Comparison of auditory evoked potential thresholds in three shark species

Elasmobranchs (sharks, skates, and rays) are generally regarded as having a poor sense of hearing. This perception is based upon the relatively poor acoustic sensitivity and narrow frequency detection bandwidth (<1 kHz) compared with most teleosts (Maisey and Lane, 2010; Popper et al., 2021). Auditory research in elasmobranchs has received little attention, with basic acoustic sensitivity measurements published for only 12 of the more than 1150 extant species of elasmobranch (Chapuis and Collin, 2022; Ebert et al., 2021; Wiernicki et al., 2020). Yet, sound may be of greater importance to this ancient and diverse group of fishes than previously thought. Behavioural evidence suggests that elasmobranchs use sound for prey detection (Backus, 1963; Banner, 1972; Myrberg, 1972; Myrberg et al., 1976; Nelson and Gruber, 1963; Nelson and Johnson, 1972; Richard, 1968), predator avoidance (Chapuis et al., 2019; Fetterplace et al., 2022; Klimley and Myrberg, 1979; Myrberg et al., 1978) and potentially also reproduction (Carrier, Pratt, and Martin, 1994). Despite the behavioural evidence, there is relatively little physiological data on understanding sound perception in elasmobranchs and their close relatives (Chapuis and Collin, 2022; Mickle and Higgs, 2022).

Underwater sound consists of two components: sound pressure, which is a scalar and omnidirectional, and particle motion, which is a vector and directional (Rogers and Cox, 1988). In fish, the lateral line and the inner ear detect particle motion (Kalmijn, 1988). The ear has three semi-circular canals that are involved in determining the angular movements of the fish (Lowenstein, 1971). The ear also has three otoconial organs, the saccule, lagena and utricle, that are involved in both determining the orientation of the fish relative to gravity and detecting sound (Lowenstein and Roberts, 1950, 1951; Corwin, 1981b; Kalmijn, 1988; Popper and Fay, 1977). The otoconial endorgans contain sensory epithelia (maculae) composed of supporting cells and sensory hair cells that are overlaid with a dense otoconial mass (Mulligan and Gauldie, 1989) (Mulligan and Gauldie, 1989; Tester et al., 1972). At the apical ends of the hair cells are ciliary bundles, composed of many stereocilia that are organized in a stepwise arrangement of increasing height, leading to a single kinocilium (Flock, 1971). The vibrations generated by a sound source efficiently propagate through the water and move the fish as water and fish have similar densities (Rogers and Cox, 1988). The denser otoconia lag the motions of surrounding soft tissue and bend the hair cell bundles, which activates the auditory system (Flock and Wersall, 1963). Fishes with internal gas-filled structures (e.g. swim bladder, auditory bullae, branchial bubbles) can detect sound pressure (Popper and Fay, 1977). The vibrations of the gas bubble transform the pressure signal into particle motion, which is then transmitted to the inner ear (Sand and Enger, 1973). Elasmobranchs lack any known pressure-transducing structure; therefore, it is thought that they can only detect the particle motion component of the sound field (Banner, 1967; Kelly and Nelson, 1975; Popper and Hawkins, 2021).

Elasmobranchs also detect sound using a fourth endorgan that is not loaded with an otoconial mass, the macula neglecta, located inside the posterior canal duct (Corwin, 1977, 1981b; Fay et al., 1974; Lowenstein and Roberts, 1951; Retzius, 1881). The macula neglecta is particularly well developed in sharks and is thought to function in directional hearing in carcharhinid sharks (Corwin, 1977, 1978). The hair cells are embedded in a gelatinous cupula, which is suspended in the endolymphatic fluid of the posterior canal duct (Tester et al., 1972). The hair cells of the macula neglecta are very likely stimulated by movements of the cupula (Tester et al., 1972); however, the modus operandi, including the adequate stimulus have not been unequivocally demonstrated (Popper and Fay, 2011).

Regardless of the mechanism, when bent in the appropriate direction, the sensory hair cells of the otoconial endorgans and the macula neglecta, convert acoustic energy into electric potentials that can be measured from the eighth nerve and higher auditory centres in the brainstem (Bullock and Corwin, 1979; Fay and Popper, 2000; Flock and Wersall, 1963). The auditory evoked potential (AEP) technique has commonly been used to measure basic acoustic abilities in fish (Ladich and Fay, 2013), including sharks (Bullock and Corwin, 1979; Casper and Mann, 2009; Corwin et al., 1982). It is a relatively quick, non-invasive method to record compound field potentials of the entire auditory system (Bullock and Corwin, 1979; Corwin et al., 1982; Kenyon et al., 1998). The method has its limitations, because hearing is a complex cognitive process that requires signal integration at the level of the whole animal and the AEP technique only represents one part of that (Popper et al., 2019; Popper and Hawkins, 2021). Nonetheless, the AEP technique has proven extremely useful to compare frequency detection range and best sensitivity among different species (Ladich and Fay, 2013; Vetter et al., 2018; Wysocki et al., 2009a). A review of AEP-generated detection threshold curves includes only two species of ray and six shark species (Ladich and Fay, 2013), highlighting the need to assess more species from different families, habitats and ecological roles.

The goal of this study was to measure auditory detection thresholds in the New Zealand carpet shark (Cephaloscyllium isabellum), the rig shark (Mustelus lenticulatus) and the school shark (Galeorhinus galeus) using the AEP technique in combination with an underwater speaker tank setup. The carpet shark is a nocturnal slow-swimming benthic shark. The benthopelagic rig shark swims most of the time and specialises in crushing crustaceans close to the seafloor (Francis et al., 2012). The school shark is a benthopelagic species that continuously swims and preys on small fishes and invertebrates (Francis and Mulligan, 1998). The results of the present study extend current knowledge of elasmobranch hearing abilities and will be useful for future integration with morphological and behavioural studies.

Six rig sharks (Mustelus lenticulatus Phillipps 1932) (3 male, 3 female; TL range 51.5–70 cm) and seven school sharks (Galeorhinus galeus Linnaeus 1758) (4 male, 3 female; TL range 49.5–75 cm) were caught using circle-hook and line in the Kaipara Harbour (New Zealand, North Island). Eight carpet sharks (Cephaloscyllium isabellum Bonnaterre 1788) (5 male, 3 female; TL range 58.5–75 cm) were obtained from local commercial fishermen (Table 1). Animals were housed in flow-through holding tanks, supplied with ambient seawater, and maintained on a mixed diet of squid and fish three times a week. Sharks were acclimated for at least 1 week prior to experimentation. All procedures were conducted in accordance with ethics protocols #002066/#AEC23071, approved by the University of Auckland Animal Ethics Committee (AEC).

A 4360 litre circular tank (made of high-density polyethylene, 2100 mm inside diameter, 1260 mm water depth) was isolated from ground vibrations by sitting it on rubber tyres. A monopole underwater loudspeaker (UW-30, Lubell Labs Inc., Columbus, OH, USA) was wrapped in aluminium foil and grounded to reduce signal interference during AEP recordings. The speaker was suspended from a wooden plank using nylon and bungee cords and positioned in the centre of the tank 65 cm below the surface (Fig. 1). The plank was isolated from the tank with multiple sheets of vibration- and noise-dampening mats. The water temperature and salinity ranged from 14.7 to 22.5°C and 35–36 ppt, respectively.

Auditory stimuli were produced by a sound module (Tucker-Davis Technologies, TDT, Gainesville, FL, USA) operated by a laptop (Lenovo ThinkPad X270) running SigGen® (v. 4.4.9) and BioSig® (v. 4.4.11) software. Signals were digitised (RP 2.1, TDT) attenuated (PA5, TDT) and amplified (Pyle®, PLA2378, Sonic Electronix, Louisville, KY USA) before being played through the speaker. Auditory signals consisted of pulsed tone bursts (25–50 ms duration, with a 3 ms rise/fall time gated through a Hanning window) at 80, 100, 150, 200, 300, 450, 600, 800 and 1200 Hz (Table 2). The frequency specificity of acoustic tone pips in confined tank setups can depend on the stimulus duration (e.g. Christensen et al., 2015; Lauridsen et al., 2021). Initial examinations of the frequency spectra of all test stimuli in our tank setup revealed that optimal frequency specificity was achieved for pip durations of 50 ms for the lower frequency tone-bursts (<450 Hz), and 25 ms for frequencies 450–1200 Hz. Pressure waveforms, spectral levels and particle acceleration magnitude spectra are shown in Fig. 2.

Prior to experiments the shark was anesthetized by immersion (∼10–15 min) in a salt-water bath of MS-222 (ethyl 3-aminobenzoate methanesulfonate, PubChem Substance ID: 24894382, Sigma Aldrich) and sodium hydrogen bicarbonate in a 1:2 ratio. When there was no response to tail-pinching, the shark was firmly wrapped (from pectoral fins to tail, leaving the gills exposed for breathing to occur normally) into a piece of stocking, and positioned ventrally onto a custom-made plastic mesh holder, where it was secured with Velcro straps. Stainless steel reusable subdermal needle electrodes (27 gauge, 13 mm, Rochester Electromedical Inc.; Coral Springs, FL, USA) were used to record AEPs. The recording electrode was inserted dorsally, underneath the skin, at the border of the largest lateral-medial diameter of the parietal fossa. The reference electrode was inserted into the cartilage at the tip of the snout, and the ground electrode was placed in plasticine next to the shark's body. The electrodes were insulated with nail varnish, apart from the tips. The electrode wires were tightly wrapped in foil tape and surrounded by a grounded stainless steel mesh sleeve to minimise interference and electrical noise.

The animal holder was then carefully lowered into the tank to ∼70 cm under the waterline, such that the inner ear of the shark was positioned 55 cm in front, and 5 cm above the underwater speaker (Fig. 1). Carpet and rig sharks are buccal pumpers, therefore were allowed to recover from the initial anaesthesia. School sharks are ram ventilators and therefore needed to be ventilated with a dilute mixture of anaesthetic (0.075 g L⁻¹ MS-222 and 0.115 g L⁻¹ sodium hydrogen bicarbonate) that was dripped through the mouth and over the gills during the experiment to maintain a mild anaesthetised state. Previous research (Chapius et al., 2019) has shown that MS222 has no significant effect on the shark's AEP responses at the concentrations used in this study. The presentation order of the frequencies was conducted randomly. An average of 1200 responses (600 sweeps from stimuli presented at 0 deg and 600 sweeps from stimuli presented at 180 deg) was taken for each sound pressure level (SPL) at each frequency. However, to verify AEPs close to threshold level 2000 sweeps were undertaken (e.g. 1000 sweeps at each polarity). AEPs were first elicited using a SPL above threshold (Table 2). The SPLs were then decreased in 5 dB steps for each frequency until an AEP could no longer be visually identified. Then, one to three additional measurements, at 5–15 dB below this roughly estimated threshold were made to ensure responses were not missed. The presence of an AEP was verified visually through (1) observation of the characteristic wave visible above the background noise, and (2) by FFT analysis to screen for peaks at twice the stimulus frequency. This method is commonly used in fish AEP studies (Casper et al., 2003; Kenyon et al., 1998) and is based on the theory that the opposed orientation of the hair cells in the sacculus of the inner ear gives rise to the characteristic frequency response at twice the stimulus frequency (Fay, 1974). The visual estimate of the hearing threshold was defined as the lowest SPL that generated an AEP response in both the averaged trace and the FFT (Vetter et al., 2018).

Dead controls (10 in total) were frozen for a minimum of 24 h and defrosted prior to the control experiment. Control experiments were run according to the same protocol, including the positioning of the electrodes. No AEPs were elicited from any of these control specimens, confirming that the AEPs recorded in this study represent electrophysiological responses from live sharks and were not artefacts from the experimental setup.

As suggested by Sisneros et al. (2016), we included an objective AEP threshold determination method, because visual methods alone were shown to be subject to observer bias (Xiao and Braun, 2008). The ‘x-intercept’ estimation method applied here has previously been used in hearing studies with marine mammals (Nachtigall et al., 2004, 2007), cephalopods (Mooney et al., 2010) and crustaceans (Dinh and Radford, 2021; Jézéquel et al., 2021). First, the 2048-point fast Fourier transform (FFT) power spectra was calculated for 5–12 averaged waveforms per frequency, using a custom-written MATLAB script (see Fig. 5 for example). As with fish AEPs, the spectra revealed peaks at twice the stimulus frequency at suprathreshold and a decrease in FFT peak amplitude corresponded with decreasing SPLs (Maruska and Sisneros, 2016). Second, the maximum FFT values (peak value) were found across five FFT bins greater than the presented frequencies and five FFT bins less than twice the presented frequencies and were plotted against the presented SPL (see Fig. 5D). Third, a series of regressions was run, that included 3, then 4, 5, to the ith peak value. The regression line that yielded the highest r²-value (best fitting line) was selected to calculate the x-intercept. The x-intercept of the best fitting line served as estimate of the animal's probable hearing threshold (Nachtigall et al., 2007).

Acoustic stimuli were calibrated at the beginning of every experimental day using a miniature reference hydrophone (TC4013, sensitivity −211 dB re. 1 V µPa⁻¹, Teledyne Reson Ltd, Slangerup, Denmark) placed inside the empty animal holder and held in place by rubber bands at the same location where the animal's inner ear would be located during the experiment. A digital oscilloscope (Tektronix DPO2014 digital phosphor oscilloscope) was used to measure the sound pressure level (SPL) at each frequency, which was then attenuated through SigGen to output the desired decibel levels. Filtered ambient noise readings were taken to ensure that background sound levels were similar between test days.

Acoustic particle motion is likely to be the most relevant stimulus for hearing in sharks, we subsequently measured the particle acceleration levels, associated with each determined pressure threshold. A waterproofed triaxial accelerometer (DeltaTron®, Type 4524, National Instruments, Austin, TX, USA; sensitivities: x=9.770 mV ms⁻², y=9.977 mV ms⁻², z=10.02 mV ms⁻²) was placed into the fish holder and held in place by rubber bands in the position of the shark's inner ear. The sensors were aligned such that ‘x’ would be the along-body axis (head to tail), ‘y’ the left-right axis and the ‘z’ was the vertical axis. Data from the accelerometer were amplified with a Brüel & Kjær signal conditioning amplifier (NEXUS Type 2690-OS4, Nærum, Denmark) and the peak-to-peak voltage (V_pk–pk) measured on an oscilloscope (Tektronix DPO 2014). The V_pk–pk was converted to V_rms and the acceleration was calculated for the x, y and z planes and followed by the particle acceleration magnitude (µms⁻², calculated as ⁠).

As suggested by Popper et al. (2019) and Popper and Fay (2011), we report the acoustic impedance of the testing environment for all stimulus frequencies and compare it to the acoustic impedance in water in a free field. The acoustic impedance is defined as the ratio between sound pressure (Pa) and particle velocity (ms⁻²). In contrast to a free-field environment, the ratio between the pressure and particle motion is unpredictable for a confined tank environment (Rogers et al., 2016). This is because the pressure to particle motion ratio changes with distance from the sound source, the water surface, the tank walls and the dimensions and material of the tank itself. Because sharks are sensitive to particle motion, their hearing thresholds would also depend on the acoustic impedance of the test environment. Therefore, the acoustic impedance of the experimental tank was determined at the location of the shark's inner ear at three relevant sound pressure levels (120, 135 and 140 dB re. 1 µPa SPL_rms) for all frequencies examined. A hydrophone (TC4013, sensitivity −211 dB re. 1 V µPa⁻¹; Teledyne Reson Ltd, Slangerup, Denmark) and a waterproofed, neutrally buoyant, triaxial accelerometer (DeltaTron®, Type 4524, National Instruments, Austin, TX, USA; sensitivities: x=9.770 mV ms⁻², y=9.977 mV ms⁻², z=10.02 mV ms⁻²) were placed inside the empty animal holder and held in place by rubber bands in the same location as the shark's inner ear. Data from the accelerometer were amplified by a Brüel & Kjær signal conditioning amplifier (NEXUS Type 2690-OS4, Nærum, Denmark) and data from the hydrophone were filtered (HP: 50 Hz; LP: 1 kHz) and amplified by a charge amplifier (VP2000, Teledyne Reson Ltd, Slangerup, Denmark). The V_pk–pk for both the pressure wave and the particle acceleration wave of the x-axis (along-body axis, head to tail) were measured on an oscilloscope (Tektronix DPO2014). The x-axis was chosen because the acoustic power was higher along this axis than along the y-axis (left–right) and z-axis (vertical), due to the position of the speaker 55 cm in front of the shark. Particle acceleration was transformed into particle velocity using the formula v=a/2πf (Nedelec et al., 2016). Then, the impedance for each frequency was calculated in MRayl, where 1 Rayl=1 Pa s m⁻¹. These values were then compared with the theoretical free-field impedance of seawater with a salinity of 35 ppt and 15°C and represented on a log scale (dB re. 1.5597 MRayl) (Vetter et al., 2019). The phase of the impedance was estimated by measuring the phase difference (Δɸ) between the particle acceleration and the sound pressure wave. Based on the assumption that in the acoustic nearfield the phase of the particle velocity waveform leads the phase of the particle acceleration waveform by 90 deg, the phase of the impedance was then calculated as Δɸ_p,v=Δɸ_p,a+90 deg (Vetter et al., 2019).

The distributions of the response variables (pressure and PAL based AEP thresholds) were normal; hence no transformation of the data was needed. However, with only one individual showing a threshold at 800 Hz, this observation was removed from the analysis, because no comparisons could be made for that frequency.

The test tank was supplied with ambient seawater, and the water temperature ranged from 14.7 to 22.5°C over the course of the experimental period, depending on the season. Difference in water temperature can affect the latency and amplitude of the AEP response in marine invertebrates (Jézéquel et al., 2021; Mooney et al., 2010) and has been shown to affect AEP thresholds in some fish (Wysocki et al., 2009b). Sex and total length of the animal may potentially affect hearing abilities and to increase over the course of ontogeny in some species of elasmobranchs (Corwin, 1981c, 1983; Parmentier et al., 2020; Sauer et al., 2022). To test for any effects of water temperature, total length, and sex on AEP thresholds, linear mixed effects analyses were firstly conducted. PAL-based thresholds were fitted against temperature, total length and sex as the main effects and random intercepts for each shark subject. Data for each species were analysed in separate models and checked for linear and quadratic curve relationships, respectively. There was no association between threshold and temperature (carpet shark, t₄=0.87, P=0.43; rig shark, t₃=1.35, P=0.27; school shark, t₃₇=1.99, P=0.053), threshold and total length (carpet shark, t₄=0.13, P=0.91; rig shark, t₂=0.59, P=0.62; school shark, t₃₇=0.92, P=0.36), and threshold and sex (carpet shark, t₄=0.1, P=0.92; rig shark, t₂=−0.05, P=0.97; school shark, t₃₇=0.003, P=0.99) (Fig. S1). Therefore, all data could be grouped to increase sample size and power of the analysis.

To compare the two threshold determination methods (visual and statistical) within each species and each frequency a linear mixed-effects analysis was performed (Bates et al., 2015). Pressure and PAL based thresholds were fitted respectively against a factor, termed ‘group’ as the main effect and random intercepts for each shark subject. The group factor represents all possible combinations of the variables species, frequency, and method (e.g. Carpet.Shark_80Hz_visual, Carpet.Shark_80Hz_statistical, Rig.Shark_80Hz_visual, etc.). This factor was needed for the models to run properly, as there was not enough overlap in the distribution of frequencies, because the carpet shark did not show any responses at 450, 600 Hz. Post hoc pairwise comparisons were used to explore differences between threshold means between methods (within each species and each frequency). Only comparisons within the same species and frequency were examined (e.g. Carpet.Shark_80Hz_visual vs. Carpet.Shark_80Hz_statistical). All between species and ‘nonsensical’ comparisons (e.g. Carpet.Shark_80Hz_visual vs. Rig.Shark_100Hz_statistical) were ignored.

To compare statistically derived thresholds between species within each frequency, linear mixed-effects analyses were performed taking only the data from the statistical method into account. Pressure- and PAL-based thresholds were fitted against a factor, termed ‘group’ (representing all possible combinations of the variables species, frequency, e.g. Carpet.Shark_80 Hz, Rig.Shark_80 Hz, School.Shark_80 Hz, etc.) as the main effect and random intercepts for each shark subject. Post hoc pairwise comparisons were used to explore differences between threshold means between species (within frequencies). Only comparisons between species (within frequency) were examined (e.g. Carpet.Shark_80 Hz vs. Rig.Shark_80 Hz). All ‘nonsensical’ or within-species comparisons (e.g. Carpet.Shark_80 Hz vs. Carpet.Shark_100 Hz) were ignored.

P-values were adjusted using the FDR method and all statistical tests were considered significant at P<0.05. All statistical analyses were performed in R (v.4.1.1; https://www.r-project.org/) using lmer in lme4 (https://CRAN.R-project.org/package=lme4) and emmeans (v.1.6; https://CRAN.R-project.org/package=emmeans) packages.

Fast Fourier transformation analyses of the presented stimuli waveforms reveal clear spectral peaks at the respective stimulus frequencies (Fig. 2). No acoustic energy was contained in any higher-order harmonics, except for 1200 Hz, which was well outside the hearing range of these three animals.

For lower signal levels (e.g. 120 dB re. 1 µPa), the acoustic energy was equally distributed among the three axes, whereas for higher signal levels (e.g. 135 dB re. µPa) the along-body axis (x: 88.5±4.4 dB re. 1 µms⁻²) had the highest accelerations compared with the y-axis (79.3±3.6 dB re. 1 µms⁻²) and z-axis (83.4±2 dB re. 1 µms⁻²) across all test frequencies (Fig. 3). As expected, the impedance values of our test tank were much lower than in a free-field environment (Fig. 4A). This means that the particle acceleration levels, associated with any given sound pressure level, were much higher than what would be expected in the acoustic far field in an unbound medium. Furthermore, these results also indicate that there were no major resonances in the tank at any of the test frequencies (Fig. 4B).

The AEP waveforms of the three shark species were similar in shape and time course and showed a sharp peak at twice the stimulus frequency in the FFT analysis at suprathreshold levels (Fig. 5). A typical suprathreshold AEP response consisted of a series of downward and upward peaks superimposed over a slow negative deflection, that was generally followed by a slow positive deflection, as it is typically described for other fishes (Bullock and Corwin, 1979; Corwin et al., 1982; Kenyon et al., 1998).

The shape and slope of visually and statistically determined threshold curves were similar; however, the statistically derived thresholds were consistently below the visually derived thresholds. For both the pressure and PAL determined thresholds, the statistical estimates were on average ∼4 dB re. 1 µPa below the visual estimates (Fig. 6, Tables S1,S2).

Detection threshold curves for pressure and PAL are similar in shape and slope within each species, with sensitivity maxima and minima at the same frequencies (Fig. 7). Pressure thresholds were similar between the three species (Fig. 7A; Table S3). However, PAL thresholds showed that at 150 Hz the rig shark was more sensitive (P=0.004; 74.8±1.8 dB re. 1 µms⁻² PAL_rms) than the carpet shark (82.8±2.5 dB re. 1 µms⁻² PAL_rms), but not the school shark (79.8±1.6 dB re. 1 µms⁻² PAL_rms) (Fig. 7B; Table S4). All sharks except one responded at frequencies from 80 to 300 Hz, while less than half responded at higher frequencies, and only one rig and none of the school sharks responded at 800 Hz (Fig. 8).

The carpet shark showed the narrowest hearing bandwidth (80–300 Hz), with the lowest mean threshold observed at 100 Hz (SPL_rms: 125.5±2.5 dB re. 1 µPa; PAL_rms: 82.3±1.5 dB re. 1 µms⁻²). The rig shark showed the broadest hearing bandwidth (80–800 Hz) and was most sensitive to sounds below 300 Hz, with the lowest mean threshold observed at 150 Hz (119.8±2.5 dB re. 1 µPa SPL_rms/ 74.8±1.8 dB re. 1 µms⁻² PAL_rms). Finally, the school shark responded from 80 to 600 Hz and was most sensitive to frequencies below 200 Hz, with lowest mean thresholds observed at 100 Hz (122.4±2.6 dB re. 1 µPa SPL_rms/ 79.3±1.6 dB re. 1 µms⁻², PAL_rms).

The background SPL_rms and PAL_rms in the tank (47–1200 Hz) ranged from 81 to 69 dB re. 1 µPa SPL_rms and from 65.3 to 52 dB re. 1 µms⁻², respectively, and was below the shark detection thresholds determined here.

This study provides novel baseline auditory sensitivity data for three ecologically distinct species of sharks, a fish group of which hearing abilities are largely unknown. Auditory evoked potential detection thresholds were quantified in response to pure tone acoustic stimuli from an underwater speaker. It was found that the three sharks were similar in their overall sensitivities but differed with respect to their upper frequency limits. The bottom dwelling carpet shark had the narrowest frequency detection range (80–300 Hz). Bandwidths were broader for the benthopelagic rig shark (80–800 Hz) and school shark (80–600 Hz). These results show that although hearing abilities were restricted to low frequencies, they do vary between species, suggesting that sound likely plays different roles in these species. Further studies comparing the anatomical structures of the inner ears in the three species are needed to determine which morphological adaptations may be responsible for the observed difference in detectable frequency range.

It is difficult to compare AEP thresholds with behaviourally derived thresholds (Hawkins, 1981). The AEP technique requires the fish to be restrained, often by use of anaesthesia, thus providing information on the response properties of the auditory system at the level of the brainstem, but not at the level of the whole animal (Sisneros et al., 2016). AEP thresholds in fish were shown to be above (∼10 dB) behavioural thresholds at frequencies below 1 kHz (Ladich and Fay, 2013; Popper et al., 2019). In addition, absolute hearing thresholds vary greatly between different studies, because of differences in experimental setups and acoustic environments, making quantitative comparisons of threshold levels impossible (Ladich and Wysocki, 2009; Ladich and Fay, 2013). However, AEP threshold curves and behavioural audiograms can be used to estimate the auditory bandwidth of a species (Ladich and Fay, 2013; Vetter et al., 2018).

It is important to note that we were not able to test frequencies below 80 Hz, owing to tank acoustics, and speaker limitations. Therefore, it is very likely that the species examined here can detect frequencies well below 80 Hz, as has been shown for other species of sharks. For instance, the horn shark (Heterodontus francisi) (Casper and Mann, 2007a), bamboo sharks (Chiloscyllium plagiosum and Chiloscyllium punctatum) (Casper and Mann, 2007a,b) and the Atlantic sharpnose shark (Rhizoprionodon terranovae) (Casper and Mann, 2009) showed AEP responses to frequencies as low as 20 Hz. And the lemon shark (Negaprion brevirostris) has been shown to respond to frequencies as low as 10 Hz in behavioural experiments (Nelson, 1967).

The habitat and lifestyle of a species (e.g. motionless on the seafloor versus swimming in midwater) was proposed as a potential determining factor of the upper hearing limit in elasmobranchs (Corwin, 1981a; Mickle and Higgs, 2022). The present study showed that the demersal carpet shark has a low upper frequency limit of 300 Hz, compared with the more active, benthopelagic school and rig sharks that respond at 600 and 800 Hz, respectively. Physiological and behavioural evidence in other benthic sharks, such as C. pliagiosum, H. francisi (Casper and Mann, 2007a; Kelly and Nelson, 1975) and the cloudy catshark (Scyliorhinus torazame) (Ahn et al., 2011) suggest equally low upper frequency limits (max 300 Hz). However, Corwin's habitat-niche hypothesis to predict hearing abilities in elasmobranchs (Corwin, 1977, 1981a,b,c, 1989) does not apply to all bottom-dwelling species, as some strictly benthic species have been shown to detect frequencies of up to 1 kHz. For instance, AEP measurements in the bottom-dwelling, small spotted catshark (Scyliorhinus canicula) showed an upper frequency limit at 600 Hz (Parmentier et al., 2020) and nurse shark (Ginglymostoma cirratum) 1000 Hz (Casper and Mann, 2006). Similarly, benthic batoid species have all showed upper frequency limits of ∼800–1000 Hz (Casper et al., 2003; Casper and Mann, 2006; Corwin, 1983; Mickle et al., 2020). Habitat and feeding strategies likely contribute to hearing abilities in elasmobranchs, but there may be other factors (e.g. predation risk, reproduction strategies) driving interspecies differences in elasmobranch hearing abilities. There is currently a lack of audiometric data for any of the more basal Squalomorphii sharks, deep-sea or purely oceanic species as these are very difficult to keep in captivity (Corwin, 1989). Further comparative physiological, behavioural and anatomical studies in more species representative of different lifestyles and habitats are needed to resolve the ecological drivers for differences in hearing abilities in elasmobranchs.

The bandwidths and peak sensitivities of AEP detection thresholds of the sharks tested here are comparable with teleost species that only detect particle motion (Ladich and Fay, 2013; Popper and Fay, 2011). Examples are the flatfishes, such as common dab (Limanda limanda) and European plaice (Pleuronectes platessa), that have narrow hearing ranges (30–250 Hz) with best hearing at or below 110–160 Hz (Chapman and Sand, 1974); the common triple fin (Forsterygion lapillum) with peak sensitivity of 100–200 Hz (Radford et al., 2012); the red-mouthed goby (Gobius cruentatus) with upper limit at 700 Hz and peak sensitivity at 100–200 Hz (Wysocki et al., 2009a) and kawakawa (Euthynnus affinis) with upper limit at 800 Hz and peak sensitivity at 500–600 Hz (Iversen, 1969). The low frequency bandwidth observed in the sharks support previous experimental evidence suggesting that sharks only detect particle motion (Banner, 1967; Kelly and Nelson, 1975).

The results of this study must be evaluated with respect to the limitations of the acoustical setup and the measurement approach. Tank acoustics are very complicated, due to complex interactions of the sound with the water–air boundary and the tank walls (Parvulescu, 1964; Rogers et al., 2016). Therefore, the acoustic condition in the test tank experienced by the shark is very different from the natural free-field environment (Gray et al., 2016; Larsen and Radford, 2018; Popper and Hawkins, 2021). A small sound pressure generated within the tank will produce much larger particle motion than would be the case in the natural environment (Parvulescu, 1964). As a result, the hearing abilities of the sharks may be unnaturally extended to higher frequencies in a tank setting (Chapman and Sand, 1974). Ideally, hearing thresholds should be measured in the free field (e.g. Chapman and Sand, 1974; Hawkins and Chapman 1975; Hawkins and Johnstone, 1978). It would be particularly interesting to conduct free field experiments within both the acoustic near field and far field, to solve the currently unanswered question of how far from a sound source sharks can detect sound (Casper and Mann, 2006, 2009). In this study, the shark was positioned within the acoustic nearfield [assuming the boundary of the acoustic nearfield for a monopole source is approximately equal to λ/2π (Yan, Anraku, & Babaran, 2010)], which the impedance results confirm (Vetter et al., 2019). In addition, in the nearfield, there exists hydrodynamic flow that is generated by the vibrations of the underwater loudspeaker (Kalmijn, 1988). In some teleost fishes, the lateral line responds to these flows at frequencies up to 200 Hz. (Braun and Coombs, 2000; Harris and van Bergeijk, 1962; Higgs and Radford, 2013; Hueter et al., 2004; Maruska and Sisneros, 2016). Therefore, the AEP responses measured at frequencies below 200 Hz may potentially contain contributions from the lateral line system.

In summary, this study assessed AEP sensitivities to acoustic stimuli from an underwater speaker in the benthic New Zealand carpet shark, the benthopelagic rig and the school shark. The three species have similar sensitivities but differ in their upper frequency limit. Our results indicate that the hearing abilities of the three shark species were most consistent with those known for hearing ‘non-specialists’ fishes that detect only the particle motion component of the sound field. Furthermore, our results provide evidence that benthopelagic sharks can hear higher frequencies (max. 800 Hz) than some of the bottom dwelling sharks (max. 300 Hz). Finally, the AEP detection thresholds presented here provide a good first approximation of the basic hearing abilities for the three shark species. However, further behavioural and morphological studies are needed to test which ecological factors are driving the observed interspecific differences in hearing bandwidth.

The authors would like to thank Errol Murray, Gavin Perry, Derek Sauer, Stefan Spreitzenbarth, and Stefano Schoene from the Leigh Marine Lab for their help with the animal collections. We would like to express our gratitude to Jenni Stanley and Paul Caiger for their help troubleshooting the AEP setup. We are grateful to Jessica McLay and Jason P. Dinh for help with statistics.

This article has an associated ECR Spotlight interview with Carolin Nieder.