The Scent of Deception: Does the Hornet Clearwing Sesia apiformis Chemically Mimic Vespa crabro? Authors: Marta Skowron Volponi, Claudia Bruschini, Leonardo Dapporto, Francesca R Dani Published: 2026 ABSTRACT AbstractAvoiding attack requires harmless species to escape, prevent recognition as prey, or remain undetected. Clearwing moths display hymenopteran mimicry that may involve multiple sensory modalities—from visual and locomotor to acoustic cues. Here, we tested whether the morphological and acoustical resemblance of the hornet clearwing Sesia apiformis to the hornet Vespa crabro also extends to chemical traits by comparing their cuticular hydrocarbon (CHC) profiles. Compositional and quantitative analyses revealed no evidence of chemical mimicry between S. apiformis and V. crabro; instead, the moth exhibited an approximately sixfold reduction in total CHC abundance in comparison to the hornet and a complete absence of alkenes, compounds known to function as key recognition cues in insect predator–prey interactions. These results suggest that S. apiformis does not converge on the chemical profile of its model and are consistent with the possibility that its reduced CHC profile contributes to lowered detectability by predators. However, the ecological significance of our findings remains to be tested through behavioural assays with potential predators. We discuss the likely predator guild of S. apiformis and the selective pressures these predators may exert on chemical detectability. Comparative analyses across closely related lepidopteran families (Cossidae, Zygaenidae, Heterogynidae, and Castniidae) revealed CHC profiles that differed markedly in both composition and abundance from that of S. apiformis. Together, these findings indicate that the reduced and simplified CHC profile in S. apiformis is unusual among related taxa and may represent a mechanism contributing to reduced predator recognition, highlighting the potential importance of signal suppression, alongside signal elaboration in predator–prey communication.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s10886-026-01759-7.Keywords: Animal communication, Chemical insignificance, Cuticular hydrocarbons, Predator-prey interactions, Recognition cues, Signal suppression Introduction The evolutionary push and pull between predators and prey has led to the development of sophisticated, multimodal mimicry systems. In Batesian mimicry, the harmless mimic gains fitness as a result of a potential predator (i.e., the signal receiver) identifying it as an unpalatable species (the mimicry model) (Bates 1862; Dettner and Liepert 1994). Mimicry may include not only visual traits, but also behaviours (e.g., Rota and Wagner 2006), sounds (e.g., Gough et al. 2024), or even odours (e.g., Mitchell et al. 2017). Imitating multiple sensory signals may provide increased protection from various predators with diverse means of recognizing prey, for example, birds are mostly visual predators (Caves et al. 2024), whereas lizards and snakes depend heavily on chemoreception (Schwenk 1995). Clearwing moths (Lepidoptera: Sesiidae) are striking hymenopteran mimics that look nothing like regular lepidopterans: they lack most of the scales from their wings, making them at least partially transparent. The bodies of sesiids feature alternating patterns of dark or metallic and bright stripes of yellow, orange, red, or white, creating a sharp contrast (Gorbunov 2023; Špatenka et al. 1999; Sáfián and Pühringer 2016; Skowron Volponi 2019). Additionally, their antennae are simple and often clavate, further enhancing their mimicry of wasps or bees. It has recently been demonstrated that sesiids also imitate the flight trajectories of their model hymenopterans (Skowron Volponi et al. 2018) and even produce buzzing sounds resembling those of wasps or bees (Skowron Volponi et al. 2021, 2025). For example, the species Malayomelittia pahangensis Skowron, 2015, a mimic of the bee Tetragonilla collina Smith, 1857, exhibits at least three sensory levels of mimicry: morphological, locomotor, and acoustical (Skowron et al. 2015). Complex, multisensory mimicry may be widely distributed across this poorly studied family. An entirely unexplored area of research on clearwing moth mimicry is the use of chemical cues, which are crucial for the survival and interactions of many insects. Insects are covered with a lipid-based layer of waxes, mainly composed of cuticular hydrocarbons (CHCs). The primary role of these compounds is to protect against desiccation, mechanical injuries, and infections (Lorenzi and d’Ettorre 2020). However, insect species have developed different CHC signatures which vary with gender, age, population, kin and, in social insects, also among colonies. For this reason, they have also become crucial for communication. These unique chemical “signatures” enable intra- and interspecific communication by conveying information that allows individuals to identify potential mates and parasitoid hosts (Goldberg et al. 2021; Millar 2010) and to distinguish friends from foes in social insects (Drijfhout et al. 2009; Bruschini et al. 2011; Leonhardt et al. 2009). There is emerging evidence that CHCs also serve a recognition role for chemically-oriented predators locating their prey (Goldberg et al. 2021; Rutledge et al. 2014; Xue et al. 2018). To avoid detection by unwanted signal receivers, some insects, mainly parasites of social species, have evolved chemical mimicry (Howard and Blomquist 2005). This survival strategy, although rare, is significant in parasitic and mutualistic relationships where mimicking the host’s profile offers survival advantages. For instance, Myrmecophilus beetles (Lenoir et al. 2013) and Rhyncomya dipterans (Schär et al. 2025) possess cuticular signatures that closely resemble those of ants and termites, respectively, allowing them to live undetected within colonies and exploit resources. Similarly, Maculinea butterfly larvae biosynthesise some CHCs in common with Myrmica ants, deceiving their hosts to the extent that they get adopted and carried into the nest, and then adsorb the missing cuticular compounds post-adoption to fine-tune their chemical mimicry, effectively compelling ants to care for them as if they were ant offspring (Akino et al. 1999). Social parasites such as Polistes atrimandibularis Zimmermann, 1930 employ this tactic, mimicking CHCs of their host species, Polistes biglumis Linnaeus, 1758, to integrate undetected into host colonies. By mimicking the CHCs, P. atrimandibularis disrupts the host colony’s chemical profile, or “nest signature,” thus complicating nestmate recognition and potentially impacting colony dynamics over time (Bruschini et al. 2011; Elia et al. 2017). Similar strategies are employed by other parasitic Polistes species, such as P. sulcifer Zimmermann, 1930 (Sledge et al. 2001) and P. semenowii Morawitz, 1889 (Lorenzi et al. 2004a, b), as well as by species of ants, which avoid detection by replicating the CHCs composition of their hosts to blend seamlessly within colonies (Dettner and Liepert 1994; Cini et al. 2011; Bagnères and Lorenzi 2010). A rare example of chemical mimicry of a non-social insect was described by Strohm and colleagues (2008): the specialized parasitic cuckoo wasp, Hedychrum rutilans Dahlbom, 1854, mimics the CHC composition of its host, the solitary European beewolf Philanthus triangulum Fabricius, 1775. To date, there is no indication that chemical integration is widespread among parasites of solitary species—for example, some cuckoo bees that lay their eggs in open cells of megachilids, do not employ any chemical strategy to deceive their hosts (Maggioni et al. 2023). In addition to chemical mimicry, in which organisms reproduce the specific cuticular hydrocarbon profiles of their hosts to evade detection, some species adopt an alternative strategy known as chemical insignificance (or chemical hiding, von Beeren et al. 2012; Day-Dell’Olio et al. 2025). Rather than imitating a specific chemical signature, these organisms reduce the expression of recognition cues to levels that presumably fall below the detection threshold of the host’s sensory system or shift their chemical profile towards compounds that are more difficult to perceive (Lorenzi and d’Ettorre 2020). This strategy can effectively render the subject chemically invisible to heterospecifics, minimizing the likelihood of eliciting predatory, aggressive or defensive responses. Chemical insignificance has been documented in a range of ant-associated insects (Lorenzi and d’Ettorre 2020; Longhurst et al. 1979) and other social parasites, and may be particularly advantageous in contexts where the complexity or variability of the host’s chemical profile makes accurate mimicry costly or unreliable. An example of this strategy occurs in the velvet ant Mutilla europaea Linnaeus, 1758, a cleptoparasite of the paper wasp Polistes biglumis Linnaeus, 1758. These velvet ants sneak into the wasps’ nests to steal resources, yet avoid attack thanks to their low overall amounts of cuticular hydrocarbons in comparison with their hosts (Uboni et al. 2012). The cuckoo bee Sphecodes molinilicornis Kirby, 1802, displays very simple chemical profiles in comparison to its hosts, Lasioglossum bees (Polidori et al. 2020). In behavioural experiments, these chemically insignificant cuckoo bees did not evoke aggressive behaviours in Lasioglossum bees. Among European clearwing moths, Sesia apiformis Clerck, 1759, or the hornet clearwing, stands out as perhaps the most spectacular example of hymenopteran mimicry (Fig. 1). Its morphology and buzzing sounds closely resemble those of its model, the hornet Vespa crabro Linnaeus, 1761, to the point that, in flight, the mimic becomes nearly indistinguishable from the real hornet (Skowron Volponi et al. 2025; Clifford 1879). Building on these findings, we asked if S. apiformis might also mimic its model at the chemical level - specifically, in its CHCs profile - potentially enhancing its protection from predators that rely on chemosensory cues. The hornet clearwing, Sesia apiformis. Female (top) and male (bottom) mating Initially, our study aimed to compare the cuticular hydrocarbons of the mimic (S. apiformis) and its model (V. crabro). However, preliminary results led us to expand our research to other lepidopterans to assess whether the sesiid’s cuticular hydrocarbons profile represents a functional adaptation associated with mimicry or whether it primarily reflects phylogenetic conservatism, being retained across closely related lineages. To address this question, we included additional lepidopteran species in our analyses, focusing on representatives of the two superfamilies most closely related to Sesioidea: Cossoidea and Zygaenoidea (Kawahara et al. 2019), for compositional and quantitative comparisons of their CHC profiles. In this study, we tested whether S. apiformis represents an exceptional case of Batesian mimicry that integrates morphological, acoustic, and chemical resemblance to its hornet model, V. crabro, while also evaluating the possibility that it exhibits chemical insignificance to reduce detectability to predators using chemical cues. Methods and Materials The cuticular lipids of S. apiformis (n = 20), a clearwing moth morphologically resembling V. crabro (n = 9), the European hornet, were analysed. Insects were killed by freezing and stored in aluminium foil at -20℃ until analysis. CHC extracts were obtained by washing a pair of wings in a 2 mL glass vial containing enough pentane to cover the wings, for 10 min. The extracts were then dried under a stream of nitrogen and re-suspended, due to differences in specimen size and expected CHC yield, in 100 µL (V. crabro) and 10 µL (lepidopterans), respectively, of a solution containing 15 ng/µL of 1-hexadecanol in heptane, used as an internal standard for coupled Gas Chromatography – Mass Spectrometry analyses (GC-MS). The wings after extraction were kept for further measurements. As reported in previous studies, the chemical profile obtained from insect wings corresponds to that of the entire body, while also preventing contamination by internal fluids (Cappa et al. 2016). Additionally, the planarity of wings allows for precise measurements of their area and thus quantitative analyses of cuticular hydrocarbons per surface unit. Two µL of the extract were injected into a Hewlett Packard (Palo Alto, CA, USA) 7820A gas chromatograph (GC) coupled to an HP 5977B mass selective detector (with a 70 eV electronic ionization source), using a 7693 autosampler (Agilent Technologies). The GC was equipped with a J&W DB-5ms fused silica capillary column (30 m × 0.25 mm x 0.10 μm, Agilent Technologies). The injector port and transfer line temperatures were set at 270°C, and the carrier gas was helium (at 15 psi head pressure). The temperature protocol was programmed as follows: from 70 to 150°C at a rate of 15°C/min (held for 3 min) and from 150°C to 310°C at a rate of 5°C/min (held for 8.667 min) to a final temperature of 325°C. Injections were performed in splitless mode (1 min purge valve off). Data acquisition and analysis were performed using the Agilent MassHunter Workstation software (version B.07.00). We used the area of the internal standard as a reference for the consistency of the injected sample volume. Chromatograms were aligned manually based on their retention times and their mass spectra. While the identification of methyl alkanes and alkenes was performed manually using both their mass spectra and retention indices (Lockey 1988; Carlson et al. 1998), the mass spectra of most other compounds were searched against electronic mass spectral libraries (NIST11; Wiley275); compounds were considered identified if both their mass spectra and their retention index matched the experimental data reported in the NIST11 library and Pherobase, or in the MACE library (Schulz and Möllerke 2025), for the epoxyalkanes reported here for Jordanita tenuicornis. The same extraction and analysis protocol used for S. apiformis was applied to selected representatives of families most closely-related to Sesiidae (diurnal, mimetic, plant borer): eight species of Zygaenidae (diurnal, aposematic, n = 16), Zygaena carniolica Scopoli, 1763, Zygaena purpuralis Brünnich, 1763, Zygaena punctum Ochsenheimer, 1808, Zygaena viciae Denis & Schiffermüller, 1775, Zygaena filipendulae Linnaeus, 1758, Zygaena transalpina Esper, 1780, Jordanita tenuicornis Zeller 1847, Adscita sp. Retzius, 1783; one Heterogyniidae (diurnal, non-mimetic, n = 1), Heterogynis penella Hübner, 1819; one Castniidae (diurnal, non-mimetic, plant borer, n = 2), Paysandisia archon Houlbert, 1912; and two Cossidae species (nocturnal, non-mimetic, plant borer, n = 5), Zuezera pyrina Linnaeus, 1761, Parahypopta caestrum Hübner, 1804. Collection details of all analysed samples are provided in Table S1. Compounds identified as common contaminants from plasticware or GC column stationary phase polymers were excluded from further analyses. We report all other detected compounds in Tables 1 and 2, S2 and S3, however statistical analyses were based on acyclic hydrocarbons only. We also excluded C21 and C22 from analyses, as these two compounds coeluted with fatty acids in most samples and it was impossible to separate the hydrocarbon peak areas from those of the contaminants. The CHC profile of each specimen was described by calculating the percentage of the peak area given by each acyclic hydrocarbon relative to the sum of the peak areas of all the considered hydrocarbons. Table 1Cuticular hydrocarbons of the hornet, Vespa crabron°AbbreviationFull nameMolecular formulaMolecular weightRIMean ± S.D. (Males)Mean ± S.D. (Females)1n-C21n-HeneicosaneC21H4429621.0077658.7±3010.30.0±0.02C23:1TricoseneC23H4632222.57100198.1±57979.3355634.7±217391.63C23:1TricoseneC23H4632222.7356764.0±4121.142711.4±85422.84n-C23n-TricosaneC23H4832423.002953292.7±203354.14114253.1±1548629.153-meC233-MethyltricosaneC24H5033823.63765843.8±99455.61615720.2±611688.66n-C24n-TetracosaneC24H5033824.00246352.2±27687.7188441.8±118884.57C25:1Pentacosene C25H5035024.7417939346.9±2064344.027614190.1±7220133.98C25:1Pentacosene C25H5035024.79412345.1±22610.1330259.9±152960.39C25:1Pentacosene C25H5035024.9185936.8±28129.81649578.8±3299157.610n-C25n-PentacosaneC25H5235225.005878989.7±418407.83345871.5±2398492.31111-meC2511-MethylpentacosaneC26H5436625.32277385.8±53244.40.0±0.0127-meC257-MethylpentacosaneC26H5436625.38100670.2±45499.10.0±0.0135-meC255-MethylpentacosaneC26H5436625.4782232.7±49837.40.0±0.0143-meC253-MethylpentacosaneC26H5436625.722024205.0±267492.62860951.1±832997.215n-C26n-HexacosaneC26H5436626.0087107.2±48793.44837722.2±1280202.816C27:1Heptacosene C27H5437826.593702117.2±131935.70.0±0.017C27:1Heptacosene C27H5437826.70106562.5±24678.50.0±0.018C27:1Heptacosene C27H5437826.8773535.2±11467.80.0±0.019n-C27n-HeptacosaneC27H5638027.00923516.1±31416.8678046.0±169951.42013-meC2713-MethylheptacosaneC28H5839427.34477865.3±85205.1111401.7±10840.9213-meC273-MethylheptacosaneC28H5839427.771218359.8±217898.31606030.2±446377.322C29:1NonacoseneC29H5840628.75324569.4±24388.9369025.4±150800.323n-C29n-NonacosaneC29H6040829.00163723.7±23862.7160338.9±64607.12413-meC2913-MethylnonacosaneC30H6242229.14352034.0±54842.234562.0±69124.025meC29Methyl-nonacosane C30H6242229.65202392.1±71998.3150483.1±106991.2263-meC293-MethylnonacosaneC30H6242229.65177017.8±47142.60.0±0.02713-meC3113-MethylhentriacontaneC32H6645031.33200456.6±42379.40.0±0.028meC31Methyl-hentriacontane C32H6645031.33102250.0±40519.60.0±0.0 Cuticular hydrocarbons of the hornet, Vespa crabro Table 2Cuticular hydrocarbons of the hornet clearwing, Sesia apiformisn°AbbreviationFull nameMolecular formulaMolecular weightRIMean ± S.D. (Males)Mean ± S.D. (Females)4n-C23n-TricosaneC23H48324233552098.2±610638.87464306.1±5221493.06n-C24n-TetracosaneC24H5033824169603.2±42837.1328464.5±125829.810n-C25n-PentacosaneC25H52352253541123.1±710852.06290579.1±1966016.115n-C26n-HexacosaneC26H5436626495640.7±87394.21202548.1±471959.419n-C27n-HeptacosaneC27H56380278328687.8±1287110.717522139.7±5228726.3an-C28n-OctacosaneC28H5839428274628.1±40047.9667514.9±175922.123n-C29n-NonacosaneC29H60408294334050.1±607076.99254259.5±2938549.9bn-C31n-HentriacontaneC31​H64​43631442522.6±96845.01568099.1±1165846.7c11,y-dimeC3711,y-DimethylheptatriacontaneC39​H80​54837.59135405.0±42499.2471981.2±498675.2d13,17,21-trimeC3713,17,21-TrimethylheptatriacontaneC40​H82​56237.821302753.1±197373.34609639.2±4744058.5euh-tri-tetra-meC37Unidentified Tri- and TetramethylheptatriacontaneC40​H82​−C41​H84​562–57637.82 – 38.13280121.8±96502.9402231.0±272626.2fuh_trimeC38Unidentified TrimethyloctatriacontaneC41​H84​57639.05403217.7±130718.2976430.8±803928.2gce-dimeC39Centrally-branched DimethylnonatriacontaneC41​H84​57840.289436312.6±76628.31666072.9±1960837.1hce_trimeC39Centrally-branched TrimethylnonatriacontaneC42​H86​59240.3854076729.1±353042.112029600.8±9781999.2i11,x,y,z-tetrameC3911,x,y,z-TetramethylnonatriacontaneC43​H88​60640.63602234.7±110486.31300643.3±850012.0j13,y-dimeC4113,y-DimethylhentetracontaneC43​H88​60443.271265967.1±329449.11513804.1±1031292.1k13,17,21-trimeC4113,17,21-TrimethylhentetracontaneC44​H90​61843.661634976.5±234756.34306690.2±2597794.6 Cuticular hydrocarbons of the hornet clearwing, Sesia apiformis Calibration curves were calculated to estimate the analyte amount by injecting solutions of linear alkanes and alkenes with the same or similar chain length as those identified in the biological samples, at known concentrations (5, 10, 20, 40, 100, 200 ng of each component) under the same conditions as described above. Since only standards of linear alkanes were available, the amount of methyl-branched alkanes was calculated by applying the curves of the linear alkane corresponding to their main chain. Alkanes and alkenes present in the biological samples, for which standards were not available, were estimated by merging data of standards with the chain length immediately shorter and longer than the identified compounds (e.g., n-C33 and n-C35, for compounds with the main chain containing 34 Carbon atoms). The calculated CHC quantities were compared between each analysed Lepidoptera family and V. crabro using a Kruskal–Wallis test, followed by Dunn’s post hoc pairwise comparisons with Bonferroni correction. Wings used for extractions were scanned and their areas measured using ImageJ software (Schneider et al. 2012). The total amount of hydrocarbons estimated for each specimen was then divided by the measured surfaces (summed dorsal and ventral areas) of the wing. Statistical analyses were performed, and plots were drawn in RStudio. The pattern of compositional similarity among GC/MS samples was assessed by converting peak areas of hydrocarbons only into relative fractions for each specimen (i.e. each chromatogram was normalized to sum to one). Pairwise Bray-Curtis dissimilarities were then calculated among specimens using the vegdist function in the R package vegan. Differences in chemical composition between S. apiformis and V. crabro were then tested using a permutational multivariate analysis of variance (PERMANOVA) based on Bray–Curtis dissimilarities (adonis2, vegan package), with statistical significance assessed using 9,999 permutations. Homogeneity of multivariate dispersions was evaluated using the betadisper function followed by a permutation test to verify that significant differences reflected shifts in group centroids rather than differences in within-group dispersion. A hierarchical cluster was constructed using the UPGMA algorithm implemented in the native R function hclust using a Bray Curtis dissimilarity matrix based on chemical composition of all the samples belonging to Vespidae and to five different Lepidopteran families. The pattern of compositional dissimilarity was also visualized by principal coordinates analysis (PCoA) using the cmdscale function in R. Results Cuticular lipids of V. crabro (Table 1; Fig. 2A, B) are represented by saturated, both non-branched and methyl-branched hydrocarbons, with chains from 21 to 31 carbons, as well as by straight monoenes with chains ranging from 23 to 29 carbons, and cholesterol, as previously reported (Butts et al. 1991; Ruther et al. 1998). As in most insects, odd-numbered chains form the majority of detected hydrocarbons. S. apiformis’ cuticular hydrocarbons (Table 2; Fig. 3A, B) are represented by saturated, unbranched chains ranging from 21 to 31 carbons (both even and odd-numbered chains, but lacking triacontane in all cases) and mono-, di-, tri- and tetramethyl methyl branched hydrocarbons from 37 to 41 carbon chains. Notably, no alkenes were detected (squalene was present in several samples, however we believe this was a contaminant from human skin, possibly transferred to the moths at the time of collection). There is a gap in the cuticular hydrocarbon profile, with no alkanes detected containing between 32 and 36 carbon atoms. Apart from hydrocarbons, we detected cholesterol, triacontanediol (possibly 1,30-triacontanediol) and two plant sterols: tocopherol and sitosterol (however, phytosterols could not be identified with certainty due to lack of data in available libraries and our lack of standards of these compounds), as well as linear aldehydes: octacosanal, triacontanal and dotricontanal (Table S2). As it often occurs in GC-MS analyses of insect cuticle extracts, fatty acids (palmitic, linoleic and oleic acid), their methyl esters and the corresponding amides were present in some samples, mostly those that had been stored for a longer time in the freezer before the wings were removed for analysis, and thus we believe these are contaminations from the hemolymph. Representative GC-MS chromatograms of S. apiformis and V. crabro are shown in Figs. 2 and 3, highlighting the lower overall abundance of CHCs in the mimic compared to its model. PERMANOVA based on Bray–Curtis dissimilarities revealed a strong difference in chemical composition between S. apiformis and V. crabro (pseudo-F = 130.22, R² = 0.828, P = 0.0001; 9,999 permutations). Multivariate dispersion did not differ between species (PERMDISP: F = 1.20, P = 0.282), indicating that the observed separation reflects differences in centroid location rather than within-group variability. Representative GC-MS chromatograms of Vespa crabro (A) V. crabro male, (B) V. crabro female. V. crabro extracts were re-suspended in 100 µL. Numbers above the peaks indicate compounds identified in V. crabro, including those shared with S. apiformis (see Tables 1 and 2) Representative GC-MS chromatograms of Sesia apiformis. (A) S. apiformis male, (B) S. apiformis female. Extracts of S. apiformis were re-suspended in 10 µL, whereas V. crabro extracts were re-suspended in 100 µL; chromatograms of the model (V.crabro) are therefore 10-fold diluted compared to the mimic (S. apiformis), which provides clear evidence of the substantially higher quantity of CHCs in V. crabro. Numbers above the peaks indicate compounds identified in V. crabro, including those shared with S. apiformis (see Tables 1 and 2). Letters above the peaks indicate compounds identified exclusively in S. apiformis (see Table 2) Compositional analysis of relative CHC profiles of lepidopteran families other than Sesiidae (Table S3, Figs. S1-S12, Fig. 4A, B) revealed distinct CHC profiles among them, with Zeuzera pyrina (Cossidae family, Fig. S1) showing the greatest similarity to S. apiformis, which is visualized by the placement of this cossid close to the sesiid in the dendrogram in Fig. 4A. Analogous to the sesiid, Z. pyrina has no alkenes. Detected hydrocarbons include non-branched alkanes with 25 to 31 carbons, but without triacontane, and methyl-branched alkanes with 39 to 41 carbons. This cossid has a small hydrocarbon amount per unit of wing area (median 0.665 µg/cm2), which falls within the range displayed by S. apiformis. Other detected compounds include cholesterol and aldehydes: hexacosanal and octacosanal. Similarly to Z. pyrina, the other cossid species analysed, Parahypopta caestrum (Fig. S2), also has a small amount of CHCs (1.35 µg/cm2) and lacks alkenes. This species has different methylated alkanes than Z. pyrina and no aldehydes, however we found campesterol and sitosterol. (A) Cluster dendrogram based on Bray-Curtis dissimilarities to calculate distances between CHC profiles of all analysed samples (hydrocarbons only). (B) Principal coordinates analysis (PCoA) showing the pattern of compositional similarity between species. (C) Quantity of acyclic hydrocarbons per unit of wing area [µg/cm2] by family. C21 and C22 were excluded from all analysed species due to peak overlap with fatty acids Representatives of the family Zygaenidae (Figs. S3-S10) are highly variable in both the composition and amount of cuticular hydrocarbons, yet they all produced both unbranched alkanes and alkenes (with up to 31 carbon atoms), as well as mono-, di- and trimethyl branched alkanes ranging from 23 to 42 carbon atoms (details in Supplementary Table S3). Aldehydes, cholesterol, sitosterol, desmosterol, lathosterol, stigmasterol as well as tocopherol (the latter only in Adscita sp.), were also detected. Noteworthy are two epoxides found in Jordanita tenuicornis: 9,10-epoxytricosane and 9,10-epoxypentacosane, which have never been reported for Lepidoptera before (Table S2). The former is a housefly sex pheromone (Dillwith et al. 1983). Median hydrocarbon amount per wing area ranges from 1.43 µg/cm2 in J. tenuicornis to 5.04 µg/cm2 in Z. carniolica, with a median value for the family of 2.33 µg/cm2. In Heterogynis penella (Zygaenoidea: Heterogynidae, Fig. S11), we detected unbranched alkanes with 25 to 30 carbon atoms, and several methyl-branched alkanes of even and odd-numbered chains in the range of 25–31 carbon atoms, including several terminally branched methyl alkanes, and just one alkene, triacontene. Other detected compounds include cholesterol, tocopherol, stigmasterol and sitosterol. Median wing hydrocarbon amount in H. penella reaches 2.35 µg/cm². Alkenes were also not detected in Paysandisia archon (Fig. S12), a known pest of palm trees, invasive in Europe. This species is, however, exceptionally rich in mono- and dimethyl branched long-chain alkanes (with 39 to 42 carbons). We also detected 2,13-octadecadien-1-ol, the P. archon male sex pheromone, 1,30-triacontanediol, cholesterol, sitosterol and some unidentified acetates. This castniid moth has a high median amount of cuticular hydrocarbons (7.11 µg/cm2), falling within the range of V. crabro (as visualized in the boxplot in Fig. 4C) and greatly exceeding that of S. apiformis. A thorough evaluation of potential intraspecific variability would require GC–MS analyses of a larger number of specimens. Nevertheless, we chose to maximize species representation rather than specimen replication within species, as we considered this strategy more relevant to the objectives of the study and better aligned with the hypotheses under investigation. Quantitative analyses (Fig. 4C) revealed significant differences between the six analysed families in hydrocarbon amount per unit of wing area, as indicated by a Kruskal-Wallis test (χ²(5) = 34.63, p < 0.001). Dunn’s post-hoc test with the Bonferroni correction (Table S4) indicated that Sesia apiformis had significantly lower CHC amounts than its model Vespa crabro (Z = -5.15, adjusted p < 0.001, Fig. 4C): the median CHC load of V. crabro (5.87 µg/cm2) was approximately 6.4 times higher than that of S. apiformis (0.916 µg/cm2). Similarly, the family Cossidae has significantly less CHCs than Vespidae (Z = -3.59, adjusted p < 0.01). The families Zygaenidae (Z = 2.65, adjusted p > 0.05), Heterogynidae (Z = -1.00, adjusted p = 1) and Castniidae (Z = 0.26, adjusted p = 1) did not differ significantly from Vespidae in CHC quantity, although Zygaenidae exhibited lower median hydrocarbon levels (2.32 µg/cm2). Among the Lepidoptera examined, hydrocarbon quantities varied considerably, with Paysandisia archon, Zygaena carniolica, and Z. viciae reaching levels comparable to those of the hornet. In contrast, the remaining Zygaenidae species generally fell within or only slightly above the range observed for S. apiformis. Nevertheless, the median hydrocarbon quantity across all other Lepidoptera except Cossidae remained notably higher than that of the sesiid but lower than that of the vespid. Discussion The cuticular lipid profile of the hornet clearwing (Sesia apiformis, mimic) is markedly distinct from that of its model, the hornet (Vespa crabro), both qualitatively and quantitatively, indicating that chemical mimicry does not occur in this system. Why does S. apiformis resemble hornets morphologically and acoustically, but not chemically? There are at least two possible explanations. Firstly, selection for mimicry in S. apiformis may have been driven by visually or acoustically (or both) directed predators but not by one that relies on chemical cues. Secondly, the evolution of a finely-tuned hornet chemical profile is extremely improbable, if not completely impossible, for a lepidopteran that does not have close interactions with its model. In favour of the first hypothesis is a study (Newland and Sawyer 2016) demonstrating that birds, which rely primarily on vision and hearing, are in fact predators of S. apiformis. Newland & Sawyer observed heavy predation on the hornet moth by great tits and blue tits in a single year and reported that, during that year, vespid wasps were notably absent from the study area. This absence may have meant that local birds had no previous unpleasant encounters with true hornets, reducing the effectiveness of S. apiformis’ mimicry. In a recent study (Skowron Volponi et al. 2025), wild European robins were deterred to a similar extent by hornets and hornet moths, indicating that morphological and acoustical mimicry displayed by S. apiformis is efficient in misleading these avian predators, even though there is no chemical resemblance between mimic and model. Supporting the second hypothesis is the fact that chemical signatures of social wasps are often highly complex, consisting of dozens of compounds in strictly defined proportions, enabling not only intraspecific recognition but also accurate nestmate recognition (Lorenzi et al. 2004a, b). The evolution of perfect mimicry of a model’s CHC profile is unlikely due to this trait’s multicomponent nature. Close chemical resemblance usually occurs in mimics that are able to acquire compounds either directly from their host or from the nest, as in Maculinea larvae (Akino et al. 1999). In some cases, however, common foraging habits of model and mimic seem to have been enough for chemical mimicry to evolve, as is the case of the wasp-resembling Cerambycidae beetles Megacyllene caryae Gahan, 1908, M. robiniae Forster, 1771 and Callisphyris apicicornis Fairmaire & Germain, 1859 which have been demonstrated to produce blends of spiroacetals similar to those released by wasps as alarm pheromones (Mitchell et al. 2017). These are however, volatile compounds and producing some chemicals in common with wasps is perhaps more probable than developing a complete CHC profile resemblance, especially since spiroacetals are quite widespread semiochemicals in several orders of insects, and they may be involved in intraspecific communication, including sexual signalling (Booth et al. 2006). One may speculate that, since mimicking the complex CHC profile of a model species may have strong evolutionary constraints, it may be more likely to evolve similarities on other channels—namely, visual and acoustic—while reducing both the quantity and variability of one’s CHC profile. If S. apiformis looked and sounded like a hornet but retained an appealing, prey-like odour, the risk of predation would remain high, depending on which sensory modalities a given predator relies on most. Conversely, if it looks and sounds like a hornet but emits no detectable odour, then no additional attractive cues remain, and potential predators may simply ignore it. Our quantitative CHC analyses revealed that S. apiformis carries less than one-sixth the amount of CHCs per unit wing area compared with V. crabro. Their CHC profile lacks alkenes and is relatively rich in long methyl-branched alkanes. Alkenes have been experimentally shown to serve as important recognition cues, eliciting aggression in honeybee workers when applied to nestmates (Dani et al. 2005). Their absence in Sesiidae may therefore reduce the risk of triggering aggressive responses. Linear compounds are less likely to engage in specific binding with receptors, and heavy, less volatile compounds, such as those present in S. apiformis, may be more difficult to detect (Dani et al. 2005). This strongly suggests that the hornet clearwing is chemically insignificant, yet direct behavioural assays (e.g. aggression or detection trials with V. crabro and chemosensory predators) would be required to confirm whether low CHC quantities in Sesiidae reduce their detectability or aggression. In sensory communication, the effective range of a signal is crucial: vision and hearing operate over larger distances than chemoreception, particularly for non-volatile compounds. Accordingly, visual and acoustic mimicry of hymenopterans in clearwing moths may have evolved under selective pressures distinct from those shaping their chemical traits. The former is most probably an anti-avian strategy (Skowron Volponi et al. 2025), given that birds rely heavily on visual and acoustical cues (Martin 2017), although olfaction also plays a significant and often underestimated role in avian foraging behaviour (Abankwah et al. 2020). Lizards and snakes, on the other hand, sample their surroundings with their tongues and can detect non-volatile compounds through their vomeronasal system (Schwenk 1995). Lizards have been shown to recognize their prey based on cuticular hydrocarbons (Goldberg et al. 2021). Chemical insignificance in S. apiformis could therefore reduce detection by reptilian predators. To date, no instances of lizard predation on clearwing moths have been reported. However, it is worth noting that during our studies of S. apiformis in their natural habitat, we repeatedly saw lizards basking on the sunlit trunks of the same poplar trees from which adult hornet moths emerged. Females of S. apiformis often remain almost motionless on these trunks for hours, while releasing sex pheromones. Although we did not witness direct interactions with lizards, such encounters are plausible. Whether having a chemically insignificant CHC profile enhances hornet moths’ chances of survival in case of potential attacks from chemically-oriented predators such as lizards, remains to be verified. Another possibility is that clearwing moths fall below the detection threshold of the hymenopteran models they resemble. As shown by Cini et al. (2009) on Polistes wasps, vespids react aggressively to lures with extracts of foreign female wasps, but when the amount of CHCs is reduced to one-third of an intruder’s concentration, aggressive responses are greatly reduced - to the point where they do not differ from control, blank lures. Since S. apiformis carries only about one-sixth of the total CHC quantity found in hornets, it is plausible that it would not elicit aggressive responses from insect predators, including V. crabro—although this remains to be verified through behavioural trials. Could hymenopterans have been the selective force driving the evolution of chemical insignificance in clearwing moths? While this hypothesis is appealing, it is difficult to test, as there is no evidence that S. apiformis interacts directly with hornets. The imagos possess reduced proboscides and do not feed, ruling out trophic competition with vespiids. Nevertheless, given that hornets are ferocious predators, the sesiid’s chemical invisibility toward their own visual model species could confer a significant advantage to S. apiformis. The idea would be even more interesting to explore in clearwing moths that do share foraging grounds with their hymenopteran models. We additionally placed the hornet clearwing within a wider phylogenetic context with the aim of disentangling adaptive signals of chemical insignificance from background patterns of CHC evolution in Lepidoptera. The sesiid’s CHC profile differed qualitatively from representatives of the most closely related families (Zygaenidae, Castniidae and Heterogyniidae), indicating that the CHC pattern displayed by Sesiidae is probably not a conserved trait shaped by phylogenetic history. However, the cossid moth Zeuzera pyrina displayed a low quantity of CHCs in comparison to other species, and fell within the range displayed by S. apiformis. Z. pyrina’s overall CHC pattern was also the closest to S. apiformis out of all considered species. This may be associated with these moths’ similar biology - both of these species’ larvae are plant borers, whereas the imagos do not feed. S. apiformis larvae exploit firstly the phloem of poplar trees and then bore deeper towards the sapwood (Špatenka et al. 1999). Z. pyrina larvae feed on all layers of live wood under the bark and may also bore into hardwood (Waring et al. 2018). One may argue that xylophages, having a poorer diet than folivores, have limited access to metabolic precursors, which in turn may constrain both the quantity and diversity of their cuticular hydrocarbons. The studied castniid moth, P. archon, is also a plant borer, yet, in contrast to S. apiformis and Z. pyrina, it displays a rich CHC profile (both quantitatively and qualitatively), but it is worth noting that the larva forages not only inside the trunk of the palm, but also in leaves and fruit. Our study provides novel insight into the surface chemical profiles of representatives of the lepidopteran families Sesiidae, Cossidae, Zygaenidae, Castniidae and Heterogyniidae. The relative roles of predation pressure and phylogeny in shaping the composition and abundance of lepidopteran cuticular lipids require further investigation. Whether the CHC profile of S. apiformis is representative of the family Sesiidae, and what the biological meaning of the displayed low quantity of CHCs is, remain open questions. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Figures S1-S12 (PDF 1.60 MB) Supplementary Table S1 (XLSX 12.2 KB) Supplementary Table S2 (XLSX 34.1 KB) Supplementary Table S3 (XLSX 34.1 KB) Supplementary Table S4 (XLSX 10.1 KB) Acknowledgements We are deeply grateful to Dr Giuseppe Pieraccini, Director of the Mass Spectrometry Center (CISM) in Florence, for his invaluable advice and assistance during the GC–MS analyses. We thank Paolo Volponi for his dedicated assistance in the field and for capturing numerous hornets. We also thank Dr Federico Cappa for providing male specimens of Vespa crabro, Marek Hołowiński for specimens of Sesia apiformis from Poland, and Leo Serra for specimens of Paysandisia archon. We are further indebted to Marco Bonifacino for identifying part of the Zygaenidae specimens used in this study and his help in locating S. apiformis in Abruzzo. Author contributions Conceptualization MSV, LD & FRD; investigation MSV & CB; methodology MSV, LD & FRD; analysis and interpretation of data MSV, CB, LD & FRD; visualization MSV, CB & LD; writing – original draft MSV & CB; writing – review and editing MSV, CB, LD & FRD. All authors contributed critically to the drafts and gave final approval for publication. Funding This study was carried out as part of M.S.V.’s postdoctoral scholarship no. BPN/BEK/2021/1/00318 received from the Polish National Agency for Academic Exchange Bekker Programme. F.R.D. and L.D. acknowledge the support of the NBFC to the University of Florence, funded by the Italian Ministry of University and Research, PNRR, Missione 4 Componente 2, “Dalla ricerca all’impresa”, Investimento 1.4, Project CN00000033, spoke 3 and spoke 7. Data availability All data supporting the findings of this study are available within the paper and its Supplementary Information. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Figures S1-S12 (PDF 1.60 MB) Supplementary Table S1 (XLSX 12.2 KB) Supplementary Table S2 (XLSX 34.1 KB) Supplementary Table S3 (XLSX 34.1 KB) Supplementary Table S4 (XLSX 10.1 KB) Data Availability Statement All data supporting the findings of this study are available within the paper and its Supplementary Information.