This study investigates the variations in body size, bioacoustic traits, and cuticular hydrocarbon (CHC) profiles in different populations of the bush cricket species Isophya autumnalis Karabağ, 1962. Within-population body size variations, particularly those associated with distinct habitat differences and climate shifts within their local distribution ranges, suggest that ecological factors affect morphological characteristics. Sexual size dimorphism (SSD) in I. autumnalis may affect reproductive behavior and strategies, potentially influenced by the bioacoustic environment. While male calling songs exhibit temporal variations across populations, suggesting differences among allopatric populations, CHC profiles, known to undergo selection under various climatic conditions, also vary noticeably across local populations. These findings highlight the importance of understanding within-species variations for the conservation of Isophya and similar taxa in the face of habitat threats. Overall, this study contributes to a comprehensive understanding of how morphology and bioacoustic behavioral traits are shaped over short distances in allopatric populations of species with limited mobility, such as I. autumnalis, providing insights into adaptation processes and highlighting the urgency of conservation efforts for endemic species in Anatolia.

biodiversity, chemical signals, ecological variation, evolution, male calling song, Orthoptera, plump bush-cricket, speciation

There are four main speciation types: allopatric, peripatric, parapatric, and sympatric (Howard and Berlocher 1998, Tregenza et al. 2000, Butlin et al. 2008, Fitzpatrick et al. 2009). The most critical factor accelerating allopatric speciation is topographic isolation that prevents reproductive unity between populations. Many studies have shown genetic and morphological differentiation from the ancestral population due to genetic drift and divergent selection caused by various ecological conditions due to the reduction or complete cessation of gene exchange between geographically separated local populations of the common ancestral population (Hemp et al. 2010, Langerhans and Riesch 2013, Taylan and Şirin 2016, Sevgili et al. 2018). However, many questions remain unanswered. For example, questions regarding the extent to which ecological factors (such as climate, vegetation, predator-prey relationships, etc.) lead to reproductive isolation (habitat, temporal, or gametic) in local populations within a species’ range (Sobel et al. 2010, Langerhans and Riesch 2013); the degree of variation that may arise in any trait within local populations as they adapt to varying ecological conditions; and which traits (morphology, behavior, etc.) are more plastic if allopatric speciation occurs, remain insufficiently addressed.

Understanding the degree of variation in biochemical, behavioral, and morphological characters within populations is crucial. Insect mating recognition systems often rely on species-specific chemical compound profiles driven by strong sexual selection (Chapman et al. 2000, Kather and Martin 2012, Moore et al. 2021). In contrast, the evolutionary rates of the taxonomic characters used to define species are not uniform. Environmental conditions, population density, selection pressure, behavioral differences, and mutation rates can vary within local populations, influencing the speed of change (Bromham 2009, Adams 2013). For example, behavioral characters (social behaviors or bioacoustic characters) can evolve more rapidly than morphological traits (Losos 1990, Heller 2006, Pitchers et al. 2014). Orthoptera is one of the insect groups that provides the best evidence to support this claim.

In particular, research on meadow grasshoppers of the genus Chorthippus Fieber, 1852 (Acrididae) shows that whereas its species are very difficult to distinguish morphologically, genetically, and ecologically, they can be clearly differentiated through species-specific bioacoustic traits (Sirin et al. 2010, Nolen et al. 2020, Tarasova et al. 2021). In contrast, studies of the bush cricket species of the genus Poecilimon Fischer, 1853 have identified significant interspecies differences in their morphological, genetic, ecological, and behavioral characters (Heller 1988, Sevgili et al. 2008, Kociński et al. 2022). In some bush cricket genera, such as Eupholidoptera Maran, 1953 and Parapholidoptera Maran, 1953, bioacoustic characters are quite similar across species, while morphological and molecular characters are more likely to be selected for (Heller 2006, Çıplak et al. 2021). Specifically, it is important to study local populations of species with short wings and limited mobility to determine intraspecific variation in characters because these local populations have a fragmented distribution, so there are only narrow zones where gene exchange can occur. In populations with limited mobility due to geographical isolation, selection for pre-breeding isolation mechanisms can be strong, thereby accelerating allopatric and parapatric speciation (Eweleit et al. 2015, Sevgili 2018).

Turkey has a highly diverse geography and rich biological diversity with different genetic patterns across short distances. In the present study, we investigated the variability of several taxonomic characters in local subpopulations of Isophya autumnalis Karabağ, 1962 (IA). The genus Isophya Brunner von Wattenwyl, 1878 is distributed in the Palaearctic region but is especially concentrated in the Balkans and Anatolia and includes species with short wings and limited mobility. Out of 90 species globally, the 44 species and subspecies distributed in Anatolia include 77% local endemic species with narrow distributions (Sevgili 2004, 2018, 2020, Sevgili et al. 2006, 2012). We tried to reach other potential local IA populations previously reported from the type locality, namely the Zigana Mountains, Trabzon. We hypothesized that there are differences in morphological/morphometric, bioacoustic, and biochemical (cuticular hydrocarbons, CHC) characters between IA subpopulations at localities that are somewhat distant and have different climatic conditions and vegetation structures. Body size, which is a vital trait for organisms, is highly variable among local populations because of its high sensitivity to changes in biotic and abiotic factors (Chown and Gaston 2010). In addition to having different distribution patterns, local population body sizes may also vary due to different climate (Bidau and Martí 2007, Eweleit and Reinhold 2014), food/water resources (Stillwell et al. 2007), habitat type/width and vegetation structure (König and Krauss 2019), size-dependent predation pressure (Remmel and Tammaru 2009), population density (White et al. 2007), and general life history traits (Roff 2002, Chown and Gaston 2010).

Isophya’s distinctive species-level taxonomic characters derive from morphology, bioacoustics (Heller et al. 2004, Sevgili 2004, Sevgili et al. 2006, Ünal 2010), mitochondrial (cyt b, COI, COII, and ND2), and nuclear (ITS1 and ITS2) molecular markers (Grzywacz-Gibała et al. 2010, Chobanov et al. 2017). However, the molecular findings and taxonomy based on morphology and bioacoustic data (species-specific male calls) do not fully align regarding the phylogeny and differentiation of Isophya species (Grzywacz-Gibała et al. 2010, Chobanov et al. 2017). IA belongs to the I. zernovi species group (Sevgili 2004, 2020, unpublished data Sevgili et al.). While ecological differentiation plays a crucial role in speciation, it is important to recognize that sexual selection can also interact with it, further influencing the process (Weissing et al. 2011).

CHCs, which consist of an intricate blend of unsaturated hydrocarbons, methyl-branched alkanes, and n-alkanes secreted from arthropod cuticles, are a biochemical taxonomic character in social insects (Martin et al. 2008, Kather and Martin 2012, Peña-Carrillo et al. 2021) and some other insect groups (Chorthippus sp., Finck et al. 2016; Drosophila sp., Davis et al. 2021; ladybugs, Redjdal et al. 2023). CHCs have many functions. In terrestrial environments, they primarily prevent dehydration while also serving as sexual signal molecules in a wide variety of chemical communication systems (Drijfhout et al. 2010, Ingleby 2015, Holze et al. 2021).

Rather than addressing these functions, we investigated the taxonomic utility of CHC profiles based on their variation across local populations of IA. However, we also hypothesize that CHC profiles in Isophya populations will be less variable than morphological/morphometric and bioacoustic characters. This is because Isophya species are not eusocial insects, and male calling songs are used for communication, such as finding mates. We also hypothesized that males and females within a population would have dimorphic CHC profiles. The results of this study are important for understanding which characters are particularly important for speciation processes in narrowly distributed and endemic populations. The results can also aid efforts to conserve narrowly distributed endemic species by clarifying which traits are more variable in allopatric populations. Given that Isophya species have a fragmented distribution pattern due to their limited mobility and unique habitat preferences (Nuhlíčková et al. 2023), and that gene exchange is largely restricted, it is important to determine how much the multiple taxonomical characters differ between local populations.

The plump bush crickets.—The model for our study, IA, is an endemic species described by Karabağ in 1962 based on material collected during a field study in September from the Zigana Mountains (Trabzon, Turkey). In species of the genus Isophya, the wings of both sexes are shortened and do not reach half of the abdomen, and the wings, especially in males, have a species-specific sound-producing organ (stridulatory organ) (Heller and von Helversen 1986, Sevgili et al. 2012, Iorgu et al. 2013). All Isophya species are plump-green bush-crickets with limited mobility (Sevgili 2004, Nuhlíčková et al. 2023). They are generally among the earliest bush-cricket species to emerge in the field at the end of winter. They hatch earlier (around February) in the southern parts of the genus’s distribution range (e.g., Southeastern Anatolia, Isophya sikorai Ramme, 1951 and later (around June) in northern Anatolia and the high mountains (e.g., Mountains in Black Sea Region, I. rizeensis Sevgili, 2003 (Sevgili 2004). Field observations show that males hatch earlier than females. Isophya species generally reach adulthood after five molts during the nymphal stage (Can 1958). Males first produce a calling song at 2–3 days old, while successful mating generally begins after the seventh day (when sperm transfer occurs) (Uma and Sevgili 2015). IA is a typical mountain species, and the northern population is distributed within alpine zones at high altitudes in Zigana Mountain. Nymphs can be found in suitable habitats in June and adults until mid-September. The species was known only from the type locality, but we found populations further south as a result of our field studies. We observed several morphological and phenological differences in local populations and investigated the variation in taxonomic characters.

Spatial differences of the sampling locations.—Species’ presence in a habitat, their ecological niche, and distribution patterns, including habitat quality, seasonality, and life cycles, can often be understood by utilizing remote sensing variables (Leitão and Santos 2019, Schwager and Berg 2021). In this study, we determined some environmental characteristics of the five sampled localities using variables provided by satellite remote sensing (S-RS) products, which are used to measure species–environment relationships that predict species distributions. In particular, it was found that S-RS products provide more reliable spatial resolution when used together with bioclimatic data to distinguish the habitat characteristics of species (Pinto-Ledezma and Cavender-Bares 2020). Normalized difference vegetation index (NDVI), which measures green vegetation cover, and normalized difference moisture index (NDMI), which monitors vegetation water content and drought, were used to compare the locations studied. NDVI, which ranges from -1 to 1, is a measure of the health of vegetation based on how plants reflect light at certain wavelengths. Values close to zero indicate barren areas, such as rocky and sandy areas, while a positive 1 indicates forested areas with mild temperatures and abundant rainfall. NDMI also ranges from -1 to 1 and provides important data on the water content and drought status of vegetation. Negative values of NDMI correspond to barren soil, values around zero indicate water stress, and high positive values represent canopy without water stress.

For this purpose, the Sentinel-2B level-2A product was downloaded using the European Space Agency (ESA) data portal (https://browser.dataspace.copernicus.eu/). Satellite images were downloaded after applying a search constraint to ensure cloud cover was below 5%. The files were downloaded with high resolution selected. All viewing and analysis processes were conducted using the “raster” and “ggplot2” R packages. Satellite data from July 15, 2018 to August 31, 2018 were evaluated, which is the period when the adult stages of IA passed through and samples were collected. NDVI and NDMI data were downloaded for both July and August for a snow-free and cloud-free 2 km² area over the location where the specimens were collected. In addition, Sentinel -2 bands BO4 (red) and BO8 (NIR-near-infrared) were used to calculate the NDVI index. The normalized difference vegetation index, NDVI = (NIR-RED)/(NIR+RED) equation was used to calculate the normalized difference moisture index, NDMI = (NIR-B11)/(NIR+B11) (Piedelobo et al. 2019, Lastovicka et al. 2020). B11 and 12 short-wave infrared (SWIR) bands are often used for examining surface features such as vegetation, soil, and water. Calculations and plots were performed in the R environment using raster and rasterVis packages (Hijmans and van Etten 2012, Perpiñán and Hijmans 2023). SWIR band measurements, a technique for classifying species, utilizes either the continuous spectrum from visible to shortwave infrared, specific spectral curve shape-based features, or vegetation indices. Linking spectral data with biochemical and physiological ancillary data can enhance classification accuracy and relate species’ spectral separability to their characteristic ecophysiological traits. In this context, vegetation indices and spectral features can be considered semi-quantitative biochemical parameters of the reflectance spectrum, aiding in assessing the physiological status of vegetation (Große-Stoltenberg et al. 2016). Monthly/average annual temperature and relative humidity data for the studied regions by single point method were obtained from Prediction of Worldwide Energy Resources (POWER) (https://power.larc.nasa.gov/beta/data-access-viewer/) by filtering for “temperature at 2 meters” and “relative humidity at 2 meters.” A table was created by adding one year before and after the sampling year (2018). Since no specific data for Pekün Mountain could be produced from this website, it was not included in the table.

Field study.—During field studies conducted in Trabzon and Gümüşhane provinces in June–July 2018, populations were found at five different locations in, approximately, a line from north to south (Fig. 1). From each location, 60 male and female nymphs of the last stage were collected and transported live to the laboratory in 30 × 20 × 20 cm cages with a piece of plant from the habitats where they were collected for feeding (Suppl. material 1: fig. S1).

Fig. 1. 

Map of sampling localities of Isophya autumnalis. The sampling sites are also the latest known distribution points of the species.

Maintenance of the bushcrickets.—In the laboratory, the samples were transferred to larger 50 × 30 × 30 cm cages, with males and females separated. The cages were cleaned every two days and supplied with the same amount of nettle (Urtica dioica), Taraxacum sp., lettuce, and apple slices. The adult males and females were transferred to smaller individual cages and given the above-mentioned food. Water was sprayed into the cages every day to provide moisture. Adult males began producing sounds on the second or third day, and calling songs were recorded after the third day. Specimens were collected from field locations at different times due to differences in aspect, vegetation structure, and distance between locations. However, feeding conditions were standardized, and care was taken to ensure that the specimens were of the same age.

Bioacoustics.—To record the males’ two-syllable calling songs, each male was placed in a small cage and taken to an isolated room where it could not hear the calling songs of other males. Songs were recorded from each individual for at least 3 minutes. All recordings were made under the same temperature conditions (25–27°C). For song recordings, a condenser ultrasonic microphone (Avisoft Bioacoustics CM16/CMPA, sampling rate 96 kHz) connected to a digital recorder, TASCAM HD-P2, was used. The recorded songs were analyzed using Raven Pro software (Raven Sound Analysis 2022) by taking the necessary spectral measurements. The males were then preserved at -80°C together with the females (all 8 days old) based on their field locations to examine the CHC profiles and morphological measurements. Song methodology and terminology follows Heller et al. (2004), Sevgili et al. (2006) and Sevgili et al. (2012).

Morphology/morphometry.—The taxonomically important anatomical parts of each specimen, such as the head, fastigium, pronotum, wing, hind femur, cercus, subgenital plate, epiproct, and ovipositor of females, were photographed with a stereo zoom microscope (Suppl. material 1: fig. S2). The morphology, number of teeth, and shape of the male stridulatory organs differ significantly between species (Sevgili 2004). Therefore, to identify variations in the stridulatory organs between local populations within the species, photographs of the organ on the upper left wing of the males from each population were taken using a scanning electron microscope (SEM Hitachi, SU-1510). Morphometric measurements of all photographs were taken using the ImageJ program (https://imagej.nih.gov/ij/download.html). In measurements of the stridulatory file, the focus was on the shortest distance of the file, the distance to the wing proximal and distal edges, and the number of teeth (Suppl. material 1: fig. S3).

Regarding morphology, measurements were taken of the fastigium length, fastigium base width, interocular distance, dorsal pronotum length, pronotum width, lateral pronotum length, right posterior femur length, epiproct length and width, subgenital plate length, shortest proximal and distal cercus lengths in males, and length of the ovipositor and subgenital plate of females (see Suppl. material 1: fig. S2).

To identify potential size dimorphism in local populations, sexual size dimorphism (SSD) was estimated for each population based on the following formula: SSD = (size of the larger sex/size of the smaller sex). A positive result indicates that females are the larger sex and vice versa (Lovich and Gibbons 1992, Stillwell et al. 2007). The hind femur length was used in the formula because it is the best predictor of body size in bush-crickets (Eweleit and Reinhold 2014).

CHC extraction and analysis.—Determination of the specimens’ CHC profiles followed the methods traditionally used for research on Orthoptera (Neems and Butlin 1994, Thomas and Simmons 2008, 2010, Thomas et al. 2011, Veltsos et al. 2012, Steiger et al. 2013). CHC data were obtained using a Shimadzu GCMS QP2010 Ultra gas chromatograph-mass spectrometer (GC/MS). Individuals previously frozen at -80°C were kept in a desiccator for 45 minutes to reach room temperature and to get rid of excess water. Each individual was then placed in its entirety in a separate conical-based glass centrifuge tube and 4 ml, 10 ppm pentadecane (internal standard) containing hexane was added. After waiting for 30 seconds in hexane and 5 minutes including vortex, the upper layer of the solution was taken from the vial using disposable glass pipettes and placed in the GC/MS device for analysis. To determine the CHC profiles in the bush-crickets, each sample (1 µl) was injected into the device for split-mode (10:1) analysis in a Stabilwax column (Restek/Rtx-5) with an inner diameter of 30 m × 0.25 mm and chromatographic purity under helium gas. The analysis started at 50°C for 1 min, and then the temperature was raised to 250°C at 15°C per min. The final temperature was increased to 320°C at 3°C per min, and the sample was held at this temperature for 5 min. The extract was separated using the temperature difference in the column. The transfer of components to GC/MS was done at 250°C. Hexane was added to the analysis every day to prevent contamination. Peak numbers were coded taking into account the retention time for data analysis. The names of basic hydrocarbons were identified by examining GC/MS libraries (NIST-17 and Willey libraries) and other published data. The fragmentation profiles were acquired and analyzed with theMS and MSD ChemStation software.

Statistics.—Principal component analysis (PCA) was applied to the measurement data from the morphological parts for both sexes (66 females; 57 males), as shown in Suppl. material 1: fig. S2. Cercus and ovipositor measurements were included for males and females, respectively.

PCA was applied to number of teeth, distal distance, proximal distance, and stridulatory file length, taken from the SEM image of the stridulatory organs of the male wings (Fig. 6; Suppl. material 1: fig. S3). All components with eigenvalues greater than 1 were included in the statistical analysis. The CHC phenotypes of the 66 females and 57 males collected from all localities were compared. Hydrocarbons were identified based on retention times for their mass spectra. The CHC data were evaluated statistically following the methods of related studies (Thomas and Simmons 2009, Steiger et al. 2013, Finck et al. 2016).

Before the analysis, the area under each peak on the chromatograph was divided by the area of the internal standard (pentadecane) to control for variations in CHC extraction efficiency between specimens. These ratios were then transformed using log10 to ensure that the data were normally distributed. The data for each peak were then analyzed by PCA using the “Factoextra” and “FactoMineR” R packages (Lê et al. 2008, Kassambra and Mundt 2020). The first two components (PC1, PC2) were compared based on locality and sex and visualized using the “ggplot2” R package (Wickham 2016). The first five PC scores with eigenvalues greater than 1 (PC1) were analyzed multivariately using the dependent variables of locality and sex and the interaction variable sex*localities. Tukey’s HSD post hoc analysis was applied to compare each PC score between localities. All analyses were performed in R Studio (Rstudio Team 2021).

Spatial differences of the localities.—The Zigana and Krom Valley regions were evaluated together due to their geographical proximity and the availability of similar climate data. All the regions evaluated exhibited different patterns of temperature, humidity, and precipitation (no data for Pekün) (Suppl. material 2: table S1).

The pixel values of the July and August plots of NDVI, NDMI, and SWIR data vary across localities (Suppl. material 1: figs S4–S8; Suppl. material 2: table S2). The mean NDVI, NDMI, and SWIR data for the locations are also different in both months. Zigana has very low mean values for all three data (for both months), while Vauk and Pöske have the highest values (Suppl. material 1: figs S9–S11).

Morphometry.—Suppl. material 2: table S3 provides the mean and standard deviation values for various morphological and bioacoustic measurements of the species. Regarding body size (specifically hind femur length), males from the Zigana and Krom Valley populations were significantly larger than those from the southern populations (ANOVA, F_4,51 = 23.72, p < 0.001). However, there were no significant body size differences between the other groups. For females, the pattern was different. Only the Zigana population exhibited larger body size (ANOVA; F_4,61 = 12, p < 0.001). There were no significant body size differences among the other female populations.

Regarding the PCA of the morphometric measurements of males, the first four components, which collectively explained 67.67% of the variance, all had eigenvalues greater than 1. Comparing populations based on PC1, the Zigana population was significantly distinct from the other populations (ANOVA, F_4,51 = 21.15, p < 0.001) (Fig. 3). There were no significant differences between the mountain populations of Krom, Vauk, and Pöske (p_Krom-Vauk = 0.226, p_Krom-Pöske = 0.363, p_{Pöske-Pekün} = 0.066), whereas the Vauk and Pekün populations were distinct (p_Vauk-Pekün = 0.029).

PC1 was predicted by the following variables in descending order of importance: interocular distance (15.72%), hind femur length (14.46%), epiproct width (14.46%), lateral pronotum length (13.77%), and cercus length (13.04%). The PC2 component was significantly influenced by fastigium length (30.16%), proximal width of fastigium (38.11%), and epiproct length (26.86%).

Regarding the PCA of the female morphological measurements, the first three components, collectively explaining 68.72% of the variance, had eigenvalues greater than 1. Comparing the populations based on PC1 and PC2, only the Zigana population was significantly distinct from the others (ANOVA; F_4,61 = 28.91, p < 0.001) (Fig. 4).

In terms of common morphological structures measured in both males and females (fastigium, interocular distance, pronotum, hind femur, and subgenital plate), PC1, which explained 45.49% of the variance, exhibited clear sexual dimorphism, with females being significantly larger (t-test; t = -17.71, df = 119.91, p < 0.001). However, for pronotum length, which is an important indicator of body size, like hind femur length in bush crickets (Eweleit and Reinhold 2014), there was no significant size difference between males and females (t-test, t = 0.924, df = 99.492, p = 0.358).

Sexual size dimorphism.—Although there was a decreasing SSD trend from north to south in the local IA populations, this difference was not statistically significant (ANOVA; F_1,54 = 1.897, t = -1.377, R² = 0.034, p = 0.174) (Fig. 5).

Male calling song.—The main syllable typically exhibits a crescendo structure, transitioning to a decrescendo slightly past the midpoint. Sometimes, an “after-click” impulse is created at the end of the syllables (Fig. 2). Temporal measurements of the songs were made as shown in Fig. 2.

Fig. 2. 

Schematic oscillograms showing the studied male calling song characters and measurement points.

Fig. 3. 

Principal component analysis (PCA) plots based on some morphological characteristics of allopatric populations of the male Isophya autumnalis (A). Boxplots of morphological traits among different populations (B). Different letters on top of boxplots indicate significant differences. Those with the same letters do not have a statistically significant difference from each other.

Fig. 4. 

Principal component analysis (PCA) plots based on some morphological characteristics of allopatric populations of the female Isophya autumnalis (A). Boxplots of morphological traits among different populations (B). Different letters on top of boxplots indicate significant differences. Those with the same letters do not have a statistically significant difference from each other.

Fig. 5. 

Sexual size dimorphism in Isophya autumnalis based on hind femur length.

Stridulatory organ and bioacoustics.—SEM photos of stridulatory files of the populations are given in Fig. 6. PC1 was most determined by stridulatory organ length (46.36%), followed by proximal distance (27.92%), number of teeth in the stridulatory organ (16.16%), and distal distance (9.55%). PC2 was most determined by number of teeth (37.58%), followed by right proximal distance (36.36%), left proximal distance (19.55%), and shortest distance length of stridulatory organ (6.50%).

Fig. 6. 

Scanning electron microscope (SEM) images showing examples of the male stridulatory file organ by location.

Stridulatory organ length was significantly positively correlated with number of teeth (R² _adj = 0.33, p < 0.001). Regarding the number of teeth, the Krom Valley population had significantly fewer than the others except the Pöske Dağı population (ANOVA, F_4,51 = 8.419, p < 0.001). There were no significant differences between the other populations. The average number of teeth across all populations was 156±13.11 (mean±sd) (range: 128 to 185).

Regarding the PCA analysis of the male stridulatory organ measurements, the first two components, which explained 80% of the variance, both had eigenvalues greater than 1. Comparing populations based on PC1, there were significant differences (ANOVA; F_4,51 = 13.47, p < 0.001) (Fig. 7). Post-hoc analysis (TukeyHSD) revealed that, similar to the other morphological measurements, the Zigana population differed significantly from all the others (p_Zigana-Vauk = 0.008, p_{Zigana-Pöske} < 0.001, p_{Zigana-Pekün} = 0.003, p_Zigana-Krom < 0.001). Among the others, only the Pöske and Pekün populations differed significantly (p = 0.002).

Fig. 7. 

Principal component analysis (PCA) plot based on the variables proximal-distal shortest length, distance to the distal and proximal edges, and number of sound teeth of male stridulatory organs of Isophya autumnalis (A). Boxplot of the measurements among different populations (B). Different letters on top of boxplots indicate significant differences. Those with the same letters do not have a statistically significant difference from each other.

PCA was also conducted for 12 variables measuring the temporal features of the male calling songs. These included five temporal variables and seven impulse-related variables during crescendos and decrescendos, and second syllable impulse counts. The first four PCs, explaining 77.65% of the variance, all had eigenvalues greater than 1. PC1, which explained the most variance (32.29%), differed significantly across populations (ANOVA, F_4,1302 = 919.9, p < 0.001). Post-hoc analysis indicated significant differences between all populations except the Vauk and Pöske Mountain populations (p = 0.396) (Fig. 8).

Fig. 8. 

Principal component analysis (PCA) plot based on temporal characteristics of the male calling song of Isophya autumnalis (A). Boxplots of the measurements among different populations (B). Different letters on top of boxplots indicate significant differences. Those with the same letters do not have a statistically significant difference from each other.

The variables contributing the most to PC1 were number of impulses and crescendos in the first syllable (13.04% and 17.64%, respectively), the duration of first syllable (11.63%), and the duration between the two syllables (15.52%). The variables contributing most to PC2 were total duration of the sound (23.66%), duration of the first syllable (11.06%), duration between the two syllables (8.56%), duration of song period (12.20%), and number of impulses in the first syllable (13.04%).

Cuticular hydrocarbon analysis.—The CHC profile analysis identified four types of hydrocarbon structures: alkanes, monomethyl alkanes, dimethyl alkanes, and alkadienes. The following straight-chain alkanes were the most frequently represented: hexatriacontane (C36 – 13.55%), tritriacontane (C33 – 8.18%), hentriacontane (C31 – 7.10%), nonacosane (C29 – 6.61%), tetratriacontane (C34 – 6.39%), heptacosane (C27 – 6.61%), and octacosane (C28 – 5.20%). Of the other peaks, the major components were 3,11-dimethylnonacosane (7.21%) and 2-methylhexacosane (7.10%).

GC/MS showed that the CHC profile of IA contained different number of compounds for each sex (Figs 9, 10). Except for the Zigana population (16 each), females and males in each locality had different numbers of CHC components: 18 and 19 in Krom Valley; 11 and 17 in Pekün Mountain; 14 and 19 in Pöske Mountain; 11 and 22 in Vauk Mountain. Certain alkanes were found exclusively in each sex: docosane (C22) and dotriacontane (C32) in females; 3,11-dimethyloctacosane and unidentified 26, 27, 28, and 37 carbon atom peaks in males. Furthermore, small proportions of alkene structures were detected in males (1%) but not females. The proportions of alkane structures were 71% in males and 82% in females. The proportion of branched monomethyl alkane profiles was higher in males (17%) than females (5%), whereas the proportion of branched dimethyl alkanes was the same (7%). Alkadiene structures were barely detected in either sex (1%).

Fig. 9. 

Chromatogram and peak numbering of the cuticular hydrocarbon profiles from both sexes.

Fig. 10. 

Distribution of carbon atom numbers in cuticular hydrocarbon profiles of males and females across all populations studied.

PC1, PC2, and PC3 clearly separated the local populations, while PC1 and PC2 separated individuals according to sex and interaction of sex and locality (Table 1, Fig. 11). However, this difference was observed only between the of Vauk Mountain population and other local populations. Significant differences were found in terms of the PC3 component between other locations (Table 1).

Fig. 11. 

Principal component analysis (PCA) plot of cuticular hydrocarbon (CHC) phenotypes of males and females from populations in different localities (A). Sexually dimorphic CHC differentiation in the species when all locations are included (B). The ellipses in the figure are the confidence ellipses around the group means in the principal components space of the function.

Body size.—As in many animal groups, body size (measurements/mass) is one of the most important traits directly or indirectly affecting physiological processes and adaptability in Orthoptera species (Roff 2002, Whitman 2008). Many studies have investigated the variation of body size in populations living in areas where abiotic and biotic factors are influenced and consequently changed by different environmental conditions (Whitman 2008, Eweleit and Reinhold 2014). The fragmentation of a species’ distribution range exposes local populations to varying environmental conditions. This variation can lead to changes in physiological constraints, with increased or decreased spatial and temporal habitat heterogeneity playing a potential role. For example, studies have shown a decrease in bee body size variation (heterogeneity) at higher altitudes (Classen et al. 2017). However, the interplay between these environmental and physiological factors is complex and context-dependent. The physiological constraints in high-altitude and harsh environments may filter out extreme phenotypes. Thus, the lack of significant variability in body size variables in IA populations other than the Zigana population may indicate a lack of directional selection in local populations exposed to similar altitudes and ecological conditions. Because IA cannot fly or jump long distances, they are expected to be more sensitive to specific environmental conditions and exhibit morphological, genetic, and behavioral diversity (i.e., Amara alpina (Paykull, 1790) (Carabidae ground beetle), Beckers et al. 2020). Macropterous morphs disperse over longer distances than brachypterous morphs (Poniatowski and Fartmann 2011, Nuhlíčková et al. 2023). This model is supported by the weak body size variability within local populations of IA, with the exception of the population in Zigana Mountain, which is found in habitats with north-facing slopes and typical Black Sea high mountain climatic characteristics.

An analysis of over 1,500 Orthoptera species indicated a general female-biased body size dimorphism, which is stronger in Caelifera than in Ensifera (Hochkirch and Gröning 2008). The PC1 obtained from the measured body parts of all our IA populations revealed a pattern consistent with this analysis, although no significant sex differences were found in hind femur lengths. This suggests that SSD, in terms of general body size parameters, cannot always be determined based on a single character. The larger body size of females is usually an evolved response to intraspecific competition focused on reproductive success whereby increased body size is associated with more eggs and reproductive success (Hochkirch and Gröning 2008). As with other Isophya and Poecilimon species, IA males are more mobile and therefore spend more energy searching for females who respond bioacoustically in order to mate (McCartney et al. 2008). SSD and intraspecific body size variation can also determine male reproductive behavior strategies, thereby shaping certain immune system parameters, such as the transfer of more sperm and larger spermatophores to larger females (Uma and Sevgili 2015, Sevgili 2022). In the wing-dimorphic bush-cricket Metrioptera roeselii (Hagenbach, 1822), which has mostly long- but also short-winged individuals, males are more mobile than females (Poniatowski and Fartmann 2011). While there may be other reasons for this difference, the bioacoustic environment is thought to be the most important factor (Berggren 2005).

Although the trend of our SSD results indicated that the Zigana Mountain population may be more dimorphic than the southern populations, this difference was not statistically significant. Some insect and many other animal species have been reported to have higher SSD in areas with higher humidity and rainfall (Stillwell et al. 2007). However, wider sampling is needed to determine why SSD is weak in IA and to understand its relationships with the ecological factors affecting populations in different geographical areas. It is possible that no specific life-history traits emerged because the sampled individuals of both sexes were in the same young age group in the laboratory and experienced similar feeding and growth environments.

Bioacoustics and stridulatory file.—As in most Orthoptera groups, the male calling song and stridulatory organ morphology of Isophya species have species-specific characteristics (Zhantiev and Dubrovin 1977, Heller 1988, Heller et al. 2004, Sevgili 2004, Zhantiev et al. 2017) that are particularly important in identifying cryptic species (Sevgili et al. 2006, 2012, Iorgu 2012, Szövényi et al. 2012, Sevgili 2018, 2020). Although the male calling song in Orthoptera has largely species-specific features, song temporal structure, sound signal transmission, and hearing are also affected by environmental conditions, such as ambient temperature and vegetation structure, and anthropogenic factors, such as noise and habitat fragmentation (Hall and Robinson 2021). Isophya species generally exhibit a fragmented habitat distribution pattern, and intraspecific variation is inevitably common given their extremely limited mobility. For example, I. modestior has a wide and fragmented distribution in the Balkan Peninsula and has variations in the duration of song, the number of stridulatory organ teeth, and the impulse of the main syllable, while populations in the northern (Pannonian) and southern (Balkan) regions form two groups (Ivković et al. 2022). Another study using two mitochondrial (COI and 12S) and two nuclear (ITS2 and H3) molecular markers, the presence of two clades identified based on acoustic data was supported (Ivković et al. 2023). However, it was shown that the Pannonian-North clade, which includes Austria, Serbia, Hungary, and Slovenia, also has subgroups.

We found that both PC1 and PC2 are determined mainly by particular temporal features of the song and the structure of the stridulatory organ (Figs 7, 8). This is important in revealing variations in secondary sexual characteristics in the partially isolated local populations of the same species. The observed environmental variations inferred from remote sensing data, including temperature, humidity, precipitation, NDVI, NDMI, and SWIR, among different localities suggest the exposure of local populations to distinct selective pressures, thereby indicating potential adaptation processes. In tettigonids, increased ambient temperature significantly differentiates features such as calling song duration, repetition rate, and number of impulses in each syllables (Arias et al. 2012, Cooper et al. 2013, Cusano et al. 2016). These studies are important in showing that some environmental factors induce variation in some parameters of calling songs and point to possible causes of differences between at least some of the IA populations. For example, Cooper et al. (2013) reported regional differences for Tettigonia viridissima (Linnaeus, 1758) in both morphology and bioacoustic parameters (e.g., minimum stridulation frequency). However, in another study, T. viridissima populations were found to exhibit relatively uniform song patterns over a wide morphological range (Grzywacz et al. 2017). It is also worth noting that this species has long wings and is highly mobile.

Cuticular hydrocarbon profiles.—Regarding variation among local populations, CHC profiles did not show significant variation in the morphological structures examined for both sexes, as well as in the stridulatory organ located of the left wing and bioacoustic characters of males.

Several studies on CHC profiles have been conducted on various insect orders, such as Coleoptera (Niogret et al. 2019), Formicidae (Hymenoptera) (Martin et al. 2008, Morrison and Witte 2011), and Diptera (Moore et al. 2021, Kula et al. 2023). The findings suggest that CHC profile differentiation can be used as a species-level chemotaxonomic tool. Although CHCs are highly stable (Drijfhout 2010), CHC chemical profiles vary significantly across geographical regions in some insect species, while in others, the differentiation is weak or absent (Suvanto et al. 2000, Moore et al. 2022). In addition, CHC profiles may vary to protect different populations of the same species from drying out or from cold/heat due to abiotic environmental variation (such as habitats having different climatic characteristics) (Veltsos et al. 2012, Kárpáti et al. 2023). CHC profiles may also vary with the age and seasonal morphs of the insect (Kárpáti et al. 2023). In our study, in which the IA individuals were of similar age, there were no significant differences in CHC profile compositions between all local populations except for one.

Although CHCs can be useful in distinguishing between species, they may not differentiate as strongly as morphological, bioacoustic, and genetic characters (Morrison and Witte 2011, Kather and Martin 2012, Veltsos et al. 2012, Dos Santos and do Nascimento 2015, Kárpáti et al. 2023). On the other hand, close relatives and sympatric species of the meadow grasshoppers Chorthippus (Glyptobothrus) biguttulus (Linnaeus, 1758) and C. (Glyptobothrus) mollis (Charpentier, 1825) show clear differences in the position of the first methyl group in major methyl-branched CHCs (Finck et al. 2016). Through an analysis of the molecular (COI gene) and CHC profiles of six Acridid (Orthoptera) species, Sofrane et al. (2022) demonstrated interspecific differences using both molecular and biochemical (weaker) data. Since pre-mating communication in Isophya species occurs through male call signals and short female responses, pheromonal communication, possibly associated with the CHC profile, may not develop strongly. In Isophya, male calling songs exhibit clear distinctions between species.

Studies on Isophya and Tettigonia bushcrickets have reported the molecular data largely matching the bioacoustic and morphological characters (but see; Grzywacz-Gibała et al. 2010, Chobanov et al. 2017, Grzywacz et al. 2017). In our study, there was a clear SSD in CHC composition in the IA individuals collected from each location. Similar findings have been reported in many insect species, such as the Australian bushcricket Kawanaphila nartee Rentz, 1993. These CHC profile differences can be used to indicate relatedness, mate quality, body size, or fertility (Hare et al. 2022). For the first three components of the CHC profiles, we found significant differences between the Vauk–Zigana mountain populations and the others. Genes regulating melanization, which provides resistance to water loss and desiccation in insects, can also affect CHC profiles (Lamb et al. 2020). For example, it has been shown in Isophya rizeensis that melanization, which exhibits an altitude/temperature dependent change, is also affected by solar radiation and vegetation height in the area (Kuyucu et al. 2018). Differences in NDVI between the studied regions do not seem to be related to differences in IA CHC profiles. Given the distinct environmental conditions on Mount Vauk, including temperature, relative humidity, and average precipitation, compared to other regions, these factors likely contribute to the notable divergence in CHC profiles observed in the Vauk population compared to other local populations, particularly regarding PC1 (Table 1). Kota et al. (2021) reported differences in both sexual dimorphism and the effect of geographic isolation on the CHC profiles of different island populations of the Pacific field cricket Teleogryllus (Teleogryllus) oceanicus (Le Guillou, 1841), although these differences were not significantly affected by local, environmental, and climatic variables. Similarly, Dupraz et al. (2022) found different CHC profiles for separate populations of the seabird tick Ixodes uriae White, 1852, and Kula et al. (2023) found Calliphora vicina (Robineau-Desvoidy, 1830) (Diptera) to have a distinct CHC profile in European (Germany and England) and Turkish populations.

Finck et al. (2016) detected 34 different peaks with 25–33 carbon atom components in the meadow grasshoppers Chorthippus biguttulus and C. mollis. On the other hand, greater diversity in chain length and carbon atom number (males: 15–38; female: 22–38) was found in IA. Hare et al. (2022) identified 23 CHC components with chain lengths ranging from 14 to 39 in Kawanaphila nartee (Orthoptera: Tettigoniidae). Studying the flesh fly Sarcophaga Meigen, 1826 from Diptera, Moore et al. (2021) reported a profile containing alkanes, methyl and di-methyl alkanes, and alkenes with chain lengths varying between C23–C33, which differed between species and sexes within each species. Overall, these results indicate a wide variation in CHC profiles both within and between sexes in the genus Isophya.

Our findings reveal a high degree of uniformity across morphological, bioacoustic, and CHC profiles in the studied populations of IA. However, noteworthy intraspecific variations were observed in certain localities, particularly regarding male stridulatory organ and bioacoustic parameters. Notably, the Vauk population exhibited unique CHC profiles compared to the other populations. This variability is evident across different geographic regions, although it did not consistently follow the same pattern for all studied characteristics.

Furthermore, our study provides preliminary insights into intraspecific sexual dimorphism and sexual selection in IA. However, further research is necessary to fully understand these variations across different IA populations.

Thus, this study contributes to a deeper understanding of intraspecific variability within IA. As the habitats of narrow-range endemic species such as IA face threats from human impacts, such as fragmentation, degradation, and even elimination, understanding the variations between local populations becomes crucial for the conservation of these ecologically diverse endemic populations in Anatolia.

Conceived and designed the field study and experiments: EKÖ, HS. Analyzed data and wrote the paper: HS. Conceived of the methodology of GC/MS to analyze CHC: EB.

Permission for the collection of the bush-crickets from the field was obtained from the Directorate of Nature Conservation and National Parks of Türkiye.

The authors declare no competing interests.

We are thankful to TUBITAK (The Scientific and Technological Research Council of Türkiye, Project Number: 117Z068) and Ordu University, TR (ODUBAP, No: B-1807) for their financial support. The authors would like to thank Hülya Önal Özdemir, Elif Açikel, and Bekir Gökçen Mazi for assistance with lab work.