ± 7.02) and duration varies between 2.27 and 2.77 s (2.55 ± 2.54). Amplitude of the impulses of each syllable from beginning to end of it is getting louder (Figure 17d). Therefore general song shape shows crescendo type (Figure 17c, 17d). The number of syllables in 100 ms is 4-5 (4.18 ± 0.19). Syllable duration varies between 14 and 28 ms (20.92 ± 1.79) with an interval of 3-7 ms (5.62 ± 0.11). Oscillographic analyses showed that each syllable consists of a single element and includes easily countable impulses 16-22 (19.72 ± 1.87) in a crescending structure (Figure 17d).

Discussion
This study is the first one to reveal the descriptions of the acoustic parameters (amplitude-temporal pattern) of 14 endemic and two sub-endemic species in Anatolia. Also, this data could be used as an archive to determine the species in the field (Oliveira et al. 2001) which is important for species having local distribution in nature, among which the endemics studied herewith.
Remarks on song patterns: This part of our discussion focuses only on Parapholidoptera, Pezodrymadusa, Psorodonotus and Saga, because we posses sufficient amount of comparative data only on these genera. The four species of genus Pezodrymadusa show very similar song patterns, similar to the song pattern in genus Eupholidoptera (Heller 2006(Heller , Çıplak et al. 2009). Eupholidoptera is a well known genus and all the species in the genus have uniform song pattern, but a different morphology (Heller 2006(Heller , Çıplak et al. 2009). All four species, Pezodrymadusa konowi (Bolivar, 1899), Pezodrymadusa kurmana (Ramme, 1939), Pezodrymadusa lata Karabağ, 1961, andPezodrymadusa subinermis Karabağ, 1961, produce a multi-syllable song with syllable groups (Figures 8b,  9b, 10b, 11b). Within these four species, Pezodrymadusa subinermis has the lowest syllable number (Figures 11b-c), whereas Pezodrymadusa lata has the highest syllable number in a phrase (Figures 10b-c). Pezodrymadusa subinermis shows partially differences in the fine structure of syllables than the other species in this study.
Parapholidoptera is the second species-rich genus of the tribe Pholidopterini (including Pholidoptera, Eupholidoptera, Apholidoptera, Uvarovistia, Parapholidoptera, Exopholidoptera) in Anatolia (Çıplak 2000(Çıplak , Eades et al. 2013. Genus Parapholidoptera was studied morphologically by Çıplak (2000) and acoustically by Heller (2006). Songs of six species have already been described in the P. castaneoviridis and P. distincta groups (Heller 2006). Song records of three Parapholidoptera species are the members of the P. castaneoviridis group according to cladograms obtained based on the morphological data (Çıplak 2000). Heller (2006) indicated that only P. salmani presumably has a differentiated song pattern within P. castaneoviridis group. However, the results of this study show that the general song pattern of P. salmani is similar to the P. castaneoviridis group song pattern (Figures 7b-c). On the other hand, P. castaneoviridis group syllable pattern consists of coupled pulses; however, the syllables of P. salmani song consist of one continuous impulse series without any interval (Figure 7d).
Genus Psorodonotus has 11 species and eight of them are endemic/subendemic to Anatolia (especially north-east Anatolia) (Çıplak 2008(Çıplak , Ünal 2013a(Çıplak , Eades et al. 2013). The recorded song data in this study show interspecific differences. According to the song results in this study, Psorodonotus ebneri Karabağ, 1952 and Psorodonotus rugulosus Karabağ, 1952 exhibit a similar song pattern (Figures 13c-d, 14c-d). The song of Psorodonotus davisi Karabağ, 1956 shows different syllable composition ( Figure 12). However, the song pattern indicates a close relationship between Psorodonotus ebneri and Psorodonotus rugulosus, although they are not close geographically. On the other hand, P. davisi and P. rugulosus are located close to each other and far away from the P. ebneri.
The saw-legged bush-crickets are among the largest insect species in the Palaearctic. The range of most species of this genus covers the Balkan Peninsula and Asian Turkey (Kaltenbach 1967(Kaltenbach , 1970. Seven of these species are found in Asian Turkey (S. beieri, S. cappadocica, S. longicaudata, S. ephippigera, S. natoliae, S. puella and S. rhodiensis) (Karabağ 1958(Karabağ , Çıplak et al. 2002. Five European taxa (S. campbelli, S. gracilis, S. helenica, S. rammei and S. natoliae) were discussed in detail using songs characteristics (Kolics et al. 2008). The songs of Saga rhodiensis Salfi, 1929 and Saga cappadocica Werner, 1903 are described in this study. The song of S. rhodiensis shows a similar song pattern with these five species, but differs in length and impulse number of the syllables from them. On the other hand, S. cappadocica shows distinct syllable elements composition (Figures 16b-d, 17b-d). According to the results of this study and Kolics et al. (2008), the song patterns in genus Saga are distinct between taxa and they could be used for taxonomic purposes.
Remarks on the relation of distribution and song diversity: Genus Pezodrymadusa is distributed in Anatolia with 14 endemic taxa, in Caucasia with P. magnifica, and in Iran with P. grisea (Eades et al. 2013). Anatolian species are distributed in a narrow area separated by short distances especially in the eastern part of central Anatolia and the eastern Anatolia (Karabağ 1961, Sevgili et al. 2012b, Ünal 2013b). This distribution gives us a hint on why the species of this genus have a uniform-like song pattern but a different morphology. Heller (2006) mentions a similar situation in different allopatric groups or genera, such as Psorodonotus fieberi ssp. or Eupholidoptera, and suggests as the most possible explanation the fact that the changes in song in these groups appear slower than the changes in morphology. Similarly, Pezodrymadusa shows the same pattern for the recorded taxa.
Parapholidoptera castaneoviridis species-group has 16 members (Eades et al. 2013). Up to now, the songs of the four species of this group have been described showing identical pattern (Heller 2006). The song pattern of the three species described in the present study and belonging to the same species-group, also corresponds to the latter. The general distribution of the species in genus Parapholidoptera shows allopatric pattern, but only P. distincta and P. signata occur parapatrically (Heller 2006). These two parapatric species have a very different song pattern (see detail in Heller 2006). When the distribution and song diversity of Parapholidoptera species are considered, parapatric taxa develop stronger acoustic specializations than allopatric taxa.
However, though in P. castaneoviridis species-group and genus Pezodrymadusa "changes in song appear more slowly than changes in morphology" (Heller 2006), genus Psorodonotus tells us a different story. The distribution of Psorodonotus shows in general an allopatric pattern, but P. davisi and P. specularis occur parapatrically (Karabağ 1958, Ünal 2006, Ünal 2013b. Heller (2006) shows that the three subspecies of Psorodonotus fieberi songs do not differ. However, the song diversity of Psorodonotus species recorded in this study discloses that the song is applicable to species identification. Besides, the males of the studied species also differ in matter of comparison with the titilators (Karabağ 1952, Karabağ 1956. These data about Psorodonotus appear in the case of "allopatric forms differ in song and also morphology".