Volume 92, Issue 12 p. 2363-2372
Open Access

Acoustic camouflage increases with body size and changes with bat echolocation frequency range in a community of nocturnally active Lepidoptera

Ralph Simon

Ralph Simon

Department of Ecological Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Behavioral Ecology and Conservation Lab, Nuremberg Zoo, Nuremberg, Germany

Machine Learning and Data Analytics Lab, Friedrich-Alexander Universität Erlangen-Nürnberg, Erlangen, Germany

Search for more papers by this author
Alrike Dreissen

Alrike Dreissen

Department of Ecological Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Search for more papers by this author
Helene Leroy

Helene Leroy

Department of Ecological Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Search for more papers by this author
Matty P. Berg

Matty P. Berg

Department of Ecological Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Groningen Institute of Evolutionary Life Sciences, University of Groningen, Groningen, The Netherlands

Search for more papers by this author
Wouter Halfwerk

Corresponding Author

Wouter Halfwerk

Department of Ecological Sciences, Vrije Universiteit, Amsterdam, The Netherlands


Wouter Halfwerk

Email: [email protected]

Search for more papers by this author
First published: 26 October 2023
Handling Editor: David Berger


  1. Body size is an important trait in predator–prey dynamics as it is often linked to detection, as well as the success of capture or escape. Larger prey, for example, often runs higher risk of detection by their predators, which imposes stronger selection on their anti-predator traits compared to smaller prey.
  2. Nocturnal Lepidoptera (moths) vary strongly in body size, which has consequences for their predation risk, as bigger moths return stronger echoes for echolocating bats. To compensate for increased predation risk, larger moths are therefore expected to have improved anti-predator defences. Moths are covered by different types of scales, which for a few species are known to absorb ultrasound, thus providing acoustic camouflage. Here, we assessed whether moths differ in their acoustic camouflage in a size-dependent way by focusing on their body scales and the different frequency ranges used by bats.
  3. We used a sonar head to measure 3D echo scans of a total of 111 moth specimens across 58 species, from eight different families of Lepidoptera. We scanned all the specimens and related their echo-acoustic target strength to various body size measurements. Next, we removed the scales covering the thorax and abdomen and scanned a subset of specimens again to assess the sound absorptive properties of these scales.
  4. Comparing intact specimens with descaled specimens, we found almost all species to absorb ultrasound, reducing detection risk on average by 8%. Furthermore, the sound absorptive capacities of body scales increased with body size suggesting that larger species benefit more from acoustic camouflage. The size-dependent effect of camouflage was in particular pronounced for the higher frequencies (above 29 kHz), with moth species belonging to large-bodied families consequently demonstrating similar target strengths compared to species from small-bodied families. Finally, we found the families to differ in frequency range that provided the largest reduction in detection risk, which may be related to differences in predation pressure and predator communities of these families.
  5. In general, our findings have important implications for predator–prey interactions across eco-evolutionary timescales and may suggest that acoustic camouflage played a role in body size evolution of nocturnally active Lepidoptera.


Evolutionary arms races, such as between predator and prey, are characterised by rapid phenotypic change, which often leads to trade-offs and novel functional traits (Abrams, 2000; Cortez & Ellner, 2010; ter Hofstede & Ratcliffe, 2016). Predators may, for example, evolve adaptations to detect and capture their prey, which in turn selects for counter-adaptations that help the prey to escape or stay undetected. Body size plays a central role in these arms races, as it is linked to many traits involved in detection, capture and escape and often co-evolves between predator and prey (Carbone et al., 2011; Jakobsen et al., 2013; Naisbit et al., 2011; Troost et al., 2008). Understanding how body size relates to functional traits is therefore crucial to assess the evolution of adaptations and counter-adaptions in a broader comparative framework.

The interaction between echolocating bats and nocturnally active moths is a classic example of an evolutionary arms race and provides great opportunities to study eco-evolutionary processes in a comparative way (Corcoran & Conner, 2016; Goerlitz et al., 2020; Ratcliffe & Nydam, 2008; Ter Hofstede et al., 2013; ter Hofstede & Ratcliffe, 2016). Bats have evolved specialised echolocation systems or biosonar, to detect, identify and close in to capture their prey, whereas moths have evolved all sorts of counter-adaptations to evade, deceive or escape their predators (Barber et al., 2015; Barber & Conner, 2007; Corcoran et al., 2009; Goerlitz et al., 2010; Rubin et al., 2018).

All nocturnal moths are faced with an important trade-off that involves the relationship between their body size and their average target strength (Barber et al., 2015; Surlykke et al., 1999). Simply put, bigger moths return stronger echoes and are therefore at higher risk of detection when compared to smaller moths. Furthermore, when given a choice, bats would prefer to chase and attack echo cues representative of larger prey (Koselj et al., 2011), showing they base optimal foraging decisions on prey size (Andreas et al., 2012; Jones, 1990). All in all, bigger moths are thus experiencing higher predation risk.

Sound absorptive scales, a recently discovered anti-predator trait thought to function as acoustic camouflage (Ntelezos et al., 2017; Shen et al., 2018), may provide nocturnal moths an ideal all-round solution to escape the target strength–size relationship. All Lepidoptera possess specialised scales on their bodies and wings, which function in thermoregulation, aerodynamics, communication and visual camouflage (Cuthill, 2019; Hegna et al., 2013; Heinrich, 1987; Kingsolver, 1983; Michalis et al., 2017; Stevens & Ruxton, 2019; Van Dyck & Wiklund, 2002). In many moths, these scales may also function in reducing their detectability to echolocating bats by absorbing or diffracting incident ultrasound (Neil et al., 2022; Shen et al., 2018; Zhang et al., 2011), in particular scales that cover their body (Neil et al., 2020b). In general, insect bodies are protected by a cuticle that is highly reflective to ultrasound, making them an easy, omnidirectional target to bats (Neil et al., 2020b). Many moth bodies are, however, covered by a dense layer of hair-like piliform scales that may act as broadband sound absorbing structure in the 20–160 kHz frequency range covered by most bat echolocation calls (Neil et al., 2020b). Bigger bodies are supporting a larger and possibly thicker layer of sound absorptive scales to escape bat predation. We therefore predict an allometric relationship between body size and sound absorption across moth species that are under natural selection pressure imposed by echolocating bats.

We focused on the sound-absorptive properties of the body scales of a community of nocturnal moths belonging to eight different families. We covered a wide range of species that differed in several morphological characteristics, including the amount and type of body scales, body size and weight, as well as wing width and span. Furthermore, we covered species that were known to differ in anti-predator strategies other than acoustic camouflage, such as presence or absence of ultrasound sensitive ears or fast and erratic flight. So far, only very few species from the same clade have been tested on ultrasound absorption. Our approach thus sheds light on the importance of this trait across different families of moths and its possible role in acoustic camouflage as anti-predator strategy.


2.1 Study sites and species

Moth specimens were collected using a UV lamp and white sheet from different locations (Schiermonnikoog, Friesland, 53°29′ N, 6°10′ E; Veluwe, Gelderland, 52.1° N, 5.9° E; Meijendel, South Holland, 52°6′ N, 4°20′ E) in the Netherlands between April 2019 and April 2020. After collection, specimens were first chilled at 4°C for 2–5 h until we were back at the lab. At the lab, the moths were euthanised in a freezer at −20°C and stored in plastic tubes in the freezer until preparation. The moths were pinned with the costal margins of the forewings aligned (Figure S1a). To reduce echo from the needle on which the moth was pinned, the small spherical tip of the needle was cut off. The antenna and legs were removed to prevent changes in echo if they were to fall off later. The specimens were dried at room temperature on a pinning board for at least 72 h. In total, 111 specimens belonging to 58 species were prepared from the families Cossidae (N = 1), Drepanidae (N = 2), Erebidae (N = 16), Geometridae (N = 21), Lasiocampidae (N = 6), Noctuidae (N = 46), Notodontidae (N = 13) and Sphingidae (N = 7). Body weight (incl. wings, excl. antenna and legs) was measured directly after we took the moth out of the freezer and thawed them. After preparation, the following six morphological measurements were taken (Figure S1a): wingspan (from base of the wing to the tip), wing width, body width and body length by using a calliper and wing surface and body surface by calculation.

2.2 Echo measurements

The setup to measure ultrasonic echoes bouncing off from targets, here moth specimens, consisted of two main components, the biomimetic sonar head (Figure S1c) and the unit where objects could be fixed and rotated. Both were installed on guides made out of aluminium profiles (item Industrietechnik GmbH). Moth specimens were pinned on a piece of ultrasound absorbing Basotect® foam at a distance of 20 cm from the sonar head. The Basotect® was mounted on a metal rod which was fixed on the axis of a bipolar stepping motor (Nema 17, 17HS13-0404S-PG27). The stepping motor was controlled by a custom-build software program (see below) and turned the moth in 3° or 6° steps. The sonar head (representing a ‘bat head’) consisted of a centrally located loudspeaker and left and right above the loudspeaker two microphones where installed in a solid aluminium body which was fixed on a vertical guide (Figure S1c). The distance between the centre of the microphone and the centre of the loudspeaker was 25 mm. For the measurements described here, we only analysed the recordings from the right microphone, which was a ¼” free-field microphone Type 40BF in combination with the preamplifier 26AB connected to the power module 12AA (all from G.R.A.S. Sound & Vibration). As speaker, we used an ultrasonic sensor from SensComp (Series 7000). The sonar head and the stepping motor were controlled via data acquisition card (USB-6361, National Instruments) and operated by a custom-made LabVIEW program (ECHOlab v.14.0, Ralph Simon).

Each moth specimen was repeatedly ensonified with a 12-ms long, frequency-modulated signal that swept downwards from 160 to 12 kHz. For each recorded sequence, the echo of the moth (impulse response) was extracted via cross-correlation of broadcasted signal and recorded echo. To improve the signal to noise ratio, 10 consecutive impulse response measurements were averaged, for more details, see Simon et al. (2021). The moth specimens were scanned (Figure S1b) in 6° steps and in four circular planes (360°) resulting in a total of 240 measurements from different positions per specimen. A selection of moths was measured in 3° steps, but a preliminary analysis revealed no differences in overall target strength, after which all specimens were measured in steps of 6°.

To obtain spectra independent of the frequency response of the loudspeaker, an additional measurement was performed for calibration. We installed a 25 cm × 25 cm plexiglass plate in place of the moth, used a laser pointer to ensure perpendicular placement of the plate and measured its echo (Figure S1d). To obtain the moths' spectra independent of the frequency characteristics of our system, we calculated the differences between their spectra and the spectrum of the plate measurement, thereby we received the spectral target strength (TS); for more details, see Simon et al. (2021).

We obtained overall target strength for the frequency range covering echolocation calls of the bat community found in North-western Europe (19–144 kHz), as well as the spectral target strength for five different frequency bands. We chose the following bands 19–29 kHz, 29–43 kHz, 43–65 kHz, 65–96 kHz, 96–144 kHz. We analysed target strength values over all incidence angles using the overall mean in the respective frequency band as a measure of the target strength of a specimen. As we wanted to focus on the TS returned form the moth bodies, we subsetted some of our data to exclude values from angles (60–120° for one of the 4 planes, so <8% of the data) in which the wings were perpendicular positioned to the angle of sound incidence, as these returned strong echoes, but could swamp any echoes returning from the body. After our first scanning round, we first partially removed body scales on a subset of specimens using 20 small sticky tapes, in an attempt to weigh the scales. During this tape treatment, each moth specimen was waxed with 13 cm2 of masking tape divided into 20 parts. We attempted to remove even amounts of hair from the dorsal and ventral side of the thorax as well as abdomen. Each collection of tapes was weighted with a precision of 0.0001 mg before and after waxing to determine the weight of the removed hair. Due to the COVID-19 pandemic, we did not have access to the lab containing the microscale, so tapes with hairs on them were kept in the freezer for 2 months. We scanned these tape-treated specimens again and subtracted the TS value of the intact from the treated specimens to obtain a value on sound absorption (expressed as TS difference) and to relate this value to the amount of hairs removed.

Specimen were not completely descaled using the tape treatment, in particular the larger ones. We therefore continued complete descaling all specimen using a small paintbrush and repeated the scanning procedure for a third time (see Figure S1e for an intact and completely descaled specimen). All TS values were averaged (after delogarithmizing) for all incidence angles and measurement planes per specimen and frequency band for further analysis.

2.3 Data analyses

All measurements were statistically analysed in RStudio (version 1.3.959). Our six morphological measures were highly correlated and we therefore collapsed them into a single principal component (using prcomp) which was representative of body size. We lacked sufficient phylogenetic coverage of our sampled specimens to include phylogenetic distance in our analyses. Instead, we added species nested in family as random factor to our models in an attempt to correct for relatedness among our samples as best as we could.

For follow-up analyses, we used body width as a proxy for body size, as we focused on scales on the body and not the wings and because body width showed the largest size range across specimens. We constructed linear mixed effects models (LMM) using the package lme4. We added species nested in family as random effect and body size (log-transformed), frequency range (one of five consecutive bands) and their interaction as fixed effect. We first analysed the relationship between body size and overall TS, as well as the interaction between body size and frequency range. We repeated this procedure for the descaled specimens, using TS difference as dependent variable.

Finally, we calculated the impact of sound absorbing body scales on changes in detection volume, defined as the 3D area in which a prey flying in mid-air can be detected by a bat approaching from all angles. To determine the detection distance of a hypothetical bat species, we need to know the source level and frequency range of the echolocation call, frequency-dependent attenuation due to atmospherically conditions, target strength of the moth, as well as the signal to noise levels. We calculated detection distances (after Bass et al., 1995; Kinsler et al., 1999; Goerlitz, 2018) assuming a source level of 120 dB SPL (sound pressure level at 10 cm), a signal to noise detection ratio of 20 dB, an average temperature of 24°C and a humidity of 70%. We transformed detection distances to a spherical metric (4 π r) in order to obtain detection volume. Finally, we subtracted the spherical values of intact specimens from scale removed specimens to obtain changes in detection volume that could be ascribed to the removal of body scales.


We scanned a total of 111 specimen, belonging to 58 species from eight families, from 240 different angles in four circular planes around the specimens. We collapsed all morphological measurements into a single principal component (PC1) that explained 85% of the variation (Table S1). All morphological measurements had high loadings with PC1, caused by their high level of covariance. Body width showed the broadest size range (1.9–10.1 mm), the largest species being more than five times bigger than the smallest one. Body length, wing width and length differed to a lesser extent, whereas wing surface and body surface differed to a larger extent among species (Table S2). Moths of different families differed substantially in size with largest species belonging to Lasiocampidae and Sphingidae and smallest ones to Geometridae (Table S2, Figure 1). Moths differed in their overall target strength (TS) from −32.9 to −39.6 dB (covering the full frequency range of 19–144 kHz and averaged over all angles and measurement planes), although families mostly overlapped in their TS (Figure 1).

Details are in the caption following the image
Differences in target strength, body size and acoustic camouflage between Lepidoptera families. (a) Moth families differed greatly in body size, measured as body width (in cm). Note that the biggest and smallest families have overlapping target strength. (b) Overall target strength (expressed as dB values relative to an optimal reflector) of moth specimen differed between families of macromoths, with Sphingidae on average returning the strongest echoes and Geometridae the weakest. Overall target strength was calculated by ensonifying each of the 110 specimens from 240 different angles and four different measurement planes (see Figure S1). (c) Families differed greatly in their level of acoustic camouflage, with Sphingidae absorbing on average most sound in the range of 19–144 kHz and Geometridae the least. Boxplots depict median values and upper and lower 25% quartiles. Number of species per Family are Cossidae (N = 1), Drepanidae (N = 1), Erebidae (N = 3), Geometridae (N = 16), Lasiocampidae (N = 2), Noctuidae (N = 21), Notodontidae (N = 9) and Sphingidae (N = 4).

We assessed the relation between target strength across different frequency ranges and size (by focusing specifically on body width). We determined TS for five frequency bands corresponding to typical frequency ranges of different predator call types. Frequency range (FR) had a clear impact on TS, with higher frequencies returning weaker echoes (linear mixed model [LMM]; Nsamples = 555, Nspecies = 58, χ2 = 338.94, p < 0.001; Figure 2), but the effect of echolocation frequency strongly depended on the size of the target specimen (interaction body width x frequency range; χ2 = 21.02, p < 0.001, Table S3). At higher frequencies, larger moths return relatively weaker echoes when compared to their target strengths at lower frequencies and when compared to smaller moths (Figure 2). For example, at 19–29 kHz, a fivefold increase in body size results in an estimated TS increase of roughly 6 dB, whereas at 96–144 kHz, the TS increase is less than 2 dB (Figure 2).

Details are in the caption following the image
Frequency-dependent effect of body size on target strength. Bigger moths returned stronger echoes, in particular for lower frequencies. Shown are relative target strength data (TS in dB-values, on the y-axis) across a range of body widths (in cm, on the x-axis) for different frequency bands (a–e), as well as (f) average across the full range of frequencies (19–144 kHz, used by bat species in the area from which we obtained moth specimen). Note that the regression slope decreases with increasing frequencies, indicating that at higher frequencies, larger moths return relatively weaker echoes when compared to their target strengths at lower frequencies.

We focused on the sound absorptive properties of moth body scales for a subset of our specimen. We repeated the scanning procedure and first compared target strength before and after partial scale removal (expressed as TS difference in dB-values). The effect of partially removing body scales depended strongly on frequency range (LMM; Nsamples = 205, Nspecies = 29, χ2 = 48.32, p < 0.001, Table S4), with echolocation call frequency ranges below 43 kHz having no clear effect on TS difference (Figure S2). The quantity of removed body scales (assessed by weighing sticky tapes before and after the scale removal) tended to affect TS differences, but again only for the higher frequencies (interaction between scale weight and FR; χ2 = 3.76, p = 0.052; Figure S2).

We repeated the scanning procedure after removing all remaining scales with a fine brush (see echo-acoustic fingerprints of a scale removed and intact specimen in Figure 3 and Figure S1e for a close-up of the same specimen scale removed and intact). All but six specimen (belonging to the Geometridae family) returned stronger echoes after the full scale removal (median TS difference = 1.3 dB and range = −0.6–3.2 dB, see also Figure 3). Note that averaging over all measurement angles and planes led to comparatively low TS values as for many incidence angles moths return very weak echoes. The body should only under certain angles return strong echoes (e.g. when the centre line of thorax and abdomen was perpendicular to the direction of sound propagation) and under these angles' TS difference can be much higher (compare e.g. Figure 3a,c and note that sound absorption of the body can range between −10 and −15 dB for a medium-sized moth specimen).

Details are in the caption following the image
Effect of removing body scales on echo-acoustic reflectivity. (a) Echo fingerprint of an intact specimen (Colocasia coryli) scanned with its body tilted −45° in relation to on-axis impulse sound from the sonar head. A colour reference of the target strength (TS) for a reference distance of 10 cm in dB is provided on the left. (b) Comparison of intact versus scale removed specimen revealed that removing body scales increases target strength for all but six specimens. (c) Echo fingerprint of the same specimen as in (a) after scale removal. Note the high target strength (yellow colour) of echoes returning when the body is perpendicular to the sound beam, demonstrating that the body scales function in sound absorption.

We related target strength differences between the intact and scale removed specimen to body and wing size measurements as well as weight of our specimen. We found body weight (intercept = 156.6 mg, estimate = 54.7 ± 21.3 SE; χ2 = 5.89, p = 0.015), body width (intercept = −0.19 cm, estimate = 0.22 ± 0.06 SE; χ2 = 13.61, p < 0.0001), body length (intercept = −0.06 cm, estimate = 0.07 ± 0.03 SE; χ2 = 5.84, p = 0.021) to correlate positively with TS difference, whereas wing width, span and surface did not show a statistically significant relation with TS difference (all p > 0.05).

Next, we tested for the interaction between body width and frequency range, which we found to have a large effect on TS difference (LMM; interaction BW × FR; Nsamples = 280, Nspecies = 35, χ2 = 16.31, p < 0.001, Figure 4, Table S5), with larger moths showing higher TS differences due to removal of their body scales compared to smaller moths with increasing frequencies. The effect of removing body scales for larger specimen was most prominent for frequency ranges above 29 kHz and the estimated change in target strength over the full size range could be as high as 4 dB (compare e.g. slopes of Figure 4a–e).

Details are in the caption following the image
Size-dependent effect of body scales on target strength (TS). Shown is acoustic camouflage (expressed as TS difference before and after scale removal) as a function of moth body size (expressed as body width) for different echolocation frequency bands as well as across all frequencies. Positive TS difference values indicate an increase in reflectivity for shaved moth bodies and therefore an acoustic camouflage effect. The scale removal led almost always to a TS increase demonstrating acoustic camouflage to occur for all but the lowest frequency range (a). Larger moths returned stronger echoes after removal of their body scales across all frequency ranges individually (a–e) as well as for the full range (19–144 kHz range; f). All but five specimens displayed acoustic camouflage when assessing overall TS differences (f).

To understand the impact on acoustic camouflage in reducing predation detection, we calculated the changes in detection volume between scale removed and intact specimen per frequency band. The detection volume is an estimation of the volume around the moth in which the echo of the moth is salient enough to be perceived by a bat. Detection volume was on average reduced by 19 m3 (8%) due to the presence of body scales, but varied largely between species and frequency ranges (Figure S3). Across all moth specimens, acoustic camouflage provided the best protection in the frequency ranges between 29 and 96 kHz, with highest decrease for the 29–43 kHz range in terms of absolute volume change (average decrease of 53 m3 [5%] in detection volume), intermediate by 43–65 kHz (decrease of 40 m3 [10%]) and highest relative reduction (13%) in the 65–96 kHz range (but low absolute reduction [20 m3]). When comparing species from different families, we found our general pattern to deviate in particular for Sphingidae, as their camouflage decreased detection volumes best for the lowest frequency range (19–29 kHz, Figure S3), in one case even up to 365 m3 (10%) for D. porcellus.


We assessed the presence of acoustic camouflage provided by body scales for a whole community of nocturnally active moths. Almost all of our specimen returned stronger echoes after we removed the scales from their abdomen and thorax. The amount of sound absorption provided by the scales depended strongly on the frequency range of interest as well as the body size of the specimen. Larger and heavier species and specimen demonstrated higher levels of acoustic camouflage, in particular for the higher frequency ranges. When translating the level of acoustic camouflage to changes in absolute and relative detection volume, an ecologically relevant measure related to predation risk, we found moth specimen to benefit mostly from acoustic camouflage in the mid frequency ranges, except for Sphingidae species, who benefitted mostly in the lowest frequency range.

From an evolutionary perspective, detection risk is determined by the body parts that provide the strongest echoes from angles that are relevant to bats during the search and capture phase. The wings of moths provide a much larger reflective surface compared to the body and may during flight often block the direct sound path to the thorax and abdomen. Using acoustic tomography, Neil et al. (2020b) estimated the proportion of angles during which a moth body provides the strongest echoes still to be 38%. This estimate assumes ensonification from all angles equally, whereas in reality bats would approach flying moths from their front, back or their side, and to a lesser extent from below or above for moths (when they are flying in a straight line across the landscape). Interestingly, Neil et al. found their estimate to increase to 84% after complete removal of body scales. Although they studied two moth species that were almost twice the size than our largest specimen, their overall level of acoustic camouflage (3.7–5.6 dB) would fit our allometric relationship, and we therefore predict, for most of our specimen, the descaled body to be the most salient target, especially when we take into account that bats may approach more often from angles where the body and not the wings provide the strongest echo.

Bigger bodies reflect more echoes (Barber et al., 2015; Surlykke et al., 1999) and provide a more profitable meal to bats (Koselj et al., 2011), so larger moth species should experience stronger predation risk in the absence of protective sound absorptive structures. Our results suggest that acoustic camouflage occurs across several families of moths and that it can provide some level of protection against bat detection, especially for the largest species, as we had predicted. When intact, the largest moths from our study were found to return echoes that were only ~2 dB louder compared to smallest moths (at frequencies above 29 kHz). When we descaled our specimens, this difference in sound absorption across the size range increased to ~6 dB. To put these differences in perspective, moths that are fivefold larger in terms of body width are roughly 25% more at risk of detection, whereas without the ~4 dB of acoustic camouflage, the increase in risk would have been around 100%. From an evolutionary point, acoustic camouflage may thus enable moths to grow to a larger size and achieve increased fecundity (Tammaru et al., 1996), while reducing the cost of size-dependent predation risk.

Our findings are much in line with findings on other anti-predator strategies that may additionally enable moths to escape the relationship between body size and target strength and ultimately predation risk. In the family of noctuid moths, larger species have shown to possess lower auditory thresholds and lower frequency tuning to echolocation calls when compared to smaller species (Surlykke et al., 1999). When focusing on noctuids of the same size range between studies, the largest species would have ~5 dB lower detection thresholds and ~2.5 dB higher sound absorption compared to the smallest species. Combined, these anti-predator strategies probably put large noctuids at the same detection risk compared to small noctuids, although experimental evidence from staged encounters between free-flying bats and moths would be needed to confirm this. Furthermore, it would be interesting to study the size–camouflage allometry relationship within different lineages that differ in their ear morphology and physiology. Based on our findings, we would, for example, predict the effect of camouflage on reducing size-dependent target strength to be largest for earless species, such as any of the Lasiocampidae.

Variation in the level of acoustic camouflage between our studied species of different families can be largely explained by their body size, and possibly by the amount, type and orientation of scales. Species belonging to Geometridae, for example, have on average the thinnest bodies compared to other families, but show target strength values comparable to other larger bodied families. Most species within this family do not appear very furry, especially when compared to other families such as Notodontidae and Lasiocampidae. However, the few larger bodied species within Geometridae, such as the peppered moth (Biston betularia), also display higher levels of camouflage, supporting the general allometric scaling rule.

We found acoustic camouflage to provide on average the best absolute protection in the mid frequency ranges (29–65 kHz, highest reduction in absolute detection volume), which covers the range of most bat species found within our study area. This also includes the Western Barbastelle (Barbastella barbastellus), which possesses a specialised stealth mode to hunt eared moths (Goerlitz et al., 2010). However, we found the strongest relative reduction in target strength at the highest frequencies (65–144 kHz, highest reduction in relative detection volume), in particular for the biggest moth species. Within our study area, only two bat species rely on echolocation peak frequencies inside these ranges, the Greater and Lesser Horseshoe Bat (Rhinolophus spec.). Bats of this genus use, however, an echolocation strategy that is remarkably different from other bat genera. They employ very long echolocation calls and thus can detect (Doppler shifts induced by) prey wing movements. It would thus be interesting to assess how wing movements interact with body camouflage, in particular at these very high frequencies (Kober & Schnitzler, 1990). Determining whether relative or absolute reduction in detection volume holds greater ecological relevance is complex. If bat density is equally distributed, the absolute reduction in detection volume should be more relevant. This is because the probability of a bat being present within a 50-m3 volume is much higher than within a 20-m3 volume, irrespective of the relative reduction in volume. However, it is also clear that acoustic camouflage works better for higher frequencies, but if this is adaptive or just a physical consequence of a hairy surface still has to be tested. In a study on the reflectance of a hairy inflorescence zone of a bat-pollinated cactus, it also turned out that absorption worked better for higher frequencies (Simon, Matt, et al., 2023).

Moth ears most likely evolved before insectivorous bats took to the air, suggesting that these structures evolved only secondarily to detect echolocation calls (Kawahara et al., 2019). The specialised scales on body and wings likewise have evolved prior to the bat–moth arms race. Day-active butterflies and moths may lack scales that function in acoustic camouflage, at least on their wings (Shen et al., 2018; Zeng et al., 2011). Scales of day-active species may function in particular in visual signalling, and thus need to provide optimal surface to create bright and sharply contrasting colours in order to warn off predators or attract mates (Cuthill et al., 2017). Perhaps, these visual functions trade-off with acoustic functions, which may explain why the few species of butterfly that have so far been tested seem to lack acoustic camouflage on their wings.

The scales on the body and wings of Lepidoptera may originally have evolved to function in thermoregulation (Rydell & Lancaster, 2000). Structures that are characterised by a low density and high porosity increase the surface area of insulating material, as well as sound absorptive material. Sound absorption and insulation against colder night temperatures may thus have evolved hand in hand. The positioning of the scales appears to create a trade-off between these two functions, as erect scales would provide more insulation (Metwally et al., 2019), whereas uniformly layered scales trap more air and thus absorb more sound (Neil et al., 2020a).

Trade-offs between different scale functions can be in particular important for fast-flying species, such as hawk-moths (Sphingidae). Body insulation may trade-off with aerodynamics, as well as sound absorption. Faster flight requires higher body temperatures (Rydell & Lancaster, 2000), so more erect and flurry scales would be favoured, in particular at colder temperatures. Hawk-moths typically show fast and erratic flight, presumably an adaptation against fast-flying bats. They possess, however, a thick layer of flat scales on their bodies, which is optimal for aerodynamics (Dudley, 2002) but also sound absorption (Neil et al., 2020a). Interestingly, we found that only hawk-moths benefitted from acoustic camouflage at the lowest frequency band. These moths typically fly over open fields where they are at risk of predation by fast-flying, low-frequency echolocating bats, such as the Noctule bat (Nyctalis noctula) at our study sites. Scales on thorax and abdomen of hawk moths lay relatively flat (giving them less of a furry appearance compared to Notodontidae and Lasiocampidae). It would be interesting to study the morphological and developmental mechanisms that give rise to the differences among families.

In conclusion, we have shown that many moth species belonging to a range of different families appear to possess some level of acoustic camouflage provided by the scales on their thorax and abdomen. Whether acoustic camouflage functions in reducing predation risk to echolocating bats needs to be experimentally tested, but our estimates indicate at least that larger species may benefit more compared to smaller species, especially those that are at risk of predation by bats calling at high frequencies. Our observed relationship between body size and sound absorptive properties of moth scales therefore has important eco-evolutionary implications and future studies, using phylogenetic comparative methods should test if acoustic camouflage has enabled body size evolution in nocturnally flying insects experiencing strong predation from echolocating bats.


Ralph Simon and Wouter Halfwerk conceived the study and designed the experiments. Wouter Halfwerk collected the specimens, Alrike Dreissen and Helene Leroy carried out experiments. Alrike Dreissen, Helene Leroy, Ralph Simon, Matty P. Berg and Wouter Halfwerk analysed and discussed the data. Ralph Simon and Wouter Halfwerk wrote the paper.


We are grateful to Dr. Peter Lindenburg for help with collecting moth specimen and to Dr. Thomas Neil for helpful suggestions for our analyses.


    The authors report no conflict of interest.


    Raw data on moth specimens are available on Dryad Digital Repository: https://doi.org/10.5061/dryad.66t1g1k73 (Simon, Dreissen, et al., 2023).

      Journal list menu