JHR 98: 1087-1106 (2025) Bp Sy JOURNAL OR She tentonicatiante doi: 10.3897/jhr.98.170415 RESEARCH ARTICLE () I Tymenopter a 9 https://jhr.pensoft.net The Imarasional Society of Hymenopeeriss, RESEARCH Vibroacoustic signalling of the Maltese Messor ants (Hymenoptera, Formicidae), with geographical insights into M. ibericus Albena Lapeva-Gjonova', Ilia Gjonov', Rumyana Kostova', David Mifsud* I Faculty of Biology, Sofia University, 8 Dragan Tsankov str, 1164 Sofia, Bulgaria 2 Institute of Earth Systems, Division of Rural Sciences and Food Systems, University of Malta, Msida MSD 2080, Malta Corresponding author: Albena Lapeva-Gjonova (gjonova@gmail.com) Academiceditor: Francisco Hita Garcia | Received 28 August 2025 | Accepted 24 October2025 | Published 17 November2025 https./zoobank. org/B3C7233E-6A66-4D 1 E-BC1B-954DAF98404D Citation: Lapeva-Gjonova A, Gjonov I, Kostova R, Mifsud D (2025) Vibroacoustic signalling of the Maltese Messor ants (Hymenoptera, Formicidae), with geographical insights into M. ibericus. Journal of Hymenoptera Research 98: 1087-1106. https://doi.org/10.3897/jhr.98.170415 Abstract This study presents vibroacoustic recordings of the three Messor species in the Maltese archipelago, M. cap- itatus (Latreille, 1798), M. bouvieri Bondroit, 1918, and M. ibericus Santschi, 1925, and their signal characteristics were analysed. The observed patterns are consistent with the hypothesis that acoustic traits in Messor ants are subject to species-specific constraints. To gain further insight into the sound parameters of M. ibericus, geographically isolated populations from Malta, Gozo, Sicily, Crete, and Bulgaria were compared, revealing variation and significant differences at the population level. Additionally, the study provides new occurrence data for M. ibericus (queens and workers), which was previously reported as M. structor (Latreille, 1798) for Malta, along with information on its habitat preferences and activity. Keywords Biogeography, harvester ants, Mediterranean, vibroacoustics Introduction Over the last decades, increasing information has accumulated on the unexpectedly wide distribution of hidden taxonomic diversity in European ants (Seifert 2009; Csdsz et al. 2014; Seifert 2018; Steiner et al. 2018). In this context, one of the emblematic examples in this line are the ants of the genus Messor Forel, 1890, known as harvester Copyright Albena Lapeva-Gjonova et al.. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 1088 Albena Lapeva-Gjonova et al. / Journal of Hymenoptera Research 98: 1087-1106 (2025) ants, in which intraspecific variability, polymorphism, cases of hybridisation and even xenoparity are well documented (Schlick-Steiner et al. 2006; Steiner et al. 2011, 2018; Orou et al. 2023; Juvé et al. 2025a, 2025b). The application of complementary analyses from the genetics, cuticular hydrocar- bons, ethology and ecology to standard comparative morphology and morphometrics in ants, known as integrative taxonomy, is the most convincing approach to reveal true biodiversity (Seifert 2009). In recent years, acoustic communication studies in a number of ants have shown that they can often be species-specific and has been suc- cessfully used to detect and confirm new species, especially those that are morphologi- cally indistinguishable from other closely related species (Pavan et al. 1997; Ferreira et al. 2010; Pefa Carrillo 2021). Species of Messor are known to possess stridulatory organs that produce vibrational signals by rubbing the tergal portion of the postpetiolus, known as the scraper (or plec- trum), against a finely striated region called the file (or pars stridens), located dorsally to the first metasomal segment (Grasso et al. 1998; Hernandez et al. 2002). However, only in a few species stridulation fields (SF) have been studied and documented by scanning electron microscopy. Even more limited are available records of the acoustic signals produced and their role in behaviour (e.g. Hahn and Maschwitz 1985; Priano et al. 1997; Grasso et al. 2009). Vibroacoustic signals have been analysed in a few species of the genus Messor, including M. barbarus (Linnaeus, 1767), M. capitatus (La- treille, 1798), MZ. minor (André, 1883), M. structor (Latreille, 1798), and M. wasmanni Krausse, 1910 (Grasso et al. 1999, 2000; Gjonov et al. 2025). The majority of species in the genus Messor occur in the Palaearctic, with the Medi- terranean subregion having high species diversity (Borowiec 2014). The historical ac- cumulation of many synonyms and species names makes the genus taxonomically challenging, and recent taxonomic revisions are available for two of the species groups, M. structor and M. semirufus (Steiner et al. 2018; Salata et al. 2023). To date, three species of the genus Messor have been documented from the Maltese archipelago, recorded from a considerable number of localities. These include Messor bouvieri Bondroit, 1918, Messor capitatus, and M. structor. A fourth species of the ge- nus from the island of Comino was reported by Baroni Urbani (1968) and Schembri and Collingwood (1981) and identified as M. meridionalis ssp. wasmanni Krausse and M. caducus (Victor, 1839). In the Mediterranean Region, both taxa are now recognised as M. wasmanni. However, the presence of this species in Malta raise questions about its natural occurrence, particularly given the absence of this species from the islands of Malta, Gozo, and Sicily (Baroni Urbani 1968). This study presents vibroacoustic recordings of all three Messor species typical of the Maltese archipelago, M. capitatus, M. bouvieri, and M. ibericus Santschi, 1925, the latter previously reported as M. structor for this region. In order to gain further insight into the sound parameters of workers of the species M. ibericus, a comparison between geographically isolated populations from Malta, Gozo, Sicily, Crete and Bulgaria was conducted. In addition, new occurrence data of M. ibericus from Malta and Gozo are provided. Notably, a recent discovery indicates that M. ibericus workers are hybrids Vibroacoustic signalling in Maltese Messor ants 1089 derived from queens of this species and males of M. structor (either wild-type or clonal lineages), even though /M. structor itself may be absent from the area, as has been dem- onstrated for Sicily Juvé et al. 2025a). Materials and methods Samples collection Specimens of Messor species were collected by hand throughout the islands of Malta and Gozo to assess species occurrence within the archipelago and were stored in 96% ethanol for morphological examination. They were sampled by the first author in 2017 (November), 2024 (April) and 2025 (March). Identification of the ants follow Schem- bri and Collingwood (1981) and Seifert (2018). Specimens used for vibroacoustic recordings were collected in Malta (April 2024), Crete (September 2023), Bulgaria (May 2024), and Sicily (May 2025). They were kept alive until processing, which was carried out indoors. The studied material is stored in the Zoological collection of Sofia University (BFUS). Data on the occurrence by species and nesting sam- ples of the specimens used for sound recordings are provided below, and additional details on the presence of M. ibericus in Malta are provided in Suppl. material 1. Considering the genetic complexity observed in M. ibericus (including the hybrid origin of workers) by Juvé et al. (2025a), we use this name for the worker caste as well, for consistency. Although we acknowledge that the worker caste represents hybrid individuals (MZ. ibericus x M. structor), our study is not genomic or taxonomic in scope, and therefore the use of the established species name ensures comparability with previous ecological and morphological research. Messor ibericus Santschi, 1925 Fig» 1A, B Sound recordings. Marta — Malta isl. * Gnejna Bay; 35°55.224'N, 14°20.640'E; alt. 9 m; 18 Apr. 2024; sound records: R264, R265, R266 (major workers); — Gozo isl. ¢ Ta’ Pinu; 36°03.630'N, 14°12.546'E; alt. 127 m; 20 Apr. 2024; sound records: R285, R288, R289 (major workers), R287 (queen); Butcaria * Plovdiv; 42°8.820'N, 24°45.096'E; alt. 205 m; 27 May 2024; sound records: R314, R315, R316 (major workers); GREECE ¢ Crete; Thronos; 35°16.188'N, 24°37.086'E; alt. 372 m; 19 Sep. 2023, sound records: R161, R164, R166 (major workers); [tary ¢ Sicily; Vizzini; 37°8.572'N, 14°44.682'E; alt. 395 m; 10 May 2025; sound records: RO1, RO2, ROG (major workers). Messor bouvieri Bondroit, 1918 Fig. 1C Sound recordings. Matta — Malta isl. * Ghadira; 35°58.157'N, 14°20.960'E; 4 m; 19 Apr. 2024; sound records: R276, R278, R279 (major workers). 1090 Albena Lapeva-Gjonova et al. / Journal of Hymenoptera Research 98: 1087-1106 (2025) Figure |. Habitus of Maltese Messor ants A Messor ibericus (worker) B M. ibericus (queen) C M. bouvieri (worker) D M. capitatus (worker). Scale bars: 1 mm. Messor capitatus (Latreille, 1798) Fig. 1D Sound recordings. Marta — Malta isl. * Il-Kalanka; 35°49.483'N, 14°33.388'E, 18 m; 16 Apr. 2024; sound records: R242, R243, R248 (major workers). Photographic images Images of the specimens were taken with Nikon DS-Ri2 through Nikon SMZ 1270i stereomicroscope and then aligned and stacked using Helicon Focus 8.3.6. Pro. Stridu- lation fields were imaged by scanning electron microscope JSM-5510 at 10 kW. Sound recording and analysis The terminology in Grasso et al. (2000) was used for signal parameters. The recorded signal (presumed disturbing) consists of trains of chirps, each composed of two parts (forward and backward). The maximum frequency was measured not across the entire chirp, but separately for each part. Since the experimental setup did not allow for cali- brated measurement of absolute amplitude values, the ratio between measured ampli- tudes under constant conditions was used for the purposes of this study. Vibroacoustic signalling in Maltese Messor ants 1091 The following signal parameters were selected for comparing acoustic signals: chirp rate (chirps/s), chirp duration (ms), forward chirp duration (ms), backward chirp dura- tion (ms), maximum frequency of the forward chirp (Hz), maximum frequency of the backward chirp (Hz), and the ratio of the amplitude at maximum frequency between the forward and backward parts of the chirp. Signals were recorded using a DIY piezoelectric contact microphone, following the methodology described by Gjonov et al. (2025), and a Zoom F3 digital recorder. GNU Audacity software was used to process and measure the parameters. Noise in the record- ings was suppressed by 24 dB, using a noise sample taken from parts of the recording where no signal was present. After confirming that no recorded signals were within this frequency range, a high-pass filter at 100 Hz with a roll-off of 24 dB per octave was also applied. For each recording, a section containing a relatively long chirp train was selected. Its duration was measured in seconds, and the chirps were counted to obtain the chirp rate value. The other parameters were measured from three chirps within this section. For each of the three Messor species, across the major worker caste and differ- ent locations, recordings from three specimens were analysed, except for the queen of M. ibericus, for which only a single specimen was available; its measurements are presented but were not included in the analysis. Selected recordings are included in the Suppl. materials 2-5. Sample oscillograms and spectrograms of trains and chirps were prepared for each species, caste, and location. Statistics Principal Component Analysis (PCA) was used to visualise the variance among the samples and to reduce the dimensionality of the dataset. Variables with correlations to the axes greater than r = 0.2 were retained in the analysis. Since most of the measured acoustic variables did not follow a normal distribution, a non-parametric Permutational Multivariate Analysis of Variance (PERMANOVA) was employed, incorporating a Monte Carlo permutation procedure appropriate for small sample sizes. Two sets of PERMANOVAs were conducted to assess both over- all and pairwise differences in acoustic characteristics: (1) between Messor ant species found in Maltese Islands, and (2) among five geographically distant populations of Messor ibericus (Malta and Gozo, Sicily, Crete, Bulgaria). All data were z-score standardised to account for differences in scale and units. Analyses were based on an Euclidean distance matrix, with an unrestricted permutation of raw data— suitable for single-factor designs—as the permutation method. A total of 9999 permuta- tions were performed. To test for differences in variation around group centroids, a homo- geneity of multivariate dispersions test (PERMDISP) was applied, as variation in dispersion can influence PERMANOVA results. All groups showed significant differences (p < 0.01). Nevertheless, as PERMANOVA is robust to heterogeneity in balanced designs (i.e., equal sample sizes), the results were interpreted with caution (Anderson and Walsh 2013). To identify the most influential variables distinguishing the groups and for classifi- cation purposes, a Stepwise Discriminant Analysis (SDA) was performed. At each step, 1092 Albena Lapeva-Gjonova et al. / Journal of Hymenoptera Research 98: 1087-1106 (2025) the variable that minimised the overall Wilks’ lambda—the proportion of total vari- ance in discriminant scores not explained by group differences—was included. Vari- ables were entered into the model if their P-value significance was < 0.05 and removed if > 0.10. A cross-validation procedure was also performed. The statistical analyses were conducted using PRIMER w7 (Clarke and Gorley 2015) and SPSS 17 (IBM 2007). Results Sound recordings — stridulation, oscilograms and spectograms, acoustic char- acteristics The stridulatory apparatus of Messor species is positioned on the dorsal surface of the postpetiole and the first metasomal tergite (Fig. 2A—D). The pars stridens, located in the mid-dorsal region of the first metasomal pretergite, is distinctly oval. Except M. bou- vieri, the stridulatory field (SF) is wider than long, with transverse parallel ridges that become denser posteriorly. The surrounding tegument displays an irregular strigulate texture of short transverse ridges, either convex or concave. Well-developed “pillars” 1eku OIE 264m JSM-s5 16 1eakW 1eku Figure 2. Stridulatory fields in Maltese Messor ants A, B MV. ibericus (worker) C Messor bouvieri (worker) D ™. capitatus (worker). Vibroacoustic signalling in Maltese Messor ants 1093 beneath the pretergite margin are clearly visible in SEM images of M. ibericus, while in M. capitatus the pars stridens is sharply delimited from the adjacent integument. When restrained, ants emit stridulatory signals through characteristic dorsoventral abdominal movements, produced either in brief bursts or in extended sequences. ‘The resulting vibrations are transmitted via the pincer. A single stroke of the plectrum across the stridulatory field generates a series of pulses, termed “chirps,” each arising from contact between the plectrum and an individual ridge. In Messor, the sequential passage of the plectrum produces a distinctive disyllabic chirp. Oscillograms and spectrograms of portions of trains and chirps from the vibroa- coustic signals of major workers of Messor bouvieri, M. capitatus, and M. ibericus are presented below (Figs 3-6), and audio files, after noise reduction and high-pass filtering, are also provided (see Suppl. materials 2-5). For M. ibericus, recordings were obtained 0.0 os 1.0 LS Messor ibericus Malta 0.005 A “=e” we” nsec dad fe Nada hh sa ALAAMAAAUIAMD ALLA eeteeeecghhihbbhidt MMMM iii oy eee bial Figure 3. Oscillogram and sonogram of stridulation distress signals in Messor ibericus (worker) A train B chirp. 1094 Albena Lapeva-Gjonova et al. / Journal of Hymenoptera Research 98: 1087-1106 (2025) Messor ibericus queen, Malta nS ih eat IR AT Wh sa bbl Aish ¢ a Wi Lik Pinte anna a ma (a a Ihe lh tie chi i oe | CECE TTEPET ETE EE ATTY ited a i“ FRR O RT ial Ua Figure 4. Oscillogram and sonogram of stridulation distress signals in Messor ibericus (queen) A train B chirp. from a queen collected on the island of Gozo (Fig. 4) as well as from major workers sam- pled across different populations, with material from Malta, Sicily, Crete, and Bulgaria. Acoustic characteristics of the three Messor species found in Malta Selected parameters of acoustic signals were used to characterise the three species of the genus Messor found on the Maltese Islands (Table 1). PERMANOVA results indicated significant differences in the acoustic character- istics of the three Maltese Messor species (F = 15.96, p(MC) < 0.001), with post-hoc pairwise tests confirming differences between all species pairs (Suppl. material 6: tables 1, 2). The separation was clearly evident in the PCA plot (Fig. 7), where the first two axes explained 84.7% of the total variation. SDA identified chirp duration, chirp rate, Vibroacoustic signalling in Maltese Messor ants 1095 Messor bouvieri Malta Messor bouvieri Malta wate vi ie —nerapenmanemenganeneceie Tre | Figure 5. Oscillogram and sonogram of stridulation distress signals in Messor bouvieri (worker) A train B chirp. the amplitude ratio between the forward and backward chirp and backward chirp du- ration as the most important discriminative variables. Cross-validation achieved 100% classification accuracy (Suppl. material 6: tables 3, 4). Although the sample size was relatively small and may not capture the full range of variation in population-level acoustic traits, distinct species-specific patterns were apparent. M/. bouvieri exhibited the longest chirp duration (428.22 ms), the slowest chirp rate (2.31 chirps/s), and the highest forward-to-backward chirp amplitude ratio (1.17), followed by M. capitatus and M. ibericus which demonstrated the highest chirp rate of 2.98 chirps/s (Table 1). Both M. bouvieri and M. capitatus had a longer backward chirp, whereas M. iberi- cus showed a longer forward chirp. M. capitatus had the highest frequencies in both chirp parts (forward — 3777.89 Hz and backward — 3723.44 Hz), while MZ. bouvieri and M. ibericus had lower and more similar frequencies (Table 1). The lowest overall 1096 Albena Lapeva-Gjonova et al. / Journal of Hymenoptera Research 98: 1087-1106 (2025) | Messor capitatus Malta i 7 | Figure 6. Oscillogram and sonogram of stridulation distress signals in Messor capitatus (worker) A train B chirp. variation in acoustic parameters was observed in M. bouvieri, except for chirp rate, where M. ibericus showed the lowest variability. The greatest variation in the duration of both chirp parts occurred in M. capitatus, whereas M. ibericus displayed the highest variation for all other parameters (Table 1). Acoustic characteristics of the studied Messor ibericus populations The same parameters listed in Table 1 were used to characterise the acoustic signals of Messor ibericus specimens (major workers and a queen) from geographically distant populations from Malta, Gozo, Sicily, Crete, and Bulgaria (Table 2). Acoustic parameters of M. ibericus major workers varied significantly among geographically distant populations (F = 60.76, p(MC) < 0.001), with post-hoc tests confirming differences between all pairs except the Malta-Bulgaria pair (Sup- pl. material 6: tables 5, 6). SDA identified chirp rate, backward frequency peak, Vibroacoustic signalling in Maltese Messor ants 1097 Table |. Descriptives of the studied acoustic parameters of the three Maltese Messor species. Acoustic parameter Species Sample size Range Mean Std. Deviation Std. Error Median Coeff. of (N) variation % Chirp rate (chirp/s) M. bouvieri 3 [2.09,2.48] 2.31 0.2 0.11 2.36 8.66 M. capitatus 3 [2.25,3.01] 2.73 0.42 0.24 2.93 15.38 M. ibericus 6 [2.85,3.15] 2.98 0.12 0.03 2.93 3.90 Chirp duration (ms) M. bouvieri 9 [373,488 428.22 31.44 10.48 430 7.34 M. capitatus 9 328,398 359.78 28.31 9.44 360 7.87 M. ibericus 18 [253,333 295 26.53 De 293 8.99 Forward chirp M. bouvieri 9 155,217 180.56 18.04 6.01 179 9.99 duration (ms) M. capitatus 9 109,227 150.56 50.05 16.68 130 33.24 M. ibericus 18 [151,220 180 22.08 4.82 179 12.28 Backward chirp M. bouvieri 9 [204,258 236.11 18.2 6.07 245 7.71 duration (ms) M. capitatus 9 [101,253 197.67 63.04 21.01 218 31.89 M. ibericus 18 [62,126] 96 16.92 3.69 99 17.58 Forward chirp M. bouvieri 9 3174,3284 3229 39.9 13.3 3246 1.24 frequency peak (Hz) =v. capitatus 9 [3727,3893] 3777.89 68.86 22.95 3732 1.82 M. ibericus 18 [2904,4036 3387 397.02 86.64 3283 11.72 Backward chirp M. bouvieri 9 3167,3249 3223.22 24.35 8.12 3228 0.76 frequency peak (Hz) WV. capitatus 9 [3301,3863] 3723.44 168.83 56.28 3741 4.53 M. ibericus 18 [2900,4039 3432 414.87 90.53 3601 12.09 Forward chirp M. bouvieri 9 -49.6,-44.7 -46.84 1.47 0.49 -47.1 3.13 amplitude M. capitatus 9 [-59.9,-52.3] -56.83 2.66 0.89 -57.6 4.68 M. ibericus 18 [-72,-49.6] -62.04 5.63 1.23 -62.9 14.83 Backward chirp M. bouvieri 9 [-56.1,-52.9 -54.68 0.94 0.31 -55 1.72 amplitude M. capitatus 9 [-64.6,-53.9] -59.42 3.96 1.32 -58.2 6.66 M. ibericus 18 [-75.6,-58.9 -65.45 5.44 1.19 -63.2 14.31 Amplitude forward / M. bouvieri 9 [1.13,1.22] 1.17 0.03 0.01 1.17 2.33 backward chirp ratio VY. capitatus 9 [0.99,1.13] 1.05 0.05 0.02 1.03 4.48 M. ibericus 18 [0.94,1.19] 1.06 0.07 0.01 1.05 6.62 Forward freq. peak Backward freq. peak A Backward duration ifs] 3 OD Arfiplitude ratio } } er duration © PC2 (38.5%) Wi ™. capitatus