AM. ZOOLOGIST, 6:371-377(1966). Evolution of the Sense of Hearing in Vertebrates WlLLEM A. VAN BERGEIJK* Bell Telephone Laboratories, Inc., Murray Hill, Neio Jersey SYNOPSIS: This paper briefly recapitulates work on the preterrestrial phase of the evolution oE hearing which was published in detail elsewhere. Following this, the problem of the origin of the terrestrial middle ear is examined in some detail, and it is demonstrated that Eusthenopteron, and probably other rhipidistians as well, possessed an eardrum. This drum was formed by the fusion of the distal surface of a spiracular diverticulum and a ligament between parietal shield and squamosal. It is further demonstrated that the geometry of this eardrum in relation to the middle-ear cavity was adequate to allow the fish to make the transition from aquatic hearing to terrestrial hearing without loss of sensitivity. As the various papers presented in this paper of R. M. Denison in this Symposium), Symposium demonstrate, the evolution of we pick up the thread of evolution at a the ear as a morphological entity has been point where a hypothetical early vertebrate intensively studied, and a considerable de- (or pre-vertebrate) had a functional lateralgree of understanding has been attained line system. We assume that this lateralis with regard to the homologies of structure system remained, as far as its functional that exist. The same cannot be said of aspects arc concerned, essentially the same the evolution of hearing as a physiological up to the present. Thus, we can bring to phenomenon. The reasons for this state of bear the results of Cenozoic research on affairs are obvious: morphological hypothe- Paleozoic problems. The lateral line (see ses can be tested quite readily by examina- Dijkgraaf, 1963, and van Bergeijk, 1966, for tion of fossil material, but hypotheses about extensive discussions) is a sensory system function are virtually untestable, since that responds to water motions. It is not physiology demands a living body for the sensitive to pressure or pressure waves, and test. What information we do have about thus not to sound in the conventional sense. the evolution of function is, therefore, However, a real sound source in water based heavily on physical principles, morph- produces water motions as well as pressure ological evidence, logical inference, and waves in its near field. These near-field analogy with the known physiology of con- water motions excite the lateral line (Hartemporary species. ris and van Bergeijk, 1962), and thus it In a recent paper (van Bergeijk, 1966) can be said that the lateral line is sensitive I proposed an outline for a theory on the to near-field sound. The old controversy of evolution of hearing, which was based on whether the lateral line does or does not just such a variety of sources. The present respond to sound is thus laid to rest by paper will expand somewhat on that work; the simple expedient of denning "sound" specifically the evolution of the terrestrial more accurately. Moreover, the lateral line middle ear will be considered. In order is directionally sensitive (van Bergeijk, to be properly understood, however, I must 1964), and enables the animal to locate a briefly restate the salient points of my source of disturbance at a distance. One can argue endlessly whether "distant touch" earlier paper. or "nearby hearing" is the more appropriate term to describe this sensory system; PRETERRESTRIAL EVOLUTION1 the distinction so far is largely semantic, Although the origin of the lateral line is since neither description leads to concluby no means a solved question (see the • Present address: Center for Neural Sciences, Indiana University, Bloomington, Indiana 47401. 1 In order to minimize clutter, only key references will be given; I refer to my earlier paper for more thorough discussion and further references. (371) 372 WlLLEM A. VAN BERGEIJK sions which the other does not also imply. Harris, 1964). As a result of the changes Hearing in the conventional sense, viz., the in volume of the bladder, near-field waterperception of pressure waves, is not deriv- motions (as well as pressure waves) radiate able directly from lateral-line perception from it. The inner ear is sufficiently close anyway, since an additional structure (hav- to this source of near-field motions to being nothing to do with the lateral line) is come excited by them, and a new sensory required for the transformation of pressure channel is opened: thanks to the pressureto water motion. to-displacement transformer-action of the But first came the inner ear, the laby- swimbladder, the inner ear is now sensitive rinth. Initially, it had nothing to do with to pressure waves. It hears. hearing, even near-field hearing. In fact, Various adaptations evolved that appear it appears that the labyrinth is a successful to increase the sensitivity of this new sensory attempt to isolate part of the fluid-motion channel. The bladder becomes connected detector-apparatus of the lateral line from to the inner ear by means of a chain of outside disturbances, and put it to work ossicles (Weberian ossicles of Ostariophysi), to detect movements of the animal's own bladder extensions (in clupeids, for exbody. The simple ears of Myxine and ample), or other specializations. The effect Petromyzon, for instance, make sense as of such connections is primarily to minimize angular and linear accelerometer systems, transmission loss, though some equalization and the evolution from one to two to three of frequency and acoustic filtering are possemicircular canals is a very nice example sibly important secondary functions. Data of increasing precision in a biological available in the literature (see van Bergeijk, measuring device.2 By being sunk deeply 1966, for discussion) appear to support into the skull, the inner ear is much less these notions. The inner ear evolves specialsensitive to near-field motions in the ex- ized sensory sites that deal exclusively with ternal environment than the lateral line, fluid motions originating from the swimand it is, of course, certainly not more bladder, and contributes its part to a lowsensitive to pressure waves. loss transmission path by the development The structure that converts pressure vari- of a perilymphntic system. ations into water motions is the swimbladThe pressure-sensitive ear of fishes is, der. It can be regarded as essentially a gas however, not sensitive to the direction of bubble, having a primary function for sound. Whereas the lateral line is capable buoyancy perhaps, or respiration; the origi- of locating the source of near-field motions nal function of the organ, or the selection quite accurately, the ear of fishes is inpressures to which it is (presumably) an capable of locating the source of a pressure adaptation, do not concern us here. Being wave (van Bergeijk, 1964). a gas bubble, a swimbladder will obey the The swimbladder with its extensions or gas laws and change volume when pressure is changed. It will follow variations in ossicles may, quite properly, be considered pressure over a considerable range of fre- a middle ear; its function and mode of quencies because, although it has a definite action are the same as those of the terresresonance frequency, damping is so high trial middle ear, which also is essentially a that the response curve is almost flat (see trapped air bubble, changing volume under the influence of pressure waves. First, however, we must examine the origin of the s I t is sometimes proposed that the myxinoid ear is a degenerative version, rather than a primitive terrestrial middle-ear gas-bubble, for it is one; this seems unlikely to me, for why should the clear that it does not derive from any sort ear degenerate in an actively swimming hagfish of swimbladder. Instead, the middle ear but remain beautifully intact in such an almost is derived from the first, or spiracular, sedentary animal as a flatfish? Moreover, in degeneration it is size and differentiation that are pouch of the pharynx. There are some modern, teleost, fishes, notably members of reduced, not dimensionality. EVOLUTION OF HEARING IN VERTEBRATES the genus Ophicephalus (now, Channa), which possess air-filled diverticula of the spiracular pouch that serve as auxiliary respiratory organs. Other modern fish, such as the anabantids, have similar air-pouches although these do not derive from the first, but from one of the other gill pouches. These air-filled sacs have been demonstrated to be important for hearing in those fish (Schneider, 1941). Some Devonian Rhipidistia, notably Eusthenopteron, possessed diverticula of the spiracle; some cogent arguments can be made for (and none against) the assumption that this rhipidistian pouch also was filled with air, and served as a middle ear (van Bergeijk, 1966; see also K. S. Thomson's paper in this symposium). Thus, as the earliest amphibians evolved from the rhipidistians, the terrestrial middle-ear cavity was already present morphologically and was, as we shall see, apparently functional. Pharyngeal middle ears are distinctly superior to a swimbladder middle ear in fish. For one thing, the air bubbles rest almost directly against the inner ears, and no special chains of ossicles are required for efficient transmission; more importantly, there are now two sound receptors which, at least in principle, allow directional hearing in the far field. The emergence on land, however, poses some special problems for this hearing apparatus which we shall now examine in some detail. EVOLUTION OF THE TERRESTRIAL MIDDLE EAR The problems encountered by the rhipidistian hearing apparatus with the emergence on dry land are traceable to just one factor: the acoustic impedance of air. Acoustic impedance for plane waves is given by the product of density Q>) of the medium and the velocity of sound (c) in the medium. The impedance of water (and, to a first approximation, of living tissue), pcwater, is about 153,000 g/cm2/sec (the exact value depends on temperature, atmospheric pressure, solutes, and other factors; the value given is only intended as representative). The acoustic impedance of air 373 (pcair)> however, is only 39.5 g/cm2/sec» or about 1/4000 that of water. One important consequence of this difference is that a sound source in air produces only 1/63 of the pressure3 developed by an underwater source of the same intensity (power transmitted/cm2). This ratio, conventionally expressed in decibels (in this case, some 36 dB), represents the "hearing loss" suffered by a rhipidistian that crawled out of the water onto land. For the middle ear is a pressure-sensitive device, and it is the sound pressure that is diminished in air as compared to water.4 This loss of sensitivity is, however, at least partially, and perhaps largely, offset by a peculiarity in the anatomy of the spiracular diverticulum and the bones surrounding it. We shall take Eusthenopteron as an example, because there is a great deal of detailed information available about this fish; whether other rhipidistians show similar structures is not known to me. In Figure 1, redrawn from Jarvik (1954), is shown part of the skull of Eusthenopteron. The dermal bones are removed (although some are shown in outline by a heavy dashed line), and the palatoquadrate is similarly left off. Thus the view is into the medial wall of the spiracular diverticulum. The lightly dashed line, running along the dorsal margin of the hyomandibular, outlines the dorsal extent of the removed half of the diverticulum. The diverticulum is a rather flat and broad one, sandwiched as it were between the hyomandibular and parotic dental plates and the palatoquadrate. According to Jarvik (op. cit., p. 50) the diverticulum probably opens to the outside along this cleft. If that a The ratio of pressures in air and water is proportional to the square root of the ratio of the impedances. 4 We assume here that the power output of sound sources is about equivalent in air and water, since the power of a source is a function of the energy available to it, as well as of the radiation impedance. For biological sources (mates, prey, predators) this is quite probably the case; if anything, they are less efficient in air because of poorer coupling, and thus the effective power of these sources may be less in air than in water. WrLLEM A. VAN BERfiEIJK 374 SPIRACULAR PARIETAL CANAL INTERTEMPORAL 1 SUPRATEMPORAL DIVERTICULUM - < - SQUAMosAi HYOMANDIBULAR DENTAL PLATE 10MM FIG. 1. Otic region ot the skull of Euslhenopteron. Palatoquadrate removed to show hyomandibular side of spiracular diverticulum. Heavy dashed lines: outlines of dermal bones covering the skull; names of these bones are underlined. Redrawn from Jarvik (1954). were true, then no air could be contained in it, and it would act perhaps as an extra spiracle. It seems, however, rather unlikely that this is the case, because the spiracle itself opens, via a separate canal, to the outside at the extreme front end of the cleft (see Fig. 2). Why should a diverticulum develop that is just a parallel channel to an already existing one? Jarvik offers no support for his statement and, in fact, makes it essentially en passant; he appears to have given it no further consideration. Thomson (this symposium), on the other hand, states that there must have been a Jigamentous connection of the squamosal with the bones of the parietal shield (supratemporal and postspiracular, 5 see Fig. 2, A), because without such a ligament the skull would have fallen apart. These dermal bones overlie the hyomandibular and palatoquadrate, and the ligament between them allows no place for the diverticulum to open up to the surface. More significantly, as shown in Figure 1, the margins of the dermal bones (supratemporal-postspiracular and squamosal) coincide precisely with the dorso-lateral margins of the hyomandibular and palatoquadrate, and thus with the distal margins of the diverticulum, so that the ligament covers the diverticulum exactly. In other crossopterygians (say, coelacanths or dipnoans) this coincidence of dermal and visceral bone-margins is not found; of course, there also is no evidence of a diverticulum in these forms. In Osteolepis, however, there appears to have been a diverticulum present (Thomson, personal communication), and the external skull matches this by showing a slit similar to that of Eusthenopteron (Fig. 2, B). The conclusion is now inescapable that the diverticulum indeed ends blindly, as I had previously only assumed. 5 1 used Jarvik's terminology here, although it is somewhat different from standard nomenclature; what Jarvik calls intertemporal and supratemporal, for instance, are usually known as siipratenipoial and tabular, respectively. EVOLUTION OF HEARING IN VERTEBRATES 375 There appears to be no reason why we cannot take the situation one step further, and propose that the diverticulum applied itself intimately to the ligament, thus forming a double-layered membrane. The inner layer of this membrane is then the entoderm of the diverticulum, while the outer layer is an ectodermal ligament. This hypothetical membrane is a precise homologue of the tympanic membrane that is recognized in tetrapods. It would appear, therefore, that we may hypothesize that in the rhipidistians SPIRACLE FIG. 3. Schema to explain the effect of asymmetrical placement ol an air bubble in tissue. See text. FRONTAL INTERTEMPORAL PARIETAL SUPRATEMPORAL POSTSPIRACULAR OPERCULAR SQUAMOSAL MAXILLARY MANDIBLE SUBORBITAL 10 MM A SPIRACLE KIG. 2. A. Reconstructed skull of Eusthenopteron, largely after Jarvik. (1954) and Goodrich (1930). The black area behind the spiracle is the eardrum. B. Skull of Osleolepis (after Goodrich, 1930). also showing drum area in black. As noted in the text, the black areas represent the maximal possible extent of the drum; it is probably smaller. the eardrum was already established as a long, narrow, double-layered membrane between the squamosal and parietal shields; this membrane constituted the distal surface of a spiracular diverticulum which was probably filled with air, and thus constituted a middle-ear cavity. The size of this presumed drum membrane, as shown in Figure 2, is what I would consider its maximal possible extent; Thomson (personal communication) thinks it should be much sinaller, perhaps only half as long. As we shall see in a moment, there is a functional advantage to this also. At this point it is well to return briefly to the question of how this drum membrane can offset the disadvantage suffered by a rhipidistian emerging into air where the sound pressure is 36 dB less than it is in water. Consider a spherical air bubble in water with a surface area, A, as shown in Figure 3, upper left. Subjecting this bubble to a pressure change will change its volume by AV. The resulting displacement of the bubble's surface, d, is the quotient of AV volume and area: d = . If we embed A this bubble inside a mantle of non-elastic living tissue, and suspend the entire assembly in water (Fig. ?>, upper right), nothing changes in the behavior of the bubble under 376 WlLLEM A. VAN BERGEIJK pressure, because the density of the tissue is very nearly that of water, and the resulting inertial load is symmetrically disposed. Even when the bubble is moved to an asymmetrical position within the tissue (Fig. 3, lower left), so that a thin membrane with surface area, a, is created, nothing changes in the bubble's behavior, because the inertial load is still symmetrical. But when the asymmetrical assembly is put in air (Fig. 3, lower right), the situation changes drastically. The inertial load on the membrane is now virtually nil, while on the rest of the bubble the load of the tissue is still present. The result is that in dynamic operation essentially the entire change of volume is taken up by the membrane, producing a disAV placement, d' = , where d' is greater a than d by a factor A/a. This amplification of d can be seen to increase as a becomes smaller with respect to A. Of course, this would only be true if the tissue mantle had an effective mass that is infinite compared to the mass of the membrane, or if the mantle were completely rigid. In actuality, those conditions may be met with varying degrees of exactitude. What is important, however, is that some degree of asymmetry allows a fairly small part of the bubble's surface to move with considerably greater ease than the rest. What, then, should the ratio, A/a, be if the drop in pressure encountered in going from water to air is to be compensated? Clearly, if for the same change in pressure (and so, the same AV) the asymmetrical bubble shows a greater displacement of the moveable surface in air, then less pressure is necessary in air to produce the same displacement that would ensue in water for a given pressure. The calculation is simple: a difference in pressure of 36 dB is a factor of about 63, as we have seen already, and thus A/a should be about 63. Let us now return to Eusthetwpleron and see how well-adapted its system is to the transition from water to air. We note immediately that we have indeed an asymmetrical bubble with only a small part of FIG. 4. Same part of skull as in Fig. 1, but with rectangles superimposed to calculate approximate ratio of drum surface, C, to diverticulum surface, 2(A+B). See text. its surface exposed, and the rest enclosed in what clearly is a rather massive and perhaps rigid jacket consisting of the hyomandibular and parotic dental plates, and the palatoquadrate. The surface ratio, A/a, has, of course, never been measured, but a ballpark approximation can be made on the basis of Jarvik's figure. In Figure 4, this drawing is shown again, but now with three rectangles superimposed. Rectangles A and B approximate my best guess about the outline of the diverticulum; small rectangle C represents the drum membrane. The total surface of A -|- B, multiplied by two to take in the other side of the sac, and then divided by the area of C, yields a quotient of about 50. This corresponds to an increase of pressure-sensitivity of 34 dB. This figure is gratifyingly close to the theoretical value, and, in view of the very rough approximations we have had to make (especially the probably exaggerated size of the drum) in the absence of precise measurements, may be taken to indicate that Eusthenopteron could have made the transition from water to air with practically no loss in acuity of hearing. Having demonstrated that Eusthenopteron had an eardrum and was, at least on physical grounds, quite capable of hearing in air, we may next inquire how the drummembrane displacements were communi- EVOLUTION OF HEARING IN VERTEBRATES cated to the inner ear. The most obvious (and, indeed, the only serious) candidate for this function is the hyomandibular. It not only forms the caudo-dorsal margin of the endothelial side of the drum, but it also articulates directly with the otic capsule. The homology of the hyomandibular and the stapes of tetrapods is well-known and needs no belaboring (see Goodrich, 1930, pp. 449485, for an extensive discussion). Until now, however, the eardrum and stapes were assumed to have made their first appearance in the primitive amphibians (Labyrinthodontia), since no recognizable precursor had been demonstrated. The evidence presented here pushes the origin of the eardrum-stapes complex back to the ancestral fish, where it appears to have been entirely functional for hearing in air as well as underwater. Thus, continuity in both structural and functional evolution is demonstrated. ACKNOWLEDGMENT For advice in matters physical I owe a debt to G. G. Harris, whose patience with my ignorance is immense. 377 REFERENCES Dijkgraaf, S. 1963. The functioning and significance of the lateral-line organs. Biol. Rev. 38:51-105. Goodrich, E. S. 1930. Studies on the structure and development of vertebrates. Macmillan, London, lxix + 837 p. Harris, G. G. 1964. Considerations on the physics of sound production by fishes, pp. 233-247. In \V. N. Tavolga, [ed.], Marine-Bio-acoustics. Per gamon Press, New York. Harris, G. G., and W. A. van Bergeijk. 1962. Evidence that the lateral-line organ responds to near-field displacements of sound sources in water. J. Acoust. Soc. Amer. 34:1831-1841. Jarvik, E. 1954. On the visceral skeleton in Eusthenopteron with a discussion of the parasphenoid and palatoquadrate in fishes. Kungl. Svenska Vetenskapsakad. Handl. (4) 5: (1)1-104. Schneider, H. 1941. Die Bedeutung der Atemhohle der Labyrinthfische fur ihr Horvermogen. Vergl. Physiol. 29:172-194. van Bergeijk, W. A. 1964. Directional and nondirectional hearing in fish. pp. 281-299. In W. N. Tavolga, [ed.], Marine-Bio-acoustics. I'ergamon Press, New York. van Bergeijk, W. A. 1966. The evolution of vertebrate hearing. Contrib. Sensory Physiol. 2: (in press).
© Copyright 2026 Paperzz