Introduction

Insect flight evolved around 400 million years ago (Grimaldi and Engel 2005; Misof et al. 2014), probably only once (Hovmöller et al. 2002; Misof et al. 2014). This way of moving around was a key evolutionary innovation in insects and is one of the reasons for their success (Engel et al. 2013; Nicholson et al. 2014). Flight is advantageous for dispersal and migration (Bowler and Benton 2005) as it allows the quick exploitation of a wide range of interspersed habitats (Kingsolver and Koehl 1994; Denno et al. 1996; Langellotto and Denno 2001). Well-developed wings enable insects to disperse widely and easily in search of mates, food, and new habitats. In some species, flight is primarily an adaption for dispersal, with important consequences for gene flow, speciation, and evolution. The ability to disperse likely makes a major contribution to the fitness of individuals (Mayhew 2007). Powered flight is energetically costly as individuals have to produce and sustain lift and overcome drag. Reviews of flight costs have revealed an energy partitioning conflict between reproduction and flight, resulting in a trade-off: the so-called oogenesis-flight syndrome (Dixon et al. 1993; Guerra 2011). Even at rest, long-winged insects (capable of flight) need extra energy for their flight muscles (Reinhold 1999) and store metabolic resources as readily available flight fuels (Zera 2005; Zera and Zhao 2006). In contrast, flightless individuals can invest more energy into reproduction (Guerra 2011; Steenman et al. 2015).

The development of wings and ability to fly are characteristic traits of all higher insect orders. Nonetheless, a multitude of insect taxa secondarily lost the ability to fly (Roff 1990, 1994a; Wagner and Liebherr 1992), which is interpreted as an evolutionary adaptation to environmental factors (Roff 1994b; Hodkinson 2005). There have been debates on the evolutionary forces driving the occurrence of regressive traits, including flightlessness in insects (Fong et al. 1995; Lahti et al. 2009). The development of flight organs and their physiological maintenance imposes metabolic costs, and any individual must weigh the benefits of flight against the costs. When dispersal does not bring ecological benefits, flightlessness is the natural consequence; this is the case for cave, subterranean or small island insects, and for phoretic species like fleas (Roff 1990, 1994a; Wagner and Liebherr 1992). Habitat stability in forests also seems to support the evolution of flightlessness (Roff 1994b), at least in females (Hunter 1995; Snaell et al. 2007). Furthermore, short-winged species are common in the temperate zones (north and south) and high mountains of the tropics (Roff 1990). Winglessness in these cases seems to be better explained by the costs, where losing the ability to fly is an adaptation for energy saving in less favorite climates. Regardless of the ecological reasons, flightlessness can boost species diversity (Ikeda et al. 2012) and correlates with relaxed molecular evolution in energy-related mitochondrial genes (Mitterboeck and Adamowicz 2013; Mitterboeck et al. 2017).

The bushcricket or katydid subfamily Phaneropterinae is a suitable taxon for studying the origins of flight loss. This subfamily is the most species-rich group within the Tettigonioidea and has approximately 2500 species distributed worldwide, mostly in the tropics (Cigliano et al. 2018). In line with the general pattern for wing reduction, short-winged Phaneropterinae occur in tropical mountains (Braun 2010; Massa 2015), where brachypterism increases with altitude (Braun 2011). In the Western Palaearctic, the situation is completely different: there are just a few long-winged Phaneropterinae and a very large number of short-winged species. All short-winged Phaneropterinae worldwide were originally placed in the single tribe Odonturini (Brunner von Wattenwyl 1878, 1891). The European perspective made it necessary to separate several genera into the tribe Barbitistini (Bey-Bienko 1954). These have radiated into a vast number of allo- and parapatric species (Lehmann 1998) with moderate to small distribution areas in the Eastern Mediterranean (Bey-Bienko 1954; Heller 1984; Willemse and Heller 1992), comprising around 300 described species in 15 genera (Cigliano et al. 2018). This radiation in the Eastern Mediterranean is linked to extraordinary diversification in acoustic communication systems (Heller 1984, 1990, 2006) and corresponding sensory ecology (Stumpner and Heller 1992; Strauß et al. 2012, 2014). From the plesiomorphic state of bidirectional acoustic communication in Phaneropterinae, where males sing and females answer (Heller et al. 2015), some species within the Barbitistini genus Poecilimon Fischer, 1853 reduced the female’s wings, which thus became non-functional for acoustic communication (Heller 1984, 1992; Heller and von Helversen 1993; Anichini et al. 2017). Furthermore, one species, P. intermedius (Fieber, 1853), has switched to obligate parthenogenesis (Lehmann et al. 2011), which occurs in less than ten bushcricket species worldwide, and also has sensory reduction due to its missing sexual communication (Lehmann et al. 2007; Strauß et al. 2014). The reproduction of the Barbitistini, especially the genus Poecilimon (McCartney et al. 2008, 2010), is well-studied due to the extremely large nuptial gifts transferred during mating (Lehmann 2012). One could speculate that flightlessness is a primer for the extraordinary, large nuptial gifts in this tribe, as variations in the sex that searches for mates (unidirectional versus bidirectional species) correlate with spermatophore size (McCartney et al. 2012). Moreover, no other Orthoptera group has been more intensively studied with respect to chromosomal evolution (both chromosome numbers and structures) (Warchałowska-Śliwa 1998; Warchałowska-Śliwa and Heller 1998; Warchałowska-Śliwa et al. 2000, 2008, 2011, 2013; Grzywacz et al. 2011). Astonishingly, all studies reveal low chromosome differentiation between species and genera (Warchałowska-Śliwa et al. 2013; Grzywacz et al. 2014a) and show small phylogenetic signal. The few published molecular studies on Barbitistini bushcrickets are limited to species groups (Lehmann 1998), included in barcoding analyses (Hawlitschek et al. 2017), or have mainly concentrated on the genera Poecilimon (Ullrich et al. 2010) and Isophya Brunner von Wattenwyl, 1878 (Chobanov et al. 2017). In a previous study, we were unable to clarify the position of the Barbitistini relative to Odonturini, due to restricted taxon sampling (Grzywacz et al. 2014b).

Species that radiated in the Western Mediterranean show low lineage diversification, with 17 taxa placed in the single genus Odontura Rambur, 1838 (Supplementary Table 1). They show little variation in their bidirectional acoustic communication system (Heller 1988; Grzywacz et al. 2014b), with the exception of O. microptera; the tegmina in females of this species do not touch each other; therefore, this species might have secondarily returned to a unidirectional acoustic communication system. However, despite their restricted species number, their chromosome organization is very differentiated, with autosome numbers ranging from 26 to 30 in males and sex chromosomes having evolved multiple times (Warchałowska-Śliwa et al. 2011; Grzywacz et al. 2014b). There is little geographic overlap between the many Barbitistini genera in the east and the genus Odontura in the west (Harz 1969; Heller 1988). To add further complication, a multitude of short-winged genera around the world without geographic overlap and very distinct morphologies are still included in the Odonturini, despite the cautionary comments of Braun (2011) and others (Supplementary Table 2, Cigliano et al. 2018).

Here, we performed a molecular phylogenetic analysis of the flightless West-Palaearctic Barbitistini and Odonturini alongside a diverse subset of long-winged genera, including all European genera. We were also able to include the short-winged Cohnia andeana (Hebard, 1924) from South America. We selected three nuclear genes (small subunit ribosomal RNA gene-18S rDNA, histone 3-H3, internal transcribed spacer 2-ITS2), which have been used successfully to resolve Orthoptera phylogenies (Svenson and Whiting 2004; Jost and Shaw 2006; Ullrich et al. 2010; Mugleston et al. 2013; Song et al. 2015; Chobanov et al. 2017). The present study focuses on the evolutionary origin of flightlessness in Phaneropterinae and clarifies the status of the Odonturini.

Material and methods

Taxon sampling

A total of 101 specimens belonging to 42 species of 17 genera of the Phaneropterinae (Orthoptera, Tettigonioidea) were selected for this study (Table 1). From the tribe Barbitistini, which comprises only short-winged species, we sampled 31 species from 10 out of the 15 genera currently recognized. From the tribe Odonturini, which also only comprises short-winged species, we included four species from the genus Odontura and the tentatively placed South American species Cohnia andeana. We also added six long-winged Phaneropterinae species covering selected genera occurring in the Western Palaearctic and one from East Asia hypothesized to be closely related to Odontura (compare Ragge 1980). Three taxa from the genus Tettigonia Linnaeus, 1758, representing a different bushcricket family (Orthoptera, Tettigonioidea, Tettigoniidae) were selected as the outgroup.

Table 1 Taxonomic information and GenBank accession numbers for taxa included in this study

DNA extractions, PCR amplification, and sequencing

Genomic DNA was extracted from a hind leg of individuals using the NucleoSpin® Tissue kit (Macherey-Nagel, Germany), following the manufacturer’s instructions. Polymerase chain reaction (PCR) was carried out to amplify three nuclear genes: a fragment of the small subunit ribosomal RNA (18S rDNA), histone 3 (H3), and internal transcribed spacer 2 (ITS2). The primers used for the amplifications were 18Sai [5′-CCT GAG AAA CGG CTA CCA CAT C-3′] and 18Sbi [5′-GAG TCT CGT TCG TTA TCG GA-3′] for 18S rDNA (Whiting et al. 1997), H3fwd [5′-ATG GCT CGT ACC AAG CAG ACG GC-3′] and H3rev [5′-ATA TCC TTG GGC ATG ATG GTG AC-3′] for H3 (Colgan et al. 1998), and ITS2-28S [5′-GGA TCG ATG AAG AAC G-3′] and 28S–18S [5′-GCT TAA ATT CAG CGG-3′] for ITS2 (Weekers et al. 2001).

The PCR reaction was performed in a 30-μl reaction volume containing 3.0 μl of 10× PCR buffer, 25 mM MgCl2, 10 mM dNTP mixture, 15 μM forward and reverse primers, 1 μl of genomic DNA, 0.2 μl of Taq DNA polymerase (EURx, Poland), and sterile deionized water. The general PCR profile run on the Thermocycler Mastercycler EP (Eppendorf, Germany) consisted of an initial denaturation step at 95 °C for 4 min, followed by 34 cycles at 95 °C for 30 s, 50 °C for 1 min, and 72 °C for 2 min, and a final extension step of 10 min at 72 °C. The cycling conditions for the 18S rDNA amplification consisted of an initial denaturation for 3 min at 94 °C followed by 30 cycles at 94 °C for 1 min, 51 °C and 72 °C for 1.30 min, with a 10-min final extension at 72 °C. PCR products were purified with the GeneMATRIX PCR/DNA Clean-Up Purification kit (EURx, Poland; following the standard protocol) and were sequenced using the ABI Prism BigDye® Terminator kit version 3.1 (PE Applied Biosystems, Foster City, CA) and ABI 3730XL sequencer. DNA sequences for each gene were deposited in GenBank under the accession numbers listed in Table 1.

Phylogenetic analyses

The obtained nucleotide sequences were aligned and edited in Sequencher v. 4.1 (Gene Codes Corporation). Ambiguously aligned regions were identified following the method proposed by Lutzoni et al. (2000). An unambiguous alignment of a 111 bp portion of ITS2 could not be achieved, similar to the ambiguous characters in those nuclear markers found in a previous study of Phaneropterinae bushcrickets (Ullrich et al. 2010). Therefore, this region was excluded from further analysis. The partition homogeneity test (Farris et al. 1995) implemented in PAUP 4.0a (Swofford 2002) was used to determine the validity of combining 18S, H3, and ITS2 genes into a single analysis.

Phylogenetic inference analyses were conducted using maximum likelihood (ML) and Bayesian inference (BI). Best fit models for ML and BI analyses were calculated in MrModeltest (Nylander 2004) using the Akaike information criterion (AIC). SYM with gamma distribution (SYM + G) represented the best fitting model of nucleotide substitution for the combined datasets. Maximum likelihood (ML) analyses were conducted in PAUP 4.0a. Bootstrap support (BS) was calculated with 1000 replicates. Bayesian analysis was performed in MrBayes v 3.1 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003) with four independent runs, each having three heated and one cold chain. Analyses were run for 10 million generations with trees sampled every 100 generations. Bayesian posterior probabilities (PP) were calculated using a Metropolis-coupled, Markov Chain Monte Carlo (MCMC) sampling approach. The first 25% of each run was discarded as burn-in. Convergence among the runs was assessed using Tracer v1.5 (Rambaut and Drummond 2009). Figtree (http://tree.bio.ed.ac.uk/software/figtree) was used to visualize the trees. Pairwise genetic distances were calculated in Mega v 6.0 (Tamura et al. 2013).

Results

Sequence data

The final DNA sequence dataset comprised 1177 bp. The sequences of the ITS2 gene (289 bp) were more polymorphic than those of H3 and 18S: for all taxa, 70% of sites were variable and about 40% parsimony informative. The tribe Barbitistini had around 60% of sites variable and 47% parsimony informative for the ITS2. The sequences of the H3 gene (328 bp) were less variable: for all taxa, 47% of sites were variable and 18% parsimony informative. In this case, the tribe Barbitistini had 40% of sites variable and 7% parsimony informative. Sequences for 18S rDNA (560 bp) showed a much lower rate of polymorphism, of just 9% across all taxa with 4% of sites parsimony informative, and for Barbitistini, the values were 3 and 1%, respectively. The partition homogeneity test did not detect significant incongruence between genes; so, our analyses were conducted on the combined dataset. For the rDNA genes, in-group genetic distances within Phaneropterinae tribes were largest in Barbitistini (35%) and Odonturini (34%) (Table 2).

Table 2 Mean pairwise distances within studied tribes of Phaneropterinae calculated from combined three nuclear genes (18S + H3 + ITS2)

Phylogenetic analyses

The analyses of maximum likelihood and Bayesian inference resulted in similar tree topologies (Fig. 1). The Bayesian posterior probability values for the nodes were generally higher than the bootstrap values. The outgroup taxon sampling with the three species of the genus Tettigonia, members of the bushcricket subfamily Tettigoniinae, clearly defined and routed the tree. The phylogenetic analysis divided the subfamily Phaneropterinae into four lineages. The first major lineage (short-wing group I) is the sister group to all the rest and comprises solely the species of the genus Odontura, whose monophyly was supported in all analyses, with a posterior probability of 0.7 and bootstrap values of 68%. Our phylogeny suggests that the short-winged Cohnia andeana from the Andes of South America, tentatively placed within Odonturini (Braun 2011), is phylogenetically separated from Odontura and in our analysis represents the sole species of an additional, New World lineage (short-wing group II), paraphyletic to the rest of the Phaneropterinae. The “long-winged” taxa of Phaneropterinae included in this study do not form a homogenous group: they are separated by long branches and have unresolved basal relationships. The latter aspect shows their distinctiveness and supports their traditional placement in four separate tribes: the Acrometopini, Ducetiini, Phaneropterini, and Tylopsidini. Within the fourth major lineage, members of the species-rich tribe Barbitistini form a well-supported monophyletic lineage (short-wing group III) with 100% posterior probability but low bootstrap values (< 50%). The typically brachypterous groups, especially the Barbitistini and the Odonturini, are well separated, which suggests that these groups lost flight independently of each other.

Fig. 1
figure 1

Bayesian tree from three nuclear gene analyses as performed in MrBayes. The two values on each branch represent the following: (1) Bayesian posterior probability (PP) and (2) maximum likelihood bootstrap support (BS) (only support values above 50%) as PP/BS

Taxonomic conclusions

Based on our analysis, we restrict the tribe Odonturini Brunner von Wattenwyl, 1878 solely to the West-Palaearctic genus Odontura Rambur, 1838 (Supplement Table S1), well separated from the tribe Barbitistini. Other flightless genera from the Americas, tropical Africa and Papua New Guinea and previously included in Odonturini are placed here as incertae sedis (Supplement Table S2) until their phylogenetic position can be established.

Discussion

The phylogenetic analyses of the sequences of the three nuclear DNA genes created a single most parsimonious tree for the Phaneropterinae. Within this tree, the genus Odontura is a clear monophyletic entity, which can be classified as the tribe Odonturini (Supplementary Table 1). However, the long-lasting inclusion of other short-winged genera from all over the world into the Odonturini is rejected here (Supplementary Table 2). The South American genus Cohnia branches outside the Odonturini, in full agreement with our previous phylogenetic results for this genus (Grzywacz et al. 2014b). The phylogenies of two additional genera formerly placed inside the Odonturini have been analyzed before by Mugleston et al. (2013, 2016). Firstly, the Austrodontura capensis (Walker, 1869) (Naskrecki and Bazelet 2011) from the Fynbos Flora of South Africa branches out with a Madagascan long-winged species Parapyrrhicia dentipes Saussure, 1899 (Mugleston et al. 2016) arbitrarily placed in the Phaneropterini, but certainly belonging to an unnamed tribe (Hemp et al. 2017a). The second genus Monticolaria Sjöstedt, 1910 from the mountain arc of East Africa (Hemp et al. 2009; Massa 2015) clusters deep inside a branch containing a number of long-winged genera, traditionally embedded in the Phaneropterini and Tylopsidini (Mugleston et al. 2016). On the basis of our results in combination with the mentioned molecular phylogenetic studies, we formally restrict the tribe Odonturini to the 17 taxa of the genus Odontura and exclude all other short-winged genera from the Odonturini (see Supplementary Table 2). A future study with more comprehensive coverage from the megadiverse Phaneropterinae might clarify their phylogenetic affinities. The current knowledge suggests that the relatives of the short-winged Phaneropterinae may be best searched for in the fully winged genera of the regions of their occurrence. The wing reduction in Odontura and Cohnia is supposed to have evolved independently as both genera live on different continents, with Odontura in the Western Mediterranean, and Cohnia in the Central Andes of South America.

The Eastern Mediterranean genera form a third, well-supported monophyletic group of flightless species, which corresponds with their classification as the tribe Barbitistini. This tribe or at least a subset of genera belonging to it has been repeatedly supported as a monophyletic group (Mugleston et al. 2013, 2016). All species are short-winged and speciation may well have been related to the limited dispersal capacities in the oro-geographic diverse landscapes of Southeastern Europe and Anatolia (Lehmann 1998). This may have contributed to the sheer number of species (around 300), with little geographic overlap between the closely related taxa (Lehmann 1998; Boztepe et al. 2013; Kaya et al. 2015; Chobanov et al. 2017). Similarly to the western Odonturini, the nearest relatives of the Barbitistini are unknown: they probably also originated from some long-winged species that found refugia in the Eastern Mediterranean and radiated into allo- and parapatrically distributed species. The radiation resulting in the huge species number may be the result of geographic separation (discussed in Lehmann 1998) caused by climate cycles including multiple ice-ages (see Hewitt 2000, 2004). The genus Leptophyes has been repeatedly found to branch with the Poecilimon-cluster (Ullrich et al. 2010; Mugleston et al. 2013, 2016; Grzywacz et al. 2014b), which is also supported here for L. punctatissima. Interestingly, the other three species, Leptophyes albovittata (Kollar, 1833), L. boscii Fieber, 1853 and L. discoidalis (Frivaldszky, 1868), nested in another subgroup. This split is coherent with differences in general morphology and bioacoustic data (Bey-Bienko 1954; Kleukers et al. 2010; Sevgili 2004); however, we leave this question open until more Leptophyes species are studied. The genus Andreiniimon Capra, 1937, based on its overall appearance, used to be related to Leptophyes (Bey-Bienko 1954), but our study found it genetically clustered with Metaplastes Ramme, 1939, Barbitistes Charpentier, 1825, and Ancistrura Uvarov, 1921. Interestingly, highly modified external male genitalia are a shared trait of Ancistrura–Andreiniimon–Barbitistes–Metaplastes, which is coupled to a unique sperm-removal ability in the genus Metaplastes (von Helversen and von Helversen 1991; Foraita et al. 2017). Therefore, the modified external male genitalia might be a synapomorphic character for the group and the overall similarity of Andreiniimon with Leptophyes either results from plesiomorphy or convergent evolution. The genus Isophya, with the second highest species-number (Chobanov et al. 2013, 2017; Grzywacz et al. 2014a), has an ancestral position within the Barbitistini.

The full development of wings is the plesiomorphic character state in the Phaneropterinae. Therefore, it was well expected that the European long-winged genera are rather unrelated, separated by long branching axes. This is in complete agreement with the morphologically classified traditional system, where Acrometopa Fieber, 1853, Ducetia Stål, 1874, Phaneroptera Serville, 1831, and Tylopsis Fieber, 1853 are placed in different tribes (Cigliano et al. 2018). The deep splits between the genera studied here are also supported by a much broader phylogenetic approach (Mugleston et al. 2013, 2016). Interestingly, none of the long-winged genera that currently occur in the Mediterranean region seem to be closely related to the short-winged Odonturini in the Southwest or the short-winged Barbitistini in the Southeast. As concluded by several authors (Braun 2011; Naskrecki and Bazelet 2011; Grzywacz et al. 2014b; Massa 2015), wing size reduction resulting in flightlessness must have occurred multiple times in the Phaneropterinae, which is supported by the recent discovery of the short-winged East African genus Peronurella Hemp, 2017, belonging to the tribe Acrometopini (Hemp et al. 2017b). The conclusions of our study are in line with such a notion and support two evolutionary events within the Western Palaearctic and a probable third in South America. The overall number of times this type of concerted evolution has occurred is obviously higher, as worldwide many more short-winged Phaneropterinae genera of unknown tribal affinities occur (Supplementary Table 2). Based on their distinct appearance, we can easily conclude that the short-winged genera do not form a single worldwide group, and instead likely developed regionally on different continents. Phaneropterinae generally lack genital titillators as a shared character (Vahed et al. 2011; Lehmann et al. 2017), but it occurs independently in central and North American taxa of the Dichopetala group (Cohn et al. 2014; Rocha-Sánchez et al. 2015; Barrientos-Lozano et al. 2016), the South African genus Brinckiella Chopard, 1955 (Naskrecki and Bazelet 2009), and the Asian Letana inflata Brunner von Wattenwyl, 1878 (Heller and Liu 2015). However, the convergent development of titillators is linked to mating-related features such as copulation duration (Vahed et al. 2011; Lehmann et al. 2016) or polyandry (Lehmann et al. 2017), but is not coupled to flightlessness. In conclusion, the evolutionary transition to flightlessness seems to be moderately common in bushcrickets and the five events mentioned in the global analysis by Mugleston et al. (2013) are an underestimation.

Brachypterism can be interpreted to be an evolutionary adaptation resulting from a variety of different environmental conditions. At least three conditions could support flightlessness: first, habitat stability is often an explanation, as dispersal becomes less favorable with increasing stability. The chance of ending up at inferior places after emigration may counter its positive effects, such as outbreeding opportunities. Second, the same is true for isolated habitats, where travel by air is either impossible, like in caves, or dangerous as on islands or mountains (Roff 1990; Wagner and Liebherr 1992). Third, whatever the ecological circumstances, an individual is more likely to disperse when the potential benefits exceed the risks and travel costs (Roff 1984, 1990; Wagner and Liebherr 1992).

In line with the general picture for insects (Roff 1990; Guerra 2011), the loss of flight in tropical bushcrickets occurs predominantly at higher altitudes (Braun 2011), either to save metabolic energy at lower temperatures or because long-distance dispersal provides fewer opportunities to colonize new habitats on mountains than in flat landscapes. However, the situation in the Western Palaearctic needs closer examination. Barbitistini genera have an early seasonal occurrence coupled with a rather short and synchronized life history (Lehmann and Lehmann 2006; Lehmann 2012). This might be a leftover from multiple ice-cycles in which the flightless species adapted to local climates. In line with such a scenario, many species nowadays occur on mountains and abundantly thrive in mesophilic meadows. Alternatively, the pattern of a short and early adult season may allow the species to avoid the hot and dry Mediterranean summers. Chobanov et al. (2017) estimated the common ancestor of Barbitistini to have evolved during the so-called Middle Miocene climatic transition, characterized by a global drop in temperatures and dry climates. The loss of flight and fast life cycle may therefore be connected with the necessity to avoid water loss (Chobanov et al. 2017). For example, the species Poecilimon thessalicus Brunner von Wattenwyl, 1891 is known to suffer from the early onset of the dry Mediterranean summer, which leads to the populations living on the dryer, eastern mountain slopes to have a smaller body size than those on the wetter, western mountain slopes (Lehmann and Lehmann 2008).

Whatever the evolutionary forces that led to the multiple loss of flight, the reduced dispersal capacities have promoted speciation, as in other insects (Ikeda et al. 2012; Vogler and Timmermans 2012; Sota et al. 2014). The overall diversity in Barbitistini with almost 300 species is an impressive example of rapid speciation (Lehmann 1998) within a restricted temporal (Ullrich et al. 2010; Chobanov et al. 2017) and spatial frame (Heller 1984; Willemse and Heller 1992). Thus, it could be speculated that brachyptery in Barbitistini bushcrickets has contributed to the impressive species number—mainly by allopatric separation events during multiple ice-cycles in an oro-geographic diverse landscape (Lehmann 1998).

Conclusion

Based on a phylogenetic reconstruction using three nuclear markers, we found strong evidence for multiple flight loss in Phaneropterinae bushcrickets. In the temperate zone of the Western Palaearctic, flightlessness originated twice. In the first group, the Barbitistini, speciation led to an impressive number of mainly allo- and parapatrically distributed species in Southeastern Europe, Anatolia, and the Middle East. The second group, the Odonturini, occurs as a limited number of species in Southwestern Europe and Northern Africa. Interestingly, the closest relatives of the Odonturini are still unknown but may be best searched for in afro- or asiotropical species, which invaded Europe during a warmer climatic period, while the Barbitistini probably originated from long-winged species found their refugia in the Eastern Mediterranean.