This discussion presents a convergence of two complementary approaches to explaining mammalian hearing. The proximate approach is that species hear the way they do because of the mechanical structure of their ears. The ultimate approach refers to selective pressures that favor particular hearing abilities, and the structures that enable those abilities, that confer an advantage for survival and reproduction.
The middle ear and hearing range
The middle ear of the African pygmy hedgehog shows characteristics generally regarded as ancestral features of therian mammals. These features include a ‘loose’ U-shaped ectotympanic which does not form a complete auditory bulla, a relatively large pars flaccida of the tympanic membrane, and the retention of both middle ear muscles (Henson, 1961; Segall, 1969; Novacek, 1977; Fleischer, 1978). The malleus in such ears has what Fleischer (1978) referred to as ‘ancestral-type’ morphology, including a long anterior process which is tightly connected to the ectotympanic, and a manubrium approximately parallel to that process. Fleischer’s ‘microtype’ malleus, found in small mammals such as mice, bats and shrews, is very similar but for the development of the orbicular apophysis, a projection from near the base of the manubrium. The malleus of African pygmy hedgehogs conforms to this microtype morphology, although its orbicular apophysis (Fig. 6) is not as prominent as in some of the smaller ‘microtype’ species. Other eulipotyphlans, including other hedgehogs, shrews, solenodons and some talpid moles, share similar microtype middle ear features to Atelerix (Wassif, 1948; McDowell, 1958; Henson, 1961; Segall, 1970; Mason, 2006; Wible, 2008; Zaytseva et al., 2015), with variable levels of development of the orbicular apophysis. Although African pygmy hedgehogs are somewhat larger than other species possessing microtype ears, their middle ear measurements fit closely with the regression relationships for microtype species reported by Mason (2013). Hedgehogs like Atelerix can be regarded as relatively large microtype species, with otherwise primitive therian middle ear characteristics.
At low frequencies, sound transmission through the middle ear is affected by acoustic compliance (see Mason, 2016b for a review). The main determinants of compliance likely include the volume of the middle ear cavity and the stiffness of the conducting apparatus, i.e. tympanic membrane and ossicular chain. Low-frequency sound transmission could then be improved by a larger cavity volume and a less-stiff tympanic-ossicular chain, but their relative contributions can differ. Compared to non-microtype mammals of similar body mass, microtype species have smaller middle ear cavities, a difference especially evident in smaller microtype species (Mason, 2016a). Microtype species also have a particularly stiff articulation between an expanded anterior process of the malleus and the tympanic bone: in Atelerix we found the two bones to be fused, i.e., synostosed. It is unclear whether the low tympanic cavity volume or the high tympanic-ossicular stiffness is more limiting to low-frequency hearing in microtype species.
Low-frequency hearing
The hearing of the African pygmy hedgehog differs in both absolute sensitivity and frequency range from that of Hemiechinus auratus, the only other hedgehog the hearing of which has been behaviorally determined. (Ravizza et al., 1969). Fig. 8 compares these two species and shows that the long-eared hedgehog hears 1.87 octaves further into the low frequencies and is also more sensitive throughout its hearing range. Nevertheless, both species of hedgehogs fall into the group of mammals that do not hear frequencies below about 400 Hz (at a level of 60 dB or less), as illustrated in Fig. 9 (cf. Heffner et al., 2020). This bimodal distribution of low-frequency hearing (unlike the approximately normal distribution of high-frequency hearing, Heffner et al., 2001b) suggests that it would be instructive to consider environmental factors that might exert selective pressure on low-frequency hearing, as well as anatomical mechanisms.
The middle ear morphologies of most terrestrial mammals fall on a spectrum between two extreme types, the small and stiffly-articulated microtype referred to earlier, and a more compliant ear referred to as ‘freely mobile’ (Fleischer, 1978; Mason, 2013). Being small and stiff, microtype ears (and the ancestral ears, which are similar) should be good transmitters of high frequencies from tympanic membrane to inner ear, while transitional ears and freely-mobile ears should be better than microtype ears at transmitting low frequencies (Mason, 2016b). Thus, it is of interest to consider whether grouping the ears in this way is congruent with the bimodal distribution of mammalian low-frequency hearing shown in Fig. 9 and Table 1 (Heffner et al., 2001b).
Fig. 9 illustrates the bimodal distribution of low-frequency hearing limits of terrestrial mammals, excluding the aquatic/amphibious species with their highly specialized ears. The upper mode consists of mammals, including both species of hedgehogs, that do not hear low frequencies (below approximately 400 Hz at 60 dB SPL). It includes 24 terrestrial species, 22 of which have been classified as having a microtype ear, plus two marsupials with ancestral middle ears, which are very similar to the microtype except for the lack of the orbicular apophysis. None of these 24 terrestrial species have freely mobile ears. Thus, the microtype/ancestral morphology seems to be found in species that do not hear low frequencies.
On the other hand, there are 43 terrestrial species known to hear below 400 Hz (the lower mode in Fig. 9). Of these 43, all have been classified as having more flexible middle ears (often labeled either freely mobile or transitional)—and despite the good high-frequency hearing of most of them, none has a microtype ear. We do not suggest that non-microtype ears preclude good high-frequency hearing, only that, because of their higher compliance, they are likely to be better transmitters of low-frequency sound. This is illustrated by relatively large, non-microtype species with good high-frequency hearing, such as pigs (42 Hz to 40.7 kHz, Heffner and Heffner, 1990) and cats (55 Hz to 79 kHz, Heffner and Heffner, 1985). Perhaps more surprisingly, this group also includes many small species that not only hear below 50 Hz, but also frequencies as high as 60 kHz (gerbil and least weasel, both weighing about 80 g). It would appear, then, that the broad categorization of an ear as freely mobile/transitional as opposed to microtype/ancestral allows us to predict whether the animal in question can or cannot hear sound below 400 Hz. However, high-frequency hearing limits are normally distributed and no clear division is found in the upper frequency limits of hearing that can be linked to ear type.
Both the hedgehogs, A. albiventris and H. auritus, fall within the group of mammals with limited low-frequency hearing, and both possess microtype middle ears. One key difference between the two species relates to the volume of the middle ear cavity. In A. albiventris, cavity volume was measured here at around 40 μl, while the volume for two specimens of H. auritus was 120 and 140 μl (Packer, 1987). Because acoustic compliance is directly proportional to cavity volume, acoustic compliance will be three times greater in H. auritus. This is consistent with the superior low-frequency hearing of H. auritus compared to A. albiventris(Fig. 8), although other differences in their ears, including tympanic-ossicular stiffness, are also potential contributors. In speculating on an evolutionary explanation for their difference in hearing and anatomical configuration, we would look to differences in their habitat or lifestyle. It has long been known that many small mammals living in arid habitats have enlarged auditory bullae, which are associated with improved low-frequency hearing in an environment in where that may be ecologically advantageous (see Mason, 2016a for a review). Whether the differences between the anatomy and audiograms of Hemiechinus and Atelerix relate in a similar way to differences in their natural habitats remains unexplored.
High-frequency hearing
The high-frequency hearing limit of A. albiventrisisconsistent with selective pressure on small species to hear high frequencies in order to localize sound using interaural intensity differences (e.g., Heffner and Heffner, 2016). High-frequency hearing is strongly correlated with functional interaural distance, measured as the time required for a sound to travel from one auditory meatus to the other, whether in air or in water (r = – .76, p < .0001, n = 74). The correlation is based on the apparent selective pressure on species with smaller functional interaural distances to hear frequencies high enough to be shadowed by their head and pinnae and produces a binaural intensity/spectral-difference cue that can be used to localize sound sources. With a functional interaural distance of 135 μs, the 46-kHz high-frequency hearing limit of A. albiventrisdoes not deviate significantly from the mammalian regression line (p > .05). Hedgehogs, despite their unusual anti-predator behavior of rolling into a ball which seemingly relies more on detecting predators than requiring a directional escape response, have high-frequency hearing consistent with this overall mammalian pattern.
Sound-Localization
Fig. 10 shows that the localization thresholds of both hedgehogs fall near the midrange of the distribution of 41 mammals (mean 13.1°, median 14.0°). At 14°, the African pygmy hedgehog is slightly better than the only other hedgehog with known localization acuity, Paraechinus hypomelas (Brandt’s hedgehog, previously known asHemiechinus hypomelas)with its threshold of 19° (Chambers, 1971). As such their localization acuity is not unusual among mammals as a whole.
Despite fairly typical acuity for localizing broadband noise sources that provide both time and intensity binaural cues to locus, the African pygmy hedgehog,like Brandt’s hedgehog, does not use both of these cues for localization (cf. Masterton et al., 1975). With a functional interaural delay of approximately 135 μs, the interaural phase-difference cue is expected to become physically ambiguous for African pygmy hedgehogsabove about 7.7 kHz, forcing a reliance on an interaural intensity difference for pure tones above that frequency—frequencies that were localized easily. Similarly, it is estimated that animals with this functional interaural distance would produce little or no intensity difference between the two ears at frequencies below about 11.6 kHz (based on the simple spherical head model), thereby forcing reliance on any available interaural phase differences to localize pure tones at lower frequencies. However, the ability of the nervous system to phase-lock serves to limit the use of the phase cue at frequencies higher than about 5 kHz even when that cue is physically present (cf. Heffner and Heffner, 1987; Heffner et al., 1994, 2001c). Unlike the good performance at high frequencies where the interaural intensity difference was available, localization of lower frequencies deteriorated and fell to chance as pure tones of 4 kHz and below were tested, indicating that this species cannot localize sound sources when forced to rely on the interaural phase-difference cue alone. This result was further supported by the inability to localize a clearly audible 3-kHz tone when it was amplitude-modulated. Such an amplitude-modulated envelope provides a basis for a time comparison between the two ears that enables even small species possessing the ability to use time cues to localize frequencies that are not localizable as pure carrier tones (such as the common vampire bat, Desmodus rotundus, Heffner et al., 2015). However, the envelope did not improve localization performance by African pygmy hedgehogs. Thus, like Brandt’s hedgehog, the African pygmy hedgehog showed no evidence of being able to use interaural phase differences in either the carrier wave, or envelope of a sound.
The inability to use the phase cue at low frequencies suggests that the African pygmy hedgehog also should not be able to localize tones of 5.6 and 8 kHz because the spherical model suggests the only cue would have been a phase difference. There are two reasons for inferring that those frequencies were not localized using the phase cue, but instead relied on an intensity difference. First, even though a phase difference might be physically available at frequencies as high as 7.7 kHz to a small species with a head the size of Atelerix, mammalian auditory systems are not known to phase lock at frequencies above about 5 kHz, hence they are not able to use the cue even though it is physically present. Indeed, the Egyptian fruit bat (Rousettus aegyptiacus) and the Jamaican fruit bat (Artibeus jamaicensis), both echolocators, are the only species known to localize pure tones at such high frequencies—5.6 kHz and 6.3 kHz, respectively, and both of those species use the phase cue at lower frequencies as well (Heffner et al., 1999, 2001c). We know of no species that uses the phase cue at relatively high frequencies, but not at lower frequencies where it is most effective.Instead, the observation that localization was, in fact, possible at5.6 and 8 kHz, islikely owing to an enhancement of interaural intensity differences by the animal’s relatively large pinnae and the non-spherical shape of its head. A simple spherical model cannot take into account the directional properties of the pinnae and head, and provides only an estimate of the lowest frequency at which the head shadow should become effective. More definitive tests would require the placement of headphones on trained hedgehogs.
The inability of both species of hedgehogsto use the binaural phase cue is noteworthy. It demonstrates that even though a sound can be easily audible, it may not be localizable. Perhaps more importantly, it shows that at least average localization acuity for more natural broadband sounds (that produce both time and intensity cues) can be achieved using only one of the binaural cues. Thus, one more species of hedgehog (from a different genus) now joins several species of rodents and bats known not to use the binaural phase cue for sound localization (domestic Norway rats, Rattus norvegicus, Wesolek et al., 2010; domestic house mouse, Mus musculus, Heffner et al., 2001a; spiny mouse, Acomys cahirinus, Mooney, 1992; straw colored fruit bat, Eidolon helvum, and dog-faced fruit bat, Cynopterus brachyotis, Heffner et al., 2010a; greater spear-nosed bat, Phillostomus hastatus, and short-tailed fruit bat, Carollia perspicillata, Heffner et al., 2010b; big brown bat, Eptesicus fuscus, Koay et al., 1998). It might be thought that all of these species that do not use the binaural time cues simply have heads too small to generate useful time delays, but this is contradicted by at least five even-smaller species that do use time cues (Fig. 10: Egyptian fruit bat, Rousettus aegyptiacus (Heffner et al., 1999); kangaroo rat, Dipodomys merriami (Heffner and Masterton, 1980); gerbil, Meriones unguiculatus (Heffner and Heffner, 1988); least weasel, Mustela nivalis (Heffner and Heffner, 1987); common vampire bat, Desmodus rotundus (Heffner et al., 2015); Jamaican fruit bat, Artibeus jamaicensis(Heffner et al., 2001c). It seems the mammalian nervous system is capable of analyzing small interaural time differences for localization, but it does not do so in every species.
Finally, we have noted (Heffner et al., 2010b) that all of the species that do not use the binaural time cue fall into the group of mammals that do not hear low frequencies (Fig. 9, Table 1). As such, we might be tempted to conclude that they simply do not hear low enough to use time cues. Again, this possibility is contradicted by species such as Egyptian fruit bats with a low-frequency hearing limit of 2.25 kHz (Koay et al., 1998), and Jamaican fruit bats with a low-frequency hearing limit of 2.8 kHz (Heffner et al., 2003). These species have similar low-frequency hearing to that of African pygmy hedgehogs, and even poorer low-frequency hearing than Brandt’s hedgehog,yet they retain the ability to use the binaural phase-difference cue even though the frequency range over which it can be used is narrow due to their limited low-frequency hearing.
In summary, the overall localization acuity of both species of hedgehogs is unremarkable compared to other mammals. However, their inability to use the time cue for localization is somewhat unusual. Foregoing the use of time cues for localization, first noted in Brandt’s hedgehog (Paraechinus hypomelas), is not restricted to hedgehogs since it is also found among several rodents and bats. Yet, sound localization that relies on only one of the binaural cues to locus seems surprising when so many species use both cues. This remains unexplained—we still do not have an indication of what selective pressures or neural mechanisms might account for this condition or even how widespread it is among mammals. Indeed, we might ask why so many species devote resources to the maintenance of the dual neural apparatus to process both binaural cues when either one alone seems sufficient for good localization.